Louisiana State University LSU Digital Commons
LSU Doctoral Dissertations Graduate School
2005 An evaluation of the enantiomeric recognition of amino acid based polymeric surfactants and cyclodextrins using spectroscopic and chromatographic methods Bertha Cedillo Valle Louisiana State University and Agricultural and Mechanical College, [email protected]
Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_dissertations Part of the Chemistry Commons
Recommended Citation Valle, Bertha Cedillo, "An evaluation of the enantiomeric recognition of amino acid based polymeric surfactants and cyclodextrins using spectroscopic and chromatographic methods" (2005). LSU Doctoral Dissertations. 3349. https://digitalcommons.lsu.edu/gradschool_dissertations/3349
This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Doctoral Dissertations by an authorized graduate school editor of LSU Digital Commons. For more information, please [email protected]. AN EVALUATION OF THE ENANTIOMERIC RECOGNITION OF AMINO ACID BASED POLYMERIC SURFACTANTS AND CYCLODEXTRINS USING SPECTROSCOPIC AND CHROMATOGRAPHIC METHODS
A Dissertation
Submitted to the graduate Faculty of the Louisiana State University and Agricultural and Mechanical College In partial fulfillment of the Requirements for the degree of Doctor of Philosophy
in
The Department of Chemistry
By Bertha Cedillo Valle B.S. Texas Tech University, 1998 December 2005
DEDICATION
I want to dedicate this work to my husband, Larry, and my immediate siblings, Octavio
Jorge, Martha, and Silvia, along with their respective spouses. You guys were a constant driving force for me and were always there to support me financially and/or emotionally. I appreciate your support more than words can say. I hope that I can return the favor one day.
ii ACKNOWLEDGEMENTS
I would like to graciously acknowledge the support of the many people that made the completion of this work possible.
My mother and father, Octavio and Mercedes Cedillo for leaving your home country to establish a new life so that I may have a better chance at excelling in life. Thank you also for encouraging me to follow my dreams no matter where they took me.
My parents-in-law, Abel and Gloria Valle, for your constant words of encouragement and emotional support.
Dr. Isiah Warner, for all your support both financially and professionally but most of all for your fatherly concern for my future. I could not have done it without your undoubted belief in me.
Dr. Shahab A. Shamsi, for our collaborations in research and in manuscript writing.
Dr. Rezik A. Agbaria, for teaching me everything about fluorescence and life in general.
Dr. Kevin Morris, for a seemingly effortless collaboration involving the understanding of chiral interactions using PFG-NMR and 2D-NMR.
Constantina Kapnissi, PhD., for your faithful friendship that I will remember always.
Karen A. Prado, for your hard work and dedication to our research and especially for allowing me into your home. Thank you for being the “sister” I needed.
Judson L. Haynes, PhD., for your constant personal and professional mentoring.
Charles W. Henry, PhD., for being my “big brother” during and after grad school.
Dr. Matthew McCarroll, for teaching me all about instrumentation and electronics.
iii Kristin A. Fletcher, PhD., for the speedy return of my manuscripts as well for listening and caring about my problems both professional and personal. I could not have done it with out your friendship.
Warner Research Group, for all the experiences that made me to grow both as a scientist and as an individual.
Dr. Purnendu P. Dasgupta, for allowing me to be a part of your group and giving me the opportunity to learn how to do research.
Dr. Dennis L. Shelly, the wonderful time I had working for you and for making me come to
LSU for a visit.
Dr. Carl J. Popp, for the incredible summer I spent in New Mexico working on M mountain and for your excellent letter of recommendation that eventually got me in to LSU.
My husband, Larry A. Valle, for hanging in there with me while we lived apart for over four years of our marriage. Thank you for putting your life on hold for me. I hope to make our future worth the wait.
iv TABLE OF CONTENTS
DEDICATION…………………………………………………………………………… ii
ACKNOWLEDGEMENTS……………………………………………………………… iii
LIST OF TABLES……………………………………………………………………….. viii
LIST OF FIGURES……………………………………………………………………….ix
LIST OF ABBREVIATIONS……………………………………………………………. xiii
ABSTRACT……………………………………………………………………………… xvi
CHAPTER 1.INTRODUCTION………………………………………………………… 1 1.1 Chirality and Its Significance……………………………………. 1 1.2 Fundamentals of Capillary Electrophoresis………………………5 1.3 Capillary Zone Electrophoresis (CZE)……………………………7 1.4 Micellar Electrokinetic Chromatography………………………... 11 1.5 Chiral Selectors for Chiral CE…………………………………… 17 1.5.1 Polymeric Surfactants……………………………………. 17 1.5.2 Native Cyclodextrins and Their Derivatives……………... 21 1.5.3 Ligand-Exchange Metal Complexes……………………... 24 1.5.4 Crown Ethers……………………………………………...26 1.5.5 Macrocyclic Antibiotics………………………………….. 28 1.6 Theory of Fluorescence Spectroscopy…………………………… 29 1.6.1 Polarization of Photoluminescence………………………. 33 1.7 Nuclear Magnetic Resonance (NMR)……………………………. 36 1.7.1 Pulsed-Field Gradient NMR……………………………... 39 1.7.2 Nuclear Overhauser Effect Spectroscopy NMR…………. 42 1.8 Scope of Dissertation…………………………………………….. 45 1.9 References………………………………………………………... 47
CHAPTER 2. COMBINATION OF CYCLODEXTRINS AND POLYMERIC SURFACTANTS FOR CHIRAL SEPARATIONS……………………... 54 2.1 Introduction………………………………………………………. 54 2.2 Experimental……………………………………………………... 56 2.2.1 Materials and Reagents…………………………………... 56 2.2.2 Synthesis of Polymeric Chiral Surfactant………………... 58 2.2.3 Capillary Electrophoresis Procedure…………………….. 60 2.3 Results and Discussion…………………………………………... 61 2.3.1 Elution Order of Binaphthyl Derivatives as a Function of CD Cavity Size and the Stereochemical Configuration of Polymeric Surfactants…………………………………. 61 2.3.2 Enantioseparation of BNA……………………………….. 62 2.3.3 Enantioseparation of BOH……………………………….. 65
v 2.3.4 Enantioseparation of BNP………………………………... 70 2.4 Conclusions………………………………………………………. 72 2.5 References………………………………………………………... 73
CHAPTER 3. POLYSODIUM N-UNDECANOYL-L-LEUCYLVALINATE: A VERSATILE CHIRAL SELECTOR FOR MICELLAR ELECTROKINETIC CHROMATOGRAPHY………………………….. 75 3.1 Introduction………………………………………………………. 75 3.2 Experimental……………………………………………………... 77 3.2.1 Materials…………………………………………………. 77 3.2.2 Instrumentation…………………………………………... 77 3.2.3 Synthesis of Polysodium N-undecanoyl-L-leucylvalinate.. 78 3.2.4 Determination of Partial Specific Volume……………...... 78 3.2.5 Preparation of MEKC Buffers and Analyte Solutions…… 79 3.2.6 Capillary Electrophoresis Procedure…………………….. 79 3.3 Results and Discussion…………………………………………... 79 3.3.1 Enantioseparation of Class I Analytes…………………… 80 3.3.2 Enantioseparation of Class II Analytes………………….. 86 3.3.3 Enantioseparation of Class III Analytes…………………. 94 3.4 Conclusions………………………………………………………. 99 3.5 References………………………………………………………... 100
CHAPTER 4. UNDERSTANDING CHIRAL SEPARATIONS INVOLVING P-(SULV) USING STEADY-STATE FLUORESCENCE ANISOTROPY, CAPILLARY ELECTROPHORESIS, AND NMR……………………... 102 4.1 Introduction……………………………………………………… 102 4.1.1 Nuclear Magnetic Resonance Theory……………………. 103 4.1.2 α vs. β Theory……………………………………………. 104 4.2 Experimental…………………………………………………….. 106 4.2.1 Materials and Reagents…………………………………... 106 4.2.2 Dansyl Amino Acid (Dns-AA) Derivatization…………... 107 4.2.3 Prep-TLC………………………………………………… 108 4.2.4 Synthesis of Polysodium N-Undecanoyl- (L,L)-Leucylvaline, p-(L-SULV), and Its Diastereomers………………………………………….. 109 4.2.5 Centrifugal Filtration…………………………………….. 109 4.2.6 Capillary Electrophoresis Instrumentation and Procedures………………………………………………... 109 4.2.7 Background Electrolyte (BGE) and Capillary Electrophoresis Standard Preparation……………………. 110 4.2.8 Selectivity Calculation…………………………………… 110 4.2.9 Fluorescence Spectroscopy Instrumentation and Procedures………………………………………………... 111 4.3 Results and Discussion…………………………………………... 113 4.3.1 MEKC Results…………………………………………… 113 4.3.2 Nuclear Overhauser Effect Spectroscopy (NOESY)
vi Results ………………………………………………….. 117 4.3.3 Steady-State Fluorescence Spectroscopy Result………… 124 4.4 Conclusions………………………………………………………. 132 4.5 References………………………………………………………... 133
CHAPTER 5. CONCLUSIONS AND FUTURE DIRECTIONS………………………. 137
APPENDIX I PLOT OF THE CORRELATION BETWEEN THE α VALUES DERIVED FROM MEKC EXPERIMENTS AND β-VALUES DERIVED FROM FLUORESCENCE………………………………………………. 142
APPENDIX II DIGITAL PHOTOGRAPHS OF PREP TLC…………………... 143
APPENDIX III MEKC ENANTIOSEPARATION OF THE ENANTIOMERS OF SEVEN ANALYTES …………………………………………………….. 146
APPENDIX IV TITRATION CURVES SHOWING THE A) ENANTIOMER ANISOTROPY AND B) CHIRAL SELECTIVITY OF BOH AND DNS-PHE………... 153
APPENDIX V α VS. β PLOTS OBTAINED USING p-(L,D)-SULV AND p-(D,L)-SULV AS THE CHIRAL SELECTORS…………………………………………………….. 155
APPENDIX VI EXPERIMENTAL CONDITIONS USED IN THE PFG NMR STUDIES………………………………………………………… 157
APPENDIX VII RAW 2-D NOE PLOTS FOR ANALYTES IN CHAPTER 4.….. 158
VITA……………………………………………………………………………………... 163
vii LIST OF TABLES Table Page
1.1 Activities of select chiral compounds……………………………………………. 3
1.2 Brief description of six modes of capillary electrophoresis……………………… 6
1.3 Critical micelle concentrations (CMC) and aggregation numbers (AN) of select surfactants [18]…………………………………………………………. 14
1.4 Important characteristics of α-, β-, and γ-CDs [46]……………………………… 22
2.1 Enantiomeric selectivity pattern of binaphthyl derivatives towards neutral CDs and PSs……………………………………………………………... 62
3.1 Physicochemical Properties of sodium N-undecenoyl-L-leucylvalinate (L-SULV) and polysodium N-undecanoyl-L-leucylvalinate (p-SULV)…………. 76
3.2 Class I: Examples of compounds exhibiting strong chiral interactions with p-(L-SULV) and requiring only 5-20 mM PS for enantioseparation………………………………………………………… 81
3.3 Class II: Examples of compounds exhibiting moderate chiral interactions with p-(L-SULV) and requiring 30-50 mM PS for enantioseparation………………………………………………………… 87
3.4 Class III: Examples of compounds exhibiting weak chiral interactions with p-(L-SULV) and requiring 75-85 mM PS for enantioseparation………………………………………………………… 95
4.1 Enantiomeric selectivity pattern of several analytes towards p-(SULV) using MEKC………………………………………………………….. 116
4.2 Relative NOE intensities observed between all BNP protons and those of p-(L-SULV)………………………………………………………… 120
4.3 Minimum p-(L-SULV) concentration required to achieve differences in anisotropy and selectivity for three model analytes………………. 128
5.1 List of spectrophysical properties observed for various chiral analytes in the presence of p-(L-SULV) determined by PFG-NMR. Conditions shown in Appendix VI………………………………………………. 141
viii LIST OF FIGURES
Figure Page
1.1 Structures of the three major types of chiral molecules………………………….. 2
1.2 Three point interaction model used to describe the diastereomeric analyte-chiral selector complex…………………………………………………...4
1.3 Schematic diagram of a typical capillary electrophoresis system……………….. 7
1.4 Schematic represenatation of the double layer formation……………………….. 8
1.5 (a) Electrically driven flow (e.g., EOF) and corresponding solute zone profile, (b) Pressure driven (or hydrodynamic) flow and corresponding solute zone profile………………………………………………... 9
1.6 Solute elution order in capillary zone electrophoresis (CZE). Solvent ions have been omitted for simplicity…………………………………… 11
1.7 Schematic of micelle formation in aqueous solution as the surfactant monomer concentration is increased………………………………….. 13
1.8 Elution window for neutral solutes in MEKC…………………………………… 15
1.9 A schematic representation of the torus shape of CDs (side viewof CD in blue) embedded within the molecular structure of β-CD (top view of CD in black)….. 23
1.10 Configuration of the Cu(II) ternary complex formed with aspartame and an amino acid…………………………………………………….. 26
1.11 Molecular structure of 18-crown-6-tetracarboxylic acid ether (18C6H4)………... 27
1.12 Structural formula of two representative macrocycles. Vancomycin consists of three fused rings while rifamycin B contains a single fused ring…… 29
1.13 Molecular electronic states showing the difference between fluorescence (singlet/singlet transition) and phosphorescence (singlet/triplet transition)……... 31
1.14 Jablonski diagram depicting the transitions giving rise to fluorescence emission spectra………………………………………………………………….. 32
1.15 The phenomenon of photoselection where excitation with polarized light results in a cos2 θ probability distribution of excited fluorophores [90]……………………………………………………… 34
ix 1.16 Schematic diagram depicting a typical fluorometer used for the measurement of fluorescence anisotropy………………………………… 37
1.17 Magnetic moments and energy levels for a proton with a spin quantum number of ½……………………………………………………………. 38
1.18 The experimental pulse sequence for a PFG-NMR experiment ………………… 41
1.19 Representative 1-D 1H NMR spectrum showing the ambiguous nature of single dimension NMR a 2-D 1H correlation map…………………….. 43
1.20 A portion of a NOESY spectrum of a BOH–p-(L-SULV) mixture from reference [101]…………………………………………………………………... 44
2.1 Basic structure of the amino acid-based polymeric surfactants.…………………. 57
2.2 Structure of binaphthyl derivatives………………………………………………. 57
2.3 General scheme for the synthesis of amino acid-based surfactants……………… 59
2.4 Enantioseparation of the non-racemic mixture (excess R) of BNA enantiomers using a 6 mM p-(D-SUA) with (a) 0 mM β-CD, (b) 3 mM β-CD, (c) 5 mM β-CD, and (d) 8 mM β-CD. MEKC conditions: 100 mM TRIS / 10 mM borate buffer, pH 10, 30 kV, 254 nm……………………………………………………………. 63
2.5 Bar plots showing the a) selectivity (α) and b) resolution (Rs) trends for the separation of BNA enantiomers while adding incremental amounts of γ-CD to a fixed concentration (6 mM) of D-PS………………………………………………………………….. 65
2.6 Enantioseparation of the non-racemic mixture (excess R) of BNA enantiomers using a 6 mM p-L-SUA with (a) 0 mM β-CD, (b) 3 mM β-CD, (c) 5 mM β-CD, and (d) 8 mM β-CD………………………….. 66
2.7 Bar plots showing the a) selectivity (α) and b) resolution (Rs) trends for the separation of BNA enantiomers while adding incremental amounts of γ-CD to a fixed concentration (6 mM) of L-PS…………………………………………………………………………… 67
2.8 Bar plots showing the selectivity (α) and resolution (Rs) trends for BOH enantiomeric separation when adding incremental amounts of β-CD (panels a and b) or γ-CD (panels c and d) to 6 mM D-PS…………………………………………………... 68
x 2.9 Bar plots showing the selectivity (α) and resolution (Rs) trends for BOH enantiomeric separation when adding incremental amounts of β-CD (panels a and b) and γ-CD (panels c and d) to 6 mM L-PS…………………………………………………... 69
2.10 Bar plots showing the selectivity (α) and resolution (Rs) trends for the separation of BNP using 30 mM D-PS and incremental amounts of β-CD (panels a and b) or γ-CD (panels c and d)…………………………………………………………………... 71
2.11 Bar plots showing the selectivity (α) and resolution (Rs) trends for the separation of BNP using 30 mM L-PS and incremental amounts of β-CD (panels a and b) and γ-CD (panels c and d)…………………………………………………………………... 71
2.12 Enantioseparation of BNP using a) 20 mM γ-CD; b) 30 mM p-(D-SUL); c) 20 mM γ-CD and 3 mM p-(D-SUL); and d) 20 mM γ-CD and 10 mM p-(D-SUL)…………………………………….. 72
3.1 a) Simultaneous separation and enantioseparation of binaphthyl derivatives. b) Simultaneous separation and enantioseparation of paveroline derivatives. c) Simultaneous separation and enantioseparation of benzodiazepines: temazepam (Tzp), oxazepam (Oxp), lorazepam (Lzp)………………………………………………. 84
3.2 a) Simultaneous separation and enantioseparation of four phenylthiohydantoin amino acids (PTH-AAs). b) Simultaneous separation and enantioseparation of 0.2 mg/mL each of the coumarinic derivatives c) Simultaneous separation and enantioseparation of barbiturates (0.4 mg/mL)………………………………….. 90
3.3 Simultaneous separation and enantioseparation of 0.5 mg/mL each of the benzoin derivatives (hydrobenzoin, benzoin, and benzoin methyl ether). Conditions: same as Figure 2(b) except detection was performed at 220 nm……………………………………… 93
3.4 a) Simultaneous separation and enantioseparation of 0.5 mg/mL each of β-blockers b) Simultaneous separation and enantioseparation of 0.5 mg/mL each of aromatic
xi amines c) Simultaneous separation and enantioseparation of 0.5 mg/mL each of aminoglutethimide and glutethimide…………………….. 98
4.1 Molecular structure of the eight analytes investigated. The four dansyl amino acids and TB possess an asymmetric center denoted by the red asterisk, whereas the binaphthyl derivatives possess an asymmetric plane………………………………………… 107
4.2 MEKC enantioseparation of the enantiomers of BNP in the presence of the four diastereomers of poly (SULV)…………………………. 116
4.3 Molecular structure of a) a BNP molecule and b) a p-(L-SULV) surfactant unit indicating the different protons on each molecule……………….. 119
4.4 Plots of NOE intensities observed between the S and R enantiomers of BNP and chiral centers of p-(L-SULV)……………………………………….. 122
4.5 NOESY data for R-BOH and p-(L-SULV) indicating that the R enantiomer of BOH interacts more strongly with the leucine chiral center(left) and the molecular structure of BOH (right)………….. 123
4.6 NOESY data for (+)-TB and p-(L-SULV) indicating that the (+) isomer of TB interacts more strongly with the valine chiral center (left) and the molecular structure of TB (right)……………………. 124
4.7 Titration curves showing the a) enantiomer anisotropy and b) chiral selectivity of BNP as a function of p-(L-SULV) concentration. Peaks shown in boxes correspond with the associated MEKC separation…………………………………………………….. 127
4.8 An α vs. β plot obtained using p-(L-SULV) as the chiral selector. The standard deviation of all β-values presented was less than 1.8…………………………………………………………………. 130
4.9 An α vs. β plot obtained using p-(D-SULV) as the chiral selector. The standard deviation of all β-values presented was less than 1.8…………………………………………………………………. 131
xii LIST OF ABBREVIATIONS
Abbreviation Name
β-CD beta-cyclodextrin
∆∆G° free energy of binding
γ-CD gamma-cyclodextrin
BGE background electrolyte
BNA 1,1′-binaphthyl-2,2′-diamine
BNP 1,1′-binaphthyl-2,2′-diyl hydrogen phosphate
BOH 1,1′-bi-2-naphthol
CD cyclodextrin
CE capillary electrophoresis
CEC capillary electrochromatography
CD-MEKC cyclodextrin-modified micellar electrokinetic chromatography
CMC critical micelle concentration
CZE capillary zone electrophoresis
EOF electroosmotic flow
EKC electrokinetic chromatography
FA fluorescence anisotropy
GC gas chromatography
HPLC high performance liquid chromatography k´ capacity factor
MEKC micellar electrokinetic chromatography
MWCO molecular weight cut-off
xiii NMR nuclear magnetic resonance
NOESY Nuclear Overhauser Effect Spectroscopy
OT-CEC open-tubular capillary electrochromatography
PEM polyelectrolyte multilayers
PFG-NMR pulsed field gradient – nuclear magnetic resonance p-(D-SUA) poly (sodium N-undecanoyl D-alaninate) p-(D-SUL) poly (sodium N-undecanoyl D-leucinate) p-(D-SUV) poly (sodium N-undecanoyl D-valinate) p-(L-SUA) poly (sodium N-undecanoyl L-alaninate) p-(L-SUG) poly (sodium N-undecanoyl L-glycinate) p-(L-SUL) poly (sodium N-undecanoyl L-leucinate) p-(L-SULL) poly (sodium N-undecanoyl L-leucylleucinate) p-(L-SULV) poly (sodium N-undecanoyl-L-leucylvalinate) p-(D-SULV) poly (sodium N-undecanoyl-(D,D)-leucylvalinate) p-(D,D)-SULV poly (sodium N-undecanoyl-(D,D)-leucylvalinate) p-(D,L)-SULV poly (sodium N-undecanoyl-(D,L)-leucylvalinate) p-(L,D)-SULV poly (sodium N-undecanoyl-(L,D)-leucylvalinate) p-(L-SUV) poly (sodium N-undecanoyl L-valinate) p-(L-SUVL) poly (sodium N-undecanoyl L-valylleucinate) p-(L-SUVV) poly (sodium N-undecanoyl L-valylvalinate)
Prop propranolol
Rs resolution
TB Troger’s Base
xiv Trif 2,2,2-Trifluoroanthrylethanol
Warf warfarin
xv ABSTRACT
The scope of this dissertation is to explore the enantiomeric recognition properties of amino acid based polymeric surfactants using a combination of analytical techniques including chromatographic and spectroscopic methods. Chapter 1 includes an introduction to chirality and chiral separations using capillary electrophoresis. Using cyclodextrin modified micellar electrokinetic chromatography (CD-MEKC), the enantioseparation of three binaphthyl derivatives using native β- and γ-cyclodextrins in combination with various diastereomers of chiral polymeric surfactants (PS) was examined. A reversal of enantiomeric order was observed with the enantiomers of (±)1,1′-binaphthyl-2,2′-diamine (BNA) and (±)1,1′- binaphthol (BOH) due to a competition between the two chiral selectors for the same enantiomer. Overall trends observed for the analytes were discussed in terms of selectivity and resolution values for each possible combination of CD and PS. In a subsequent study, the dipeptide micelle polymer, poly (sodium N–undecanoyl-L-leucylvalinate) p-(L-SULV), was used as the sole chiral discriminator in an extensive chiral screening program for the separation of 75 racemic drug compounds. P-(L-SULV) was developed in our lab based on information gleaned from previous studies and our understanding of chiral separation mechanisms. A total of 58 out of 75 compounds were resolved using p-(L-SULV) thereby establishing the dipeptide as a broadly applicable chiral selector for micellar electrokinetic chromatography (MEKC). Complementary analytical methods (MEKC, nuclear magnetic resonance (NMR), and steady-state fluorescence anisotropy) were used to elucidate the chiral separation mechanisms involving the dipeptide p-(L-SULV). An MEKC technique developed by our group was used to determine the primary site of interaction which leads to
xvi the enantioseparation of a select group of analytes. Subsequently, NOESY NMR used as a validation tool in certain instances where MEKC results were inconclusive. For example, while MEKC results were ambiguous regarding the primary site of interaction, NOESY
NMR results strongly suggested that the analyte Troger’s Base interacted primarily with the valine chiral center. In addition, the α vs. β relationship, first introduced by our group, was validated using p-(L-SULV) as the chiral selector. The fluorescence studies revealed a correlation between the chromatographic parameter, selectivity (α) and the spectroscopic parameter, β.
xvii CHAPTER 1.
INTRODUCTION
1.1 Chirality and Its Significance
The root of the word chiral derives from the greek cheir meaning hand. Hence, chirality was first introduced by Louis Pasteur in 1848 to describe the peculiar left and right
“handedness” of sodium ammonium tartrate crystals which demonstrated optical inactivity when mixed and showed equal but opposite optical activity when analyzed independently [1,
2]. Chirality was later more rigorously defined by Lord Kelvin in 1906 as the non- superimposability of a molecule on its mirror image [3]. The individual mirror images that make up a chiral molecule are called optical isomers because solutions of individual enantiomers rotate plane-polarized light in equal but opposite directions. The optical isomer or enantiomer which rotates plane-polarized light in the clockwise direction is designated as the dextro-rotatory (d) or (+)-enantiomer. In contrast, its antipode (e.g., opposite enantiomer) which rotates plane-polarized light in the counterclockwise direction is designated as the levorotatory (l) or (–)-enantiomer. An equal mixture of each of the enantiomers is known as a racemic mixture [4, 5].
Chiral compounds can be divided into three major classes depending on the type of chirality. A simple representation of each type is shown in Figure 1.1. They include chirality due to an asymmetrically substituted atom, as with the amino acid glycine, an asymmetric plane, and an asymmetric helix. In general, compounds are most often designated as chiral due to the presence of an sp3-carbon covalently bound to four different substituents. However, possessing a tetrahedral carbon is not critical for a molecule to exist in different enantiomeric forms. In fact, any asymmetrically substituted atom with
1 tetrahedral geometry can be chiral. Another major class of asymmetry results from the hindered rotation about a central bond and dissymmetric ring substitution, as shown with the biphenyl compound where the two aromatic rings are forced to lay in different planes [4].
These molecules are also commonly called atropisomers, which means molecules that prevent rotation. The last major class of chiral molecules includes those compounds that assume a helical shape, as with a DNA helix, which may have left- and right-handed orientations resulting in chirality. Other examples of helical chirality include the tertiary structures of proteins and polysaccharides.
NO 2 HOOC
COOH (S)
(S) H H3C NH2 O N COOH 2 asymmetric atom asymmetric plane asymmetric helix
Figure 1.1 Structures of the three major types of chiral molecules.
With the exception of rotating plane-polarized light in opposite directions, enantiomers possess identical physical properties in an achiral environment. However, in a chiral environment, such as in living organisms, enantiomers often interact differently due to differences in their three dimensional structures [4]. For example, drug efficacy may reside with a single enantiomer while its antipode may be pharmacologically inactive, demonstrate different pharmacological activity, or may even be toxic.
Table 1.1 is a list of several chiral compounds and the associated activities of their individual enantiomers, some of which are markedly different. The most notorious compound, thalidomide, was developed in 1957 as a sedative used to stem the effects of
2 morning sickness. Soon after its release, a large surge in birth defects swept through
Europe. Infants were born with flipper-like appendages rather than fully developed limbs.
The common denominator in all the cases was that the mothers of these infants had all consumed thalidomide, which was sold as a racemic mixture at the time, early in their pregnancy. It was later determined that the sedative property of thalidomide resides exclusively in the R enantiomer, while the teratogenic activity (i.e., to preclude the normal development of a fetus or embryo) resides in the S enantiomer [6].
Table 1.1 Activities of select chiral compounds [3, 4, 7].
Name Activity
thalidomide* (R)-enantiomer relieves morning sickness; (S)-enantiomer shows teratogenic activity
propranolol (S)-enantiomer has 100 times the adrenogenic blocking activity of the (S)- enantiomer
limonene (R)-enantiomer smells like lemons; (S)-enantiomer smell like oranges
asparagine (D)-enantiomer tastes sweet; (L)-enantiomer tastes bitter
* undergoes in vivo unidirectional conversion from (R)- to (S)-enantiomer
Consequently, a high demand for more effective analytical methods to investigate chiral discrimination of enantiomers was introduced. Because of the demand for more effective and pure drugs, in 1992 the Food and Drug Administration mandated that chiral drugs be marketed as single isomers wherever possible [8]. Since then the pharmaceutical industry has become greatly involved in chiral separations and, as a result, in 2001 single- enantiomer drug sales accounted for 36% of the $410 billion revenues in the worldwide sale of final formulation products [9].
3 Chiral discrimination between enantiomers of a drug requires that a chiral selector recognize both enantiomers stereoselectively (i.e., with different binding constants). While the exact mechanism of chiral discrimination is not well understood, the “three point model”
(Figure 1.2) describes the most accepted scenario of events [10, 11]. In general, the model requires a minimum of three simultaneous interactions between chiral selector and enantiomer for chiral discrimination to take place (Figure 1.2 top). This interaction leads to the formation of transient diastereomeric complexes which result in chiral recognition. Due to spatial restraints, however, the second enantiomer (Figure 1.2 bottom) can achieve one or at most two, of these interactions.
Enantiomers Chiral Selector
A A C C B B
D D
A A B C C B
D D
Figure 1.2 Three point interaction model used to describe the diastereomeric analyte-chiral selector complex.
Historically, chiral separations have been achieved using mature chromatographic methods such as thin layer chromatography (TLC), gas chromatography (GC), and high performance liquid chromatography (HPLC). However, capillary electrophoresis offers a number of practical advantages over these methods such as minimal sample and selector
4 consumption, inexpensive column replacement, high separation efficiency, a diverse application range, and relatively rapid analysis time.
1.2 Fundamentals of Capillary Electrophoresis
Electrophoresis is a separation technique used for separating mixtures of compounds into individual components under the influence of an applied electric field. The separation can be carried out in a buffer solution containing a supporting media, such as paper, cellulose acetate, starch, agarose, and polyacrylamide, or in free aqueous solution [6]. The latter, more contemporary separation method, when carried out within narrow bore capillary tubes is referred to as capillary electrophoresis (CE). CE consists of a family of six related techniques that offer different mechanisms, or modes, of separation. As a result, CE can be used for the separation of a wide variety of analytes including chiral molecules, vitamins, organic and inorganic ions, amino acids, peptides, proteins, and nucleic acids [12]. Table
1.2 describes the six modes of CE and their mechanism of separation. The fundamental properties of electrophoretic separations are similar for nearly all modes of CE and therefore only the theory for the simplest and most commonly used mode (i.e., capillary zone electrophoresis) will be further discussed.
A typical instrument for conducting electrophoresis experiments, illustrated in Figure
1.3, consists of four primary components: a capillary, high-voltage power supply, buffer and sample reservoirs, and a detector. A polyimide coated fused silica capillary serves as the flow-through detector cell. The ends of the capillary are submerged within two buffer reservoirs. Also submerged within the buffer reservoirs are two electrodes (anode/cathode) which serve to complete the electrical circuit between the capillary and the high-voltage power supply [13]. A detection window is sandwiched between a UV source, typically a
5 deuterium lamp, and a photodiode array detector. Common to most commercial CE instruments is an automated sample tray or carousel which holds, positions, and repositions all buffer and analyte vials. Sample introduction is achieved via hydrodynamic (i.e., pressure, vacuum, or gravity) or electrokinetic (i.e., voltage) injection. The separation is initiated by the application of a voltage between 0-30 kV across the capillary. The separation is subsequently monitored via a data acquisition system (e.g., personal computer).
Table 1.2 Brief description of six modes of capillary electrophoresis.
