Copyright by John Chance Rainwater 2008

Copyright by John Chance Rainwater 2008

The Dissertation Committee for John Chance Rainwater Certifies that this is the approved version of the following dissertation:

Dyes and Indicators in Molecular Sensing Ensembles: Progress toward Novel Uses of Dendrimers and Reactands in Optical Sensing Methods

Committee:

Eric V. Anslyn, Supervisor Jason Shear Hung-Wen Liu Christopher Bielawski George Georgiou

Dyes and Indicators in Molecular Sensing Ensembles: Progress toward Novel Uses of Dendrimers and Reactands in Optical Sensing Methods

by John Chance Rainwater, B.S.; M.A.

Dissertation Presented to the Faculty of the Graduate School of The University of Texas at Austin in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

The University of Texas at Austin December 2008

Dedication

To my family for their love and support.

Acknowledgements

Thanks to all the members of the Anslyn group who have served as an endless resource of knowledge, inspiration, and humor. I am especially indebted to Suzanne Tobey for her helpful instruction during the early days of my graduate career and Alona Umali and Shagufta Shabbir for their assistance near the end. Chelsea Martinez is thanked for proofreading several sections of this work. Colin Kubarych also deserves special mention for supplying the ambient works that soundtracked the preparation of this dissertation. I would also like to thank my supervisor, Prof. Eric V. Anslyn, for his patience, encouragement, and enthusiasm during my time here. His insight and support were invaluable in the completion of this work, and his contributions to my scientific education are without equal.

v

Dyes and Indicators in Molecular Sensing Ensembles: Progress toward Novel Uses of Dendrimers and Reactands in Optical Sensing Methods

Publication No._____________

John Chance Rainwater, Ph.D. The University of Texas at Austin, 2008

Supervisor: Eric V. Anslyn

Over the past two decades, the field of molecular sensing has developed into a mature offshoot of molecular recognition, and sensing protocols based on optical signal modulations have enjoyed particularly great success. Such sensing methods are the focus of this dissertation, in which efforts toward the integration of dendrimers and reactands into separate, optically-based sensing platforms are described. To this end, Chapter 1 provides a brief introduction to molecular sensing and its supramolecular underpinnings. The remainder of Chapter 1 is dedicated to dendrimers and their application to molecular recognition and sensing. A discussion of the physicochemical properties of dendrimers is also included to lend perspective on the structure, size, and shape of these macromolecules. The role of dyes and indicators in the elucidation of dendritic structure and function is given special consideration. Finally, selected reports of dendrimers in molecular recognition and optical sensing are summarized. vi

Chapter 2 details original research directed toward the incorporation of dendrimers into molecular sensing ensembles. This use of dendrimers in molecular recognition and sensing is distinguished from those examples described in Chapter 1 by its modular nature. This modularity is achieved through the use of a non-covalent sensing motif based on indicator displacement. The identification and optimization of the appropriate components for use in such dendrimer-based sensing ensembles represents a contribution of the research described herein. An evaluation of indicator dyes for their incorporation into an enantioselective indicator displacement assay (eIDA) for common organic molecules is the subject of the research discussed in Chapter 3. The selected indicator dyes were assessed for use in a novel eIDA that relies on covalent bond formation for the enantioselective signaling of monofunctional organic analytes. A survey of colorimetric methods for the identification and discrimination of amines is included because these compounds served as an initial target in the proposed assay. Optical enantiosensing strategies are also reviewed in light of their relevance to the present work.

vii

Table of Contents List of Figures .......................................................................................................xii
 List of Schemes ..................................................................................................... xv
 Chapter 1: Dendrimers in Molecular Recognition and Sensing............................. 1
 1.1 Introduction and Scope............................................................................. 1
 1.2 Supramolecular Chemistry ....................................................................... 2
 1.2.1 Molecular Recognition ................................................................. 3
 1.2.1.1 Electrostatic Interactions .................................................. 3
 1.2.1.2 Hydrogen Bonding ........................................................... 5
 1.2.1.3 Metal Coordination........................................................... 7
 1.2.1.4 Solvophobic Effects ......................................................... 8 1.2.2. From Molecular Recognition to Molecular Sensing ................... 9
 1.2.2.1 Colorimetric Chemosensors ........................................... 10
 1.2.2.2 Fluorometric Chemosensors........................................... 11
 1.3 Synthetic and Natural Dendrimers ......................................................... 13
 1.3.1 Dendrimer Synthesis .................................................................. 14
 1.4 Physicochemistry of Dendrimers ........................................................... 17
 1.4.1 Conformational Analysis of Dendrimers ................................... 18
 1.4.1.1 The influence of generation............................................ 19
 1.4.1.2 The influence of pH........................................................ 20
 1.4.1.3 The influence of solvent ................................................. 22
 1.5 Dyes and Indicators as Structural Probes of Dendrimers....................... 24
 1.5.1 Probing the Dendritic Interior .................................................... 25
 1.5.2 Probing the Dendritic Surface .................................................... 30
 1.6 Dyes and Indicators in Applications of Dendrimers .............................. 31
 1.6.1 Solubilization of Guest Molecules ............................................. 31
 1.6.2 Extraction of Guest Molecules ................................................... 34
 1.6.3 Controlled Release of Guest Molecules ..................................... 34
 viii

1.7 Dendrimers in Molecular Recognition ................................................... 39
 1.7.1 Recognition at the Core .............................................................. 40
 1.7.2 Recognition within the Branches ............................................... 44
 1.7.3 Recognition at the Surface ......................................................... 44
 1.7.3.1 Binding of Tripeptides.................................................... 46
 1.7.3.2 Metal Coordination......................................................... 48
 1.8 Dendrimers in Molecular Sensing .......................................................... 50
 1.8.1 Signal Amplification .................................................................. 52
 1.9 Conclusion and Outlook ......................................................................... 54
 1.10 References ............................................................................................ 55
 Chapter 2: Progress Toward Templated Assembly of Dendritic Receptors ........ 66
 2.1 Introduction and Scope........................................................................... 66
 2.2 Background and Significance................................................................. 67
 2.3 Research Design ..................................................................................... 75
 2.4 Identification of a Binding Subunit ........................................................ 79
 2.4.1 Screening and Binding of Indicators .......................................... 80
 2.4.2 Evaluation of Tripeptide Binding via Indicator Displacement .. 86
 2.4.3 Development of Binding Subunit............................................... 89
 2.4.3.1 Synthetic Refinement of Binding Subunit...................... 92
 2.4.3.2 Conjugation of Binding Subunit 2.23 to Peptides .......... 93
 2.5 Dendrimers as Scaffolds in Non-Covalent Sensing Ensembles ............. 94
 2.5.1 Preparation of a Boronic Acid-Based Peptide............................ 95
 2.5.2 Dendrimer Binding/Uptake of Tagged Peptide.......................... 97
 2.5.3 Development of Signaling Protocol ......................................... 105
 2.5.3.1 Attachment of Binding Tag to Esculetin Derivative. ... 109
 2.5.3.2 Binding and FRET Signaling Studies Using Compound 2.36 ............................................................ 111
 2.5.4 Synopsis.................................................................................... 119
 2.6 Metal Binding by AT-PAMAM Dendrimers ....................................... 120 2.7 Conclusion and Outlook ....................................................................... 123 ix

2.8 Experimental Details for Chapter 2...................................................... 124
 2.8.1 General Considerations ............................................................ 124
 2.8.2 Cuvette-based Titrations........................................................... 125
 2.8.3 Plate Reader-based Parallel Titrations ..................................... 125
 2.8.4 Solid-Phase Peptide Synthesis.................................................. 126
 2.8.4.1 General Automated Methods........................................ 126
 2.8.4.2 Manual Solid-Phase Synthesis of Boronic Acidbased Peptide 2.33 ........................................................ 127
 



















2.8.5 General Synthesis ..................................................................... 128
 2.9 References ............................................................................................ 132
 Chapter 3: Evaluation of Indicator Dyes for Use in a Novel Enantioselective Indicator Displacement Assay.................................................................... 138
 3.1 Introduction and Scope......................................................................... 138
 3.2 Colorimetric Sensing and Discrimination of Amines .......................... 138
 3.2.1 Classical Approaches ............................................................... 139
 3.2.2. Array-based Approaches ......................................................... 141
 3.2.3 Amine Sensing Based on Covalent Bond Formation ............... 144
 3.2.3.1 Trifluoroacetyl-based Chromoreactands ...................... 146
 3.2.3.2 Tricyanovinyl-based Chromoreactands........................ 147
 3.3 Optical Enantiosensing......................................................................... 149
 3.3.1 Fluorescent Methods ................................................................ 150
 3.3.2 Colorimetric Methods............................................................... 150
 3.3.2.1 Classical Single Molecule Enantiosensors ................... 150
 3.3.2.2 Liquid Crystals as Colorimetric Indicators of Chirality ..................................................... 151 3.3.2.3 Enantioselective Indicator Displacement Assays......... 152
 3.4 The Need for Improved High Throughput Screening Assays .............. 154
 3.5 Research Design ................................................................................... 155
 3.5.1 Tricyanovinyl Azo Dye ............................................................ 159
 3.5.2 Tricyanovinyl Ferrocene .......................................................... 162
 x

3.5.3 Tricyanoethynylethene ............................................................. 164
 3.6 Summary .............................................................................................. 171
 3.7 Experimental Details for Chapter 3...................................................... 172
 3.7.1 General Considerations ............................................................ 172
 3.7.2 UV-Visible Titrations............................................................... 172
 3.7.3 Synthesis................................................................................... 173 3.8 References ............................................................................................. 175
 Works Consulted ................................................................................................. 182 Vita 183


xi

List of Figures Figure 1.1 Crown ether complexes of alkali metal cations. ................................. 5
 Figure 1.2 A simple model of the hydrophobic effect. .......................................... 9
 Figure 1.3 An optical chemosensing strategy. .................................................... 10
 Figure 1.4 Synthesis of dendrimers. .................................................................... 16
 Figure 1.5 Commerically available dendrimers.. ................................................ 17
 Figure 1.6 The dimensions of selected PAMAM dendrimers............................. 19
 Figure 1.7 The influence of pH on PAMAM dendrimer conformation............. 21
 Figure 1.8 The influence of solvent on PAMAM dendrimer conformation. ..... 23
 Figure 1.9 Dyes as solvatochromic probes. ......................................................... 26
 Figure 1.10 Dendrimers as unimolecular micelles. ............................................ 32
 Figure 1.11 Meijer’s dendritic box. ..................................................................... 36
 Figure 1.12 Topological regions for dendrimer-based molecular recognition ....................................................................................... 40 Figure 1.13 Coordination modes of Cu(II) binding to PAMAM dendrimers.... 50 Figure 2.1 Colorimetric indicators screened for binding to PAMAM dendrimers ....................................................................................... 81 Figure 2.2 Absorbance modulations of selected indicators upon addition of AT-PAMAM dendrimers............................................................ 82
 Figure 2.3 Binding of 2.4 to various generations of AT-PAMAM dendrimers ....................................................................................... 84 Figure 2.4 Binding of 2.7 to various generations of AT-PAMAM dendrimers. ...................................................................................... 85


xii

Figure 2.5 Binding of 2.6 to various generations of AT-PAMAM dendrimers ...................................................................................... 86
 Figure 2.6 The importance of anionic groups in binding AT-PAMAM dendrimers ....................................................................................... 87
 Figure 2.7 Tripeptide displacement of 2.4 from G5 AT-PAMAM dendrimer.... 89
 Figure 2.8 “Simulated” fluorescence titration of G6 AT-PAMAM dendrimer into 5-carboxyfluorescein............................................ 100 Figure 2.9 Addition of tagged peptide 2.26 to 5-carboxyfluorescein/ dendrimer ensembles at various “loading ratios.” ....................... 102 Figure 2.10 Overlay of absorption and emission bands of fluorescent indicators 2.34 and 2.5. ................................................................ 107
 Figure 2.11 Probing FRET in the presence of dendrimer................................ 108
 Figure 2.12 Binding of 2.36 to a G6 AT-PAMAM dendrimer.......................... 112
 Figure 2.13 Spectral evidence for FRET between dendrimer-bound fluorophores.................................................................................. 113
 Figure 2.14 Control experiments to validate FRET between dendrimerbound fluorophores. ...................................................................... 115
 Figure 2.15 Excitation scan of the G6/CF/MEL-COOH ensemble. ................ 116
 Figure 2.16 Spectral modulations of G6/CF/peptide ensemble upon addition of MEL-COOH................................................................ 118
 Figure 2.17 Binding of Cu+2 to G6 AT-PAMAM dendrimer. ........................... 121
 Figure 2.18 Verification of the spectral response associated with copper binding. .............................................................................. 121
 Figure 2.19 Spectral evidence for the uptake of copper ions by the CF/G6 ensemble. ....................................................................................... 122
 xiii

Figure 3.1 Colorimetric molecular sensing using covalent bond formation. .. 145
 Figure 3.2 Absorption spectra of tricyanovinyl ferrocene 3.15. ....................... 163
 Figure 3.3 UV-vis absorbance modulations and associated isotherm for addition of benzylamine to 3.16. .................................................. 165
 Figure 3.4 UV-vis absorbance modulations and associated isotherm for addition of methanol to 3.16. ....................................................... 167
 Figure 3.5 UV-vis absorbance modulations upon addition of NaCN to a solution containing MeOH and 3.16.. ................................... 169
 Figure 3.6 UV-vis absorbance modulations upon addition of NaCN to 3.16. ........................................................................................... 170


xiv

List of Schemes Scheme 1.1 Hydrogen bonding in molecular recognition. ................................... 6
 Scheme 1.2 The dual utility of metal coordination in molecular recognition. .... 8
 Scheme 1.3 A non-covalent “twist” on the dendritic box. .................................. 38
 Scheme 1.4 Equilibration of the two forms of dianionic fluorescein................. 39
 Scheme 1.5 Molecular recognition at the dendritic surface............................... 47
 Scheme 2.1 Molecular imprinting with polymers. .............................................. 68
 Scheme 2.2 Selection of a molecular host using a dynamic combinatorial library. ....................................................................................... 70 Scheme 2.3 A homogeneous nanoscale imprinting system. ............................... 72
 Scheme 2.4 Molecularly imprinted dendrimers. ................................................. 73
 Scheme 2.5 A strategy for templated assembly of dendritic receptors................ 77
 Scheme 2.6. Optical indicator displacement assay............................................. 79
 Scheme 2.7 Ionization Equilibria Responsible for Observed Spectral Changes of 2.4 and 2.5 ............................................................. 83
 Scheme 2.8 Attempted synthesis of a gallic acid-derived binding subunit......... 90
 Scheme 2.9 Attempted selective saponification routes to 2.15............................ 92
 Scheme 2.10 Oxidative approach to the synthesis of the binding subunit. ........ 93
 Scheme 2.11 Dendrimers as scaffolds in non-covalent sensing ensembles. ...... 95
 Scheme 2.12. Incorporation of a boronic acid into tagged peptide 2.25 via solid-phase reductive amination.. ............................................. 97 Scheme 2.13 Indicator displacement and peptide uptake regimes observed upon addition of a competitive binder to an indicator/ dendrimer ensemble................................................................. 105
 xv

Scheme 2.14 Proposed synthesis of a tagged MEL derivative.......................... 110
 Scheme 3.1 Colorimetric discrimination of molecular length.......................... 140
 Scheme 3.2 Covalent bond formation between tricyanovinyl-based reactand 3.5 and a primary amine. ....................................................... 149
 Scheme 3.3 Equilibria processes in an enantioselective indicator displacement assay (eIDA). ..................................................... 153
 Scheme 3.4 Equilibria processes in the proposed ee assay............................... 157
 Scheme 3.5 Schematic representation of the proposed equilbria processes. ... 158
 Scheme 3.6 Synthesis of tricyanovinyl azo dye 3.11. ........................................ 160
 Scheme 3.7 Alternative routes to the installation of the tricyanovinyl group.. 161
 Scheme 3.8 Contributing resonance structures of tricyanovinyl ferrocene 3.15. ........................................................................ 164
 Scheme 3.9 Possible reactions of reactand 3.16 with cyanide.......................... 169


xvi

Chapter 1: Dendrimers in Molecular Recognition and Sensing 1.1 INTRODUCTION AND SCOPE Vogtle’s initial report of a branched “cascade” molecule1 prompted subsequent elaborations of dendritic architecture on a molecular scale. Early synthetic contributions by Newkome2 and Tomalia3 expanded Vogtle’s molecular motif into finite, macromolecular dendritic structures.

These seminal reports on the controlled

construction of “starburst polymers” also posited applications for this new molecular topology. Vogtle framed his initial success in terms of molecular recognition, arguing that “for the construction of large molecular cavities and pseudocavities that are capable of binding ionic guests … synthetic pathways allowing a frequent repetition of similar steps would be advantageous” and that Lehn’s neutral ligand syntheses4 did not adequately exploit this strategy. Newkome cast his synthetic contributions to the field as efforts toward “unimolecular micelles” capable of surface inclusion, and Tomalia suggested that the controlled molecular arrays of dendrimers might offer the synergistic effects of potential enzyme mimics. Despite their remarkable prescience, the pioneers of dendrimer chemistry are likely surprised at the innovation their work has inspired. Over the first five years of the 1990’s the number of publications on dendrimers increased from the dozens to the thousands.5 As predicted in Vogtle, Newkome, and Tomalia’s seminal publications, dendrimers have found vast application in supramolecular chemistry.

These uses,

however, are not limited to their predicted utility in molecular recognition6,7,8 and catalysis9,10,11 but rather span applications ranging from molecular sensing12,13 to energy harvesting14,15 to drug delivery.16,17,18,19 1

Because of this surfeit of literature concerning the preparation and application of dendrimers, the work presented in this chapter is ultimately directed to the use of dendrimers in molecular sensing applications. A thorough review of these applications requires a brief discussion of the governing molecular recognition phenomena and a survey of the physicochemical nature of dendrimers. A significant portion is dedicated to describing the role of dyes and indicators in the elucidation of dendritic structure and function.

A brief summary of the synthesis of dendrimers is also included to define

nomenclature, to provide insight into the nature of these structures, and to stress their unique applicability to supramolecular chemistry. 1.2 SUPRAMOLECULAR CHEMISTRY Supramolecular chemistry is a discipline concerned with the “chemistry of molecular assemblies and of the intermolecular bond.”20 The key distinction between molecular chemistry and supramolecular chemistry is the former’s focus on the covalent bond and the latter’s reliance on reversible intermolecular forces to generate new molecular systems.

These intermolecular forces include electrostatic interactions,

hydrogen bonding, metal coordination, and solvophobic effects, and the realization of supramolecular chemistry depends on the effective summation of these various forces to achieve robust intermolecular assemblies. Biology provides the most elegant examples of supramolecular systems as noncovalent interactions mediate many of life’s essential processes. Non-covalent forces drive the formation of DNA’s double helix, regulate enzyme-substrate complexes, and direct antigen-antibody association. The remarkable efficiency and selectivity associated with these intermolecular binding events impart unique functionality in applications such as catalysis and information storage.

A goal of the supramolecular chemist is the

replication of such functionality in abiotic systems through an efficient manipulation of 2

the various non-covalent forces that govern the binding event.21 The engineering of a selective complexation predicated on the concerted action of relatively weak binding forces is within the purview of molecular recognition. 1.2.1 Molecular Recognition Molecular recognition lies at the heart of supramolecular chemistry and describes the selective association between two chemical species based on their electronic and geometric complementarity.22

Emil Fischer first articulated this notion of

complementarity between interacting partners in the description of his lock and key principle.23 Fischer formulated this principle based on his studies of the interactions between enzymes and their substrates. From these studies, he surmised that the size, shape, and position of the enzyme’s binding sites determine the specificity of substrate recognition. While this is now known to be an overly simplistic view of biological receptors, the lock and key analogy is a landmark concept in supramolecular chemistry that laid the foundation for abiotic molecular recognition. This section provides a brief, non-exhaustive survey of non-covalent forces that are particularly relevant to the work described in this dissertation.

The described

interactions represent forces that supramolecular chemists exploit in the generation of supramolecular assemblies, and these interactions can be differentiated by factors such as strength and directionality.

While these forces can be employed in isolation, their

concerted power is usually harnessed to provide maximal affinity and selectivity and can potentially impart tunability to the host-guest interaction. 1.2.1.1 Electrostatic Interactions Electrostatic interactions are based on the Coulombic attraction between opposite charges and include ion-ion, ion-dipole, and dipole-dipole interactions. 3

These

interactions range in strength from 5 kJmol-1 (for a weak dipole-dipole interaction) to 250 kJmol-1 (for ion-ion interaction). While ion-ion interactions are the strongest of these forces, they are non-directional in nature. The weaker ion-dipole and dipole-dipole forces, however, require a specific orientation of both partners to be operational and are therefore useful in achieving selectivity. Seminal examples of molecular recognition are based on directional ion-dipole interactions. Crown ethers are macrocyclic polyether-based receptors that rely on these interactions in their selective complexation of metal ions; the negative portion of the dipole associated with the ether group interacts with a cation via electrostatic attraction. When these dipoles are arrayed in a highly organized fashion, selectivity is imparted to the recognition event based on the relative size of the macrocyclic host and the atomic radius of the cationic guest. This selectivity is evidenced by the eventual development of a series of crown ethers with varying affinity for a variety of cationic guests (Figure 1.1).24

4

18-Crown-6 1.1

15-Crown-5 1.2

12-Crown-4 1.3

Figure 1.1 Crown ether complexes of alkali metal cations. Crown ethers 1.1-1.3 demonstrate selective binding of cations based on the diameter of the ring. The smaller crowns preferentially bind smaller cations, while the larger crowns demonstrate selectivity for larger cations. This observation illustrates the importance of geometric complementarity in molecular recognition. 1.2.1.2 Hydrogen Bonding A hydrogen bond is the result of a dipole-dipole interaction that occurs when a molecule bearing an acidic hydrogen atom (the donor) is brought within the van der Waals radius of an atom with available lone pair electrons (the acceptor). The strength of a hydrogen bond is highly contextual and depends both on the nature of the donor and acceptor and the local environment, as solvation effects and secondary interactions are known to modulate the strength of these interactions. Most hydrogen bonds are relatively weak interactions, so the success of their use in supramolecular chemistry often depends on the judicious arrangement of multiple donor and acceptor sites to afford a cumulative effect that powers the molecular recognition event.

5

O

O N

H

H

N

O

N

N O

N H

O

N H

H

N

O

+

N

N

O

1.4

N

H O

H

N H O

O

O

O

O

O

H O H N

N

N H O

N H

N

O

O

1.5

Scheme 1.1 Hydrogen bonding in molecular recognition. Receptor 1.4 binds a complementary barbituric acid core (1.5) in chloroform through the action of six hydrogen bonds. A report from Hamilton illustrates the utility of exploiting multiple hydrogen bonding interactions in molecular recognition.25

The synthesis of the macrocyclic

receptor enforces a presentation of six hydrogen-bonding groups in which the combination of donor/acceptor sites is directed towards the center of the cavity. As a result, receptor 1.4 is expected to bind substrates with a complementary arrangement of donor/acceptor sites. Substrates containing a barbituric acid core were found to bind to the receptor through the action of multiple hydrogen bonds (Scheme 1.1.) Experiments using “impaired” guests suggest a strong selectivity for barbituric acid-based guests that can fully engage all six of their hydrogen bonding sites. Affinities as high as 6 x 105 M-1 were observed in CDCl3 for guests containing the complementary barbituric acid core. This high association constant confirms the power of the concerted action of hydrogen bonds in host-guest chemistry, and Hamilton’s work remains a hallmark of hydrogenbond mediated molecular recognition. 6

1.2.1.3 Metal Coordination While not strictly a non-covalent interaction, classical coordination chemistry has enjoyed extensive use in supramolecular chemistry.26 Chelate effects can serve to bring multiple binding sites together, thereby facilitating and organizing receptor formation. Work from the Anslyn group provides an example of this strategy.27 A receptor for phosphate anions is formed with the aid of a stoichiometric amount of Cu(II). The metal ion preorganizes the pyridine ligands to provide a C3ν symmetric cavity that is decorated with three positively charged groups that accentuate the binding of anions in highly competitive media (Scheme 1.2).

In addition to templating the formation of the

receptor’s binding cavity, control experiments reveal that the metal center also contributes to the binding of the anionic guests. Cooperative effects between the host’s ion-pairing recognition sites and metal center are noted. While the ion-pairing sites are necessary to produce high affinity anion receptors, inclusion of Cu(II) provides increased affinity and selectivity for small tetrahedral anions.

7

HN NH N H

N

NH N

N H

N H HN

N Cu+2

HN

NH

HN

N

HN

HN CuII

N

HN

NH

N

NH NH

N

N HN

NH

1.6:Cu+2

1.6

Scheme 1.2 The dual utility of metal coordination in molecular recognition. The addition of Cu(II) to receptor 1.6 induces the formation of a binding cavity that is complementary to small tetrahedral oxyanions such as phosphate. The metal center also enhances the receptor’s affinity for these anionic targets through charge-pairing interactions to the oxygen atoms of the guest. 1.2.1.4 Solvophobic Effects The most widely investigated solvophobic effect involves the aggregation of apolar organic molecules in aqueous solutions. This phenomenon is termed known as the hydrophobic effect, and it is a vital contributor to molecular recognition in biological systems.28 As a result of its prominence in nature, the hydrophobic effect has been the subject of intense scrutiny by chemists, and abiotic analogues have been reported.29 Unlike other non-covalent forces, aggregation governed by the hydrophobic effect is often entropically driven. Interaction between two molecules of an organic solute in aqueous solvent leads to an overall decrease in the exposed hydrocarbon surface. The concomitant release of water molecules from the hydrophobic faces of the organic solutes increases the disorder of the system, leading to a favorable change in entropy that drives the association (Figure 1.2). Chemists have exploited this effect in abiotic molecular 8

recognition using molecular hosts that feature a hydrophobic core. Cyclodextrins30 and cyclophanes31 are the two molecular constructs that have been most widely investigated and applied in molecular recognition involving a large hydrophobic component.

Figure 1.2 A simple model of the hydrophobic effect. The aggregation of two organic solutes ( ) in an aqueous solution decreases the hydrocarbon surface area exposed to solvent molecules ( ). This association is entropically favorable because the accompanying release of highly structured solvent increases the disorder of the system. 1.2.2. From Molecular Recognition to Molecular Sensing Lehn argues that “mere binding is not recognition” and that true recognition often implies the emergence of a specific function due to a well-defined pattern of intermolecular interactions.32 Indeed, molecular recognition is the basis of technologies such as selective catalysis,33 molecular transport,34 and molecular logic gates.35

In

addition to these applications, molecular sensing has become a particularly successful descendant of instructed molecular recognition within the past two decades. Czarnik coined the term chemosensor to describe an abiotic receptor whose reversible binding to a targeted analyte results in an appreciable change of the system’s physical properties.36 The challenge of designing a successful chemosensor, therefore, is two-fold:

the

underlying molecular recognition phenomena must provide sufficient discrimination between structurally similar analytes (an issue of selectivity), while the requisite change in its physical property must be perceptible at the desired concentration ranges (an issue 9

of sensitivity). Some factors governing selectivity were highlighted in the previous section. The present section will therefore focus on the modulation of physical properties as it relates to chemosensing.

This modulation, which can be optical37 or

electrochemical38 in nature, relies on signal transduction mechanisms to relay the presence of an analyte with an easily observable signal. This section will focus on the various modes of signal transduction exploited in optical chemosensors. A.

Signaling unit

B.

Recognition unit

Figure 1.3 An optical chemosensing strategy. The signaling unit and recognition site are covalently attached by an appropriate linker. (A.) The signaling unit is unperturbed in the absence of analyte ( ). (B.) The recognition unit selectively binds the analyte, resulting in a modulation of the optical properties of the attached chromophore via signal transduction mechanisms. Optical chemosensors require a recognition site and a signaling moiety and as well as a mechanism for communication between these two units (Figure 1.3). For optical chemosensors, the signaling moiety is a chromophore or fluorophore whose optical properties are altered upon binding an analyte. Depending on the nature of the signaling group, optical chemosensors are considered colorimetric (when a change in absorbance results) or fluorometric (when a change in emission results). 1.2.2.1 Colorimetric Chemosensors Colorimetric chemosensors have found widespread application in the detection of myriad analytes including transition metal cations39 and inorganic anions.40 The reported functionality of some these chemosensors in highly competitive media is a testament to 10

the field’s increasing applicability to real-world biological samples.41

Examples of

classical chemosensors that target organic analytes are more limited, but colorimetric detection of monosaccharides,42 amino alcohols,43 and dicarboxylic acids44 has been described. Guest-induced absorbance modulations typically result from a change in the molecular structure of the chromophore. These changes include proton transfer,45 charge transfer,39 and isomerization46 events.

If such phenomena effect a change in the

chemosensor’s UV-vis absorption spectrum, then they can serve as signaling mechanisms in absorbance-based optical chemosensors. Signaling mechanisms that modulate optical signals within the visible spectrum lead to a particularly attractive feature of colorimetric chemosensors:

the potential for “naked-eye” detection of the presence of analytes.

Naked-eye detection is desirable in many proposed applications of optical chemosensing because it lessens the need for expensive instrumentation and provides an immediate qualitative read-out of analyte presence. 1.2.2.2 Fluorometric Chemosensors47 The signaling mechanisms incorporated into fluorometric chemosensors encompass the structural changes exploited in colorimetric chemosensors but also include numerous photophysical processes such as energy48 and electron49 transfer. In contrast to absorbance, fluorescence is very sensitive to subtle changes in the geometry and electronic structure of the ground state. This sensitivity permits the use of additional optical response mechanisms such as those based on conformational restriction50 and allows detection at concentrations up to one million times smaller than those used in absorbance-based colorimetric sensing. The complementary nature of colorimetric and fluorometric sensing is evidenced by work from Gunnlaugsson et al in which they report the dual action of a single 11

molecule that serves as both a fluorometric and colorimetric sensor over two different concentration ranges.51 Sensor 1.7 incorporates a thiourea-based recognition unit for anions and a naphthalimide signaling moiety. Binding of anions such as fluoride to the receptor quenched the fluorescence emission of the naphthalimide group via a photoinduced electron transfer mechanism. The relatively low concentrations of fluoride that resulted in fluorescence quenching only caused minor changes in the chemosensor’s absorption spectra. However, upon increasing the fluoride concentration above 30 mM, dramatic changes in the absorption spectrum were observed due to deprotonation of the naphthalimide unit.

This dually responsive system represents a novel approach to

chemosensing and harnesses the advantages of both colorimetric and fluorimetric sensing: the fluorimetric response is characterized by its high sensitivity, while the colorimetric response allows for naked eye detection of analytes.

O

N

O S

N H N

H

HN

1.7 The intent of the previous sections was to provide a brief account of molecular recognition and its role in the evolution of molecular sensing. The remainder of this chapter will shift its focus to dendrimers and will be directed toward their use in molecular recognition and application in molecular sensing. Many of the aforementioned binding forces remain at play in dendrimer-based molecular recognition, and the selected dendritic sensors rely on the modulation of an optical signal that is often mediated by the same processes described in the context of small molecule chemosensors. 12

1.3 SYNTHETIC AND NATURAL DENDRIMERS Dendrimers are highly branched, three-dimensional macromolecules whose bonds emanate from a central core. This class of macromolecules is named for their arboreal appearance. The word “dendrimer” comes from the Greek words dendron (tree) and meros (part), and Tomalia cites his interest in horticulture as inspiration for the synthesis of these tree-like molecules.5 Dendrimers are prepared in a controlled, stepwise fashion by an iterative sequence of reactions; the result is macromolecules with defined shape and size. Unlike traditional polymers whose synthesis results in a Gaussian distribution of molecular weights, dendrimers are considered monodisperse. This ability to create discrete, homogenous high molecular weight polymers is a hallmark of dendrimer synthesis. While traditional organic chemistry offers a mastery of the covalent bond, it is not practical for molecular construction of nanoscale dimensions. Polymer science offers facile entry to nanoscale structures, but suffers from imperfect control over product size and mass. Dendrimer synthesis represents a hybrid of these approaches and offers intellectual control of the size, mass, and shape of nanoscale architecture. Examples from nature suggest such a precise control of structure confers functionality. Nature is replete with dendritic structures at every dimension. Trees, neurons, blood vessels, and polysaccharides all exhibit pervasive branching. The ubiquity of dendritic networks in nature suggests that this topology offers some unique advantage over unbranched analogues and may be the result of evolution. These architectures are often found in natural systems whose function requires amplification or enhancement. For example, the branching of trees ensures maximum interaction with the environment for efficient energy harvesting, and the dendritic nature of neurons maximizes connections in the brain for optimum information storage. A recent study on gecko’s 13

feet52 demonstrates the power of a dendritic array on functional performance. The hairs of the gecko’s foot are divided into a branched network that allows the gecko to adhere to dry surfaces. This adhesion is the result of the summation of weak intermolecular forces between the foot hair and the surface. The dendritic presentation of foot hairs amplifies each of these discrete interactions and parlays this enhancement into remarkable functionality. On the molecular level, nature employs branched polysaccharides such as glycogen for energy storage as their structure provides enzymatic access for the rapid release of large amounts of sugar.53 The replication of such functionality on a molecular scale is the impetus for much dendrimer research. As examples from nature illustrate, an exact control over structure offers the potential to generate functionality. The precise placement of chemical functionality and the control of size and inner volume offered by their synthesis make dendrimers attractive candidates to rival the functionality of their natural counterparts. 1.3.1 Dendrimer Synthesis Two distinct strategies for the synthesis of dendrimers have emerged (Figure 1.4). In the divergent approach, growth of the dendrimer originates from the core.

The

alternative convergent method starts construction at the surface and ends at the core. Early efforts in dendrimer synthesis relied upon a divergent strategy. Frechet described a convergent approach to dendrimer synthesis in 1990 to address problems associated with the divergent methodology.54 In a divergent approach, synthesis of the dendrimer occurs in a stepwise fashion through an iterative coupling/deprotection of monomeric building blocks. Each iteration results in a deprotected shell of monomers termed generations, and each generation offers a new reactive surface that can be further elaborated. The coupling/deprotection 14

protocol is reminiscent of the solid-phase synthesis of peptides.55 In contrast to the synthesis of linear peptides, however, the iterations of divergent dendrimer synthesis result in an exponential increase of terminal surface groups as a function of generation. Synthesis of subsequent generations requires an exponentially increasing number of reactions to be performed on one molecule. As a result, divergent syntheses require efficient, selective transformations to avoid “deletions” that lead to structural defects at higher generations. Regardless of reaction efficiency, however, it is virtually impossible to divergently synthesize structurally homogenous dendrimers beyond six generations. For example, even an average coupling selectivity of 99.5% provides purities of only 8% for sixth generation poly(propyleneimine) dendrimers.56 Given that chromatographic purification of intermediate generations is obviated by the similarity of these byproducts to the desired macromolecule, higher generation dendrimers prepared by divergent methods should be considered structurally imperfect molecules whose purity is governed by statistics. Frechet described a convergent approach to dendrimer synthesis to complement divergent methods. The convergent method builds the dendrimer from the periphery and proceeds inward with a final coupling of the dendritic segments (dendrons).

This

approach offers enhanced structural control because the low, constant number of reaction sites at each step limits the formation of side products and provides the possibility of purification of intermediate generations. Unlike divergently prepared dendrimers, those constructed via convergent syntheses are considered defect free.

Application of a

convergent route, however, is limited to dendrimers of lower generations because increased steric crowding at the reactive focal point of the dendrons leads to decreased reaction rates and yields.

15

A.

B.

Figure 1.4. Synthesis of dendrimers. Two general synthetic strategies are employed in dendrimer synthesis. (A.) A divergent synthesis is initiated from the core and terminates at the dendrimer surface through a series of iterative reaction steps. (B.) Convergent methods proceed from the surface and rely on a final coupling of dendritic wedges to generate the macromolecule. These two seemingly disparate synthetic paradigms are complementary in practice. The nature of the monomeric building blocks and the features of the desired macromolecule often dictate the choice of methodology in dendrimer synthesis.

In

general, the convergent method provides better structural control, but the divergent method is more amenable to large scale synthesis and preparation of higher generation dendrimers.

As a result, the commercially available polyamidoamine (PAMAM)

dendrimers that are the subject of the work described in Chapter 2 are prepared through divergent protocols. A final analysis of dendrimer synthesis, however, should note that the minute structural differences between divergently and convergently prepared dendrimers are not expressed in their macroscopic behavior.23 Dendrimers, therefore, represent the most defined macromolecular structures currently known regardless of their method of 16

preparation. This level of intellectual control in their preparation makes dendrimers attractive subjects in molecular recognition studies and molecular sensing applications.

1.8

1.9

Figure 1.5 Commerically available dendrimers. Polypropylene imine (PPI) and polyamido amine (PAMAM) dendrimers are two commercially available families synthesized by divergent methodologies. Structures 1.8 and 1.9 represent a fifth generation PPI dendrimer and a third generation PAMAM dendrimer, respectively. Dendrimer generation is defined by the number of cascade points between the core and the surface. 1.4 PHYSICOCHEMISTRY OF DENDRIMERS A comparision of dendrimers to proteins is an instructive entry into a discussion of their physicochemical properties.

