<<

Coupling Ambient Spectrometry with Liquid and

Electrochemistry and Their Applications

A dissertation presented to

the faculty of

the College of Arts and Sciences of Ohio University

In partial fulfillment

of the requirements for the degree

Doctor of Philosophy

Yi Cai

December 2016

© 2016 Yi Cai. All Rights Reserved. 2

This dissertation titled

Coupling Ambient Ionization with Liquid Chromatography and

Electrochemistry and Their Applications

by

YI CAI

has been approved for

the Department of and

and the College of Arts and Sciences by

Hao Chen

Associate Professor of Chemistry and Biochemistry

Robert Frank

Dean, College of Arts and Sciences 3

ABSTRACT

CAI, YI, Ph.D., December 2016, Chemistry

Coupling Ambient Ionization Mass Spectrometry with Liquid Chromatography and

Electrochemistry and Their Applications

Director of Dissertation: Hao Chen

Ambient ionization methods allow the ionization of untreated samples in the open

environment. In this dissertation, two different ambient ionization techniques, desorption ionization (DESI) and probe (PESI), has been developed and coupled with liquid chromatograph (LC) and electrochemistry (EC) and their analytical applications have been explored and discussed.

Liquid sample DESI generally employs a DESI probe to spray with high to ionize sample as the sample is delivered to the source by a piece of fused silica transfer capillary. A new splitting interface, a PEEK capillary tube with a micro-orifice drilled in the capillary wall, was used to connect with LC column for applying

DESI ionization. A small portion of LC eluent emerging from the orifice can be directly ionized by DESI with negligible time delay while the remaining analytes can be online collected. Furthermore, online derivatization using reactive DESI is possible for additional application such as supercharging .

Since splitting via an orifice introduces negligible dead volume and back pressure, the performance of the LC/DESI-MS with the focus of using ultra-fast LC for analyzing

sample was further evaluated. Using a monolithic C18 column, metabolites in urine can be

separated within 1.6 min, online monitored by DESI and collected as purified samples. 4

Negative can be directly generated for acidic analytes in acidic LC eluent by DESI during the LC/MS analysis process using a spray solvent with alkaline pH. In addition,

DESI-MS is found to be compatible with ultra-performance liquid chromatography

(UPLC) for the first time. The 45 s separation of drugs can be achieved via UPLC/DESI-

MS under high .

The combination with EC further broadens LC/MS applications. UPLC-MS combined with EC via DESI was first developed for the structural analysis of proteins/ that contain disulfide bonds. Using this combined UPLC/EC/DESI-MS method, peptides containing disulfide bonds can be differentiated from those without disulfide bonds, as the former are electroactive and reducible. MS/MS analysis of disulfide-reduced ions provides increased information about the peptide sequence and disulfide-linkage pattern. In addition, upon online electrolytic reduction and reactive

DESI, supercharged proteins showed increased charges distribution which is of value for

MS/MS sequencing application.

PESI, another ambient ionization technique, employs a conductive solid probe to ionize samples directly on the with the aid of applied high voltage. Due to the high salt tolerance of PESI, the detection of electrochemical reaction products in room- temperature ionic liquids is realized, for the first time. Furthermore, PESI-MS allows the detection the electrochemical reaction products on different or multiple surfaces.

In addition, peptides and proteins fractionated through isoelectric focusing (IEF) can also be directly analyzed by PESI-MS.

5

DEDICATION

I dedicate this dissertation to my friends, parents, and husband

6

ACKNOWLEDGMENTS

I acknowledge my advisor, Dr. Hao Chen, for his great mentoring. His dedication to science deeply impacts me and encourages me to become a good researcher. He is also generous to support me for attending conferences.

I acknowledge to all my dissertation committee members, Dr. Peter de B.

Harrington, Dr. Shiyong Wu and Dr. Shigeru Okada for their guidance.

I acknowledge to National Science Foundation Career Award (to Dr. Hao Chen,

CHE-1149367), National Science Foundation Instrument Development for Biological

Research (CHE 1455554), National Natural Science Foundation of China (Grant

21328502), Edison Institute Faculty Fellowship, Merck Research

Laboratories New Review & Licensing Committee, Ohio Third Frontier

Technology Validation and Start-Up Fund, and Center for Intelligent Chemical

Instrumentation, Department of Chemistry and Biochemistry, Ohio University for the financial support.

I acknowledge to all former and current members in Dr. Hao Chen’s group, Dr.

Zhixin Miao, Dr. Yun Zhang, Zongqian Yuan, Pengyuan Liu, Mei Lu, Dr. Qiuling Zheng,

Si Cheng, He Xiao, Meihong Hu, Chang Xu, Yuexiang Zhang, Najah Almowalad,

Amanda Forni, Sabrina Cramer, Fengyao Li, Denial Adams, David Hu, Prof. Dr. Ping Li,

Dr. Ning Pan, Prof. Dr. Kehua Xu, Prof. Dr. Jun Wang, Prof. Dr. Qiuhua Wu and Prof.

Dr. Zhi Li for their assistance and support.

I acknowledge to all my collaborators, Dr. Howard D. Dewald (Ohio University);

Dr. Michael Held (Ohio University); Dr. Huifang Yao (Merck & Co., Inc., Rahway, 7

New Jersey), Dr. Yong Liu (Merck & Co., Inc., Rahway, New Jersey), Dr. Roy Helmy

(Merck & Co., Inc., Rahway, New Jersey) for their helpful discussion and suggestions.

I acknowledge to Dr. Andrew Tangonan, Bascom French, Paul Schmittauer,

Aaron Dillon, Carolyn Khurshid, Marlene Jenkins, Jackie Bennett-Hanning, Dr.

Zhengfang Wang, Dr. Mengliang Zhang, Xue Zhao, Xinyi Wang, and Dr. Lei Wang for their generous help. 8

TABLE OF CONTENTS Page

Abstract ...... 3 Dedication ...... 5 Acknowledgments...... 6 List of Figures ...... 11 List of Schemes ...... 16 List of Abbreviations ...... 17 Chapter 1: Introduction ...... 19 1.1 Mass Spectrometry ...... 19 1.2 Mass Spectrometer ...... 19 1.3 Ionization Methods ...... 20 1.3.1 ESI-MS ...... 21 1.3.2 DESI-MS ...... 22 1.3.3 Liquid Sample DESI-MS ...... 24 1.3.4 Probe Electrospray Ionization (PESI) ...... 25 1.4 EC-MS ...... 26 1.4.1 EC/ESI-MS ...... 28 1.4.2 EC/DESI-MS ...... 28 1.5 LC/DESI-MS ...... 30 Chapter 2: A New Splitting Method for Both Analytical and Preparative LC/MS ...... 33 2.1 Introduction ...... 33 2.2 Experimental ...... 35 2.2.1 Chemicals ...... 35 2.2.2 LC Separation Condition ...... 36 2.2.3 DESI-MS Detection ...... 36 2.3 Results and Discussion ...... 37 2.3.1 Detection ...... 37 2.3.2 Saccharide Detection...... 46 2.3.3 Sulfonamides Detection ...... 49 2.4 Conclusions ...... 51 9

Chapter 3: Coupling of Ultrafast LC with Mass Spectrometry by DESI ...... 53 3.1 Introduction ...... 53 3.2 Experimental ...... 55 3.2.1 LC/DESI-MS Instrument ...... 55 3.2.2 LC Separation Conditions ...... 56 3.3 Results and Discussion ...... 57 3.3.1 LC/DESI-MS Analysis of Drugs ...... 57 3.3.2 LC/DESI-MS analysis of Acidic Compounds ...... 61 3.3.3 UPLC/DESI-MS Analysis ...... 63 3.4 Conclusions ...... 66 Chapter 4: Integration of Electrochemistry with Ultra Performance Liquid Chromatography/Mass Spectrometry (UPLC/MS)...... 68 4.1 Introduction ...... 68 4.2 Experimental ...... 70 4.2.1 Chemicals ...... 70 4.2.2 Apparatus ...... 70 4.2.3 LC Separation Condition ...... 71 4.3 Results and Discussion ...... 72 4.3.1 Reduction of Somatostatin 1-14 ...... 72 4.3.2 Reduction of Insulin ...... 75 4.3.3 Reduction of Proteins ...... 86 4.4 Conclusions ...... 90 Chapter 5: Coupling Electrochemistry with Probe Electrospray Ionization Mass Spectrometry ...... 91 5.1 Introduction ...... 91 5.2 Experimental ...... 92 5.2.1 Chemicals ...... 92 5.2.2 Apparatus ...... 93 5.3 Results and Discussion ...... 94 5.3.1 PESI-MS Detection of Electrochemical Reactions in RTILs ...... 94 5.3.2 PESI-MS Detection of Electrochemical Reaction on Different ... 101 5.3.3 PESI-MS Detection of Separated Proteins/Peptides in IEF ...... 107 10

5.4 Conclusions ...... 111 Chapter 6: Summary and Future Work ...... 113 References ...... 115 Appendix: Publications ...... 129

11

LIST OF FIGURES

Page

Figure 1-1. Basic components of a mass spectrometer...... 20

Figure 1-2. Scheme showing an ESI source.[16] Reproduced with permission from

WILEY-VCH, Copyright 2001...... 22

Figure 1-3. Scheme showing the apparatus of DESI.[17] Reproduced with permission from Science, Copyright 2004...... 23

Figure 1-4. Scheme showing the apparatus of liquid sample DESI.[24] Reproduced with permission from Springer, Copyright 2009...... 25

Figure 1-5. Scheme showing the apparatus of PESI.[36] Reproduced with permission from

WILEY-VCH, Copyright 2007...... 26

Figure 1-6. Timeline for the development of EC/MS with different ionization method.[49]

Reproduced with permission from RSC, Copyright 2013...... 27

Figure 1-7. Scheme showing the apparatus for online coupling of a thin-layer electrochemical flow with DESI-MS.[28] Reproduced with permission from ACS,

Copyright 2009...... 30

Figure 1-8.Scheme showing the apparatus for LC/DESI-MS.[63] Reproduced with permission from ACS, Copyright 2011...... 32

Figure 2-1. EICs of (a) insulin (+4 ion) and (b) ubiquitin (+6 ion) from LC/DESI-MS analysis and the corresponding DESI-MS spectra of (c) insulin and (d) ubiquitin...... 38

Figure 2-2. EICs of (a) insulin (+4 ion) and (b) ubiquitin (+6 ion) from LC/ESSI-MS analysis...... 40 12

Figure 2-3. EICs of (a) insulin (+5 ion) and (b) ubiquitin (+7 ion) from LC/reactive

DESI-MS analysis and the corresponding reactive DESI-MS spectra of (c) insulin and (d) ubiquitin...... 42

Figure 2-4. EICs of (a) insulin (+4 ion) and (b) ubiquitin (+6 ion) from LC/DESI-MS analysis (100 μm i.d. orifice) and the collected (c) insulin and (d) collected ubiquitin. .. 45

Figure 2-5. Calibration curves of insulin and ubiquitin (UV absorption peak area vs. protein )...... 46

Figure 2-6. EICs of (a) N-acetyl-D-glucosamine and (b) maltohexaose eluted out of the

C18 column and detected by DESI-MS. EICs of (c) N-acetyl-D-glucosamine and (d) maltohexaose eluted out of the C18 column and detected by ESSI-MS. MS spectra of (e) the collected N-acetyl-D-glucosamine and (f) the collected maltohexaose. The re-analysis of the collected proteins was performed using ESSI-MS. The peak at m/z 223 in (f) came from solvent background...... 49

Figure 2-7. EIC spectra of a mixture of sulfadiazine, sulfamerazine, and sulfaquinoxaline

(a) from LC/DESI-MS analysis and (b) from LC/UV analysis...... 51

Figure 3-1. EICs of (a) dopamine, (b) 3-methoxytyramine, (c) L-kynurenine, and (d) L- tryptophan acquired by LC/DESI-MS...... 58

Figure 3-2. EICs of (a) dopamine, (b) 3-methoxytryamine, (c) L-kynurenine, and (d) L- tryptophan acquired by LC/ESI-MS...... 59

Figure 3-3. Calibration curves of 3-MT (UV absorption peak area vs. 3-MT concentration) ...... 61 13

Figure 3-4. EICs of (a) ketoprofen, (b) fenoprofen, and (c) ibuprofen acquired by

LC/negative ion DESI-MS. EICs of (a') ketoprofen, (b') fenoprofen, and (c') ibuprofen acquired by LC/negative ion ESI-MS...... 63

Figure 3-5. EICs of a) codeine, b) cocaine and c) flunitrazepam acquired by

UPLC/DESI-MS with MS/MS spectra shown in the figure insets. The red lines shown in the structures indicate the bond cleavage upon CID...... 64

Figure 3-6. EICs of a) codeine, b) cocaine and c) flunitrazepam acquired by high temperature UPLC/DESI-MS (80 oC column temperature). Re-analysis of collected d) codeine, e) cocaine and f) flunitrazepam after high temperature UPLC separation...... 66

Figure 4-1. a) The apparatus of UPLC/EC/DESI-MS. b) Extracted ion chromatograms

(EICs) showing the UPLC separation of b) AGCK TFTSC (upper panel) and NFFWK

(lower panel). DESI-MS spectra of AGCK TFTSC c) when cell was off and d) when cell was on...... 74

Figure 4-2. CID MS/MS spectra of a) [AGCK TFTSC+H]+ (m/z 933), b) reduced peptide [AGCK+H]+ (m/z 378) and c) reduced peptide [TFTSC+H]+ (m/z 558)...... 75

Figure 4-3. EICs of the pepsin-digested insulin...... 77

Figure 4-4. DESI-MS spectra of the digested insulin peptides carrying no disulfide bonds

YTPKA: a) cell off and b) cell on; FVNQ: c) cell off and d) cell on; GIVE: e) cell off and f) cell on; YQLEN: g) cell off and h) cell on; YQLE: i) cell off and j) cell on; and VEAL: k) cell off and l) cell on...... 78 14

Figure 4-5. DESI-MS spectra of P3 with a) cell off and b) cell on, c) CID MS2 spectrum of doubly charged P3 at m/z 1037, d) CID MS/MS spectrum of singly charged P3 chain A at m/z 1311, and d) CID MS/MS spectrum of singly charged P3 chain B at m/z 766...... 81

Figure 4-6. CID MS/MS spectra of a) [P1+2H]2+ (m/z 1082) and b) [P4+3H]3+ (m/z 854).

...... 83

Figure 4-7. DESI-MS spectra of P2 with a) cell off and b) cell on. CID MS/MS spectra of

[P2+2H]2+ (m/z 769) and [LVCGERGFF+2H]2+ (m/z 514)...... 85

Figure 4-8. EICs of the protein mixture containing insulin, myoglobin and α-lactalbumin.

...... 87

Figure 4-9. DESI-MS spectra of the UPLC-separated insulin a) when cell was off without supercharging, b) when cell was off with supercharging and c) when cell was on with supercharging...... 88

Figure 4-10. DESI-MS spectra of the UPLC-separated myoglobin a) when cell was off without supercharging, b) when cell was off with supercharging and c) when cell was on with supercharging...... 88

Figure 4-11. DESI-MS spectra of the UPLC-separated α-lactalbumin a) when cell was off without supercharging, b) when cell was off with supercharging and c) when cell was on with supercharging...... 89

Figure 5-1. a) Scheme showing the PESI-MS and the configuration.

PESI-MS spectra of b) CoII salen before and c) CoII after electrolysis. The inset in c) shows the oxidation reaction of CoII salen...... 96 15

Figure 5-2. CID MS/MS spectrum of the electrochemical oxidation product CoIII salen ion (m/z 325)...... 97

Figure 5-3. PESI-MS spectra of a) before electrolysis and b) ferrocene after electrolysis. The inset in c) shows the oxidation reaction of ferrocene...... 98

Figure 5-4. PESI-MS spectra of a) TEMPO before electrolysis and b) TEMPO after electrolysis. The inset in c) shows the oxidation reaction of ferrocene. CID MS/MS spectrum of the electrochemical oxidation product TEMPO ion (m/z 156)...... 100

Figure 5-5. a) Equations showing the electrochemical reduction of flutamide and the electrochemical oxidation of dopamine; EC/PESI-MS spectra of the electrolyzed mixture sampled from b) the surface and c) the surface...... 102

Figure 5-6. Traditional EC/MS spectrum of the mixture after electrolysis...... 104

Figure 5-7. a) Scheme showing the BPE cell configuration; b) the oxidation process of

CLZ. PESI-MS spectra of the electrolyzed CLZ solution sampled from four different electrode surfaces: c) position 1 (driving electrode anode), d) position 2 (BPE cathode), e) position 3 (BPE anode), and f) position 4 (driving electrode cathode)...... 106

Figure 5-8. a) Scheme illustrating the process of the IEF/PESI-MS; b) PESI-MS spectrum of the mixture sample in the IEF buffer before electrofocusing; PESI-MS spectra of c) angiotensin II (fraction #12), d) angiotensin II (1-4, fraction #9), and e) angiotensin II antipeptide (fraction #6)...... 109

Figure 5-9. a) PESI-MS spectrum of the mixed protein sample in the IEF buffer prior to isoelectric focusing; PESI-MS spectra of b) cytochrome c (fraction #20), c) ubiquitin

(fraction #12) and d) β-lactoglobulin A (fraction #6)...... 111 16

LIST OF SCHEMES

Page

Scheme 2-1. Apparatus of a new LC/DESI-MS post column splitting method...... 35

17

LIST OF ABBREVIATIONS

CE ...... counter electrode

CI......

CID ...... collision-induced

CLZ ...... clozapine

CRM ...... charge residue mode

CSD ...... charge state distribution

DESI ...... desorption electrospray ionization

EC ...... electrochemistry

ECD...... capture dissociation

EI ......

EIC ...... extracted ion chromatogram

EID ...... electron ionization dissociation

ESI...... electrospray ionization

ESSI ...... electrosonic spray ionization

ETD ...... dissociation

FAB ...... fast bombardment

FAPA ...... flowing atmospheric pressure afterglow

GC ......

HPLC ...... high performance liquid chromatography

IEF...... isoelectric focusing

IEM ...... ion mode 18

IR...... infrared

LC ...... liquid chromatography

LS-DESI ...... liquid sample desorption electrospray ionization

MALDI ...... matrix-assisted desorption ionization m-NBA ...... m-nitrobenzyl alcohol

MS ...... mass spectrometry

MS/MS ......

MT...... methoxytryamine m/z ...... mass-to-charge ratio nano-DESI...... nanospray desorption electrospray ionization

PESI ...... probe electrospray ionization

RE ...... reference electrode

RTIL ...... room temperature ionic liquid

S/N ...... signal-to-noise ratio

TIC ...... total ion chromatogram

TFA ......

UPLC ...... ultra performance liquid chromatography

UV ......

WE ......

