Gas Phase Acidity and Proton Affinity Studies of Organic

GAS PHASE ACIDITY AND PROTON AFFINITY STUDIES OF ORGANIC

SPECIES USING MASS SPECTROMETRY

MIN LIU

A Dissertation submitted to the

Graduate School-New Brunswick

Rutgers, The State University of New Jersey

in partial fulfillment of the requirements

for the degree of

Doctor of Philosophy

Graduate Program in Chemistry and Chemical Biology

written under the direction of

Professor Jeehiun K. Lee

and approved by

______

New Brunswick, New Jersey

[October, 2011]

ABSTRACT OF THE DISSERTATION

GAS PHASE ACIDITY AND PROTON AFFINITY STUDIES OF ORGANIC

SPECIES USING MASS SPECTROMETRY

By MIN LIU

Dissertation Director: Professor Jeehiun K. Lee

DNA damaged bases are linked to carcinogenesis, aging and cell death. One of our main focuses is to examine the intrinsic reactivity of normal and damaged nucleobases in order to find out how damaged bases are different from normal bases. 1,N6-ethenoadenine (eA) is one of the damaged nucleobases which can be excised by alkyladenine DNA glycosylase (AAG) in humans. We find that the N9-H of eA is more acidic than adenine and guanine, which indicates that AAG may cleave certain damaged nucleobases as anions and deprotonated damaged bases are better leaving groups than deprotonated adenine and guanine. Furthermore, differentiation of damaged from normal bases may be enhanced by the nonpolarity of the active site. In addition we find that the N9-H of O6- methylguanine (OMG) is less acidic than adenine and guanine. This result is consistent with the fact that OMG is not an AAG substrate. We also studied the thermochemical properties and tautomerism of normal pyrimidine bases cytosine and thymine. Another of our focuses is the study of gas phase acidity studies of organic silanols and some known hydrogen-bonding organocatalysts. It is generally accepted that silanols

are more acidic than their carbon analogs, but we have found that the theoretical carbon diol analogs are actually more acidic than silicon diols depending on substitution and structure. Polarizability versus induction, gas phase versus solution phase, catalysis and molecular recognition are also discussed. We are also interested in the proton affinity and reactivity of N-heterocyclic carbenes (NHCs). Stable NHCs are widely used as novel ligands for transition-metal- catalyzed reactions. The dialkylimidazolium salts (protonated carbenes) are also an important class of ionic liquids. It has been found that the second generation of Grubbs metathesis catalysts, which contain NHCs, are more acitive than the first generation of catalysts, which contain tricyclohexyl phosphine (PCy3) only. More basic carbenes presumably will be more effective ligands. Therefore we are interested in the proton

affinity of carbenes versus PCy3. We find that both dimethyl carbene and ethyl methyl

carbene are more basic than PCy3. Proton-bound dimers of carbenes and PCy3 are also found to exhibit interesting reactivity.

iii

DEDICATION

To my daughter Renee Tongyu Yang

ACKNOWLEDGEMENTS

I would like to express my sincere appreciation to my advisor, Dr. Jeehiun K. Lee, for her numerious help and guidance over these years. I would also like to thank Dr. Ralf

Warmuth and Dr. Karsten Kroph-Jespersen in Chemistry and Chemical Biology

Department, and Dr. Brian Buckley from Environmental & Occupational Health Sciences

Institute, for serving on my committee and for their valuable time and helpful suggestions.

Especially thank to Dr. Brian Buckley and Dr. Ill Yang in EOHSI of Rutgers for their precious instrument time.

I would like to thank the previous and current members of Lee group: Su Pan,

Xiaofeng Shi, Meng Xu, Xuejun Sun, Anna Michelson, Daisy Cardoso, James Lim, Mu

Chen, Kai Wang, Sisi Zhang, Yuan Tian and Jorge Pavon for their friendship and help in both research and life. Many thanks to Dr. Alexei Ermakov for his incredible expertise in instrumentation. I also thank Dr. Laurence Romsted and Dr. Gene Hall for their helpful discussion.

I am very grateful to my parents and sister for their unconditional support and love at every step of my life.

At last, I thank to my dear husband Lin Yang for his love, patience and support.

TABLE OF CONTENTS

ABSTRACT OF THE DISSERTATION ...... ii

DEDICATION ...... iv

ACKNOWLEDGEMENTS ...... v

TABLE OF CONTENTS ...... vi

LIST OF FIGURES ...... x

LIST OF TABLES ...... xiii

1.1 Overview ...... 1

1.1.1 Gas phase acidity and proton affinity of nucleobases ...... 1

1.1.2 Gas phase acidity studies of organic silanols ...... 6

1.1.3 Gas phase proton affinity of N-heterocyclic carbenes (NHCs) ...... 7

1.2 Instrumentation ...... 8

1.2.1 FTMS ...... 8

1.2.2 ESI and ion trap mass spectrometer ...... 11

1.3 Methodology ...... 15

1.3.1 Bracketing method ...... 15

1.3.2 Cooks Kinetic Method ...... 18

1.3.3 Computational method ...... 20

Chapter 2 The Acidity and Proton Affinity of the Damaged Base 1,N6-

Ethenoadenine in the Gas Phase versus in Solution: Intrinsic Reactivity and

Biological Implications ...... 21

2.1 Introduction ...... 21

2.2 Experimental ...... 24

2.3 Results ...... 27

2.3.1 Computational results: tautomers ...... 27

2.3.2 Computational results: acidity ...... 27

2.3.3 Computational results: proton affinity ...... 28

2.3.4 Experimental results: acidity ...... 30

2.3.5 Experimental results: proton affinity ...... 33

2.4 Biological Implications ...... 40

2.5 Conclusions ...... 42

Chapter 3 The Gas-Phase Thermochemical Properties of the Damaged Base O6-

Methylguanine versus Adenine and Guanine ...... 43

3.1 Introduction ...... 43

3.2 Experimental ...... 44

3.3 Results ...... 44

3.3.1 Computational results: tautomers ...... 44

3.3.2 Computational results: acidities ...... 45

3.3.3 Computational results: proton affinities ...... 45

3.3.4 Experimental results: acidities ...... 46

3.3.5 Experimental results: proton affinities ...... 48

3.4 Discussion ...... 49

3.4.1 O6-methylguanine (OMG): tautomers, acidities and proton affinities ...... 49

vii

3.4.2 Comparison of adenine and guanine properties with O6-methylguanine

(OMG)...... 51

3.5 Conclusions ...... 57

Chapter 4 The Gas-Phase Thermochemical Properties of Pyrimidine Nucleobases

...... 60

4.1 Introduction ...... 60

4.2 Experimental ...... 61

4.3 Results ...... 62

4.3.1 Cytosine ...... 62

4.3.2 1-Methyl Cytosine ...... 67

4.3.3 Thymine ...... 70

4.4 Discussions ...... 74

4.4.1 Cytosine ...... 74

4.4.2 1-Methyl Cytosine ...... 81

4.4.3 Thymine ...... 81

4. 5 Conclusions ...... 82

Chapter 5 Gas Phase Acidity Studies of Dual Hydrogen-Bonding Organic Silanols and Organocatalysts ...... 85

5.1 Introduction ...... 85

5.2 Results ...... 87

5.2.1 Computational results ...... 87

5.3 Discussion ...... 100

5.3.1 Comparison of silicon versus carbon analogs ...... 102

viii

5.3.2 Gas phase versus solution ...... 107

5.4 Conclusions ...... 109

5.5 Experimental...... 110

Chapter 6 Proton Affinity of Phosphines versus N-heterocyclic Carbenes and

Reactivity of Carbene•Phosphine Dimers ...... 111

6.1 Introduction ...... 111

6.2 Results and Discussion ...... 112

6.2.1 Proton Affinity ...... 112

6.2.2 Dimer fragmentation...... 125

6.3 Conclusion ...... 128

6.4 Experimental ...... 129

Reference: ...... 132

CURRICULUM VITAE ...... 147

LIST OF FIGURES

Figure 1.1 Watson-Crick 5’-AGCT-3’/3’-TCGA-5’ duplex ...... 2

Figure 1.2 Structures of four nucleobases ...... 2

Figure 1.3 Silanol acidity ...... 7

Figure 1.4 Formation of carbene from imidazolium ion ...... 7

Figure 1.5 The motion of a charged ion in a magnetic field, B ...... 9

Figure 1.6 A cubic analyzer cell in FTMS ...... 11

Figure 1.7 Electrospray ionization ...... 12

Figure 1.8 Schematic of a 3D ion trap mass spectrometer ...... 13

Figure 1.9 Basic linear trap structure ...... 14

Figure 1.11 FTICR-MS ...... 15

Figure 1.12 Schematic of FT-ICR dual cell bracketing experiments ...... 16

Figure 2.1. Relative enthalpies (∆H in kcal mol-1) of the three most stable tautomers of

1,N6-ethenoadenine, calculated at B3LYP/6-31+G* (298 K) ...... 27

-1 Figure 2.2. Calculated acidities (∆Hacid; ∆Gacid in parentheses; all values in kcal mol ) of the two most stable tautomers of 1,N6-ethenoadenine at B3LYP/6-31+G* (298 K) 28

Figure 2.3. Calculated proton affinities (∆H; gas phase basicity values (∆G) in parentheses; all values in kcal mol-1) of the two most stable tautomers of 1,N6- ethenoadenine at B3LYP/6-31+G* (298 K) ...... 29

Figure 2.4. Calculated proton affinities (∆H; gas phase basicity values (∆G) in parentheses; all values in kcal mol-1) of 9-methyl-1,N6-ethenoadenine at B3LYP/6-

31+G* (298 K) ...... 30

Figure 3.1. Relative enthalpies (∆H in kcal mol-1) of the three most stable OMG tautomers and acidities (red values; ∆Hacid, with ∆Gacid in parentheses; all values in kcal mol-1), and proton affinities (blue values; PA, with GB in parentheses; all values in kcal mol-1) , calculated at B3LYP/6-31+G* (298 K) ...... 45

Figure 3.2. Rotation of the methyl group in the canonical O6-methylguanine structure, calculated at 298 K at B3LYP/6-31+G* ...... 50

Figure 3.3. Calculated proton affinity (PA, blue) and acidity (∆Hacid, red) values

(B3LYP/6-31+G*, 298 K, in kcal mol-1) for the atoms of the nucleobases involved in the G•C, OMG•C, and OMG•T base pairs ...... 52

Figure 3.4. Calculated N9-H ∆Hacid values for normal and damaged purines

(B3LYP/6-31+G*, 298 K) ...... 55

Figure 4.1. Relative enthalpies (∆H in kcal mol-1) of six possible tautomers of cytosine,

-1 and the acidities (red values; ∆Hacid, with ∆Gacid in parentheses; all values in kcal mol ), and proton affinities (blue values; PA, with GB in parentheses; all values in kcal mol-1) of the four most stable tautomers, calculated at B3LYP/6-31+G* (298 K) ...... 63

Figure 4.2. Relative enthalpies (∆H in kcal mol-1) of five possible tautomers of 1- methyl cytosine, and acidities (red values; ∆Hacid, with ∆Gacid in parentheses; all values in kcal mol-1), and proton affinities (blue values; PA, with GB in parentheses; all values in kcal mol-1) of the two most stable tautomers, calculated at B3LYP/6-31+G* (298 K)

...... 68

Figure 4.3. Relative enthalpies (∆H in kcal mol-1) of the two most stable thymine tautomers and acidities (red values; ∆Hacid, with ∆Gacid in parentheses; all values in kcal

mol-1), and proton affinities (blue values; PA, with GB in parentheses; all values in kcal

mol-1) of the canonical thymine tautomer, calculated at B3LYP/6-31+G* (298 K) ..... 71

Figure 5.1. Silanols and other hydrogen-bonding molecules examined in this paper.

Some carbon analogs were also examined and will be indicated by a prime; for

example, 2', the carbon analog of 2, is (Me)2PhCOH ...... 86

-1 Figure 5.2. Summary of data for substrates examined in this study (∆Hacid, kcal mol )

...... 101

Figure 5.3. Deprotonated diol structures, calculated at B3LYP/6-311++G(2df, p) .. 105

Figure 5.4. Organocatalyst and organosilicon neutral diol structures, calculated at

B3LYP/6-311++G(2df,p) ...... 107

Figure 6.1. N-Heterocyclic singlet carbenes ...... 112

Figure 6.2. Proton-bound dimer of PCy3 and NHC ...... 116

Figure 6.3. Calculated (B3LYP/6-31+G(d)) acidities of the 1-ethyl-3-

methylimidazolium ion (in kcal mol-1)...... 118

Figure 6.4. Possible paths by which proton-bound dimer of the methyl ethyl carbene

1b and PCy3 could form protonated carbene and protonated phosphine ...... 119

xii

LIST OF TABLES

Table 1.1 Representative gas-phase acidities ...... 4

Table 1.2 Representative gas-phase proton affinities ...... 4

Table 2.1. Summary of results for acidity bracketing of more acidic site of 1,N6- ethenoadenine...... 31

Table 2.2. Summary of results for acidity bracketing of less acidic site of 1,N6- ethenoadenine...... 32

Table 2.3. Summary of results for proton affinity bracketing of more basic site of 1,N6- ethenoadenine ...... 36

Table 2.4. Summary of results for proton affinity bracketing of more basic site of 9- methyl-1,N6-ethenoadenine ...... 38

Table 3.1. Summary of results for acidity bracketing of more acidic site of OMG. .... 46

Table 3.2. Summary of results for acidity bracketing of less acidic site of OMG...... 47

Table 3.3. Summary of results for PA bracketing of more basic site of OMG...... 48

Table 4.1. Summary of results for acidity bracketing of more acidic site of cytosine. 64

Table 4.2. Summary of results for acidity bracketing of less acidic site of cytosine.... 65

Table 4.3. Summary of results for PA bracketing of more basic site of cytosine...... 66

Table 4.4. Summary of results for acidity bracketing of more acidic site of 1-methyl cytosine...... 69

Table 4.5. Summary of results for PA bracketing of more basic site of 1-methyl cytosine...... 70

Table 4.6. Summary of results for acidity bracketing of more acidic site of thymine. . 72

xiii

Table 4.7. Summary of results for acidity bracketing of less acidic site of thymine. ... 72

Table 4.8. Summary of results for PA bracketing of more basic site of thymine...... 73

Table 4.9. Expected acidity bracketing table if cytosine tautomers 1, 1a, 1b, and 1c are

present...... 78

Table 5.1. Computational results for acidity of silanols and their carbon analogs ...... 87

Table 5.2. Computational results for acidity of known hydrogen-bonding catalysts ... 88

Table 5.3. Summary of results for acidity bracketing of triethylsilanol (1)...... 89

Table 5.4. Summary of results for acidity bracketing of dimethylphenylsilanol (2). ... 90

Table 5.5. Summary of results for acidity bracketing of triphenylsilanol (3)...... 91

Table 5.6. Summary of results for acidity bracketing of di-tert-butylsilanediol (4). .... 92

Table 5.7. Summary of results for acidity bracketing of diphenylsilanediol (5)...... 93

Table 5.8. Summary of results for acidity bracketing of (4-

fluorophenyl)(mesityl)silanediol (6)...... 93

Table 5.9. Summary of results for acidity bracketing of 1,1,3,3-

tetraisopropyldisiloxane-1,3-diol (8)...... 94

Table 5.10. Summary of results for acidity bracketing of 2-methyl-phenyl-ethanol (2').

...... 96

Table 5.11. Summary of results for acidity bracketing of triphenylmethanol (3’)...... 97

Table 5.12. Summary of results for acidity bracketing of BINOL (10)...... 97

Table 5.13. Summary of results for acidity bracketing of TADDOL (11)...... 98

Table 5.14. Summary of results for acidity bracketing of N,N-diphenylthiourea (12). 99

Table 5.15. B3LYP/6-311++G(2df,p) calculated ∆Hacid values for silanols X and

carbon analogs X'...... 102

xiv

Table 5.16. Comparison of catalytic activation of carbonyl compounds in a Diels Alder reaction using various silanols and alcohols...... 108

Table 6.1. Proton Affinity Calculations on 1-Ethyl-3-methylimidazol-2-ylidene 1b . 113

Table 6.2. Summary of results for acidity bracketing of PCy3 ...... 114

Table 6.3. Summary of results for acidity bracketing of HPCy2 ...... 114

Table 6.4. Experimental and computational proton affinity values for various phosphines and 1-ethyl-3-methylimidazol-2-ylidene 1b ...... 115

Table 6.5. Natural logarithm of product ion abundance ratios R for CID of the proton- bound dimers of 1-ethyl-3-methylimidazol-2-ylidene 1b with various diamines to obtain PAa ...... 121

Table 6.6. Summary of results for PA bracketing (in quadrupole ion trap) of more basic site of carbene 1b ...... 122

Table 6.7. Summary of results for PA bracketing (in quadrupole ion trap) of more basic site of carbene 1a ...... 123

Table 6.8. Computational (B3LYP/6-31+G*) proton affinities and TEP values for selected carbenes...... 124

Chapter 1 Introduction

1.1 Overview

1.1.1 Gas phase acidity and proton affinity of nucleobases

Deoxyribonucleic acid (DNA) is a nucleic acid which is responsible for storing genetic information for the biological development of all known cellular life and some viruses. DNA consists of two complementary strands,1 which is also known as a

“polynucleotide”. Each monomer unit, the nucleotide, contains a 5-carbon sugar, a nucleobase attached to the sugar and a phosphate group. The esterification of the 3’ hydroxyl group with the phosphate on the adjacent nucleotide (Figure 1.1) leads to the polymerization of the DNA building blocks. Four nucleobases are found in DNA (Figure

1.2): adenine (A), cytosine (C), guanine (G) and thymine (T). In a DNA double helix, the two strands are associated through hydrogen bonding (base pairing) and pi stacking. A always bonds with T and C always bonds with G. A Watson-Crick model2 of two tetramer strands (AGCT/TCGA) is shown in Figure 1.1.

OH O- O O P O N NH2 O T O N N HN - A O O O P O O- N N O O H2N N O P O O O - N C O N N - G O O N O O P O N NH2 O O H2N N N O P O O O - G O N HN O- C N O O O P O N NH2 O N N O- O P O O O A 5' end O- N N N O T H2N N O O P O O O- OH 3' end

Figure 1.1 Watson-Crick 5’-AGCT-3’/3’-TCGA-5’ duplex

Figure 1.2 Structures of four nucleobases

Acid-base equilibria involving nucleobases have been widely investigated in polar and apolar solvents by mass spectrometry.3 The determination of acidity and proton

affinity of nucleobases is essential for understanding the chemical processes that DNA

molecules undergo,4 because the interaction energy between two complementary strands

depends on the intrinsic basicity of the acceptor atoms and on the acidity of the NH donor

groups.5,6 Also, the intrinsic acidity and basicity of nucleobases play significant roles in biosynthetic reactions for which nucleobases are substrates.

The gas phase is a particularly valuable environment where to examine the intrinsic properties and reactivity of molecules, which can be quite different in solution. For example, ion-molecule reactions have different mechanisms in gas phase and solution phase. In solution phase, ions and molecules are strongly solvated and energy is required to break the solvation cage to form the activated complex, while in the gas-phase, the formation of ion-molecule complexes is exothermic.7 Many biological media, from

intracellular environs to the interior of proteins, are nonaqueous. The gas phase is the

“ultimate” nonpolar environment in which any interference of solution effects can be excluded.

o The gas phase acidity is defined as the positive enthalpy change ( ΔH acid ) associated

with the deprotonation of a molecule HA to form H+ and A- ions (Eq. 1.1). The gas phase

proton affinity (PA) is defined as the negative enthalpy change associated with the

protonation of a molecule B to form the HB+ ion (Eq. 1.2).

Eq. 1.1

Eq. 1.2

The gas phase acidities and proton affinities of a series of prototypical compounds

are listed in Table 1.1 and Table 1.2, respectively.8 The gas phase acidity ranges from

314~417 kcal mol-1 with the higher value corresponding to the lower acidity. The gas phase proton affinity ranges from 130~291 kcal mol-1 with the higher value

corresponding to the higher proton affinity.

Table 1.1 Representative gas-phase acidities

o -1 Class of molecules Specific example ΔH acid (kcal mol )

Alkanes CH4 416.7

Vinylic C2H4 409.4

Amines NH3 403.6

Aromatic C6H6 400.7

Allytic CH2CHCH3 390.8

Water H2O 390.7

Alcohols CH3OH 381.7

Benzylic PhCH3 380.8

Acetylenes C2H2 377.8

Nitriles CH3CN 372.9

α to Carbonyls CH3COCH3 369.1

Thiols CH3SH 356.9 Phenols PhOH 349.2

Carboxylic acids CH3COOH 348.6 Benzoic acids PhCOOH 340.2

Strong acids HI 314.4

Table 1.2 Representative gas-phase proton affinities

Class of molecules Specific example PA (kcal mol-1)

Alkanes CH4 129.9 Strong acids HI 150.0

Acetylene C2H2 153.3

Vinylic C2H4 162.6

Water H2O 165.2

Formaldehyde CH2O 170.4

Aromatic C6H6 179.3

Alcohols CH3OH 180.3

Carboxylic acids CH3COOH 187.3

Thiols C2H5SH 188.7

Nitriles CH3CNC2H5CN 189.8

Benzonitrile C6H5CN 194.0

Ethers CH3OCH2CH2CH3 194.8

Amides C2H5NO 203.5

Amines NH3 204.0

Aniline C6H5NH2 210.9

Pyridine C5H5N 222.3 Hydride HNa 261.7 Monoxide BaO 290.5

In previous work, we have reported the gas phase thermochemical properties of

adenine and uracil as well as the damaged base hypoxanthine.9-13 Damaged DNA bases differ in structure and properties from normal nucleobases and intervene with gene replication and expression, leading to cell death, aging, and carcinogenesis; therefore, they must be repaired.14-17 Our studies are motivated by understanding the mechanisms

by which mutated bases are cleaved, focusing on the glycosylase enzymes that excise

damaged bases.14,17,18 In Chapter 2 and Chapter3, we will describe our gas phase property

studies of damaged bases, 1,N6-ethenoadenine and O6-methylguanine, from which we

seek to uncover how damaged bases are different from normal bases and why

glycosylases recognize certain damaged bases, but leave other damaged or normal bases

untouched. In Chapter 4, we will focus on pyrimidine nucleobases. Characterizing the

naturally occurring normal bases is the first step to understand how damaged bases

different from normal bases.

1.1.2 Gas phase acidity studies of organic silanols

Silanols (SiOH), analogues of alcohols, are compounds containing Si atom with one

or more –OH groups attached to it. Inorganic silanols provide reactive hydroxyl groups

on the surface of silicon-containing materials, for example, silicate rocks, zeolites and

silica gel. These hydrogen bonds play an important role in materials, catalysts and

chromatographic separations.

Silanols have applications ranging from organic synthesis to biological activity.19

Since silanols have both donor (-OH) and acceptor groups (-OH), they can easily form hydrogen bonds with themselves or with other molecules. In solution, silanols are always more acidic than their carbon analogs (alcohols).19,20 The enhanced acidity is due to both

the back-bonding interactions from the oxygen into the σ* C-Si vacant orbital and into

the silicon d-orbitals as well (Figure 1.3), while the latter back-bonding interaction is not

accessible for alcohols. Silanols are expected to be used as good hydrogen-bonding

catalysts for enantioselective carbon-carbon bond forming reactions, such as Strecker reaction and Diels-Alder reaction. We collaborate with Prof. Annaliese Franz at UC

Davis, who is developing a new class of hydrogen-bonding catalysts based on organic silanols. In Chapter 5, we provide a comprehensive examination of the gas phase acidity of various types of silanols including monosilanols, silanediols and disiloxanediols. We

are interested in the acidity of silanols versus that of alcohols in both gas phase and solution phase and how the different substituents affect the acidity of silanols and alcohols. At last we will assess the relationship between thermochemical properties and

catalyst activity.

Figure 1.3 Silanol acidity

1.1.3 Gas phase proton affinity of N-heterocyclic carbenes (NHCs)

Stable N-Heterocyclic carbenes (NHCs) were first reported by Arduengo and

coworkers in 1991 and have demonstrated broad applications in organic synthesis.21-28

For example, NHCs are effective novel ligands for transition-metal-catalyzed reactions, such as Grubbs ruthenium olefin metathesis reaction, palladium-catalyzed cross-coupling reactions and nickel-catalyzed cycloadditions.29-33 In addition, the protonated NHCs are

known as imidazolium-type ionic liquids which are an important class of “green”

nonvolatile solvents that have been increasingly used in organic synthesis.34-39 Figure 1.4

shows the formation of NHCs from dialkylimidazolium salts. Also, such species have

biological counterparts: the catalytically active part of coenzyme thiamine (vitamin B1) is

a nucleophilic carbene.40-43

Figure 1.4 Formation of carbene from imidazolium ion

Despite the broad applications of NHCs, the experimental thermochemical

properties of these carbenes are surprisingly little known, both in solution and in the gas

phase.26,44 We are particularly interested in their proton affinity (acidity of imidazolium cations) in the gas phase. More basic carbenes will presumably be more effective ligands for olefin metathesis, because it is generally accepted that NHCs bind to metals via σ-

bonding with negligible π-back-bonding.45-47 In terms of ionic liquids, it has been found that imidazolium moieties tend to be deprotonated in the base-catalyzed reactions. So the acidity of imidazolium ions or the proton affinities of NHCs is highly desired. In Chapter

6, we will report our computational and experimental study of the gas phase proton affinity of NHCs. Another very important aspect of our study is to compare the proton affinity of NHCs with phosphines. Phosphines comprise the ligands of the first generation of Grubbs olefin metathesis catalysts, but are not as active as the NHCs. In

Chapter 6, we will compare the proton affinity of carbenes with the most commonly used triphenylphosphine (PCy3) in olefin metathesis catalysis. Also some interesting reactivity

is found for the proton-bound dimers of carbenes and phosphine.

1.2 Instrumentation

1.2.1 FTMS

Mass spectrometry has recently become a powerful tool for investigating

biomolecules due to newly developed ionization methods, such as electrospray ionization

(ESI)48 and matrix-assisted laser desorption and ionization (MALDI).49 Fourier

Transform Ion Cyclotron Resonance (FT-ICR), also known as Fourier Transform Mass

Spectrometry (FTMS), is a type of mass analyzer whose design is based on the cyclotron

frequency of the ions in a fixed magnetic field. It is evaluated as the most complex

method in terms of mass analysis and detection. Recently FT-ICR has attracted great

attention for its ability to make mass measurements with a combination of higher resolving power and accuracy than other mass spectrometers. However, it requires an experienced operator and high cost.50,51

In the 1930’s, Lawrence developed the theory of cyclotron resonance. Later in the

early 1970’s, Alan Marshall and Melvin Comisarow applied the Fourier Transform to this

technique and it soon became an important academic tool.52 FT-ICR theory is based on

the motion of a charged ion (including positive charge and negative charge) in a magnetic

field, B (Figure 1.5). Herein, v is velocity, r is radius of gyration, F1 is magnetic force (or

Lorentz force), F2 is centrifugal force.

F2 F2 v v

F1 F1

⊗

Figure 1.5 The motion of a charged ion in a magnetic field, B

Lorentz force FqvB1 = Eq. 1.3

2 Centrifugal force F2 = mv r Eq. 1.4

where q is the charge on the ion and m is the mass of the ion. When the two forces are

balanced by each other (Eq. 1.5), ions can be stabilized on a circular trajectory with a

frequency of f ( f = vr2π ) (Eq. 1.6).

qvB= mv2 / r Eq. 1.5

f = qB2π m Eq. 1.6

Eq. 1.6 shows that the m/z ratio only depends on the cyclotron frequency, f, and magnetic strength, B. However, at this stage, no signal is observed because the radius of

the motion is very small. Therefore, ions must be excited to be detected. Figure 1.6

shows one type of analyzer cell, a cubic cell, which is oriented in the magnetic field and

composed of three pairs of plates: trapping plates, detection plates and excitation plates.

Trapping plates are perpendicular to the magnetic field. The hole in the center of each of

the trapping plates allows ions or electrons to enter the cell along the magnetic field lines.

Excitation plates and detection plates are parallel to the magnetic field. A swept radio

frequency pulse is applied across the two excitation plates. Each individual excitation

frequency is associated with the ions’ natural motion. Ions with the same m/z are always

grouped as “packets”. When they are excited to a larger trajectory and get closer to the detection plates, an alternating current is induced between the detection plates. This

alternating current is usually called the image current. Then the image current is

amplified, quantified and transformed into frequency domain signals.

Detection plate

Excitation plate

B Trapping plate

Figure 1.6 A cubic analyzer cell in FTMS

1.2.2 ESI and ion trap mass spectrometer

Electrospray ionization (ESI) is one of the atmospheric pressure ionization (API)

sources, which ionize the sample at atmospheric pressure and then transfer the ions into

the mass spectrometer. ESI technique is so called “soft” ionization, which means the

sample typically does not fragment. Therefore it is especially useful in ionizing polymers

and biological macromolecules such as proteins/peptides and DNA/RNA. The process of

ESI is illustrated in Figure 1.7.53 The sample is dissolved in a polar and volatile solvent

(water, methanol or acetonitrile, etc.) and injected through a narrow silica or stainless steel capillary at a flow rate of 1 µL/min~1 mL/min. By applying a strong electric field

(3~6kV) at the capillary tip under atmospheric pressure, a charge accumulation at the

liquid surface is induced. Taking the positive mode as an example (Figure 1.7), cations are enriched on the liquid surface and anions move towards to conductive tip. As the effect of the electric field overcomes the surface tension, the highly charged droplet deforms to a conical shape rather than spherical shape which is caused by surface tension effect.54,55 This cone is so-called “Taylor cone”.56 Then the tip of the cone elongates into a filament, which breaks apart and emits a stream of charged droplets. The charge density is increased with the evaporation of the solvent and Coulombic forces caused by the

repulsion of cations further produces even finer droplets. The process of evaporation and

repulsion repeats until fully desolvated cations are released.

Figure 1.7 Electrospray ionization

The quadrupole ion trap (QIT) is a mass spectrometer with high sensitivity and specificity.57 It was invented by Wolfgang Paul who received the Nobel Prize in 1989 for his work.58 It is also known as 3D (Paul) trap. The 3D trap consists of three hyperbolic electrodes (one ring and two endcap electrodes, Figure 1.8). Firstly, ions are generated by ESI (Ion Generation) and focused using two octapole transmission systems (Ion

Focusing), and then the ions can be trapped, excited and ejected in the cavity formed by these three electrodes which are applied by both AC and DC (Ion Analysis). Each endcap electrode has a small hole through which ions can travel in and travel out. Damping gas

(helium, ~1m Torr) is filled in the ion trap and dampens the kinetic energy of “hot” ions.

And then the ions move toward the center of the ion trap with the aid of an AC applied to the ring electrode. By altering the amplitude of the AC, the ions are destabilized and ejected through the holes in the endcap electrodes, and at last, detected by a collision dynode and electron multiplier system (Detection).