Many dendrimers are considered “artificial

proteins” based on their electrophoretic57 and biomimetic58 properties and their similarity on dimensional length scales (Figure 1.6). PAMAM dendrimers of generations 3, 4, and 5 are approximately the same size of insulin, cytochrome C, and hemoglobin, respectively; but unlike their natural counterparts whose tertiary structure results from 17

noncovalent interactions of linear peptide chains, dendrimers are covalently fixed structures. The result is that dendrimers are somewhat more rigid than proteins and less susceptible to environmental changes. Dendrimers are also less compact than proteins of comparable molecular weight. This observation suggests that dendrimers have a less congested interior than proteins. The possibility of channels and cavities within the dendrimer’s interior has prompted intense computational and experimental scrutiny of the conformations of dendrimers and aroused the interest of supramolecular chemists. 1.4.1 Conformational Analysis of Dendrimers The existence of internal voids within dendrimers is a topic of debate in the literature. The presence of a permanent, rigid cavity is not necessarily a prerequisite for the known encapsulation of guest molecules by dendrimers. As a result, researchers have performed extensive experimental and theoretical studies to probe the topology of the dendritic state.

Elucidating the presence of internal voids or cavities within the

dendrimer’s interior has been a major driving force for these studies, and the conformational behavior of these molecules has been studied as a function of generation, pH, and solvent.

18

Figure 1.6 The dimensions of selected PAMAM dendrimers. Comparison of a series of PAMAM dendrimers to proteins reveals that various generations of these synthetic structures are comparable in size and shape to many biological macromolecules (image adapted from reference 16). 1.4.1.1 The influence of generation A pioneering theoretical study by De Gennes and Hervet59 described the intramolecular behavior of PAMAM dendrimers. Using a self-consistent field model, De Gennes demonstrated that the molecular density of the interior remains low during dendrimer growth while the periphery becomes increasingly crowded. This study also posited that lower generation dendrimers exist in a relatively open form, while dendrimers of higher generation adopt a three-dimensional globular structure. This more ordered, spherical structure is often observed beginning at the fifth generation.7 The De Gennes study, however, neglects a fundamental aspect of dendrimer conformation – backfolding. Backfolding refers to the entropy driven folding of the dendrimer segments (dendrons) into the interior of the dendrimer. This phenomenon is 19

the result of two factors: 1) the lack of favorable interactions between the surface groups and 2) the presence of favorable interactions between the surface groups and branches. In these situations, the resulting mobility of the dendrons disfavors the idealized De Gennes model. Subsequent theoretical studies60,61,62 demonstrate that the surface groups are distributed throughout the volume of the dendrimer, predict an increase in backfolding with generation, and suggest a monotonic decrease in molecular density progressing from the dendrimer core.

These theoretical assertions are borne out in experimental

investigations using size exclusion measurements and NMR techniques.63,64,65 The existence of the molecularly dense cores proposed by these studies might render dendrimers unsuitable for guest encapsulation and applications predicated on molecular recognition. The extent of this undesired backfolding, however, is dependent on dendrimer structure and can be precluded completely by the presence of secondary interactions between the surface groups. An example from the Percec lab shows that the fluorophobic effect promotes segregation between the dendritic branches and surface groups in the solid state.66 Experimental consideration of PAMAM dendrimers conclude that, in contrast to other unmodified dendrimers, these molecules have a hollow interior and densely packed periphery.67,68 A small-angle neutron-scattering (SANS) study confirms that the surface groups of a PAMAM dendrimer are located at the periphery even at the seventh generation.69 These experimental findings are consistent with the De Gennes model which predicts a low molecular density at the core. 1.4.1.2 The influence of pH The PAMAM dendrimers used in the studies described in Chapter 2 are functionalized at the surface with amino groups. The primary amines at the surface and 20

the tertiary amines located in the branches give these PAMAM dendrimers a basic periphery and a basic interior.

As a result of these arrayed, densely packed basic

functionalities, conformational changes are observed upon variation of pH. Molecular dynamics calculations show that PAMAM dendrimers assume extended conformations at low pH but adopt a more spherical structure at high pH.70 At pH values less than 4, electrostatic repulsions between the protonated surface primary amines and interior tertiary amines force the dendrimer branches apart, ensuring a highly ordered extended structure with an interior that becomes increasingly hollow with generation. A recent report also documents fluorescence emission from G4 amino-terminated PAMAM dendrimers at low pH; the authors’ suggest that strong electrostatic repulsions present at pH 2.5 rigidify the structure thereby promoting luminescence.71

Figure 1.7 The influence of pH on PAMAM dendrimer conformation. Aminoterminated PAMAM dendrimers become increasingly compact as the pH is raised because the electrostatic repulsion between the protonated surface and interior amines is mitigated (figure adapted from reference 17).

21

At higher pH values (pH > 10), the dendrimer contracts because the repulsive interactions are minimized as the molecule becomes increasingly neutral (Figure 1.7). The resulting globular conformation features a higher degree of backfolding due to the lack of repulsive forces between the interior and surface amines. Backfolding is also present to some extent at the neutral pH values at which the studies described in this dissertation were conducted. This backfolding is a consequence of hydrogen bonding between uncharged tertiary amines in the branches and positively charged surface amines, an interaction that is possible at neutral pH based on the relative pKa’s of the amino groups. Crooks et al determined the pKa’s of G4 PAMAM dendrimers via potentiometric pH titrations.72 These experiments determined the pKa of the internal tertiary amines to be 6.30 and that of the peripheral primary amines to be 9.23. The value determined for the surface primary amines does not differ significantly from that of a non-dendritic structural analog. The pKa of the internal tertiary amines, however, is nearly two pKa units lower than that found for a tertiary amine in a structurally similar compound. Crooks ascribes this discrepancy to the hydrophobic microenvironment created within the interior of the dendrimer. Such a shift in pKa would be analogous to that observed in the hydrophobic pocket of some proteins.73 A final contribution of this work is the assessment of the extent of protonation as a function of pH. A proton binding curve for the G4 PAMAM dendrimer revealed that at pH 7 most primary amines are protonated but most tertiary amines remain neutral. 1.4.1.3 The influence of solvent Solvent also influences dendrimer conformation.

Multidimensional NMR

experiments74 suggest that poorly solvated (poly-(propylene imine)) (PPI) dendrimers adopt backfolded conformations. Apolar solvents such as benzene promote secondary, 22

noncovalent interactions between dendrimer segments. increased molecular densities at the core.

These interactions lead to

In contrast, polar solvents disrupt these

intramolecular interactions, and an extended conformation is preferred when the dendrimer is effectively solvated (Figure 1.8). A.

B.

Figure 1.8 The influence of solvent on PAMAM dendrimer conformation. (A.) PAMAM dendrimers prefer extended conformations in polar solvents that disrupt the intramolecular forces responsible for backfolding. (B.) In apolar solvents, PAMAM dendrimers are poorly solvated and adopt a more compact conformation. The increased molecular density results from significant backfolding of the surface groups via noncovalent interactions. De Backer et al used fluorescence depolarization techniques to determine the hydrodynamic volumes of dendrimers based on Frechet-type dendrons.75 These polybenzyl ether-based dendrimers have a less polar interior than PAMAM or PPI dendrimers; thus, the solvent dependent behavior of these macromolecules is expected to differ from that observed with PPI dendrimers. Four generations of a Frechet-type dendrimer with a fluorescent rubicene core was prepared to investigate the relationship between dendrimer volume and solvent quality. Anisotropy measurements revealed that toluene effected complete expansion of all generations at room temperature as a consequence of solvation via π-interactions.

In contrast, acetonitrile, a more polar

solvent, is unable to effectively solvate the benzyl ether-based dendrons and promotes a 23

decrease in hydrodynamic volume due to an increase in intramolecular π-π interactions. The authors note that the measured volume of the fourth generation dendrimer indicates a complete collapse of the structure. In conclusion, dendrimer conformation is governed by free energy with thermodynamic parameters such as entropy and enthalpy determining the position of the surface groups. Environmental factors such as pH and solvent polarity modulate entropic and enthalpic contributions to give appreciable changes in conformation. Conformation is also influenced by dendrimer generation, and the following generalization can be made about PAMAM dendrimers from the reported theoretical and experimental evidence: as a dendrimer generation increases, the structure becomes increasingly globular, and the interior becomes increasingly isolated from the surroundings by the proliferating surface groups.

The result is a unique microenvironment at the dendrimer core that is the

product of the limited diffusion of solvent molecules.

This microenvironment is

analogous to the hydrophobic pockets of proteins and can modulate the optical and redox properties76 of encapsulated guests. As a result, dendrimers with a defined interior and congested periphery are attractive candidates in molecular recognition and chemosensing applications. 1.5 DYES AND INDICATORS AS STRUCTURAL PROBES OF DENDRIMERS Solvatochromic probes offer insight into the structure of dendrimers by relaying information from the dendritic microenvironment.

Solvatochromic effects refer to

changes in the absorbance or emission properties of a compound due to changes in solvent polarity.77,78 Because higher generation dendrimers have interiors that are shielded from bulk solvent, the polarity of the medium is modulated at the dendrimer core. The result is differential solvation of a chromophore upon its encapsulation in a dendritic microenvironment. This differential solvation leads to stabilization or destabilization of 24

the chromophore’s electronic states, and this phenomenon is manifested by changes in absorbance and fluorescence properties.

Researchers have exploited these solvent

dependent optical changes to probe the polarity, accessibility, and size of the dendritic milieu. 1.5.1 Probing the Dendritic Interior Frechet first demonstrated the modulation of physical properties within the dendritic microenvironment.79 Using a convergent methodology, the researchers attached a solvatochromic probe to the focal point of a poly-benzyl ether-based dendrimer. Six generations were produced to assess the nature of the microenvironment as a function of dendrimer size.

UV-vis absorption analysis of the dendrimers in CCl4 showed a

bathochromic shift of λmax with increasing generation. The authors attribute this shift to the increased shielding of the solvatochromic probe from the apolar bulk solvent (Figure 1.9). In effect, the shielding provided by the dendritic wedges increases the polarity of the local medium and results in a change in absorption maxima of the chromophore. A plot of λmax versus generation number with CCl4 as the solvent showed a discontinuity in progressing from the third to fourth generation. The discontinuity corresponds to the transition from an open, extended structure at the third generation to a spherical, globular structure at the fourth generation.

This assertion agrees with intrinsic viscosity

measurements that show a similar transition at the fourth generation.31

25

Figure 1.9 Dyes as solvatochromic probes. Space-filling models of dendritic pnitroanilines illustrate the increased shielding of the chromophore from bulk solvent. The increase in the local polarity of the solvatochromic probe’s microenvironment induces a change in its absorption spectra that is generation dependent (image adapted from reference 58.) A report by Moore also notes an anomalous shift in the fluorescence spectra of higher generation phenylacetylene dendrimers equipped with an electron donating moiety at the focal point.80 Modulations of the fluorescence maximum of the charge transfer state were shown to be dependent on dendrimer generation in nonpolar hydrocarbon solvents. While change in the position of the charge transfer emission is small up to generation 4, an abrupt change was observed in the spectra at the fifth generation. The onset of an encapsulated, globular conformation at higher dendrimer conformations is again implicated in the modulation of the photophysical properties of the core. David Smith demonstrated that individual dendritic branches are sufficient to encapsulate dyes in a microenvironment that can alter optical properties.81 Dendritic branches constructed from BOC-protected lysine building blocks with a carboxylic acid focal point solubilized proflavine, an amine-containing hydrophilic dye, in CH2Cl2. The size of the dendritic branch influenced the absorption maximum of proflavine as a 26

bathochromic shift was observed with increasing branch generation. Since the absorption properties of proflavine depends on its protonation state, encapsulation within the dendritic branch likely promotes an increase in the degree of proton transfer between the carboxylic acid and amine, leading to the observed bathochromic shift. This result suggests that, like fully elaborated spherical dendrimers, judiciously functionalized dendritic branches themselves can encapsulate a dye molecule and offer a microenvironment that modulates absorbance properties. Solvatochromic dyes have also been used to probe the interior of PAMAM dendrimers of various generations.

Pistolis et al described the use of pyrene as a

solvatochromic probe of low generation PAMAM interiors.82 This study determined that the larger the dendrimer, the better the fluorophore is sequestered from bulk aqueous solutions. However, this investigation was limited to PAMAM dendrimers of less than three generations, structures that are known to exist in extended, open conformations. The inherent fluorescence of PAMAM dendrimers, the potential formation of excitedstate complexes, and fluorescence quenching by the amine groups of the dendrimer also complicate fluorescence-based probing of PAMAM dendrimers. Reichardt’s dye (1.10), an absorption-based solvent polarity probe, has also provided insight into the nature of PAMAM dendrimers.83 A decrease in local polarity at the dendritic core in aqueous media was responsible for a shift in the solvatochromic band of the dye’s absorption spectrum, and the resulting λmax corresponded to a microenvironmental polarity analogous to 1-decanol. While the authors demonstrated that Reichardt’s dye penetrated the dendritic surface, their next study sought to provide a more complete description of the interior microenvironment and employed a smaller probe that could be fully encapsulated by the PAMAM dendrimer.

27

O N

N

O

N

1.10

1.11

In this subsequent study, the investigators chose the smaller phenol blue dye (1.11) as the solvent polarity probe.84 In addition to a solvatochromic absorption band, phenol blue also has a solvatochromic fluorescence emission which, unlike that of pyrene, is not quenched by amino-terminated PAMAM dendrimers. Such persistent fluorescence allows for a more detailed analysis of the nature of these dendrimers. The authors observed an increase in phenol blue’s absorption band centered at 560 nm with increasing dendrimer generation. The increase in absorption intensity at this wavelength suggests the probe’s transition to a rather nonpolar microenvironment. While the absorption spectrum of phenol blue with the smallest generation dendrimers (G0 and G1) was similar to that observed in water, the intensity of the solvatochromic band at 560 nm steadily increased for generations 2-4 and suggested a local polarity similar to that of furan. This is much less polar than the 1-decanol-like microenvironment elucidated in the previous study using Reichardt’s dye. Because both studies were conducted under same experimental conditions, the observation of a less polar microenvironment in this case suggests that the smaller phenol blue penetrates further into the dendrimer’s interior. The absorption intensity at 560 nm was mostly constant for generations 4-7 but experienced a relative decrease at generation 8.

28

This anomaly is attributed to the

increased surface group density at the eighth generation that limits the dye’s ability to effectively penetrate the dendrimer periphery. In the same study, a blue shift in the fluorescence emission of phenol blue was also observed in the presence of PAMAM dendrimers, suggesting a decrease in local polarity. In parallel to the absorbance studies, larger generation dendrimers showed similar fluorescence profiles with the exception of generation 8, which had decreased emission intensity. Anisotropy values also indicate minimal rotational motion for the higher generation dendrimers (G6-G8), but no correlation between dendrimer generation and microenvironmental rigidity was elucidated. Based on these results, the following generalizations about amino-terminated PAMAM dendrimers can be made: 1.

Low generation PAMAM dendrimers (G0-G1) offer little protection of a nonpolar solute from aqueous media.

2.

A dendritic microenvironment of decreasing polarity is observed up to generation 3.

3.

Generations 4-7 have roughly equal capacities to solvate a nonpolar solute in water.

4.

Higher generation dendrimers (G8) exhibits anomalous spectral behavior as the surface group density nears its theoretical limit.27 These generalizations indicate that a balance exists between PAMAM dendrimer

size and indicator uptake. While there is a generational threshold for adoption of the globular conformation that induces true encapsulation, the increased surface density of high generations eventually hinders the dye’s ability to penetrate the dendrimer. It is also worthwhile to note that PAMAM dendrimers of intermediate generation (4-7) all provide a microenvironment of similar polarity and offer no increased protection of indicator as a function of size.

29

1.5.2 Probing the Dendritic Surface Information about the surface properties of dendrimers has also been gleaned from spectroscopic analyses.

In these studies, the interaction between probe and

dendrimer is usually based on electrostatics. An abundance of charged surface groups allows dendrimers to serve as polyions that can aggregate oppositely charged dyes on their surface.

If the nature of the dye allows this aggregation to be detected

spectroscopically, then the charged dye serves as a probe of the dendrimer surface. Tomalia and Turro have investigated polyanionic half-generation PAMAM dendrimers with cationic structural probes [a “half-generation” PAMAM dendrimer (G=n.5) refers to one that terminates in a carboxylate group rather than the amine termini of full generation (G=n.0) PAMAM dendrimers.]85 Using methylene blue (1.12) as the spectroscopic probe in UV-vis analysis, the researchers determined that the extent of dye aggregation at the surface strongly depends on dendrimer generation. The authors note that the observed aggregation becomes more pronounced at higher generations but state that this effect is not solely due to the greater number of surface groups since aggregration behavior tracks similarly with both dendrimer concentration and surface group concentration. This study also documented a dramatic break in the aggregation effects at generation 3.5. This observation parallels previously described evidence of a fundamental change in dendritic structure at the third generation, and the authors speculate that the larger separation of surface groups found in low generation dendrimers precludes aggregate formation. A subsequent study examined the amino-substituted surfaces of full generation PAMAM dendrimers.86 Negatively charged fluorescein (1.13) was selected to probe the surface of these cationic polyelectrolytes. Results were similar to those observed in the study of half-generation dendrimers with methylene blue. A generational dependence on 30

fluorescein surface aggregation was described, with no aggregation observed for generations 0-2 and a discontinuity in fluorescence intensity at the third generation. These results are again attributed to the open structure of lower generation PAMAM dendrimers and the transition to a closed surface structure at generation three. This report is especially interesting in light of the generational dependence described in Chapter 2 for the interaction of fluorescein and its derivatives with amino-terminated PAMAM dendrimers (see section 2.4.1).

N N

S

CO2 N Cl

1.12

HO

O

O

1.13

1.6 DYES AND INDICATORS IN APPLICATIONS OF DENDRIMERS In addition to providing insight into their structure, dyes and indicators have also been integral in the demonstration of early applications of dendrimers.

These

applications again exploit the dendrimer’s ability to encapsulate dyes and provide sanctuary from bulk media. Enhanced solubilization of guests, liquid-liquid extraction of guests, and controlled entrapment and release of guests are all examples of dendrimer applications whose initial demonstration involved spectroscopic probes. 1.6.1 Solubilization of Guest Molecules Dendrimers are often referred to as “unimolecular micelles” capable of serving as containers for small molecular guests. Micelles are aggregates of amphiphilic molecules in which a polar “head” group and a hydrophobic “tail” adopt an arrangement in water that permits solvation of the polar head group while minimizing contact between the 31

aqueous solvent and the hydrophobic tails.87 The result is a globular structure that is reminiscent of higher generation dendrimers; and like dendrimers, micelles have enjoyed success in drug solubilization and delivery applications.88,89 Micelles, however, require a threshold concentration of surfactant because their formation is governed by free energy. This

critical

micelle

concentration

(CMC)

renders

micelle-based

systems

thermodynamically unstable under certain conditions. Dendrimers represent alternative micelle-like constructs that are covalent in nature and therefore not constrained by CMC considerations (Figure 1.10). A.

B.

Figure 1.10 Dendrimers as unimolecular micelles. (A.) Non-polar guests are encapsulated by traditional polymolecular micelles. (B.) Dendrimers can be considered “unimolecular micelles” also capable of encapsulating and solubilizing molecular cargo (image adapted from C.R. Chimie 2003, 6, 715-724). Hawker and Frechet provide an early analogy of dendrimers to traditional micelles.

In their report, the researchers investigate the capability of carboxylate-

terminated polyether-based dendrimers to serve as unimolecular micelles.90

Like

traditional micelles, these dendrimers offer charged end groups surrounding a relatively apolar interior and should likewise assist in the solubilization of hydrophobic molecules in polar solvents. Indeed, the presence of these amphiphilic dendrimers provided a 120fold increase in solubilized pyrene compared to that observed in pure water.

This

solubilizing power is comparable to that seen for sodium dodecyl sulfate micelles, and 32

the linear relationship between the solubilizing ability of the dendrimer and its concentration suggests there is no operational concentration threshold as in micelles. This study documents that dendrimers can offer solubilization properties analogous to traditional micelles while remaining active over wider concentration ranges. Dendrimer synthesis also allows the creation of “inverted” unimolecular micelles. In contrast to traditional micelles, inverted micelles feature a polar interior decorated with an apolar periphery and thus should enhance the solubility of hydrophilic compounds in organic media. Meijer reports the synthesis of such a dendrimer.91 By functionalizing the surface of hydrophilic poly(propylene imine) dendrimers with alkyl chains, an inverted unimolecular micelle (1.14) is created. The micellar nature of this structure was confirmed by its trapping of the hydrophilic dye Rose Bengal in ethanol and the improved compatibility between the dendrimer encapsulated dye and hexane.

1.14 33

1.6.2 Extraction of Guest Molecules The improved solubility imparted by dendritic micelle mimics soon led to applications in liquid-liquid extractions. Meijer, using the same alkylated PPI dendrimers employed in Rose Bengal solubilization, studied the extraction properties of other anionic xanthene-based dyes.92

The results indicated that solutes such as fluorescein were

efficiently transferred to the organic layer at low to neutral pH values. Because such extraction was not observed for cationic probes, the authors suggest the dendrimer’s extractant properties are the result of its basic functionalities and identify a direct relationship between the number of interior tertiary amines and the number of dye molecules extracted per dendrimer. These results inspired the subsequent synthesis of PPI dendrimers functionalized with a fluorinated shell.93 The resulting dendrimer was insoluble in water and most organic solvents but soluble in liquid CO2.

In a remarkable demonstration, this

dendrimer transferred methyl orange, a CO2-insoluble dye, from aqueous solution into liquid CO2. Complete extraction was effected after 150 minutes, and a maximum of 12 dye molecules per dendrimer was calculated. This seminal demonstration, along with Meijer’s work with xanthene-based dyes, laid the foundation for practical applications of similar systems. For example, an amphiphilic, hyperbranched polyester with the ability to encapsulate polar organic guests is reported to serve as a necessary dye carrier in the colorization of polypropylene fibers.94 1.6.3 Controlled Release of Guest Molecules The aforementioned studies prove that judiciously functionalized surfaces can imbue dendrimers with unique extraction properties.

In addition to this solubility

regulation, widely referenced reports by Meijer suggests that surface groups can also modulate the dendrimer’s ability to serve as a molecular container.95,96 The previously 34

described examples of dye encapsulation were subject to reversibility, with the encapsulated guest free to depart the dendrimer’s interior based on free energy considerations. Meijer’s seminal study, however, demonstrated that dye molecules could not only be encapsulated but also physically entrapped by high generation dendrimers. Meijer and coworkers coupled tBoc-L-phenylalanine to the surface of a fifth generation PPI dendrimer. When this protected amino acid shell was constructed in the presence of a dye such as Rose Bengal, the dye molecules were physically trapped in the interior voids of the dendrimer, and their diffusion out of the dendrimer was immeasurably slow (Figure 1.11). Entrapment of the dye is ascribed to the increased molecular density of the dendrimer surface that is augmented by intramolecular hydrogen bonding. The authors term this locked molecular container a “dendritic box” and report a solid phase-like character of the surface shell based on 13C-NMR relaxation data.

35

Figure 1.11 Meijer’s dendritic box. Large dye molecules such as Rose Bengal are physically entrapped within the voids of the dendrimer upon coupling of the protected amino acid surface groups (image adapted from reference 7). Meijer’s following study demonstrates a shape-selective release of entrapped guests from this dendritic box. Such ability to tune the permeability of the dendritic surface has important implications in drug delivery applications.

Simultaneous

entrapment of Rose Bengal and p-nitrobenzoic acid in the dendritic box was accomplished. The dendritic box contained four molecules of Rose Bengal and 8-10 molecules of the smaller p-nitrobenzoic acid as determined by UV-vis spectroscopy. Subsequent hydrolysis of the BOC groups followed by dialysis promoted release of all the p-nitrobenzoic acid molecules from the box while the larger Rose Bengal molecules remained confined to the interior. Further hydrolysis of the amide bonds, however, recreated the unmodified PPI dendrimer with a relatively open surface structure that 36

allowed the liberation of Rose Bengal. This shape-based release of molecular guests is shown to be a general principle as both FMOC-protected amino acids and various mixtures of guest molecules are also amenable to the creation of a selectively permeable dendritic box. In addition, variation of the amino acid shell component offers tunability to the shape selective release of nitrophenol and methylene violet.

While not

immediately applicable to drug delivery applications, Meijer’s illustration of a two-step selective liberation of dendrimer guests provided a conceptual basis for subsequent practical systems. An example of a system seemingly inspired by Meijer’s work is Balogh’s dendrimer-encapsulated nanoparticles which show in vitro antibacterial activity due to a controlled release of silver ions.97 Minghui Chai offers a non-covalent “twist” on the dendritic box that is particularly relevant to the work described in Chapter 2. In this report, Chai describes the use of self-assembly on the dendrimer surface in the entrapment of fluorescein.98 Using an amino-terminated, third generation PPI dendrimer, Chai created a water-soluble selfassembly by associating dibasic adipic acid onto the positively charged dendrimer surface. The “capped” surface “locks” fluorescein in the interior voids of the dendrimer (Scheme 1.3). This non-covalent assembly was stable over the pH range 6.0-9.0 but could be disrupted at extremely high or low pH values.

37

Scheme 1.3 A non-covalent “twist” on the dendritic box. Association of adipic acid on the surface of a PPI dendrimer effectively traps fluorescein within the cavities of the dendrimer. The authors of this study also document a change in fluorescein structure upon encapsulation. A small red shift in the UV-vis spectrum of fluorescein upon addition of dendrimer suggests an increase in the conjugation of the dye.

13

C NMR experiments

confirmed fluorescein’s structural change upon encapsulation in the dendrimer. Fluorescein is known to exist in two equilibrating forms: a lactone and a carboxylate (Scheme 1.4). As a result of sp2 hybridization at C7, the carboxylate form is the more conjugated structure. Therefore, the recorded spectral data suggests a shift to the carboxylate form upon interaction with the dendrimer. The authors propose that this transition to a more conjugated structure provides a flatter molecule that is more easily incorporated into the dendrimer based on steric considerations. Once encapsulated, the fluorescein molecule is further sequestered by the addition of adipic acid which restricts the motion of the dendritic branches, thereby reducing the dye’s rate of exchange between the dendritic microenvironment and bulk media.

38

7

7

1.15

1.16

Scheme 1.4 Equilibration of the two forms of dianionic fluorescein. In solution, fluorescein exists in one of two forms: a carboxylate (1.16) or a lactone (1.15). A shift to the more conjugated carboxylate form is proposed upon binding to the dendrimer. 1.7 DENDRIMERS IN MOLECULAR RECOGNITION The survey of literature provided thus far has been mostly limited to non-specific interactions of guest molecules with the dendrimer core and branches.

The

aforementioned unimolecular micelles and Meijer’s dendritic box rely on ill-defined hydrophobic effects and lack the specific binding sites found within enzymes and pervasive in abiotic molecular recognition. Dendrimer synthesis, however, offers the possibility of a precise placement of specific binding sites within the dendritic architecture. Molecular recognition can occur inside the dendrimer’s core, among the dendritic branches, or on its multivalent surface (Figure 1.12). Functionalization of such areas with defined recognition sites should confer unique binding properties that result from the nature of the dendritic core or the array of surface binding sites. The following section will highlight selected examples of instructed molecular recognition within each of the dendritic regimes (core, branches, and surface) and will also provide an entry to a discussion of dendrimers in molecular sensing applications.

39

Figure 1.12 Topological regions for dendrimer-based molecular recognition. Dendrimers feature three distinct regions to exploit molecular recognition phenomena: an encapsulated core, the branched repeat units, and the multivalent surface. 1.7.1 Recognition at the Core The suggested use of “shell” molecules in molecular recognition occurred early in the chronology of dendrimers.99 Numerous examples have been reported since then, and this topic has been the subject of several reviews.6,7,8,100 A seminal report in this area comes from the Diederich laboratory and describes the synthesis and molecular recognition properties of dendrophanes.101 Inspired by the apolar binding sites buried within globular proteins, Diederich functionalized cyclophane cores with poly(ether amide)-based dendritic wedges that conferred water solubility. Cyclophanes are cyclic compounds containing aromatic rings that have enjoyed success in the molecular recognition of hydrophobic guests.102 By incorporating specific cyclophanes into dendrimers, Diederich sought to achieve specific complexation of arenes and steroids at the dendritic core in an effort to study the influence of the dendritic microenvironment on

40

the affinity and kinetics of these interactions. His results represent the first example of a water-soluble dendrimer with a defined recognition site at the core (1.17). 1

H NMR titrations revealed a specific binding of aromatic molecules to the

cyclophane moieties in contrast to the nonspecific apolar inclusion of hydrophobic guests by unimolecular micelles. Dendrophanes based on a tetraoxa[6.1.6.1]paracyclophane initiator core formed 1:1 complexes with benzene and naphthalene in aqueous media while those based on a [6.1.6.1]cyclophane core bound testosterone in a 1:1 mixture of water and methanol. Three generations of each series were investigated to determine the availability and polarity of the core binding site as a function of dendrimer size. The researchers determined that the binding sites remain accessible at all investigated generations, and that the resulting inclusion complexes have comparable stability to those involving the non-dendritic cyclophanes. Fluorescence titrations using a spectroscopic probe showed the expected decrease in the local polarity of the medium at the dendrimer core. This effect became more pronounced up to the third generation. Despite the decreased polarity provided by the dendritic microenvironment, no increase in binding affinity was observed relative to the parent cyclophanes (~103 M-1). This paradox is explained by a relatively open dendrophane structure that is confirmed by kinetic measurements.

Each investigated dendrophane showed surprisingly fast host-guest

exchange rates, but a decrease in the rate of exchange was observed upon increasing generation. The rapid host-guest exchange observed with these dendrophanes stand in sharp contrast to Meijer’s sterically enforced dendritic box which maintained substrate encapsulation for hours. Because fast exchange rates render a rate-limiting product release unlikely, dendrophanes soon found application as enzyme mimics.103

41

HELVETICA CHIMICA ACTA- Vol. 80 (1997)

Diederich also targeted polar

2393

3 R=COOH 12 1.17 R=COOMe 13 R = CONHC(CH20CH2CH2COOMe)3 14 R = CONHC(CH20CH2CH2COOH), compounds such as sugars in studies

involving

recognition at the core of dendritic hosts. His so-called “dendroclefts” featured polyether groups arrayed around an optically active core, and the effect of the dendritic branching on the stereoselectivity of the recognition event was investigated by 1H NMR titrations in CDCl3.104 Optically pure dendrocleft 1.18 was synthesized from a 9.9’-spirobi[9Hfluorene] initiator core containing 2,6-di(carboxamido)pyridine moieties.

t

0 4 R=OH

42 23 R = OMe 5 R = NHC(CH~~CH~CH~COZH)~ 24 R = NHC(CH20CH2CH2C02Me)3

R

of a functional core.2 Hydrogen-bonding dendritic hosts have been reported, but the branching does not appear to play an active role in modulating guest recognition.3 In addition chiral recognition inside a dendrimer is as yet unknown, but possesses great scientific and technological potential.4 Here we report enantiomerically pure dendritic cleft-type receptors (dendroclefts) of first [(2)-G1] and second [(2)-G2] generation for the chiral recognition of monosaccharide guests via Hbonding. O O O O O O O

O O

O O

O

O

O

O

O O

O O

O O

O

O

O

O

O

O

O O

O O

O O

O

O

O

O

O

O

O O

HN

O

O

O N

(–)-G2

H O N N

H

H

N N

O

H

N

O

O

O

NH

O O

O

O

O

O

O O

O

O

O O O

O O

O O

O

O

O

O O

O O

O

complexes are summarized in Table 1. T sions can be drawn. (i) The complexes formed by the den (2)-G2, and core (2)-G0 are of similar 100 and 600 M21). Hydrogen bonds betw sugars and the NH groups of the receptors guest interactions in all complexes as e complexation-induced downfield shifts saturation binding) of the NH resonan amido)pyridine moieties. Apparently, the in (2)-G1 and (2)-G2 does not preven from penetrating the receptor and interac bonding sites. It is actually quite remarka (2)-G1 and (2)-G2 is not weakened b which contains a relatively high densit petitive donor oxygen atoms. (ii) The degree of enantioselectivity in the enantiomeric a-glucosides 6 and attachment of the dendritic shells. The between diastereoisomeric complexes D 3.6 kJ mol21 [(2)-G0], to 0.8 kJ mol21 [ mol21 [(2)-G2]. (iii) On the other hand, the diastereo plexation are remarkably enhanced by dendritic branches. Thus, the difference i complexes of the diastereoisomeric gue from 0.7 kJ mol21 [(2)-G0], to 1.4 kJ m 2.3 kJ mol21 [(2)-G2]. HO O

O O

O

O

O

O O O

O

O O O

O O

HO H17C8O

OH OH OH

Octyl !-L-glucoside 6

HO HO

O OH OC8H17

Octyl !-D-glucoside 7

O O

These results indicate that the dendrit the selectivity of complexation at the co result. There are two plausible reaso modulation of binding selectivity whic flexible branches to a rigid, optically pure 9,9A-spirobi[9HBased on NMR data, the recognition event is predicated on hydrogen bonding investigation. Firstly, the steric dema fluorene] initiator core bearing 2,6-di(carboxamido)pyridine 5 branching may disfavour certain comp moieties in the 2,2A-positions. The resulting dendrimer possesses buried H-bonding cleft selected suitable glucopyranoside for complexing oxygen donor atoms in the dendritic she interactions between theaoxygen atoms of the targets and the carbohydrate guests.5,6 Its periphery is functionalised with the formation of a hydrogen bonding ne neutral polyether groups, which provide excellent solubility in both changing the binding selectivity.‡ NH donor groups of the core’s di(carboxamido)pyridine groups. While dendroclefts a wide range of solvents, including H2O. The use of such dendroclefts as ch For the synthesis of (2)-G1 and (2)-G2 by the convergent currently under active investigation. Pro G1 and G2 (as wellapproach, as the non-dendritic core core molecule G0)prepared did notfrom exhibit circular any significant 7 the optically active (2)-3 was dichroism (CD) spectra are obser 5,8 dicarboxylic acid (2)-(R)-1 via (2)-2 and (2)-G0 (Scheme glucopyranoside guests, the response

1.18 Dendrocleft (2)-G2 was targeted by the attachment of

changes in overall complex stability as a result of increased branching, the researchers

Chem. Commun., 1998

did note a diastereoselectivity in the recognition of the anomers of octyl-D-glucoside. This diastereoselectivity was enhanced with increased branching, an observation the authors attribute to the steric requirements imparted by the dendritic branches and/or the involvement of the oxygen atoms of the polyether periphery in a hydrogen-bonding network with guest. Regardless of the source of diastereoselectivity, this report provides a novel demonstration of the dendritic surface shell controlling the selectivity of a recognition event at the core. 43

1.7.2 Recognition within the Branches Defined molecular recognition processes can also occur within the repeating units of the dendritic branches, and the multiple recognition sites offer the potential generation of novel recognition properties such as cooperative binding. Shinkai appreciated these possibilities and synthesized “crowned arborols,” dendritric receptors containing multiple crown ether binding sites within the branches.105

Subsequent investigation into the

binding properties of these crowned arborols, however, revealed that the crown ether moieties behave independently in their molecular recognition of cations and do not show cooperative complexation effects.106 The absence of cooperative metal complexation was established by a determination of 1:1 stoichiometry between the crown ether group and metal cation guest. This result stands in contrast to linear polymeric crown ethers which form 1:2 cation:crown ether complexes by adopting a sandwich conformation. While Shinkai failed to observe any emergent binding properties of his crown ether-based dendrimers, Sanders describes interesting cooperative binding effects using a metalloporphyrin dendrimer.107 Nine porphyrin rings are connected by a combination of rigid and flexible linkers, and the resulting independent dendritric arms fold in a predictable manner in the presence of a coordinating, bifunctional ligand such as DABCO.

The result is a cooperative binding event as the first coordination event

accentuates the binding of a second equivalent of DABCO by providing a preorganized recognition site. The authors speculate that such a conformational sensitivity that leads to cooperative binding might have significant effects on the electrochemical and photophysical properties of other classes of dendrimers. 1.7.3 Recognition at the Surface The work previously described in this chapter demonstrates that the dendrimer’s surface can modulate the selectivity of molecular recognition processes at the core 44

(dendroclefts), govern macromolecular properties such as solubility (unimolecular micelles), and affect the dendrimer’s ability to serve as a molecular container (the dendritic box). In addition to regulating such properties, the dendritic surface itself provides a unique molecular landscape for parlaying individual binding events into enhanced recognition phenomena. This enhancement arises from the surface’s ability to form multiple host-guest interactions, a feature that could conceivably amplify the strength of interactions that are weaker in isolation. Such amplification of binding affinity is observed in multivalent biological systems such as antibodies and viruses and is therefore expected to occur in synthetic analogues such as dendrimers.108 Astruc provides an example of the “dendritic effect” in surface binding using an amido-ferrocene terminated dendrimer (1.19) in the recognition of inorganic anions.109,110 Cyclic voltammetry experiments using small anionic guests such H2PO4- and HSO4provide evidence of strong dendrimer-anion binding that results from an electrostatic interaction between the ferricinium cation and the anionic guest that is complemented by the hydrogen bonding capacity of the amide NH. Enhanced recognition of H2PO4- and HSO4-, as ascertained by cyclic voltammetry, occurs upon increasing the number of ferricinium surface groups from 3 to 9 to 18.