19

CHAPTER 1: INTRODUCTION

1.1 Mass Spectrometry

Mass spectrometry (MS) is an analytical chemical technique that identifies chemical compounds based on their mass-to-charge ratios (m/z). Nowadays it has become a very powerful for both qualitative and quantitative analysis. In addition, tandem mass spectrometry (MS/MS) could be used to produce fragment ions for structural determination.1-2 With its capability of providing structural and analytical information, MS has been applied to the characterization of a variety of compounds, from ranging small organic molecules3 to large biological complexes.4 Coupling other chromatographic separation methods, such as gas chromatography (GC) and liquid chromatography (LC), with MS further extends its analytical applications for more complicated samples. Currently, MS has played an important role in many different analysis disciplines, including biology,5 medicine,6 ,7 environmental sciences,8 etc.

1.2 Mass Spectrometer

Figure 1-1 shows the basic components of a mass spectrometer including a sample inlet, an , ion optics, a mass analyzer, a detector, and vacuum system. A sample inlet could be connected with LC or GC for separation purpose. Alternatively, samples in various physical states (solid, liquid or gas state) could be directly introduced.

An ion source is used to convert analyte to gaseous ions for detection by mass spectrometer. Ion optics, connecting the ion source and the mass analyzer, focused and transferred the ions beams generated by ion source to the mass analyzer. A mass analyzer 20 is utilized to measure their m/z of the generated ions. A detector is used to detect the ions ejected by the mass analyzer and recorded the relative abundance of each of the resolved ionic species. Finally, a computer is required to control the instrument, acquire and manipulate data. Among these components, the ion source is one of the most crucial parts so that the development of novel ionization methods is a highly active research area.

Ion Data Inlet Source Ion Optics Analyzer Detector System

Vacuum System

Figure 1-1. Basic components of a mass spectrometer

1.3 Ionization Methods

Back to the early 20th Century, classical ionization methods such as electron ionization (EI)9 and chemical ionization (CI)10 were developed and used. However, these methods have their limitations. They are only suitable for volatile and non-polar analytes.

In addition, they all work in vacuum condition that makes the instrument operating and sample handling difficult. These limitations pushed the researchers to develop the new ionization techniques. With the development of modern ionization methods, such as atmospheric pressure chemical ionization (APCI),11 electrospray ionization (ESI),12 and matrix-assisted laser desorption ionization (MALDI),13 the scope of the ionizable analytes was extended. The MS analytes are no longer limited to volatile ; even larger protein molecule can be ionized. Moreover, the vacuum issue has been solved, the 21 ion source could work under atmospheric pressure. Although these atmospheric ionization methods expanded the MS application, the limitations, like the complicated process, still exist and encourage the further improvement of the ionization methods. Recently, the advent of ambient ionization methods allow the ionization of untreated samples in the open environment.14-15

1.3.1 ESI-MS

ESI was first introduced by John Fenn and Masamichi Yamashita in 1984,12 which brought MS great advance in various application areas. For this seminal work and contribution, John Fenn won the in Chemistry in 2002. As shown in Figure

1-2,16 in ESI, the sample solution is usually infused through a piece of capillary with high voltage applied (2~5 kV). With the assistance of nebulizing gas and heating, a mist of highly charged droplets will be formed at the end of the capillary. As the droplets travel toward the MS inlet, the solvent evaporates; the charged droplets will shrink into smaller one; finally the dried analytes ions are formed and enter into the MS for detection.

The ionization mechanism is still under debate: (i) the ion evaporation model (IEM) suggests that droplets shrink by evaporation until the field strength at their surface is sufficiently large to assist the of solvated ions; (ii) the charge residue model (CRM) assumes that electrospray droplets undergo evaporation and fission cycles, eventually leading progeny droplets that contain on average one analyte ion or less. In general, ESI is a powerful ionization technique that can ionize a wide range of analytes, from small organic compounds to large biomolecules. As a soft ionization method, ESI produce very little fragments during ionization and works well at atmospheric pressure. 22

Furthermore, ESI as an interface could be used to couple LC with MS based on its liquid sample analysis characteristics. In this case, ESI become the most commercially equipped ion source nowadays.

Figure 1-2. Scheme showing an ESI source.16 Reproduced with permission from

WILEY-VCH, Copyright 2001.

1.3.2 DESI-MS

A ambient ionization refers to generate ions directly under ambient condition with little or no sample preparation.14 In 2004 desorption electrospray ionization (DESI), one of the most well-known ambient ionization methods, was introduced by Cooks group, which regarded as a milestone in the development of ionization methods.17 As shown in

Figure 1-3, the DESI apparatus consists of a DESI sprayer, sample surface, and a mass spectrometer. The DESI sprayer is an electrospray emitter used to create charged 23 microdroplets under high voltage and high pressure nebulization gas. The electrically charged microdroplets beam is directly onto the solid sample surface to desorb and ionize sample molecule. The resulting ions travel through the atmospheric pressure interface into the mass spectrometer. With the capability of surface desorption with little or no sample preparation, detection, and high throughput analysis, numbers of applications have been explored with the development of DESI. In general, these can be classified into the following categories: (I) biological and clinical applications by the examination of biological surfaces, such as imaging of different tissues, even the intact bio-specimens, to help diagnosis; (II) high-throughput analysis in pharmaceutical and environmental applications; (III) the forensic security application by the examination of native surfaces.18-23

Figure 1-3. Scheme showing the apparatus of DESI.17 Reproduced with permission from

Science, Copyright 2004. 24

1.3.3 Liquid Sample DESI-MS

With the advent of DESI, many research groups have made much effort to improve and modify the DESI apparatus to improve the performance of DESI for different applications. In typical DESI experiment, liquid samples are died on surface before ionization. For the purpose of direct analysis a continuous-flowing liquid sample, our group developed the liquid sample DESI (LS-DESI) in 2008. The prototype apparatus for liquid sample DESI is shown in Figure 1-4.24 The charged microdroplets from DESI sprayer are also needed. Instead of analyzing a solid sample, the liquid sample is continuously infused through a piece of silica capillary onto a surface, the ionization occurs via the interaction of liquid sample with charged droplets generated by the DESI sprayer and the generated sample ions are transferred to the MS for detection. Later, the apparatus was improved by removal of the sample surface. The ionization still occurs when the charged microdroplets interacts the sample on the tip of the silica capillary.25-27

It has the capability of direct analysis of liquid sample in their native condition, not only for small molecules but also for large proteins and protein/protein complex. In addition, it is possible to couple liquid sample DESI-MS with LC,28-29 ,30 and electrochemistry (EC).31-35 25

Figure 1-4. Scheme showing the apparatus of liquid sample DESI.24 Reproduced with permission from Springer, Copyright 2009.

1.3.4 Probe Electrospray Ionization (PESI)

It was well recognized that operating at lower liquid flow rates improves the detection limits of the sample in electrospray. Since low-flow electrospray uses a capillary with a small bore diameter of a few μm, care must be taken to avoid capillary clogging and breaking. To circumvent these problems, the term PESI was introduced by

Hiraoka and his coworkers.36 Unlike ESI using capillaries for sample introduction, a sample is prepared in a droplet of solution on a surface and loaded by dipping the PESI probe, a conductive solid probe, into the droplet. Then the droplet of analyte solution is deposited on the PESI probe. When a high voltage is applied to the probe, the deposited droplet will deform to form a Taylor cone at the tip of the probe and then emits a spray 26

(Figure 1-5). Compared to ESI or nano-ESI, PESI has the advantages of high salt 37 and detergent 38 tolerance. In addition, it is rapid and consumes very little of the total sample solution.39 Several interesting applications of PESI-MS have been found, including imaging biological tissues,40 monitoring chemical reactions in solution,41 and analyzing living biological samples in real time.42-43 Very recently, single cell analysis with PESI for detection of metabolites at cellular and subcellular level has been reported.44

Figure 1-5. Scheme showing the apparatus of PESI.36 Reproduced with permission from

WILEY-VCH, Copyright 2007.

1.4 EC-MS

Both EC and MS play an important role in . EC coupled with

MS enable identifications of the products or intermediates of electrochemical reactions 45. 27

EC/MS is also used for the mechanistic studies of reactions of interest to biology 31,

46 and drug discovery 47-48. As shown in Figure 1-6, the development of the EC/MS coupling interface always follows after the progress of different ionization methods.49

The first combination of EC with MS has been introduced in 1971 for in-situ detection of volatile electrochemical oxidation products by Bruckenstein and Gadde.50 Later, with the increasing detection requirement of wide scale of sample, such as polar or even charged analytes that are not suitable for EI-MS, different ionization techniques were continually developed to combine EC with MS in need, including (TS),51 (FAB),52 ESI,53 and DESI,31 etc.

Figure 1-6. Timeline for the development of EC/MS with different ionization method.49

Reproduced with permission from RSC, Copyright 2013.

28

1.4.1 EC/ESI-MS

In particular, ESI has been widely used to couple EC with MS for various analytical applications because ESI could be used for the detection of both small compounds and large biomolecules. As the high voltage is applied to the capillary and the

MS inlet is grounded, ESI itself is an “electrochemical cell” and could induce electrochemical reactions. 54 The inherent EC which occurs in ESI source could help to enhance MS signals of species. The metal-containing analytes, such as metal porphyrins,55 and ferrocene-based derivatives,56 could undergo an oxidation process during the ionization to be converted to ionic species and to obtain good signals.

However, the first direct coupling of EC with ESI-MS was reported by Van

Berkel and his co-worker in 1995.53 In their study, three types of EC flow cells were discussed. The electrochemical cell should be floated or decoupled from the ESI high voltage.53 The utility of EC/ESI-MS was demonstrated by ionization of neutral compounds, analysis of EC reaction products and signal enhancement of metal complex detection.53 With the generation of various types of electrochemical cells to combine with

MS, more applications have been achieved, including monitoring electrochemical reaction intermediates,57 mimicking metabolic pathway of drug compounds,58 online chemical derivatization59 , and so on.

1.4.2 EC/DESI-MS

Although EC/ESI-MS has been developed and widely used for years, there are still some inherent problems of the coupling of EC with ESI-MS. First is voltage conflict.

High voltage is used for ESI to generate charged microdroplets. But for electrolysis, a 29 small voltage is needed for oxidation or reduction. In this case, the small voltage for EC should be floated decoupled with high voltage for ESI. So, the instrumentation is complex for EC/ESI-MS. The second problem is the compatibility of solvent and used. MS compatible and volatile such as , ammonium hydroxide, and ammonium acetate are used in EC/ESI-MS studies while it is difficult to employ nonvolatile salts, such as KCl, as an electrolyte due to the possible ion signal suppression effect of nonvolatile salts and clogging of the capillary.

Once liquid sample DESI is used as the interface for coupling EC with MS, the major problems mentioned above would be avoided. In EC/DESI-MS (Figure 1-7), sample solution is introduced by a syringe pump and flows through the EC cell. When either and oxidation or reduction potential is applied to the working electrode (WE), analytes will undergo electrolytic reaction. The reaction products will flow out via a short piece of silica capillary; be desorbed and ionized by charged droplets generated by DESI probe. In this case, the small potential applied to the electrochemical cell and the high voltage used for spray ionization are physically separated in EC/DESI-MS, there is no voltage conflict issue. Moreover, DESI appears to have tolerance for high salt concentration. Previous results show that a good MS signal can be obtained even when

10 mM NaCl is used as an electrolyte.28 Furthermore, EC/DESI-MS could also avoid the inherent EC reaction occurring in the ESI source, which can prevent the background signal from the oxidation of analytes.

With the advantage of combination EC with DESI-MS, several electrochemical reactions were investigated, including small molecule redox reaction60 and 30 protein/peptide reduction.32, 34 It could be used for the structural analysis of disulfide bonding-containing proteins in either top-down or bottom approaches.32, 34-35 Recently,

DESI-MS was also used to capture the fleeting electrochemical reaction intermediates on the electrode surface by using a rotating working electrode.61-62

Figure 1-7. Scheme showing the apparatus for online coupling of a thin-layer electrochemical flow cell with DESI-MS.28 Reproduced with permission from ACS,

Copyright 2009.

1.5 LC/DESI-MS

The combination of LC with MS has become one of the most powerful techniques for the analysis of biomolecules and pharmaceuticals, as it has the capability of LC separation and the power of mass analysis. Nowadays, ESI is the most commonly interface for LC/MS. However, there are limitations of LC/ESI-MS. First, the ESI

“friendly” solvent should be used as a mobile phase for ionization. Second, ESI could not tolerate high flow rate, which would cause a “flood” in the ion source. Normally, under 31 the high mobile phase flow rate, splitting is needed for LC/ESI-MS. Our has used DESI for coupling LC with MS which could increase the LC/MS performance.63

Figure 1-8 shows the prototype coupled LC/DESI-MS apparatus. A short piece of fused- silica capillary was connected to the outlet of the LC column. A liquid jet was formed and directly ionized by charged microdroplets produced by DESI probe. This is the first time for DESI to be used as an interface to analyze the LC eluent at the flow rate of 1.8 mL/min. In addition, reactive DESI could also be adopted for online derivatization. The derivatizing reagent can be doped into the spray solvent to react with the analyte during the ionization process, which could increase the signal of the analyte. Not only for small molecules, like LC-separated carbohydrates using N-methyl-4-pyridienboronic acid iodide as the derivatzing reagent,63 but also for large molecular, separated proteins using m-nitrobenzyl alcohol (m-NBA) or sulfolane as a supercharging reagent 29 could be derivatized. Integration of an EC with LC/ESI-MS is also possible, a small peptide, somatostatin-14, was successfully separated by LC and flowed through a thin-layer flow cell for electrochemical reduction prior to DESI-MS detection. However, LC/DESI-MS and LC/EC/DESI-MS still have potential to be modified to achieve better performance which is what I focused in my PhD study. Now, more and more interesting applications of liquid sample DESI and other ambient ionization method are studied in our laboratory. 32

Figure 1-8. Scheme showing the apparatus for LC/DESI-MS.63 Reproduced with permission from ACS, Copyright 2011.

33

CHAPTER 2: A NEW SPLITTING METHOD FOR BOTH ANALYTICAL AND

PREPARATIVE LC/MS

Adapted from Cai, Y. Adam, D. and Chen, H., J. Am. Soc. Mass Spectrom. 2014, 25,

286-292. Copyright 2014, Springer.

2.1 Introduction

Splitting the eluent in LC/MS experiment, in some cases, is needed. Electrospray ionization, a common ionization method for LC/MS, requires an optimal sample infusion rate at μL/min level. In addition, post-column splitting is also necessary when preparative purpose is required in the experiment. Typically, the post-column splitting of LC eluent can be achieved simply using a Tee splitter, in which one stream channel goes to MS for detection and the other one goes to the waste or other detector.64-65 However, the connection Tee splitting would introduce dead volume and backpressure which could cause the peak broadening or time delay for MS detection.66-67

DESI was originally developed by Cooks and co-workers,17 which has been introduced to provide direct ionization of analytes with little or no sample preparation.

Besides the analyses of solid samples on surface, in our and other , we extended the conventional DESI method for direct liquid sample analysis.24 Our results show that liquid DESI is a soft ionization method which can be used to ionize high-mass proteins/protein complexes in solution,68 24. Furthermore, it is possible to couple liquid

DESI-MS with electrochemical cells33, 69-70 or other separation devices such as micro- extraction.71 In the previous coupling of LC with MS by liquid DESI,29, 63 LC/DESI-MS was shown to apply to the analysis of both small organic molecules and large protein 34 molecules. So this LC/DESI-MS is different from LC/DART-MS as DART is typically limited to small molecule analysis. In these previous experiments,29, 63 LC eluent flowed out of the LC column as a free jet with a high flow rate, which was ionized via interaction with a pneumatically assisted DESI spray of a chosen solvent. Therefore, to some extent, it is hard to operate and easy to cause source “flooding”. Moreover, the collection of remaining analytes in the eluent after ionization would be difficult due to the of the jet sample into a plume during the DESI ionization process. The significance of the collection of fully separated compounds via LC separation (i.e., preparative liquid chromatography) is hard to overstate. As a very important protocol to obtain purified samples, it has been widely adopted in organic and biological research laboratories. Thus, a new splitting method with fast MS detection is in need to assist both online detection and online sample collection.

In this study, we present a new splitting interface for LC/MS application based on fast DESI ionization capability. In this approach, a PEEK capillary tube with a micro- drilled orifice on the side wall is used to connect with LC column outlet (Figure 2-1). In this case, a small aliquot of LC eluent emerges out of the orifice and can be directly sampled and ionized by DESI, while the remaining analytes can be collected from the tube outlet. There are several advantages of such a new splitting LC/DESI-MS method.

First, it minimizes the dead volume, because there is no need to use connection tubing to bridge the splitter and MS. In this case, the peak broadening can be avoided. Second, it is convenient to collect the purified analytes from the PEEK tube outlet for the preparative purpose. Third, reactive DESI72-73 can be used for online derivatization without 35 introducing an extra dead volume. Lastly, the application of this proposed splitting

LC/DESI-MS method can tolerate high mobile phase flow rates for shorter run times

DESI spray

N2 probe

LC column . . Sample .. PEEK tubing mixture

DESI MS spray ... 530 µm .. Orifice (i.d. 350 µm) 510 µm Outlet

PEEK tubing

Scheme 2-1. Apparatus of a new LC/DESI-MS post column splitting method.

2.2 Experimental

2.2.1 Chemicals

Insulin (from bovine pancreas, HPLC grade), ubiquitin (from bovine erythrocytes), trifluoroacetic acid (TFA), 3-nitrobenzyl alcohol (m-NBA,≥99.5%), N- acetyl-D-glucosamine (≥99%), maltohexaose, sulfadiazine (HPLC grade), sulfamerazine

(HPLC grade, ≥98.8%), and sulfaquinoxaline (HPLC grade, ≥96%) were purchased from Sigma-Aldrich (St. Louis, MO). Acetic acid was purchased from Fisher Chemicals

(Pittsburgh, PA). HPLC-grade methanol was purchased from Fisher Chemicals

(Pittsburgh, PA).HPLC-grade acetonitrile (ACN) was purchased from EMD Chemicals

Inc. (Billerica, MA). 36

2.2.2 LC Separation Condition

A commercial PerkinElmer HPLC system (Perkin Elmer, Shelton, CT) was used throughout the experiments. For the protein mixture separation, a XB Bridge TM

300 C4 column (4.6 mm×150 mm) was employed with a trifluoroacetic acid (TFA)- containing mobile phase of ACN:H2O:TFA (30:70:0.1 by volume) being used. The mobile phase flow rate was 1.0 mL/min. For the saccharide mixture separation, an

Agilent ZORBAX ODS C18 column (4.6 mm×250 mm) was used with 0.1% FA in H2O as the mobile phase at the flow rate of 1.0 mL/min. For the mixture of the sulfonamides, a monolithic C18 column (Phenomenex, Onyx, 4.6 mm ×100 mm) was adopted. A gradient program from 100% A down to 90% A in 4 min was used (mobile phase composition A: and B: ACN) with 4 mL/min flow rate. A 20 μL injection loop was used for sample loading.