Figure 1.8 Schematic of a 3D ion trap mass spectrometer

Another configuration of the ion trap is the 2D linear trap.59 Unlike the 3D ion trap,

the linear trap itself is a square array of precision-machined and precision-aligned

hyperbolic electrodes (Figure 1.9). Each hyperbolic electrode is cut into three sections

with lengths of 12 mm, 37 mm and 12 mm, respectively. Ions are confined radially in the

center section by applying a RF voltage and axially by DC potentials on end electrodes.

As illustrated in Figure 1.10, ions are transmitted to the entrance of the ion trap after ionization by ESI, and then focused. After acceleration, ions are moving into the linear ion trap, where the ions are detected by two detectors. Then the ion current signals are

amplified and passed on to the data system. Compared with 3D ion trap, 2D linear ion

trap has many advantages, such as larger ion storage capacity due to the linear

configuration, faster scan time and higher sensitivity.

Figure 1.9 Basic linear trap structure

Figure 1.10 Schematic of a 2D linear ion trap mass spectrometer

1.3 Methodology

1.3.1 Bracketing method

The gas phase acidities and proton affinities are measured using bracketing methods.

The experiments are conducted on the Finnigan 2001 Fourier Transform Ion Cyclotron

Resonance Mass Spectrometer (FTICR-MS) with a dual cell setup (Figure 1.11).

Finnigan 3 T Magnet 2001 Batch Inlet Batch Inlet

Electron Solids Probe ESI beam

Pulsed Valve Inlet

Dual Cell Diffusion Pump

Figure 1.11 FTICR-MS

In the FT-ICR, the two adjoining 1-inch cubic cells are positioned colinearly with the magnetic field produced by a 3.3 T superconducting magnet. The cells are called, traditionally, the source cell (on the “left” side as you face the instrument) and the analyzer cell (on the “right” side). The pressure of the dual cell is pumped down to less than 1×10-9 torr. Ions can be transferred between the source and analyzer cells via a 2- mm hole in the center of the trapping plate. We use argon to cool down the transferred

ions. Neutral samples are introduced into the FT-ICR using heated solids probe, heated batch inlet, pulsing valves and leaking valves. Acidity and proton affinities are assessed using bracketing experiments in the FTMS. For acidity bracketing, hydroxide ions are generated first by pulsing water (via a pulsed valve system) into the FTMS cell and sending an electron beam (8 eV, 6 µA) through the center of the cell. The hydroxide ions deprotonate neutral molecules “M” (either unknown or reference acids) to yield the [M-

H]– ions. The [M-H]– ion is allowed to react with the unknown or reference acid. The

reactions for both directions are shown in Figure 1.12a and 1.12b. The same procedure is used for bracketing proton affinity, where hydronium ions (20 eV, 6 µA) are used for protonation. The occurrence of proton transfer is regarded as evidence that the reaction is

exothermic.

Figure 1.12 Schematic of FT-ICR dual cell bracketing experiments

We have developed an FTMS method for the bracketing of the acidity and basicity of

less acidic and basic sites in nucleobases that have multiple acidic and basic sites.9-12

Acidity measurement is used here as an example. In this setup, nucleobase ions produced after reaction of the corresponding neutral with hydroxide ions are immediately removed from the first cell and transferred into the second cell. Reference acids are then introduced into the second cell via a batch inlet system and allowed to react with nucleobase ions. The first reaction cell is high in neutral nucleobase concentration, and over time, neutral-catalyzed isomerization leads to survival of only the most acidic ions.

Transferring ions into the second cell immediately after their generation allows us to carry out the reaction between reference acids and nucleobase ions in the absence of neutral nucleobase. The same procedure can be applied to the bracketing of the basicity of less basic sites as well.

Reaction efficiency is used to assess occurrence or non-occurrence of a proton transfer reaction (Eq. 1.7); the cutoff is 10%.

k efficiency = exp ×100% Eq. kcoll

1.7

If the efficiency is higher than 10%, proton transfer occurs. On the other hand, if

the efficiency is less than 10%, we assume there is no proton transfer for that reaction. In

Eq. 1.7, the theoretical ion-molecule collision rate constant kcoll is obtained from the

Average Dipole Orientation (ADO) program60,61 and represents the rate constant for a

reaction in which every ion-molecule collision rate constant for either the reaction of

deprotonated (protonated) unknown reacting with the neutral reference acid (base), or the

conjugate base (acid) ion reacts with neutral unknown. This program estimates kcoll based

18 on the dipole moment, polarizability, mass of the neutral molecule and the mass of the ion (Eq. 1.8).

1/ 2 1/ 2 1/ 2 kcoll = (2πq μ )[α + Cμ D (2 πkT) ] Eq. 1.8 where μD is the dipole moment of the neutral molecule, μ is the reduced mass of the ion- molecule system, q is the charge of the ion, C is the dipole locking constant, and α is the polarizability of the neutral.62-64

The experimental rate constant kexp is calculated by using Eq. 1.9:

− slope kexp = Eq. 1.9 Pneutral × Φ where Φ = conversion factor = 3.239 × 1016 molecule·cm-3·torr-1. The slope in Eq. 1.9 is obtained by plotting the natural logarithm of the relative intensity of reactant ions vs. reaction time. In our experiments, we have pseudo-first-order conditions, where the amount of the neutral substrates is in excess relative to the reactant ions. Taking acidity measurement as an example, we “back out” the neutral pressure Pneutral from a control reaction where hydroxide reacts with neutral substrate. Because hydroxide is very basic,

9,10,60,61 we assume this reaction proceeds at the theoretical collision rate kcoll’. We can then use the calculated kcoll’ to “back out” Pneutral (Eq. 1.10). The slope’ is achieved by plotting the disappearance of hydroxide ions vs. reaction time.

− slope' Pneutral = Eq. 1.10 kcoll '×Φ

1.3.2 Cooks Kinetic Method

We also used the Cooks kinetic method in a quadrupole ion trap (LCQ) or linear ion trap (LTQ) to measure the acidities and proton affinities.65-68 For PA experiments, this

19 method involves the formation of a proton-bound complex, or dimer, of the unknown

(“A” in Eq. 1.11) and a reference base of known proton affinity (“Bi” in Eq. 1.11).

Collision-induced dissociation (CID) of this dimer leads to the formation of either the protonated unknown or the protonated reference base. The ratio of these two products yields the relative proton affinities of the two compounds of interest, assuming that the dissociation has no reverse activation energy barrier and that the transition state is late thus indicates the stabilities of the products. These assumptions are generally true for proton-bound systems.69-71 Similar procedures can be applied for acidity measurements.

++k1 (Eq. 1.11) []AHBii⎯⎯→+ AH B

k2 + ⎯⎯→+Bi HA (Eq. 1.12)

++⎯⎯K → BiiHA+ ←⎯⎯ AHB+ (Eq. 1.13)

Kkk≈ 12/ (Eq. 1.14)

PA() A− PA ( Bieff )≈ RT ln K (Eq. 1.15)

1 ln(kk12 / )=− ( PAAPAB ( ) (i ) (Eq. 1.16) RTeff

Teff is the effective temperature of the proton bound complex in Kelvin. A plot of ln(k1/k2) versus the proton affinity of a series of reference bases (PA(Bi)) will yield Teff from the slope; this value varies depending on the substrate. For all the measurements we repeat at least three times to ensure reproducibility.

1.3.3 Computational method

Theoretical calculations are conducted to predict the tautomerism, acidities and proton affinities of molecules.

The GAUSSIAN03 and GAUSSIAN09 programs were used.72 The B3LYP method and the 6-31+G(d) basis set is used for all the gas phase calculations. Additionally, the larger basis set 6-311++G(2df,p) basis set is used as well for calculating the acidities of silanols and their carbon analogs. In chapter 6, besides B3LYP/6-31+G(d), other levels such as RHF/6-31+G(d), M06-2X/6-31+G(d), MP2/6-311+G(2d,p)//B3LYP/6-31+G(d),

M06-2X/aug-cc-pVTZ//B3LYP/6-31+G(d) and CBS-QB3 are used to calculate the proton affinity of 1-ethyl-3-methyl carbene. Also we calculated PCy3 PA using MP2/6-

311+G(2d,p)//B3LYP/6-31+G(d). All the geometries are fully optimized and frequencies are calculated and used without using any scaling factor. The relative stabilities of tautomers in nucleobases, the acidities and PAs are reported at 298 K. Solvation studies are conducted using the conductor-like polarizable continuum model (CPCM), where molecules are optimized at B3LYP/6-31+G(d) and UAKS radii are used.73,74 A dielectric constant of 78.4 is used to simulate an aqueous environment.

Note: Major parts of the following chapter have been published: Liu, M.; Xu, M.; Lee, J.

K. J. Org. Chem. 2008, 73, 5907-5914.

Chapter 2 The Acidity and Proton Affinity of the Damaged Base 1,N6-

Ethenoadenine in the Gas Phase versus in Solution: Intrinsic Reactivity

and Biological Implications

2.1 Introduction

1,N6-Ethenoadenine (εA, 1) is a highly mutagenic DNA lesion linked to cytotoxicity and carcinogenesis.15,75-80 Exogenous sources leading to εA formation include vinyl chloride, ethyl carbamate, crotonaldehyde, N-nitrosopyrrolidine, chloroethylene oxide, and chloroacetaldehyde.18,75,81-84 The εA adduct may also be produced in vivo via lipid peroxidation.81-83

As we addressed in the first chapter, damaged DNA bases differ in structure and properties from normal nucleobases. They intervene with gene replication and expression and must be repaired.14,17 The εA lesion is repaired in humans by the base excision repair

(BER) pathway, which involves DNA glycosylase enzymes.14,17,85 The glycosylase responsible for εA excision in humans is alkyladenine DNA glycosylase (AAG). AAG excises a wide variety of damaged nucleobases from double-stranded DNA, including εA

(1), hypoxanthine (2), 3-methyladenine (3) and 7-methylguanine (4).14,17,86-94 An outstanding mechanistic question is how AAG cleaves such a broad range of bases, yet leaves normal bases adenine (5) and guanine (6) untouched.14,17,89,91-98

10 11 N O NH2 O H3C 7 6 12 7 H 7 H H N N 1 N 1 N 7 1 N1 N N N H H H H 8 2 8 2 8 2 8 2 N N N N 9 N H 9 N H 9 N3 H 9 N NH2 H 3 H 3 H H 3 CH3 1 2 3 4

NH O 7 2 7 1 1 H N N N N H H 8 2 8 2 N N 9 N H 9 N NH2 H 3 H 3 5 6

Possible mechanisms of depurination include cleavage via the departure of deprotonated 1,N6-ethenoadenine or cleavage via departure of neutral 1,N6-ethenoadenine

(which requires pre-protonation of the damaged nucleobase).9,10,14,17,91,93,99-103

Scheme 2.1

10 N 7 N N1 N 10 O 9 N3 O N O P O O P O 7 O AAG O N DNA O- 1' DNA O- 1' N1 O O N 9 N3 OOP - OOP - O O DNA DNA

We have set forth the mechanistic proposal that AAG cleaves mutated purine substrates such as εA as anions (Scheme 2.1).9,10,99,100,103 Such a mechanism has also been proposed for a related enzyme, thymine DNA glycosylase (TDG), which cleaves pyrimidine nucleobases.102 We postulate that AAG differentiates among substrates by cleaving those nucleobases which are the most facile to remove. For the mechanism shown in Scheme 2.1, the ease of excision should be related to how good of a leaving group the anionic, deprotonated nucleobase is. Our theory is that deprotonated damaged bases are better leaving groups than deprotonated normal nucleobases. Furthermore, we propose that the differences among the leaving group abilities of various anionic, deprotonated nucleobases are enhanced in a nonpolar environment; by providing such an environment, AAG is able to execute its broad specificity.9,91,92,94,102,103

Our goal is therefore to ascertain whether the deprotonated forms of the damaged bases excised by AAG are in fact better leaving groups than the normal bases. We assess leaving group ability by ascertaining the acidity of the damaged bases: the more acidic the base at N9, the better a leaving group its conjugate base should be. Furthermore, we assess the acidity in the gas phase, which is the 'ultimate' nonpolar environment; our theory is that any differences in acidity between 1,N6-ethenoadenine and normal bases should be greater in the gas phase than in solution. To our knowledge, the gas phase acidity of 1,N6-ethenoadenine is unreported. In this chapter, we describe a study of the acidic and basic properties of 1,N6-ethenoadenine in an effort to both characterize this damaged base and also to probe how it might be recognized by AAG.

2.2 Experimental

All chemicals except 9-methyl-1,N6-ethenoadenine are commercially available and were used as received. 9-Methyl-1,N6-ethenoadenine was synthesized following literature procedure.104,105

Acidity and proton affinity experiments were conducted using bracketing method on Fourier Transform Ion Cyclotron Resonance Mass Spectrometer (FTMS), which has been described previously.9,11 The occurrence of proton transfer is regarded as evidence that the reaction is exothermic (“+” in Tables, see Results section). In our experiments, we have pseudo-first order conditions, where the amount of the neutral substrate is in excess relative to the reactant ions. Reading the pressure from an ion gauge is often unreliable, both because of the gauge's remote location as well as varying sensitivity for different substrates.10,106 We therefore "back out" the neutral pressure from a control reaction. Briefly, we obtain the pseudo-first order rate constant for the reaction of hydroxide and the neutral substrate. Because hydroxide is very basic, we assume this reaction proceeds at the theoretical collision rate.60,61 We can then use the calculated collisional rate constant to “back out” the neutral pressure.

We also used the Cooks kinetic method in a quadrupole ion trap (LCQ) mass spectrometer48,67,69,107-109 to measure the proton affinities of the nucleobases and to conduct a relative acidity study of εA, adenine and guanine. For the proton affinity studies, the proton-bound complex ions are generated by electrospray (ESI). For each experiment, a solution of the nucleobase and reference base is prepared (10-3 to 10-4 M solutions in methanol; a small amount of acetic acid is also added). An electrospray needle voltage of ~4.5 kV was used. The flow rate is 25 μL/min. The proton-bound

25 complex ions were isolated and then dissociated by applying collision-induced dissociation (CID); the complexes were activated for about 30 ms. Finally, the dissociation product ions are detected to give the ratio of the protonated analyte and protonated reference base. A total of forty scans was averaged for the product ions.

The Cooks kinetic method involves the formation of a proton bound complex, or dimer, of the unknown A and a reference base of known proton affinity (eq. 2.1),

⊕ k1 ⊕ k2 ⊕ AH + Bi ←⎯⎯[A⋅⋅⋅ H ⋅⋅⋅ Bi ] ⎯⎯→ A + Bi H Eq. 2.1

6 where A represents 1,N -ethenoadenine and Bi denotes a series of reference bases with

67,69,107-109 ⊕ known proton affinities. The proton bound dimer [A⋅⋅⋅ H ⋅⋅⋅ Bi ] is dissociated via collision-induced dissociation (CID). The rate constants k1 and k2 are for the two different dissociation pathways. The relationship of these rate constants to PA is shown in eq. 2.2:

⊕ k1 AH ΔG(Bi ) − ΔG(A) ΔH (Bi ) − ΔH (A) PA(A) − PA(Bi ) ln = ln ⊕ = ≈ = Eq. 2.2 k2 Bi H RTeff RTeff RTeff

110 where R is the gas constant and Teff is the effective temperature of the activated dimer67,69,107-109. The ratio of the amounts (intensities) of the two protonated products yields the relative proton affinities of the two compounds of interest, assuming the dissociation has no reverse activation energy barrier and that the dissociation transition structure is late and therefore indicative of the stability of the two protonated products.

These assumptions are generally true for proton bound systems.69-71 In order to obtain the proton affinity of compound A, the natural logarithm of relative intensity ratios is plotted versus the proton affinities of a series of reference bases, where the slope is (-1/RTeff) and

26 the y-intercept is (PA(A)/ RTeff). The Teff is obtained from the slope. The proton affinity of compound A, (PA(A)), is calculated from either eq. 2.2 or the y-intercept.

For the εA acidity studies, the Cooks kinetic method was used on the deprotonated dimers of εA•adenine and εA•guanine to assess relative values. The deprotonated

εA•adenine or εA•guanine complex ions are generated by ESI.48 For each experiment, a solution of εA and adenine or εA and guanine is prepared (10-3 M solutions in water mixed with 20% ethanol; a small amount of formic acid is needed to dissolve guanine).

As with the proton affinity measurements, an electrospray needle voltage of ~4.5 kV was used and the flow rate is 25 μL/min. CID activation of the isolated deprotonated dimers is for 30 ms and a total of forty scans was averaged for the product ions. A Teff of 434 K, obtained from calibration experiments with guanine, was used.

6 We used two methods to measure the pKa of 1,N -ethenoadenine. In the first, we simply prepared a solution of a known concentration of εA, and then added a 0.5 stoichiometric amount of sodium hydroxide; we then assume that pH = pKa. In the second method, we utilized changes in absorbance at λ=240 nm versus pH.111 We measured the pKa five times and report the averaged value.

The B3LYP method and the 6-31+G* basis set as implemented in Gaussian03 were used for all the gas phase calculations.112-116 All the geometries are fully optimized and frequencies are calculated; no scaling factor is applied. Reported values herein are at 298

K.117 Solvation studies were conducted using the CPCM method (full optimizations at

B3LYP/6-31+G*; UAKS radii as implemented in Gaussian03).73,74 A dielectric constant of ε=78.4 was used to simulate an aqueous environment.

2.3 Results

2.3.1 Computational results: tautomers

As is common with nucleobases, 1,N6-ethenoadenine has several possible tautomers.

We calculated the relative enthalpies of all eight possible tautomers of 1,N6- ethenoadenine at B3LYP/6-31+G*. The most stable tautomer in the gas phase is calculated to be the “N7” tautomer 7 (proton resides on the N7). The canonical structure

1, which is the biologically relevant “N9” tautomer, is calculated to be less stable than the

“N7” tautomer by 0.7 kcal mol-1. The next most stable “N10” tautomer 8 (proton resides on the N10) is 12.8 kcal mol-1 less stable than the N7 tautomer. Therefore our calculations indicate that most likely under our gas phase conditions only the N7 and/or N9 tautomers will be present (Figure 2.1).

H H 10 10 H 10 H N 11 H N 11 N 11 7 6 6 7 6 N 5 H 7N 5 H 5 H N N N N H 1 H 1 H 1 8 8 8 N 4 N 2 H N 4 N 2 H N 4 N 2 H H 9 3 9 3 9 3

1 (N9) 7 (N7) 8 (N10) 0.7 0 12.8

Figure 2.1. Relative enthalpies (∆H in kcal mol-1) of the three most stable tautomers of 1,N6-ethenoadenine, calculated at B3LYP/6-31+G* (298 K)

2.3.2 Computational results: acidity

The acidities of the eight possible tautomers were calculated, but herein we only discuss the most stable tautomers 1 and 7 (Figure 2.2). The most acidic site of the N9

-1 -1 tautomer 1 is N9-H (∆Hacid = 330.7 kcal mol ; ∆Gacid = 323.2 kcal mol ). The remaining protons are all C-H protons that are much less acidic. For the N7 tautomer 7, the most

-1 acidic N7-H proton has a computed acidity of 331.3 (∆Hacid)/323.9 (∆Gacid) kcal mol , which is quite close to the acidity of the N9-H site of the N9 tautomer 1. The remaining less acidic sites are all C-H protons (Figure 2.2).

H 391.2 (383.2) H 387.0 (379.1) 10 331.3 (323.9) 10 N 11 H N 11 7 6 6 372.4 (364.7) N 5 H 7 5 H N N N 370.4 (362.7) H 1 373.0 (365.3) H 1 8 8 N 4 N 2 H 371.5 N 4 N 2 H H 9 3 368.3 (360.5) (363.7) 9 3 372.0 (364.0) 330.7 (323.2)

1 (N9) 7 (N7)

-1 Figure 2.2. Calculated acidities (∆Hacid; ∆Gacid in parentheses; all values in kcal mol ) of the two most stable tautomers of 1,N6-ethenoadenine at B3LYP/6-31+G* (298 K)

2.3.3 Computational results: proton affinity

To provide a more complete picture of εA reactivity, we also calculated the proton affinities (PA; ∆H) and gas phase basicities (GB; ∆G) of the N9 tautomer 1 and the N7 tautomer 7 (Figure 2.3). For the N9 tautomer 1, the most basic site is the N10 (PA =

232.6 kcal mol-1; GB = 224.9 kcal mol-1); the next most basic site is the N7 (PA = 223.7 kcal mol-1; GB = 216.1 kcal mol-1) and the third most basic site is the N3 (PA = 207.1 kcal mol-1; GB = 199.8 kcal mol-1). The least basic sites are N1 and N9. For the N7 tautomer 7, the most basic N9 site is calculated to have a proton affinity of 223.0 kcal mol-1 (GB=215.4 kcal mol-1), which is much less basic than the most basic N10 site of

29 the N9 tautomer 1. The N10 site of the N7 tautomer is comparable in basicity to the N9 site. The remaining PAs/GBs on the N7 tautomer are shown in Figure 2.3.

232.6 (224.9) 222.4 (215.0) H H 10 11 177.8 (170.5) 10 223.7 (216.1) N N 11 6 12 H 12 N 5 H 5 6 H 7 N1 7 N H 162.1 (154.9) N1 8 H 160.5 (153.4) N 8 4 N3 2 H N 4 N 2 H H 9 9 3 207.1 (199.8) 178.5 (171.2) 223.0 (215.4) 218.1 (210.5)

1 (N9) 7 (N7)

For reasons delineated later, we also calculated the proton affinity of an alkylated derivative of εA, 9-methyl-1,N6-ethenoadenine (9, 9-me-εA, Figure 2.4). This methylated derivative is slightly more basic than the parent 1,N6-ethenoadenine. The proton affinity of the most basic site of 9-methyl-1,N6-ethenoadenine is calculated to be

235.0 kcal mol-1 (GB = 227.6 kcal mol-1), which is just under 3 kcal mol-1 more basic than that of εA. The other basic sites are about 2~5 kcal mol-1 more basic than the corresponding sites on 1,N6-ethenoadenine.

235.0 (227.6) H 10 N 11 228.0 (220.2) 12 5 6 H 7 N N1 H 164.8 (157.8) 8 N 4 N 2 H 182.9 (174.7) 9 3 H3C 209.4 (201.6)

Figure 2.4. Calculated proton affinities (∆H; gas phase basicity values (∆G) in parentheses; all values in kcal mol-1) of 9-methyl-1,N6-ethenoadenine at B3LYP/6-31+G* (298 K)

2.3.4 Experimental results: acidity

Calculations indicate that the most acidic site of the N9 tautomer is the N9-H and the most acidic site of the N7 tautomer is the N7-H (Figure 2.2). The two acidities are not expected to be differentiable experimentally since they differ by less than 1 kcal mol-1.

Table 2.1 summarizes the acidity bracketing results for the most acidic site of 1,N6- ethenoadenine. We find that deprotonated 1,N6-ethenoadenine deprotonates per-fluoro-

-1 tert-butyl alcohol ((CF3)3COH, ∆Hacid = 331.6 ±2.2 kcal mol ; ∆Gacid = 324.0 ± 2.0 kcal

-1 -1 mol ), but not hydrogen chloride (HCl, ∆Hacid = 333.4 ± 0.1 kcal mol ; ∆Gacid = 328.1 ±

0.2 kcal mol-1). In the opposite direction, chloride (Cl-) deprotonates 1,N6-ethenoadenine,

- 6 but ((CF3)3CO ) does not. Therefore we bracket the most acidic site of 1,N -

-1 -1 ethenoadenine to be ∆Hacid = 332 ± 2 kcal mol (∆Gacid = 325 ± 3 kcal mol ). This value is consistent with both the calculated acidity of the N9-H of the N9 tautomer 1 (∆Hacid =

-1 330.7; ∆Gacid = 323.2 kcal mol ) as well as the N7-H site of the N7 tautomer 7 (∆Hacid =

-1 331.3; ∆Gacid =323.9 kcal mol ).

Table 2.1. Summary of results for acidity bracketing of more acidic site of 1,N6- ethenoadenine. a a Reference compound ΔHacid ΔGacid Proton

transferb

(kcal mol-1) (kcal mol-1) Ref. Con

acid j. base

1,1,1-trifluoro-2,4- 328.3 ± 2.9 322.0 ± 2.0 + -

pentanedione

3,5 bis(trifluoromethyl)phenol 329.8 ± 2.1 322.9 ± 2.0 + -

difluoroacetic acid 331.0 ± 2.2 323.8 ± 2.0 + -

per-fluoro-tert-butyl alcohol 331.6 ± 2.2 324.0 ± 2.0 + -

hydrogen chloride 333.4 ± 0.1 328.1 ± 0.2 - +

pyruvic acid 333.5 ± 2.9 326.5 ± 2.8 - +

malononitrile 335.8 ± 2.1 328.1 ± 2.0 - +

2-bromopropionic acid 336.8 ± 2.1 329.8 ± 2.0 - +

trifluoro-m-cresol 339.3 ± 2.1 332.4 ± 2.0 - +

acetic acid 348.1 ± 2.2 341.1 ± 2.0 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton

transfer.

We also bracketed the gas phase acidity of the less acidic site of 1,N6-ethenoadenine,

using a method developed in our lab (Table 2.2). We find that the conjugate base of 1,N6- ethenoadenine deprotonates aniline (C6H5NH2, ∆Hacid = 366.4 ± 2.1; ∆Gacid = 359.1 ± 2.0

-1 kcal mol ), but does not deprotonate p-toluidine (CH3C6H4NH2, ∆Hacid = 367.3 ± 2.1;

-1 ∆Gacid = 360.1 ± 2.0 kcal mol ). We therefore bracket the gas phase acidity of the less

6 -1 acidic site of 1,N -ethenoadenine to be: ∆Hacid = 367 ± 3 kcal mol ; ∆Gacid = 360 ± 3 kcal

mol-1.

Therefore, in terms of acidity, we bracket two sites for εA (∆Hacid/∆Gacid): 332/325

and 367/360 kcal mol-1. Because both the N9 and N7 tautomers have acidic sites in the

vicinity of both these experimental values (Figure 2.2), these experiments alone do not reveal which tautomer(s) is (are) present. We therefore embarked on proton affinity experiments to help answer this question.

Table 2.2. Summary of results for acidity bracketing of less acidic site of 1,N6- ethenoadenine.

a a Reference compound ΔHacid ΔGacid Proton

transferb

(kcal mol-1) (kcal mol-1) Ref. acid

formic acid 345.3 ± 2.2 338.3 ± 2.0 +

acetic acid 348.1 ± 2.1 341.1 ± 2.0 +

pyrrole 359.6 ± 2.9 351.8 ± 2.0 +

2-fluroaniline 362.6 ± 2.2 355.3 ± 2.0 +

N-ethylaniline 364.1 ± 2.1 356.8 ± 2.0 +

aniline 366.4 ± 2.1 359.1 ± 2.0 +

p-toluidine 367.3 ± 2.1 360.1 ± 2.0 -

acetone 369.1 ± 2.1 361.9 ± 2.0 -

3-ethyl-3-pentanol 370.9 ± 2.8 364.3 ± 2.7 -

benzylbromide 372.1 ± 2.1 364.9 ± 2.0 -

3-3-dimethyl-1-butanol 372.5 ± 2.8 365.9 ± 2.7 -

acetonitrile 372.9 ± 2.1 365.2 ± 2.0 -

2-ethyl-1-butanol 373.1 ± 2.0 366.5 ± 2.1 -

4-chlorotoluene 374.0 ± 2.1 366.8 ± 2.0 -

2-butanol 374.1 ± 2.0 367.5 ± 2.1 -

1-propanol 376.0 ± 2.1 369.4 ± 2.0 -

ethanol 378.3 ± 1.0 371.7 ± 1.1 -

methanol 381.7 ± 1.0 375.1 ± 1.1 -

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

2.3.5 Experimental results: proton affinity

As the calculations indicate, the proton affinities (PA = ∆H) of the most basic site of the N9 tautomer 1 and the N7 tautomer 7 are 232.6 and 223.0 kcal mol-1, respectively.

Based on our computational and acidity bracketing results, particularly the calculation that tautomers 1 and 7 are close in terms of stability (ΔΔH = 0.7 kcal mol-1), we envision three possible scenarios for our PA bracketing experiments: i) only the N9 tautomer 1 is present; ii) only the N7 tautomer 7 is present; iii) both tautomers 1 and 7 are present. If only the N9 canonical tautomer 1 is present, protonation by hydronium will yield the

N10-protonated species (calculated PA = 232.6 kcal mol-1). Depending on the accuracy of the calculations, we would expect to bracket a proton affinity around 233 kcal mol-1.

The second scenario is that only the N7 tautomer 7 is present. Again, if the calculations are accurate, the bracketing experiments should target a proton affinity of about 222-223 kcal mol-1 (corresponding to the proton at N9 and/or N10 on 7, Figure 2.3). If both the

N9 and N7 tautomers 1 and 7 are present under our gas phase conditions, then the bracketing experiments should yield more complex results, since the N9 and N7 tautomers have such different PAs. Usually when bracketing experiments are conducted,

34 there is a clear "crossover point" as is seen in our acidity experiments, between HCl and per-fluoro-tert-butyl alcohol (Table 2.1). However, if both the N9 tautomer, with a PA of about 233, and the N7 tautomer, with a PA of about 222, are present, then reactions with reference bases in the 222-233 range will yield intriguing results, without a clean crossover point. For example, let us assume that in fact both the N9 and N7 tautomers are present, and that the former has a PA of 233 and the latter a PA of 222. What would happen if we utilized a reference base in between the range of 222 and 233, such as 4- picoline (PA = 226.4 ± 2.0 kcal mol-1)? The expected reactions of protonated εA with 4- picoline and protonated 4-picoline with εA are shown in Scheme 2.2.

In Reaction A, 4-picoline is basic enough to deprotonate the protonated N7 tautomer but not basic enough to deprotonate the protonated N9 tautomer.118 One would call this reaction a "+" since one would see proton transfer, even though only one tautomer, the

N7 tautomer, is reacting. In the opposite direction (Reaction B), the N7 tautomer cannot deprotonate protonated 4-picoline but the N9 tautomer can. Again, one would call this reaction a "+" even though now only the N9 tautomer is reacting. The mass spectrometer cannot differentiate protonated N9 tautomer from protonated N7 tautomer; the m/z ratio is the same. Therefore, the reaction with 4-picoline would be marked as occurring in both directions. The interesting feature is that any reference base between the PAs of the two tautomers will appear as a "+,+" since in one direction, the N7 tautomer will react, and in the other, the N9 tautomer will. Thus, rather than a clean "crossover" point, one would see a range (if calculations are accurate, between about 222 and 233 kcal mol-1) where proton transfer occurs in both directions! Therefore, if a mixture of the N7 and N9 tautomers is present, the bracketing experiments will reveal this via a large "crossover

35 range". If only the N7 or the N9 tautomer is present, then the crossover point will be clean, and we can bracket a corresponding PA for whichever tautomer prevails.

