Advanced Vibrational Spectroscopic Studies of Biological Molecules

ADVANCED VIBRATIONAL SPECTROSCOPIC STUDIES OF BIOLOGICAL MOLECULES

A thesis submitted to the University of Manchester for the degree of Doctor in Philosophy in the Faculty of Life Sciences

2012

Saeideh Ostovar pour

Table of Contents

List of Contents ...... 2 List of tables ...... 6 List of figures ...... 7 Abstract ...... 11 Declaration ...... 12 Copyright statement...... 13 Abbreviations List ...... 14 Glossary ...... 15 Acknowledgements ...... 17

Chapter 1

1.0 Introduction ...... 19 1.1 Conformational analysis of biological molecules ...... 19 1.2 Infrared and Raman spectroscopies ...... 20 1.2.1 Principles of Raman spectroscopy ...... 23 1.2.2 Resonance Raman spectroscopy ...... 24 1.2.3 Raman spectroscopy of biomolecules ...... 25 1.3 Raman Optical Activity (ROA) ...... 28 1.3.1 ROA spectroscopy of biomolecules ...... 32 1.4 Surface-Enhanced Raman Scattering ...... 34 1.4.1 The Electromagnetic Mechanism...... 35 1.4.2 The Chemical Mechanism (CE) ...... 36 1.4.3 Biological applications of SERS ...... 37 1.5 Surface enhanced Raman optical activity (SEROA) ...... 39 1.6 References ...... 45

Chapter 2

Use of a Hydrogel Polymer for Reproducible Surface Enhanced Raman Optical Activity (SEROA)

2.0 Declaration ...... 50 2

Table of Contents

2.1 Abstract ...... 51 2.2 Introduction ...... 51 2.3 Experimental ...... 53 2.4 Results and Discussion ...... 54 2.5 Conclusion ...... 57 2.6 References ...... 59 2.7 Supplementary Information ...... 63 2.7.1 Colloid Preparation ...... 63 2.7.2 Sample Preparation for Raman and ROA Measurements ...... 63 2.7.3 Atomic Force Microscopy ...... 64 2.7.4 SERS Time Dependence ...... 66

Chapter 3

Induced Chirality to Non-chiral Surfaces of Silver Silica Nanotags

3.0 Declaration ...... 71 3.1 Abstract ...... 72 3.2 Introduction ...... 73 3.3 Experimental ...... 75 3.4 Results and Discussion ...... 77 3.5 Conclusion ...... 82 3.6 References ...... 83 3.7 Supplementary Information ...... 88

Chapter 4

Phosphorylation Detection and Characterization in Ribonucleotides Using Raman and Raman Optical Activity (ROA) Spectroscopies

4.0 Declaration ...... 90 4.1 Abstract ...... 91 4.2 Introduction ...... 91 4.3 Experimental ...... 93 4.4 Results and Discussion ...... 94 4.5 Conclusion ...... 99 3

Table of Contents

4.6 Acknowledgement...... 100 4.7 References ...... 101 4.8 Supplementary Information ...... 107 4.8.1 Colloid Preparation ...... 107 4.8.2 Surface enhanced Raman spectroscopy (SERS) ...... 107 4.8.3 References ...... 109

Chapter 5

Study of Experimental and Computational Raman and Raman Optical Activity (ROA) Spectra of Cyclic and Linear L-Ala-L-Ala in Solution

5.0 Declaration ...... 113 5.1 Abstract ...... 114 5.2 Introduction ...... 115 5.3 Experimental ...... 117 5.4 Computational methods ...... 118 5.5 Results and Discussion ...... 119 5.6 Conclusion ...... 125 5.7 References ...... 129 5.8 Supplementary Information ...... 135

Chapter 6

Combined Experimental and Computational Study of Raman and Raman Optical Activity (ROA) Spectra of Linear and Cyclic L-Ser-L-Ser in Solution

6.0 Declaration ...... 140 6.1 Abstract ...... 141 6.2 Introduction ...... 142 6.3 Experimental ...... 143 6.4 Computational methods ...... 144 6.5 Results and Discussion ...... 146 6.6 Conclusion ...... 150 6.7 References ...... 152 6.8 Supplementary Information ...... 161

Table of Contents

Chapter 7 7.0 Conclusion ...... 166 7.1 Future work ...... 170 7.2 References ...... 172

Chapter 8 Appendix 8.0 Declaration ...... 174

37,600 words

List of Figures

List of Tables

Table 1.1 Advantages and disadvantages of Raman spectroscopy use ...... 26 Table 1.2 Important spectral regions of protein vibrations ...... 27 Table S2.1 Raman and ROA band assignments of L- and D-ribose in aqueous solution ...... 68 Table 4.1 Raman band assignments of adenosine, AMP, ADP, ATP, A(2)MP, A(2,3)MP, A(3)MP and A(3,5)MP ...... 105 Table 4.2 ROA band assignments of adenosine, AMP, ADP, ATP, A(3,5)MP, A(2,3)MP, A(2)MP and A(3)MP ...... 106 Table S4.1 SERS band assignments of adenosine, AMP, ADP, ATP, A(2)MP, A(2,3)MP, A(3)MP and A(3,5)MP ...... 111 Table 5.1 Calculated and experimental wavenumber band assignments for

Raman and ROA of cyclic and linear L-Ala-L-Ala in H2O...... 131 Table 5.2 Calculated and experimental wavenumber band assignments for

Raman and ROA of cyclic and linear L-Ala-L-Ala in D2O...... 133 Table S5.1 Calculated and experimental bond lengths (Å) for cyclic and linear L-Ala-L-Ala ...... 135 Table S5.2 Calculated and experimental bond angles (o) for cyclic and linear L-Ala-L-Ala ...... 136 Table S5.3 Calculated and experimental torsion angles (o) for cyclic and linear L-Ala-L-Ala ...... 137 Table 6.1 Calculated and experimental Raman and ROA bands for cyclic and

linear L-Ser-L-Ser in H2O ...... 157 Table 6.2 Calculated and experimental Raman and ROA bands for cyclic and

linear L-Ser-L-Ser in D2O ...... 159 Table S6.1 Calculated and experimental bond lengths (Å) for cyclic and linear L-Ser-L-Ser ...... 161 Table S6.2 Calculated bond angles (o) for cyclic and linear L-Ser-L-Ser .... 161 Table S6.3 Calculated torsion angles (o) for cyclic and linear L-Ser-L-Ser . 163

List of Figures

Figure 1.1 An energy level diagram showing the transitions involved in Raman scattering ...... 22 Figure 1.2 Schematic diagram of the basic ROA experiment which measures a small difference in the intensity of Raman scattering in right (R) and left (L) circularly polarized light from chiral molecules ...... 28

Figure 2.1 Raman (IR + IL, top) and SCP ROA (IR – IL, bottom) spectra of D- and L-ribose in aqueous. Sample concentrations were 2.66 M at pH 5.46 (D-) and 5.60 (L-ribose), data collection time was 237.02 min, and laser power at the sample 0.625 W for each ...... 60 Figure 2.2 Raman spectrum of polycarbopol in solution (A), SERS spectra before (B) and after addition of L- and D-ribose (0.25 mg ml-1) in

the presence of silver citrate reduced colloid and K2SO4 at 0.020 M concentration, data collection time: 20 min (C), SERS spectra of L- and D-ribose (0.25 mg ml-1) in the presence of polycarbopol polymer, data collection time: 20 min (D), ROA spectrum of polycarbopol polymer in solution, sample concentration 40 mg ml- 1, data collection time: 218 min (E), SEROA spectra of silver citrate reduced colloids in presence of aggregating salt before (F) and after addition of L- and D-ribose, data collection time: 35 min (G), SEROA spectra of L- and D-ribose with addition of polycarbopol, datacollection time of 35 min (H) ...... 61 Figure S2.1 A schematic diagram of SEROA sample preparation with use of the polycarbopol polymer ...... 63 Figure S2.2 Dual views of AFM images of silver citrate reduced colloids only (A), polycarbopol polymer only (B) and a mixture of silver citrate reduced colloids and polycarbopol polymer (C) ...... 65 Figure S2.3 Time dependence SERS study for D-ribose molecule with and without addition of polymer to SERS solution ...... 66 Figure S2.4 Example repeats of SEROA spectra for D- (top) and L-ribose (bottom) with the same conditions for each spectrum, where

concentration of each sample was 0.25 mg/ml and K2SO4

List of Figures

concentration was 0.020 M, with addition of 20 mg/mL of polycarbopol, data collection time of 35 mins ...... 67

Figure S2.5 Raman (IR + IL, A and B) and SCP ROA (IR – IL, E and F) spectra of D- and L-tryptophan in aqueous. Sample concentrations were 50 mg/ml at pH 1.90 (D-) and 1.58 (L-tryptophan), data collection time was 4-8hrs, and laser power at the sample 0.625 W for each. SERS spectra after addition of D- and L-tryptophan (C and D), (0.0002 M) in the presence of silver hydroxyalamine reduced

colloid and MgSO4 at 0.050 M concentration, data collection time: 5 min, SEROA spectra of D- and L-tryptophan (G and H) with addition of polycarbopol, data collection time of 35 min ...... 69 Figure 3.1 SERRS spectra of nanotag (tri-functional benzotriazole dye) without silica coated silver colloids (A), with silica coated silver colloids (B) and SERROA spectra of A (C) and B (D), data collection time of 35 min and laser power at source 0.20 W ...... 85 Figure 3.2 SERRS spectra of D- and L-ribose that attached to silver silica nanotag (A), SERROA of D- and L-ribose replicates 1 (B) and batch 2 (C), data collection time of 35 min and laser power at source 0.20 W ...... 86 Figure 3.3 SERRS spectra of D- and L-tryptophan that attached to silver silica nanotag (A), SERROA spectra of D- and L-tryptophan (B), data collection time of 35 min and laser power at source 0.20 W ...... 87 Figure S3.1 Structure of tri-functional benzotriazole dye ...... 88 Figure 4.1 Raman spectra of adenosine (pH= 12.95), AMP (pH= 6.02), ADP (pH= 5.18), ATP (pH= 4.20), A(2)MP (pH= 3.13), A(2,3)MP (pH=5.54), A(3)MP (pH= 8.10) and A(3,5)MP (pH=6.67) in solution. The concentration for each sample was 100 mg/ml and laser power was 0.6 W at the sample ...... 103 Figure 4.2 ROA spectra of adenosine (pH= 12.95), AMP (pH= 6.02), ADP (pH= 5.18), ATP (pH= 4.20), A(2)MP (pH= 3.13), A(2,3)MP (pH=5.54), A(3)MP (pH= 8.10) and A(3,5)MP (pH=6.67) in solution. The concentration for each sample was 100 mg/ml and laser power at the sample was 0.6 W ...... 104

List of Figures

Figure S4.1 SERS spectra of adenosine ribonucleotides and adenosine in the presence of silver citrate reduced colloid. Analyte concentrations -5 were 1x10 mg/ml, K2SO4 concentration was 0.020 M, data collection time was 50 seconds, and laser power was 0.20 W at the laser ...... 110 Figure 5.1 The chemical structure with atom numbering scheme (left) and calculate minimum energy conformation (right) of linear L-Ala-L- Ala ...... 126 Figure 5.2 The chemical structure with atom numbering scheme (left) and calculate minimum energy conformation (right) of cylic L-Ala-L- Ala ...... 126 Figure 5.3 Experimental and computed Raman (top) and ROA (bottom) spectra of linear (pH= 7.0) and cyclic L-Ala-L-Ala (pH= 7.0) in aqueous solution. The concentration for each sample was 50 mg/ml and laser power was 0.6 W at the sample. The marker bands that are induced upon cyclization are highlighted by shading ...... 127 Figure 5.4 Experimental and computed Raman (top) and ROA (bottom) spectra of linear (pH= 7.0) and cyclic L-Ala-L-Ala (pH= 7.0) in

D2O. The concentration for each sample was 50 mg/ml and laser power was 0.6 W at the sample. The marker bands that are induced upon cyclization are highlighted by shading ...... 128 Figure 6.1 The chemical structure with atom numbering scheme (left) and calculate minimum energy conformation (right) of linear L-Ser-L- Ser ...... 154 Figure 6.2 The chemical structure with atom numbering scheme (left) and calculate minimum energy conformation (right) of cyclic L-Ser-L- Ser ...... 154 Figure 6.3 Experimental and computed Raman (top) and ROA (bottom) spectra of linear (pH= 7.0) and cyclic L-Ser-L-Ser (pH= 7.0) in aqueous solution. The concentration for each sample was 50 mg/ml and laser power was 1.2 W at the laser. The marker bands that are induced upon cyclization are highlighted by shading ...... 155

List of Figures

Figure 6.4 Experimental and computed Raman (top) and ROA (bottom) spectra of linear (pH= 7.0) and cyclic L-Ser-L-Ser (pH= 7.0) in

D2O. The concentration for each sample was 50 mg/ml and the laser power was 0.6 W at the sample. The marker bands that are induced upon cyclization are highlighted by shading ...... 156

Abstract

Abstract Raman optical activity (ROA) is a powerful probe of the structure and behaviour of biomolecules in aqueous solution for a number of important problems in molecular biology. Although ROA is a very sensitive technique for studying biological samples, it is a very weak effect and the conditions of high concentration and long data collection time required limit its application for a wide range of biological samples. These limitations could possibly be overcome using the principle of surface enhanced Raman scattering (SERS). The combination of ROA with SERS in the form of surface enhanced ROA (SEROA) could be a solution for widening the application of ROA. In the last few years, the generation of reliable SEROA spectra of biomolecules has been problematic due to non-homogenous colloidal systems forming and low signal-to-noise ratios which complicated detection of the true SEROA signal from the analyte. L- and D-enantiomers give full or partially mirror image chiroptical spectra, this property of enantiomers can be employed to prove the chiroptical activity of the SEROA technique. In this thesis we employed a hydrophilic polycarbopol polymer as stabilising media which has led to the first report of mirror image SEROA bands for enantiomeric structures. This new technique of incorporating the hydrogel polymer as a means to stabilise the colloidal system has proven to be reliable in obtaining high quality SEROA spectra of D- and L-enantiomers of ribose and tryptophan. In an extension of the hydrogel-stabilised SEROA work, we also demonstrate that single nanoparticle plasmonic substrate such as silver silica nanotags can enhance the weak ROA effect. These dye tagged silica coated silver nanoparticles have enabled a chiral response to be transmitted from a chiral analyte to the plasmon resonance of an achiral metallic nanostructure. The measurement of mirror image SERROA bands for the two enantiomers of each of ribose and tryptophan was confirmed for this system. The generation of SEROA for both systems was achieved and confirmed SEROA as a new sensitive tool for analysis of biomolecular structure. In a related project, Raman and ROA spectra were measured for adenosine and seven of its derivative ribonucleotides. Both of these spectroscopic techniques are shown to be sensitive to the site and degree of phosphorylation, with a considerable number of marker bands being identified for these ribonucleotides. Moreover, the SERS studies of these ribonucleotides were also performed. The obtained SERS spectra were shown similar features that confirm these analytes interact with the surface in a similar manner, hence limiting the structural sensitivity of this method towards phosphate position. Short dipeptides such as diketopiperazine (DKP) have been investigated during the last decades as both natural and synthetic DKPs have a wide variety of biological activities. Raman and ROA spectra of linear and cyclic dialanine and diserine were measured to charecterize their solution structures. Density functional theory (DFT) calculations were carried out by a collaborator to assist in making vibrational band assignments. Considerable differences were observed between the ROA bands for the cyclic and linear forms of both dialanine and diserine that reflect large differences in the vibrational modes of the polypeptide backbone upon cyclicization. In this study, the ROA spectra of cyclic dialanine and diserine have been reported for the first time which demonstrated that ROA spectroscopy when utilised in combination with computational modelling clearly provides a potential tool for characterization of cyclic peptides.

Declaration

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

Copy right Statement

1. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes.

2. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made.

3. The ownership of certain Copyright, patents, designs, trade marks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions.

4. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://www.campus.manchester.ac.uk/medialibrary/ policies/intellectualproperty.pdf), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regulations) and in The University’s policy on presentation of These.

Abbreviations List

AMP Adenosine 5’-monophosphate

ADP Adenosine 5’-diphosphate

A(2)MP Adenosine 2’-monophosphate

A(2,3)MP Adenosine cyclic 2,3’- monophosphate

A(3)MP Adenosine 3’-monophosphate

A(3,5)MP Adenosine 3’,5’-cyclic monophosphate

ATP Adenosine 5’-triphosphate

CID Circularly intensity difference

CP Circular polarization

Cyclic L-Ala-L-Ala Cyclic L-Alanine-L-Alanine

Cyclic L-Ser-L-Ser Cyclic L-Serine-L-Serine

DCP Dual circularly polarised

DNA Deoxyribonucleic acid

EM Electromagnetic

IR Infrared Linear L-Ala-L-Ala Linear form of L-Alanine-L-Alanine

Linear L-Ser-L-Ser Linear form of L-Serine-L-Serine

NMR Nuclear magnetic resonance

SCP Scattered circular polarisation

SEROA Surface-enhanced Raman optical activity

SERROA Surface enhanced resonance Raman optical activity

SERRS Surface enhanced resonance Raman scattering

SERS Surface-enhanced Raman scattering

Glossary

Anti-stokes scattering Raman scattering event where the scattered photon has more energy than incident radiation

Chirality Property of molecules with a handedness to their chemical Structure (i.e. molecules which cannot super impose on its mirror image) and possessing optical activity

Colloids Type of chemical mixture when one substance is dispersed evenly throughout another

D-configuration Dextrorotatory- configuration rotating the plane of vibration of polarized light to the right

Elastic scattering Specific form of scattering where the energy of incident photon is equal to the energy of scattered photon.

Enantiomers Molecules that are optical isomers, or mirror images, of one another. Enantiomers can be distinguished by the direction in which they rotate the plane of polarization of polarized light

Inelastic scattering Raman scattering where the energy of scattered photon is not equal to Incident photon

L-configuration Laevorotatory-Rotating the plane of vibration of polarized light to the left

Nanoparticles Particles whose diameter is between 1 and 100 nm

Nucleoside are glycosylamines consisting of a nucleobase (often referred to simply base) bound to a ribose or deoxyribose sugar Nucleotides Nucleic acids base, sugar and phosphate

Purine Heterocyclic aromatic organic compound which is consisting of a pyrimidine ring fused to an imidazole ring

Pyrimidine Nitrogenous organic base, containing two nitrogen atoms at position 1 and 3 of the six-member ring

Raman scattering The phenomenon change in wavelength of a photon as result of inelastic scattering of light by molecule

Rayleigh scattering Occurs as result of elastic collision between the photons and molecules in the samples (no energy change in energy level of the analyte)

Glossary

ROA Raman optical activity one form of vibrational spectroscopies which is reliant on the difference in intensity of Raman scattered right and left circularly polarized light due to molecular Chirality

Single plasmonic Metallic nanoparticles which have localized surface plasmon nanoparticles substrate resonances that match the excitation wavelengths of lasers used in

Stokes scattering Raman scattering event where those photon scattered with less energy than the incident radiation

Surface plasmons The oscillations that occur as result of the interaction of light beam with conduction electrons held in a lattice by the presence of positive charge from the metal centre

Acknowledgements

I would like to thank to my supervisor Dr Ewan Blanch for his endless enthusiasm for making this project feasible. It was a great experience working with him. To all group mates, Clare, Myra, Ben, Christian, Grant, Nicola, Lorna, Heather, Kieaibi and my friends Zahra and Soumya for making my last three years here a wonderful experience, to Steve Prince, David Ellis and Elon Correa for giving me advice on different subjects. I am truly fortunate to know you all.

To my parents, Jafar, Suosan and my grandmother for their endless support, encouragement and love that made me believe in myself and pursue my dreams.

Without you all this would not be achievable.

My final and greatest appreciation is for my sister Soheyla, without her nothing was possible.

Chapter 1

Introduction

Chapter 1

1.0 Introduction

1.1 Conformational analysis of biological molecules

The characterisation of the three-dimensional (3D) structures of biological molecules and their relationship to their functions has made a tremendous impact on all subsequent biochemical investigations [1,2]. X-ray protein crystallography is currently the primary methodology used for determining the 3D structure of biological molecules at near-atomic or atomic resolution, the other notable atomic resolution technique being nuclear magnetic resonance (NMR) [1-4]. However, typically around 20–40% of all protein molecules, including many important proteins, are difficult or impossible to crystallize, and hence their structures have not been accessible by crystallography [1,2]. NMR has limitations on the size and type of molecules that can be structurally characterized [3,4]. Overcoming these limitations has therefore necessitated the development of novel approaches for structural biology.

One of these novel approaches is the employment of vibrational spectroscopy which is concerned with the interaction of electromagnetic radiation with matter.

Vibrational spectroscopic techniques, such as infrared (IR) [5-7] and Raman [8-10] spectroscopies have been widely applied in different fields of science such as biochemistry and biomolecular structure analysis. Infrared spectroscopy measures the light intensity of light absorbed in the infrared region of the electromagnetic spectrum. The electric dipole of the molecule must change during a molecular vibration in order for the molecule to absorb infrared radiation [5]. In the mid- and far-infrared spectral regions, the absorption occurs where the frequencies of light and molecular vibrations are equal which causes promotion of the molecule to a vibrationally excited state [5]. Raman spectroscopy is also based on vibrational

Chapter 1 transitions that give rise to narrow spectral features characteristic of the investigated sample [11] and measures the intensity of light that is inelastically scattered from the molecule.

Raman and infrared spectroscopies possess several advantages in contrast to other analytical techniques since they are non-invasive, the requirement for sample mass/volume is minimal and more importantly, there is no requirement for chemical labelling/probes. Protein aggregation, stability, conformational changes induced by different factors, the accurate prediction of structure and folding can all be assessed by vibrational spectroscopy [5-11].

1.2 Infrared and Raman spectroscopies

Light is a form of electromagnetic (EM) radiation composed of electric and magnetic waves that are oriented perpendicular to each other when they oscillate in single planes. The interaction of electric waves with matter can either lead to the absorbance or scattering of the incident light. If the energy of the photon matches that of the energy difference between the ground and excited states of a molecule, the photon may be absorbed and the molecule is promoted to a higher energy excited state. The energy change resulting from this phenomenon is measured by absorbance spectroscopy through the detection of the energy lost from the initial radiation.

Infrared spectroscopy is a technique derived from the vibrations of atoms of a molecule. An infrared spectrum is commonly obtained by passing infrared radiation through a sample and determining what fraction of the incident radiation is absorbed at a particular energy. The energy at which any peak in an absorbance spectrum appears corresponds to the frequency of a vibration of the sample molecule. This

Chapter 1 interaction of infrared radiation with matter can be explained by changes in the molecular dipole associated with vibrations and rotations [5-7,11].

Light scattering occurs when there is no necessity for a photon to possess an energy matching the energy difference between the ground and excited vibrational levels.

This phenomenon of change in wavelength as a result of inelastic scattering of light by matter was first observed by the Indian scientist C.V. Raman in 1928 [12]. When the incident photon collides with a molecule, the electron distribution is perturbed which results in the Raman effect. The majority of photons are scattered elastically which is known as Rayleigh scattering, with no change in wavelength of the scattered photon [11]. However, a transfer of energy can occur either from the incident photon to the molecule or from the molecule to the incident photon, if nuclear motion is induced. This results in inelastic scattering of the photon for which the energy of the incident photon is different to that of the scattered photon, termed

Raman scattering [13].

In the energy level diagram shown in Figure 1.1, at room temperature most molecules are present in the lowest energy vibrational level [11], and with a monochromatic light beam of photon energy hv and wavelength λ the Raman scattering process from the ground vibrational state m leads to the absorption of energy to a higher excited vibrational state n. This difference in energy between the incident and scattered photons is represented by Stokes and anti-Stokes lines. The scattered photon for Stokes lines has a lower energy than the incident photon, and for anti-Stokes lines the incident photon has less energy than the scattered photon [11].

Variation in energy from the excitation state is correlated to the vibrational energy spacing in the ground state of the molecule, hence, the vibrational energy of the molecule can be a probe of molecular chemistry of the sample via quantification of

Chapter 1 the frequencies of Stokes and anti-Stokes lines [14]. Vibrational motion is sensitive to chemical modification and therefore the molecular chemistry of samples can be studied. In Raman spectroscopy typically the Stokes scattering is used for analysis of molecular structure since the anti-Stokes scattering is weak.

Virtual State

Stokes Rayleigh Anti Stokes

Vibrational Levels

n m

Ground Electronic State

Figure 1.1: An energy level diagram showing the transitions involved in Rayleigh and Raman scattering, adapted from [11].

Chapter 1

1.2.1 Principles of Raman spectroscopy:

As an oscillating beam of light interacts with a molecule, the electron cloud of the molecule is perturbed periodically with the same frequency as the electric field of the incident wave. The perturbation of the electron cloud results in a periodic separation of charge within the molecules that is termed an induced dipole moment. The oscillating induced dipole moment is a source of EM radiation, thereby resulting in scattered light [15].

As discussed above, the electric field associated with the laser radiation induces a dipole moment in the molecule which is proportional to the electrical field strength E and to the molecular polarizability α (the ability of the electron charge distribution to be distorted by an electric field) that depends on the molecular structure and the nature of the molecular bonds.

The strength of the induced dipole moment μ is given by [15],

(1)

Because of the vector nature of the dipole moment and electric field, α is not a simple constant, but can be written as a tensor by taking account of the contributions with respect to the three Cartesian axes x, y and z. All three components of E contribute to each of the three components of μ shown in tensor form, as they refer to the Cartesian axis directions [11].

(2)

Chapter 1

In the case of aqueous solution samples, there is no ordering of the axes of the molecule to the polarization direction of the light but it is possible to measure these from polarization measurements. In practical situations, the ratio of depolarization is determined where the intensity of a given band is measured with respect to the plane of polarization of the incident light being parallel or perpendicular to the scattered light. The average polarizability can also be described in terms of isotropic (with the analyzer parallel to the plane of the incident radiation) and anisotropic (with the analyzer perpendicular to the plane) components of the tensors as represented in equations (3) and (4), where and represent isotropic and anisotropic terms, respectively [11,14],

(3)

(4)

1.2.2 Resonance Raman spectroscopy

Historically, coloured compounds were avoided by most Raman spectroscopists.

This was mainly due to decomposition of the sample by the powerful lasers used that prevents Raman analysis as a result of strong fluorescence [16,17]. However, if the frequency of the laser beam is close to the frequency of an electronic transition, scattering enhancement of up to 106 can be observed [11,16-17]. This results in a more sensitive technique in contrast to conventional Raman spectroscopy since a chromophore provides more efficient scattering that is selective for the molecule involving the chromophore [16,17]. The resonance scattering can provide both electronic and vibrational information concerning the molecule of interest.

Chapter 1

Resonance Raman scattering can occur when an incident laser beam has an excitation frequency close to that of an electric transition. A tuneable laser beam can be used for excitation, and the frequency would correspond to the energy difference between the ground vibrational state and the first or second vibronic state of the excited state.

The resonance condition is met when the energy difference between the lowest vibrational state of the ground electronic state and the resonant vibronic state is of the same energy as the excitation resulting from the incident light. The obtained enhancement of Raman scattering is mainly due to an increase in polarizability

[16,17]. Elucidation of structural information from deep within complex biological samples was enabled through development of this technique [18-20].

1.2.3 Raman spectroscopy of biomolecules

The molecular information provided by Raman spectroscopy is the same as that from infrared spectroscopy. However, the Raman effect has advantages over IR absorption for aqueous environments since less interference occurs from the solvent [5-10]. This advantage is beneficial for studying biological samples in solution since water in most cases is a pre-requisite for functioning in the surrounding physiological environment. Raman spectroscopy has a number of advantages and disadvantages compared to other analytical techniques for studying biological samples, which are summarised in Table 1.1.

Raman spectroscopy still remains a practical method for probing the interplay between structure, dynamics and function of biomolecules [21-25]. Understanding the precise structure of biomolecules in terms of their vibrational spectra can have a large impact on discovery of exact physiological function in living systems. The

Chapter 1 vibrational modes of biomolecules that can be studied by vibrational spectroscopies such as Raman are characteristic of their molecular structure. However, due to the large number of vibrational modes in biomolecules, it is a complex task to elucidate detailed information based on the measurements of vibrational spectra. Even so, important information on secondary structure elements may frequently be derived

[21-25].

Table 1.1: Advantages and disadvantages of Raman spectroscopy [5-11].u Column2

Advantages Disadvantages

Distinct characteristic vibrations that can be Weak effect used as finger prints for qualitative/ quantitative identification

Lack of interference with other vibrational Interference from fluorescence bands which results in narrower absorption bands from the laser beam

No or minimal sample preparation required Decomposition of coloured samples as a result of heating

Minimal volume

Minimum absorption by water molecules Non-invasive

Can be used for a wide range of conditions e.g. aqueous, gas, solid, tissue

Table 1.2, displays the most notable spectral regions from protein vibrations, which are called amide I, II and III [26-28]. The spectral information obtained in these regions is a sensitive indicator of the presence of secondary structure within biomolecules such as proteins and peptides where they have been used to estimate the amount of α-helix and β-sheet content [26-28].

Chapter 1

Table 1.2: Important spectral regions for protein vibrations [26-28]

Band Wavenumber (cm-1) Vibrational Assignment

Amide I 1630-1700 C=O Stretching Amide II 1480-1575 N-H bend/ C-N stretching Amide III 1230-1330 N-H bend/ C-N stretching/ Cα-H deformation

Bands in the vicinity of 1655-1659, 1300 and 1340 cm-1 in the amide I and III regions indicate α-helical conformations whereas bands in the vicinity of 1670, 1700 and 1229-1240 cm-1 usually indicate β-sheet conformations [26-28].

Chapter 1

1.3 Raman Optical Activity (ROA)

The vibrational optical activity of chiral molecules, exemplified by Raman optical activity (ROA), was predicted by Atkins and Barron in 1969 [29]. They noticed a new optical process involving interference between light waves scattered through the polarizability and optical activity tensors of a chiral molecule that was first experimentally measured by Barron, Bogaard and Buckingham, in 1973, who observed a small difference in the intensity of Raman scattering in right- and left- circularly polarized light from α-phenylethylamine and α-phenylethanol [30]. This observation was independently verified by Hug in 1975 (Figure 1.2).

Raman IR+ IL

R L z ω y

 

ROA IR- IL

Figure 1.2: Schematic diagram of the basic SCP ROA experiment which measures as a small circular component in the scattered light in right (R) and left (L) using unpolarized incident light (adapted from [11]).

The ROA measurement can be represented in terms of the circular intensity difference (CID) that is defined by;

(5)

Chapter 1 where I R and I L are scattered Raman intensities in right and left circularly polarized light, respectively.

ROA measures the optical activity related to Raman scattering and the chirality associated with molecular vibrational transitions [32], where a chiral molecule is one that is not super-imposable on its mirror image. The two mirror image forms of a chiral molecule are referred to as enantiomers [33]. Chiral molecules scatter left- and right-circularly polarized light to different degrees which leads to the resultant ROA spectrum. Unlike conventional Raman spectroscopy, in which only the electric dipole interacts with the incident light, in ROA spectroscopy, contributions from magnetic dipole and electric quadrupole optical activity tensors must also be considered.

The oscillating electric dipole, magnetic dipole and electric quadrupole moments are characteristic of the scattered radiation field induced in a molecule by the incident

light. The electric dipole, magnetic dipole m and electric quadrupole moments  are described by,

(6)

(7)

(8)

where particle i at distance ri has charge ei , mass mi , linear momentum pi and the

Kronecker delta, , is a function of two variables which is equal to 1 if they are equal and 0 otherwise [31,32].