2.2.3 DESI-MS Detection

A Thermo Finnigan LCQ DECA ion trap mass spectrometer was used throughout the experiments. The commercial ion source of the mass spectrometer was removed to accommodate a home-built liquid DESI ion source. As shown in Figure 2-1, a short piece of PEEK tube (i.d. 510 μm; wall thickness: 530 μm; length: 3 cm) with a micro-drilled orifice (i.d. 350 μm) was connected to the LC column. The orifice was located in the tube

2 cm downstream from the LC column and the tube outlet was slightly bent downward to facilitate sample collection. The sample eluent flowing out of the orifice underwent interactions with the charged microdroplets generated from DESI spray for ionization.

Unless specified, the spray solvent for DESI probe was CH3OH/H2O/HOAc (50:50:1 by 37 volume). The injection flow rate is 10 μL/min for DESI probe with 5 kV applied. The

DESI spray probe was placed above the orifice and the distance between the probe and orifice was about 1-2 mm. The orifice was placed approximately 2 cm away from the MS inlet.

2.3 Results and Discussion

2.3.1 Protein Detection

To demonstrate the feasibility of the proposed splitting LC/DESI-MS method

(Scheme 2-1), a protein mixture of insulin and ubiquitin was first chosen. For separation, a C4 column with a mobile phase of ACN:H2O:TFA (30:70:0.1 by volume) was employed. In peptide and small protein separation, TFA not only serves to adjust pH but also uses as an ion-pairing agent to promote protein separation. However, during the ionization process, TFA anions can form strong ion pairs with analytes that could neutralize or decrease the charges of the protonated analytes. In this case, TFA has been reported to suppress MS ionization.29 The flow rate is 1.0 mL/min. When the eluent from

LC column flows through the PEEK tube, a small portion of the eluent emerging from the

PEEK orifice was ionized by DESI probe with the assistance of high voltage and nebulizing gas. As shown in the extracted ion chromatograms (EICs, Figures 2-1a and 2-

1b), two proteins were well separated. The resulting DESI-MS spectra (Figures 2-1c and

2-1d) also clearly display the ionized individual proteins with multiple charge distribution.

In addition, with the help of DESI spray solvent, the TFA suppression effect was mitigated. It only takes 10 ms for the sample eluent to go through the micro-drilled orifice (calculation based on the flow rate of eluent through the orifice (300 μL/min) and 38 the orifice dead volume of 50 nL). Therefore, DESI-MS used in this experiment provides a “near-real time” monitoring of LC eluent-stream in the PEEK tube.

LC/DESI-MS

+4 100 a) EIC of insulin NL: 1.76×106

0 100 b) EIC of ubiquitin +6 NL: 2.69×106

0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 Time (min) +4 100 c) Insulin NL : 2.42×105 Relative Abundance Relative 50

+5 +3

0 600 800 1000 1200 1400 1600 1800 2000 +6 100 d) Ubiquitin +7 NL : 2.26×105

50 +8 +5 +9 +10 +11 +12 0 600 800 1000 1200 1400 1600 1800 2000 m/z

Figure 2-1. EICs of (a) insulin (+4 ion) and (b) ubiquitin (+6 ion) from LC/DESI-MS analysis and the corresponding DESI-MS spectra of (c) insulin and (d) ubiquitin.

For comparison purpose, electrosonic spray ionization74 (ESSI, a variant form of

ESI which employs an ESI source with supersonic nebulization gas) was also used to detect the LC separated analytes. In the experiment, the LC eluent flow rate was reduced 39 to an optimized flow rate at 10 μL/min using the ASI adjustable commercial splitter

(Analytical Scientific Instruments, Richmond, CA), because the high flow rate (1.0 mL/min) could cause the flood for the ESSI ion source. LC separation conditions were kept the same as mentioned above. As revealed by the resulting EIC spectra (Figures 2-2a and 2-2b) from ESSI-MS detection, the retention times for two proteins are both delayed by about 4 min in comparison to DESI detection. This delay is very likely to be caused by the increased dead volume from the traditional splitting method used in the LC/ESSI-MS experiment. Also, the two EIC peaks of the proteins are much wider than those recorded using LC/DESI-MS, each of which lasts about 4 minutes. In addition, the two protein’s

EIC profiles overlap with each other in the retention time window of 8-11 min. Moreover,

Because of the peak broadening and more severe TFA ion suppression effect, the EIC intensities from ESSI-MS detection (Figures 2-2a and 2-2b) are lower than those from

DESI-MS detection (Figures 2-1a and 2-1b). 40

LC/ESSI-MS 100 a) EIC of insulin +4 NL: 6.42×105

0 100 b) EIC of ubiquitin +6 NL: 1.65×106 Relative Abundance Relative

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Time (min)

Figure 2-2. EICs of (a) insulin (+4 ion) and (b) ubiquitin (+6 ion) from LC/ESSI-MS analysis.

Reactive DESI for online derivatization is an additional benefit in this LC/DESI-

MS method. Reactive DESI is performed simply by doping a chosen chemical reagent into the DESI sprays solvent for supercharging proteins and for enhancing their signals.

As the supercharging occurs during DESI ionization, the online supercharging would not introduce an extra dead volume. The charge enhancement for proteins by supercharging has great value for further structural analysis via top-down approaches for increasing the

75-76 fragmentation efficiency. In this study, DESI spray solvent of CH3OH:H2O:HOAc

(50:50:1 by volume) was doped with 50 mM m-NBA to effect “supercharging” proteins eluted from LC column. The m-NBA, an effective supercharging reagent,28 was used in 41 the experiment. As m-NBA was doped in DESI spray solvent, the maximum charge state of insulin shifted from +5 to +6 and the charge with the highest abundance shifted from

+4 to +5 (Figure 2-3c) compared to the regular DESI data (Figure 2-1c). Likewise, the maximum charge state of ubiquitin shifted from +12 to +13 and the charge with the highest abundance shifted from +6 to +7 (Figure 2-3d). In addition, compared LC/DESI-

MS (Figures 2-1a and 2-1b) and ESSI ionization (Figures 2-2a and 2-2b), the intensity of resulting protein ions in this LC/reactive DESI-MS (Figures 2-3a and 2-3b) are significantly enhanced. Presumably, m-NBA reduces TFA dissociation77 so that the decreased concentration of trifluoroacetate anions reduces their ion pairing interaction with protein cationic sites, a mechanism responsible for signal suppression of TFA. 42

LC/Reactive DESI-MS, with m-NBA

+5 100 a) EIC of insulin NL: 2.83×107

0 100 b) EIC of ubiquitin +7 NL: 2.02×107

0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 Time (min) +5 100 c) Insulin NL : 2.15×106 +4

50

+6 +3

0 600 800 1000 1200 1400 1600 1800 2000

d) Ubiquitin +7 100 6 NL : 1.30×10 +8 +9 +6

50 +10 +11 +12 +13 +5 0 600 800 1000 1200 1400 1600 1800 2000 m/z

Figure 2-3. EICs of (a) insulin (+5 ion) and (b) ubiquitin (+7 ion) from LC/reactive

DESI-MS analysis and the corresponding reactive DESI-MS spectra of (c) insulin and (d) ubiquitin.

Besides the analytical strength, another important feature of this method is to collect isolated samples following LC/MS analysis. Indeed, with the aid of online DESI monitoring as mentioned above, the major portion of insulin and ubiquitin exiting from the PEEK tube outlet (70%) were collected. The collected samples were re-analyzed by 43

MS (directly infusion) and gave rise to the spectra of isolated insulin and ubiquitin. As can be seen in Figures 2-4a and 2-4b, there is no cross-contamination. The results confirm the two samples were completely separated and successfully collected.

The eluent splitting ratio provided by the PEEK tube through the orifice can be adjusted simply by reducing the orifice i.d.. When the orifice i.d. decreased from 350 μm down to 100 μm, the splitting ratio of 3:7 can be reduced to 4:96 for which more sample can be collected. An experiment was conducted to evaluate the performance of the PEEK tube with the 100 μm i.d. orifice. For this case, the time for the eluent to flow through the orifice channel was 6 ms (the split flow rate: 40 μL/min; orifice dead volume: about 4 nL). By using a mixture of insulin and ubiquitin as a test sample, the EIC spectra acquired show good protein separation (Figures 2-4a and 2-4b). In comparison to Figures

2-1a and 2-1b, the signals of both proteins are even higher. Also, the proteins after LC separation were collected and re-analysis of the collected samples gave pure spectra

(Figures 2-4c and 2-4d). We further quantified the yield of protein collection using UV (PERKIN ELMER 785A UV/VIS Detector). The UV detection wavelength was selected at 220 nm. Insulin and ubiquitin standard stock were diluted into the appropriate concentration ranges (0.625-10.5 μM for insulin and 0.55-8.8 μM for ubiquitin) using the mobile phase solvent to establish the calibration curves (Figure 2-5).

The peak areas of collected insulin and ubiquitin solution were used to calibrate the of the collected protein solutions. Then, based on the collected volume, the moles of collected samples can be obtained to compare with the original amount of injected proteins for LC separation (2 nmol) to get the protein collection yields, which 44 was 94.2±0.8% average yield for insulin and 94.6±1.9% average yield for ubiquitin.

These yields are fairly close to the theoretical value of 96% based on the splitting ratio, suggestion the feasibility and potential of this approach for sample purification.

45

+4 100 c) Collected insulin 100 μM i.d. orifice +3 LC/DESI-MS, 100 μm i.d. orifice 50

100 a) EIC of insulin NL: 5.11×106 +5 0 800 1000 1200 1400 1600 1800 2000

0 +6 100 100 b) EIC of ubiqutin d) Collected ubiquitin NL: 3.34×106 100 μM i.d. orifice +7 Relative Abundance Relative +8 +5 50 +9 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 Time (min) +10 +11 +12 0 800 1000 1200 1400 1600 1800 2000 m/z Figure 2-4. EICs of (a) insulin (+4 ion) and (b) ubiquitin (+6 ion) from LC/DESI-MS analysis (100 μm i.d. orifice) and the collected (c) insulin and (d) collected ubiquitin.

46

800000

700000 y = 54901x + 93988 R² = 0.9992 600000

500000

400000

Peak Area 300000

200000

100000

0 0 2 4 6 8 10 12 Insulin Concentration (μM)

700000 y = 62847x + 67220 600000 R² = 0.9993

500000

400000 Peak Area

300000

200000

100000

0 0 1 2 3 4 5 6 7 8 9 10 Ubiquitin Concentration (μM)

Figure 2-5. Calibration curves of insulin and ubiquitin (UV absorption peak area wrt. protein concentration).

2.3.2 Saccharide Detection

Conventionally, in the preparative LC experiment, the collection of LC-separated eluent is often enabled by placing a UV detector in between the LC separation column and the sample collection reservoir. However, such an approach is limited to compounds with chromophores. For those without chromophores, derivatization is in need, which is time consuming and troublesome. By using online DESI monitoring in our method, this problem can be solved as MS is a detector. 47

As a demonstration, a saccharide mixture consisting of N-acetyl-D-glucosamine

(NAG) and maltohexaose (no or weak UV absorption) was chosen to test feasibility of the LC/DESI-MS. A C18 column was used for separation. As shown in the recorded EIC spectra, NAG and maltohexaose were well separated (Figures 2-6a and 2-6b). Although there is no organic solvent used in the mobile phase (only 0.1% FA in H2O was used as the mobile phase), the saccharides can still be detected by DESI-MS because of the freedom of DESI probe to choose favorable spray solvent of CH3OH/H2O/FA (50:50:1 by volume) for sample ionization. Again, in contrast, when ESSI was used for detection

(in conjunction with the employment of the commercial splitter to obtain an optimized flow rate of 20 μL/min for ionization), wider peaks with lower intensities were resulted.

As shown in Figures 2c and 2d, the peaking broadening occurred and both of the saccharide compounds have a peak width of 2~3 min. As a result, the EIC profiles of

NAG and maltohexaose overlap with each other (Figures 2-6c and 2-6d) and the ESSI-

MS spectra recorded also show the presence of both NAG and maltohexaose (data not shown). Lower intensities are also observed in the ESSI-MS detection (Figures 2-6c and

2-6d) in comparison to DESI-MS detection (Figures 2-6a and 2-6b). Two factors might contribute to the low intensities in the LC/ESSI-MS experiment; one is due to peak broadening as mentioned above and the other one might be owing to the low ionization efficiency of analytes in aqueous mobile phase without organic solvent (0.1% FA in H2O in this case). By contrast, in DESI-MS detection, the spray solvent

(CH3OH/H2O/HOAc=50:50:1 by volume) yielded favorable for ionization. This phenomenon is also in agreement with previous reports.29 Following the LC/DESI-MS 48 analysis, the remaining portion of saccharides flowing out of the PEEK tube outlet was collected and re-tested with ESSI, which show clearly the full separation of two saccharide compounds (Figures 2-6e and 2-6f).

49

[NAG+H]+ [NAG+H]+ [2NAG+H]+ 100 a) LC/DESI-MS 100 e) Collected N-acetyl-D-glucosamine

EIC of N-acetyl-D-glucosamine NL: 2.70×106

0 + 50 [Maltohexaose+H] + 100 b) LC/DESI-MS [NAG-H2O+H]

EIC of maltohexaose NL: 1.75×106 [3NAG+H]+ [4NAG+H]+

0 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 200 300 400 500 600 700 800 900 1000

100 + [NAG+H]+ f) Collected maltohexaose [Maltohexaose+H] 100 c) LC/ESSI-MS

Relative Abundance Relative EIC of N-Acetyl-D-glucosamine NL: 2.53×105

50 + 0 [Maltohexaose+H] 100 d) LC/ESSI-MS [Maltohexaose+Na]+ EIC of Maltohexaose 5 NL: 7.16×10 223

0 0 200 300 400 500 600 700 800 900 1000 0 1 2 3 4 5 6 7 8 Time (min) m/z

Figure 2-6. EICs of (a) N-acetyl-D-glucosamine and (b) maltohexaose eluted out of the

C18 column and detected by DESI-MS. EICs of (c) N-acetyl-D-glucosamine and (d) maltohexaose eluted out of the C18 column and detected by ESSI-MS. MS spectra of (e) the collected N-acetyl-D-glucosamine and (f) the collected maltohexaose. The re-analysis of the collected proteins was performed using ESSI-MS. The peak at m/z 223 in (f) came from solvent background

2.3.3 Sulfonamides Detection

Besides the applications to LC separation using regular analytical columns, we also examined the compatibility of our splitting method to monolithic column-based ultra-fast LC separation. The monolithic column is a kind of “single-piece column” that has been developed to tolerate fast high-throughput analysis.78 It allows performing high- 50 throughput and maintaining good separation performance close to that of ultra-high performance LC. The key factor in such an LC experiment is that an extremely high flow rate (up to 9 mL/min) can be used with monolithic columns without causing significantly high back pressure. Thus the separation can be completed in a short period of time. In our experiment, a mixture of sulfonamides including sulfadiazine, sulfamerazine, and sulfaquinoxaline, drugs commonly used for eliminating bacteria and treating urinary tract , was chosen as a test sample for demonstration of the application of our splitting method to the monolithic column-based ultra-fast LC separation. As shown in

EIC spectrum recorded by DESI-MS (Figure 2-7a), the sulfadiazine, sulfamerazine, and sulfaquinoxaline were all separated and eluted within 3 min. While, when a regular reversed-phase C18 column was used at 1 mL/min flow rate separation, the total elution time was 11 min (data not shown).This result shows that the high flow rate does help fast elution and separation. This result shows that the high flow rate does help fast elution and separation. We also recorded the chromatogram using an UV detector (detection wavelength: 254 nm) for comparison (Figure 2-7b). The DESI-MS recorded chromatogram has slightly better resolution than that recorded by UV detector. In figure

2-6b, there is about 10% overlap of the sulfadiazine and sulfamerazine peaks in the UV chromatogram, while the two peaks are well resolved in the DESI-MS chromatogram.

However, attempt to split the high flow eluent of 4 mL/min down to 10 μL/min for ESSI-

MS detection by the commercial splitter caused column failure due to high backpressure that exceeded the tolerance of the monolithic column used. These results prove that in our 51 split method in conjunction with DESI-MS detection, there is small dead volume and negligible backpressure.

[Sulfamerazine + H]+ 100 a) LC/DESI-MS [Sulfaquinoxaline + H]+ [Sulfadiazine + H]+

0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 b) LC/UV Sulfamerazine Sulfadiazine Sulfaquinoxaline 100 Relative Abundance Relative

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Time (min)

Figure 2-7. EIC spectra of a mixture of sulfadiazine, sulfamerazine, and sulfaquinoxaline

(a) from LC/DESI-MS analysis and (b) from LC/UV analysis.

2.4 Conclusions

This study suggests a new and versatile splitting method for LC/MS coupling, as an addition example of combining LC with MS using ambient ionization methods. The

LC/DESI-MS method via PEEK tubing with a micro-drilled orifice has the advantages including premium analytical performance of DESI and successful sample purification with high collection yield. The presented methodology has striking features involving narrow peak width, high sensitivity, no back pressure issue and low cost. The freedom to 52 choose DESI spray solvents allows online supercharging separated proteins and for enhancing their signals. Owing to DESI appears to be compatible with both small organic molecules and protein/peptide, this LC/DESI-MS method would have wide applications in bioanalysis.

53

CHAPTER 3: COUPLING OF ULTRAFAST LC WITH MASS SPECTROMETRY BY

DESI

Adapted from Cai, Y. Liu, Y. Helmy, R. and Chen, H., J. Am. Soc. Mass Spectrom. 2014,

25, 1820-1823. Copyright 2014, Springer.

3.1 Introduction

LC/MS is powerful for analysis of complicated samples and there is a growing demand for fast separation to reduce cost and time. To improve the separation speed and efficiency, several innovative strategies have been developed, including monolithic or fused-core columns, high-temperature LC and ultra-performance liquid chromatography

(UPLC).79 Among these innovations, monolithic LC employing a stationary phase formed from a monolithic porous rod is one of the breakthroughs in the column design.