Scheme 2.2

The experimental PA bracketing results are summarized in Table 2.3. The protonated 1,N6-ethenoadenine protonates the first four reference bases: 1-pyrrolidino-1- cyclopentene, tributylamine, N-N-dimethylcyclohexylamine and triethylamine. The conjugate acids of these four reference bases do not protonate neutral 1,N6-ethenoadenine.

These results indicate that 1,N6-ethenoadenine is less basic than these particular reference bases. We find that proton transfer occurs in both directions for 1-methyl piperidine and

1-methyl pyrrolidine. For piperidine and weaker bases, protonated 1,N6-ethenoadenine does not react with the neutral bases, but 1,N6-ethenoadenine can deprotonate protonated piperidine and weaker bases. Therefore we bracket the gas phase proton affinity of the most basic site of 1,N6-ethenoadenine to be PA = 232 ± 4 kcal mol-1 (GB= 224 ± 4 kcal mol-1). Comparison of this experimental result with our calculations (Figure 2.3) implies that we have bracketed the N10 site on the N9 tautomer 1.

Table 2.3. Summary of results for proton affinity bracketing of more basic site of 1,N6- ethenoadenine Reference compound Proton affinitya Gas phase Proton transferb

basicitya

(∆H, kcal mol-1) (∆G, kcal mol-1) Ref. Conj.

base acid

1-pyrrolidino-1- 243.6 ± 2.0 236.2 ± 2.0 + -

cyclopentene

tributylamine 238.6 ± 2.0 231.3 ± 2.0 + -

N, N- 235.1 ± 2.0 227.7 ± 2.0 + -

dimethylcyclohexylamine

triethylamine 234.7 ± 2.0 227.0 ± 2.0 + -

1-methyl piperidine 232.1 ± 2.0 224.7 ± 2.0 + +

1-methyl pyrrolidine 230.8 ± 2.0 223.4 ± 2.0 + +

piperidine 228.0 ± 2.0 220.0 ± 2.0 - +

pyrrolidine 226.6 ± 2.0 218.8 ± 2.0 - +

4-picoline 226.4 ± 2.0 218.8 ± 2.0 - +

3-picoline 225.5 ± 2.0 217.9 ± 2.0 - +

cyclohexanamine 223.3 ± 2.0 215.0 ± 2.0 - +

pyridine 222.0 ± 2.0 214.7 ± 2.0 - +

N-methyl propionamide 220.0 ± 2.0 212.6 ± 2.0 - +

aniline 210.9 ± 2.0 203.3 ± 2.0 - +

aProton affinities and gas phase basicities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

In order to mitigate the ambiguity associated with the possibility of more than one tautomer of 1,N6-ethenoadenine present, we also examined the 9-methyl-1,N6- ethenoadenine derivative 9.104,105 This compound is a “frozen” form of the N9 tautomer, with an alkyl group at N9. Proton affinity bracketing studies indicate that proton transfer occurs when protonated 9-methyl-1,N6-ethenoadenine 9 and 2,2,6,6-tetramethylpiperidine

(PA= 235.9 kcal mol-1) are allowed to react; however, proton transfer does not occur in the reverse direction (Table 2.4). Furthermore, no proton transfer reaction is observed between protonated 9-methyl-1,N6-ethenoadenine and 1-methyl piperidine (PA = 232.1 kcal mol-1), but proton transfer is observed when 9-methyl-1,N6-ethenoadenine is allowed to react with protonated 1-methyl piperidine. When the reference bases N, N- dimethylcyclohexylamine (PA = 235.1 kcal mol-1) and triethylamine (PA = 234.7 kcal mol-1) are used, proton transfer is observed in both directions. Therefore we bracket the

PA of the most basic site in 9-methyl-1,N6-ethenoadenine to be 235 ± 4 kcal mol-1 (GB =

227 ± 4 kcal mol-1). We believe we have bracketed the proton affinity of N10 site of the methylated derivative 9, because our experimental result is consistent with the calculated

-1 proton affinity of the N10 site (PAcalc = 235.0; GBcalc = 227.6 kcal mol ). Comparison of this experimental result (235 ± 4 kcal mol-1) with our experimental bracketing result for the parent 1,N6-ethenoadenine (232 ± 4 kcal mol-1) leads us to conclude that the most basic site we bracketed on εA is the N10 site of N9 tautomer 1.

Table 2.4. Summary of results for proton affinity bracketing of more basic site of 9- methyl-1,N6-ethenoadenine Reference compound Proton affinitya Gas phase Proton transferb

basicitya

(∆H, kcal mol-1) (∆G, kcal mol-1) Ref. Conj.

base acid

1-pyrrolidino-1- 243.6 ± 2.0 236.2 ± 2.0 + -

cyclopentene

tributylamine 238.6 ± 2.0 231.3 ± 2.0 + -

N,N- 237.6 ± 2.0 230.3 ± 2.0 + -

diisopropylethylamine

2,2,6,6- 235.9 ± 2.0 228.0 ± 2.0 + -

tetramethylpiperidine

N,N- 235.1 ± 2.0 227.7 ± 2.0 + +

dimethylcyclohexylamine

triethylamine 234.7 ± 2.0 227.0 ± 2.0 + +

1-methyl piperidine 232.1 ± 2.0 224.7 ± 2.0 - +

1-methyl pyrrolidine 230.8 ± 2.0 223.4 ± 2.0 - +

piperidine 228.0 ± 2.0 220.0 ± 2.0 - +

aProton affinities and gas phase basicities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

We also measured the proton affinities of 1,N6-ethenoadenine and 9-methyl-1,N6- ethenoadenine using the Cooks kinetic method, as a check of our bracketing results. For

1,N6-ethenoadenine, we used 2,4-lutidine (PA = 230.1 kcal mol-1), 1-methyl pyrrolidine

(PA = 230.8 kcal mol-1), 1-methyl piperidine (PA = 232.1 kcal mol-1), triethylamine (PA

= 234.7 kcal mol-1), N,N-dimethylcyclohexanamine (PA = 235.1 kcal mol-1) and 2,2,6,6- tetramethylpiperidine (PA = 235.9 kcal mol-1) as the reference bases, yielding a proton affinity of 233 ± 2 kcal mol-1, which is consistent with the bracketing results (232 ± 4 kcal mol-1). For 9-methyl-1,N6-ethenoadenine, seven reference bases were used: 1-methyl piperidine, triethylamine, N,N-dimethylcyclohexanamine, 2,2,6,6-tetramethylpiperidine, tripropylamine (PA = 236.9 kcal mol-1), N,N-diisopropylethylamine (PA = 237.6 kcal mol-1) and tributylamine (PA = 238.6 kcal mol-1). The proton affinity is measured to be

236 ± 2 kcal mol-1, which is consistent with the bracketing result (235 ± 4 kcal mol-1). It therefore appears that under our conditions, only the N9 tautomer 1 of 1,N6- ethenoadenine is present. Our acidity and proton affinity experiments are consistent with this conclusion.

One interesting difference in the bracketing versus the Cooks method experiment is that in the former, the neutral εA is added via the heatable solids probe, whereas in the latter, the proton bound dimers are prepared in solution and then electrosprayed. It therefore appears from our acidity and proton affinity experiments that gas phase εA, whether vaporized from the solid form or from solution, under our conditions, exists as the canonical tautomer 1.

2.4 Biological Implications

Adenine alkyl glycosylase (AAG) is a broadly specific enzyme that cleaves several mutated purine bases, including 1,N6-ethenoadenine, hypoxanthine, 3-methyladenine and

7-methylguanine (1-4) from double-stranded DNA.91-93 The exact mechanism by which

AAG recognizes and excises the damaged bases, but leaves adenine (5) and guanine (6) untouched, is unknown. Our previous studies with hypoxanthine and 3-methyladenine have led us to propose a mechanism involving anionic cleavage (Scheme 2.1).9,10,14,17,91-

93,99,102,103 This mechanism is also proposed for a related enzyme, thymine DNA glycosylase (TDG), which cleaves mutated pyrimidine bases from DNA.102 In this mechanism, a nucleophile such as an activated water attacks the C1' and the damaged nucleobase leaves in its deprotonated form. We postulate that AAG discriminates the damaged bases from the normal bases because the deprotonated damaged bases are better leaving groups than the deprotonated normal bases. Furthermore, our hypothesis is that the differences in leaving group ability of the damaged versus the normal bases are enhanced in the gas phase. We recently found this to be true with hypoxanthine; hypoxanthine is more acidic than adenine and guanine (thus the conjugate base of hypoxanthine would be the best leaving group) and the differences in acidity are even greater in the gas phase than in aqueous solution. We therefore propose that AAG provides a hydrophobic environment that allows it to easily differentiate damaged bases from normal bases.

We sought to ascertain whether εA fits into this picture. Is εA more acidic than adenine and guanine? How do those acidities differ in the gas phase versus in solution?

6 The ∆Hacid of the N9-H of 1,N -ethenoadenine (N9 tautomer 1) is calculated to be 330.7

-1 -1 kcal mol , which is 4.1 kcal mol more acidic than adenine (calculated ∆Hacid = 334.8

-1 -1 kcal mol ) and 3.6 kcal mol more acidic than guanine (calculated ∆Hacid = 334.3 kcal mol-1) in the gas phase.103 The error bars on known experimental values of adenine and

-1 guanine acidity render comparison to εA difficult (∆Hacid (εA) = 332 ± 2 kcal mol ;

-1 - ∆Hacid (adenine) = 332-333 ± 2-3 kcal mol ; ∆Hacid (guanine) = 331-332 ± 3-4 kcal mol

1).9,119 Therefore, we conducted Cooks kinetic experiments on the εA•A and εA•G deprotonated dimers and find that εA is 3.42 ± 0.11 kcal mol-1 more acidic than adenine and 2.54 ± 0.01 kcal mol-1 more acidic than guanine. These results indicate that if the damaged base leaves in a deprotonated anionic form when the enzyme catalyzed excision occurs, deprotonated 1,N6-ethenoadenine is the best leaving group among adenine, guanine and 1,N6-ethenoadenine. The next question is, what are the relative acidities of

εA versus adenine and guanine in the gas phase versus in solution? Enhanced acidity differences in the gas phase would be consistent with our theory that AAG provides a hydrophobic active site to discriminate damaged from normal bases.

We measured the aqueous pKa of εA to be 9.9, which is comparable in acidity to both adenine (9.8) and guanine (10.0).120-122 We also conducted solvent dielectric calculations to ascertain how the relative acidities change in a more polar environment.73,74 The calculated acidity of the N9-H of 1,N6-ethenoadenine (N9 tautomer

1) in water is 294.1 kcal mol-1, which is comparable to that of adenine (294.8 kcal mol-) and guanine (294.6 kcal mol-). The differences in acidity in solution among εA, adenine and guanine are thus significantly lower than the differences found in the gas phase. This result, coupled with our similar result for hypoxanthine, implies that the nonpolar active

42 site in AAG could contribute to specificity by enhancing the differences in acidity among adenine, guanine and damaged bases.101,103

2.5 Conclusions

Our experimental and computational study of 1,N6-ethenoadenine indicates that although both the N9 tautomer 1 and the N7 tautomer 7 may be close in gas phase stability, we measure the acidities and proton affinity of only the N9 tautomer 1.

Comparison of the acidic properties of 1,N6-ethenoadenine to those of the normal nucleobases adenine and guanine, both in the gas phase and in solution, supports our theory that AAG cleaves damaged nucleobases as anions and that the active site may take advantage of a nonpolar environment to favor deprotonated damaged bases such as 1,N6- ethenoadenine as a leaving group versus deprotonated adenine or guanine. Our earlier studies on hypoxanthine are consistent with this theory as well.

Note: Major parts of the following chapter have been published: Zhachkina, A.; Liu, M.;

Sun, X.; Amegayibor, F. S.; Lee, J. K. J. Org. Chem. 2009, 74, 7429-7440.

Chapter 3 The Gas-Phase Thermochemical Properties of the Damaged

Base O6-Methylguanine versus Adenine and Guanine

3.1 Introduction

In Chapter 2, we reported the gas phase thermochemical properties of the damaged nucleobase 1,N6-ethenoadenine. In this chapter we are going to focus on another damaged base, O6-methylguanine (3, OMG), and also compare it with normal purine nucleobases, adenine (1) and guanine (2). As previously addressed in this thesis, damaged DNA bases differ in structure and properties from normal nucleobases and therefore intervene with gene replication and expression, leading to cell death, aging and carcinogenesis.14,15,17,93

Our studies are motivated by understanding the mechanisms by which mutated bases are cleaved.9,11,12,102,103,123 Herein, we provide a comprehensive examination of the gas phase thermochemical properties of the damaged guanine base O6-methylguanine. We find

OMG to be less basic at O6 than guanine and less acidic at N9 than adenine and guanine; these results have interesting biological implications which are discussed.

3.2 Experimental

All chemicals are commercially available and were used as received.

Acidity and proton affinity bracketing experiments were conducted using a Fourier

Transform Ion Cyclotron Resonance Mass Spectrometer (FTMS) with a dual cell set up which has been described previously.11,12,123 More details may be found in Chapter 2. We also used the Cooks kinetic method in a quadrupole ion trap (LCQ) mass spectrometer67,69,107-109 to measure the proton affinities and acidities. The procedure for conducting these experiments in our lab has been described previously103 and in Chapter

The B3LYP method and the 6-31+G* basis set as implemented in Gaussian03 were used for all the gas phase calculations.72,112-116 This level has been previously shown to be reasonably accurate for gas phase acidity and proton affinity calculations of nucleobases.9-13,103,123 All the geometries are fully optimized and frequencies are calculated; no scaling factor is applied. Reported values herein are at 298 K.117

3.3 Results

3.3.1 Computational results: tautomers

At B3LYP/6-31+G*, the canonical "N9H" structure of OMG is the most stable (3,

Figure 3.1). The rotamer wherein the methyl group is "pointing" toward the N7 (rather

45 than the N1, 3a) is 2.6 kcal mol-1 higher in energy; the methyl should freely rotate (we calculate the barrier to rotation to be 6.9 kcal mol-1) so at room temperature, the majority of molecules will be in the more stable minimum-energy structure.124 The next most stable tautomer is the N7H structure (3b), which is 2.7 kcal mol-1 less stable than the canonical. All remaining tautomers are more than 10 kcal mol-1 less stable than the canonical N9H structure.

Figure 3.1. Relative enthalpies (∆H in kcal mol-1) of the three most stable OMG

tautomers and acidities (red values; ∆Hacid, with ∆Gacid in parentheses; all values in kcal mol-1), and proton affinities (blue values; PA, with GB in parentheses; all values in kcal mol-1) , calculated at B3LYP/6-31+G* (298 K)

3.3.2 Computational results: acidities

The most acidic site of the canonical tautomer 3 is the N9-H (∆Hacid = 337.4 kcal

-1 -1 mol ; ∆Gacid = 329.9 kcal mol , Figure 3.1, in red). The rotamer 3a is slightly more

-1 -1 acidic at N9 (∆Hacid = 335.3 kcal mol ; ∆Gacid = 327.7 kcal mol ), and comparable in

-1 -1 acidity to the N7 tautomer (N7-H, ∆Hacid = 335.2 kcal mol ; ∆Gacid = 327.6 kcal mol ).

3.3.3 Computational results: proton affinities

The most basic site of the most stable form of OMG (canonical form 3) has a calculated PA of 228.4 kcal mol-1 (GB = 220.9 kcal mol-1), at N7 (Figure 3.1, in blue).

The rotamer 3a has a comparable basicity (at N1, PA = 228.7 kcal mol-1, GB = 221.7 kcal mol-1). The most basic site on the N7H tautomer is the N3 (PA = 233.0 kcal mol-1; GB =

225.7 kcal mol-1).

3.3.4 Experimental results: acidities

The acidity bracketing results for the more acidic site of OMG are shown in Table

3.1. We find that deprotonated OMG cannot deprotonate trifluoro-m-cresol (∆Hacid =

-1 -1 339.3 kcal mol ; ∆Gacid = 332.4 kcal mol ) but that the opposite reaction does occur.

Deprotonated OMG does react with 2-chloropropanoic acid (∆Hacid = 337.0; ∆Gacid =

330.4 kcal mol-1), but 2-chloropropanoate does not deprotonate OMG. We therefore

-1 - bracket the acidity of OMG to be ∆Hacid = 338 ± 3 kcal mol (∆Gacid = 331 ± 3 kcal mol

1).

Table 3.1. Summary of results for acidity bracketing of more acidic site of OMG. a a b Reference compound ΔHacid ΔGacid Proton transfer

Ref. Conj.

acid base

2,4-pentadione 343.8 ± 2.1 336.7 ± 2.0 – +

2-chlorophenol 343.4 ± 2.3 337.1 ± 2.0 – +

ethoxyacetic acid 342.0 ± 2.2 335.0 ± 2.0 – +

trifluoro-m-cresol 339.3 ± 2.1 332.4 ± 2.0 – +

2-chloropropanoic acid 337.0 ± 2.1 330.4 ± 2.0 + –

2-bromopropanoic acid 336.8 ± 2.1 329.8 ± 2.0 + –

pyruvic acid 333.5 ± 2.9 326.5 ± 2.8 + –

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

We also used the methodology developed in our lab to bracket the less acidic site of

OMG (Table 3.2). Deprotonated OMG under "less acidic" conditions does not

-1 -1 deprotonate 2-fluoroaniline (∆Hacid = 362.6 kcal mol ; ∆Gacid = 355.3 kcal mol ) but

-1 does deprotonate 2,2,2-trifluoroethanol (∆Hacid = 361.7 kcal mol ; ∆Gacid = 354.1 kcal

-1 -1 mol ). We therefore bracket the less acidic site of OMG to be ∆Hacid = 362 ± 3 kcal mol

-1 (∆Gacid = 355 ± 3 kcal mol ).

Table 3.2. Summary of results for acidity bracketing of less acidic site of OMG.

a a b Reference compound ΔHacid ΔGacid Proton transfer

Ref. acid

acetone 368.8 ± 2.0 361.9 ± 2.0 –

aniline 366.4 ± 2.1 359.1 ± 2.0 –

4-fluoroaniline 364.3 ± 2.1 357.0 ± 2.0 –

2-fluoroaniline 362.6 ± 2.2 355.3 ± 2.0 –

2,2,2-trifluoroethanol 361.7 ± 2.5 354.1 ± 2.0 +

pyrrole 359.5 ± 0.3 350.9 ± 2.0 +

4-(trifluoromethyl)-aniline 353.3 ± 2.1 346.0 ± 2.0 +

butanoic acid 346.8 ± 2.0 339.5 ± 2.0 +

aAcidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

We also measured the acidity of OMG using the Cooks kinetic method. The

-1 reference acids 4-hydroxybenzoic acid (∆Hacid = 335.9 ± 2.1 kcal mol ; ∆Gacid = 328.9 ±

-1 -1 2.0 kcal mol ), 3-(chloromethyl)benzoic acid (∆Hacid = 336.8 ± 2.1 kcal mol ; ∆Gacid =

-1 -1 329.8 ± 2.0 kcal mol ), 2-fluorobenzoic acid (∆Hacid = 338.1 ± 2.2 kcal mol ; ∆Gacid =

-1 -1 330.6 ± 2.0 kcal mol ), and 3-hydroxybenzoic acid (∆Hacid = 338.6 ± 2.1 kcal mol ;

-1 -1 ∆Gacid = 331.6 ± 2.0 kcal mol ) were used, yielding a ∆Hacid = 337 ± 3 kcal mol (∆Gacid

= 330 ± 3 kcal mol-1).

3.3.5 Experimental results: proton affinities

The proton affinity results for OMG are summarized in Table 3.3. We find that 1- methylpyrrolidine (PA = 230.8 kcal mol-1; GB = 223.4 kcal mol-1) does deprotonate protonated OMG, but the opposite reaction does not occur. Piperidine, however (PA =

228.0 kcal mol-1; GB = 220.0 kcal mol-1), is not basic enough to deprotonate protonated

OMG; OMG does deprotonate protonated piperidine. The PA of OMG is therefore bracketed to be 229 ± 4 kcal mol-1 (GB = 222 ± 4 kcal mol-1).

Table 3.3. Summary of results for PA bracketing of more basic site of OMG.

Reference compound PAa GBa Proton transferb

Ref. Conj.

base acid

triethylamine 234.7 ± 2.0 227.0 ± 2.0 + –

1-methylpiperidine 232.1 ± 2.0 224.7 ± 2.0 + –

1-methylpyrrolidine 230.8 ± 2.0 223.4 ± 2.0 + –

piperidine 228.0 ± 2.0 220.0 ± 2.0 – +

3-picoline 225.5 ± 2.0 217.9 ± 2.0 – +

pyridine 222.3 ± 2.0 214.7 ± 2.0 – +

propylamine 219.4 ± 2.0 211.3 ± 2.0 – +

aValues are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

We also measured the proton affinity and gas phase basicity of OMG using the

Cooks kinetic method. The reference bases piperidine (PA = 228.0 ± 2.0 kcal mol-1; GB

= 220.0 ± 2.0 kcal mol-1), 2,4-lutidine (PA = 230.1 ± 2.0 kcal mol-1; GB = 222.5 ± 2.0 kcal mol-1), 1-methylpyrrolidine (PA = 230.8 ± 2.0 kcal mol-1; GB = 223.4 ± 2.0 kcal mol-1), N, N-dimethylbenzylamine (PA = 231.5 ± 2.0 kcal mol-1; GB = 224.0 ± 2.0 kcal mol-1), N, N-dimethylisopropylamine (PA = 232.0 ± 2.0 kcal mol-1; GB = 224.6 ± 2.0 kcal mol-1), and 1-methylpiperidine (PA = 232.1 ± 2.0 kcal mol-1; GB = 224.7 ± 2.0 kcal mol-1) were used. By this method, the PA for OMG is measured to be 231 ± 3 kcal mol-1

(GB = 223 ± 3 kcal mol-1).

3.4 Discussion

3.4.1 O6-methylguanine (OMG): tautomers, acidities and proton affinities

To our knowledge, there have been no prior gas phase experimental studies of OMG.

In the gas phase, the N9H and N7H tautomers of O6-methylguanine are calculated to be somewhat close in energy (less than 3 kcal mol-1 apart), and thus both might coexist

(Figure 3.1). In Figure 3.2, we also show the rotamer of the N9 tautomer 3 (3a), which is 2.6 kcal mol-1 higher in energy than the canonical form. The methyl group in structure

3 is referred to as "distal" while that in 3a is "proximal" (relative to the N9). We would expect the methyl group to rotate fairly freely in this tautomer -- our calculated transition structure for rotation is only 6.9 kcal mol-1 higher than the most stable ground state

(Figure 3.2). The distal structure of the N9H-tautomer (3) is also found to be preferred in the solid state and aqueous phase (over the proximal structure (3a)), in agreement with our gas phase calculations.125-131

Figure 3.2. Rotation of the methyl group in the canonical O6-methylguanine structure, calculated at 298 K at B3LYP/6-31+G*

-1 In terms of acidity, the N9H tautomer 3 has a calculated ∆Hacid of 337.4 kcal mol

(the N9 proton). Its rotamer 3a and the N7H tautomer 3b both have a calculated ∆Hacid of

-1 335 kcal mol (at, respectively, the N9 and N7 sites). The measured OMG ∆Hacid is 338 ±

3 kcal mol-1 by bracketing and 337 ± 3 kcal mol-1 by the Cooks kinetic method. While this value is certainly closest to the N9H tautomer 3, it is still consistent with structures

3a and 3b, so we cannot discount their presence under our conditions.

-1 We also measured a less acidic site on OMG; ∆Hacid= 362 ± 3 kcal mol . This could correspond to the exocyclic -NH2 protons on any of the three tautomers 3-3b. Thus, based on our calculations and experiments, we may have the OMG structures 3, 3a and/or 3b present under our gas phase conditions, though 3 should predominate. Also, the

51 measurements from the bracketing and Cooks kinetic experiments are very close; in the

FTMS bracketing experiment, OMG is heated in a solids probe in vacuo while for the

Cooks method the OMG is electrosprayed as a proton-bound dimer with a reference acid.

We could therefore conceivably have different tautomeric mixtures under the different experimental conditions in which the substrate is generated.132,133

In terms of PA, we bracket the most basic site to be 229 ± 4 kcal mol-1 (GB = 222 ±

4 kcal mol-1), which could correspond to 3 (the N7), 3a (the N1) or 3b (the N3). Again, based on our calculations, 3 should predominate in the gas phase. By the Cooks kinetic method, we measure the most basic site to be 231 ± 3 kcal mol-1 (GB = 223 ± 3 kcal mol-

1); again, although this value could also correspond to any of the low-lying tautomers, we use electrospray to generate the OMG•reference-base protonated dimers from aqueous solution, so the canonical structure 3 favored in solution may predominate. These are the first studies probing tautomer composition of OMG in the gas phase.

3.4.2 Comparison of adenine and guanine properties with O6-methylguanine (OMG).

O6-Methylguanine arises from the alkylation of guanine, and is highly mutagenic.134-143 Its mutagenicity arises from a "GC to AT" transition that occurs due to different hydrogen bonding preferences for normal guanine versus OMG.144-146 In DNA replication, guanine base pairs preferentially with cytosine. Once guanine is alkylated to form OMG, however, the preferred hydrogen bonding pattern is to thymine rather than cytosine.147,148 The result is that a DNA sequence that originally contained a G•C base pair will become an OMG•T base pair; when the strand containing the "T" replicates, the end result will be an A•T base pair; this is the "GC to AT" transition. Since the exact

52 sequence of DNA is necessary for proper life function (coding for proteins, signaling), the mutation to OMG is known to be highly carcinogenic.149-153

Two aspects in particular regarding OMG's biological activity intrigue us. First is the OMG preference for T over C, the provenance of which is unknown.136,147,148 Jones et al. conducted aqueous studies of the stabilities of 9-mer duplexes containing OMG•C versus OMG•T base pairs and found that the OMG•C duplexes are more stable than those containing OMG•T.154-156 We found this to be true in the gas phase as well in previously published work.157,158 Thus, intrinsically, OMG binds more strongly to C, but in Nature, when replication occurs, OMG binds to T.

Because proton affinities and acidities are related to the hydrogen bonding acceptor and donor ability, we can use these values to try to understand OMG binding to C and to

T. In Figure 3.3, we show the G•C, OMG•C and OMG•T base pairs with the relevant proton affinity and acidity values.159 Our goal here is to not focus on geometries, but just to look at the proton affinities of acceptor atoms and the acidities of donor atoms of each individual substrate, in the context of hydrogen bonding.

Figure 3.3. Calculated proton affinity (PA, blue) and acidity (∆Hacid, red) values (B3LYP/6-31+G*, 298 K, in kcal mol-1) for the atoms of the nucleobases involved in the G•C, OMG•C, and OMG•T base pairs

The G•C base pair has three hydrogen bonds, from the O6 in guanine to the N4H in cytosine; from the N1-H in guanine to the N3 in cytosine; and from the N2H in guanine to the O2 in cytosine. The heteroatoms all have PA values higher than 220, and the acidic

-1 protons all have ∆Hacid values below 353 kcal mol .

The configuration we chose to draw of the OMG•C pair is based on crystallographic, NMR and computational data (Figure 3.3).125,127,129,160-163 The two main hydrogen bonds are the N1 on OMG to the exocyclic NH2 on cytosine, and the exocyclic

NH2 on OMG to the N3 of cytosine. In both experiment and computation, it appears that sometimes the OMG methyl is proximal (pointing toward the "N7" side), and sometimes it is distal (toward the "N1" side). We utilized the PA associated with the latter, since that structure is the more stable by our calculations.

Alkylation of the O6 drops its basicity considerably, from 220.4 kcal mol-1 to 187.3 kcal mol-1. Therefore, any hydrogen bond to this site would be fairly weak. Alkylation

-1 -1 also decreases the acidity at the exocyclic -NH2, from 336 kcal mol to 357.5 kcal mol , thus decreasing the ability of that proton to act as a hydrogen bond donor. The N1 changes from an acidic to a basic site, leading to the hydrogen bonding pattern shown in

Figure 3.3. Two main hydrogen bonds versus three means that the OMG•C base pair will be less stable than the parent G•C. The proton donors in the OMG•C base pair have

-1 ∆Hacid values of 353 and 357.5 kcal mol while the proton acceptors have PA values of

219 and 227 kcal mol-1. Compared to the G•C base pair, the PA values are comparable, but the OMG -NH2 acidity is considerably less than that of G (357.5 versus 336.0 kcal mol-1), so not only does the OMG•C base pair have one less hydrogen bond than the G•C base pair, but depending on proximity, one of the hydrogen bonds may also be less stable.

Calculations and experiment (solid state and solution phase) are not wholly in agreement on the structure of the OMG•T base pair.125,127,130,162,164,165 By NMR, the methyl group is distal (pointing toward N1); by X-ray, it is proximal. As we did with the

OMG•C base pair, we utilize the PA associated with the distal configuration. Of the two hydrogen bonds, the one between the N1 of OMG and the N3-H of T seems reasonable

-1 -1 based on the associated PA (219.2 kcal mol ) and ∆Hacid (344.8 kcal mol ) values.

However, the second one, between the rather nonacidic OMG N2H (∆Hacid = 357.5 kcal mol-1) to the only slightly basic O2 of thymine (PA = 196.1 kcal mol-1) seems as if it would be much weaker, especially in comparison to the hydrogen bonds on the OMG•C pair. Our values are consistent with earlier studies in supporting that intrinsically, the

OMG•C base pair is likely to be favored energetically over the OMG•T base pair. The preference in vivo for the OMG•T base pair must therefore be controlled by other features, such as the enzyme involved in replication.

The second aspect that intrigues us regarding OMG is how it is repaired in DNA.

DNA is inevitably damaged and Nature has devised various ways of repairing damaged nucleobases.14,17 We have, for the past several years, been studying the glycosylase family of enzymes, which, as part of the base excision repair (BER) pathway, excises a wide range of oxidized and alkylated bases. In humans, alkyl adenine glycosylase (AAG) is responsible for excising a wide range of bases, including 3-methyladenine, hypoxanthine, and 1, N6-ethenoadenine (structures shown in Figure 3.4; possible AAG mechanism shown in Scheme 3.1).14,17 One mechanistic question is how AAG achieves this "broad specificity"; many of the damaged bases are very similar in structure to adenine and guanine, yet the latter normal nucleobases do not get cleaved. We have

55 proposed that AAG excises the damaged nucleobases in deprotonated form (as the N9- deprotonated anions), and that discrimination is achieved because the damaged bases are more acidic at the N9 position, and therefore are better anionic leaving groups and more prone to cleavage (Scheme 3.1). We have shown in previous studies that in fact, the damaged bases 3-methyladenine, hypoxanthine, and 1, N6-ethenoadenine (Chapter 2) are indeed more acidic than adenine and guanine, which provides support for our hypothesis.9,10,103,166 Furthermore, we have shown that the difference in acidity between the normal and damaged bases is enhanced in the gas phase.9,10,103 We therefore propose that AAG might provide a hydrophobic active site that enhances the discrimination of normal from damaged bases.