Chapter 1

The molecular multipole moments and quantum mechanical expressions for the dynamic molecular property tensors can be defined by the fields and field gradients that are assessed at the origin of the molecule. The field and field gradients are derived from the time-dependent perturbation theory and are defined as,

Electric dipole-electric dipole tensor:

(9)

Electric dipole-magnetic dipole optical activity tensor:

(10)

Electric dipole-electric quadrupole tensor:

(11)

where n and j represent, respectively, the initial and virtual intermediate states of the

molecule,  jn  j n is their angular frequency separation and is Plank’s constant [31]. Circularly polarized light scattering and diffraction are caused by the

electric dipole-electric dipole tensor , optical rotation in aqueous solution is

generated by the electric dipole-magnetic dipole tensor G ; and A is the electric dipole-electric quadrupole tensor which leads to additional contributions to optical rotation in oriented samples [34,35].

By averaging the different polarizability-polarizability and polarizability-optical activity tensors components for all possible orientations of a molecule, we can

Chapter 1 consider their tensorial components that are invariant to axis rotations as shown in equations 12-15.

The isotopic invariants are defined as,

and (12)

and the anisotropic invariants as [34, 35],

(13)

(14)

(15)

The scattering angle can be varied, for example we can have forward (0) or backward (180) scattering. Right angle scattering can also be measured using a linear polarization analyzer in the scattered beam to select either the perpendicular

(x) or parallel (z) transmission axis to the scattering plane (yz). CID expressions for

different scattering geometries can be written in terms of  , G and A ,

(16)

(17)

Chapter 1

(18)

(19)

When a molecule consists of idealized axially symmetric bonds, for which

 (G)2  (A)2 and  G0, a simple bond polarizability theory explains that ROA is generated entirely by anisotropic scattering in which case the CID expressions reduce to :

(20) and

(21)

These equations illustrate that ROA scattering intensities are maximized in the backward direction and are zero in forward scattering. This is unlike the case of conventional Raman spectroscopy, where the forward and backward scattering intensities are the same [34,35].

1.3.1 ROA spectroscopy of biomolecules

By understanding the connection that exists between protein structure and function, the behaviour of proteins can be studied. A range of techniques have been applied to the clarification of 3D structures of proteins, ranging from prediction based on the sequence and physico-chemical properties of the constituent amino acids to high resolution methods for the detection of atoms and determination of their molecular

Chapter 1 coordinates. As a result, investigations of protein structure (at primary, secondary, tertiary and quaternary levels) are important as probes of protein function in living organisms [36]. For a number of important problems in molecular biology such as protein folding, protein-protein interactions, and protein-nucleic acid interactions, quantitative measurement of the secondary structure provides significant insight into structural features critical to biological function [36,37].

Almost all biological compounds are chiral, so it is logical to investigate them not only by Raman spectroscopy but also by ROA, where additional features can be determined. Although ROA is a very weak effect with  -values typically being ~

103 105 , it provides more structural information than Raman spectroscopy as

ROA spectral details are more sensitive to stereochemistry [31].

ROA has been measured for a wide range of biological molecules including proteins, carbohydrates [37], nucleic acids [38, 39] and viruses [40]. The ROA spectra of proteins are dominated by bands originating in the peptide backbone which directly reflect their solution conformations [37]. The bands from side chains are usually not as significant in the ROA spectra of polypeptides and proteins as they are in the conventional Raman spectra [41] since the largest ROA signals are often associated with vibrational coordinates from the most rigid and chiral parts of the biomolecules

[31].

Chapter 1

1.4 Surface-Enhanced Raman Scattering

Surface-enhanced Raman scattering (SERS) is a powerful tool for determining chemical information about molecule substrates. The enhanced Raman signals in

SERS are due to enhanced electromagnetic fields that result from adsorption of molecules on nanotextured metallic surfaces [42]. The SERS observation was first reported by Fleischman and co-workers, Hendra and McQuillan in 1974 [43]. In

1977 two separate papers by Jeanmaire and Van Duyne, and by Albrecht and

Creighton confirmed the observation of a surface Raman spectrum of pyridine adsorbed on electrochemically roughened silver electrodes as a result of successive oxidation-reduction cycles [44,45].

The SERS method can increase the intensity of the Raman signal with enhancement factors of 102 106 in scattering efficiency over conventional Raman scattering [45].

It has been reported that the enhancement of Raman signals can be up to 1011 or greater for some experiments which proposes the possibility of single molecule detection levels [46,47]. Silver and gold are the typical substrates used in SERS technique and in various metal forms, for instance colloids, roughened electrodes, deposited layers and nanoshells [48]. While the type and preparation methods of metal substrates have an effect on the outcome of SERS signals, other factors such as temperature, pressure, nature of the analyte, aggregating agents and laser power can also have significant influence on SERS signals [49-51].

Several theories have been suggested to explain the mechanisms involved in SERS enhancement. The enhancement occurs due to increases in both the molecular polarizability of adsorbed species and the local electric field in the vicinity of the metallic surface [52]. However, the exact nature of SERS is still unknown, though it

Chapter 1 is now accepted that the electromagnetic and charge transfer (chemical) enhancements are the two most important mechanisms [50].

1.4.1 The Electromagnetic Mechanism

Various electromagnetic (EM) theories have been developed over the past decades.

Complete electrodynamic calculations have been performed for simpler systems and the effects of dielectric responses have been discussed by Moskovits [52]. EM enhancement only depends on the characteristics and morphology of the metal surface [53]; therefore the same enhancement factor of vibrational modes should be obtained for the same surface morphology [54]. In order to explain the theory, the morphologies of roughened metal surface need to be understood. Electrons circulate on the metal surface which is held in a lattice by the presence of positive charges.

This electron density on the surface expands in a significant distance from the surface which has the freedom of movement in the lateral direction [55,56]. As the incident electromagnetic radiation interacts with the electron density that surrounds the atomic lattice sites of the metal, vibrations in the molecule are initiated, resulting in a collective oscillation which is known as a plasmon [50,57-58]. Surface plasmons from small uniform particles or from single periodic roughness features of a surface have a resonance frequency which results in scattering of electromagnetic radiation

[57]. The dielectric constant of the metal has a direct effect on the frequency of the surface plasmon oscillation. The resonance frequency should match with the visible frequencies of Raman scattering in order to generate the SERS enhancement [57]. To facilitate enhancement it is necessary for the oscillation of the plasmon to be perpendicular to the surface plane which is usually achieved by roughening of the surface [58]. This results in an increase of the coupling concentration of the electromagnetic field in certain regions on metallic surfaces [58]. An analyte

Chapter 1 molecule on or near the metal surface interacting with a surface plasmon experiences a large electromagnetic field, resulting in enhancement of the vibrational modes in the Raman spectrum [58].

EM theory cannot entirely explain the mechanisms involved in SERS; in particular it predicts the uniform enhancement of all Raman active bands. However, this is not the case in practice since some intense bands that can be observed in Raman spectra weaken or disappear in their corresponding SERS spectra. Therefore, mechanisms other than EM must be implicated in the SERS phenomenon in order to fully explain

SERS enhancement.

1.4.2 The Chemical Mechanism (CE)

Other studies suggest that there is a second enhancement mechanism for SERS which operates independently from the EM mechanism [48,59]. Different molecules with identical polarizability, adsorbed onto the same metallic substrate under the same experimental conditions, demonstrate different enhancement factors [48]. This would suggest that an EM mechanism is not the only mechanism involve in SERS enhancement. Further evidence in support of the chemical mechanism comes from potential-dependent electrochemical experiments. When an electrode potential is scanned at a fixed laser frequency, or the laser frequency is scanned at fixed potential, broad resonance is observed [48].

Enhancement resulting from the chemical mechanism occurs due to the formation of a chemical bond between the atomic scale of metal roughness and the adsorbed analyte [10]. This chemical contribution generates surface species which consist of the analyte and surface metal atoms [10]. This in turn enables feasibility of charge transfer from the metal surface into the analyte which causes an increase in the

Chapter 1 molecular polarizability of the analyte due to interaction with the metal’s electrons

[59]. Basically, these observations can be explained by resonant intermediates in

Raman scattering in which either the electronic states of the adsorbate are shifted and broadened by their interaction with the surface, or the formation of a new electronic state arising from chemisorption [50]. The highest occupied molecular orbital and lowest unoccupied molecular orbital of the adsorbate are symmetrically disposed in energy with respect to the Fermi level of the metal [50]. In this case, charge transfer excitations can occur either from the metal substrate to the molecule or from the molecule to the metal substrate at the about half the energy of the intrinsic intramolecular excitation of the adsorbate [10,50]. Most charge transfer excitations in

SERS take place at visible wavelengths since the molecule’s lowest-lying excitation energy is in the near ultraviolet region [59].

As discussed above, it is very difficult to separate the contribution effects resulting from EM and chemical mechanisms on systems which support SERS enhancement.

Although the majority of evidence suggests that both mechanisms play a key role in

SERS enhancement, EM enhancement has the greater effect on enhancement as the charge transfer enhancement of Raman signals drops off by 1 r 3 with distance ‘r’ from the surface [10].

1.4.3 Biological applications of SERS

Metal substrates can be applied to obtain more precise information for structural determination of nucleic acids and peptides as they can give more enhanced signals in short illumination periods (less than 1s), lower laser power and sample concentrations. SERS spectra of different biomolecules such as amino acids, nucleotides, peptides, enzymes, DNA and RNA have been reported using different

Chapter 1 metal substrates e.g. Ag and Au [60-77]. One of the main advantages of SERS for analysis of biomolecules is the reduction of the luminescent background that often obscures Raman scattering from biological molecules [78]. The compatibility of the metal substrate with biomolecules and the morphology of the surface play an important role in SERS activation. This compatibility facilitates a better coupling between adsorbed sample molecules and the metallic surface for enhancement of

Raman scattering. The potential for application of SERS for analysis of biological components is illustrated by its use in the diagnosis of tissue lesions [79,80], analysis of blood components and study of tissues [79,80]. A highly sensitive and selective

SERS detection has been reported for DNA using plasmonic nanoparticle substrates, highlighting the potential of this approach as a rapid genetic analysis tool for understanding biological process, for unlocking the underlying molecular cause of diseases and for development of biosensors [81].

Chapter 1

1.5 Surface enhanced Raman optical activity (SEROA)

Existing problems with ROA spectroscopy for studying biological molecules are principally that the conditions required include high sample concentrations (10-100 mg/ml) and long acquisition periods in comparison to those required for Raman and

SERS [82]. Although ROA spectroscopy gives more incisive information about stereochemistry of biomolecules, it suffers from being a weak effect hence preventing the application of ROA to a wider range of target molecules at present.

Alternatively, SERS generates signals that can be ~106 larger than the conventional

Raman signals [45]. The combination of ROA with SERS in the form of surface enhanced Raman optical activity (SEROA) is a promising solution for widening the application of ROA to other biomolecules that are not easily accessible to other structural methods, such as unfolded proteins and viruses.

In recent decades, various theoretical aspects of surface enhancement of ROA signals have been investigated. Efrima proposed that measurements of ROA are possible for a molecule adsorbed on the metal surface which contains information on the local electric field, their gradients and, in general, local dielectric properties of the metal- solution interface [83,84]. The model relies upon large electric field gradients close to the metal surface where a molecule is subjected to a larger electric field than in the bulk solution. Enhancement of the ROA signal can be achieved by this model as the electric dipole-electric quadrupole contribution is predicted to be large. According to

Efrima’s calculations, SEROA spectra can be influenced by several properties of electromagnetic radiation once it interacts with metal surface. These include the magnitude, direction, spatial dependence and polarization of the electromagnetic radiation. In summary, Efrima proposed that SEROA can be obtained if certain conditions are met [83,84]. These are the existence of an electromagnetic field

Chapter 1 gradient near the surface of the metal surface, a phase difference between the electric field and its gradient and finally induction of resonance where the interaction between the molecule and the metal surface occurs [83,84].

Hecht and Barron also considered enhancement of the ROA signal using a metal substrate, through the approximation of a pure electric dipole surface ROA model

[85,86]. They also investigated the possibility of a SEROA spectrum being generated by an achiral molecule. They postulated that when an achiral molecule adsorbs onto the metal surface and is randomly oriented, no ROA signal can be achieved. This is mainly due to cancellation of the signal as a result of different enantiomeric projections. However, if they align in a manner to form an ordered surface, obtaining a SEROA signal from an achiral molecule is feasible.

A decade later Janesko and Scuseria considered the effect of averaging over all orientations of the metal surface, using three different models; a dipolar sphere, quadrupole sphere and a dipolar nanorod [87]. In contrast with Efrima’s models, they predicted significantly smaller CID values.

Etchegoin et al. [88] have proposed the generation of SEROA signals by modelling of single SERS experiments. In their theoretical work, they have predicted that the high enhancements associated with ‘hot spots’ for SERS single molecule detection affect the behaviour of circularly polarized light in the vicinity of the surface plasmons [88]. Given that ROA is a weaker effect than conventional Raman, they predicted that SEROA signals may be small and difficult to distinguish from background noise. Also, the detected SEROA signal comes from the electric field of a number of different molecules that may have different polarization directions which may result in cancelling of the SEROA signals and prevent reliable measurement of a SEROA spectrum.

Chapter 1

Various theoretical studies modelling ROA responses in the vicinity of metal substrates have facilitated a better understanding of the fundamental phenomenon of

SEROA [89]. The first simulation of SEROA using the time-dependent density function theory has recently been proposed by Janesko and Scuseria where interaction of adenine with a Ag cluster was investigated [90]. They predicted an enhancement factor of 104 for both SERS and SEROA and concluded that the observed enhancement due to charge transfer is larger than that of SERS. The chemical effect of analyte-colloid binding on the metal surface was also calculated

[90]. The results suggest that observed SEROA bands, in terms of signs and intensities, are very sensitive to analyte-metal orientation. As a result, they proposed that future SEROA experiments may require utilising ordered monolayers of chiral analytes to minimise this orientation effect [90]. Yang et al. [91] have studied the interaction of electromagnetic fields with light and the phenomena of enhancement of magnetic and electric field gradients. They have highlighted the potential applicability of SEROA as a chiroptical analytical tool.

Halas et al. [92] also developed a SEROA model for molecules moving near spherical metal nanoshells where the excitation profiles for a simple chiroptical model was analyzed in detail and suggested a preferred excitation wavenumber. Very recently, the matrix polarization theory was employed to model the SEROA spectra of ribose and cysteine molecules and enabled comparison with experimental results

[93]. Findings showed a strong distance dependence of enhancement between molecules and the metal surface along with dependence of the ROA ratio and Raman intensities (CID) on distance and rotational averaging. Findings from this study confirmed that maximum enhancement can be obtained by colloidal aggregates rather than asymmetry of individual particles and validated the experimental

Chapter 1 observation of L- and D-ribose which was reported as part of this thesis [94]. The importance of controlling the colloid-sugar distance was emphasised which was done in the experimental set up for SEROA measurement of the ribose molecule.

Currently, few experimental studies have been reported to validate the technique with a high degree of certainty. Kneipp et al. [95] have claimed to observe a SEROA spectrum of adenine molecule adsorbed onto silver colloidal nanoparticles. They correlated two SEROA bands to the most enhanced peak in the SERS spectrum of adenine. Since adenine is an achiral molecule, they suggested that symmetry of the molecule can be lost once it is adsorbed onto the metal surface which then becomes chiral. They have also postulated that the adenine molecule aligns in the same orientation on the metal surface that then result in prevention of signal cancelling from various pro-chiral attachments. Abdali and Blanch [82] in their review article dismissed these results and suggested that they are artefacts as Kneipp et al. neglected the problems associated with changes in the polarization state modelled by

Etchegoin et al. [84]. These problems arise from reflection of electromagnetic light from the metal surface that modify circularly polarized light and create elliptically polarized light, which may result in the generation of birefringent artefacts [84].

Also, the generation of an ROA signal under these conditions would require a significant proportion of circularly polarized light, since a significant proportion of linearly polarized light would result in SEROA artefacts.

Abdali et al. reported SEROA spectra of two resonant molecules; cyctochrome c and myoglobin as well as a nonresonant molecule; Met-enkephalin [96-98]. However, it is difficult to verify these results since there may be artefacts arising from the parent

SERS bands and also there were no enantiomers of these compounds available to

Chapter 1 verify SEROA mirror image band responses which are very important for validating of chiroptical techniques such as SEROA.

More recently, Osinska et el. have reported the measurement of mirror image

SEROA spectra of L- and D-cysteine using an electrochemically roughened solid silver based system [99]. Although this was an interesting observation, this study did not consider a number of details concerning the experimental procedure utilised.

These results cannot be deemed reliable since the authors reported that SERS spectra of L- and D-cysteine for the same experiment could not be observed. Instead they reported SERS spectra for a different colloid-based experiment with no corresponding SEROA. Since any SEROA measurement must logically be weaker than the corresponding SERS measurement, and therefore more difficult, this calls into doubt the reliability of these spectra. It appears that they did not measure

SEROA but rather the solution-phase ROA, as sample concentrations were very high in their study, complicated by the presence of large reflection-based artefacts from the metal surface.

In 2008 [100] and 2009 [101], two PhD theses presented attempts to prove SEROA as a technique, but which were unsuccessful in both cases due to the complexity of the process. However, both illustrated the importance of controlling the experimental

SERS conditions such as colloidal type, analytes, aggregating agents, pH and concentration since they have a direct effect on obtaining not only reliable SERS but also any associated SEROA spectra. The time-dependent nature of the enhancement process was shown to have a significant effect on the obtained results laying the platform for this thesis.

SEROA can still be considered an unproven technique. The main objective of this

PhD thesis is to prove and develop reliable SEROA. Obtaining a more consistent

Chapter 1 colloidal system may assist to validate SEROA as a feasible technique and this could be achieved by controlling the colloidal aggregation over long time periods using a hydrogel polymer as a stabilising agent. The measurement of mirror image bands for two enantiomers of each of ribose and tryptophan were undertaken in order to prove the validity of SEROA measurements, and this work is presented in Chapter 2. Silver silica nanotags, which have been proven to provide strong SERS enhancement, as well as stable metal nanoparticles were both used to assess the potential application of SERROA as a nanoprobe for biomolecule analysis. The chiroptical properties of these nanotags were confirmed by the measurement of mirror image surface enhanced resonance Raman optical activity (SERROA) spectra of the two enantiomers of each of ribose and tryptophan which is demonstrated in Chapter 3. As part of the PhD research undertaken a number of other studies were performed to investigate outstanding problems in biomolecular structure using analytical spectroscopies. Chapter 4 presents work in which adenosine and seven of its derivative ribonucleotides were studied by Raman, ROA and SERS in order to identify spectral markers of site-specific phosphorylation in nucleic acids. Raman and ROA spectroscopies in combination with computational modelling were used to study the structural changes in short linear dipeptides, specifically diserine and dialanine, due to cyclization; are presented in Chapters 5 and 6. The conclusion chapter highlights the importance of optimization of the correct experimental protocols for obtaining reliable SERS and SEROA spectra. The conclusion also discusses how that combination of four different spectroscopic techniques, Raman,

ROA, SERS and SEROA is more advantageous for studying of biological samples since they can provide more structural information on the nature of biomolecules along with their chirality.

Chapter 1

1.6 References

1. T.Takahashi , S. Nakatani, Surface Science. 1995, 326, 347-360. 2. F. H. C. Crick, Methods in Enzymology. 1957, 4, 127-146. 3. A. Krushelnitsky, D. Reichert, Progress in Nuclear Magnetic Resonance Spectroscopy. 2005, 47, 1-25. 4. M. W. F. Fischer, A. Majumdar, E. R. P. Zuiderweg, Progress in Nuclear Magnetic Resonance Spectroscopy. 1998, 33, 207-272. 5. M. Subirade, D. Marion, M. Pézolet, Thin Solid Films. 1996, 284-285, 326- 329. 6. I. Martin, E. Goormaghtigh, J. M. Ruysschaert, Biochimica et Biophysica Acta (BBA) Biomembranes. 2003, 1614, 97-103. 7. I. T. Arkin, Current Opinion in Chemical Biology. 2006, 10, 394-401. 8. V. A. Shashilov, M. Xu, V. V. Ermolenkov, I. K. Lednev, Journal of Quantitative Spectroscopy and Radiative Transfe. 2006, 102, 46-61. 9. S. W. Ellepola, S. M. Choi, D. L. Phillips, C. Y. Ma, Journal of Cereal Science. 2006, 43, 85-93. 10. Jr. V. Kopecký , V. Baumruk, Vibrational Spectroscopy. 2006, 42, 184-187. 11. E. Smith, G. Dent, Moder Raman Spectroscopy: A Practical Approach. 2005: Wiley. 12. C. V. Raman, K. S. Krishnan, Nature. 1928, 121, 501-502. 13. P. R. Carey, Biochemical Application of Raman and Resonance Raman Spectroscopies. 1982: Academic Press. 14. R. L. McCreery, Raman Spectroscopy for Chemical Analysis. 2000: Wiley. 15. H. T. Helmut, A. G. Urena, J. Robert, Laser Chemistry: Spectroscopy, Dynamics and Aplication. 2007: Wiley. 16. B. Robert, Photosynthesis Research. 2009, 101, 147-155. 17. A. Albrecht, Journal of Chemical Physics. 1961, 34, 1476-1484. 18. A. V. Ruban, R. Berera, C. Ilioaia, Ivo. H. M. van Stokkum, J. T. M. Kennis, A. A. Pascal , H. V. Amerongen, B. Robert, P. Horton, R. V. Grondelle, Nature. 2007, 450, 575–578. 19. A. M. Dokter, M. C. Van Hemert, C. M. In’t Velt, K. van der Hoef, J. Lugtenburg, H. A. Frank, E. J. J. Groenen, Journal of Physical Chemistry A. 2002, 106, 9463-9469. 20. M. Cecarelli, M. Lutz, M. Marchi, Journal of the American Chemical Society. 2009, 122, 3532-3533. 21. L. J. Lis, J. W. Kauffman, D. F. Shriver, Biochimica et Biophysica Acta Biomembranes. 1975, 406, 453-464. 22. L. J. Lis, J. W. Kauffman, D. F. Shriver, Biochimica et Biophysica Acta Biomembranes. 1976, 436, 513-522. 23. A. Schneemann, Annual Review of Microbiology. 2006, 60, 51-67. 24. G. Cerchiaro, A. C. Sant'Ana, M. L. A. Temperini, A. M. da Costa Ferreira, Carbohydrate Research. 2005, 340, 2352-2359. 25. G. Zhang, S. Shuang, C. Dong, J. Pan, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy. 2003, 59, 2935-2941. 26. T. M. Theophanides, Infrared and Raman Spectroscopy of Biological Macromolecules. 1979, D. Rediel Pub: Boston. 27. B. G. Frushour, J. K. Koeing , Advances in IR and Raman Spectroscopy. 1975, Heyden & son: Berkshire.

Chapter 1

28. R. W. Williams, Enzymology. 1986, Academic Press. 311. 29. P. W. Atkins, L. D. Barron, Molecular Physics. 1969, 16, 453-466. 30. L. D. Barron, M. P. Bogaard, A. D. Buckingham, Journal of the American Chemical Society. 1973, 95, 603-605. 31. L. D. Barron, L. Hecht. I. H. McColl, E. W. Blanch, Molecular Physics. 2004, 102, 731-734. 32. L. D. Barron, F. Zhu, L. Hecht, G. E. Tranter, N. W. Isaacs, Journal of Molecular Structure. 2007, 834, 7-16. 33. L. D. Barron, L. Hecht, E. W. Blanch, A. F. Bell, Progress in Biophysics & Molecular Biology. 2000, 73, 1-49. 34. L. D. Barron, E. W. Blanch, L. Hecht, Unfolded Proteins Studied by Raman Optical Activity, in Advances in Protein Chemistry. 2002, Elsevier Science & Technology Academic Press Inc. 35. E. W. Blanch, L. Hecht, L. D. Barron, Methods. 2003, 29, 196-202. 36. J. King, Protein and Nucleic Acid Structure and Dynamics. 1985, Benjamin/Cummings Pub. Co. 587. 37. L. D. Barron, E. W. Blanch, H. I. McColl, C. D. Syme, L. Hecht, K. Nielsen, Spectroscopy: An International Journal. 2003, 17, 101-126. 38. W. L. Peticolas, Raman spectroscopy of DNA and proteins, in Methods in Enzymology, S. Kenneth, Editor. 1995, Academic Press. p. 389-416. 39. A. J. Hobro, M. Rouhi, E. W. Blanch, G. L. Conn. Nucleic Acids Research. 2007, 35, 1169-1177. 40. E. W. Blanch, L. Hecht, C. D. Syme, V. Volpetti, G. P. Lomonossoff, K. Nielsen, L. D. Barron, Journal of General Virology. 2002, 83, 2593-2600. 41. L. D. Barron, Current Opinion in Structural Biology. 2006, 16, 638-643. 42. B. Pettinger, U. Wenning, Chemical Physics Letters. 1978, 56, 253-257. 43. M. Fleishmann, P. J. Hendra, A. McQuillan, Journal of Chemical Physics Letter. 1974, 24, 163-166. 44. D. L. Jeanmaire, R. P. Van Duyne, Journal of Electroanalytical Chemistry and Interfacial Electrochemistry. 1977, 84, 1-20. 45. M. G. Albrecht, J. A. Creighton, Journal of the American Chemical Society. 1977, 99, 5215-5217. 46. S. Nie, S. R. Emory, Science. 1997, 275, 1102-1106. 47. K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. R. Dasari, M. S. Feld, Physical Review Letters. 1997, 78, 1667-1670. 48. A. Campion, P. Kambhampati, Chemical Society Reviews. 1998, 27, 241-250. 49. S. H. Macomber, T. E. Furtak, Solid State Communications. 1983, 45, 267- 271. 50. A. G. Brolo, D. E. Irish, B. D. Smith, Journal of Molecular Structure. 1997, 405, 29-44. 51. R. K. Chang, B. L. Laube, Critical Reviews in Solid State and Materials Sciences. 1984, 12, 1-73. 52. M. Moskovits, Reviews of Modern Physics. 1985, 57, 783-826. 53. W. E. Smith, Chemical Society Reviews. 2008, 37, 955-964. 54. H. Seki, Raman Spectra of Molecules Considered to be Surface Enhanced, in Studies in Surface Science and Catalysis. D. A. King, N. V. Richardson and S. Holloway, Editors. 1986, Elsevier. p. 289-310. 55. J. A. Creighton, Surface Science. 1983, 124, 209-219. 56. M. Kerker, Accounts of Chemical Research. 1984, 17, 271-277.

Chapter 1

57. E. J. Zeman, G. C. Schatz, The Journal of Physical Chemistry. 1987, 91, 634- 643. 58. J. C. Tsang, J. R. Kirtley, J. A. Bradley, Physical Review Letters. 1979, 43, 772-775. 59. A. Otto, Journal of Raman Spectroscopy. 2005, 36, 497-509. 60. Z. M. Pan, S. H. Ueda, A. Mu, R. Cui, Y. Guo, M. Burger, A. Yeh Yin, Journal of Raman Spectroscopy. 2005, 36, 1082-1087. 61. E. Kai, S. Sawata, K. Ikebukuro, T. Iida, T. Honda, I. Karube, Analytical Chemistry. 1999, 71, 796–800. 62. P. A. van der Merwe, A. N. Barclay, Current Opinion in Immunology. 1996, 8, 257–261. 63. M. Besenicar, P. Macek, J. H. Lakey, G. Anderluh, Chemistry and Physics of Lipids. 2006, 141, 169–178. 64. H. Celia, E. Wilson-Kubalek, R. A. Milligan, L. Teyton, Proceedings of the National Academic Sciences. U. S. A. 1999, 96, 5634–5639. 65. E. Hutter, M. P. Pileni, Journal of Physical Chemistry B. 2003, 107, 6497– 6499. 66. B. P. Nelson, T. E. Grimsrud, M. R. Liles, R. M. Goodman, R. M. Corn, Analytical Chemistry. 2000, 73, 1–7. 67. T. T. Goodrich, H. J. Lee, R. M. Corn, Journal of the American Chemical Society. 2004, 126, 4086–4087. 68. H. J. Lee, A. W. Wark, R. M. Corn, Langmuir. 2006, 22, 5241–5250. 69. C. S. Thaxton, D. G. Georganopoulou, C. A. Mirkin, Clinica Chimica Acta. 2006, 363, 120–126. 70. H. Li, L. J. Rothberg, Proceeding of the National Academic Sciences. U. S. A. 2004, 101, 14036–14039. 71. S. Stewart, P. M. Fredericks, Spectrochimica Acta Part A. 1999, 55, 1615– 1640. 72. E. Podstawka, Y. Ozaki, L. M. Proniewicz, Applied Spectroscopy. 2005, 59, 1516–1526. 73. E. Podstawka, Y. Ozaki, L. M. Proniewicz, Applied Spectroscopy. 2004, 58, 570–580. 74. A. E. Aliaga, I. Osorio-Roman, C. Garrido, P. Leyton, J. Carcamo, E. Clavijo, J. S. Gomez-Jeria, G. Diaz, M. M. Campos-Vallette, Vibrational Spectroscopy. 2009, 50, 131–135. 75. E. J. Bjerneld, P. Johansson, M. Kall, Single Molecules. 2000, 1, 239–248. 76. A. Kudelski, Vibrational Spectroscopy. 2008, 46, 34–38. 77. T. Deckert-Gaudig, E. Bailo, V. Deckert, Physical Chemistry Chemical Physics. 2009, 11, 7360–7362. 78. V. P. Drachev, M. D. Thoreson, V. Nashine, E. N. Khaliullin, D. Ben-Amotz, V. Jo Davisson, V. M. Shalaev, Journal of Raman Spectroscopy. 2005, 36, 648-656. 79. S. C. Pînzaru, L. M. Andronie, I. Domsa, O. Cozar, S. Astilean, Journal of Raman Spectroscopy. 2008, 39, 331-334. 80. T. Vo-Dinh, L. R. Allain, D. L. Stokes, Journal of Raman Spectroscopy. 2002, 33, 511-516. 81. Y. C. Cao, R. Jin, C. A. Mirkin, Science. 2002, 297, 1536-1540. 82. S. Abdali, E. W. Blanch, Chemical Society Reviews. 2008, 37, 980-992. 83. S. Efrima, Chemical Physics Letters. 1983, 102, 79-82. 84. S. Efrima, The Journal of Chemical Physics. 1985, 83, 1356-1362.

Chapter 1

85. L. Hecht, L. D. Barron, Journal of Molecular Structure. 1995, 348, 217-220. 86. L. Hecht, L. D. Barron, Chemical Physics Letters. 1994, 225, 525-530. 87. B. G. Janesko, G. E. Scuseria, Journal of Chemical Physics. 2006, 125, 124704-124710. 88. P. G. Etchegoin, C. Galloway, E. C. Le Ru, Physical Chemistry Chemical Physics. 2006, 8, 2624-2628. 89. P. Bour, The Journal of Chemical Physics. 2007, 126, 136101-136103. 90. B. G. Janesko, G. E. Scuseria, The Journal of Physical Chemistry C. 2009, 113, 9445-9449. 91. N. Yang, Y. Tang, A. E. Cohen, Nano Today. 2009, 4, 269-279. 92. R. Lombardini, R. Acevedo, N. J. Halas, B. R. Johnson, Journal of Physical Chemistry C. 2010, 114, 7390-7400. 93. V. Novak, J. Sebestík, P. Bour, Journal of Chemical Theory and Computation. 2012, 8, 1714-1720. 94. S. Ostovar Pour, S. E. J. Bell, E. W. Blanch, Chemical Communications. 2011, 47, 4754-4756. 95. H. Kneipp, J. Kneipp, K. Kneipp, Analytical Chemistry. 2006, 78, 1363-1366. 96. S. Abdali, C. Johannessen, J. Nygaard, T. Nørbygaard, Journal of Physics: Condensed Matter. 2007, 19, 285205. 97. S. Abdali, Journal of Raman Spectroscopy. 2006, 37, 1341-1345. 98. C. Johannessen, P. C. White, S. Abdali, Journal of Physical Chemistry A, 2007. 111, 7771–7776 99. K. Osińska, M. Pecul, A. Kudelski, Chemical Physics Letters. 2010, 496, 86- 90. 100. A. J. Hobro, PhD thesis, Structural Investigation of RNA through Application of Raman, Raman Optical Activity and Surface Enhanced Spectroscopies. 2008, University of Manchester, UK. 101. N. R. Yaffe, PhD thesis, Raman Spectroscopic Studies of Biological Molecules. 2009, University of Manchester, UK.