The porous rod structure provides high permeability, numerous channels, and a high surface area for interacting with analytes.80 Therefore the column can have a low hydraulic resistance to allow very high elution flow rate (e.g., 9 mL/min) for ultrafast separation (in few minutes) or high-throughput analysis. In addition, the column is tolerant with untreated raw or viscous samples because of its porous structure. In the case of UPLC, the packing particle size is decreased to less than 2.0 µm, therefore, the analysis time can be reduced and chromatographic resolution can be enhanced.81

We previously presented a DESI interface to combine LC with MS using a new

LC eluent splitting strategy via a tiny orifice drilled on a capillary tube.82 In the experiment, a small portion of LC eluent leaking out of the capillary orifice can be instantaneously desorbed and ionized by DESI while the remaining majority of LC eluent 54 can be collected for preparative purpose. Splitting the eluent in LC/MS experiments is also necessary under the circumstance when the mobile phase flow rate is too high for direct MS ionization. In contrast to using tradition splitter, splitting via a capillary orifice introduces negligible dead volume and back pressure, thus enabling nearly real-time

DESI-MS detection and online MS-directed collection of LC-separation species.82

This study focuses on further investigating the performance of our DESI-MS, in conjunction with the capillary orifice splitting technique, for coupling with ultrafast LC based on either monolithic column or UPLC-based separation. The demonstrated examples include LC/DESI-MS analysis of metabolites in human urine, drugs-of-abuse in a cola drink, acidic anti-inflammatory drug mixture and insulin peptic digest. Several advantages of our LC/DESI-MS in this study have been revealed. First, high elution flow rate used in the monolithic column LC experiments (3 mL/min) led to a very short total separation time (1.6~3.5 min). Second, DESI-MS-directed purification allowed us to online collect LC-separated metabolites for further structural elucidation by NMR. Third, with reducing the capillary orifice to 50 µm, the splitting ratio was lowered down to ca.

1:99 (i.e., only 1% sample was used for DESI-MS detection), offering “nearly non- destructive” MS detection with simultaneous online collection of 99% purified analytes.

Fourth, we demonstrated a novel way to carry out “wrong-way around” ionization for

LC/MS. Acidic compounds that were separated using acidic mobile phase solvent could be directly ionized into negative ions by DESI using basic spray solvent. It is advantageous as acidic compounds generally have high ionization efficiency in the negative ion spray ionization mode rather than in the positive ion mode. In addition, for 55 the first time, DESI was shown to be applicable for combining MS with UPLC. For the analysis of drugs-of-abuse in cola drink, the total was completed within 2 min. Comparable sensitivity was obtained with DESI interface in comparison with using commercial Waters ESI.

3.2 Experimental

3.2.1 LC/DESI-MS Instrument

The schematic of the home-built liquid DESI ion source and the PEEK capillary tube with a drilled micro-orifice is shown in Scheme 2-1 and details were mentioned above. Briefly, a short piece of PEEK capillary tube (i.d. 510 μm; wall thickness: 530 μm; length: 2.5-3 cm) with a micro-orifice (i.d. 100 μm) drilled through the tube wall 0.5~1 cm upstream to the tube outlet was used for LC eluent splitting. In the case of monolithic column separation, the PEEK tube was directly connected to the LC column. A commercial PerkinElmer HPLC system (Perkin Elmer, Shelton, CT) was adopted for LC separation. MS and MS/MS analysis were performed employing a Thermo Finnigan LCQ

DECA-MAX ion trap mass spectrometer (San Jose, CA).

In the case of UPLC separation, a Waters ACQUITY UPLC® System was used and MS analysis was performed using a Waters Xevo QTOF (Milford, MA). A piece of extension connection capillary (i.d. 25.4 μm; length: 30 cm), originally used for connecting UPLC column with ESI source was used to bridge the PEEK capillary tube with UPLC column due to the distance between UPLC column and the MS inlet. Unless otherwise specified, the spray solvent for DESI was CH3OH/H2O/HOAc (50:50:1 by volume) for the positive ion mode and CH3OH/H2O/NH4OH (50:50:2 by volume) for the 56 negative ion mode. The spray solvent for DESI was injected at 10 μL/min with + 5 kV applied for the positive ion mode or -5 kV applied for the negative ion mode.

3.2.2 LC Separation Conditions

In all monolithic column LC experiments, the mobile phase elution flow rate was kept at 3.0 mL/min and a 20 μL injection loop was used for sample loading. Under such a flow rate, the splitting ratio was measured to be 4:96 when the orifice i.d. was 100 μm i.d.

Specifically, Phenomenex Onyx Monolithic C18 column (4.6×100 mm) was employed for the separation of dopamine, 3-methoxytyramine, L-tryptophan and L-kynurenine in urine.

The urine sample doped with these compounds was diluted 1:1 (v/v) with water to have the final concentration of 60 μM for each species, and then filtered with 0.2 μm Nylon

Membrane for removing possible particulates before LC injection. Solvent A was 0.1%

FA in H2O, and solvent B was 0.1% FA in ACN with elution program as such: 0-1min, 1%

B; 1-3min, 1% B was ramped to 50%. For the separation of acidic anti-inflammatory drugs, isocratic mobile phase of H2O: MeOH: FA (30:70:0.05% by volume) was used and the Phenomenex Onyx Monolithic C18 column (100×4.6 mm) was also chosen.

® For the UPLC experiments, ACQUITY UPLC BEH C18 column (50×2.1 mm) was employed for the separation. For analysis of drugs in Pepsi, the Pepsi sample contained codeine, cocaine, and flunitrazepam (diluted using water to 15 μM). The mobile phase consisted of A: 0.1% FA in H2O and B: 0.1% FA in ACN. A linear gradient program ran from 23% to 90% solvent B in 3 min. With the mobile phase elution flow rate kept at 0.3 mL/min, the splitting ratio using the PEEK capillary tube carrying the 100

μm i.d. orifice was 1:1. 57

3.3 Results and Discussion

3.3.1 LC/DESI-MS Analysis of Drugs

First we showed the capability of our LC/DESI-MS for direct analysis of neurotransmitters, amino acids and their metabolites in biological matrices with minimum sample clean-up. In this experiment, dopamine, 3-methoxytyramine, L- tryptophan and L-kynurenine were chosen as test samples (note that 3-methoxytyramine and L-kynurenine are the metabolites of neurotransmitter dopamine and L- tryptophan, respectively) and dissolved in human urine. As shown in the extracted ion chromatograms (EICs, Figures 3-1), these compounds were well separated in less than

1.6 min. In the resulting DESI-MS spectra shown in Figure 1 insets, all protonated molecules are clearly observed. No interference from urine was noted, probably because salts in urine matrix would elute out without retention. Also, a fragment ion by loss of -

NH3 group from the protonated ion molecule was observed in both dopamine and 3- methoxytyramine DESI-MS spectra (Figure 3-1a and 3-1b insets). 58

100 154 a) EIC of dopamine in urine 100 [ +H ]+ NL: 1.71×106 50 [154-NH3]

0 150 200 250 m/z 0 100 b) EIC of 3-methoxytyramine in urine 6 168 [ +H ]+ NL: 7.11×10 100

[168-NH3] 50

0 150 200 250 0 m/z 100 c) EIC of L-Kynurenine urine 209 100 NL: 1.47×106 [ +H ]+

50

Relative Abundance Relative 0 150 200 250 m/z 0 d) EIC of L-tryptophan in urine 205 100 100 NL: 8.70×105 [ +H ]+ 50

0 150 m/z 200 250 0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 Time (min) Figure 3-1. EICs of (a) dopamine, (b) 3-methoxytyramine, (c) L-kynurenine, and (d) L- tryptophan acquired by LC/DESI-MS.

In comparison, the commercial ESI source was used. The high flow of LC eluent of 3.0 mL/min that would “flood” the ESI ion source, splitting was in need. A home-built

Tee splitter was used to reduce the flow rate to an optimal value of 100 μL/min for the

ESI detection without causing too much back pressure. ESI-MS data was acquired with other experimental conditions kept the same as those used in LC/DESI-MS analysis mentioned above. As revealed by the resulting EIC spectra (Figures 3-2), the signals of the four compounds recorded by ESI are 20~40 fold lower than those detected by DESI.

It is probably caused by the high proportion (ca. 85~95%) of water in the mobile phase which disfavored ion desolvation. The regular spray solvent for DESI could help to 59 increase the analyte ionization efficiency under such a circumstance. The sensitivity of the LC/DESI-MS was good, with using the LCQ DECA MAX mass spectrometer in the experiment. The evaluation of the LC/DESI-MS sensitivity was performed using single ion monitoring (SIM) mode. With the injection concentration of 250-620 ng/mL

(corresponding to ca. 0.2-0.5 ng sample injected into the DESI ion source), a signal to noise ratio (S/N) of 5-14 was obtained, which are similar to the reported results by

Schiavo S. et al.

LC/ESI-MS

100 a) EIC of dopamine NL: 2.40×104

0 100 b) EIC of 3-methoxytyramine NL: 1.77×105

0 100 c) EIC of L-kynurenine NL: 4.49×104

0 100 Relative Abundance Relative d) EIC of L-tryptophan NL: 3.53×104

0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 Time (min) Figure 3-2. EICs of (a) dopamine, (b) 3-methoxytryamine, (c) L-kynurenine, and (d) L- tryptophan acquired by LC/ESI-MS.

In order to enhance the MS-directed sample collection yield, it would be beneficious to minimize the sample consumed for MS detection. This can be achieved by 60 reducing the orifice i.d.. In the past, the splitting ratio was adjusted from 30:70 to 4:96 by reducing the orifice i.d. from 350 μm to 100 μm.82 In this study, with further reducing the orifice i.d. from 100 μm down to 50 μm, the splitting ratio of 0.9:99.1 (approximately

1:99) was obtained. This ratio would favor improving the sample collection yield up to

99% with only 1% samples being consumed for DESI-MS (i.e., “nearly non-destructive”

MS detection). The experiment was conducted to testify this hypothesis and the mixture of dopamine, 3-methoxytyramine, L-tryptophan and L-kynurenine wer again chosen as the test sample. The four samples were well separated by the C18 monolithic column and successfully detected online by DESI-MS using the smaller orifice (50 μm i.d., data not shown). During the run of LC/DESI-MS, we collected 3-methoxytyramine from the

PEEK capillary tube outlet and quantified the collection yield using UV absorption spectroscopy. LC with a UV detector (detection wavelength was selected at 210 nm) was used for the quantification of the DESI-MS-directed collection yield. 3-Methoxytyramine

(3-MT) standard stock solutions were diluted into the standard solutions with five different concentrations (1, 2, 5, 10, and 20 μM), which were analyzed three times by

LC/UV using the same LC separation condition described previously for monolithic C18 column procedure to establish the (Figure 3-3). The collected 3-MT solution was also injected for LC/UV analysis in triplicate measurements. The peak areas of collected 3-MT solution were brought into the regression equation of the calibration curve to get the concentrations of the collected 3-MT solutions. Then, based on the collected volume, the moles of collected samples can be obtained to compare with the original amount of injected 3-MT for LC separation to obtain the collection yield of 98.6% 61

± 0.6 on average. This collection yield is fairly close to the actual splitting ratio value of

99.1%.

100000 90000 y = 4381.8x - 2146.8 80000 R² = 0.9963 70000 60000 50000 40000 Peak Area 30000 20000 10000 0 0 5 10 15 20 25 3-MT Concentration (μM)

Figure 3-3. Calibration curves of 3-MT (UV absorption peak area vs. 3-MT concentration)

3.3.2 LC/DESI-MS analysis of Acidic Compounds

Acidic mobile phases are often used in LC. However, the acidic mobile phase represents a dilemma for ionization of acidic analytes by ESI, as acidic analytes tend to form negative ions rather than positive ions. Typically, the eluent needs to be adjusted to alkaline pH using a Tee junction to introduce base which would to increased post- column dead volume. In contrast, DESI-MS can provide a simpler approach to tackling this challenge, simply by spraying basic solvent to achieve the so-called “wrong-way around” ionization to generate negative ions. In this study, we investigated LC/DESI-MS analysis of acidic anti-inflammatory drug mixture consisting of ketoprofen, fenoprofen, and ibuprofen with LC separation using traditional acidic mobile phase and online DESI- 62

MS detection in the negative ion mode. As shown in EIC spectra recorded by DESI-MS employing a spray solvent of MeOH:H2O:NH4OH (50:50:2 by volume, Figures 3-4), the ketoprofen, fenoprofen, and ibuprofen were all separated and eluted within 2.5 min by isocratic elution using mobile phase of H2O:MeOH:FA (30:70:0.05 by volume). In comparison, negative ion mode ESI-MS was directly used for the detection of the separated compounds under the same separation condition except LC eluent flow rate was reduced using a home-built Tee splitter to an optimal rate of 100 μL/min. The signals of ketoprofen and fenoprofen were weaker than those of DESI-MS detection and ibuprofen could not be detected (Figure 3-4), probably because the formic acid present in the mobile phase suppressed the deprotonation of the analytes in negative mode ESI ionization process. 63

LC/DESI-MS LC/ESI-MS 100 a) EIC of Ketoprofen 100 NL: 5.30×105 a’) EIC of Ketoprofen NL: 2.03×105

0 0 100 b) EIC of Fenoprofen 100 NL: 3.64×105 b’) EIC of Fenoprofen NL: 1.25×105

0 0 100 c) EIC of Ibuprofen 100 5 Relative Abundance Relative NL: 1.15×10 c’) EIC of Ibuprofen

0 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Time (min) Time (min)

Figure 3-4. EICs of (a) ketoprofen, (b) fenoprofen, and (c) ibuprofen acquired by

LC/negative ion DESI-MS. EICs of (a') ketoprofen, (b') fenoprofen, and (c') ibuprofen acquired by LC/negative ion ESI-MS.

3.3.3 UPLC/DESI-MS Analysis

Besides monolithic column LC, DESI-MS is also applicable to UPLC applications. For the analysis of a drug mixture of cocaine, codeine, and flunitrazepam in cola, the three drugs were well separated in less than 2 min using UPLC (Figure 3-5). The

CID data were also obtained during LC separation, for ion structural confirmation. CID of the protonated codeine at m/z 300 (Figure 3-5a inset) yielded the fragment ions m/z

243 and 215 by consecutive losses of C3H5O and CO. Direct loss of one molecule of H2O from m/z 300 produced fragment ion at m/z 282. For cocaine, fragment ion of m/z 182 was generated by CID of the protonated cocaine at m/z 304 via the loss of PhCOOH

(Figure 3-5b inset). In the flunitrazepam MS/MS spectrum (Figure 3-5c inset), CID of the 64 protonated flunitrazepam ion at m/z 314 yielded the fragment ions of m/z 268 and 240 by consecutive losses of NO2 and CO, in agreement with our previous observation. Direct loss of CO due to ring contraction was also noted to form a fragment ion of m/z 286. This result shows the strength of UPLC as it could take over 10 min for the separation using traditional HPLC.83

300 m/z 300 100 a) EIC of Codeine in Cola × 3 NL: 5.10 10 215 -C3H5O - CO 243 282

-H2O

200 220 240 260 280 300 320 340 0 m/z 304 182 100 b) EIC of Cocaine in Cola NL: 7.34×103

- PhCOOH 304

160 200 240 280 320 0 Relative Abundance Relative m/z 314 268 100 c) EIC of Flunitrazepam in Cola 4 314 NL: 2.63×10 - NO2

286 - CO

240 - CO

200 220 240 260 280 300 320 340 0 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 Time (min) Figure 3-5. EICs of a) codeine, b) cocaine and c) flunitrazepam acquired by

UPLC/DESI-MS with MS/MS spectra are given in the figure insets. The red lines shown in the molecule structures indicate the bond cleavage upon CID.

When the temperature of the column was elevated to 80 oC, the elution flow rate for UPLC separation could be increased to 1.0 mL/min without causing backpressure 65 beyond the system limit. Under this condition, the LC separation was completed in 45 s, and the analytes were all well detected by DESI-MS (Figure 3-6 a-c). In addition, the separated drugs were also successfully online collected with the aid of DESI-MS detection (Figure 3-6 d-f).

66

HTUPLC/DESI-MS 100 a) EIC of Codeine in Cola NL: 2.37×104

0 100 b) EIC of Cocaine in Cola NL: 4.82×103

0 100 c) EIC of Flunitrazepam in Cola

Relative Abundance Relative NL: 3.69×103

0 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 Time (min) 300 100 d) Re-analysis of codeine

0 304 100 e) Re-analysis of Cocaine

0 314 100

Relative Abundance Relative f) Re-analysis of Flunitrazepam

0 295 300 305 310 315 320 Figure 3-6. EICs of a) codeine, b) cocaine and c) flunitrazepam acquired by high temperature UPLC/DESI-MS (80 oC column temperature). Re-analysis of collected d) codeine, e) cocaine and f) flunitrazepam after high-temperature UPLC separation.

3.4 Conclusions

This study demonstrates the combination of DESI-MS with ultrafast LC separation using monolithic and UPLC columns. The splitting via a capillary orifice allows fast LC elution for rapid separation and online MS-directed purification with recovery yield up to 99%. The DESI also allows direct and efficient ionization of acidic 67 compounds in the acidic eluent. These strengths of DESI-MS would lead to many useful bioanalytical applications.

68

CHAPTER 4: INTEGRATION OF ELECTROCHEMISTRY WITH ULTRA

PERFORMANCE LIQUID CHROMATOGRAPHY/MASS SPECTROMETRY

(UPLC/MS)

Adapted from Cai, Y.; Zheng, Q.; Liu, Y. Helmy,; R. Loo. J,; and Chen, H., Eur J Mass

Spectrom. 2015, 21, 341–351. Copyright 2015, IM Publications.

4.1 Introduction

Redox-active disulfide bonds are one of the most common protein post- translational modifications. The disulfide bond can maintain three-dimensional protein structures and their biological activities.33 However, the presence of disulfide bone linkages increases the complexity protein structure analysis determined by MS. EC has been introduced as a new protocol to couple with DESI/MS for the structural analysis of disulfide bond containing proteins/peptides in either top-down or bottom-up. However, in such a study, the reduced linear peptide chains are hard to be assigned to their precursors and sequenced. The EC/DESI-MS method would be further benefited if UPLC could be coupled for fast separation.

The combination with LC further broadens EC/MS applications. Previous studies showed that post-column electrochemical conversion in LC/MS could be used to increase the MS detection sensitivity of the target compounds. In such an experiment, an electrochemical flow cell is placed in between the LC and MS to convert the LC- separated compounds to more polar or even charged products, which are suitable for MS detection with increased ionization efficiency.84-86 But no disulfide bond reduction was studied using LC/EC/MS system. In addition, the mobile phase flow rate should be 69 considered to be compatible with the use of the electrochemical cell.86-87 In this case,

UPLC is a good choice to for this combination. UPLC does not require a high elution flow rate to achieve fast separation, thus making it a better fit for LC/EC/MS experiments than HPLC. It also can significantly shorten the separation time by 10 times in comparison to conventional HPLC. However, UPLC has not been reported to couple with

EC yet.

In this study, we developed the LC/EC/MS method using UPLC for the first time.