Capillary Electrophoresis Mode Mechanism of Separation
Capillary Zone Performed in aqueous buffer; solutes migrate in discrete Electrophoresis (CZE) zones at different velocities due to differences in charge-to- size ratio
Capillary Ioselectric Separates peptides and proteins in aqueous buffer on the Focusing (CIEF) basis of their isoelectric point (pI)
Capillary Gel Performed in a polymeric network that acts as a molecular Electrophoresis (CGE) sieve
Capillary Isotachophoresis Solutes are separated into zones which are sandwiched (CITP) between leading and terminating electrolytes of known mobility
Micellar Electrokinetic Essentially the same as CZE with the addition of Chromatography (MEKC) hydrophobic/ionic interactions with the micellar phase
Capillary Similar to HPLC in that the solute partitions between an Electrochromatography immobilized stationary phase and a mobile phase (CEC)
6 1.3 Capillary Zone Electrophoresis (CZE)
The primary driving force for separations in all modes of CE is the electroosmotic flow (EOF). In addition, the separation mechanism governing CZE is the electrophoretic mobility (mobility of a charged solute due to an electric field) of a solute molecule. The following section describes how the above-mentioned forces are derived and their effects on solute separation.
Capillary
High-Voltage Detector Power Supply + window - UV Source i
Buffer Reservoirs
Figure 1.3 Schematic diagram of a typical capillary electrophoresis system.
The EOF is defined as the bulk movement of solvent resulting from the presence of an electric double layer, known as the zeta potential (ζ ), at the silica–water interface
(Figure 1.4). Three important events lead to the development of the EOF. First, under basic conditions, the silanol groups (SiOH) lining the inner surface of the capillary are hydrolyzed to the deprotonated form (SiO−). Next, in order to maintain charge balance, the counterions from the buffer or electrolyte solution electrostatically adsorb to the negatively charged wall.
This rigid double layer, often referred to as the Stern layer, reaches equilibrium with an
7 outer, more diffuse layer aptly named the diffuse layer or the Debye-Hückel layer. Finally, when voltage is applied, an excess of cations found in the bulk solution migrate toward the cathode producing a net flow from anode to cathode. Since the cations are solvated by water molecules, their movement drags the bulk solution toward the cathode as well. The magnitude of the EOF is described in terms of velocity (ν ) or mobility ( µ ) as follows;
vEOF = (/εζ η)E (1.1)
µEOF = (/εζη) (1.2) where ε is the dielectric constant,ζ is the zeta potential, η is the solvent viscosity, and E is the electric field generated upon application of a constant voltage across the length of the capillary [13].
+ _
Capillary Surface = SiO- ______+ + + + Stern Layer + + + + + + + + + + + + + + _ + _ _ + ______+ _ + _ + +
+ + Diffuse Layer + _ _ + + _ + + + EOF _ _ _ + +
+ + ______
+ + + + + + + + + + + + + + + + + + + + _ _ + ______
Figure 1.4 Schematic representation of the double layer formation.
8 Figure 1.5 demonstrates one of the advantages of electrically driven systems (i.e.,
CE) over pressure driven systems (i.e., HPLC). In CE, the EOF flow stream originates at the walls of the capillary, producing a plug-like plow that, with the exception of a few nanometers near the surface, has a flat velocity profile across the capillary diameter [6, 14,
15]. It is this flat laminar profile that facilitates higher peak efficiencies in electrokinetically driven separation techniques. Conversely, in pressure driven systems, the frictional forces at the solid–liquid interface cause large pressure drops across the capillary resulting in broad peaks with low efficiency.
a) Electrically Driven Flow
Laminar profile
b) Pressure Driven Flow
Parabolic profile
Figure 1.5 (a) Electrically driven flow (e.g., EOF) and corresponding solute zone profile, (b) Pressure driven (or hydrodynamic) flow and corresponding solute zone profile.
In the presence of an electric field, ions travel at a constant velocity that is simply a
product of the electrophoretic mobility of the ion ( µe ) and the electric field strength (E).
v= µe E (1.3)
9 In addition, the ions experience two opposing forces in the presence of the static electric
field. Ionic species experience an electrostatic force ( Fe ) which is proportional to the electric field strength and the charge (q ) of the particular ion:
Fe = qE (1.4)
This electrostatic force ( Fe ) accelerates ions toward the oppositely charged electrode. In a
viscous medium, a counteracting frictional force ( Ff ) retards the mobility of the ions.
According to Stokes’ law for spherical particles, this frictional force can be expressed as follows,
Ff = 6πηrv (1.5) where η is the viscosity of the liquid, r is the radius of the ion, and v is the migration velocity [16]. The ions experience two opposing forces that reach equilibrium in the presence of a steady state electric field and can be expressed as
qE = 6πηrv (1.6)
Substituting Equation 1.6 into Equation 1.3 gives rise to the electrophoretic mobility ( µe ) of a solute molecule.
q µ (cm21V −−s 1)= (1.7) e 6πηr
Equation 1.7 suggests the electrophoretic mobility is determined by the charge-to-size ratio
( qr). That is, small, highly charged solutes will migrate faster than larger solutes of similar charge. Moreover, for neutral solutes, with charge q = 0 , the corresponding electrophoretic mobility is zero.
From the previous discussion we learned that both the EOF and electrophoretic mobility affect the overall mobility of solute molecules. Therefore, the apparent mobility
10 ( µapp ) is given by the algebraic sum of the electrophoretic mobility of the ion and its mobility due to EOF as shown below.
µapp = µe + µEOF (1.8)
Since the magnitude of the EOF is often more that an order of magnitude greater than the electrophoretic mobility of the ion, all species, regardless of charge, are forced toward the cathode [14]. The order of elution, however, depends on the actual charge of the ion (Figure
1.6). Cations migrate fastest because both the electrophoretic mobility and the EOF are
positive. Since neutrals have an electrophoretic mobility equal to zero, µapp = µEOF and neutrals migrate at the same rate as the EOF. Neutrals therefore cannot be distinguished from one another using CZE. The last solutes to elute are the anionic molecules, since their electrophoretic mobility is negative.
______
+ ++ + +
_ N + _ _ + _ _ _ + _ N + + + _ _ _ _ + + ++ _ _ _
_ + +
_ + + _ EOF + + + _ _ + + _ + _ +
_ _ +
_ + _ _ N + anode _ _ + _ cathode _ + _ _ + + _ _ _ N N + _ +
+ + ______
Figure 1.6 Solute elution order in capillary zone electrophoresis (CZE). Solvent ions have been omitted for simplicity.
1.4 Micellar Electrokinetic Chromatography (MEKC)
The theory described thus far for CZE is only applicable for the separation of charged solutes. Terabe et al., however, describe a technique that allows for the separation of neutral species [17]. This method combines CZE with the idea of the differential
11 partitioning of a solute between a mobile and stationary phase borrowed from conventional
HPLC. In doing so, molecular aggregates, known as micelles, are introduced into a CZE buffer solution. The micelles serve as a slow moving stationary phase or so-called
“pseudostationary” phase. Neutral, hydrophobic solutes can then partition between the pseudostationary micellar phase and the surrounding aqueous mobile phase, thus producing a separation. Terabe later named the technique micellar electrokinetic chromatography
(MEKC).
Micelles are aggregates of surfactant monomer units. Surfactants are composed of two parts – a hydrophilic “head” group and a long hydrophobic “tail” (Figure 1.7). This property makes them soluble in both nonpolar and polar reagents. In general, surfactants are classified according to the charge of the head group (anionic, cationic, zwitterionic, and nonionic). Table 1.3 lists representative surfactants from the four major classes as well as select properties [18]. At low concentrations, surfactants position themselves in such a way so as to reduce the negative interactions between a polar solvent and the nonpolar tail groups. However, at high surfactant concentrations, exceeding a value known as the critical micelle concentration (CMC), the surfactant units are forced to spontaneously organize into spherical micelles or aggregates of different shape in order to reduce the unfavorable interaction between solvent and tail groups. As a result, the surface tension of the bulk solution is often dramatically altered. In aqueous solution, the surfactant units aggregate such that the hydrophilic head groups form an outer shell. The tails, in turn, shielding themselves from the aqueous solution, reside within a hydrophobic core available for solubilizing non-polar species. Once formed, micelles are in a state of dynamic equilibrium where they rapidly interconvert between a single surfactant unit and a spherical micelle.
12 The average number of surfactant units per micelle, or aggregation number, and CMC depend on the nature of the head group, length of the hydrophobic tail, solvent composition, and temperature.
As mentioned previously, the separation mechanism of MEKC is based on column chromatographic principles involving true stationary phases. In MEKC however, micelles are not truly stationary. Hence, modified chromatographic expressions are used to describe the theory of differential partitioning.
Surfactant monomer Surfactant monomers Micelles aggregating above CMC
Head group
Tail
concentration
Figure 1.7 Schematic of micelle formation in aqueous solution as the surfactant monomer concentration is increased.
The separation of two solutes is best achieved when there are differences in the partitioning of solutes between the micellar and aqueous mobile phases. Consider a surfactant with an anionic polar head group. Due to its charge, the corresponding anionic micelles will electrophorese slowly to the anode while being dragged toward the detector by the bulk flow. Thus, the overall velocity of the micelle and anything it solubilizes is reduced
13 Table 1.3 Critical micelle concentration (CMC) and aggregation number (AN) of select surfactants [19]. Surfactant CMC, mM AN Anionic Deoxycholic acid, sodium salt 5 4-10 Sodium dodecyl surfate (SDS) 8.27 62 Taurocholic acid, sodium salt 10-15 4 Cationic Cetyltrimethylammonium chloride 1 ND Cetyltrimethylammonium bromide 1.3 78 Dodecyltrimethylammonium bromide 14 50 Hexadecyltrimethylammonium bromide 0.0026 169 Zwitterionic CHAPSa 8 10 CHAPSOb 8 11 Nonionic N-Decyl-β-D-glucopyranoside 2.2 ND Triton X-100 0.24 140 a3-[3-cholamidopropyl)dimethylammonio]-1-propanesulfonate b3-[3-cholamidopropyl)dimethylammonio]-2-hydroxy-1-propanesulfonate ND, not determined
compared to bulk flow. Since the solute will partition into and out of the micellar phase, its mobility will also be slowed. The distribution coefficient ( K ) describes this partitioning of a solute molecule between the two phases in terms of concentrations,
[Ss ] K = (1.9) [Sm ]
where [S s ] is the molar concentration of the solute in the micellar pseudostationary phase
and [S m ] is the molar concentration of solute in the mobile phase.
The distribution coefficient, in turn, is used to describe the capacity factor ( k ′ ) or retention factor as follows. In addition, k ′ can be described as a function of adjusted retention times of the solute,
14 ⎛ V ⎞ (ttr − 0 ) kK′ = ⎜ S ⎟ = (1.10) ⎝ VM ⎠ ⎛ tr ⎞ t0 ⎜ 1 − ⎟ ⎝ tmc ⎠
where VS and VM are the respective volumes of the micellar and aqueous phases, tr is the
migration time of the analyte, t0 is the retention time of the unretained solute and tmc is the retention time of the solute completely solubilized by the micelle that migrates at the same
velocity as the micelle. Hence, the retention time of a given solute tr should appear in the
range tt0 < Elution W indow Solutes M icelle EOF 0t0 t 1 t 2 t 3 t mc Time Figure 1.8 Elution window for neutral solutes in MEKC. The ratio of the adjusted retention times (tr − t0 ) of two adjacent solutes gives the selectivity factor tt− α = 20 (1.11) tt10− 15 where t1 and t2 are the migration times of the first and second eluting solutes, respectively. The selectivity describes the ability of a system to selectively interact with one solute over the other. A selectivity of unity suggests two solutes are indistinguishable to the system and have identical migration times. Resolution, Rs , expresses the ability of the system to separate two closely eluting species. A commonly used equation based on chromatographic parameters can be used to compute resolution. ⎛ t ⎞ ⎜ 1− 0 ⎟ ⎛ N ⎞⎛ α −1⎞⎛ k′ ⎞⎜ t ⎟ ⎜ ⎟ ⎜ 2 ⎟ mc Rs = ⎜ ⎟⎜ ⎟⎜ ⎟ (1.12) ⎜ 4 ⎟⎝ α ⎠ k′ +1 ⎛ t ⎞ ⎝ ⎠ ⎝ 2 ⎠⎜1− ⎜ 0 ⎟k′ ⎟ ⎜ ⎜ ⎟ 1 ⎟ ⎝ ⎝ tmc ⎠ ⎠ Efficiency Selectivity Retention In Eq. 1.12, the efficiency, N, refers to the number of theoretical plates, and k1′ and k2′ are the capacity factors for the first and second eluting peaks, respectively. The efficiency, in turn, can be obtained experimentally by 2 ⎛ t ⎞ N = 5.54⎜ n ⎟ (1.13) ⎜ ⎟ ⎝ w1 2 ⎠ where w12 is the peak width at half its maximum height, and tn refers to the elution time for peak n [13]. Equation 1.12 states that the resolution between two adjacent species can be improved by optimizing the efficiency, selectivity, and/or retention factor. Selectivity and efficiency are dependent on a number of parameters (i.e., zone dispersion, ionic strength and pH of buffer, strength of applied field, and temperature) and should be optimized to obtain 16 an optimum Rs . However, when calculating peak resolution experimentally, as opposed to theoretically as above, a more practical equation is used 2(tt2− 1) Rs = (1.14) w12+ w where w1 and w2 are the widths at the base of peaks 1 and 2, respectively, in units of time. 1.5 Chiral Selectors for Chiral CE The resolution of optically active molecules is an area of immense interest to many areas of the natural sciences, in particular to the pharmaceutical industry. Considering that enantiomers of a given compound possess identical physical properties (e.g., melting point, boiling point, molecular weight, density, vapor pressure, etc.) in an achiral environment, the act of separating enantiomers into single isomers is a massive undertaking. For example, it is only when the enantiomers are in the presence of a chiral additive or chiral selector that they have the potential to form transient diastereomeric complexes which result in differences in their observed mobilities. There are a number of chiral selectors available for enantioseparation studies in CE, many of which have been borrowed from more conventional methods. In fact, many chiral selectors currently available for use in CE were initially used as HLPC and/or TLC chiral selectors [20-25]. Polymeric surfactants, cyclodextrins, ligand-exchange metal complexes, crown ethers, and macrocyclic antibiotics are some of the most commonly used chiral selectors and will be discussed in greater detail in the sections that follow. 1.5.1 Polymeric Surfactants Larrabee and Sprague introduced polymeric surfactants also known, as molecular micelles, to circumvent the problems associated with limited reproducibility observed in 17 MEKC separations resulting from the dynamic equilibrium between conventional micelles and surfactant monomers [26]. The authors chose to eliminate the dynamic equilibrium found in conventional micelles through the formation of covalent bonds between surfactant monomers. As a result, polymeric surfactants are defined as high molecular weight macromolecules containing a polymerizable group(s) (e.g., vinyl group) that result(s) from the polymerization of conventional surfactants concentrations above the CMC. There are several methods available for polymerization that include chemical initiation, ultraviolet light initiation, and γ-irradiation. The γ-irradiation method, for example, involves exposing a surfactant containing a polymerizable group to γ-radiation resulting in polymerization via free-radical formation. Three major factors directly influence the polymerization kinetics and shape of the resulting polymerized micelles: concentration of surfactant solution, nature of the solvent (e.g., temperature, pressure, ionic strength, presence of additives), and the molecular structure of the surfactant. For example, as the concentration of the surfactant solution increases, the shape of the micelle changes from spherical to cylindrical, to hexagonal, and finally to a lamellar geometry [27]. While the true shape of polymeric surfactants has not been determined, the polymeric surfactants used in this dissertation are presumed to be similar to conventional micelles, spherical or nearly so. Polymeric surfactants bearing an amino acid head group were first introduced in the early 1990s by Hara and Dobashi [28] and later by Wang and Warner [29]. This pseudostationary phase is preferred over conventional micelles primarily because the former possess a covalent linkage between adjacent surfactant monomers and, as a result, solutes are not able to penetrate as deeply into the core of the micelle. This, in turn, facilitates a 18 faster rate of mass transfer and, consequently, an increase in the separation efficiency is observed [30]. In addition, Billiot et al. observed enhanced resolution in the separation of the enantiomers of lorazepam, temazepam, 1,1'-bi-2-naphthol, and propranolol when using one of the 18 polymeric surfactants as the pseudostationary phase in MEKC compared to their monomeric counterparts [31]. Compared to traditional micelles, polymeric surfactants exhibit: i) enhanced stability especially in presence of organic solvents, ii) rigid size and structure resulting from the covalent bond rather than the weak forces that result in self- assembly iii) elimination of the CMC requirement, and iv) the ability to be coupled with mass spectrometric detection [32]. Since polymeric surfactants used in this dissertation were amino acid based, the following discussion will focus solely on studies employing amino acid based polymeric surfactants. Warner and co-workers compared two polymeric, chiral, anionic surfactants [poly (sodium N-undecanoyl L-valinate) p-(L-SUV) and poly (sodium N-undecanoyl-L-valiyl valinate) p-(L-SUVV)] for the enantioseparation of basic, acidic, and neutral enantiomers [33]. In this study, the authors showed that chiral recognition was significantly enhanced with the dipeptide, poly-L-SUVV, as compared with that of the monopeptide, poly-L-SUV, presumably due to the additional chiral center present in the former. In a later study, Billiot et al. studied the enantiorecognition ability of the dipeptide polymers of sodium N- undecanoyl L-leucylleucinate) (L-SULL), sodium N-undecanoyl L-valylleucinate (L-SUVL), and sodium N-undecanoyl-L-leucylvalinate (L-SULV), compared to the single amino acid polymers of sodium N-undecylenyl-L-leucine (L -SUL) and sodium N-undecylenyl-L-valine (L-SUV) [34]. The results mirrored those of the earlier work where the performance of the dipeptide polymers was greater than that of the single amino acid polymers. 19 While the latter two studies established that dipeptide polymeric surfactants are more effective in separating basic, acidic, and neutral enantiomers than the corresponding single amino acids surfactants, subsequent studies concentrated on the effect, if any, of the position and number of chiral centers [34, 35], amino acid order [34], steric effects [36, 37], and the polydispersity of the polymeric surfactants [38]. The studies demonstrated that the polymeric surfactants were polydispersed macromolecules and the most favorable enantiorecognition ability was observed using dipeptide polymeric surfactants with the bulkier amino acid located in the innermost position (i.e., the N-terminus) and the less bulky amino acid located in the outermost position (i.e., the C-terminus). Information gleaned from these previous studies served as the impetus for developing a single, robust chiral selector that could resolve a wide array of analytes independently of additional modifiers. The dipeptide chosen for this endeavor, poly (sodium N-undecanoyl-L- leucylvalinate) [p-(L-SULV)] was shown to have a 77% overall success rate in separating 75 neutral, cationic, and anionic drug compounds which were randomly pulled off the shelf [39]. Furthermore, poly (L-SULV) was most successful in separating the enantiomers of anionic and neutral compounds with a 92 and 81% success rate, respectively, whereas the anionic compounds were separated with a 29% success rate seemingly due to electrostatic repulsions with the anionic head group. This work is described in detail in much greater detail in Chapter 3 of this dissertation. One of the latest applications of amino acid based polymeric surfactants has been in the area of polyelectrolyte multilayer (PEM) coatings. Such PEM coatings are used in the fabrication of open tubular columns for use in open-tubular capillary electrochromatography (OT-CEC). In this format, stable coatings are constructed on the inner walls of the capillary 20 through the layer-by-layer deposition of positively and negatively charged polymers which are held together via electrostatic forces. Kapnissi-Christodoulou et al. first introduced a PEM coating using poly (sodium N-undecanoyl L-glycinate) as the anionic polymer for the achiral separation of seven benzodiazepines [40]. This study established the robustness of the coating which gave a relative standard deviation value for the EOF retention time of less than 1% for the 200+ runs accomplished using a single column. Later PEM-coated capillaries were investigated for the chiral separation of the enantiomers of various compounds including several barbitals and benzodiazepines using p-(L-SULV) as the anionic polymer [41]. In this case, the endurance of the PEM coatings proved to be as robust as with the achiral coatings however, modifications to the polyelectrolyte solutions were required. The interested reader is directed to selected excellent reviews and articles that provide the most recent developments in the areas of OT-CEC and PEMs [41-46]. 1.5.2 Native Cyclodextrins and Their Derivatives Undoubtedly, the most commonly used family of chiral selectors in CE are cyclodextrins (CDs). CDs are naturally occurring cyclic oligosaccharides derived from the enzymatic digestion of starch. Their basic structure is that of a torus-shaped cylinder comprising six, seven, or eight glucopyranose units attached by α-(1,4)-linkages and are referred to as α-, β-, and γ-CD, respectively. The physical properties of the three native CDs differ in width of the cavity, solubility, molecular mass, etc. with the exception of their depth which is the same for all three. The main physicochemical properties of native CDs are summarized in Table 1.4 while Figure 1.9 shows a schematic drawing of a CD molecule embedded within the molecular structure of β-CD. The schematic demonstrates the ability of CDs to form 21 inclusion complexes with solute molecules provided the size and shape of the solute fits within the dimensions of the torus. The inner cavities of CDs are lined with hydrogen atoms and glycosidic oxygen bridges which favor hydrogen bonding and hydrophobic interactions between a guest (i.e., a solute molecule) and the CD host. Due to their rather large dipole moments, CDs also possess the ability to bind other molecules via dipole-dipole interactions [47]. Still another characteristic of CDs is that each glucose unit that comprises the molecule possesses five chiral centers. Hence, not only can CDs form intermolecular complexes involving hydrophobic and dipole-dipole interactions, but they can also form stereoselective intermolecular complexes. Table 1.4 Important characteristics of α-, β-, and γ-CDs [47]. Characteristics α-CD β-CD γ-CD Number of glucose units 6 7 8 Molecular weight (g/mol) 972 1135 1297 Solubility in weight (g/100 mL) 14.5 1.85 23.2 Cavity diameter (Å) 4.7–5.3 6.0–6.5 7.5–8.3 Height of Torus (Å) 7.9 ± 0.1 7.9 ± 0.1 7.9 ± 0.1 pKa 12.33 12.20 12.08 Number of chiral centers 30 35 40 22 HO OH HO O O O OH O HO OH HO O OH O HO OH O O HO OH OH HO O O O H OH O HO OH O O HO OH O HO Figure 1.9 A schematic representation of the torus shape of CDs (side view of CD in blue) embedded within the molecular structure of β-CD (top view of CD in black). Native CDs possess 18 (α-CD), 21 (β-CD), or 24 (γ-CD) substitutable hydroxyl groups which makes them amenable to derivatization methods. Neutrally derivatized CDs such as dimethylated- (DM), trimethylated- (TM), hydroxyethylated- (HE), hydroxypropylated- (HP), acetylated- (Ac)-CDs are commonly used in CE for the enantioseparation of charged basic [48-50] and acidic analytes [51, 52]. Snopek et al. were the first to report the use of CDs as chiral reagents [53]. In this study, neutral β-CD, DM-β-CD, and TM-β-CD were used for the enantioseparation of ephedrine and its analogs in the CITP mode of CE. While TM-β-CD did show sufficient enantioselective behavior toward pseudoephedrine and ephedrine, the simultaneous separation of the enantiomers of both compounds were observed upon the addition of 10 mM DM-β-CD. In addition, they found that the addition of β-CD to the leading electrolyte led to the successful enantioseparation of the enantiomers of pseudoephedrine but not of ephedrine. 23 Later, Koppenhoefer and co-workers examined the effect of cavity size on the enantioselectivity of three methylated CDs [54]. They found the enantioselectivity of methylated β-CD to be markedly greater than that of methylated α- and γ-CD, presumably due to its cavity dimension. Furthermore, also due to its dimensions, β-CD was found to readily accommodate benzene ring, aryl, and heterocycle containing compounds [48, 49]. Charged CDs often exhibit higher chiral resolving capabilities than neutral CDs and are typically used to separate uncharged enantiomers. Anionic derivatives of β-CD, such as methylamino- (MA), sulphobutylether- (SBE), carboxymethylated- (CM), heptakis(2,3-di- O-acetyl-6 sulfo)- (HDAS), sulphated- and phosphated-CDs, are the most commonly used charged CDs [48, 55-62]. 1.5.3 Ligand-Exchange Metal Complexes Chelating agents, which are based on the idea of ligand-exchange, are still another type of chiral selector. The concept of ligand-exchange chromatography (LEC) was suggested as early as 1961 by Helfferich [63]. Nearly a decade later, in the early 1970s, Davankov and Rogozhin further developed the technique and transformed it into a powerful chiral chromatographic method for use in HPLC analysis [20]. It was not until the mid-to- late 1980s that Zare’s group successfully applied the ligand exchange (LE) technology to CE and also performed the very first enantioseparation in CE [64, 65]. Zare and co-workers demonstrated the great potential of chiral CE when they used the Cu(II) complexes of D- and L-histidine as chiral selectors for the enantioseparation of dansyl amino acids. Later the same group used a more efficient chiral selector (i.e., Cu(II)-aspartame) to successfully resolved 14 racemic amino acid derivatives in a single run in under 12 minutes. The 24 interested reader is directed to several review papers describing recent developments in LE- CE [66-68]. LEC separations are based on the swapping of analyte and ligand in the coordination sphere of a central ion. Analytes following this criterion include amino acids, hydroxyl acids, some dipeptides and polypeptides, amino alcohols, and diamines [47]. The most commonly used central ion is the Cu(II) cation because it forms metastable complexes with amino acids. Furthermore, various chiral selectors including histidine, proline, and tartaric acid derivatives are commonly used in LEC. Cu(II) ions readily react with aspartame molecules to form a Cu(II)-(aspartame)2 complex (Figure 1.10). The Cu(II)-(aspartame)2 complex comprises a six-membered ring formed via the α-amino and the β-carboxy groups of the aspartyl residue of the aspartame molecule. However, in the presence of an amino acid, a more stable five-membered ring involving the α-amino and the β-carboxy groups of the amino acid is preferred over the six- membered ring (Figure 1.10). Therefore, when an amino acid is added to an electrolyte solution containing the Cu(II)-(aspartame)2 complex, it can easily replace one aspartame ligand, leading to the formation of a ternary complex [65]. Chiral resolution stems from the difference between the complex stability constants of the two mixed ternary complexes with each analyte enantiomer [69]. The stability of each ternary complex, in turn, depends on a variety of factors (i.e., steric, electrostatic, hydrogen bonding, and π-π interactions) [47]. The limitations of ligand-exchange selectors arise due to their limited stability and UV detection difficulties. 25 R R R O NH O HN NH O O NH2 O NH H N H 2 2 Cu Cu O N H2 R' O O O O O Aspartame Aspartame Aspartame Amino acid O H C R = OCH3 C H 6 5 Figure 1.10 Configuration of the Cu(II) ternary complex formed with aspartame and an amino acid. 1.5.4 Crown Ethers Crown ethers are another viable option for synthetic chiral selectors for CE. First discovered in 1967 by Pedersen, crown ethers are synthetic macrocycles composed of ethylene groups bonded by ether bridges [70]. They typically contain electron-donating heteroatoms such as N, O, P, S, and Se. In addition, they possess the ability to form stable and selective complexes with alkali, alkaline-earth, and primary ammonium cations. Figure 1.11 shows the molecular structure of the most common crown ether used to date for CE, 18-crown-6-tetracarboxylic acid (18C6H4). Due to the central cavity, the separation mechanism of crown ethers is similar to that of CDs in that they are able to form guest-host complexes. This ability allows crown ethers to form diastereomers with different complex formation constants [71]. 26 O HOOC O O COOH O O COOH HOOC O Figure 1.11 Molecular structure of 18-crown-6-tetracarboxylic acid ether (18C6H4). In 1984, Stover used crown ethers to separate the hydrochloric acid and chloride salts of potassium, sodium, ammonium, lithium, and rubidium using isotachophoresis [72]. Later, Kuhn et al. introduced, 18C6H4, as new chiral selector for the CE separation of several underivatized amino acids, drug racemates, and peptides [73-75]. Their studies demonstrated the suitability of crown ethers in separating protonated primary amines. Secondary, tertiary amino groups and other functional groups proved to lack the adequate complexation ability necessary for chiral recognition. Furthermore, they found that the performance of the separation was strongly influenced by buffer composition. Dutton et al. later found that the carboxylic groups of the macrocycles have pKa values of 2.13, 2.84, 4.29, and 4.88, causing an increase the net negative charge of the crown ether with increasing pH [76]. Since carboxylate anions help stabilize the complex by cooperative crown ether and carboxylate binding interactions, an increase in pH will increase the complexation of a chiral analyte with 18C6H4. However, an increase in pH will cause a decrease in the net charge of the amine, thereby hindering the complexation ability of C18C6H4. The method has, therefore, been limited to separating only underivatized amino acids under acidic conditions where the amine is protonated. 