The authors attribute this observed

dendritic effect to the increased surface packing of the recognition moieties at higher generations. This enhancement, however, was found to be maximal for 18 surface recognition groups as the periphery becomes sterically saturated between 18 and 36 end groups. Thus, shape selectivity defines the extent of the dendritic effect. The requisite hydrogen bonding to the amide group demands that the anionic guest penetrate the dendrimer surface to some degree, a requirement that becomes increasingly difficult to meet as the surface groups proliferate with each dendritic generation.

45

1.19 1.7.3.1 Binding of Tripeptides Reports from the Meijer lab provide examples of surface group recognition that are particularly germane to the dissertation research described in the following chapter. PPI dendrimers modified with adamantyl surface groups attached via a thiourea linker served as multivalent hosts for guest molecules containing a terminal urea-glycine unit.111 Based on NMR experiments, the guests bind to the dendrimer through hydrogen bonding interactions between the urea group of the guest and the thiourea moiety of the dendritic host (Scheme 1.5). Ionic bonding between the carboxylate terminus of the guest and the tertiary amines of the dendritic branches is also implicated in the recognition event, and isothermal calorimetry measurements determined binding constants ~ 2 x 104 M-1 in CHCl3. 46

Scheme 1.5 Molecular recognition at the dendritic surface. Electrostatic interactions and hydrogen bonding drive the association of guests containing terminal urea-glycine groups with the thiourea moieties of the modified PPI dendrimer-based host. Recognizing the structural similarities between the urea-glycine “tails” of the guests and the carboxy termini of peptides, Meijer extended the use of these dendrimers to the multivalent recognition of short amino acid sequences.112 The authors note that addition of the thiourea-modified PPI dendrimer to a suspension of tripeptide in CHCl3 afforded a soluble complex in CHCl3, and IR and 1H NMR experiments confirm the binding of tripeptides such as N-Boc-Gly-Gly-Gly-OH to the dendrimer via hydrogen bonding and ion pairing. Investigation of a series of seven (Boc)-protected tripeptides revealed that the extent of hydrogen bonding to the dendrimer decreased with increasing bulkiness of the side chains. Meijer exposed the dendrimer to a mixture of the seven peptides to assess its ability to serve as a host for peptides with varying side chains.

After mixing the

dendrimer with four equivalents of each tripeptide, the resulting complex was treated with a 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) cleavage mixture to release the bound peptides from the dendrimer host. HPLC analysis of the cleavage mixture indicated the presence of all seven tripeptides, suggesting that the dendrimer successfully binds peptides regardless of side chain size (Biobeads filtration removed loosely bound and adhered peptides from the dendrimer prior to HFIP cleavage.) In fact, based on analysis 47

of the cleavage mixture, the bulkier peptides were determined to be present in larger quantities than the less sterically demanding peptides. This observation contradicts the NMR data which suggested that the bulkier side chains weakened the tripeptidedendrimer interaction through steric repulsion. The authors argue that the hydrophobic side chains of the bulkier peptides impart an increased solubility in CHCl3; therefore, the bulkier, more soluble tripeptides are present in solution at greater concentrations, a situation that leads to the observed binding of all seven tripeptides. 1.7.3.2 Metal Coordination Dendrimers also offer the potential for the formation of multiple metal ion complexes on their surfaces. Interest in surface-metallated dendrimers stems from prior uses of metal ions in catalytic and medicinal applications. The catalytic efficacy of various metal ions has been well documented for many synthetic transformations, and lanthanide complexes are used as contrast agents in magnetic resonance imaging (MRI).113

Surface-metallated dendrimers are envisioned as ideal macromolecular

catalysts since their spherical nature ensures accessibility of catalytic sites while their high molecular weight offers the recyclability of heterogeneous catalysts; and the slow diffusion associated with dendrimer-based contrast agents should allow for improved imaging of vascular structures. The studies described below, however, are limited to fundamental investigations into surface-metallated PAMAM dendrimers, and discussion of any related applications of these structures is neglected. Ottaviani et al described the binding of Cu(II) ions to amino-terminated PAMAM dendrimers through the use of electron paramagnetic resonance (EPR).114

Because

protons compete with Cu(II) ions for binding to the surface amine groups, no complexation was observed at pH < 3.5. An increase in pH allows for deeper penetration of the metal ions into the dendrimer’s structure, and several binding modes were 48

elucidated between pH 4 and 5.

These complexes included a Cu(II)-N2O2 species

comprised of two amino termini and two water molecules and a Cu(II)-N4 complex involving two internal tertiary amines and two surface primary amines (Figure 1.13). When the pH was increased > 6.0, the Cu(II)-N4 complex was the only observed species, and the authors conclude that high pH is necessary to effect inclusion of the metal ions. Copper laden PAMAM dendrimers have also been isolated by ultrafiltration and studied by atomic absorption spectrophotometry.115 Compared to traditional chelating agents which often bind only one metal ion per molecule, the amino-terminated PAMAM dendrimers were shown to have significantly larger binding capacities. The researchers report maximum extents of binding of 8, 13, and 29 ions per dendrimer for generations 3,4, and 5, respectively, while generations 6, 7, and 8 showed maximum uptake of 46, 83, and 153 Cu(II) ions. The relationship between maximum extent of binding and number of terminal NH2 groups was also investigated to provide insight into the nature of the coordination event. The resulting plots were compared to those associated with known bidentate and tetradentate copper coordination, and the data suggests a tetradentate coordination mode for dendrimers G3, G4, G5. Neither coordination model, however, fit the experimental data obtained using higher generation dendrimers (G7 and G8). The authors insist that while both primary surface amino groups and tertiary interior amines are involved in Cu(II) binding, their relative roles are modulated by factors such as dendrimer generation and solution pH.

49

Figure 1.13 Coordination modes of Cu(II) binding to PAMAM dendrimers. Ottaviani described two binding modes of Cu(II) to the surface of PAMAM dendrimers: one in which the Cu(II) is complexed to two surface amino groups and two water molecules [Cu(II)-N2O2] and another in which the metal ion is coordinated with two primary amino groups and two tertiary amino groups [Cu(II)-N4]. Crooks studied the partitioning of Cu(II) ions into PAMAM dendrimers in an effort to prepare metal nanoclusters.116 The addition of dendrimer to a Cu(II) solution promoted a shift in the λmax associated with the metal’s d-d transition. This binding was again determined to be pH dependent as the ions could be released from the interior of a hydroxy-terminated PAMAM dendrimer at low pH.

Crooks reports that an amino-

terminated G4 PAMAM dendrimer bound a maximum of 36 copper ions, primarily through the primary amine surface groups. As a result, copper nanoclusters prepared from amino-terminated PAMAM dendrimers are larger in size than those prepared from their surface-hydroxylated cousins. 1.8 DENDRIMERS IN MOLECULAR SENSING The examples highlighted in the previous section illustrate that chemoselective, diastereoselective, and shape selective molecular recognition behavior can be achieved 50

through the rational design of dendritic host molecules. True functionality, however, demands that these requisite recognition events be parlayed into a tangible result. In the realm of molecular sensing, that response takes the form of optical signals. To this end, the following section will detail successes in extending dendrimer-based molecular recognition processes into platforms for the optical detection of various substrates. Discussion of dendritic sensors based on other detection modes such as quartz microbalance117 and conductance118,119 is omitted. James and Shinkai offered one of the earliest reports of optical dendritic sensors.120 Extending their previous work on photoinduced electron transfer (PET)-based fluorescence sensors for saccharides, James and Shinkai appended boronic acid moieties onto the surface of a PAMAM dendrimer to give a dendritic fluorescent sensor featuring eight boronic acid recognition sites and eight anthracene signaling units (1.20). Increase of the dendrimer’s fluorescence emission at 423 nm accompanied the binding of various saccharides. When compared to a non-dendritic model sensor, the dendritic saccharide “sponge” showed enhanced binding towards sugars such as D-galactose and D-fructose. In the case of D-galactose, this increased affinity is attributed to the formation of a 2:1 boronic acid:sugar complex whose creation is assisted by the presence of multiple recognition sites.

Because of the dendrimer’s inherent flexibility, binding of the

saccharide to one boronic acid recognition site allows a second interaction to occur between the bound sugar and any one of the remaining seven boronic acid groups.

51

B(OH)2

B(OH)2

N

N

NH

HN

O

O

O

O N (HO)2B

N N

N H

O

O

N

N H

B(OH)2

NH

HN N

N

NH HN

O

O (HO)2B N

H N

H N

N

N

B(OH)2 N

O

O

O O

NH

HN N

N B(OH)2

B(OH)2

1.20 1.8.1 Signal Amplification Work by Balzani and Vogtle illustrates another unique advantage offered by dendritic sensors:

signal amplification.

In many traditional fluorescent sensors,

coordination of a guest quenches the excited state of an appended fluorophore, leading to attenuation of the fluorescent signal.121 Dendrimer synthesis permits the presentation of an array of fluorescent signaling units whose simultaneous quenching by a guest molecule leads to signal amplification. In Balzani and Vogtle’s report, the surface of a fourth generation PPI dendrimer is decorated with 32 dansyl units.122 Coordination of Co+2 to the dendrimer’s interior tertiary amines resulted in quenching of the dansyl fluorophores.

Detailed investigation of this quenching showed a linear relationship

between fluorescence intensity and Co+2 concentration at low metal ion concentrations. 52

This linear relationship suggests that a single cobalt ion effectively quenches the fluorescence of the 32 dansyl units. The result is a putative 32-fold increase in sensitivity relative to a mono-dansylated sensor. Lin Pu reports analogous sensitivity effects observed with his dendritic sensors. Incorporation of an optically active binapthol (BINOL) core into a phenyleneethynylene dendrimer afforded an enantioselective fluorescent sensor whose sensitivity is increased relative to an analogous small molecule sensor.123,124 While various chiral amino alcohols quenched the fluorescence of the dendritic sensor upon interaction with the (S)-BINOL core, the enantiomers of phenylalaninol did so at significantly different rates, with the S enantiomer quenching more efficiently than the R enantiomer. Quenching efficiency profiles generated using the R dendrimer showed the opposite trend, confirming that the observed differences in quenching arise solely from chiral interactions. The heightened sensitivity to quenching observed for the dendritic sensor relative to a parent binapthol is attributed to a signal amplification provided by the conjugated dendritic branches. These branches serve as light harvesting antennas that funnel energy to the BINOL recognition unit where it is quenched by the amino alcohol guest. Analysis of the Stern-Volmer constants also suggests that fluorescence quenching of the dendritic BINOL by phenylalaninol is more enantioselective than that observed with an isolated parent BINOL sensor. So not only is the fluorescence of the dendritic sensor more sensitive to amino alcohol quenchers than the parent binapthol, the dendritic structure also provides a relative increase in the enantioselectivity of the chiral recognition event. The observed signal amplification and increased selectivity of this sensor represents two advantages offered by the use of dendritic scaffolds in molecular sensing.

53

The advantages associated with dendritic sensors have led to their directed use in the detection of “real-world” analytes. A recent example is a fluorescent dendritic system that responds to the presence of TNT.125 In analogy to Balzani and Vogtle’s work, the researchers functionalized the dendritic surface with a chromophore known to be sensitive to the presence of electron deficient quenchers such as TNT. The fluorescence of a G4 dendrimer containing 96 signaling units on its surface was effectively quenched in the presence of TNT. The associated quenching constant (Ksv) for this process was twice as large as that calculated for a previously reported polymeric sensor under similar conditions.

Time-resolved fluorescence spectroscopy and anisotropy measurements

determined that this amplified quenching likely originates from exciton migration from the excitation site to the recognition site via the dendritic surface. Other examples of “real-world” dendritic sensors include systems used for the optical detection of bacteria,126 glucose,127 and tissue oxygen.128 1.9 CONCLUSION AND OUTLOOK Dendrimers occupy a unique place in supramolecular chemistry. This standing derives from the inherent properties of dendrimers that create a veritable playground for the supramolecular chemist. The multivalent surfaces and diverse microenvironments, along with a well-defined presentation of functional groups at a high molecular density, render this class of macromolecules especially suited for novel forays into the areas of molecular recognition and sensing.

Insight into the nature of dendritic structure,

however, is required for a thorough understanding of these phenomena. Much of the current

understanding

of

dendrimer

conformations

and

their

associated

microenvironments comes from studies that employ dyes and indicators as structural probes of both the dendritic interior and surface.

Dyes and indicators also played

important roles in verifying some early applications of this class of macromolecules. In 54

the following chapter, indicator dyes are exploited in a sensing application.

This

application relies on some of the extraordinary features of dendrimers highlighted in this chapter but is distinguished from previous dendrimer-based sensing strategies by its modular nature.

1.10 REFERENCES 1

“’Cascade’ and ‘Nonskid-Chain-like’ Syntheses of Molecular Cavity Topologies” Buhleier, E.; Wehner, W.; Vogtle, F. Synthesis. 1978, 155-158.

2

“Cascade Molecules: A New Approach to Micelles. A [27]-Arborol” Newkome G.; Yao, Z.; Baker, G.; Gupta, V. J. Org. Chem. 1985, 50, 2004-06.

3

“A New Class of Polymers: Starburst-Dendritic Macromolecules” Tomalia, D.A.; Baker, H.; Dewald, J.; Hall, M.; Kallos, G.; Martin, S.; Roeck, J.; Ryder, J.; Smith, P. Polym. J. 1985, 17, 117-132.

4

“Cryptates: The Chemistry of Macropolycylic Inclusion Complexes” Lehn, J.M. Acc. Chem. Res. 1978, 11, 49-57.

5

Frechet, J.M.; Tomalia, D.A. Dendrimers and other Dendritic Polymers (Wiley 2001).

6

“Host-Guest Chemistry of Dendritic Molecules” Baars, M.; Meijer, E.W. Top. Curr. Chem. 2000, 210, 132-182.

7

“Dendrimers in Supramolecular Chemistry: From Molecular Recognition to Self Assembly” Zeng, F.; Zimmerman, S. Chem. Rev. 1997, 97, 1681-1712.

8

“Supramolecular Dendrimer Chemistry: A Journey Through the Branched Architecture” Smith, D.K.; Diederich, F. Top. Curr. Chem. 2000, 210, 183-227.

9

“Peptide Dendrimers as Artificial Enzymes, Receptors, and Drug-Delivery Agents” Darbre, T.; Reymond, J.-L. Reymond. Acc. Chem. Res. 2006, 39, 925-934.

10

“Functional Dendrimers: Unique Biological Mimics” Smith, D.K.; Diederich, F. Chem. Eur. J. 1998, 4, 1353-1361.

11

“The Synthesis and Properties of Novel Functional Dendritic Molecules” Chow, H-F.; Mong, T.K.; Nongrum, M.; Wan, C-W. Tetrahedron 1998, 54, 8543-8660 55

12

“New Organic Materials Suitable for Use in Chemical Sensor Arrays” Crooks, R.; Ricco, A. Acc. Chem. Res. 1998, 31, 219-227.

13

“Dendrimers: A Review of Their Appeal and Applications” Dykes, G.M. J. Chem. Technol. Biotechnol. 2001, 76, 903-918.

14

“Light-harvesting Dendrimers” Adronov, A.; Frechet, J.M. Chem. Commun. 2000, 1701-1710.

15

“Approaches for Biological and Biomimetic Energy Conversion” Lavan, D.A.; Cha, J. Proc. Nat. Acad. Sci. U.S.A. 2006, 103, 5251-5255.

16

“Poly(amidoamine) (PAMAM) dendrimers: from biomimicry to drug delivery and biomedical applications” Esfand, R.; Tomalia, D.A. Drug Discovery Today 2001, 6, 427-436.

17

“Dendrimers in Drug Research” Boas, U.; Heegaard, P. Chem. Soc. Rev. 2004, 33, 4363.

18

“Designing Dendrimers for Drug Delivery” Liu, M.; Frechet, J.M. Pharm. Sci. Technol. Today 1999, 2, 393-401.

19

“Dendrimers in Biomedical Applications – Reflections on the Field” Svenson, S.; Tomalia, D.A. Adv. Drug Delivery Rev. 2005, 57, 2106-2129.

20

“Cryptates: Inclusion Complexes of Macropolycyclic Receptor Molecules” Lehn, J.M. Pure & Appl. Chem. 1978, 50, 871-892.

21

“Biomimetic Chemistry and Artificial Enzymes” Breslow, R. Acc. Chem. Res. 1995, 28, 146-153.

22

“Design of Complexes between Synthetic Hosts and Organic Guests” Cram, D.J.; Cram, J.M. Acc. Chem. Res. 1978, 11, 8-14.

23

Behr, J.P. (Ed.) The Lock and Key Principle: The State of the Art – 100 Years On (Wiley, Chichester, 1994)

24

Anslyn, E.V.; Dougherty, D.A. Modern Physical Organic Chemistry. University Science Books, Sausalito, 2004.

25

“Hydrogen Bonding and Molecular Recognition: Synthetic, Complexation, and Structural Studies on Barbiturate Binding to an Artficial Receptor” Chang, S.K.; Van Engen, D.; Fan, E.; Hamilton, A.D. J. Am. Chem. Soc. 1991, 113, 7640-7645. 56

26

“Formation of Artificial Receptors by Metal-Templated Self-Assembly” Linton, B.; Hamilton A.D. Chem. Rev. 1997, 97, 1669-1680.

27

“C3v Symmetric Receptors Show High Selectivity and High Affinity for Phosphate” Tobey, S.L.; Jones, B.D.; Anslyn, E.V. J. Am. Chem. Soc. 2003, 125, 4026-4027.

28

“Dominant Forces in Protein Folding” Dill, K.A. Biochemistry 1990 29, 7133-7155.

29

“Solvophobically Driven Folding of Nonbiological Oligomers” Nelson, J.C.; Saven, J.G.; Moore, J.S.; Wolynes, P.G. Science 1997, 277, 1793-1796.

30

“Cyclodextrin Inclusion Compounds in Research and Industry” Saenger, W. Angew. Chem. Int. Ed. 1980, 19, 344-362.

31

“Complexation of Neutral Molecules by Cyclophane Hosts” Diederich, F. Angew. Chem. Int. Ed. 1988, 27, 362-386.

32

“Supramolecular Chemistry – Scope and Perspectives. Molecules, Supermolecules, and Molecular Devices (Nobel Lecture)” Lehn, J.M. Angew. Chem. Int. Ed. 1988, 27, 89-112.

33

“Molecular Recognition in Homogeneous Transition Metal Catalysis: A Biomimetic Strategy for High Selectivity” Das, S.; Brudvig, G.W.; Crabtree, R.H. Chem. Commun. 2008, 413-424.

34

“Molecular Umbrella-Assisted Transport of an Oligonucleotide across ChloesterolRich Phospholipid Bilayers” Janout, V.; Jing, B.; Regen, S.L. J. Am. Chem. Soc. 2005, 127, 15862-15870.

35

“Molecular Logic and Computing” de Silva, A.P.; Uchiyama, S. Nat. Nanotechnol. 2007, 2, 399-410.

36

“Chemical Communication in Water Using Fluorescent Chemosensors” Czarnik, A.W. Acc. Chem. Res. 1994, 27, 302-308.

37

“Supramolecular Optical Chemosensors for Organic Analytes” Bell, T.W.; Hext, N.M. Chem. Soc. Rev. 2004, 33, 589-598.

38

“Anion induced binding electrochemical signal transduction in ferrocenyl benzolimidazolium podands” Bai, Y.; Zhang, B.; Duan, C.; Dang, D.; Meng, Q. New J. Chem. 2006, 30, 266-271.

57

39

“Highly Selective Colorimetric Naked-Eye Cu(II) Detection Using an Azobenzene Chemosensor” Gunnlaugsson, T.; Leonard, J.P.; Murray, N.S. Org. Lett. 2004, 6, 1557-1560.

40

“Chromogenic Anion Sensors” Suksai, C.; Tuntulani, T. Chem. Soc. Rev. 2003, 32, 192-202.

41

“Colorimetric ‘Naked Eye’ Sensing of Anions in Aqueous Solution” Gunnlaugsson, T.; Kruger, P.E.; Jensen, P.; Tierney, J.; Ali, H.D.P.; Hussey, G.M. J. Org. Chem. 2005, 70, 10875-10878.

42

“Molecular Color Sensors for Monosaccharides” Ward, C.J.; Patel, P.; James, T.D. Org. Lett. 2002, 4, 477-479.

43

“Colorimetric Chiral Recognition by a Molecular Sensor” Kubo, Y.; Maeda, S.; Tokita, S.; Kubo, M. Nature 1996, 382, 522-524

44

“Self-Assembling, Chromogenic Receptors for the Recognition of Dicarboxylic Acids” Goodman, M.S.; Hamilton, A.D.; Weiss, J. J. Am. Chem. Soc. 1995, 117, 84478455.

45

“Oxdized Bis(indolyl)methane: A Simple and Efficient Chromogenic-Sensing Molecule Based on the Proton Transfer Signaling Mode” He, X.; Hu, S.; Liu, K.; Guo, Y.; Xu, J.; Shao, S. Org. Lett. 2006, 8, 333-336.

46

“A Selective Colorimetric Chemosensor for Lanthanide Ions” Liu, Z.; Jiang, L.; Liang, Z.; Gao, Y. Tetrahedron 2006, 62, 3214-3220.

47

“Signaling Recognition Events with Fluorescent Sensors and Switches” de Silva, A.P.; Gunaratne, H.Q.N.; Gunnlaugsson, T.; Huxley, A.J.M.; McCoy, C.P.; Rademacher, J.T.; Rice, T.E. Chem. Rev. 1997, 97, 1515-1566.

48

“Photoinduced Electronic Energy Transfer in a Supramolecular Complex between a Coumarin Bichromophoric Molecule and Lead(II)” Valeur, B.; Bourson, J.; Poujet, J. J. Lumin. 1992, 52, 345-347.

49

“A New Fluorescent PET Chemosensor for Fluoride Ions” Kim, S.K.; Yoon, J. Chem. Commun. 2002, 770-771.

50

“Efficient Fluorescent ATP-sensing Based on Coordination Chemistry under Aqueous Neutral Conditions” Ojida, A.; Park, S.; Mito-oka, Y.; Hamachi, I. Tetrahedron Lett. 2002, 43, 6193-6195. 58

51

“Dual Responsive Chemosensors for Anions: The Combination of Fluorescent PET and Colorimetric Chemosensors in a Single Molecule” Gunnlaugsson, T.; Kruger, P.E.; Lee, T.C.; Parkesh, R.; Pfeffer, F.M.; Hussey, G.M. Tetrahedron Lett. 2003, 44, 6575-6578.

52

“Adhesive Force of a Single Gecko Foot-Hair” Autumn, K.; Liang, Y.; Hsieh, S.T.; Zesch, W.; Chan, W.P.; Kenny, T.; Fearing, R.; Full, R.J. Nature 2000, 405, 681685.

53

“Controlling the Growth of Polymer Trees: Concepts and Perspectives for Hyperbranched Polymers” Sunder, A.; Heinemann, J.; Frey, H. Chem Eur. J. 2000, 6, 2499-2506.

54

“Preparation of Polymers with Controlled Molecular Architecture – A New Convergent Approach to Dendritic Macromolecules” Hawker, C.J.; Frechet, J.M. J. Am. Chem. Soc. 1990, 112, 7638-7647.

55

“Solid Phase Peptide Synthesis. I. The Synthesis of a Tetrapeptide” Merrifield, R.B. J. Am. Chem. Soc. 1963, 85, 2149-2154.

56

“About Dendrimers: Structure, Physical Properties, and Applications” Bosman, A.W.; Janssen, H.M.; Meijer, E.W. Chem. Rev. 1999, 99, 1665-1688.

57

“Slab-gel and Capillary Electrophoretic Characterization of Polyamidoamine Dendrimers” Brothers, H.M; Piehler, L.T.; Tomalia, D.A. J. Chromatogr. A 1998, 814, 233-246.

58

“Dendritic Encapsulation of Function: Applying Nature’s Site Isolation Principle from Biomimetics to Materials Science” Hecht, S.; Frechet, J.M. Angew. Chem. Int. Ed. 2001,40, 74-91.

59

“Statistics of Starburst Polymers” de Gennes, P.G.; Hervet, H. J. Physique Lett. 1983, 44, 351-360.

60

“Configurational Characteristics and Scaling Behavior of Starburst Molecules: A Computational Study” Lescanec, R.L.; Muthukumar, M. Macromolecules 1990, 23, 2280-2288.

61

“A Self-Consistent Mean Field Model of a Starburst Dendrimer: Dense Core vs Dense Shell” Boris, D.; Rubinstein, M. Macromolecules 1996, 29, 7251-7260.

62

“Molecular Dynamics Study of Dendrimer Molecules in Solvents of Varying Quality” Murat, M.; Grest, G. Macromolecules 1996, 29, 1278-1285. 59

63

“Unique Behavior of Dendritic Macromolecules: Intrinsic Viscosity of Polyether Dendrimers” Mourey, T.H.; Turner, S.R.; Rubinstein, M.; Frechet, J.M.; Hawker, C.J.; Wooley, K.L. Macromolecules 1992, 25, 2401-2406.

64

“Shapes of Dendrimers from Rotational-Echo Double Resonance NMR” Wooley, K.L.; Klug, C.A.; Tasaki, K.; Schaefer, J. J. Am. Chem. Soc. 1997, 119, 53-58.

65

“Use of a Paramagnetic Core to Affect Longitudinal Nuclear Relaxation in Dendrimers – A Tool for Probing Dendrimer Conformation” Macromolecules 1998, 31, 815822.

66

“Fluorophobic Effect Induces the Self-Assembly of Semifluorinated Tapered Monodendrons Containing Crown Ethers into Supramolecular Columnar Dendrimers Which Exhibit a Homeotropic Hexagonal Columnar Liquid Crystalline Phase” Percec, V.; Johansson, G.; Ungar, G.; Zhou, J. J. Am. Chem. Soc. 1996, 118, 9855-9866.

67

“Covalently Fixed Unimolecular Assemblages Reminiscent of Spheroidal Micelles” Tomalia, D.A.; Berry, V.; Hall, M.; Hedstrand, D.M. Macromolecules 1987, 20, 1167-1169

68

“Carboxylated Starburst Dendrimers as Calibration Standards for Aqueous Size Exclusion Chromatography” Dubin, P.L.; Edwards, S.L.; Kaplan, J.I.; Mehta, M.S.; Tomalia, D.A.; Xia, J. Anal. Chem. 1992, 64, 2344-2347

69

“SANS Study of Labeled PAMAM Dendrimer” Amis, E.J.; Topp, A.; Bauer, B.J. Tomalia, D.A. Polym. Mater. Sci. Eng. 1997, 77, 183-184

70

“Structural Molecular Dynamics Studies on Polyamidoamine Dendrimers for a Therapeutic Application: Effects of pH and Generation” Lee, I.; Athey, B.D.; Wetzel, A.W.; Meixner, W.; Baker, J.R. Macromolecules 2002, 35, 4510-4520.

71

“Fluorescence Emission from Dendrimers and Its pH Dependence” Wang, D.; Imae, T. J. Am. Chem. Soc. 2004, 126, 13204-13205.

72

“Determination of the Intrinsic Proton Binding Constants for Poly(amidoamine) Dendrimers via Potentiometric pH Titration” Niu, Y.; Sun, L.; Crooks, R.M. Macromolecules 2003, 36, 5725-5731.

73

“The Role of Hydrophobic Microenvironments in Modulating pKa Shifts in Proteins” Mehler, E.L.; Fuxreiter, M.; Simon, I.; Garcia-Moreno, B. Proteins 2002, 48, 283292. 60

74

“Structure and Conformation of DAB Dendrimers in Solution via Multidimensional NMR Techniques” Chai, M.; Niu, Y.; Youngs, W.; Rinaldi, P.L. J. Am. Chem. Soc. 2001, 123, 4670-4678.

75

“Solvent Dependence of the Hydrodynamical Volume of Dendrimers with a Rubicene Core” De Backer, S.; Prinzie, Y.; Verheijen, W.; Smet, M.; Desmedt, K.; Dehaen, W.; De Schryver, F.C. J. Phys. Chem. A 1998, 102, 5451-5455.

76

“Effects of Dendrimer Generation on Site Isolation of Core Moieties: Electrochemical and Fluorescence Studies with Metalloporphyrin Core Dendrimers” Pollack, K.W.; Leon, J.W.; Frechet, J.M.; Maskus, M.; Abruna, H.D. Chem. Mater. 1998, 10, 30-38.

77

Suppan, P.; Ghoneim, N. Solvatochromism (Springer-Verlag, New York, 1997)

78

“Solvatochromic Dyes as Solvent Polarity Indicators” Reichardt, C. Chem. Rev. 1994, 2319-2358.

79

“Solvatochromism as a Probe of the Microenvironment in Dendritic Polyethers: Transition from an Extended to a Globular Structure” Hawker, C.J.; Wooley, K.; Frechet, J.M. J. Am. Chem. Soc. 1993, 115, 4375-4376.

80

“Anomalous Shift in the Fluorescence Spectra of a High-Generation Dendrimer in Nonpolar Solvents” Devadoss, C.; Bharathi, P.; Moore, J.S. Angew. Chem. Int. Ed. 1997, 36, 1633-1635.

81

“Supramolecular Dendritic Solubilisation of a Hydrophilic Dye and Tuning of its Optical Properties” Smith, D.K. Chem. Commun. 1999, 1685-1686.

82

“Study of Poly(amidoamine) Starburst Dendrimers by Fluorescence Probing” Pistolis, G.; Malliaris, A.; Paleos, C.M.; Tsiourvas, D. Langmuir 1997, 13, 5870-5875.

83

“Spectroscopic Investigations of Polyamido Amine Starburst Dendrimers with Reichardt’s ET-30 Dye” Egger-Richter, D.L.; Hong, L.; Tucker, S.A. Appl. Spectrosc. 2000, 54, 1151-1156.

84

“Spectroscopic Investigations of Polyamido Amine Starburst Dendrimers Using the Solvatochromic Probe Phenol Blue” Egger-Richter, D.L.; Landry, J.C.; Tesfai, A.; Tucker, S.A. J. Phys. Chem. A 2001, 105, 6826-6833.

85

“Aggregation of Methylene Blue Adsorbed on Starburst Dendrimers” Jockusch, S.; Turro, N.; Tomalia, D.A. Macromolecules 1995, 28, 7416-7418. 61

86

“Comparison of Nitrogen Core and Ethylenediamine Core Starburst Dendrimers Through Photochemical and Spectroscopic Probes” Jockusch, S.; Ramirez, J.; Sanghvi, K.; Nociti, R.; Turro, N.; Tomalia, D.A. Macromolecules 1999, 32, 4419-4423.

87

Voet, D.; Voet, J.G. Biochemistry, 3rd Ed. (Wiley, International 1994) 390-391.

88

“Polymeric Micelles for Drug Delivery” Croy, S.R.; Kwon, G.S. Curr. Pharm. Des. 2006, 12, 4669-4684.

89

“Micellar Solubilization of Drugs” Rangel-Yagui, C.O.; Pessoa, A.; Tavares, L.C. J. Pharm. Pharmaceut. Sci. 2005, 8, 147-163.

90

“Unimolecular Micelles and Globular Amphiphiles: Dendritic Macromolecules as Novel Recyclable Solubilization Agents” Hawker, C.J.; Wooley, K.; Frechet, J.M. J. Chem Soc. Perkin Trans. 1993, 1287-1297.

91

“Synthesis, Characterization, and Guest-Host Properties of Inverted Unimolecular Dendritic Micelles” Stevelmans, S.; van Hest, J.C.M.; Jansen, J.F.G.A.; van Boxtel, D.A.F.J.; de Brabander-van den Berg, E.M.M.; Meijer, E.W. J. Am. Chem. Soc. 1996, 118, 7398-7399.

92

“Liquid-liquid Extractions Using Poly(propylene imine) Dendrimers with an Apolar Periphery” Baars, M.W.P.L.; Froehling, P.E.; Meijer, E.W. Chem. Commun. 1997, 1959-1960.

93

“Extraction of a Hydrophilic Compound from Water into Liquid CO2 Using Dendritic Surfactants” Cooper, A.I.; Londono, J.D.; Wignall, G.; McClain, J.B.; Samulski, E.T.; Lin, J.S.; Dobrynin, A.; Rubinstein, M.; Burke, A.L.C.; Frechet, J.M.; DeSimone, J.M. Nature 1997, 389, 368-371.

94

“Blends of Amphiphilic, Hyperbranched Polyesters and Different Polyolefins” Schmaljohann, D.; Potschke, P.; Hassler, R.; Voit, B.I.; Froehling, P.E.; Mostert, B.; Loontjens, J.A. Macromolecules 1999, 32, 6333-6339.

95

“Encapsulation of Guest Molecules into a Dendritic Box” Jansen, J.F.G.A.; de Brabander-van den Berg, E.M.M.; Meijer, E.W. Science 1994, 266, 1226-1229.

96

“The Dendritic Box: Shape-Selective Liberation of Encapsulated Guests” Jansen, J.F.G.A.; Meijer, E.W.; de Brabander-van den Berg, E.M.M. J. Am. Chem. Soc. 1995, 117, 4417-4418.

62

97

“Dendrimer-Silver Complexes and Nanocomposites as Antimicrobial Agents” Balogh, L.; Swanson, D.R.; Tomalia, D.A.; Hagnauer, G.L.; McManus, A.T. Nano Lett. 2001, 1, 18-21.

98

“Encapsulating Fluorescein Using Adipic Acid Self-Assembly on the Surface of PPI-3 Dendrimer” Chai, M.; Holley, A.K.; Kruskamp, M. Chem. Commun. 2007, 168170.

99

“Concepts of Trapping Topologically by Shell Molecules” Maciejewski, M. J. Macromol. Sci. Chem. 1982, 17, 689-703.

100

“Supramolecular Chemistry of Dendrimers” Zimmerman, S.C.; Lawless, L.J. Top. Curr. Chem. 2001, 217, 95-120.

101

“Dendrophanes: Water-Soluble Dendritic Receptors as Models for Buried Recognition Sites in Globular Proteins” Mattei, S.; Wallimann, P.; Kenda, B.; Amrein, W.; Diederich, F. Helv. Chim. Acta 1997, 80, 2391-2417.

102

“Cyclophanes as Synthetic Receptors” Sutherland, I.O. Pure Appl. Chem. 1990, 62, 499-504.

103

“Catalytic Dendrophanes as Enzyme Mimics: Synthesis, Binding Properties, Micropolarity Effect, and Catalytic Activity of Dendritic Thiazolio-cyclophanes” Habicher, T.; Diederich, F.; Gramlich, V. Helv. Chim. Acta 1999, 82, 1066-1095.

104

“Dendritic Hydrogen Bonding Receptors: Enantiomerically Pure Dendroclefts for the Selective Recognition of Monosaccharides” Smith, D.K.; Diederich, F. Chem. Commun. 1998, 2501-2502.

105

“’Crowned’ Arborols” Nagasaki, T.; Ukon, M.; Arimori, S.; Shinkai, S. J. Chem. Soc., Chem. Commun. 1992, 608-610.

106

“Synthesis, Metal-binding Properties and Polypeptide Solubilization of ‘Crowned’ Arborols” Nagasaki, T.; Kimura, O.; Ukon, M.; Arimori, S.; Hamachi, I.; Shinkai, S. J. Chem. Soc. Perkin Trans. 1 1994, 75-81.

107

“Metalloporphyrin Dendrimers with Folding Arms” Mak, C.C.; Bampos, N.; Sanders, J.K.M. Angew. Chem. Int. Ed. 1998, 37, 3020-3023.

108

“Polyvalent Interactions in Biological Systems: Implications for Design and Use of Multivalent Ligands and Inhibitors” Mammen, M.; Choi, S.; Whitesides, G.M. Angew. Chem. Int. Ed. 1998, 37, 2754-2794. 63

109

“The Dendritic Effect in Molecular Recognition: Ferrocene Dendrimers and Their Use as Supramolecular Redox Sensors for the Recognition of Small Inorganic Anions” Valerio, C.; Fillaut, J-L.; Ruiz, J.; Guittard, J.; Blais, J-C.; Astruc, D. J. Am. Chem. Soc. 1997, 119, 2588-2589.

110

“Metallocenyl Dendrimers and Their Applications in Molecular Electronics, Sensing, and Catalysis” Astruc, D.; Ornelas, C.; Ruiz, J. Acc. Chem. Res. 2008, 41, 841856.

