<%BANNER%>

Phosphoproteomics of Arabadopsis thaliana


PAGE 1

PHOSPHOPROTEOMICS OF Arabidopsis thaliana By CAMILLE STRACHAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Camille Nicola Strachan

PAGE 3

This is dedicated to my parents, John and Ma rjorie Strachan, who always gave me their loving support and encouragem ent; to my second parents, Leon and Daphney Strachan, for their love and support throughout the year s; and to my brothers, Wayne, Ken and Stephen, for their love and friendship.

PAGE 4

iv ACKNOWLEDGMENTS This work would not have been possible without the help of many individuals, namely, my three supervisory committee memb ers: Dr. James Winefordner, Dr. Nancy Denslow, and Dr. Alice Harmon. I thank Dr. Winefordner for welcoming me into his group and allowing me to do a collaboration project even though it was outside of his field. Despite the fact that this project was not in his area, he always made the effort to gain an understanding of proteomics. Being a part of his group truly meant a lot to me and it was an inspiration to see his determin ation to keep his research group family by planning group trips. I thank Dr. Denslow for taking the time to meet with me as I tried to find my ideal project (application of mass spectrometry to a biological project) which at the time seemed impossible to find. She discussed the various projects available and welcomed me into her lab to begin what seemed an im possible but exciting tas k. I am truly grateful for her invaluable guidance as I entered the world of research, an area with which I had very limited experience. Her never-ending en thusiasm and encouragement always gave me hope that anything is possible; you just have to figure out th e right approach. How many approaches do I have to try, to figure out the right one? Her unending enthusiasm showed me that in research you have to keep trying until you find it; and once you do, youll be guaranteed one of her famous ha ndshakes in congratulations! Her passion for science was contagious and I will always try to emulate it. The knowledge she imparted to me in our weekly group meetings was invalu able and for that I am also truly grateful. I

PAGE 5

v am also thankful to her for starting me on the collaboration project with Dr. Harmon, to whom I am eternally grateful. Dr. Harmons unending patience with me (as I tried to grasp the biological concepts behind the projects) and with my never-ending questions about biological sample preparation was greatly appreciated. During my initial stages of working in her lab, she actually took the time to perform se veral experiments with me. I am eternally grateful to her for those learning experiences and the privilege of performing experiments with her. Working in her lab was a tremendous pleasure as I watched her enthusiasm for research and her patience with her students as she explained various approaches and also performed various experiments with them. She always made time to sit with me and discuss any results or approaches that I had questions about. She was an advisor and also a mentor, always lending a listening ear and giving advice as I vented my frustrations with research or daily life. She was a true inspiration to me and has encouraged me to continue my career in research and devel opment. I am eternally indebted to her. I also owe my gratitude to members of the three groups that I am associated with. I thank the Winefordner group for their frie ndship and for always making me feel welcomed even though I was not based in the Winefordner lab. I am grateful to Dr. Ben Smith and Dr. Nico Omenetto for their input and interest in group meetings. I thank the Protein Core Facility for the use of its mass spectrometers and for always making me feel welcomed in the lab. Most especially I tha nk Dr. Stanley Stevens and Marjorie Chow. I thank Dr. Stevens for sharing his knowle dge of mass spectrometry and database searching, as well as his renditions of catchy tunes during a frustrating day of lab work. I am truly grateful to Marjorie Chow for her tremendous help and support with sample-

PAGE 6

vi preparation techniques, and fo r always giving me a listening ear as I vented about one thing or another. I thank Scott McMill en for MALDI-TOF-MS training. I extend my gratitude to Alfred Chung fo r giving me insight into synthesis and for always encouraging me to choose the best career pa th. I am truly grateful to Scott McLung for always ensuring that the QITMS was running optimally. Our friendly banter in the lab always brought a smile to my face, especia lly with the constant chocolate supply that Scott always provided. I also thank Marg aret Joyner from the Harmon lab for her constant encouragement and reminiscing conve rsations, and for alwa ys ensuring that I had anything I needed in th e lab to achieve my goals. Last but not least, I would like to extend my gratitude to my family and friends for their unending love, support, encouragement, a nd patience. Each day I give thanks for my good fortune of having such a wonderful family. My mothers unending strength and determination are only a few of the qualities that I try to emulate. My fathers organizational qualities and atten tion to detail I hope to one day be able to achieve. I wish he could be here to see what I have achieved. I know that he would be proud. The presence of newly made friends and colleague s made my time here pleasurable. Finally, I thank my Heavenly Father for giving me the ab ility to achieve all that I have achieved so far. Without Him I would never have gotten th is far. He has given me the strength and perseverance to strive for excellence even when times we re rough. If there is anyone I have left out, I thank them.