27 Because complexes between alkali metal ions and crown ethers are approximately an order of magnitude more stable than those with primary amines, careful attention must be taken in choosing a background electrolyte (BGE) that does not contain alkali metal ions [77]. TRIS-citric acid buffer is the BGE of choice when using 18C6H4 as a chiral selector. The enantiomers of amino alcohols [78], as well as the analogues of DOPA and γ- aminobutyric acid (GABA) [79] were separated with crown ethers. The effect of chemical structure of analytes on CE separations was investigated by Kuhn [80]. Separations involving dipeptides [81] and tripeptides [82] have also been reported. Kuhn studied the synergistic effect on the resolution of chiral amines when crown ethers and a CD were dissolved in the same BGE [75]. The primary limitation of using crown ethers as selectors instead of cyclodextrins is that the separation efficiency of the former is about 5-fold lower than the latter [73]. 1.5.5 Macrocyclic Antibiotics Macrocyclic antibiotics are a relatively new class of chiral selectors introduced by the Armstrong group for the enantioseparation of amino acids [83]. They possess the following general characteristics: i) medium molecular weight, ii) complex structure, iii) are of biological origin, iv) were once used as drugs, and v) show high UV absorbance [84]. The latter requires that experiments performed using macrocyclic antibiotics be done using indirect detection or partial filling mode. This class of compounds includes vancomycin, rifamycin B, ristocetin A, teicoplanin, and others. The structures of two common macrocycles, vancomycin and rifamycin B, are shown in Figure 1.12. Vancomycin has a characteristic basket shape with three internal macrocyclic cavities and two side chains. In addition, this molecule contains five aromatic 28 rings, 18 stereogenic centers, nine hydroxyl groups, two amine groups, seven amido groups, and two chlorine moieties. Rifamycin B, however, possesses one each of carboxylic acid, carboxymethyl, hydroxyl, amide, and aromatic groups; nine stereogenic centers; and a ring structure formed with the chromophore spanned by an aliphatic chain. Because of the diversity of functionality of this and other macrocycles, enantioseparation is possible via several different mechanisms such as π-π interactions, hydrogen bonding, peptide and carbohydrate bonding [85], inclusion complex formation [86], and affinity interaction [87]. HO HO OH O HOOC O O O COOH NH2 O H O HN N O O H N N H O HN O O OH Cl Cl HN OH OH OH NH O O O O OH O O HO O O O OH NH2 HOH2C OH OH Vancomycin Rifamycin B Figure 1.12 Structural formula of two representative macrocycles. Vancomycin consists of three fused rings while rifamycin B contains a single fused ring. 1.6 Theory of Fluorescence Spectroscopy To understand fluorescence spectroscopy one must first understand the idea of luminescence and the associated Jablonski diagram, both of which will be discussed in detail in the following section. 29 Luminescence is defined as the emission of light occurring from an electronically excited state from any substance [88]. Since a system emitting luminescence is continually losing energy, an external source of energy must be supplied in order for the luminescence to persist. The source from which the energy is derived is typically used to classify most types of luminescence. The light emitted from a gallium arsenide laser for example, is called electroluminescence and is derived by the passage of an electric current through the semiconductor p-n junction. Radioluminescence and chemiluminescence are induced by the high energy particles of a radioactive material and the energy from a chemical reaction, respectively. Similarly, photoluminescence is fueled by the absorption of infrared, visible, or ultraviolet light. Since the excitation of fluorescence and phosphorescence is brought about by the absorption of photons, both are collectively known as photoluminescence events. Fluorescence differs from phosphorescence in the nature of the excited state. For instance, consider a pair of electrons in the ground state (Figure 1.13). If the transition of one electron to a higher electronic level is paired (i.e., have opposite spin) to the second electron in the ground state, a singlet state is formed. Consequently, the return of the photon to the ground state is said to be spin-allowed and therefore occurs rapidly (~10-8s). Phosphorescence, however, arises from the emission of a photon from a triplet excited state where the electron in the excited state has the same spin as that of the ground state electron. Subsequent transitions to the ground state are deemed “forbidden”, thus the emission rate is several orders of magnitude slower for phosphorescence (10-4 to 102s) compared with fluorescence [88]. 30 Excited singlet Excited triplet Ground singlet Figure 1.13 Molecular electronic states showing the difference between fluorescence (singlet/singlet transition) and phosphorescence (singlet/triplet transition). Only processes leading to fluorescence will be discussed in greater detail. A Jablonski diagram shown in Figure 1.14 best depicts the transitions giving rise to absorption and fluorescence emission spectra. The singlet ground, first, and second excited states as well as the triplet excited electronic state are depicted by the thicker horizontal lines labeled S0, S1, S2, and T1, respectively. Several vibrational levels are associated with each of these four electronic states and are represented by the thinner horizontal lines, denoted by ν0, ν1, ν2, ν3, and ν4. For simplicity, the rotational energy levels are not shown. Transitions between the different energy levels are shown as vertical lines. At room temperature, thermal energy is generally not sufficient to populate the higher energy vibrational levels of the ground state therefore most electrons reside in the lowest vibrational level of the ground electronic state, S0. Electrons can be promoted to any of several excited singlet states during the absorption process as shown in Figure 1.14. If an electron is excited to a higher energy excited state, such as S2 or higher, it subsequently relaxes to the lowest vibrational level of S1 via radiationless internal conversions and vibrational relaxation. In addition, when in 31 Singlet Excited State (S) Internal conversion (10-14 -10-11 s) Vibrational Radiationless relaxation deactivation S2 S1 y erg n E Absorbance Fluorescence (10-15 s) (10-9-10-7 s) ν 4 Vibrational ν3 levels of ν2 ν1 ground state S 0 ν0 (ν ) 0 Figure 1.14 Jablonski diagram depicting the transitions giving rise to fluorescence emission spectra. The approximate rate of transitions is shown in parenthesis. Solid lines denote transitions involving the absorption of a photon, dotted lines denote transitions involving the emission of a photon, and curved lines denote radiationless transitions [12]. solution, excess vibrational energy is lost via vibrational relaxation (10-12s) due to collisions between molecules of the excited species and those of the solvent. It is because of these processes that solution phase fluorescence nearly always involves a transition from the lowest vibrational level of the first excited singlet state to the ground state (So). Radiationless deactivation to the ground state offers another means for electrons to return to the ground state without the emission of a photon. 32 1.6.1 Polarization of Photoluminescence There are many photoluminescence applications available in the physical, chemical, biological, and medical sciences. In fact, fluorescence techniques are routinely applied for quantitative studies, qualitative identification, and structural characterization of organic and inorganic compounds [89]. Furthermore, fluorescent probes are frequently used to explore the polarity, molecular mobility, pH, electric potential, or fluidity of organized media such as polymeric and micellar systems, biological membranes, proteins, nucleic acids, and living cells. One technique in particular, fluorescence anisotropy, can provide useful information on molecular mobility, size, shape, and flexibility of molecules, fluidity of a medium, and order parameters (e.g., in a lipid bilayer) [90]. When a population of fluorophores is illuminated by linearly polarized light (e.g., along the z-axis), molecules whose absorption transition moments are parallel to the electric vector of the incident light are preferentially excited. In order to absorb polarized light, the transient dipole moment of the fluorophore need not be precisely aligned about the z-axis, but the probability of absorption is proportional to cos2 θ, where θ is the angle the absorption dipole makes with the z-axis [90]. Consequently, excitation with polarized light results in a population of excited fluorophores which are symmetrically distributed about the z-axis (shown by the cones in Figure 1.15). This phenomenon of selective excitation is called photoselection. 33 z θ x y Figure 1.15 The phenomenon of photoselection where excitation with polarized light results in a cos2 θ probability distribution of excited fluorophores [91]. Since the distribution of excited fluorophores is anisotropic (i.e., not occurring in all directions), the emitted fluorescence is also anisotropic. Depolarization effects, or changes in the direction of the transition moment during the lifetime of the excited state, will result in changes in the anisotropic emission. Therefore, the term anisotropy (r) is used to describe the average angular displacement of the fluorophore that occurs between the absorption and emission transition moments. 31cos2 θ − r = (1.15) 2 The angular difference observed between excitation and emission transition moments is called an intrinsic depolarization effect because it results in a decrease in anisotropy due to inherent properties of the fluorophore. Anisotropy can also decrease due to extrinsic factors which act during the lifetime of the excited state [88, 92]. The primary source of depolarization results from the rotational diffusion of the fluorophores, but can also be due to resonance energy transfer (RET) of energy among fluorophores and other factors. When the rotational movements of a fluorophore occur within the lifetime of the excited state, 34 information on its rotational rate, and subsequently, the fluidity of its microenvironment, can be determined. However, if the movements are slow compared to the excited state lifetime no information is available and depolarization effects are not observed. RET occurs in concentrated solutions where the average distance between fluorophores is small enough to induce the transfer of energy between a donor and acceptor molecule. The resulting radiationless transfer of energy reduces the anisotropy considerably. Minimizing the effects of RET simply requires the dilution of the system under investigation. In contrast to anisotropy ( r ), intrinsic or fundamental anisotropy (ro) is the loss of polarization observed in the absence of other depolarizing processes. Assuming collinear absorption and emission dipoles and recalling that anisotropy varies as a function of cos2 θ, the maximum value of cos2 θ = 0.6. By substituting this into Equation 1.15, one arrives at a maximum in r0, max of 0.4. Therefore, due to photoselection, the anisotropy is reduced by 0.4 (i.e., 2/5). In addition, in the above example the absorption and emission dipoles are assumed to be collinear. Generally, transition moments change upon the absorption of radiation. Therefore, it is rare to find a transition moment in the ground state that coincides entirely with that of the excited molecule. Hence for a typical sample, with non-collinear dipoles, their angular displacement is described by a parameter, β. 2⎛ 31cos2 β− ⎞ r = ⎜ ⎟ (1.16) 0 5⎝ 2 ⎠ The measurement of fluorescence anisotropy (r) is best described schematically as shown in Figure 1.16. A sample is excited with vertically polarized light (V) and the intensity of the emission is measured with a second polarizer oriented parallel (VV) and 35 perpendicular (VH) to the excitation. The intensity of these signals is used to calculate the anisotropy (r) and/or polarization (P). II− r = VV VH (1.17) IIVV + 2 VH II− P = VV VH (1.18) IIVV + VH Anisotropy and polarization are both expressions for the same phenomenon. However, anisotropy values are preferred as most theoretical expressions are considerably simpler when using this parameter, especially in the case involving mixtures of fluorophores where the total emission anisotropy is the weighted sum of the individual anisotropies [90]. In order to compensate for fluctuations in the sensitivity of the detection system for vertically and horizontally polarized light, a parameter called the G factor must also be determined. This value is derived with the excitation polarizer set to the horizontal (H) position and collecting both the vertically (V) and horizontally (H) polarized component of the emission. The G factor is given by I G = HV (1.19) I HH where IHV and IHH are the intensity of the vertically and horizontally polarized component of the emission, respectively. Hence, an alternative formula for to Equation 1.17 for the calculation of anisotropy is I − GI r = VV VH (1.20) IVV + 2GIVH 1.7 Nuclear Magnetic Resonance Unlike other spectroscopic techniques such as fluorescence where the electronic component of electromagnetic radiation was considered, in nuclear magnetic resonance 36 (NMR) it is the magnetic component of electromagnetic radiation that is responsible for a sample absorbing in the radio-frequency (RF) region [12]. As a result, the nuclei of atoms, rather than electrons, are involved in the NMR absorption process. Monochromator Sample Polarizer Ivv Ivv c n Ar mp Xeno La PMT LSU Monochromator Data Acquisition Figure 1.16 Schematic diagram depicting a typical fluorometer used for the measurement of fluorescence anisotropy. An NMR spectrum is the result of nuclei with either odd atomic number ( Z ) or mass possessing the property of spin. An angular momentum ( P ) arising from the nucleus spinning on its axis generates a magnetic dipole. The magnitude of the generated dipole is expressed in terms of the nuclear magnetic moment, µ µ = γ P (1.21) where γ is the magnetogyric ratio, a proportionality constant characteristic of the isotope being examined. The angular momentum of the spinning charge can be described in terms of spin quantum numbers ( I ) which may have values greater than or equal to zero and 37 occur in multiples of ½. It is important to note that nuclei with an even atomic mass and atomic number have zero spin and therefore I = 0. Consider a proton with uniform spherical charge distribution (i.e., I = ½). When such a nuclei is placed in an external static magnetic field (Bo) its magnetic moment (i.e., P and µ) aligns itself in a discrete number of m= ± ½ orientations. A nuclide may assume 2 I +1 possible orientations. Hence for I = ½ there exist two possible orientations, called magnetic quantum states, aligned either parallel (m = – ½) or antiparallel (m = + ½) to the static field as shown in Figure 1.17. Note that in the absence of a magnetic field, the energies of the magnetic quantum states of the nuclei are degenerate or identical. NO FIELD APPLIED FIELD (Bo) m = − ½ µ Y G m = ± ½ R 0 γh ∆E = Bo ENE 2π µ m = + ½ Figure 1.17 Magnetic moments and energy levels for a proton with a spin quantum number of ½. The potential energy (E) of a nucleus in the two orientations can be computed using the following expression γmh E=− B (1.22) 2π o For m = + ½ and m = – ½, E is given by 38 γh E =− B (1.23) + 12/ 4π o γh E = B , (1.24) −12/ 4π o respectively. The difference in energy (∆E) is simply the difference between Equations 1.23 and 1.24, γh ∆ E= B (1.25) 2π o where h is Planck’s constant. Therefore, it is possible to introduce energy corresponding to ∆E to induce a transition between these energy levels. If one substitutes Planck’s relationship ∆E = hνo into the Equation 1.25, the frequency of radiation required to evince the transition, and hence for nuclear magnetic resonance to occur, can be determined from Equation 1.26. γB v = o (1.26) o 2π This frequency, known as the Larmor frequency, generally falls within the radio-frequency region of the electromagnetic spectrum. For protons, frequencies up to approximately 500 MHz are typical. 1.7.1 Pulsed-Field Gradient NMR An advanced application of NMR spectroscopy is the pulsed-field gradient spin echo technique referred to as PFG-NMR. This technique, first described by Hahn in 1950 [93], has been used to study molecular dynamics such as diffusion constants, determination of aggregation states of proteins and peptides [94-97], measurement of bulk movement of hemoglobin [98], and the determination of the monomeric nature of RNA samples [99]. In 39 this dissertation, PFG-NMR was used to determine translational diffusion coefficients as well as free energies of binding for chiral selector-selectand systems. The following paragraphs describe in the detail the anatomy of a typical PFG-NMR experiment beginning with a brief introduction to diffusion. In a liquid, the movement of particles, or the translational diffusion of particles, due to Brownian motion gives rise to the translational diffusion coefficient (Dt) as described by the Stokes-Einstein equation kT Dt = (1.27) 6πηrs where k is the Boltzmann constant, T is temperature in Kelvin, η is viscosity of the solvent, and rs is the hydrodynamic radius of the molecule. This equation assumes the hydrodynamic shape of the molecule in solution is a sphere, prolate ellipsoid, or a symmetric cylinder [99]. PFG-NMR spectroscopy offers an alternative approach to calculating Dt of molecules in solution. To understand the sequence of events needed to arrive at Dt, consider the pulse sequence depicted in Figure 1.18 where the spins are shown as thin disks. The spins are initially randomly oriented (Figure 1.18a) until a 90° pulse is applied in order to rotate the magnetization and align the individual magnetic moments. The sum of the aligned magnetic moments results in a sharp resonance peak (Figure 1.18b). Next, a linear field gradient (along the +z axis) of strength Gz is imposed for a discrete period of time τg. As a result, all the chemically equivalent spins experience a different applied field and therefore precess with different frequencies. The degree with which they rotate is dictated by their position in the sample. The sum of the spins results in a zero net magnetization. Consequently, the gradient pulse is said to have been defocused, yielding no significantly 40 detectable resonance signal (Figure 1.18c). If molecules do not diffuse during the time of the pulse sequence, application of a second gradient pulse of equal intensity and duration but opposite in direction (along the –z axis) refocuses the individual vectors, thereby recovering the original resonance signal (Figure 1.18d). If, however, the molecules diffuse to a different position, the second PFG pulse fails to exactly reverse the effect of defocusing, resulting in an attenuated NMR signal (Figure 1.18e). The change in intensity is given by 2 2 2 −Dt γ GZτ g (∆−δ / 3) A = Aoe (1.28) (a) (b) (c) (d) Refocused 90º Gz, τg −Gz, τg (e) Focused Defocused Attenuated Figure 1.18 The experimental pulse sequence for a PFG-NMR experiment showing a) a sample with randomly oriented magnetic moments, b) the sample after 90° pulse, c) magnetic moments after a linear field gradient of strength Gz is applied, d) after the application of a second gradient pulse of equal but opposite intensity and if the moments do not diffuse, and e) after application of a second gradient pulse of equal but opposite intensity and the moments do diffuse (modified figure from ref. [100]). 41 where Ao and A are the maximum and measured peak intensity, respectively, Dt is the 2 translational diffusion coefficient (cm /s), γ is the magnetogyric ratio of the nuclei, τg is the duration of the gradient, ∆ is the time between gradients and Gz is the strength of the 222 gradient (G/cm). The slope of a plot of − ln( AA/ o ) versus γτGZg(∆−δ/3) yields the translational diffusion (Dt). 1.7.2 Nuclear Overhauser Effect Spectroscopy NMR For complex samples, a typical one dimensional (1-D) NMR spectrum can be ambiguous due to signal overlap of protons in similar environments. For instance, consider a 1H NMR spectrum of a mixture of two compounds, compound A and compound B, having identical coupling constants (i.e., J-couplings) as shown in Figure 1.19a. Upon inspection, the 1-D 1H NMR spectrum gives no indication as to which NMR peaks belong to the same spin system and which originate from a different spin system. The given spectrum can have a number of interpretations. One interpretation is that the first two pairs of peaks correlate with compound A and the remaining peaks with compound B. Another interpretation is that second and fourth pairs of bands are associated with compound A while the first and third pairs are associated with compound B. Hence, examination of a 1-D spectrum alone does not always provide sufficient information to make definitive peak assignments. However, so-called ‘correlation maps’ can, in fact, indicate pictorially peaks that are related, or coupled, to one another. An example of such a correlation map is shown in Figure 1.19b. The schematic demonstrates that a two dimensional (2-D) NMR technique allows one to correctly assign peaks. In this example, the first two sets of peaks (Figure 1.19a) shown in the 1-D 1H spectrum belong to one component while the second sets of peaks along the vertical axis belong to a second component in the mixture. 42 (a) 1-D NMR spectrum (b) 2-D NMR spectrum Figure 1.19 Representative 1-D 1H NMR spectrum showing the ambiguous nature of single dimension NMR a 2-D correlation map. The 2-D technique that provides experimental correlation maps indicating the correct assignment of peaks is referred to as correlation spectroscopy (COSY). In addition to NOESY, there are a number of related 2-D techniques (e.g., TOCSY and ROESY) that aid in providing information on the three-dimensional molecular geometry of complex mixtures. The 2-D technique used in this dissertation is called nuclear Overhauser effect spectroscopy, or NOESY. A 2-D NOESY spectrum contains all the information of a COSY spectrum with the addition of off-diagonal peaks resulting from protons close in space. Nuclear Overhauser effect spectroscopy is a powerful technique for structural studies. In essence, the nuclear Overhauser effect (NOE) is the cross-relaxation of protons by the transfer of resonance energy from one excited proton to an unexcited proton. However, only those proton spins that are close in space (typically ≤ 5 Å) are capable of demonstrating significant NOE effects [101]. For this reason, a major application of NOE is in high resolution NMR where the magnitude of the effect provides information about 43 intramolecular distances of molecules. Knowing approximate distances between certain protons in a molecule can help establish the conformation of an entire molecule. 7.0 H5(H5’) and H6(H6’) 7.2 H7(H7’) ) 7.4 H3(H3’) (ppm resdiual N-H 7.6 7.8 H8(H8’) H4(H4’) 8.0 4.0 3.0 2.0 1.0 (ppm) Figure 1.20 A portion of a NOESY spectrum of a BOH–p-(L-SULV) mixture from reference [102]. BOH was prepared in a 50.0 mM sodium borate buffer at pD 10.2. Shown on the x-axis are the polymeric surfactant resonances and on the y-axis are the BOH resonances. In a NOE experiment a specific group is selectively irradiated with a pre-saturation pulse. After the selective saturation of a given group in the sample, the number of protons in the excited state and unexcited state becomes equal. After the pre-saturation pulse is switched off, a non-selective excitation pulse is applied. Consequently, the peak corresponding to the targeted group disappears from the NMR spectrum, and the intensities of other peaks either increase or decrease. The peaks with variable intensity correspond to protons that are close in space to the irradiated group. The changes in intensity are due to the ability of the irradiated group to cross-relax with neighboring protons, which cause 44 localized partial saturation [101]. To obtain a single NOESY spectrum, hundreds of spectra are measured then assembled into a two dimensional spectrum by Fourier transformation (Figure 1.20). 1.8 Scope of Dissertation In this dissertation, the enantiomeric recognition ability of amino acids based polymeric surfactants is explored using spectroscopic and chromatographic methods. The primary aim of this research was to better understand the underlying mechanism(s) of chiral separations involving analyte molecules and polymeric surfactants. In Chapter 2, a study on the use of a combination of cyclodextrins and polymeric surfactants to separate model compounds is reported. The retention behavior of three binaphthyl derivatives under optimal electrophoretic conditions using a single chiral additive (polymeric surfactant or CD) was discussed. In addition, the effect of CD cavity size and stereochemical configuration of polymeric surfactants on selectivity (α) and resolution (Rs) were investigated. A combination of 3 mM γ-CD and 30 mM D-PS was resulted in enhanced enantioselectivity of (±)1,1’-binaphthyl-2,2’-diyl hydrogen phosphate (BNP) compared to using either chiral selector alone. The overall trends observed for the three analytes were discussed in terms of selectivity and resolution values for each possible combination of CD and PS. The use of p-(L-SULV) as a versatile chiral selector for MECK is introduced in Chapter 3. The negatively charged molecular micelle is a high molecular weight macromolecule with large countercurrent mobility, zero critical micelle concentration, low aggregation number, and high solubility in aqueous solvents. In an extensive chiral screening program p-(L-SULV) was found to be a broadly applicable chiral selector for 45 micellar electrokinetic chromatography. Although the anionic chiral analytes proved more difficult to resolve (i.e., 29 % success rate), the success rate (%) for the chiral resolution of cationic and neutral racemates was extraordinarily high at 77% and 85%, respectively. Of the 75 compounds examined, an astounding 58 of them were resolved after selecting the appropriate concentration of p-(L-SULV). This study established this particular dipeptide polymeric surfactant as a broadly applicable chiral selector for MEKC. Chapter 4 is an outline of an investigation that probes the chiral interactions of p-(L- SULV) with select chiral analytes by the use of capillary electrophoresis, NMR, and fluorescence anisotropy. MEKC studies revealed that BNP, BOH, and BNA enantiomers interact primarily with the leucine chiral center of the dipeptide polymeric surfactant, whereas the dansyl amino acids, Dns-Phe and Dns-Nor interact primarily with the valine chiral center. NOESY NMR results agreed with MEKC results and further confirmed that TB, Dns-Trp, and Dns-Leu interact primarily with valine. The correlation between the chromatographic parameter (α) and the spectroscopic parameter (β) was validated using the selected fluorophores. The α vs. β plots obtained revealed that two chiral separation mechanisms were involved based on data from eight fluorophores. According to the plots, BNP and BOH follow one chiral separation mechanism, while the Dns-AAs follow another mechanism. This result was in agreement with the former two techniques which suggests that the α vs. β correlation serves as a viable alternative for studying chiral interactions. Chapter 5 is a summary of the work contained in this dissertation and highlights future areas to be explored in the realm of chiral separations and fluorescence spectroscopy. Time-resolved fluorescence anisotropy will be explored in order to confirm the assumption 46 that the average time an enantiomer interacts with the polymer is related to the strength of binding. 1.9 References (1) Pasteur, L. Ann. Chim. Phys. 1848, 24, 442-446. (2) Pasteur, L. C. R. Acad. Sci. 1848, 26, 535-540. (3) Jacques, J., Ed. The Molecule and Its Double; McGraw-Hill, Inc.: New York, 1993. (4) Ahuja, S. Chiral Separations- Applications and Technology: Washington D.C., 1997. (5) Jacques, J.; Collet, A.; Wilen, S. H. Enantiomers, Racemates, and Resolutions; Krieger Publishing Co.: Marabar, 1981. (6) Camilleri, P. Capillary Electrophoresis - Theory and Practice; CRC Press: Boca Raton, 1993. (7) Crossley, R., Ed. Chirality and the Biological Activity of Drugs; CRC Press: New York, 1995. (8) U. S. Food and Drug Administration Chirality 1992, 4, 338-340. (9) Rouhi, A. M. Chem. Eng. News 2002, 80, 43-53. (10) Easson, E. H.; Stedman, E. J. Biochem. 1933, 27, 1257-1260. (11) Dalgliesh, C. E. J. Chem. Soc. 1952, 47, 3940-3942. (12) Skoog, D. A.; Holler, F. J.; Nieman, T. A. Principles of Instrumental Analysis, 5th ed.; Harcourt Brace & Company: Orlando, 1998. (13) Heiger, D. N. High Performance Capillary Electrophoresis- An Introduction; Hewlett Packard Company, 2000. (14) Jorgenson, J. W.; Lukacs, K. D. Anal. Chem. 1981, 53, 1298-1302. (15) Jorgenson, J. W.; Lukacs, K. D. J. High Resolut. Chromatogr. Chromatogr. Commun. 1981, 4, 230-231. 47 (16) Reist, P. C. Aerosol Science and Technology, 2nd ed.; McGraw-Hill, Inc.: New York, 1993. (17) Terabe, S.; Otsuka, K.; Ichikawa, K.; Tsuchiya, A.; Ando, T. Anal. Chem. 1984, 56, 111-113. (18) Haynes III, J. L.; Warner, I. M.; Shahab, S. Rev. Anal. Chem. 1999, 18, 317-382. (19) Pramuaro, E.; Pelizzetti, E. In Wilsons and Wilsons Comprehensive Analytical Chemistry; Weber, S. G., Ed.; Elsevier: Amsterdam, 1996; Vol. XXXI. (20) Davankov, V. A.; Rogozhin, S. V. J. Chromatogr. 1971, 60, 280-283. (21) Debowski, J.; Sybilska, D.; Kuraczak, J. J. Chromatogr. 1982, 237, 303-306. (22) Armstrong, D. W. J. Liq. Chromatogr. 1980, 3, 895-900. (23) Armstrong, D. W.; Terrill, R. Q. Anal. Chem. 1979, 51, 2160-2163. (24) LePage, J. N.; Lindner, W.; Davies, G.; Seitz, D. E.; Karger, B. L. Anal. Chem. 1979, 51, 433-435. (25) Dotsevi, G.; Sogah, Y.; Cram, D. J. J. Am. Chem. Soc. 1976, 98, 3038-3041. (26) Larrabee, C. E., Jr.; Sprague, E. D. J. Polym. Sci., Polym. Lett. Ed. 1979, 17, 749- 751. (27) Paleos, C. M.; Malliaris, A. Macromol. Chem. Phys. 1988, 28, 403-409. (28) Hara, S.; Dobashi, A. In Chem. Abstr.; Pat. 04149205: Japan, 1993; Vol. 118. (29) Wang, J.; Warner, I. M. Anal. Chem. 1994, 66, 3773-3776. (30) Palmer, C. P.; Tanaka, N. J. Chromatogr., A 1997, 792, 105-124. (31) Billiot, F. H.; Billiot, E. J.; Warner, I. M. J. Chromatogr., A 2001, 922, 329-338. (32) Yarabe, H. H. Ph. D. Dissertation, Louisiana State University, Baton Rouge, 2000. (33) Shamsi, S. A.; Macossay, J.; Warner, I. M. Anal. Chem. 1997, 69, 2980-2987. 