111

“Synthesis and Properties of New Thiourea-Functionalized Poly(propylene imine) Dendrimers and Their Role as Hosts for Urea Functionalized Guests” Boas, U.; Karlsson, A.J.; de Waal, B.F.M.; Meijer, E.W. J. Org. Chem. 2001, 66, 21362145.

112

“New Dendrimer-Peptide Host-Guest Complexes: Towards Dendrimers as Peptide Carriers” Boas, U.; Sontjens, S.H.M.; Jensen, K.J.; Christensen, J.B.; Meijer, E.W. ChemBioChem 2002, 3, 433-439.

113

“Small Molecular Gadolinium (III) Complexes as MRI Contrast Agents for Diagnostic Imaging” Chan, K.W.; Wong, W. Coord. Chem. Rev. 2007, 251, 2428-2451.

114

“Characterization of Starburst Dendrimers by the EPR Technique. Copper(II) Ions Binding Full Generation Dendrimers” Ottaviani, M.F.; Montalti, F.; Turro, N.J.; Tomalia, D.A. J. Phys. Chem. B 1997, 101, 158-166.

115

“Poly(amidoamine) Dendrimers: A New Class of High Capacity Chelating Agents for Cu(II) Ions” Diallo, M.S.; Balogh, L.; Shafagati, A.; Johnson, J.H.; Goddard, W.A.; Tomalia, D.A. Enviro. Sci. Technol. 1999, 33, 820-824.

116

“Preparation of Cu Nanoclusters Within Dendrimer Templates” Mingqi, Z.; Sun, L.; Crooks, R.M. J. Am. Chem. Soc. 1998, 120, 4877-4878.

117

“Polyphenylene Dendrimers as Sensitive and Selective Sensor Layers” Schlupp, M.; Weil, T.; Berresheim, A.J.; Wiesler, U.M.; Bargon, J.; Mullen, K. Angew. Chem. Int. Ed. 2001, 40, 4011-4015.

118

“Application of Dendrimer as a New Material for Electrical and Optical Devices” Song, C.; Koo, B.; Kim, C. Jpn. J. Appl. Phys. 2002, 41, 2735-2738.

119

“Electrocatalytical Properties of Polymethylferrocenyl Dendrimers and Their Applications in Biosensing” Armada, M.P.G.; Losada, J.; Zamora, M.; Alonso, B.; Cuadrado, I.; Casado, C.M. Bioelectrochemistry 2006, 69, 65-73. 64

120

“A Saccharide ‘Sponge.’ Synthesis and Properties of a Dendritic Boronic Acid” James, T.D.; Shinmori, H.; Takeuchi, M.; Shinkai, S. Chem. Commun. 1996, 705706.

121

“Fluorescent Sensors for Transition Metals Based on Electron-Transfer and EnergyTransfer Mechanisms” Fabbrizzi, L.; Licchelli, M.; Pallavicini, P.; Perotti, A.; Taglietti, A.; Sacchi, D. Chem. Eur. J. 1996, 2, 75-82.

122

“Dendrimers as Fluorescent Sensors with Signal Amplification” Balzani, V.; Ceroni, P.; Gestermann, S.; Kauffmann, C.; Gorka, M.; Vogtle, F. Chem. Commun. 2000, 853-854.

123

“The First Dendrimer-Based Enantioselective Fluorescent Sensor for the Recognition of Chiral Amino Alcohols” Pugh, V.J.; Hu, Q.; Pu, L. Angew. Chem. Int. Ed. 2000, 39, 3638-3541.

124

“Optically Active BINOL Core-Based Phenyleneethynylene Dendrimers for the Enantioselective Fluorescent Recognition of Amino Alcohols” Pugh, V.J.; Hu, Q.; Zuo, X.; Lewis, F.D.; Pu, L. J. Org. Chem. 2001, 66, 6136-6140.

125

“Detection of TNT Using a Sensitive Two-Photon Organic Dendrimer for Remote Sensing” Narayanan, A.; Varnavski, O.; Mongin, O.; Majoral, J.; BlanchardDesce, M.; Goodson, T. Nanotechnology 2008, 19, 1-6.

126

“Real-Time Detection of Bacterial Contamination in Dynamic Aqueous Environments Using Optical Sensors” Ji, J.; Schanzle, A.; Tabacco, M.B. Anal. Chem. 2004, 76, 1411-1418.

127

“Competitive Binding Assay for Glucose Based on Glycodendrimer-Fluorophore Conjugates” Ibey, B.L.; Beier, H.T.; Rounds, R.M.; Cote, G.L. Anal. Chem. 2005, 77, 7039-7046.

128

“Oxygen Distribution in Murine Tumors: Characterization Using Oxygen-Dependent Quenching of Phosphorescence” Ziemer, L.S.; Lee, W.M.F.; Vinogradov, S.A.; Sehgal, C.; Wilson, D.F. J. Appl. Physiol. 2005, 98, 1503-1510.

65

Chapter 2: Progress Toward Templated Assembly of Dendritic Receptors 2.1 INTRODUCTION AND SCOPE Much of the research described in this chapter is directed toward novel uses of dendrimers in molecular recognition and sensing applications. Efforts directed toward the dynamic, templated assembly of dendritic receptors are summarized. The novel approach outlined in this chapter is distinct from those examples of rationally designed dendritic receptors/sensors described in Chapter 1 (see sections 1.7 and 1.8).

Because

this lofty goal serves as the inspiration for some preliminary studies described herein, one “big-picture” incarnation is described in detail to grant the reader some perspective on the roles of the experiments conducted in the advancement of these objectives; and because the selected embodiment relies on aspects of both dynamic combinatorial chemistry (DCC) and molecularly imprinted polymers (MIPs), a brief description of these chemistries, along with relevant examples of their confluence, is included. While much of the introductory matter pertains to the realization of the “bigpicture” goals, the selected original research represents several supporting studies that extend the scope of the investigated systems in the realm of molecular sensing (see section 1.2.2). To this end, a discussion of indicator displacement assays (IDAs) and their use in the highlighted studies is also included. The new direction of this project into molecular sensing culminates with the description and development of a fluorescencebased signaling protocol that relies on a sensing ensemble of dendrimers, indicators, and peptides.

66

2.2 BACKGROUND AND SIGNIFICANCE An original embodiment of these goals envisions dendrimers as scaffolds in the non-covalent assembly of templated molecular receptors. The “molding” of a dendrimerbased receptor around a guest “template” might provide synthetic macromolecular receptors that rival the specificity and function of biological constructs such as antibodies and enzymes. Linus Pauling recognized the elegant economy of such a process in his ruminations on the formation of antibodies.1 Pauling hypothesized that antigens induce a binding site within the polypeptide of an antibody and suggested that the “antigen molecule, after its desertion by the newly-formed antibody molecule, may serve as the pattern for another.” While this hypothesis for the formation of antibodies has since been rejected, Pauling’s compelling suggestion has found value beyond its original intent. The guest-induced formation of binding sites within a molecular receptor is now an active area of research, with the pursuit of “tailor-made” synthetic hosts for specific targets as the goal. This templated approach to the construction of molecular receptors is an instructed, target-driven process that complements the traditional rational design of synthetic hosts (see Section 1.2.1) Two strategies have emerged to generate artificial molecular receptors in this manner. Molecularly imprinted polymers2 (MIPs) and dynamic combinatorial libraries (DCLs)3,4 both allow facile entry to receptors “molded” to specific guests. Molecularly imprinted polymers are formed when functional and cross-linking monomers are polymerized in the presence of the target analyte. Prior to polymerization, the functional monomers complex the analyte. This complexation can occur via covalent or noncovalent interactions.

Subsequent polymerization leads to a highly cross-linked

macromolecule that casts the binding sites in an orientation that is complementary to the size and shape of the templated guest. Removal of the template molecule exposes these 67

recognition sites and creates a polymer with a binding cavity tailored to the imprinted molecule (Scheme 2.1).

This imprinting process constitutes an induced “molecular

4 Chemical Reviews, 2002, Vol. 102, No. 1

memory” that allows the tight and selective re-binding of the molecular template (or Scheme 4. Schematic Representation of the reader is referred to the Imprinting of Specific Cavities in a Cross-Linked structural analogs). Recent advances in MIP technology have lead researchers to paragraph some In this Polymer by a Template (T) with Three Different are especially importa 2a,5,6 ref 38) from speculate on theirBinding status as Groups “synthetic(adapted antibodies.” printed polymers as ca Polymerizable functio by covalent or noncova template molecule (se monomer is then copoly in the presence of a la agent and a certain amo as a porogen). This proc polymers with a perman inner surface area. U template molecules can ages. Optically active t most cases during opti accuracy of the structu with binding sites) cou for racemic resolution, batch procedure or by u as chromatographic su treated in detail by Davis.39,40 One of his special Scheme 2.1 Molecular imprinting with polymers. Functional monomers that can of the early exam One interests concerned the preparation of catalysts by interact with template molecule T are polymerized in its presence. The resulting in Scheme 5. In this cas imprinting in silica and zeolites. macromolecule contains of complementary and size to the template. side 2 acts as the tem Most binding papers sites on catalytically activeshape imprinted polySubsequent removal the template these functionalities, a 4-vinylbenzeneboronic mers of appeared during unmasks the last 10 years. As there are providing macromolecular very receptor capablethis of period, rebinding thetoo analyte high affinity and linkages to this diester few before they will bewith included polymerized with a larg specificity (imageinadapted from Chem. Rev. 2002, 102, 1-27). this discussion. acrylate as cross-linker II. Some Basic Considerations on Molecular split off by the addition Despite their Imprrelative inting insuccess, PolymMIPs eric Msuffer ateriafrom ls forseveral the limitations including extent of up to 95%, the batch procedure with a P r e p a r a t i on o f C a t a l ys t s incomplete template removal, the lack of an intrinsic signaling mode, and binding the site template. The enric A. An Example of Covalent Imprinting heterogeneity. This heterogeneous nature of MIP binding sites is in contrast topolymer those and in solut separation factor R, i.e. Numerous reviews on the molecular imprinting coefficients of D and L procedure have been published the last yearsthousands observed in many biological receptors. While during MIPs may contain of 57,58 and solution, is calculat The (see, e.g., refs 41-56) as well as two books. recognition sites per molecule, biological macromolecules often just feature one binding

Scheme 5. Schematic Representation of a Cavity Obtained by Polymerization o

and site. Furthermore, the59) population of MIP binding sites is often heterogeneous due to the

multiple equilibria that govern the formation of the template-monomer complexes.6 This 68

binding site heterogeneity leads to a broad range of affinities that sometimes limits the practical application of MIPs. MIPs also lack a mechanism for dynamic error-correction. Because the most common methods of their production rely on free radical polymerization, an irreversible process, there is no possibility of equilibration of lowaffinity binding sites to high affinity ones via reversible covalent bond formation. While such reversibility would allow the energetic sampling of both templated and nontemplated binding sites, there are no known reports of dynamically synthesized MIPs in the literature. A second approach to the rapid creation of “molded” molecular receptors employs dynamic combinatorial libraries (DCLs) of potential host molecules. DCLs rely on the rapid interconversion of library members via reversible chemical processes. In effect, this approach is simply traditional combinatorial chemistry under thermodynamic control, and the result is a library whose composition is governed by the stability of each potential member and not simply populated by discrete, static, prefabricated molecules. Lehn further distinguishes dynamic combinatorial chemistry (DCC) from traditional combinatorial chemistry by suggesting DCCs offer access to “virtual combinatorial libraries” whose members at any particular moment are just “the real subset of all those that are potentially accessible.”4 Because DCLs operate under thermodynamic control, changes in experimental conditions are expected to alter the library’s composition. As a result, DCLs are responsive and adaptive toward external influences. For example, introduction of a guest molecule to a dynamic library of potential hosts perturbs the equilibria processes of the interconverting members. Binding of the guest to a particular library member will stabilize this constituent and effect its removal from the pool of interconverting members; the equilibrium will then shift towards its formation, thereby increasing its concentration through a template-induced amplification process. 69

Conversely, the equilibrium will shift away from the formation of library members that are not efficient binders to the added guest.

This process illustrates an important

distinction between traditional combinatorial libraries and the “virtual” libraries offered by DCC—while members of traditional libraries are synthesized in the absence of the target, the amplification of a selected DCL member occurs in the presence of the target! In fact, the desired constituent(s) of a DCL may not even exist in significant amounts in the absence of target, but rather are generated through an instructed, target-driven process. Lehn suggests the responsive, adaptive nature of DCLs amount to “molecular Darwinism”4 in which the compound best suited for binding the guest evolves through a “survival of the fittest” process (Scheme 2.2).

Scheme 2.2 Selection of a molecular host using a dynamic combinatorial library. The members of a DCL reversibly interact to give a dynamic mixture of potential selfassembled hosts. Addition of a guest template sequesters the most effective host from this mixture of interconverting components, and this molecular assembly is selectively amplified and expressed. (image adapted from Drug Discovery Today 2002, 7, 117-125). DCLs and MIPs are conceptually similar as both rely on molecular templates in the rapid construction of molecular receptors. Whereas MIP template molecules imprint their structure into polymeric materials, compounds added to DCLs imprint their structure within an equilibrating, “virtual” library of potential host molecules. There are some important distinctions between these methodologies however. As discussed earlier, 70

MIPs, unlike receptors generated from DCLs, are constructed using irreversible covalent bond formation. This irreversibility obviates the potential for thermodynamic sampling of incipient receptors, error-correction and rejection of poorly suited binders, and molecular “evolution” of an optimal binding entity. On the other hand, the dynamic reversibility that predicates the creation of DCLs offers a thorough energetic survey of real and “virtual” library members in the search for the supreme binder.

This

reversibility imbues DCLs with another distinct advantage over MIPs; unlike MIPs, DCLs offer entry to templated molecular receptors containing only a single binding site. MIPs may contain thousands of binding sites of varying identity and accessibility. This heterogeneous population of binding sites complicates evaluation of both structure and binding. The recent embracing of an imprinted polymeric microgel containing 10-100 binding cavities per particle7 as a significant advance toward the generation of a single binding site within a protein-sized macromolecule illustrates the dearth of such synthetic molecular constructs. Shinkai and co-workers reported a seminal example of a macromolecule containing a single, imprinted binding site.8 Appreciating the general insolubility and heterogeneous distribution of MIPs, Shinkai sought to create a macromolecular receptor with a single, imprinted binding site that is amenable to solution-phase studies. The researchers selected [60]fullerene as a nano-sized scaffold because of its moderate solubility in organic solvents and its presentation of a reactive surface functionality. A regioselective double cycloaddition of a preformed 2:1 boronic acid:saccharide complex to the surface of the fullerene scaffold affords an adduct that positions the boronic acid recognition units in an orientation complementary to the size and shape of the “templated” saccharide. Subsequent removal of the template is effected by treatment of the adduct with aqueous acid (Scheme 2.3). Solid-liquid extraction experiments proved 71

that the fullerene-based receptor is capable of re-binding templated saccharides. This work illustrates the potential for structural identification and evaluation of individually imprinted binding sites within a macromolecular receptor.

Scheme 2.3 A homogeneous nanoscale imprinting system. The double cycloaddition between a boronic acid:saccharide complex and a fullerene scaffold creates a cleft that complements the templated saccharide. Removal of the template allows a selective rebinding of the saccharide guest. This system is the first example of an imprinted macromolecular receptor that operates in homogeneous solution (image adapted from reference 8b). Zimmerman and Suslick married aspects of DCL and MIP technologies to dynamically imprint a single binding site within a macromolecule. Their investigations employed dendrimers as scaffolds in the construction of dynamically templated macromolecular receptors.9 Dendrimers are attractive macromolecular scaffolds for such an application because their monodispersity (see Section 1.3) facilitates receptor characterization, and their presentation of a large number of surface groups (see Section 1.3.1) permits extensive cross-linking.

When this cross-linking is performed in a

reversible fashion in the presence of a template guest, a dynamic imprinting of the macromolecule is possible (Scheme 2.4). Grubbs has previously employed ring closing metathesis (RCM) reactions to produce macrocyclic structures.10 Alkene cross-links 72

prepared in this fashion are robust yet reversible. Furthermore, RCM’s potential for dynamic error-correction is illustrated in the ruthenium-mediated cyclization of 1,2polydienes where undesirable ring closings were rejected and cleaved until adjacent alkenes were coupled exclusively and quantitatively.11 This precedent led Zimmerman and Suslick to consider functional dendrons containing end-groups that could be reversibly cross-linked by established RCM protocols.

Scheme 2.4 Molecularly imprinted dendrimers. Functional dendrons are covalently attached to a template molecule. A reversible cross-linking process allows dynamic imprinting of the guest within the structure of the macromolecule. Hydrolytic cleavage effects template removal and allows for selective re-binding of the templated guest (image adapted from reference 9b). A Frechet-type dendrimer containing an ester-linked porphyrin core12 and decorated with 64 allyl surface groups was subjected to RCM to provide a highly crosslinked dendrimer surface.

Quantitative removal of the porphyrin template was

accomplished by basic hydrolysis, and the molecularly imprinted dendrimer (MID) was 73

isolated in 43% yield. The binding of 11 different porphyrin guests to this MID was investigated. It is important to note that although the template was imprinted using covalent ester linkages to the dendrimer, re-binding of guests occurs via non-covalent interactions. A corollary of such covalent imprinting processes is that the imprinted template is not necessarily an ideal guest upon re-binding. Guest binding was assessed by size exclusion chromatography and also monitored by UV-vis spectroscopy. The associated titration data suggested the formation of 1:1 complexes. Structural inspection of the investigated porphyrins revealed that successful binders featured at least four hydrogen-bonding contacts. Size selectivity was also observed, but the imprint was not highly shape selective. Importantly, there was no evidence of binding site heterogeneity within the MID.

One equivalent of a porphyrin guest almost fully complexed an

equimolar amount of the MID receptor, indicating that nearly all of the binding sites were effectively imprinted. This approach therefore represents a unique strategy that advances the art of imprinting macromolecules. For the first time, a single, discrete binding site was imprinted within a large molecular scaffold (Mw ~ 10kDa). Further development of dynamically imprinted dendritic receptors might soon provide synthetic macromolecules that rival the remarkable functionalities of their natural counterparts. Ironically, the realization of such a goal would reify Pauling’s original vision in an unexpected way. In a serendipitous conflation of this methodology with the colorimetric sensing strategy described in Chapter 3, Zimmerman incorporated a trifluoroacetyl-based “chromoreactand”13 (see section 3.2.3.1) into the dynamic imprinting of dendritic receptors.14 Trifluoroacetyl-functionalized azo dyes are known colorimetric reporters for the presence of amines via reversible covalent bond formation. The amine-reactive, trifluoroacetyl-based dye was covalently attached to dendritic segments via a bis-imine linkage. The allyl surface groups of the dendrons were then cross-linked through RCM 74

protocols. Subsequent acidification of the imprinted dendrimer effected removal of the diamine template by imine hydrolysis, thereby re-forming the reactive trifluoroacetyl moieties for the re-binding of amine guests. Qualitative color changes upon interaction with amine guests indicated selectivity for the templated diamine over monofunctional analytes such as butylamine. However, an ensuing report15 offered evidence that this observed selectivity arises not form the two-point covalent imprinting process but rather originates from a kinetic binding effect. Regardless of the origin of its selectivity, this system represents a conceptual unification of work described in both the present and following chapters.

Zimmerman’s fusion of these strategies, however, is directed

towards a goal that is incidental to the research described in this dissertation. 2.3 RESEARCH DESIGN An original incarnation of this project envisions the use of dendrimers as scaffolds in the dynamic assembly of templated receptors. This goal is somewhat analogous to the successful imprinting of dendrimers reported by Zimmerman because it also invokes aspects of dynamic combinatorial chemistry in the creation of an imprinted binding site within a macromolecule.

Our proposed method, however, might offer specific

advantages over Zimmerman’s methodology in some applications.

Zimmerman’s

imprinting process, while dynamic, occurred on the dendritic surface through reversible alkene metathesis. While a single imprinted site capable of selective guest recognition was created, the template molecule was covalently fixed to the core distal to the responsive imprinting functionalities. Recall that a hallmark of Zimmerman’s strategy is its appropriation of aspects of both DCL and MIP technologies to dynamically imprint a template molecule within the core of a dendrimer. The potential for receptors generated using Zimmerman’s approach is likely mitigated by the structural distance between the template molecule and imprinting functionality. Although Zimmerman’s own reported 75

preparation of dendrimers with interior cross-links16 might alleviate this concern, the preparation of the requisite cored dendrimers demands multiple synthetic steps. The method proposed in this section is geared toward the rapid generation of dendrimer-based receptors via a dynamic, non-covalent imprinting process. In its ideal implementation, the proposed method would employ commercial dendrimers and easily synthesized components and permit intimate contact between the template molecule and imprinting functionality.

The general strategy is represented in Scheme 2.5.

A

dendrimer reversibly interacts with members of a library containing variable recognition units. The variable recognition units non-covalently bind to the dendrimer through the action of a complementary binding subunit “tag” covalently appended to each member of the library.

The non-covalent nature of the interaction between tag and dendrimer

ensures a dynamic process that allows each library member to bind reversibly to the dendrimer.

This process creates a situation analogous to Lehn’s so-called “virtual

library” in which the instantaneous composition of the mixture is merely a subset of all potential combinations.

Under these dynamic conditions, the concentration of each

potential member is governed by its relative stability.

This thermodynamic control

renders the distribution of constituents susceptible to external influences. Upon introduction of a guest, the assemblies that are the most effective binders will be selectively stabilized by the resulting molecular recognition event, and the equilibria will shift to the formation of this stabilized species, thereby amplifying its production. In other words, the variable recognition units best suited for binding the guest will be sequestered by the dendrimer in the creation of an optimal supramolecular entity. Recognition units that do not strongly interact with the added guest will remain in the equilibrating pool and will not be expressed in the amplified assembly. Subsequent covalent bond formation between the tag and the dendrimer will lock the sequestered 76

components in an orientation that is complementary to the size and shape of the templated guest.

Removal of the template allows re-binding of the guest by a

dynamically imprinted macromolecular receptor with a discrete binding site.

Scheme 2.5 A strategy for templated assembly of dendritic receptors. Addition of a guest (

) to a mixture of variable recognition units (

) reversibly interacting

with a dendritic scaffold ( ) templates the formation of an ideal receptor that can be covalently “fixed” by reaction between the dendritic surface and recognition unit. Polyamidoamine (PAMAM) dendrimers were selected as dendritic scaffolds due to their commercial availability and reported success in diverse applications17 (an aminoterminated G2 PAMAM dendrimer is represented by structure 2.0.)

The variable

recognition units are envisioned as oligopeptides. Previous reports from the Anslyn group demonstrate the efficacy of variable peptide arms as differential recognition units in the analysis of tripeptide18 and protein19 mixtures. Established protocols also allow facile entry to combinatorial peptide libraries from split-and-pool methods20 or designed 77

peptidic arms21 from automated synthesis.

Identification and conjugation of an

appropriate subunit (the tag) that effects reversible binding of the peptides to the dendrimer embodies the initial challenge of this project. NH2 O H2N

H2N

O

O

N

O

H2N

H2N

NH2

O NH2

NH2

O

N

O

O

HN

N HN

O

N

NH

O

O

O N

N

HN

O

O

N H

O

N

NH O

N H

N O

NH

HN

O

O HN

NH

O

N

N

N H

N

NH2

O HN

O HN

HN

O

N

O H2N

H2N

O O O HN

N O

NH2 N

O NH2

O

NH2

O

O

NH2

N H 2

2.0 Before discussing the development and evaluation of an appropriate binding subunit, the reader should note that the scenario envisioned in Scheme 2.5 represents merely one reification of the potential contributions offered by investigations into such a system. In many scientific endeavors, the avenues leading to the ultimate goal are fraught with intersections and detours; and an original ideal, much like Pauling’s hypotheses on antibody formation, can inspire additional routes wholly unanticipated upon its initial conception. In this spirit, the progress described in the remainder of this chapter represents several ancillary studies that are distinct from (yet crucial to) the realization of the previously stated goals. A discussion of the original research design

78

was included because some of the following studies were conceived and executed with this final application in mind. 2.4 IDENTIFICATION OF A BINDING SUBUNIT The evaluation of potential binding subunits was conducted through indicator displacement assays (IDAs).22 IDAs exploit the multiple equilibria established when an analyte competes with an indicator for binding to the host (Scheme 2.6). Displacement of the indicator by the analyte leads to a change in the indicator’s spectroscopic properties; as a result, the binding efficacy of potential analytes can be assessed by absorbance and/or fluorescence measurements. The studies described in this section are directed towards the identification of an appropriate binding subunit (a functional tag for the library members) via an indicator displacement strategy. In the following experiments, the selected dendrimer serves as the receptor, and the compounds investigated as potential tags are considered the analytes. However, before the binding of tags can be probed by this method, the appropriate indicators must be identified. The details of this selection are described in the following section.

Scheme 2.6. Optical indicator displacement assay. Introduction of an analyte to a receptor:indicator (i.e., host:guest) complex establishes a competitive equilibrium that displaces the indicator from the receptor’s binding pocket, thereby modulating the indicator’s spectroscopic properties in a manner related to analyte concentration.

79

2.4.1 Screening and Binding of Indicators Two classes of commercially available PAMAM dendrimers were investigated in indicator binding studies.

Various generations of amino-terminated (AT-PAMAM)

dendrimers and hydroxy-terminated PAMAM (HT-PAMAM) dendrimers were employed in conjunction with various colorimetric indicators to identify dendrimer/indicator pairs amenable to future displacement methods. Potential dendrimer/indicator ensembles were quickly screened with the aid of a plate reader. Solutions of HT-PAMAM and ATPAMAM dendrimers of generations 2-4 were individually added to wells of a microplate containing 500 µM solutions of indicators in buffered 1:1 MeOH:H2O. The structures of the employed indicators are shown below (Figure 2.1). The indicators that gave the greatest color changes upon dendrimer addition are highlighted in Figure 2.1. Of the screened indicators, the best binders (i.e., the indicators that gave the greatest visible color changes) included xanthene-based dyes carboxyfluorescein (2.5), fluorescein (2.6), and pyrogallol red (2.7) and triarylmethane-derived pyrocatechol violet (2.4). The observed binding of fluorescein derivatives was particularly attractive because its permits the incorporation of fluorescent signaling elements in potential applications. No visible color changes were observed for the investigated azo dyes. Also, no HTPAMAM dendrimer promoted a color change for any of the indicators. The associated color changes were all observed with AT-PAMAM dendrimers. This makes intuitive sense because the indicator screenings were conducted in solutions buffered to pH ~ 7, where the amine surface is significantly protonated.23

A significant contribution of

electrostatic forces is implicated in the binding of anionic indicators to AT-PAMAM dendrimers.24

80

Figure 2.1 Colorimetric indicators screened for binding to PAMAM dendrimers. These indicators were screened for binding to PAMAM dendrimers. The indicators that gave the greatest visible color changes are highlighted in boxes. The indicators are shown in their fully protonated forms. Having identified several dyes that could serve as potential indicators, more rigorous titrations of these indicators with AT-PAMAM dendrimers were performed. In these titration experiments, a solution of a specific AT-PAMAM dendrimer was added to a solution of indicator. The indicator solution was held at constant concentration to obviate spectral changes due to concentration, and all titrations were carried out in buffered solutions to rule out indicator color changes based solely on pH fluctuations. Comparison of the pH values of the indicator solutions before and after complete addition of dendrimer revealed negligible changes of < 0.1 pH units. Representative spectral 81

modulations of indicators 2.4-2.7 with selected AT-PAMAM dendrimers are shown in Figure 2.2. These spectral changes are consistent with specific deprotonation events associated with the indicators (Scheme 2.7). 25,26,27 Because constant pH was maintained throughout the experiments, these deprotonation events are taken as evidence of dendrimer binding.

A.

B.

C.

D.

Figure 2.2 Absorbance modulations of selected indicators upon addition of ATPAMAM dendrimers. (A.) Spectral modulations of 15 µM 2.4 with increasing concentrations of G7 dendrimer. (B.) Spectral modulations of 20 µM 2.7 with increasing concentrations of G3 dendrimer. (C.) Spectral modulations of 20 µM 2.6 with increasing concentrations of G5 dendrimer. (D.) Spectral modulations of 6 µM 2.5 with increasing concentrations of G6 dendrimer. All experiments were performed in 1:1 MeOH:H2O, 25mM HEPES buffer, pH 7.3, except for (C.) which was done in 95:5 MeOH:H2O at pH 6.9 in 25mM HEPES. 82

Scheme 2.7 Ionization equilibria responsible for observed spectral changes of 2.4 and 2.5. (A.) The spectral modulations observed in Figure 2.2A are consistent with the formation of the dianion of pyrocatechol violet (2.4).25 (B.) The spectral modulations observed in Figure 2.2D are consistent with the formation of the trianion of 5carboxyfluorescein (2.5).26,27 The generational dependence of AT-PAMAM dendrimers on indicator uptake was assessed by comparing the binding isotherms of indicators with dendrimers of various generations.

These comparisons were made under consistent experimental

conditions, and the results are presented in Figures 2.3-2.5. Dendrimers of generations 36 were employed in these comparisons (the generational comparison for indicator 2.4 includes an isotherm associated with binding a G7 AT-PAMAM dendrimer; the high cost and limited amount of this dendrimer, however, restricted its use in additional comparative studies). Inspection of the resulting plots shows a definite generational dependence on uptake for each investigated indicator. Furthermore, the plots’ saturation points suggest that the larger dendrimers (G5, G6) are binding up to 10 indicator molecules per dendrimer. This observation is consistent with Meijer’s reports on his 83

“dendritic box” in which a generation 5 PPI dendrimer bound 8-10 molecules of pnitrobenzoic acid.28

Figure 2.3 Binding of 2.4 to various generations of AT-PAMAM dendrimers. An increase in uptake of 2.4 (pyrocatechol violet) was observed with increasing generations of dendrimers (G3-G7). These binding isotherms correspond to absorbance changes at 609 nm and were generated under identical experimental conditions: 15 µM 2.4 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.2.

84

Figure 2.4 Binding of 2.7 to various generations of AT-PAMAM dendrimers. An increase in uptake of 2.7 (pyrogallol red) was observed with increasing generations of dendrimer (G3-G6). These binding isotherms correspond to absorbance changes at 559 nm and were generated under identical experimental conditions: 20 µM 2.7 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.1.

85

Figure 2.5 Binding of 2.6 to various generations of AT-PAMAM dendrimers. An increase in uptake of 2.6 (fluorescein) was observed with increasing generations of dendrimer (G3-G6). These binding isotherms correspond to absorbance changes at 500 nm and were generated under identical experimental conditions: 20 µM 2.6 in 95:5 MeOH:H2O, 25 mM HEPES, pH 6.8. 2.4.2 Evaluation of Tripeptide Binding via Indicator Displacement While the studies presented above illustrate distinct generational trends in indicator binding, the identification of a suitable indicator for evaluation of tags via IDA was the original driving force behind these experiments.

As previously discussed,

peptides are cast as variable recognition units in the scenario illustrated in Scheme 2.5. As a result, the binding tag was originally envisioned to be a tripeptide sequence because established peptide coupling techniques29 allow for its facile covalent attachment. With this goal in mind, the binding efficacies of a series of tripeptides were investigated through indicator displacement. Tripeptides containing anionic and hydrophobic residues were selected for study.

The previously described studies demonstrate that anionic 86

indicators effectively bind to the dendrimers; this importance of negative charge in ATPAMAM dendrimer binding is illustrated by a comparison of the binding isotherms of fluorescein and 5-carboxyfluorescein (Figure 2.6). The uptake of fluorescein, which lacks the additional carboxylate of carboxyfluorescein, is much less dramatic than that of 5-carboxyfluorescein under similar conditions.

Hydrophobic residues were also

incorporated into the tripeptide segments because solvatochromic probes suggest a local polarity analogous to furan within the interior of G4 PAMAM dendrimers.30 Therefore, hydrophobic inclusion might assist electrostatic interactions in tripeptide-dendrimer binding.

Figure 2.6 The importance of anionic groups in binding AT-PAMAM dendrimers. Comparison of the binding profiles of 2.5 (5-carboxyfluorescein) and 2.6 (fluorescein) illustrates the impact anionic groups exert on binding the dendrimers. Both experiments were conducted under similar conditions: 20 µM indicator in 95:5 MeOH:H2O, 25 mM HEPES, pH 6.9. 87

Of the indicators, only pyrocatechol violet (2.4) was displaced to any appreciable extent upon addition of tripeptide; the tripeptides that effected displacement of 2.4 from a G5 AT-PAMAM dendrimer included Asp-Asp-Asp, Asp-Asn-Gln, and Glu-Gly-Phe. Gly-Gly-Glu and Gly-Leu-Phe did not displace 2.4 to any measurable extent. Binding of hydrophobic variant Phe-Phe-Phe could not be assessed because of its insolubility under the employed conditions.

While displacement of 2.4 was observed for several

tripeptides, Figure 2.7 shows the large excesses of Asp-Asn-Gln and Glu-Gly-Phe (relative to indicator) required to promote relatively small absorbance changes. Even polyanionic Asp-Asp-Asp required addition of nearly 300 equivalents before saturation behavior was established.

Interestingly, displacement of pyrogallol red 2.7 and

fluorescein 2.6 from the G5 dendrimer was not observed with any added tripeptide. Because no investigated tripeptide demonstrated the efficient dye displacement indicative of strong dendrimer binding, we looked at other anionic alternatives to serve as the binding tag. The indicators themselves were posited as potential tags due to their established dendrimer binding properties. The putative ease of conjugation of fluorescein derivatives to peptides31 led to some work directed towards their use as binding subunits. These efforts, however, were soon superseded by the preparation of an unnatural tricarboxylate. The development of this synthetic tag is described in the following section.

88

Figure 2.7 Tripeptide displacement of 2.4 from G5 AT-PAMAM dendrimer. Only polyanionic Asp-Asp-Asp showed appreciable displacement of pyrocatechol violet 2.4 from the dendrimer, and this displacement required hundreds of equivalents (relative to indicator) of added peptide. These experiments were conducted under analogous conditions: 15 µM 2.4 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.2. 2.4.3 Development of Binding Subunit A derivative of gallic acid (2.15) was originally envisioned as a tricarboxylate that could serve as the binding tag for members of the peptide library. The key features of 2.15 include its free carboxylic acid for facile attachment to the N-termini of peptides and the latent carboxylates (protected as their t-butyl ester derivatives) that are unmasked upon standard resin cleavage conditions.

The synthesis of 2.15 was attempted via

substitution of 2.14 by gallic acid 2.13 under basic conditions (Scheme 2.8). These reaction conditions are similar to an analogous transformation described in the 89

literature.32 Under these conditions, however, the present synthesis yielded the undesired tetra-alkylated product 2.16, while desired product 2.15 was not formed. The use of sodium hydride is reported to effect the selective alkylation of a phenoxide with 2.14 in the presence of an aromatic carboxylic acid.33 However, tetra-alkylated product 2.16 was again formed under these conditions, and there was no evidence for formation of 2.15.

O O

O

O

OH

O

O

OH +

HO

O Br

OH

K2CO3 O

DMF, heat

O

O O

O

O O

O

O

O O

OH 2.13

O

2.14

O

2.15

O

O O O

O O

2.16

Scheme 2.8 Attempted synthesis of a gallic acid-derived binding subunit. The synthesis of 2.15 was complicated by the formation of tetra-alkylated product 2.16. Shorter reaction times, cooling of the reaction mixture, and the use of sodium hydride did not circumvent the exclusive formation of 2.16. A selective saponification of 2.16 was then attempted to produce desired compound 2.15.

The t-butyl esters of tetra-alkylated product 2.16 should be less

susceptible to basic hydrolysis than the esters produced from alkylation of the carboxylate.34

Therefore, appropriate saponification conditions should afford 2.15.

Treatment of 2.16, however, with LiOH in a refluxing THF:H2O solvent mixture (1:10) only afforded significant amounts of starting material. Selective saponification of methyl ester derivative 2.17 was then considered (Scheme 2.9A). Compound 2.17 was prepared in analogy with the synthesis shown in Scheme 2.8. The selective unmasking of the carboxylate of the methyl ester to give desired tag 2.15, however, did not proceed as planned.

In this case, saponification conditions led to the cleavage of all ester

functionalites to give tetra-acid 2.18. NMR and mass spectral data associated with 90

hydrolysis of 2.17 with KOH in H2O:EtOH35 indicate the presence of both tetra-acid 2.18 and deprotected methyl ester 2.19. This finding suggests that the t-butyl esters are cleaved more readily than the electron-rich, aromatic methyl ester, an observation that precludes the synthesis of 2.15 by this approach. A final strategy involving selective saponification invoked the use of the benzyl ester of gallic acid (2.20 in Scheme 2.9B). As protecting groups, benzyl esters benefit from orthogonal cleavage conditions via hydrogenolysis.34 Therefore, use of the benzyl ester of gallic acid not only circumvents tetra-alkylation but also permits its selective removal by conditions that do not cleave the t-butyl esters. The benzyl ester of gallic acid 2.13 was successfully prepared through alkylation with benzyl iodide, but the yield of this derivative suffered from additional benzylation at the phenolic position para to the carboxylic acid in 2.13 to give dibenzylated product 2.21. Further investigation into this approach was curtailed by the successful parallel development of a regioisomeric binding tag through an oxidative strategy. The refinement and subsequent development of this binding tag is discussed in the following section.