We adopted DESI as the interface to couple the electrochemical cell with a mass spectrometer. There are several advantages of using DESI as the interface. First, the conflict between the electrochemical redox potential and the high voltage for spray ionization is avoided using DESI as the interface to combine EC and MS, since the DESI high voltage is separated from the cell potential. Second, the cell and DESI source can be connected with a very short piece of the capillary as a conduit, thus minimizing the post- column dead volume in this LC/EC/MS apparatus. Third, taking advantage of the freedom to choose different DESI spray solvents, reactive DESI can be directly performed in the coupling of UPLC/EC/DESI-MS for post-column derivatization. Using the new apparatus, we demonstrated the application of LC/EC/MS in protein/peptide structural analysis. The disulfide bond-containing peptides can be differentiated from those without disulfide bonds because the disulfide bond-containing peptides are electroactive and electrochemically reducible. Moreover, the tandem MS analysis of the reduced peptide ions can provide more information for sequencing and pinpoint the disulfide bond linkages. In reactive DESI-MS experiment using the spray solvent doped 70 with supercharging reagents, online electrolytic reduction of disulfide-bond containing proteins in combination with supercharging to higher protein charges. In this study, we focused on its application for the structural analysis of disulfide bond-containing proteins/peptides, taking advantage of the platform’s capability for fast separation, online electroreduction and MS detection.

4.2 Experimental

4.2.1 Chemicals

Somatostatin 1-14 was purchased from American Peptide Company (Sunnyvale,

CA). TPCK-treated trypsin from bovine pancreas, pepsin from porcine gastric mucosa, insulin from the bovine pancreas, α-lactalbumin from bovine (type III, calcium depleted), formic acid (FA) and HPLC-grade acetonitrile (ACN) were all purchased from

Sigma-Aldrich (St. Louis, MO). HPLC-grade methanol was obtained from Fisher

Scientific (Fairlawn, NJ) and deionized water used for sample preparation was obtained using a Nanopure Diamond Barnstead purification system (Barnstead International,

Dubuque, IA).

4.2.2 Apparatus

A Waters ACQUITY UPLC® System with a Waters Xevo QTOF mass spectrometer (Milford, MA) was used in this experiment. Our UPLC/EC/DESI-MS assembly employs a thin-layer μ-PrepCellTM electrochemical flow cell. The cell equipped with a magic diamond (MD) electrode served as the working electrode (WE), a electrode served as the reference electrode (RE) and a electrode served as an (AE). The cell was connected to the UPLC column using a piece of 71

PEEK connection tubing (i.d.: 200 μm, length: 25 cm). An ACQUITY UPLC BEH C18 column (1.7 µm, 2.1 mm x 50 mm) was used for peptide separation. An ACQUITY

UPLC Protein BEH C4 column (1.7 µm, 2.1 mm x 50 mm) were employed for protein separation. The flow rate of the mobile phase was 200-300 μL/min. After UPLC separation, compounds flowed through the electrochemical cell and underwent electrochemical reduction. Then the electrolyzed species flowed out of the thin-layer cell via a short PEEK tube (i.d. 510 μm; wall thickness: 530 μm, length: 3 cm) carrying a micro-orifice (i.d. 100 μm) that was located in the tube 2 cm downstream the electrochemical cell. About one-third of the eluent emerged out of the micro-orifice and then was subject to DESI ionization. Unless specified, the DESI spray solvent was

CH3OH/H2O/HOAc (50:50:1 by volume). The flow rate was 10 μL/min. A 5 kV potential

TM was applied to the spray solvent with N2 nebulization (160 psi). A Roxy was used to apply a potential (pulsed mode, E1= -2.0 V for 1990 ms, E2 = -1.5 V for 1010 ms and E3= 0 V for 20 ms) to the electrochemical flow cell to trigger redox reaction.

4.2.3 LC Separation Condition

Somatostatin 1–14 was trypsin digested in 25 mM ammonium bicarbonate at a molar ratio of 1:25 (enzyme/protein) for 12 h at 37 °C. The digest products separation condition were solvent A was 0.1% FA in H2O, and solvent B was 0.1% FA in ACN; 10%

B was ramped to 40% in 3 min. The mobile phase flow rate was 300 μL/min.

Insulin was pepsin digested in water containing 1% acetic acid at a molar ratio of

1:25 (enzyme/protein) for 12 h at 37 °C . The digested insulin was diluted to 20 μM with water containing 0.1% FA and 6 μL of the sample was injected into the UPLC system 72 using an autosampler. Solvent A was 0.1% FA in H2O, and solvent B was 0.1% FA in

ACN. Peptides were eluted by using a 5 min linear elution from 5% to 7% B, and from 7% to 15% B in 1 min, then from 15% to 30% B in 10 min. The flow rate of mobile phase was 300 μL/min.

For the protein mixture separation, 6 μL of a protein mixture of insulin, myoglobin, and α-lactalbumin was loaded onto the UPLC C4 column. The elution flow rate is 200 μL/min. The LC gradient program: solvent A was 0.1% FA in H2O, and solvent B was 0.1% FA in ACN; 28% B was ramped to 32% in 3 min, and then increased to 45% in 3 min.

4.3 Results and Discussion

4.3.1 Reduction of Somatostatin 1-14

The home-built apparatus for coupling a thin-layer electrochemical flow cell with

UPLC/DESI-MS was used (Figure 4-1a). A disulfide-containing peptide somatostain 1-

14 (MW 1637.9 Da) was first chosen to test the feasibility of the UPLC/EC/DESI-MS method. After trypsin digestion, Somatostain 1-14 was produced a peptide mixture,

AGCK TFTSC and NFFWK. The digested mixture (6 μL, 20 μM) was loaded onto the

UPLC for separation. As shown in Figure 1b, the retention times of the two peptides are

1.20 and 2.08 min, respectively. Figure 4-1 c-d shows the DESI-MS spectra of the peptide AGCK TFTSC. When the cell was off, m/z 933 and m/z 467 were observed corresponding to the singly and doubly charged peptide AGCK TFTSC ion, respectively.

Tandem MS/MS was performed to elucidate the ion structures. Upon collision induced dissociation (CID), the fragment ions of the singly charged AGCK TFTSC of m/z 933 are 73 limited, only producing B(b2), A/B(y1) (A/B(y1) refers to a fragment with y1 ion of B chain linked with an intact A chain; the notation also is applicable to other fragment ions),

A/B(y2), A/B(y3), and B/A(y2) (Figure 4-2a). Once the reduction potential was applied, two new ions of m/z 378 and m/z 558 were observed, resulting from the reduction of the

AGCK TFTSC. Moreover, CID of the electro-generated ions provides more fragment ions that cover all of the backbone cleavage sites. CID of the ion at m/z 378 gives rise to b3, b3-H2O, y2 and y3 fragment ions and CID of the ion at m/z 558 yields b2, b3, b4, b4-H2O, y3 y3-H2O and y4-H2O fragment ions. According to the CID fragment ions, the two peptide sequences are assigned as AGCK and TFTSC (Figure 4-2 b-c). In addition, the sum of the MWs of the two products (378.2 + 558.2 – 2.0 = 934.4 Da) is higher than that of the precursor peptide (932.4 Da) by 2.0 Da, suggesting that the precursor peptide

AGCK TFTSC has one disulfide bond. The results show that the electrochemical reduction removes the disulfide bond linkage and produce linear peptides. In this case, more sequences information can be identified by MS/MS. 74

b) a) 1.20 AGCK 100 TFTSC

0 0.50 1.00 1.50 2.00 2.50 3.00 2.08 100 NFFWK

0 Time 0.50 1.00 1.50 2.00 2.50 3.00 100 AGCK AGCK 933.4 c) Cell off [ +H]+ [ +2H]2+ TFTSC TFTSC 467.2 AGCK

% AGCK [ +2Na]2+ [ +Na]+ TFTSC TFTSC 489.6 955.4

0 AGCK + 933.4 100 [ +H] d) Cell on AGCK TFTSC [ +2H]2+ TFTSC AGCK [ +Na]+ % + 467.2 [TFTSC+H]+ [AGCK+H] + TFTSC 378.2 558.2 [TFTSC+Na] 580.2 955.4

0 m/z 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 Figure 4-1. a) The apparatus of UPLC/EC/DESI-MS. b) Extracted ion chromatograms

(EICs) showing the UPLC separation of b) AGCK TFTSC (upper panel) and NFFWK

(lower panel). DESI-MS spectra of AGCK TFTSC c) when the cell was off and d) when the cell was on.

Furthermore, the result shown above is also helpful for pinpointing the disulfide bond linkage between the 3rd residue of one chain AGCK with the 5th residue of the other chain TFTSC. However, the protonated peptide NFFWK containing no disulfide bonds 75 remained unchanged with and without potential applied. This result suggests that

UPLC/EC/DESI could differentiate disulfide bond-containing peptides from others.

100 a) m/z 933

A G C K [ +H]+ A/B(y ) T F T S C A/B(y ) 3 A/B(y1) 2 B/A(y2) B(b2) 0 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 b2 100 b) m/z 558

[ T F T S C +H]+

y3 b b3 4 y -H O 3 2 b4-H2O y4-H2O -H2O Relative Abundance Relative 0 230 250 270 290 310 330 350 370 390 410 430 450 470 490 510 530 550 570 y 100 c) 2 m/z 378

b3 y3 + b3-H2O [ A G C K +H]

m/z 0 200 220 240 260 280 300 320 340 360 380 400 Figure 4-2. CID MS/MS spectra of a) [AGCK TFTSC+H]+ (m/z 933), b) reduced peptide [AGCK+H]+ (m/z 378) and c) reduced peptide [TFTSC+H]+ (m/z 558).

4.3.2 Reduction of Insulin

After the successful trial with somatostain 1-14 digest, we further tested the method for the analysis of a protein pepsin digest. The bovine pancreatic insulin was 76 chosen as the test sample. It is known that Insulin is composed of A chain and B chain linked by two inter-peptide disulfide bonds; and the A-chain has an additional intra- peptide disulfide bond. As shown in the acquired EICs (Figure 3-2), the peptides were well separated within 15 min. After separation, The MS spectra (Figure 4-4) obtained in the positive ion show the ions of the first six eluted peptides, including [YTPKA+H]+

(m/z 579), [FVNQ+H]+ (m/z 507), [GIVE+H]+ (m/z 417), [YQLEN+H]+ (m/z 666),

[YQLE+H]+ (m/z 552), and [VEAL+H]+ (m/z 431). When the reduction potential was applied, there were no new peaks observed (), which indicates that these are peptides without disulfide bonds. CID MS/MS analysis was applied to gain further structural information of these peptide ions except GIVE in which the first and second peptide bond cleavages were missing. 77

100 YTPKA % 0 2.00 4.00 6.00 8.00 10.00 12.00 14.00

100 FVNQ % 0 2.00 4.00 6.00 8.00 10.00 12.00 14.00

100 GIVE % 0 2.00 4.00 6.00 8.00 10.00 12.00 14.00

100 YQLEN % 0 2.00 4.00 6.00 8.00 10.00 12.00 14.00

100 YQLE % 0 Time 2.00 4.00 6.00 8.00 10.00 12.00 14.00

100

% VEAL 0 2.00 4.00 6.00 8.00 10.00 12.00 14.00 P1: QCCASVCSL Relative Abundance Relative 100

% FVNQHLCGSHL 0 2.00 4.00 6.00 8.00 10.00 12.00 14.00 P2: NYCN 100

% LVCGERGFF 0 2.00 4.00 6.00 8.00 10.00 12.00 14.00 P3: GIVEQCCASVCSL 100

% HLCGSHL 0 2.00 4.00 6.00 8.00 10.00 12.00 14.00 P4: GIVEQCCASVCSL 100

% FVNQHLCGSHL 0 Time 2.00 4.00 6.00 8.00 10.00 12.00 14.00 Figure 4-3. EICs of the pepsin-digested insulin. 78

[YTPKA + H]+ [FVNQ + H]+ 100 579 507 a) Cell off 100 c) Cell off [FVNQ + Na]+ % [YTPKA + Na]+ % 529 601 0 0 [YTPKA + H]+ [FVNQ+ H]+ 100 579 507 b) Cell on 100 d) Cell on [FVNQ + Na]+ %

+ %

Relative Abundance Relative [YTPKA + Na] 529 601 0 m/z 0 m/z 100 150 200 250 300 350 400 450 500 550 600 650 700 100 150 200 250 300 350 400 450 500 550 600 [GIVE + H]+ [YQLEN + H]+ 666 100 e) Cell off 417 100 g) Cell off

[YQLEN + Na]+ % % [GIVE + Na]+ 688 439 0 0 [GIVE + H]+ [YQLEN + H]+ 417 666 100 f) Cell on 100 h) Cell on

+ % % [YQLEN + Na] Relative Abundance Relative + GIVE+ Na] 688 439 0 m/z 0 m/z 300 320 340 360 380 400 420 440 460 480 500 400 450 500 550 600 650 700 750 800 [YQLE + H]+ [VEAL + H]+ 552 431 100 i) Cell off 100 k) Cell off

+ + [VEAL + Na] % [YQLE + Na] % 574 453

0 0 [YQLE + H]+ [VEAL + H]+ 552 431 100 j) Cell on 100 l) Cell on + + [VEAL + Na] %

[YQLE + Na] %

Relative Abundance Relative 574 453

0 m/z 0 m/z 300 350 400 450 500 550 600 650 700 400 410 420 430 440 450 460 470 480 490 500 Figure 4-4. DESI-MS spectra of the digested insulin peptides carrying no disulfide bonds

YTPKA: a) cell off and b) cell on; FVNQ: c) cell off and d) cell on; GIVE: e) cell off and f) cell on; YQLEN: g) cell off and h) cell on; YQLE: i) cell off and j) cell on; and VEAL: k) cell off and l) cell on.

For the additional four peptides that eluted later (denoted as P1, P2, P3, and P4,

Figure 4-3), new peaks were observed once the potential was applied to the cell. This 79 indicates that the peptides carried disulfide bonds. Since the UPLC separation was used before electroreduction, these new peptide ions would correspond to the reduced peptides of the UPLC-separated precursor peptides. Importantly, CID MS/MS analysis of the reduced peptides provides useful information for the sequencing and disulfide bond mapping of the precursor peptide. As shown in Figure 4-5, in the case of P3, the sum of the MWs of the two reduced products GIVEQCCASVCSL and HLCGSHL (2075.9 Da, calculated from the measured m/z of the corresponding product ions) is higher than that of P3 (2071.9 Da) by 4.0 Da, which indicates that the precursor peptide P3 has two disulfide bonds. The singly charged ion of chain B generated from electrolysis (m/z 766) gave rise to fragment ions b2, b3, b4, b5, b6, y2, y3, y4, y5, and y6 upon CID (Figure 4-5e).

This set of fragment ions covers all of the backbone cleavages and gives the sequence of the peptide as HLCGSHL with one cysteine residue located at the 3rd amino acid site.

Likewise, CID spectrum of the single charged ion of chain A from P3 reduction (m/z

1311) yielded b3 b4, b5, b6, b6-H2O, b7, b8, b9, b10, b10-H2O, b11, and b12 (Figure 4-5d). The chain A determines the most of the sequence for XXXEQCCASVCSL (X means the first three amino acids are unknown). All three cysteine residues of chain A are known to be located at the 6th, 7th, and 11th amino acid sites. Thus, there are three possible disulfide bond linkages for P3 (the chain B sole cysteine residue links with one of the three cysteine residues of chain A). Upon CID, the doubly charged P3 ion (m/z 1037, Figure 4-

2+ 2+ 5c) dissociated into A(b2), A(b3), A(b4) B/A(b11) , B/A(b12) , B/A(y8), B/A(y10), B(b2),

6 11 B(y2), and B(y4), in which the backbone cleavage between Cys and Cys of the chain is missing. This result reveals that the Cys6 in the chain A is paired up with Cys11 of the 80 chain A to form an intra-peptide bond and Cys7 of chain A links with Cys3 of chain B in

P3. The appearance of fragment ions A(b2) and A(b3) in Figure 4-5c shows the 3rd residue of A is valine. Thus, chain A can be determined as XXVEQCCASVCSL. These results show that both the locations of disulfide bonds and most of the P3 sequence can be determined using the information acquired from this UPLC/EC/DESI-MS method. 81

[GIVEQCCASVCSL + 2H]2+ 100100 1037 HLCGSHL a) P3 Cell off

% [GIVEQCCASVCSL + 3H]3+ HLCGSHL 692

00

100100 1037 b) P3 Cell on [GIVEQCCASVCSL + 2H]2+ HLCGSHL Relative Abundance Relative [GIVEQCCASVCSL + 3H]3+

% HLCGSHL + 692 766 [HLCGSHL+H] [GIVEQCCASVCSL+H]+ 1311

00 m/zm/z 700700 7508 80000 8509 90000 9501000 1000 10501100 1100 11501200 1200 12501300 1300 13501400 1400

x2 ×2 100 c) m/z 1037

G I V E Q C C A S V C S L [ +2H]2+ H L C G S H L % B(y2) B/A(b )2+ Relative Abundance Relative B(b2) 12 A(b3) 2+ A(b2) B/A(b ) A(b ) B(y4) 11 4 B/A(y8) B/A(y10) 0 m/z 200 400 600 800 1000 1200 1400 1600 1800

100 d) m/z 1311

[G I V E Q C C A S V C S L +H]+ %

b b10-H2O b6 8 b9 b b6-H2O b7 4 b5 b10 b11 b3 b12

0 m/zm/z 200 400 600 800 1000 1200 1400 1600 100 e) m/z 766 Relative Abundance Relative

[H L C G S H L+H]+

% b2 y5 y2 b 5 y6

y3 b4 b b6 3 y4

0 m/zm/z 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800

Figure 4-5. DESI-MS spectra of P3 with a) cell off and b) cell on, c) CID MS2 spectrum of doubly charged P3 at m/z 1037, d) CID MS/MS spectrum of singly charged P3 chain A at m/z 1311, and d) CID MS/MS spectrum of singly charged P3 chain B at m/z 766. 82

Since the low digestion specificity of pepsin, two additional peptides, P1 and P4, were generated. Their structures (shown in Figure 4-3) are similar to that of P3 and can be determined by MS/MS analysis. The fragmentation patterns of P1 and P4 ions also agree with the disulfide bond assignment for P3 (Figure 4-6).

83

100100 a) m/z 1082 Q C C A S V C S L [ +2H]2+ F V N Q H L C G S H L

% A(y2) 2+ B(b2) B/A(b8) 2+ A/B(y10) B(b ) 3 A/B(y6)-NH3 B(y2) B(b5) B(y4) B(b6) A/B(y )-NH A/B(y5)-NH3 7 3 A/B(b7) 00 m/zm/z 200200400 400 600 600 800 8001000 1000 1200 12001400 14001600 16001800 1800 100 b) m/z 854

Relative Abundance Relative A(b2) G I V E Q C C A S V C S L [ +3H]3+ F V N Q H L C G S H L A(y2) % A/B(y9)

B/A(b11) A(y1) B/A(y9) A(b3) A/B(y ) orB(y ) B(y ) 7 1 A(b ) 4 B/A(y10) B/A(y11) B(b2) 4 B(b3) A(b ) B(b ) B/A(y8) 5 B(b5) 6

0 m/z 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 Figure 4-6. CID MS/MS spectra of a) [P1+2H]2+ (m/z 1082) and b) [P4+3H]3+ (m/z 854).