48 (34) Billiot, E.; Macossay, J.; Thibodeaux, S.; Shamsi, S. A.; Warner, I. M. Anal. Chem. 1998, 70, 1375-1381. (35) Billiot, E.; Thibodeaux, S. J.; Shamsi, S. A.; Warner, I. M. Anal. Chem. 1999, 71, 4044-4049. (36) Thibodeaux, S. J.; Billiot, E.; Torres, E.; Valle, B. C.; Warner, I. M. Electrophoresis 2003, 24, 1077-1082. (37) Thibodeaux, S. J.; Billiot, E.; Warner, I. M. J. Chromatogr., A 2002, 966, 179-186. (38) Tarus, J.; Agbaria, R. A.; Morris, K.; Mwongela, S.; Numan, A.; Simuli, L.; Fletcher, K. A.; Warner, I. M. Langmuir 2004, 20, 6887-6895. (39) Shamsi, S. A.; Valle, B. C.; Billiot, F. H.; Warner, I. M. Anal. Chem. 2003, 75, 379- 387. (40) Kapnissi, C. P.; Akbay, C.; Schlenoff, J. B.; Warner, I. M. Anal. Chem. 2002, 74, 2328-2335. (41) Kapnissi, C. P.; Valle, B. C.; Warner, I. M. Anal. Chem. 2003, 75, 6097-6104. (42) Kamande, M. W.; Fletcher, K. A.; Lowry, M.; Warner, I. M. Journal of Separation Science 2005, 28, 710-718. (43) Kamande, M. W.; Kapnissi, C. P.; Zhu, X.; Akbay, C.; Warner, I. M. Electrophoresis 2003, 24, 945-951. (44) Kamande, M. W.; Zhu, X.; Kapnissi-Christodoulou, C. P.; Warner, I. M. Anal. Chem. 2004, 76, 6681-6692. (45) Kapnissi-Christodoulou, C. P.; Lowry, M.; Agbaria, R. A.; Lei, G.; Warner, I. M. Electrophoresis 2005, 26, 783-789. (46) Kapnissi-Christodoulou, C. P.; Warner, I. M. Electrophoresis 2003, 24, 3917-3934. (47) Chankvetadze, B. Capillary Electrophoresis in Chiral Analysis; John Wiley & Sons: New York, 1997. (48) Wedig, M.; Holzgrabe, U. Electrophoresis 1999, 20, 2690-2704. 49 (49) Tahara, S.-I.; Okayama, A.; Kitada, Y.; Watanabe, T.; Nakazawa, H.; Kakehi, K.; Hisamatu, Y. J. Chromatogr., A 1999, 848, 465-471. (50) Amini, A. Electrophoresis 2001, 22, 3107-3130. (51) Fischer, C.; Schmidt, U.; Oeheme, G. J. Chromatogr., A 1999, 845, 273-283. (52) Kodama, S.; Yamamoto, A.; Matsunaga, A. Analyst 1999, 124, 55-59. (53) Snopek, J.; Jelinek, I.; Smolkova-Keulemansova, E. J. Chromatogr. 1988, 438, 211- 218. (54) Koppenhoefer, B.; Zhu, X.; Jakob, A.; Wuerthner, S.; Lin, B. J. Chromatogr., A 2000, 875, 135-161. (55) Tamisier-Karolak, S. L.; Stenger, M.-A.; Bommart, A. Electrophoresis 1999, 20, 2656-2663. (56) Mateus, L.; Cherkaoui, S.; Christen, P.; Veuthey, J.-L. J. Chromatogr., A 1999, 868, 285-294. (57) Zhou, L.; Johnson, B. D.; Miller, C.; Wyvratt, J. M. J. Chromatogr., A 2000, 875, 389-401. (58) Kurth, B.; Blaschke, G. Electrophoresis 1999, 20, 555-563. (59) Ferrara, G.; Santagati, N. A.; Aturki, Z.; Fanali, S. Electrophoresis 1999, 20, 2432- 2437. (60) Martin-Biosca, Y.; Garcia-Ruiz, C.; Marina, M. L. Electrophoresis 2000, 21, 3240- 3248. (61) Verleysen, K.; Van den Bosch, T.; Sandra, P. Electrophoresis 1999, 20, 2650-2655. (62) Rudaz, S.; Veuthey, J.-L.; Desiderio, C.; Fanali, S. Chromatographia 1999, 50, 369- 372. (63) Helfferich, F. Nature 1961, 189, 1001-1002. (64) Gassmann, E.; Kuo, J. E.; Zare, R. N. Science 1985, 230, 813-814. (65) Gozel, P.; Gassmann, E.; Michelsen, H.; Zare, R. N. Anal. Chem. 1987, 59, 44-49. 50 (66) Schmid, M. G.; Grobuschek, N.; Lecnik, O.; Gubitz, G. J. Biochem. Biophys. Methods 2001, 48, 143-154. (67) Chen, Z.; Hobo, T. Recent Res. Devel. Chem. Pharm. Sciences 2002, 2, 15-30. (68) Davankov, V. A. J. Chromatogr., A 2003, 1000, 891-915. (69) Chen, Z. Chromatography 2004, 25, 9-23. (70) Pedersen, C. J. J. Am. Chem. Soc. 1967, 89, 7017-7036. (71) Wan, H.; Blomberg, L. G. J. Chromatogr., A 2000, 875, 43-88. (72) Stover, F. J. Chromatogr. 1984, 298, 203-210. (73) Kuhn, R.; Stoecklin, F.; Erni, F. Chromatographia 1992, 33, 32-36. (74) Kuhn, R.; Erni, F.; Bereuter, T.; Hausler, J. Anal. Chem. 1992, 64, 2815-2820. (75) Kuhn, R.; Sreinmetz, C.; Bereuter, T.; Haas, P.; Erni, F. J. Chromatogr., A 1994, 666, 367-373. (76) Dutton, P. J.; Fyles, T. M.; McDermid, S. J. Can. J. Chem. 1988, 66, 1097-1108. (77) Behr, J.-P.; Lehn, J.-M.; Vierling, P. Helv. Chim. Acta 1982, 65, 1853-1867. (78) Hoehne, E.; Krauss, G. J.; Guebitz, G. J. High Res. Chromatogr. 1992, 15, 698-700. (79) Walbroehl, Y.; Wagner, J. J. Chromatogr., A 1994, 685, 321-329. (80) Kuhn, R.; Wagner, J.; Walbroehl, Y.; Bereuter, T. Electrophoresis 1994, 15, 828- 834. (81) Schmid, M. G.; Guebitz, G. J. Chromatogr., A 1995, 709, 81-88. (82) Kuhn, R.; Riester, D.; Fleckenstein, B.; Weismuller, K. H. J. Chromatogr., A 1995, 716, 371-379. (83) Armstrong, D. W.; Rundlett, K. L.; Chen, J. R. Chirality 1994, 6, 496-509. (84) Vespalec, R.; Bocek, P. Chem. Rev. 2000, 100, 3715-3753. 51 (85) Strege, M. A.; Huff, B. E.; Risley, D. S. LC-GC 1996, 14, 144-150. (86) Rundlett, K. L.; Armstrong, D. W. Anal. Chem. 1995, 67, 2088-2095. (87) Ward, T. J.; Dann III, C.; Blaylock, A. J. Chromatogr., A 1995, 715, 337-344. (88) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd ed.; Plenum: New York, 1999. (89) Schulman, S. G., Ed. Molecular Luminescence Spectroscopy-Methods and Applications: Part 3; Wiley: New York, 1993. (90) Valeur, B. Molecular Fluorescence-Principles and Applications; Wiley-VCH: Weinheim, Germany, 2002. (91) Lakowicz, J. R. Principles of Fluorescence Spectroscopy; Plenum Press: New York, 1983. (92) Weber, G. In Fluorescence and Phosphorescence Analysis; Hercules, D. M., Ed.; Wiley: New York, 1966, pp 217-240. (93) Hahn, E. L. Phys. Rev. 1950, 80, 580-594. (94) Altieri, A. S.; Hinton, D. P.; Byrd, R. A. J. Am. Chem. Soc. 1995, 117, 7566-7567. (95) Dingley, A. J.; Mackay, J. P.; Chapman, B. E.; Morris, M. B.; Kechel, P. W.; Hambly, B. D.; King, G. F. J. Biomol. NMR 1995, 6, 321-328. (96) Mansfield, S. L.; Gotch, A. J.; Harms, G. S.; Johnson, C. K.; Larive, C. K. J. Phys. Chem. B 1999, 103, 2262-2269. (97) Mansfield, S. L.; Jayawickrama, D. A.; Timmons, S. J.; Larive, C. K. Biochim. Biophys. Acta 1998, 1382, 257-265. (98) Kuchel, P. W.; Chapman, B. E. J. Magn. Reson. 1991, 94, 574-580. (99) Lapham, J.; Rife, J. P.; Moore, P. B.; Crothers, D. M. J. Biomol. NMR 1997, 10, 255- 262. (100) Claridge, T. D. W. High-Resolution NMR Techniques in Organic Chemistry; Pergamon: Amsterdam, 1999. 52 (101) Morrissey, D.; www.bch.bris.ac.uk/staff/pfdg/teaching/nmr.htm, 2001. (102) Morris, K. F.; Becker, B. A.; Valle, B. C.; Warner, I. M.; Larive, C. K. In Preparation. 2005. 53 CHAPTER 2. COMBINATION OF CYCLODEXTRINS AND POLYMERIC SURFACTANTS FOR CHIRAL SEPARATIONS 2.1 Introduction Cyclodextrins (CDs) can be described as conical shaped molecules with hydrophobic cavities which allow them to serve as hosts to many hydrophobic molecules. This property, as well as their low cost, high solubility, ease of derivatization, and UV transparency are characteristics that make CDs among the most versatile chiral selectors [1, 2]. While neutral CDs are the most widely used chiral selectors, their separation abilities are limited to only charged species when using CE as the separation technique. Therefore, dual chiral selector systems, where neutral CDs were used in combination with other neutral CDs [3], charged CDs [4-7] and charged polymeric surfactants [8-13], were introduced to overcome this limitation. Various charged micelles have been used in CD-MEKC (e.g., conventional micelles of sodium dodecyl sulfate (SDS) or bile salts) for the separation of various chiral and achiral compounds. The polymeric surfactant (PS), poly sodium undecenyl sulfate was recently studied in combination with neutral or derivatized cyclodextrins for the separation of geometrical isomers of benz[α]anthracenes [14, 15] and polychlorinated biphenyl congeners [16]. In addition to MEKC and CD-MEKC applications, PS have proven to be very robust when used as polyelectrolytes in polyelectrolyte multilayer (PEM) coatings for achiral and chiral separations in open tubular capillary electrochromatography [17-19]. A useful phenomenon of chiral separations is the observation of reversal of enantiomer migration order (RMO). This phenomenon is especially important to the pharmaceutical industry where enantiomerically pure drugs are favored over racemic mixtures. Because of the common occurrence of peak tailing in CE, it is often preferred that 54 the undesired component (the minor component) elutes before the desired, or major, component. If the reverse is true, and the minor component elutes second, then it may be overlooked due to possible overlap with the tailing peak of the first eluting enantiomer. Therefore, the intentional RMO is important in lowering the limits of detection and quantitation [20]. In chiral CE, a RMO is not uncommon; in fact, it can be frequently achieved via various methods, many of which have been well summarized in the literature [21-23]. Some more common methods include: i) reversal of EOF via the addition of various additives [20, 21]; ii) use of pH dependent mobilities of chiral selectors [4, 24]; iii) use of CDs with differing cavity size, charge, concentration, and substituent pattern [20, 25]; iv) use of chiral micelles of opposite configuration [26]; and v) use of combinations of different chiral selectors [16, 27]. However, the choice of chiral selector is important in performing such studies. Since synthetic PSs are available in two forms (the D and L configuration), they offer the most common way by which one can achieve a RMO. In contrast, CDs and their derivatives are naturally occurring oligosaccharides composed of α-D-glucopyranose units and are therefore only available in one stereochemical configuration. While the goal of this work is not to discuss additional methods by which RMO of enantiomers can occur, this study will focus, to some extent on the combined use of two versatile chiral selectors, (neutral CDs and PSs), and their effect on the RMO of the enantiomers of three binaphthyl derivatives. Previously, our group has shown that the combination of γ-CD and poly (sodium N-undecanoyl-D-valinate) [p-(D-SUV)] enhanced the separation of D/L-laudanosine, (±)1,1'-bi-2-naphthol (BOH), (±)-verapamil, and (±)1,1′- binaphthyl-2,2'-diyl hydrogen phosphate (BNP) [28] resulting in improved selectivity and 55 resolution. The present work will expand on this previous study to include β- and γ-CD in combination with both D and L configuration amino acid-based polymeric surfactants. The PSs studied include poly (sodium N-undecanoyl D-alaninate) [p-(D-SUA)], poly (sodium N- undecanoyl D-leucinate) [p-(D-SUL)], poly (sodium N-undecanoyl D-valinate) [p-(D-SUV)], and their respective antipodes poly (sodium N-undecanoyl L-alaninate) [p-(L-SUA)], poly (sodium N-undecanoyl L-leucinate) [p-(L-SUL)], and poly (sodium N-undecanoyl L- valinate) [p-(L-SUV)]. The generalized structures of the surfactant monomer and corresponding PSs are illustrated in Figure 2.1. The binaphthyl derivatives (±)1,1'-binaphthyl-2,2'-diamine (BNA), BOH, and BNP, are atropisomers which possess a chiral plane rather than a chiral carbon (Figure 2.2). Under basic pH conditions, the three analytes differ in charge states as follows: BNA is neutral, BOH is neutral to partially anionic, and BNP is anionic. Our group previously optimized the separation conditions for this set of analytes and they determined the optimum BGE to be a 100 mM TRIS / 10 mM borate buffer at pH 10 [29]. The retention behavior, migration order, resolution, and selectivity values of binaphthyl derivatives using PSs and CDs as chiral selectors are discussed. 2.2 Experimental 2.2.1 Materials and Reagents Gamma cyclodextrin (γ-CD) was obtained from Cerestar (Hammond, IN, USA) and β-CD was a gift from American Maize Products (Hammond, IN, USA). Both the background electrolytes (BGEs) (e.g., Tris(hydroxymethyl)aminomethane (TRIS) and sodium borate (Na2B4O7)) and the chiral analytes were purchased from the Sigma-Aldrich Corp. (St. Louis, MO). Chemicals used for the synthesis of surfactants, N,N '- 56 Na O O Na O O R R NH O HN 60Co γ -radiation O R = amino acid Polymeric Surfactant Alanine = CH3 Valine = CH(CH3)2 Leucine = CH CH(CH ) 2 3 2 Figure 2.1 Basic structure of the amino acid-based polymeric surfactants investigated. ± BNA ± BOH ± BNP O NH2 O OH P NH2 OH O OH (±)-1,1’-binaphthyl- (±) 1,1’-bi-2-naphthol (±)-1,1’-binaphthyl- 2,2’-diamine 2,2’-diyl hydrogen phosphate Figure 2.2 Structure of binaphthyl derivatives. 57 dicyclohexylcarobodiimide (DCC), N-hydroxysuccinimide, undecylenic acid, and the single amino acids (i.e., L-valine, D-valine, L-leucine, D-leucine, L-alanine, and D-alanine), were also obtained from the Sigma-Aldrich Corp. All chemicals were of the highest purity or of analytical reagent grade and were used as received. 2.2.2 Synthesis of Polymeric Chiral Surfactants All surfactant monomers were synthesized according to a procedure reported by Lapidot et al. [30] with modifications made in our laboratory [31, 32]. Figure 2.3 describes the general synthetic procedure followed for the synthesis of all surfactants described in this dissertation. Scheme I describes the synthesis of the ester of undecylenic acid, which is the common intermediate for all the surfactants synthesized. Scheme II shows the reaction of the aforementioned ester with an amino acid of choice (e.g., monopeptide or a dipeptide) to give the N-hydroxysuccinimide amino acid surfactants. Undecylenic acid (I), is reacted with N-hydroxysuccinimide (II) in the presence of the catalyst, dicyclohexylcarbodiimide (III), in dry ethyl acetate. The reaction was allowed to proceed for at least 12 hours under anhydrous conditions. The above condensation reaction yields the N-hydroxysuccinimide ester of undecylenic acid (IV) as well as a precipitate byproduct, dicyclohexylurea (V), which is removed by filtration. The solvent is next evaporated under reduced pressure giving a yellow waxy product that is recrystallized from isopropanol and water, respectively and lipophylized to complete dryness. The amino acid or dipeptide of choice (VI) is dissolved in a sodium bicarbonate solution and allowed to react with a solution of N-hydroxysuccinimide ester of undecylenic acid (IV) in THF. The reaction proceeds for at least 12 hours at room temperature. The corresponding N-acylamino acid (VII) is formed and the solvent is evaporated under 58 vaccuum. The resulting solution is acidified by the addition of 1M HCl to give a pH of about 2. At this pH, the product crashes out of solution giving a white sticky solid which is washed twice with excess water to remove any acid residue. This acidification procedure is repeated twice. Finally, the acid form of the acid surfactant is reacted with excess sodium bicarbonate, filtered, and lyophilized to complete dryness. Polymerization of the resulting sodium salt of N-undecylenyl amino acid surfactant was achieved by exposing a 0.1 M aqueous surfactant solution to an in-house 60Co γ-source (0.7 krad/hr) for seven days to give the poly sodium undecanoyl amino acid surfactants Scheme I O O OH N + HO N + C H2C=HC(H2C)8 N O I II III O O O HN NH H2C=HC(H2C)8 + O N O IV V Scheme II O O OH H2C=HC(H 2C)8 C IV + H2N C CH O CH O N R R H VI VII R = amino acid Figure 2.3 General scheme for the synthesis of amino acid-based surfactants. Surfactant polymerization was monitored using 1H-NMR. The disappearance of the vinyl proton NMR signal at approximately 6.0-5.0 ppm and the broadening of the upfield peaks 59 were used as confirmation that polymerization had taken place. Following polymerization, the polymers were filtered and lyophilized to complete dryness. 2.2.3 Capillary Electrophoresis Procedure A Beckman P/ACE 5510 Series automated CE system (Fullerton, CA, USA) was used for all CD-MEKC experiments. An IBM (486) computer utilizing P/ACE software was used for instrument control and data collection. All separations were conducted using uncoated fused silica capillaries, 57 cm (50 cm effective length) × 50 µm I.D. × 375 µm O.D., purchased from Polymicro Technologies (Phoenix, AZ, USA). For all experiments, the BGE was a 100 mM Tris/10 mM borate buffer adjusted to pH 10 using 1 M sodium hydroxide (NaOH). BGE solutions were next filtered using a nylon syringe filter with 0.45 µm pore size (Nalgene, Rochester, NY). MEKC solutions were prepared by using the equivalent monomer concentration (EMC), where the EMC is the molecular weight of an individual surfactant unit (i.e., the EMC of SUA is 277 g/mol, SUV is 305 g/mol, and SUL is 319 g/mol). Final MEKC or CD-MEKC BGEs were prepared by dissolving the appropriate amounts of CDs and PS, respectively, followed by sonication. All separations were conducted at 25 oC and 30 kV with UV detection at 254 nm at the cathodic end. Samples were prepared in concentrations of 0.1 mg/mL in a (50:50) methanol/water mixture and introduced into the capillary using the pressure option for 3 s. Enantiomeric elution orders were determined by spiking a single pure R-enantiomer into the solution of the corresponding racemic mixture. New bare fused silica capillaries were conditioned successively, under high pressure, with 1 M NaOH for 120 min, 0.1 M NaOH for 30 min and triply distilled deionized water for 15 min prior to use. In addition, the capillary was conditioned with the MEKC or CD-MEKC solutions for 3 min between injections. 60 2.3 Results and Discussion 2.3.1 Elution Order of Binaphthyl Derivatives as a Function of CD Cavity Size and the Stereochemical Configuration of Polymeric Surfactants The elution order of BNA, BOH, and BNP enantiomers in the presence of two native cyclodextrins (β- and γ-CD) and six PS [p-(D-SUA), p-(D-SUV), p-(D-SUL), p-(L-SUA), p- (L-SUV), p-(L-SUL)] was investigated. Table 2.1 summarizes the enantiomeric elution order in all cases. It is well known that the longer the retention time of an analyte, the stronger the interaction with the anionic PS, whereas the reverse is true for native CDs. Since both β- and γ-CDs are neutral and possess no electrophoretic mobility, no separation was observed for neutral compounds such as BNA. Due to the partial anionic charge of BOH, γ-CD afforded the enantioseparation. In contrast, no enantioseparation of BOH was observed using β-CD. The anionic BNP enantiomers were resolved using both β- and γ-CD, where the R enantiomer eluted second with both CDs. For the BNA and BOH enantiomers, the R-enantiomer has a longer retention time in the presence of L-PS (i.e. p-(L-SUA), p-(L-SUV), p-(L-SUL)). When the configuration of the PS is changed to the D-form (i.e. p-(D-SUA), p-(D-SUV), p-(D-SUL)), the migration order switches and the S-enantiomer elutes second. The above results indicate that if the PS exhibits enantioselective behavior toward an analyte, a simple change in the configuration of the PS results in a reversal of enantiomer elution order. No apparent separation was observed for the enantiomers of BNP using either D- or L-PS at pH 10. 61 Table 2.1 Enantiomeric selectivity pattern of binaphthyl derivatives towards neutral CDs and PS Chiral analytea Chiral selector and the second eluting enantiomerb β-CD γ-CD D-SUA L-SUA D-SUV L-SUV D-SUL L-SUL BNA (n) nsc nsc S R S R S R BOH (−/n) nsc S S R S R S R c c c c c c BNP (−) R R ns ns ns ns ns ns aSymbol in parenthesis indicates charge of analyte under given experimental conditions where (n) is neutral and (−) is negative; bConditions: 10 mM β-CD, 20 mM γ-CD, 30 mM PS for BNA, 6 mM PS for BOH and BNP in 100 mM TRIS / 10 mM borate buffer (pH 10); cns = no separation. 2.3.2 Enantioseparation of BNA 2.3.2.1 D-Polymeric Surfactants in the Presence of β-CD or γ-CD Previous studies by our group determined the optimal PS concentration for the separation of BNA enantiomers to be 6 mM [29]. The CD-MEKC BGE solutions were prepared by slowly titrating β-CD into 6 mM p-(D-SUA), p-(D-SUV), or p-(D-SUL). As shown in Figure 2.3a, the enantiomeric elution order was R before S in the presence of 6 mM p-(D-SUA) alone. The enantiomers co-elute, with shorter retention time, upon theaddition of 3 mM β-CD to p-(D-SUA) (Figure 2.3b). As the β-CD concentration is increased to 5 mM (Figure 2.3c), a RMO of enantiomers is observed (i.e., S enantiomer elutes before R). Upon further increase in β-CD concentration to 8 mM (Figure 2.3d), even shorter retention times are observed and the resolution (Rs) and efficiency continue to improve. It should be noted that a similar reversal was observed for the other two D- configuration PS (i.e., p-(D-SUL) and p-(D-SUV)). This type of RMO, that is dependent on the concentration of one chiral selector (e.g., β-CD), can be rationalized as follows. At 0 62 1.15 a 1.10 b c α 0.008 α 1.05 R (a) 0 mM β-CD 0.004 1.00 S 6 mM p-(D-SUA) 0358 R = 1.35 β-CD concentration [mM] s 0.000 0.008 R,S β 0.004 (b) 3 mM -CD 6 mM p-(D-SUA) Rs = 0.00 0.000 0.008 AU AU R 0.004 S (c) 5 mM β-CD 6 mM p-(D-SUA) R = 1.42 s 0.000 0.008 R (d) 8 mM β-CD 0.004 S 6 mM p-(D-SUA) Rs = 1.43 0.000 02468 migration time [min] Figure 2.4 Enantioseparation of the non-racemic mixture (excess R) of BNA enantiomers using a 6 mM p-(D-SUA) with (a) 0 mM β-CD, (b) 3 mM β-CD, (c) 5 mM β-CD, and (d) 8 mM β-CD. MEKC conditions: 100 mM TRIS / 10 mM borate buffer, pH 10, 30 kV, 254 nm. Inset shows the (α) as a function of β-CD concentration using p-(D-SUA), p-(D- SUL), p-(D-SUV), denoted a, b, c, respectively. 63 mM β-CD, the migration order of the BNA enantiomers is determined by their binding capability with D-PS, and the preferentially complexed S-enantiomer migrates more slowly than the more weakly complexed R-enantiomer. The peak coalescence at an intermediate β- CD concentration (i.e., 3 mM β-CD) indicates both chiral selectors (D-PS and β-CD with anionic and neutral charge states, respectively) compete for the same S-enantiomer resulting in Rs and selectivity (α) values of 0 and 1, respectively. As the concentration of the β-CD approaches that of the PS, a reversal of enantiomeric order takes place along with an increase in Rs. At a higher concentration of β-CD (e.g. 8 mM), the complex of the preferentially bound S-enantiomer migrates faster to the detector compared to the transient complex of the more weakly bound R-enantiomer. Observed trends in α agree with Rs trends and are shown in the inset of Figure 2.3. The effect of cavity size on the enantiomeric Rs and α of BNA was investigated by using γ-CD (Figure 2.4), which has a larger cavity than β-CD. As mentioned above, the enantiomer elution order was R before S in the presence of all three D-PS, and remained so as the concentration of γ-CD was increased up to a maximum of 8 mM. The larger cavity of γ-CD, and thus the change in distance between noncovalently interacting selector-selectand groups, seemed to decrease the enantioselectivity of the system as compared to β-CD. 2.3.2.2 L-Polymeric Surfactants in the Presence of β-CD or γ-CD When using the p-(L-SUA) in conjunction with β-CD or γ-CD, no reversal of enantiomeric elution order took place (Figures 2.5 and 2.6). Instead, there was a general increase in both α and Rs with increasing β-CD or γ-CD concentration for all three L-PSs. The data in Table 2.1 confirms that all three L-PS interact more strongly with the R enantiomer. Therefore, it is reasonable to conclude the S-enantiomer of BNA has a greater 64 affinity for the neutral β-CD, resulting in enhanced Rs and α values with increasing β-CD concentration. This demonstrates an enhancement in chiral recognition of BNA enantiomers by use of a dual chiral selector system. (a) 1.06 p-(D-SUA) p-(D-SUV) 1.04 p-(D-SUL) α 1.02 1.00 0 5 10 15 20 (b) 2.5 2.0 1.5 s 1.0 R 0.5 0.0 0 5 10 15 20 γ-CD concentration [mM] Figure 2.5 Bar plots showing the a) selectivity (α) and b) resolution (Rs) trends for the separation of BNA enantiomers while adding incremental amounts of γ-CD to a fixed concentration (6 mM) of D-PS. EKC conditions are identical to Figure 2.4. 2.3.3 Enantioseparation of BOH 2.3.3.1 D-Polymeric Surfactants in the Presence of β-CD or γ-CD Under the BGE conditions investigated (pH 10), the enantiomers of BOH, which are neutral to partially anionic, were successfully resolved using all three D-PSs and the elution order was R before S. However, since β-CD shows little or no enantioselective behavior 65 1.3 b a c 1.2 α 0.008 α (a) 0 mM β-CD 1.1 6 mM p-(L-SUA) R R = 1.40 s 0.004 1.0 S 0358 β-CD concentration [mM] 0.000 0.008 (b) 3 mM β-CD R 6 mM p-(L-SUA) R = 2.57 0.004 S s 0.000 0.008 R (c) 5 mM β-CD AU AU 6 mM p-(L-SUA) S R = 3.24 0.004 s 0.000 0.008 R (d) 8 mM β-CD 6 mM p-(L-SUA) S R = 3.51 0.004 s 0.000 02468 migration time [min] Figure 2.6 Enantioseparation of the non-racemic mixture (excess R) of BNA enantiomers using a 6 mM p-L-SUA with (a) 0 mM β-CD, (b) 3 mM β-CD, (c) 5 mM β-CD, and (d) 8 mM β-CD. Insets show the selectivity (α) as a function of β-CD concentration using p-(L-SUL), p-(L-SUV), p-(L-SUA), denoted a, b, c, respectively. EKC conditions are identical to Figure 2.4. 66 p-(L-SUA) (a) 1.20 p-(L-SUV) p-(L-SUL) 1.15 1.10 α 1.05 1.00 0 5 10 15 20 (b) 3.5 3.0 2.5 2.0 s 1.5 R 1.0 0.5 0.0 0 5 10 15 20 γ-CD concentration [mM] Figure 2.7 Bar plots showing the a) selectivity (α) and b) resolution (Rs) trends for the separation of BNA enantiomers while adding incremental amounts of γ-CD to a fixed concentration (6 mM) of L-PS. EKC conditions are the same as in Figure 2.4. toward BOH, a general decrease in selectivity was observed upon the addition of β-CD (Figure 2.7, panels a and b). In the case of p-(D-SUL) and p-(D-SUV), the addition of 5 mM β-CD significantly hindered the interaction between BOH and PS resulting in a complete loss of selectivity. In contrast to β-CD, the addition of γ-CD to all three D-PS resulted in an increase in chiral recognition of BOH (Figure 2.7, panels c and d). In addition, the use of γ- CD alone exhibits some chiral recognition ability toward the enantiomers of BOH (elution order is R before S). Therefore, it is reasonable to expect that the addition of a second chiral selector (i.e., D-PS), that also exhibits a similar tendency for the same analyte, would result in an enhancement in enantioseparation. 67 2.3.3.2 L-Polymeric Surfactants in the Presence of β-CD or γ-CD A RMO of BOH, unseen for the D-PS/CD system, was observed when the antipode, L-PS, was used. There was an observed synergistic effect on the enantioseparation of BOH by the use of the following combinations: L-PS/β-CD (Figure 2.8, panels a and b) and L- PS/γ-CD (Figure 2.8, panels c and d). (a) 1.16 (b) 3.5 p-(D-SUA) 1.14 p-(D-SUV) 3.0 p-(D-SUL) 1.12 2.5 1.10 2.0 1.08 s α R 1.5 1.06 1.0 1.04 1.02 0.5 1.00 0.0 035810 035810 β-CD concentration [mM] β-CD concentration [mM] (c ) 1.16 (d) 3.5 1.14 3.0 1.12 2.5 1.10 2.0 1.08 s R α 1.5 1.06 1.04 1.0 1.02 0.5 1.00 0.0 0 5 10 15 20 05101520 γ-CD concentration [mM] γ-CD concentration [mM] Figure 2.8 Bar plots showing the selectivity (α) and resolution (Rs) trends for BOH enantiomeric separation when adding incremental amounts of β-CD (panels a and b) or γ-CD (panels c and d) to 6 mM D-PS. EKC conditions are the same as in Figure 2.4. 68 p-(L-SUA) p-(L-SUV) (a) 1.16 (b ) 5.5 p-(L-SUL) 5.0 1.14 4.5 1.12 4.0 1.10 3.5 1.08 3.0 s 2.5 α 1.06 R 2.0 1.04 1.5 1.02 1.0 0.5 1.00 0.0 035810 035810 (c) 1.16 β-CD concentration [mM] (d ) 5.5 β-CD concentration [mM] 1.14 5.0 4.5 1.12 4.0 1.10 3.5 1.08 3.0 s 2.5 α 1.06 R 2.0 1.04 1.5 1.02 1.0 * 0.5 1.00 * 0.0 0 5 10 15 20 0 5 10 15 20 γ-CD concentration [mM] γ-CD concnetration [mM] Figure 2.9 Bar plots showing the selectivity (α) and resolution (Rs) trends for BOH enantiomeric separation when adding incremental amounts of β-CD (panels a and b) and γ-CD (panels c and d) to 6 mM L-PS. EKC conditions are the same as in Figure 2.4. However, after the addition of 5 mM γ-CD, the selectivity of the p-(L-SUA) and p-(L-SUL) systems was reduced and a RMO was observed only in the case of p-(L-SUV). A reasonable explanation for this RMO is that, at higher CD concentrations, both chiral selectors exhibit a stronger interaction with R-BOH, thus, counteracting each other and resulting in an overall loss of selectivity (indicated by the asterisk in Figure 2.8c). These findings agree with previous results by Wang and Warner [31] where the combination of 10 mM γ-CD and 0.5% p-(L-SUV) resulted in a loss of resolution and selectivity of BOH enantiomers. They showed that a combination of 10 mM γ-CD and 0.5% p-(D-SUV) resulted in a far superior selectivity value. While a RMO of BOH was not observed for p-(L-SUA) or p-(L-SUL) 69 under the concentration range studied, it is speculated that a reversal should occur at higher concentrations of γ-CD. 2.3.4 Enantioseparation of BNP 2.3.4.1 D-Polymeric Surfactants in the Presence of β-CD or γ-CD The previously determined optimum MEKC conditions (i.e. 30 mM PS, pH 10) was used for enantioseparation of BNP [29] enantiomers. Due to electrostatic repulsion, the anionic polar head groups of the PSs likely repel the anionic enantiomers of BNP; hence, no separation (α =1) was observed for all PS (D or L configuration) in the absence of CDs. Although, an examination of Table 2.1 reveals that the CD cavity diameter has little influence on the chiral recognition of BNP enantiomers, Figures 2.9 and 2.10 (panels a and b), show a marked increase in selectivity for BNP upon the addition of β- or γ-CD to 30 mM D- or L-PS. It is important to note that in all cases S-BNP eluted before the R-enantiomer. This suggests that the R-enantiomer of BNP has little to no affinity for the PS and a much higher affinity for both β- and γ-CDs. Although slight resolution of BNP enantiomers was achieved with only 20 mM γ-CD (Figure 2.11a), no chiral separation was observed using 30 mM p-(D-SUL) alone (Figure 2.11b). However, the resolution improved when increasing concentrations of PS were added to the fixed concentration of γ-CD (Figure 2.11c). As the p-(D-SUL) concentration was increased from 3 to 10 mM, the retention time increased as well as the Rs value (Figure 2.11d). While the selectivity of the system did not change significantly, an enhancement in the efficiency of the peaks is observed resulting in an improvement in chiral resolution. Therefore, a synergistic effect was observed using a combination of PS and γ-CD compared with that obtained by using either γ-CD or p-(D-SUL) alone. 