PAGE 7

vii TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...............................................................................................................x LIST OF FIGURES...........................................................................................................xi ABSTRACT.....................................................................................................................xv i CHAPTER 1 INTRODUCTION........................................................................................................1 Calcium-Dependent Protein Kinases............................................................................1 14-3-3s........................................................................................................................ ..6 Detection and Analysis of Protein Phosphorylation.....................................................7 Mass Spectrometric-based Methods......................................................................9 Ionization techniques....................................................................................10 Mass analysis................................................................................................13 Mass analysis of peptides and proteins........................................................19 Data interpretation........................................................................................23 Isolation and enrichment techniques............................................................26 Immobilized metal-ion affinity chromatography.........................................26 Chemical derivatization................................................................................28 Qiagen phosphoprotein purification kit........................................................28 2 PHOSPHORYLATION DETECTION AND ENRICHMENT.................................42 Experimental Methods................................................................................................44 Materials and Instruments...................................................................................44 Phosphoprotein Detection...................................................................................45 Matrix Optimization............................................................................................45 Phosphopeptide Enrichment................................................................................46 Immobilized metal-ion affinity chromatography.........................................46 Chemical derivatization................................................................................47 Results and Discussion...............................................................................................47 Phosphoprotein Detection...................................................................................47 Matrix Optimization............................................................................................48 Phosphopeptide Enrichment................................................................................48

PAGE 8

viii 3 INVESTIGATING THE AUTOPHO SPHORYLATION SITES OF A CALCIUM-DEPENDENT PROTEIN KINASE.......................................................61 Experimental Methods................................................................................................62 Materials and Instruments...................................................................................62 Kinase Assay and Protein Preparation................................................................63 Phosphopeptide Enrichment................................................................................63 -Elimination.......................................................................................................64 Data-dependent LC/MS/MS on the QIT-MS......................................................64 MALDI-TOF-MS Analysis of the -eliminated Digests.....................................65 Precursor Scanning..............................................................................................65 Data-dependent LC/MS/MS on the QqTOF-MS................................................66 Results and Discussion...............................................................................................66 Kinase Autophosphorylation...............................................................................66 Ion Trap: Data-dependent LC/MS/MS of the Autophosphorylated CPK5 Digest...............................................................................................................66 MALDI-TOFMS.................................................................................................67 QqTOF-MS analysis............................................................................................68 4 IDENTIFICATION OF IN VITRO SUBSTRATES OF A CALCIUMDEPENDENT PROTEIN KINASE...........................................................................78 Experimental Methods................................................................................................80 Materials and Instruments...................................................................................80 Method Development..........................................................................................81 Protein extract preparation...........................................................................81 Dephosphorylation of the protein extract.....................................................82 Phosphatase inhibition..................................................................................82 In vitro phosphorylation of Arabidopsis thaliana extract with CDPK4......83 Phosphoprotein enrichment..........................................................................83 Data-dependent LC/MS/MS on the ion trap................................................84 Results and Discussion...............................................................................................84 5 14-3-3 INTERACTORS FROM ARABIDIOPSIS THALIANA...............................95 Experimental Methods................................................................................................96 Materials and Instruments...................................................................................96 Protein Extract Preparation..................................................................................96 14-3-3 Affinity Purification.................................................................................97 Amino Acid Sequencing by nanoESI QqTOF MS Analysis...............................98 Protein Identification...........................................................................................98 Results and Discussion...............................................................................................99