In the case of P2, one of its reduced products (A-chain) was missing in the detected by UPLC/DESI-MS (Figure 4-7 a-b) It was caused by the lack of the basic amino acid residues that will decrease the ionization efficiency. Only B-chain

LVCGERGFF was detected. Similar phenomena of missing chain A in the spectra of reduced insulin has been reported before.88 Even though the chain A was missing, the 84 sequence of P2 and its disulfide bond location still can be inferred based on examination of MS/MS data and MW analysis. CID spectrum of doubly charged P2 m/z 769 (Figure

2+ 4-7c) yields fragment ions B/A(y2), B/A(y2)-NH3, B/A(y3), B/A(b3) , A(b2), B(y1), B(y4),

2+ A/B(y7), A/B(y7) , and B(b2). CID spectrum of the doubly charged chain B of P2 (m/z 514) was further examined (Figure 4-7d), and its fragment ions of b2, b3, b6, b7, y1, y4, y5, y6, y7 and y8. In this case, the sequence of the reduced peptide, chain B, was determined to be

LVCGERGFF. Assuming that P2 has one disulfide bond, the MW of chain A is calculated to be 512.1 Da based on the MWs of P2 and chain B. For sequencing chain A, the first two amino acids of chain A can be determined as asparagine and tyrosine, which was confirmed by both the MW of the A-chain and the CID MS/MS analysis of

2+ [P2+2H] (m/z 769) by observing fragment ions B/A(y2) and B/A(y3) (Figure 4-7c). The

2+ appearance of A(b2) and B/A(b3) in Figure 4-7c further suggests that the last two amino acid residues of chain A are cysteine (modified with chain B) and asparagine residues.

Thus the chain A sequence is determined as NYCN. In this case, chain A and B in P2 peptide were connected with one inter-peptide disulfide bond. The Cys3 of chain A links with Cys3 of chain B in P2, which is in agreement with our assumption above. This result reveals that the sequence of P2 can also be identified. Note that another peak of m/z 815 was observed with the cell on (Figure 4-7b), which is caused by in-source CID of chain B ions. Thus, the UPLC/EC/DESI-MS with MS/MS could provide rich information for insulin sequencing (98% coverage) and disulfide bond location determination (three disulfide bonds can be located). 85

769 2+ 100 [NYCN+2H] a) Cell off LVCGERGFF %

0 2+ 815[CGERGFF+H]+ [LVCGERGFF+H]+ 100 [NYCN+2H] b) Cell on 769 1027 LVCGERGFF Relative Abundance Relative [LVCGERGFF+2H]2+ 514 %

0 m/zm/z 500 550 600 650 700 750 800 850 900 950 1000 1050 1100 1150 1200

100 c) m/z 769

N Y C N [ +2H]2+ L V C G E R G F F %

B(b2) 2+ B/A(b3) B(y1) A(b2) B(y4) 2+ A/B(y7) B/A(y2) A/B(y7) B/A(y2)-NH3 B/A(y3) 0 m/zm/z 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400

100 m/z 514 d) y7 [L V C G E R G F F +2H]2+ Relative Abundance Relative %

b2 b y1 6 y b7 5 y6 y8 b3 y4 0 m/zm/z 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 Figure 4-7. DESI-MS spectra of P2 with a) cell off and b) cell on. CID MS/MS spectra of

[P2+2H]2+ (m/z 769) and [LVCGERGFF+2H]2+ (m/z 514).

86

4.3.3 Reduction of Proteins

The UPLC/EC/DESI-MS method is not only for small peptides, but also work for intact proteins. It can be used for determining the presence of disulfide bonds in proteins.

The higher charging of protein ions can also be obtained using this method via online electroreduction combined with online supercharging. From previous work, electrochemical reduction can remove disulfide bond bridges and then the protein would unfold with increased charges observed.32, 89-90 Also, supercharging reagents can be used to increase ion charges.91 It would be interesting that the electroreduction of disulfide coupled with a supercharging reagent to increase protein charging. So we introduce our method to supercharge proteins following UPLC separation and electrochemical reduction. The charge state distributions (CSDs) of protein ions will shift to higher charges with the appearance of a new population which could indicate a conformational change after disulfide bond reduction.

In the experiment, we employed a mixture of insulin (containing two inter-peptide and one intra-peptide disulfide bond), myoglobin (containing no disulfide bonds) and α- lactalbumin (containing four inter-peptide disulfide bonds) to test the capability of the method. The mixture was well separated within 5.5 min (Figure 4-8). 87

100100 insulin %

0 2.00 4.00 6.00 8.00 10.00 100100 myoglobin %

0 0 2.00 4.00 6.00 8.00 10.00 100100 α-lactalbumin Relative Abundance Relative %

00 TimeTime 2.002 4.004 6.006 8.00 8 10.0010 Figure 4-8. EICs of the protein mixture containing insulin, myoglobin and α-lactalbumin.

Comparing the mass spectra recorded of insulin before and after adding the supercharging reagent m-NBA (50 mM) into the DESI spray, the maximum charge state shifted from +5 to +6 (Figures 4-9 a-b). Insulin has three disulfide bonds as mentioned above. Once the reduction potential was applied, two new peaks at m/z 851 and 1134 corresponding to the +4 and +3 charge states of chain B, respectively, could be observed

(Figure 4-9c), suggesting the disulfide bonds existed. The missing of A chain ions in the spectrum was probably caused by the lack of sufficient numbers of basic amino acid residues in chain A. As shown in Figure 4-10, for myoglobin, After supercharging reagent m-NBA (50 mM) doped the DESI spray, the maximum charge state of myoglobin shifted from +22 to +24, the average charge state increased from +15.1 to +18.9, and the highest abundance ion was shifted from +18 to +20. But since no disulfide bond was involved in myoglobin, the mass spectra had no obvious differences with and without cell on (shown in Figure 4-10 b-c). 88

Insulin 100 a) Cell off +4 without supercharging +5 +3 %

0

+4 100 b) Cell on +5 with supercharging

% +3 +6 0 Relative Abundance Relative c) Cell on +4 100 with supercharging 4+ +5 Chain B Chain B3+ +3 % +6 0 m/z 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900

Figure 4-9. DESI-MS spectra of the UPLC-separated insulin a) when cell was off without supercharging, b) when cell was off with supercharging and c) when cell was on with supercharging.

Myoglobin +18 100100 +19 +17 a) Cell off +20 +16 without supercharging

% +15 +21 +14 +22 +13 +12 +11 +10 +9 00 +20 100100 +21 +19+18 b) Cell off +22 +17 with supercharging

% +16 +23 +15 +24 +14 +13 00 +20+19 100100 +21 c) Cell on

Relative Abundance Relative +18 +22 +17 with supercharging % +23 +24 +16 +15 +14 +13 00 m/z 700800 800 900 1000 1000 11001200 1200 13001400 1400 15001600 1600 17001800 1800 1900

Figure 4-10. DESI-MS spectra of the UPLC-separated myoglobin a) when cell was off without supercharging, b) when cell was off with supercharging and c) when cell was on with supercharging.

89

For the protein α-lactalbumin, upon adding the supercharging reagent (Figure 4-

11), the maximum charge state of α-lactalbumin shifted from +11 to +14. Moreover, when the reduction potential was applied, interestingly, a new peak corresponding to +15 ion appeared (Figure 4-11c). Compared the mass spectra of α-lactalbumin before and after cell on with supercharging reagent (Figure 4-11 b-c), the average charge state increased from +9.9 to +11.2; and the ion with the highest abundance was shifted from

+8 to +12. α-Lactalbumin has 123 amino acid residues and contains four disulfide cross- links to maintain and stabilize its structure, so it is likely that the reduction of the disulfide bonds shall unfold the protein. In this case, unfolded proteins have a greater capability to carry a larger number of charges on its surface than that of folded proteins.29,

88 This result shows the feasibility of combinng supercharging and disulfide bond reduction for increasing protein charge states92. Increased charging is useful for increasing ion dissociation efficiency, especially for proteins constrained by disulfide bonds.29, 93

α-lactalbumin +8 100100 a) Cell off

% without supercharging +9 +11 +10 00 +8 100100 b) Cell off +9 +10 with supercharging +11 % +13 +12 +14 00 +12 100100 +13 c) Cell on Relative Abundance Relative +11 with supercharging +14 % +15 +10 +9 +8 00 m/z 900 10001000 11001200 1200 13001400 1400 15001600 1600 17001800 1800 Figure 4-11. DESI-MS spectra of the UPLC-separated α-lactalbumin a) when cell was 90 off without supercharging, b) when cell was off with supercharging and c) when cell was on with supercharging.

4.4 Conclusions

This study suggests a new approach using UPLC/EC/DESI-MS for the structural elucidation of different types of disulfide-bond containing proteins/peptides. Disulfide bond-containing peptides in enzymatic digest mixtures can be identified. After UPLC separation, the linear peptide chains from electrochemical reduction can be easily assigned to their precursors and sequenced. In addition, upon the CID MS/MS analysis, sequencing peptides and pinpointing the disulfide linkages are possible. The combination of electrolysis treatment and online supercharging reactive-DESI could be used to increase charges for the proteins carrying intra-disulfide bonds. As disulfide bonds play an important role in protein conformation and function, this UPLC/CE/DESI-MS method would become a powerful tool in proteomics research.

91

CHAPTER 5: COUPLING ELECTROCHEMISTRY WITH PROBE ELECTROSPRAY

IONIZATION MASS SPECTROMETRY

Adapted from Cai, Y.; Liu, P. Held, M. A.; Dewald, H. D., and Chen, H.,

ChemPhysChem 2016, 17, 1104-1108. Copyright 2016, WILEY-VCH.

5.1 Introduction

EC-MS has become a powerful method for structural identification of electrochemical reaction products or intermediates.49, 69 Ionization techniques are the key component to join these two techniques and many ionization methods such as EI,50 TS,51

FAB,52 and ESI53, 94 have been used as the EC/MS coupling interface. Recently, with the advent of ambient MS, EC/MS coupling has been diversified,49, 69 using techniques such as DESI,28, 88 nanospray DESI,95 and flowing atmospheric-pressure afterglow (FAPA).96

The direct sampling capability of ambient ionization techniques simplifies the instrumentation of EC/MS and allows the examination of electrolyzed samples in various media.34-35, 60, 69, 95 However, although EC/MS techniques have found extensive applications in proteomics34-35, 88 and drug simulation,97 no investigation of the location of the electrolytic redox reaction occurrence in the cell by MS has been reported to the date. Electrochemical cell MS imaging, telling what electrochemical reaction takes place and where it occurs, would be of high value in elucidating electrochemical reaction mechanisms.

PESI is a soft ionization method which operates at atmospheric pressure and employs a conductive solid probe for a microliter droplet of analyte solution to be deposited. When a high voltage is applied to the probe, the droplet will be charged and 92 sprayed to produce ions. Early work relevant to PESI was conducted by Shiea and co- workers98-100 and more recently by Hiraoka.36, 40, 101 Compared to ESI or nano-ESI, PESI has the advantages of higher salt tolerances.37-38 In addition, it is rapid and consumes only a small fraction of total sample solution (in microliters) for ionization.39 In particular,

PESI-MS is capable of performing chemical imaging analysis of biological tissues,41, 43,

102 providing information about the spatial distribution of analyte of interest. Very recently, single cell analysis with PESI-MS for the detection of metabolites at cellular and subcellular levels was reported.44 In addition, PESI was shown to be capable of directly monitoring solvent-free reactions and organometallic catalysts in room- temperature ionic liquids (RTILs) in our laboratory.103

In this study, PESI-MS was employed to study electrochemistry. First PESI-MS was used to monitor directly various electrochemical reactions occurring in RTILs, including the electrochemical oxidation of ferrocene, N,N’- bis(salicylidene)ethylenediaminocobalt(II) (Co(II) salen), 2,2,6,6-tetramethylpiperidine

1-oxyl (TEMPO). Second, PESI-MS was introduced to detect electrochemical reactions on different or multiple electrode surfaces. Furthermore, peptides/proteins separated in an isoelectric focusing (IEF) cell were also successfully detected by PESI-MS.

5.2 Experimental

5.2.1 Chemicals

Clozapine, 1-butyl-3-methylimidazolium hexafluorophosphate ([BMIM][PF6]) ferrocene, N,N’-bis(salicylidene)ethylenediaminocobalt(II) (Co(II) salen), formic acid

(FA), 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO), β-lactoglobulin A from bovine, 93 ubiquitin from bovine erythrocytes, cytochrome c from bovine heart, 3-nitrobenzoic acid

(m-NBA, HPLC) and HPLC-grade acetonitrile (ACN) were all purchased from Sigma-

Aldrich (St. Louis, MO). Angiotensin II (human), angiotensin II (1-4), and angiotensin II antipeptide were all purchased from American (Sunnyvale, CA). Bio-lyte ampholytes

(pH 3-10) was purchased from Bio-Rad (Hercules, CA). HPLC-grade methanol was obtained from Fisher Scientific (Fairlawn, NJ) and deionized water used for sample preparation was obtained using a Nanopure Diamond Barnstead purification system

(Barnstead International, Dubuque, IA).

5.2.2 Apparatus

The mass spectrometer used in this study was either an AB SCIEX Q-trap 2000 triple-quadrupole-linear ion trap mass spectrometer (Concord, Canada) or an LCQ DECA ion trap mass spectrometer (San Jose, CA). The commercial ion source of the mass spectrometer was removed to accommodate the PESI ion source. As illustrated in Figure

1a, the PESI probe used was a solid stainless needle (Fisher Scientific, San Jose,

CA). Unless specified otherwise, a small amount of sample solution (1-2 μL) from the electrochemical cell was directly loaded onto the probe tip and then ionized with a 3.5-4 kV high voltage applied to the probe. The sample loading was carried out by inserting the probe into the specific location of interest in the for 20s. The distance between the probe tip and the mass spectrometer inlet varied from 0.5 to 1.0 cm. The ions generated from PESI were collected and detected using the mass spectrometer mentioned above. Collision-induced dissociation (CID) spectra were also collected for ion structure confirmation. 94

5.3 Results and Discussion

5.3.1 PESI-MS Detection of Electrochemical Reactions in RTILs

RTILs has become popular the electrolysis media for electrochemistry because no additional electrolyte is needed. RTILs have a wide potential window and excellent solubility for either polar or nonpolar compounds. In this case, many classical electrochemical reactions can be explored in RTILs. However, RTILs have strong ion suppression effect for MS examination. In addition heated curtain/desolvation gas was required to reduce the sample viscosity,104 and the contamination are often the major problem for MS detection.105 PESI-MS has the advantage of high salt tolerance, so the challenging experiment of detecting electrochemical reactions in RTILs was explored.

To validate the feasibility of this method, the organometallic compound CoII salen, a catalyst used for dehalogenation reaction, was first chosen as a test sample. 500 μL of 1

II mM Co salen was dissolved in ionic liquid [BMIM][PF6] and added to a petri dish (i.d.

2.5 cm). Two Pt (i.d. 0.4 mm) were served both as the working electrodes (WE) and the counter electrode (CE) and inserted into the solution for electrolysis (Figure 5-1a).

The PESI probe (a stainless steel needle, Fisher Scientific) was dipped in the sample solution (about ~30 s) and then ionized with a high voltage (+3 kV) in front of the mass spectrometer. When no voltage was applied to the cell, there is no electrooxidation

II + product of Co salen was observed. The ion m/z 423 was [BMIM]2[PF6] (Figure 5-1b).

But as shown in Figure 5-1c, when 5 V was applied across the two Pt wires for 30 V and the sample solution on the Pt WE was sampled and ionized by PESI, the CoII salen ion at m/z 325 was detected. Upon CID, m/z 325 gave rise to fragment ions m/z 204, m/z 192 95 and m/z 179 by the loss of C7H5NO, C8H7NO, and C9H8NO, respectively (Figure 5-2), consistent with its assigned structure. When PESI probe was used to replace Pt WE for experimental simplicity, a similar result was obtained.

96

Figure 5-1. a) Scheme showing the PESI-MS and the electrochemical cell configuration.

PESI-MS spectra of b) CoII salen before electrolysis and c) CoII after electrolysis. The inset in c) shows the oxidation reaction of CoII salen.

97

Figure 5-2. CID MS/MS spectrum of the electrochemical oxidation product CoIII salen ion (m/z 325).

In addition, ferrocene, another organometallic compound, was also tested using the same apparatus. After electrolyzed, the oxidation product ferrocenium cation at m/z

186 was detected by PESI-MS (Figure 5-3b). As a control, before electrolyzed no oxidation product of ferrocene was observed (Figure 5-3a). 98

a)

b)

Figure 5-3. PESI-MS spectra of a) ferrocene before electrolysis and b) ferrocene after electrolysis. The inset in c) shows the oxidation reaction of ferrocene.

Besides the electrolysis of organometallic compounds, the electrochemical oxidation of TEMPO in RTIL was also examined. TEMPO loses on electron and undergoes a reversible redox reaction to form the catalytically active oxoammonium species (Figure 5-4b inset). As shown in Figure 5-4a, only m/z 139 and m/z 157 was

+ + observed which corresponding to [BMIM] and [BMIM+H2O] . Without a voltage applied on the cell, no oxidation product of TEMPO was observed by PESI-MS. When a

5 V potential was applied to the cell, the oxidation product oxoammonium species at m/z

156 was seen in the PESI-MS spectrum (Figure 5-4b). CID of m/z 156 gave a fragment 99 ion of m/z 139 by the loss of NH2OH. These results suggest that PESI-MS has the feasibility of studying electrochemical reaction using an ionic liquid as a media.

100

156 m/z 156 7.5e4 123 -HONH2

3.5e4 Intensity

0 100 150 200 m/z Figure 5-4. PESI-MS spectra of a) TEMPO before electrolysis and b) TEMPO after electrolysis. The inset in c) shows the oxidation reaction of ferrocene. CID MS/MS spectrum of the electrochemical oxidation product TEMPO ion (m/z 156).