70 p-(D-SUA) p-(D-SUV) p-(D-SUL) (a) 1.26 (b) 6.5 1.24 6.0 1.22 5.5 1.20 5.0 1.18 4.5 1.16 4.0 1.14 3.5 1.12 s 3.0 α R 1.10 2.5 1.08 2.0 1.06 1.5 1.04 1.0 1.02 0.5 1.00 0.0 035810 035810 β-CD concentration [mM] β-CD concentration [mM] (c) 1.14 (d) 3.5 1.12 3.0 1.10 2.5 1.08 2.0 1.06 s α R 1.5 1.04 1.0 1.02 0.5 1.00 0.0 0 5 10 15 20 0 5 10 15 20 γ-CD concnetration [mM] γ-CD concentration [mM] Figure 2.10 Bar plots showing the selectivity (α) and resolution (Rs) trends for the separation of BNP using 30 mM D-PS and incremental amounts of β-CD (panels a and b) or γ-CD (panels c and d). EKC conditions are the same as in Figure 2.4. p-(L-SUA) p-(L-SUV) p-(L-SUL) (a) 1.24 (b) 7.0 1.22 6.5 1.20 6.0 1.18 5.5 5.0 1.16 4.5 1.14 4.0 1.12 3.5 s α 1.10 R 3.0 1.08 2.5 1.06 2.0 1.04 1.5 1.0 1.02 0.5 1.00 0.0 035810 035810 β-CD concentration [mM] β-CD concentration [mM] (c) 1.10 (d) 3.5 3.0 1.08 2.5 1.06 2.0 s α 1.04 R 1.5 1.0 1.02 0.5 1.00 0.0 0 5 10 15 20 0 5 10 15 20 γ-CD concentration [mM] γ-CD concentration [mM] Figure 2.11 Bar plots showing the selectivity (α) and resolution (Rs) trends for the separation of BNP using 30 mM L-PS and incremental amounts of β-CD (panels a and b) and γ-CD (panels c and d). EKC conditions are the same as in Figure 2.4. 71 300 R = 1.35 (a) 20 mM γ-CD s R α = 1.08 200 S 100 0 (b) 30 mM p-(D-SUL) 500 R s = 0.00 400 α = 1.00 R,S 300 200 100 ) 0 -5 300 (c) 20 mM γ-CD + 10 R x R = 1.49 s 3 mM p-(D-SUL) S ( 200 α = 1.06 AU 100 0 400 R = 1.83 (d) 20 mM γ -CD + R s 300 α = 1.07 10 m M p-(D-SUL) S 200 100 0 2345 migration time [min] Figure 2.12 Enantioseparation of BNP using a) 20 mM γ-CD; b) 30 mM p-(D-SUL); c) 20 mM γ-CD and 3 mM p-(D-SUL); and d) 20 mM γ-CD and 10 mM p-(D- SUL). EKC conditions are the same as in Figure 2.4. 2.4 Conclusions The results obtained in this study clearly demonstrate that chiral recognition in CD- MEKC is influenced by factors such as the charge-state of the analyte, type, concentration, and size of CD used, as well as the optical configuration of the polymeric surfactant. A reversal of enantiomeric order was observed for BNA and BOH due to a competition between the D-PS and the CDs for the same enantiomer. An example where a combination of CD and PS gave an α value much greater than that of either chiral selector alone was shown for BNP. The overall trends observed for the three analytes were as follows: i) for BNA, the combination of L-PS with either β- and γ-CDs provided an improvement in chiral recognition; ii) for BOH, the L-PS worked better when combined with β-CD, while the D-PS 72 worked better when combined with γ-CD; iii) for BNP, a combination of 30 mM D-PS with 3 mM γ-CD gave the best α and Rs values. This combined approach, using binary chiral selectors, suggests the possibility of enhanced enantioseparation for select chiral analytes in capillary electrophoresis. 2.5 References (1) Chankvetadze, B. Capillary Electrophoresis in Chiral Analysis; John Wiley & Sons: New York, 1997. (2) Gubitz, G.; Schmid, M. G. J. Chromatogr., A 1997, 792, 179-225. (3) Zhu, X.; Ding, Y.; Lin, B.; Jakob, A.; Koppenhoefer, B. J. Chromatogr., A 2000, 888, 241-250. (4) Sabbah, S.; Scriba, G. K. E. Electrophoresis 2001, 22, 1385-1393. (5) Fillet, M.; Hubert, P.; Crommen, J. J. Chromatogr., A 2000, 875, 123-134. (6) DeSilva, K.; Kuwana, T. Biomed. Chromatogr. 1997, 11, 230-235. (7) Magnusson, J.; Wan, H.; Blomberg, L. G. Electrophoresis 2002, 23, 3013-3019. (8) Mendez, S. P.; Gonzalez, E. B.; Sanz-Medel, A. Anal. Chim. Acta 2000, 416, 1-7. (9) Jabor, V. A. P.; Bonato, P. S. Electrophoresis 2001, 22, 1399-1405. (10) Terabe, S.; Miyashita, Y.; Shibata, O.; Barnhart, E. R.; Alexander, L. R.; Patterson, D. G.; Karger, B. L.; Hosoya, K.; Tanaka, N. J. Chromatogr., A 1990, 516, 23-31. (11) Schmitt, P.; Garrison, A. W.; Freitag, D.; Kettrup, A. J. Chromatogr., A 1997, 792, 419-429. (12) Salami, M.; Otto, H. H.; Jira, T. Electrophoresis 2001, 22, 3291-3296. (13) Gotti, R.; Fiori, J.; Hudaib, M.; Cavrini, V. Electrophoresis 2002, 23, 3084-3092. (14) Akbay, C.; Warner, I. M.; Shamsi, S. A. Electrophoresis 1999, 20, 145-151. 73 (15) Akbay, C.; Shamsi, S. A.; Warner, I. M. J. Chromatogr., A 2001, 910, 147-155. (16) Edwards, S. H.; Shamsi, S. A. Electrophoresis 2002, 23, 1320-1327. (17) Kapnissi, C. P.; Akbay, C.; Schlenoff, J. B.; Warner, I. M. Anal. Chem. 2002, 74, 2328-2335. (18) Kamande, M. W.; Kapnissi, C. P.; Zhu, X.; Akbay, C.; Warner, I. M. Electrophoresis 2003, 24, 945-951. (19) Kapnissi, C. P.; Valle, B. C.; Warner, I. M. Anal. Chem. 2003, 75, 6097-6104. (20) Schmitt, T.; Engelhardt, H. J. Chromatogr., A 1995, 697, 561-570. (21) Chankvetadze, B.; Schulte, G.; Blaschke, G. J. Pharm. Biomed. Anal. 1997, 15, 1577-1584. (22) Sabbah, S.; Scriba, G. K. E. J. Chromatogr., A 1999, 833, 261-266. (23) Chankvetadze, B. Electrophoresis 2002, 23, 4022-4035. (24) Sabbah, S.; Suss, F.; Scriba, G. K. E. Electrophoresis 2001, 22, 3163-3170. (25) Chankvetadze, B.; Schulte, G.; Blaschke, G. J. Chromatogr., A 1996, 732, 183-187. (26) Mazzeo, J. R.; Grover, E. R.; Swartz, M. E.; Peterson, J. S. J. Chromatogr., A 1994, 680, 125-135. (27) Terabe, S.; Miyashita, Y.; Ishihama, Y.; Shibata, O. J. Chromatogr. 1993, 636, 47- 55. (28) Wang, J.; Warner, I. M. J. Chromatogr., A 1995, 711, 297-304. (29) Billiot, F. H.; Billiot, E. J.; Warner, I. M. J. Chromatogr., A 2001, 922, 329-338. (30) Lapidot, Y.; Rappoport, S.; Wolman, Y. J. Lipid Res. 1967, 8, 142-145. (31) Wang, J.; Warner, I. M. Anal. Chem. 1994, 66, 3773-3776. (32) Macossay, J.; Shamsi, S. A.; Warner, I. M. Tetrahedron Lett. 1999, 40, 577-580. 74 CHAPTER 3. POLYSODIUM N-UNDECANOYL-L-LEUCYLVALINATE: A VERSATILE CHIRAL SELECTOR FOR MICELLAR ELECTROKINETIC CHROMATOGRAPHY 3.1 Introduction Molecular micelles (a.k.a. polymeric surfactants or micelle polymers) have recently received considerable attention as stable pseudostationary phases for achiral [1-6] and chiral separations [7-11] in micellar electrokinetic chromatography (MEKC). The use of molecular micelles in capillary electrophoresis (CE) appears to offer distinct advantages over conventional surfactant micelles. Unlike conventional micelles, polymeric surfactants can be purified and can be used at concentrations below the critical micelle concentration (CMC) because they are covalently linked. This in turn provides much higher efficiencies, lower Joule heating and rapid analysis in CE as compared to equivalent concentrations of monomeric surfactant. Another important advantage of polymeric surfactants is that after the separation of charged and polar compounds, sensitive detection of such compounds can be achieved using more mass spectrometric-friendly conditions [12-15] than those required using cyclodextrins [16, 17] or conventional micelles [18, 19]. Dipeptide micelle polymers belong to a relatively new class of chiral selectors for MEKC that were introduced by Shamsi and coworkers in 1997 [20]. Since then, various dipeptide micelle polymers with different chiral centers [21, 22], chiral combinations [23- 25], and configurations [26] have been synthesized and evaluated. Based on information gleaned by previous studies by our group, it was determined that optimal separations were observed when the innermost amino acid is bulkier than that of the outermost amino acid. This arrangement was found to aid in the diffusion of the analyte into and out of the polymeric surfactant, thereby encouraging more analyte-polymeric surfactant interactions. 75 The dipeptide polymeric surfactant, poly sodium N-undecanoyl-L-leucylvalinate p-(L- SULV), was chosen for this study because it fit the above criterion and it appeared to have great potential as a versatile chiral selector. Table 3.1 lists some important physicochemical properties of the chiral surfactant, L-SULV, and its corresponding polymer, p-(L-SULV). As noted in Table 3.1, the monomer of L-SULV contains two chiral centers of the same L- configuration with N-terminal leucine and C-terminal valine amino acids attached to a polymerizable C11 hydrocarbon chain. Comparison of the physicochemical properties of the monomer and polymer of L-SULV reveals that the latter is a high molecular weight, highly charged species with higher electrophoretic mobility. Although the partial specific volume and electrophoretic mobility of unpolymerized L-SULV and p-(L-SULV) are very similar, p- (L-SULV) has a significantly lower aggregation number than L-SULV, as well as a CMC of zero. Table 3.1 Physicochemical properties of sodium N-undecenoyl-L-leucylvalinate (L- SULV) and poly sodium N-undecanoyl-L-leucylvalinate p-(L-SULV) Characteristic L-SULV p-(L-SULV) molecular weight (g/mol) 418.6 14,650 (+ 625)a Critical micelle concentration (CMC) [mM]b 7 0 Optical rotation -0.39 -0.39 effective electrophoretic mobility (cm2V-1/s-1) -3.74 X 10-4(c) -3.76 X 10-4(c) -4.19 X 10-4(d) -4.65 X 10-4(d) -3.89 X 10-4(e) Aggregation numberb 39 18 d d Partial specific volume[ V ] (mL/g) 0.83 0.80 aObtained from ref 27. bObtained from ref 28. cDetermined using 20 mM borate buffer, 20 mM (L-SULV or p-(L-SULV)); pH 9.3, 30 kV; 254 nm; micelle marker: pyrene. dDetermined using 25 mM borate buffer, 20 mM (L-SULV or p-(L-SULV)); pH 9.3; 30 kV; 254 nm; micelle marker: tert-butylanthracene. eDetermined using 20 mM borate buffer with direct injection of polymer solution at 30 kV. 76 In this work, the chiral recognition properties of p-(L-SULV) were investigated using 75 racemic compounds with various structural features. Chiral MEKC parameters (e.g., pH, type and concentration of background electrolyte, and concentration of micelle polymers) were optimized for each class of compounds. The enantioseparation of various classes of cationic, anionic, and neutral analytes is correlated with their stereoselective interactions with p-(L-SULV). The major interest in this study focuses on defining or discovering reliable rules that affect the chiral recognition of various enantiomers using p-(L-SULV) as a chiral pseudostationary phase. 3.2 Experimental 3.2.1 Materials The analytes were obtained either as racemic mixtures or in pure enantiomeric form. The background electrolytes (BGEs), such as sodium phosphate monobasic (NaH2PO4), sodium phosphate dibasic (Na2HPO4), boric acid (H3BO3), tris(hydroxymethyl)aminomethane (TRIS), sodium borate (Na2B4O7), and cyclohexylamino propanesulfonate (CAPS), were of analytical reagent grade. All chemicals used for the surfactant synthesis including, N,N’-dicyclohexylcarobodiimide (DCC), N- hydroxysuccinimide, undecylenic acid, BGEs, and analytes were all obtained from the Sigma-Aldrich Corp. (St. Louis, MO). The dipeptide L-leucylvalinate was purchased from BACHEM Bioscience Inc. (King of Prussia, PA) and was used as received. 3.2.2 Instrumentation All enantioseparations were performed using an Agilent CE system. The CE instrument was equipped with (i) a high voltage power supply (0-30 kV), (ii) a diode array detector for UV detection, and (iii) ChemStation software for system control and data 77 handling. Fused silica capillaries (50 µm ID, 320 µm OD) obtained from Polymicro Technologies (Phoenix, AZ, USA) were employed in all experiments. The effective length of the capillaries was 56 cm and the total length was 64.5 cm. The temperature used was 25°C unless otherwise stated. 3.2.3 Synthesis of Poly Sodium N-undecanoyl-L-leucylvalinate The dipeptide micelle polymer was synthesized from enantiomerically pure L- leucylvalinate using a procedure previously outlined in chapter 2.2.2. The molecular structure is similar to the single amino acid micelle shown in chapter 2, (Figure 2.1) with the exception of the polar headgroup. Table 3.1 lists some major physicochemical properties of the monomer and corresponding polymer of SULV. 3.2.4 Determination of Partial Specific Volume Partial specific volume (V ) is defined as the increase in volume when one gram of dry solute is dissolved in a large volume of solvent. This parameter is used to estimate the approximate volume of a particle since the exact volume of a particle is difficult to measure. The partial specific volume measurements were obtained as follows. Five solutions containing 0, 5, 10, 15, 20, and 25 mg of pseudostationary phase were dissolved in 50 g of solvent (100 mM NaCl). A high precision densitometer (model DMA58) from Anton Paar USA (League City, TX) was used to determine density measurements. Air and water were used for calibration. The V values were determined by plotting the reciprocal of the density (1/ρ) of the solutions as a function of the weight fraction (W) of the chiral pseudostationary phase, according to the following equation. 1 (1 ρ) =+VW∂ (1) ρ ∂W 78 3.2.5 Preparation of MEKC Buffers and Analyte Solutions BGEs at pH 6.5 and 7.0 were prepared by mixing 25 mM solutions of each sodium phosphate mono and dibasic or 30 mM sodium phosphate dibasic with 275 mM boric acid, respectively. The BGE at pH 8.5 contained 25 mM TRIS and 25 mM sodium borate or 300 mM CAPS and 50 mM sodium borate. The basic BGEs at pH 10.2 and 11.2 were prepared using either 25 mM TRIS / 25 mM sodium phosphate (dibasic) or 50 mM sodium phosphate (dibasic). In all cases, the desired pH values of the BGEs were achieved by using either 1 M sodium hydroxide (NaOH) or 1M phosphoric acid (H3PO4). After the pH was adjusted, an appropriate equivalent monomer concentration (EMC) of p-(L-SULV) was added to the BGE solutions. The EMC for p-(L-SULV) and L-SULV is 418.6 g/mol. Finally, the BGEs containing p-(L-SULV) were filtered through a 0.45 µm nylon syringe filter (Nalgene, Rochester, NY), and sonicated 10 min to ensure properly degassed micellar running buffers. The analytes were dissolved either in a 50:50 methanol/water mixture or an 80:20 methanol/water mixture to give the final desired concentration. 3.2.6 Capillary Electrophoresis Procedure New capillaries were washed with 1 M NaOH for 1hr at 60 °C followed by a 10 min rinse with triply distilled deionized water. Between injections the capillary was flushed for two min each with water and micellar BGE. The separation parameters such as retention factor (k´), selectivity factor (α), and resolution (Rs) were calculated using the equations described in chapter 1.4. 3.3 Results and Discussion Various experimental conditions such as pH, concentration and type of BGE, concentration of p-(L-SULV), sample concentration, and injection size were optimized for 79 the screening of 75 racemic compounds. Although a complete optimization of the aforementioned experimental factors was independently conducted for each racemate, the impact of p-(L-SULV) concentration on chiral resolution of the analytes served as the basis for analyte classification. Therefore, in this study, analyte enantiomers were classified into three groups according to the concentration range of p-(L-SULV) required to achieve enantioseparation 3.3.1 Enantioseparation of Class I Analytes As shown in Table 3.2, a total of 18 racemic compounds (i.e., 24% of all analytes investigated) could be resolved with Rs values reported ranging from 1.12 to 4.27. Such racemic compounds can be classified as class I analytes (i.e., those exhibiting strong chiral interactions) because relatively low concentrations (between 5-20 mM) of p-(L-SULV) were needed for the chiral separation of these analytes. Class I compounds include binaphthyl derivatives, benzodiazepines, paveroline derivatives, several members of the β-blocker family, and a variety of other classes of cationic and neutral compounds. Racemates of neutral (e.g., BNA = 1,1'-binaphthylamine) and partially anionic (e.g., BOH = 1,1'-bi-2-naphthol) binaphthyl derivatives are particularly easy to resolve and showed baseline resolution at concentrations as low as 5 mM p-(L-SULV). In contrast, chiral separation of the anionic binaphthyl derivative, BNP = 1,1'-binaphthyl-2,2'-diyl hydrogen phosphate, requires a slightly higher concentration, i.e., 7 mM p-(L-SULV) (Table 3.2). However, the simultaneous enantioseparation of all three binaphthyl derivatives requires at least 15 mM p-(L-SULV) (Figure 3.1a). This is not so surprising, since an increase in polymeric surfactant concentration essentially serves to extend the chiral elution window. 80 Table 3.2 Class I: Examples of compounds exhibiting strong chiral interactions with p-(L-SULV) and requiring only 5-20 mM PS for enantioseparation. Charge of mM Compound i compound Rs k2′ a under given conditions 1. Partially a 5 2.05 1.54 1.034 BOH Anionic OH (±0.003) OH 2. Neutral BNAa 5 1.74 1.74 1.023 (±0.005) NH2 NH2 3. Anionic 7 1.52 1.22 1.015 a BNP (±0.002) O O P O OH 4. Neutral Temazepamb 12 3.43 0.85 1.023 O N (±0.006) * OH Cl N O H 5. Neutral N 12 1.70 1.25 1.011 b * OH Oxazepam (±0.009) Cl N H O 6. Neutral N 12 2.53 1.44 1.013 b * OH Lorazepam (±0.008) Cl N Cl 81 Table 3.2 Class I compounds (cont'd) Charge of mM Compound i compound Rs k2′ a under given conditions 7. Cationic Laudanosolinec 20 2.36 1.14 1.044 HO (±0.004) HO HO * N HO 8. Cationic Norlaudanosolinec 10 4.27 1.37 1.072 HO (±0.007) HO HO * NH HO 9. Cationic Laudanosinec 20 1.44 1.35 1.014 O (±0.005) O O * N O 10. Partially Oxprenolold 12.5 1.41 0.65 1.014 O Cationic (±0.006) * O N H OH e 11. Partially Alprenolol 12.5 1.52 0.82 1.016 Cationic (±0.003) * O N H OH e 12. Partially Propranolol 12.5 1.41 1.32 1.015 Cationic (±0.005) NH * O OH 82 Table 3.2 Class I compounds (cont'd) Charge of mM Compound i compound Rs k2′ a under given conditions f 13. Neutral Cl Chlorothalidone 6 1.84 2.56 1.032 O OH (±0.003) S * NH H2N O O g 14. Neutral O Diltiazem 6 1.12 4.22 1.021 (±0.006) S * * O N O O N HO CF3 15. Neutral * 2,2,2- 6 3.14 2.57 1.044 Trifluoro- (±0.007) anthrylethanolg 16. Neutral Troger’s Basef 5 2.20 1.67 1.026 N (±0.005) N O PTH- 17. Neutral H 10 2.66 4.23 1.082 α- (±0.006) N * NH aminocaprylic S acidh O H 18. Cationic NH 10 1.55 4.55 1.056 N * HN (±0.008) NH NH2 S PTH-Arginineh a100 mM TRIS / 10 mM borate buffer, pH 10.2, 90 mbar*s, 30 kV, 220 nm; b25 mM TRIS / 25 mM borate buffer, pH 8.5, 30 mbar*s, 30 kV, 220 nm; c50 mM phosphate buffer, pH 7, 20 mbar*s, 30 kV, 220 nm; d50 mM borate buffer, pH 9.2, 5 mbar*s, 30 kV, 220 nm; e300 mM CAPS / 50 mM borate, pH 8.5, 5 mbar*s, 220 nm; f30 mM phosphate buffer, pH 7.0, 12 °C, 30 mbar*s, 30 kV; g50 mM borate buffer, pH 11.2, 12 °C, 30 mbar*s, 30 kV, 254 or 280 nm; h275 mM boric acid / 30 mM phosphate (dibasic), 10 mM triethylamine, pH 7.2; 50 mbar*s, 30 kV, 269 nm; i90% confidence interval calculated per four injections; (*) indicates chiral carbon. 83 mAU O O (a) Binaphthyl derivatives 40 P 30 O OH OH OH 20 NH 10 BNP BNA 2 NH2 0 BOH -10 min -20 -30 22 24 26 28 30 32 34 36 mAU HO (b) Paveroline derivatives 15 O HO 10 HO * O N 5 O * 0 Laudanosoline HO N -5 O Laudanosine -10 min 4 5 6 7 8 9 10 11 mAU (c) Benzodiazepines 15 O H O N N 10 O * OH * OH H N Cl N N Cl 5 * OH N Cl Cl 0 -5 Tzp Oxp Lzp -10 min 13 14 15 16 17 18 19 20 Figure 3.1 a) Simultaneous separation and enantioseparation of binaphthyl derivatives. Conditions: 15 mM p-(L-SULV); 14 °C; pH 10.0; 100 mM TRIS / 10 mM borate buffer; 90 mbar*s; 30 kV; 220 nm; 0.2 mg/mL of BNP = 1,1'- binaphthyl-2,2'-diyl-hydrogen phosphate, BOH = 1,1'-bi-2-naphthol, BNA = 1,1'-bi-2-naphthylamine. b) Simultaneous separation and enantioseparation of paveroline derivatives. Conditions: 20 mM p-(L-SULV), pH 7.0, 50 mM phosphate buffer. c) Simultaneous separation and enantioseparation of benzodiazepines: temazepam (Tzp), oxazepam (Oxp), lorazepam (Lzp). Conditions: 25 mM p-(L-SULV), pH 8.5; 100 mM TRIS / 10 mM borate buffer Conditions for Figure 3.1 b and c are same as Figure 3.1a unless otherwise stated. Simultaneous chiral resolution of laudanosoline and laudanosine derivatives was achieved using 20 mM p-(L-SULV) (Figure 3.1b). A comparison of the migration times and enantioresolution of laudanosoline and laudanosine demonstrates the ability of p-(L-SULV) to discriminate not only on the basis of hydrophobicity but also on the basis of hydrogen- 84 bonding interactions. The hydrophobically driven complexation between p-(L-SULV) and laudanosine resulted in a weaker chiral interaction than the hydrogen-bonding interactions offered by the multiple hydroxyl groups found on laudanosoline or norlaudanosoline (Table 3.2). Moreover, norlaudanosoline possesses a secondary amine group that offers an additional site for hydrogen-bonding, producing much greater enantioresolution than that observed for either laudanosine or laudonsoline. Another interesting comparison of class I analytes can be made for the chiral separation of the benzodiazepines. The three racemates of benzodiazepines separated in Figure 3.1c possess a similar aromatic skeleton. Structural differences lie in the number and type of functional groups attached to the aromatic ring. Relative to oxazepam, temazepam has an extra methyl group located on one of the ring nitrogens, and an additional Cl group is located in the ortho position of the lower benzene ring of lorazepam. As shown in Table 3.2, p-(L-SULV) individually resolved the enantiomers of temazepam, oxazepam and lorazepam with resolution values of 3.44, 1.70 and 2.53, respectively. Surprisingly, temazepam, which has fewer hydrogen-bonding sites than oxazepam or lorazepam, provides the fastest separation with greatest enantioresolution. Evidently, the hydrogen-bonding moiety at the ring nitrogen of benzodiazepines seems to play a minor role in chiral recognition since the presence of a methyl group on this ring nitrogen of temazepam does not sterically inhibit chiral recognition. Since the chiral resolution and chiral selectivity values of the benzodiazepines (temazepam, oxazepam, and lorazepam) do not follow any particular trends, no clear explanation is available at this time for the preferential sites of interaction of these analytes with p-(L-SULV). 85 In general, the enantioselectivity of p-(L-SULV) for the separation of class I analytes seems to be similar to single amino acid surfactants such as poly sodium N-undecanoyl-L- leucinate p-(L-SUL) and poly sodium N-undecanoyl-L-valinate p-(L-SUV) [25, 26]. However, some chiral anionic compounds (e.g., BNP) or neutral compound (e.g., oxazepam) can be resolved with p-(L-SULV) that could not be separated (under neutral to basic conditions) using either p-(L-SUV) or p-(L-SUL) alone. In addition, for all concentrations of p-(L-SULV), the migration times for cationic compounds are significantly lower than those obtained using either p-(L-SUV) or p-(L-SUL). 3.3.2 Enantioseparation of Class II Analytes This class of analytes includes phenylthiohydantoin amino acids (PTH-AAs), barbiturates, phenylhydantoin derivatives, coumarinic derivatives, and benzoin derivatives. In general, class II analytes require moderate concentrations (i.e., 30-50 mM) of p-(L- SULV) for chiral separations to be achieved. Table 3.3 shows that 19 of the 20 racemic class II analytes, representing ~25% of all analytes investigated, were successfully resolved with, Rs values ranging from 1.44 to 4.9. Figure 3.2 and Table 3.3 show four comparisons of structurally related compounds of class II analytes. The enantiomeric pairs of eight PTH-AAs were successfully separated as shown in Table 3.3 with Rs values over the range of 1.46 to 4.92. Although the racemate of each PTH-AA was well separated using 50 mM p-(L-SULV) in a 30 mM phosphate (dibasic) / 275 mM boric acid buffer (pH 7.2), the simultaneous separation of all eight racemates was not achieved in a single run due to overlapping k´ values (e.g., PTH-tyrosine, PTH- norleucine, and PTH-phenylalanine). 86 Table 3.3 Class II: Examples of enantiomers exhibiting moderate chiral interactions with p-(L-SULV) and requiring 30-50 mM PS for enantioseparation. Charge of mM Compound i compound Rs k2′ a under given conditions O H N R N S Phenylthiohydantoin (PTH); R = amino acid chain 1. Neutral PTH- 50 2.33 0.94 1.062 R= CH a valine (±0.002) 2. Neutral H2 PTH- 50 4.92 1.23 1.085 R= C norvaline H2 (±0.003) C OH R= 3. Neutral PTH-tyrosinea 50 1.46 2.23 1.033 (±0.007) 4. Neutral H2 PTH- 50 2.82 2.47 1.056 R= C a norleucine (±0.008) 5. Neutral PTH- 50 3.02 1.56 1.063 R= CH a isoleucine (±0.002) 6. Neutral PTH- 50 1.84 2.54 1.042 a tryptophan (±0.006) CNH R= 7. Neutral H2 PTH- 50 2.63 2.25 1.046 R= C phenylalanine (±0.007) a 8. Cationic R= C PTH- 50 1.47 1.43 1.035 NH histidinea N (±0.006) 87 Table 3.3 Class II compounds (cont'd) Charge of mM Compound i compound Rs k2′ a under given conditions 9. Neutral 30 1.44 1.26 1.016 5-methyl-5-phenyl- (±0.004) hydantoina HN * O N O H OH 10. Neutral 5-(4-hydroxyphenyl)- 30 2.23 1.56 1.053 5-phenylhydantoina (±0.002) HN * O N O H 11. Neutral 30 1.86 2.17 1.032 5-(4-methylphenyl)- (±0.006) 5-phenylhydantoina HN * O N O H b 12. Neutral Hydrobenzoin 50 1.45 1.72 1.017 OH (±0.002) * HO 13. Neutral O OH Benzoinb 50 2.95 2.32 1.054 * (±0.006) 88 Table 3.3 Class II compounds (cont'd) Charge of mM Compound i compound Rs k2′ a under given conditions O O 14. Neutral Benzoin 50 2.33 3.56 1.056 b * methyl ether (±0.007) 15. Neutral O O Benzoin 50 0.62 3.82 1.017 ethyletherb * (±0.005) 16. Neutral Benzoin 50 0.30 4.22 1.011 O O Isobutyl (±0.005) b * ether O a 17. Anionic O Warfarin 50 1.51 0.88 1.023 * (±0.005) OH O 18. Anionic O O Cl Coumachlora 50 3.12 1.21 1.038 * (±0.002) OH O H O O 19. Partially N Pentobarbitalc 50 1.48 1.33 1.014 Anionic NH (±0.003) * O O H 20. Partially N SecobarbitalO c 50 1.52 1.60 1.018 Anionic NH (±0.003) * O a275 mM boric acid / 30 mM phosphate (dibasic) /10 mM triethylamine buffer, pH 7.2, 30- 150 mbar*s, 269 nm; b50 mM phosphate, pH 7.0, 12 °C, 30 kV, 220 nm; c50 mM phosphate, pH 7.2, 12 °C, 30 kV, 220 nm; d90% confidence interval calculated per four injections. 89 mAU 11’ (a) PTH-AAs 17.5 1. valine CH 3’ 3 H 2 2 2’ 2. norvaline 12.5 C H 2 3. tyrosine C OH 7.5 4. α -aminocaprylic acid H 2 2.5 C 4 4’ min 0 -2.5 10 15 20 25 30 35 40 45 mAU30 (b) Coumarinic derivatives OO 25 OO Cl * 20 * OH OH 15 O 10 O Warfarin Coumachlor 5 min 0 14 15 16 17 18 19 mAU H O H O N O (c) Barbiturates N O 25 NH NH 20 * 15 O * O 10 5 Pentobarbital Secobarbital 0 -5 10 11 12 min 13 14 Figure 3.2 a) Simultaneous separation and enantioseparation of four phenylthiohydantoin amino acids (PTH-AAs). Conditions: 0.2 mg/mL; 50 ° mM p-(L-SULV); 15 C; pH 7.2; 275 mM boric acid / 30 mM Na2HPO4 / 10 mM triethylamine buffer; pH 7.2; 150 mbar*s; 30 kV; 269 nm. b) Simultaneous separation and enantioseparation of 0.2 mg/mL each of the coumarinic derivatives. Conditions: 50 mM p-(L-SULV); 15 °C; 50 mM phosphate (mono/dibasic), pH 7.0; 280 nm; 150 mbar*s; 30 kV. c) Simultaneous separation and enantioseparation of barbiturates (0.4 mg/mL). Conditions: 50 mM p-(L-SULV); 15 °C; pH 7.2; 30 mM phosphate (dibasic) / 300 mM boric acid buffer; 269 nm; 150 mbar*s; 30 kV. Figure 3.2a shows the simultaneous enantioseparation of four PTH-AAs. Note that PTH-α- aminocaprylic acid is also classified under the class I analytes since it provided high resolution (Rs = 2.66) at lower concentrations of p-(L-SULV) (i.e., 10 mM, Table 3.2). However, the resolution of this analyte was even higher (Rs = 9.3) when separated using the 90 conditions described in Figure 3.2a. Although the hydrophobicity of the PTH-AAs dictates the elution order, as shown in Figure 3.2a, a similar trend was not observed for their enantiomeric resolution. For example, PTH-α-aminocaprylic acid, the most hydrophobic PTH-AA with the longest retention time, had an exceptionally high Rs value of 9.3. However, under similar conditions PTH-tyrosine, with the third longest migration time, yielded a Rs value of only 1.4 which is much lower than the resolution obtained for the earlier eluting PTH-norvaline (Rs = 4.9). In general, anionic chiral compounds are difficult to separate using an anionic chiral selector under neutral-to-basic conditions. Among the several anionic classes of compounds examined in this study, two classes (coumarinic, barbiturate derivatives) showed significant enantioresolution with p-(L-SULV). Presumably, favorable chiral interactions between p- (L-SULV) and these anionic compounds likely arise from multiple hydrogen-bonding sites present (see Table 3.3, compounds 17-20). However, in this case, electrostatic repulsions cannot be ruled out as an interaction promoting chiral recognition. The simultaneous enantioseparation of two coumarinic derivatives (warfarin and coumachlor) is shown in Figure 3.2b. Warfarin is used as an anticoagulant in the treatment of thromboembolic disease and as a rodenticide while coumachlor is solely used as a rodenticide [27,28]. Both of these compounds possess multiple hydrogen-bonding moieties and are electronegative due to their keto-enol group. The phenolic functional group of warfarin has a pKa of 5.1 [9]. Therefore, these compounds typically exist as anions under the conditions used for the separation (pH 7, 50 mM phosphate buffer). Higher resolution and a longer retention time were achieved with coumachlor relative to warfarin. 91 This is likely attributable to the larger electronegativity due to the presence of a chloride group on coumachlor. Barbiturates (secobarbital, pentobarbital, mephobarbital, and hexobarbital) are acidic compounds (pKa 7.6-8.0) that exist as partially anionic compounds in a phosphate BGE with pH 7.0 as was used for these separations. In general, barbiturate compounds can possess two centers of chirality if (i) two different C5 substituents are present and (ii) one of the keto-enol ring nitrogens is substituted. However, in the case of pentobarbital and secobarbital, the keto-enol ring system is not chiral, and only one stereogenic center is present, as indicated in Table 3.3 (compounds 19 and 20). The enantioseparation of these compounds resulted in broad peaks under the given conditions (see Figure 3.2c) which may be the result of the partial dissociation of the barbiturates to their corresponding anionic conjugate bases. Another plausible explanation for the peak broadening is a mobility mismatch between the analytes and the BGE. In contrast, separations of mephobarbital and hexobarbital (compounds 19-20, Table 3.4) resulted in little to no chiral resolution under all pH conditions investigated. Unlike pentobarbital and secobarbital, mephobarbital and hexobarbital possess a stereogenic center within the keto-enol ring system. It appears that the presence of a phenyl group and a cyclohexene group attached to the chiral carbon in mephobarbital and hexobarbital, respectively, causes considerable steric hindrance that inhibits favorable micelle polymer-analyte interactions. Therefore, only slight chiral resolution with much lower k´ values (k´ = 1.18 - 1.22), was obtained for mephobarbital and hexobarbital even when the concentration of p-(L-SULV) was as high as 85 mM. The enantiomeric separation of a series of benzoin derivatives was studied to show the effects of steric, hydrophobic, and hydrogen-bonding capabilities of these compounds 92 mAU O OH 50 * O O 40 * Benzoin 30 OH * * 20 Benzoin methyl ether HO Hydrobenzoin 10 0 5 10 15 20 25 30 min Figure 3.3 Simultaneous separation and enantioseparation of 0.5 mg/mL each of the benzoin derivatives (hydrobenzoin, benzoin, and benzoin methyl ether). Conditions: same as Figure 2(b) except detection was performed at 220 nm. with p-(L-SULV). Figure 3.3 shows the electropherogram for the simultaneous separation of hydrobenzoin, benzoin, and benzoin methyl ether. Differences in chiral resolution values of all three benzoin derivatives demonstrates that p-(L-SULV) exhibits chiral binding interactions that are sensitive to the substituent present at the stereogenic centers. In the case of benzoin and benzoin methyl ether, the presence of the carbonyl group (which acts as a hydrogen bond acceptor) decreases congestion near the stereogenic center. This decrease in congestion results in additional rigidity and a stronger complex with p-(L-SULV) relative to hydrobenzoin. Furthermore, a comparison of Rs values for the benzoin derivatives in Table 3.3 also indicates that, in the case of hydrobenzoin, the presence of an additional hydroxyl group (which can act as a hydrogen-bond acceptor) was not necessary for chiral 93 recognition. This is shown by the lower resolution observed for hydrobenzoin (Rs = 1.45) compared to that of benzoin (Rs = 2.95) and benzoin methyl ether (Rs =2.33). It should be noted that the more hydrophobic benzoin derivatives (benzoin ethyl ether and benzoin isobutyl ether, Table 3.3), demonstrated unfavorable steric interactions and showed only slight enantioselectivity. The separation of hydantoin racemates in which the stereogenic center forms a part of the ring system reveals some interesting trends as shown in Table 3.3. Rs and k´ values obtained for the enantiomers of 5-phenylhydantoin derivatives (compounds 9-11, Table 3.3) show a marked decrease as the number of aromatic ring systems attached to the stereogenic center drops from two to one. Further comparison shows that the presence of a hydroxyl substituent on one of the aromatic rings (e.g., 5-(4-hydroxyphenyl)-5- phenylhydantoin, Rs = 2.23) is more favorable than the addition of a methyl group to the same position of the aromatic ring (e.g., 5-(4-methylphenyl)-5-phenylhydantoin, Rs = 1.86). 