PAGE 9

ix 6 HIGH-THROUGHPUT PHOSPHOPROTEOMICS OF ARABIDOPSIS THALIANA..............................................................................................................121 Experimental Methods..............................................................................................124 Materials and Instruments.................................................................................124 Sample Preparation............................................................................................125 2-Dimensional Gel Electrophoresis...................................................................126 Protein Visualization and Analysis...................................................................126 Automated Spot Picking and Digestion............................................................127 Amino Acid Sequencing by nanoESI QqTOF MS Analysis.............................127 Protein Identification.........................................................................................127 Results and Discussion.............................................................................................128 Protein Visualization and Analysis...................................................................128 Automated Spot Picking and Digestion............................................................128 Protein Identification.........................................................................................128 7 RESEARCH OVERVIEW.......................................................................................147 LIST OF REFERENCES.................................................................................................152 BIOGRAPHICAL SKETCH...........................................................................................161

PAGE 10

x LIST OF TABLES Table page 2-1 Matrix solvent systems for evaluation.....................................................................46 4-1 List of proteins identified in only the CDPK4 treated sample.................................94 5-1 Proteins identified from 143-3 affinity chromatography of Arabidopsis thaliana proteins...................................................................................................................114 6-1 List of proteins iden tified from 2-DE spots...........................................................143 7-1 List of proteins identified as both CDPK substrates and 14-3-3 interactors..........150 7-2 Correlating proteins identified in 2-DE and 14-3-3 experiments...........................151

PAGE 11

xi LIST OF FIGURES Figure page 1-1 The MALDI process.................................................................................................30 1-2 Commonly used MALDI matrices for peptide and protein analysis.......................31 1-3 Taylor cone formation..............................................................................................31 1-4 The electrospray ionization process.........................................................................32 1-5 A linear time-of-fli ght mass spectrometer...............................................................32 1-6 A reflectron time-of-flight mass spectrometer.........................................................33 1-7 Three-dimensional ideal ion trap showing the asymptotes and the dimensions r0 and z0....................................................................................................................33 1-8 Trajectory of a trapped ion.......................................................................................34 1-9 Mathieu stability diagram.........................................................................................35 1-10 Tandem QqTOF mass spectrometer.........................................................................36 1-11 Peptide fragmentation..............................................................................................37 1-12 Peptide mass mapping..............................................................................................38 1-13 Tandem mass spectrometric sequencing..................................................................39 1-14 Two of the most frequently used chelating ligands for IMAC................................40 1-15 Chemical derivatization method for the affinity purification of a phosphorylated peptide......................................................................................................................41 2-1 One-Dimensional gel electrophores is of four protein standards..............................53 2-2 Pro-Q stained gel of varying con centrations of protein standards...........................54 2-3 Comparison of matrix cond itions for phosphopeptide spectra.................................55 2-4 -casein digest spectrum..........................................................................................56

PAGE 12

xii 2-5 Comparison of different metal ions on the ZipTipMC for isolation of -casein phosphopeptides. A) Fe3+. B) Ga3+. C) Cu2+. D) Ni3+..............................................57 2-6 Spectra of -casein phosphopeptides isolated by the Pierce Phosphopeptide Isolation Kit..............................................................................................................58 2-7 Chemical derivatization of phosphopeptide.............................................................59 2-8 Spectrum of chemical derivatization label...............................................................60 3-1 Autoradiography of SDS-PAGE se parated autophosphorylated CPDKs................70 3-2 Autophosphorylated CPK5 digest............................................................................71 3-3 IMAC enriched phosphopeptides fr om autophosphorylated CPK5 digest..............72 3-4 Spectra of -casein digest.........................................................................................73 3-5 Spectra of CPK5 digest............................................................................................74 3-6 Precursor ion scan (400 500 m/z ) of autophosphorylated CPK5 digest.................75 3-7 MS/MS spectrum and correspondin g MASCOT search result of the phosphopeptide TSTTNLSSNSDHSPNAADIIAQEFSK ( m/z 939).......................76 3-8 Overlap of phosphopeptides identified by all methods............................................77 4-1 Pro-Q Diamond Phosphoprotein Gel Stain of Arabidopsis thaliana extract dephosphorylation with phosphatases......................................................................89 4-2 Inhibiting Calf Intestinal Alkaline Phosphatase.......................................................90 4-3 Steps for identifying in vitro CDPK4 substrates from Arabidopsis thaliana ..........91 4-4 SDS gel electrophoresis separation of the control and CDPK4 treated Qiagen samples.....................................................................................................................92 4-5 Base peak chromatogram and full MS spectrum of tryptic digest of band 20,21....93 5-1 General procedure for affinity chromatography.....................................................100 5-2 Gel images of 14-3-3 affinity purified Arabidopsis thaliana proteins...................101 5-3 MASCOT search results of the phosphopeptide NAGpSRLVVR (m/z 1050.5895) from photosystem I subunit PSI-E-like protein (gi|7269730) identified in band 2.................................................................................................102 5-4 MASCOT search results of th e phosphopeptide QERFSQILpTPR (m/z 1453.7791) from Nuf2 family protein (gi|15219846) identified in band 3............102