101

5.3.2 PESI-MS Detection of Electrochemical Reaction on Different Electrodes

For conventional EC/MS experiment, both oxidation and reduction products would be detected together. But using PESI probe, the products can be detected separately. To demonstrate the possibility, a mixture of dopamine and flutamide was used to test. The reduction of flutamide and the oxidation of dopamine process were shown in

Figure 5-5a. In this experiment, the mixture (200 μM each) in MeOH/H2O/FA (50:50:1 by volume) with 10 mM KCl was added into the petri-dish cell. Two wires were used as both the WE and CE. A 5 V voltage was applied across the two electrodes. The

PESI probe was dipped in solution on the cathode surface and on the anode surface to load sample, respectively. Then the PESI probe was ionized with high voltage. The MS spectra acquired from the electrolyzed solution sample on the cathode and anode are given in Figures 5-5 b-c. As shown in Figure 5-5 b for the cathode sample, the nitro group of flutamide undergoes a two e- reduction process with the loss of one molecule of water to form nitroso intermediate ion observed at m/z 261, a further two e- reduction generates the hydroxylamine intermediate ion detected at m/z 263 which further undergoes a two e- reduction process to yield the final amine product ion appearing at m/z 247. But only the protonated dopamine (m/z 154) along with its in-source fragmentation ion at m/z 137 was observed and no oxidation product was detected. In contrast, as shown in Figure 5-5c, for the anode sample, the oxidized dopamine product, dopamine quinone, was observed at m/z 152. The ion at m/z 123 results from the

CH2=NH loss from m/z 152 due to in-source dissociation. But only the protonated flutamide ion at m/z 277 was observed without any reduction products ion. 102

Figure 5-5. a) Equations showing the electrochemical reduction of flutamide and the electrochemical oxidation of dopamine; EC/PESI-MS spectra of the electrolyzed mixture sampled from b) the anode surface and c) the cathode surface.

103

As a comparison, the EC/MS experiment was also run in the traditional way. A commercial thin-layer flow cell was used as the electrochemical cell, and desorption electrospray ionization (DESI) were used as the ionization method for EC/MS coupling.

The commercial flow cell employed a gold (Au) electrode as the WE (i.d. 8 mm) and a

HyREF electrode as the RF and a titanium electrode the CE. A mixture of dopamine and flutamide (100 μM each) in MeOH: H2O: FA (50:50:0.1 by volume) was infused into the cell. The flow rate was 5 μL/min. A -1.5 V was applied to the cell using a Roxy™ potentiostat. The electrochemically oxidized and reduced samples flowed out of the cell via a short piece of silica capillary and interacted with the charged microdroplets from the

DESI spray for ionization. The spray solvent for DESI was MeOH : H2O : FA (50:50:1, by volume). The flow rate for DESI spray solvent was 5 μL/ min, and a high voltage + 5 kV was applied to the spray probe. When the mixture sample flowed out of the cell after electrolysis and was ionized by MS, the acquired DESI-MS spectrum is shown in Figure

5-6. Both the reduced flutamide products and the oxidized dopamine product were detected simultaneously.

104

Figure 5-6. Traditional EC/MS spectrum of the mixture after electrolysis.

In addition to studying two electrodes, a bipolar electrode (BPE) was employed.

A typical BPE cell involves two driving electrodes and a BPE that is placed between the driving electrodes. When sufficient potential is applied across the driving electrodes, the potential differences between the two ends of the BPE built up. In this case, the electrochemical reaction can be driven. In previous studies, the color indicators or fluorescent tracer are needed to monitor the occurrence of the electrochemical redox reaction in BPE cell. But in this case, no chemical structure information can be obtained.

PESI-MS can be a good method to detect the electrochemical reactions in BPE cells without adding visualization reagents.

Two Pt wires were inserted into the dish serving as two driving electrodes. A U- shaped Pt serving as the BPE was placed between the driving electrodes (Figure 5-

7a). 0.5 mM clozapine (CLZ) was chosen as the demonstrated sample. ACN/H2O (30:70, by volume) containing 20 mM ammonium formate was used as an electrolyte. The CLZ tends to lose two to form the CLZ cation (Figure 5-7b). So after 30 V was 105 applied to the two driving electrodes, four positions in the cell were sampled and analyzed by PESI-MS. The acquired spectra are shown in Figure 5-7 c-e, the oxidized product CLZ cation at m/z 325 was observed in both position 1 and 3 which corresponding to BPE anode and driving electrode anode. CID of m/z 325 gives a fragment ion of m/z 277 by the loss of C5H10N2, which is consistent with its structure.

Only protonated unoxidized CLZ (m/z 327) was observed on position 2 and position 4

(Figures 5-7 d-f). No oxidation product was detected. These results suggest that PESI-MS can be used to detect the electrochemical reaction on different/multiple electrodes. 106

Figure 5-7. a) Scheme showing the BPE cell configuration; b) the oxidation process of

CLZ. PESI-MS spectra of the electrolyzed CLZ solution sampled from four different electrode surfaces: c) position 1 (driving electrode anode), d) position 2 (BPE cathode), e) position 3 (BPE anode), and f) position 4 (driving electrode cathode).

107

5.3.3 PESI-MS Detection of Separated Proteins/Peptides in IEF

PESI-MS are not only for detecting small molecular, but also for protein/peptide analysis. PESI-MS was further employed to analyze IEF separated protein/peptide. IEF typically uses the migration of ampholyte buffer, a mixture of aliphatic polyaminopolycarboxylic acids, to establish a pH gradient in the presence of an applied electric field. It is a technique for separating different molecules (e.g., proteins or peptides) by differences in their isoelectric points (pIs). A charged analyte starts to migrate in an IEF cell (configuration is shown in Figure 5-8a) and stops migration when it reaches the region where the pH is equal to its pI. For IEF experiments, the fractionated samples are usually examined by LC/ESI-MS and MALDI-MS. However, further time- consuming preparation is needed as the high concentration of ampholytes suppress the ion signal. In this study, PESI-MS was attempted to be used for the detection in IEF experiments, in the consideration of the fact that IEF cell is analogous to the electrochemical cell and the high salt tolerance of PESI-MS.

Mini Rotofor® cell with 18 mL sample volume (Bio-Rad, Hercules, CA) was used for liquid-phase IEF to separate a mixture of proteins and peptides. Bio-Lyte ampholyte (pH 3-10, 40% solution, Bio-Rad) was diluted to 1.5% (for peptides) or 2.0% w/v (for proteins) by water as the buffer solution. A 700-800 V voltage was applied across the cell to drive the separation. The total power was remained at 12 W for 2 during the separation. After isoelectric focusing, 20 fractions were collected through tubings into 20 vials enclosed in the harvesting box driven by a vacuum source. Finally, the PESI probe was dipped into the sample for 20s and then taken out for detection. 108

In our experiment, a peptide mixture of angiotensin II (pI 6.7), angiotensin II (1-4)

(pI 5.8), and angiotensin II antipeptide (pI 5.2, 0.25 mg/mL each) was separated. After separation, twenty fractions were numbered and collected and then subjected to PESI-MS analysis. A drop of the solvent of MeOH: H2O: FA (50: 50: 1 by volume) was added to the solution adsorbed on the PESI probe tip for dilution and then a 4 kV voltage was applied to the probe for ionization. Good ion signal was obtained (it turns out that the dilution helps to obtain stable signal probably due to the reduced viscosity of the sample).

Figures 5-8 c-e show PESI-MS spectra of the fractionated angiotensin II (fraction #12), angiotensin II (1-4, fraction #9) and angiotensin II antipeptide (fraction #6), respectively.

The ions of all the three separated peptides were clearly seen. In comparison to the signals of ions detected by PESI-MS before IEF (Figure 5-8b), the intensity of the separated peptide ions increased by 5-20 folds, showing the enrichment effect of isoelectric focusing. 109

Figure 5-8. a) Scheme illustrating the process of the IEF/PESI-MS; b) PESI-MS spectrum of the mixture sample in the IEF buffer before electrofocusing; PESI-MS spectra of c) angiotensin II (fraction #12), d) angiotensin II (1-4, fraction #9), and e) angiotensin II antipeptide (fraction #6). 110

Besides peptides, proteins were also tested in this experiment. A mixture of β- lactoglobulin A (pI 9.6) ubiquitin (pI 6.8) and cytochrome c (pI 5.1, 0.3 mg/mL each) in

Bio-Lyte ampholyte (pH 3-10, 2.0% w/v, Bio-Rad) aqueous solution was separated in the

IEF cell. After separation, twenty fractions were collected. The PESI probe was dipped into each fractionated sample solution for 20 s. After adding a drop of MeOH: H2O: FA

(50: 50: 1 by volume) containing 1% m-NBA (a compound found to help obtain protein ion signal from ion suppressing matrices) to the probe tip for sample dilution and applying a 4 kV to the probe, protein samples were successfully ionization. As shown in

Figures 5-9 b-d, cytochrome c (fraction #20), ubiquitin (fraction #12) and β-lactoglobulin

A (fraction #6) were clearly detected by PESI-MS. In contrast, no protein signal was seen from PESI of the mixture sample before isoelectric focusing (Figure 5-9a). Apparently, these results demonstrate that PESI-MS can serve as a detection technique for IEF experiments, which would have extensive applications. 111

Figure 5-9. a) PESI-MS spectrum of the mixed protein sample in the IEF buffer prior to isoelectric focusing; PESI-MS spectra of b) cytochrome c (fraction #20), c) ubiquitin

(fraction #12) and d) β-lactoglobulin A (fraction #6).

5.4 Conclusions

This study presents the development of EC/PESI-MS along with its new application for electrochemical cell. Benefitting from its high salt tolerance, PESI-MS could be used to investigate electrochemical reactions taking place in RTILs, a challenging problem in the past. PESI-MS detection of electrochemical reactions on 112 different electrodes also provides spatial information for the reaction occurrence and the structural information of the reaction products, which was not achieved by using other ionization methods. Equally importantly, PESI-MS provides an easy and fast way to detect the fractionated samples from isoelectric focusing. It can be seen that PESI-MS has a great potential in the study of electrochemistry.

113

CHAPTER 6: SUMMARY AND FUTURE WORK

Base on the original liquid sample DESI work done in our laboratories, further modification with the sample introduction part has been carried out, resulting in the development of the new DESI orifice for LC coupling. In this case, three advantages over traditional LC/MS could be achieved: 1) The separation can be operated in high flow rate without causing ion source flooding; 2) Reactive DESI can be achieved for online derivatization. 3) The continuous MS response could be utilized for further coupling of

LC/MS with EC.

To validate the new orifice of liquid sample DESI, a wide range of samples were tested from small organic molecules, peptides, protein digests and proteins from which high quality and reproducible mass spectra were obtained. Versatile application for coupling with EC has been explored using the LC/DESI-MS interface carrying the new capillary orifice. For disulfide bond-containing peptides, fragmentation was greatly enhanced following the reduction. After separation, the reduced linear peptide chains can be assigned to their precursors and used for sequencing based on ion dissociation. Both electrolysis treatment and online supercharging reactive-DESI experiments could be used to increase charges for the proteins carrying intra-disulfide bonds.

For the future works related LC/DESI-MS and EC/LC/DESI-MS, by applying specific chemistry, the protein reaction could be carried out, such as using selenium reagent. Then together with LC separation, and online EC reduction, protein or protein/protein complex 3D structure could be obtained. 114

In the research of PESI, PESI was used as a new interface for coupling EC/MS for its high salt tolerance. More electrochemical reactions employ a high concentration of salts as electrolyte could be explored. PESI-MS provides an easy and fast way to detect fractionated samples from isoelectric focusing. In this case, for further application, real biological samples can be separated by IEF and analyzed using PESI-MS.

115

REFERENCES

1. McLafferty, F., Tandem mass spectrometry. Science 1981, 214 (4518), 280-287.

2. McLafferty, F. W., Tandem mass spectrometry: From infancy to maturity in twenty-five years. Organic Mass Spectrometry 1993, 28 (12), 1403-1406.

3. McLafferty, F., Analytical information from mass spectrometry, past and future. J

Am Soc Mass Spectrom 1990, 1 (1), 1-5.

4. McLafferty, F. W., Tandem mass spectrometric analysis of complex biological mixtures. International Journal of Mass Spectrometry 2001, 212 (1–3), 81-87.

5. Aebersold, R.; Mann, M., Mass spectrometry-based proteomics. Nature 2003, 422

(6928), 198-207.

6. Beck, A.; Sanglier-Cianférani, S.; Van Dorsselaer, A., Biosimilar, Biobetter, and

Next Generation Characterization by Mass Spectrometry. Analytical Chemistry

2012, 84 (11), 4637-4646.

7. Bruins, A. P., Atmospheric-pressure-ionization mass spectrometry: II.

Applications in pharmacy, biochemistry and . TrAC Trends in

Analytical Chemistry 1994, 13 (2), 81-90.

8. Picó, Y.; Blasco, C.; Font, G., Environmental and food applications of LC– tandem mass spectrometry in pesticide-residue analysis: An overview. Mass

Spectrometry Reviews 2004, 23 (1), 45-85.

9. Zeleny, J., The Electrical Discharge from Liquid Points, and a Hydrostatic

Method of Measuring the Electric Intensity at Their Surfaces. Physical Review 1914, 3

(2), 69-91. 116

10. Munson, M. S. B.; Field, F. H., Chemical Ionization Mass Spectrometry. I.

General Introduction. Journal of the American Chemical Society 1966, 88 (12), 2621-

2630.

11. Carroll, D. I.; Dzidic, I.; Stillwell, R. N.; Haegele, K. D.; Horning, E. C.,

Atmospheric pressure ionization mass spectrometry. ion source for use in a liquid chromatograph-mass spectrometer-computer analytical system. Analytical

Chemistry 1975, 47 (14), 2369-2373.

12. Yamashita, M.; Fenn, J. B., Electrospray ion source. Another variation on the free-jet theme. The Journal of 1984, 88 (20), 4451-4459.

13. Karas, M.; Hillenkamp, F., Laser desorption ionization of proteins with molecular exceeding 10,000 daltons. Analytical Chemistry 1988, 60 (20), 2299-2301.

14. Ifa, D. R.; Wu, C.; Ouyang, Z.; Cooks, R. G., Desorption electrospray ionization and other ambient ionization methods: current progress and preview. Analyst 2010, 135

(4), 669-681.

15. Cooks, R. G.; Ouyang, Z.; Takats, Z.; Wiseman, J. M., Ambient Mass

Spectrometry. Science 2006, 311 (5767), 1566-1570.

16. Cech, N. B.; Enke, C. G., Practical implications of some recent studies in electrospray ionization fundamentals. Mass Spectrometry Reviews 2001, 20 (6), 362-387.

17. Takáts, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G., Mass Spectrometry

Sampling Under Ambient Conditions with Desorption Electrospray Ionization. Science

2004, 306 (5695), 471-473. 117

18. Shin, Y.-S.; Drolet, B.; Mayer, R.; Dolence, K.; Basile, F., Desorption

Electrospray Ionization-Mass Spectrometry of Proteins. Anal. Chem. (Washington, DC, U.

S.) 2007, 79, 3514-3518.

19. Kauppila, T. J.; Wiseman, J. M.; Ketola, R. A.; Kotiaho, T.; Cooks, R. G.;

Kostiainen, R., Desorption electrospray ionization mass spectrometry for the analysis of pharmaceuticals and metabolites. Rapid Commun. Mass Spectrom. 2005, 20, 387-392.

20. Chen, H.; Talaty, N. N.; Takats, Z.; Cooks, R. G., Desorption electrospray ionization mass spectrometry for high-throughput analysis of pharmaceutical samples in the ambient environment. Anal. Chem. 2005, 77, 6915-6927.

21. Rodriguez-Cruz, S. E., Rapid analysis of controlled substances using desorption electrospray ionization mass spectrometry. Rapid Commun. Mass Spectrom. 2005, 20,

53-60.

22. Talaty, N.; Takats, Z.; Cooks, R. G., Rapid in situ detection of in plant tissue under ambient conditions using desorption electrospray ionization. Analyst 2005,

130, 1624-1633.

23. Takats, Z.; Wiseman, J. M.; Cooks, R. G., Ambient mass spectrometry using desorption electrospray ionization (DESI): Instrumentation, mechanisms and applications in forensics, chemistry, and biology. J. Mass Spectrom. 2005, 40, 1261-1275.

24. Miao, Z.; Chen, H., Direct Analysis of Liquid Samples by Desorption

Electrospray Ionization-Mass Spectrometry (DESI-MS). J Am Soc Mass Spectrom 2009,

20 (1), 10-19. 118

25. Miao, Z.; Wu, S.; Chen, H., The Study of Protein Conformation in Solution Via

Direct Sampling by Desorption Electrospray Ionization Mass Spectrometry. J Am Soc

Mass Spectrom 2010, 21 (10), 1730-1736.

26. Zhang, Y.; Chen, H., Detection of saccharides by reactive desorption electrospray ionization (DESI) using modified phenylboronic acids. International Journal of Mass

Spectrometry 2010, 289 (2–3), 98-107.

27. Pan, N.; Liu, P.; Cui, W.; Tang, B.; Shi, J.; Chen, H., Highly efficient ionization of phosphopeptides at low pH by desorption electrospray ionization mass spectrometry.

Analyst 2013, 138, 1321-1324.

28. Li, J.; Dewald, H. D.; Chen, H., Online Coupling of Electrochemical Reactions with Liquid Sample Desorption Electrospray Ionization-Mass Spectrometry. Anal. Chem.

2009, 81 (23), 9716-9722.

29. Liu, Y.; Miao, Z.; Lakshmanan, R.; Loo, R. R. O.; Loo, J. A.; Chen, H., Signal and charge enhancement for protein analysis by liquid chromatography–mass spectrometry with desorption electrospray ionization. International Journal of Mass

Spectrometry 2012, 325–327, 161-166.

30. Ma, X.; Zhao, M.; Lin, Z.; Zhang, S.; Yang, C.; Zhang, X., Versatile platform employing desorption electrospray ionization mass spectrometry for high-throughput analysis. Anal. Chem. 2008, 80, 6131-6136.

31. Zhang, Y.; Dewald, H. D.; Chen, H., Online Mass Spectrometric Analysis of

Proteins/Peptides Following Electrolytic Cleavage of Disulfide Bonds. J. Res.

2011, 10 (3), 1293-1304. 119

32. Zhang, Y.; Cui, W.; Zhang, H.; Dewald, H. D.; Chen, H., Electrochemistry-

Assisted Top-Down Characterization of Disulfide-Containing Proteins. Analytical

Chemistry 2012, 84 (8), 3838-3842.

33. Zheng, Q.; Zhang, H.; Chen, H., Integration of online digestion and electrolytic reduction with mass spectrometry for rapid disulfide-containing protein structural analysis. International Journal of Mass Spectrometry 2013, 353, 84-92.

34. Zheng, Q.; Zhang, H.; Tong, L.; Wu, S.; Chen, H., Cross-Linking Electrochemical

Mass Spectrometry for Probing Protein Three-Dimensional Structures. Anal. Chem. 2014,

86 (18), 8983-8991.

35. Zheng, Q.; Zhang, H.; Wu, S.; Chen, H., Probing Protein 3D Structures and

Conformational Changes Using Electrochemistry-Assisted Isotope Labeling Cross-

Linking Mass Spectrometry. J Am Soc Mass Spectrom 2016, 27 (5), 864-875.

36. Hiraoka, K.; Nishidate, K.; Mori, K.; Asakawa, D.; Suzuki, S., Development of probe electrospray using a solid needle. Rapid Commun. Mass Spectrom. 2007, 21, 3139-

3144.