3.3.3 Enantioseparation of Class III Analytes The data for the chiral separations of class III analytes, those requiring 75-85 mM p- (L-SULV) are tabulated in Table 3.4. Interestingly, all of the analytes that are enantioseparated in this class (with the exception of nefopam) are cationic and exhibit only weak chiral interactions with p-(L-SULV). Weak chiral interactions are said to be exhibited by this class because high concentrations of p-(L-SULV) (e.g., 75-85 mM) are required to achieve simply borderline baseline separations with Rs values of 1.33 - 1.72 and low retention factors. All β-adrenergic blockers (β-blockers) contain at least one aromatic ring and an alkanolamine side chain terminating in a secondary amino group, which includes the chiral 94 Table 3.4 Class III: Examples of enantiomers exhibiting weak chiral interactions with p-(L-SULV) and requiring 75-85 mM PS for enantioseparation Charge of mM Compound i compound Rs k2′ a under given conditions 1. Partially Pindolola 75 1.33 2.84 1.028 OH Cationic O (±0.006) H * N HN a 2. Partially O Metoprolol 75 1.42 3.23 1.023 Cationic (±0.006) O OH * NH O 3. Partially Atenolola 75 0.50 2.30 1.010 H2N Cationic O OH (±0.002) * NH 4. Partially O O Acebutol 75 0.20 2.72 1.010 a Cationic -ol (±0.001) HN O OH * NH 5. Partially Terbutalinea 75 2.02 3.06 1.052 HO Cationic (±0.008) NH HO * OH 6. Partially Epinephrinea 75 1.34 1.25 1.022 HO Cationic (±0.003) NH HO * OH a 7. Partially Isoproternol 75 1.41 1.64 1.033 Cationic HO (±0.006) NH HO * OH 95 Table 3.4 Class III compounds (cont'd) Charge of mM Compound i compound Rs k2′ a under given conditions OH a 8. Partially Ephedrine 75 1.55 2.25 1.019 Cationic * * (±0.00) NH 9. Partially OH Pseudo- 75 1.41 2.35 1.020 Cationic ephedrinea * (±0.006) * NH a 10. Partially OH Oxyephedrine 75 1.41 2.38 1.020 Cationic (±0.006) * * HN O 11. Partially OH Methyl 75 1.52 2.75 1.028 a Cationic ephedrine (±0.004) * N 12. Neutral Cl 75 1.72 1.45 1.029 O a HN Ketamine (±0.003) * 13. Cationic Homatropinea 75 1.52 1.33 1.019 N (±0.009) O O * HO 96 Table 3.4 Class III compounds (cont'd) Charge of mM Compound i compound Rs k2′ a under given conditions a 14. Cationic N Atropine 75 0.72 1.55 1.025 (±0.006) O O * OH Norephedrinea 15. Partially OH 75 1.50 1.92 1.029 Cationic (±0.008) * NH2 16. Partially Aminoglutethimideb 80 5.40 0.88 1.060 O Cationic (±0.007) NH HN * 2 O 17. Partially O Glutethimideb 80 1.43 1.42 1.013 Cationic (±0.008) HN * O b 18. Neutral Nefopam 80 0.45 1.46 1.011 (±0.006) O * N 19. Partially Mephobarbitalc 85 0.42 1.18 1.012 Anionic (±0.002) 20. Partially Hexobarbitalc 85 0.22 1.22 1.011 Anionic (±0.002) a50 mM phosphate buffer, pH 6.5, 30 kV, 220 nm, 13 °C, 20 mbar sec; b25 mM TRIS / 25 mM borate buffer, pH 8.5, 30 kV, 220 nm, 13 °C, 90 mbar*s; c50 mM phosphate buffer, pH 7.2, 30 kV, 220 nm, 90 mbar sec; d90% confidence interval calculated per four injections. 97 mAU O O Pindolol (a) β-Blockers 25 20 OH O O OH Metoprolol 15 H * N 10 * NH HN 5 0 -5 min -10 mAU , 10 14 HO 11 12 13 50 OH NH (b) Aromatic amines HO H 45 HO * HO N NH * OH 40 HO * OH Epinephrine HO 35 Isoproterenol Terbutaline 30 min mAU 6 7 8 9 10 11 12 13 O (c) Glutethimide and 25 O aminoglutethimide 20 NH HN 2 15 * HN * 10 5 O Glutethimide 0 Aminoglutethimide O -5 min 12 14 16 18 20 22 24 -10 Figure 3.4 a) Simultaneous separation and enantioseparation of 0.5 mg/mL each of β- blockers. Conditions: 75 mM p-(L-SULV); 14 °C; pH 6.5; 50 mM phosphate (mono and dibasic) buffer; 30 mbar*s, 30 kV; 220 nm. b) Simultaneous separation and enantioseparation of 0.5 mg/mL each of aromatic amines. Conditions: same as Figure 3.4a except detection at 214 nm. c) Simultaneous separation and enantioseparation of 0.5 mg/mL each of aminoglutethimide and glutethimide. Conditions: 80 mM p-(L-SULV); 12 °C; pH 8.5; 25 mM TRIS / 25 mM borate; 90 mbar*s, 30 kV; 254 nm. center. Although the three β-blockers (propranolol, alprenolol, and oxprenolol) were baseline resolved using only 12.5 mM p-(L-SULV) (Table 3.2), a much higher concentration (i.e., 75 mM) of p-(L-SULV) was required for the chiral separations of the other two structural analogs – pindolol and metoprolol (Figure 3.4a). In addition, it appears that p-(L- SULV) exhibits better chiral resolution for β-blockers in which the substituents on the aromatic ring are positioned ortho to the alkanolamine side chain containing the chiral carbon. For example, propranolol, alprenolol, oxprenolol, metoprolol, and pindolol (see 98 Table 3.3 and Table 3.4 for molecular structures) all provided baseline resolution either at low or high concentration of p-(L-SULV). In contrast, acebutolol and atenolol, in which the substituent on the aromatic ring is positioned para to the alkanolamine side chain (compounds 3-4, Table 3.4), showed only slight chiral resolution. The chiral resolution, chiral selectivity, and k´ values for epinephrine, isoproterenol and terbutaline are illustrated in Figure 3.4b and tabulated in Table 3.4. All three secondary amines contain two hydroxyl groups on the aromatic ring adjacent to the chiral carbon. Examination of Figure 3.4b shows that both the retention time increases and the enantiomeric resolution of the three amines decreases as a function of the bulkiness on the amine functionality in the order: terbutaline > isoproterenol > epinephrine. The greater resolution obtained for terbutaline relative to the other two chiral amines suggests the position of the two hydroxyl groups on the aromatic ring may also play an active role in the relative binding of these compounds with p-(L-SULV). A comparison of chiral resolution observed for glutethimide and its structurally similar analog, aminoglutethimide (Table 3.4, Figure 3.4c), indicates that aminoglutethimide exhibits higher enantioresolution (Rs = 5.40) than glutethimide (Rs = 1.43), despite the shorter retention factors (k´ of 0.88 and 1.42, respectively). This comparison of electrophoretic data again points out the importance of electrostatic and hydrogen-bonding interactions with p-(L-SULV) that allow for improved chiral resolution. 3.4 Conclusions According to our current chiral MEKC database, a total of 58 out of 75 racemic compounds of different structural features have been resolved using p-(L-SULV). This represents an overall success rate of ~ 77%. Among the 58 resolved compounds, 19 chiral 99 compounds (~33%) provided superior chiral resolution (Rs > 2.00), 31 chiral compounds (~54%) provided satisfactory chiral resolution (1.00 < Rs < 1.90), and only 8 chiral compounds (~13%) provided Rs values less than 1.0. 3.5 References (1) Palmer, C. P.; McNair, H. M. J. Microcolumn Sep. 1993, 4, 509-514. (2) Palmer, C. P.; Khaled, M. Y.; McNair, H. M. J. High Res. Chromatogr. 1992, 15, 756-762. (3) Palmer, C. P.; Terabe, S. J. Microcolumn Sep. 1996, 8, 115-121. (4) Palmer, C. P.; Terabe, S. Anal. Chem. 1997, 69, 1852-1860. (5) Shamsi, S. A.; Akbay, C.; Warner, I. M. Anal. Chem. 1998, 70, 3078-3083. (6) Edwards, S. H.; Shamsi, S. A. J. Chromatogr., A 2000, 903, 227-236. (7) Wang, J.; Warner, I. M. Anal. Chem. 1994, 66, 3773-3776. (8) Wang, J.; Warner, I. M. J. Chromatogr., A 1995, 711, 297-304. (9) Agnew-Heard, K. A.; Sanchez-Pena, M.; Shamsi, S.; Warner, I. M. Anal. Chem. 1997, 69, 958-964. (10) Yarabe, H. H.; Shamsi, S. A.; Warner, I. M. Anal. Chem. 1999, 71, 3992-3999. (11) Shamsi, S. A.; Warner, I. M. Electrophoresis 1997, 18, 853-872. (12) Ozaki, H.; Itou, N.; Terabe, S.; Takada, Y.; Sakairi, M.; Koizumi, H. J. Chromatogr., A 1995, 716, 69-79. (13) Rundlett, K. L.; Armstrong, D. W. J. Chromatogr., A 1996, 721, 173-186. (14) Lu, W.; Shamsi, S. A.; McCarley, T. D.; Warner, I. M. Electrophoresis 1998, 19, 2193-2199. (15) Shamsi, S. A. Anal. Chem. 2001, 73, 5103-5108. 100 (16) Toussaint, B.; Palmer, M.; Chiap, P.; Hubert, P.; Crommen, J. Electrophoresis 2001, 22, 1363-1372. (17) Cherkaoui, S.; Rudaz, S.; Varesio, E.; Veuthev, J.-L. Electrophoresis 2001, 22, 3308-3315. (18) Lu, W.; Poon, G. K.; Carmichael, P. L.; Cole, R. B. Anal. Chem. 1996, 68, 668-674. (19) Kirby, D.; Greve, K. F.; Foret, F.; Vourous, P.; Karger, B. L. In 42nd ASMS Conference on Mass Spectrometry: Chicago, IL, 1994, pp 1014-1015. (20) Shamsi, S. A.; Macossay, J.; Warner, I. M. Anal. Chem. 1997, 69, 2980-2987. (21) Macossay, J.; Shamsi, S. A.; Warner, I. M. Tetrahedron Lett. 1999, 40, 577-580. (22) Billiot, E.; Macossay, J.; Thibodeaux, S.; Shamsi, S. A.; Warner, I. M. Anal. Chem. 1998, 70, 1375-1381. (23) Billiot, E.; Agbaria, R. A.; Thibodeaux, S.; Shamsi, S. A.; Warner, I. M. Anal. Chem. 1999, 71, 1252-1256. (24) Haddadian, F.; Billiot, E. J.; Shamsi, S. A.; Warner, I. M. J. Chromatogr., A 1999, 858, 219-227. (25) Billiot, E.; Warner, I. M. Anal. Chem. 2000, 72, 1740-1748. (26) Billiot, E.; Thibodeaux, S. J.; Shamsi, S. A.; Warner, I. M. Anal. Chem. 1999, 71, 4044-4049. (27) www.hclrss.demon.co.uk./coumachlor.html. (28) http://inchem.org/documents/hsg/hsg/hsg096.htm. 101 CHAPTER 4. UNDERSTANDING CHIRAL SEPARATIONS INVOLVING P-(SULV) USING STEADY-STATE FLUORESCENCE ANISOTROPY, CAPILLARY ELECTROPHORESIS, AND NMR 4.1 Introduction Chiral separations have become an area of ever-increasing attention in the last decade due to the vast number of chiral drug compounds found to demonstrate different physiologic properties between their respective enantiomers [1, 2]. In the past, the “gold standard” for studying chiral separations fell within the realm of chromatography (e.g., HPLC, CE, and some GC), and numerous publications have demonstrated the chiral separation of countless drug compounds using the above methods [3-13]. Recently, however, several broad-ranging analytical techniques such as calorimetry, circular dichroism, crystallography, NMR and fluorescence spectroscopy have demonstrated their ability for studying chiral interactions as well [14-31]. While achieving chiral separations is not a trivial undertaking, it has become more commonplace as seen by the number of publications demonstrating chiral separations of large groups of analytes. The next hurdle separation scientists need to surpass is to fully understand the mechanism(s) that give rise to chiral separations which is vital in predicting how to undertake a separation problem. In addition to MEKC studies, NMR and fluorescence techniques have been used to better define mechanisms leading to the separation of the analytes employed here with a chiral polymeric surfactant. NOESY NMR is used to identify the regions of a primary structure (analyte) involved in an interaction with a macromolecule (polymeric surfactant) through the use of binding interaction mapping techniques [31, 32] Fluorescence spectroscopy has also shown much progress in the study of chiral recognition studies 102 because of its high sensitivity and selectivity, together with its ability to probe molecular interactions such as selector-selectand binding interactions through the use of fluorescence anisotropy (FA). The study reported here has a three-fold purpose. The first goal is to use MEKC and NOESY NMR methods to gain a better understanding of chiral separations involving dipeptide polymeric surfactants (PSs) by determining the primary site of interaction between a select group of chiral analytes and a given PS. The second is to determine if MEKC could be used as an additional tool for investigating chiral interactions compared with the more established method of NMR. The third goal of this chapter was to validate a previous study by our group that suggested a direct relationship between chiral selectivity, α, and the spectroscopic parameter β as well as to provide preliminary evidence that the slope is related to the separation mechanism. 4.1.1 Nuclear Magnetic Resonance Theory Various NMR experiments (e.g., NOESY, NOE difference, and pulsed field gradient NMR) can be used to generate binding interaction maps, also called group epitope maps, involving several chiral analytes and chiral selectors. The resulting maps offer an additional tool with which to explore the chiral analyte–chiral selector intermolecular interactions responsible for chiral resolution. In this study, NOESY NMR was used as a validation tool to verify the primary site of interaction between several chiral analytes and the PS, p-(L- SULV). In NOESY, cross peak volume (i.e., peaks with different frequencies in two dimensions) is directly proportional to the distance between protons experiencing an NOE effect (Chapter 1.7.2) [31]. Hence, comparison of the relative NOE strengths between all 103 analyte resonances and a given PS peak helps to identify those analyte protons in closest proximity and thus most likely experiencing the strongest interactions with the PS. NOE difference experiments, in turn, yield peaks corresponding to analyte protons experiencing NOEs to saturated protons on the PS where the intensity of the peaks indicates the relative distances between the saturated proton and the protons experiencing NOEs. 4.1.2 α versus β Theory Previously, our group discovered an unexpected correlation between chromatographic selectivity and fluorescence anisotropy data gathered from a single chiral selector–selectand system. Our initial thoughts were that this discovery may lead to a new approach to further elucidate the chiral recognition mechanism involved using polymeric surfactants. In this initial study, we found a linear correlation (i.e., R2 = 0.994) between the selectivity observed in MEKC experiments and steady-state fluorescence anisotropy experiments (α- and β-values, respectively) using four model compounds (Appendix I) [33]. As noted in Chapter 1, selectivity is defined as t α = R (4.1) tS where tR and tS are the adjusted retention times for the R and S enantiomers, respectively. In addition, the parameter β is defined as a function of the ratio of the anisotropy ( rS rR ) of the S and R enantiomers in the presence of a chiral selector. r 3cos2 β −1 S = (4.2) rR 2 Recalling the linear relationship observed between α and β as illustrated in Appendix I, it is apparent that the two terms are related by 104 α = mβ + I (4.3) Equation 4.3 expresses the relationship between chiral selectivity determined from capillary electrophoretic separations (α ) to that observed using steady-state fluorescence anisotropy (β ), where m is the slope and I is the intercept. If chiral recognition is not observed from CE experiments (i.e., from Eq. 4.1, if tR = tS, then α = 1), then differences in fluorescence anisotropy should not be observed either (i.e., from Eq. 4.2, rS rR = 1 then β = 0). If this is true, the intercept, I, must equal unity. Substituting for I and applying the natural logarithm to Eq. 4.3 gives rise to the following expression: ln(α) = ln(mβ +1) (4.4) Furthermore, according to the literature, the difference in free energy of association between two enantiomers for a given chiral selector is related to the logarithm of the selectivity as described below. ∆(∆G)S ,R = −RT ln(α) (4.5) Therefore, substituting mβ +1 for α into Eq. 4.5 gives ∆(∆G)S ,R = −RT ln(mβ +1) (4.6) where m is the slope of the plot of α versus β and is also believed to be related to the separation mechanism. After our initial publication, we postulated that if the observed relationship between α and β is related to different chiral recognition mechanisms, then different classes of chiral analytes will have different slopes, thus reflecting the difference in chiral separation mechanisms. Our initial study as performed with a dipeptide containing two leucine, i.e., one type of chiral selector. The focus of the study presented here is to examine a larger group of chiral analytes and examine their α, β relationships with two 105 different chiral centers. The classes of compounds examined this study are shown in Figure 4.1 and include four dansyl amino acids, three binaphthyl derivatives, and Troger’s base. Chiral separations employing four diastereomers of poly sodium N-undecanoyl leucylvalinate (SULV) as chiral selectors are probed by the use of MEKC, steady-state fluorescence anisotropy, and NMR. By employing diastereomers, and thus altering the stereochemistry of a single amino acid in a systematic way, one may control the enantiorecognition ability of the chiral selector. As a result, one can gain a better understanding of the mechanisms of chiral recognition for the two classes of neutral or anionic chiral analytes studied. An evaluation of the chiral interactions leading to chiral separations confirmed our earlier observation of a strong relationship between the selectivity (α) observed using a chromatographic separation technique (MEKC) and that determined from the spectroscopic parameter β [33]. A linear α versus β relationship was observed for the polymeric surfactant p-(L-SULV) with all eight analytes included in this study. However, as we predicted, different groups of analytes had different slopes, i.e., values of ″m″. An evaluation of the data allowed a grouping of the analytes according to the primary site of chiral interaction with the leucine or valine moiety of PS chiral head group. 4.2 Experimental 4.2.1 Materials and Reagents The pure enantiomers of Troger’s base (TB), 1,1'-binaphthyl-2,2'-diamine (BNA), 1,1'-bi-2-naphthol (BOH), 1,1'-binaphthyl-2,2'-diyl hydrogen phosphate (BNP), dansyl- phenylalanine (Dns-Phe), tryptophan (>98%), leucine (>98%), and norleucine (>98%) were obtained from the Sigma-Aldrich Corporation (Milwaukee, WI). The analytical grade reagents: dansyl chloride (>99%), sodium bicarbonate, ethyl acetate, and methanol were also 106 obtained from the Sigma-Aldrich Corporation. The dipeptide, (L,L)-leucine-valine (LV), was obtained from Bachem (King of Prussia, PA) while the (D,D)-, (D,L)-, and (L,D)-LV diastereomeric forms were from Genscript Corporation (Piscataway, NJ). All were of 99% purity or better. Sodium phosphate dibasic (Na2HPO4) and sodium borate (NaB4O7) were from Mallinckrodt and Baker, Inc. (Paris, KY) and Fisher Scientific (Fair Lawn, NJ), respectively. Normal phase TLC plates (100 µm) were obtained from Sorbent Technologies (Atlanta, GA). All reagents were used as received. N N O O OH NH2 P O OOS OOS O OH OH NH2 NH NH HO * * BNP BOH BNA Dns-Phe HO O Dns-Leu N N N N O OOS OOS TB NH NH HO * * HO O HN Dns-Nor Dns-Trp Figure 4.1 Molecular structure of the eight analytes investigated. The four dansyl amino acids and TB possess an asymmetric center denoted by the red asterisk, whereas the binaphthyl derivatives possess an asymmetric plane (not denoted). 4.2.2 Dansyl Amino Acid (Dns-AA) Derivatization Dansyl amino acids (Dns-AA) are generally not commercially available in enantiomerically pure form. Therefore, derivatization of amino acid residues (i.e., leucine 107 (Leu), tryptophan (Trp), and norleucine (Nor)) with dansyl chloride was performed using the method of Tsai and co-workers [34]. Briefly, 100 µL of each: 20 mM dansyl chloride, 500 mM sodium bicarbonate, and 0.4 mg/mL amino acid were allowed to react in an amber scintillation vial at 65°C for 40 minutes in a water bath. This derivatization procedure yielded the desired chiral Dns-AA as well as a highly fluorescent achiral by-product, the 5- dimethylamino-naphthalene-1-sulfonic acid anion, also known as dansyl sulphonate. Consequently, preparative thin layer chromatography (Prep–TLC) was used to purify the Dns-AAs. 4.2.3 Prep–TLC Derivatized D and L forms of the amino acids were separately spotted onto prep-TLC plates and a mobile phase consisting of 72.9/15/12.1 (v/v/v) ethyl acetate, methanol, and water was used to separate them from the derivatization by-product. The resulting bands were visualized under a handheld UV light source at 254 nm, identified, and collected by physically scraping the silica from the glass plate. Methanol was used to extract the purified Dns-AAs from the silica. After sonicating the methanol–silica slurry for 30 min, the silica was allowed to settle out of solution. To ensure that no silica remained with the Dns-AA, the solution was subsequently transferred into a clean, pre-weighed scintillation vial using a syringe equipped with a 0.45 µm nylon filter (Nalgene, Rochester, NY). Next, methanol was purged under a stream of ultra high purity N2. Finally, stock solutions of an appropriate concentration (ca. 5 mg/mL) were prepared for each Dns-AA. The identification procedure and digital photographs of the TLC separations are shown in Appendix II. 108 4.2.4 Synthesis of Poly(Sodium N-Undecanoyl-L-Leucylvalinate), p-(L-SULV), and Its Diastereomers The surfactants were synthesized by reacting the N-hydroxysuccinimide ester of undecylenic acid with the dipeptides, (L,L)-LV, (D,D)-LV, (D,L)-LV, and (L,D)-LV, respectively, to form the corresponding N-alkyl chiral surfactants. Subsequent irradiation of the monomers with 60Co γ-radiation for seven days yielded the corresponding polymerized surfactants. A more detailed description of the above synthetic procedure is found in Chapter 2 of this dissertation. For simplicity, a conventional naming system for the PSs is as follows: when the configuration of both amino acids is identical (i.e., L,L or D,D), the polymer will be named either p-(L-SULV) or p-(D-SULV); when the configuration of both amino acids differs, (i.e., D,L or L,D), the polymer will be respectively named p-(D,L)- SULV or p-(L,D)-SULV. 4.2.5 Centrifugal Filtration Disposable ultrafree Millipore centrifugation filters, with 30,000 molecular weight cut off (MWCO) membranes, were used to reduce the polydispersity of the polymeric surfactants. Typically, 15 mL aliquots of a 100 mM solution of each polymeric surfactant were placed into a centrifugation filter and subsequently placed into a Centra GP8 (Thermo IEC, Philadelphia, PA) centrifuge at a speed of 3400 RPM for 20 min. The higher molecular weight surfactant concentrate was then collected and lyophilized to complete dryness, yielding polymeric surfactants of a specific size range (i.e., 30,000 MW +). 4.2.6 Capillary Electrophoresis Instrumentation and Procedures MEKC experiments were conducted on a Beckman Coulter MDQ CE instrument (Fullerton, CA) equipped with a photodiode array detector. Separations were performed using a 60 cm (52 cm effective length) × 50 µm i.d. uncoated fused silica capillary 109 (Polymicro Technologies, Phoenix, AZ). The cartridge temperature was maintained at 15°C using a liquid cooling system. UV detection was performed at the cathodic end at 254 nm and the applied voltage was 30 kV. Analytes were introduced into the capillary via pressure injection at 0.5 psi for 5 s. 4.2.7 Background Electrolyte (BGE) and Capillary Electrophoresis Standard Preparation The BGE used as the running buffer for all analytes was a 50 mM phosphate (dibasic) / 25 mM borate buffer (pH 9.0). The pH of the BGE was adjusted with 1 M sodium hydroxide. Next, the solution was filtered through a 0.45 µm nylon syringe filter and degassed for 10 minutes. Finally, 24 mM equivalent monomer concentration (EMC) of each monodispersed LV surfactant was added to individual BGE solutions. Analyte stock solutions were prepared in methanol followed by subsequent dilution to arrive at the appropriate final concentration. The final concentration for the binaphthyl compounds and TB was 0.2 mg/mL in a 1:1 methanol-water mixture while the final concentration for Dns-Phe, Dns-Trp, Dns-Leu, and Dns-Nor was ca. 0.25 mg/mL in a 1:4 methanol–BGE mixture. 4.2.8 Selectivity Calculation When making direct comparisons of selectivity values of charged and neutral species, as is done in this chapter, the issue related to differences in inherent electrophoretic mobility of the different species must be addressed. This is because the true α-values should be derived from the interaction of the analytes with the pseudostationary phase (polymeric surfactant in this case) and therefore α must be corrected for analyte electrophoretic mobility to obtain the true α. In order to do so, the selectivity formula was slightly modified in the following manner, 110 (t − t + t ) α′ = 2 0 sol (anions) (4.7) (t1 − t0 + tsol ) (t − t − t ) α′′ = 2 0 sol (cations) (4.8) (t1 − t0 − tsol ) where tsol refers to the retention time of the solute in the absence of PS, i.e., the adjusted mobility of the solute in a capillary zone electrophoresis (CZE) experiment (Chapter 1). While the standard calculation for selectivity is the ratio of the adjusted retention times of two adjacent solutes (Eq. 1.11), the modified equation either increases (for anions) or reduces (for cations) the retention time of charged species accordingly. For example, under the influence of an electric field, the electrophoretic mobility of an anionic solute is negative, or opposite, that of the EOF; therefore, adding the absolute value of the adjusted retention time would allow for direct comparison with neutral species which are devoid of electrophoretic mobility. In contrast, for cationic species the electrophoretic mobility of the solute is subtracted in order to correct for its enhanced effective mobility relative to the EOF. In the case of neutral species, the equation is unaffected because, tsol = 0. Selectivity data set reflected an average of 3-5 replicate MEKC measurements for each α-value presented with relative standard deviation (RSD) values less than 4.5%. 4.2.9 Fluorescence Spectroscopy Instrumentation and Procedures Steady-state fluorescence excitation/emission spectra and anisotropy measurements were acquired using a Spex Fluorolog®-3 spectrofluorometer model FL3-22TAU3 (Jobin Yvon-Spex, Edison, NJ) equipped with a 450-W xenon arc lamp, double grating excitation and emission monochromators, automated Glan-Thompson polarizers, and DatamaxTM for WindowsTM driving software. Anisotropy measurements were collected by use of the continuous wavelength analysis (CWA) mode using an L-format configuration. The 111 excitation/emission spectra collected served two purposes. First, they allowed for the determination of excitation and emission λmax values essential for anisotropy data acquisition in CWA mode and the second purpose was to ensure that the samples were free of contaminants. In addition, all data were automatically corrected for the G-factor by use of the DatamaxTM software. The temperature was maintained at 15°C for all samples using a VWR 1160 circulating water bath (West Chester, PA). Two FA studies were conducted: 1) an enantiomer binding affinity study and 2) an α vs. β study. For both studies, stock analyte solutions of a single enantiomer (2 mg/mL) were prepared in methanol. A 10-µL aliquot of stock enantiomer solution was delivered into a scintillation vial and purged with ultra high purity N2 to remove the methanol. In the first study, to a solution containing a single enantiomer of a given analyte and concentration (ca. 10-5 M), increasing amounts of PS were added. For the second study, sufficient BGE containing 24 mM PS was added to each single enantiomer sample to give a final analyte concentration of ca. 10-5 M. In both studies, samples were ultrasonicated, and allowed to equilibrate overnight before analysis. It is important to note that for both FA studies, all attempts were made to maintain conditions identical to MEKC studies such as BGE and temperature, with the exception of the applied voltage since applied voltage is not required for fluorescence anisotropy. The data reflect an average of 28 replicate anisotropy measurements that were averaged to arrive at each β-value presented. Fluorescence polarization and FA both provide a measure of the polarized emission of fluorescence and provide information about the rotational behavior of molecules. However, FA is generally preferred over polarization because many theoretical expressions are considerably easier to manipulate compared with polarization. For instance, the 112 additivity law of anisotropy offers a simple method for determination of the fractional contribution of a single fluorophore amongst a mixture [35]. Hence, in this chapter, FA was calculated according to the following equation: I − GI r = VV VH (4.9) IVV + 2GIVH where IVV and IVH are the vertically and horizontally polarized components of the emission, respectively. In this expression, G refers to the G-factor, an instrumental and wavelength correction factor, used to compensate for the polarization bias of the detection system, G = IHV/IHH. 4.3 Results and Discussion 4.3.1 MEKC Results Our group has developed an MEKC method for facile and unambiguous determination of the preferential site of interaction of an analyte with a polymeric diastereomeric dipeptide surfactant [36]. Due to the two optical configurations possible for each chiral center (D or L), dipepide amino acid-based polymeric surfactants having two asymmetric carbons, such as p-(L-SULV), can have at most four different optical configurations (D,D; D,L; L,L; L,D). According to the technique developed by our group, a single chiral analyte is first enantioseparated using the above-mentioned configurations (known as diastereomers) for a given dipeptide surfactant. The contribution of each chiral center to the overall enantioseparation is determined by monitoring the elution order of the enantiomers in the resulting electropherogram, where elution orders were determined by spiking a single pure R or D enantiomer into a solution of the corresponding racemic mixture. Consequently, a more intense peak corresponds to either the R- or D-enantiomer of 113 a given analyte. An example of the use of this technique for determining the primary site of interaction is given for the model compound, BNP in the following section. It should be noted that the conditions selected for MEKC separation of these eight analytes are not necessarily the optimum conditions for the individual analytes. Instead, conditions were selected such that near baseline resolution of analytes could be obtained for each analyte using p-(L-SULV) as a chiral selector. By maintaining constant conditions, (i.e., buffer pH and composition, temperature, PS concentration, etc.) experimental uncertainties are minimized and the charge state of the PS is kept constant. Thus, any observed differences in chiral selectivity should be due solely to differences in analyte–PS interactions. 