Chapter 2

Use of a Hydrogel Polymer for Reproducible Surface Enhanced Raman Optical Activity (SEROA)

Chapter 2

2.0 Declaration

This chapter consists of one published full paper: S. Ostovar Pour, S. E. J. Bell, E.

W. Blanch, Chemical Communication. 2011, 47, 4754–4756.

The manuscripts have been incorporated in a format identical to that for journal submission, except for minor adjustments to incorporate them into this thesis. As first author on this publication I carried out all of the associated experimental and spectroscopic analysis. The polymer was provided by Dr Steve Bell at the Queen’s

University of Belfast.

Chapter 2

Use of a Hydrogel Polymer for Reproducible Surface

Enhanced Raman Optical Activity (SEROA)

Saeideh Ostovar Pour,*a Steven E. J. Bellb and Ewan W. Blancha

Received 30th November 2010, Accepted 25th February 2011 a Faculty of Life Sciences, Manchester Interdisciplinary Biocentre, The University of Manchester, 131

Princess Street, Manchester, UK M1 7DN.

E-mail: [email protected]; Fax: +44 (0)161 236 0409; Tel: +44 (0)161 306 5819 b School of Chemistry and Chemical Engineering, The Queen’s University of Belfast, Belfast, UK

BT9 5AG.

E-mail: [email protected] ; Fax: +44(0) 2890 976524; Tel: +44(0)2890 974470

2.1 Abstract

We present surface enhanced Raman optical activity (SEROA), as well as Raman,

SERS and ROA, spectra of D- and L-ribose. By employing a gel forming polyacrylic acid to control colloid aggregation and associated birefringent artefacts we observe the first definitive proof of SEROA through measurement of mirror image bands for the two enantiomers.

2.2 Introduction

As a result of its sensitivity to chirality, Raman optical activity (ROA), which measures a small difference in the intensity of vibrational Raman scattering from chiral molecules in right- and left-circularly polarized light [1,2] is a powerful probe of the structure and behaviour of biomolecules in aqueous solution [3–7].

However, ROA is a very weak effect being ~3–5 orders of magnitude smaller than the parent Raman scattering. The conditions of high concentration and long data collection time required for ROA currently limit its application for a wide range of

Chapter 2 biological samples. These limitations could possibly be overcome using the principles of surface enhanced Raman scattering (SERS) [8–10] in which a sample in the presence of surface plasmons localized on a neighbouring nanostructured feature of a metal surface can interact with the incident light leading to large enhancement of the

Raman scattering. However, the generation of reliable SEROA spectra of biomolecules has been problematic due to difficulties in controlling spectral artefacts and low signal-to-noise ratios which complicate detection of true SEROA signals.

Although several papers have presented possible SEROA spectra [11, 12] currently a proof demonstrating mirror image SEROA spectra from opposite enantiomers has not been reported. Recently, observation of SEROA spectra for the L- and D-enantiomers of cysteine has been claimed [13]; however the authors stated that no corresponding

SERS spectra could be measured under the same conditions. Therefore, no surface enhancement had occurred in the earlier study [13] either SERS or SEROA with the observed spectral features probably being reflection associated birefringent artefacts.

Thus, SEROA has still not been confirmed as an experimental technique. SEROA spectral features depend on SERS experimental conditions, since they reflect the stability of colloids over time periods longer than those typically used in conventional

SERS [14] Contributing factors, including the concentration of analyte and aggregating agents, pH, type of colloid and time dependence, have been studied in order to determine the effects of these parameters on SERS spectra [15–17] as they should also be optimal for measuring SEROA [18]. However, it has proven difficult to stabilise the extent of aggregation in colloidal systems sufficiently to control the fluctuation of bands in SEROA experiments, complicating validation of observed spectral features. Etchegoin et al. modelled the effect of surface plasmons on circularly polarized light [19]. Their calculations suggest that large artefacts in

Chapter 2

SEROA spectra would be highly sensitive to the nature of colloid–colloid interactions, explaining the origin of the intense and fluctuating features often observed in SEROA experiments. Slowing down changes in the aggregation state to minimize changes in colloidal interactions should improve the reliability and reproducibility of SEROA spectra.

In this study we employ a hydrophilic polyacrylic acid ‘‘polycarbopol’’ polymer as a stabilising medium. This polymer has small Raman and surface-enhanced Raman cross sections, minimising interference from background signals, does not significantly change the UV-vis absorption spectra when added to silver colloids and is known to stabilise even aggregated colloids for extended periods of time [20,21].

We report SERS and SEROA spectra, along with the Raman and ROA spectra of D- and L-ribose measured in the presence of citrate-reduced silver colloids and the polycarbopol polymer, providing the first definitive observation of SEROA.

2.3 Experimental

Silver nitrate (99%), sodium borohydride (99%), sodium citrate (99%), sodium hydroxide (99%), potassium sulphate (99%), D- and L-ribose (99%) were purchased from Sigma- Aldrich UK and used without further purification.

Citrate reduced silver colloids were prepared by reduction of silver nitrate with citrate ions [22] see supplementary information for details. The polycarbopol polymer was purchased from B.F. Goodrich Ltd and used without further purification to form the polymer-sol mixture.

The Raman (IR + IL), scattered circularly polarized (SCP) ROA (IR - IL), SERS (IR +

IL) and SCP SEROA (IR - IL) spectra were all measured using a ChiralRAMAN SCP spectrometer (BioTools Inc., Jupiter FL) operating in the backscattering configuration

Chapter 2 at an excitation wavelength of 532 nm with spectral resolution of 7 cm-1. Raman and

ROA spectra were taken with laser power of 0.625 W at the sample with data collection times of 4–6 h. The laser power for SERS and SEROA was 0.25 W at the sample with data collection times of 35 min. The details of sample, aggregating agent and colloid concentration are given in each figure legend.

All SERS samples were prepared to 1 ml, the sample was left to sit for 15 min in order to obtain maximum SERS enhancement, which was determined from time dependence measurements, and then 20 mg of polycarbophil polymer powder was added and stirred vigorously for a few seconds, then left for 60 min in order to allow full hydration and swelling of the polymer prior to data collection.

2.4 Results and Discussion

The Raman and ROA spectra in aqueous solution obtained for D- and L-ribose are shown in Figure 2.1. All spectra (Raman, ROA, SERS and SEROA) presented in this study are raw data without any smoothing, base lining, normalization or any other data pretreatment. The Raman and ROA band assignments for both enantiomers of ribose are summarized in Table S2.1 in supplementary information [23–25]. The

Raman and ROA spectra of D-ribose are in excellent agreement with those reported by Wen et al. [23] and those measured recently by Dr C. Johannessen in Glasgow

(personal communication). We have repeated the Raman and ROA spectra for L- ribose, but these have not been previously reported. Mirror image responses are observed for most ROA bands, though it is not known why no ROA band appears near 877 cm-1 for L-ribose, though this spectrum is reproducible.

Figure 2.2 A and E presents the Raman and ROA spectra, respectively, of polycarbopol in solution, measured at the same concentration as used in the SEROA

Chapter 2 experiments, with SERS and SEROA spectra of D- and L-ribose shown in Figure 2.2

C, D, G and H, respectively, before and after addition of polycarbopol polymer. The

Raman spectrum of polycarbopol shows that the polymer does not generate any significant Raman signal as this polymer has a very small Raman cross section [20], and only the spectrum of water is evident. Although the polycarbopol subunit is chiral, it has a low Raman cross-section [21] so helping to minimise its ROA spectrum. Together, this leads to the ROA spectrum of polycarbopol being very weak, barely above the noise level. Figure 2.2 B and F shows spectra for the combination of silver colloids, aggregating salt and polycarbopol (no analyte). The two stronger bands at ~1394 and 1452 cm-1 in the SERS spectrum are a fingerprint of the sol with polycarbopol.

The corresponding SEROA spectrum has negative features which are very noisy, that arise from both the polycarbopol and the interaction of plasmon resonances with circularly polarized light. Figure 2.2 C and D presents the experimental SERS spectra of D- and L-ribose before and after, respectively, the addition of the polymer. The

SERS spectra for D- and L-ribose shown in Figure 2.2 C and D were obtained using the optimum type of aggregating agent (K2SO4), its concentration (20 mM) and pH

(8.7). The optimum concentration of the polycarbopol polymer was found to be 20 mg ml-1 which generated a viscous solution that was dilute enough to pipette but thick enough to control the aggregation of colloids, and gave rise to strong SERS signals for an extended period of time. The spectra demonstrate that the SERS signals for the two enantiomers are similar both in the presence (Figure 2.2 D) and the absence (Figure 2.2 C) of the polymer. All bands measured in the conventional SERS experiments appear at the same position in the presence of the polymer with only small differences in relative intensities of bands, confirming that the polymer does

Chapter 2 not interfere with signals from ribose molecules. Time dependence measurements, see Figure S2.3 in supplementary information, show that SERS intensity is stable for over 35 min with the polymer, but for only 10 min without polymer. We conclude, therefore, that the addition of the polymer increases the stability of the aggregated colloids, allowing measurement of reliable SERS signals from the analyte. The

SEROA spectra of D- and L-ribose measured in the absence of polymer are shown in

Figure 2.2 G. These spectra present a common problem that can occur in attempts to measure SEROA spectra. The SEROA spectrum of D-ribose gives rise to a mix of

+ve and -ve bands, which are what may be expected in a chiroptical measurement, but in the spectrum of L-ribose all of the bands are negative in sign, due to difficulties in controlling the highly birefringent background signal. Therefore, we do not observe a mirror image response in Figure 2.2 G for any of the purported SEROA bands generated by the two enantiomers due to the large birefringence generated by the surface plasmons from the aggregating colloids, making it difficult to have confidence in the reliability of either of these two spectra. Furthermore, though the

SERS spectra presented in Figure 2.2 C, which are insensitive to this problem, could be reproduced many times, the corresponding SEROA spectra demonstrated poor reproducibility both from sample-to-sample and as a function of time.

Figure 2.2 H shows the SEROA spectra of D- and L-ribose with polycarbopol polymer. Both D- and L-ribose give highly reproducible SEORA spectra (see Figure

S2.4 in supplementary information for replicate measurements) with positive and negative bands. Critically, despite baseline variations, mirror image bands are now observed for the two enantiomers. The SEROA spectrum of D-ribose displays a number of bands that clearly show the opposite sign to their L-ribose counterparts.

The +ve SEROA bands for D-ribose at 1247, 1273 and 1315 cm-1 correspond to the -

Chapter 2 ve SEROA bands for L-ribose at 1242, 1270 and 1310 cm-1, respectively. A complex

-ve/+ve/-ve triplet exhibited by D-ribose from ~1100–1230 cm-1 is nicely replicated as a +ve/-ve/+ve triplet by L-ribose with similar band shapes and intensities, as is the

+ve/-ve couplet for D-ribose from ~1430–1500 cm-1. A strong +ve SEROA band at

1571 cm-1 for D-ribose gives rise to an equivalent -ve feature for L-ribose. The regions between 1300–1400 cm-1 and below 1000 cm-1 reveal a number of weak features that appear to show opposite sign for the two enantiomers, though variations in local baselines due to residual birefringent background signals complicate their analysis. However, several features in the SEROA spectrum for D-ribose do not lead to a mirror image for the L-enantiomer, most notably the +ve bands at ~1014 and

1539 cm-1 , while there are also no counterpart features to the -ve SEROA bands displayed by L-ribose at ~1699 and 1739 cm-1 . The reasons for these differences are not known, but they are reproducible (Figure S2.4, supplementary information) so do not originate from variable birefringent artefacts or shot noise.

In order to verify the reliability of this method further, L- and D-tryptophan were also measured, Figure S2.5 (spectra provided in supplementary information). In both case

L- and D-tryptophan provided mirror image response in SEROA Spectra.

2.5 Conclusion

We have demonstrated the first experimental proof of SEROA by recording SEROA spectra for two enantiomers, D- and L-ribose, along with their corresponding SERS and ROA spectra. Addition of the polycarbopol polymer provides a solution to the problem of how to stabilize the aggregated colloids, and so reduce the effect of plasmon resonance induced changes in circularly polarized light that typically plague

Chapter 2

SEROA experiments. This strategy will allow the potential of SEROA to be more effectively explored.

Chapter 2

2.6 References

1. P. W. Atkins, L. D. Barron, Molecular Physics. 1969, 16, 453–466. 2. L. D. Barron, M. P. Bogaard, A. D. Buckingham, Journal of the American Chemical Society. 1973, 95, 603–605. 3. L. D. Barron, L. Hecht, E. W. Blanch Molecular Physics. 2004, 102, 731– 744. 4. L. D. Barron, Current Opinion in Structural Biology. 2006, 16, 638–643. 5. T. Uchiyama, M. Sonoyama, Y. Hamada, R. K. Dukor, L. A. Nafie, F. Hayashi, K. Oosawa, Vibrational Spectroscopy. 2008, 48, 65–68. 6. L. D. Barron, E. W. Blanch, I. H. McColl, C. D. Syme, L. Hecht, K. Nielsen, Spectroscopy. 2003, 17, 101–126. 7. E. W. Blanch, L. Hecht, L. D. Barron, Methods. 2003, 29, 196–202. 8. D. L. Jeanmaire, R. P. Van Duyne, Journal of Electroanalytical Chemistry. 1977, 84, 1–20. 9. K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. Dasari, M. S. Feld, Physical Review Letters. 1997, 78, 1667–1670. 10. E. Koglin, H. H. Lewinsky, J. M. Sequaris, Surface Science. 1985, 58, 370– 380. 11. C. Johannessen, P. C. White, S. Abdali, Journal of Physical Chemistry A. 2007, 111, 7771–7776. 12. N. A. Brazhe, A. R. Brazhe, O. V. Sosnovtseva, S. Abdali, Chirality. 2009, 21, E307–E312. 13. K. Osinska, M. Pecul, A. Kudelski, Chemical Physics Letters. 2010, 496, 86– 90. 14. S. Abdali, Journal of Raman Spectroscopy. 2006, 37, 1341–1345. 15. A. J. Hobro, S. Jabeen, B. Z. Chowdhry, E. W. Blanch, Journal of Physical Chemistry C. 2010, 114, 7314–7323. 16. N. R. Yaffe, E. W. Blanch, Vibrational Spectroscopy. 2008, 48, 196–201. 17. N. R. Yaffe, A. Ingram, D. Graham, E. W. Blanch, Journal of Raman Spectroscopy. 2009, 41, 618–623. 18. S. Abdali, E. W. Blanch, Chemical Society Reviews. 2008, 37, 980–992. 19. P. G. Etchegoin, C. Galloway, E. C. Le Ru, Physical Chemistry Chemical Physics. 2006, 8, 2624–2628. 20. S. E. J. Bell, S. J. Spence, Analyst. 2001, 126, 1–3. 21. S. E. J. Bell, N. M. S. Sirimuthu, Analyst. 2004, 129, 1032–1036. 22. P. C. Lee, D. Meisel, Journal of Physical Chemistry. 1982, 86, 3991. 23. Z. Q. Wen, L. D. Barron, L. Hecht, Journal of the American Chemical Society. 1993, 115, 285–292. 24. P. Carmona, M. Molina, Journal of Raman Spectroscopy. 1990, 21, 395–400. 25. M. Mathlouthi, A. M. Seuvre, J. L. Koenig, Carbohydrate Research. 1983, 122, 31–47.

Chapter 2

Raman D - Ribose 10 L - Ribose

8.8x10

1467

1270

879

1127 1083

10 L

6.6x10 1640

547

970

1014

805

420

+ I

601

919 R

10 729 682 I 4.4x10

877 ROA

7.3x106 1135

550 D- Ribose

964

598 1167

1009 L- Ribose

455

510

1048

1260

1363

652

1069 1105

0.0

- I -

544

874

598 510

-7.3x106 652

452

1167

1467

1105

1048

967

1135

1363

1069

1262 1009 400 600 800 1000 1200 1400 1600 1800 wavenumber (cm-1)

Figure 2.1: Raman (IR + IL, top) and SCP ROA (IR – IL, bottom) spectra of D- and L- ribose in aqueous. Sample concentrations were 2.66 M at pH 5.46 (D-) and 5.60 (L- ribose), data collection time was 237.02 min, and laser power at the sample 0.625 W for each.

Chapter 2

6.9x108 A

4.6x108

1.5x108 1394 1452 B

L 7.5x107 + I

R 1366 1624 I D-ribose 9 2.4x10 L- ribose C

1.2x109

0.0 7.0x108 D

3.5x108

4 7.8x10 E

0.0

0.0 F

-5.0x104 L -1.0x105 - I - G

R 0.0 I

-6.5x105

1247 1539 H 2.5x105 1273 1014 1315

0.0 1242 1699 1739 400 600 800 1000 1200 1400 1600 1800 Wavenumber (cm-1)

Figure 2.2: Raman spectrum of polycarbopol in solution (A), SERS spectra before

(B) and after addition of L- and D-ribose (0.25 mg ml-1) in the presence of silver

Chapter 2 citrate reduced colloid and K2SO4 at 0.020 M concentration, data collection time: 20 min (C), SERS spectra of L- and D-ribose (0.25 mg ml-1) in the presence of polycarbopol polymer, data collection time: 20 min (D), ROA spectrum of polycarbopol polymer in solution, sample concentration 40 mg ml-1, data collection time: 218 min (E), SEROA spectra of silver citrate reduced colloids in presence of aggregating salt before (F) and after addition of L- and D-ribose, data collection time: 35 min (G), SEROA spectra of L- and D-ribose with addition of polycarbopol, data collection time of 35 min (H).

Chapter 2

2.7 Supplementary Information

* n Polycarbopol

Figure S2.1: A schematic diagram of SEROA sample preparation with use of the polycarbopol polymer.

2.7.1 Colloid Preparation

Citrate-reduced silver colloids were prepared by reduction of silver nitrate with

21 citrate ions (Lee and Meisel method), where 0.094 g of AgNO3 was dissolved in

500 ml of distilled H2O and heated to boiling point, then 10 ml of 1% trisodium citrate solution was added drop wise to the mixture. Heating was continued for another hour with constant stirring and then the solution was allowed to cool to room temperature. Approximately 300 ml of a green-grey solution was obtained at ~0.5 M concentration. All glassware used to prepare the colloids was washed prior to use with aqua regia followed by gentle scrubbing with a 2% Helmanex solution and thorough rising with distilled water.

2.7.2 Sample Preparation for Raman and ROA Measurements

Samples of D- and L-ribose for Raman and ROA spectra were prepared by dissolving into distilled water at 100 mg/ml, then were microcentrifuged for 5 minutes at 3000 rpm (1000 g) to minimize dust particles prior to loading into quartz microflourescence cells.

Chapter 2

2.7.3 Atomic Force Microscopy

Micrographs were obtained using a Veeco Picoforce Multimode AFM with standard extender module, Nanoscope IIIA controller and a Picoforce scanner.

Each AFM plate was prepared by adding 50 µl of sample to freshly cleaved mica and left at room temperature for 30 minutes. The mica was then rinsed carefully under distilled water for approximately 10 seconds and dried under a gentle stream of nitrogen. AFM was carried out in air in tapping mode with a scan size of 5 microns and a scan rate of 0.5 Hz using a Silicon ‘TAP300’ AFM cantilever and tip (oscillated at approximately 260 kHz). Figure S2.2 presents AFM images of silver citrate reduced colloids before and after addition of polycarbopol polymer. The micrograph of polycarbopol without metal colloids (Figure S2.2 B) shows a very smooth surface whereas the metal colloids give rise to a very rough surface (Figure S2.2 A). Addition of polymer to metal colloids does not induce significant change to morphologies of the metal particles that are observed in

Figure S2.2 C and confirms that aggregation is significantly reduced.

Chapter 2

Figure S2.2: Dual views of AFM images of silver citrate reduced colloids only (A), polycarbopol polymer only (B) and a mixture of silver citrate reduced colloids and polycarbopol polymer (C).

Random regions of different micrographs were selected and the diameters of nanoparticles contained within were measured by using the measuring tool in the

Nanoscope 7.2 software. The average particle sizes of silver colloids with and without polycarbopol polymer were ~67 and 70 nm, respectively, so the particle size in the polymer gel was very similar to that observed for normal silver colloids.

However, the individual colloidal particles are much more distinct in the AFM images upon addition of the polymer. This is not an issue as the maximum SERS enhancement was obtained before controlling the aggregation process. The micrographs in combination with the time dependent SERS data (Figure S2.3), which are discussed below, confirm that the addition of polycarbopol to silver colloid controls the aggregation process of the nanoparticles.

Chapter 2

2.7.4 SERS Time Dependence

The time dependence study was performed by measuring the intesities of bands at

1366 and 1409 cm-1. Figure S2.3 shows the intensity changes in both normal and polymer-SERS solution over a 35 mins time interval. The intensity in normal SERS experiments decreased significantly after 10 mins. As is shown in Figure S2.3, the enhancement was kept constant in the polymer-SERS sample for a substantially longer time.

D-ribose in polymer-SERS solution 1.8x109 D-ribose in normal SERS solution

1.6x109

1.4x109

1.2x109

1.0x109

8.0x108

6.0x108 SERS Intensity SERS

4.0x108

2.0x108

0.0 5-10 10-15 15-20 20-25 25-30 30-35 Time Intervals

Figure S2.3: Time dependence SERS study for D-ribose molecule with and without addition of polymer to SERS solution.

Chapter 2

2.3x105 D-Ribose 0.0 2.3x105 0.0 2.8x105 0.0 2.8x105

L 0.0 1.3x105

- I - L-Ribose R

I 0.0 1.4x105 0.0 1.4x105 0.0 1.8x105 0.0

400 600 800 1000 1200 1400 1600 1800 Wavenumber (cm-1)

Figure S2.4: Example repeats of SEROA spectra for D- (top) and L-ribose (bottom) with the same conditions for each spectrum, where concentration of each sample was

0.25 mg/ml and K2SO4 concentration was 0.020 M, with addition of 20 mg/mL of polycarbopol, data collection time of 35 mins.

Chapter 2

Table S2.1: Raman and ROA band assignments of L- and D-ribose in aqueous

solution

Raman ROA D-ribose (cm-1) L-ribose (cm-1) D-ribose (cm-1) L-ribose (cm-1) Assignments 420 420 420 (+ve) 408 (-ve) δ CCO +δ CCC23 467 467 452 (-ve) 455 (+ve) δ CCO +δ CCC23 547 547 550 (+ve) 544 (+ve) Sym, ring bend23 601 601 598 (+ve) 595 (-ve) δ CCO +δ CCC23 652 652 652 (-ve) 652 (+ve) δ OCO, β anomer22,23 682 682 682w (+ve) 682w (+ve) δ OCO, α anomer22,24 729 729 726 (+ve) 732 (+ve) ν CC, δ CCO +δ CCC23 805 805 802 (+ve) 802 (+ve) ν CC+ ν CO+ δ OH24 879 879 877 (+ve) 874 (-ve) ν CC+ ν CO+ δ OH24 919 919 914 (-ve) 914 (+ve) ν CC+ ν CO24 970 970 964 (+ve) 967 (-ve) ν CC+ ν CO + δ CH24 1014 1014 1009 (-ve) 1009 (+ve) ν CC+ ν CO24 - - 1048 (+ve) 1048 (-ve) ν CC+ ν CO24 1083 1083 1069 (-ve) 1069 (+ve) ν CC+ ν CO24 - - 1105 (-ve) 1105 (+ve) ν CC+ ν CO24 1127 1127 1135 (+ve) 1135 (-ve) ν CO24 23 1270 1270 1260 (+ve) 1262 (-ve) τ CH2 + δ OH 23 1327 1327 1363 (-ve) 1363 (+ve) δ CH+ ω CH2 23 1467 1467 1467 (-ve) 1467 (-ve) δ CH2 23 1640 1640 - - δ CH2

*ν = stretching mode, δ = bending, τ = torsion, ω = wagging

Chapter 2

1.20x1010

0.00 2.80x109 B

1.40x109

L 0.00 C 7

+ I 6.0x10

3.0x107

1.18x108 D

5.90x107

6 E

2.20x10

0.00

1.30x106

0.00 - I -

R 4 I 3.20x10 G

0.00

2.70x104 H

0.00

400 600 800 1000 1200 1400 1600 1800

Wavenumber (cm-1)

Figure S2.5: Raman (IR + IL, A and B) and SCP ROA (IR – IL, E and F) spectra of D- and L-tryptophan in aqueous. Sample concentrations were 50 mg/ml at pH 1.90 (D-) and 1.58 (L-tryptophan), data collection time was 4-8hrs, and laser power at the sample 0.625 W for each. SERS spectra after addition of D- and L-tryptophan (C and

D), (0.0002 M) in the presence of silver hydroxyalamine reduced colloid and MgSO4 at 0.050 M concentration, data collection time: 5 min, SEROA spectra of D- and L- tryptophan (G and H) with addition of polycarbopol, data collection time of 35 min.

Chapter 3

Induced Chirality to Non-chiral Surfaces of Silver Silica Nanotags

Chapter 3

3.0 Declaration

This chapter consists of one draft paper awaiting submission to Journal of the

American Chemical Society: S. Ostovar pour 1*, L. Rocks 2, K. Faulds 2, D. Graham2 and E.W. Blanch1, Journal of the American Chemical Society. 2012

The manuscript has been incorporated in a format identical to that for journal submission except for minor adjustments to incorporate them into this thesis. As the first author on this publication I have carried out all associated spectroscopic measurements and analysis. The experimental work involving the silica nanotag preparation was carried out at the University of Strathclyde by Ms. Louise Rocks.

Chapter 3

Induced Chirality to Non-chiral Surfaces of Silver Silica

Nanotags

S. Ostovar pour 1*, L. Rocks 2, K. Faulds 2, D. Graham2 and E.W. Blanch1

1 Faculty of Life Sciences, Manchester Interdisciplinary Biocentre, The University of Manchester, 131

Princess Street, Manchester, UK, M1 7DN.

2 Centre for Molecular Nanometrology, WestCHEM, Department of Pure and Applied Chemistry,

University of Strathclyde, 295 Cathedral Street, Glasgow, UK, G1 1XL.

* Corresponding author: [email protected]; Fax: +44 (0)161 236 0409; Tel: +44 (0)161 306

5819

3.1 Abstract

Nanoprobes can offer advantages for chemical analysis, since spectroscopic detection of biomolecules is limited by the inherently weak Raman effect. In the present work, we demonstrate that silver silica nanotags can provide an enhancement for the chirally sensitive but even weaker Raman optical activity (ROA) effect enabling highly sensitive detection of chiral biomolecules. These single nanoparticle plasmonic substrates function as achiral plasmonic nanomaterials that produce intense optical activity. This system enables a chiral response to be transmitted from a chiral analyte to the plasmon resonance of an achiral metallic nanostructure. The chiroptical properties of these nanotags were confirmed by the measurement of mirror image surface enhanced resonance Raman optical activity (SERROA) spectra of the two enantiomers of each of ribose and tryptophan. This highly sensitive probe of chiral molecules can provide a new approach for studying biomolecules in solution. This is the first report of colloidal metal nanoparticles in the form of single plasmonic substrates displaying an intrinsic chiral sensitivity once attached to a chiral

Chapter 3 molecule. Therefore, the observed mechanism is a novel and remarkable fundamental effect which can provide a new route for engineering chiral plasmonic nanomaterials.

3.2 Introduction

Single nanoparticle plasmonic substrates, such as hollow gold nanospheres [1], silver triangles [2], gold nanorods [3] and gold/silver silica nanoshells [4], have localized surface plasmon resonances that match the excitation wavelengths of lasers used in, for example, Raman spectroscopy. They provide a strong electromagnetic field that increases the Raman cross-section giving rise to the technique of surface-enhanced

Raman scattering (SERS), as opposed to relying upon high conjunction potentials or

‘hot spots’ that are characteristic of aggregated metal colloids. Additionally, they provide large scattering cross-sections that allow colorimetric detection of analytes at relatively low concentrations [5].

In previous studies, silver silica nanotag particles conjugated to dye molecules have been used as an approach to enhance the sensitivity of SERS [6]. Silica coated silver colloids that have been conjugated to dyes such as benzotriazole have an electronic transition that can be tuned to the excitation frequency of the laser, leading to surfaced-enhanced resonance Raman scattering (SERRS). Although the Raman signal enhancement by single nanoparticle plasmonic substrates can detect and provide useful information about the conformations of biomolecules, this technique is blind to chirality. Determination of molecular stereochemistry and the consequent impact of chirality on bioactivity are of particular significance owing to the fact that

99% of biomolecules are chiral. The aim of this paper is to determine the possibility of identification of chiral molecules by nanotags, in the form of chirally sensitive nanoprobes, using Raman optical activity (ROA) spectroscopy. ROA measures the

Chapter 3 intensity difference between Raman scattering of left- and right-circular polarized light from chiral molecules) [7-10]. No ROA signals are detected if a molecule possesses mirror-symmetry planes or centres of symmetry, and spectral bands of opposing signs are detected for enantiomers [11]. Although ROA is a very sensitive technique for studying biological samples it is also a very weak effect, being approximately 3-5 orders of magnitude less than the parent Raman scattering. The conditions of high concentration and long data acquisition time required for ROA currently limit its application for a wide range of biological samples. These limitations could possibly be circumvented using the application of SERRS [12-14] due to the large increase in sensitivity which makes the technique worthy of consideration for the enhancement of ROA spectra.

Recently, a class of nanomaterials, known as chiral plasmonic nanomaterials [15-17], has attracted much attention due to their ability to be used as broad band circular polarisers and to generate superchiral electromagnectic fields for ultrasensitive conformational detection of biomolecules [18]. Plasmonic nanomaterials such as metallic nanoparticles experience strong absorption in the visible wavelength range but are achiral with no inherent chiroptical properties [15-17]. It has previously been demonstrated that when biomolecules are assembled with nanomaterials, chirality from the biomolecule can be imparted to the nanomaterial, resulting in the artificial yield of a plasmon-induced CD signal in the visible spectral region [19]. This was investigated for a limited number of chiral biomolecules/nanomaterial complexes, examples of which include DNA and peptide nanotubes decorated with gold/silver nanoparticles [21-23]. Optical chirality of nanostructured systems is a remarkable area of research that is receiving increasing interest as a result of its potential

Chapter 3 applications in optically active component devices for biomedical science, environmental sensing and bioterrorism detection.

In view of the fact that plasmonic substrates offer huge electromagnetic fields for

SERS measurements, in this report fluorescent dye-labelled silver colloids have been investigated to ascertain whether they can provide the same level of enhancement for chiroptical spectroscopy, in this particular case ROA. Here, we have demonstrated chiroptical behaviour of silver silica nanotags that possess an achiral plasmonic nanostructure. Furthermore, we show that chirality was induced into the achiral plasmonic surface of the substrate by binding to L- and D-enantiomeric analytes.