91

O

O

O

LiOH O

A.

O O

O

8:2 H2O:THF reflux

O

O

HO

O

O O

O

O

OH

OH

O O

HO

O

1:2 H2O:EtOH reflux

HO

2.17

O

O

O

O O

2.19 O

OH

OH

O

HO

2.18

O

O O

KOH

O

O

O

O

BnI

B. HO

OH OH

2.13

DIPEA DMF

HO

OH O

HO

OH OH

2.20 2.21

Scheme 2.9 Attempted selective saponification routes to 2.15. (A.) An alternative route to binding tag 2.15 involved selective saponification of methyl ester derivative 2.17. Spectral data of the reaction mixture, however, indicated the presence of both fully hydrolyzed tetra-acid 2.18 and methyl ester tri-acid 2.19. (B.) The benzyl ester of gallic acid (2.20) obviates over-alkylation of gallic acid 2.13 and offers the potential for its orthogonal removal in the synthesis of 2.15. The formation of dibenzylated product 2.21 was also observed under these reaction conditions. 2.4.3.1 Synthetic Refinement of Binding Subunit The oxidative approach to synthesis of a binding subunit is outlined in Scheme 2.10. Compound 2.22 was prepared by alkylation of 2,3,4-trihydroxybenzaldehyde with 2.14.

Subsequent oxidation of the aromatic aldehyde provides entry to 2.23, a

regioisomer of 2.15 that might also be conjugated to peptides and deprotected under standard resin cleavage conditions. Oxidation of 2.22 was attempted with Oxone® (the potassium salt of peroxymonosulfuric acid), a known efficient and mild oxidant of aldehydes.36 The oxidation of aldehyde 2.22 with Oxone, however, was complicated by the formation of phenolic Dakin product 2.24.37 The formation of this undesired phenol 92

is a known side reaction in the Oxone-mediated oxidations of electron-rich aldehydes such as 2.22. Travis et al report that 4-hydroxy and 4-methoxy-substituted aldehydes give significant amounts of the phenolic Dakin products upon exposure to Oxone.38 While the desired acid was indeed observed in Oxone oxidation, the preponderance of phenol 2.24 prompted the search for more efficient oxidants in the preparation of 2.23. OH

O O

O

O

+

O O

O

O O

2.24

Scheme 2.10 Oxidative approach to the synthesis of the binding subunit. 2,3,4trihydroxybenzaldehyde derivative 2.22 can be oxidized to compound 2.23 to give a functional binding subunit amenable to peptide conjugation. Phenolic Dakin product 2.24 was formed upon oxidation of 2.22 with Oxone®. Sodium chlorite was determined to be the optimal oxidant for this transformation. The use of basic hydrogen peroxide has been reported in the oxidation of electron-rich aromatic aldehydes.39 However, application of these harsh conditions (the reaction is conducted in 50% aqueous KOH) to the oxidation of 2.22 did not afford the desired acid, and spectral data suggested hydrolysis of the t-butyl esters. A surfactantassisted KMnO4 oxidation40 of 2.22 was also attempted before sodium chlorite41 gave 2.23 in satisfactory yields. 2.4.3.2 Conjugation of Binding Subunit 2.23 to Peptides Attachment of binding tag 2.23 to the N-terminus of peptides was accomplished through standard solid-phase peptide coupling techniques. Tagged tetrapeptides 2.25 and 93

2.26 were prepared through automated synthetic protocols, and the employed resin cleavage conditions provided concomitant deprotection of the t-butyl esters of the “tag.” This facile attachment and deprotection scheme led us to consider an alternative role of dendrimer-bound peptides that serves as interesting and useful offshoot of the efforts heretofore described in the development of dynamically templated dendritic receptors. CO2H

CO2H O H Ala-Lys-Phe-Asp-N

O

O O

H Ala-His-Phe-Asp-N

CO2H

O O

CO2H

O

O

CO2H

CO2H

2.25

2.26

2.5 DENDRIMERS AS SCAFFOLDS IN NON-COVALENT SENSING ENSEMBLES This new direction involves the use of dendrimers as scaffolds in the assembly of non-covalent sensing ensembles. This sensing strategy, depicted in Scheme 2.11, was partly inspired by the prodigious abilities of AT-PAMAM dendrimers to bind various indicator dyes (section 2.4.1).

The sensing protocol shown below relies on the

competitive displacement of a peptide-bound indicator by an analyte of interest. Before this displacement, this peptide-bound indicator is held in proximity to a separate, dendrimer-bound indicator by virtue of the non-covalent dendrimer-peptide interactions. A judicious selection of indicators allows for a distance-dependent electronic communication between the dendrimer-bound and peptide-bound indicator species. Displacement of the peptide-bound indicator through the addition of a competitively binding analyte affects this intermolecular distance and therefore modulates the extent of this communication.

The resulting attenuation of intermolecular communication is 94

manifested in the indicators’ spectral properties, thereby establishing the basis of the analyte sensing protocol.

Scheme 2.11 Dendrimers as scaffolds in non-covalent sensing ensembles. An ATPAMAM dendrimer is pre-loaded with an indicator ( ), and peptides ( ) binds the dendrimer through the action of the covalently attached tag ( ). The bound peptide is non-covalently associated with an additional indicator ( ) capable of communication with the dendrimer-bound indicator. Introduction of an analyte ( ) disrupts this communication between dendrimer-bound indicator and peptide-bound indicator via a competitive indicator displacement, thereby providing the signal modulation associated with analyte presence. The work described in this section pertains to the selection of appropriate indicators for this sensing protocol. Related to this selection process is the preparation of a peptide functionalized with a discrete binding site capable of specific, non-covalent interactions with the selected indicator. Also, the elucidation of conditions that permit the requisite, simultaneous binding of indicator and peptide by the employed dendrimer is discussed. 2.5.1 Preparation of a Boronic Acid-Based Peptide Boronic acids were chosen as the recognition elements for incorporation into the tagged peptide. The established molecular recognition processes of boronic acids remain operational in aqueous solutions, and boronic acid-based receptors have been successfully integrated into indicator displacement assays.42 Reductive amination of a boronic acid-containing aldehyde by the primary amine of a lysine side-chain provides 95

entry to the desired functionalized peptide.43 Tagged peptide 2.25 was subjected to reductive amination conditions with aldehyde 2.27 in the solution phase; however, no product of reductive amination was observed by LC-MS.

Solution phase reductive

amination of 2.27 with FMOC-protected lysine derivative 2.28 was also attempted. If successful, the resulting protected, boronic acid-based lysine derivative could be directly incorporated into automated, solid-phase peptide synthesis methods; but again, no evidence for the formation of the desired amino acid derivative was observed.

O HO

H N

O H3C

O

2.28

H N

2.29

NH2

These failures prompted the consideration of a reductive amination performed on a resin-bound, lysine-containing peptide. This solid-phase installation of the boronic acid group requires an orthogonal protection of the lysine side-chain that allows selective unmasking of the aldehyde-reactive amine while avoiding wholesale cleavage of the peptide from the resin.

Methyltrityl (Mtt) protecting groups (2.29) promise such a

protection scheme.44 The solid-phase reductive amination strategy using Mtt-protected lysine is shown in Scheme 2.12. Boronic acid-based, tagged peptide 2.33 was prepared in this manner.

96

O O

O

O

O

O

H Ala Lys Phe Asp N

O

O O

H Ala Lys Phe Asp N

1.8% TFA

O

O

O

O O

O

O

in DCM

O

O

O

NH(Mtt)

2.31

2.30 O O 1) 2.27

O

NH2

O

O

H Ala Lys Phe Asp N

O O

2) NaBH4

95 % TFA

O

2.5 % H2O 2.5% HSi(Et)3

O O HN

O

OH B OH

2.32 O

O

O H Ala Lys Phe Asp N

HO

OH

O

O O

OH

O OH HN

OH B OH

O

2.33

Scheme 2.12. Incorporation of a boronic acid into tagged peptide 2.25 via solid-phase reductive amination. The lysine side-chain of resin-bound tagged peptide 2.30 was selectively deprotected for subsequent solid-phase reduction amination with boronic acidbased aldehyde 2.27. Standard peptide cleavage conditions severed the resin from the bead and fully unmasked the carboxylic acid groups of the peptide tag. 2.5.2 Dendrimer Binding/Uptake of Tagged Peptide To this point in the description of the current research project, a considerable fraction of the discussion has been devoted to the design and synthesis of an unnatural 97

binding subunit and its attachment to peptides. The binding efficacy of the peptide tag derived from carboxylic acid derivative 2.23 to AT-PAMAM dendrimers has thus far been neglected.

However, robust non-covalent interactions between peptide and

dendrimer are essential to the molecular sensing protocol shown in Scheme 2.11. The experiments described in this section were performed to validate the binding between a tagged peptide and the AT-PAMAM dendrimer.

Although crucial to the strategy

envisioned in Scheme 2.11, mere binding of tagged peptide to dendrimer is not the sole requirement of these partners in our proposed sensing protocol. As shown in Scheme 2.11, the employed dendrimer must bind the tagged peptide while simultaneously encapsulating multiple indicator molecules. As a result of this requirement, this section will also include a discussion of the conditions that effect such simultaneous binding and provide experimental evidence of this requisite molecular scenario. Binding of the tagged peptide 2.26 was assessed by titrating a solution of it into an ensemble of indicator and dendrimer. Peptide 2.26 lacks a boronic acid moiety but was selected simply as a representative of “tagged” peptides. Boronic acid-containing, tagged peptide 2.33 was not used in these initial binding studies because of its relative paucity.

5-Carboxyfluorescein (CF, 2.5) was selected as the indicator due to its

fluorescent nature. The use of fluorescent indicators enables the use of fluorescence resonance energy transfer (FRET)45 as the communication mechanism between dendrimer-bound and peptide-bound fluorophores. In the absence of FRET, ratiometric fluorescent signaling methods46 might also be applied. A G6 AT-PAMAM dendrimer (G6) was chosen as the dendrimer scaffold because of its efficient binding of 5carboxyfluorescein and relative cost. Because of its inherent sensitivity, fluorescence spectroscopy was used to monitor the binding the tagged peptide. The fluorescencebased experiments detailed in this section were obtained on a filter-based plate reader. 98

The use of a plate reader permitted automated solution dispensing that provided the reproducibility absent in our cuvette-based fluorescence measurements.

This

reproducibility is evidenced in the isotherm associated with the fluorescence titration of a G6 AT-PAMAM dendrimer into 5-carboxyfluorescein (2.5) shown in Figure 2.8. This isotherm was generated by plotting the individual fluorescence readings of solutions in a single row of a 96-well microplate. The amount of G6 dendrimer solution added to the wells increased across the microplate row, while the concentration of indicator was held constant. Data from two analogous fluorescence experiments are plotted in Figure 2.8 to demonstrate the reproducibility of this “simulated” titration. The vertical lines represent points along the curve that correspond to specific molar ratios of constant concentration indicator with respect to added dendrimer. The absolute fluorescence values associated with these points along the x-axis will be important in the following experiments that probe peptide binding by the dendrimer-indicator ensembles. When compared to the absorption changes caused by the addition of G6 ATPAMAM dendrimer to carboxyfluorescein (Figure 2.2D), the basis of fluorescent signal modulation can be inferred. As described in Section 1.6.3, a recent study by Chai on the encapsulation of fluorescein by a third generation PPI dendrimer suggests that dendrimer binding promotes a shift to the carboxylate form of fluorescein.47 A conflicting report published in 2008 asserts that fluorescein’s tautomeric equilibrium shifts to the colorless lactone upon interaction with low-generation, cationic PPI dendrimers.48 The observed dendrimer-promoted decrease in carboxyfluorescein’s emission illustrated below could therefore result from either a shift to the non-fluorescent lactonized tautomer or from PET-based quenching by unprotonated tertiary amines located in the dendritic branches. In the absence of supporting data, neither mechanism can be differentiated. Inspection of the corresponding absorption modulations, however, shows an increase in the intensity of 99

the long-wavelength absorption maxima. This observation rules out a transition to the less conjugated lactonized form upon dendrimer binding. As a result, the data presented in this dissertation agrees with Chai’s contention of a shift to the carboxylate form upon interaction with cationic dendrimer. In light of this evidence, the dramatic decrease in carboxyfluorescein’s emission in the presence of dendrimer must derive from a quenching process not structural isomerism.

Figure 2.8 “Simulated” fluorescence titration of G6 AT-PAMAM dendrimer into 5carboxyfluorescein. A decrease in the fluorescence intensity of 5-carboxyfluorescein (CF, 2.5) was observed upon increasing concentrations of dendrimer. A replicate data set is included to demonstrate the reproducibility of these parallel, filter-based fluorescence measurements. Both isotherms were generated under identical experimental conditions: 40 µM 2.5 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.3. Ex. Filter/Bandwidth = 485/20; Em. Filter/Bandwidth = 540/25.

100

Two scenarios can be anticipated upon addition of peptide 2.26 to the carboxyfluorescein/dendrimer

ensemble:

1)

competitive

displacement

of

carboxyfluorescein by the tagged peptide or 2) simultaneous uptake of the peptide by the carboxyfluorescein-containing dendrimer.

Because the realization of our proposed

sensing protocol requires that both the fluorescent indicator and the tagged peptide simultaneously associate with the dendritic scaffold, various “loading ratios” of dendrimer to carboxyfluorescein were investigated to determine which relative concentrations cause indicator displacement (scenario 1 above) and which promote simultaneous uptake (scenario 2 above) upon peptide addition. These “loading ratios” correspond to concentrations that give an excess of carboxyfluorescein with respect to dendrimer. These relative concentrations ranged from a 5-fold excess ([CF] / [G6] = 5) to a 100-fold excess ([CF] / [G6] = 100) of carboxyfluorescein relative to dendrimer. The results of this experiment are presented in Figure 2.9. The results presented in Figure 2.9 establish two salient points concerning the addition of tagged peptide to ensembles of dendrimer and 5-carboxyfluorescein: 1) at relatively low dendrimer loadings ([CF] / [G6] = 100 and 40), global displacement of carboxyfluorescein is observed and 2) at relatively high dendrimer loadings ([CF] / [G6] = 10 and 5), uptake of the tagged peptide by the dendrimer occurs in the presence of bound indicator. The inflection in the curve generated using an intermediate loading value of 25 might indicate a hybrid indicator displacement/peptide uptake behavior.

101

Figure 2.9 Addition of tagged peptide 2.26 to 5-carboxyfluorescein/dendrimer ensembles at various “loading ratios.” Isotherms of simulated fluorescence titrations of peptide 2.26 into indicator/dendrimer ensembles were generated by plotting fluorescence data of each well across a row of the microplate. The concentration of peptide was increased across the row of the 96-well plate, and each isotherm corresponds to a unique initial ratio of indicator to dendrimer. 7 µM 2.5 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.5, Ex. Filter/Bandwidth = 485/20; Em. Filter/Bandwidth = 540/25. Displacement of carboxyfluorescein at low dendrimer loading ratios ([CF] / [G6] = 100 and 40) is confirmed by the absolute value of the solution’s fluorescence at the end of the titration. This absolute fluorescence value corresponds to that of free, unbound 5carboxyfluorescein under the current conditions. Also, the indicator-binding isotherm shown in Figure 2.8 has several key reference points for the current experiment. In this plot, the point on the x-axis corresponding to a loading ratio of 100 CF / G6 shows a 102

fluorescence decrease of ~15,000 units from that of free, unbound 5-carboxyfluorescein. Addition of tagged peptide to this ensemble shows a maximum fluorescence change of ~15,000 units, suggesting a total displacement of bound 5-carboxyfluorescein from dendrimer. A similar situation is observed for the loading ratio of 40 CF / G6. The corresponding isotherm in Figure 2.9 shows saturation behavior at a maximal fluorescence change of ~25,000 units upon peptide addition at this indicator/dendrimer ratio. The magnitude of this fluorescence change is comparable to the absolute value of the change in indicator fluorescence observed upon addition of G6 dendrimer to 5carboxyfluorescein to give a loading ratio of 40 CF/G6 (see Figure 2.8).

Because

addition of peptide 2.26 to this ratio of dendrimer and indicator shows an ultimate restoration of free 5-carboxyfluorescein by saturation at 25,000 fluorescence units, the global displacement of indicator by the added peptide is also proposed as the dominant regime under these conditions. This observed indicator displacement makes intuitive sense under these conditions employing large excesses of indicator. Previous indicator binding studies (see Section 2.4.1) suggest that a single AT-PAMAM dendrimer can sequester several 5carboxyfluorescein molecules. When the dendrimer concentration is very low relative to that of 5-carboxyfluorescein (as it is with loading ratios of [CF] / [G6] = 100 and 40), an excess of indicator molecules are forced into only a few dendrimers. Displacement of these indicator molecules must occur to accommodate the addition of a competitive binder such as tagged peptide 2.26. This situation is represented in Scheme 2.13A. While this scenario is not commensurate with the goals of this project, the binding experiments performed under these conditions confirm that the tagged peptide is indeed binding to the dendrimer, an interaction that is crucial to the ultimate assay.

103

Uptake of the added peptide by the indicator/dendrimer ensemble is effected at higher relative concentrations of dendrimer to indicator ([CF] / [G6] = 10 and 5). At these dendrimer concentrations, there are fewer indicator molecules per dendrimer. As a result, the added peptide can be accommodated by the dendrimer without concomitant indicator displacement (Scheme 2.13B). In contrast to the absolute fluorescence values observed after complete peptide addition to solutions containing lower relative dendrimer concentrations, those noted at CF/G6 loading ratios of 10 and 5 did not approach a value corresponding to free, unbound 5-carboxyfluorescein. While the x-axis of the indicator binding isotherm shown in Figure 2.8 does not extend to relative dendrimer concentrations corresponding to [CF] / [G6] loading ratios of 10 and 5, the saturation behavior seen in this plot beyond loading ratios of 25 is assumed to also be operative for these even lower loading ratios. Comparing the magnitude of the overall fluorescence decrease extrapolated for CF/G6 loading ratios of 10 and 5 (from Figure 2.8) to the small overall fluorescence increases seen upon addition of tagged peptide to solutions at these ratios (see Figure 2.9), one can conclude that indicator displacement by the added peptide is minimal at these initial loading ratios. This assertion is based on the observation that, under these conditions, the added peptide does not restore the fluorescence associated with unbound 5-carboxyfluorescein as it did with loading ratios of 100 and 40. Again, this circumvention of an undesired indicator displacement regime is attributed to the higher initial dendrimer concentrations (relative to indicator) that result in fewer bound indicators per dendrimer, thereby freeing up binding sites for simultaneous peptide uptake. Inspection of Figure 2.9 suggests that CF displacement is minimal even at a 10fold excess of peptide relative to dendrimer (at a loading ratio of CF/G6 = 5, two equivalents (relative to indicator) of added peptide represents a 10-fold excess of dendrimer with respect to peptide.) Binding of the peptide to the dendrimer is still 104

assumed to occur at these increased dendrimer loadings since the previous displacement analyses confirmed peptide binding at even lower dendrimer concentrations.

Scheme 2.13 Indicator displacement and peptide uptake regimes observed upon addition of a competitive binder to an indicator/dendrimer ensemble. (A.) At high indicator to dendrimer ratios, the binding sites of the dendritic scaffold are saturated with indicator. Addition of a competitive binder (such tagged peptide 2.26) can only be accommodated by the displacement of indicator from the dendrimer, resulting in an observable change in the indicator’s optical properties. (B.) At low indicator to dendrimer ratios, there are fewer indicator molecules per dendrimer, and additional binding sites remain available. Addition of a competitive binder is accommodated without global indicator displacement, leading to a scenario referred to in this dissertation as “peptide uptake.” 2.5.3 Development of Signaling Protocol Once the binding of the peptide to the dendrimer was established and the conditions that promote peptide uptake by indicator-bound dendrimers were elucidated, 105

the development of the signaling protocol represented in Scheme 2.11 became the next hurdle in the progress towards dendrimer-based, non-covalent sensing ensembles. Because the sensing strategy proposed in Scheme 2.11 invokes a change in proximity between two fluorophores upon analyte addition, a FRET-based signaling mechanism was investigated. The efficiency of energy transfer via FRET mechanisms is known to depend on the distance between the donor and acceptor fluorophores.49 To probe the potential for FRET signaling with our system, methylesculetin (MEL, 2.34) was selected to serve as the peptide-bound indicator (

in Scheme 2.11). It was chosen for the

following studies because it is a catechol-based fluorescent indicator that will likely serve as a FRET donor when paired with 5-carboxyfluorescein. Peptide 2.33 is expected to bind indicator 2.34 because catechols are known to interact with boronic acids50; and methylesculetin is a derivative of coumarin, a fluorophore known to serve as FRET donor for fluorescein-based acceptors.51

A normalized plot that overlays experimentally

obtained absorbance and emission spectra of 5-carboxyfluorescein and methylesculetin 2.34 is shown in Figure 2.10. The good overlap between the emission of 2.34 (green line) with the excitation wavelengths of 2.5 (blue line) augurs well for the establishment of a FRET-based signaling method.

106

Figure 2.10 Overlay of absorption and emission bands of fluorescent indicators 2.34 and 2.5. The yellow and green plots correspond to the absorption band and emission spectrum of MEL 2.34; the blue and red plots correspond to the absorption band and emission spectrum of CF 2.5. The gray boxes refer to the available bandwidths for the filter-based fluorescence measurements. The structure of 5-carboxyfluorescein (2.5) and methylesculetin (2.34) are also shown. An initial control study was performed to probe the energy transfer between dendrimer-bound CF (the acceptor) and MEL (the donor) in the absence of boronic acidbased peptide 2.33. In this control experiment, MEL concentration was increased across a row of a microplate containing constant concentrations of G6 dendrimer and CF at a loading ratio of CF/G6 = 5. Using the filter-based system of the plate reader, the solutions were excited near MEL’s λmax, and emission was monitored at both CF and MEL’s emission maxima. Figure 2.11 shows that fluorescence emission from MEL increases linearly. A deviation from this linearity would be expected if the fluorophores are participating in FRET, but the observed linear increase of MEL emission is simply a reflection of its increasing concentration. Similarly, the emission from the CF acceptor 107

shows only a slight enhancement as the MEL donor concentration is increased, and this increase is likely due to the slight overlap of the CF absorption band in the excitation region (see Figure 2.10). The observed absence of FRET between CF and MEL in the presence of dendrimer is not surprising, however. Spectrophotometric titrations have shown that CF binds to the dendrimer (and is likely encapsulated to some extent47) while similar UV-Vis titrations prove that MEL does not bind to AT-PAMAM dendrimers. As a result, the donor and acceptor fluorophores are isolated from each other, thereby obviating energy transfer.

Figure 2.11 Probing FRET in the presence of dendrimer. The potential for FRET between dendrimer-bound CF and MEL was investigated in the presence and absence of tagged peptide 2.33. Energy transfer between the two fluorophores did not occur in either case. Filter-based excitation was performed with a bandwidth of 380/20 (nm). CF and MEL emission were monitored with a 540/25 filter and a 460/40 filter, respectively. 6.6 µM CF, 1.2 µM G6 dendrimer, 12 µM 2.33 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.3.

108

However,

addition

of

boronic

acid-based

peptide

2.33

to

these

indicator/dendrimer ensembles was expected to promote FRET between the MEL donor and the CF acceptor. The binding of tagged peptides to the dendrimers has already been established. Dendrimer-binding peptide 2.33 also contains a boronic acid group that can bind MEL, sequestering this donor in proximity to the dendrimer-bound CF to turn on the FRET signal. The experiments described above were repeated in the presence of 10 equivalents (relative to dendrimer) of peptide 2.33.

Unfortunately, the addition of

boronic acid-based peptide did not alter the emission properties of either fluorophore upon comparison to the control experiment (Figure 2.11). The lack of a FRET signal in the presence of 2.33 might indicate that MEL is not effectively binding to the peptide under the employed conditions. To account for this possibility, a control titration of peptide 2.33 into MEL was performed at the same concentrations used in the previous experiments that probed FRET. Indeed, no binding of MEL to the peptide was observed under these conditions.

Repetition of this

experiment at 50x higher MEL concentrations did show evidence of peptide binding; however, the concentrations required for this interaction to be operational are prohibitive given the high cost of dendrimer and the available amount of 2.33. 2.5.3.1 Attachment of Binding Tag to Esculetin Derivative. The use of a MEL fluorophore functionalized with binding tag 2.23 was proposed to overcome these untenable concentration requirements. Direct binding of a MEL derivative to the dendrimer might position its catechol functionality in an orientation that enhances binding to the dendrimer-bound peptide. The simultaneous binding of MEL and peptide to dendrimer might offer an increase in the effective molarity of the interacting groups, leading to complex formation between the catechol and boronic acid moieties at more amenable concentrations. Introduction of an appropriate analyte would 109

then disrupt this interaction, providing the basis of the signaling mechanism. To this end, model compound 2.35 was synthesized by alkylation of a 4-bromomethyl coumarin derivative, but it was not successfully deprotected. O O O

O

O

O

O O

O

O O O

O O

O

2.35

An alternative route to a tagged esculetin derivative is shown in Scheme 2.14. 4Carboxymethyl coumarin 2.36 (MEL-COOH) can be coupled to a hemi-protected bisamine linker. Subsequent deprotection of the linker and amide coupling to binding tag 2.23 would yield tagged MEL derivative 2.38. Use of PyBop as the coupling reagent afforded the presence of intermediate 2.37 as ascertained by LC-MS. This product, however, could not be purified by flash chromatography. HO

O

O

HO O HO

O

O

HO O OH

2.36

H2N

NHBoc

PyBop, NMM, DMF 20 h, RT

HO

O

O

NH

HO O

O

NH

HN

2.37

O

NHBoc

2.38

CO2H

O O

CO2H

CO2H

Scheme 2.14 Proposed synthesis of a tagged MEL derivative. Coupled product 2.37 was observed by LC-MS, but could not be purified by flash chromatography. 110

Protection of the catechol moiety of 2.36 was attempted to improve the efficacy of the coupling reaction and to aid in the isolation of the amine coupled product. Protected derivatives 2.39 and 2.40 were targeted for preparation. Formation of a cyclic ethyl orthoformate (as in 2.39), a known catechol protecting group, was attempted as described in the literature52, but this reaction only gave a decarboxylated product corresponding to methylesculetin 2.34. Putative protection of the catechol as an acetal derived from cyclohexanone (2.40) also yielded only decarboxylated product 2.34. O

O

O

O

O

EtO

O

O

O

O

O

2.39

2.40

OH

OH

2.5.3.2 Binding and FRET Signaling Studies Using Compound 2.36 Following these synthetic difficulties, the use of 4-carboxymethyl coumarin 2.36 (MEL-COOH) was investigated for dendrimer binding. While 2.36 lacks the multiple carboxylates of tagged binders derived from 2.23, perhaps the presence of a single carboxylate is sufficient to promote its binding to dendrimer. If this is the case, then the covalent attachment of an exogenous binding tag is unnecessary. Indeed, Figure 2.12 shows the absorbance modulations (and associated isotherm) that occur upon addition of G6 dendrimer to a solution of MEL-COOH (2.36). This behavior is consistent with the binding of esculetin derivatives reported in the literature,53 but it was not observed upon addition of dendrimer to MEL 2.34.

111

Figure 2.12 Binding of 2.36 to a G6 AT-PAMAM dendrimer. Change in the absorption spectrum of 2.36 upon increasing dendrimer concentration is shown on the left. The isotherm derived from spectral changes at 399 nm is presented on the right. 20 µM 2.36 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.5. After binding of MEL-COOH to the G6 dendrimer was validated, several studies were designed to probe FRET between the dendrimer-bound fluorophores 2.34 and 2.5. The potential for FRET between these two fluorophores exists because they will be proximally sequestered upon binding to the dendrimer. Energy transfer processes have been previously observed between fluorophores in aqueous micellar solutions.54 In these systems, micellar assemblies encapsulate hydrophobic fluorophores, thereby holding them in the proximity required for efficient FRET. The role of the micelle in these examples is analogous to the proposed role of the dendrimer in the current studies. An initial investigation into FRET between dendrimer-bound fluorophores was a cuvettebased titration of an ensemble of CF and dendrimer (at a loading ratio of 5 to 1) into a solution of MEL-COOH (Figure 2.13). The resulting spectral changes were promising for the existence of FRET between CF and MEL-COOH. Emission from the MELCOOH donor (~ 460 nm) decreased as the concentration of CF increased in the presence of dendrimer (to rule out spectral changes due to dilution, the concentration of MEL112

COOH was held constant throughout the titration). Conversely, emission from the CF acceptor (~ 530 nm) increased upon addition to MEL-COOH donor. Plotting the change in emission at this wavelength against the amount of titrant added shows that this fluorescence enhancement is not simply the result of increased CF concentration. The isotherm plotted for the emission changes at 530 nm shows saturation behavior at increasing amounts of the G6/CF ensemble (Figure 2.13). This deviation from linearity indicates an interaction between the fluorophores that might correspond to FRET. The isotherm generated from the emission decrease at 457 nm also showed saturation behavior with respect to increasing amounts of CF/G6 ensemble (not shown).

Figure 2.13 Spectral evidence for FRET between dendrimer-bound fluorophores. The fluorescence data collected from the titration of a G6/CF ensemble into a solution of MEL-COOH shows behavior indicative of energy transfer between the MEL-COOH donor and CF acceptor. Excitation was performed at 346 nm. The number of added equivalents of dendrimer is relative to MEL-COOH. 100 µM MEL-COOH in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.3. Control experiments were then performed to complement and confirm the results discussed above. A fluorescence emission scan of the G6/CF/MEL-COOH ensemble (at a respective molar ratio of 1/5/13) was recorded using an excitation wavelength of 346 nm (the blue data series in Figure 2.14). The emission bands of MEL-COOH donor and 113

CF acceptor are both observed in this composite spectrum. It remained unclear, however, whether the observed CF emission ~530 nm is due to FRET from the donor MELCOOH, or simply derives from a competing absorbance of CF at the incident excitation wavelength (i.e., cross-talk55). To address this concern, a fluorescence emission scan of the CF/G6 ensemble in the absence of the MEL-COOH donor was recorded (the red data series in Figure 2.14) using the same excitation wavelength as in the previous composite spectrum. If the CF emission seen in the previous composite spectrum originates from FRET, then removal of the donor fluorophore should result in the absence (or at least attenuation) of the CF emission band in the associated spectrum. If the CF emission observed in the presence of the donor results solely from the acceptor’s competitive absorbance at the incident wavelength, then removal of the donor fluorophore should lead to an increase in the CF emission band. The latter scenario was indeed observed upon recording the emission scan in the absence of MEL-COOH donor. As seen in Figure 2.14, the intensity of CF emission is dramatically increased in the absence of donor, suggesting significant cross-talk between the donor and acceptor fluorophores. These results suggest that the observed CF emission does not originate from FRET but rather stems from overlap of the absorption spectra of CF and MEL-COOH. A composite absorption spectrum of the CF/G6 ensemble did have a slight absorbance at 346 nm (not shown).

114

Figure 2.14 Control experiments to validate FRET between dendrimer-bound fluorophores. A composite emission spectra of the G6/CF/MEL-COOH ensemble shows both MEL-COOH emission and CF emission using an excitation wavelength of 346 nm. When the MEL-COOH donor is removed from the sensing ensemble, an increase in the CF emission is observed, suggesting significant cross-talk between the two fluorophores. 100 µM MEL-COOH, 40 µM CF, 7 µM G6 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.3. To reconcile these seemingly conflicting FRET studies, an excitation experiment was then performed. In an excitation study, the excitation wavelength is varied while the resulting fluorescence emission is monitored at a single wavelength.

This approach

allows a selective examination of fluorescence emission from those excitation wavelengths exclusive to MEL-COOH absorption.

This experiment also permits

monitoring of fluorescence emission of the CF acceptor as a function of excitation wavelength.

If the exclusive excitation of MEL-COOH modulates the fluorescence

intensity of the CF emission band, then this will be manifested in the excitation spectrum to provide evidence for FRET between the dendrimer-bound fluorophores. 115

The excitation experiment was performed using the G6/CF/MEL-COOH ensemble at a respective ratio of 1:5:13. Emission was monitored at 533 nm (the CF emission maxima in the composite spectra shown in Figure 2.14). A scan of excitation wavelengths ranging from 300 – 420 nm shows a maximum centered near 400 nm (Figure 2.15). Referring back to Figure 2.12, this excitation maximum corresponds to the absorption maximum of dendrimer-bound MEL-COOH.

This finding suggests that

excitation of dendrimer-bound MEL leads to a fluorescence enhancement in the emission of CF. It is important to note that the absorption band corresponding to free MEL-COOH (~ 350 nm) does not appear in the excitation spectrum, indicating that unbound MELCOOH donor does not participate in FRET with the CF acceptor. An absorbance control of only the CF/G6 ensemble shows an absorption minimum at 400 nm, suggesting that the emission observed in the excitation spectrum largely results from the exclusive excitation of the bound donor MEL-COOH.

Figure 2.15 Excitation scan of the G6/CF/MEL-COOH ensemble. An excitation scan of the sensing ensemble reveals a band that corresponds to the absorption maximum of dendrimer-bound MEL-COOH. 100 µM MEL-COOH, 40 µM CF, 7 µM G6 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.3. 116

The promising result of the excitation scan encouraged further investigations into this sensing strategy.

A fluorescence titration of MEL-COOH into a CF/G6/peptide

ensemble was performed to assess the signal modulation that occurs when the MELCOOH donor binds to the dendrimer-bound, boronic acid-containing peptide 2.33. A control study that mimicked the conditions of this titration in the absence of peptide 2.33 was also conducted. Both studies employed an excitation wavelength of 346 nm, and the results are presented in Figure 2.16. In the control study (Figure 2.16A), the addition of MEL-COOH to the indicator and dendrimer in the absence of peptide 2.33 promoted an increase in the fluorescence emission centered at 444 nm due to the increasing concentration of MEL-COOH. Conversely, the signal corresponding to CF emission (538 nm) decreases with increasing amounts of MEL-COOH because the added fluorophore competitively absorbs the excitation radiation, leading to an attenuation of the CF emission intensity.

Similar behavior is observed for CF emission in the

complementary fluorescence titration featuring added peptide 2.33, but a subtle difference is noted in the response of the MEL-COOH emission band in this experiment (Figure 2.16B). Upon addition of MEL-COOH to the G6/CF/peptide ensemble, the fluorescence signal associated with MEL-COOH grows in at 435 nm, a value that differs from that observed in the control titration by 9 nm. Increasing concentrations of MELCOOH causes a subsequent decrease in this emission, and the fluorescence maximum undergoes a bathochromic shift back to 444 nm, the emission maximum associated with MEL-COOH in the previous control experiment. While the cause of this behavior is somewhat unclear, this spectral modulation in the presence of peptide 2.33 proved reproducible and may serve in the sensing of analytes that can competitively interact with the boronic acids side-chain of 2.33. Subtle spectral variations have previously proven sufficient for effective discrimination of a class of analytes.56 117

A.

B.

Figure 2.16 Spectral modulations of G6/CF/peptide ensemble upon addition of MELCOOH. (A.) Addition of MEL-COOH to a solution containing 5:1 CF:G6 promotes the growth of an emission maximum centered at 444 nm. (B.) Addition of MEL-COOH to a solution of 5:1:10 CF:G6:peptide promotes the growth of an emission maximum centered at 435 nm with a subsequent red-shift to 444 nm at higher MEL-COOH concentrations. 24 µM G6, 125 µM CF, 240 µM peptide 2.33 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.2. 118

2.5.4 Synopsis The facile conjugation of an unnatural binding subunit to multiple peptides inspired the use of dendrimer-bound peptides in non-covalent molecular sensing ensembles. The binding of this tagged peptide to a AT-PAMAM dendrimer (G6) was achieved in the presence of dendrimer-bound indicator 5-carboxyfluorescein (CF). Conditions mitigating the global displacement of dendrimer-bound CF by the added peptide were elucidated by a series of experiments using varying molar ratios of indicator to G6. These experiments showed that relatively low CF/G6 ratios (e.g., 5 and 10) accommodated both indicator and peptide binding within the dendrimer, while the peptide-promoted indicator displacement observed for higher CF/G6 ratios (e.g., 40 and 100) confirmed the dendrimer-binding properties of the binding subunit. A boronic-acid containing peptide bearing the unnatural binding “tag” was also prepared through solidphase reductive amination of a lysine side-chain. The boronic acid moiety serves as a supramolecular “handle” for the implementation of an indicator displacement assay in which the components are non-covalently assembled by a dendritic scaffold. FRET-based signaling protocols were then assessed using a CF acceptor with catechol-based coumarin derivatives as the donors. The dendrimers are proposed to serve as organizing scaffolds that promote this requisite energy transfer between the fluorophores. When experimental conditions precluded the binding of methylesculetin to the functionalized side-chain of a dendrimer-bound peptide, the direct dendrimer binding of a coumarin-derived donor fluorophore was investigated. Evidence for FRET between dendrimer-bound fluorophores 4-carboxymethyl coumarin (MEL-COOH) and CF was observed in both emission and excitation experiments.