PAGE 13

xiii 5-5 MASCOT search results of th e phosphopeptide SRLSSAAAKPSVpTA (m/z 1424.7812) from ribosomal protein S6 (gi|2662469) identified in band 4............103 5-6 MASCOT search results of the phosphopeptide AMAVpSGAVLSGIGSSFLpTGGKR (m/z 2225.2202) from Lhcb6 protein (gi|4741960) identified in band 4...........................................................................103 5-7 MASCOT search results of the phosphopeptide pTWEKLQMAAR (m/z 1312.7833) from laminin receptor homologue (gi|16380) identified in band 6.....104 5-8 MASCOT search results of th e phosphopeptide LEAIEpTAK (m/z 953.5758) from cysteine synthase (gi| 1488519) identified in band 6.....................................104 5-9 MASCOT search results of th e phosphopeptide SRLpSSAAAKPSVTA (m/z 1424.7800) from ribosomal protein S6 (gi|2662469) identified in band 7............105 5-10 MASCOT search results of th e phosphopeptide LpSSAAAKPSVTA (m/z 1181.6309) from ribosomal protein S6-lik e (gi|7270073) identified in band 8.....105 5-11 MASCOT search results of th e phosphopeptide SRLpSSAAAKPSVTA (m/z 1424.7857) from ribosomal protein S6-lik e (gi|7270073) identified in band 8.....106 5-12 MASCOT search results of the phosphopeptide LpSSAPAKPVAA (m/z 1090.6041) from ribosomal protein S6 (gi|2224751) identified in band 8............106 5-13 MASCOT search results of the phosphopeptide pSRLSSAPAKPVAA (m/z 1333.7495) from ribosomal protein S6 (gi|2224751) identified in band 8............107 5-14 MASCOT search results of th e phosphopeptide SLGGSRPGLPpTGR (m/z 1333.7944) from unknown (gi|21592536) identified in band 8.............................107 5-15 MASCOT search results of th e phosphopeptide IKLPSGpSK (m/z 908.6211) from 60S ribosomal protein L2 (gi|22135870) identified in band 10....................108 5-16 MASCOT search results of th e phosphopeptide IKLPSGpSK (m/z 908.6211) from putative ribosomal protein L8 (gi|7270565) identified in band 10................108 5-17 MASCOT search results of the phosphopeptide AESLNPLNFpSSSKPK (m/z 1697.9166) from ATP-dependent Clp prot ease proteolytic subunit ClpR4, putative (gi|21593086) identified in band 11.........................................................109 5-18 MASCOT search results of the phosphopeptide ALVTLIEKGVAFEpTIPVDLMK (m/z 2366.3912) from Glutathione Stransferase (gi|27363352) identified in band 12.....................................................109 5-19 MASCOT search results of the phosphopeptide KVEMLDGVpTIVR (m/z 1454.8631) from putative ribosomal protei n L9 (gi|12642868) identified in band 12............................................................................................................................1 10