Figure 3.4. Calculated N9-H ∆Hacid values for normal and damaged purines (B3LYP/6- 31+G*, 298 K)

Scheme 3.1

10 N 7 N N1 N 10 O 9 N3 O N O P O O P O 7 O AAG O N DNA O- 1' DNA O- 1' N1 O O N 9 N3 OOP - OOP - O O DNA DNA

We have long been intrigued by the fact that OMG is not excised by AAG.167 Repair of an OMG lesion is carried out by a methylguanine methyl transferase (MGMT), which dealkylates OMG (Scheme 3.2).168-170

Scheme 3.2

The studies herein show that OMG is markedly different from the other damaged bases we have studied (3-methyladenine, hypoxanthine, and 1, N6-ethenoadenine) in that it is less acidic than adenine and guanine at the N9 position. The calculated values for the canonical structures indicate that while adenine and guanine have a ∆Hacid of about 335

-1 -1 kcal mol , that of OMG is 337 kcal mol . In contrast, the ∆Hacid values for 3-

57 methyladenine, hypoxanthine, and 1, N6-ethenoadenine are, respectively, 323, 331, and

331 kcal mol-1 (Figure 3.4).9,10,103 We also confirmed the relative acidity ordering

(hypoxanthine and 1, N6-ethenoadenine are more acidic than adenine and guanine which are more acidic than OMG) by conducting relative Cooks kinetic experiments among these substrates. That is, in addition to our Cooks experiments, where we measure properties by dissociating proton-bound dimers of the nucleobases with reference compounds, we also did "relative" experiments where we generated proton-bound dimers of, for example, deprotonated adenine and deprotonated guanine ([A–•••H+•••G–]) in order to ascertain their relative acidities. These sets of experiments confirm that OMG is the least acidic nucleobase, and that the other damaged bases are more acidic than adenine and guanine.

Based on these acidity values, it makes sense that AAG does not cleave OMG; if the enzyme achieves discrimination by cleaving those bases that are the best leaving groups,

OMG would be even less likely than the normal bases to be excised. Because OMG has such a low acidity and therefore is not prone to anionic cleavage, Nature presumably devised an alternate method of repair, namely, dealkylating the OMG via a methyl transfer reaction (Scheme 3.2).

3.5 Conclusions

DNA is inevitably damaged by environmental mutagens as well as chemotherapeutics; such mutations are linked to carcinogenesis and aging.14,15,17We study the mechanisms by which enzymes might cleave damaged bases from DNA, thereby protecting our genome.9,11,12,102,103,123 Previous results have shown that the properties of normal versus damaged bases lend insight into the mechanisms by which the damaged

58 bases are cleaved.9,11,12,102,103,123 We have calculated and measured the acidity and proton affinity of one of the damaged nucleobases, O6-methylguanine (OMG) to probe its intrinsic reactivity and examine how it is different from normal bases. O6-methylguanine

(OMG) is a highly mutagenic and carcinogenic damaged form of guanine that has not been heretofore examined in the gas phase before.

-1 We measure the ∆Hacid of the more acidic site to be 338 ± 3 kcal mol ( ∆Gacid = 331

-1 -1 ± 3 kcal mol ) and that of the less acidic site to be 362 ± 3 kcal mol ( ∆Gacid = 355 ± 3 kcal mol-1), and the PA to be 229 ± 4 kcal mol-1 (GB = 222 ± 4 kcal mol-1); these values are in agreement with Cooks kinetic method experiments that we also conducted. In terms of OMG tautomerism, we may have a mix of the canonical N9H and the N7H structures present, though the N9H most likely predominates.

OMG is less acidic at the N9 position than both adenine and guanine, and also, less acidic than other damaged bases we have studied (3-methyladenine, hypoxanthine, 1, N6- ethenoadenine).9,10,103 This is interesting because unlike many other damaged bases, the

OMG lesion is repaired in the genome by demethylation rather than enzyme-catalyzed excision at N9. The fact that OMG is not as acidic at the N9 position is consistent with the fact that OMG is repaired by a different method than the other damaged bases; that low acidity indicates that its conjugate base anion would be a worse leaving group, and demethylation might have evolved as a more efficient method for repair.

We also used the proton affinity and acidity of OMG to explore the issue of hydrogen bonding to cytosine (C) and thymine (T). Although the normal base guanine pairs preferentially with C in replication, OMG pairs with T. Experiments indicate that duplexes containing OMG•C base pairs are more stable than those containing OMG•T

59 base pairs, so it is puzzling as to why OMG•T base pairs are formed in Nature.154-158 Our examination of the proton donating and acceptor abilities are consistent with the in vitro experimental results; we would expect the OMG•C base pair to be more stable than the

OMG•T base pair. Therefore, the preference of OMG to pair with T in replication is controlled by features other than pure intrinsic stability.

Liu, M; Li, T.; Amegayibor, F. S.; Cardoso, D. S.; Fu, Y.; Lee, J. K. J. Org. Chem. 2008, 73, 9283-9291.

Chapter 4 The Gas-Phase Thermochemical Properties of Pyrimidine

Nucleobases

4.1 Introduction

In previous work and in Chapters 2 and 3, we have reported the gas phase thermochemical properties of uracil and adenine, as well as several damaged purine nucleobases.9-13,103,123,157 Damaged DNA bases differ in structure and properties from normal nucleobases and therefore intervene with gene replication and expression, leading to cell death, aging and carcinogenesis.14,15,17,93 Our studies are motivated by understanding the mechanisms by which mutated bases are cleaved, focusing on the glycosylase enzymes that excise damaged bases.9,11,12,102,103,123 The first step toward understanding how normal bases differ from damaged bases is to characterize the naturally occurring normal compounds. The pyrimidine nucleobases have previously been the subject of some study both theoretically and experimentally; to our knowledge, there have been two experimental studies of the gas phase proton affinity of the most basic site and one on the most acidic site of cytosine and thymine.6,8,119,171-181 Herein, we provide a comprehensive examination of the gas phase thermochemical properties of the naturally occurring pyrimidine bases thymine and cytosine as well as of the 1-methyl derivative of cytosine. We measure multiple acidities (more and less acidic sites, not

61 heretofore accomplished) as well as the proton affinity of thymine and cytosine. We also provide new data for 1-methyl cytosine which, unlike cytosine, does not have several possible stable tautomeric states in the gas phase.

4.2 Experimental

All chemicals except 1-methyl cytosine are commercially available and were used as received. 1-Methyl cytosine was synthesized following literature procedure.182

Acidity and proton affinity bracketing experiments (more site and less site) were conducted using a Fourier Transform Ion Cyclotron Resonance Mass Spectrometer

(FTMS) with a dual cell set up which has been described previously.11,12,123 More details may be found in Chapter 2.

In addition, we also used the Cooks kinetic method in a quadrupole ion trap (LCQ) mass spectrometer67,69,107-109 to measure the proton affinity and acidity of cytosine. The procedure for conducting these experiments in our lab has been described previously and in Chapter 2.103 In contrast to Chapters 2 and 3, analysis of the data was conducted using the Cooks "extended" kinetic method.183-187 This method has been well-described and involves acquiring ion abundance ratios at different collision energies (and therefore different effective temperatures (vide infra)), which allows for deconvolution of the enthalpic and entropic contributions. The data were worked up using a method developed by Armentrout, which is related to a method developed by Fenselau and Wesdemiotis.183-

185 Eq. 4.2-4.5 summarize the data analysis. Teff is the effective temperature of the dissociating proton bound complex in Kelvin. The term "∆(∆S)" is the difference in the

∆S associated with the two channels in Eq. 4.1. In the Fenselau-Wesdemiotis method, a

+ + plot of ln(k1/k2) (which is equal to ln([AH ]/[BiH ]), Eq. 4.3) versus PA(Bi) yields the Teff from the slope (Eq. 4.2) and the "GBapp(A)" from the intercept (Eq. 4.2, Eq. 4.4).

Plotting Eq. 4.4 at different values of Teff yields the proton affinity and ∆∆S for the

Cooks measurement. However, Armentrout noted that a more statistically rigorous analysis would involve using Eq. 4.1-4.4, but instead of plotting ln(k1/k2) versus PA(Bi)

(from Eq. 4.2), one would plot ln(k1/k2) versus [PA(Bi) - PAavg], where PAavg is equal to the average PA of the reference bases used. This would result in a "new" y-intercept of y01' (Eq. 4.5). Plotting the y01' values versus 1/RTeff (Eq. 4.5) will yield ∆∆S/R (from the y-intercept) and PA(A) (from the slope, which equals PA(A)-PAavg).

ln(k1 / k 2 ) = [(PA(A) / RTeff ) − Δ(ΔS) / R] − PA(Bi ) /(RTeff ) Eq. 4.2

+ + n(k1 / k 2 ) = ln([AH ]/[Bi H ] Eq. 4.3

app GB (A) / RTeff = PA(A) /(RTeff ) − Δ(ΔS) / R Eq. 4.4

y01 '= [(PA(A) − PAavg ) / RTeff ] − Δ(ΔS) / R Eq. 4.5

4.3 Results

4.3.1 Cytosine

4.3.1.1 Computational results: tautomers

Cytosine, as with all nucleobases, has several possible tautomeric forms (six most stable shown in Figure 4.1).6,19,173,175-179,181,188-196 At B3LYP/6-31+G*, we find that the canonical tautomer "N1H" (1) is the most stable, but that three other tautomers ("O2H" enols 1a and 1b and "N3H" imine 1c) are very close in energy to the canonical form. The next most stable tautomers are 4.0 and 6.7 kcal mol-1 less stable than the canonical; the remaining tautomers are all quite a bit higher in energy.

195.2(187.9) 197.0(190.1) 194.9(188.3) 228.8(221.2) 354.1(346.7) 349.9(342.6) 348.5(341.2) 352.6(344.6)H H 346.6(338.8) H H 351.8(344.5)H H H N N N 350.4(342.0) N 226.5(218.6)3 3 208.9(201.8)3 3 H 218.8(211.2) H H 182.7(175.7)H H 225.5(217.4) N N N N 180.1(173.0) 340.8(333.1) 198.8(191.4) 1 1 H 1 1 O N H O N H O N H O N H 218.7(211.3) 180.9(173.8) H343.3(336.0) H 228.0(220.3) H335.4(328.1) 341.6(333.9) 197.5(190.2) 217.0(209.1) 0.0 1.7 2.5 2.2 N1H (1) O2Ha (1a) O2Hb (1b) N3Ha (1c)

H N NH2 H 3 H H 3 H N N 1 1 O N H O N H H 4.0 6.7

Figure 4.1. Relative enthalpies (∆H in kcal mol-1) of six possible tautomers of cytosine,

4.3.1.2 Computational results: acidities

The calculated values for the acidity of the four most stable tautomers of cytosine

197 are shown in red in Figure 4.1. The ∆Hacid of the most acidic site for the canonical and

-1 -1 enol tautomers appears to be around 341-343 kcal mol (∆Gacid = 333-336 kcal mol ),

64 whereas for the imine tautomer 1c, the most acidic site is calculated to be closer to ∆Hacid

-1 -1 ~335 kcal mol (∆Gacid ~ 328 kcal mol ).

4.3.1.3 Computational results: proton affinities

The calculated values for the proton affinity (PA; ∆H) and gas phase basicity (GB;

∆G) of the four most stable tautomers of cytosine are shown in blue in Figure 4.1. The

PA/GB of the most basic site of cytosine varies by tautomer. For the canonical tautomer, there are two sites that are close in basicity: the O2 (PA = 225.5 kcal mol-1; GB = 217.4 kcal mol-1) and the N3 (PA = 226.5 kcal mol-1; GB = 218.6 kcal mol-1). The most basic sites for the enol 1b and the imine 1c have PAs similar to that of the canonical (1b: PA =

228.0; GB = 220.3 kcal mol-1; 1c: PA = 228.8; GB = 221.2 kcal mol-1). In contrast, the most basic site of the enol tautomer 1a has a calculated PA of 219 kcal mol-1 (GB = 211 kcal mol-1), corresponding to either the N1 or the N3 site.

4.3.1.4 Cytosine experimental results: acidities

Table 4.1. Summary of results for acidity bracketing of more acidic site of cytosine.

a a b Reference compound ΔHacid ΔGacid Proton transfer

(kcal mol-1) (kcal mol-1) Ref. acid Conj.

base

acetic acid 348.1 ± 2.2 341.1 ± 2.0 – +

formic acid 345.3 ± 2.2 338.3 ± 2.0 – +

2,4-pentadione 343.8± 2.1 336.7± 2.0 – +

ethoxyacetic acid 342.0 ± 2.2 335.0± 2.0 + +

3-chloropropanoic acid 340.8 ± 2.7 333.8 ± 2.0 + –

trifluoro-m-cresol 339.3 ± 2.1 332.4 ± 2.0 + –

methyl cyanoacetate 340.8 ± 0.6 334.5 + –

2-chloropropanoic acid 337.0 ± 2.1 330.4 ± 2.0 + –

difluoroacetic acid 331.0 ± 2.2 323.8 ± 2.0 + –

1,1,1-trifluoro-2,4- 328.3 ± 2.9 322.0 ± 2.0 + –

pentadione

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

The bracketing results for the more acidic site of cytosine are summarized in Table

4.1.8 We find that deprotonated cytosine can deprotonate neutral ethoxyacetic acid

-1 (∆Hacid = 342.0; ∆Gacid = 335.0 kcal mol ) and more acidic compounds, but cannot

-1 deprotonate neutral 2,4-pentadione (∆Hacid = 343.8; ∆Gacid = 336.7 kcal mol ) nor less acidic compounds. In the opposite direction, ethoxyacetate deprotonates neutral cytosine, as do more basic bases. Less basic bases do not deprotonate neutral cytosine. We are therefore able to bracket the acidity of the more acidic site of cytosine as ∆Hacid = 342 ± 3

-1 -1 9-13,103,123 kcal mol (∆Gacid= 335 ± 3 kcal mol ).

Table 4.2. Summary of results for acidity bracketing of less acidic site of cytosine.

a a Reference compound ΔHacid ΔGacid Proton

transferb

(kcal mol-1) (kcal mol-1)

pyrrole 359.6 ± 2.9 351.8 ± 2.0 –

chloroacetonitrile 357.7 ± 2.2 350.0 ± 2.0 –

1-propane thiol 354.2 ± 2.2 347.9 ± 2.0 –

2-propane thiol 353.4 ± 2.2 347.1 ± 2.0 –

4-(trifluoromethyl)-aniline 353.3 ± 2.1 346.0 ± 2.0 –

p-cresol 350.3 ± 2.1 343.4 ± 2.0 +

m-cresol 349.6 ± 2.1 342.7 ± 2.0 +

acetic acid 348.1 ± 2.2 341.1 ± 2.0 +

formic acid 345.3 ± 2.2 338.3 ± 2.0 +

2,4-pentadione 343.8 ± 2.1 336.7 ± 2.0 +

methyl cyanoacetate 340.8 ± 0.6 334.5 +

a Acidities are in kcal mol-1. bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

Using methodology developed in our lab, we also measured the less acidic site of

cytosine (Table 4.2).9-13,103,123 We find that deprotonated cytosine under "less acidic conditions" cannot deprotonate acids equal to and less acidic than 4-(trifluoromethyl)-

-1 aniline (∆Hacid = 353.3; ∆Gacid = 346.0 kcal mol ) but can deprotonate acids equal to and

-1 more acidic than p-cresol (∆Hacid = 350.3; ∆Gacid = 343.4 kcal mol ). We therefore

-1 bracket the less acidic site of cytosine as ∆Hacid = 352 ± 4 kcal mol (∆Gacid = 345 ± 4 kcal mol-1).

4.3.1.5 Experimental results: proton affinities

Table 4.3. Summary of results for PA bracketing of more basic site of cytosine.

Reference compound PAa GBa Proton transferb

(kcal mol-1) (kcal mol-1) Ref. Conj.

base acid

1-methyl piperidine 232.1 ± 2.0 224.7 ± 2.0 + _

1-methyl pyrrolidine 230.8 ± 2.0 223.4 ± 2.0 + –

piperidine 228.0 ± 2.0 220.0 ± 2.0 + +

pyrrolidine 226.6 ± 2.0 218.8 ± 2.0 – +

3-picoline 225.5 ± 2.0 217.9 ± 2.0 – +

pyridine 222.3 ± 2.0 214.7 ± 2.0 – +

N-methyl propanamide 220.0 ± 2.0 212.6 ± 2.0 – +

aValues are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

The bracketing results for the proton affinity of the most basic site of cytosine are shown in Table 4.3. We find that the reaction between cytosine and piperidine (PA =

228.0; GB = 220.0 kcal mol-1) proceeds in "both directions" (cytosine deprotonates protonated piperidine and piperidine deprotonates protonated cytosine); with 1-methyl

-1 pyrrolidine (PA = 230.8; GB = 223.4 kcal mol ), reaction with protonated cytosine does occur, but the opposite reaction does not. With pyrrolidine (PA = 226.6; GB = 218.8 kcal mol-1), cytosine can deprotonate protonated pyrrolidine, but the opposite reaction does not proceed. These data point to a PA for cytosine as 228 ± 3 kcal mol-1 (∆G = gas phase basicity (GB) = 220 ± 3 kcal mol-1).

4.3.2 1-Methyl Cytosine

4.3.2.1 Computational results: tautomers

As with cytosine, 1-methyl cytosine has more than one tautomer that is low-lying in energy, but fewer than with cytosine (five most stable tautomers shown in Figure

4.2).180,198 The canonical tautomer is calculated to be 2.9 kcal mol-1 more stable than the next nearest (imine) tautomer; the other imine tautomer lies 4.6 kcal mol-1 higher in energy than the canonical structure.

199.0(191.7) 354.1(345.9) 348.3(340.3) 232.9(224.9) H H H N 350.7(344.5) N 376.4(370.7) 230.0(222.4)3 H 377.3(369.8) H H N N 228.6(220.6) 201.9(194.5) 1 O N H 369.6(362.5) O N H 367.8(360.9) CH3 CH3 219.6(211.8) 200.8(192.9) 0.0 2.9

H N H H N N H H N H H N N O N H H H O N H O N H CH3 CH3 CH3 4.6 21.4 24.8 Figure 4.2. Relative enthalpies (∆H in kcal mol-1) of five possible tautomers of 1- methyl cytosine, and acidities (red values; ∆Hacid, with ∆Gacid in parentheses; all values in kcal mol-1), and proton affinities (blue values; PA, with GB in parentheses; all values in kcal mol-1) of the two most stable tautomers, calculated at B3LYP/6-31+G* (298 K)

4.3.2.2 Computational results: acidities

The acidities of the two lowest-lying tautomers of 1-methyl cytosine are shown in

Figure 4.2. The acidity of both tautomers is similar (for canonical: ∆Hacid = 348.3, ∆Gacid

-1 - = 340.3 kcal mol ; for higher-energy tautomer: ∆Hacid = 350.7, ∆Gacid = 344.5 kcal mol

1).

4.3.2.3 Computational results: proton affinities

The proton affinities of 1-methyl cytosine are also in Figure 4.2. There are two sites on the canonical tautomer that are comparable in PA: the O2 (PA = 228.6, GB = 220.6 kcal mol-1) and the N3 (PA = 230.0, GB = 222.4 kcal mol-1). The imine tautomer is slightly more basic; the calculated PA value is 232.9 kcal mol-1 (GB = 224.9 kcal mol-1).

4.3.2.4 Experimental results: acidities

Table 4.4. Summary of results for acidity bracketing of more acidic site of 1-methyl cytosine.

a a b Reference compound ΔHacid ΔGacid Proton transfer

(kcal mol-1) (kcal mol-1) Ref. acid Conj.

base

acetone 369.1 ± 2.1 361.9 ± 2.0 – +

pyrrole 359.6 ± 2.9 351.8 ± 2.0 – +

4-(trifluoromethyl)-aniline 353.3 ± 2.1 346.0 ± 2.0 – +

m-cresol 349.6 ± 2.1 342.7 ± 2.0 – +

acetic acid 348.1 ± 2.2 341.1 ± 2.0 + –

butanoic acid 346.5 ± 2.2 339.5 ± 2.0 + –

formic acid 345.3 ± 2.2 338.3 ± 2.0 + –

2,4-pentadione 343.8 ± 2.1 336.7 ± 2.0 + –

trifluoro-m-cresol 339.3 ± 2.1 332.4 ± 2.0 + –

aAcidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

The acidity bracketing results for 1-methyl cytosine are shown in Table 4.4.

Deprotonated 1-methyl cytosine can deprotonate acetic acid (∆Hacid = 348.1; ∆Gacid =

-1 341.1 kcal mol ) and more acidic acids, but cannot deprotonate m-cresol (∆Hacid = 349.6;

-1 ∆Gacid = 342.7 kcal mol ) nor acids that are less acidic. Likewise, m-cresolate can deprotonate 1-methyl cytosine, but acetate cannot. We therefore bracket a ∆Hacid of 1-

-1 -1 methyl cytosine of 349 ± 3 kcal mol (∆Gacid = 342 ± 3 kcal mol ).

4.3.2.5 Experimental results: proton affinities

Table 4.5. Summary of results for PA bracketing of more basic site of 1-methyl cytosine.

Reference compound PAa GBa Proton transferb

(kcal mol-1) (kcal mol-1) Ref. Conj.

base acid

triethylamine 234.7 ± 2.0 227.0 ± 2.0 + –

1-methyl pyrrolidine 230.8 ± 2.0 223.4 ± 2.0 + –

2,4-lutidine 230.1 ± 2.0 222.5 ± 2.0 + +

piperidine 228.0 ± 2.0 220.0 ± 2.0 – +

4-picoline 226.4 ± 2.0 218.8 ± 2.0 – +

pyridine 222.0 ± 2.0 214.7 ± 2.0 – +

2,4-pentadione 208.8 ± 2.0 200.0 ± 2.0 – +

aValues are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

The proton affinity bracketing results for 1-methyl cytosine are summarized in

Table 4.5. 1-Methyl pyrrolidine (PA = 230.8; GB = 223.4 kcal mol-1) is basic enough to deprotonate protonated 1-methyl cytosine, but piperidine (PA = 228.0; GB = 220.0 kcal mol-1) is not. In the opposite direction, 1-methyl cytosine can deprotonate protonated piperidine but cannot deprotonate protonated 1-methyl pyrrolidine. The reaction with 2,4- lutidine (PA = 230.1; GB = 222.5 kcal mol-1) proceeds in both directions. We therefore bracket the PA of 1-methyl cytosine to be 230 ± 3 kcal mol-1 (GB = 223 ± 3 kcal mol-1).

4.3.3 Thymine

4.3.3.1 Computational results: tautomers

The canonical structure of thymine is the most stable by far; it is calculated to be about 12 kcal mol-1 more stable than the next nearest tautomer (Figure 4.3; higher energy tautomers are in Supporting Information).

201.0(193.4) 203.7(196.0) O O 3 344.8(336.9) H CH3 H 3 CH N N 3 196.1(188.8) 1 1 O N H O N H H H 332.2(324.8) 194.9(187.6) 0.0 11.9 Figure 4.3. Relative enthalpies (∆H in kcal mol-1) of the two most stable thymine

4.3.3.2 Computational results: acidities

-1 The more acidic site of thymine has a calculated ∆Hacid of 332.2 kcal mol (∆Gacid =

324.8 kcal mol-1), at N1 (Figure 4.3, in red).

4.3.3.3 Computational results: proton affinities

The more basic site of thymine has a calculated PA of 203.7 kcal mol-1 (GB = 196.0 kcal mol-1), at O4 (Figure 4.3, in blue).

4.3.3.4 Experimental results: acidities

The acidity bracketing results for the more acidic site of thymine are shown in Table

4.6. We find that deprotonated thymine cannot deprotonate 2-bromopropionic acid

-1 (∆Hacid = 336.8; ∆Gacid = 329.8 kcal mol ) but that the opposite reaction does occur.

Deprotonated thymine does react with pyruvic acid (∆Hacid = 333.5; ∆Gacid = 326.5 kcal

72 mol-1), but pyruvate does not deprotonate thymine. We therefore bracket the acidity of

-1 -1 thymine to be ∆Hacid = 335 ± 4 kcal mol (∆Gacid = 328 ± 4 kcal mol ).

Table 4.6. Summary of results for acidity bracketing of more acidic site of thymine.

a a b Reference compound ΔHacid ΔGacid Proton transfer

(kcal mol-1) (kcal mol-1) Ref. Conj.

acid base

2,4 pentadione 343.8 ± 2.1 336.7 ± 2.0 – +

trifluoro-m-cresol 339.3 ± 2.1 332.4 ± 2.0 – +

2-bromopropionic acid 336.8 ± 2.1 329.8 ± 2.0 – +

pyruvic acid 333.5 ± 2.9 326.5 ± 2.8 + –

per-fluoro-tert-butanol 331.6 ± 2.2 324.0 ± 2.0 + –

difluoroacetic acid 331.0 ± 2.2 323.8 ± 2.0 + –

1,1,1-trifluoro-2,4- 328.3 ± 2.9 322.0 ± 2.0 + –

pentadione

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

We have also bracketed the less acidic site of thymine (Table 4.7).8 Deprotonated thymine under "less acidic" conditions does not deprotonate butyric acid (∆Hacid = 346.5;

-1 ∆Gacid = 339.5 kcal mol ) but does deprotonate isovaleric acid (∆Hacid = 345.5; ∆Gacid =

-1 338.5 kcal mol ). We therefore bracket the less acidic site of thymine to be ∆Hacid = 346

-1 -1 ± 3 kcal mol (∆Gacid = 339 ± 3 kcal mol ).

Table 4.7. Summary of results for acidity bracketing of less acidic site of thymine.

a a Reference compound ΔHacid ΔGacid Proton

transferb

(kcal mol-1) (kcal mol-1)

pentane thiol 352.5 ± 2.3 346.2 ± 2.5 –

m-cresol 349.6 ± 2.1 342.7 ± 2.0 –

acetic acid 348.1 ± 2.2 341.1 ± 2.0 –

butyric acid 346.5 ± 2.2 339.5 ± 2.0 –

isovaleric acid 345.5 ± 2.1 338.5 ± 2.0 +

formic acid 345.3 ± 2.2 338.3 ± 2.0 +

2,4-pentadione 343.8 ± 2.1 336.7 ± 2.0 +

methyl cyanoacetate 340.8 ± 0.6 334.5 +

trifluoro-m-cresol 339.3 ± 2.1 332.4 ± 2.0 +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

4.3.3.5 Experimental results: proton affinities

The proton affinity results for thymine are summarized in Table 4.8. We find that

pyrimidine (PA = 211.7; GB = 204.5 kcal mol-1) does deprotonate protonated thymine,

but the opposite reaction does not occur. Pyrrole, however (PA = 209.2; GB = 201.7 kcal

mol-1), is not basic enough to deprotonate protonated thymine; thymine does deprotonate protonated pyrrole. The PA of thymine is therefore bracketed to be 211 ± 3 kcal mol-1

(GB = 203 ± 3 kcal mol-1).

Table 4.8. Summary of results for PA bracketing of more basic site of thymine.

Reference compound PAa GBa Proton transferb

(kcal mol-1) (kcal mol-1) Ref. Conj.

base acid

2-chloropyridine 215.3 ± 2.0 208.0 ± 2.0 + –

o-toluidine 212.9 ± 2.0 205.3 ± 2.0 + –

pyrimidine 211.7 ± 2.0 204.5 ± 2.0 + –

pyrrole 209.2 ± 2.0 201.7 ± 2.0 – +

2,4-pentadione 208.8 ± 2.0 200.0 ± 2.0 – +

m-chloro-aniline 207.5 ± 2.0 199.9 ± 2.0 – +

methyl styrene 206.0 ± 2.0 199.0 ± 2.0 – +

diethyl sulfide 204.8 ± 2.0 197.7 ± 2.0 – +

4-methyl-cyclohexanone 201.9 ± 2.0 194.3 ± 2.0 – +

cyclohexanone 201.0 ± 2.0 193.9 ± 2.0 – +

2-butanone 197.7 ± 2.0 190.1 ± 2.0 – +

acetone 194.0 ± 2.0 186.9 ± 2.0 – +

aValues are in kcal mol-1. bA “+” indicates the occurrence and a “–” indicates the absence of proton

transfer.

4.4 Discussions

4.4.1 Cytosine

We have measured the most acidic site of cytosine to have a ∆Hacid = 342 ± 3 kcal

-1 -1 mol (∆Gacid = 335 ± 3 kcal mol ). This value is in agreement with that previously

measured by Chen and Chen (340 ± 2 kcal mol-1) using a variety of methods: derivation from DMSO pKa values; electron impact; negative chemical ionization mass

119,171 spectrometry. We have also measured the less acidic site to have a ∆Hacid = 352 ± 4

kcal mol-1. The proton affinity is bracketed to 228 ± 3 kcal mol-1. This PA value is also in

agreement with previous measurements (227.0 kcal mol-1).8,172,174,199,200

In aqueous solution and the solid state, the canonical tautomer of cytosine (1) is the predominant form.188-190,192,194-196 One of the issues with the exploration of cytosine in the gas phase, however, is what tautomer(s) might be present. Our calculations (Figure 4.1) and others' indicate that there are four tautomers (1, 1a, 1b, and 1c, Figure 4.1) that are close in stability.6,173,175,176,178,179,194,201 Turecek and Wesdemiotis and coworkers have done a superb job of assessing the energy orderings of cytosine versus calculational level and conclude that while the relative energies are very sensitive to the type of basis set used (with B3LYP calculations, for example, consistently favoring 1), it is clear that 1, 1a and 1b are all very close in energy and are likely to coexist in the gas phase, with possibly a small amount of 1c.178 Experiments in the gas phase are consistent with calculations in so far as mixtures of tautomers are found. Resonance enhanced multiphoton ionization (REMPI) experiments indicated that 1, 1a, and possibly 1b coexist in the gas phase.191 This same mixture plus 1c were observed in IR matrix isolation studies.193 Molecular beam microwave (MW) spectroscopy studies showed the presence of 1, 1a, and 1c.202

Can we use our bracketed acidity and proton affinity values and compare them to the calculated values (Figure 4.1) to discern which tautomers are present under our gas phase conditions? If we presume that we have a mixture of 1, 1a, 1b, and 1c, and we conduct the "more acidic" gas phase bracketing experiment, we would expect the reactivity depicted in Scheme 4.1. The bracketing reaction is conducted in two directions: deprotonated reference acid (A–) plus cytosine, and deprotonated cytosine plus reference acid HA. In the former, tautomers 1, 1a and 1b are indistinguishable: any A– that is basic

-1 enough (a ∆Hacid value for HA of at least around 341-343 kcal mol , corresponding to the

76 most acidic site of all three tautomers) will deprotonate each. The experiment cannot differentiate among those three tautomers. Tautomer 1c is different: its predicted acidity

-1 – is ∆Hacid ~ 335 kcal mol . If 1c is present, an anion A whose ∆Hacid (HA) is equal to or greater than 335 kcal mol-1 should result in proton transfer. Therefore, in this direction,

– – when a series of conjugate bases of reference acids A is introduced, if an A with a ∆Hacid value (for the conjugate acid HA) lower than ~342 deprotonates cytosine, then 1c is present. Interestingly, the reaction in the opposite direction is not as informative (Scheme

4.1). Deprotonation of 1, 1a, or 1b results in the same anion. Reaction with HA could protonate N1 or O2 to produce any of those three tautomers; for all three, the reaction

-1 should proceed if ∆Hacid (HA) is equal to or less than about 342 kcal mol . If 1c is also present, the reactivity will not change; the reaction turns "on," or results in a "+," for any

-1 HA with a ∆Hacid value of 342 kcal mol or lower, including an HA with an acidity of

335 kcal mol-1.