3.3 Experimental

Synthesis of EDTA-reduced silver colloid (AgEDTA)

Silica encapsulated silver nanotags were synthesised following the method outlined in Graham et al. [6]. Nanoparticles of approximately 40 nm diameter were synthesised according to methods described by Fabrikano et al. [24]. Briefly, sodium hydroxide (0.4 M, 10 mL) was injected into a boiling 1 L solution of EDTA (1.62 x

10-4 M). Silver nitrate (0.026 M, 10 mL) was added to the boiling solution in 2.5 mL aliquots. Following 15 minutes of continued heating, the solution was allowed to equilibrate with room temperature. Stirring was maintained throughout.

Conjugation of dye to silver nanoparticles

Three separate replicates of nanoparticle-dye conjugates were synthesised for analysis. The tri-functional benzotriazole dye shown in Figure S3.1 (refer to supplementary information) was added to AgEDTA (1 mL, 1.00 x 10-10 M,) to a final

Chapter 3 concentration of 10-7 M. Samples were agitated prior to centrifugation (6000 rpm

(3500 g), 20 mins) and resuspension in 500 μL dH2O.

Silica encapsulation of silver – precursor conjugates

The nanoparticle (NP) – dye conjugates were prepared to 1 mL with the slow addition of ethanol. Silica growth was initiated by the addition of triethylamine (10

µL, 1% v/v in ethanol) and the sequential addition of tetraethyl orthosilicate (TEOS)

(10 µL, 4% v/v in ethanol) over a three hour period until the final concentration of

TEOS was 5.4 mM.

Functionalisation of silica encapsulated nanotags

Based on the original NP concentration, the nanotags were functionalised with approximately 4000x molar excess of the selected enantiomer, L- or D-, of the analyte, ribose or tryptophan. This was achieved by reacting 1.21 molar equivalents of triethoxysilylpropyl isocyanate with the required enantiomer; L- or D-ribose, or L- or D-tryptophan, in NaHCO3 buffer (0.1 M at pH 9) at 4°C overnight. The

“silanised” molecules were added to unwashed nanotags and agitated prior to centrifugation (7000 rpm (4700 g), 20 mins) and resuspension (500 μL dH2O). Due to the formation of dimers, trimers and possibly small aggregates it was difficult to determine the actual nanotag:analyte molar ratio, therefore a 1:500 ratio is quoted based on the initial nanoparticle concentration that had been used to prepare each sample.

All Raman and ROA spectra were measured using a ChiralRaman spectrometer

(BioTools Inc., Jupiter FL, USA) configured in the backscattering geometry and operating at a wavelength of 532 nm and spectral resolution of 7 cm-1. The laser

Chapter 3 power was set to 0.20 W, with laser power at the sample being approximately 0.10

W, and data acquisition times ranged from 5 min to 2 hour.

3.4 Results and Discussion

Silver silica nanotags were designed to possess maximum absorption corresponding to their plasmon resonance at ~514 nm. The silver silica nanoshells have been previously linked with benzotriazole dye molecules to construct SERRS nanoprobes

[6]. Certain dyes such as benzotriazole can act as both a resonant reporter for SERRS and a precursor for the formation of silica shells around silver nanoparticles. They have also been used to stabilise silver nanoparticle cores. To determine whether the silica shell interferes with the ROA signal, spectra of silver nanoparticles bound to the benzotriazole dye with and without the silica shell were recorded and are presented in Figure 3.1. This serves as a control experiment since obtaining a true

ROA signal from an analyte can be challenging when either it is a resonant molecule or silver nanoparticles are present [25,26]. The SERRS spectra of silver nanotags both with and without the silica nanoshell coating are shown in Figure 3.1 (A) and

(B). These SERRS spectra demonstrate an identical profile with strong SERRS bands attributed to the resonance effect of the dye. It is clear that the silica shell does not interfere with the SERRS signal generated. The corresponding SERROA spectra of silver nanotags with and without a silica nanoshell, Figure 3.1 (C) and (D), also show similar spectral features being dominated by positive bands. The positive SERROA bands resemble the parent SERRS bands with a lower signal to noise ratio which indicates in both cases that the observed bands principally originate from interaction of linear contaminants in the scattered circularly polarized light with the SERRS bands rather than being a true measure of optical activity, therefore indicating that

Chapter 3 they are artefacts. This is a known problem in attempts to measure the SEROA effect

[26,27] and signifies non-optical activity of the silver silica nanotags as expected due to the achiral structures of the dye and silver nanotags.

Figure 3.2 displays the SERRS and SERROA spectra of L- and D-ribose that were attached covalently to silver silica nanotags. As is clearly shown, the SERRS spectra of the two enantiomers of ribose have identical profiles to each other as well as the

SERRS spectra of the silver silica nanotag (Figure 3.1), since the resonance effect of the dye dominates the spectra. The SERS spectra of L- and D-ribose previously reported [25] are significantly different from the SERRS spectra of the same analytes shown here since direct interaction of ribose molecules with the metal nanoparticle surfaces occurred in the previous study which is not the case here. The SERROA spectra presented in Figure 3.2 (B) and (C) which represent different replicates of these experiments for ribose, exhibit a mirror image response and very different features to their corresponding SERRS bands, unlike the already discussed spectra obtained from the achiral silver silica nanotags. The corresponding SERRS spectra for Figure 3.2 (C) were identical to 3.2 (B), therefore are not shown here. SERROA spectra from the two enantiomers of ribose have both positive and negative bands, the positions of which correlate well with each other. Strong peaks that appear above

1200 cm-1 in both cases can clearly be resolved from the background noise verifying the detection of the optical activity of L- and D-ribose, and can be used to confirm the stereochemistry of each analyte. For example, the strong +ve/-ve/+ve SERROA bands at 1391, 1431 and 1450 cm-1 for L-ribose have corresponding –ve/+ve/-ve bands at 1389, 1431 and 1450 cm-1 for the D-enantiomer. In addition the band above

1600 cm-1 shows a strong mirror image response at 1616 and 1623 cm-1 for L- and

D-ribose, respectively. The observed SERROA bands are probably due to the effect

Chapter 3 of ribose chirality on the plasmon at the silver nanotag surface as it modifies the surface state. This chiral response then appears to be imprinted on the SERRS spectrum of the bezotriazole dye, leading to the enantiomeric-sensitivity observed for the SERRS bands.

In order to further evaluate these results, the SERRS and SERROA spectra of L- and

D-tryptophan were also measured and are presented in Figure 3.3. Once again, the

SERRS profiles for both enantiomers are identical to the aforementioned SERRS profile of the benzotriazole dye-tagged silver colloids. The corresponding SERROA spectra for the two enantiomers of this amino acid display mirror image responses, in particular for the bands at 1317, 1347 and 1390 cm-1. The SEROA band intensities observed for the two enantiomers of both ribose and tryptophan are similar. This indicates that both molecules provide around the same level of chiral response in the

SERRS bands from the dyes despite having different molecular weights, chemical structures and number of chiral centres. More mirror image bands are observed for

L- and D-ribose in contrast to L- and D-tryptophan within the region of 1400-1600 cm-1, mainly the two +ve doublet bands at 1531 and 1564 cm-1 for D-ribose.

The observation of mirror image SERROA bands originating from the SERRS spectrum of the benzotriazole dye can be explained on the basis that chiral molecules, in this study L- and D-ribose or tryptophan, once attached to a silver surface possess the enantioselectivity required to break the symmetric environment in the achiral metallic cluster, which is reported here by the SERRS response from the dye molecules which are in close proximity to the surface plasmons. Thus, induced dissymmetry in the surface plasmon interaction with the benzotriazole molecules is responsible for the observed mirror image bands in the SERROA spectra of L- and D-ribose and tryptophan. The SERROA phenomenon measured

Chapter 3 here is fundamentally different from that responsible for our previously reported

SEROA spectra of L- and D-ribose [25] and the SERROA spectra of two resonant proteins myoglobin and cyctochrome c [28,29], since the direct interaction between the chiral molecules and the surface plasmons from metal nanoparticles in those cases was responsible for the enhancement of ROA signals. The SEROA spectral details in those previous studies also originate directly from the analyte investigated, whereas in the SERROA spectra presented in this study we observe a chiral influence on the SERRS spectrum of the benzotriazole dye. Therefore, the molecular dissymmetry has a direct effect on the observed SEROA spectra mainly due to the field gradient generated by the plasmon resonance. Chirality observed in such systems mainly originates from dipolar interactions with chiral molecules [19]. The current observation of chiroptical behaviour monitored by our silver silica nanotags is principally due to radiative electromagnetic coupling between metallic particles and nanotag plasmons and the surrounding chiral molecules interacting over a long range [30]. The chiral electromagnetic currents generated by the perturbation existent in the presence of chiral chromophores have induced optical activity to the metal nanoparticles.

The mechanism proposed as being responsible for these SERROA results is comparable to that recently reported for a class of hybrid plasmonic nanomaterials

[30] exhibiting nanolithographed achiral gold crosses on the surface utilised with indirect adsorption of chiral molecules. Abdulrahman et al. [30] demonstrated that a chiral response was induced into the plasmonic resonance of the achiral nanostructure using circular dichroism, so via measurement of electronic excitation.

In this paper we have observed a similar response but through monitoring vibrational excitation. As proposed by Abdulrahman et al., this far-field effect that we have

Chapter 3 observed occurs due to the radiative electromagnetic interaction between a non- absorbing isotropic chiral medium and a strongly absorbing metallic plasmon resonance. In our case the plasmon resonance from the silver surface induces the

SERRS signal from the tethered benzotriazole dye molecules, with the chiral analytes then interacting with that SERRS signal to generate an enantiomerically- sensitive response.

Although a mirror image response has been obtained in this work for at least the main SERROA bands for L- and D-ribose and L- and D-tryptophan, it is not possible to assign these bands to the vibrational modes of ribose and tryptophan molecules.

This is, as explained above, because we are monitoring the induction of a chiral influence on the SERRS spectrum of benzotriazole which acts here as a reporter molecule. It is as yet unknown whether the differences observed in the details of the

SERROA band structures shown in Figures 3.2 and 3.3 are due to the still less than perfect control of birefringent artefacts arising from distortions induced in the scattered circularly polarized light by the surface plasmons, or whether they are indicative of a difference in the interactions between the different chiral analytes and the surface plasmons. However, this will be pursued in future work and the aim of this current study is to confirm measurement of the optical activity of chiral analytes through monitoring vibrational excitation in sensitive achiral nanostructures. These spectra clearly show that SERROA bands of opposing sign are obtained from the two enantiomers of chiral molecules and that these are signatures of the chiroptical nature of the interaction between these analytes and the surface plasmons of these dye- tagged nanoprobes.

Chapter 3

3.5 Conclusion

Molecular adsorption onto a nanostructured surface is of fundamental importance to many processes involving separation, bio-sensing, surface processing, lubrication and heterogeneous catalysis [31]. Therefore, the conformations of surface molecules may have a pronounced effect on the physiochemical properties of biomolecules.

However, gaining information about surface organization at a molecular level is often rather difficult and the use of spectroscopic sensing probes such as silver silica nanotags has proven indeed to be very useful.

The results obtained here have demonstrated that achiral plasmonic substrates can be used to detect optical activity through a chiral signature present in a SERRS spectrum and that this is due to the novel mechanism of induced dissymmetry in a plasmonic resonance monitored through vibrational excitation. Although much further work needs to be performed to investigate this new phenomenon, and to optimize it, this study already raises interesting possibilities for the enantioselective detection of chiral molecules, and in particular biomolecules, hence extending the scope of nanoplasmonic devices.

Chapter 3

3.6 References:

1. H. Chon, S. Lee, S. W. Son, C. Oh, H. Choo, Journal of Analytical Chemistry. 2009, 81, 3029–3034. 2. J. P. Camden, J. A. Dieringer, J. Zhao, R. P. Van Duyne, Accounts of Chemical Research. 2008, 41, 1653–1661. 3. E. Temur, I. Boyaci, U. Tamer, H. Unsal, N. Aydogan, Analytical and Bioanalytical Chemistry. 2010, 397, 1595–1604. 4. J. B. Jackson, S. L. Westcott, L. R. Hirsch, J. L. West, N. Halas, Journal of Applied Physics Letter. 2003, 82, 257–259. 5. A. W. H. Lin, N. A. Lewinski, M. H. Lee, R. A. Drezek, Journal of Nanoparticles Research. 2006, 8, 681–692. 6. L. Rocks, K. Faulds, D. Graham, Chemical Communication. 2011, 47, 4415- 4417. 7. L. D. Barron, L. Hecht, Bimolecular conformational studies with vibrational Raman optical activity. In Biomolecular Spectroscopy; Clark, R. J. H., Hester, R. E., Editors. Wiley: Chichester, 1993; Part B, pp 235. 8. T. B. Freedman, L. A. Nafie, T. A. Keiderling, Biopolymers. 1995, 37, 265- 279. 9. L. A. Nafie, Applied Spectroscopy. 1996, 50, 14A-26A. 10. L. D. Barron, L. Hecht, Vibrational Raman optical activity: From fundamentals to biochemical applications. In Circular Dichroism, Principles and Applications; Nakanishi, K., Berova, N., Woody, R. W., Eds.; VCH Publishers: New York, 1994; pp 179. 11. L. D. Barron, L.Hecht. I. H. McColl, E. W. Blanch, Molecular Physics. 2004, 102, 731-734. 12. D. Van Duyne, L. Jeanmaire, Journal of Elecroanalytical Chemistry. 1977, 84, 1-20. 13. K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. R. Dasari, M. S. Feld, Physical Review Letters. 1997, 78, 1667-1670. 14. E. Koglin, H. H. Lewinsky, J. M. Sequaris, Surface Science. 1985, 158, 370- 380. 15. J. K. Gansel, M. Thiel, M. S. Rill, M. Decker, K. Bade, V. Saile, G. von Freymann, S. Linden, M. Wegener, Science. 2009, 325, 1513-1515. 16. A. S. Schwanecke, A. Krasavin, D. M. Bagnall, A. Potts, A. V. Zayats, N. I. Zheludev, Physical Review Letters. 2003, 91, 247404-247409. 17. M. Kuwata-Gonokami, N. Saito, Y. Ino, M.Kauranen, K. Jefimovs, T. Vallius, J. Turunen, Y. Svirko, Physical Review Letters. 2005, 95, 227401- 227404. 18. E. Hendry, T. Carpy, J. Johnston, M. Popland, R. Mikhaylovskiy, A. J. Lapthorn, S. M. Kelly, L. D. Barron, N. Gadegaard, M. Kadodwala, Nature Nanotechnology. 2010, 5, 783-787. 19. J. M. Slocik, A. O. Govorov, R. R. Naik, Nano Letters. 2011, 11, 701–705. 20. G. Shemer, O. Kruchevski,G. Markovich, T. Molotsky, I. Lubitz, A. B. Kotlyar, Journal of the American Chemistry Society. 2006, 128, 11006– 11007. 21. J. George, G. Thomas, Journal of the American Chemistry Society. 2010, 132, 2502–2503.

Chapter 3

22. N. Shukla, M. A. Bartel, A. J. Gellman, Journal of the American Chemistry Society. 2010, 132, 8580–8575. 23. P. Rezanka, A. Zaruba, V. Kral, Colloids Surface A: Physicochemical and Engineering Aspects. 2011, 374, 77-83. 24. V. Fabrikanos, S. Athanassiou, K. Leiser, Zeitschrift für Naturforschung B: A Journal of Chemical Sciences. 1963, 186, 612-615. 25. S.Ostovar Pour, S.E.J. Bell, E. W. Blanch, Chemical Communications. 2011, 37, 4754-4756. 26. S. Abdali, E. W. Blanch, Chemical Society Reviews. 2008, 19, 980-992. 27. P. G. Etchegoin, C. Galloway, E.C. Le Ru, Physical Chemistry Chemical Physics. 2006, 8, 2624-2628. 28. S. Abdali, C. Johannessen, J. Nygaard, T. Nørbygaard, Journal of Physics: Condensed Matter. 2007, 19, 285205-285212. 29. C. Johannessen, P. C. White, S. Abdali, Journal of Physical Chemistry A, 2007. 111, 7771–7776 30. N. A. Abdulrahman, Z. Fan, T. Tonooka, S. M. Kelly, N. Gadegaard, E. Hendry, A. O. Govorov, M. Kadodwala, Nano Letters. 2012, 12, 977-983. 31. S. K. Basiruddin, A. Saha, N. Pradhan, N. R. Jana, Journal of Physical Chemistry C. 2010, 114, 11009-11017.

Chapter 3

5x108

+ I + R I B

5x108

5x104

- I - 0.0

R I D

5x10

0.0 200 400 600 800 1000 1200 1400 1600 1800 Wavenumber (cm-1)

Figure 3.1: SERRS spectra of nanotag (tri-functional benzotriazole dye) without silica coated silver colloids (A), with silica coated silver colloids (B) and SERROA spectra of A (C) and B (D), data collection time of 35 min and laser power at source

0.20 W.

Chapter 3

A D-ribose 8

L L-ribose 7x10

+ I + R

I B

1616

1391 2.5x105

1431

- I - 1433

R I

C 1389 1623

2.5x105

200 400 600 800 1000 1200 1400 1600 1800

-1 Wavenumber (cm )

Figure 3.2: SERRS spectra of D- and L-ribose that are attached to silver silica

nanotags (A), SERROA of D- and L-ribose replicates 1 (B) and batch 2 (C), data

collection time of 35 min and laser power at source 0.20 W.

Chapter 3

A D-tryptophan L-tryptophan

1x109

+ I +

R I

1317

B 1347 1393 1637

4x104

- I -

R I

1320 1347 1630 1390

200 400 600 800 1000 1200 1400 1600 1800

-1 Wavenumber (cm )

Figure 3.3: SERRS spectra of D- and L-tryptophan that are attached to silver silica nanotags (A), SERROA spectra of D- and L-tryptophan (B), data collection time of

35 min and laser power at source 0.20 W.

Chapter 3

3.7 Supplementary Information

Figure S3.1: Structure of tri-functional benzotriazole dye.

Chapter 4

Phosphorylation Detection and Characterization in Ribonucleotides Using Raman and Raman Optical Activity (ROA) Spectroscopies

Chapter 4

4.0 Declaration

This chapter consists of one published full paper: S. Ostovar Pour, E. W. Blanch, Applied

Spectroscopy. 2012, 289-293.

The manuscripts have been incorporated in a format identical to that for journal submission, except for minor adjustments to incorporate them into this thesis. As first author on this publication, I carried out all of the associated experimental and spectroscopic analysis.

Chapter 4

Phosphorylation Detection and Characterization in

Ribonucleotides Using Raman and Raman Optical Activity

(ROA) Spectroscopies

Seideh Ostovar Pour and Ewan W. Blanch*

Received 8 September 2011; accepted 23 November 2011

Manchester Interdisciplinary Biocentre and Faculty of Life Sciences, The University of Manchester,

131 Princess Street, Manchester M1 7DN, United Kingdom

* Author to whom correspondence should be sent. E-mail: [email protected].

4.1 Abstract

Raman and Raman optical activity (ROA) spectra are presented for adenosine and seven of its derivative ribonucleotides. Both of these spectroscopic techniques are shown to be sensitive to the site and degree of phosphorylation, with a considerable number of marker bands being identified for these ribonucleotides. ROA spectra are shown to provide the most sensitive diagnostic tool for phosphorylation characterization and quantification.

Index Headings: Raman spectroscopy; Raman optical activity; ROA;

Ribonucleotides; Phosphorylation detection.

4.2 Introduction

Protein phosphorylation is an important regulatory mechanism that plays a role in a wide range of cellular processes, causing or preventing the mechanisms of diseases such as cancer and diabetes [1–3]. Although phosphorylation pathways have been studied widely, the detection and quantification of phosphorylated species in complex mixtures has proven difficult due to the limited availability of suitable

Chapter 4 methods. The methods that are available for characterization of phosphorylated biomolecules include specific antibody staining, two-dimensional (2D) gel electrophoresis with specific fluorescent probe labeling, and, most recently, mass spectrometry [3–12]. Although mass spectrometry is a very accurate technique compared to other available methods, it is not intrinsically sensitive to the site of phosphorylation. Fluorimetric methods have in part replaced these problems by attaching fluorescent markers to specific locations of the active molecule. This technique can provide high spatial and temporal resolution; however, the addition of substantial structural elements such as fluorescent markers will affect the kinetic and thermodynamic behavior of biomolecules and may even influence the outcome of a complex enzymatic process [13,14]. Due to the limitations inherent to the above techniques, there is a need to develop a structurally sensitive and label-free method for characterizing the degree and location of phosphorylation.

In recent years, Raman spectroscopy has been widely applied in different fields of science for recognition of chemical and biological samples as well as for the classification of molecular structure [15–17]. Another promising technique that is even more sensitive to structural changes than Raman spectroscopy is Raman optical activity (ROA), which measures a small difference in the intensity of vibrational

Raman scattering from chiral molecules in right and left circularly polarized light

[18,19]. As a result of its sensitivity to chirality, ROA is a powerful probe of the structure and behavior of biomolecules in aqueous solution [20]. Recently, an investigation of protein phosphorylation using Raman and ROA spectroscopies has highlighted the sensitivity of these methods when studying phosphorylation patterns in both amino acids and proteins [21].

Chapter 4

We present here a detection platform based on Raman and ROA spectroscopies for sensitive detection of phosphate groups in adenosine ribonucleotides with positional specificity. A considerable benefit in using Raman and ROA lies in their ability to provide information about molecular structure with label-free sensitivity, making them promising candidates for detecting, both qualitatively and quantitatively, phosphorylation. Ribonucleotides present a diverse range of phosphorylated variants and are known to be amenable to spectroscopic study, and thus provide a suitable model system for testing sensitivity to phosphorylation. To our knowledge, this is the first study to present combined Raman and ROA spectra of adenosine ribonucleotides containing one or more phosphate groups at different positions on the ribose ring, obtained under the same experimental conditions.

4.3 Experimental

Samples of adenosine, adenosine 2’,3’-cyclic monophosphate sodium salt

[A(2,3)MP], adenosine 3’,5’-cyclic monophosphate sodium salt monohydrate

[A(3,5)MP], adenosine 2’-monophosphate [A(2)MP], adenosine 3’-monophosphate

[A(3)MP], adenosine 5’-monophosphate [AMP], adenosine 5’-diphosphate [ADP], and adenosine 5’-triphosphate [ATP] were obtained from Sigma-Aldrich Ltd (Poole,

Dorset, UK) and used without further purification. For Raman and ROA measurements, each sample was prepared by dissolving dry powder in distilled deionized H2O at a concentration of ~100 mg/mL, which was then transferred to a quartz microflourescence cell. Each cell was then microcentrifuged at 3000 rpm

(1000 g) for 5 min in order to minimize dust particles prior to spectral measurements.

The Raman and ROA measurements were performed using a ChiralRaman spectrometer (BioTools Inc., Jupiter, FL) configured in the backscattering geometry,

Chapter 4 with an excitation wavelength of 532 nm and spectral resolution of 7 cm-1. The

Raman and ROA experimental conditions were as follows: power at the laser head was 0.60 W, data collection times of 4–12 h. A baseline subtraction was performed for each spectrum prior to analysis using MATLAB 7.6 software by employing the inbuilt bioinformatics toolbox.

4.4 Results and Discussion

The Raman spectra of the seven adenosine ribonucleotides, along with that of adenosine, are presented in Figure 4.1. In Table 4.1, the Raman assignments associated with these molecules are summarized. The results are in agreement with previous studies on adenosine, AMP, A(3,5)MP, ADP, and ATP molecules [22–24].

In general, the spectral features of all samples except ATP are similar in terms of band intensity and profile, but several differences are observed. For example, the band observed around 754 cm-1, which is assigned to a ring breathing mode in adenosine molecules [22], shifts to a lower wavenumber with the addition of a phosphate group. This band remains at 731 cm-1 for both AMP and ADP, whereas for

ATP the band shifts lower to 693 cm-1. As the position of a phosphate group changes along the ribose ring from C2 to C5, the peak becomes more intense as well as shifting to a lower wavenumber.

The Raman bands that are associated with phosphate group vibrations principally occur in the region between 900 and 1200 cm-1 [23,24]. The peaks observed in this region indicate that Raman spectra are responsive to phosphate group numbers and position around the ribose ring. Therefore, the marker band for each analyte under study can be found in this region. There are weak/no bands present in this region for adenosine, so that any intense band within this region for any ribonucleotide studied

Chapter 4 here can be assigned to a phosphate group vibrational mode. Bands at 1083 and 1113 cm-1, assigned to the phosphate group vibration, are noted in the Raman spectra of

AMP and ADP, respectively, and so serve as the sole marker bands for distinguishing between these two analytes as the rest of their Raman spectra are virtually identical. The most intense band arising from this phosphate group vibration can be observed for ATP at 1092 cm-1, presumably because it has the highest number of phosphates [23,24]. Previous researchers [25–33] pointed out that the band near

~1100 cm-1 in the spectra for polyribonucleotides is due to the symmetric stretching

2- vibration of the PO3 moiety and shows similar relative intensities to our results. As the phosphate group moves from C5 to C3 around the ribose ring, the spectral appearance changes and the peak at 1083 cm-1 shifts to lower Wavenumber at 946 cm-1.

The signature band of the phosphate group vibration for A(2)MP molecule appears at

1048 cm-1 while this band is upshifted to 1081 cm-1 for A(2,3)MP. The most significant difference for A(3)MP and A(3,5)MP appears in the bands at 946 and

1056 cm-1, respectively, which are assigned to phosphate vibrations, while other bands for these two molecules are almost identical.

The Raman spectra of all analytes investigated in this work are dominated by the adenine ring breathing C–H, C–N, and C=C stretching modes [34], which appear as a triplet of bands from 1275 to 1392 cm-1, apart from the case of ATP, which shows a doublet of peaks at 1295 and 1377 cm-1. The relative intensity of these triplet bands varies for adenosine, ATP, A(2)MP, and A(2,3)MP but not for AMP, A(3)MP,

A(3,5)MP, and ADP, for which the positions and intensities of these bands are identical. Raman bands observed at 1499 and 1524 cm-1 for adenosine are associated with ribose CH2 bending and C–C and C–N stretching coordinates [33], while these

Chapter 4 bands are downshifted to 1487 and 1514 cm-1 for AMP and ADP. The same bands shift to 1477 and 1529 cm-1 for ATP, which illustrates the significant changes in the vibrational modes upon addition of the third phosphate group. The Raman spectra for

A(2)MP, A(2,3)MP, and A(3)MP show identical bands at 1455 and 1480 cm-1 for these vibrational modes, which shift by 3 to 5 cm-1 for A(3,5)MP. There is a band observed at 1584 cm-1 in the Raman spectra of AMP and ADP that is assigned to a combined C=C and C=N stretching mode that appears at 1582 cm-1 for ATP and

1550 cm-1 for A(2)MP, A(2,3)MP, A(3)MP, and A(3,5)MP [23,35].

In addition to Raman spectra measurement, surface enhanced Raman spectroscopy

(SERS) was also performed (see Supplemental Material, available online). It was found that SERS generated very similar spectra for A(2,3)MP, A(3)MP, and

A(3,5)MP, suggesting that all of these analytes interact with the surface in a similar manner where the adenine moiety interacts most directly with the metal surface while the phosphate group is located further away, hence limiting the structural sensitivity of this method towards phosphate position.

Figure 4.2 presents the ROA spectra of the seven adenosine ribonucleotides and that of adenosine in aqueous solution. The ROA band assignments for AMP, ADP, ATP,

A(2)MP, A(2,3)MP, A(3)MP, and A(3,5)MP are listed in Table 4.2. In general, all spectra show very distinct profiles for each individual species and a number of marker bands are obvious. Notable differences are apparent between the ROA bands in the region of 600–800 cm-1, which involves the adenosine backbone vibration. The

ROA spectrum of adenosine gives a single band at 793 cm-1 that is due to the ribose ring breathing mode [26–28].

The +ve/-ve couplet band at 761/782 cm-1 for AMP and ADP corresponds to the couplet at 723/749 cm-1 for A(2)MP, presumably because the adenosine backbone

Chapter 4 appears to be in a more stereochemically restricted environment and has less freedom to rotate, since the ROA spectra of ATP, A(2,3)MP, A(3)MP, and A(3,5)MP show only a weaker single band at 738, 744, 735, or 732 cm-1, respectively. From the ROA spectra of all seven ribonucleotides in this region it is noted that more bands are detected in contrast to the case of adenosine; therefore, the addition of phosphate to the ribose ring causes considerable change to its conformation and vibrational dynamics. The region of 900–1200 cm-1 includes bands corresponding to the

2- vibrations of PO3 , the ribose moiety, C–O, and C–C in the ribose ring [26–28]. The intense band at 925 cm-1 for adenosine that is mainly assigned to a combined C–O and C–C stretching vibration in the ribose ring is found at the same position for AMP and ADP but shifts to 888 cm-1 for ATP. The ROA spectra of AMP and ADP in this region are very similar except for appearance of the weak bands at 1180 and 1050

-1 2- cm that are associated with the PO3 symmetric vibration mode [26–28].

The absence of bands at 1075 and 1127 cm-1 in the ROA spectrum of adenosine, but their presence in the spectrum for ATP, indicates that these bands arise from phosphate groups. As the phosphate group moves from C5 to C2 and C3 positions within these nucleotides, more intense ROA bands are detected. The +ve/-ve/+ve triplet at 879/905/931 cm-1 for A(2)MP, which appears at 882/914/936 cm-1 for

A(3)MP, evidently is sensitive to the location of the phosphate group on the ribose ring. The -ve/+ve couplet at 1064/1110 cm-1 corresponding to the phosphate group vibration in the ROA spectrum of A(2,3)MP undergoes an inversion of sign and shifts to 1078/1113 cm-1 for A(3,5)MP. This striking change in spectral signature from the phosphate group when attached to two different positions on the ribose ring, so giving rise to a more rigid and chiral structure upon cyclization, indicates that the signals coming from phosphate are stronger for A(2,3)MP and A(3,5)MP. We note

Chapter 4 the absence of this couplet in the ROA spectra of AMP, ADP, ATP, A(3)MP, and

A(2)MP for which the phosphate groups have a more flexible conformation.

The higher wavenumber bands are the most intense and mainly correspond to purine ring stretching modes, for which adenosine gives an intense peak at 1361 cm-1. Upon phosphorylation of the ribose ring of adenosine more changes are noted within this region that distinctly indicate the effect of phosphorylation upon the purine ring. The set of +ve/-ve/+ve bands at 1299/1348/1381 cm-1 for AMP shift slightly to

1299/1345/1376 cm-1 for ADP, with these bands arising from combined C–C, C=N, and C–H stretching modes. The relative intensities of these triplet bands for AMP vary significantly in comparison to those measured for ADP. However, the higher peak in the triplet changes sign and shifts downward by 10 cm-1 for ATP, which indicates that the addition of the third phosphate has a large and direct effect on the purine ring’s vibrations. As the phosphate group transfers from C2,3 to C3,5 linkages, more intense bands become apparent. The +ve/-ve/+ve triplet peaks at

1312/1351/1384 cm-1 for A(2,3)MP correspond to the +ve/-ve/+ve triplet bands at

1309/1345/1386 cm-1 for A(3,5)MP. The +ve/-ve couplet at 1207/1265 cm-1 for

A(3,5)MP is absent for A(2,3)MP; however, this couplet appears for A(2)MP at

1210/ 1268 cm-1. The intense band at 1422 cm-1 in the ROA spectrum of AMP, assigned to a combined C–C and C–N stretching mode, is downshifted to 1417 and

1419 cm-1 for ATP and ADP, respectively. No band of similar relative intensity can be observed for the other analytes with fewer phosphate groups or where the phosphate group is moved to a different position of the ribose ring, that is from the

C5 to C2 or C3 position.