PAGE 14

xiv 5-20 MASCOT search results of th e phosphopeptide LApTGEPLR (m/z 935.5762) from putative protein (gi| 7573368) identified in band 12......................................110 5-21 MASCOT search results of the phosphopeptide SFGLDSpSQAR (m/z 1146.6510) from putative protein 1 photosystem II oxygen-evolving complex (gi|4835233) identified in band 13.........................................................................111 5-22 MASCOT search results of the phosphopeptide IGTADVLAFFLPGVVpSQVFK (m/z 2187.3049) from unknown protein (gi|3152582) identified in band 14.........................................................................111 5-23 MASCOT search results of th e phosphopeptide SAGSVGKSAGpSEK (m/z 1243.7179) from putative TNP1-like trans poson protein (gi|4734013) identified in band 14...............................................................................................................112 5-24 MASCOT search results of th e phosphopeptide SpSGIALpSSRLHYASPIK (m/z 1946.0432) from peptidylprolyl isomer ase ROC4 (gi|6899901) identified in band 15...................................................................................................................112 5-25 MASCOT search results of the phosphopeptide IDCEpSACVAR (m/z 1259.5482) from GAST1-like protein (gi| 21618022) identified in band 19..........113 6-1 Pro-Q Diamond Phosphoprotein Gel Stai n image indicating potential phosphorylated proteins from a two-dime nsional gel electrophoresis separation of Arabidopsis thaliana protein extract..................................................................131 6-2 Colloidal Gel Stain image indicating a ll proteins from a two-dimensional gel electrophoresis separation of Arabidopsis thaliana protein extract.......................132 6-3 InvestigatorTM ProPicTM image of Colloidal staine d gel and corresponding spots for excision.............................................................................................................133 6-4 MS/MS spectrum and correspondin g MASCOT search results of the phosphopeptide QTGpSLYpSDWDLLPAK ( m/z 1852.9480) from the unknown protein (gi|30725696) iden tified in spot 1..............................................................134 6-5 MS/MS spectrum and correspondin g MASCOT search results of the phosphopeptide QpTGSLYpSDWDLLPAK ( m/z 1852.9516) from the unknown protein (gi|30725696) iden tified in spot 2..............................................................135 6-6 MS/MS spectrum and correspondin g MASCOT search results of the phosphopeptide SGpSGDDEEGSYGR ( m/z 1394.4946) from the unknown protein (gi|23308191) iden tified in spot 3..............................................................136 6-7 MS/MS spectrum and correspondin g MASCOT search results of the phosphopeptide NRSGpSGDDEEGSYGR ( m/z 1665.6520) from the unknown protein (gi|23308191) iden tified in spot 3..............................................................137

PAGE 15

xv 6-8 MS/MS spectrum and correspondin g MASCOT search results of the phosphopeptide pTTGEEEKK ( m/z 1000.5043) from the low temperatureinduced protein (gi|509262) identified in spot 8....................................................138 6-9 MS/MS spectrum and correspondin g MASCOT search results of the phosphopeptide SFGLDSpSQAR ( m/z 1146.6510) from putative protein 1 photosystem II oxygen-evolving complex (g i|4835233) identified in spot 13......139 6-10 MS/MS spectrum and correspondin g MASCOT search results of the phosphopeptide GTGTANQCPpTI DGGSETFSFKPGKYAGK ( m/z 2955.3193) from the 33 kDa polypeptide of oxygen-evolving complex (gi|10177538) identified in spot 14................................................................................................140 6-11 MS/MS spectrum and correspondin g MASCOT search results of the phosphopeptide SPASDpTYVIFGEAK ( m/z 1563.7827) from the unknown protein (gi|48310641) iden tified in spot 15............................................................141 6-12 MS/MS spectrum and correspondin g MASCOT search results of the phosphopeptide RSPpSPPPAR ( m/z 1043.5207) from the RSZp22 protein (gi|2582645) identified in spot 19..........................................................................142