This complicated situation is well summarized in a hypothetical acidity bracketing table (Table 4.9). In essence, if multiple tautomers with differing acidities are present in appreciable amounts, one will not see a clean "crossover" point in an acidity bracketing table. Instead, there will be a large section where there are two "+" indicators, spanning the acidities of the different tautomers (in this hypothetical example, 335-343 kcal mol-1).

We do not see such a pattern (Table 4.1), which would imply that under our

-1 conditions, the imine tautomer 1c (∆Hacid(calc) = 335 kcal mol ) is not present. However, it is also possible that the imine tautomer is simply present in very small amounts (a calculation by Turecek, Wesdemiotis and coworkers estimates 5% at 473 K, the

77 temperature of a typical ion source), and we therefore do not see its contribution to reactivity.178 Thus, we cannot rule out the presence of 1c.

Scheme 4.1

Deprotonated reference acid (AŠ) + cytosine H H N 3 H N 1 O N H H 343.3 1

H H N 3 H N AŠ Š 1 reaction will occur for any A for which the ĘHacid (HA) value is O N H about equal to or greater than 341-343 kcal mol-1 H 341.6 1a

H H N 3 H 340.8 N H 1 O N H

1b H N 3 H H AŠ N Š reaction will occur for any A for which the ĘHacid (HA) value is 1 -1 O N H about equal to or greater than 335 kcal mol H 335.4 1c

Deprotonated cytosine + reference acid (HA) H H N 3 H N 1 O N H

HA Deprotonated 1, 1a and 1b reaction will occur for any HA for which the ĘHacid (HA) value is about equal to or less than 341-343 kcal mol-1

H N H 3 H N 1 O N H

Deprotonated 1c

Table 4.9. Expected acidity bracketing table if cytosine tautomers 1, 1a, 1b, and 1c are present.

a b ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

347 – +

345 – +

343 + +

341 + +

339 + +

337 + +

335 + +

333 + –

331 + –

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer.

We can also compare our less acidic site acidity measurement with calculations.

From previous experiments, we have found that when measuring the acidity of a "less acidic" site, we measure the "least acidic ion" present; thus, any tautomer with a proton that has an acidity near 352 kcal mol-1 could contribute to the measurement of that value.11-13,103 Therefore, the less acidic measurement could be attributable to any of the four most stable tautomers 1, 1a, 1b, and 1c. For PA, the measured proton affinity of cytosine is 228 kcal mol-1. This could correspond to tautomers 1 (probably both the O2 and N3 sites, which will be indistinguishable), 1b, and/or 1c. Tautomer 1a has a much

79 lower calculated PA of ~219 kcal mol-1. Since protonated cytosine under our conditions does not react with reference bases whose PAs are 226.6 kcal mol-1 and lower, it is possible that 1a is not present. However, once 1a is protonated, a series of proton transfers could occur that would simply produce an isomeric protonated ion, which would appear as a lack of reaction (shown in Scheme 4.2 with initial N1-protonated ion of 1a).

This is simply a limitation of the mass spectrometric experiment.

Scheme 4.2

H H H H N N H H N B B N 1 O N H PA(B) = 220 kcal mol-1 O N H H H 218.7 H H

H H H H N N H BH H B N N H 225.5 1 O N H O N H H H 217.0 1

H H N H B + N H O N H H

Thus, under our FTMS bracketing conditions, wherein cytosine is introduced via a heated solids probe, our results are consistent with a cytosine mixture of 1, 1a, 1b, and possibly 1c. This is also consistent with the previous gas phase results.191,193,202

We followed up the bracketing experiments with Cooks extended kinetic method measurements of the acidity and proton affinity of cytosine. In this experiment, a proton- bound dimer of cytosine and a reference acid or base is formed in solution and vaporized via electrospray. It is interesting that the cytosine is electrosprayed from aqueous solution as opposed to sublimed from the solid state; most likely the canonical form of cytosine is therefore the predominant reactant.180,189 For the Cooks kinetic measurement of cytosine acidity, the following reference acids were used: trifluoro-m-cresol (∆Hacid = 339.3 ± 2.1

-1 -1 kcal mol ), methoxyacetic acid (∆Hacid = 341.9 ± 2.1 kcal mol ), ethoxyacetic acid

-1 -1 (∆Hacid = 342.0 ± 2.2 kcal mol ) and 2-chlorophenol (∆Hacid = 343.4 ± 2.3 kcal mol ).

For the cytosine PA measurement, we used 1-methyl pyrrolidine (PA = 230.8 ± 2.0 kcal mol-1), piperidine (PA = 228.0 ± 2.0 kcal mol-1), pyrrolidine (PA = 226.6 ± 2.0 kcal mol-

1), 3-picoline (PA = 225.5 ± 2.0 kcal mol-1), pyridine (PA = 222.3 ± 2.0 kcal mol-1), and

1-octanamine (PA = 222.0 ± 2.0 kcal mol-1). The Cooks extended kinetic method

-1 -1 183- experiments yield a ∆Hacid of 343 ± 3 kcal mol and a PA of 227 ± 3 kcal mol .

186,203,204 These data are in agreement with the calculated values for the most acidic and most basic site of the canonical cytosine tautomer 1 (as well as, of course, 1b and 1c, though given the aqueous conditions, we consider the presence of those tautomers unlikely).

4.4.2 1-Methyl Cytosine

We synthesized and examined this derivative of cytosine because methylation of the

1-position decreases the number of energetically accessible tautomers. We find that at

B3LYP/6-31+G*, the canonical tautomer is the most stable, by about 3 kcal mol-1

(Figure 4.2). This value is in agreement with calculations conducted at B3-MP2/6-

311++G**//B3LYP/6-31+G** by Turecek and coworkers.180 These same authors also predict that at 298 K, the canonical tautomer will predominate in the gas phase (98.3% equilibrium fraction); at 473 K, the canonical tautomer still predominates (92%). We would therefore expect to see prevalent reactivity of the canonical tautomer of 1-methyl

-1 -1 cytosine. We measure a ∆Hacid = 349 ± 3 kcal mol (∆Gacid = 342 ± 3 kcal mol ) and a

PA of 230 ± 3 kcal mol-1 (GB = 223 ± 3 kcal mol-1). These values are in agreement with the calculated acidity and basicity values for most acidic and most basic site of the canonical tautomer, and provide benchmarking data that this level of calculation is reasonable for the acidity and proton affinity of cytosine and derivatives.

4.4.3 Thymine

The canonical tautomer of thymine is by far the most stable; by our calculations, the next-nearest tautomer is about 12 kcal mol-1 less stable (Figure 4.3).6,173,175 The acidity of

-1 thymine at the most acidic site, N1, is calculated to be ∆Hacid = 332.2 kcal mol ; ∆Gacid =

-1 -1 324.8 kcal mol . We measure the acidity to be ∆Hacid = 335 ± 4 kcal mol (∆Gacid = 328

± 4 kcal mol-1), which is also in agreement with a previously measured gas phase value

-1 119,171 (∆Hacid = 333 ± 2 kcal mol , using different methods). The next acidity brackets to

-1 -1 346 ± 3 kcal mol (∆Hacid; ∆Gacid = 339 ± 3 kcal mol ), which is in agreement with the

-1 -1 calculated value for N3-H acidity of 344.8 kcal mol (∆Hacid; ∆Gacid = 336.9 kcal mol ).

The proton affinity measurement yields a value of PA = 211 ± 3 kcal mol-1 (GB = 203 ± 3 kcal mol-1), which agrees with previous measurements (PA = 210.5 kcal mol-

1).8,172,174,200,205 The value of 211 kcal mol-1 is somewhat surprising in light of the calculated value of 203.7 kcal mol-1. We are not certain why the calculated and experimental values are quite different; we saw a similar result when we examined uracil, which differs from thymine only by lack of the methyl group on C5.13,175,178 Higher level calculations do bring the calculated values closer to the experimental; we find that at

B3LYP/6-311++G**, the PA of thymine is calculated to be 206.6 kcal mol-1.6,206

4. 5 Conclusions

We have calculated and measured the acidity and proton affinity of cytosine, 1- methyl cytosine, and thymine, to probe the intrinsic reactivity of these pyrimidine nucleobases. We are interested in particular in how damaged bases differ from normal bases. DNA is inevitably damaged by environmental mutagens as well as chemotherapeutics; such mutations are linked to carcinogenesis and aging.14,15,17 Our lab studies the mechanisms by which enzymes (primarily glycosylases) might cleave damaged bases from DNA, thereby protecting our genome.9,11,12,102,103,123 Previous results from our lab have shown that the properties of normal versus damaged bases lend insight into the mechanisms by which the damaged bases are cleaved.9,11,12,102,103,123 The first step toward understanding how normal bases differ from damaged bases is to characterize the naturally occurring normal compounds, which motivates the study herein.

Cytosine has two measurable acidic sites, measured using the acidity bracketing

-1 -1 method: ∆Hacid = 342 ± 3 kcal mol and 352 ± 4 kcal mol (∆Gacid = 335 ± 3 and 345 ±

4 kcal mol-1). The PA of cytosine brackets to 228 ± 3 kcal mol-1 (GB = 220 ± 3 kcal mol-

1). Comparison of these values to theoretical data indicates that under our conditions, we probably have a mixture of the canonical tautomer 1, the enol tautomers 1a and 1b, and possibly the imine tautomer 1c. We also measured the acidity of the most acidic and the proton affinity of the most basic site of cytosine using the Cooks extended kinetic method; in these experiments, protonated dimers with cytosine are electrosprayed from aqueous solution and are therefore more likely to consist predominantly of the canonical tautomer.

-1 Using the Cooks method, we measure the ∆Hacid to be 343 ± 3 kcal mol and the PA to be 227 ± 3 kcal mol-1. We also examined the properties of 1-methyl cytosine, which is predicted to be predominantly the canonical tautomer in the gas phase and therefore allows benchmarking of the calculational level. We measure the ∆Hacid to be 349 ± 3 kcal

-1 -1 mol (corresponding to the exocyclic -NH2; ∆Gacid = 342 ± 3 kcal mol ) and the PA to be

230 ± 3 kcal mol-1 (corresponding to the O2 and/or N3; GB = 223 ± 3 kcal mol-1); both these values are in agreement with calculation. Last, we examined the pyrimidine nucleobase thymine. Thymine exists as the canonical tautomer in the gas phase and has a

-1 -1 ∆Hacid of 335 ± 4 kcal mol (∆Gacid = 328 ± 4 kcal mol ) for the more acidic N1 site and

-1 -1 ∆Hacid of 346 ± 3 kcal mol (∆Gacid = 339 ± 3 kcal mol ) for the less acidic N3 site. The

PA brackets to 211 ± 3 kcal mol-1 (corresponding to the O4 site; GB = 203 ± 3 kcal mol-1).

Thus, it is much easier to deprotonate neutral thymine in the gas phase than it is to deprotonate neutral cytosine, which is also true in solution (pKa of thymine = 9.9; pKa of cytosine = 12.2).207 For protonation of the neutral bases, cytosine is more basic than thymine in the gas phase; this is also true in solution (pKa of protonated thymine = 0; pKa of protonated cytosine = 4.45).207 Studies are currently underway on adenine and guanine

84 to measure the thermochemical properties, which will allow for a full comparison of all five RNA/DNA nucleobases in the gas phase and in solution.

Note: Major parts of the following chapter have been published: Liu, M.; Tran, T. N.; Franz, A. K.; Lee, J. K. J. Org. Chem., Article ASAP

Chapter 5 Gas Phase Acidity Studies of Dual Hydrogen-Bonding

Organic Silanols and Organocatalysts

5.1 Introduction

Organic silanols and silanediols have great potential in molecular recognition and catalysis, due in large part to the ability to be both a proton donor and a proton acceptor.19,208-213 A few studies of acidity exist in the gas phase and in solution for simple organic silanols, but no systematic study has been conducted, and no studies exist for more complex systems.19,214 In this paper, we focus on a series of monosilanols, silanediols, and disiloxanediols, in comparison to several known hydrogen-bonding catalysts, in an effort to characterize their fundamental properties. Silanediols R2Si(OH)2 are of particular interest because they contain a geminal-diol bonding motif that is not commonly accessible for carbon analogs, and has the potential to serve as a dual hydrogen-bonding group. Recent studies demonstrate that silanols can function as isosteres and transition state analogs in drug design, where the enhanced acidity of the silanol can improve binding to a receptor.215-218 Antimicrobial activities of monosilanols have been reported where silanols exhibit greater biocidal properties relative to carbon analogs due to enhanced acidity and lipophilicity.219 Studying silanol and silanediol groups may also be useful for understanding local surface sites and reactivity of silica materials for heterogeneous catalysis. The chosen substrates for this study (Figure 5.1) are targets for molecular recognition and catalysis.

Et OH Ph OH Ph OH Si Si Si Et Et CH3 CH3 Ph Ph

1 2 3

t-Bu Mes OH Ph OH Mes OH F C OH Si Si Si 3 Si t-Bu OH Ph OH OH OH 4 5 CF3 F 6 7

i-Pr i-Pr Ph Ph O O i-Pr Si Si i-Pr Ph Si Si Ph HO OH HO OH 8 9

Ph Ph O OH OH CH3

OH CH3 O OH Ph Ph

10 11 CF3

S CH3 R S Ph Ph Ph N N N N N CF3 H H O H H

12 13a: R = t-Bu 13b: R = CH3

Figure 5.1. Silanols and other hydrogen-bonding molecules examined in this paper. Some carbon analogs were also examined and will be indicated by a prime; for example,

2', the carbon analog of 2, is (Me)2PhCOH

5.2 Results

5.2.1 Computational results

The O-H acidity for the silanols and their carbon analogs were calculated at two levels: B3LYP/6-31+G(d) and B3LYP/6-311++G(2df,p) (Table 1).29 The carbon analogs are of particular interest to examine by computation: carbon analogs of the monosilanols are known, stable compounds, whereas those of the diols can be computed, but are generally unstable in aqueous solution, favoring structures that contain carbonyl groups. Although the carbon analogs are not experimentally accessible, calculating their properties is of interest in order to better understand the fundamental similarities and differences between analogous silicon and carbon species. We also calculated the acidities of some known hydrogen-bonding catalysts at the same computational level.

Table 5.1. Computational results for acidity of silanols and their carbon analogs

Silanols B3LYP/6-31+G(d) B3LYP/6-311++G(2df,

(∆H, kcal mol-1) p)

(∆H, kcal mol-1)

silanols carbon silanols carbon

analogs analogs

triethylsilanol (1) 355.3 367.4 358.5 370.3

dimethylphenylsilanol (2) 351.9 361.8 355.3 365.1

triphenylsilanol (3) 344.5 351.3 347.6 354.0

di-tert-butylsilanediol (4) 348.4 348.6 352.2 351.9

diphenylsilanediol (5) 344.2 347.5 348.4 350.8

(4-fluorophenyl)-

(mesityl)silanediol (6) 340.3 339.1 345.0 343.1

(2,6-bis(trifluoromethyl)-

phenyl)(mesityl)silanediol (7) 339.0 335.1 342.8 338.5

1,1,3,3-tetraisopropyldisiloxane-1,3-diol (8) 336.7 339.6 338.8 342.3

1,1,3,3-tetraphenyldisiloxane-1,3-diol (9) 329.4 331.7 333.1 334.6

Table 5.2. Computational results for acidity of known hydrogen-bonding catalysts

Commercial catalyst B3LYP/6-31+G(d) B3LYP/6-

(∆H, kcal mol-1) 311++G(2df, p)

(∆H, kcal mol-1)

BINOL (10) 321.4 323.1

TADDOL (11) 328.5 330.4

N,N-diphenylthiourea (12) 328.2 330.5

2-[[3, 5-Bis(trifluoromethyl)phenyl]thioureido]-N- 322.2 325.2 benzyl-N-methylbutanamide (13b)

5.2.2 Experimental results

The acidity of the silanols was measured in the gas phase using both bracketing and Cooks kinetic methods.67,69,107-109,220,221

5.2.2.1 Monosilanol

Triethylsilanol (1): The acidity of triethylsilanol was measured previously by

-1 Damrauer and coworkers to be between pyrrole (∆Hacid = 359.5 ± 0.5 kcal mol ) and

-1 2,2,2-trifluorethanol (∆Hacid = 361.7 ± 2.5 kcal mol ), using a flowing afterglow-selected ion flow tube mass spectrometer.20 In Damrauer's study, the reaction of the siloxide and the reference acid was monitored. We have repeated this measurement in our Fourier

89 transform mass spectrometer; in our case, we were able to look at the reaction in both "directions": siloxide plus reference acid and the conjugate base of the reference acid plus the silanol (Table 5.3). Our results are consistent with Damrauer's results. Table 5.3. Summary of results for acidity bracketing of triethylsilanol (1).

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

pyrrole 359.5 ± 0.5 + -

2,2,2-trifluoroethanol 361.7 ± 2.5 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

Dimethylphenylsilanol (2): We measured the acidity of dimethylphenylsilanol using two complementary methods: acidity bracketing and the Cooks kinetic method. In the bracketing experiment (Table 4), dimethylphenylsiloxide deprotonates 3-

-1 methylpyrazole (∆Hacid = 356.0 ± 2.1 kcal mol ), but not 3-(trifluoromethyl)aniline

-1 (∆Hacid = 356.9 ± 2.1 kcal mol ). In the opposite direction, 3-(trifluoromethyl)anilide deprotonates dimethylphenylsilanol, but 3-methylpyrazolide does not. Therefore we bracket the acidity of dimethylphenylsilanol to be 356 ± 3 kcal mol-1.

Using the Cooks kinetic method with reference acids p-cresol (∆Hacid = 350.3 ± 2.1

-1 -1 kcal mol ), 3-nitroaniline (∆Hacid = 352.3 ± 2.1 kcal mol ), 4-(trifluoromethyl)aniline

-1 -1 (∆Hacid = 353.3 ± 2.1 kcal mol ), 3,5-dimethylpyrazole (∆Hacid = 353.8 ± 2.1 kcal mol )

-1 and benzamide (∆Hacid = 354.0 ± 2.1 kcal mol ) yields a slightly lower acidity of 354 ± 3 kcal mol-1. Damrauer also measured the acidity of dimethylphenylsilanol.20 The reaction was only followed in one direction (siloxide plus reference acid) and only two reference acids were used: he reported that dimethylphenylsiloxide deprotonates pyrrole but does not

90 deprotonate trifluoroethanol, placing the acidity between 359.5 and 361.7 kcal mol-1. As can be seen from Table 5.4, our results are different; we see no reaction between dimethylphenylsiloxide and pyrrole. Because we also examined the reaction in the reverse direction (deprotonated pyrrole plus dimethylphenylsilanol) and do see a reaction, we think it is likely that dimethylphenylsilanol is more acidic than pyrrole. Furthermore, the acidity measured by our secondary method (Cooks) is in agreement with our bracketed value. Table 5.4. Summary of results for acidity bracketing of dimethylphenylsilanol (2).

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

acetic acid 347.4 ± 0.5 + -

p-cresol 350.3 ± 2.1 + -

1-pentanethiol 352.5 ± 2.3 + -

1-propanethiol 354.2 ± 2.2 + -

3-methylpyrazole 356.0 ± 2.1 + -

3-(trifluoromethyl)aniline 356.9 ± 2.1 - +

fluoroacetone 357.7 ± 3.6 - +

pyrrole 359.5 ± 0.5 - +

2,2,2-trifluoroethanol 361.7 ± 2.5 - +

2-fluoroaniline 362.6 ± 2.2 - +

N-ethylaniline 364.1 ± 2.1 - +

aniline 366.4 ± 2.1 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

Triphenylsilanol (3): The reaction of triphenylsilanol with 3-methoxyphenol

-1 (∆Hacid = 348.0 ± 2.1 kcal mol ) proceeds in both directions, which indicates the acidities

91 are comparable, ~348 kcal mol-1 (Table 5.5). The Cooks kinetic experiments with

-1 reference acids 2-fluorophenol (∆Hacid = 345.3 ± 2.2 kcal mol ), 2-tert-butylphenol

-1 -1 (∆Hacid = 345.8 ± 2.2 kcal mol ), 2-isopropylphenol (∆Hacid = 347.5 ± 2.2 kcal mol ), 3-

-1 methoxyphenol (∆Hacid = 348.0 ± 2.1 kcal mol ) and 4-tert-butylphenol (∆Hacid = 348.5 ±

-1 -1 2.1 kcal mol ) give a comparable ∆Hacid of 347 ± 3 kcal mol . Table 5.5. Summary of results for acidity bracketing of triphenylsilanol (3).

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

2,4-pentanedione 343.8 ± 2.1 + -

formic acid 345.3 ± 2.3 + -

acetic acid 347.4 ± 0.5 + -

3-methoxyphenol 348.0 ± 2.1 + +

m-cresol 349.6 ± 2.1 - +

p-cresol 350.3 ± 2.1 - +

4-(trifluoromethyl)aniline 353.3 ± 2.1 - +

1-propanethiol 354.2 ± 2.2 - +

pyrrole 359.5 ± 0.5 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

5.2.2.2 Silanediol

The acidities of silanediols have not previously been measured. The silanediols provide a unique opportunity to access a stable geminal-diol motif, and limited studies have been performed with silanediols. Di-tert-butylsilanediol (4): Deprotonated di-tert-butylsilanediol can deprotonate p- cresol, but not 1-pentanethiol; 1-pentanethiolate deprotonates neutral di-tert-

92 butylsilanediol, but p-cresolate does not (Table 5.6). The bracketed acidity of di-tert- butylsilanediol is thus 352 ± 3 kcal mol-1. The Cooks kinetic method with reference acids

-1 -1 m-cresol (∆Hacid = 349.6 ± 2.1kcal mol ), p-cresol (∆Hacid = 350.3 ± 2.1 kcal mol ), 4-

-1 hydroxyphenol (∆Hacid = 350.4 ± 2.1 kcal mol ), 3-aminophenol (∆Hacid = 350.6 ± 2.1

-1 -1 kcal mol ) and 4-aminophenol (∆Hacid = 352.5 ± 2.1 kcal mol ), yields the same ∆Hacid of 352 ± 3 kcal mol-1. Table 5.6. Summary of results for acidity bracketing of di-tert-butylsilanediol (4).

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

2,4-pentanedione 343.8 ± 2.1 + -

acetic acid 347.4 ± 0.5 + -

1,1,1-trifluoroacetone 349.2 ± 2.1 + -

m-cresol 349.6 ± 2.1 + -

p-cresol 350.3 ± 2.1 + -

1-pentanethiol 352.5 ± 2.3 - +

1-propanethiol 354.2 ± 2.2 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

Diphenylsilanediol (5): While deprotonated diphenylsilanediol can deprotonate

-1 neutral acetic acid (∆Hacid = 347.4 ± 0.5 kcal mol ) and more acidic compounds, it cannot

-1 deprotonate neutral m-cresol (∆Hacid = 349.6 ± 2.1 kcal mol ) or less acidic compounds (Table 5.7). In the opposite direction, m-cresolate deprotonates neutral diphenylsilanediol as do more basic bases. We therefore bracket the acidity of diphenylsilanediol as ∆Hacid = 349 ± 2 kcal mol-1. By the Cooks kinetic method, we obtain an acidity of

-1 diphenylsilanediol of ∆Hacid = 347 ± 3 kcal mol , using reference acids pivalic acid

-1 -1 (∆Hacid = 344.6 ± 2.1 kcal mol ), isovaleric acid (∆Hacid = 345.5 ± 2.1 kcal mol ), valeric

-1 -1 acid (∆Hacid = 346.2 ± 2.1 kcal mol ), butyric acid (∆Hacid = 346.8 ± 2.0 kcal mol ) and

-1 m-cresol (∆Hacid = 349.6 ± 2.1 kcal mol ). Table 5.7. Summary of results for acidity bracketing of diphenylsilanediol (5).

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

trifluoro-m-cresol 339.3 ± 2.1 + -

2,4-pentanedione 343.8 ± 2.1 + -

formic acid 345.3 ± 2.3 + -

butyric acid 346.8 ± 2.0 + -

acetic acid 347.4 ± 0.5 + -

m-cresol 349.6 ± 2.1 - +

p-cresol 350.3 ± 2.1 - +

4-(trifluoromethyl)aniline 353.3 ± 2.1 - +

chloroacetonitrile 357.7 ± 2.2 - +

pyrrole 359.5 ± 0.5 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

(4-fluorophenyl)(mesityl)silanediol (6): Proton transfer occurs in both "directions" with acetic acid, which implies that the acidity of (4- fluorophenyl)(mesityl)silanediol is close to the acidity of acetic acid (∆Hacid = 347.4 ± 0.5 kcal mol-1; Table 5.8). Cooks kinetic studies were difficult due to the inability to generate enough protonated dimer signal with reference acids. Table 5.8. Summary of results for acidity bracketing of (4- fluorophenyl)(mesityl)silanediol (6).

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

2,4-pentanedione 343.8 ± 2.1 + -

formic acid 346.0 ± 0.5 + -

acedic acid 347.4 ± 0.5 + +

m-cresol 349.6 ± 2.1 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer. (2,6-bis(trifluoromethyl)phenyl)(mesityl)silanediol (7): We were unable to vaporize 7 for FTMS bracketing experiments. The Cooks kinetic method is prohibited by the high molecular mass of the protonated dimers; dissociation of those protonated dimers yields peaks at too low m/z values to observe (due to instrumental software limitations).

5.2.2.3 Disiloxanediol

1,1,3,3-tetraisopropyldisiloxane-1,3-diol (8): Deprotonated 1,1,3,3- tetraisopropyldisiloxane-1,3-diol can deprotonate neutral methyl cyanoacetate(∆Hacid =

-1 -1 340.8 ± 0.6 kcal mol ) , but not neutral ethoxyacetic acid (∆Hacid = 342.0 ± 2.2 kcal mol , Table 5. 9). In the opposite direction, 1,1,3,3-tetraisopropyldisiloxane-1,3-diol reacts with ethoxyacetate, but there is no proton transfer observed in the reaction of 1,1,3,3- tetraisopropyldisiloxane-1,3-diol and deprotonated methyl cyanoacetate. Thus, we bracket the acidity to be 342 ± 2 kcal mol-1. Measurement by the Cooks kinetic method

-1 gives the same acidity, using L-phenylalanine (∆Hacid = 336.5 ± 3.1 kcal mol ), 3-

-1 hydroxybenzoic acid (∆Hacid = 338.6 ± 2.1 kcal mol ), trifluoro-m-cresol (∆Hacid = 339.3

-1 -1 ± 2.1 kcal mol ) and 2-chlorophenol (∆Hacid = 343.4 ± 2.3 kcal mol ).

Table 5.9. Summary of results for acidity bracketing of 1,1,3,3-tetraisopropyldisiloxane- 1,3-diol (8).

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

α,α,α-trifluoro-m-cresol 339.3 ± 2.1 + -

methyl cyanoacetate 340.8 ± 0.6 + -

ethoxyacetic acid 342.0 ± 2.2 - +

2,4-pentanedione 343.8 ± 2.1 - +

formic acid 346.0 ± 0.5 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

1,1,3,3-tetraphenyldisiloxane-1,3-diol (9): Attempts to bracket the acidity of 9 were hindered by its proclivity to fragment under mass spectrometry conditions. For example, with α,α,α-trifluoro-m-cresol, ions at m/z 377 and 335 are observed; possible structures 9a and 9b are shown below (where HA is the cresol).

Ph OH Ph O Ph Ph Si HA Ph Si Si O O O

m/z = 377 m/z = 335 9a 9b

We therefore measured the acidity using the Cooks kinetic method. Using

-1 reference acids perfluoro-tert-butanol (∆Hacid = 331.6 ± 2.2 kcal mol ), 3-

-1 (trifluoromethyl) benzoic acid (∆Hacid = 332.2 ± 2.1 kcal mol ), 4-acetylbenzoic acid

-1 -1 (∆Hacid = 334.3 ± 2.1 kcal mol ), 3, 5-dichlorophenyl (∆Hacid = 334.4 ± 2.1 kcal mol )

-1 and L-phenylalanine (∆Hacid = 336.5 ± 3.1 kcal mol ), the acidity of 1,1,3,3-

-1 tetraphenyldisiloxane-1,3-diol is measured to be ∆Hacid = 334 ± 3 kcal mol .

5.2.2.4 Silanol carbon analogs

Two carbon analogs (2’ and 3’) of monosilanols 2 and 3 are known and the acidity can be measured for comparison. 2-methyl-2-phenyl-ethanol (2').: 2-Methyl-2-phenyl-ethanol is the carbon analog of dimethylphenylsilanol. 2-Methyl-2-phenyl-ethoxide deprotonates aniline (∆Hacid =

-1 -1 366.4 ± 2.1 kcal mol ), but not acetone (∆Hacid = 368.8 ± 2.0 kcal mol ); also deprotonated acetone deprotonates neutral 2-methyl-2-phenyl-ethanol, but the anilide does not (Table 5. 10). We therefore bracket the acidity of 2-methyl-2-phenyl-ethanol to be 368 ± 3 kcal mol-1. Table 5.10. Summary of results for acidity bracketing of 2-methyl-phenyl-ethanol (2').

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

pyrrole 359.5 ± 0.5 + –

2,2,2-trifluoroethanol 361.7 ± 2.5 + –

4-fluoroaniline 364.3 ± 2.1 + –

aniline 366.4 ± 2.1 + –

acetone 368.8 ± 2.0 – +

3-ethyl-3-pentanol 370.9 ± 2.8 – +

aAcidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “–” indicates the absence of proton transfer. Triphenylmethanol (3'). Triphenylmethanol is the carbon analog of triphenylsilanol. Triphenylmethoxide deprotonates 3-methylpyrazole (∆Hacid = 356.0 ±

-1 -1 2.1 kcal mol ), but not 3-(trifluoromethyl)aniline (∆Hacid = 356.9 ± 2.1 kcal mol ; Table 5. 11). In the opposite direction, 3-(trifluoromethyl)anilide deprotonates triphenylmethanol, but 3-methylpyrazolide does not. Therefore we bracket the acidity of triphenylmethanol to be 356 ± 3 kcal mol-1.