The sharp +ve/-ve couplet at 1509/1560 cm-1, from a combined C–C, C–N, C=C, and

C=N stretching mode, for ATP is reduced in relative intensity and is accompanied by

Chapter 4 inversion of sign in the cases of AMP and ADP. This couplet also shifts to

1512/1580 and 1489/1585 cm-1 for A(2)MP and A(3)MP, respectively. A ROA feature above 1600 cm-1 that is due to C=N, C–C, and C–N stretching is only apparent as an intense band for AMP and ADP at 1626 and 1621 cm-1, respectively, while only a very weak or no ROA band is observed in this region for ATP, A(2)MP,

A(2,3)MP, A(3)MP, and A(3,5)MP.

4.5 Conclusion

Raman and ROA spectra of adenosine ribonucleotides in aqueous solution have been investigated to study the conformational changes of these analytes upon phosphorylation. It was found that Raman spectra were sensitive to the number of phosphate groups but not to their position around the ribose ring. The general features of the ROA spectra vary dramatically among these samples and provide a fingerprint sensitive to both the number and position of phosphate groups. Although the ROA spectra of AMP and ADP are very similar, there are several bands that can be used as markers for differentiating between them. This is notable as only an achiral phosphate group is added and no direct structural change is made to a chiral center. Therefore, several corresponding ROA bands for AMP and ADP show apparent sensitivity to an indirect change upon a chiral center. Upon cyclization of the phosphate group in A(2,3)MP and A(3,5)MP, which appears to lead to a more rigid structure in each case, unique ROA signatures are observed, which also indicate that signals for the phosphorus groups in these two cyclic ribonucleotides are opposite in sign and reflect opposite stereochemistry in the environment of the phosphorus atom in each case.

Chapter 4

The ROA technique is more insightful for investigating phosphorylation of adenosine ribonucleotides because it provides more informative details on the phosphate groups. ROA signals associated with phosphate groups evidently can distinguish between achiral and chiral environments of these phosphate groups. By contrast, Raman spectra are, of course, blind to stereochemistry. The ROA spectra reported here have demonstrated that it is feasible to distinguish between adenosine ribonucleotides that vary in the occurrence and position of phosphorylation even when the ribonucleotides have very similar structure. An earlier study from our group has reported Raman and ROA spectra for uridine and its mono and triphosphates, which showed similar strong marker band differences as a function of phosphorylation [36]. We therefore expect that all other nucleotides will possess unique phosporylation-sensitive marker bands, with ROA spectra being particularly informative.

In conclusion, we have shown that Raman and ROA spectra are very sensitive to structural differences between these adenosine ribonucleotides and specifically to the site and degree of their phosphorylation. Furthermore, of these two spectroscopic techniques ROA provides the most sensitive diagnostic of phosphorylation.

4.6 Acknowledgement

The authors thank Dr. Christian Johannessen and Professor Laurence Barron for helpful discussions. Additionally, the authors would like to thank the Royal Society of Chemistry and Analytical Trust Fund for a studentship to S.O.

100

Chapter 4

4.7 References

1. P. Cohen, Trends in Biochemical sciences. 2000, 25, 596-601. 2. T. H. Steinberg, K. R. Agnew, K. R. Gee, W. Leung, T. Goodman, B. Schulenberg, J. Hendrickson, J. M. Beechem, R. P. Haugland, W. F. Patton, Proteomics. 2003, 3, 1128-1144. 3. G. Guy, R. Philip, Y. Tan, Electrophoresis. 1994, 15, 417-440. 4. V. Sockic, M. Gorlach, S. Poznanovic, F. D. Boehmer, J. Godovac- Zimmermann, Biochemistry. 1999, 38, 1757-1764. 5. H. Kaufmann, J. E. Fussenegger, Proteomics. 2001, 1, 194-199. 6. M. Mann, O. N. Jensen, Nature. 2003, 21, 255-261. 7. R. R. E. Schweppe, C. E. Haydon, T. S. Lewis, K. A. Resing, N. G. Ahn, Accounts of Chemical Research. 2003, 36, 453-461. 8. R. D. Aebersold, R. Goodlet, Chemical Reviews. 2001, 101, 269-295. 9. S. P. Gygi, B. Rist, S. A. Gerber, F. Turecek, M. H. Gelb, R. Aebersold, Nature Biotechnology. 1999, 17, 994-999. 10. K. Zhang, H. Tang, L. Haung, J. W. Blankenship, P. R. Jones, F. Xiang, P. M. Yau, A. L. Burlingame, Biochemistry. 2002, 306, 259-269. 11. A. G. Marshall, C. L. Hendrickson, G. S. Jackson, Mass Spectrometry Reviews. 1998, 17, 1-35. 12. S. E. Martin, J. Shabanowits, D. F. Hunt, J. A. Marol, Analytical Chemistry. 2000, 72, 4266-4274. 13. R. J. Beynon, M. Pratt, Molecular & Cellular Proteomics. 2005, 4, 857-872. 14. S. E. Ong, M. Mann, Nature Chemical Biology. 2005, 1, 252-262. 15. E. B. Hanlon, R. Manoharan, T. W. Koo, K. E. Shafer, J. T. Motz, M. Fitzmaurices, J. R. Kramer, I. Itzkan, M. S. Feld, Physical Medical Biology. 2000, 45, R1-R59. 16. D. Zhang, Y. Xie, M. F. Mrozek, C. Ortiz, V. J. Davisson, D. Ben- Amotz, Analytical Chemistry. 2003, 75, 5703-5709. 17. K. Kneipp, H. Kneipp, V. B. Kartha, R. Manoharan, G. Deinum, I. Itzkan, R. R. Dasari, M. S. Feld, Physical Review E: Statistical Physics. 1998, 57, R6281-R6284. 18. P. W. Atkins, L. D. Barron, Molecular Physics. 1969, 16, 453-466. 19. L. D. Barron, M. P. Bogaard, A. D. Buckingham, Journal of the American Chemical Society. 1973, 95, 603-605. 20. L. D. Barron, L. Hecht, E. W. Blanch, A. F. Bell, Progress in Biophysics and Molecular Biology. 2000, 73, 1-49. 21. L. Ashton, C. Johannessen, R. Goodacre, Analytical Chemistry. 2011, 83, 7978-7983. 22. A. Lee, W. Anderson, R. Smith, H. Griffey, V. Mohan, Journal of Raman Spectroscopy. 2001, 32, 795-802. 23. L. Rimai, T. Cole, J. L. Parsons, J. T. Hickmott, E. B. Carew, Biophysical Journal. 1969, 9, 320-329. 24. R. F. Steiner, R. F. Beers, Polynucleotides: Natural and Synthetic Nucleic Acids , Elsevier, Amsterdam, 1961, viii, pp. 305. 25. Y. Liao, Y. Meng, H. Lei, Y. Wang, Chinese Optics Letter. 2008, 6, 61-63. 26. A. F. Bell, L. Hecht, L. D. Barron, Journal of the American Chemical Society. 1997, 116, 6006-6013.

101

Chapter 4

27. A. F. Bell, L. Hecht, L. D. Barron, Journal of Raman Spectroscopy. 1999, 30, 651-656. 28. A. F. Bell, L. Hecht, L. D. Barron, Chemistry A European Journal. 1997, 3, 1292-1298. 29. A. F. Bell, L. Hecht, L. D. Barron, Journal of Chemistry Society: Faraday Transactions. 1997, 93, 553-562. 30. A. F. Bell, L. Hecht, L. D. Barron, Journal of the American Chemical Society. 1998, 120, 5820-5821. 31. A. F. Bell, L. Hecht, L. D. Barron, Biospectroscopy. 1998, 4, 107-111. 32. G. Felsenfeld, D. R. Davies, A. Rich, Journal of the American Chemical Society. 1957, 79, 2023-2024. 33. M. Mathlouthi, A. M. Seuvre, J. L. Koenig, Carbohydrate Research. 1984, 131, 1-15. 34. T. Ueda, K. Ushizawa, M. Tsuboi, Spectrochimica Acta. 1994, 50, 1661- 1674. 35. E. Koglin, J. M. Sequaris, P. Valenta, Journal of Molecular Structure. 1980, 60, 421-425. 36. A. J. Hobro, PhD thesis, Structural Investigation of RNA through the Application of Raman, Raman Optical Activity and Surface Enhanced Spectroscopies, 2008, Manchester, University of Manchester, UK.

102

Chapter 4

Adenosine

AMP

ADP

ATP

A(2)MP Raman Intensity

A(2,3)MP

A(3)MP

A(3,5)MP

600 800 1000 1200 1400 1600 1800 Wavenumber (cm-1)

Figure 4.1: Raman spectra of adenosine (pH= 12.95), AMP (pH= 6.02), ADP (pH=

5.18), ATP (pH= 4.20), A(2)MP (pH= 3.13), A(2,3)MP (pH=5.54), A(3)MP (pH=

8.10) and A(3,5)MP (pH=6.67) in solution. The concentration for each sample was

100 mg/ml and laser power was 0.6 W at the sample.

103

Chapter 4

Adenosine

AMP

ADP

ATP

A(2)MP ROA Intensity

A(2,3)MP

A(3)MP

A(3,5)MP

600 800 1000 1200 1400 1600 1800 Wavenumber (cm-1)

Figure 4.2: ROA spectra of adenosine (pH= 12.95), AMP (pH= 6.02), ADP (pH=

5.18), ATP (pH= 4.20), A(2)MP (pH= 3.13), A(2,3)MP (pH=5.54), A(3)MP (pH=

8.10) and A(3,5)MP (pH=6.67) in solution. The concentration for each sample was

100 mg/ml and laser power at the sample was 0.6 W.

104

Chapter 4

Table 4.1: Raman band assignments of adenosine, AMP, ADP, ATP, A(2)MP,

A(2,3)MP, A(3)MP and A(3,5)MP [23-32].

Assignment Adenosine AMP ADP ATP A(2)MP A(2,3)MP A(3)MP A(3,5)MP Backbone A bend - 643 640 637 - - - 631 Ring breathing mode 754 731 731 693 699 699 696 693 OH bend - - - 789 769 772 766 743 (C-O),(C-C)(R) str - - - - 809 - - 807 “ - 821 827 815 - - - 830 “ - 855 855 858 841 850 847 - “ - 886 884 - 881 881 881 881 2- Asym PO3 bend - 920 - - - 912 - 918 “ ------946 - 2- Sym PO3 str - - - 985 979 976 976 976 2- Sym PO3 str - 1007 1012 - 1012 - - - 2- Sym PO3 str - - 1045 - 1053 1048 - 1056 2- Sym PO3 str - 1083 - 1092 1100 1081 - - 2- Sym PO3 str - - 1113s 1127s - 1143 1145 1145 Phosphate vibration - 1180 1180 - 1183 1183 1183 1185 2- Asym PO3 str 1230 1222 1225 - 1222 1222 1220 1220 C-N, C-H str 1277 1259 1256 - 1277 1275 1275 1275 Pyr ring 1326 1311 1311 1295 1306 1306 1306 1306 In-plane ring vibrations of adenine 1354 1342 1342 - - 1347 1347 1344 residue Imidazole ring str - - - - 1347 - - C-N, C=C str 1392 1382 1382 1377 - 1397 1397 1392 C-N, C-H 1442 1427 1425 1430 1427 1430 1427 - Ribose CH2 bend 1475 1465 1465 1447 1455 1455 1455 1450 C2H,N-C,C-H str 1499 1487 1487 1477 1480 1480 1480 1477 C-C, C-N (Imidazole ring) 1524 1514 1514 1529 - - - - C=C str, C=N str ------„ 1594 1584 1584 1582 1550 1550 1550 1550 C-C, C-N str 1660 1651 1653 1667 1615 1620 1617 1608

*str= stretching, bend= bending, sym= symmetric, A-sym= anti-symetric, A=adenosine, pyr= pyrimidine.

105

Chapter 4

Table 4.2: ROA band assignments of adenosine, AMP, ADP, ATP,

A(3,5)MP,A(2,3)MP,A(2)MP and A(3)MP [25-32].

Assignment Adenosine AMP ADP ATP A(2)MP A(2,3)MP A(3)MP A(3,5)MP Backbone A bending - +694 +697 +652 +679 +643 +667 +673 “ - +741 +744 +702 -723 +708 - - Ring breathing mode +793 -758 -761 +738 +749 +744 +735 +732 OH bending - +779 +782 - -796 -773 -799 -779 Phosphate vib ------“ - - - - - +808 - - - - +842 - +842 -848 - +839 asym O-P-O vib Adenosine base - -885 +882 +888 +879 +882 +882 -871 Ribose moiety and phosphate - - - - -905 -911 -914 +914 vib Ribose moiety +925 +922 +925 +946 +931 +942 +936 - 2- - - +992 -984 - +975 +984 - Sym str PO3 vib “ - - - +1023 +1006 +1009 +1012 +1025 “ - - +1050 +1075 -1094 -1064 -1064 +1078 2- - -1121 - +1127 -1121 +1110 - -1113 PO3 vib Phosphate vib - +1180 - - +1154 +1183 - +1170 Couples base and sugar ring - +1215 +1210 - +1244 +1210 +1249 +1207 vib C-N str - -1260 -1257 - -1268 - - -1265 Pyr ring - - - +1289 +1283 - -1296 - C-N,C-H str +1296 +1299 +1299 - +1320 +1312 +1320 +1309 Ribose moiety and phosphate +1338 +1348 +1345 +1330 -1348 -1351 - +1354 group C-N, C=C str +1361 +1381 +1376 -1366 - - -1361 - “ +1386 - - - +1394 +1384 +1386 +1386 C-N, C-H str -1404 +1422 +1419 +1417 - +1434 +1427 +1422 C-C-H,N-C,C-H str +1474 - +1464 +1475 +1474 +1454 +1489 -1462 C-C, C-N (Imidazole ring) +1530 -1512 - +1509 +1512 +1489 - +1494 C=C str,C=N str - +1553 +1551 -1560 -1580 +1556 - - “ - - - +1590 - -1585 -1585 -1590 C=N, C-C, C-N str -1628 +1626 +1621 +1623 - - +1604 -

*str= stretching, bend= bending, sym= symmetric, A-sym= anti-symetric, A=adenosine, pyr= pyrimidin

106

Chapter 4

4.8 Supplementary Information

4.8.1 Colloid Preparation

Citrate-reduced colloids were prepared by reduction of silver nitrate by citrate ions

(Lee and Meisel method) [1], where 0.094 g of AgNO3 was dissolved in 500 ml of

H2O and heated to boiling point, then 10 ml of 1% trisodium citrate solution was added drop wise to the mixture. Heating was continued for another hour with constant stirring and then the solution was allowed to cool at room temperature.

Approximately 300 ml of green-grey solution was obtained at 0.5 M concentration.

All glassware used to prepare the colloids was washed prior to use with aqua regia followed by gentle scrubbing with a 2% Helmanex solution and then thoroughly rinsed. Distilled water was used for all colloid and sample preparations.

Sample Preparation for SERS Measurements

The sample, aggregating agent salt (K2SO4) and colloid concentration details are given in the figure legend below. All samples were prepared to 1 ml, the sample being left to sit for 60 min in order to obtain maximum SERS enhancement.

4.8.2 Surface enhanced Raman spectroscopy (SERS)

SERS spectra of adenosine and its ribonucleotides in the presence of silver nanoparticles in an aqueous environment are shown in Figure S4.1. In Table S4.1 we present SERS band assignments for the seven adenosine ribonucleotides studied.

Under the same experimental conditions, the SERS spectra of AMP, A(2)MP,

A(2,3)MP, A(3)MP and A(3,5)MP are all seen to be very similar to each other. This postulates the possibility that all AMP derivative analytes interact with the metal surface through the same sites of the adenosine moiety. The phosphate group for

107

Chapter 4 each adenosine monophosphate appears to be located too far away from the metal surface to show any sensitivity to position. Thus, it is more difficult to differentiate between phosphate group vibrational modes in these SERS spectra.

Previous studies of adenosine, AMP, ADP and ATP at metal surfaces, mainly gold and silver, have also proposed that the binding site occur through adenine ring and the adenine ring vibration dominate the spectra [2-6]. This interpretation is supported by the fact that there are no obvious SERS bands from 1000-1100 cm-1, the region corresponding to bands arising from distinctive phosphate group vibrations in the

Raman spectra of these species.

It was found that Raman and SERS spectra were sensitive to the number of phosphate groups but not to their position around the ribose ring. Previous publications for detection of phosphorylation using SERS have shown that there are variations on results as they are very dependent on experimental conditions such as aggregating salt and analyte concentration [7, 8]. There is also a time dependent decline in colloidal based SERS signals during the measurement. Although it is possible to differentiate between individual adenosine ribonucleotides, the phosphate region barely shows reliably enhanced SERS bands [7,8].

108

Chapter 4

4.8.3 References

1. P.C. Lee, D. Meisel, Journal of Physical Chemistry. 1982, 86, 3391-3395. 2. T. Watanabe, O. Kawanami, H. Katoh, K. Honda, Y. Nishimura, M. Tsuboi, Surface Science. 1985, 158, 341-351. 3. J. S. Suh, M. Moskovits, Journal of the American Chemical Society. 1986, 108, 4711-4718. 4. C. Otto, T. J. J. Van den Tweel, F. F. M. De Mul, J. Greve, Journal of Raman Spectroscopy. 1986, 17, 289-298. 5. S. K. Kim, T. H. Joo, S. W. Suh, M.S. Kim, Journal of Raman Spectroscopy. 1986, 17, 381-386. 6. K. Itoch, K. Minami, T. Tsujino, M. Kim, Journal of Physical Chemistry. 1991, 95, 1339-1345. 7. J. Moger, P. Gribbon, A. Sewing, C. P. Winlove, Biochimica Et Biophysica Acta. 2007, 91170, 12-918. 8. B. L. Mitchell, A. J. Patwardhan, S. M. Ngola, S. Chan, N. Sundararajan, Journal of Raman Spectroscopy. 2008, 39, 380-388.

109

Chapter 4

Adenosine

AMP

ADP

ATP

A(2)MP

SERS Intensity A(2,3)MP

A(3)MP

A(3,5)MP

600 800 1000 1200 1400 1600 1800 Wavenumber (cm-1)

Figure S4.1: SERS spectra of adenosine ribonucleotides and adenosine in the presence of silver citrate reduced colloid. Analyte concentrations were 1x10-5 mg/ml,

K2SO4 concentration was 0.020 M, data collection time was 50 seconds, and laser power was 0.20 W at the laser.

110

Chapter 4

Table S4.1: SERS band assignments of adenosine, AMP, ADP, ATP, A(2)MP,

A(2,3)MP, A(3)MP and A(3,5)MP [2-6].

Assignment Adenosine AMP ADP ATP A(2)MP A(2,3)MP A(3)MP A(3,5)MP

Backbone A bend - - 616 616 - - - -

Ring breathing mode 738 735 735 735 735 738 738 738

OH bend - - 776 - - - - -

(C-O),(C-C)(R) str - - - - - 799 799 -

“ ------“ ------“ ------2- - 925 916 911 931 931 928 - Asym PO3 bend

“ - 956 959 956 - - - 950 2------984 981 - Sym PO3 str 2------Sym PO3 str 2- - 1031 1028 1034 1028 1026 1028 1034 Sym PO3 str 2- - - - - 1078 - - - Sym PO3 str 2- - - 1129 - 1143 - - - Sym PO3 str

Phosphate vibration - 1178 1183 1181 - - - - 2------Asym PO3 str

C-N, C-H str 1260 1252 1249 1247 - - - 1255

Pyr ring 1335 1328 1325 1325 1273 1328 1328 -

In-plane ring vibrations - - - - 1338 - - 1335 of adenine residue

Imidazole ring str - 1361 1364 1364 1376 1361 - 1366

C-N, C=C str 1389 1394 1394 1394 - 1399 1397 1394

C-N, C-H ------

Ribose CH2 bend 1469 1465 - - 1457 1462 1462 1462

C2H,N-C,C-H str - - 1472 1472 1495 - - -

C-C, C-N (Imidazole - 1507 1512 1512 - 1502 1504 - ring)

C=C str, C=N str - 1557 1575 - - - - - „ - 1585 1624 1578 1549 1563 1563 1558

C-C, C-N str - 1659 1652 1652 1645 1640 1643 1638

*str= stretching, bend= bending, sym= symmetric, A-sym= anti-symetric, A=adenosine, pyr= pyrimidine.

111

Chapter 5

Study of Experimental and Computational Raman and Raman Optical Activity (ROA) Spectra of Cyclic and Linear L-Ala-L-Ala in Solution

112

Chapter 5

5.0 Declaration

This chapter consists of one draft paper awaiting submission to Journal of Raman

Spectroscopy: S. Ostovar pour 1*, T. J. Dines 2, C. Levene 1, B.Z. Chowdhry 3 and

E.W. Blanch1, Journal of Raman spectroscopy. 2012

The manuscripts have been incorporated in a format identical to that for journal submission, except for minor adjustments to incorporate them into this thesis. As first author on this publication I carried out all of the associated experimental and spectroscopic analysis. The calculation were carried out by T. J. Dines and B. Z.

Chowdhry and provided here for purpose of comparison with experimental results.

113

Chapter 5

Study of Experimental and Computational Raman and

Raman Optical Activity (ROA) Spectra of Cyclic and

Linear L-Ala-L-Ala in Solution

S. Ostovar pour 1*, T. J. Dines 2, C. Levene 1, B.Z. Chowdhry 3 and E.W. Blanch1

1 Faculty of Life Sciences, University of Manchester, MIB 131 Princess Street, Manchester, M1 7DN,

2 Division of Electronic Engineering and Physics, University of Dundee, Dundee, DD1 4HN, UK

3 School of Science, University of Greenwich at Medway, Central Avenue, Chatham Maritime, Kent,

ME4 4TB, UK

E-mail: [email protected]; Fax: +44 (0)161 236 0409; Tel: +44 (0)161 306 5819

5.1 Abstract

Raman spectroscopy and Raman optical activity (which measures the small difference between the intensities of left and right scattered polarized light from chiral molecules) have been utilized to determine peptide conformations in solution as a result of their sensitivity to molecular conformation. Short dipeptides such as diketopiperazine (DKP) have been investigated past decades since both natural and synthetic DKPs have a wide variety of desired biological effects including anti- tumour, anti-viral, anti-fungal and anti-bacterial activities. Density functional theory

(DFT) calculations were carried out using the Gaussian 09 program and the IEF-

PCM solvation method with the Karplus and York continuous surface charge formalism, with water as the solvent. The starting geometries for geometry optimization were based upon those obtained from the X-ray crystal structures of L-

Ala-L-Ala. The calculated spectra of both linear and cyclic L-Ala-L-Ala are in good agreement with our experimental ROA and Raman spectra; the comparison of which

114

Chapter 5 for both forms showed that ROA is more sensitive to structural changes where it provided more marker bands. In particular, Raman bands at 1317, 1459 and 1524 cm-

1 for the cyclic form with their corresponding ROA bands at 1322, 1452 and 1521 cm-1 are not observed in the spectra for the linear form which suggests that these bands are unique for the cyclic form of dipeptides. There are considerable differences between the observed ROA bands for the cyclic and linear forms of dialanine that reflect large differences in the vibrational modes of the polypeptide backbone upon cyclization. In this study, the ROA spectrum of cyclic L-Ala-L-Ala has been reported for the first time which demonstrated that ROA spectroscopy when utilised in combination with computational modelling clearly provides a potential tool for characterization of cyclic peptides.

5.2 Introduction

Cyclic dipeptides, particularly diketopiperazine (DKP), have been extensively investigated during the last few decades as they are a large and important group of compounds in medicinal chemistry due to their pharmacological activity [1]. In some cases, cyclic compounds appear to have more interesting biological properties in comparison to their linear equivalents [2]. The small size, hydrophobicity and lack of charge exhibited by cyclic dipeptides render them more cell membrane permeable, thereby promoting the bioavailability of these peptides [3]. They are also of interest in studies on the thermodynamic behaviour of non-ionic compounds in water as a result of sharing the capability of establishing hydrogen bonds with the solvent [4].

Because of their simplicity and limited conformational freedom along with a higher probability of conformational homogeneity, cyclic dipeptides are readily used as model compounds for longer peptide molecules [5]. Both natural and synthetic DKPs

115

Chapter 5 display a wide variety of biological activities including anti-tumour, anti-viral, anti- fungal, and anti-bacterial effects [6-8]. Due to their rigidity, chiral nature and varied side chains they are also an attractive scaffold for drug design. Infrared, CD, NMR and Raman spectra of linear and cyclic L-Ala-L-Ala and its N-deuterated form have been obtained by several groups in attempts to characterize their structures in solution [9-15].

Raman optical activity (ROA), which measures the small difference between the intensities of Raman scattering using right and left circularly polarized light [16-18], has been utilized in the determination of peptide conformation in solution as a result of its sensitivity to molecular structure [19]. ROA provides more structural information than Raman spectroscopy as ROA spectral details are directly responsive to stereochemistry. Despite the considerable volume of work reported on linear dialanine structures using both Raman and ROA [20-24], some important aspects of their structures are not yet determined. Therefore, in this paper we have used both

Raman and ROA spectroscopies to investigate the differences in conformation between cyclic and linear forms of L-Ala-L-Ala.

Experimental and calculated Raman and ROA spectra of linear L-Ala-L-Ala have both been widely reported [25], and the Raman spectrum of cyclic L-Ala-L-Ala has been presented [26] but not its ROA spectrum. The combination of quantum mechanics/molecular mechanics modelling with experimental ROA spectroscopy provides a uniquely sensitive tool for investigating the conformations of unusual peptides. This approach was utilised by Bour and co-workers [25], who used the

B3LYP functional and the 6-31+G** Pople-type basis set, with the CPCM dielectric correlation in the Gaussian version of the COSMO solvent model in order to better characterize the spectra of the linear form [25]. Clusters of linear L-Ala-L-Ala with

116

Chapter 5 water molecules were obtained from molecular dynamics simulations that gave qualitatively correct inhomogeneous broadening of Raman spectra lines but did not bring a convincing improvement of ROA signals. There were also differences in the calculated ROA intensities compared to the experimental case [25]. In this work, we report the simulated and experimental ROA spectra for cyclic L-Ala-L-Ala for the first time, as well as the ROA spectrum of the linear form, in order to improve analysis of experimental spectra and so better understand the structural constraints imposed by cyclization.

5.3 Experimental methods

Cyclic and linear L-Ala-L-Ala were purchased from Bachem Ltd. (Saffron Walden,

Essex, UK) and used without further purification. Deuterium oxide (99.98 atom %),

Na2HPO4 and NaH2PO4 were obtained from Sigma-Aldrich Ltd (Poole, Dorset, UK).

The buffer and sample concentrations used are given in the corresponding figure captions. Samples for Raman and ROA spectroscopy were prepared by dissolving lyophilized material into distilled and deionised phosphate buffer solution; where low solubility was exhibited the sample was heated to 70 oC and then allowed to cool to room temperature. Each solution was then centrifuged for 5 minutes at 3000 rpm

(1000 g) in order to minimize dust particles from the environment prior to loading into a quartz microflourescence cell for spectroscopic measurement.

All Raman and ROA spectra were measured using a chiralRaman spectrometer

(BioTools Inc., Jupiter FL, USA) configured in the backscattering geometry and operating with a wavelength of 532 nm and spectral resolution of 7 cm-1. The laser power at the laser was 1.20 W and data collection times were 6-24 h.

117

Chapter 5

5.4 Computational methods

DFT calculations were carried out using the Gaussian 09 program [27] with the

B3LYP method [28, 29] and the AUG-cc-pVDZ basis set [30]. All calculations were performed with the IEF-PCM solvation method [30] using the Karplus and York continuous surface charge formalism [31, 32], with a Polarizability Continuum

Model (PCM) of the water solvent with default PCM parameters being used, except that Pauling atomic radii were substituted for the default UFF radii. The starting geometries for geometry optimisation were those previously obtained at the B3-

LYP/cc-pVDZ level for cyclic (L-Ala-L-Ala) [33], assuming C2 symmetry. The starting geometries for linear L-Ala-L-Ala geometry optimisation were based upon its reported X-ray crystal structure [34]. Vibrational spectra were calculated at the optimised geometries and Raman and ROA activities were computed dynamically for an excitation wavelength of 532 nm. Relative Raman and ROA intensities were calculated from the computed Raman and ROA activities using the following equations;

4 Raman 0  fi    22 Iafi 45 7 hc fi  fi 1 exp kT 

4 ROA 0  fi    22 I fi48 G'' 16 A hc fi  fi 1 exp kT

-1 where 0 = 18,797 cm and T = 298.15 K.

In these equations the ytensor invariants are defined as follows: a2 is the isotropic invariant of the electric-dipole/electricdipole polarizability tensor,

 is the symmetric anisotropic invariant of the electric-dipole/electricdipole

118

Chapter 5

polarizability tensor, is the anisotropic invariant of the cross-product of the electric-dipole/electricdipole polarizability tensor with the electric-dipole/magnetic-

dipole polarizability tensor, is the anisotropic invariant of the cross-product of the electric-dipole/electricdipole polarizability tensor with the tensor A obtained by contracting the electric-dipole/electric-quadrupole polarizability tensor with the antisymmetric-unit tensor of Levi-Civita.

The Cartesian force constants obtained from the Gaussian 09 output were converted to force constants expressed in terms of internal coordinates using a normal coordinate analysis program derived from that of Schachtsneider [34]. A full set of internal coordinates, including all bond angles and torsion angles, was reduced to a set of 3N-6 symmetry-adapted internal coordinates. Normal coordinate analyses were performed without scaling of force constants, producing potential energy distributions for harmonic wavenumbers. Simulated Raman and ROA spectra were constructed by convolution with a Lorentzian lineshape function of 10 cm-1 fwhm.

5.5 Results and Discussion

The atom numbering scheme along with computed molecular geometry of linear and cyclic L-Ala-L-Ala are shown in Figure 5.1 and 5.2. The experimental and calculated

Raman and ROA spectra of both linear and cyclic L-Ala-L-Ala are shown in Figure

5.3, with their corresponding vibrational band assignments in Table 5.1. The band lengths, angles and selected torsion angles are shown in Tables S5.1, S5.2 and S5.3

(refer to supplementary information). The predicted spectra of linear and cyclic L-

Ala-L-Ala agree well with both experimental Raman and ROA spectra presented here, in addition to matching those spectra reported in previous studies for the linear form [20-25]. The calculated wavenumber values in some regions are slightly

119

Chapter 5 overestimated by less than 1%; the relative intensities in the spectra of linear L-Ala-

L-Ala are predicted accurately for most of the Raman and ROA bands, while those of cyclic L-Ala-L-Ala are reasonably well predicted. To assist in gaining reliable assignment of features, the Raman and ROA spectra of deuterated versions of these dipeptides were also investigated and are presented in Figure 5.4. The Raman and

ROA band assignments for both types of L-Ala-L-Ala are listed in Table 5.2. The experimental and calculated ROA spectra of the cyclic form of L-Ala-L-Ala are reported here for the first time.