Incorporation of the tagged

boronic acid-based peptide into the CF/MEL-COOH/G6 ensemble also provided a small yet reproducible spectral modulation that might be exploited in array sensing platforms. 119

2.6 METAL BINDING BY AT-PAMAM DENDRIMERS The binding of Cu+2 to AT-PAMAM dendrimers was also investigated for use in our system. The interaction of copper complexes with catechol-based indicators has been previously exploited in indicator displacement assays57 and implemented in sensor arrays based on these IDAs58; and the interaction of divalent metals with esculetin derivatives has been parlayed into a ratiometric fluorescent probe for zinc.52 The binding of Cu+2 by PAMAM dendrimers has also been documented through EPR and atomic absorption spectrometry.59 (see section 1.7.3.2). In his preparation of copper nanoclusters within PAMAM dendrimers, Crooks provides evidence of these dendrimers’ Cu+2 binding capacities based on UV-vis absorption data.60 He notes a shift of the metal ion’s d-d transition from ~800 nm in the absence of dendrimer to ~600 nm in the presence of a hydroxy-terminated G4 PAMAM dendrimer. Because the introduction of divalent metals such as Cu+2 might lead to enhanced sensor diversification in future array-based applications of our sensing ensemble, the binding of Cu+2 to the G6 AT-PAMAM dendrimer was assessed. As in Crooks’ report, the addition of dendrimer promoted an increase in the solution’s long-wavelength absorption with a maximum centered ~600 nm (Figure 2.17). Based on the previous literature report, the position of this λmax suggests that the added Cu+2 is binding to the dendrimer. A control study confirmed that the emergence of this band is due solely to the presence of dendrimer (Figure 2.18).

120

Figure 2.17 Binding of Cu+2 to G6 AT-PAMAM dendrimer. Addition of a solution of CuCl2 to dendrimer promotes an increase in the absorption ~600 nm. This finding is consistent with the reported binding of Cu+2 to hydroxy-terminated PAMAM dendrimers. 5 µM G6 in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.3.

Figure 2.18 Verification of the spectral response associated with copper binding. The absorbance spectrum of a solution of 110 µM CuCl2 in the buffered 50:50 MeOH:H2O solvent system undergoes a subtle yet reproducible change upon exposure to the G6 dendrimer. 121

A titration of CuCl2 into the CF/G6 ensemble also gives evidence of copper binding. While no absorption modulations at long wavelengths (e.g., ~600 nm) were observed upon addition, surveillance of the CF absorption band shows a hyperchromic increase in intensity as the concentration of metal ion is increased (Figure 2.19A). This spectral

change

suggests

a

change

in

the

dendrimer-bound

chromophore’s

microenvironment as the copper ions also begin to occupy binding sites within the dendrimer. Significant modulation of the bound CF’s fluorescence emission also occurs with increasing Cu+2 concentration (Figure 2.19B). These absorbance and fluorescence changes are not consistent with CF displacement, and therefore provide evidence of a simultaneous binding of indicator and metal by the dendritic scaffold.

A.

B.

Figure 2.19 Spectral evidence for the uptake of copper ions by the CF/G6 ensemble. (A.) The absorbance of the CF/G6 ensemble at a respective ratio of 5/1 is hyperchromically increased as CuCl2 is titrated in. (B.) The emission intensity of the CF/G6 ensemble at a respective ratio of 5/1 is decreased as CuCl2 is titrated in. These experiments suggest that the binding of the added copper is accommodated by the CF/G6 ensemble. The slight spectral modulations suggest a change in the CF microenvironment that is not consistent with its displacement from the dendrimer. 1.2 µM G6, 6 µM CF in 1:1 MeOH:H2O, 25 mM HEPES, pH 7.3.

122

2.7 CONCLUSIONS AND OUTLOOK Dynamically templated dendritic receptors featuring well-defined binding sites represent

synthetic

biomacromolecules.

constructs

that

can

potentially

rival

the

functions

of

Successful generation of such macromolecular structures must

appropriate aspects of established imprinting strategies such as those employed in MIPs and DCLs.

As a result, this chapter detailed efforts directed toward the templated

assembly of imprinted receptors based on a dendritic scaffold. While the notion of this novel templating strategy informed many of the initial studies described in this chapter, the subsequent extension of this system into the realm of molecular sensing constitutes a significant portion of the described progress. The identification of appropriate indicators as potential signaling elements is one contribution of the work described herein. Subsequent comparison of indicator binding with dendrimer generation offered general trends in indicator binding, and comparison of the binding profiles of various indicators with a selected dendrimer provided insight into the functional requirements of an effective binding subunit.

With these functional

requirements in mind, an unnatural binding subunit was conceived and synthesized and easily conjugated to the N-termini of peptides. The binding of these “tagged” peptides to AT-PAMAM dendrimers was confirmed by a series of experiments that also elucidated the conditions necessary for simultaneous dendrimer binding of tagged peptide and indicator. This molecular scenario is necessary for the envisioned sensing protocol. The proposed sensing methodology also required the use of a tagged peptide functionalized with a discrete binding site. To this end, a boronic acid-based tagged peptide was prepared for incorporation into the sensing ensemble. The use of this system in dendrimer-based molecular sensing ensembles was prompted by several features gleaned from these initial studies. 123

The dendrimers’

prodigious indicator binding abilities and their demonstrated potential for simultaneous peptide uptake portends their use as scaffolds in molecular sensing ensembles. In the advancement of such goals, several potential sensing modes involving dendritic scaffolds were evaluated. These modes involved FRET-based signaling between dendrimer-bound indicators and optical modulations of a fluorescent indicator due to its interaction with a dendrimer-bound functionalized peptide. Preliminary metal binding studies involving AT-PAMAM dendrimers were also carried with the intention of identifying an additional mode of sensor diversification for future incorporation into array-based protocols. 2.8 EXPERIMENTAL DETAILS FOR CHAPTER 2 2.8.1 General Considerations All PAMAM dendrimers were purchased as methanol solutions from Dendritech. The tripeptides used in section 2.4.2 were obtained from Bachem. 5-carboxyfluorescein was obtained from Molecular Probes as a single isomer. Coumarin derivatives were purchased from Indofine. The suitably protected amino acids were purchased from NovaBiochem. Other organic starting materials were purchased from Aldrich or Acros and used without further purification.

Flash chromatography was performed using

Whatman 60 A 23-400 mesh silica gel. All products were dried for at least 3 hours prior to spectral analysis. A Varian Unity Plus 400 MHz spectrometer recorded 1H and 13C NMR spectra. The resulting NMR spectra were referenced to the solvent peaks. Highresolution mass spectra data was obtained by a MicroMass AutoSpec-Ultima spectrometer. LC-MS was performed with aid of a Thermo LTQ-XL linear ion trap mass spectrometer. UV-vis absorption data was collected on a Beckman DU-800 spectrophotometer. Cuvette-based fluorescent measurements were recorded by a Photon Technology

International

QuantaMaster 124

spectrofluorimeter.

Microplate-based

fluorescent measurements were made using a Biotek Synergy 2 Multi-detection Microplate Reader. Microplates were filled with the aid of a Biotek Precision Automated Pipetting System.

Absorbance and fluorescence data were collected at ambient

temperature. Water used in the preparation of buffer solutions was distilled, deionized, and filtered.

Automated peptide synthesis was accomplished using both Protein

Technologies’ Prelude parallel peptide synthesizer and a CEM Liberty microwave peptide synthesizer. pH values were recorded with an Orion 720A pH meter equipped with a glass electrode. 2.8.2 Cuvette-based Titrations Although concentrations varied, the following represents a typical cuvette-based titration: a solution of 2.36 (20 µM) in 1:1 MeOH:H2O containing 25 mM HEPES was prepared, and a measured amount was transferred to an appropriate cuvette. A second buffered solution containing 20 µM 2.36 and 3 µM G6 dendrimer was added in aliquots to the cuvette containing the original indicator solution.

The absorption and/or

fluorescence spectrum of the sample was recorded between each of these additions. Data collected at wavelengths associated with maximal fluorescence/absorbance changes was plotted as a function of dendrimer concentration. In the case of absorbance titrations, blanks were performed on the employed solvent system prior to data collection. 2.8.3 Plate Reader-based Parallel Titrations Although concentrations were altered as necessary, the following represents a typical titration performed on a multi-detection plate reader: 50 µL of 40 µM CF (2.5) and 60 µL of 6 µM G6 dendrimer in buffered 1:1 MeOH:H2O were added by an automated pipetting system to the 12 wells across a row of a 96-well microplate. Aliquots of a 20 µM solution of MEL (2.34) were added in increasing amounts across the 125

12 wells of the row (added volumes ranged from 0-140 µL). Decreasing amounts of buffer solution were then dispensed across the microplate to bring each well to a total volume of 300 µL, providing consistent concentrations of G6 and CF in each well. The solutions in each well differed only in MEL concentration.

Both absorbance and

fluorescence data could be collected in a single experiment. Absorbance measurements were obtained by a wavelength scan that generated absorption spectra when plotted. These spectra were overlayed, and absorption data at wavelengths associated with maximal spectral changes was used in plotting isotherms relating absorbance change to MEL concentration. Fluorescence data was collected by a filter-based method. As a result, emission spectra of the solutions could not be collected on the plate reader. Instead, single point fluorescence emission values of the contents of each well were obtained by the applicable excitation/emission filter sets. These values were plotted versus the increasing MEL concentration to provide the isotherm. Because of the parallel nature of the titrations conducted on the plate reader (as opposed to the standard serial addition used in cuvette-based titrations), they are referred to in this dissertation as “simulated” titrations. 2.8.4 Solid-Phase Peptide Synthesis 2.8.4.1 General Automated Methods Automated peptide synthesis was performed on a Wang resin pre-functionalized with an Fmoc-protected Ala residue, and subsequent couplings were performed in DMF with HBTU as the activating agent. Coupling was carried out in the presence of N,Ndiisopropylethylamine. Fmoc deprotection was effected by washing the resin in 20% v/v piperdine in DMF containing 0.1 M HOBt. After the automated synthesis, the beads were collected from the synthesizer, washed several times with glacial AcOH followed 126

by MeOH, and dried under vacuum for at least 3 hours. The beads were then transferred to a round bottom flask, and the peptide was cleaved overnight from the resin by a mixture of 95% TFA containing 2.5% H2O and 2.5% triethylsilane as scavengers (10-25 mL cleavage solution per gram of resin). The cleavage solution was separated from the resin by filtration on a fritted syringe. The resin was washed with TFA several times, and the washings were combined with the filtrate. The volume of the filtrate was reduced by evaporation with nitrogen, and ~10 volumes of ethyl ether was added to the remaining solution. The resulting white precipitate was collected by centrifugation and lyophilized. Peptides were characterized by LC-MS. 2.8.4.2 Manual Solid-Phase Synthesis of Boronic Acid-based Peptide 2.33 After automated synthesis of the peptide backbone of 2.33, the beads were transferred to a solid phase reaction vessel and dried under vacuum. The Mtt protecting group of the lysine side-chain was selectively deprotected by washing the resin in DCM containing 1.8% TFA for 3 minutes. The solution was drained from the beads, and this treatment was repeated until the washings ran clear and the beads were decolorized. The resin was again dried under vacuum before it was soaked in triethyl orthoformate (TEOF) with 1% glacial AcOH for 30 minutes. After draining the reaction vessel, 7 equivalents of 2-formyl phenylboronic acid 2.27 was added to the resin in 1:1:1 MeOH:TEOF:NMP containing 1% glacial acetic acid. The contents of the reaction vessel were agitated in this solution overnight. The beads were then drained and washed with MeOH for 2 min (this washing was repeated 5 times) and briefly dried under vacuum. Ten equivalents of NaBH4 was dissolved in 1:1:1 MeOH:TEOF:NMP added to the resin, and the reaction vessel was again agitated for ~4 hrs. The beads were drained and washed with MeOH for 2 min 127

(this washing was repeated 5 times) and cleaved and treated as described in Section 2.8.4.1. 2.8.5 General Synthesis Tert-butyl-2,2’,2’’-(5-((2-tert-butoxy-2-oxoethoxy)carbonyl)benzene-1,2,3triyl)tris(oxy)triacetate (2.16): This product was isolated from a reaction between gallic acid 2.13 (2.00 g, 11.8 mmol) and t-butylbromoacetate 2.14 (6 mL, 41 mmol) in 55 mL of DMF at 60 degrees C in the presence of K2CO3 (9.8 g, 71 mmol).

Silca gel

chromatography in 2:1 hexanes:EtOAc eluted compound 2.16 which was then recrystalized from MeOH. A percent yield was not calculated for this compound.

1

H

NMR (d-DMSO): 1.40 (m, 36 H) 4.68 (s, 2H), 4.73 (s, 2H), 4.75 (s, 2H), 7.16 (s, 2H). 13

C NMR (CDCl3): 27.76, 61.38, 66.49, 69.45, 81.32, 82.16, 104.69, 109.52, 123.50,

141.68, 150.45, 164.73, 166.32, 167.13, 167.76.

HRMS (CI+) C31H46O13Na+1 m/z:

649.28307; calcd: 649.2835.

Tert-butyl-2,2’,2’’-(5-(methoxycarbonyl)benzene-1,2,3-triyl)-tris(oxy)triacetate (2.17): 3,4,5-trihydroxymethyl benzoate (0.5 g, 2.7 mmol) was dissolved in 10 mL DMF, and K2CO3 (2.25 g, 16.3 mmol) was added in one solid portion. After stirring this mixture for ~ 5 minutes, t-butylbromoacetate 2.14 (1.5 mL, 10.3 mmol) was delivered by syringe. The reaction flask was equipped with a condenser and stirred at 60 degrees C for 14 h. The mixture was then poured into H2O and extracted with EtOAc. The organic layer was dried with MgSO4, filtered, and subjected to rotary evaporation to give an amber residue (1.40 g, 2.66 mmol) that corresponded to product in a 98% yield. 1H NMR (CDCl3): 1.42 (m, 27H), 3.81 (s, 3H), 4.56 (s, 4H), 4.72 (s, 2H), 7.18, (s, 2H).

13

C NMR

(CDCl3): 28.02, 52.20, 66.81, 69.79, 81.63, 82.45, 105.00, 109.57, 124.64, 141.64, 128

150.75, 166.14, 167.52, 168.12. HRMS (CI+) C26H38O11Na+1 m/z: 549.23064; calcd: 549.2313.

2,2’,2’’-(5-carboxybenzene-1,2,3-triyl)tris(oxy)triacetic acid (2.18): This product was isolated from an attempt to prepare 2.15 by selective ester hydrolysis. Trialkylated methyl ester 2.17 (0.5 g, 0.95 mmol) was dissolved in 8 mL THF. Lithium hydroxide (0.026 g, 1.1 mmol) was dissolved in 1 mL H2O and added dropwise to the reaction flask. The contents of the reaction flask were heated to reflux overnight. Starting material 2.17 was still present after 24 h at reflux. Two additional equivalents of lithium hydroxide were added as solid portions. After an additional 16 h at reflux, TLC analysis indicated consumption of starting material. The reaction mixture was acidified (< pH 5) with 2M HCl and extracted with EtOAc. The organic layer was dried with MgSO4, filtered, and subjected to rotary evaporation to afford a white solid that corresponded to tetra-acid 2.18. A percent yield was not calculated for the formation of this product. 1H NMR (dDMSO): 4.65 (s, 2H), 4.77 (s, 4H), 7.12 (s, 2H).

13

C NMR (CD3OD): 66.89, 70.15,

110.49, 127.32, 142.46, 152.49, 168.76, 172.07, 172.93.

Benzyl-3,4,5-trihydroxybenzoate (2.20):

Gallic acid 2.13 (0.20 g, 1.2 mmol) was

dissolved in 7 mL DMF. Hunig’s base (0.25 mL, 1.3 mmol) was added followed by addition of benzyl iodide (0.15 mL, 1.3 mmol). The contents of the reaction flask were stirred at room temperature overnight. H2O was then added, and the resulting mixture was extracted with EtOAc. The organic layer was collected, and dried with MgSO4, filtered, and its volume was reduced on a rotary evaporator. The crude product was subjected to silica gel chromatography in 4:1 to 1:1 hexanes:EtOAc to give 0.05 g (0.20 mmol) of compound 2.20 for a 17% yield. 1H NMR (CD3OD): 5.14 (s, 2H), 6.97 (s, 2H), 129

7.18-7.30 (m, 5H). HRMS (CI+) C14H13O5+1 m/z: 261.07575; calcd. 261.0760. Benzyl4-(benzyloxy)-3,5-dihydroxybenzoate (2.21) was also isolated from this reaction. 1H NMR (CD3OD): 5.03 (s, 2H), 5.17 (s, 2H), 6.95 (s, 2H), 7.16-7.32 (m, 8H), 7.39 (d, 2H). HRMS (CI+) C21H19O5+1 m/z: 351.12270; calcd. 351.1232.

Tert-butyl-2,2’,2’’-(4-formylbenzene-1,2,3-triyl)tris(oxy)triacetate (2.22): Commercially available 2,3,4-trihydroxybenzaldehyde (0.25 g, 1.6 mmol) was dissolved in 10 mL DMF to give a dark brown solution. K2CO3 (1.33 g, 9.6 mmol) was added in one solid portion, whereupon the heterogeneous mixture became bluish green. After stirring at room temperature for several minutes, t-butylbromoacetate 2.14 (0.8 mL, 5.5 mmol) was added dropwise via syringe. The reaction mixture was heated to 60 degrees C for 12 h. After cooling, the reaction mixture was poured into H2O and extracted with EtOAc. The organic layer was dried with MgSO4, filtered, and reduced on a rotary evaporator to 0.69 g (1.4 mmol) of a brown oil corresponding to 2.22 in an 86% yield. 1H NMR (CDCl3): 1.45 (m, 27H), 4.59 (s, 4H), 4.90 (s, 2H), 6.55 (d, 1H), 7.57 (d, 1H), 10.45 (s, 1H). 13C NMR (CDCl3): 27.53, 31.12, 36.22, 65.66, 70.04, 70.41, 81.88, 82.71, 107.77, 123.21, 123.75, 139.15, 154.45, 156.26, 162.25, 166.50, 167.62, 168.03, 189.32. HRMS (CI+) C25H37O10 m/z: 497.2385; calcd: 497.2387.

2,3,4-tris(2-tert-butoxy-2-oxoethoxy)benzoic acid (2.23): Aldehyde 2.22 (12.8 g, 25.8 mmol) was dissolved in 40 mL CH3CN. A separate solution of NaClO2 (6.59 g) was prepared in 70 mL H2O and slowly added to the solution of 2.22. After stirring at room temperature for 14 h, the reaction mixture was acidified with 2 M HCl and extracted into EtOAc. The organic layer was dried with MgSO4, filtered, and reduced on a rotary evaporator. Silica gel chromatography was performed in 5:1-1:1 hexanes:EtOAc, and 3.6 130

g (7.0 mmol) of white residue corresponding to 2.23 was recovered for 27% yield. 1H NMR (CDCl3): 4.40 (s, 4H), 4.82 (s, 2H), 6.41 (d, 1H), 7.49 (d, 1H). 13C NMR (CDCl3): 27.31, 65.19, 69.73, 70.68, 81.36, 82.14, 82.69, 107.51, 115.62, 127.56, 138.40, 150.40, 154.54, 164.53, 166.13, 167.24, 168.54. HRMS (CI+) C25H37O11 m/z: 513.2334; calcd: 513.2336. Formation of phenolic Dakin product 2.24 was also observed in oxidation of 2.22 using Oxone as the oxidant. 1H NMR (CDCl3): 1.45 (m, 27H), 4.51 (s, 2H), 4.61 (s, 2H), 4.68 (s, 2H), 6.52 (d, 1H), 6.76 (d, 1H), 8.24 (s, 1H).

Tert-butyl-2,2’,2’’-(4-(((6,7-dimethoxy-2-oxo-2H-chromen-4yl)methoxy)carbonyl)benzene-1,2,3-triyl)tris(oxy)triacetate (2.35): 2.23 (0.23 g, 0.45 mmol) was dissolved in 5 mL DMF. K2CO3 (0.17 g, 1.2 mmol) was added to this solution in one solid portion.

Commercially available 4-bromomethyl-6,7-

dimethoxycoumarin (0.13, 0.45 mmol) was then added as a solid, and the heterogeneous reaction mixture was stirred at room temperature for 8 h. The reaction mixture was washed with H2O and extracted with EtOAc. The organic layer was dried with MgSO4, filtered, and reduced on a rotary evaporator. The crude product was purified by silica gel chromatography using a gradient of 2:1 to 1:1 hexanes:EtOAc as the eluant. Compound 2.35 was isolated as a solid in a 60% yield (0.20 g, 0.27 mmol). 1H NMR (CDCl3): 1.341.41 (m, 27H), 3.84 (s, 3H), 3.87 (s, 3H), 4.52 (s, 2H), 4.54 (s, 2H), 4.65 (s, 2H), 5.39 (s, 2H), 6.39 (s, 1H), 6.49 (d, 1H), 6.79 (s, 1H), 6.87 (s, 1H), 7.53 (d, 1H).

131

2.9 REFERENCES

1

“A Theory of the Structure and Process of Formation of Antibodies” Pauling, L. J. Am. Chem. Soc. 1940, 62, 2643-2657.

2

(a.) “Molecular Imprinting in Cross-Linked Materials with the Aid of Molecular Templates—A Way Towards Artificial Antibodies” Wulff, G. Angew. Chem. Int. Ed. 1995, 34, 1812-1832. (b.) “Molecularly Imprinted Polymers and Their Use in Biomimetic Sensors” Haupt, K.; Mosbach, K. Chem. Rev. 2000, 100, 2495-2504.

3

“Dynamic Combinatorial Chemistry” Corbett, P.T.; Leclaire, J.; Vial, L.; West, K.R.; Wietor, J.; Sanders, J.K.M.; Otto, S. Chem. Rev. 2006, 106, 3652-3711.

4

“Dynamic Combinatorial Chemistry and Virtual Combinatorial Libraries” Lehn, J.M. Chem. Eur. J. 1999, 5, 2455-2463.

5

“Synthetic Hosts via Molecular Imprinting—Are Universal Synthetic Antibodies Realistically Possible?” Zimmerman, S.C.; Lemcoff, N.G. Chem. Commun. 2004, 5-14.

6

“Drug Assay Using Antibody Mimics Made by Molecular Imprinting” Vlatakis, G.; Andersson, L.I.; Muller, R.; Mosbach, K. Nature 1993, 361, 645-647.

7

“The Synthesis, Characterization and Molecular Recognition Properties of Imprinted Microgels” Biffis, A.; Graham, N.B.; Siedlaczek, G.; Stalberg, S.; Wulff, G. Macromol. Chem. Phys. 2001, 202, 163-171.

8

(a.) “Regioselective Introduction of Two Boronic Acid Groups into [60]fullerene Using Saccharides as Imprinting Templates” Ishi-i, T.; Nakashima, K.; Shinkai, S. Chem. Commun. 1998, 1047-1048. (b.) “D/L Selective Re-binding of SaccharideImprinted [60]Fullerene-Bisadducts Based on a Saccharide-Boronic Acid Interaction: Development of a Molecular Imprinting Technique Useful in a Homogeneous System” Ishi-i, T.; Iguchi, R.; Shinkai, S. Tetrahedron, 1999, 5, 3883-3892.

132

9

(a.) “Synthetic Hosts by Monomolecular Imprinting Inside Dendrimers” Zimmerman, S.C.; Wendland, M.S.; Rakow, N.A.; Zharov, I.; Suslick, K.S. Nature 2002, 418, 399-403. (b.) “Molecular Imprinting Inside Dendrimers” Zimmerman, S.C.; Zharov, I.; Wendland, M.S.; Rakow, N.A.; Suslick, K.S. J. Am. Chem. Soc. 2003, 125, 13504-13518. (c.) “Creating a Good Impression” Haupt, K. Nat. Biotechnol. 2002, 20, 884-885.

10

“Formation of Macrocycles via Ring-Closing Olefin Metathesis” Lee, C.W.; Grubbs, R.H. J. Org. Chem. 2001, 66, 7155-7158.

11

“Quantitative Ring-Closing Metathesis of Polyolefins” Coates, G.W.; Grubbs, R.H. J. Am. Chem. Soc. 1996, 118, 229-230.

12

“Synthesis of Cored Dendrimers” Wendland, M.S.; Zimmerman, S.C. J. Am. Chem. Soc. 1999, 121, 1389-1390.

13

“Chromo- and Fluororeactands: Indicators for Detection of Neutral Analytes by Using Reversible Covalent-Bond Chemistry” Mohr, G.J. Chem. Eur. J. 2004, 10, 10821090.

14

“Cross-Linked Dendrimer Hosts Containing Reporter Groups for Amine Guests” Mertz, E.; Zimmerman, S.C. J. Am. Chem. Soc. 2003, 125, 3424-3425.

15

“Kinetics and Thermodynamics of Amine and Diamine Signaling by a Trifluoroacetyl Azobenzene Reporter Group” Mertz, E.; Beil, J.B.; Zimmerman, S.C. Org. Lett. 2003, 5, 3127-3130.

16

“Synthesis of Cored Dendrimers with Internal Cross-Links” Schultz, L.G.; Zhao, Y.; Zimmerman, S.C. Angew. Chem. Int. Ed. 2001, 40, 1962-1966.

17

“Poly(amidoamine) (PAMAM) Dendrimers: From Biomimicry to Drug Delivery and Biomedical Applications” Esfand, R.; Tomalia, D.A. Drug Discovery Today 2001, 6, 427-436.

18

“A Differential Array of Metalated Synthetic Receptors for the Analysis of Tripeptide Mixtures” Wright, A.T.; Anslyn, E.V.; McDevitt, J.T. J. Am. Chem. Soc. 2005, 127, 17405-17411.

19

“Differential Receptors Create Patterns That Distinguish Various Proteins” Wright, A.T.; Griffin, M.J.; Zhong, Z.; McCleskey, S.C.; Anslyn, E.V.; McDevitt, J.T. Angew. Chem. Int. Ed. 2005, 44, 6375-6378.

133

20

“The ‘One-Bead-One-Compound’ Combinatorial Library Method” Lam, K.S.; Lebl, M.; Krchank, V. Chem. Rev. 1997, 97, 411-448.

21

“Cooperative Metal-Coordination and Ion Pairing in Tripeptide Recognition” Wright, A.T.; Anslyn, E.V. Org. Lett. 2004, 6, 1341-1344.

22

(a.) “Teaching Old Indicators New Tricks” Wiskur, S.L.; Ait-Haddou, H.; Lavigne, J.J.; Anslyn, E.V. Acc. Chem. Res. 2001, 34, 963-972. (b.) “Indicator-Displacement Assays” Nguyen, B.T.; Anslyn, E.V. Coord. Chem. Rev. 2006, 250, 3118-3127. (c.) “Supramolecular Analytical Chemistry” Anslyn, E.V. J. Org. Chem. 2007, 72, 687-699.

23

“Determination of the Intrinsic Proton Binding Constants for Poly(amidoamine) Dendrimers via Potentiometric pH Titration” Niu, Y.; Sun, L.; Crooks, R.M. Macromolecules 2003, 36, 5725-5731.

24

“Comparison of Nitrogen Core and Ethylenediamine Core Starburst Dendrimers Through Photochemical and Spectroscopic Probes” Jockusch, S.; Ramirez, J.; Sanghvi, K.; Nociti, R.; Turro, N.; Tomalia, D.A. Macromolecules 1999, 32, 4419-4423.

25

“Pyrocatechol Violet in Spectrophotometric and Novel Optical Methods” Ivanov, V.M.; Kochelayeva, G.A. Russ. Chem. Rev. 2006, 75, 255-266.

26

“Spectral Properties of the Prototopic Forms of Fluorescein in Aqueous Solution” Klonis, N.; Sawyer, W.H. J. Fluor. 1996, 6, 147-157.

27

Haugland, R.P. Handbook of Fluorescent Probes and Research Chemicals, 6th Ed. (Molecular Probes, Eugene, Oregon 1996).

28

“The Dendritic Box: Shape-Selective Liberation of Encapsulated Guests” Jansen, J.F.G.A.; Meijer, E.W.; de Brabander-van den Berg, E.M.M. J. Am. Chem. Soc. 1995, 117, 4417-4418.

29

“Amide Bond Formation and Peptide Coupling” Montalbetti, C.A.G.N.; Falque, V. Tetrahedron 2005, 61, 10827-10852.

30

“Spectroscopic Investigations of Polyamido Amine Starburst Dendrimers Using the Solvatochromic Probe Phenol Blue” Egger-Richter, D.L.; Landry, J.C.; Tesfai, A.; Tucker, S.A. J. Phys. Chem. A 2001, 105, 6826-6833.

31

(A.) “Preparation and Utilization of Fluorescent Synthetic Peptides” Chersi, A.; Sezzi, M.L.; Romano, T.F.; Evangelista, M.; Nista, A. Biochim. Biophys. Acta 1990, 134

1034, 333-336. (B.) “Extending the Applicability of Carboxyfluorescein in SolidPhase Synthesis” Fischer, R.; Mader, O.; Jung, G.; Brock, R. Bioconjugate Chem. 2003, 14, 653-660. (C.) “A Fast and Inexpensive Method for N-Terminal Fluorescein-Labeling of Peptides” Weber, P.J.A.; Bader, J.E.; Folkers, G.; BeckSickinger, A.G. Bioorg. Med. Chem. Lett. 1998, 8, 597-600. 32

“A New Strategy for the Synthesis of Pendant Benzodiazacoronands and Their Use as Components of Chromatographic Stationary Phases” Piatek, P.; Gryko, D.T.; Szumna, A.; Jurczak, J. Tetrahedron 2004, 60, 5769-5776.

33

“Spirocyclic Nonpeptide Glycoprotein IIb-IIIa Antagonists. Part 3: Synthesis and SAR of Potent and Specific 2,8-Diazaspiro[4.5]decanes” Mehrotra, M.M.; Heath, J.A.; Rose, J.W.; Smyth, M.S.; Seroogy, J.; Volkots, D.L.; Ruhter, G.; Schotten, T.; Alaimo, L.; Park, G.; Pandey, A.; Scarborough, R.M. Bioorg. Med. Chem. Lett. 2002, 12, 1103-1107.

34

Greene, T.W.; Wuts, P.G.M. Protective Groups in Organic Synthesis, 3rd Ed. (Wiley, New York 1999).

35

“An Enantiomeric Nanoscale Architecture Obtained from a Pseudoenantiomeric Aggregate: Covalent Fixation of Helical Chirality Formed in Self-Assembled Discotic Triazine Triamides by Chiral Amplification” Ishi-i, T.; Kuwahara, R.; Takata, A.; Jeong, Y.; Sakurai, K.; Mataka, S. Chem. Eur. J. 2006, 12, 763-776.

36

“Oxidation of Aromatic Aldehydes Using Oxone” Gandhari, R.; Maddukuri, P.P.; Vinod, T.K. J. Chem. Ed. 2007, 84, 852-854.

37

“Cleavages and Rearrangements Involving Oxygen Radicals and Cations” Leffler, J.E. Chem. Rev. 1949, 45, 385-417.

38

“Facile Oxidation of Aldehydes to Acids and Esters with Oxone” Travis, B.R.; Sivakumar, M.; Hollist, G.O.; Borhan, B. Org. Lett. 2003, 5, 1031-1034.

39

“Efficient and Rapid Method for the Oxidation of Electron-Rich Aromatic Aldehydes to Carboxylic Acids Using Improved Basic Hydrogen Peroxide” Cong, Z.; Wang, C.; Chen, T.; Yin, B. Synth. Commun. 2006, 36, 679-683.

40

“Surfactant Assisted Permanganate Oxidation of Aromatic Compounds” Jursic, B. Can. J. Chem. 1989, 67, 1381-1383.

41

“Sodium Chlorite-Hydrogen Peroxide – A Mild and Selective Reagent for the Oxidation of Aldehydes to Carboxylic Acids” Raach, A.; Reiser, O. J. Prakt. Chem. 2000, 342, 605-608. 135

42

(A.) “A Colorimetric Boronic Acid Based Sensing Ensemble for Carboxy and Phospho Sugars” Zhang, T.; Anslyn E.V. Org. Lett. 2006, 8, 1649-1652. (B.) “Using an Indicator Displacement Assay to Monitor Glucose Oxidase Activity in Blood Serum” Zhang, T.; Anslyn, E.V. Org. Lett. 2007, 9, 1627-1629.

43

(A.) “Boronic Acid Based Peptidic Receptors for Pattern-Based Saccharide Sensing in Neutral Aqueous Media, an Application in Real-Life Samples” Edwards, N.Y.; Sager, T.W.; McDevitt, J.T.; Anslyn, E.V. J. Am. Chem. Soc. 2007, 129, 1357513583. (B.) “Solid-Phase Library Synthesis, Screening, and Selection of TightBinding Reduced Peptide Bond Inhibitors of a Recombinant Leishmania mexicana Cysteine Protease B” J. Med. Chem. 2002, 45, 1971-1982.

44

(A.) “The Deprotection of Lys(Mtt) Revisited” Bourel, L.; Carion, O.; Masse-Gras, H.; Melnyk, O. J. Peptide Sci. 2000, 6, 264-270. (B.) “The Kinetics of the Removal of the N-methyltrityl (Mtt) Group during the Synthesis of Branched Peptides” Li, D.; Elbert, D.L. J. Peptide Res. 2002, 60, 300-303.

45

“Fluorescence Resonance Energy Transfer” Clegg, R.M. Curr. Opin. Biotechnol. 1995, 6, 103-110.

46

“Practical Design Criteria for a Dynamic Ratio Imaging System” Tsein, R.Y.; Harootunian, A.T. Cell Calcium 1990, 11, 93-109.

47

“Encapsulating Fluorescein Using Adipic Acid Self-Assembly on the Surface of PPI-3 Dendrimer” Chai, M.; Holley, A.K.; Kruskamp, M. Chem. Commun. 2007, 168170.

48

“Nature of Cationic Poly(propylenimine) Dendrimers in Aqueous Solutions as Studied Using Versatile Indicator Dyes” Mchedlov-Petrossyan, N.O.; Bryleva, E.Y.; Vodolazkaya, N.A.; Dissanayake, A.A.; Ford, W.T. Langmuir 2008, 24, 56895699.

49

(A.) “Fluorescence Ratio Imaging of Cyclic AMP in Single Cells” Adams, S.R.; Harootunian, A.T.; Buechler, Y.J.; Taylor, S.S.; Tsein, R.Y. Nature, 1991, 349, 694-697. (B.) “Recent Advances in the Design of Small Molecule-Based FRET Sensors for Cell Biology” Kikuchi, K.; Takakusa, H.; Nagano, T. Trends Anal. Chem. 2004, 23, 407-415.

50

“Selective Amine Recognition: Development of a Chemosensor for Dopamine and Norepinephrine” Secor, K.E.; Glass, T.E. Org. Lett. 2004, 6, 3727-3730.

51

(A.) “Coumarin-Fluorescein Pair as a New Donor-Acceptor Set for Fluorescence Energy Transfer Study of DNA” Mitsui, T.; Nakano, H.; Yamana, K. Tetrahedron 136

Lett. 2000, 41, 2605-2608. (B.) “A Peptide-Cyclodextrin Hybrid System Capable of Detecting Guest Molecules Utilizing Fluorescence Resonance Energy Transfer” Macromol. Rapid Commun. 2004, 25, 577-581. 52

“Protection of 3,4-Dihydroxyphenylalanine (DOPA) for Fmoc Solid-Phase Peptide Synthesis” Hu, B.; Messersmith, P.B. Tetrahedron Lett. 2000, 41, 5795-5798.

53

“Fluorescent Dyes of the Esculetin and Alizarin Families Respond to Zinc Ions Ratiometrically” Zhang, L.; Dong, S.; Zhu, L. Chem. Commun. 2007, 1891-1893.

54

(A.) “Intramolecular Energy Transfer from Phenanthrene to Europium in Aqueous Micellar Solution” Darwent, J.R.; Flint, C.D.; Sharpe, N.W. J. Chem. Soc., Chem. Commun. 1988, 747-748. (B.) “A Pattern Recognition Based Fluorescence Quenching Assay for the Detection and Identification of Nitrated Explosive Analytes” Hughes, A.D.; Glenn, I.C.; Patrick, A.D.; Ellington, A.; Anslyn, E.V. Chem. Eur. J. 2008, 14, 1822-1827.