37. Mandal, M. K.; Chen, L. C.; Hashimoto, Y.; Yu, Z.; Hiraoka, K., Detection of biomolecules from solutions with high concentration of salts using probe electrospray and nano-electrospray ionization mass spectrometry. Analytical Methods 2010, 2 (12),

1905-1912.

38. Mandal, M. K.; Chen, L. C.; Yu, Z.; Nonami, H.; Erra-Balsells, R.; Hiraoka, K.,

Detection of protein from detergent solutions by probe electrospray ionization mass spectrometry (PESI-MS). J. Mass Spectrom. 2011, 46, 967-975. 120

39. Hiraoka, K.; Nishidate, K.; Mori, K.; Asakawa, D.; Suzuki, S., Development of probe electrospray using a solid needle. Rapid in Mass Spectrometry

2007, 21 (18), 3139-3144.

40. Chen, L. C.; Yoshimura, K.; Yu, Z.; Iwata, R.; Ito, H.; Suzuki, H.; Mori, K.;

Ariyada, O.; Takeda, S.; Kubota, T.; Hiraoka, K., Ambient imaging mass spectrometry by electrospray ionization using solid needle as sampling probe. J. Mass Spectrom. 2009,

44, 1469-1477.

41. Yu, Z.; Chen, L. C.; Erra-Balsells, R.; Nonami, H.; Hiraoka, K., Real-time reaction monitoring by probe electrospray ionization mass spectrometry. Rapid Commun.

Mass Spectrom. 2010, 24, 1507-1513.

42. Yu, Z.; Chen, L. C.; Suzuki, H.; Ariyada, O.; Erra-Balsells, R.; Nonami, H.;

Hiraoka, K., Direct profiling of phytochems. in tulip tissues and in vivo monitoring of change of carbohydrate content in tulip bulbs by probe electrospray ionization mass spectrometry. J. Am. Soc. Mass Spectrom. 2009, 20, 2304-2311.

43. Yoshimura, K.; Chen, L. C.; Yu, Z.; Hiraoka, K.; Takeda, S., Real-time analysis of living animals by electrospray ionization mass spectrometry. Anal. Biochem. 2011,

417, 195-201.

44. Gong, X.; Zhao, Y.; Cai, S.; Fu, S.; Yang, C.; Zhang, S.; Zhang, X., Single Cell

Analysis with Probe ESI-Mass Spectrometry: Detection of Metabolites at Cellular and

Subcellular Levels. Anal. Chem. 2014, 86, 3809-3816. 121

45. Volk, K. J.; Yost, R. A.; Brajter-Toth, A., Electrochemistry on line with mass spectrometry. Insight into biological redox reactions. Anal. Chem. 1992, 64, 21A-26A,

28A, 30A-33A.

46. Deng, H.; Van Berkel, G. J., A thin-layer electrochemical flow cell coupled online with electrospray-mass spectrometry for the study of biological redox reactions.

Electroanalysis 1999, 11, 857-865.

47. Permentier, H. P.; Bruins, A. P.; Bischoff, R., Electrochemistry-mass spectrometry in drug metabolism and protein research. Mini-Rev. Med. Chem. 2008, 8,

46-56.

48. Lohmann, W.; Karst, U., Biomimetic modeling of oxidative drug metabolism.

Anal. Bioanal. Chem. 2008, 391, 79-96.

50. Bruckenstein, S.; Gadde, R. R., Use of a porous electrode for in situ mass spectrometric determination of volatile electrode reaction products. J. Amer. Chem. Soc.

1971, 93, 793-4.

51. Hambitzer, G.; Heitbaum, J., Electrochemical thermospray mass spectrometry.

Anal. Chem. 1986, 58, 1067-70.

52. Bartmess, J. E.; Phillips, L. R., Electrochemically assisted fast atom bombardment mass spectrometry. Anal. Chem. 1987, 59, 2012-14.

53. Zhou, F.; Van Berkel, G. J., Electrochemistry Combined Online with Electrospray

Mass Spectrometry. Anal. Chem. 1995, 67, 3643-9.

54. Blades, A. T.; Ikonomou, M. G.; Kebarle, P., Mechanism of electrospray mass spectrometry. Electrospray as an electrolysis cell. Anal. Chem. 1991, 63, 2109-14. 122

55. Vandell, V. E.; Limbach, P. A., Electrospray ionization mass spectrometry of metalloporphyrins. J. Mass Spectrom. 1998, 33, 212-220.

56. Dupont, A.; Gisselbrecht, J.-P.; Leize, E.; Wagner, L.; Van Dorsselaer, A.,

Electrospray mass spectrometry of electrochemically ionized molecules: application to the study of . Tetrahedron Lett. 1994, 35, 6083-6.

57. Castagnoli, N.; Bissel, P.; Jurva, U.; Ashraf-Khorasani, M., Electrochemical mass spectrometric studies on 1-cyclopropyl-4-phenyl-1,2,3,6-tetrahydropyridine. Rapid

Commun. Mass Spectrom. 2008, 22, 2089-2096.

58. Faber, H.; Vogel, M.; Karst, U., Electrochemistry/mass spectrometry as a tool in metabolism studies-A review. Anal. Chim. Acta 2014, 834, 9-21.

59. Roussel, C.; Rohner, T. C.; Jensen, H.; Girault, H. H., Mechanistic aspects of on- line electrochemical tagging of free L-cysteine residues during electrospray ionisation for mass spectrometry in protein analysis. ChemPhysChem 2003, 4, 200-206.

60. Lu, M.; Liu, Y.; Helmy, R.; Martin, G. E.; Dewald, H. D.; Chen, H., Online

Investigation of Aqueous-Phase Electrochemical Reactions by Desorption Electrospray

Ionization Mass Spectrometry. J. Am. Soc. Mass Spectrom. 2015, 26, 1676-1685.

61. Brown, T. A.; Chen, H.; Zare, R. N., Detection of the Short-Lived Radical Cation

Intermediate in the Electrooxidation of N,N-Dimethylaniline by Mass Spectrometry.

Angew. Chem., Int. Ed. 2015, 54, 11183-11185.

62. Brown, T. A.; Chen, H.; Zare, R. N., Identification of Fleeting Electrochemical

Reaction Intermediates Using Desorption Electrospray Ionization Mass Spectrometry. J.

Am. Chem. Soc. 2015, 137, 7274-7277. 123

63. Zhang, Y.; Yuan, Z.; Dewald, H. D.; Chen, H., Coupling of liquid chromatography with mass spectrometry by desorption electrospray ionization (DESI).

Chem. Commun. 2011, 47 (14), 4171-4173.

64. Schiavo, S.; Ebbel, E.; Sharma, S.; Matson, W.; Kristal, B. S.; Hersch, S.; Vouros,

P., Metabolite Identification Using a Nanoelectrospray LC-EC-array-MS Integrated

System. Anal. Chem. 2008, 80, 5912-5923.

65. Andrews, C. L.; Li, F.; Yang, E.; Yu, C.-P.; Vouros, P., Incorporation of a nanosplitter interface into an LC-MS-RD system to facilitate drug metabolism studies. J.

Mass Spectrom. 2006, 41, 43-49.

66. Camenzuli, M.; Goodie, T. A.; Bassanese, D. N.; Francis, P. S.; Barnett, N. W.;

Ritchie, H.; LaDine, J.; Shalliker, R. A.; Conlan, X. A., The use of parallel segmented outlet flow columns for enhanced mass spectral sensitivity at high chromatographic flow rates. Rapid Commun. Mass Spectrom. 2012, 26, 943-949.

67. Andrews, C. L.; Yu, C.-P.; Yang, E.; Vouros, P., Improved liquid chromatography-mass spectrometry performance in quantitative analysis using a nanosplitter interface. J. Chromatogr. A 2004, 1053, 151-159.

68. Ferguson, C. N.; Benchaar, S. A.; Miao, Z.; Loo, J. A.; Chen, H., Direct

Ionization of Large Proteins and Protein Complexes by Desorption Electrospray

Ionization-Mass Spectrometry. Analytical Chemistry 2011, 83 (17), 6468-6473.

69. Liu, P.; Zheng, Q.; Dewald, H. D.; Zhou, R.; Chen, H., The study of electrochemistry with ambient mass spectrometry. TrAC, Trends Anal. Chem. 2015, 70,

20-30. 124

70. Lu, M.; Wolff, C.; Cui, W.; Chen, H., Investigation of some biologically relevant redox reactions using electrochemical mass spectrometry interfaced by desorption electrospray ionization. Analytical and Bioanalytical Chemistry 2012, 403 (2), 355-365.

71. Sun, X.; Miao, Z.; Yuan, Z.; Harrington, P. d. B.; Colla, J.; Chen, H., Coupling of single droplet micro-extraction with desorption electrospray ionization-mass spectrometry. Int. J. Mass Spectrom. 2011, 301, 102-108.

72. Perry, R. H.; Splendore, M.; Chien, A.; Davis, N. K.; Zare, R. N., Detecting

Reaction Intermediates in Liquids on the Millisecond Time Scale Using Desorption

Electrospray Ionization. Angew. Chem., Int. Ed. 2011, 50, 250-254, S250/1-S250/3.

73. Chen, H.; Cotte-Rodriguez, I.; Cooks, R. G., cis-Diol recognition by reactive desorption electrospray ionization (DESI). Chem. Commun. 2006, 597-599.

74. Takats, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G., Electrosonic Spray

Ionization. A Gentle Technique for Generating Folded Proteins and Protein Complexes in the Gas Phase and for Studying Ion-Molecule Reactions at Atmospheric Pressure. Anal.

Chem. 2004, 76, 4050-4058.

75. Zubarev, R. A.; Kelleher, N. L.; McLafferty, F. W., Electron Capture Dissociation of Multiply Charged Protein Cations. A Nonergodic Process. J. Am. Chem. Soc. 1998,

120, 3265-3266.

76. Stutzman, J. R.; McLuckey, S. A., Ion/Ion Reactions of MALDI-Derived Peptide

Ions: Increased Sequence Coverage via Covalent and Electrostatic Modification upon

Charge Inversion. Anal. Chem. 2012, 84, 10679-10685. 125

77. Streuli, C. A., in nonaqueous solvents. Anal. Chem. 1964, 36, 363R-

369R.

78. Sklenarova, H.; Chocholous, P.; Koblova, P.; Zahalka, L.; Satinsky, D.; Matysova,

L.; Solich, P., High-resolution monolithic columns-a new tool for effective and quick separation. Anal. Bioanal. Chem. 2013, 405, 2255-2263.

79. Guillarme, D.; Ruta, J.; Rudaz, S.; Veuthey, J.-L., New trends in fast and high- resolution liquid chromatography: a critical comparison of existing approaches. Anal.

Bioanal. Chem. 2010, 397, 1069-1082.

80. Guiochon, G., Monolithic columns in high-performance liquid chromatography. J.

Chromatogr. A 2007, 1168, 101-168.

81. Nguyen, D. T. T.; Guillarme, D.; Rudaz, S.; Veuthey, J.-L., Fast analysis in liquid chromatography using small particle size and high pressure. J. Sep. Sci. 2006, 29, 1836-

1848.

82. Cai, Y.; Adams, D.; Chen, H., A New Splitting Method for Both Analytical and

Preparative LC/MS. J. Am. Soc. Mass Spectrom. 2014, 25, 286-292.

83. Fernandez, P.; Seoane, S.; Vazquez, C.; Bermejo, A. M.; Carro, A. M.; Lorenzo,

R. A., A rapid analytical method based on -assisted extraction for the determination of drugs of abuse in vitreous humor. Anal. Bioanal. Chem. 2011, 401,

2177-2186.

84. Seiwert, B.; Henneken, H.; Karst, U., Ferrocenoyl piperazide as derivatizing agent for the analysis of isocyanates and related compounds using liquid 126 chromatography/electrochemistry/mass spectrometry (LC/EC/MS). J. Am. Soc. Mass

Spectrom. 2004, 15, 1727-1736.

85. Hayen, H.; Karst, U., Analysis of phenothiazine and its derivatives using

LC/electrochemistry/MS and LC/electrochemistry/fluorescence. Anal. Chem. 2003, 75,

4833-4840.

86. van Leeuwen, S. M.; Hayen, H.; Karst, U., Liquid chromatography- electrochemistry-mass spectrometry of polycyclic aromatic hydrocarbons. Anal. Bioanal.

Chem. 2004, 378, 917-925.

87. Karst, U., Analytical methods: Electrochemistry/mass spectrometry (EC/MS) - a new tool to study drug metabolism and reaction mechanisms. Angew. Chem., Int. Ed.

2004, 43, 2476-2478.

88. Zhang, Y.; Dewald, H. D.; Chen, H., Online Mass Spectrometric Analysis of

Proteins/Peptides Following Electrolytic Cleavage of Disulfide Bonds. J. Proteome Res.

2011, 10, 1293-1304.

89. Kraj, A.; Brouwer, H.-J.; Reinhoud, N.; Chervet, J.-P., A novel electrochemical method for efficient reduction of disulfide bonds in peptides and proteins prior to MS detection. Anal. Bioanal. Chem. 2013, 405, 9311-9320.

90. Nicolardi, S.; Giera, M.; Kooijman, P.; Kraj, A.; Chervet, J.-P.; Deelder, A. M.; van der Burgt, Y. E. M., On-Line Electrochemical Reduction of Disulfide Bonds:

Improved FTICR-CID and -ETD Coverage of Oxytocin and Hepcidin. J. Am. Soc. Mass

Spectrom. 2013, 24, 1980-1987. 127

91. Iavarone, A. T.; Williams, E. R., Mechanism of Charging and Supercharging

Molecules in Electrospray Ionization. J. Am. Chem. Soc. 2003, 125, 2319-2327.

92. Sterling, H. J.; Cassou, C. A.; Trnka, M. J.; Burlingame, A. L.; Krantz, B. A.;

Williams, E. R., The role of conformational flexibility on protein supercharging in native electrospray ionization. Phys. Chem. Chem. Phys. 2011, 13, 18288-18296.

93. Zhang, J.; Loo, R. R. O.; Loo, J. A., Increasing fragmentation of disulfide-bonded proteins for top-down mass spectrometry by supercharging. Int. J. Mass Spectrom. 2015,

377, 546-556.

94. Bond, A. M.; Colton, R.; D'Agostino, A.; Downard, A. J.; Traeger, J. C., A Role for Electrospray Mass Spectrometry in Electrochemical Studies. Anal. Chem. 1995, 67,

1691-5.

95. Liu, P.; Lanekoff, I. T.; Laskin, J.; Dewald, H. D.; Chen, H., Study of

Electrochemical Reactions Using Nanospray Desorption Electrospray Ionization Mass

Spectrometry. Anal. Chem. 2012, 84, 5737-5743.

96. Smoluch, M.; Mielczarek, P.; Reszke, E.; Hieftje, G. M.; Silberring, J.,

Determination of psychostimulants and their metabolites by electrochemistry linked on- line to flowing atmospheric pressure afterglow mass spectrometry. Analyst 2014, 139,

4350-4355.

97. Bischoff, R.; Karst, U., Electrochemistry-mass spectrometry: Fundamentals and applications in pharmaceutical and environmental sciences. TrAC, Trends Anal. Chem.

2015, 70, 3. 128

98. Shiea, J.; Huang, M.-Z.; Hsu, H.-J.; Lee, C.-Y.; Yuan, C.-H.; Beech, I.; Sunner, J.,

Electrospray-assisted laser desorption/ionization mass spectrometry for direct ambient analysis of solids. Rapid Communications in Mass Spectrometry 2005, 19 (24), 3701-

3704.

99. Jeng, J.; Lin, C.-H.; Shiea, J., Electrospray from Nanostructured Tungsten Oxide

Surfaces with Ultralow Sample Volume. Anal. Chem. 2005, 77, 8170-8173.

100. Hong, C.-M.; Lee, C.-T.; Lee, Y.-M.; Kuo, C.-P.; Yuan, C.-H.; Shiea, J.,

Generating electrospray from solutions predeposited on a wire. Rapid Commun.

Mass Spectrom. 1999, 13, 21-25.

101. Lee Chuin, C.; Zhan, Y. U.; Hiroshi, N.; Yutaka, H.; Kenzo, H., Application of

Probe Electrospray Ionization for Biological Sample Measurements. Environ. Control

Biol. 2009, 47 (2), 73-86.

102. Laskin, J.; Heath, B. S.; Roach, P. J.; Cazares, L.; Semmes, O. J., Tissue Imaging

Using Nanospray Desorption Electrospray Ionization Mass Spectrometry. Anal. Chem.

2012, 84, 141-148.

103. Liu, P.; Forni, A.; Chen, H., Development of solvent-free ambient mass spectrometry for applications. Anal Chem 2014, 86, 4024-32.

104. Jackson, G. P.; Duckworth, D. C., Electrospray mass spectrometry of undiluted ionic liquids. Chem. Commun. 2004, (5), 522-523.

105. Dyson, P. J.; Khalaila, I.; Luettgen, S.; McIndoe, J. S.; Zhao, D., Direct probe electrospray (and nanospray) ionization mass spectrometry of neat ionic liquids. Chem.

Commun. 2004, (19), 2204-2205. 129

APPENDIX: PUBLICATIONS

1. Yi Cai, Daniel Adams and Hao Chen*, A New Splitting Method for Both Analytical and Preparative LC/MS, J. Am. Soc. Mass Spectrom. 2014, 25, 286-292.

2. Yi Cai, Yong Liu, Roy Hemly and Hao Chen*, Coupling of Ultrafast LC Separation with Mass Spectrometry by DESI, J. Am. Soc. Mass Spectrom. 2014, 25, 1820-1823.

3. Yi Cai, Qiuling Zheng, Yong Liu, Roy Helmy, Joseph A. Loo and Hao Chen*,

Integration of Electrochemistry with Ultra Performance Liquid Chromatography/Mass

Spectrometry (UPLC/MS), Eur. J. Mass Spectrom. 2015, 21, 341–351.

4. Yi Cai, Pengyuan Liu, Michael Held, Howard Dewald, Hao Chen*, Coupling

Electrochemistry with Probe Electrospray Ionization Mass Spectrometry (PESI-MS),

ChemPhysChem. 2016, 8, 1104–1108.

5. Si Cheng, Jun Wang,* Yi Cai, Joseph A. Loo,* and Hao Chen*, Enhancing

Performance of Liquid Sample Desorption Electrospray Ionization Mass Spectrometry

Using Trap and Capillary Columns, Int. J. Mass Spectrom. 2015, 392, 73-79.

6. Zhi Li, * Shuaihua Zhang, Yi Cai, Qiuhua Wu, Hao Chen*, Hollow fiber-based solid- liquid phase microextraction combined with theta capillary electrospray ionization mass spectrometry for sensitive and accurate analysis of methamphetamine, Anal. Methods, submitted. ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

! ! Thesis and Dissertation Services