ROA and Raman spectra of peptides are often considered in terms of three distinct

-1 regions: the backbone skeletal stretch region ~870-1150 cm which originates from

-1 Cα-C, Cα-Cβ, Cα-N stretch coordinates; the amide III region from ~1230-1340 cm in which bands are involved mainly from N-H in-plane deformations with Cα-N stretching and contributions from Cα-H deformations; and the amide I region ~1630-

1700 cm-1 which is associated mostly with C=O stretching [35]. The amide III region is very important for ROA study as the coupling between N-H and Cα-H deformation is very sensitive to geometry which provides information on ROA band structure

[35]. However, the spectra of cyclic dipeptides may contain larger features in other spectral regions than are typically observed for the corresponding linear forms; therefore here we discuss the features observed across a much broader spectral range.

A comparison of solution models in various isotopomers used in the calculation of the Raman and ROA spectra of linear L-Ala-L-Ala was reported by Jalkanen et al.

[20]. Most recently, the conformational changes in ROA spectra of isotopic substitution of C and N for linear L-Ala-L-Ala has been investigated by employing a

CPCM solvent model; however it did not bring a convincing improvement of ROA signals when matched up to standard dielectric solvent correlation [25]. We discuss

120

Chapter 5 the differences between the spectra of the two dipeptides, as well as the level of agreement between experiment and computation over the full spectral range measured, in detail below.

400-800 cm-1 region

In previous reports of Raman and ROA spectra of linear L-Ala-L-Ala [25] only a few peaks below 580 cm-1 were assigned while in the present study all of the vibrational mode assignments from 400-800 cm-1 are well predicted that enables a full vibrational mode analysis as shown in Table 5.1. Overall there is good agreement between the simulated and experimental spectra of both dipeptides despite the over prediction of ROA band intensities from 600-800 cm-1 for cyclic L-Ala-L-Ala which also corresponds to the over prediction of the corresponding Raman intensities.

The peaks in the 400-800 cm-1 region of the Raman and ROA spectra for both dipeptides are due to a variety of ring, C=O and N-H vibrations as well as to C-C bending modes. The –ve/+ve ROA couplet at 433/473 cm-1 does not correspond to any similar ROA band for linear L-Ala-L-Ala and is predicted as a +ve band in the calculation that arises from the DKP ring vibration. This is the first ROA signature of cyclic dipetide structures observed at low wavenumber.

In the experimental ROA spectrum of linear L-Ala-L-Ala two bands, one +ve and one –ve, can be observed at 458 and 580 cm-1, respectively, that are due to C-N and

C-C stretching vibrations. For cyclic L-Ala-L-Ala there are strong negative ROA peaks predicted at 678 and 616 cm-1 as well as a +ve peak at 597 cm-1 which correspond to either very weak or negligible signals in the experimental ROA spectrum of cyclic L-Ala-L-Ala. This may be explained by overlap of several of these oppositely signed bands leading to their cancellation. The Raman band

121

Chapter 5 observed at 688 cm-1 is assigned to a combination of C-C stretching, C-C-N and

C=O bending that appears stronger for the cyclic form, which is also predicted by the calculations, this band is shifted upwards by about 12 cm-1 upon deuteration for both linear and cyclic L-Ala-L-Ala which verifies that there is a similar degree of coupling in linear and cyclic forms of the C=O stretch vibration with N-H bending.

The calculated ROA bands at 473, 601, 720 and 790 cm-1 for the cyclic form are mainly due to the DKP ring and amide VI vibrations which are either very weak or negligible in the experimental ROA spectrum.

800-1400 cm-1 region

The relative intensities of cyclic and linear L-Ala-L-Ala bands in both experimental

Raman and ROA spectra are in excellent agreement with the calculated intensities.

Both simulated ROA spectra of linear and cyclic L-Ala-L-Ala demonstrate the best level of correlation with the corresponding experimental spectra in the region from

800-1400 cm-1. This is significant since the extended amide III region, from 1230-

1340 cm-1 is included which is a sensitive fingerprint of peptide conformation.

Therefore, differences in ROA spectra within this region are particularly useful for investigating structural differences.

Additionally, the most notable differences between the experimental Raman spectra of linear and cyclic L-Ala-L-Ala can be observed within this region. The observed

Raman bands for linear L-Ala-L-Ala are more intense in contrast to those for cyclic

L-Ala-L-Ala, for which the same sample concentration and conditions were applied.

The intense Raman peaks for linear L-Ala-L-Ala at ~885, 1281 and 1409 cm-1 that are mainly assigned to C-N, C-C and C-H stretching also appear as intense bands

-1 when dissolved in D2O, albeit with a 3 cm upward shift. The corresponding ROA

122

Chapter 5 bands are observed at 885, 1283 and 1417 cm-1 while these bands are not present in either the Raman or ROA spectra for the cyclic form. These latter vibrational modes are thus greatly changed by cyclization and formation of the second peptide linkage.

The observed intensities in the region of 1000-1400 cm-1 appear to be sensitive to the structural changes induced by cyclization in both Raman and ROA spectra as they reduce by a factor of two for the cyclic form. The experimental ROA features for linear L-Ala-L-Ala at 1025, 1078, 1097, 1121 and 1384 cm-1 reverse their signs for the cyclic form. These bands are mainly assigned to CH3, C-C and C-N stretching mode which indicates that stereochemical changes induced by cyclization for these vibrational modes are indeed present. The only corresponding Raman bands for cyclic and linear L-Ala-L-Ala appear at 1105 and 1097 cm-1. These sign changes in band appearance illustrate the influence of cyclization on the vibrational coordinates of the peptide group and emphasize that ROA is more sensitive than Raman spectroscopy for studying the effects of cyclization upon conformation. Some features in the experimental ROA spectra of linear and cyclic L-Ala-L-Ala are not predicted by the computed spectra e.g. the bands at 1078 and 1327 cm-1 for linear L-

Ala-L-Ala and 925 and 1080 cm-1 for the cyclic form. Nevertheless, there is a good match between both calculated and experimental Raman and ROA spectra of cyclic and linear L-Ala-L-Ala apart from the slight under prediction of ROA bands in the region from 1000- 1150 cm-1, which mainly originate from C-C, C-H and C-N vibrational modes. The different appearance of both Raman and ROA spectra for cyclic and linear L-Ala-L-Ala in both the backbone skeletal stretch region ~ 1050-

1200 cm-1 and the extended amide III region ~1250-1350 cm-1 suggests that the backbone conformation is very different in these two dialanines.

123

Chapter 5

1400- 1800 cm-1 region

A number of spectral differences also occur within the high wavenumber region as a consequence of cyclization. The positive ROA peak for linear L-Ala-L-Ala at 1417 cm-1 does not correspond to any peak in the spectra of the cyclic form in either the hydrated or deuterated states. The ROA band due to CH3 bending vibrations, which appears as a sharp and unusually intense negative peak at 1452 cm-1 for cyclic L-Ala-

L-Ala and flips its sign to a positive and smaller peak at 1457 cm-1 for linear L-Ala-

-1 L-Ala in H2O, is up shifted by 5 cm in D2O for the linear form but remains at the same position for the cyclic form. The amide I Raman band at ~1650 cm-1 and amide

II mode at ~1550 cm-1 materialize in this region and two bands are present in the experimental Raman spectra at 1524 and 1661 cm-1 for cyclic L-Ala-L-Ala but only at 1642 cm-1 for linear L-Ala-L-Ala. The corresponding Raman peak for 1661 cm-1

-1 in the cyclic form appears as two distinct bands at 1594 and 1668 cm in D2O. There are no corresponding Raman or ROA bands around 1524 cm-1 for the linear form which postulates that this is a marker band for cyclic L-Ala-L-Ala. In the previous study by Bour et al. [21] the amide II mode remained unobserved in the computed spectra of both normal and labelled linear L-Ala-L-Ala whereas, in our present study the Raman and ROA amide II bands are successfully simulated. The computed

Raman and ROA spectra for both linear and cyclic L-Ala-L-Ala show a similar peak at 1656 cm-1 in the amide I region, originating from the C=O stretching mode, but such a feature is present only in the experimental ROA spectrum of the linear form.

Interestingly, no such amide I ROA band appears for the cyclic form, possibly because the two carbonyl groups have opposing orientation so that their individual

ROA signatures effectively cancel each other. More ROA bands can be observed experimentally for this region in contrast with the Raman spectra which confirms the

124

Chapter 5 enhanced sensitivity of ROA over Raman spectra to the structural changes upon cyclization for L-Ala-L-Ala. We also note that the observed bands in the ROA spectrum of cyclic L-Ala-L-Ala in this region differ to those of the linear form that is found for the other spectral regions. In addition, there is no ROA signal observed in the amide II region for linear L-Ala-L-Ala, whereas there are small but clear +ve amide II bands at 1477 and 1521cm-1 for the cyclic form.

5.6 Conclusion

Cyclization is an important structural modification of biological peptides, as is shown by the increasing number of cyclic peptides that are now being discovered to perform physiological function, but the conformational constraints imposed by this process are not well understood. This study has identified several Raman and ROA marker bands that identify cyclization in dialanine. ROA spectra, being sensitive to stereochemistry, are found to be more sensitive to cyclization with ROA marker bands being observed at 433, 790, 845, 987, 1067, 1304, 1322, 1407, 1452, 1521 and

1541 cm-1. Although less sensitive, Raman spectra also appear to contain bands that identify cyclization, these principally being 989, 1183, 1317, 1459 and 1524 cm-1.

Even though explicit solvation was not utilized in the calculations of the Raman and

ROA spectra, we obtained a good level of agreement with experimental data that has allowed the vibrational modes responsible for the spectral features to be identified.

The combination of ROA spectroscopy with computational modelling clearly provides an incisive tool for characterizing cyclic peptides.

125

Chapter 5

Figure 5.1: The chemical structure with atom numbering scheme (left) and calculate minimum energy conformation (right) of linear L-Ala-L-Ala.

H5 H12 H2 O H17 14 H18 15 H C C H20 3 10 N C13 C4 1 C C N 8 11 6 H16 H19 O H 7 9

Figure 5.2: The chemical structure with atom numbering scheme (left) and calculate minimum energy conformation (right) of cyclic L-Ala-L-Ala.

126

Chapter 5

Exp. Linear L-Ala-L-Ala

9 Calc. Linear L-Ala-L-Ala 4.5x10

Exp. Cyclic L-Ala-L-Ala

+ I + R I

2x109

Calc. Cyclic L-Ala-L-Ala

400 600 800 1000 1200 1400 1600 1800

Exp. Linear L-Ala-L-Ala

1.7x106 Calc. Linear L-Ala-L-Ala

L Exp. Cyclic L-Ala-L-Ala

- I - R

5.6x105 Calc. Cyclic L-Ala-L-Ala

400 600 800 1000 1200 1400 1600 1800 -1

Wavenumber (cm )

Figure 5.3: Experimental and computed Raman (top) and ROA (bottom) spectra of linear (pH= 7.0) and cyclic L-Ala-L-Ala (pH= 7.0) in aqueous solution. The concentration for each sample was 50 mg/ml and laser power was 0.6 W at the sample. The marker bands that are induced upon cyclization are highlighted by shading.

127

Chapter 5

Exp. Linear L-Ala-L-Ala

5x109

Calc. Linear L-Ala-L-Ala

+ I + Exp. Cyclic L-Ala-L-Ala

R I 5x109

Calc. Cyclic L-Ala-L-Ala

400 600 800 1000 1200 1400 1600 1800

Exp. Linear L-Ala-L-Ala

2x106 Calc. Linear L-Ala-L-Ala

Exp. Cyclic L-Ala-L-Ala

- I -

R I 1x106 Calc. Cyclic L-Ala-L-Ala

400 600 800 1000 1200 1400 1600 1800 Wavenumber cm-1

Figure 5.4: Experimental and computed Raman (top) and ROA (bottom) spectra of linear (pH= 7.0) and cyclic L-Ala-L-Ala (pH= 7.0) in D2O. The concentration for each sample was 50 mg/ml and laser power was 0.6 W at the sample. The marker bands that are induced upon cyclization are highlighted by shading.

128

Chapter 5

5.7 References:

1. S. C. Brauns, P. Milne, R. Naude, M. Van de Venter, Anticancer Research. 2004, 24, 1713-1719. 2. P. J. Milne, A. L. Hunt, K. Rostoll, J. J. Van der Walt, C. J. M. Graz, Journal of Pharmacy and Pharmacology. 1998, 50, 1331-1337. 3. C. B. Unal, M. D. Owen, W. R. Millington, Brain Research. 1997, 47, 52-59. 4. V. Crescenzi, A. Cesaro, E. Russo, International Journal of Peptide and Protein Research. 1973, 5, 427-434. 5. M. J. O. Anteunis, Bulletin des Sociétés Chimiques Belges. 1978, 87, 627- 650. 6. K. McCleland, P. J. Milne, F. R. Lucieto, C. Frost, S. C. Brauns, M. V. D. Venter, J. Du Plessis, K. Dyason, Journal of Pharmacy and Pharmacology. 2004, 56, 1143-1153. 7. S. W. Yang, T. M. Chan, J. Terracciano, D. Loebenberg, G. D. Chen, M. Petal, V. Gullo, B. Pramani, M. Chu, Journal of Antibiotics. 2004, 57, 345- 347. 8. S. C. Brauns, G. Dealtry, P. Milne, R. Naude, M. Van De Venter, Anticancer Research. 2005, 25, 4197-4202. 9. D. B. Davies, M. A. Khalad, Journal of the Chemical Society, Perkin Transactions 2. 1976, 11, 1238-1244. 10. K. D. Kopple, V. Narutis, International Journal of Peptide and Protein Research. 1981, 18, 33-40. 11. R. L. Bowman, M. Kellerman, W. C. Johnson, Biopolymers. 1983, 22, 1045- 1070. 12. F. L. Bettens, R. P. A. Bettens, R. D. Brown, P. D. Godfrey, Journal of the American chemical society. 2000, 122, 5856-5860. 13. Y. Zhu, M. Tang, X. Shi, Y. Zhao, International Journal of Quantum Chemistry. 2007, 107, 745-753. 14. T. C. Cheam , S. Krimm, Spectrochimica Acta Part A: Molecular Spectroscopy. 1984, 40, 481-501. 15. T. C. Cheam, S. Krimm, Spectrochimica Acta Part A: Molecular Spectroscopy. 1988, 44, 185-208. 16. L. D. Barron, Molecular Light Scattering and Optical Activity. Cambridge University Press. 2004. 17. F. Zhu, N. W. Isaacs, L. Hecht, L. D. Barron, Structure. 2005, 13, 1409-1419. 18. T. A. Keiderling, Current Opinion in Chemical Biology. 2002, 6, 682-688. 19. S. J. Ford, Z. Q. Wen, L. Hecht, L. D. Barron, Biopolymers. 1994, 34, 303- 313. 20. K. J. Jalkanen, R. M. Nieminen, M. Knapp-Mohammady, S. Suhai, International Journal of Quantum Chemistry. 2003, 92, 239-259. 21. P. Bour, J. Kapitan, V. R. Baumruk, Journal of Physical Chemistry A. 2001, 105, 6362-6368. 22. M. Knapp-Mohammady, K. J. Jalkanen, F. Nardi, R. C. Wade, S. Suhai, Chemical Physics. 1999, 240, 63-77. 23. A. F. Weir, A. H. Lowrey, R. W. Williams, Biopolymers. 2001, 58, 577-591. 24. P. Mukhopadhyay, G. R. Zuber, D. N. Beratan, Biophysical Journal. 2008, 95, 5574-5586. 25. J. Sebek, J. Kapitan, J. Sebestik, V. Baumruk, P. Bour, The Journal of Physical Chemistry A. 2009, 113, 7760-7768.

129

Chapter 5

26. A. P. Mendham, T. J. Dines, M. J. Snowden, B. Z. Chowdhry, R. Withnall, Journal of Raman Spectroscopy. 2009, 40, 1478-1497. 27. Gaussian 09, Revision A.1, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox, Gaussian, Inc., Wallingford CT, 2009. 28. A. D. Becke, Journal of Chemical Physics. 1993, 98, 5648-5652 29. C. Lee, W. Yang, R. G. Parr, Physical Review B. 1988, 37, 785-789. 30. T. H. Dunning, Journal of Chemical Physics. 1989, 90, 1007-1021. 31. M. T. Cances, V. Mennucci, J. Tomasi, Journal of Chemical Physics. 1997, 107, 3032-3041. 32. D. M. York, M. Karplus, Journal of Physical chemistry A. 1999, 103, 11060- 11079. 33. R. J. Fletterick, C. Tsai, R. E. Hughes, Journal of physical chemistry. 1971, 75, 918-922. 34. J. A. Schachtschneider, Vibrational Analysis of Polyatomic Molecules, Parts V and VI, Technical Report Nos. 231 and 57, Shell Development Co., Houston TX, 1964 and 1965. 35. L. D. Barron, Current Opinion in Structural Biology. 2006, 16, 638-643.

130

Chapter 5

Table 5.1: Calculated and experimental wavenumber band assignments for Raman and ROA of cyclic and linear L-Ala-L-Ala in H2O Column1 Column2 Column3 Column4 Column5 Column6 Column7 Column8 Experimental Calculated Band assignments Linear Cyclic Linear Cyclic Linear Cyclic ROA Raman ROA Raman

- - - - 402 - (N5C4C10) (41), ip(CO) (11) - - 433(-ve) - - 421 A - ip(ring-1) (12), ip(ring-2) (58) 458 (+ve) 458 - - 454 456 B (N9C8C13) (57) (CCH) (21),ip(CO) (29),op(CO) (15) - 473 473 (+ve) 473 - - - -

- - - 485 484 A - (NC) (18), ip(ring-1) (43), ip(ring-2) (12) 550 (+ve) - - - 561 - op(NH) (36),(N5C6) (44) - 580 (-ve) - - - 577 - (C1C4) (10),(N5C6C8) (17), (CO2) (13),(CO2) (14) ------597 A - (NC) (18), ip(ring-1) (43), ip(ring-2) (12) - - 601 (-ve) - - 600 B - op(NH) (84) [amide V] 616 (+ve) - - 616 - 613 A - (CC) (31), ip(ring-1) (13), ip(CO) (14), op(NH) (35) 658 (-ve) - - - 660 - (N5C4C1) (13), ip(CO) (12),(CO2) (23),(CO2) (10) - - - - - 678 678 A (C6C8) (17),(C6C8N9) (13), ip(CO) (20),(CO2) (10) (NC) (10), (CC) (21), op(NH) (22), op(CO) (25) 699 (+ve) 688 - 688 - 684 B - (C3C4) (17), ip(ring-3) (32), op(CO) (37) [amide VI] - - 720 (-ve) - 760 - op(CO) (29),(CO2) (37) - 776 (-ve) 782 - 761 767 - op(CO) (33),(CO2) (25) ------780 B - (CC) (10), ip(ring-3) (25), op(CO) (25) - - 790 (+ve) 793 - 786 A - op(CO) (47), ip(CH3) (18) [amide VI] 834 (+ve) - - - 840 - s(CO2) (10),(C1C4) (19), (CO2) (22) - 859 (-ve) - 845 (-ve) 844 - 844 B - (C3C4) (11),(CC) (37), op(CH3) (12) 885 (+ve) 885 - - 878 - (C8N9) (24), (C8C13) (16) -

905 (-ve) 928 - 925 914 - (C4N5) (10),ip(CH3) (18),op(CH3) (18) - 931 (+ve) - - - - 963 B - (NC) (18), ip(ring-3) (16), ip(CH3) (13), op(CH3) (25)

953 (-ve) 956 953 (-ve) - 956 - (C4C10) (13), (C6C8) (19),op(CH3′) (15) - 964 (+ve) 995 ------

984 (+ve) 1012 987 (+ve) 989 994 1001 A (C8N9) (11), (C8C13) (30),ip(NH3) (30), op(CH3′) (16) (NC) (10), op(CH3) (54) 1009 (-ve) ------1025(+ve) - 1023 (-ve) - 1019 - op(NH3) (36), ip(CH3′) (34) ------1035 A - (C3C4) (31),ip(CH3) (20)

131

Chapter 5

1053 (-ve) 1050 1067 (+ve) - 1051 1061 B (C4N5) (13),(C4C10) (15),′(C4H11) (14),op(CH3) (20) (CCH) (12),ip(CH3) (26), op(CH3) (16) 1078(+ve) - 1080 (-ve) - - - - - 1097 (-ve) 1105 1094 (+ve) 1097 1109 1109 A (C4C10) (27), ip(CH3) (17), op(CH3′) (10) (C3C4) (39),(CCH) (11),ip(CH3) (33) 1121(+ve) - 1129 (-ve) - 1120 1129 B (C8N9) (17), (C8C13) (13),ip(NH3) (18) (C3C4) (46),ip(CH3) (17) - - - - 1138 - op(NH3) (10),′(C8H14) (15), op(CH3′) (16) - - - 1156 (+ve) - - 1160 B  (NC) (28), (CCH) (20), op(CH3) (24) 1175 (-ve) 1172 - - 1163 - (C4N5) (25),(N5C4C10) (11), op(CH3) (20) - - - - 1186 - 1194 A - (NC) (31), (CCH) (21), op(CH3) (13) 1247(+ve) - - - 1233 - op(NH3) (27), (N9C8C13) (10), ip(CH3′) (16) - 1270 (-ve) - - - 1280 - (C6N5) (27), ip(NH) (29),′(C8H14) (11) - 1283(+ve) 1281 - - 1286 1294 A ′(C4H11) (72), ip(CH3) (10) (NC) (17), (CCH) (41),(CCH) (17) - - 1304 (+ve) 1317 - 1314 B - (CCH) (49),(CCH) (20) - 1327 1322 (-ve) - - 1322 B - (CCH) (57),(CCH) (14) 1343(+ve) 1343 - - 1335 1332 A (C4H11) (65),′(C8H14) (13) (CCH) (64),(CCH) (16) - - 1351 (+ve) - 1365 - (C8H14) (14),′(C8H14) (32) -

- 1374 - - 1369 - s(CO2) (17),s(CH3) (59) - 1384 (-ve) - 1389 (+ve) 1386 - 1388 A,B - s(CH3) (84) (91) - - - - - 1395 A - (NC) (13), ip(NH) (40),s(CH3) (12) [amide III] 1399 (-ve) - 1407 (+ve) - 1403 - s(CO2) (40),s(CH3) (34) - 1417(+ve) 1409 - - 1410 - s(CH3′) (88) - 1437 (-ve) - 1437 (+ve) - 1428 1436 B (C8H14) (45),s(CH3′) (11), as(CH3′) (21) (NC) (31), (CC) (10),ip(NH) (15),as(CH3) (26) - - - - - 1452 B - (CO) (10), (NC) (22), ip(NH) (11),as(CH3) (41) [amide III] 1457(+ve) - - - 1456 1454 A as(CH3) (89), ip(CH3) (10) as(CH3) (83) - - 1452 (-ve) 1459 - 1459 B - as(CH3) (70) - 1469 - - 1460 1460 A as(CH3) (84) as(CH3) (86) 1472 (-ve) - - - 1471 - as(CH3′) (82), ip(CH3′) (13) - - - 1477 (+ve) - 1474 1477 B (C8H14) (16), as(CH3′) (62), op(CH3′) (10) (CO) (14), ip(NH) (36), as(CH3) (24) [amide II] - - 1492 (+ve) - 1495 - s(NH3) (90) - - 1521 (+ve) 1524 - 1521 A - (CO) (17), (NC) (22), (CC) (12), ip(NH) (20) [amide II] 1568(+ve) - 1541 (-ve) - 1546 - as(CO2) (91) - - - 1553 (+ve) - 1589 - (C6N5) (30), ip(NH) (63) - 1628 (-ve) - 1567 (+ve) - 1612 - as(NH3) (74) - - 1642 1640 (+ve) - 1650 1652 B (CO) (33), as(NH3) (45) (CO) (60), (NC) (10), ip(NH) (15) [amide I] 1680(+ve) - 1654 (-ve) 1661 1656 1656 A (CO) (32), as(NH3) (47) (CO) (61), (NC) (12), ip(NH) (15) [amide I]

132

Chapter 5

Table 5.2: Calculated and experimental wavenumber band assignments for Raman and ROA of cyclic and linear L-Ala-L-Ala in D2O.

Experimental Calculated Band assignments Linear Cyclic Linear Cyclic Linear Cyclic ROA Raman ROA Raman

- 436 - - 431 420 B (N9C8C13) (52) ip(CO) (13), op(ND) (73) [amide V] 445 (-ve) - - - 443 - (N5C4C10) (18),op(NH) (36),(N5C6) (20) - 470 (-ve) - 461 (+ve) - - 473 A - (NC) (16), (CCH) (21), ip(ring-1) (13), ip(ring-2) (17),op(ND) (30) - - - 482 - 486 B - (CCH) (18),ip(CO) (20), op(ND) (26),op(CO) (14)

495 (-ve) - 504 (+ve) - - 498 A - ip(ring-1) (34), op(NH) (46) [amide V] 550 (+ve) - - - 558 - (C1C4) (10),(N5C6C8) (19), (CO2) (10),(CO2) (20) - - 571 ------

- - 601 (+ve) 607 - 590 A - (CC) (21), ip(CO) (50) [amide IV] 613 (+ve) 631 622 (+ve) - 629 - (C6C8) (24),ip(ND3) (15), ip(CO) (29)  - 640 (+ve) - - - - 661 A - (NC) (10), (NC) (10), (CC) (29), op(CO) (26) - - - 676 661 678 B (N5C4C1) (21), (CO2) (36) (C3C4) (19), op(CO) (36), ip(ring-3) (31) [amide VI] - 694 - - - -

738 (+ve) 755 729 (+ve) - 754 757 B (C6C8C13) (13), op(CO) (60) (NC) (11), (CC) (25), ip(ND) (15), ip(ring-3) (20) - - 770 (+ve) 767 763 - (C1C4C10) (11),(CO2) (56) - - 782 - - - 786 A - op(CO) (47), ip(CH3) (18) [amide VI] - - - - 811 812 B (C8N9) (12),(C8C13) (26), ip(ND3) (14), op(ND3) (33) (CC) (16), ip(ND) (11), op(CO) (24), ip(ring-3) (10),ip(CH3) (13)

825 (-ve) 828 822 (-ve) - 833 - ip(ND3) (29), (CO2) (12) - - - - - 856 - (C1C4) (13), ip(ND3) (11), op(ND3) (10), ip(CH3) (13), (CO2) (10) - 877 (+ve) 877 885 (+ve) 882 877 - (C8N9) (19),op(ND3) (19),ip(CH3′) (23) - 916 (+ve) 911 919 (+ve) - 908 917 B (C4N5) (13), ip(CH3) (16),op(CH3) (19) (NC) (28), ip(ND) (15),op(CH3) (36) 975 (+ve) 945 - 942 950 931 A (C8N9) (11), (C6C8) (19),op(CH3′) (20) (NC) (19), ip(ND) (41),op(CH3) (30) [amide III]

1012 (+ve) 1003 989 (+ve) 992 1012 1017 B (C4C10) (19),ip(ND) (66) (C3C4) (25), ip(ND) (14), ip(ring-3) (13), ip(CH3) (25) 1048 (-ve) - - 1037 1049 1028 A (C4N5) (12), (C4C10) (11),′(C4H11) (14), ip(CH3) (12), op(CH3) (17)(C3C4) (35), (CCH) (10), ip(CH3) (21) 1067 (+ve) 1067 1072 (+ve) - 1078 1083 A (C8C13) (23),(C6C8N9) (10) ip(ND) (21), (CCH) (10),op(CH3) (38) 1116(+ve) - 1091 (-ve) 1094 1108 1101 B s(ND3) (12), ip(CH3′) (11), op(CH3′) (28) (C3C4) (20),ip(CH3) (22), op(CH3) (16) - 1100 1113 (+ve) - 1126 1118 A (C4N5) (11), (C4C10) (18) (C3C4) (34),ip(CH3) (29) 1159 (-ve) - - - 1145 - (C8N9) (13), (C8C13) (15),s(ND3) (20), as(ND3) (21) - - - - 1151 1156 1151 B (C4N5) (14),(N5C4C10) (16), ip(CH3) (15), op(CH3) (20) (C3C4) (10),(NC) (11), (CCH) (18), ip(CH3) (14), op(CH3) (24)

133

Chapter 5

- - - - 1165 - s(ND3) (26), as(ND3) (33) - - - - - 1184 - as(ND3) (95) - - 1204 - 1207 1194 - (C8N9) (10),s(ND3) (18), as(ND3) (10), ip(CH3′) (10) ------1229 B - (NC) (25), ip(ND) (26), (CCH) (10),ip(CO) (16) [amide III]

- - 1244 (-ve) - - 1230 A - (NC) (20), ip(ND) (10), (CCH) (12), (CCH) (20) 1281 (+ve) 1281 - - 1286 - ′(C4H11) (73) -

- 1312 1304 (+ve) - - 1309 A - (NC) (18), (CCH) (33),(CCH) (11) - - - - - 1319 B - as(CH3) (20), (CCH) (36),(CCH) (19) 1330 (+ve) 1333 - 1322 1329 1335 A (C4H11) (38),′(C8H14) (37) (CCH) (60),(CCH) (16)

1358 (-ve) 1358 1340 (+ve) - 1351 1347 B (C4H11) (35),′(C8H14) (30) as(CH3) (36), (CCH) (12),(CCH) (10) 1379 (+ve) 1371 1384 (+ve) 1386 1369 1388 B s(CO2) (20),s(CH3) (62) s(CH3) (94) - - - - - 1389 A - s(CH3) (96) - - - - 1400 - s(CO2) (25),(C8H14) (31),s(CH3) (27) - - - - - 1407 - s(CO2) (14),(C8H14) (44) - 1412 (+ve) 1412 - - 1411 - s(CH3) (93) ------1441 B - (NC) (54) [amide II]

- - 1452 (+ve) - 1456 1452 A as(CH3) (90), ip(CH3) (10) as(CH3) (76) - - 1457 1457 1458 A, B as(CH3) (48),as(CH3′) (29) as(CH3) (88),as(CH3) (89) 1462 (-ve) 1462 - - 1463 1461 B as(CH3) (35),as(CH3′) (39) as(CH3) (80) 1487 (+ve) 1487 1479 (+ve) - 1473 - as(CH3′) (81), ip(CH3′) (13) - - - 1502 (+ve) 1507 1502 1495 A (C6N5) (32), (C6C8) (11), ip(ND) (11) (NC) (35), (CC) (14) [amide II] - 1556 - 1546 - as(CO2) (91) - - 1594 - - 1634 B - (CO) (72), (NC) (10) [amide I] 1668 (+ve) 1668 1647 (+ve) 1649 1640 1637 A (CO) (72), (C6N5) (16) (CO) (75), (NC) (10) [amide I]

134

Chapter 5

5.8 Supplementary information

Table S5.1: Calculated and experimental bond lengths (Å) for cyclic and linear L-

Ala-L-Ala.

Cyclic L-Ala-L-Ala Calculated X-raya

(N1C13) 1.3429 1.331 r(N1H2) 1.0147 0.926 r(C3N1) 1.4662 1.454 r(C3C4) 1.5281 1.514 r(C4H15) 1.0975 0.964 r(C4H16) 1.0976 0.963 r(H5C3) 1.1032 0.977 r(C6C3) 1.5249 1.518 r(O7C6) 1.244 1.239 r(C4H17) 1.0946 1.035

Linear L-Ala-L-Ala Calculated X-raya

r(C1O2) 1.2661 1.2422 r(C1C4) 1.5533 1.5409 r(O3C1) 1.2628 1.2279 r(C4H11) 1.0947 0.8961 r(N5C6) 1.3381 1.3462 r(N5C4) 1.4608 1.4561 r(C6O7) 1.2445 1.2252 r(C8N9) 1.5007 1.4953 r(C8C6) 1.5351 1.5306 r(N9H17) 1.0229 0.8721 r(N9H16) 1.0269 0.7652 r(C10H20) 1.096 - r(C10H19) 1.0975 0.8983 r(C10C4) 1.5345 1.5181 r(H12N5) 1.0127 0.9433 r(C13C8) 1.5285 1.5231 r(C13H22) 1.097 1.1875 r(H14C8) 1.0934 0.9581 r(H15N9) 1.0223 0.866 r(H18C10) 1.0992 1.131 r(H21C13) 1.0966 1.0212 r(H23C13) 1.0956 0.8012

135

Chapter 5

Table S5.2: Calculated and experimental bond angles (o) for cyclic and linear L-Ala-

L-Ala.