55

“GFP-Based FRET Analysis in Live Cells” Takanishi, C.L.; Bykova, E.A.; Cheng, W.; Zheng, J. Brain Res. 2006, 1091, 132-139.

56

“Cross-Reactive Conjugated Polymers: Analyte-Specific Aggregative Response for Structurally Similar Diamines” Nelson, T.L.; O’Sullivan, C.; Greene, N.T.; Maynor, M.S.; Lavigne, J.J. J. Am. Chem. Soc. 2006, 128, 5640-5641.

57

“Colorimetric Enantiodiscrimination of α-Amino Acids in Protic Media” FolmerAndersen, J.F.; Lynch, V.M.; Anslyn, E.V. J. Am. Chem. Soc. 2005, 127, 79867987.

58

“Pattern-Based Discrimination of Enantiomeric and Structurally Similar Amino Acids: An Optical Mimic of the Mammalian Taste Response” Folmer-Andersen, J.F.; Kitamura, M.; Anslyn, E.V. J. Am. Chem. Soc. 2006, 128, 5652-5653.

59

(A.) “Characterization of Starburst Dendrimers by the EPR Technique. Copper(II) Ions Binding Full Generation Dendrimers” Ottaviani, M.F.; Montalti, F.; Turro, N.J.; Tomalia, D.A. J. Phys. Chem. B 1997, 101, 158-166. (B.) Poly(amidoamine) Dendrimers: A New Class of High Capacity Chelating Agents for Cu(II) Ions” Diallo, M.S.; Balogh, L.; Shafagati, A.; Johnson, J.H.; Goddard, W.A.; Tomalia, D.A. Enviro. Sci. Technol. 1999, 33, 820-824.

60

“Preparation of Cu Nanoclusters within Dendrimer Templates” Zhao, M.; Sun, L.; Crooks, R.M. J. Am. Chem. Soc. 1998, 120, 4877-4878.

137

Chapter 3: Evaluation of Indicator Dyes for Use in a Novel Enantioselective Indicator Displacement Assay 3.1 INTRODUCTION AND SCOPE The chapter details progress in the development of a colorimetric assay for ee of molecules containing common organic functional groups. The operational design of the proposed assay is intended to facilitate its future transition to high-throughput screening protocols. As a result, this chapter highlights the need for improved high-throughput screening methods. The method for rapid determination of ee proposed in this chapter represents a hybrid approach, combining an existing strategy for the colorimetric sensing of amines with an established enantiosensing platform. Because amines were an initial chiral target of the proposed assay, a brief history of their colorimetric discrimination is provided. A survey of optical enantiosensing strategies is also included, with special consideration given to colorimetric methods. Finally, progress towards the development of the proposed assay is described through an evaluation of the suitability of three compounds responsible for the colorimetric nature of the proposed assay. 3.2 COLORIMETRIC SENSING AND DISCRIMINATION OF AMINES The development of reliable methods for the accurate determination of ee of organic analytes has become an active area of research. Of these organic compounds, amines represent attractive “real-world” targets because their extensive use in manufacturing contributes to their status as environmental pollutants.1 Biogenic amines also play important physiological roles in regulating sleep2, emotion3, and body temperature,4 and they have been implicated in conditions such as cancer5, bacterial infection6, and food poisoning.7,8 138

The detection of amines also presents a unique design challenge to the supramolecular chemist. There is a dearth of classical hosts for uncharged amines. As neutral molecules, amines lack the electrostatic handle exploited in ion-pairing and hydrogen-bonding driven molecular recognition.

Indeed, the classical approaches

highlighted below rely on recognition of the amine’s positively charged ammonium form. New sensing strategies, however, have emerged to signal the presence of neutral amines. These strategies include array-based sensing approaches and the use of covalent bond formation. 3.2.1 Classical Approaches One strategy for the colorimetric detection of amines exploits established supramolecular host-guest chemistry. In this approach, the non-covalent association of a molecular guest modulates the absorbance properties of the host molecule (see section 1.2.2.1). This change in the absorbance profile is the basis of the colorimetric response, and several research groups have utilized these guest-induced changes in optical properties to develop colorimetric chemosensors for various amine-containing organic analytes. Because these systems rely on the rational design of discrete molecules for the selective recognition and sensing of specific targets, they are considered “classical” chemosensors.

The classical nature of this approach is evidenced by its use of

established supramolecular constructs such as crown ethers and cyclodextrins. Fuji reports the use of a modified phenolphthalein receptor that develops color in the presence of selected diamines and triamines.9 Diamines form a 2:1 complex with the dianion of receptor 3.1. One equivalent of diamine guest bridges the appended crown ether moieties, while the other equivalent acts as the counterion for the carboxylate generated from the γ-lactone (Scheme 3.1). Because the color change depends on the length of the diamine, receptor 3.1 provides direct visualization of molecular length based 139

on the selective recognition of diamines. The most pronounced color changes were observed for terminal diamines containing 8 and 9 carbons, while diamines shorter than 1,5-diaminopentane did not result in any appreciable coloration. Triamines such as spermidine formed a 1:1 complex with the dianion of receptor 3.1 because the additional internal imino group interacts with the carboxylate upon association. Receptor 3.1 also effectively discriminated between triamines on the basis of color development. As with diamine guests, the extent of coloration was proportional to the product of the association constants and absorptivity coefficients of the complexes.

3.1 Scheme 3.1 Colorimetric discrimination of molecular length. Receptor 3.1 visually discriminates terminal diamines on the basis of length. Experimental evidence suggests a host:diamine guest ratio of 1:2, with one equivalent of diamine bridging the crown ether units and a second equivalent serving as a counterion to the incipient carboxylate. Colorimetric versions of the Hinsberg test10 that allow for discrimination of primary, secondary, and tertiary amines are also described.

An azophenol dye

functionalized with a crown ether and a permethylated cyclodextrin (3.2) provides discrimination of these classes of amines.11 While the addition of primary and secondary amines resulted in bathochromic shifts of 195 nm and 145 nm, addition of tertiary amines 140

did not effect a change in color. The bathochromic shifts observed for primary and secondary amines is attributed to their deprotonation of the azophenol. Analysis of structurally impaired analogs of 3.2 proved that both the appended cyclodextrin and crown ether groups are integral to the selective colorimetric discrimination of these amine families. This selectivity is based on the relative hydrogen-bonding abilities and steric demands of the amine guests. Hydrophobic forces between the cyclodextrin and the lipophilic tails of the investigated amines also contribute to the observed selectivity and consequent colorimetric response. O O

O

O O

O O

O OO

O O O O

O

O

O

O O O O O

O

O

O O O

O

O

O O

O

O

O O OH

O

O N N

NO2

3.2 3.2.2. Array-based Approaches The aforementioned classical approaches to molecular recognition and colorimetric sensing often adhere to a “lock and key” model. The success of 141

chemosensing protocols based on this model often depends on the highly selective association of a specific, targeted analyte. This model, therefore, demands some level of rational design to engineer the requisite selectivities necessary for successful discrimination; as a corollary, it requires the individual design and synthesis of a discrete, specific receptor for each analyte targeted for detection. As a result, classical approaches are not readily amenable to the analysis and classification of complex mixtures composed of multiple analytes. Array-based sensing protocols have emerged to address these shortcomings.12,13 Inspired by the mammalian senses of taste and smell, array-based molecular recognition and discrimination actually benefits from the employed receptors’ lack of selectivity. The synthetic receptors incorporated into the array are inherently cross-reactive with a number of potential analytes; and the selected receptors should ideally comprise a wide swath of chemical diversity to effect the array’s response to a large number of analytes,. A promiscuous response of the elements of the sensor array ensures a composite signal (a “fingerprint”) that can be interpreted by pattern-recognition protocols.14 Such an approach transforms a major weakness of classical supramolecular sensors – competition or interference from structurally similar analytes – into an advantage as responses from multiple analytes and natural interferants, both known and unknown, contribute to the array’s composite signal. Suslick and co-workers developed a colorimetric array capable of discriminating various amines.15 They incorporated four families of responsive dyes that would probe a large cross-section of chemical diversity.

Metallated porphyrins were used to

differentiate the amine analytes on the basis of metal-selective coordination; bis-pocketed zinc porphyrins were incorporated to provide size and/or shape-based discrimination; pH indicators provided differentiation via Bronsted basicity; and solvatochromic dyes offered an indication of the polarity of the analyte. 142

Upon exposure to the array, a series of 12 amines comprising various molecular shapes (linear, branched, cyclic) each provided unique, colorimetric composite responses that were distinguishable by eye. Further processing by pattern recognition protocols such as principal component analysis (PCA) and hierarchical cluster analysis (HCA) documented the array’s ability to resolve the investigated amines based on their structural and electronic properties.

HCA effectively segregated the branched amines, linear

amines, and cyclic amines on the basis of their qualitative structures. The single aromatic amine analyte (pyridine) was also separated from all other clusters produced by HCA. The power of this technique is demonstrated in its ability to resolve very similar compounds such as structural isomers. The authors assert that such a high level of discrimination is the result of the wide range of “chemical space” being probed by the array. Molecularly imprinted polymers16 (MIPs) have also been exploited in array-based sensing formats targeting amines. MIPs are crosslinked polymers formed in the presence of a template molecule.

Subsequent removal of the template creates a cavity of

complementary shape and size to the template molecule, generating a macromolecular receptor ideally capable of rebinding the templated analyte with a very high specificity (see Section 2.2). In practice, however, MIPs often suffer from high levels of cross reactivity. While this cross-reactivity is undesirable when the MIP is used in isolation, such molecular promiscuity renders MIPs attractive candidates for integration into sensor arrays. MIPs, however, lack inherent signaling elements, and an initial report on their use in sensor arrays relied on the UV absorbance of the unbound analyte.17 Shimizu later integrated a colorimetric response to an MIP sensor array by using a dye-displacement strategy.18 The array consisted of seven molecularly imprinted polymers, each templated 143

to seven structurally similar amines. The colorimetric response to the amine analytes was generated by the displacement of a benzofurazan-based amine dye from the polymer upon amine binding.

The resulting response pattern for each amine analyte was

evaluated using linear discriminant analysis, and the amines were classified with 94% accuracy. The differentiated amines included diastereomers and compounds differing only by a methyl group. The classification accuracy is comparable to that obtained in the earlier report that relied on direct observation of binding of the analyte17; this agreement suggests that, despite being an indirect assessment, indicator displacement measurements provide accurate quantification of binding efficiency. A final important contribution of this report is the successful classification of a non-templated amine analyte. This success expands the scope of potential analytes beyond those specifically templated for MIPbased sensor arrays. 3.2.3 Amine Sensing Based on Covalent Bond Formation Sensor designs predicated on covalent bond formation are also prevalent in the literature.19 This design is particularly useful in the detection of neutral organic analytes where the inherent Lewis basicity or acidity of the analyte is exploited in a chemical reaction with functionalized dyes. The resulting covalent bond formation between the analyte and the dye affects the dye’s electronic properties and thereby promotes a change in its optical properties (Figure 3.1). Therefore, the selectivity and sensitivity of these sensors is governed by their different chemical reactivity towards competing analytes and can therefore be tuned by introduction of specific substituents.20

Sensors based on

covalent bond formation have been developed for a number of analytes, including hydrazines21, saccharides22, aldehydes23, alcohols24, amino acids25, and cyanide.26 This type of sensing strategy does have some caveats, however, as these sensors are prone to cross-reactivity and are often designed to merely signal families of compounds. 144

Figure 3.1 Colorimetric molecular sensing using covalent bond formation. Reversible covalent bond formation between a chromogenic indicator and targeted analyte produces a change in the absorbance properties of the dye. This strategy has been employed in the detection of various families of organic compounds. Several groups report the selective colorimetric signaling of amines via covalent bond formation. For example, electrophilic indicator 3.3 serves as a chromogenic probe for primary amines. While no response is observed for secondary and tertiary amines, 3.3 suffers from cross-reactivity in the presence of an excess of amines and is not specific for any particular aliphatic primary amine. To address these problems, Torroba and Martinez-Manez covalently anchored 3.3 onto a mesoporous solid featuring nanoscopic hydrophobic pockets.27 These hydrophobic pockets are primitive analogues of those found in biological systems and might offer additional interactions that improve analyte selectivity. This proposed increase in selectivity was indeed observed for fairly lipophilic amines such as n-nonylamine and n-decylamine.

Small hydrophilic amines (e.g., n-

propylamine) and larger aliphatic compounds (e.g., n-dodecylamine) did not provide a colorimetric response. Cross-reactivity is also remedied by attachment to the inorganic support. The authors attribute this enhanced selectivity to the three-dimensional, solidstate architecture offered by the nanoscopic pores of the mesoporous solid.

145

N Cl NC

CN

3.3

3.4

Functionalization of the 3-position of a chlorin macrocycle with a trifluoroacetyl group yielded the first chlorophyll-based chemosensor (3.4).28 Trifluoroacetyl groups are known covalent sensing elements for amines and alcohols, and the formation of a hemiaminal upon reaction between 3.4 and n-butylamine resulted in visible color changes. This reaction took several hours in THF, but afforded a blue shift of 31 nm in the absorption maximum of the dye. An association constant of 7.8 x 103 M-1 was determined for this reaction in CDCl3, indicating the strong binding affinity of the trifluoroacetyl group. Ratiometric fluorescent sensing of n-butylamine was also possible using 3.4. 3.2.3.1 Trifluoroacetyl-based Chromoreactands Gerhard Mohr has also extensively used dyes modified with trifluoroacetyl groups in the colorimetric detection of amines. Because these dyes marry the utility of chemical reagents with the selectivity of ligands, Mohr coined the term “reactand”29 to describe these compounds. The most successful of his trifluoroacetyl reactands are based on azobenzene chromophores functionalized with terminal donor substituents.

The

trifluoroacetyl group is appended as a terminal electron-accepting group that is conjugated to the donating group via the azo chromophore. 146

Conversion of the

trifluoroacetyl group into a hemiaminal reduces its electron accepting ability and results in a hypsochromic shift in the reactand’s absorption spectrum. By embedding them in plasticized PVC membranes, Mohr extended the use of these trifluoroacetyl reactands into the realm of optical sensor layers.30 For example, exposure of a plasticized PVC sensor layer containing a bisazo trifluoroacetyl reactand underwent dramatic color changes upon exposure to amphetamine.31 The blue color associated with the reactand’s trifluoroacetyl form changes to a deep red color in the presence of this analyte. This fully reversible response was also noted for other aliphatic amines, although the sensor layer was much more sensitive to the presence of amphetamine and methamphetamine. Trifluoroacetyl reactands capable of amine detection have also been covalently fixed into copolymerized sensor materials32 and integrated into molecularly imprinted dendrimers.33 3.2.3.2 Tricyanovinyl-based Chromoreactands The remarkable electron-accepting ability of tricyanovinyl groups is manifested in its high Hammett substituent constant. Its σp value of 0.98 is significantly higher than that for trifluoroacetyl groups (σp=0.80).34 When conjugated to an election donor through an azobenzene linker, tricyanovinyl acceptors imbue the resulting dye with long wavelength absorption properties. Long wavelength absorption is desired for biological applications of chemosensors to avoid interference from the intrinsic absorption of the sample. The λmax of tricyanovinyl-based reactand 3.5 in CH3CN is 593 nm, but the absorption maximum red shifts to 642 nm when 3.5 is dispersed in plasticized PVC sensor layers. These absorption properties also make the probe compatible with cheap light sources and detectors.

147

CN C8H17

N N

N

C8H17

CN CN

3.5 The sensor layer composed of reactand 3.5 showed a decrease in its long wavelength absorption and an increase in absorption around 470 nm upon interaction with 1-propylamine.35 This spectral modulation is attributed to nucleophilic attack by the amine on the tricyanovinyl group (Scheme 3.2).

The presence of isosbestic points

confirms that the system is undergoing only one type of chemical reaction, and the proposed nucleophilic attack is supported by IR spectroscopy. Because this chemical reaction liberates hydrogen cyanide, however, the addition of 1-propylamine to 3.5 is irreversible.

This irreversibility stands in contrast to the dynamic reactivity of the

trifluoroacteyl-based reactands and obviates the use of tricyanovinyl-based reactands in continuous monitoring of amine concentration. Despite this limitation, tricyanovinylbased reactands are less promiscuous than those based on trifluoroacetyl groups. Whereas trifluoroacetyl groups react with various nucleophilic species (e.g., alcohols, thiols, phenols), tricyanovinyl derivatives only react with primary and secondary amines. The present study successfully quantified amine concentration through a kinetic evaluation of the absorption spectra of 3.5 as a function of time and analyte concentration. Mohr reports a detection limit of 0.1 mM for these sensor layers and suggests their future application in optical test strips for food quality.

148

R CN C8H17

N N

CN

N

C8H17

R NH2

CN

NH C8H17 C8H17

N N

N

+

CN CN

HCN

Scheme 3.2 Covalent bond formation between tricyanovinyl-based reactand 3.5 and a primary amine. Reaction with primary and secondary amines attenuates the electronaccepting power of the terminal group and changes the color of the reactand solution from blue to orange. This process is the basis of Mohr’s amine-responsive sensor layers. 3.3 OPTICAL ENANTIOSENSING Molecular chirality has fascinated chemists since Pasteur first separated the enantiomers of tartaric acid by hand.36 The ubiquity of chiral recognition in biological systems37 soon prompted interest in the underlying principles of this phenomenon, and abiotic chiral recognition emerged to elucidate these fundamentals.38,39 The formation of diastereomeric complexes between a chiral host and the enantiomers of a chiral guest is the basis of any enantioselective recognition event. The efficiency of the enantioselective complexation process is directly related to the energy difference between the diastereomeric transition states that lead to these complexes. These energy differences are often quite small; thus the challenge presented by chiral recognition is the control and manipulation of these thermodynamic disparities through steric interactions, structural complementarity, and molecular symmetry.40 Enantioselective molecular recognition processes provide the foundation for optical enantiosensors. In addition to their academic appeal, the design of practical enantiosensors is currently driven by the need for real-time analytical assays of enantiomeric excess (ee).41

Despite intense philosophical interest and developing

economic incentive, pragmatic examples in the field of enantioselective sensing remain somewhat rare. Like their achiral counterparts, enantiosensors can be fluorometric or 149

colorimetric in nature; and as with amine sensing and discrimination, new methods have recently evolved to complement classical single molecule-based sensor designs. These new strategies include enantioselective indicator displacement assays and the use of doped liquid crystals. 3.3.1 Fluorescent Methods42 The majority of optical enantiosensors reported in the literature are based on fluorescence, but they will be only be briefly highlighted here.

Lin Pu developed

enantioselective fluorescent sensors based on chiral bisbinaphthyl macrocycles.43 These sensors rely on the molecular recognition of α-hydroxycarboxylic acids and an enantioselective fluorescence enhancement of the macrocycle in the presence of these guests. Under certain conditions, Pu’s chiral macrocycle showed a 2-3-fold fluorescence enhancement in the presence of (S)-mandelic acid; negligible fluorescence enhancement was observed for the macrocycle upon treatment with the corresponding R enantiomer. Pu has explored the use of similar macrocycles in the determination of enantiomeric composition and has suggested their application in chiral catalyst screening.44 3.3.2 Colorimetric Methods While fluorescence-based methods generally benefit from increased sensitivity, colorimetric enantiosensors offer the potential for macroscopic expression of chirality. A direct visualization of the stereochemical outcome of a reaction would aid in the naked-eye determination of “hits” in a combinatorial library.41 3.3.2.1 Classical Single Molecule Enantiosensors Kubo reported a seminal example of colorimetric chiral recognition using chromogenic calixarene 3.6.45

An optically active binapthyl scaffold provides

enantiodiscrimination, while indophenol chromophores signal the recognition event. A 150

critical design feature is the unsymmetrical ether linker that chemically distinguishes the two appended chromophores. The result is a dual sensory mode in which the two chromophores are differentially affected upon binding a chiral substrate. Chemosensor 3.6 proved effective in the enantiosensing of amino alcohols:

addition of (R)-

phenylglycinol to 3.6 effected a visual color change not observed in the presence of 1000 equivalents of its enantiomer. Subsequent competition experiments confirmed that 3.6 was capable of signaling low concentrations of (R)-phenylglycinol even in the presence of excess S enantiomer, validating its status as a colorimetric enantiosensor. This work represents an early example of the translation of a chiral recognition event into a perceptible color change. Several examples of such enantioselective signal transduction have been since reported, including a chiral version of Fuji’s colorimetric sensor (see Section 3.2.1) that gives visual enantiomeric recognition of alanine derivatives.46

O

N

OH X HO

N

O

X= O

O O

O O

3.6

3.3.2.2 Liquid Crystals as Colorimetric Indicators of Chirality The induction of a phase change in liquid crystalline (LC) materials by chiral dopants has also been utilized in color-based assessments of ee. Non-racemic chiral dopants effect a transition from nematic to cholesteric liquid crystalline phases. The resulting phases reflect light at wavelengths that can be tuned to the visible spectrum 151

through judicious selection of dopant. These reflected wavelengths also depend on the pitch of the material; this property is, in turn, inversely proportional to the ee of the chiral dopant. The reflected wavelengths can therefore be related to the ee of the sample, and a visual read-out of stereochemical composition is possible when these wavelengths are in the visible range. The magnitude of the pitch is also dependent on the helical twisting power of the dopant. Helical twisting power is a measure of the intrinsic ability of the chiral molecule to provoke helical induction in the LC phase. A challenge associated with the use of liquid crystals in the colorimetric determination of ee is the low helical twisting powers of the common organic molecules envisioned as dopants. As a result, typical asymmetric reaction products may not provide the requisite phase transition.

Feringa initially

addressed this problem by functionalizing the chiral analytes with a mesogenic group bearing a structural resemblance to the LC materials.47

These derivatives induced

cholesteric textures that reflected light at increasingly longer wavelengths as the ee of the analytes decreased. This relationship between ee and the nature of the reflected light is the foundation of a visual chirality indicator based on a doped LC matrix. Feringa successfully extended this methodology to the evaluation of six phosphoramidite ligands in an asymmetric Michael addition.48

While this is a novel approach to chiral sensing,

this platform requires a prior derivatization step or the use of structurally contrived “benchmark” reagents for ee analysis. 3.3.2.3 Enantioselective Indicator Displacement Assays The Anslyn group has pioneered the use of enantioselective indicator displacement assays (eIDAs). These assays are conceptually similar to achiral IDA’s (see Section 2.4) but require chiral host molecules. Interaction of a chiral host molecule with a mixture of enantiomeric analytes establishes multiple solution equilibria leading to 152

the formation of two diastereomeric host-guest complexes.

Differential indicator

displacement by the chiral guests occurs because the equilibrium constants leading to each host-guest complex differ in magnitude (Scheme 3.3, equations (2) and (3)). The energetic differences in the diastereomeric host-guest (i.e., receptor-analyte) ensembles are consequently related to the observed absorbance of the solutions, and thus the color of the solutions can be correlated with stereochemical composition.

Scheme 3.3 Equilibria processes in an enantioselective indicator displacement assay (eIDA). H* = chiral host/receptor; I = indicator; GR, GS = analyte/guests; KI, KR, KS = association constants The first example of an eIDA was reported in 2004.49 In this study, a boronic acid-based host and colorimetric, diol-containing indicators were employed in the quantification of ee of α-hydroxyacids. The concentration and ee of phenyllactic acid was determined with respective accuracies of 10% and 20%. Concentration and ee values were determined in two absorption measurements; an advantage of this assay is that no calibration curve was needed for ee evaluation at the determined analyte concentration.

The subsequent use of a dual-chamber cuvette allowed both ee and

concentration to be ascertained by a single absorption measurement.50 An eIDA that successfully discriminated the enantiomers of 13 amino acids was also recently reported.51 153

3.4 THE NEED FOR IMPROVED HIGH THROUGHPUT SCREENING ASSAYS The facile and accurate determination of the stereochemical composition of reaction products is an economic necessity in the combinatorial development of enantioselective enzymes and asymmetric catalysts.41 Traditional methods such as chiral HPLC and GC are not amenable to such analyses because they only accommodate one sample at a time and require the physical separation of enantiomers. As a result, the determination of ee is generally the rate-limiting step in the evaluation of catalyst libraries. Several approaches have been described to address these shortcomings in an attempt to provide true high throughput screening (HTS) assays for ee. These methods include the use of fluorescent antibodies52, reaction microarrays53, enzymes54, and indicator displacement assays.55

Expansion and automation of chromatographic and

NMR-based methods have also been described.41,56 These recent reports highlight the connection between the emergence of combinatorial chemistry and the development of HTS assays. In fact, no HTS techniques for ee existed prior to 1997. None of these reported methods, however, offer a solution to the major limitation of combinatorial libraries—the lack of rapid, general HTS protocols for ee. Current strategies require derivatization of analytes, expensive chromatographic equipment, existing enzymatic reactions, or extensive calibration curves. Additionally, very few of these assays are colorimetric and therefore do not allow for “naked-eye” evaluation of stereochemical composition. The work described in the remainder of this chapter represents efforts toward the development of a colorimetric assay for ee of chiral analytes containing common organic functional groups. An advantage of our proposed method over current assays is its inherent simplicity.

Our approach obviates the need for extensive instrumentation,

laborious syntheses of complex receptors, and prior derivatization of analytes. Our 154

strategy’s proposed ease-of-use combined with an instant visual read-out of ee would make this method attractive in both HTS applications and general organic synthesis. The color-based nature of an enantioselective assay would also constitute a unique contribution to the field. Examples of the direct visualization of molecular chirality remain rare, and the development of enantioselective receptors and sensors has been a fundamental goal long before the advent of combinatorial technologies rendered HTS methods desirable. 3.5 RESEARCH DESIGN The immediate goal of this project is the enantioselective colorimetric signaling of chiral amines. Two major challenges must be addressed for this objective to be realized: (1) the assay must incorporate a colorimetric sensing mode for the analyte of interest, and (2) a chiral bias must be engineered into the system to achieve visual discrimination of enantiomers. The examples described earlier in this chapter provide potential strategies to surmount these challenges. Colorimetric signaling of amines has been accomplished through various strategies; colorimetric enantiosensing has also been realized using several sensing platforms. A successful approach to the goals of the current project relies on inspiration from these disparate strategies, and an appropriate starting point is the identification of a documented colorimetric sensing strategy that can be easily elaborated into an enantioselective assay for chiral compounds containing common organic functional groups such as amines. In evaluating potential strategies, it is important to remember that a successful platform will be amenable to HTS applications and accessible to all chemists.

With these considerations in mind, the possible

enantioselective detection and signaling modes can be narrowed down to those commensurate to our goals. 155

Preparation of the macrocyclic hosts found in classical molecular recognition requires considerable synthetic effort; the use of array-based sensors presumes specialized knowledge of pattern recognition protocols; and color tests based on liquid crystalline materials require an inefficient derivatization step. While these colorimetric sensing strategies might yet be applied in a high-throughput setting, it is unlikely they will ever find widespread adoption among chemists seeking facile determination of ee. Because of these technologies’ high barriers to adoption in the general chemistry community, the operational simplicity of IDAs casts them as attractive alternatives in the determination of ee. Furthermore, their use of commercially available indicators and synthetically accessible hosts significantly lowers the barrier to widespread adoption of these protocols. It is this facile implementation of IDAs that serves as an inspiration to the envisioned assay for chiral amines. In keeping with this goal of accessibility, the signaling mode will rely on reversible covalent bond formation. Mohr has already documented the signaling of amines via chemical reaction with indicator dyes containing tricyanovinyl groups.30 As a result, the synthesis of many of these indicators have been reported in the literature, while the modular nature of IDAs renders them potentially interchangeable. While this sensing mode’s failure to discriminate among different amines might be considered a weakness in some applications, it is preferred in the present assay. A promiscuous response to a wide variety of amines ensures the method’s amenability to a large number of amine analytes. Given the long-term goals of the project, we envisioned a modified IDA based on the reversible covalent bond formation between amine analytes and a tricyanovinyl-based indicator and a chiral “receptor.” The general strategy is depicted below:

156

Scheme 3.4 Equilibria processes in the proposed ee assay. I = indicator; GR,GS = analytes/guests; H* = chiral host/receptor; KR, KS, KR,H*, KS,H* = association constants. This process is represented schematically in Scheme 3.5. Equilibrium will be established between the tricyanovinyl-based indicator (I) and the targeted chiral analytes (GR and GS, see Scheme 3.4, equation (1)). Because no chiral influence is present, GR and GS should bind to I with equal affinities. A structural analogue of the indicator (host compound H*) will serve as the “receptor.” In analogy to indicator I, host H* will also feature an amine-reactive tricyanovinyl moiety; H* will also impart asymmetry to the assay through the direct attachment of a stereogenic center to the electrophilic carbon of its tricyanovinyl group. Upon the introduction of chiral host molecule H*, the solution equilibria will be perturbed as H* competes with I in binding the enantiomeric guests GR and GS. Interaction between the enantiomers and chiral host leads to diastereomeric complexes H*:GR and H*:GS (Scheme 3.4, equations (2) and (3). Because these complexes are formed with differing free energies, this process should effect the concentrations of each enantiomer available to bind with I, thereby modulating the amount of “free” I in solution and providing an ee dependent color change. It should be noted that this design is a modification of traditional IDAs in which a pre-equilbrium is established between a receptor and indicator before addition of guest analyte (Scheme 3.3). In contrast, the indicator I does not interact with H* in our 157

system; rather, I is alternatively binding the enantiomers of the chiral analyte.

A

schematic representation of the described equilbria processes is depicted in Scheme 3.5. Y NC

CN

RS

NH2

Y

CN

RR

NH2

HN RS

CN

CN

indicator

Z*

CN

HN RR

CN

+

+

Y

CN

HN RR

CN

Z*

CN

NC

CN

= chiral "host" H

Z*

CN

HN RS

CN

+

Y

CN

NC

CN

diastereomers

Scheme 3.5 Schematic representation of the proposed equilbria processes. Diastereomers are differentially formed upon introduction of a tricyanovinyl-based chiral host to an equilibrating mixture of indicator and enantiomeric amines. This process modulates the concentration of enantiomers available to react with the tricyanovinylbased indicator. The amount of free (blue) indicator in solution is regulated in a manner related to the stereochemical composition of the amine analyte. While the general reaction between amines and tricyanovinyl reactands is known to produce dramatic color changes, this covalent bond formation is reported to be irreversible, and previous use of this reactand in the quantification of amines relied on kinetic evaluation.35 The design and operation of the present assay, however, is predicated on dynamic reversibility between the indicator and enantiomeric analytes. Addition of exogeneous cyanide and modification of reaction conditions were expected to confer the requisite reversibility.

158

3.5.1 Tricyanovinyl Azo Dye Before elucidation of the conditions effecting reversibility, an appropriate indicator needed to be synthesized and its behavior with amines validated. Tricyanovinyl azo dye 3.11 remained attractive for the present study because of its known synthesis and large color changes in the presence of amines. While the synthesis of 3.11 has been previously reported57, installation of the tricyanovinyl unit proved more difficult than anticipated. The preferred method in the literature involves formal addition of hydrogen cyanide to the alkene followed by dehydrogenation (step iii in Scheme 3.6).58,59 Addition of HCN across the alkene was accomplished by addition of potassium cyanide followed by acetic acid. To limit the amount of hydrogen cyanide evolved, a roughly equimolar amount of cyanide was added to the malononitrile 3.9. The amount of added cyanide never exceeded 2.3 equivalents relative to the malononitrile. It is likely this precaution severely limited the efficacy of the reaction. Successful literature methods often employ large excesses of cyanide. One synthesis even used commercial hydrogen cyanide to provide 6 equivalents of cyanide.60 Even these conditions only gave a moderate yield of 62%, and the authors state that use of potassium cyanide resulted in an “impure product in low yield.”

159

NO2

NH2

NO2 i.

H

NH2

ii.

3.8

iv.

N

CN CN

3.10

3.9

Bu Bu

NC

CN

CN

3.7

CN

H

CN

H

O

iii.

CN

N N

3.11

CN CN

Scheme 3.6 Synthesis of tricyanovinyl azo dye 3.11. (i.) NCCH2CN, cat. AcOH/pyridine, EtOH, 76% (ii.) SnCl2, EtOAc, 57% (iii.) 1) KCN, DMF 2) AcOH 3) Pb(OAc)4, 14% (iv.) 1) NaNO2 / HCl 2) N,N-dibutylaniline, 13%. As per literature preparations, the cyanide addition product of step iii was not isolated. Rather, an oxidizing agent was added directly to the reaction mixture to effect dehydrogenation. Lead tetraacetate and N-bromosuccinimde (NBS) were investigated separately as oxidants in this transformation. Literature reports suggest the use of Nbromosuccinimide gives increased yields of the tricyanovinyl product relative to those observed with lead tetraacetate.60 While each oxidant proved equally effective in the present synthesis, the highest observed yield over the addition/oxidation steps was 14%. Again, this low yield is attributed to the use of only small excesses of potassium cyanide. Literature methods employing hydrogen cyanide and NBS in the addition-oxidation step provided overall yields of 40-52% for installation of a tricyanovinyl group onto a benzaldehyde derivative.60 In the current study, however, safety considerations curtailed the use of large excesses of potassium cyanide and obviated the use of hydrogen cyanide.

160

Installation of the tricyanovinyl group was also attempted via electrophilic aromatic substitution of acetanilide with tetracyanoethylene (3.12).

The N-acetyl

protecting group is necessary because formation of N-tricyanovinyl derivatives is a known reaction between unprotected anilines and tetracyanoethylene.60 Attempts to prepare the tricyanovinyl compound using this method proved ineffective, as only starting material was recovered. Condensation of malononitrile with acyl cyanides also provides entry to tricyanovinyl compounds. However, use of tributyltin cyanide61 as a cyanating agent of p-nitrobenzoyl chloride did not produce the desired acyl cyanide product, and the traditional preparation of p-nitrobenzoyl cyanide 3.13 has been reported as unsuccessful.62 NHAc

NC NC

NH2

CN CN

3.12

NO2

CN

NC

NC CN

3.10

O

3.13

Scheme 3.7 Alternative routes to the installation of the tricyanovinyl group. Synthesis of 3.10 was also pursued using tetracyanoethylene 3.12 and acyl cyanide 3.13. Despite these difficulties, tricyanovinyl azo reactand 3.11 was prepared in small amounts via standard azo coupling procedures (this step also suffered from poor yields as a result of the formation of numerous side products, including homo-coupled substitution product 3.14.) The behavior of 3.11 in the presence of primary amines such as t-butylamine was qualitatively confirmed. A solution of 3.11 in CH3CN went from deep blue to bright orange within seconds after addition of t-butylamine. This color change also took place upon addition of N,N-diisopropylethylamine, albeit over 15 161

minutes. This observation is surprising given the behavior reported by Sausen et al who observed no reaction between tertiary aliphatic amines and tricyanovinyl derivatives.60 However, Mohr reports a small response of his tricyanovinyl-based sensor layers to triethylamine. He attributes this unexpected response to a slow hydrolysis of the reactand caused by hydroxide ion formed by interaction between the amine base and adventitious water in the sensing membrane.35 CN NC

CN

N CN

N CN CN

3.14 An advantage of traditional IDAs is their modular nature. This feature makes the selected indicators easily interchangeable, and this benefit extends to the present assay. While the colorimetric modulation of 3.11 in the presence of amines was confirmed, synthetic difficulties led us to consider alternative compounds that could also serve as reactand indicators in the proposed assay. Any compound that features a tricyanovinyl group as an integral part of the chromophore could serve as a potential indicator. A survey of the literature offered additional tricyanovinyl derivatives for further investigation. 3.5.2 Tricyanovinyl Ferrocene The preparation of tricyanovinyl ferrocene 3.15 has been previously reported.63 3.15 exhibits a strong charge-transfer maximum at 628 nm that is responsible for its brilliant blue color. While the synthesis of 3.15 was straightforward, complications arose because of its instability in polar solvents (Figure 3.2). While its blue color persisted for weeks in CHCl3, solutions of 3.15 in degassed THF, CH3CN, DMF, and DMSO all 162

changed color within 5 minutes. Light accelerated this decomposition, and the formation of a dark precipitate suggested that Fe+2 was being reduced to Fe0.

Figure 3.2 Absorption spectra of tricyanovinyl ferrocene 3.15. A 5 µM solution of 3.15 in CH3CN changed from blue to yellow over the course of three hours. Exposure of the solution to light accelerated this process. These spectral changes were not observed for solutions of 3.15 in non-polar solvents such as CHCl3. As a result of this sensitivity, the titration of benzylamine with 3.15 was carried out in the dark in CH3CN. No reaction was observed by UV-Vis spectroscopy. Similar UV-Vis titrations with stronger nucleophiles such as azide and thiolate also provided no observable reaction. This lack of reactivity with nucleophiles can be explained by a resonance argument: the negative charge of the cyclopentadienyl ring can be delocalized onto the tricyanovinyl moiety (Scheme 3.8).

The resulting resonance structure

contributes to 3.15’s resonance hybrid, thereby mitigating the electrophilicity of the reactive carbon of the tricyanovinyl group.