PAGE 16

xvi Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PHOSPHOPROTEOMICS OF Arabidopsis thaliana By Camille Nicola Strachan August 2005 Chair: James D. Winefordner Cochair: Alice C. Harmon Major Department: Chemistry Reversible protein phosphoryla tion on serine, threonine a nd tyrosine residues is one of the most common and important regulatory modifications of intracellular proteins, playing a role in many biological and biom edical phenomena such as cellular signal transduction, cell growth, cell differentiati on, cell division, metabolism and cancer. Mass spectrometry has emerged as the method of choice for identifying phosphorylation sites in phosphopeptides because of its advantages over previous methods (high performance liquid chromatography separation of radiolabelled proteins with 32P or 33P followed by Edman degradation) including its increased sensitivity and speed, and because it eliminates the need for protein radiolabelling. This research focused on the appl ication of mass spectrometry to phosphoproteomic analyses of Arabidopsis thaliana We demonstrated the application of mass spectrometry to four phosphoproteomic proj ects. Before working on these projects, we examined the developmen t of methods for phosphorylation enrichment and analysis.

PAGE 17

xvii These methods were then applied to th e various projects. Project 1, identifying autophosphorylation sites of a calcium-dependent protein ki nase, demonstrated the use of several complementary methods for identify ing numerous autophosphorylation sites of the protein. Project 2, identifyi ng substrates of a calcium-dep endent protein kinase from Arabidopsis thaliana demonstrated the application of several newer technologies for identifying numerous substrates of the ki nase. Project 3, identifying 14-3-3 interactors from Arabidopsis thaliana examined the identification of numerous protein interactors, several of which were proven to be phosphoryl ated. These interactors were then shown to overlap with the substrates identified for the kinase, possibly i ndicating interaction between the two families of proteins. Fina lly, project 4, the application of robotic instrumentation was demonstrated as a means for high-throughput phosphoproteomic analysis, which resulted in identifying several phosphorylated proteins.

PAGE 18

1 CHAPTER 1 INTRODUCTION Reversible protein phosphoryla tion on serine, threonine a nd tyrosine residues is one of the most common and important regulatory modifications of intracellular proteins, playing a role in many biological and biom edical phenomena such as cellular signal transduction, cell growth, cell differentia tion, cell division, metabolism, and cancer.1, 2 Highlighting its importance is the fact that up to one-third of all proteins in a cell are phosphorylated at any given time, and as much as 5% of all the genes in a vertebrate genome code for enzymes involved in phosphor ylation (kinases) or dephosphorylation (phosphatases).2 Due to its importance, research ha s been initiated in many areas of biomedical research towards the understandi ng of the regulatory properties of protein phosphorylation. Included in these studies are the investigation of the function of protein phosphorylation in cell cycle regulation, enzy me activation/deactivation, and proteinprotein association. Calcium-Dependent Protein Kinases In eukaryotes, protein kinases regulate key aspects of cellular function (such as metabolism) and responses to external signals by catalyzing the transfer of the terminal group of ATP to seryl or threonyl residue s in a variety of protein substrates.3 Recent mapping of the Arabidopsis genome provides the first opportunity to identify all the protein kinases present in a plant model a nd to begin to understand their physiological roles. The Arabidopsis genome encodes 1085 typi cal protein kinases, which is about 4% of the predicted 25,500 genes.4 These plant kinases differ from animal kinases, since in

PAGE 19

2 plants they phosphorylate only serine and threon ine residues; while in animals, tyrosine is the predominant residue that is phosphorylate d. Moreover, a number of kinase families in plants are either not found in animals or yeast, or are highly divergen t. Some of these are calcium-dependent protein kinases (CDPK s) found in vascular a nd nonvascular plants, green algae, and certain protozoa (ciliates and apicomplexans).5 These enzymes are proposed to be involved in all aspects of plant development and physiology, and participate in the coupling of cellular res ponses to environmental and developmental signals.6 Regulation of the CDPK kinase activity depends on calcium signaling and possibly autophosphorylation on Ser/Thr re sidues of the kinase itsel f; however, the regulatory effects of autophosphorylat ion still remain unclear.7-9 Autophosphorylation of a CDPK from groundnut was suggested to be a prerequisite for its activation,10 while inhibition of activity was seen after autophos phorylation of a winged bean.11 On the other hand, preautophosphorylation of a CD PK from sandalwood had no effect on kinase activity.12 Also, conflicting results were seen for a CDPK from ice plant whereby mutation of either one of the two autophosphorylation sites in creased activity, but mutations at both residues dramatically decreased activity.9 This regulatory process needs further study to learn if common autophosphorylatio n sites exist among this family of kinases that will eventually lead to a greater understandi ng in the role of this mechanism. Another aspect of CDPKs th at is not understood and re quires further investigation is their recognition of substr ate proteins. It is expected that CDPKs have access to hundreds of potential substrates in the cyto sol and nucleus since they are found as both soluble and membrane-anchored isoforms.13 The expectation is that most isoforms will be