Cyclic L-Ala-L-Ala Calculated Experimental θ(C13N1H2) 115.69 111.35 θ(C13N1C3) 126.49 126.16 θ(N1C13O14) 122.29 122.53 θ(H2N1C3) 116.57 121.82 θ(N1C3C4) 109.82 109.65 θ(N1C3H5) 109.05 107.41 θ(N1C3C6) 111.29 110.57 θ(C3C4H15) 109.65 112.88 θ(C3C4H16) 110.75 109.5 θ(C4C3H5) 108.71 110.51 θ(C4C3C6) 112.26 112.12 θ(C3C4H17) 110.05 108.11 θ(H15C4H16) 108.95 112.58 θ(H15C4H17) 108.28 109.06 θ(H16C4H17) 109.13 104.265 θ(H5C3C6) 105.55 106.11 θ(C3C6N8) 116.78 110.53 θ(C3C6O7) 120.93 120.56

Linear L-Ala-L-Ala

θ(O2C1C4) 115.45 118.16 θ(O2C1O3) 125.7 123.86 θ(C4C1O3) 118.84 117.98 θ(C1C4H11) 107.29 103.74 θ(C1C4N5) 112.8 109.16 θ(C1C4C10) 110.59 110.67 θ(H11C4N5) 106.9 111.18 θ(H11C4C10) 109.38 111.77 θ(C6N5C4) 124.31 122.88 θ(N5C6O7) 124.82 125.93 θ(N5C6C8) 115.49 113.16 θ(C6N5H12) 117.72 119.19 θ(N5C4C10) 109.75 110.14 θ(C4N5H12) 117.54 117.83 θ(O7C6C8) 119.68 120.89 θ(N9C8C6) 106.53 108.33 θ(C8N9H17) 111.75 110.23 θ(C8N9H16) 109.12 110.46 θ(N9C8C13) 110.16 107.45 θ(N9C8H14) 107 108.46 θ(C8N9H15) 111.87 112.33 θ(C6C8C13) 111.25 111 θ(C6C8H14) 110.88 109.28 θ(H17N9H16) 107.24 102.91 θ(H17N9H15) 107.81 113.92 θ(H16N9H15) 108.91 106.49 θ(H20C10H19) 108.63 - θ(H20C10C4) 109.91 - θ(H20C10H18) 108.44 - θ(H19C10C4) 110.29 117.14 θ(H19C10H18) 108.66 120.6

136

Chapter 5

θ(C4C10H18) 110.85 104.99 θ(C8C13H22) 110.8 107.07 θ(C13C8H14) 110.85 112.21 θ(C8C13H21) 110.42 108.22 θ(C8C13H23) 109.19 110.12 θ(H22C13H21) 109.21 117.46 θ(H22C13H23) 108.73 105.4 θ(H21C13H23) 108.43 108.44

Table S5.3: Calculated and experimental torsion angles (o) for cyclic and linear L-

Ala-L-Ala.

Cyclic L-Ala-L-Ala  Calculated Experimental (C10C13N1H2)  175.65 - (O14C13N1H2)  -3.77 - (C10C13N1C3)  8.96 - (O14C13N1C3)  -170.46 - (C13N1C3C4)  -154.91 - (C13N1C3H5)  86.06 - (C13N1C3C6)  -29.98 - (N1C13C10N8)  17.2555 - (N1C13C10C11)  143.63 - (N1C13C10H12)  -98.09 - (H2N1C3C4)  38.5 - (H2N1C3H5)  -80.53 - (H2N1C3C6)  163.44 - (N1C3C4H15)  -56.84 - (N1C3C4H16)  63.41 - (N1C3C4H17)  -175.84 - (N1C3C6N8)  20.08 - (N1C3C6O7)  -160.5 - (H5C3C4H15)  62.39 - (C6C3C4H15)  178.78 - (H5C3C4H16)  -177.35 - (C6C3C4H16)  -60.96 - (H17C4C3H5)  -56.61 - (H17C4C3C6)  59.78 - (C4C3C6O7)  -36.94 - (H5C3C6O7)  81.33 -

Linear L-Ala-L-Ala

τ (O2C1C4H11) 45.15 164.66 τ (O2C1C4N5) 162.6 -76.74 τ (O2C1C4C10) -74.07 44.64 τ (H11C4C1O3) -136.17 -15.62 τ (N5C4C1O3) -18.72 102.98 τ (C10C4C1O3) 104.61 -135.64 τ (C1C4N5C6) -92.69 -112.98 τ (C1C4N5H12) 79.58 70.86 τ (C1C4C10H20) 56.62 - τ (C1C4C10H19) 176.36 167.9 τ (C1C4C10H18) -63.25 -55.26 τ (H11C4N5C6) 25 0.87

137

Chapter 5

τ (H11C4N5H12) -162.74 -175.3 τ (H11C4C10H20) -61.34 - τ (H11C4C10H19) 58.41 52.81 τ (H11C4C10H18) 178.8 -170.35 τ (O7C6N5C4) -4.83 -2.74 τ (C8C6N5C4) 173.7 175.73 τ (C6N5C4C10) 143.52 125.32 τ (H12N5C6O7) -177.08 173.37 τ (H12N5C6C8) 1.45 -8.16 τ (N5C6C8N9) 151.25 165.4 τ (N5C6C8C13) -88.67 -76.86 τ (N5C6C8H14) 35.18 47.41 τ (H12N5C4C10) -44.22 -50.84 τ (N5C4C10H20) -178.31 - τ (N5C4C10H19) -58.56 -71.3 τ (N5C4C10H18) 61.82 65.54 τ (O7C6C8N9) -30.15 -16.05 τ (O7C6C8C13) 89.93 101.69 τ (O7C6C8H14) -146.22 -134.03 τ (H17N9C8C6) -75.59 -56.2 τ (H16N9C8C6) 42.84 56.86 τ (H15N9C8C6) 163.42 175.59 τ (C13C8N9H17) 163.63 -176.19 τ (H14C8N9H17) 43.06 62.3 τ (C13C8N9H16) -77.94 -63.13 τ (H14C8N9H16) 161.49 175.36 τ (H15N9C8C13) 42.64 55.61 τ (N9C8C13H22) 57.89 62.08 τ (N9C8C13H21) -63.25 -65.44 τ (N9C8C13H23) 177.63 176.19 τ (H15N9C8H14) -77.93 -65.9 τ (C6C8C13H22) -60.01 -56.19 τ (C6C8C13H21) 178.85 176.29 τ (C6C8C13H23) 59.72 57.91 τ (H14C8C13H22) 176.13 -178.79 τ (H21C13C8H14) 54.99 53.69 τ (H23C13C8H14) -64.14 -64.68

138

Chapter 6

Combined Experimental and Computational Study of Raman and Raman Optical Activity (ROA) Spectra of Linear and Cyclic L-Ser-L- Ser in Solution

139

Chapter 6

6.0 Declaration

This chapter consists of one draft paper awaiting submission to Journal of Raman

Spectroscopy: S. Ostovar pour 1*, T. J. Dines 2, C. Levene 1, B.Z. Chowdhry 3 and

E.W. Blanch1, Journal of Raman spectroscopy. 2012

The manuscripts have been incorporated in a format identical to that for journal submission, except for minor adjustments to incorporate them into this thesis. As first author on this publication, I carried out all of the associated experimental and spectroscopic analysis. The calculation were carried out by T. J. Dines and B. Z.

Chowdhry and provided here for purpose of comparison with experimental results.

140

Chapter 6

Combined Experimental and Computational Study of

Raman and Raman Optical Activity (ROA) Spectra of

Linear and Cyclic L-Ser-L-Ser in Solution

S. Ostovar pour, 1 T. J. Dines, 2 C. Levene, 1 B.Z. Chowdhry3 and E.W. Blanch1*

1 Faculty of Life Sciences, University of Manchester, MIB 131 Princess Street, Manchester, M1 7DN,

2 Division of Electronic Engineering and Physics, University of Dundee, Dundee, DD1 4HN, UK

3 School of Science, University of Greenwich at Medway, Central Avenue, Chatham Maritime, Kent,

ME4 4TB, UK

* Corresponding author: [email protected]; Fax: +44 (0)161 236 0409; Tel: +44 (0)161 306

5819

6.1 Abstract

A study of the conformations of cyclic and linear L-Ser-L-Ser in aqueous solution has been carried out using Raman and Raman optical activity (ROA) spectroscopies and quantum mechanical calculations. Raman and ROA spectra of linear and cyclic

L-Ser-L-Ser in H2O and D2O were measured and assignment of the observed bands has been proposed from DFT calculations at the B3LYP/cc-pVDZ level assuming C2 symmetry in both cases. We found that ROA spectra are more sensitive than Raman spectra to the structural changes induced by cyclization. Specifically, Raman bands at 670, 1164 and 1317 cm-1 along with ROA bands at 682, 716, 782, 894, 1064,

1080, 1154, 1345, 1482 and 1683 cm-1 have been identified as marker bands for the cyclic form. In addition, we observe an unusually intense amide II band in both the

Raman and ROA spectra at 1519 cm-1 for cyclic L-Ser-L-Ser that is not present in the spectra of the linear form.

141

Chapter 6

6.2 Introduction

The cyclization of peptides has recently been the subject of a number of studies in peptide chemistry in relation to their biological activity [1,2]. Their potential application as antibiotic and anticancer drugs has been highlighted by several publications [3-5]. Appropriate pharmaceutical delivery systems for these molecules have not been fully understood untill recent times owing to the lack of available information concerning their structural behaviour [3-5]. Whilst being the subject of intrinsic interest and significant for the study of biologically active short dipeptides, they also appear to be good model compounds for analysis of certain properties such as secondary structure of similar residues within polypeptide chains [6], since modelling of longer peptides at a similar level of accuracy would be very difficult due to the high computational requirements.

Raman and Raman optical activity (ROA) spectroscopies have been previously utilized for the determination of peptide conformation in aqueous environments.

ROA spectroscopy measures an intensity difference in the Raman scattering of right- and left-circularly polarized light from chiral molecules [7-10]. ROA spectroscopy is particularly sensitive to conformational changes, thereby enabling characterization of the natural behaviour of biological molecules in solution. However, the relationship between the ROA spectrum and the molecular conformation is complex, therefore theoretical simulations of ROA spectra are very useful for providing a more detailed interpretation of the vibrational modes measured [11,12].

The combination of quantum mechanics/molecular mechanics modelling with experimental ROA spectroscopy provides a uniquely sensitive tool for investigating the conformations of unusual peptides. Previous X-ray, NMR, CD and IR studies have been used to characterize the structure of cyclic L-Ser-L-Ser [13]. The Raman

142

Chapter 6

[14-15] and ROA [16] spectra of L-serine have been simulated using DFT and the

B3LYP hybrid functional in several studies in which the influence of side chain conformation on the vibrational spectrum was investigated in ROA spectra [16], though the authors made no comparison with experimental results. The Raman and

ROA spectra of tri-L-serine in aqueous solution were also investigated experimentally and theoretically in the neutral and zwitterionic states [17]. The spectra were calculated with DFT and B3LYP hybrid exchange along with a consideration of implicit solvation effects using a Polarizability Continuum Model

(PCM). Due to the limitations in modelling of hydrogen bond interactions of water with small peptides [17] inherent to the PCM, the experimental and computational spectra obviously differed in some details [17].

To our knowledge, no experimental or calculated ROA spectra of linear and cyclic

L-Ser-L-Ser have been reported previously. Therefore, we report here the simulated and experimental Raman and ROA spectra for linear and cyclic L-Ser-L-Ser for the first time, in order to improve understanding of their conformations in solution and the structural constraints imposed by cyclization. This is an extension of our recent study on cyclization effects in dialanine (Chapter 5).

6.3 Experimental methods

Cyclic and linear L-Ser-L-Ser were purchased from Bachem Ltd. (Saffron Walden,

Essex, UK) and used without further purification. Deuterium oxide (99.98 atom %),

Na2HPO4 and NaH2PO4 were obtained from Sigma-Aldrich Ltd (Poole, Dorset, UK).

The phosphate buffer and sample concentrations are given in the corresponding figure captions. Samples for Raman and ROA spectroscopy were prepared by dissolving lyophilized material into the buffer solution; in any case of insolubility the

143

Chapter 6 sample was heated to 70 oC, and then allowed to cool to room temperature. Each solution was centrifuged for 5 minutes at 3000 rpm (1000 g) in order to minimize the presence of dust particles from the environment prior to loading into a quartz microflourescence cell for spectroscopic measurement.

All Raman and ROA spectra were measured on a chiralRaman spectrometer

(BioTools Inc., Jupiter FL, USA) at a wavelength of 532 nm, configured in the backscattering geometry and with a spectral resolution of 7 cm-1. The laser power was 1.2 W and data collection times were 6-24 h.

6.4 Computational methods

DFT calculations were carried out using the Gaussian 09 program [18] with the

B3LYP method [19, 20] and the AUG-cc-pVDZ basis set [21]. All calculations were performed with the IEF-PCM solvation method [22] using the Karplus and York continuous surface charge formalism [23], with a Polarizability Continuum Model

(PCM) of the water solvent with default PCM parameters being used except that the

Pauling atomic radii were substituted for the default UFF radii. The starting geometries for geometry optimization were those previously obtained at the

B3LYP/cc-pVDZ level for cyclic (L-Ser-L-Ser) [24], assuming C2 symmetry in both cases. Vibrational spectra were calculated at the optimized geometries and Raman and ROA activities were computed dynamically for an excitation wavelength of 532 nm. Relative Raman and ROA intensities were calculated from the computed Raman and ROA activities using the equations:

144

Chapter 6

4 Raman 0  fi    22 Iafi 45 7 hc fi  fi 1 exp kT

4 ROA 0  fi    22 I fi48 G'' 16 A hc fi  fi 1 exp kT

-1 where 0 = 18,797 cm and T = 298.15 K.

In these equations the ytensor invariants are defined as follows: a2 is the isotropic invariant of the electric-dipole/electricdipole polarizability tensor,

 is the symmetric anisotropic invariant of the electric-dipole/electricdipole

polarizability tensor, is the anisotropic invariant of the cross-product of the electric-dipole/electricdipole polarizability tensor with the electric-dipole/magnetic-

The Cartesian force constants obtained from the Gaussian 09 output were converted to force constants expressed in terms of internal coordinates using a normal coordinate analysis program derived from those of Schachtsneider [24]. A full set of internal coordinates, including all bond angles and torsion angles, was reduced to a set of 3N-6 symmetry-adapted internal coordinates. Normal coordinate analyses were performed without scaling of force constants, producing potential energy distributions for harmonic wavenumbers. Simulated Raman and ROA spectra were constructed by convolution with a Lorentzian lineshape function of 10 cm-1 fwhm.

145

Chapter 6

6.5 Results and Discussions

The atom numbering scheme along with computed molecular geometry of linear and cyclic L-Ser-L-Ser are shown in Figure 6.1 and 6.2. The experimental and calculated

Raman and ROA spectra of both linear and cyclic L-Ser-L-Ser are shown in Figure

6.3, and their corresponding vibrational band assignments are presented in Table 6.1.

The bond lengths, angles and selected torsion angles are shown in Tables S6.1, S6.2 and S6.3 (refer to supplementary information). In order to enable reliable assignment of spectral features, the Raman and ROA spectra of the deuterated versions of these dipeptides were also investigated and are presented in Figure 6.4. The Raman and

ROA band assignments for both forms of L-Ser-L-Ser in D2O are listed in Table 6.2.

The observed Raman and ROA bands for this region are in reasonably good agreement with previous measurements of the Raman spectrum of cyclic L-Ser-L-

Ser and Raman and ROA measurements of L-serine and tri-L-serine [14,16,17,27].

Even without the modelling of explicit hydration the Raman spectra for both linear and cyclic forms of L-Ser-L-Ser in H2O presented in Figure 6.3, are quite well simulated, with most Raman bands being predicted in the correct position and with the correct relative intensities. The most obvious, and expected, difference is due to the presence of the large peak in each spectrum between 1600 and 1700 cm-1 that originates from O-H bending motions of solvent water molecules. This difference is, of course, due to the PCM not incorporating individual water molecules. In the case of the ROA spectrum for linear L-Ser-L-Ser, shown in Figure 6.3, most bands are correctly predicted in terms of position, sign and even relative intensity. The details of the computed ROA spectrum of cyclic L-Ser-L-Ser in H2O are not in as close agreement with those of the experimental ROA spectrum, particularly in terms of

146

Chapter 6 their relative intensities. However, even here most ROA features are correctly predicted in terms of position and sign.

Although solvent interactions have been approximated using a PCM, the calculations are clearly accurate enough to allow the origins of specific bands to be determined.

These are detailed in Tables 6.1 and 6.2, for linear and cyclic L-Ser-L-Ser in H2O and D2O, respectively. The most significant structural information which is obtained in terms of the vibrational modes is commonly considered in terms of the amide I, II and III vibrations. The C=O stretching mode of peptides is largely responsible for bands appearing in the amide I region from 1600-1700 cm-1. The amide II region at

~1510-1570 cm-1, arising from the combination of C-N and N-H in-plane bending is not usually as rich in structural information for peptides and proteins in contrast to the amide I and III regions. Commonly for peptides, the most informative region is the amide III from ~1230- 1340 cm-1, in which bands originate from the coupling between N-H deformations and Cα-H stretching [25,26]. However, the spectra of cyclic dipeptides appear to contain larger features in other spectral regions than are typically observed for the corresponding linear forms; therefore here we discuss the features observed across a much broader spectral range.

In the Raman spectra of linear and cyclic L-Ser-L-Ser the bands at 1642 and 1671 cm-1 are both assigned to the amide I vibration. Upon deuteration these bands are shifted, upwards in wavenumber for the linear form by 10 cm-1 and down shifted by

22 cm-1 in the case of the cyclic form, due to the amide I mode coupling with N-H bending vibrations as shown previously [14,27]. The corresponding ROA bands for the amide I region appear as a doublet of positive peaks at 1666 and 1638 cm-1 for linear L-ser-L-ser that appears with a negative sign at 1642 cm-1 in the calculated spectrum. The amide I band appears as a weak negative peak at 1683 cm-1 for the

147

Chapter 6

-1 cyclic form which is shifted to lower wavenumber at 1654 cm in D2O. The band at

1519 cm-1 in the Raman spectrum of cyclic L-Ser-L-Ser is assigned to a mixed vibrational mode involving C-N, C-C, C=O, N-H groups and the amide II mode and gives rise to a large positive peak at the same position in the ROA spectrum. No corresponding band is found in the Raman spectrum of the linear form which suggests that this band is unique for the cyclic form of dipeptides as it was shown previously for cyclic L-Ala-L-Ala (Chapter 5). This Raman band shifts to lower wavenumber upon deuteration by 20 cm-1, which confirms that this band has a significant contribution from N-H deformations, as was observed previously for other cyclic dipeptides [14,27]. The calculated Raman and ROA spectra for cyclic L-

Ser-L-Ser are in good agreement with the experimental results for the amide II region. The rest of the bands in the 1400- 1500 cm-1 region relate mainly to C-H bending motions. Aside from these, the Raman bands at 1469 and 1404 cm-1 for the linear and cyclic forms, respectively, can be assigned to vibrations that are mainly from CH2 and N-H bending motions. The corresponding ROA bands appear as a - ve/+ve couplet at 1417/1452 cm-1 for the linear form and appear as a negative doublet at 1427 and 1457 cm-1 for cyclic L-Ser-L-Ser. The observed peak position shifts by

-1 ~3-10 cm when going from H2O to D2O solution which confirms the contribution to this band from the N-H vibrational mode.

The strong band at 1482 cm-1 in the ROA spectrum of cyclic L-Ser-L-Ser is mainly assigned to the amide II vibration which is unusually intense compared to that observed for most peptides, and does not appear for the linear form here. This suggests that the appearance of the amide II region changes significantly when L-

Ser-L-Ser undergoes cyclization where the molecule adopts a much more rigid structure.

148

Chapter 6

For linear L-Ser-L-Ser the Raman bands at 1378 and 1320 cm-1 correspond to the three +ve ROA bands at 1384, 1343 and 1304 cm-1 that are mainly assigned to C-O-

H and C-H bending. These bands shift downward by 3 cm-1 in the Raman spectra of cyclic L-Ser-L-Ser and yield a +ve/-ve/+ve triplet of ROA peaks at 1394, 1345 and

-1 1322 cm that are assigned to CH2 wagging vibrations as well as the amide III mode, as shown in Table 6.1. The Raman spectrum of linear L-Ser-L-Ser also contains two bands at 1278 and 1252 cm-1 that give rise to two weak +ve/-ve couplets in the ROA spectrum at 1283/1268 and 1252/1239 cm-1, respectively, which are correctly predicted in the calculated spectra though with lower intensity and a more negative bias in terms of ROA bandshape. These features arise from C=O, C-H and CH2 vibrations. In contrast, the ROA spectrum of cyclic L-Ser-L-Ser exhibits a +ve/+ve/- ve triplet at 1294/1268/1239 cm-1 in this region, with the sign flipped for the peaks at

1294 and 1226 cm-1 in the deuterated sample’s ROA spectrum.

The observed bands within this region of vibrational spectra mainly originate from

N-Cα, C-C and C-O stretching along with CH2 rocking vibrations. The band at 1154 cm-1 in the Raman spectrum of linear L-Ser-L-Ser shifts upward by 10 cm-1 in the spectrum of the cyclic form, whereas in the spectra measured in D2O it is shifted downward by 27 cm-1. This band was also associated with the same vibration for the cases of cyclic Gly-Gly and L-Ala-L-Ala [12, 25]. The –ve ROA band at 1064 cm-1 that mainly originates from C-O and CH2 stretching vibrations in the spectrum of cyclic L-Ser-L-Ser changes its sign to +ve and moves to 1067 cm-1 for the linear form, in agreement with the calculated spectra of both forms. The remaining bands below 1000 cm-1 originate from a mixture of C-C-H, C-O, C-C, N-H, C-C-O and

CH2 vibrations but these only give rise to two strong ROA bands for linear L-Ser-L-

Ser at 761 and 905 cm-1. More ROA bands are observed in the case of the cyclic

149

Chapter 6 form, in particular the -ve/+ve bands at 894/782 cm-1, presumably arising because the cyclic form is more rigid.

The bands below 700 cm-1 belong to two amide group vibrations of N-H out of plane bending and C=O bending modes with contributions from the ring (i.e. in the case of cyclic form) and C-C bending. Other vibrations at this region that only occur for cyclic L-Ser-L-Ser include ring deformations, ring stretching and O-H torsions.

Though these vibrations arise in this region, in the spectrum of cyclic L-Ser-L-Ser we only observe a –ve ROA band at 717 cm-1 which is assigned to amide VI and ring stretching modes.

6.6 Conclusions

Cyclic dipeptides possess interesting and beneficial biological activities, which mostly have unknown biological functions. The application of Raman spectroscopy and Raman optical Activity (ROA) to the study of bioactive peptides has matured over the past decade to a high level of sophistication where it provides useful information regarding the biological conformations of these species. In the present study, several Raman and ROA marker bands have been identified when L-Ser-L-Ser undergoes cyclization. ROA spectra, being sensitive to stereochemistry, are found to be more sensitive to cyclization with ROA marker bands being observed at 1683,

1519, 1482, 1345, 1080, 1064, 894, 782 and 716 cm-1. Although less sensitive,

Raman spectra also appear to contain bands that identify cyclization, these principally being at 1519, 1469, 1317, 1164, 845 and 670 cm-1.

It is unusual to observe strong signals in the amide II region for dipeptides as the other amide regions generally provide much more structural information for proteins and peptides. The amide II band appeared in the ROA spectrum of cyclic L-Ser-L-

Ser at 1519 cm-1 which corresponds closely to its appearance for cyclic L-Ala-L-Ala

150

Chapter 6 at 1521 cm-1 (Chapter 5). It is apparent that this band is unique for cyclic dipeptides since it was obtained in both Raman and ROA spectra of cyclic L-Ser-L-Ser and L-

Ala-L-Ala. Other similarities were also observed in Raman and ROA spectra of cyclic L-Ala-L-Ala, in particular, similarities within the Raman bands positions at

1317 and 1459 cm-1 and ROA bands at 790 and 1067 cm-1. Together, these studies show that Raman and ROA spectra do provide sensitive markers of peptide cyclization that can be used to easily differentiate them from conventional linear peptides.

151

Chapter 6

6.7 References

1. K. McCleland, P. J. Milne, F. R. Lucieto, C. Frost, S. C. Brauns, M. V. D. Venter, J. Du Plessis, K. Dyason, Journal of Pharmacy and Pharmacology. 2004, 56, 1143-1153. 2. M. B. Martins, I. Carvalho, Tetrahedron. 2007, 63, 9923-9932. 3. J. A. Trischman, R. E. Oeffner, M. J. D. Luna, M. Kazaoka, Marine Biotechnology. 2004, 6, 215-220. 4. S. W. Yang, T. M. Chan, J. Terracciano, D. Loebenberg, G. D. Chen, M. Patel, V. Gullo, B. Pramani, M. Chu, Journal of Antibiotics. 2004, 57, 345- 347. 5. S. C. Brauns, P. Milne, R. Naude, M. Van de Venter, Anticancer Research. 2004, 24, 1713-1719. 6. L. D. Barron, Current Opinion in Structural Biology. 2006, 16, 638-643. 7. L. D Barron, L. Hecht, Bimolecular conformational studies with vibrational Raman optical activity. In Biomolecular Spectroscopy; R. J. H. Clark, R. E. Hester, Editors; Wiley: Chichester, 1993; Part B, pp 235. 8. T. B. Freedman, L. A. Nafie, T. A. Keiderling, Biopolymers. 1995, 37, 265- 279. 9. L. A. Nafie, Applied Spectroscopy. 1996, 50, 14A-26A. 10. L. D. Barron, L. Hecht, Vibrational Raman optical activity: From fundamentals to biochemical applications. In Circular Dichroism, Principles and Applications; K. Nakanishi, N. Berova, R. W. Woody, Editors; VCH Publishers: New York, 1994; pp 179. 11. K. Ruud, A. J. Thorvaldsen, Chirallity. 2009, 21, E54-E67. 12. C. Hermann, K. Ruud, M. Reiher, Chemical Physics. 2008, 343, 200-209. 13. G. G. Fava, M. F. Belicchi, Acta Crystallographica. 1981, 1337, 625-629. 14. A. P. Mendham, T. J. Dines, M. J. Snowden, B. Z. Chowdhry, R. Withnall, Journal of Raman Spectroscopy. 2009, 40, 1478-1497; 15. A. P. Mendham, T. J. Dines, M. J. Snowden, R. Withnall, B. Z. Chowdhry, Journal of Raman Spectroscopy. 2009, 40, 1508-1520. 16. M. Pecul, Chemical Physics Letters. 2006, 427, 166–176. 17. V. W. Jurgensen, K. Jalkanen, Physical Biology. 2006, 3, S63–S79 18. Gaussian 09, Revision A.1, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox, Gaussian, Inc., Wallingford CT, 2009. 19. A. D. Becke, Journal of Chemical Physics. 1993, 98, 5648-5652 20. C. Lee, W. Yang, R. G. Parr, Physical Review B. 1988, 37, 785-789.

152

Chapter 6

21. T. H. Dunning, Journal of Chemical Physics. 1989, 90, 1007-1021. 22. M. T. Cances, V. Mennucci, J. Tomasi, Journal of Chemical Physics. 1997, 107, 3032-3041. 23. D. M. York, M. Karplus, Journal of Physical chemistry A. 1999, 103, 11060- 11079. 24. J. A. Schachtschneider, Vibrational Analysis of Polyatomic Molecules, Parts V and VI, Technical Report Nos. 231 and 57, Shell Development Co., Houston TX, 1964 and 1965. 25. R. Schweitzer-Stenner, Journal of Raman Spectroscopy. 2001, 32, 711-732. 26. N. G. Mirkin, S. Krimm, Journal of Molecular Structure. 1991, 242, 143- 160. 27. T. C. Cheam, S. Krimm, Spectrochimica Acta Part A: Molecular Spectroscopy. 1984, 40, 481-501.

153

Chapter 6

Figure 6.1: The chemical structure with atom numbering scheme (left) and calculate minimum energy conformation (right) of linear L-Ser-L-Ser.

H 21 O 7 H 9 H O 16 19O 22

4 C N 6 8 H C C H18 17 11 C 3 10 C H 20 H15 1 13 H12 H5 N C

H2 O14

Figure 6.2: The chemical structure with atom numbering scheme (left) and calculate minimum energy conformation (right) of cyclic L-Ser-L-Ser.

154

Chapter 6

Exp. Linear L-Ser-L-Ser

6x109 Calc. Linear L-Ser-L-Ser

+ I + R

I Exp. Cyclic L-Ser-L-Ser

9 Calc. Cyclic L-Ser-L-Ser 4.2x10

500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800

Exp. Linear L-Ser-L-Ser

Calc. Linear L-Ser-L-Ser

Exp. Cyclic L-Ser-L-Ser

- I -

Calc. Cyclic L-Ser-L-Ser

500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 Wavenumber (cm-1) Figure 6.3: Experimental and computed Raman (top) and ROA (bottom) spectra of linear (pH= 7.0) and cyclic L-Ser-L-Ser (pH= 7.0) in aqueous solution. The concentration for each sample was 50 mg/ml and laser power was 1.2 W at the laser.

The marker bands that are induced upon cyclization are highlighted by shading.

155

Chapter 6

Exp. Linear L-Ser-L-Ser

9 Calc. Linear L-Ser-L-Ser 4.7x10

L + I +

R Exp. Cyclic L-Ser-L-Ser

Calc. Cyclic L-Ser-L-Ser 1.1x109

500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800

Exp. Linear L-Ser-L-Ser

5.7x107 Calc. Linear L-Ser-L-Ser

- I - Exp. Cyclic L-Ser-L-Ser

R 5 I 1.7x10

Calc. Cyclic L-Ser-L-Ser

500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 Wavenumber (cm-1)

Figure 6.4: Experimental and computed Raman (top) and ROA (bottom) spectra of linear (pH= 7.0) and cyclic L-Ser-L-Ser (pH= 7.0) in D2O. The concentration for each sample was 50 mg/ml and the laser power was 0.6 W at the sample. The marker bands that are induced upon cyclization are highlighted by shading.

156

Chapter 6

Table 6.1: Calculated and experimental Raman and ROA bands for cyclic and linear L-Ser-L-Ser in H2O.