Table 5.11. Summary of results for acidity bracketing of triphenylmethanol (3’). a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

1-pentanethiol 352.5 ± 2.3 + -

1-propanethiol 354.2 ± 2.2 + -

3-methylpyrazole 356.0 ± 2.1 + -

3-(trifluoromethyl)aniline 356.9 ± 2.1 - +

pyrrole 359.5 ± 0.5 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

5.2.2.5. Commercial hydrogen-bonding organocatalysts

BINOL (10): Reaction occurs in both directions for 1,1,1,-trifluoro-2,4-

-1 pentanedione (∆Hacid = 328.3 ± 2.9 kcal mol ) and difluoroacetic acid (∆Hacid = 331.0 ± 2.2 kcal mol-1, Table 5. 12). Deprotonated BINOL does not deprotonate 3,5-bis-

-1 (trifluoro-2,4-pentanedione) (∆Hacid = 329.8 ± 2.1 kcal mol ), but the reaction in the opposite direction does proceed. We therefore can bracket only an acidity range, between the pentadione and difluoroacetic acid (328-331 kcal mol-1). The Cooks kinetic

-1 method yields a ∆Hacid of 330 ± 3 kcal mol (using reference acids 2-hydroxybenzoic

-1 - acid (∆Hacid = 325.5 ± 2.2 kcal mol ), penta-fluorophenol (∆Hacid = 328.0 ± 2.2 kcal mol

1 -1 ), dichloroacetic acid (∆Hacid = 328.4 ± 2.1 kcal mol ), 3,5-bis-

-1 trifluoromethylphenol(∆Hacid = 329.8 ± 2.1 kcal mol ) and L-asparagine(∆Hacid = 331.7 ± 3.1 kcal mol-1)) Table 5.12. Summary of results for acidity bracketing of BINOL (10).

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

1,1,1,-trifluoro-2,4-pentanedione 328.3 ± 2.9 + +

3,5-bis-(trifluoromethyl)-phenol 329.8 ± 2.1 - +

difluoroacetic acid 331.0 ± 2.2 + +

perfluoro-tert-butanol 331.6 ± 2.2 - +

pyruvic acid 333.5 ± 2.9 - +

2-chloropropanoic acid 337.0 ± 2.1 - +

trifluoro-m-cresol 339.3 ± 2.1 - +

aAcidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

TADDOL (11): The conjugate base of TADDOL deprotonates perfluoro-tert-

-1 butanol (∆Hacid = 331.6 ± 2.2 kcal mol ), but not pyruvic acid (∆Hacid = 333.5 ± 2.9 kcal mol-1); also, pyruvate deprotonates TADDOL but perfluoro-tert-butoxide does not (Table 5. 13). We therefore bracket the acidity of TADDOL to be 333 ± 4 kcal mol-1. The same acidity is obtained using the Cooks kinetic method (reference acids: 3, 4, 5-

-1 trichlorophenol (∆Hacid = 330.8 ± 2.2 kcal mol ), perfluoro-tert-butanol (∆Hacid = 331.6 ±

-1 -1 2.2 kcal mol ), 3-trifluoromethyl-benzoicacid (∆Hacid = 332.2 ± 2.1 kcal mol ), 4-

-1 hydroxybenzophenone (∆Hacid = 332.9 ± 2.1 kcal mol ) and iodoacetic acid (∆Hacid = 334.7 ± 2.2 kcal mol-1)). Table 5.13. Summary of results for acidity bracketing of TADDOL (11).

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

difluoroacetic acid 331.0 ± 2.2 + -

perfluoro-tert-butanol 331.6 ± 2.2 + -

pyruvic acid 333.5 ± 2.9 - +

malononitrile 335.8 ± 2.1 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer.

N,N-diphenylthiourea (12): Proton transfer occurs in both directions with

-1 perfluoro-tert-butanol (∆Hacid = 331.6 ± 2.2 kcal mol , Table 14), yielding a bracketed

-1 -1 ∆Hacid of 332 ± 3 kcal mol . Using 3,4,5-trichlorophenol (∆Hacid = 330.8 ± 2.2 kcal mol ),

-1 perfluoro-tert-butanol (∆Hacid = 331.6 ± 2.2 kcal mol ), 3-trifluoromethyl-benzoic acid

-1 (∆Hacid = 332.2 ± 2.1 kcal mol ), 4-hydroxybenzophenone (∆Hacid = 332.9 ± 2.1 kcal

-1 -1 mol ) and iodoacetic acid (∆Hacid = 334.7 ± 2.2 kcal mol ) as reference acids in the Cooks kinetic method, we obtain an acidity of 333 ± 3 kcal mol-1. Table 5.14. Summary of results for acidity bracketing of N,N-diphenylthiourea (12).

a b Reference compound ΔHacid Proton transfer

(kcal mol-1) Ref. acid Conj. base

3,5-bis-(trifluoromethyl)-phenol 329.8 ± 2.1 + -

difluoroacetic acid 331.0 ± 2.2 + -

perfluoro-tert-butanol 331.6 ± 2.2 + +

pyruvic acid 333.5 ± 2.9 - +

malononitrile 335.8 ± 2.1 - +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer. 2-[[3, 5-Bis(trifluoromethyl)phenyl]thioureido]-N-benzyl-N, 3, 3- trimethylbutanamide (13a): Bracketing experiments were prohibited by the paucity of volatile reference acids. The Cooks kinetic method was used, using heptafluorobutyric

-1 acid (∆Hacid = 321.9 ± 2.2 kcal mol ), perfluorobenzoic acid (∆Hacid = 323.6 ± 2.1 kcal

-1 -1 mol ) and 3,5-bis(trifluoromethyl)benzoic acid (∆Hacid = 324.4 ± 2.1 kcal mol ). The acidity is measured to be 322 ± 3 kcal mol-1.

100

5.3 Discussion

All the experimental data obtained in this study are summarized, along with corresponding computational values, in Figure 5. 2. From these results, it can be seen that for the experimental data (in red), the bracketed and Cooks kinetic method acidities are generally similar, if not the same, for the silanol species and their carbon analogs. Also, for the calculations, the higher computational level (B3LYP/6-311++G(2df,p); data in blue (b)) generally yields an acidity for these compounds that is closer to the experimental value than that at the lower level (B3LYP/6-31+G*). This large number of experimental measurements therefore establish that calculations of silanol (and carbon analog) acidity at B3LYP/6-311++G(2df,p) is reasonably accurate. A summary of the calculations conducted at this level is shown in Table 5. 15.

101

Monosilanols Carbon analogs of monosilanols Et OH Ph OH Ph OH Si Si Si Ph OH Ph OH Experimental data Et Et CH3 CH3 Ph Ph C C a = bracketing CH3 CH3 Ph Ph b = Cooks kinetic method 1 2 3 2' 3' Computational data 362 ± 3 (a) 356 ± 3 (a) 348 ± 3 (a) 368 ± 3 (a) 356 ± 3 (a) a = B3LYP/6-31+G(d) 354 ± 3 (b) 347 ± 3 (b) b = B3LYP/6-311++G(2df,p)

355.3 (a) 351.9 (a) 344.5 (a) 361.8 (a) 351.3 (a) 358.5 (b) 355.3 (b) 347.6 (b) 365.1 (b) 354.0 (b)

Silanediols Disiloxanediols Mes OH Mes F C OH t-Bu OH Ph OH Si 3 i-Pr i-Pr Ph Ph Si O Si Si OH OH O t-Bu OH Ph OH i-Pr Si Si i-Pr Ph Si Si Ph CF 3 HO OH HO OH F 4 5 6 7 8 9

352 ± 3 (a) 349 ± 2 (a) 347 ± 2 (a) N/A 342 ± 2 (a) 352 ± 3 (b) 347 ± 3 (b) N/A 342 ± 3 (b) 334 ± 3 (b)

348.4 (a) 344.2 (a) 340.3 (a) 339.0 (a) 336.7 (a) 329.4 (a) 352.2 (b) 348.4 (b) 345.0 (b) 342.8 (b) 338.8 (b) 333.1 (b)

Hydrogen-bonding organocatalysts

Ph Ph CF3

O OH S CH3 R S OH CH3 Ph N OH N N CH3 O OH N N CF3 Ph Ph H H O H H

10 11 12 13a: R = t-Bu 13b: R = CH3 328~331 ± 3 (a) 333 ± 4 (a) 332 ± 3 (a) 330 ± 3 (b) 333 ± 3 (b) 333 ± 3 (b) 322 ± 3 (b) [R = t-Bu]

321.4 (a) 328.5 (a) 328.2 (a) 322.2 (a) 323.1 (b) 330.4 (b) 330.5 (b) 325.2 (b) [R=Me]

-1 Figure 5.2. Summary of data for substrates examined in this study (∆Hacid, kcal mol )

102

Table 5.15. B3LYP/6-311++G(2df,p) calculated ∆Hacid values for silanols X and carbon analogs X'.

-1 Compound ∆Hacid, kcal mol

B3LYP/6-311++G(2df, p)

silanol carbon ∆Hacid(X) –

a analog ∆Hacid(X')

triethylsilanol (1; carbon analog is 1') 358.5 370.3 –11.8

dimethylphenylsilanol (2) 355.3 365.1 -9.8

triphenylsilanol (3) 347.6 354.0 -6.4

di-tert-butylsilanediol (4) 352.2 351.9 +0.3

diphenylsilanediol (5) 348.4 350.8 -2.4

(4-fluorophenyl)-(mesityl)silanediol (6) 345.0 343.1 +1.9

(2,6-bis(trifluoromethyl)-phenyl)(mesityl)silanediol (7) 342.8 338.5 +4.3

1,1,3,3-tetraisopropyldisiloxane-1,3-diol (8) 338.8 342.3 -3.5

1,1,3,3-tetraphenyldisiloxane-1,3-diol (9) 333.1 334.6 -1.5

aRef.38

Overall, these results demonstrate the tunable acidities of organic silanols. The monosilanols vary in acidity from 348 to 359 kcal mol-1 (at B3LYP/6-311++G(2df,p)); silanediols, from 343 to 352 kcal mol-1, and disiloxanediols from 333-339 kcal mol-1.

5.3.1 Comparison of silicon versus carbon analogs

5.3.1.1 Acidity . In terms of monosilanols, Damrauer and coworkers examined a series of simple monosilanols in the gas phase and found that for the species studied, the silanols had considerable enhanced acidity relative to the corresponding alcohols (carbon analogs).20 This effect is also seen in solution, and is attributed to the lower electronegativity (more

103 positive character) of silicon versus carbon, which provides greater stabilization of the anionic oxygen.2,20,222,223 Furthermore, Damrauer found that while alkyl groups increase alcohol acidity in the gas phase, such groups had the opposite effect on silanol acidity. The explanation made for this focuses on the opposing effects of polarizability and induction. In the gas phase, for alcohols, the former prevails.224-227 Therefore, tert- butanol is more acidic than methanol because of the increased polarizability of the tert- butyl groups. Polarizability effects have an r-4 distance dependence; induction effects vary as r-2.6,7,20,224 Since bonds involving silicon are longer than those involving carbon, induction should play a larger role in silanols than it does in alcohols. Since alkyl groups have an acid-weakening inductive effect, silanols with more alkyl substitution become progressively less acidic. For our monosilanols (1-3, Table 5. 15), the alcohol analogs (1'-3') are all, as expected, less acidic. The ∆Hacid difference decreases as one moves from 1 to 3. This too would be expected: phenyl groups are polarizable but weakly inductive.20,6,228 Therefore, such groups would be expected to increase acidity more for alcohols than for silanols, and the difference in acidity between the two is lessened. The silanediols have not heretofore been studied, and are intriguing because there are no stable carbon analogs for these geminal-diol silicon species in solution. In solution, carbon analogs of the silanediols and disiloxanediols favor dehydration to form the corresponding carbonyl compound; this is one of the reasons that silanediols are of great interest, since they are often stable where the carbon analog is not. In the comparison of silanediols with the theoretical carbon analogs, a different trend is observed than for the monosilanols. The acidities of the silanediols (4-9) and carbon analogs (4'-9') are quite close, and in some cases (4, 6 and 7), unlike with the monosilanols, the carbon analog is actually more acidic than the corresponding silanol (Table 5. 15). For silanediols 4, 5, 6, and 7, this effect is probably due to two factors. First, the high polarizability of the tert-butyl and phenyl groups would, as previously

104 discussed, enhance the acidity of the carbon analog more than the silicon. Second, deprotonation of the diol results in an anionic oxide, which can potentially hydrogen bond with the remaining "OH" moiety. For these species, the hydrogen bond is significantly longer for the deprotonated silanediol than for the corresponding deprotonated carbon analog (Figure 5. 3, shown for 4 and 4'). The deprotonated product is therefore quite stabilized for the carbon analog, which greatly increases the acidity of the corresponding neutral. The greater enhanced acidity of the carbon analogs brings the difference in acidity between the silanediols and carbon analogs much closer; for 4, 6 and 7, to the point where the carbon analog is actually more acidic. Substrates 8 and 9 are slightly different, in that the internal hydrogen bond in the deprotonated species is more comparable for the Si and C compounds (Figure 5. 3, shown for deprotonated 9 and 9'). Therefore, both the deprotonated silicon and carbon analogs should get a similar benefit from the internal hydrogen bond. The influence of the four highly polarizable groups (isopropyl for 8 and phenyl for 9) must be again enhancing the acidity of the carbon analog more than that of the silicon analog (due to the distance dependence discussed earlier), so that the difference in acidity between the two is attenuated.

105

Figure 5.3. Deprotonated diol structures, calculated at B3LYP/6-311++G(2df, p)

5.3.1.2 Single versus double-point hydrogen-bonding A diol structure could, hypothetically, participate in either single-point or double- point hydrogen-bonding activation of an electrophile.229 N,N-Diphenylthiourea (12) and BINOL (10) have been proposed to act as double-point hydrogen-bonding catalysts, while TADDOL (11) organocatalysts have cooperative hydrogen bonding (i.e. an intramolecular hydrogen bond) that leads to single-point activation for carbonyl electrophiles (Figure 5. 4).230-232 The calculated structures for the aryl-containing silanediol and disiloxanediol are shown in Figure 4 for comparison (note that Figure 3, which pertains to the acidity discussion, shows anions; Figure 4, which pertains to the intermolecular hydrogen bonding of diols with electrophiles, shows neutral diols). With

106 intramolecular H-O distances greater than 3.0 Å, the diphenylsilanediol 5 is more likely to provide double-point hydrogen-bonding activation, in analogy to BINOL. The disiloxanediol 9 has a fairly weak intramolecular hydrogen bond (2.5 Å), which may possibly lead to single-point hydrogen bonding activation of electrophiles, like TADDOL.39 Most of the silicon structures we examined do not have short (less than 3.0 Å) hydrogen bonds, with the exception of 8 and 9 (the two disiloxanediols) and silanediol 7. Therefore, these three structures might be more prone to activating electrophiles via single-point hydrogen bonding.

107

Figure 5.4. Organocatalyst and organosilicon neutral diol structures, calculated at B3LYP/6-311++G(2df,p)

5.3.2 Gas phase versus solution

These particular silanediols and disiloxanediols discussed herein are under development as hosts and catalysts based on their acidity and hydrogen-bonding abilities.233-235 Many reactions that are catalyzed by non-covalent hydrogen-bonding organocatalysts are conducted in nonpolar solvents to enhance molecular recognition and substrate activation.229 Since the gas-phase is the “ultimate” nonpolar medium, gas-phase acidity values can help assess the strength of hydrogen bonding in these reactions, especially for cases where experimental pKa values and binding affinities may not be available. To date, catalytic studies (of carbonyl activation by 2, 3, 6, 7 and 9 in a Diels Alder reaction of methacrolein and Rawal's diene) were conducted in a nonpolar solvent (toluene), so the gas phase studies herein could be relevant.233,236 The observed catalysis, within a class (e.g. monosilanol, silanediol, and disiloxanediol), does track with acidity (Table 5. 16): the more acidic monosilanol 3 is a better catalyst than 2. The more acidic silanediol 7 is, likewise, a better catalyst than 6. The disiloxanediol 9 is the most acidic silicon substrate (somewhat comparable to commercial catalysts in gas phase acidity); catalytic activation by 9 was found to be quite high (55%, Table 5. 16). The somewhat high yields for the monosilanols 2 and 3 (relative to their acidities) have been attributed (in a recent study of ours) to these species having less of a propensity to self-associate than the silanediols 6 and 7, because monosilanols have one less hydroxyl and more steric bulk around that hydroxyl.233 Likewise, the yield of disiloxanediol 9 is lower than that of monosilanol 3 despite the higher acidity of 9 because 9 has two hydroxy groups that will be prone to strong self-association.

108

Table 5.16. Comparison of catalytic activation of carbonyl compounds in a Diels Alder reaction using various silanols and alcohols.

-1 a catalyst ∆Hacid, kcal mol , B3LYP/6- yield at -72˚C (%)

311++G(2df,p)

2 355.3 47

2’ 365.1 5

3 347.6 53

3’ 354.0 17

6 345.0 40

7 342.8 55

9 333.1 55

10 323.1 63

11 330.4 30b

aAt -65 ˚C, higher yields are observed for all reactions, but a potential increase in the background rate makes the comparison at -72 ˚C more representative of the intrinsic activating abilities. bYield based on literature values taken from reference 237 and this work.

The carbon analogs 2' and 3' are less acidic than the corresponding silanols, and do in fact show less catalysis, as would be expected. The organocatalysts 10 and 11 are very acidic; 10 is more acidic than 11 and has an accordingly higher yield. The gas phase acidities therefore can be correlated to activity, though the correlation is certainly not quantitative. However, examination of gas phase properties is valuable, since when activity differs from acidity, the provenance must be solvation effects, whether self-association or other influences.233 In polar solvents, catalytic activity may not track with the gas phase acidity. In general, solvation tends to decrease the difference in the acidities observed in the gas

109 phase. In polar solution, induction may prevail even more (this is why alcohol acidity decreases with increasing alkyl substitution in solution, but the opposite is true in the gas phase).20,224-227 Therefore, in terms of catalyst design, catalysts with aryl groups are likely to be more versatile in a variety of solvents than those with alkyl groups, since the latter have acid weakening inductive effects that are stronger in polar solvents than in nonpolar media.40,238 Future studies will focus on designing molecules that will allow us to test our hypotheses concerning single versus double point hydrogen bonding, self-association, and other possible solvent effects on silanol properties.

5.4 Conclusions

Much experimental work has been undertaken to characterize a series of novel silanols. Experimental gas-phase acidity values for a variety of dual hydrogen-bonding silanols and organocatalysts have been measured, and the computational methods have been optimized to B3LYP/6-311++G(2df,p). With established theory and basis-set, we have found that the acidities of organic silanols are fairly tunable: monosilanol (348 – 359 kcal mol-1), silanediol (343 - 352 kcal mol-1), and disiloxanediol (333 – 339 kcal mol- 1). Although it is generally said that silanols are more acidic than their carbon counterparts, we have found that the diol analogs show a reversal of this trend, depending on substitution and structure. Polarizability and induction have opposing effects on acidity, which differ in importance when the medium is nonpolar versus polar. Our studies have some interesting implications for these compounds as participants in molecular recognition, as transition state analogs, and in catalysis.215-218 Preliminary studies show that carbonyl activation in a Diels Alder reaction generally correlates with the acidities of

110 the various silanol and carbon analogs studied herein. The correlation is not quantitative due to effects that may dominate in solution, such as self-association.

5.5 Experimental.

Silanols 3, 4, 6, 7, 8, 9 were synthesized as previously described.233 All other silanols (1, 2, 5), hydrogen-bonding catalysts (10, 11, 12, 13) and reference acids are commercially available and were used as received. Both bracketing and Cooks kinetics methods have been described in the first chapter.

111

Note: Major parts of the following chapter have been published: Liu, M.; Yang, I.;

Buckley, B.; Lee, J. K. Org. Lett. 2010, 21, 4764–4767; Liu, M.; Chen, M.; Zhang, S.

Yang, I.; Buckley, B.; Lee, J. K. J. Phys. Org. Chem. ASAP

Chapter 6 Proton Affinity of Phosphines versus N-heterocyclic

Carbenes and Reactivity of Carbene•Phosphine Dimers

6.1 Introduction

N-Heterocyclic singlet carbenes (NHCs, such as imidazol-2-ylidenes, 1 and the saturated derivatives, 2, Figure 6. 1) are "stabilized" carbenes that were first isolated by Arduengo and coworkers in 1991.21,28 Such species are of fundamental organic interest but have also proven to be of great utility in many synthetic applications, including as versatile catalysts for a variety of organic reactions and as novel ligands for transition- metal-catalyzed reactions such as the Grubbs ruthenium olefin metathesis catalysis, palladium-catalyzed cross-coupling reactions and nickel-catalyzed cycloadditions.29,30,32,34,239-241 The dialkylimidazolium salts (which can form stabilized carbenes when deprotonated) are also an important class of room temperature ionic liquids, environmentally "clean" nonvolatile solvents that are increasingly used in organic synthesis.35-38 In addition to organic synthetic utility, these carbenes also have biological counterparts, the most well-known being thiamine pyrophosphate (the coenzyme form of vitamin B1).40-43

112

Figure 6.1. N-Heterocyclic singlet carbenes

Despite the widespread use of these carbenes as novel ligands and of the dialkylimidazolium salts as ionic liquids, the experimental acidity of the imidazolium ions (the proton affinity of the carbenes) is surprisingly little-studied, both in solution and in the gas phase.242,243 Characterizing the fundamental properties is therefore warranted. Because NHCs are more effective ligands than the classic first-generation tricyclohexylphosphine (PCy3) species in the ruthenium catalysts used for Grubbs' olefin metathesis, one interest is to compare the proton affinities of carbenes with phosphines.244

The pKa of a series of imidazolium cations at C2 (pKas ranging from 21-24; deprotonation yields the corresponding singlet N-heterocyclic carbenes) was measured in an elegant deuterium exchange reaction in water by Amyes, Diver, Richard and coworkers.242,245 The gas phase PA of 1-ethyl-3-methylimidazol-2-ylidene 1b was previously measured by Cooks and coworkers to be 251.3 ± 4 kcal mol-1.243 Cooks et al. also ascertained that the 1-ethyl-3-methyl-substituted substrate has a lower PA than the more highly substituted compound 1,3-di-tert-butylimizadol-2-ylidene, which in turn has a lower PA than 1,3-di-(2,6-isopropylphenyl)imizadol-2-ylidene. To our knowledge, there are no other measurements of N-heterocyclic diamino carbene proton affinity, in any medium.246

6.2 Results and Discussion

6.2.1 Proton Affinity

We first repeated the experimental measurement of the gas phase proton affinity of the 1-ethyl-3-methyl carbene (1b). Using the Cooks kinetic method107 and 1,5- diazabicyclo[4.3.0]non-5-ene (DBN, PA = 248.16 kcal mol-1) and 1,8-

113 diazabicyclo[5.4.0]undec-7-ene (DBU, PA = 250.45 kcal mol-1) as reference bases, we obtain a PA of 250 ± 3 kcal mol-1, which is consistent with the previous Cooks group measurement (251.3 ± 4 kcal mol-1).44 We also explored the PA of 1-ethyl-3-methyl carbene (1b) with theory to probe the accuracy of various methods and levels. For the carbene, we tried RHF/6-31+G(d), B3LYP/6-31+G(d), M06-2X/6-31+G(d), MP2/6-311+G(2d,p)//B3LYP/6-31+G(d), M06- 2X/aug-cc-pVTZ//B3LYP/6-31+G(d) and CBS-QB3.27-36 As Table 6. 1 shows, all these methods appear to overestimate the PA of the carbene, with M06-2X/6-31+G* coming the closest to experimental, although at 258 kcal mol-1, the calculated value is still some 7-8 kcal mol-1 higher than the measured value of 250-251 kcal mol-1. Table 6.1. Proton Affinity Calculations on 1-Ethyl-3-methylimidazol-2-ylidene 1b

Method/level PA

(kcal mol-1) RHF/6-31+G* 262.4

B3LYP/6-31+G* 261.4a

M06-2X/6-31+G* 257.8

MP2/6-311+G(2d,p)//B3LYP/6-31+G(d) 270.8

M06-2X/aug-cc-pVTZ//B3LYP/6-31+G* 268.3

CBS-QB3 264.5

aref44,107,247

We next examined the PA of PCy3, which has not heretofore been measured. Using

DBU and DBN as reference bases and using the Cooks kinetic method, we obtain a PCy3

-1 PA of 249 ± 3 kcal mol . Also we bracketed the PA of PCy3 on FTMS (the results are summarized in Table 6. 2). Protonated tricyclohexylphosphine undergoes proton transfer with neutral N,N,N',N'-tetramethyl-1,4-butane-diamine (PA = 250.1 kcal mol-1); the reverse reaction does not occur. Protonated PCy3 does not react with N,N,N',N'-

114 tetramethyl-1,3-propane diamine (PA = 247.4 kcal mol-1); the opposite reaction does.

-1 248 We therefore bracket the PA of PCy3 to be 249 ± 3 kcal mol .

Table 6.2. Summary of results for acidity bracketing of PCy3

Reference compound PAa Proton transferb

N,N,N',N'- tetramethyl- (kcal mol-1) Ref. base Conj. acid

1,4-butane-diamine 250.1 ± 2.0 + –

1,3-propane-diamine 247.4 ± 2.0 – +

1,8-naphthalene-diamine 245.8 ± 2.0 – +

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer. Given that the calculations of the PA of 1b are not consistent with experiment, we sought to compile a full picture of calculations versus experiments for phosphines.

Toward this end we bracketed the PA of HPCy2 as well (Table 6. 3). Piperidine (PA =

-1 228 kcal mol ) cannot deprotonate protonated HPCy2, while HPCy2 can deprotonate protonated piperidine. 2,4-Lutidine (PA = 230.1 kcal mol-1) deprotonates protonated

HPCy2 but the reverse reaction does not occur. We therefore bracket the PA of HPCy2 to be 229 ± 3 kcal mol-1.

Table 6.3. Summary of results for acidity bracketing of HPCy2

Reference compound PAa Proton transferb

N,N,N',N'- tetramethyl- (kcal mol-1) Ref. base Conj. acid

1-methylpiperidine 232.1 ± 2.0 + –

1-methylpyrrolidine 230.8 ± 2.0 + –

2,4-lutidine 230.1 ± 2.0 + –

piperidine 228.0 ± 2.0 – +

4-picoline 226.4 ± 2.0 – +

115

a Acidities are in kcal mol-1.8 bA “+” indicates the occurrence and a “-” indicates the absence of proton transfer. In Table 6. 4, we compile the computational and experimental data on the phosphines and on the ethyl methyl carbene 1b. The B3LYP/6-31+G* calculations on the methyl phosphine system (first four rows of Table 6. 4) are in agreement with experiment; in general the B3LYP/6-31+G* calculations are around 2-3 kcal mol-1 lower in value than the experimentally measured values. For the PAs of H2PCy and HPCy2, the B3LYP/6- 31+G* calculated and measured values are quite close (about 210 kcal mol-1 for the

-1 former and 229 kcal mol for the latter). However, for PCy3, the measured value is some

-1 6 kcal mol higher than the calculated. Because we measured the PCy3 value using two methods (Cooks kinetic method in a linear quadrupole ion trap and PA bracketing experiments in a Fourier Transform ion cyclotron resonance mass spectrometer), we believe the value of 249 ± 3 kcal mol-1 is accurate. Table 6.4. Experimental and computational proton affinity values for various phosphines and 1-ethyl-3-methylimidazol-2-ylidene 1b

Proton affinity (PA), kcal mol-1

substrate experimenta B3LYP/6-31+G* calculations

g PH3 188 185.7

g H2PMe 203.5 201.6

g HPMe2 218.0 215.3

g PMe3 229.2 225.8

h H2PCy 210.3 210.6

c HPCy2 229 228.6

b,c PCy3 249 243.3

1b 250-251b,d,f 261.4

116

a Experimental values have a 2-3 kcal mol-1 error bar. bThis work. Cooks kinetic method experiments. cThis work, bracketing experiments. dReference44 . eReference45. fReference247. gReference8. hReference249.

Increasing errors for DFT/B3LYP-calculated enthalpies of formation with increasing molecule size has been previously observed, with attribution to the neglect of medium-

249-251 range electron correlation effects. We therefore calculated the PCy3 PA using MP2/6-311+G(2d,p)//B3LYP/6-31+G* and find the value to be much closer to

-1 252 experiment (248.6 kcal mol ). The PCy3 system thus appears to be sufficiently electronically crowded to require adequate electron correlation for accurate description.249,251

As we can see from the experimental values, the PAs of PCy3 and 1-ethyl-3-methyl carbene appear to be very similar. Whether the PAs of 1-ethyl-3-methyl carbene and

PCy3 are slightly different is not discernible from these data, as the error on the experimental measurement is around 3 kcal mol-1. In an effort to discriminate the relative

PA of carbene 1b versus PCy3, we conducted Cooks experiments with the protonated dimers comprised of each (i.e. carbene•phosphine protonated dimer 3b, Figure 6. 2).

N a: R = R = CH HPCy3 1 2 3 N b: R1 = Et, R2 = CH3

R2 PCy3 = P 3

Figure 6.2. Proton-bound dimer of PCy3 and NHC

The protonated dimer of 1-ethyl-3-methyl carbene and PCy3 (3b) has a m/z ratio of

+ 391. CID on this dimer produces each protonated product (HPCy3 (m/z 281) and protonated carbene (m/z 111)), the ratio of which can yield the relative proton affinity. This experiment is not easy to conduct in that the protonated dimer yields a fairly weak

117 signal. Subsequent isolation and CID results in increasingly weak daughter ion signals. Furthermore, initial CID of 3b yields some unexpected fragmentation products that could affect the relative proton affinity estimation.253 We conducted this "relative PA" experiment six times, and the results consistently indicate that PCy3 is slightly more basic

-1 than 1-ethyl-3-methyl carbene 1b (by less than 2 kcal mol ). We find that PCy3 is also a little more basic than the dimethyl-substituted carbene 1a, again by less than 2 kcal mol-1. Given the difficulty of the experiment, at this point, it is probably reasonable to conclude that essentially, PCy3 and carbenes 1a and 1b have similar proton affinities. However, the discrepancy between the calculated and experimental PA values for 1- ethyl-3-methyl carbene is somewhat disconcerting and we wonder why the experimental and computational data are so disparate. Three possibilities were considered. One possibility is that the calculations are accurate, but the proton affinity measured by the Cooks kinetic method is not at the desired C2 position, but perhaps at another position on the substrate that we did not realize was less basic. The second possibility is also that the calculations are accurate, but that the Cooks kinetic method measurements of the carbene PA give us an erroneously low PA due to "technical" problems.254-256 The third is that in fact, the experimental results are correct, and the calculations are inaccurate. To address the first possibility, that we are measuring another site on the carbene 1b, we calculated (B3LYP/6-31+G*) the acidities of all the protons on the 1-ethyl-3- methylimidazolium ion (Figure 6. 3, where acidity of ion 1bH+ is equivalent to the PA of 1b). Although we and others assumed that the most acidic position on substrate 1bH+ is the C2-H,26,245 the calculations show that the path to eliminate ethylene is enthalpically preferred over deprotonation at C2-H by more than 6 kcal mol-1 (Figure 6. 3; elimination ∆H = 254.9 kcal mol-1; acidity C2-H = 261.4 kcal mol-1).