4.3.1.1 BNP Results An examination of the separation observed using the all L configuration PS (i.e., p- (L-SULV)) illustrated in Figure 4.2 shows an elution order of R-BNP before S-BNP. According to the elution order of solutes observed in normal polarity CE (cations before neutrals before anions) and the fact that the PS is anionic, the enantiomer eluting first is considered to interact less with the PS. In contrast, due to the longer retention time, the second eluting enantiomer is said to show a stronger chiral interaction with the PS. Therefore, the electropherograms in Figure 4.2 suggest that the S enantiomer of BNP interacts more strongly with p-(L-SULV). MEKC experiments employing the three other diastereomers of p-SULV (i.e., p-(D- SULV), p-(D,L)-SULV, and p-(L,D)-SULV) were also investigated and are illustrated in Figure 4.2. The interpretation of this figure requires some explanation. As described earlier, when p-(L-SULV) is used as the chiral selector, the enantiomer elution order is R 114 before S. In full agreement with previous reports [37, 38], when p-(D-SULV) is used, the enantiomer elution order is reversed (i.e., S before R) due to the change in configuration of the amino acid head group from an all L configuration to all D configuration. Moreover, when p-(L,D)-SULV is used, enantiomer elution order resembles that of the all L configuration PS, implying that changing the optical configuration of the outermost chiral center (i.e., valine) does not significantly alter the chiral separation mechanism. When the configuration is changed to p-(D,L)-SULV, the configuration of the innermost amino acid is reversed, and the elution order is also reversed. This change suggests that it is the innermost amino acid (i.e., leucine) that primarily determines the enantiomer elution order and, therefore, is interacting to a greater extent with the BNP enantiomers during the separation. Thus, an evaluation of the four electropherograms shown in Figure 4.2 leads one to conclude that leucine (i.e., the innermost amino acid) plays a greater role than valine for the separation of BNP enantiomers. 4.3.1.2 Other Analytes For simplicity, the electropherograms corresponding to the seven remaining analytes are shown in Appendix III. Table 4.1, however, summarizes the findings (i.e., second eluting enantiomer and amino acid showing stronger interaction) for all eight analytes investigated in the presence of p-(L-SULV) and its diastereomers. In brief, the binaphthyl derivatives appear to interact to a greater extent with the innermost chiral center (i.e., leucine) whereas the Dns-AAs appear to prefer the outermost chiral center (i.e., valine). Likely, because TB has two chiral centers, a conclusive determination of its primary site of interaction using MEKC alone, it was not possible. 115 -4 -4 p-(L-SU(L,LL)V)-LV (Lp-(,D)-LV,D)-SULV -6 (0.1465) (0.1649) -6 (0.1506) (0.1619) mAU -8 -8 -10 -10 15 20 25 30 15 20 25 30 -4 -4 p-(D-SU(D,DL)-LV)V p-((D,LD)-LV,L)-SULV -6 (0.1525) (0.1450) (0.1528) (0.1390) -6 U -8 mA -8 -10 -12 -10 15 20 25 30 15 20 25 30 Time [min] Time [min] Figure 4.2 MEKC enantioseparation of the enantiomers of BNP in the presence of the four diastereomers of p-(SULV). Conditions: R-enantiomer is added in excess; 24 mM p-(SULV) in a 50 mM sodium phosphate (dibasic) / 25 mM sodium borate buffer (pH 9.0); 30 kV applied voltage, 15°C, 254 nm. Number in parenthesis denotes the observed anisotropy. Table 4.1 Enantiomeric selectivity pattern of several analytes towards p-(SULV) using MEKC. a Chiral analyteb Second eluting enantiomer Amino acid with stronger in the presence of p-(L-SULV) interaction with analyte BNP (–) S Leucine BOH (–) R Leucine BNA (n) R Leucine Troger’s Base (n) (+) * Dns-Phe (–) L Valine Dns-Trp (–) L Valine* Dns-Leu (–) L Valine* Dns-Nor (–) L Valine aMEKC conditions: 24 mM p-(L-SULV) in a 50 mM sodium phosphate (dibasic) / 25 mM sodium borate buffer (pH 9.0); 30 kV applied voltage, 15°C, 254 nm; bSymbol in parenthesis indicates charge of analyte under given experimental conditions where (n) is neutral and (−) is negative; ND, not conclusive; *determination not definitive using a single MEKC condition. 116 The primary difference among the analytes investigated, besides the charge state for the given experimental conditions, is the origin of asymmetry (i.e., asymmetric plane vs. asymmetric carbon). This, as well as size and the aromatic character of the analytes, seem to influence the chiral recognition ability of the PS chiral selector. For example, a close inspection of Appendix III.F and G reveals that the resolution of Dns-Leu and Dns-Nor enantiomers is significantly decreased for separations using the (D,L)- or (L,D)-SULV configuration PSs compared to those performed with the all L- or all D-SULV. As a consequence, the primary site(s) of interaction between enantiomer and polymeric surfactant cannot be determined using the single set of experimental conditions considered here. In order to clarify with which amino acid these analytes (i.e., Dns-Leu and Dns-Nor) interact more strongly, additional MEKC experiments were conducted whereby the PS concentration was markedly increased until elution orders could be determined. It is presumed that the aliphatic character observed only with Dns-Leu and Dns-Nor seems to affect the chiral recognition ability of the mixed diastereomer configuration PS. In addition, the charge state did not seem to strongly affect the site of interaction since both neutral analytes (i.e., BNA and TB) and anionic analytes (i.e., BNP and BOH) alike appear to interact preferentially with the leucine chiral center. Instead, it is the type of chiral selector that affects the primary site of interaction (i.e., leucine vs. valine). 4.3.2 Nuclear Overhauser Effect Spectroscopy (NOESY) Results NOESY experiments can be specific and yield convincing evidence pertaining to selector–selectand interactions that CE experiments. For this reason, NOESY NMR was chosen for further probing the analyte-PS chiral interaction(s) that lead to enantiomeric recognition using p-(L-SULV). Except in the case of BNP, where NOESY data were 117 collected for both the R and S enantiomers, only the enantiomer which interacted more strongly with p-(L-SULV) according to MEKC was subsequently analyzed using NOESY. 4.3.2.1 BNP Results NOESY analyses of S- and R-BNP in the presence of p-(L-SULV) yielded very exciting results. A schematic representation of a BNP molecule showing the proton numbering used in the following table and figures is shown in Figure 4.3a. The schematic shows that BNP possesses six pairs of nonequivalent protons labeled H3–H8 and H3'–H8', respectively. Chemically equivalent protons are denoted by the same color. It is also apparent that p-(L-SULV) has three types of protons labeled a, b, and c as well as α and β protons for each of the two chiral centers (Hα and Hβ for leucine (leu) and Hα and Hβ for valine (val)). Table 4.2 shows the relative NOE intensities observed between the BNP and p-(L- SULV) protons. All BNP protons show strong NOEs to the leucine and valine CH3 side chain R groups and the methylene protons (Hc) making up the alkyl chain of p-(L-SULV). This result suggests that the BNP analyte partitions between the polar head groups of the PS and the inside core of the micelle. This is in agreement with chiral CE theory which suggests that hydrophobic analytes undergo rapid mass transfer into and out of the micelle. The weak NOEs observed for both the leucine and valine Hα of p-(L-SULV) preclude the absolute identification of the primary site of interaction for BNP but do, however, indicate that the interaction with the valine and leucine Hα is generally weak. A possible explanation for this is that the chiral recognition ability of the PS is not limited to the chiral 118 centers but rather the chirality extends out to the surrounding ligands that give rise to the chirality of the molecule. Therefore, a closer look at the NOEs for these two protons must be taken. H H (a) 5 4 H6 H3 H7 O O H8' P H8 H O 7' OH H6' H3' H H 5' 4' leucine (b) c CH3 b β CH O H2C CH3 O- H2 H2 H2 H2 H C C C C CH N α C C C C C N α CH O H2 H2 H2 H2 H2 H a O CH H3C β CH3 valine Figure 4.3 Molecular structure of a) a BNP molecule and b) a p-(L-SULV) surfactant unit indicating the different protons on each molecule. 119 Table 4.2 Relative NOE intensities observed between all BNP protons and those of p- (L-SULV). H3 (H3') H4 (H4') H5 (H5') H6 (H6') H7 (H7') H8 (H8') w w w w w w Val Hα m m m m m m Val Hβ Leu Hα w w w w w w m m m m m m Leu Hβ Leu and/or s s s s s s Val–CH3 m m m m m m Ha m m m m m m Hb Hc s s s s s s w = weak m = medium s = strong Figure 4.4 shows the normalized NOEs experienced between the Hα protons of each of the chiral centers (leu Hα and val Hα) of p-(L-SULV) and the BNP protons. The height of each bar in Figure 4.4 is proportional to the NOE observed in the NOESY spectrum. It should be noted, that the stronger the NOE the closer in proximity the protons are to one another. Therefore, stronger NOEs observed between the analyte protons and one of the polymer chiral centers over the other chiral center suggests that, on average, the analyte protons are interacting to a greater extent with that chiral atom on the polymer that is closest in proximity. The NOE results obtained for S-BNP are straightforward. Figure 4.4 shows that for every S-BNP proton a stronger NOE is observed for the leu Hα of p-(L-SULV). The stronger NOE in this case suggests that under the given conditions (i.e., 50 mM phosphate (dibasic) / 25 mM borate buffer; pH 9.0, 15ºC), BNP is in closer proximity to, and therefore 120 interacts preferentially with, the leucine chiral center. It should be noted that the same result was obtained (i.e., BNP interacts preferentially with leucine) when a 100 mM Tris / 10 mM borate BGE buffered at pH 10 was used. Unlike with S-BNP, the R enantiomer of BNP shows stronger NOEs to the leu Hα of p-(L-SULV) for all BNP protons, except the H8(H8') protons. The H8(H8') protons here show a stronger NOE to the val Hα proton instead. Remarkably, this result supports the three point model (Chapter 1) which requires a minimum of three simultaneous interactions between a chiral selector and one enantiomer for chiral discrimination to occur. Because of this restriction, its antipode can achieve at most two of these interactions. It seems that the H8(H8') protons of S-BNP fulfill one of the three point requirements with p-(L-SULV) for chiral discrimination but fails that requirement with R-BNP thereby facilitating the successful enantioseparation of BNP enantiomers. These results are in agreement with the MEKC data above which indicate that the interaction was stronger with the leucine chiral center. 4.3.2.2 BOH Results Similar interactions between BOH and TB and the Ha, Hb, and Hc of the PS were observed. Therefore, from this point forward, only data pertaining to the leucine and valine Hα protons will be discussed. The proton designations for R-BOH are similar to those of the BNP enantiomers (Figure 4.5). Figure 4.5 suggests that there is a stronger interaction between R-BOH and the leucine Hα chiral center of p-(L-SULV) since higher NOEs are consistently observed for all R-BOH protons. This observation is consistent with the conclusions drawn from using the diastereomers of p-(SULV) in MEKC. 121 1600 E 1400 R-BNP Leucine Valine NO 1200 ed 1000 z i l 800 a 600 rm 400 No 200 0 H3 H4 H5 H6 H7 H8 1600 E 1400 S-BNP NO 1200 ed 1000 z i l 800 a 600 rm 400 No 200 0 H3 H4 H5 H6 H7 H8 BNP Proton Number Figure 4.4 Plots of NOE intensities observed between the S and R enantiomers of BNP and chiral centers of p-(L-SULV). 4.3.2.3 BNA Results NOESY data were not obtained for BNA due to the limited solubility of BNA for the extended period of time required for NMR analysis. MEKC results, however, unambiguously suggests that BNA enantiomers interact preferentially with the leucine chiral center. 4.3.2.4 Troger’s Base Results Figure 4.6 shows the molecular structure of TB where chemically nonequivalent TB protons are designated numerically and the chemically equivalent protons are designated by the same color. It should be noted that the proton labeled 13* was not clearly visible due to 122 spectral overlap by the leu Hα resonance. Because the NOEs to the leucine chiral center were more than an order of magnitude smaller than those of valine, only the TB proton interactions with the val Hα are plotted on Figure 4.6. This result suggests that the (+) enantiomer of TB interacts primarily with the val Hα chiral center. Leucine (R)-BOH H5 H4 80 Valine H6 H3 E 60 NO H7 OH d e H8' z H i 8 H7' OH al 40 m Nor 20 H6' H3' H5' H4' 0 H3 H4 H5 H6 H7 H8 BOH Proton Number Figure 4.5 NOESY data for R-BOH and p-(L-SULV) indicating that the R enantiomer of BOH interacts more strongly with the leucine chiral center (left) and the molecular structure of BOH (right). 4.3.2.5 Dansyl Amino Acid Results 1 A 1-D H-NMR spectrum of Dns-Trp alone reveals that its tryptophan moiety Hα appears at approximately 3.9 ppm. Those of the leucine and valine chiral centers of p-(L- SULV) are seen at around 4.2 and 4.0 ppm, respectively. A 1H-NMR of a Dns-Trp–polymer mixture showed resonances for the leu and val Hα of the polymer but not for the Dns-Trp Hα. A likely scenario is that the Dns-Trp Hα is eclipsed by the more intense valine Hα since the valine and Dns-Trp Hα protons both have resonances around 4.0 ppm. Due to the spectral overlap between the analyte Hα and the polymer val Hα peaks, the corresponding 2- 123 D NOESY spectra failed to show any NOEs between the analyte and PS. Furthermore, because of overlap, there is not sufficient evidence to conclude where the analyte preferentially interacts. However, MEKC data did demonstrate that the primary site of interaction for all Dns-AAs was with the valine chiral center of p-(L-SULV). 25 (+)Troger’s Base Proton 20 α H H10 H1 12 H H2 line 9 N C CH3 a 15 v 13* CH2 o t H3C C N H3 10 H2 6 H7 H4 * overlapped by Leucine Hα resonance. 5 alized NOEs Norm 0 H10(H4) H9(H3) H7(H1) H10(H4) H9(H3) H7(H1) Figure 4.6 NOESY data for (+)-TB and p-(L-SULV) indicating that the (+) isomer of TB interacts more strongly with the valine chiral center (left) and the molecular structure of TB (right). NOEs were, however, observed between the aromatic protons of the tryptophan and dansyl groups and the polymer hydrocarbon chain of the PS. As with the other analytes discussed, these NMR studies support the idea that the analyte partitions into and out of the core of the micelle. The above discussion regarding Dns-Trp and its partitioning into and out of the micelle can be applied to the four Dns-AAs included in this study. 4.3.3 Steady-State Fluorescence Spectroscopy Results To better understand the following results, the effect of free and bound analyte complexes on the observed anisotropy must be explained. Low anisotropies ( r ≈ 0 ) are 124 observed for uncomplexed or free fluorophores in low viscosity solutions because of their small size and unhindered rotation. The anisotropy increases when an analyte complexes with a nonfluorescing polymer due to the larger size, and hence slower rotation, of the polymer complex relative to the unbound analyte. The increase in anisotropy of a complex can reflect the degree of binding of the analyte to the polymer. For example, for a loosely bound analyte one observes an anisotropy value closer to that of the free species, whereas a more tightly bound complex will possess an anisotropy value closer to that of the polymer. It should be noted that FA is the weighted average anisotropy of fluorophores in all states (i.e., free and bound). Therefore, a greater proportion of one enantiomer will be bound for a longer period of time when compared to the other enantiomer. Therefore, an assumption is made that the enantiomer which is bound for a longer period of time and hence possesses a larger anisotropy is said to have a stronger interaction. Investigating analyte–polymer interactions, especially in the case of chiral species, is critical to understanding the underlying chiral separation mechanism(s). Although a select few research groups have investigated chiral interactions using fluorescence anisotropy [26, 33, 39], the inherent sensitivity and selectivity of this technique should lend itself toward the study of chiral separation mechanisms. The following study was designed to (1) assess the binding of select chiral analytes to p-(L-SULV) and (2) determine the applicability of the chiral selectivity–chiral recognition (i.e., α vs. β ) relationship first recognized by this group [33]. 4.3.3.1 Enantiomer Binding Affinity for p-(L-SULV) To determine enantiomer affinities for the polymeric surfactant, p-(L-SULV), BNP, BOH, and Dns-Phe were assayed for binding by measurement of the fluorescence anisotropy 125 as a function of polymer concentration. These data were compared to MEKC data collected using a similar experimental protocol. In doing so, I was able to determine the limitation of p-(L-SULV) for distinguishing two enantiomers. In particular, the minimum PS concentration required for each method to detect differences between the fluorophores was determined. It is known that an increase in surfactant concentration leads to an increase in resolution for MEKC separations [37]. Therefore, I anticipated a similar increase in the observed anisotropy of the enantiomers as the PS concentration increased. This increase in anisotropy should be larger for one enantiomer, indicating preferential binding between the enantiomers and the PS. Individual samples of R- or S-BNP were independently mixed with p-(L-SULV) and the resulting FA and chiral selectivity (α) data were collected. An examination of Figure 4.7a reveals that, for the PS concentration range studied, the S enantiomer of BNP generally has a larger anisotropy value than its counterpart, R-BNP. Since the FA of free R and S enantiomers is identical, these results imply that the S enantiomer interacts more strongly with the PS. This supports previous FA studies which suggest that the enantiomer with the higher observed anisotropy value interacts more strongly with the polymer [40]. This anisotropy result also agrees with the MEKC study performed using p-(L-SULV) and its diastereomers from Section 4.3.1 of this chapter. Analysis of the anisotropy curves with a student t-test [41] indicated that the fluorescence technique was able to distinguish the enantiomers with a 99.0% confidence level (CL) down to a PS concentration of 1 mM, designated by the arrow in Figure 4.7a. 126 A similar titration plot, where the selectivity of the MEKC separation of BNP enantiomers is plotted as a function of p-(L-SULV) concentration is shown in Figure 4.7b. As expected, a clear trend is observed where the selectivity increases with increasing chiral selector concentration. MEKC analysis suggests that a hint of separation was first observed at 0.5 mM p-(L-SULV) (arrow on Figure 4.7b) while baseline resolution is not observed until 1.5 mM. It should be noted that the minimum α value was marked by a selectivity value greater than unity. For both FA and MEKC, a maximum PS concentration was observed beyond which no further improvement in enantiomer resolution was observed for concentrations far exceeding 2.7 mM. 0.5 mM a) anisotropy 1.0 mM b) selectivity 1.07 0.030 0.027 1.06 0.024 1.05 0.021 ) 1.04 0.018 α , r ( y p 0.015 ity 1.03 ro t o 0.012 ctiv is le 1.02 n e a 0.009 s 1.01 0.006 0.003 t-test R-BNP 1.00 S-BNP alpha 0.000 0.99 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 (Lp-,L()L-S-SUULVL Vcon) cocentncenrattrioation, n[m, [Mm]M] p-(L(,L)-L-SSUULLVV) ccononccenentrtratioation,n [ m[mMM]] Figure 4.7 Titration curves showing the a) enantiomer anisotropy and b) chiral selectivity of BNP as a function of p-(L-SULV) concentration. The arrows indicate the concentration at which each of the techniques was able to distinguish the enantiomers of BNP. Peaks shown in boxes correspond with the associated MEKC separation. Titration plots for BOH and Dns-Phe are shown in Appendix IV and summarized in Table 4.3. The widest range of α-values was observed for BOH, indicating that BOH is 127 more sensitive to PS concentration. In addition, the r curve reveals that R-BOH is more sensitive to PS concentration than S-BOH, due to the large fluctuations with small changes in PS concentration. The concentration at which the first hint of resolution of the peaks is observed according to the selectivity is seen at 0.2 mM and baseline resolution is seen at 1.25 mM. Similarly, the t-test could distinguish the enantiomers at a concentration of 1.25 with an 85% CL. In the case of Dns-Phe, much higher concentrations of PS were required for either method to distinguish the enantiomers: α = 10 mM and r = 20 mM. Table 4.3 Minimum p-(L-SULV) concentration required to achieve differences in anisotropy and selectivity for three model analytes. Chiral analyte Anisotropya Selectivitya t-test (CL) α > 1 BNP 1.0 mM (99.0%) 0.5 mM BOH 1.25 mM (85%) 0.2 mM Dns-Phe 20 mM (99.75%) 10 mM Number in parenthesis indicates the confidence level (CL) aConditions: 50 mM phosphate (dibasic) / 25 mM borate buffer (pH 9.0), 15ºC In order to confirm the assumption that the average time an enantiomer interacts with the polymer is related to the strength of binding, time-resolved fluorescence anisotropy may be useful for resolving the absolute anisotropy into the anisotropy of the free and the bound species, respectively. 4.3.3.2 α versus β Results This portion of my study was designed to investigate a previous hypothesis that the slope m correlates with the chiral separation mechanism of the analyte when plotting α vs. β. In this study, the stereochemistry of the chiral centers of the PS was systemically changed. Ordinarily, when developing chiral separations, one optimizes separation conditions for each analyte. In this case, however, separations were not performed under optimal conditions. 128 Instead they were performed using a single experimental condition so as to isolate the effect of using different surfactants on the changes observed in the slope, m. I suspect that the analytes chosen will be separated by different means due to differences in physicochemical properties. For example, BNP, BNA, and BOH are aromatic molecules with planar asymmetry while the Dns-AAs and TB possess a center of asymmetry at a carbon and nitrogen, respectively. The Dns-AA can be further classified into aliphatic (i.e, Dns-Leu and Dns-Nor) and aromatic (i.e., Dns-Phe and Dns-Trp). According to the α vs. β plot shown in Figure 4.8, two clear trends in data are observed that I associate with differences in separation mechanism as a result of their grouping. The first trend involves, BNP and BOH. According to MEKC and NMR studies, it was clear that BNP and BOH interact preferentially with leucine. Based on the above studies and the fact that both BOH and BNP fall along a single slope (R2 of 0.9889), I am confident in suggesting that these analytes in fact interact primarily with the leucine chiral center and probably with similar or identical chiral separation mechanisms. These analytes are alike in that they are all relatively large, aromatic, binaphthyl compounds possessing an asymmetric plane. In contrast, due to their positions on the second slope, the Dns-AAs appear to follow a different mechanism. MEKC unambiguously showed that Dns-Phe and Dns-Nor interact more with valine but was inconclusive for the other two Dns-AAs. Therefore, in this study, since all Dns-AAs fall along the same slope, it appears that they all follow the same separation mechanism and interact more with valine. Dns-AAs differ from the previous group, in that they possess asymmetric carbons rather than asymmetric planes, which may explain their positioning on a different slope. While all Dns-AA results proved 129 to be inconclusive, according to NMR, this FA technique is in agreement with MEKC that the Dns-AAs interact primarily with valine. The analytes TB and BNA are located in an intermediate position between the two slopes and do not appear to follow a trend of their own. While MEKC results for TB were inconclusive, NOESY NMR showed that TB interacted preferentially with valine. In addition, MEKC results showed that BNA interacts preferentially with leucine, while NMR studies were not viable due to solubility complications. It is unclear why these α vs. β results fail to substantiate either the MEKC or NMR conclusions as was observed in the case y = 0.0092x + 0.9966 1.14 R2 = 0.9889 1.12 BNP 1.10 1.08 MEKC BNA 1.06 om TB r BOH 1.04 1.02 y = 0.0009x + 1.0004 value f 2 - NOR TRP R = 0.9846 α 1.00 LEU PHE 0.98 0 5 10 15 20 25 β -value [degrees] Figure 4.8 An α vs. β plot obtained using p-(L-SULV) as the chiral selector. The standard deviation of all β-values presented was less than 1.8. of BOH and BNP. However, a likely explanation is that those analytes located in intermediate positions are interacting with both chiral centers. Consequently, their position within the plots reflects a weighted average of each of the two slopes. In the case of BNA, the inconclusive result can also be attributed to its limited solubility for extended periods of time. In addition, the misshapen peaks observed in the MEKC study for TB 130 (electropherograms in Appendix III) seem to indicate that an impurity is present which prevents the correct assignment of the primary site of interaction. It is suspected that inconclusive results observed for TB in this study are in part due to such impurity. The plot shown in Figure 4.9 illustrates a similar plot to the one shown in Figure 4.8 with the exception that p-(D-SULV) was used as the chiral selector. As was observed previously, BOH and BNP fall under one slope (i.e., R2 = 0.9737) and the Dns-AAs fall under another slope (i.e., R2 = 0.5013). Following the same logic as was used to interpret Figure 4.8, it is reasonable to assume that BOH and BNP interact preferentially with leucine and the Dns-AAs interact with valine. The only significant difference observed by using the opposite configuration PS was seen with Dns-Leu, Dns-Phe, and Dns-Trp. While these fluorophores are still believed to interact preferentially with valine, their β-value has changed arbitrarily. y = 0.0095x + 0.992 1.14 BNP R2 = 0.9737 1.12 1.10 MEKC 1.08 TB om r 1.06 BNA y = 0.0005x + 1.0054 1.04 BOH R2 = 0.5013 1.02 TRP LEU NOR -value f PHE α 1.00 0.98 0 5 10 15 20 25 β -value [degrees] Figure 4.9 An α vs. β plot obtained using p-(D-SULV) as the chiral selector. The standard deviation of all β-values presented was less than 1.8. 131 The plots that used mixed diastereomer PSs as the chiral selector gave varied results (Appendix V). In the case of p-(D,L)-SULV, the plot shows a similar result as was observed with p-(L-SULV) and p-(D-SULV). That is, BOH and BNP interact with leucine, the Dns- AAs interact with valine, and the analytes TB and BNA again could not be classified. In the case of p-(L,D)-SULV, however, BNA is classified with BOH and BNP which implies that this fluorophore interacts more with the leucine chiral center. This result is not surprising considering that BNA an aromatic binaphthyl derivative with planar asymmetry. However as discussed above, BNA is not soluble for extended periods of time therefore, results may not reflect an accurate assessment of its site of primary interaction. In summary, three of the four diasteromers of p-SULV (i.e., p-(L-SULV), p-(D- SULV), and p-(D,L)-SULV) are in accord that two different chiral separation mechanisms are responsible for the enantioseparation of this set of analytes. In addition, these results corroborate the results obtained using both MEKC and NOESY NMR. As a result, I am optimistic about the future of the α vs. β technique described as it seems to offer a viable approach to studying chiral separation mechanisms. 4.4 Conclusions Using complementary analytical methodologies proved useful in the study of chiral separations employing PSs as chiral selectors. Overall, the results of MEKC, FA, and NMR were in good agreement. Using MEKC, the primary site of interaction was determined for the eight analytes studied. NOESY NMR was able to confirm that the primary site of interaction for BNP, BOH, and TB analytes was with the leucine chiral center of p-(L- SULV). However, a limitation of NOESY is the spectral overlap of the analyte proton resonances with those of the PS peaks which precluded the measurement of NOE difference 132 binding maps for the four Dns-AAs. Therefore, in future studies the analytes must be judiciously selected in order to minimize this spectral overlap. These results suggest that MEKC may be a suitable alternative to NMR for determining the primary sites of interaction. For BNP, FA agreed with MEKC in that the S enantiomer interacted more strongly with p-(L-SULV) than its antipode, S-BNP. The use of FA alone to understand chiral interactions such as the primary sites of interaction, however, requires further investigation before implementation can occur. The α vs. β relationship was confirmed using p-(L-SULV) as the chiral selector thereby supporting the empirical relationship suggested earlier by our group. The different slopes, m, were confirmed to be related to different chiral separation mechanisms owing to preferential interaction for the two chiral centers located on the PS. The results indicate that the Dns-AAs operate under one mechanism, where their primary site of interaction is with the valine chiral center, and BNP, BOH and TB operate under a different mechanism, i.e., led by the interaction with the leucine chiral center. BNA appears to interact with both chiral centers. A determination of the extent of interaction with each chiral center will require additional studies. 4.5 References (1) Ahuja, S. Chiral Separations- Applications and Technology: Washington D.C., 1997. (2) U. S. Food and Drug Administration Chirality 1992, 4, 338-340. (3) Armstrong, D. W.; Demond, W. J. Chromatogr. Sci. 1984, 22, 411-415. (4) Liu, Y.; Gu, J.; Fu, R. J. High Resolut. Chromatogr. 