Experimental Calculated Band assignments Linear Cyclic Linear Cyclic Linear Cyclic ROA Raman ROA Raman

1666(+ve) 1642 1683(-ve) 1671 1655 1661 as(NH3) (89) (CO) (53), (NC) (15), ip(NH) (19) [amide I] 1638(+ve) - - - 1642 1658 (CO) (66), (C6N5) (13) (CO) (53), (NC) (12), ip(NH) (19) [amide I] - - - - 1612 - as(NH3) (75) -

- - - - 1562 - (C6N5) (33), ip(NH) (53) -

1565(+ve) - 1519(+ve) 1519 1558 1540 as(CO2) (91) (CO) (23), (NC) (18), (CC) (12), ip(NH) (17) [amide II] - - 1482(+ve) - 1494 1493 (CH2′) (98) (CO) (24), ip(NH) (31), (CH2) (33) [amide II] - - - 1469 1479 1476 (CH2) (98) (NC) (34), (CH2) (35) 1452(+ve) 1469 - - 1466 1472 s(NH3) (86) (CH2) (91) - - 1457(-ve) - 1446 1456 (C13O22H23) (13), (CH2′) (73) (NC) (15), ip(NH) (25), (CH2) (30), (CH2) (12) [amide III] 1417(-ve) 1404 1427(-ve) 1404 1429 1414 (C10O18H29) (18), (CH2) (58) ip(NH) (22), (CH2) (83) 1384(+ve) - 1394(+ve) - 1397 1395 (C8H14) (19),′(C8H14) (41) (COH) (22),ip(NH) (18),(CH2) (32) [amide III] - - - - - 1382 - (COH) (20),(CCH) (12),(CH2) (32),(CH2) (26) 1363(+ve) 1378 - 1374 1375 1372 s(CO2) (56),(C1C4) (10) (COH) (14),(CH2) (43),(CH2) (20) 1343(+ve) - 1345(-ve) - 1343 1332 ′(C4H11) (26),′(C8H14) (23) (NC) (10), (CCH) (12),(CCH) (36),(CCH) (15) - 1320 1322(+ve) 1317 1314 1321 ′(C4H11) (15),(C8H14) (23), (CH2′) (22) (NC) (10), (NC) (16), (CCH) (31),(CCH) (13) 1304(+ve) - - 1303 - (C4H11) (20),(CH2′) (38) - 1283(+ve) - 1294(+ve) 1281 1300 1283 (C10O18H29) (12),(C4H11) (16),′(C4H11) (12),τ(CH2′) (17) (CCH) (39),(CCH) (16) 

1268(+ve) 1278 1268(+ve) - 1267 1277 ip(NH) (16), (CH2) (35) (CCH) (45),(CCH) (16) 1252(+ve) 1252 - - 1251 - (C13O22H23) (29), (CH2) (18), (CH2′) (14) -

- - - - - 1224 - (COH) (27),(CCH) (14),(CH2) (44)

1239(+ve) - 1239(-ve) 1239 1243 1236 (C6N5) (11), ip(NH) (23),(CH2) (15) (COH) (18), (CCH) (10),(CCH) (21), (CH2) (36) - - - - 1195 - (C10O18H29) (43),(C4H11) (12),′(C4H11) (16), ω(CH2) (13) -

1178(+ve) - - - 1163 - op(NH3) (20), (CH2) (28) -

1151(+ve) 1154 1154(+ve) 1164 1143 1153 (C4N5) (29),(CH2) (23) (NC) (29),(COH) (12), (CH2) (16) 1116(+ve) - 1116(-ve) - 1108 1110 (C8N9) (11),ip(NH3) (24),(C8H14) (18) (NC) (36), (C3C4) (14),(C-O) (11),ip(CO) (10) - - 1083(+ve) 1080 - - - - 1094(+ve) 1080 1064(-ve) - 1087 1067 (C4C10) (27),(C10O18) (35) (C3C4) (19),(C-O) (25),(CH2) (13) 1067(+ve) - 1048(+ve) - 1062 1069 (C8N9) (21),op(NH3) (10) (C3C4) (28),(C-O) (59) 1039(+ve) 1048 1025(+ve) - 1040 1024 (C13O22) (83) (C-O) (40) ,(CCH) (10),(CH2) (10) 1017(+ve) 992 998(+ve) 992 1009 980 (C4N5) (12),(C10O18) (13),(CH2) (24)  (NC) (10), (C3C4) (12),(C-O) (12), (CH2) (39) 964(+ve) 964 - - 969 - (C8C13) (14),op(NH3) (27), (CH2′)(32) -

956(+ve) - - 952 954 946 (C6C8) (24), ip(NH3) (11) (C3C4) (19),(C-O) (10), (CH2) (28) 905(-ve) - 922(+ve) - 924 935 (C4C10) (17), (C10O18) (33),(CH2) (11) ip(ring-3) (28), (CH2) (33) 157

Chapter 6

877(+ve) 882 894(-ve) - 884 853 (C8N9) (36), (C8C13) (18), (CH2′) (10) (CC) (14), (C3C4) (26),(C-O) (17), (CCH) (10),op(CO) (12) - 816 - 845 848 844 (C1C4) (15),(CO2) (12) (C-O) (16),(CCH) (14),(CCH) (18),op(CO) (36) - - - - - 807 - (CC) (29), (NC) (12)

- 784 782(+ve) - 792 - (CO2) (27), (CO2) (24)  761(+ve) - 717(-ve) - 752 736 (C6C8C13) (11), ip(CO) (18), op(CO) (32)  (C3C4) (18), ip(ring-3) (32), op(CO) (32) [amide VI] - - - - 709 - ip(CO) (16), op(CO) (30)  640(-ve) 670 682(+ve) 670 641 682 (C4C10) (10), (N5C4C1) (22),(CO2) (23), (CO2) (10), (CO2) (147) op(NH) (56), op(CO) (35) [amide V,VI] - - 664(+ve) - - 645 - (CCO) (10), op(NH) (77), op(CO) (12) [amide V]

- - 631(-ve) - - 644 - (NC) (12), (CC) (38), op(NH) (23) - - 616(+ve) 622 570 621 op(NH) (34),(N5C6) (30) (CC) (11), ip(CO) (47), op(NH) (12) [amide IV] - - 577(-ve) - 575 543 (C1C4) (13),(N5C6C8) (13), (CO2) (15) (CCO) (21),op(NH) (23),op(CO) (34) 534(+ve) - - - 524 510 (C4C10O18) (31), (CO2) (15) (CCO) (23),op(NH) (12), op(CO) (20),op(ring-2) (10)

158

Chapter 6

Table 6.2: Calculated and experimental Raman and ROA bands for cyclic and linear L-Ser-L-Ser in D2O.

Experimental Calculated Band assignments Linear Cyclic Linear Cyclic Linear Cyclic ROA Raman ROA Raman 1654(+ve) 1654 1654(-ve) 1649 1646 1637 ν (CO) (72), ν(C6N5) (17) ν(CO) (72), ν(NC) (13) [amide I] - - 1575(+ve) - - 1635 - ν (CO) (68), ν(NC) (14) [amide I] 1556(+ve) 1560 1546(+ve) - 1558 1513 νas (CO2) (92) ν (NC) (35), ν (CC) (14), δ(CH2) (12) [amide II] - - - - 1495 - ν (C6N5) (31), δip(ND) (11),δ (CH2′) (26) - 1482(+ve) - - 1499 1493 - ν (C6N5) (10), δ(CH2′) (74) - 1467(-ve) 1469 1479 - 1478 1477 δ (CH2) (98) δ (CH2) (98) 1449(+ve) 1449 1457(-ve) 1469 - 1471 - δ (CH2) (87), ν(CO) (12), ν(CC) (10), ν(NC) (49) [amide II] 1429(+ve) 1427 1439(+ve) - 1429 - ω (CH2′) (98) - - - 1417(+ve) - 1410 - νs (CO2) (21), ω (CH2) (57) - 1391(+ve) - - - - 1394 - ω (CH2) (80) 1379(-ve) 1376 1384(+ve) 1386 1376 1392 δ (C8H14) (10), δ′(C8H14) (44), ω(CH2) (12) ω (CH2) (88) 1366(+ve) 1358 - - 1367 1348 νs (CO2) (39),δ′(C8H14) (14),ω(CH2) (20) ω (CCH) (25),τ(CH2) (33) 1338(+ve) 1327 1348(+ve) 1338 1332 1344 νs (CO2) (11),δ(C4H11) (15),δ′(C4H11) (40) ν (NCα) (20), ω(CCH) (17),τ(CH2) (37) 1325(-ve) - 1320(+ve) - 1301 1316 δ(C8H14) (29), τ(CH2′) (422) δ(CCH) (28),ω(CCH) (12),ρ(CCH) (11),τ(CCH) (10),τ(CH2) (16) - - - - - 1311 - δ (CCH) (32),ρ(CCH) (12),τ(CH2) (17) 1291(+ve) 1294 1294(-ve) 1294 1287 1280 δ (C4H11) (23), δ′(C4H11) (14), τ(CH2) (19),τ(CH2′) (15) δ (CCH) (16),τ(CH2) (32) - - 1268(+ve) 1270 1274 1270 δ (C8H14) (23), δ′(C8H14) (16),τ(CH2′) (20) δ (CCH) (26), τ(CH2) (27) 1249(+ve) 1249 1226(+ve) - 1252 1238 δ (C4H11) (28),τ(CH2) (56) ν (NC) (17), ν(CC) (10), δip(ND) (13), ω(CCH) (13),τ(CCH) (19) 1215(+ve) 1212 1202(+ve) 1207 - 1227 - ν (NCα) (23), δip(ND) (28), δ(CCH) (12), δip(CO) (16) [amide III] - - - - 1181 - δas (ND3) (99) - - - 1167(+ve) - 1175 - δs (ND3) (16),δas(ND3) (66) - - - - - 1152 - ν (C8N9) (13),ν(C6C13) (29),δs(ND3) (14) - 1140(+ve) 1140 1132(+ve) 1127 1144 1132 ν (C4N5) (17),δas(ND3) (20), ρ(CH2′) (18) δ (COD) (13), τ(CCH) (24),ρ(CH2) (39) - - 1108(-ve) - 1128 1103 ν (C4N5) (15),δs(ND3) (11),δas(ND3) (17), ρ(CH2) (17) δ (COD) (15), δip(ND) (19), ρ(CH2) (35) 1086(-ve) 1094 - - 1096 - ν (C4C10) (18),ν(C10O18) (25), ρ(CH2) (11) - 1064(+ve) 1064 - 1075 1074 1067 δs (ND3) (25), ρ(CH2′) (14) ν (C3C4) (26),ν(C-O) (58) - - - - 1046 - ν (C13O22) (60) - - - 1048(-ve) - 1039 - ν (C13O22) (16),δ(C10O18D29) (16) - 1003(+ve) 1006 - - 1015 - ν (C6C8) (20),ν(C10O18) (18),δ(C8H14) (10) - 987(+ve) 992 970(+ve) - 985 988 ν (C6N5) (10), δip(ND) (44) ν (C3C4) (11),ν(C-O) (12),δ(COD) (13), ρ(CCH) (13),δip(ring-3) (22) - - - 989 - 984 - δ (COD) (22), δip(ND) (18), ρ(CCH) (10) - - - 964 - 959 - ν (NCα) (23),ν(C-O) (11),δ(COD) (21),δip(ND) (21) 931(+ve) 928 922(+ve) - 931 952 ν (C4N5) (14),ν(C10O18) (13),δ(C10O18D29) (16), δip(ND) (11), ρ(CH2)(15) ν(NCα) (15), ν(C3C4) (22),ν(C-O) (18), δip(ND) (22) [amide III] - 908 - 905 912 - ν (C8N9) (11),δ(C13O22D23) (52) - - - - - 896 - ν (C1C4) (11),ν(C10O18) (19),δ(C10O18D29) (37) -

159

Chapter 6

885(-ve) 882 871(+ve) - 875 871 ρop (ND3) (16) δ (COD) (30), ρ(CH2) (29) - - - - 871 861 ν (C8N9) (20),ν(C1C4) (10),ρ(CH2′) (18) δ(COD) (35),δip(ring-3) (11), ρ(CH2) (31) 845(+ve) - 845(-ve) 845 - 835 - ν (C3C4) (20), ν(C-O) (15), δ(CCH) (13), ρ(CCH) (11), δop(CO) (16) 811(+ve) - - - 820 829 ρip (NH3) (18), δip(CO) (10), δ(CO2) (10) ν(C-O) (10),δ(COD) (16),δ(CCH) (11), ρ(CCH) (14), δop(CO) (28) 784(+ve) - 767(+ve) - 782 769 ν (C8N9) (10),ν(C8C13) (14),δ(C13O22D23) (11), ρop(ND3) (39) ν(CC) (37), ν(NC) (13), δip(ND) (20) - - - - 771 - δ (CO2) (23), ω(CO2) (31) - - 744 697(+ve) - 731 722 ν (C8C13) (14),δop(CO) (50) ν(C3C4) (19), δip(ring-3) (26), δop(CO) (32) - 676 646(+ve) 688 660 658 ν(C6C8) (14),ρip(ND3) (19),δ(C8C13O22) (11),δip(CO) (19), δop(CO) (12) ν(NCα) (10), ν(CC) (16),ν(NC) (12), ν(C3C4) (16),δop(CO) (28) [amide VI] 640(-ve) 640 - - 636 - ν (C1C4) (10), ν(C4C10) (11), δ(N5C4C1) (18),δ(CO2) (25), ω(CO2) (15) - - - 601(+ve) 601 - 617 - ν (CC) (35), δip(CO) (19) - - - - - 573 - δ (CCO) (28),δop(ND) (11),δop(CO) (46) [amide VI] 556(-ve) - 541(-ve) - 551 569 ν (C1C4) (12),δ(N5C6C8) (14),δ(CO2) (12), ρ(CO2) (14) δ(CCO) (13), δip(CO) (30), δop(ND) (29), δop(CO) (20) [amide IV] - - 519(-ve) - 519 - δ (N5C4C10) (11),δ(C4C10O18) (31), ω(CO2) (15) -

160

Chapter 6

6.8 Supplementary information

Table S6.1: Calculated and experimental bond lengths (Å) for cyclic and linear L-

Ser-L-Ser.

Linear L-Ser-L-Ser Calculated bond lengths (Å) r(C1O3) 1.2603 r(O2C1) 1.2647 r(C4N5) 1.4577 r(C4C1) 1.5598 r(N5H12) 1.0127 r(C6N5) 1.3369 r(C6C8) 1.5353 r(O7C6) 1.2429 r(C8C13) 1.5286 r(C8H14) 1.0932 r(N9H15) 1.0245 r(N9C8) 1.4955 r(C10O18) 1.4387 r(C10C4) 1.5268 r(H11C4) 1.096 r(C13O22) 1.4329 r(C13H25) 1.0977 r(H16N9) 1.0273 r(H17N9) 1.0218 r(H19O18) 0.9659 r(H20C10) 1.099 r(H21C10) 1.0974 r(O22H23) 0.9661 r(H24C13) 1.0976

Cyclic L-Ser-L-Ser r(N1-C3) 1.4605 r(H2-N1) 1.0153 r(C3-C6) 1.5216 r(C3-H5) 1.0995 r(C4-H15) 1.0997 r(C4-H17) 1.0994 r(C4-C3) 1.5395 r(C6-O7) 1.2439 r(C6-N8) 1.3407 r(O16-C4) 1.4285 r(O16-H21) 0.9679

Table S6.2: Calculated bond angles (o) for cyclic and linear L-Ser-L-Ser.

Linear L-Ser-L-Ser Calculated bond angles (o)

θ(O3C1O2) 126.09 θ(O3C1C4) 118.94 θ(O2C1C4) 114.97 θ(N5C4C1) 113.61 θ(C4N5H12) 116.88 θ(C4N5C6) 123.88

161

Chapter 6

θ(N5C4C10) 109.25 θ(N5C4H11) 107.76 θ(C1C4C10) 111.82 θ(C1C4H11) 106.72 θ(H12N5C6) 118.93 θ(N5C6C8) 115.92 θ(N5C6O7) 124.64 θ(C8C6O7) 119.44 θ(C6C8C13) 111.47 θ(C6C8H14) 111.63 θ(C6C8N9) 107.46 θ(C13C8H14) 110.07 θ(C13C8N9) 108.93 θ(C8C13O22) 106.79 θ(C8C13H25) 108.77 θ(C8C13H24) 110.15 θ(H14C8N9) 107.12 θ(H15N9C8) 109.82 θ(H15N9H16) 109.75 θ(H15N9H17) 108.36 θ(C8N9H16) 109.07 θ(C8N9H17) 111.95 θ(O18C10C4) 108.37 θ(C10O18H19) 108.19 θ(O18C10H20) 110.12 θ(O18C10H21) 110.7 θ(C10C4H11) 107.38 θ(C4C10H20) 109.41 θ(C4C10H21) 109.08 θ(O22C13H25) 111.25 θ(C13O22H23) 108.44 θ(O22C13H24) 110.64 θ(H25C13H24) 109.21 θ(H16N9H17) 107.85 θ(H20C10H21) 109.14

Cyclic L-Ser-L-Ser

θ(C3-N1-H2) 116.54 θ(N1-C3-C6) 113.27 θ(N1-C3-H5) 108.27 θ(N1-C3-C4) 111.05 θ(C3-N1-C13) 128.01 θ(H2-N1-C13) 115.42 θ(C6-C3-H5) 105.92 θ(C6-C3-C4) 111.2 θ(C3-C6-O7) 119.33 θ(C3-C6-N8) 118.32 θ(H5-C3-C4) 106.73 θ(H15-C4-C3) 108.52 θ(H15-C4-O16) 110.83 θ(H17-C4-C3) 108.86 θ(H17-C4-H15) 108.57 θ(H17-C4-O16) 106.59 θ(C3-C4-O16) 113.29 θ(O7-C6-N8) 122.25 θ(C4-O16-H21) 108.58

162

Chapter 6

Table S6.3: Calculated torsion angles (o) for cyclic and linear L-Ser-L-Ser

o Linear L-Ser-L-Ser Calculated torsion angles ( )

τ(O3C1C4N5) -7.15 τ(O3C1C4C10) 117.09 τ(O3C1C4H11) -125.76 τ(O2C1C4N5) 172.54 τ(O2C1C4C10) -63.22 τ(O2C1C4H11) 53.93 τ(H12N5C4C1) 86.21 τ(C6N5C4C1) -87.41 τ(C10C4N5H12) -39.41 τ(H11C4N5H12) -155.77 τ(C10C4N5C6) 146.97 τ(H11C4N5C6) 30.61 τ(C4N5C6C8) 175.65 τ(C4N5C6O7) -3.59 τ(N5C4C10O18) 61.15 τ(N5C4C10H20) -58.94 τ(N5C4C10H21) -178.26 τ(C1C4C10O18) -65.5 τ(C1C4C10H20) 174.41 τ(C1C4C10H21) 55.09 τ(H12N5C6C8) 2.15 τ(H12N5C6O7) -177.09 τ(N5C6C8C13) -85.71 τ(N5C6C8H14) 37.84 τ(N5C6C8N9) 155.01 τ(C13C8C6O7) 93.57 τ(H14C8C6O7) -142.88 τ(N9C8C6O7) -25.71 τ(C6C8C13O22) -175.76 τ(C6C8C13H25) 64.09 τ(C6C8C13H24) -55.57 τ(C6C8N9H15) 157.23 τ(C6C8N9H16) 36.92 τ(C6C8N9H17) -82.35 τ(O22C13C8H14) 59.81 τ(H25C13C8H14) -60.34 τ(H24C13C8H14) -180 τ(O22C13C8N9) -57.36 τ(H25C13C8N9) -177.51 τ(H24C13C8N9) 62.84 τ(C13C8N9H15) 36.33 τ(C13C8N9H16) -83.97 τ(C13C8N9H17) 156.76 τ(C8C13O22H23) 161.52 τ(H14C8N9H15) -82.7 τ(H14C8N9H16) 156.99 τ(H14C8N9H17) 37.73 τ(H19O18C10C4) -173.4 τ(O18C10C4H11) 177.75 τ(H20C10O18H19) -53.75 τ(H21C10O18H19) 67.02 τ(H20C10C4H11) 57.66 τ(H21C10C4H11) -61.66 τ(H23O22C13H25) -79.94 τ(H24C13O22H23) 41.64

163

Chapter 6

Cyclic L-Ser-L-Ser  (C6-C3-N1-H2) -171.89 (H5-C3-N1-H2) -54.73 (C4-C3-N1-H2) 62.14 (C13-N1-C3-C6) 5.93 (N1-C3-C6-O7) 173.49 (N1-C3-C6-N8) -6.34 (C13-N1-C3-H5) 123.09 (C13-N1-C3-C4) -120.04 (N1-C3-C4-H17) -174.81 (N1-C3-C4-H15) -56.75 (N1-C3-C4-O16) 66.79 (C3-N1-C13-C10) 0.57 (C3-N1-C13-O14) -179.26 (H2-N1-C13-C10) 178.42 (H2-N1-C13-O14) -1.41 (O7-C6-C3-H5) 54.97 (N8-C6-C3-H5) -124.86 (O7-C6-C3-C4) -60.61 (N8-C6-C3-C4) 119.55 (C6-C3-C4-H17) 58.07 (C6-C3-C4-H15) 176.14 (C6-C3-C4-O16) -60.33 (H5-C3-C4-H17) -57.01 (H5-C3-C4-H15) 61.06 (H5-C3-C4-O16) -175.41 (H17-C4-O16-H21) 178.27 (C3-C4-O16-H21) -62.03 (H15-C4-O16-H21) 60.23

164

Chapter 7

Conclusion and Future work

165

Chapter 7

7.0 Conclusion

Raman, ROA and SERS are spectroscopic techniques that constitute a powerful analytical toolkit for deriving quantitative and qualitative biomolecular information.

Application of these techniques in combination is more advantageous, since they can provide more structural information on the nature of biomolecules along with the chirality of the molecules. As demonstrated in different chapters of this thesis, more marker bands were obtained in the ROA spectra of the sugar, ribonucleotides and short dipeptides in contrast to the Raman and SERS spectra, which confirm the sensitivity of ROA to structural changes. However, ROA is a weak effect and limited to biological molecules that are only available at relatively low concentrations. These limitations could be overcome by employing SERS. As shown in this thesis, obtaining SERS is more complex than measuring the parent Raman spectra since a range of factors needs to be controlled in order to obtain reliable measurements.

Small changes in the experimental conditions can lead to a huge variation in the

SERS signals.

The reproducibility of SERS was one of the main issues for obtaining consistent

SEROA results as unstable substrates may cause a loss of signal enhancement over time. In order to measure reliable SEROA spectra, it was crucial to stabilise the aggregation process and indeed any other dynamic process occurring in the sample.

One solution to this problem was to suspend the colloidal particles in a gel and halt or greatly slow down the aggregation process. The time-dependence of SERS affects the optimum conditions drastically since the extent of the aggregation over the measuring period influences the fluctuations in the bands measured and the subsequent difficulty in controlling the reproducibility of spectra. This thesis presents a number of studies confirming SEROA as a chiroptical technique by controlling the

166

Chapter 7

SERS experimental setup and its time-dependence. This was mainly achieved by using citrate- and hydroxylamine-reduced silver colloids along with a polycarbophil polymer as stabilising media in solution for studying SERS and SEROA of L- and D- ribose and L- and D-tryptophan. The addition of polymer to the colloidal system was shown to control the aggregation process in SERS solution along with protecting the

Ag nanoparticles from further aggregation. SEROA spectra measured for both enantiomers of each of ribose and tryptophan clearly showed mirror image bands for the majority of spectral regions where the actual signals were clearly distinguishable from the background noise. This work presents the first observation of mirror image

SEROA bands, an essential step in the proof of this new technique [1].

Another system to provide enhancement for the ROA weak effect was that of silver silica nanotags prepared by collaborators at the University of Strathclyde which act as single plasmonic nanoparticle substrates. These silver nanoparticles which are conjugated to benzotriazole dye molecules have been used as an approach to enhance the sensitivity of SERS and allow colorimetric detection of analytes at relatively low concentrations. It was shown in this thesis that they can also function as chiral plasmonic nanomaterials by providing superchiral electromagnetic fields for ultrasensitive detection of biomolecular conformation. The chiroptical activity of this system was verified by measurement of SERROA spectra of the enantiomers of ribose and tryptophan when attached to silver silica nanotags These spectra clearly showed that SERROA bands of opposing sign were obtained from the two enantiomers of each chiral molecule and that these were signatures of the chiroptical nature of the interaction between these analytes and the surface plasmons of the dye- tagged nanoprobes, as a result of a chiral influence on the SERRS spectrum of the benzotriazole dye. Our findings demonstrated that chirality was induced into the

167

Chapter 7 achiral plasmonic surface of the substrate by binding to L- and D-enantiomeric analytes. The SERROA effect measured through the interaction of the benzotriazole dye molecules with the surface plasmons is fundamentally different to that leading to the SEROA spectra measured for D- and L-ribose and tryptophan [1] adsorbed directly onto the polymer-stabilised colloids. The direct interaction between the chiral molecules and metal nanoparticles in that case was responsible for the observed mirror image SEROA spectra which are mainly due to the field gradient generated by the plasmon resonance. The observed chirality effect mainly originates from dipolar interactions with chiral molecules. The present, SERROA, effect is a far field effect where non-direct interactions between silver nanoparticles and analyte occur and result in transmission of the chirality of the analyte to the achiral benzotriazole tags attached by linker molecules to a silver surface. The mechanism proposed as being responsible for these SERROA results is comparable to that recently reported for a class of hybrid plasmonic nanomaterials [2] with indirect adsorption of chiral molecules. Abdulrahman et al. [2] showed that a chiral response was induced into the plasmonic resonance of the achiral nanostructure using circular dichroism, so via measurement of electronic excitation. Here, we have observed a similar response but through monitoring vibrational excitation. The chiral plasmonic response from silver silica nanotags is a promising alternative which again validated

SERROA as a technique but through a different mechanism.

The detection and quantification of phosphorylated species in complex mixtures has proven to be difficult due to the limited availability of suitable methods. It has been established in this thesis that Raman and ROA spectroscopies are powerful probes, both qualitatively and quantitatively, of phosphorylation [3]. Raman and ROA spectra of adenosine and seven of its derivative ribonucleotides were measured and

168

Chapter 7 confirmed the sensitivity of both spectroscopic techniques to structural differences.

However, it was shown that ROA is more sensitive to the site and degree of phosphorylation, with a considerable number of marker bands being identified for these ribonucleotides. It was found that Raman spectra were sensitive to the number of phosphate groups but not to their position around the ribose ring. The general features of the ROA spectra varied dramatically for these adenosine nucleotides and provided a fingerprint sensitive to both the number and position of phosphate groups.

Studies of the conformations of cyclic and linear L-Ser-L-Ser and L-Ala-L-Ala in aqueous solution were carried out using Raman and ROA spectroscopies in combination with quantum mechanical calculations performed by a collaborator at the University of Dundee. The Raman and ROA spectra of the cyclic and linear forms of L-Ala-L-Ala and L-Ser-L-Ser were measured in both H2O and D2O and assignment of the observed bands was proposed from DFT calculations in both cases.

The calculated spectra of both linear and cyclic L-Ala-L-Ala and L-Ser-L-Ser were in good agreement with our experimental ROA and Raman spectra; the comparison of which for both forms showed that ROA is more sensitive to structural changes due to cyclization of dipeptides as it provided more marker bands. Considerable differences were noted between the observed ROA bands for the cyclic and linear forms of dialanine and diserine that reflect large differences in the vibrational modes of the polypeptide backbone upon cyclicization. This study demonstrated that ROA spectroscopy when utilised in combination with computational modelling clearly provides a potential tool for characterization of cyclic peptides.

169

Chapter 7

7.1 Future work

This thesis has highlighted the importance of optimization of the correct experimental protocols for obtaining reliable SERS and SEROA spectra, particularly the latter. By addition of a hydrogel polymer the aggregation of colloids was controlled which greatly assisted in obtaining reproducible SEROA spectra. The hydrogel colloidal system clearly has potential for the stereochemical analysis of a wide range of biological molecules by SEROA spectroscopy. In the present thesis it was only possible to use one type of hydrogel polymer, polycarbopol, other hydrophilic polymers should also be investigated since the current polymer used in this thesis was pH sensitive whereas other polymers are not and so may provide better stabilising media for colloid suspension.

The reported theoretical configuration of SEROA is encouraging as the right experimental setup can be predicted [4]. Bour et al. showed the dependence of CID ratio on the distance and molecular orientation for SEROA of ribose [4]. Their results were in agreement with the reported experiments in this thesis [1] and can potentially provide a suggestion for future experimental setups. Much further work is required to now understand the two mechanisms identified here for SER(R)OA and there is a need for experimental and theoretical concepts to work synergistically if the potential for SEROA as an analytical technique is to become fully realized. Other classes of nanoparticles such as hollow gold nanospheres, silver triangles, gold nanorods and gold/silver silica nanoshells provide a strong electromagnetic field that increase SERS intensities without relying upon high conjunction potentials or ‘hot spots’ that are characteristic of aggregated metal colloids. Since the nature of absorption in these nanomaterials is very different, much further work needs to be performed to investigate this new phenomenon, and to optimize it. This study already

170

Chapter 7 raises interesting possibilities for the enantioselective detection of chiral molecules, and in particular biomolecules, hence extending the scope of nanoplasmonic devices.

The potential applications of nanoparticles as plasmonic nanomaterials remains hindered due to limitations in colloidal synthesis to achieve monodisperse nanoparticles which exhibit enhanced properties. An approach to overcome this limitation is the rational design of colloidal heterostructured nanocrystals in which the chemical composition and different reactants (this refers to monomers which induce nucleation of nanocrystals and sustain their subsequent enlargement) are spatially controlled ultimately leading to induction of anisotropic growth of metallic branches and different nanoparticle junctions [5]. The use of such plasmonic nanomaterials as powerful chirality nanoprobes upon attachment to protein and DNA for in situ structural determination will be achieved by selecting appropriate nature, dimensions, morphology and functionalization of the metallic nanostructured surfaces [6,7]. The external chiral template provides a great advantage to solution- based chiral nanomaterials for biological samples such as DNA, bacteria and viruses, therefore offering a wide range of possible applications in biology and medicine.

171

Chapter 7

7.2 References

1. S. Ostovar Pour, S. E. J. Bell, E. W. Blanch, Chemical Communications. 2011, 4754-4756. 2. N. A. Abdulrahman, Z. Fan, T. Tonooka, S. M. Kelly, N. Gadegaard, E. Hendry, A. O. Govorov, M. Kadodwala, Nano Letters. 2012, 977-983. 3. S. Ostovar Pour, E. W. Blanch, Applied Spectroscopy. 2012, 289-293. 4. V. Novak, J. Sebestík, P. Bour, Journal of Chemical Theory and Computation. 2012, 1714-1720. 5. L. Carbone, P. D. Cozzoli, NanoToday. 2010, 449-493. 6. E. Hendry, T. Carpy, J. Johnston, M. Popland, R.V. Mikhaylovskiy, A. J. Lapthorn, S. M. Kelly, L. D. Barron, N. Gadegaard, M. Kadodwala, Nature Nanotechnology. 2010, 783-787 7. M. F. Garcia-Parajo, Nature Photonics. 2008, 201-203.

172

Chapter 8

Appendix

173

Chapter 8

8.0 Declaration

This chapter consists of one full paper by Catherine M. Templeton, Saeideh Ostovar pour, Jeanette R. Hobbs, Ewan W. Blanch, Steven D. Munger, Graeme L. Conn,

Chemical Senses. 2011, 1-10.

This article has been reproduced in an unchanged format. This work was a collaboration with Dr Graeme Conn’s group at the Emory University School of

Medicine, Atlanta, USA, and I am second author on this publication. In this paper I measured Raman and ROA spectra of wild type and two mutants of monellin

(MNEI), a sweet tasting protein in order to investigate secondary structures changes underlying the sweet taste receptor response. The rest of the work on this paper was carried out by the other authors.

174