118

288.9 CH3 279.7 3 H N H 261.4 279.3 1 H N H2C CH 289.1 2 H 254.9 1bH+

CH3 CH 3 3 N N + + H 1 H H N N

H2C CH2 H 254.9 + 1bH

Figure 6.3. Calculated (B3LYP/6-31+G(d)) acidities of the 1-ethyl-3- methylimidazolium ion (in kcal mol-1).

In the Cooks kinetic method experiment, the proton-bound complex of the carbene and tricyclohexylphosphine is generated (3b) and subjected to collision-induced dissociation (CID). The complex is expected to dissociate to the protonated carbene and/or the protonated phosphine, and the relative signal intensities are related to the relative proton affinities (vide infra). However, when the proton-bound dimer of the ethyl methyl carbene and PCy3 undergoes CID to produce the protonated carbene and the protonated phosphine, we cannot be certain how the latter is produced -- via straightforward deprotonation at C2 or elimination across the C-C ethyl bond (paths B and B', Figure 6. 4). The ethylene elimination is thermodynamically more favorable but presumably has a barrier whereas the proton transfer should be fairly barrierless, so it would be difficult to differentiate the two paths.257,258 The possibility of more than one + reactive path to form HPCy3 complicates the interpretation of the Cooks PA experiment.

119

CH3 A N CH H + PCy3 3 N 3N CID H PCy CH2CH3 1 3 N m/z 111 CH3 CH CH N 2 3 B + HPCy3 3b N m/z 391 m/z 281 CH2CH3

B' CH3 N + HPCy3 H + N m/z 281

Figure 6.4. Possible paths by which proton-bound dimer of the methyl ethyl carbene

1b and PCy3 could form protonated carbene and protonated phosphine

In an effort to discriminate between deprotonation at C2 (path B) versus elimination (path B'), we examined the deuteron-bound analog of 1b with the reference base N,N,N',N'-tetramethyl-1,4-butanediamine (3b-d). CID of 3b-d yields m/z ratios corresponding to the deuterated carbene (m/z 112) and the deuterated reference base (m/z 146), but no peak corresponding to the protonated reference base (m/z 145). These data indicate that paths A and B are followed, but not path B'. Therefore, for this reference base, deprotonation at C2, not elimination, takes place.222

Still, regardless of deprotonation site, the measured PA value by the Cooks kinetic method of roughly 250 kcal mol-1 seems low compared to the calculated values of 254.9 kcal mol-1 for the ethylene elimination pathway and 261.4 kcal mol-1 for the C2 PA.

120

Therefore, we considered a second possibility: a problem with the Cooks experiment. The Cooks kinetic method involves the formation of a proton-bound dimer of the unknown (A) and a reference compound (Bi, Figure 6. 5). Collision-induced dissociation of that dimer results in the proton attaching to either the unknown or the reference. The

+ + relative intensities of AH versus BiH yield the relative proton affinities of A and Bi (more details in Experimental section). In this method, one is obtaining thermochemical data from a kinetic experiment; therefore, one must assume that the dissociation has no reverse activation barrier and that the dissociation transition structure is late and therefore reflects the relative stability of the two protonated products. Another concern, especially for larger systems with polarizable components (as we have with dimer 3b), is that the proton-bound dimer is not of the expected structure. The proton might not be shared equaly, for example, between the carbene and phosphine as assumed (and depicted in

+ + structure 3b). In that case, the ratio of HPCy3 to 1bH would not reflect the relative PAs.

AH +Bi k1

AHBi

k2 BiH +A Figure 6.5. Cooks kinetic method experiment

To address this possibility, we conducted additional experiments. Initially, because of instrumental limitations that allowed us to only use the Cooks kinetic method, we re-measured the PA of 1b, but using different reference bases than our and Cooks' previous experiments, where 1,5-diazabicyclo[4.3.0]non-5-ene (DBN, PA = 248.16 kcal mol-1) and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU, PA = 250.45 kcal mol-1) were used. Using different bases might help preclude the possibility that non-specific dimers of undesired structure are formed.67,108,183,186,204,259 The reference bases N,N,N',N'- tetramethyl-1,8-naphthalenediamine, N,N,N',N'-tetramethyl-1,3-propanediamine, and

121

N,N,N',N'-tetramethyl-1,4-butanediamine were used (Table 6. 5). The ethyl methyl carbene 1b proved to be more basic than the 1,8-naphthalenediamine (PA = 245.8 kcal mol-1) and the 1,3-propanediamine (PA = 247.4 kcal mol-1), such that the Cooks experiment with these bases yields only the protonated carbene. Application of the Cooks kinetic method to carbene 1b and N,N,N',N'-tetramethyl-1,4-butanediamine (PA = 250.1 kcal mol-1), indicates that the carbene is more basic than the reference base by about 2 kcal mol-1, yielding a PA value for 1b of 252 ± 4 kcal mol-1. Given the error bars, this value is consistent with the prior measurements using DBU and DBN as reference bases. Table 6.5. Natural logarithm of product ion abundance ratios R for CID of the proton- bound dimers of 1-ethyl-3-methylimidazol-2-ylidene 1b with various diamines to obtain PAa

Reference base (N,N,N',N'- PAb ln Rc tetramethyl-)

1,4-butanediamine 250.1 –2.73

1,3-propanediamine 247.4 see only protonated carbene

1,8-naphthalenediamine 245.8 see only protonated carbene

a b -1 -1 260 c Teff = 340.3 K. Values are in kcal mol and have a ±2.0 kcal mol error bar. R = (Iref base/I1b) where I

= signal intensity.

Still not confident of the Cooks kinetic method results, we proceeded to modify our quadrupole ion trap so that proton affinity bracketing, which is a nonkinetic method (and therefore suffers from fewer caveats), could be used.261 Mu Chen in our group bracketed the PA of 1-ethyl-3-methyl carbene on this modified LCQ. The protonated carbene is introduced via electrospray, and allowed to react with the neutral reference base. The opposite direction reaction (protonated reference base plus neutral carbene) was not explored as there is no simple way to generate gaseous neutral carbene. The bracketing results are shown in Table 6. 6.

122

Table 6.6. Summary of results for PA bracketing (in quadrupole ion trap) of more basic site of carbene 1b Reference compound PAa Proton transferb Ref. base

2-tert-Butylimino-2-diethylamino-1,3-dimethylperhydro-1,3,2- 263.8 + diazaphosphorinec tert-Octylimino-tris(dimethylamino)phosphoranec 262.0 +

tert-Butylimino-tris(dimethylamino)phosphoranec 260.6 +

7-Methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene 254.0 –

1,8-Diazabicyclo[5.4.0]undec-7-ene 250.5 –

N,N,N',N'-tetramethyl-1,4-butanediamine 250.1 –

aValues are in kcal mol-1 and have a ± 2.0 kcal mol-1 error bar.260 bA “+” indicates the occurrence and a

“–” indicates the absence of proton transfer. cRef. 262

These results indicate that the PA of the ethyl methyl carbene 1b falls between 254.0 and 260.6 kcal mol-1. (There are very few reference bases in this very basic region, and the one "in between" 254.0 and 260.6 (imino-tris(dimethylamino)phosphorane (257.4 kcal mol-1)) was prone to clustering issues with carbene 1b). This range is consistent with both the calculated endothermicity of the ethylene elimination pathway (254.9 kcal mol-1) and the calculated PA at C2 (261.4 kcal mol-1, Figure 6. 3). The prior PA measurement of 250-252 kcal mol-1 for carbene 1b may thus be an artifact of the Cooks kinetic method experiment. The likelihood is that the dimer is not proton-bound at C2

+ and the structure is something less specific that, upon CID, yields more HPCy3 than would be expected based on true relative PAs. To test this hypothesis further, we examined the dimethyl carbene 1a. In our previous work, the Cooks kinetic method study of the proton-bound dimer of 1a with

PCy3 (dimer 3a) also indicated that the proton affinities of 1a and PCy3 are similar, around 250 kcal mol-1. However, the B3LYP/6-31+G* calculated PA value of the C2 position of the dimethyl carbene 1a is 259.9 kcal mol-1 (Figure 6. 5). The dimethyl

123 carbene is also a good control system since it cannot undergo ethylene elimination that the methyl ethyl can so interpretation of the results is more straightforward. 287.4 CH3 278.3 3 H N 1 H 259.9 278.3H N CH3 287.4 + 1aH

Figure 6.5. Calculated (B3LYP/6-31+G(d)) acidities of the dimethylimidazolium ion, in kcal mol-1

Mu Chen in our group bracketed the PA of dimethyl carbene 1a. The results are shown in Table 6. 7. The PA is bracketed to be between 257.4 and 260.6 kcal mol-1, which is consistent with the calculated value of 259.9 kcal mol-1. Table 6.7. Summary of results for PA bracketing (in quadrupole ion trap) of more basic site of carbene 1a

Reference compound PAa Proton transferb

Ref. base

tert-Octylimino-tris(dimethylamino)phosphoranec 262.0 +

tert-Butylimino-tris(dimethylamino)phosphoranec 260.6 +

Imino-tris(dimethylamino)phosphoranec 257.4 –

7-Methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene 254.0 –

1,8-Diazabicyclo[5.4.0]undec-7-ene 250.5 –

aValues are in kcal mol-1 and have a ±2.0 kcal mol-1 error bar.260 bA “+” indicates the occurrence and a

“–” indicates the absence of proton transfer. cRef. 262

124

CH3 CH3 3 H 3N H N 1 H 259.9 1 H 261.4 H N H N CH3 H2C CH2 H 254.9 (elimination) 1aH+ 1bH+ experimental: 257.4-260.6 254.0-260.6 Figure 6.6. B3LYP/6-31+G(d) calculated and experimental (bracketing) acidity values for dialkylimidazolium ions 1aH+ and 1bH+ (values in kcal mol-1)

The new results herein (Figure 6. 6) therefore indicate that prior measurements of the PA of carbenes 1a and 1b may suffer from issues with the Cooks kinetic method.263 These studies also show that the B3LYP/6-31+G* calculations of carbene PA do appear, at least for these systems, to be reasonably accurate.264,265 Furthermore, the carbenes are very basic, even more so than PCy3. What are the implications of the PA values in the role of these carbenes as ligands? The σ-donating ability of various NHC ligands has been studied both experimentally and computationally in order to better understand their interaction with transition metals.266 Overall, NHCs are found to be electron rich. To measure this, Tolman's electronic parameter (TEP) is often used, which correlates to the frequency of the CO stretch in a [Ni(CO)3(L)] complex; the lower the TEP, the stronger the σ-donating ability of the NHC.266,267 Our calculated proton affinity results are in agreement with TEP values for a series of alkylated carbenes (Table 6. 8, for the dimethyl, di-isopropyl and ditertbutyl systems); PA values increase along the series, while TEP values decrease. Both indicate the increasing electron donating ability of the carbene as the substitution becomes more sterically demanding. The better efficacy of the NHCs versus PCy3 may therefore be due to their higher basicity, which may lead to better donation. Table 6.8. Computational (B3LYP/6-31+G*) proton affinities and TEP values for selected carbenes.

125

Substrate PA, kcal mol-1 TEP (cm-1)a

1a (dimethyl carbene) 259.9 2052.4

1b (ethyl methyl carbene) 261.4

1c (isopropyl methyl carbene) 262.8

1d (di-isopropyl carbene) 265.6 2051.5

ditertbutyl carbene 268.9 2050.1

aReference 268

6.2.2 Dimer fragmentation.

The studies on the protonated carbene•phosphine dimers 3a yielded an additional, intriguing result. Upon CID, an ion at m/z 295 (corresponding to a loss of 82 Da) was observed in addition to the protonated carbene and protonated phosphine.269 This m/z 295 ion could correspond to either loss of cyclohexene (Scheme 6. 1, path C, to produce 4a), or to the phosphine picking up a methyl from the carbene (Scheme 6. 1, path D) to produce 5a). The former path is a cyclohexene elimination-type reaction while the latter path results from the phosphorus attacking the methyl on the imidazole ring (substitution). These correspond to the same m/z value because 1-methyl imidazole and cyclohexene have, by coincidence, the same molecular weight. This reactivity is of interest not only from a fundamental point of view but also in the use of these species as ligands with

270 PCy3 in the Grubbs ruthenium catalysts. We were curious as to which of these pathways would be favored; or more generally, what is/are the structure(s) of the m/z 295 ion?

126

Scheme 6. 1

CH3 N A H + PCy3 N m/z 97 CH3 CH3 CH N 3 N B HPCy3 + HPCyH + CH N 2 3 m/z 281 N N CID CH3 HPCy CH3 MW 82 3 CH N 3 C N CID m/z 213 CH3 + H PCy2H C' 3a N CH3 m/z 377 CH MW 82 N 3 4a + HPH3 2 m/z 295 N MW 82 CH3 m/z 131 D N + CH3PCy3 H 5a N m/z 295 CH3

D' CID

CH3PHCy2 + CH3PH2Cy + 2 m/z 213 m/z 131

Subsequent CID of the m/z 295 ion produces ions with m/z 213 and m/z 131. These m/z values correspond to successive losses of cyclohexene and can be produced from either C or D (via C' and D', Scheme 6. 1). We find that at 25% collision energy (chosen to optimize signal), the ratio of the m/z 213 to m/z 131 signal is 10 ± 4. That is, the signal intensity of m/z 213 is ten times stronger than that of m/z 131. We next sought to generate the two possible m/z 295 structures 4a and 5a independently so that each could be subjected to CID. The dimer 4a is straightforward; it can be formed from a solution of the protonated carbene and dicyclohexylphosphine. For the methyl cyclohexylphosphine cation 5a, we prepared a solution of tetramethylammonium chloride and tricyclohexylphosphine; as we hoped, the phosphine

+ acts as a nucleophile and via SN2 becomes methylated to form the CH3PCy3 species (Scheme 6. 2).

127

Scheme 6. 2

CH3 + N PCy3 CH3PCy3 + N(CH3)3 CH3 H3C CH 3 5a (m/z 295)

CID of the independently generated dimer 4a produces essentially the m/z 213 ion only. On the contrary, CID of the methylated tricyclohexylphosphine cation 5a produces a ratio of m/z 213 to m/z 131 of 10 ± 2. These results imply that the "genuine" m/z 295 ion is the cation 5a since the "213/131" ratios are similar ("213/131" is about 10 for both the genuine m/z 295 ion and the independently generated cation 5a). Of course it is also possible that cation 5a is the major product, but some 4a is also formed; this would be difficult to discern with this type of experiment. Also, we cannot be certain that the "295" produced via CID of dimer 3a would behave the same as either 4a or 5a produced independently. But, given these caveats, it does appear that the path producing 5a is certainly the main route.

To explore this further, we also synthesized the di-CD3 carbene derivative, which would allow us to differentiate between the cyclohexene elimination and substitution paths (Scheme 6. 3). CID of the deuterated species 3a-d6 yields the m/z 298 ion, but not the m/z 301 ion, implying that the substitution pathway does predominate.

Scheme 6. 3

128

CD3 cyclohexene N + elimination HPCy2H CD3 N N CID CD MW 82 3 4a-d6 H PCy3 N m/z 301 CD3 substitution N 3a-d6 H CD3PCy3 + m/z 383 N 5a-d 3 CD m/z 298 3 MW 85

In order to better understand our experimental results, we also computed the energetics for the formation of 4a and 5a from 3a. The path to form the alkylated product 5a is much more exothermic (-4.8 kcal mol-1) than the cyclohexene elimination path (4a, +11.2 kcal mol-1), consistent with the experimentally observed product.

6.3 Conclusion

Our studies establish that the dimethylimidazol-2-ylidene (1a) and the 1-ethyl-3- methylimidazol-2-ylidene (1b) are more basic than previous measurements indicate. Also, calculations at B3LYP/6-31+G(d) appear to calculate the carbene proton affinities accurately. The carbenes are nearly 10 kcal mol-1 more basic than tricyclohexylphosphine, which may provide an explanation for the improved efficacy of Ru catalysts containing

NHCs versus PCy3. We have also probed interesting reactivity that the proton bound dimers of 1a and 1b with PCy3 display upon CID. The dimethyl system (3a) undergoes phosphine methylation. Calculations also show that in the presence of a base, the ethyl methyl system (3b) could undergo elimination across the C-C ethyl bond to form ethylene.

129

6.4 Experimental

All chemicals are commercially available and were used as received, except for the di-CD3-imidazolium salt (precursor to 3a-d6), which was synthesized following

271 literature procedure. Methanol is degassed by bubbling N2 gas through for at least 30 min before use. To generate 5a, a 0.01M tetramethylammonium chloride and 0.01M tricyclohexylphosphine in methanol solution was used. Proton affinity bracketing. Proton affinity bracketing measurements were conducted using a Fourier Transform Ion Cyclotron Resonance Mass Spectrometer (FTMS) with a dual cell set up which has been described previously.11-13. We also conducted bracketing experiments in a modified quadrupole ion trap mass spectrometer.204 Protonated NHC ions were generated by electrospray (ESI) from a ~10-4 M methanol solution using a flow rate of 15~25 µL/min. The capillary temperature was 150˚C. Neutral reference bases are added with the helium gas flow. The protonated carbene ions were allowed to react with neutral reference bases for 0.03-1000 ms. The occurrence of proton transfer is regarded as an evidence that the reaction is exothermic (“+” in Tables). The typical electrospray needle voltage was ~4.5 kV. A total of 10 scans was averaged. Cooks kinetics method. Proton affinity measurement of 1b was also conducted in a linear trap quadrupole mass spectrometer using the Cooks kinetic method. 67,108,183,186,204,259 The procedure for conducting these experiments in our lab has been described previously.272,273 Briefly, for PA experiments, this method involves the formation of a proton bound complex, or dimer, of the unknown and a reference base of known proton affinity (eq 6. 1, where "A" is the unknown substrate and "Bi" is a series of reference bases). Collision-induced dissociation (CID) of this dimer leads to the formation of either the protonated unknown or the protonated reference base. The ratio of these two protonated products yields the relative proton affinities of the two compounds

130 of interest, assuming that the dissociation has no reverse activation energy barrier and that the dissociation transition structure is late and therefore indicative of the stability of the two protonated products. Both these assumptions are generally true for proton bound systems.69,70,257

ln(k1 / k 2 ) = [(PA(A) / RTeff ) − Δ(ΔS) / R] − PA(Bi ) /(RTeff ) Eq. 6.2

+ + n(k1 / k 2 ) = ln([AH ]/[Bi H ] Eq. 6.3

app GB (A) / RTeff = PA(A) /(RTeff ) − Δ(ΔS) / R Eq. 6.4

y01 '= [(PA(A) − PAavg ) / RTeff ] − Δ(ΔS) / R Eq. 6.5

Data are generally analyzed using the Cooks "extended" kinetic method.183,184,186,204 This method has been well-described and involves acquiring ion abundance ratios at different collision energies (and therefore different effective temperatures (vide infra)), which allows for deconvolution of the enthalpic and entropic contributions. Equations 6. 2-6. 4 summarize the data analysis. Teff is the effective temperature of the dissociating proton bound complex in Kelvin. The term "∆(∆S)" is the difference in the ∆S associated with the two channels in equation 6. 1. A plot of ln(k1/k2)

app versus PA(Bi) yields the Teff and the "GB (A)". Plotting equation 6. 4 at different values of Teff yields the proton affinity and ∆∆S for the Cooks measurement.

131

In this chapter, however, we only measured the relative PA of the reference base

1,4-butanediamine versus the ethyl methyl carbene 1b. A Teff of 340 K, which was obtained from the Cooks measurement of 1b with DBU and DBN, was used.

To generate the proton-bound dimers, 0.5 mM solutions in methanol were electrosprayed. To generate deuteron-bound dimer of 1b with 1,4-butanediamine, 80%

D2O/20% MeOD was used as the solvent. A needle voltage of 5 kV, capillary temperature of 100 oC, and a flow rate of about 25 µL/min were typically used. The proton bound complexes were isolated and activated for about 30 ms. 20~40 scans were averaged for the product ions. Calculations. Calculations were conducted at B3LYP/6-31+G* and MP2/6- 311+G(2d,p)//B3LYP/6-31+G* using Gaussian09; the geometries were fully optimized and frequencies were calculated.112,113,116,275 All the values reported are at 298 K. No scaling factor was applied. All calculated TS structures have one negative frequency.

132

Reference:

(1) Watson, J. D. The Double Helix: A Personal Account of the Discovery of the Structure of DNA; Norton Critical Editions, 1968. (2) Watson, J. D.; Crick, F. H. C. Nature 1953, 171, 737-738. (3) Saenger, W. Principles of Nucleic Acid Structure; Springer-Verlag: New York, 1984. (4) Lee, J. K.; Houk, K. N. Science 1997, 276, 942-945. (5) Nguyen, M. T.; Chandra, A. K.; Zeegers-Huyskens, T. Faraday Transactions 1998, 94, 1277-1280. (6) Chandra, A. K.; Nguyen, M. T.; Uchimaru, T.; Zeegers-Huyskens, T. J. Phys. Chem. A 1999, 103, 8853-8860. (7) De Hoffmann, E.; Charrette, J.; Stroobant, V. Mass Spectrometry Principles and Applications; John Wiley&Sons: Chichester, 1999. (8) NIST Chemistry WebBook, Nist Standard Reference Database Number 69, June 2005; Linstrom, P. J.; Mallard, W. G., Eds.; National Instituent of Atandards and Technology: Gaithersburg, MD 20899, 2005; http://webbook.nist.gov. (9) Sharma, S.; Lee, J. K. J. Org. Chem. 2002, 67, 8360-8365. (10) Sharma, S.; Lee, J. K. J. Org. Chem. 2004, 69, 7018-7025. (11) Kurinovich, M. A.; Lee, J. K. J. Am. Chem. Soc. 2000, 122, 6258-6262. (12) Kurinovich, M. A.; Lee, J. K. J. Am. Chem. Soc. Mass Spectrom. 2002, 13, 985-995. (13) Kurinovich, M. A.; Phillips, L. M.; Sharma, S.; Lee, J. K. Chem. Commun. 2002, 2354-2355. (14) Stivers, J. T.; Jiang, Y. Chem. Rev. 2003, 103, 2729-2759. (15) Speina, E.; Kierzek, A. M. Mutat. Res. 2003, 531, 205-217. (16) Wang, P.; B., G. A. DNA Repair 2006, 5, 23-31. (17) Berti, P. J.; McCann, J. A. B. Chem. Rev. 2006, 106, 506-555. (18) Mishina, Y.; Yang, C.-G.; He, C. J. Am. Chem. Soc. 2005, 127, 14594- 14595, and references therein. (19) Chandrasekhar, V.; Boomishankar, R.; Nagendran, S. Chem. Rev. 2004, 104, 5847-5910. (20) Damrauer, R.; Simon, R.; Krempp, M. J. Am. Chem. Soc. 1991, 113, 4431-4435.

133

(21) Arduengo, A. J. I.; Goerlich, J. R.; Marshall, W. J. J. Am. Chem. Soc. 1991, 113, 361-363. (22) Arduengo, A. J. I.; Dias, H. V. R.; Harlow, R. L.; Kline, M. J. Am. Chem. Soc. 1992, 114, 5530-5534. (23) Arduengo, A. J. I.; Dias, H. V. R.; Dixon, D. A.; Harlow, R. L.; Klooster, W. T.; Koetzle, T. F. J. Am. Chem. Soc. 1994, 116, 6812-6822. (24) Arduengo, A. J. I.; Goerlich, J. R.; Marshall, W. J. J. Am. Chem. Soc. 1995, 117, 11027-11028. (25) Dixon, D. A.; Arduengo, A. J. I. J. Phys. Chem. 1991, 95, 4180-4182. (26) Amyes, T. L.; Diver, S. T.; Richard, J. P.; Rivas, F. M.; Toth, K. J. Am. Chem. Soc. 2004, 126, 4366-4374. (27) Bourissou, D.; Guerret, O.; Gabbai, F. P.; Bertrand, G. 2000, 100, 39-91. (28) Igau, A.; Baceiredo, A.; Trinquier, G.; Bertrand, G. Angew. Chem. Int. Ed. Engl. 1989, 101, 617-618. (29) Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18-29. (30) Marion, N.; Nolan, S. P. Acc. Chem. Res. 2008, 41, 1440-1449. (31) Louie, J. N-Heterocyclic Carbenes in Synthesis 2006, 163-182. (32) Tekavec, T. N.; Louie, J. Topics in Organometallic Chemistry 2007, 21, 159-192. (33) Nolan, S. P. N-Heterocyclic Carbenes in Synthesis; Wiley-VCH: Weinheim, Germany, 2006. (34) Struble, J. R.; Kaeobamrung, J.; Bode, J. W. Org. Lett. 2008, 10, 957-960 and references therein. (35) Welton, T. Chem. Rev. 1999, 99, 2071-2083. (36) Formentin, P.; Garcia, H.; Leyva, A. J. Mol. Catal. A: Chemical 2004, 214, 137-142. (37) Wasserscheid, P.; Keim, W. Angew. Chem. Int. Ed. Engl. 2000, 39, 3772- 3789. (38) Dupont, J.; Spencer, J. Angew. Chem. Int. Ed. Engl. 2004, 43, 5296-5297. (39) Debray, J.; Leveque, J.-M.; Philouze, C.; Draye, M.; Demeunynck, M. J. Org. Chem. 2010, ASAP. (40) Walsh, C. Enzymatic Reaction Mechanisms; W. H. Freeman and Company: New York, 1979. (41) Breslow, R. J. Am. Chem. Soc. 1958, 80, 3719-3726.

134

(42) Chen, Y.-T.; Barletta, G. L.; Haghjoo, K.; Cheng, J. T.; Jordan, F. J. Org. Chem. 1994, 59, 7714-7722 and references therein. (43) Kluger, R. Chem. Rev. 1987, 87, 863-876. (44) Chen, H.; Justes, D. R.; Cooks, R. G. Org. Lett. 2005, 7, 3949-3952. (45) Lee, M.-T.; Hu, C.-H. Organometallics 2004, 23, 976-983. (46) Herrmann, W. A.; Kocher, C. Angew. Chem. Int. Ed. Engl. 1997, 36, 2162-2187. (47) Boehme, C.; Frenking, G. Organometallics 1998, 17, 5801-5809. (48) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64-71. (49) Hillenkamp, F. Anal. Chem. 1991, 63:A, 1193-1203. (50) Amster, I. J. J. Mass Spectrom. 1996, 31, 1325-1337. (51) Marshall, A. G.; Grosshans, P. B. Anal. Chem. 1991, 63A, 215-229. (52) Comisarow, M. B.; Marshall, A. G. Chemical Physics Letters 1974, 25, 282-283. (53) Sanz-Medel, A. Analyst 2000, 125, 35-43. (54) Gomez, A.; Tang, K. Fluids 1994, 6, 404-414. (55) Kebarle, P. J. Mass Spectrom. 2000, 35, 804-817. (56) Taylor, G. Proc. Roy. Soc. London Ser. A 1964, 280, 383-397. (57) March, R. E. Quadrupole Ion Trap Mass Spectrometer; John Wiley & Sons Ltd: Chichester, 2000. (58) Paul, W. Angew. Chem. 1990, 102, 780-789. (59) Douglas, D. J.; Frank, A. J.; Mao, D. Mass Spectrometry Reviews 2005, 24, 1-29. (60) Chesnavich, W. J.; Su, T.; Bowers, M. T. J. Chem. Phys. 1980, 72, 2641- 2655. (61) Su, T.; Chesnavich, W. J. J. Chem. Phys. 1982, 76, 5183-5185. (62) Bowers, M. T.; Laudenslager, J. B. J. Chem. Phys. 1972, 56, 4711. (63) Gupta, S. K. Can. J. Chem. 1967, 45, 3107. (64) Moran, T. F.; Hamil, W. H. J. Chem. Phys. 1963, 39, 1413. (65) Cooks, R. G.; Kruger, T. L. J. Am. Chem. Soc. 1977, 99.

135

(66) MuLuckey, S. A.; Cameron, D.; Cooks, R. G. J. Am. Chem. Soc. 1981, 103, 1313-1317. (67) Brodbelt-Lustig, J. S.; Cooks, R. G. Talanta 1989, 36, 255-260. (68) Green-Church, K. B.; Limbach, P. A. J. Am. Chem. Soc. Mass Spectrom. 2000, 11, 24-32. (69) Green-Church, K. B.; Limbach, P. A. J. Am. Soc. Mass Spectrom. 2000, 11, 24-32. (70) Ervin, K. M. Chem. Rev. 2001, 101, 391-444. (71) Gronert, S.; Feng, W. Y.; Chew, F.; Wu, W. Int. J. Mass Spectrom. Ion Proc. 2000, 196, 251-258. (72) Frisch, R. C.; al., e. Gaussian03, Rev. C.; Gaussian Inc: Wallingford, CT, 2004. (73) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995-2001. (74) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem. 2003, 24, 669-681. (75) Bartsch, H.; Barbin, A.; Marion, M.-J.; Nair, J.; Guichard, Y. Drug Metab. Rev. 1994, 26, 349-371. (76) Tudek, B. Mol. Aspects Med. 2007, 28, 258-275. (77) Martinez, G. R.; Loureiro, A. P. M.; Marques, S. A.; Miyamoto, S.; Yamaguchi, L. F.; Onuki, J.; Almeida, E. A.; Garcia, C. C. M.; Barbosa, L. F.; Medeiros, M. H. G.; Di Mascio, P. Mutat. Res. 2003, 544, 115-127. (78) Choi, J. H.; Pfeifer, G. P. Mutat. Res. 2004, 568, 245-256. (79) Gros, L.; Ishchenko, A. A.; Saparbaev, M. Mutat. Res. 2003, 531, 219-229. (80) Douki, T.; Odin, F.; Caillat, S.; Favier, A.; Cadet, J. Free Radical Biol. Med. 2004, 37, 62-70. (81) Chung, F.-L.; Ishchenko, A. A.; Saparbaev, M. Mutat. Res. 2003, 531, 219-229, and references therein. (82) Nair, J.; Carmichael, P. L.; Fernando, R. C.; Phillips, D. H.; Strain, A. J.; Bartsch, H. Cancer Epidemiol. Biomarkers Prev. 1998, 7, 435-440. (83) Montalto, G.; Cervello, M.; Giannitrapani, L.; Dantona, F.; Terranova, A.; Castagnetta, L. A. Ann. N.Y. Acad. Sci. 2002, 963, 13-20. (84) Nair, U.; Bartsch, H.; Nair, J. Free Radical Biol. Med. 2007, 43, 1109- 1120. (85) It has recently been discovered that E. coli Alk B directly (not via BER) repairs exocyclic DNA adducts such as εA. See ref 18.