1997, 20, 159-164. (5) Liu, Y.; Zhang, Y.-M.; Sun, S.-X.; Li, Y.-M.; Chen, R.-T. J. Chem. Soc., Perkin Trans. 2 1997, 8, 1609-1613. (6) Armstrong, D. W.; Ward, T. J.; Armstrong, R. D.; Beesley, T. E. Science 1986, 232, 1132-1135. 133 (7) Fornstedt, T.; Goetmar, G.; Andersson, M.; Guiochon, G. J. Am. Chem. Soc. 1999, 121, 1164-1174. (8) Lammerhofer, M.; Peters, E. C.; Yu, C.; Svec, F.; Frechet, J. M. Anal. Chem. 2000, 72, 4614-4622. (9) Shamsi, S. A.; Valle, B. C.; Billiot, F. H.; Warner, I. M. Anal. Chem. 2003, 75, 379- 387. (10) BingCheng, L.; Yibing, J.; Yuging, C.; Epperlein, U.; Koppenhoefer, B. Chromatographia 1996, 42, 106-110. (11) Koppenhoefer, B.; Jakob, A.; Zhu, X.; Lin, B. J. High Resolut. Chromatogr. 2000, 23, 413-429. (12) Tang, Y.; Zhou, Y.; Armstrong, D. W. J. Chromatogr., A 1994, 666, 147-159. (13) Armstrong, D. W.; Rundlett, K. L.; Nair, U. B.; Reid III, G. L. Curr. Sep. 1996, 15, 57-61. (14) Rekharsky, M.; Inoue, Y. J. Am. Chem. Soc. 2000, 122, 4418-4435. (15) Rekharsky, M. V.; Inoue, Y. J. Am. Chem. Soc. 2002, 124, 813-826. (16) Kurtan, T.; Nesnas, N.; Koehn, F. E.; Li, Y.-Q.; Nakanishi, K.; Berova, N. J. Am. Chem. Soc. 2001, 123, 5974-5982. (17) Kurtan, T.; Nesnas, N.; Li, Y.-Q.; Huang, X.; Nakanishi, K.; Berova, N. J. Am. Chem. Soc. 2001, 123, 5962-5973. (18) Clark, J. L.; Booth, B. R.; Stezowski, J. J. J. Am. Chem. Soc. 2001, 123, 9889-9895. (19) Clark, J. L.; Stezowski, J. J. J. Am. Chem. Soc. 2001, 123, 9880-9888. (20) Lipkowitz, K. B.; Raghothama, S.; Yang, J. A. J. Am. Chem. Soc. 1992, 114, 1554- 1562. (21) Maletic, M.; Wennemers, H.; McDonald, D. Q.; Breslow, R.; Still, W. C. Angew. Chem., Int. Ed. Engl. 1996, 35, 1490-1492. (22) Pirkle, W. H.; Murray, P. G.; Rausch, D. J.; McKenna, S. T. J. Org. Chem. 1996, 61, 4769-4774. 134 (23) Yashima, E.; Yamamoto, C.; Okamoto, Y. J. Am. Chem. Soc. 1996, 118, 4036-4048. (24) Kano, K.; Kato, Y.; Kodera, M. J. Am. Chem. Soc., Perkin Trans. 2 1996, 6, 1211- 1217. (25) Xu, Y.; McCarroll, M. E. J. Phys. Chem. A 2004, 108, 6929-6932. (26) Xu, Y.; McCarroll, M. E. J. Phys. Chem. B 2005, 109, 8144-8152. (27) Lustenberger, P.; Martinborough, E.; Denti, T. M.; Diederich, F. J. Chem. Soc., Perk. Trans. 2 1998, 12, 747-754. (28) Magiera, D.; Moeller, S.; Zbigniew, D.; Zbigniew, P.; Pietrusiewicz, K. M.; Duddeck, H. Chirality 2003, 15, 391-399. (29) Anczewski, W.; Dodziuk, H.; Ejchart, A. Chirality 2003, 15, 654-659. (30) Ghosh, S. K. J. Peptide Res. 1999, 53, 261-274. (31) Morris, K. F.; Becker, B. A.; Valle, B. C.; Warner, I. M.; Larive, C. K. In Preparation. 2005. (32) Meyer, B.; Weimar, T.; Peters, T. Eur. J. Biochem. 1997, 246, 705-709. (33) McCarroll, M. E.; Billiot, F. H.; Warner, I. M. J. Am. Chem. Soc. 2001, 123, 3173- 3174. (34) Tsai, C.-F.; Li, C.-F.; Chang, H.-M. J. Agric. Food Chem. 1998, 46, 979-985. (35) Valeur, B. Molecular Fluorescence-Principles and Applications; Wiley-VCH: Weinheim, Germany, 2002. (36) Billiot, E.; Thibodeaux, S. J.; Shamsi, S. A.; Warner, I. M. Anal. Chem. 1999, 71, 4044-4049. (37) Valle, B. C.; Billiot, F. H.; Shamsi, S. A.; Zhu, X.; Powe, A.; Warner, I. M. Electrophoresis 2004, 25, 743-752. (38) Wang, J.; Warner, I. M. J. Chromatogr., A 1995, 711, 297-304. (39) Al Rabaa, A. R.; Tfibel, F.; Merola, F.; Pernot, P.; Fontaine-Aupart, M.-P. J. Chem. Soc., Perkin Trans. 2 1999. 135 (40) Billiot, F. H.; McCarroll, M. E.; Billiot, E. J.; Warner, I. M. Electrophoresis 2004, 25, 753-757. (41) Skoog, D. A.; Holler, F. J.; Nieman, T. A. Principles of Instrumental Analysis, 5th ed.; Harcourt Brace & Company: Orlando, 1998. 136 CHAPTER 5. CONCLUSIONS AND FUTURE DIRECTIONS This dissertation set out to explore the enantiomeric recognition properties of amino acid based polymeric surfactants using a combination of analytical methodologies. I have successfully explored CD-MEKC, MEKC, NOESY NMR, and fluorescence anisotropy based techniques were successfully used to probe chiral interactions involving single amino acid and dipeptide polymeric surfactants. The introduction chapter discussed the fundamental theory behind many of the topics discussed throughout this manuscript. Topics discussed included: chirality, chiral selectors, different modes of CE, NMR, and fluorescence anisotropy. In Chapter 2, a CD-MEKC technique was used to study the effect of PS optical configuration on the chiral separation of three binaphthyl derivatives. The enantiomers of (±)1,1′-binaphthyl-2,2′-diamine (BNA) and (±)1,1′-binaphthol (BOH) reversed in migration order when a combination of two chiral selectors were used, presumable due to a competition between the two chiral selectors for the same enantiomer. Selectivity and resolution values were calculated and discussed for all possible combinations of PS and CD chiral selectors investigated. In Chapter 3, the dipeptide polymeric surfactant, p-(L-SULV), was used as the chiral selector in a comprehensive study where 75 chiral drug compounds were screened. Overall, the PS successfully resolved 73% of the analytes investigated, including 24 neutral analytes, 26 cationic analytes, and 5 anionic analytes. This study focused on defining reliable rules that affect the chiral recognition of various enantiomers using p-(L-SULV) as a chiral pseudostationary phase. 137 In order to better understand the chiral separation mechanism of some of the separations described in the previous chapter, Chapter 4 thoroughly probed chiral interactions of a select group of chiral analytes using three different analytical methodologies (i.e., MEKC, NMR, and fluorescence anisotropy). MEKC studies revealed that BNP, BOH, and BNA enantiomers interact primarily with the leucine chiral center, whereas the dansyl amino acids, Dns-Phe and Dns-Nor, interact primarily with the valine chiral center of the dipeptide polymeric surfactant. In the case of TB, MEKC results were inconclusive. NOESY NMR results were in good agreement with MEKC results and further confirmed that TB, Dns-Trp, and Dns-Leu interact primarily with valine, however were inconclusive for the Dns-AAs. A previously reported correlation by our group regarding the relationship between the chromatographic parameter (α) and the spectroscopic parameter (β), was validated using a select group of fluorophores. The α vs. β correlations plots revealed that two chiral separation mechanisms were involved, where BNP and BOH follow one chiral separation mechanism, while the Dns-AAs follow another mechanism. This result was in agreement with the former two techniques which suggests that the α vs. β correlation serves as a viable alternative for studying chiral interactions. Additional studies pertinent to the last chapter of this dissertation can be extended in the following manner. I. As was discussed previously, steady-state fluorescence anisotropy can only provide an average anisotropy of all fluorophores present within a selector-selectand system, which includes both free and bound fluorophores. On the other hand, time resolved fluorescence anisotropy is able to resolve the anisotropy contribution from component species, i (Equation. 5.1). In this work, it was assumed that the average time an enantiomer interacts 138 with the polymer is related to the strength of binding, which is then related to the anisotropy at a time t, r(t). Therefore, time-resolved fluorescence anisotropy can be used to confirm this assumption (Equation 5.2). −ti θ i rt()=r0e (5.1) r(t) = f(t)bound r(t)bound + f(t)free r(t)free (5.2) A consideration of the time resolved anisotropy could lead to a more linear relationship observed between α and β. II. Pulsed field gradient (PFG)-NMR can also be used as an alternative method to study chiral interactions between enantiomers and a macromolecule (polymeric surfactant). The translational diffusion coefficient, D, for each resolved peak in an NMR spectrum can be determined using PFG-NMR. In the polymeric surfactant-analyte mixtures, we assume the analyte molecules undergo fast exchange between the free and bound states on the NMR timescale. Therefore, according to Equation 5.3, the translational diffusion coefficient of the analyte in solution with the polymeric surfactant, Dobs, is the weighted average of its diffusion coefficient in free solution, Dfree, and its diffusion coefficient in the bound state, Dbound, where fb is the mole fraction of bound analyte. Assuming complexation of the relatively small analyte with the large polymeric surfactant does not change the diffusion of the uncomplexed polymeric surfactant, Dbound can be determined from the diffusion of the polymeric surfactant (Equation 5.4). The diffusion coefficient in free solution can be measured by dissolving the analyte in buffer containing no polymeric surfactant. Once the three diffusion coefficients are determined, fb is calculated and used to arrive at the analyte- polymeric surfactant association constant, K, using Equation 5.5. In this expression, [polymer] is the equivalent monomer concentration of the polymeric surfactant. Finally, 139 free energies of association for each enantiomer–polymeric surfactant complex are computed using (∆G) = -RT ln (K). Translational diffusion coefficients, fb values, association constants and free energies of association for the R and S enantiomers of five analytes (BNP, BOH, TB, warfarin, and propanolol) in the presence of poly (sodium N- undecanoyl-L-leucylvalinate), poly-L-SULV, were investigated. Representative data for a select group of analytes under optimum conditions are tabulated in Table 5.1. Data from Table 5.1 are consistent with the hypothesis that the analytes diffuse slower when they are in solution with the polymeric surfactant, thereby suggesting an association between the enantiomer and polymeric surfactant. The larger the reduction in the bound diffusion coefficient implies the analyte more strongly interacts with the polymeric surfactant. This supposition is confirmed when one calculates association constants. Dobs = fb ⋅ Dbound + (1− fb ) ⋅ D free (5.3) Dbound = Dpolymer (5.4) f K = b (5.5) (1− f b ) ⋅[polymer] III. Still another area of research that can be explored is that of achiral separations. One of the limitations in the fluorescence anisotropy and NMR studies is finding commercially available chiral compounds in enantiomerically pure form. In addition, once the enantiomers are acquired, they still must be resolved using a given chiral selector, in this case, p-(L-SULV). Therefore, exploring achiral separations can greatly increase the pool of potential analytes that can be investigated. Only by expanding the pool of analytes 140 investigated can the α vs, β correlation be truly assessed for its viability as an alternative method for studying chiral interactions. Table 5.1 List of spectrophysical properties observed for various chiral analytes in the presence of p-(L-SULV) determined by PFG-NMR. Observed association constants (K) and free energy (∆G) of association values obtained from PFG- NMR experiments of various analytes. In this case, the polymeric surfactant employed was poly-L-SULV. Conditions shown in Appendix VI. Analyte Dobs Dpolymer fb K ∆G ∆(∆G) (cm2s-1) (cm2s-1) (M-1) (kJ/mol) (kJ/mol)* (R)-Warf 3.40×10-6 1.04×10-6 0.434 ± 0.04 15.4 ± 2.3 -6.77 -1.11 2 (S)-Warf 2.90× 10-6 0.969× 10-6 0.546 ± 0.02 24.1 ± 1.6 -7.88 (+)-TB 1.51×10-6 1.16×10-6 0.931 ± 0.01 538 ± 28 -15.58 -0.48 2 (−)-TB 1.58×10-6 1.13×10-6 0.917 ± 0.01 444 ± 9 -15.10 2 (R)-BNP 1.77× 10-6 1.11× 10-6 0.837 ± 0.01 103 ± 5 -11.48 -0.48 9 (S)-BNP 1.64×10-6 1.08×10-6 0.862 ± 0.01 125 ± 3 -11.96 6 (R)-Prop 1.12×10-6 1.03×10-6 0.980 ± 0.00 993 ± 52 -17.10 -1.96 8 (S)-Prop 1.06× 10-6 1.02× 10-6 0.991 ± 0.00 2191 ± 26 -19.06 7 (R)-BOH 1.12×10-6 1.00×10-6 0.955 ± 0.02 425 ± 11 -14.99 -0.19 0 (S)-BOH 1.27×10-6 1.06×10-6 0.952 ± 0.02 393 ± 11 -14.80 0 *For warfarin, propranolol, and BNP, ∆(∆G) was calculated by subtracting ∆G(R) from ∆G(S). For Troger’s base ∆(∆G) was ∆G(+) minus ∆G(−) and with BOH ∆(∆G) was ∆G(R) minus ∆G(S). The above calculations ensured that all ∆(∆G) values had the same sign. 141 APPENDIX I PLOT OF THE CORRELATION BETWEEN THE α VALUES DERIVED FROM MEKC EXPERIMENTS AND β-VALUES DERIVED FROM FLUORESCENCE Correlation Plot Using p-SULL 1.040 TB 1.035 1.030 BNP ) (α 1.025 y it iv t 1.020 BNA Selec 1.015 1.010 BOH y = 0.0013x + 1.004 R2 = 0.994 1.005 0 5 10 15 20 25 30 β-v alue Appendix I. Plot of the correlation between the α values derived from MEKC experiments and β-values derived from fluorescence anisotropy when p- (L-SULL) is used as the chiral selector. Conditions: 30 mM phosphate buffer (pH 7.0); 1.3 (w/w) p-(L-SULL); 20°C; +30 kV. 142 APPENDIX II DIGITAL PHOTOGRAPHS OF PREP TLC Appendix II.A shows two bands visualized under UV light for each enantiomer of derivatized Dns-AA: a more polar band with a longer migration time (Dns-AA, upper band in red box) and another less polar band with a shorter migration time (dansyl sulphonate in blue box). Identification of these bands was achieved by collecting a small sample from each of the four bands. Subsections from the two upper bands were collected and combined. Similarly, subsections from the two lower bands were collected and combined. MEKC experiments were performed on each of the two mixtures using the established conditions at pH 9 and 24 mM PS. The electropherogram showing a chiral separation (i.e., two distinct analyte peaks) was determined to be the desired Dns- AA, since only the derivatized amino acid is chiral. It should be noted that the electropherogram corresponding to the less polar band (i.e., dansyl sulphonate) showed only one peak. amino acid derivative dansyl sulphonate Dns-D-Leu Dns-L-Leu Appendix II.A Digital photograph of prep–TLC plates showing the separation of Dns-leucine (Dns-Leu) (upper red boxes) from the by-product dansyl sulphonate (lower blue boxes). Conditions: 72.9/15/12.1 (v/v/v) ethyl acetate, methanol, and water with UV detection at 254 nm. 143 amino acid derivative dansyl sulphonate Dns-D-Trp Dns-L-Trp Appendix II.B Digital photograph of prep–TLC plates showing the separation of Dns-tryptophan (Dns-Trp) (upper red boxes) from the by-product dansyl sulphonate (lower blue boxes). Conditions: 72.9/15/12.1 (v/v/v) ethyl acetate, methanol, and water with UV detection at 254 nm. 144 amino acid derivative dansyl sulphonate Dns-D-Nor Dns-L-Nor Appendix II.C Digital photograph of prep–TLC plates showing the separation of Dns-norleucine (Dns-Nor) (upper red boxes) from the by-product dansyl sulphonate (lower blue boxes). Conditions: 72.9/15/12.1 (v/v/v) ethyl acetate, methanol, and water with UV detection at 254 nm. 145 APPENDIX III MEKC ENANTIOSEPARATION OF THE ENANTIOMERS OF SEVEN ANALYTES 3 12 (L,D)-LV 2 p-(L-SULV)(L,L)-LV p-(L,D)-SULV 1 0 10 -1 -2 -3 (0.2404) (0.2477) (0.2500)(0.2566) U 8 -4 mA -5 -6 -7 6 -8 -9 -10 4 20 25 30 35 20 25 30 35 4 4 (D,D)-LV 2 p-(D-SULV) 2 p-(D,(DL,L)-SULV)-LV 0 (0.2481) (0.2450)(0 0.2475) (0.2425) -2 -2 mAU -4 -4 -6 -8 -6 -10 -8 20 25 30 35 20 25 30 35 Time [min] Time [min] Appendix III.A MEKC enantioseparation of the enantiomers of BOH in the presence of the four diastereomers of poly (SULV). Conditions: R-enantiomer is added in excess; 24 mM p-(SULV) in a 50 mM sodium phosphate (dibasic) / 25 mM sodium borate buffer (pH 9.0); 30 kV applied voltage, 15°C, 254 nm. Number in parenthesis denotes the observed steady-state fluorescence anisotropy. 146 5 L (L,L)-LV 28 (LL,D)-LDV p-( -SULV) 26 p-( , )-SULV 24 (0.1286) (0.1435) 22 20 0 18 16 (0.1312)(0.1436) U 14 mA 12 -5 10 8 6 4 2 -10 0 30 32 34 36 38 40 20 25 30 35 -2 (D,D)-LV p-(D-SULV) p-(D(D,,LL)-L)-V SULV 5 (0.1456) (0.1350) (0.1325) (0.1236) -4 0 U mA -5 -6 -10 30 32 34 36 38 40 20 25 30 35 Time [min] Time [min] Appendix III.B MEKC enantioseparation of the enantiomers of BNA in the presence of the four diastereomers of poly (SULV). Conditions: Same as in Appendix III.A. Number in parenthesis denotes the observed steady-state fluorescence anisotropy. 147 3 1.0 p-(L-SULV) (L,LD)-LDV (L,L)-LV p-( , )-SULV 0.5 2 (0.2235) (0.2132) (0.1845)(0.2456) U 0.0 mA 1 -0.5 0 -1.0 20 25 30 16 18 20 0 0.0 p-(D-SULV)(D,D)-LV -0.5 p-(D(D,,LL)-L)-SULVV (0.2483) (0.2343)(-1.0 0.2042) -2 (0.2152) -1.5 U mA -2.0 -4 -2.5 -3.0 -6 -3.5 15 20 25 30 16 18 20 Time [min] Time [min] Appendix III.C MEKC enantioseparation of the enantiomers of TB in the presence of the four diastereomers of poly (SULV). Conditions: Same as in Appendix III.A. Number in parenthesis denotes the observed steady-state fluorescence anisotropy. 148 -2 -2 (L,D)-LV p-(L-SULV)(L,L)-LV p-(L,D)-SULV -4 -4 -6 (0.0598)(0.0640) (0.0707) (0.0728) U -6 mA -8 -8 -10 -10 -12 9 1011129 101112 -2 -2 p-(D-SULV)(D,D)-LV p-((DD,L,)-LLV)-SULV -4 -4 (0.0709) (0.0762) (0.0563) (0.0547) -6 U -6 mA -8 -8 -10 -10 -12 9 1011129 101112 Time [min] Time [min] Appendix III.D MEKC enantioseparation of the enantiomers of Dns-Phe in the presence of the four diastereomers of poly (SULV). Conditions: Same as in Appendix III.A.. Number in parenthesis denotes the observed steady-state fluorescence anisotropy. 149 0 2 (L,D)-LV p-(L-SULV)(L,L)-LV 0 p-(L,D)-SULV -2 -2 (0.0954) (0.0798) (0.0882) -4 -4 (0.0929) U -6 -6 mA -8 -8 -10 -10 -12 -12 -14 91011129101112 0 2 -2 p-(D-SULV)(D,D)-LV 0 p-((D,LD,)L-LV)-SULV (0.0990) (0.0960) -4 (0.0795) (0.0772) -2 U -6 -4 mA -8 -6 -10 -8 -12 -10 91011129101112 Time [min] Time [min] Appendix III.E MEKC enantioseparation of the enantiomers of Dns-Trp in the presence of the four diastereomers of poly (SULV). Conditions: Same as in Appendix III.A. Number in parenthesis denotes the observed steady-state fluorescence anisotropy. 150 -2 -8 (L,D)-LV -3 p-(L-SULV)(L,L)-LV -9 p-(L,D)-SULV -4 -10 (0.0708) (0.0664) (0.0695) -5 -11 (0.0787) U -6 -12 mA -7 -13 -8 -14 -9 -15 -10 -16 910119101112 -2 4 3 -3 p-(D-SULV)(D,D)-LV 2 p-(D,L)-(LDV ,L)-SULV -4 1 (0.0821) (0.0672)(0.0585) 0 -5 -1 (0.0822) -2 U -6 -3 mA -7 -4 -5 -8 -6 -7 -9 -8 -10 -9 910119101112 Time [min] Time [min] Appendix III.F MEKC enantioseparation of the enantiomers of Dns-Leu in the presence of the four diastereomers of poly (SULV). Conditions: Same as in Appendix III.A. Number in parenthesis denotes the observed steady-state fluorescence anisotropy. 151 -4 2 (L,D)-LV p-(L-SULV)(L,L)-LV 0 p-(L,D)-SULV -6 -2 (0.0875) (0.0855) (0.0753) -4 -6 (0.0720) U -8 -8 mA -10 -10 -12 -14 -12 -16 10 11 12 10 11 12 -4 6 4 p-(D-SULV)(D,D)-LV p-(D(D,,L)-LV)-SULV -6 2 (0.0570) (0.0731) (0.1084) (0.0850) 0 U -8 -2 mA -4 -10 -6 -8 -12 -10 10 11 12 10 11 12 Time [min] Time [min] Appendix III.G MEKC enantioseparation of the enantiomers of Dns-Nor in the presence of the four diastereomers of poly (SULV). Conditions: Same as in Appendix III.A. Number in parenthesis denotes the observed steady-state fluorescence anisotropy. 152 APPENDIX IV TITRATION CURVES SHOWING THE A) ENANTIOMER ANISOTROPY AND B) CHIRAL SELECTIVITY OF BOH AND DNS-PHE a) anisotropy 1.25 mM b) selectivity 0.2 mM 0.265 1.16 0.260 1.14 0.255 1.12 0.250 1.10 , r ] 0.245 α py 1.08 y [ t ro i t 0.240 v o 1.06 cti s i e l 0.235 e an s 1.04 0.230 1.02 0.225 R-BOH 1.00 T-test S-BOH alpha 0.220 0.98 012345 012345 poly(L,L)SULV concentration [mM] poly(L,L)SULV concentration [mM] Appendix IV.A Titration curves showing the a) enantiomer anisotropy and b) chiral selectivity of BOH as a function of p-(L-SULV) concentration. The arrows indicate the concentration at which each of the techniques was able to distinguish the enantiomers of BNP. Peaks shown in boxes correspond with the associated MEKC separation. 153 a) anisotropy 20 mM b) selectivity 10 mM 0.08 1.05 0.07 1.04 0.06 0.05 1.03 ] α , r y [ y p 0.04 o r t ivit 1.02 o 0.03 is lect n e a 0.02 s 1.01 0.01 T-test D-Phe 0.00 L-Phe 1.00 alpha 0 1020304050 0 102030405060 poly (L,L)-SULV concentration [mM] poly (L,L)SULV concentration [mM] Appendix IV.B Titration curves showing the a) enantiomer anisotropy and b) chiral selectivity of Dns-Phe as a function of p-(L-SULV) concentration. The arrows indicate the concentration at which each of the techniques was able to distinguish the enantiomers of BNP. Peaks shown in boxes correspond with the associated MEKC separation. 154 APPENDIX V α VS. β PLOTS OBTAINED USING P-(L,D)-SULV AND P-(D,L)-SULV AS THE CHIRAL SELECTORS 1.07 y = 0.0038x + 1.0021 2 1.06 BNA R = 0.9568 KC 1.05 E BNP 1.04 BOH M TB 1.03 om r 1.02 y = -0.0001x + 1.0036 1.01 PHE R2 = 0.0755 TRP LEU -value f 1.00 α NOR 0.99 0.98 0 5 10 15 20 25 β -value [degrees] Appendix V.A An α vs. β plot obtained using p-(L,D)-SULV as the chiral selector. The standard deviation of all β-values presented was less than 1.8. 155 1.07 y = 0.0046x + 0.9999 1.06 R2 = 0.9998 BNP 1.05 KC BNA 1.04 BOH ME 1.03 TB y = 0.0003x + 1.0018 1.02 R2 = 0.3766 from e 1.01 TRP PHE 1.00 NOR -valu LEU α 0.99 0.98 0 5 10 15 20 25 β -value [degrees] Appendix V.B An α vs. β plot obtained using p-(D,L)-SULV as the chiral selector. The standard deviation of all β-values presented was less than 1.8. 156 APPENDIX VI EXPERIMENTAL CONDITIONS USED IN THE PFG NMR STUDIES BNP and BOH: 100 mM Tris / 10 mM borate buffer, pH 10, 25°C, 7 mM p-(L- SULV) Trogers Base: 30 mM phosphate buffer, pH 7.0, 25°C, 5 mM p-(L-SULV) Warfarin: 50 mM phosphate buffer, pH 6.5, 25°C, 25 mM p-(L-SULV) Propranolol: 300 mM CAPS / 50 mM borate buffer, pH 8.5, 25°C, 12.5 mM p- (L-SULV) 157 APPENDIX VII RAW 2-D NOE PLOTS FOR CHAPTER 4 Appendix VII.A A NOE 2-D plot showing S-BNP on the horizontal axis and p-(L- SULV) on the vertical axis. Conditions: 50 mM phosphate (dibasic) : 25 mM borate buffer in D2O at pH 9.0. 158 Appendix VII.B A NOE 2-D plot showing (+)-TB on the horizontal axis and p-(L- SULV) on the vertical axis. Conditions: 50 mM phosphate (dibasic) : 25 mM borate buffer in D2O at pH 9.0. 159 Appendix VII.C A NOE 2-D plot showing L-Dns-Leu on the horizontal axis and p- (L-SULV) on the vertical axis. Conditions: 50 mM phosphate (dibasic) : 25 mM borate buffer in D2O at pH 9.0. 160 Appendix VII.D A NOE 2-D plot showing L-Dns-Phe on the horizontal axis and p- (L-SULV) on the vertical axis. Conditions: 50 mM phosphate (dibasic) : 25 mM borate buffer in D2O at pH 9.0. 161 Appendix VII.E A NOE 2-D plot showing L-Dns-Trp on the horizontal axis and p- (L-SULV) on the vertical axis. Conditions: 50 mM phosphate (dibasic) : 25 mM borate buffer in D2O at pH 9.0. 162 VITA Bertha Cedillo Valle was born in Houston, Texas, on February 9, 1975. She is the youngest of five children born to Octavio and Mercedes Cedillo. As a first generation American, she attended Travis Elementary, Cunningham Middle School, and graduated from North Shore Senior High School in 1993. Undecided in her career path, she enrolled at the local junior college (San Jacinto College-North Campus) near her home and took basic college courses. Soon after enrolling in college, she started a job as a physical therapist aid in a chiropractic office and started night school. She was later inspired by a television sitcom called “The Cosby Show” to go away to school as many of the Cosby children had done. In the fall of 1995, she left home to pursue a degree in chemistry at Texas Tech University in Lubbock, Texas. While at Texas Tech, she worked for Dr. Dennis Shelley (a former graduate student to whom would later become her graduate advisor) at the Leather Research Institute. She graduated with a Bachelor of Science degree in chemistry in the fall of 1998. Encouraged by Dr. Shelley, she started the doctoral program at Louisiana State University in Baton Rouge, Louisiana in the spring of 1999. After careful consideration, she joined the research group of Dr. Isiah M. Warner where she focused on investigating chiral separation mechanisms using micellar electrokinetic chromatography, fluorescence anisotropy, and nuclear magnetic resonance. While under the direction of Professor Isiah M. Warner she received a National Institutes of Health Ruth L. Kirschstein Pre-Doctoral fellowship that supported her throughout the bulk of her studies. Excerpts of her dissertation research which have been published in scientific journals or that are in preparation are listed below: Valle, B.C.; Morris, K.; Fletcher, K.A.; Agbaria, R. A.; Larive, C.K.; Warner, I.M. Spectroscopic Studies of the Diastereomers of p-(L-SULV). In preparation. 163 Morris, K.; Valle, B.C.; Warner, I.M.; Larive, C.K. Comparison of NMR Binding Interaction Mapping Techniques to Examine Interactions of Chiral Analytes with Polymeric Surfactants. In preparation. Valle, B.C.; Billiot, F.H.; Zhu, X.; Powe, A.M.; Shamsi, S. A.; Warner, I.M. A Combination of Cyclodextrins and Polymeric Surfactants for Chiral Separations. Electrophoresis 2004, 25, 743-752. Valle, B.C.; Billiot, F.H.; Zhu, X.; Powe, A.M.; Shamsi, S.A.; Warner, I.M. A Combination of Cyclodextrins and Polymeric Surfactants for Chiral Separations. Electrophoresis 2004, 25, 743-752. Kapnissi, C.P.; Valle, B.C.; Warner, I.M.; Chiral Separations Using Polymeric Surfactants and Polyelectrolyte Multilayers in Open-Tubular Capillary Electrochromatography, Anal. Chem. 2003, 75, 6097-6104. Warner, I.M.; Kapnissi, C.P.; Kamande, M.W.; Valle, B.C.; Analytical Separations with Polyelectrolyte Layers, Molecular micelles, or Zwitterionic Polymers. (US Patent Submitted 2003) Thibodeaux, S.J.; Billiot, E.; Torres, E.; Valle, B.C.; Warner, I.M. Enantiomeric Separations using Polymeric L-Glutamate Surfactant Derivatives: Effect of Increasing Steric Factors. Electrophoresis 2003, 24, 1077-1082. Shamsi, S.; Valle, B.C.; Billiot, F.H.; Warner, I.M. Polysodium N-Undecanoyl-L- Leucylvalinate: A Versatile Chiral Selector for Micellar Electrokinetic Chromatography. Anal. Chem. 2003, 75, 379-387. Zhu, X.; Thiam, S.; Valle, B.C.; Warner, I.M. Use of Colloidal Graphite Coated Emitter for Sheathless Capillary Electrochromatography Nanoelectrospray Ionization Mass Spectrometry. Anal. Chem. 2002, 74, 5405-5409. While earning her degree, she presented her findings at numerous professional meetings and they are listed below. The presenting author is underlined. "Investigating the Enantiomeric Recognition Ability of poly-SULV", Kristin A. Fletcher, Bertha C. Valle, Kevin F. Morris, Cynthia K. Larive, Isiah M. Warner, Talk presented at the 32nd Annual NOBCChE Conference, Orlando, FL, March 2005. "Investigating the Enantiomeric Recognition Ability of poly-SULV using 2D-Nuclear Magnetic Resonance and Capillary Electrophoresis", Bertha C. Valle, Kristin A. Fletcher, Kevin F. Morris, Cynthia K. Larive, and Isiah M. Warner, Poster presented at the 229th 164 National Annual Meeting of the American Chemical Society, Division of Analytical Chemistry, San Diego, CA, March 2005. This poster was also included in the Sci-Mix Poster Session. "Quantitative Analysis of Enantiomeric Compounds Using Fluorescence Anisotropy” Valle, B.C., Agbaria, R.A., and Warner, I.M. Poster presented at Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy (PITTCON), Chicago, IL, March 2004. "Chiral Separations using Polymeric Surfactant Polyelectrolyte Multilayers in Open-Tubular Capillary Electrochromatography – Morphologic Elucidation of Capillaries using Laser Scanning Confocal Microscopy” Kapnissi-Christodoulou, K; Lowry, M; Valle, B.C., Agbaria, R.A., Geng, L; and Warner, I.M. Poster presented at PITTCON, Chicago, IL, March 2004. "Molecular Studies of Chiral Recognition by Use of Fluorescence Methods” Valle, B.C., Agbaria, R.A., and Warner, I.M. Poster presented at the PITTCON, Orlando, FL, March 2003. 11. "Separation of Risperidone and the Enantiomers of 9-Hydroxyrisperidone by Capillary Electrophoresis” Cedillo, B., Zhu, X., Agbaria, R. A., and Warner, I.M Poster presented at the American Chemical Society (ACS) National Meeting. New Orleans, LA, March 2003. "Separation of Risperidone and the Enantiomers of 9-Hydroxyrisperidone by Capillary Electrophoresis” Cedillo, B., Zhu, X., Agbaria, R. A., and Warner, I.M Talk presented at the 30th Annual National Organization for the Professional Advancement of Black Chemists and Chemical Engineers (NOBCChE), Indianapolis, IN, March 2003. "Fluorescence Anisotropy as a Measure of Chiral Recognition in Homogeneous and Heterogeneous Environments” Cedillo, B., Agbaria, R.A., McLaughlin, M.L., and Warner, I.M. Talk presented at PITTCON, New Orleans, LA, March 2002. "Separation of Risperidone and the Enantiomers of 9-Hydroxyrisperidone by Capillary Electrophoresis” Cedillo, B., Zhu, X., Agbaria, R. A., and Warner, I.M Talk presented at 29th Annual NOBCChE, New Orleans, LA, March 2002. "Studies of Synergistic Effects of Cyclodextrins and Polymeric Surfactants In Chiral Recognition Using Various Spectroscopic Techniques." Cedillo, B., Zhu, X., Agbaria, R. A., and Warner, I.M Talk presented at Southwest ACS Meeting. San Antonio, TX, October 2001 "Studies of Synergistic Effects of Cyclodextrins and Polymeric Surfactants In Chiral Recognition Using Various Spectroscopic Techniques." Cedillo, B., Zhu, X., Agbaria, R. A., and Warner, I.M Poster presented at 13th International Symposium on Chirality (ISCD- 13, "Chirality-2001"), Orlando, FL, July 2001. 165 "Effect of CD in Chiral Recognition Using Various Spectroscopic Techniques." Cedillo, B., Zhu, X., Agbaria, R. A., and Warner, I.M Talk presented at PITTCON, New Orleans, LA, March 2001. "Effect of CD in Chiral Recognition Using Various Spectroscopic Techniques." Cedillo, B., Billiot, F., Shamsi, S., Warner, I. Talk presented at 28th Annual National NOBCChE, Baltimore, MD, April 2001, "Characterization of A Novel Cyclodextrin/Surfactant Complex." Cedillo, B., Agbaria, R. A., and Warner, I. M. Talk presented at the ACS 56th Joint 58th ACS Southwest Regional Conference, New Orleans, LA, December 2000. "Combinations of Cyclodextrins and Polymeric Surfactants for Chiral Separations Using Electrokinetic Chromatography." Cedillo, B., Billiot, F., Shamsi, S., and Warner, I. M. Talk presented at 27th Annual NOBCChE, Miami, FL, April 2000. "Combinations of Cyclodextrins and Polymeric Surfactants for Chiral Separations Using Electrokinetic Chromatography." Cedillo, B., Billiot, F., Shamsi, S., and Warner, I. M. Poster presented at PITTCON, New Orleans, LA, March 2000. Her academic awards include: • Alliance for Graduate Education and the Professiorate (AGEP) Dissertation Writing Fellowship made possible through Alliance For Graduate Education in Louisiana (GAELA) and funded by the National Science Foundation (Fall 2005) • The James G. Traynham Award for Most Outstanding Graduate Student. (Fall 2004). • Proctor & Gamble-Research and Technical Careers in Industry (RTCI) Scholar (June 2003) • LSU Departmental Award - Proctor & Gamble Research Scholar (Fall 2003 and 2004) • National Institutes of Health (NIH) Ruth L. Kirschtstein Pre-doctoral Fellowship (1999 - 2004) • National Science Foundation (NSF) Summer Undergraduate Research Fellowship (Summer 1998) • Golden Key National Honor Society (Spring 1996) • Dean’s List (Fall 1995; Spring 1996) 166