163

CN

CN

CN

CN

NC

NC

Fe

Fe

Scheme 3.8 Contributing resonance structures of tricyanovinyl ferrocene 3.15. Delocalization of the negative charge of the cyclopentadienyl ligand renders the tricyanovinyl ferrocene reactand non-electrophilic. This resonance argument is invoked to explain the observed lack of reactivity of 3.15 with various nucleophiles. 3.5.3 Tricyanoethynylethene Dulog et al investigated the electrochemical properties of a series of tricyanoethynylethene derivatives.64 These compounds included those with tricyanovinyl substituents conjugated to electron-donating groups. As such, compounds of this family are candidates for use as indicators in the proposed assay. Compound 3.16 is a unique donor-acceptor chromophore that has garnered interest for its potential nonlinear optical properties.65 Its intense charge-transfer band at 591 nm and integral tricyanovinyl group led to an assessment of its potential to serve as an amine-reactive indicator. N

CN

NC

CN

3.16

164

Upon addition of benzylamine to a solution of 3.16 in CH3CN, the color changed from purplish-blue to yellow within seconds. This reaction was monitored by UV-Vis spectroscopy and showed saturation behavior at one equivalent of added benzylamine (Figure 3.3). The emergence of this saturation behavior required 30-minute equilibration times between injections of benzylamine. The contents of the cuvette were also warmed in a 45° C water bath during this post-injection equilibration period. These experimental conditions were necessary because of the relatively slow covalent bond formation between the amine and tricyanoethynylethene indicator.

Titrations in which these

requirements for equilibration were neglected did not show saturation behavior even at 1.89 equivalents of benzylamine.

Figure 3.3 UV-vis absorbance modulations and associated isotherm for addition of benzylamine to 3.16. The absorption behavior of a 10 µM solution of 3.16 in CH3CN was monitored with increasing concentrations of benzylamine. After addition of each aliquot of benzylamine, the resulting solution was heated to 45° C for 30 min before collecting its absorption spectrum. The isotherm associated with the overall absorbance change at 598 nm is shown on the right. Because the goals of our project demand reversible covalent bond formation, exogenous cyanide was added to the substitution product of 3.16 and benzylamine to evaluate the system’s dynamic potential. This potential was again assessed by UV-Vis 165

spectroscopy. Compound 3.16 was saturated with benzylamine, and aliquots of sodium cyanide in CH3CN were added to the contents of the quartz cuvette. Addition of seven equivalents of sodium cyanide (relative to 3.16) did not produce a change in the absorption spectrum, indicating the system’s lack of reversibility under these conditions. This lack of reversibility is not surprising considering the relative basicities (and corresponding leaving group abilities) of amide and cyanide anions. Further modification of experimental conditions was undertaken to achieve reversibility. The use of methanolic solvents was investigated for several reasons. The use of methanol allows for buffering of the solutions at a desired pH.

Control of

experimental parameters such as pH enables “tuning” of the reaction conditions in the search for reversibility. Also, methanol was expected to dissolve sodium cyanide more readily than acetonitrile. However, a solution of 3.16 in degassed CH3OH turned from purplish-blue to yellow within 30 minutes. Unlike the polar solvent-promoted color change observed with tricyanovinyl ferrocene 3.15, the color change associated with tricyanoethynylethene 3.16 also occurred when the solution was protected from light. These observations suggested that methanol was covalently interacting with 3.16 and that this putative chemical reaction was responsible for the observed color change. This suggestion is especially interesting in light of previous reports on tricyanovinyl compounds. Mohr’s tricyanovinyl-based reactands were only employed in the detection of amines and hydroxide.29 Other groups describe the substitution reaction between tricyanovinyl groups and methanol as slow and low yielding even in the presence of catalyst and heat. The low reactivity of tricyanovinyl compounds toward methanol is evidenced by Sausen’s assertion that “reaction in the absence of catalyst was very slight; after 69 hours 84% of the tricyanovinylbenzene was recovered unchanged.”60 Perhaps the enhanced susceptibility of 3.16 to alcohols is the result of the increased electronegativity 166

of the sp-hybridized carbon of the alkyne. The amplified electronegativity of the attached carbon conceivably activates the tricyanovinyl group for attack by weaker nucleophiles even in the absence of heat and catalyst. Its observed color change in the presence of methanol renders 3.16 amenable to the detection of primary alcohols. As potential analytes, alcohols fit nicely into the current design criteria of the proposed assay.

Like amines, alcohols are pervasive

functional groups in organic chemistry, and chiral alcohols are valuable intermediates in the production of many useful compounds. Accordingly, their asymmetric preparation has been extensively studied.66 The targeting of alcohols is also more experimentally attractive. Because of alcohols’ increased leaving group abilities relative to amines, primary alcohol analytes might be more accommodating in the establishment of dynamic reversibility.

Figure 3.4 UV-vis absorbance modulations and associated isotherm for addition of methanol to 3.16. The absorption behavior of a 10 µM solution of 3.16 in CH3CN was monitored with increasing concentrations of methanol. An equimolar amount of Hunig’s base was also present in the methanolic titrant. After addition of each aliquot of methanol, the resulting solution stood for 30 min before collection of its absorption spectrum. The isotherm associated with the overall absorbance change at 583 nm is shown on the right.

167

Titration of CH3OH into 3.16 produced spectral changes similar to those observed with the addition of benzylamine (Figure 3.4). However, the reaction of 3.16 with CH3OH was much slower than the corresponding reaction with benzylamine; this discrepancy is explained by the decreased nucleophilicity of alcohols when compared to amines. Following the addition of each aliquot of titrant, the resulting solution stood at room temperature for ~40 min before recording its absorption spectrum. Despite this consideration, saturation of 3.16 was not observed upon addition of 5 equivalents of methanol. This situation is likely due to the slow kinetics of the reaction. Reversibility of the system was then probed by addition of sodium cyanide. A reversal of the previously observed spectral change upon addition of exogenous cyanide would confirm the reversibility of the process. In the absence of such reversibility, the absorbance spectrum was expected to remain unchanged. Interestingly, neither of these possibilities was observed upon addition of sodium cyanide. Instead, the absorption maxima at 581 nm decreased, and the solution became colorless (Figure 3.5). This observation suggests that the added cyanide reacts irreversibly with an electrophilic site on 3.16, and the associated loss of color indicates a disruption in conjugation. Reaction at the tricyanovinyl group would perturb the conjugation of 3.16 and mitigate the electron accepting ability of this terminal group. Alternatively, reaction at the electrophilic carbon of the alkyne would produce an allene, resulting in a loss of planarity that could account for 3.16’s loss of color (Scheme 3.9).

168

Figure 3.5 UV-vis absorbance modulations upon addition of NaCN to a solution containing MeOH and 3.16. Addition of NaCN to a solution of MeOH (20 µM), Hunig’s base (40 µM) and 3.16 (10 µM) in CH3CN at 45° C resulted in an unexpected loss of color.

N

N

N

NaCN

or

CH3CN CN

NC

CN

3.16

NC

NC NC

.

CN

CN CN

CN

3.17

CN

3.18

Scheme 3.9 Possible reactions of reactand 3.16 with cyanide. Unexpected reactivity was observed upon addition of cyanide to 3.16. The associated loss of color suggests a disruption in conjugation and/or a reduction of the electron-accepting power of the appended tricyanovinyl group. The formation of stabilized anions 3.17 and 3.18 might account for 3.16’s dramatic decoloration upon addition of NaCN. 169

Direct titration of sodium cyanide into 3.16 reproduced the spectral changes seen in the previously described experiment, providing a loss of the indicator solution’s blue color (Figure 3.6). It remains unknown whether irreversible addition of cyanide to the tricyanovinyl group or an attack on the alkyne is responsible for this loss of color. Regardless of the nature of this reaction, the behavior of 3.16 in the presence of cyanide precludes the reversible addition of alcohols and obviates its use as an indicator in assaying the ee of chiral alcohols.

Figure 3.6 UV-vis absorbance modulations upon addition of NaCN to 3.16. Direct titration of NaCN to a 10 µM solution of 3.16 in CH3CN resulted in a dramatic decrease in the absorption band at 583 nm, resulting in loss of the solution’s blue-green color. A final investigation with 3.16 invoked thiols as potential chiral analytes. Thiols have the advantage of being more nucleophilic than alcohols while their conjugate bases (thiolates) are less basic than alkoxides. As a result, thiols are expected to react faster with 3.16 than alcohols, and reversibility is more feasible given the relative leaving group ability of thiolates. Titration of ethanethiol into 3.16 in the presence of Hunig’s base 170

afforded spectral modulations analogous to those observed upon addition of benzylamine and methanol. The reaction of ethanethiol with 3.16 was qualitatively faster than the corresponding reaction with CH3OH but slower than the reaction with benzylamine. This trend is consistent with the relative Lewis basicities of these nucleophilic species. Unfortunately, titration of up to six equivalents (relative to indicator) of sodium cyanide to a solution of ethanethiol and 3.16 in the presence of Hunig’s base did not effect reversibility as monitored by UV-vis spectroscopy. 3.6 SUMMARY This chapter described efforts toward the development of a functional assay for ee of simple organic molecules.

The proposed assay was inspired by Mohr’s

chromoreactands, and an ultimate goal of this project is the development of an enantioselective version of his colorimetric sensing strategy through the use of a modified eIDA. Amines were initially targeted as potential chiral analytes, but experimental exigencies expanded the investigation to alcohols and thiols. Specifically, this chapter recounted the identification and evaluation of a series of potential indicators for use in the proposed assay. The discovery of an appropriate indicator has been a major hurdle to this point. While this selection was expected to be straightforward, prohibitively difficult synthesis (tricyanovinyl azo dye 3.11), instability and lack of reactivity with nucleophiles (tricyanovinyl ferrocene 3.15), and undesired reactivity (tricyanoethynylethene 3.16) complicated this process.

Furthermore, the proposed assay relies on dynamic

reversibility in the reaction between the indicator and analyte. This reversibility was not observed with any combination of indicator and analyte.

171

3.7 EXPERIMENTAL DETAILS FOR CHAPTER 3 3.7.1 General Considerations The required chemicals were purchased from Aldrich and Acros Organics and used without further purification unless otherwise noted. Reactions were carried out under an atmosphere of argon. Flash chromatography was performed using Whatman 60 A 230-400 mesh silica gel. All products were dried for at least 3 hours prior to spectral analysis. A Varian Unity Plus 300 MHz spectrometer recorded 1H and 13C NMR spectra. The resulting NMR spectra were referenced to the solvent peaks. High-resolution mass spectra data was obtained by a MicroMass AutoSpec-Ultima spectrometer.

UV-vis

absorption data was collected on a Beckman DU-640 spectrophotometer. These spectra were collected at ambient temperature unless otherwise noted. The CH3CN and CHCl3 solvents were degassed by sparging with argon for 1 h prior to the preparation of indicator solutions. 3.7.2 UV-Visible Titrations Although concentrations were altered as necessary, the following represents a typical UV-vis titration: a solution of 3.16 (10 µM) in CH3CN was prepared, and 1 ml of this solution was transferred to a quartz cuvette for collection of its absorption spectrum. A second CH3CN solution containing 10 µM 3.16 and 300 µM benzylamine (along with 600 µM Hunig’s base) was added in aliquots to the cuvette containing the original indicator solution. The absorption spectrum of the sample was recorded between each of these additions.

The λmax absorbance values were plotted as a function of amine

concentration to generate the binding isotherms. employed solvents prior to data collection.

172

Blanks were performed on the

3.7.3 Synthesis 2-(4-nitrobenzylidene)malononitrile (3.8): p-nitrobenzaldehyde 3.7 (10.07 g, 66.6 mmol) and malononitrile (4.42 g, 66.9 mmol) were dissolved in EtOH (30 mL). To this mixture was added 2 mL of an AcOH / pyridine catalyst mixture whose preparation is described in reference 58. The contents of the reaction flask were heated at reflux for 2 h and then cooled in an ice bath. The resulting precipitate was filtered, and the collected brown solid was the desired malononitrile derivative in 76% yield (10.02 g, 50.3 mmol). 1

H NMR (CDCl3): 7.86 (s, 1H), 8.05 (d, 2H), 8.36 (d, 2H). 13C NMR (d-DMSO): 85.9,

113.7, 124.4, 131.5, 136.7, 149.7, 159.3. HRMS (CI+) C10H5N3O2 m/z: 200.0460; calcd: 200.0460.

2-(4-aminobenzylidene)malononitrile (3.9): 3.8 (10.02 g, 50.3 mmol) was taken up in 100 mL EtOAc, and this solution was gently warmed. SnCl2 (56.75 g, 252 mmol) was added to the reaction flask in one solid portion. This mixture was heated to reflux overnight. Upon cooling to room temperature, the reddish-brown reaction mixture was poured into ice and basified (pH ~ 8) with sodium bicarbonate. This mixture was extracted numerous times with EtOAc. The organic layers were combined, dried with MgSO4, and filtered to give an amber colored solution. Silica gel chromatography (using a gradient of 2:1 to 1:1 hexanes:EtOAc) afforded pure product in 54% yield (4.82 g, 27.3 mmol). The silica column was flushed with 3% NEt3 prior to loading the sample for purification. 1H NMR (d-DMSO): 6.64 (d, 2H), 6.99 (s, 2H), 7.71 (d, 2H), 7.96 (s, 1H). 13

C NMR (d-DMSO): 67.9, 113.6, 115.6, 116.4, 118.9, 134.4, 156.0, 159.0. HRMS

(CI+) C10H8N3 m/z: 170.071388; calcd: 170.071822.

173

2-(4-aminophenyl)ethene-1,1,2-tricarbonitrile (3.10): A three-necked round bottom flask was charged with 3.9 (2.19 g, 12.9 mmol) and DMF (25 mL) was added. The reaction flask was sealed and vented to a concentrated sodium hydroxide bath through the use of an inverted funnel. An aqueous 2 M NaCN solution (7 mL) was added dropwise to the reaction flask over 1 min. The reaction mixture became red. Argon was blown through the experimental setup for 20 min. Lead tetraacetate (5.71 g, 12.9 mmol) was then added in one solid portion, and the dark reaction mixture was stirred at room temperature overnight. The contents of the reaction flask was poured into 300 mL of ice water and then filtered to collect a red residue. This residue was recrystallized from xylene to give a yield of 14% (0.34 g, 1.75 mmol). 1H NMR (CDCl3): 4.81 (s, 2H), 6.71 (d, 2H), 8.00 (d, 2H). 13C NMR (d-DMSO): 74.7, 114.7, 115.0, 116.8, 133.39, 136.51, 157.41. HRMS (CI+) C11H7N4 m/z: 195.067827; calcd: 195.067071.

2-(4-((4-(dibutylamino)phenyl)diazenyl)phenyl)ethene-1,1,2-tricarbonitrile

(3.11)

and 2-(4-((4-(1,2,2-tricyanovinyl)phenyl)diazenyl)phenyl)ethene-1,1,2-tricarbonitrile (3.14): 3.10 (0.37 g, 1.90 mmol) was dissolved in AcOH (4 mL) and concentrated HCl (0.5 mL), and the resulting solution was cooled in an ice bath. A solution of NaNO2 (0.13 g, 1.90 mmol) in a minimum amount of H2O was added dropwise to the acidic solution of 3.10. After complete addition, the solution was stirred with cooling for 10 min, and N,N-dibutylaniline (0.43 mL, 1.88 mmol) was added dropwise to the contents of the reaction flask. Upon this addition, the solution became intensely blue in color. After stirring for an additional 10 min, a solid portion (0.25 g) of sodium acetate was added. After several hours, the reaction mixture appeared dark and was difficult to stir. The reaction mixture was reduced in volume, and subjected to silica gel chromatography (2:1 CH2Cl2:hexanes) to afford a 13% yield (0.1 g, 0.24 mmol) of 3.11. 1H NMR (CDCl3): 174

0.98 (t, 6H), 1.40 (m, 4H), 1.63 (m, 4H), 3.41 (t, 4H), 6.73 (d, 2H), 7.98 (d of d, 4H), 8.15 (d, 2H). HRMS (CI+) C25H27N6 m/z: 411.2294; calcd: 411.2297. Homocoupled product 3.14 was also isolated as a product of this reaction. 1H NMR (CDCl3): 7.61 (m, 2H), 7.72 (m, 2H), 7.99 (m, 2H).

1,2,2-tricyanovinylcyclopentadienyl(cyclopentadienyl)iron (3.15): Ferrocene (5.05 g, 27 mmol) and tetracyanoethylene (3.42 g, 27 mmol) were dissolved in dry CH3CN (75 mL) to give a greenish-yellow solution which was heated to reflux for 1.5 h. The reaction was then allowed to stir at room temperature overnight.

Silica gel

chromatography was performed on the reaction mixture using a solvent gradient of 4:1 hexanes : CH2Cl2 to neat CH2Cl2. 3.15 (0.19 g, 0.66 mmol) was isolated as a dark blue crystalline solid in a 3% yield. 1H NMR (CDCl3): 4.45 (s, 5H), 5.14 (s, 2H), 5.24 (s, 2H). HRMS (CI+) C15H10N3Fe m/z: 288.0220; calcd: 288.0224. UV-Vis λmax, CHCl3: 363, 628 nm.

4-(4-(dimethylamino)phenyl)but-1-en-3-yne-1,1,2-tricarbonitrile (3.16): 4-ethynylN,N-dimethylaniline (1.08 g, 7.4 mmol) and copper(I) acetate (0.906 g, 7.4 mmol) were dissolved in 24 mL of dry THF and 6 mL of dry CH3CN, and the resulting heterogeneous reaction mixture was heated to 50° C for 30 min. A solution of tetracyanoethylene (0.89 g, 6.9 mmol) in dry THF was then added to the reaction flask. After an additional 30 min at 50° C, the mixture was allowed to cool to room temperature overnight. The volume of the reaction mixture was reduced to give a dark solid that was deposited on silica for chromatography in 3:1 hexanes : CH2Cl2 to 2:1 CH2Cl2 : hexanes. 3.16 was isolated as a bluish-green solid in a 22% yield (0.39 g, 1.58 mmol). 1H NMR (CDCl3): 3.16 (s, 6H), 6.69 (d, 2H), 7.53 (d, 2H). . UV-Vis λmax, CHCl3: 591 nm. 175

3.8 REFERENCES 1

“Aromatic Amines in the River Elbe – Determination and Behavior during Drinking Water Treatment” Bornick, H.; Hultsch, V.; Grischek, T.; Liening, D.; Worch, E. Vom Wasser 1996, 87, 305-326.

2

“Biogenic Amines and the States of Sleep” Jouvet, M. Science 1969, 163, 32-41.

3

“Biogenic Amines and Emotion” Schildkraut, J.J.; Kety, S.S. Science 1967, 156, 21-30.

4

“Impact of Centrally Applied Biogenic Amines upon the Energy Balance of Fowl” Hillman, P.E.; Scott, N.R.; van Tienhoven, A. Am. J. Physiol. Regulatory Integrative Comp. Physiol. 1977, 232, 137-144.

5

“Polyamine Profiles in Tumor, Normal Tissue of the Homologous Breast, Blood, and Urine of Breast Cancer Sufferers” Leveque, J.; Foucher, F.; Bansard, J.; Havouis, R.; Grall, J.; Moulinoux, J. Breast Cancer Res. Treat. 2000, 60, 99-105.

6

“Analysis of Bacterial Vaginosis-Related Amines in Vaginal Fluid by Gas Chromotography and Mass Spectrometry” Wolrath, H.; Forsum, U.; Larsson, P.G.; Boren, H. J. Clin. Microbiol. 2001, 39, 4026-4031.

7

“Occurrence and Formation of Biologically Active Amines in Foods” ten Brink, B.; Damink, C.; Joosten, H.M.L.J.; Huis in ‘t Veld, J.H.J. Int. J. Food. Microbiol. 1990, 11, 73-84.

8

“Biogenic Amines: Their Importance in Foods” Santos, M.H.S. Int. J. Food. Microbiol. 1996, 29, 213-231.

9

(a.) “Visualization of Molecular Length of α,ω-Diamines and Temperature by a Receptor Based on Phenolphthalein and Crown Ether” Fuji, K.; Tsubaki, K.; Tanaka, K.; Hayashi, N.; Otsubo, T.; Kinoshita, T. J. Am. Chem. Soc. 1999, 121, 3807-3808. (b.) “Visual Recognition of Triamines by Phenolphthalein Derivatives: Consideration of the Structure of the Colored Complex” Tsubaki, K.; Hayashi, N.; Nuruzzamam, M.; Kusumoto, T.; Fuji, K. Org. Lett. 2001, 3, 4067-4069.

10

Solomons, G.; Fryhle, C. Organic Chemistry, 7th ed. (Wiley & Sons, New York 2002) 975-976.

176

11

“α-CD/Crown-Appended Diazophenol for Selective Sensing of Amines” Jung, J.H.; Lee, S.J.; Kim, J.S.; Lee, W.S.; Sakata, Y.; Kaneda, T. Org. Lett. 2006, 8, 30093012.

12

“Sensing a Paradigm Shift in the Field of Molecular Recognition: From Selective to Differential Receptors” Lavigne, J.J.; Anslyn, E.V. Angew. Chem. Int. Ed. 2001, 40, 3118-3130.

13

“Cross-Reactive Chemical Sensor Arrays” Albert, K.J.; Lewis, N.S.; Schauer, C.L.; Sotzing, G.A.; Stitzel, S.E.; Vaid, T.P.; Walt, D.R. Chem. Rev. 2000, 100, 25952626.

14

“Computational Methods for the Analysis of Chemical Sensor Array Data from Volatile Analytes” Jurs, P.C.; Bakken, G.A.; McClelland, H.E. Chem. Rev. 2000, 100, 2649-2678.

15

“Molecular Recognition and Discrimination of Amines with a Colorimetric Array” Rakow, N.A.; Sen, A.; Janzen, M.C.; Ponder, J.B.; Suslick, K.S. Angew. Chem. Int. Ed. 2005, 44, 4528-4532.

16

“Molecularly Imprinted Polymers and Their Use in Biomimetic Sensors” Haupt, K.; Mosbach, K. Chem. Rev. 2000, 100, 2495-2504.

17

“Molecularly Imprinted Polymer Sensor Arrays” Greene, N.T.; Morgan, S.L.; Shimizu, K.D. Chem. Commun. 2004, 1172-1173.

18

“Colorimetric Molecularly Imprinted Polymer Sensor Array using Dye Displacement” Greene, N.T.; Shimizu, K.D. J. Am. Chem. Soc. 2005, 127, 5695-5700.

19

“New Chromogenic and Fluorogenic Reagents and Sensors for Neutral and Ionic Analytes Based on Covalent Bond Formation -- A Review of Recent Developments” Mohr, G.J. Anal. Bioanal. Chem. 2006, 386, 1201-1214.

20

“Tailoring the Sensitivity and Spectral Properties of a Chromoreactand for the Detection of Amines and Alcohols” Mohr, G.J. Anal. Chim. Acta 2004, 508, 233237.

21

“A Comparison of Immobilization Sol-gel Methods for an Optical Chemical Hydrazine Sensor” Gojon, C.; Dureault, B.; Hovnanian, N.; Guizard, C. Sens. Actuators B 1997, 38, 154-162.

22

“A Molecular Colour Sensor for Monosaccharides” Ward, C.J.; Patel, P.; Ashton, P.R.; James, T.D. Chem. Commun. 2000, 229-230. 177

23

“3,3’,5,5’-Tetramethyl-N-(9-anthrylmethyl)benzidine: A Dual-Signaling Fluorescent Reagent for Optical Sensing of Aliphatic Aldehydes” Yang, R.; Li, K.; Liu, F.; Li, N.; Zhao, F.; Chan, W. Anal. Chem. 2003, 75, 3908-3914.

24

“Development of Chromogenic Reactands for Optical Sensing of Alcohols” Mohr, G.J.; Citterio, D.; Spichiger-Keller, U.E. Sens. Actuators, B 1998, 49, 226-234.

25

“Visual Detection of Cysteine and Homocysteine” Rusin, O.; St. Luce, N.N.; Agbaria, R.A.; Escobedo, J.O.; Jiang, S.; Warner, I.M.; Dawan, F.B.; Lian, K.; Strongin, R.M. J. Am. Chem. Soc. 2004, 126, 4064-4065.

26

“A Selective Chromogenic Reagent for Cyanide Determination” Ros-Lis, J.V.; Martinez-Manez, R.; Soto, J. Chem. Commun. 2002, 2248-2249.

27

“New Chromogenic Probes into Nanoscopic Pockets in Enhanced Sensing Protocols for Amines in Aqueous Environments” Basurto, S.; Torroba, T.; Comes, M.; Martinez-Manez, R.; Sancenon, F.; Villaescusa, L.; Amoros, P. Org. Lett. 2005, 7, 5469-5472.

28

“Trifluoroacetyl-chlorin as a New Chemosensor for Alcohol/Amine Detection” Sasaki, S.; Kotegawa, Y.; Tamiaki, H. Tetrahedron Lett. 2006, 47, 4849-4852.

29

“Chromo- and Fluororeactands: Indicators for Detection of Neutral Analytes by Using Reversible Covalent-Bond Chemistry” Mohr, G.J. Chem. Eur. J. 2004, 10, 10821090.

30

“Application of Chromogenic and Fluorogenic Reactands in the Optical Sensing of Dissolved Aliphatic Amines” Mohr, G.J.; Demuth, C.; Spichiger-Keller, U.E. Anal. Chem. 1998, 70, 3868-3873.

31

“A Chromoreactand for Optical Sensing of Amphetamines” Mohr, G.J.; Wenzel, M.; Lehmann, F.; Czerney, P. Anal. Bioanal. Chem. 2002, 374, 399-402.

32

“Development of Chromogenic Copolymers for Optical Detection of Amines” Mohr, G.J.; Tirelli, N.; Lohse, C.; Spichiger-Keller, U.E. Adv. Mater. 1999, 10, 13531357

33

(a.) “Cross-linked Dendrimer Hosts Containing Reporter Groups for Amine Guests” Mertz, E.; Zimmerman, S.C. J. Am. Chem. Soc. 2003, 125, 3424-3425. (b.) “A Monomolecularly Imprinted Dendrimer (MID) Capable of Selective Binding with a Tris(2-aminoethyl)amine Guest through Multiple Functional Group Interactions” Beil, J.B.; Zimmerman, S.C. Chem. Commun. 2004, 488-489. (c.) Kinetics and Thermodynamics of Amine and Diamine Signaling by a 178

Trifluoroacetyl Azobenzene Reporter Group” Mertz, E.; Beil, J.B.; Zimmerman, S.C. Org. Lett. 2003, 5, 3127-3130. 34

“A Survey of Hammett Substituent Constants and Resonance and Field Parameters” Hansch, C.; Leo, A.; Taft, R.W. Chem. Rev. 1991, 91, 165-195.

35

“A Tricyanovinyl Azobenzene Dye Used for the Optical Detection of Amines via a Chemical Reaction in Polymer Layers” Mohr, G.J. Dyes Pigm. 2004, 62, 77-81.

36

(a.) “Chemistry at Its Most Beautiful” Freemantle, M. Chem. Eng. News 2003, 81, 2730. (b.) Ball, P. Elegant Solutions: Ten Beautiful Experiments in Chemistry (Royal Society of Chemistry, Cambridge, 2005).

37

“Chiral Molecular Recognition on Formation of a Metalloanthocyanin: A Supramolecular Metal Complex Pigment from Blue Flowers of Salvia patens” Kondo, T.; Oyama, K.; Yoshida, K. Angew. Chem. Int. Ed. 2001, 40, 894-897.

38

“Enantioselective and Diasteroselective Molecular Recognition of Neutral Molecules” Webb, T.H.; Wilcox, C.S. Chem. Soc. Rev. 1993, 22, 383-395.

39

“Judging on Host-Guest Binding Mode Uniqueness: Association Entropy as an Indicator in Enantioselection” Jadhav, V.D.; Schmidtchen, F.P. Org. Lett. 2006, 8, 2329-2332.

40

“Enantiomeric Recognition of Amine Compounds by Chiral Macrocyclic Receptors” Zhang, X.X.; Bradshaw, J.S.; Izatt, R.M. Chem. Rev. 1997, 97, 3313-3361.

41

“Combinatorial and Evolution-Based Methods in the Creation of Enantioselective Catalysts” Reetz, M.T. Angew. Chem. Int. Ed. 2001, 40, 284-310.

42

“Fluorescence of Organic Molecules in Chiral Recognition” Pu, L. Chem. Rev. 2004, 104, 1687-1716.

43

“Bisbinaphthyl Macrocycle-based Highly Enantioselective Fluorescent Sensors for αHydroxycarboxylic Acids” Lin, J.; Zhang, H.; Pu, L. Org. Lett. 2002, 4, 32973300.

44

“Enantioselective Fluorescent Recognition of a Soluble “Supported” Chiral Acid: Toward a New Method for Chiral Catalyst Screening” Li, Z.; Lin, J.; Qin, Y.; Pu, L. Org. Lett. 2005, 7, 3441-3444.

45

“Colorimetric Chiral Recognition by a Molecular Sensor” Kubo, Y.; Maeda, S.; Tokita, S.; Kubo, M. Nature 1996, 382, 522-524. 179

46

“Visual Enantiomeric Recognition Using Chiral Phenolphthalein Derivatives” Tsubaki, K.; Nuruzzaman, M.; Kusumoto, T.; Hayashi, N.; Bin-Gui, W.; Fuji, K. Org. Lett. 2001, 3, 4071-4073.

47

“Color Indicators of Molecular Chirality Based on Doped Liquid Crystals” van Delden, R.A.; Feringa, B.L. Angew. Chem. Int. Ed. 2001, 40, 3198-3200.

48

“Direct Visual Detection of the Stereoselectivity of a Catalytic Reaction” Eelkema, R.; van Delden, R.A.; Feringa, B.L. Angew. Chem. Int. Ed. 2004, 43, 5013-5016.

49

“Facile Quantification of Enantiomeric Excess and Concentration with IndicatorDisplacement Assays: An Example in the Analyses of α-Hydroxyacids” Zhu, L.; Anslyn, E.V. J. Am. Chem. Soc. 2004, 126, 3676-3677.

50

“Two Methods for the Determination of Enantiomeric Excess and Concentration of a Chiral Sample with a Single Spectroscopic Measurement” Zhu, L.; Shabbir, S.H.; Anslyn, E.V. Chem. Eur. J. 2007, 13, 99-104.

51

“Using Enantioselective Indicator Displacement Assays to Determine the Enantiomeric Excess of α-Amino Acids” Leung, D.; Folmer-Andersen, J.F.; Lynch, V.M.; Anslyn, E.V. J. Am. Chem. Soc. 2008, ASAP.

52

“High-Throughput Screening by Using a Blue-Fluorescent Antibody Sensor” Matsushita, M.; Wirsching, P.; Lerner, R.A.; Janda, K.D. Angew. Chem. Int. Ed. 2003, 42, 5984-5987.

53

“Reaction Microarrays: A Method for Rapidly Determining the Enantiomeric Excess of Thousands of Samples” Korbel, G.A.; Lalic, G.; Shair, M.D. J. Am. Chem. Soc. 2001, 123, 361-362.

54

(a.) “EMDee: An Enzymatic Method for Determining Enantiomeric Excess” Abato, P.; Seto, C.T. J. Am. Chem. Soc. 2001, 123, 9206-9207. (b.) “Using Enzyme Inhibition as a High Throughput Method to Measure the Enantiomeric Excess of a Chiral Sulfoxide” Sprout, C.M.; Seto, C.T. Org. Lett. 2005, 7, 5099-5102.

55

“Transitioning Enantioselective Indicator Displacement Assays for α-Amino Acids to Protocols Amenable to High-Throughput Screening” Leung, D.; Anslyn, E.V. J. Am. Chem. Soc. 2008, ASAP.

56

“A High-Throughput NMR-based ee-Assay Using Chemical Shift Imaging” Reetz, M.T.; Tielmann, P.; Eipper, A.; Ross, A.; Schlotterbeck, G. Chem. Commun. 2004, 1366-1367. 180

57

“Synthesis and Reactions of Cyanovinyl-Substituted Benzenediazonium Salts for NonLinear Optics” Schilling, M.L.; Katz, H.E.; Cox, D.I. J. Org. Chem. 1988, 53, 5538-5540.

58

“Cyanocarbon Chemistry. VI. Tricyanovinylamines” McKusick, B.C.; Heckert, R.E.; Cairns, T.L.; Coffman, D.D.; Mower, H.F. J. Am. Chem. Soc. 1958, 80, 28062815.

59

“Reactions of Alpha, Beta-Unsaturated Dinitriles” Corson, B.B.; Stoughton, R.W. J. Am. Chem. Soc. 1928, 50, 2825-2837.

60

“Cyanocarbon Chemistry. VII. Tricyanoethylenes” Sausen, G.N.; Engelhardt, V.A.; Middleton, W.J. J. Am. Chem. Soc. 1958, 80, 2815-2822.

61

“Tributyltin Cyanide: A Facile Cyanating Agent of Acyl Chlorides” Tanaka, M. Tetrahedron Lett. 1980, 21, 2959-2962.

62

“Photoinduced Reactions of Aroyl Cyanides” Raaen, V.F. J. Org. Chem. 1966, 31, 3310-3312.

63

“Tricyanovinylation and Cyanation of Ferrocene with Tetracyanoethylene” Nefedov, V.A.; Bespalov, B.P.; Titov, V.V. Zh. Org. Khim. 1974, 10, 1553-1554.

64

“Synthesis and Electrochemical Properties of 4-phenyl-1-buten-3-yne-1,1,2tricarbonitriles and Tricyanoacrylates” Dulog, L.; Korner, B.; Heinze, J.; Yang, J. Liebigs Ann. 1995, 1663-1671.

65

“Cyanoethynylethenes: A Class of Powerful Electron Acceptors for Molecular Scaffolding” Moonen, N.N.P.; Boudon, C.; Gisselbrecht, P.; Seiler, P.; Gross, M.; Diederich, F. Angew. Chem. Int. Ed. 2002, 41, 3044-3045.

66

“Novel Bioreduction System for the Production of Chiral Alcohols” Kataoka, M.; Kita, K.; Wada, M.; Yasohara, Y.; Hasegawa, J.; Shimizu, S. Appl. Microbiol. Biotechnol. 2003, 62, 437-445.

181

Works Consulted Atwood, J.L.; Steed, J.W. Encyclopedia of Supramolecular Chemistry (Marcel Dekker, New York, 2004) Beer, P.D.; Gale, P.A.; Smith, D.K. Supramolecular Chemistry (Oxford University Press: New York, 1999) Boas, U.; Christensen, J.B.; Heegaard, P.M.H. Dendrimers in Medicine and Biotechnology: New Molecular Tools (Royal Society of Chemistry, Cambridge, 2006) Christie, R.M. Colour Chemistry (Royal Society of Chemistry, Cambridge, 2001) Cloninger, M.J. “Biological Applications of Dendrimers” Curr. Opin. Chem. Biol. 2002, 6, 742-748. Dodd, J.S. The ACS Style Guide: A Manual for Authors and Editors, 2nd Ed. (The American Chemical Society, Washington, D.C., 1997) Fischer, M.; Vogtle, F. “Dendrimers: From Design to Application – A Progress Report” Angew. Chem. Int. Ed. 1999, 38, 884-905. Froehling, P.E. “Dendrimers and Dyes – A Review” Dyes Pigm. 2001, 48, 187-195. Lehn, J.M. Supramolecular Chemistry: Concepts and Perspectives (VCH, Weinheim, 1995) Newkome, G.R.; Moorefield, C.N. Vogtle, F. Dendritic Molecules: Concepts, Syntheses, Perspectives (VCH, New York, 1996) Steed, J.W.; Atwood, J.L. Supramolecular Chemistry (Wiley, Chichester, 2000) Steed, J.W.; Turner, D.R.; Wallace, K.J. Core Concepts in Supramolecular Chemistry and Nanochemistry (Wiley, Chichester, 2007) Tomalia, D.A.; Naylor, A.M.; Goddard, W.A. “Starburst Dendrimers: Molecular-Level Control of Size, Shape, Surface Chemistry, Topology, and Flexibility from Atoms to Macroscopic Matter” Angew. Chem. Int. Ed. 1990, 29, 138-175. Zollinger, H. Color Chemistry: Syntheses, Properties and Applications of Organic Dyes and Pigments, 2nd Ed. (VCH, Weinheim, 1991) 182

Vita John Chance Rainwater was born in Meridian, Mississippi on January 26, 1979, the son of Clayton Rainwater and Louise Rainwater. He graduated from the Mississippi School for Mathematics and Science in 1997 and entered Loyola University in New Orleans, Louisiana in the fall of that year. During the summer of 2000, he participated in an undergraduate research program at the University of Florida sponsored by the National Science Foundation. He received a Bachelor of Science from Loyola University in May 2001 and enrolled in the Graduate School of the University of Texas-Austin, from which he earned a Master of Arts in December 2004.

Permanent address:

4980 Highway 45 North, Waynesboro, Mississippi, 39367

This dissertation was typed by the author.

183

184