136

(86) Berdal, K. G.; Johansen, R. F.; Seeberg, E. Embo. J. 1998, 17, 363-367. (87) Engelward, B. P.; Weeda, G.; Wyatt, M. D.; Broekhof, J. L. M.; Wit, J. d.; Donker, I.; Allan, J. M.; Gold, B.; Hoeijmakers, J. H. J.; Samson, L. D. Proc. Natl. Acad. Sci. USA 1997, 94, 13087-13092. (88) Hang, B.; Singer, B.; Margison, G. P.; Elder, R. H. Proc. Natl. Acad. Sci. USA 1997, 94, 12869-12874. (89) Lau, A. Y.; Wyatt, M. D.; Glassner, B. J.; Samson, L. D.; Ellenberger, T. Proc. Natl. Acad. Sci. USA 2000, 97, 13573-13578. (90) O’ Connor, T. R. Nucl. Acid Res. 1993, 21, 5561-5569. (91) O'Brien, P. J.; Ellenberger, T. Biochemistry 2003, 42, 12418-12429. (92) O'Brien, P. J.; Ellenberger, T. J. Biol. Chem. 2004, 279, 26876-26884. (93) O'Brien, P. J.; Ellenberger, T. J. Biol. Chem. 2004, 279, 9750-9757. (94) Wyatt, M. D.; Allan, J. M.; Lau, A. Y.; Ellenberger, T. E.; Samson, L. D. BioEssays 1999, 21, 668-676. (95) Guliaev, A. B.; Hang, B.; Singer, B. Nucl. Acid Res. 2002, 30, 3778-3787. (96) Lau, A. Y.; Schärer, O. D.; Samson, L.; Verdine, G. L.; Ellenberger, T. Cell 1998, 95, 249-258. (97) Vallur, A. C.; Feller, J. A.; Abner, C. W.; Tran, R. K.; Bloom, L. B. J. Biol. Chem. 2002, 277, 31673-31678. (98) Xia, L.; Zheng, L.; Lee, H.-W.; Bates, S. E.; Federico, L.; Shen, B.; O'Connor, T. R. J. Mol. Biol. 2005, 346, 1259-1274. (99) Previous work indicates that glycosidic bond hydrolyses proceed via a highly dissociative SN2 or a stepwise SN1 mechanism; to simplify Scheme 2.1 we show the mechanism as SN1. Furthermore, the mechanism for enzyme cleavage also involves a “base-flipping” step, whereby the nucleobase is “flipped” into the enzyme active site; this step may also differ for different substrates. It is not known whether base flipping is a relatively fast or slow step in AAG, and herein we focus on the steps involving actual nucleobase excision. See references 14, 17 and 93. (100) Some damaged nucleobases are positively charged (such as 7-methyl guanine); these nucleobases may be cleaved as neutrals. See references 91 and 93. (101) Another prevalent mechanism is that the damaged nucleobase is protonated before cleavage. See references 91 and 93. Among εA, adenine and guanine, εA is the most basic. One might argue that therefore, under such a mechanism, εA would be most easily cleaved (since it would be most easily protonated). However, we conducted calculations to ascertain leaving group ability of neutral εA versus adenine versus guanine, and although εA is most easily protonated, it is the poorest neutral leaving group (which is consistent with its higher PA).

137

(102) Bennett, M. T.; Rodgers, M. T.; Hebert, A. S.; Ruslander, L. E.; Eisele, L.; Drohat, A. C. J. Am. Chem. Soc. 2006, 128, 12510-12519. (103) Sun, X.; Lee, J. K. J. Org. Chem. 2007, 72, 6548-6555. (104) Kochetkov, N. K.; Shibaev, V. N.; Kost, A. A. Tetrahedron Lett. 1971, 12, 1993-1996. (105) Sattsangi, P. D.; Barrio, J. R.; Leonard, N. J. J. Am. Chem. Soc. 1980, 102, 770-774. (106) Bartmess, J. E.; Georgiadis, R. M. Vacuum 1983, 33, 149-153. (107) Cooks, R. G.; Kruger, T. L. J. Am. Chem. Soc. 1977, 99, 1279-1281. (108) McLuckey, S. A.; Cameron, D.; Cooks, R. G. J. Am. Chem. Soc. 1981, 103, 1313-1317. (109) McLuckey, S. A.; Cooks, R. G.; Fulford, J. E. Int. J. Mass Spectrom. Ion Proc. 1983, 52, 165-174. (110) Drahos, L.; Vékey, K. J. Mass Spectrom. 1999, 34, 79-84. (111) Levie, R. D. Principles of Quantitative Chemical Analysis; McGraw-Hill: New York, 1996. (112) Becke, A. D. J. Chem. Phys. 1993, 98, 1372-1377. (113) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (114) Kohn, W.; Becke, A. D.; Parr, R. G. Density Functional Theory of Electronic Structure 1996, 100, 12974-12980. (115) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623-11627. (116) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785-789.

(117) To calculate the ∆H298, one must account for both the translational energy of the proton (3/2 RT) and the work term associated with the dissociation of one molecule into two (RT). For the ∆G298 values, we also account for the ∆S298 of the proton (experimental value = 26 e.u. see reference 8). (118) It is also interesting to note that the ion formed from N7 protonation of the N9 tautomer 1 is the same ion as that formed from N9 protonation of the N7 tautomer 7. Deprotonation of this ion can therefore yield either 1 (deprotonation of the N7 proton) or 7 (deprotonation of the N9 proton). Therefore, in the FTMS proton affinity bracketing experiments, there is a possibility that some of the N7 tautomer 7 gets protonated by hydronium ion, then deprotonated by neutral 1 to form another molecule of 1. We have no way of being certain that this "isomerization" is not occurring. Regardless of whether this isomerization occurs, our bracketing experiments appear to measure only the canonical tautomer 1. (119) Chen, E. C. M.; Herder, C.; Chen, E. S. J. Mol. Struct. 2006, 798, 126-133.

138

(120) Albert, A.; Brown, D. J. J. Chem. Soc. 1954, 2060-2071. (121) Jang, Y. H.; Goddard III, W. A.; Noyes, K. T.; Sowers, L. C.; Hwang, S.; Chung, S.; Chung, D. S. j. phys. Chem. B 2003, 107, 344-357. (122) Taylor, H. F. W. J. Chem. Soc. 1948, 765-766. (123) Lee, J. K. INT. J. Mass Spectrom 2005, 240, 261-272. (124) Michalczyk, R.; Russu, I. M. Biophys. J. 1999, 76, 2679-2686. (125) Cruzeiro-Hansson, L.; Parker, K.; Goodfellow, J. M. J. Chem. Soc., Faraday Trans. 1995, 91, 251-258. (126) Nakayama, N.; Tanaka, S.; Kikuchi, O. J. Theor. Biol. 2002, 215, 13-22. (127) Parker, K.; Cruzeiro-Hansson, L.; Goodfellow, J. M. J. Chem. Soc., Faraday Trans. 1993, 89, 2637-2650. (128) Parthasarathy, R.; Fridey, S. M.; Carcinogenesis 1986, 7, 221-227. (129) Patel, D. J.; Shapiro, L.; Kozlowski, S. A.; Gaffney, B. L.; Jones, R. A. Biochemistry 1986, 25, 1027-1036. (130) Patel, D. J.; Shapiro, L.; Kozlowski, S. A.; Gaffney, B. L.; Jones, R. A. Biochemistry 1986, 25, 1036-1042. (131) Pederson, L. G.; Darden, T. A.; Deerfield, D. W. I.; Anderson, M. W.; Hoel, D. G. Carcinogenesis 1988, 9, 1553-1562. (132) Tian, Z.; Kass, S. R. J. Am. Chem. Soc. 2008, 130, 10842-10843. (133) Tian, Z.; Kass, S. R. Angew. Chem. Int. Ed. 2009, 48, 1321-1323. (134) Beranek, D. T. Mutat. Res. 1990, 231, 11-30. (135) De Bont, R.; Van Larabeke, N. Mutagenesis 2004, 19, 169-185. (136) Dimitri, A.; Burns, J. A.; Broyde, S.; Scicchitano, D. A. Nucl. Acid Res. 2008, 36, 6459-6471, and references therein. (137) Liu, J. G.; Li, M. H. Carcinogenesis 1989, 10, 617-620. (138) Loechler, E. L. Chem. Res. Toxicol. 1994, 7, 277-280. (139) Rydberg, B.; Lindahl, T. Embo. J. 1982, 1, 211-216. (140) Sedgwick, B. Carcinogenesis 1997, 18, 1561-1567. (141) Souliotis, V. L.; Kaila, S.; Boussiotis, V. A.; Pangalis, G. A.; Kyrtopoulos, S. A. Cancer Res. 1990, 50, 2759-2764. (142) Taverna, P.; Catapano, C. V.; Citti, L.; Bonfanti, M.; D'Incalci, M. Anticancer Drugs 1992, 3, 401-405.

139

(143) Wyatt, M. D.; Pittman, D. L. Chem. Res. Toxicol. 2006, 19, 1580-1594. (144) Loechler, E. L.; Green, C. L.; Essigmann, J. M. Proc. Natl. Acad. Sci. USA 1984, 81, 6271-6275. (145) Singer, B.; Dosanjh, M. K. Mutat. Res. 1990, 233, 45-51. (146) Toorchen, D.; Topal, M. D. Carcinogenesis 1983, 4, 1591-1597. (147) Abbot, P. J.; Saffhill, R. Biochim. Biophys. Acta 1979, 562, 51-61. (148) Kamiya, H.; Sakaguchi, T.; Fujimoro, M.; Miura, H.; Ishikawa, H.; Shimizu, M.; Inoue, H.; Matsukage, A.; Masutani, C.; Hanaoka, F.; Ohtsuka, E. Chem. Pharm. Bull. 1992, 40, 2792-2795. (149) Galloway, S. M.; Greenwood, S. K.; Hill, R. B.; Bradt, C. I. Mutat. Res. 1995, 346, 231-245. (150) Kaina, B.; Fritz, G.; Coquerelle, T. Environ. Mol. Mutagen. 1993, 22, 283- 292. (151) Marchesi, F.; Turriziani, M.; Tortorelli, G.; Avvisati, G.; Torino, F.; De Vecchis, L. Pharmacol. Res. 2007, 56, 275-287. (152) Voigt, J. M.; Topal, M. D. Carcinogenesis 1995, 16, 1775-1782. (153) Zhang, H.; Tsujimura, T.; Bahattacharyya, N. P.; Maher, V. M.; McCormick, J. J. Carcinogenesis 1996, 17, 2229-2235. (154) Gaffney, B. L.; Jones, R. A. Biochemistry 1989, 28, 5881-5889. (155) Gaffney, B. L.; Marky, L. A.; Jones, R. A. Biochemistry 1984, 23, 5686- 5691. (156) Kuzmich, S.; Marky, L. A.; Jones, R. A. Nucl. Acid Res. 1983, 11, 3393- 3403. (157) Pan, S.; Sun, X.; Lee, J. K. j. Am. Chem. Soc. Mass Spectrom. 2006, 17, 1383-1395. (158) Pan, S.; Sun, X.; Lee, J. K. Int. J. Mass Spectrom 2006, 253, 238-248. (159) Note that we did not actually calculate the base pairs; we merely draw the configurations in Figure 3.3 to help the reader visualize the relevant hydrogen bonds. (160) Gaffney, B. L.; Goswami, B.; Jones, R. A. J. Am. Chem. Soc. 1993, 115, 12607-12608. (161) Caldwell, J. W.; Kollman, P. A. J. Biomol. Struct. Dyn. 1985, 3, 57-66. (162) Spratt, T. E.; Levy, D. E. Nucl. Acid Res. 1997, 25, 3354-3361. (163) Sriram, M.; van der Marel, G. A.; Roelen, H. L. P. F.; van boom, J. H.; Wang, A. H.-J. Biochemistry 1992, 31, 11823-11834.

140

(164) Caldwell, J. W.; Kollman, P. A. J. Biomol. Struct. Dyn. 1985, 3, 57-66. (165) Leonard, G. A.; Thomson, J.; Watson, W. P.; Brown, T. Proc. Natl. Acad. Sci. USA 1990, 87, 9573-9676. (166) When adenosine is methylated on the 3-position, a positively charged species is formed, whose pKa (to form the neutral structure), is to our knowledge unknown. In our previously published studies (reference 9), we probed the acidity at N9 of the neutral structure; that is the value reported herein. (167) Abner, C. W.; Lau, A. Y.; Ellenberger, T. E. J. Biol. Chem. 2001, 276, 13379-13387. (168) Chen, J. M.; Zhang, Y. P.; Wang, C.; Sun, Y.; Fujimoto, J.; Ikenaga, M. Carcinogenesis 1992, 13, 1503-1507. (169) Pegg, A. E.; Dolan, M. E.; Scicchitano, D. A.; Moriomoto, K. Environ. Health Perspect. 1985, 62, 109-114. (170) Pegg, A. E.; Fang, Q.; Loktionova, N. A. DNA Repair 2007, 6, 1071-1078. (171) Chen, E. C. M.; Chen, E. S. J. Phys. Chem. B 2000, 104, 7835-7844. (172) Greco, F.; Liguori, A.; Sindona, G.; Uccella, N. J. Am. Chem. Soc. 1990, 112, 9092-9096. (173) Huang, Y.; Kenttamaa, H. J. Phys. Chem. A 2003, 107, 4893-4897 and references therein. (174) Meot-Ner (Mautner), M. J. Am. Chem. Soc. 1979, 101, 2396-2403. (175) Podolyan, Y.; Gorb, L.; Leszczynski, J. J. Phys. Chem. A 2000, 104, 7346- 7352 and references therein. (176) Kobayashi, R. J. Phys. Chem. A 1998, 102, 10813-10817. (177) Turecek, F.; Yao, C. J. Phys. Chem. A 2003, 107, 9221-9231. (178) Wolken, J. K.; Yao, C.; Turecek, F.; Polce, M. J.; Wesdemiotis, C. Int. J. Mass Spectrom 2007, 267, 30-42. (179) Yang, Z.; Rodgers, M. T. Phys. Chem. Chem. Phys. 2004, 6, 2749-2757 and references therein. (180) Yao, C.; Cuadrado-Peinado, M. L.; Polasek, M.; Turecek, F. J. Mass Spectrom. 2005, 40, 1417-1428 and references therein. (181) Yao, C.; Turecek, F.; Polce, M. J.; Wesdemiotis, C. Int. J. Mass Spectrom 2007, 265, 106-123. (182) Papoulis, A.; Al-Abed, Y.; Bucala, R. Biochemistry 1995, 34, 648-655. (183) Armentrout, P. B. J. Am. Chem. Soc. Mass Spectrom. 2000, 11, 371-379.

141

(184) Cheng, X.; Wu, Z.; Fenselau, C. J. Am. Chem. Soc. 1993, 115, 4844-4848. (185) Nold, M. J.; Cerda, B. A.; Wesdemiotis, C. j. Am. Chem. Soc. Mass Spectrom. 1999, 10, 1-8. (186) Williams, T. I.; Denault, J. W.; Cooks, R. G. Int. J. Mass Spectrom 2001, 210/211, 133-146. (187) Cerda, B. A.; Wesdemiotis, C. J. Am. Chem. Soc. 1996, 118, 11884-11892. (188) Barker, D. L.; Marsh, R. E. Acta Crystallogr 1964, 17, 1581-1587. (189) Dreyfus, M.; Bensaude, O.; Dodin, G.; Dubois, J. E. J. Am. Chem. Soc. 1976, 98, 6338-6349. (190) McClure, R. J.; Craven, B. M. Acta Crystallogr 1973, B29, 1234-1238. (191) Nir, E.; Muller, M.; Grace, L. I.; de Vries, M. S. Chem. Phys. Lett. 2002, 355, 59-64. (192) Sambrano, J. R.; de Souza, A. R.; Queralt, J. J.; Andres, J. Chem. Phys. Lett. 2000, 317, 437-443. (193) Szczesniak, M.; Szcepaniak, K.; Kwiatkowski, J. S.; KuBulat, K.; Person, W. B. J. Am. Chem. Soc. 1988, 110, 8319-8330. (194) Trygubenko, S. A.; Bogdan, T. V.; Rueda, M.; Orozco, M.; Luque, F. J.; Sponer, J.; Slavicek, P.; Hobza, P. Phys. Chem. Chem. Phys. 2002, 4, 4192-4203. (195) Ueda, T.; Fox, J. J. J. Am. Chem. Soc. 1963, 85, 4024-4028. (196) Weber, H. P.; Craven, B. M.; MuMullan, R. K. Acta Crystallogr 1980, B36, 645-649. (197) If more than one value is listed for an atom, the arrows show the site of protonation (for example, the O2 of cytosine can be protonated on the "N1 side" and on the "N3 side"). (198) Szczesniak, M.; Leszczynski, J.; Person, W. B. J. Am. Chem. Soc. 1992, 114, 2731-2733. (199) Ucella and coworkers (reference 172) used the kinetic method to obtain a cytosine PA of 225.9 kcal mol-1. Mautner (reference 174) conducted high pressure mass spectrometry equilibrium measurements, obtaining a cytosine PA of 224.9 kcal mol-1. NIST subsequently evaluated these values (updating for changes in the reference acid and base scale) to report an evaluated PA of 227.0 kcal mol-1 (reference 8). (200) Wilson, M. S.; McCloskey, J. A. J. Am. Chem. Soc. 1975, 97, 3436-3444. (201) Chandra, A. K.; Michalska, D.; Wysokinsky, R.; Zeegers-Huyskens, T. J. Phys. Chem. A 2004, 108, 9593-9600. (202) Brown, R. D.; Godfrey, P. D.; McNaughton, D.; Pierlot, A. P. J. Am. Chem. Soc. 1989, 111, 2308-2310.

142

(203) We calculated the ∆∆S values for the cytosine acidity and PA Cooks extended kinetic method experiments, using the Armentrout method (see subsequent reference). For the acidity, the ∆∆S value is 2.6 cal K-1 mol-1; for the PA, it is 2.2 cal K-1 mol-1. It has been noted that the ∆∆S value is related to the accuracy of the PA value obtained by the Cooks extended method (see reference 201). Ideally, the actual ∆∆S value should be less than or equal to about 5 cal K-1 mol-1; otherwise, the extended kinetic method may underestimate the PA. Another caveat is that the ∆∆S value obtained from the extended method is often itself underestimated. There is therefore a possibility that the values obtained via our Cooks extended method experiment are underestimated; however, given that they are in agreement with the bracketing results, we are inclined to believe that the extended kinetic method value is not too low. (204) Wesdemiotis, C. J. Mass Spectrom. 2004, 39, 998-1003. (205) Ucella and coworkers used the kinetic method to obtain a thymine PA of 209.0 kcal mol-1 (reference 172). Mautner conducted high pressure mass spectrometry equilibrium measurements, obtaining a thymine PA of 210.9 kcal mol-1 (reference 174). NIST subsequently evaluated these values (updating for changes in the reference acid and base scale) to report an evaluated PA of 210.5 ± 2.0 kcal mol-1 (reference 8). (206) Chandra, A. K.; Nguyen, M. T.; Zeegers-Huyskens, T. J. Phys. Chem. A 1998, 102, 6010-6016. (207) Fasman, G. D. Handbook of Biochemistry and Molecular Biology: Physical and Chemical Data; CRC Press, Inc.: Cleveland, OH, 1976; Vol. Vol. 1. (208) Juers, D. H.; Kim, J.; Matthews, B. W.; Sieburth, S. M. Biochemistry 2005, 44, 16524-16528. (209) Ignatyev, I. S.; Partal, F.; González, J. J. L. J. Mol. Struct: THEOCHEM 2004, 678, 249-256. (210) Buttrus, N. H.; Eaborn, C.; Hitchcock, P. B.; Saxena, A. K. J. Organomet. Chem. 1985, 284, 291-297. (211) Schneider, M.; Neumann, B.; Stammler, H.-G.; Jutzi, P. Monatsh. Chem. 1999, 130, 33-44. (212) Chandrasekhar, V.; Nagendran, S.; Butcher, R. J. Organometallics 1999, 18, 4488-4492. (213) Ignatyev, I. S.; Partal, F.; González, J. J. L. Chem. Phys. Lett. 2004, 384, 326-331. (214) Lickiss, P. D. Adv. Inorg. Chem. 1995, 42, 147-262. (215) Mutahi, M. W.; Nittoli, T.; Guo, L.; Sieburth, S. M. J. Am. Chem. Soc. 2002, 124, 7363-7375. (216) Sieburth, S. M.; Chen, C.-A. Eur. J. Org. Chem. 2006, 311-322. (217) Daiss, J. O.; Burschka, C.; Mills, J. S.; Montana, J. G.; Showell, G. A.; Warneck, J. B. H.; Tacke, R. Organometallics 2006, 25, 1188-1198.

143

(218) Sieburth, S. M.; Nittoli, T.; Mutahi, A. M.; Guo, L. Angew. Chem. Int. Ed. 1998, 37, 812-814. (219) Kim, Y. M.; Farrah, S.; Baney, R. H. Int. J. Antimicro. Ag. 2007, 29, 217- 222. (220) Gronert, S. Chem. Rev. 2001, 2, 329-360. (221) DePuy, C. H.; Bierbaum, V. M. Acc. Chem. Res. 1981, 14, 146-153. (222) N,N,N',N'-Tetramethylbutanediamine was chosen as the reference base because it yielded the best signal of the deuteron-bound dimer. We cannot guarantee that all the reference bases will also favor proton transfer over elimination, but given that the bases that we examined all have similar PAs, it is likely that they behave the same as the butanediamine. (223) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17, Suppl. No. 1. (224) Taft, R. W. Prog. Phys. Org. Chem. 1983, 14, 248-350. (225) Brauman, J. I.; Blair, L. K. J. Am. Chem. Soc. 1968, 90, 6561-6562. (226) Brauman, J. I.; Blair, L. K. J. Am. Chem. Soc. 1970, 92, 5986-5992. (227) Taft, R. W.; Taagepera, M.; Abbound, J. L. M.; Wolf, J. F.; DeFrees, D. J.; Hehre, W. J.; Bartmess, J. E.; McIver, R. T. J. Am. Chem. Soc. 1978, 100, 7765-7767. (228) Taft, R. W.; Koppel, I. A.; Topsom, R. D.; Anvia, F. J. Am. Chem. Soc. 1990, 112. (229) Doyle, A. G.; Jacobsen, E. N. Chem. Rev. 2007, 107, 5713-5743. (230) Anderson, C. D.; Dudding, T.; Gordillo, R.; Houk, K. N. Org. Lett. 2008, 10, 2749-2752. (231) Akakura, M.; Kawasaki, M.; Yamamoto, H. Eur. J. Org. Chem. 2008, 4245-4249. (232) Kamberaj, H.; Low, R. J.; Neal, M. P. Mol. Phys. 2006, 104, 335-357. (233) Tran, N.; Min, T.; Franz, A. K. Angew. Chem. Int. Ed. 2011, in press (234) Kondo, S.-I.; Fukuda, A.; Yamamura, T.; Tanaka, R.; Unno, M. Tetrahedron Lett. 2007, 48, 7946-7949. (235) Kondo, S.-I.; Harada, T.; Tanaka, R. Org. Lett. 2006, 8, 4621-4624. (236) West, R.; Baney, R. H. J. Am. Chem. Soc. 1959, 81, 6145-6148. (237) Thadani, A. N.; Stankovic, A. R.; Rawal, V. H. Proc. Natl. Acad. Sci. USA 2004, 101, 5846-5850. (238) Steward, O. W.; Fussaro, D. R. J. Organomet. Chem. 1977, 129, C28-C32.

144

(239) Enders, D.; Niemeier, O.; Henseler, A. Chem. Rev. 2007, 5606-5655. (240) Lord, R. L.; Wang, H.; Vieweger, M.; Baik, M.-H. J. Organomet. Chem. 2006, 5505-5512. (241) Droge, T.; Glorius, F. Angew. Chem. Int. Ed. Engl. 2010, 6940-6952 and references therein. (242) Amyes, T. L.; Diver, S. T.; Richard, J. P.; Rivas, F. M.; Toth, K. Journal of the American Chemical Society 2004, 126, 4366-4374. (243) Chen, H.; Justes, D. R.; Cooks, R. G. Organic Letters 2005, 7, 3949-3952. (244), The improved efficacy of the second generation over the first generation ligands has been attributed to catalyst commitment differences: Sanford, M. S.; Ulman, M.; Grubbs, R. H. J. Am. Chem. Soc. 2001, 123, 749-750; Sanford, M. S.; Love, J. A.; Grubbs, R. H. J. Am. Chem. Soc. 2001, 123, 6543-6554; Adlhart, C.; Chen, P. Helv. Chim. Acta 2003, 86, 941-950 (gas phase). The fundamental basis for the differing effects by ligand is still unknown. (245) Robinson, D. R. Journal of the American Chemical Society 1970, 92, 3138-3146. (246) The gas phase PAs of other "nonstabilized" carbenes have been measured. See: Tian, Z.; Kass, S. R. Int. J. Mass Spectrom. 2007, 267, 288-294, and references therein. (247) Chen, Justes and Cooks report a calculated PA value of 260.8 kcal/mol; our value is 1.5 kcal/mol higher as we account for the enthalpy of the proton at 298 K. (248) DBU and DBN were used in the Cooks kinetic method experiment because they yielded reasonable protonated dimer signal. However, they were not volatile enough to use in the bracketing experiments so we used the bases listed in Table 6.2. (249) Ikuta, S.; Kebarle, P. Can. J. Chem. 1983, 61, 97-102. (250) Houk, K. N.; Beno, B. R.; M., N.; Black, K.; Yoo, H. Y.; Wilsey, S.; Lee, J. K. J. Mol. Struct. (THEOCHEM) 1997, 398-399, 169-179. (251) Grimme, S. Angew. Chem. Int. Ed. Engl. 2006, 45, 4460-4464. (252) Turecek, F. J. Phys. Chem. A 1998, 102, 4703-4713. (253) These fragmentation pathways are being further examined and will be the subject of a future paper. (254) Chupka, W. A. J. Chem. Phys. 1959, 191-211. (255) Rosenstock, H. M.; Wallenstein, M. B.; Wahrhaftig, A. L.; Eyring, H. Proc. Natl. Acad. Sci. USA 1952, 667-678. (256) Narancic, S.; Bach, A.; Chen, P. J. Phys. Chem. A 2007, 111, 7006-7013. (257) Gronert, S. Chemical Reviews 2001, 101, 329-360.

145

(258) Ervin, K. M. Chemical Reviews 2001, 101, 391-444 and references therein. (259) Bouchoux, G.; Sablier, M.; Berruyer-Penaud, F. J. Mass Spectrom. 2004, 39, 986-997. (260) NIST Chemistry WebBook, NIST Standard Reference Database Number 69, June 2005; Linstrom, P. J.; Mallard, W. G., Eds.; National Institute of Standards and Technology: Gaithersburg, MD 20899, 2005; Vol. http://webbook.nist.gov. (261) Gronert, S. Journal of the American Society for Mass Spectrometry 1998, 9, 845-848. (262) Kaljurand, I.; Koppel, I. A.; Kütt, A.; Room, E.-I.; Rodima, R.; Koppel, I.; Mishima, M.; Leito, I. Journal of Physical Chemistry A 2007, 111, 1245-50. (263) Haupert, L. J.; Poutsma, J. C.; Wenthold, P. G. Accounts of Chemical Research 2009, 42, 1480-88. (264) In our previous work we showed that the inaccurate calculation of the tricyclohexylphosphine PA using B3LYP is due to the neglect of medium-range electron correlation effects (a problem that is solved by using MP2 calculations). (265) Interestingly, in recent work, Turecek and co-workers have calculated a comparably high PA for the related species 4-methyl-N-3-H-imidazole radical: Turecek, F.; Chung, T. W.; Moss, C. L.; Wyer, J. A.; Ehlerding, A.; Holm, A. I. S.; Zettergren, H.; Nielsen, S. B.; Hvelplund, P.; Chamot-Rooke, J.; Bythell, B.; Paizs, B. J. Am. Chem. Soc. 2010, 132, 10728-10740. (266) Droge, T.; Glorius, F. Angew. Chem., Int. Ed 2010, 49, 6940-6952 and references therein. (267) Tolman, C. A. Chemical Reviews 1977, 77, 313–348. (268) Kelly, R. A. I.; Clavier, H.; Giudice, S.; Scott, N. M.; Stevens, E. D.; Bordner, J.; Samardjiev, I.; Hoff, C. D.; Cavallo, L.; Nolan, S. P. Organometallics 2008, 27, 202-210. (269) In general, the signal for protonated dimer 3a was quite poor and reproducibility was an issue. CID typically produced protonated carbene (path A), protonated phosphine (path B), and "m/z 295" (paths C,D), in ratios that varied somewhat. Relative PAs were measured in those cases where paths A/B were dominant. (270) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Organic Letters 1999, 1, 953-956. (271) Zhao, H.; Foss, F. W.; Breslow, R. Journal of the American Chemical Society 2008, 130, 12590-12591. (272) Liu, M.; Xu, M.; Lee, J. K. Journal of Organic Chemistry 2008, 73, 5907- 5914. (273) Sun, X.; Lee, J. K. Journal of Organic Chemistry 2007, 72, 6548-6555.

146

(274) Gaussian 09, Revision A.1, Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian, Inc., Wallingford CT, 2009.

147

CURRICULUM VITAE

Min Liu

Education:

1996 - 2000 B.S., Chemical Engineering Hebei University of Technology, P. R. China

2000 - 2003 M.S., Chemical Engineering Hebei University of Technology, P. R. China

2005 - 2011 Ph.D., Chemistry and Chemical Biology Rutgers, The State University of New Jersey

Publications:

1. Liu, M.; Tran, N. T.; Franz, A. K. and Lee, J. K. “Gas-Phase Acidity Studies of Dual Hydrogen-Bonding Organic Silanols and Organocatalysts,” J. Org. Chem. 2011, 76, 7186-7194. 2. Liu, M.; Chen, M.; Zhang, S. Yang, I.; Buckley, B.; Lee, J. K. “Reactivity of Carbene•Phosphine Dimers: Proton Affinity Revisited,” J. Phys. Org. Chem. ASAP 3. Liu, M.; Yang, I.; Buckley, B.; Lee, J. K. “Proton Affinities of Phosphines versus N-Heterocyclic Carbenes," Org. Lett. 2010, 21, 4764-4767. 4. Zhachkina, A.; Liu, M.; Sun, X.; Amegayibor, F. S. and Lee, J. K.; “Gas-Phase Thermochemical Properties of the Damaged Base O6-Methylguanine versus Adenine and Guanine,” J. Org. Chem. 2009, 74, 7429-7440. 5. Liu, M.; Li, T.; Amegayibor, F. S.; Cardoso, D. S.; Fu, Y.; Lee, J. K. “Gas-Phase Thermochemical Properties of Pyrimidine Nucleobases,” J. Org. Chem. 2008, 73, 9283-9291. 6. Liu, M.; Xu, M.; Lee, J. K. “The Intrinsic Reactivity of Ethenoadenine and Mechanism for Excision from DNA,” J. Org. Chem. 2008, 73, 5907-5914