PAGE 20

3 found associated with membranes, such as the plasma membrane,14 peroxisomes,14 endoplasmic reticulum,15 seed oil bodies,16 and mitochondria.17 In addition to their widespread subcellular distribution, there is evidence that some CDPKs can change locations in response to a stress treatment.9 This was seen when an isoform McCPK1 from ice plant ( Mesembryanthemum crystallinum ) was tagged with a fluorescent protein and transiently expressed in leaves. The tagged McCPK1 showed a pronounced shift in localization from the plasma membrane to th e nucleus in response to a salt or dehydration stress, indicating that CDPK targeting is dynamic. Again, very few substrates are known. K nowledge of the mechanism that these enzymes use to recognize their diverse substr ate proteins is even more limited. Typically, the sites phosphorylated by a particular prot ein kinase share a set of common sequence elements (its consensus sequence) whose existence is necessary and sufficient for recognition by that enzyme.18 These common sequence elements refer to the sequence elements immediately surrounding the site(s ) phosphorylated by the kinase, generally taking the form of a short linear sequence of amino acids According to Kennelly and Krebs,18 several assumptions are implied in th e formulation of a consensus sequence: 1) The existence of a consensus sequence on a protein is essential and adequate for its recognition as a substrate by a particular pr otein kinase. 2) The specificity-determining feature of the phosphorylation si te is contained in a neighbor ing sequence of amino acids around the phosphoacceptor, not including elements from different polypeptide chains or from widely scattered portions of a si ngle polypeptide chain. 3) Not all sequence positions surrounding the phosphoacceptor group ca rry equal weight in determining the recognition code.18

PAGE 21

4 Summarizing the complexities of the substr ate-recognition process as a set of short recognition sequences has its usefulness in its simplicity, which has facilitating the evaluation and application of a large body of observations. However, this sequence can be an oversimplification that can lead one to think that the primary sequence alone controls recognition; when in fact, factors su ch as secondary/tertiary structure or distant secondary recognition sites play a significant role in substrate recognition.18, 19 The secondary/tertiary structure of the protein ma y actually determine substrate specificity by denying access to potential phosphoacceptor grou ps. This means that the existence of intricate secondary/tertiary structures could be an important key to substrate recognition. That is, the more complex the determinants the more discriminating the kinase. The presence of a consensus sequence does not a ssure that a protein is a substrate of the kinase, but instead functions as a guide whose implications must be confirmed. In early studies of CDPK substrate sp ecificity, two simple phosphorylation motifs were recognized; Basic-3-x-x-[S/T]0 and S0-x-Basic+2. To date, in-depth analyses have resulted in the reporting of four consensus sequences for CDPKs with some differences apparent among the isoforms: 1) -5-x-Basic-3-x-x-S0-x-x-x+4 (minimal) or Basic-6-5x-Basic-3-x-x-S0-x-x-x+4-Basic -6 (optimal), 2) [Basic-9-Basic-8-x-7-Basic-6]-5-x-x-x-x[S/T]0-x-Basic+2, where the exact ordering of resi dues within brackets is not speci ed, 3) -1-[ST]0+1-X-Basic+3-Basic+4, and 4) [AL]-5-X-4-R-3-X-2-X-1-S0-X+1-R+2-Z+3-R+4, where is a hydrophobic residue, x is any ami no acid, and Basic is a basic amino acid residue (K or R), and Z is any residue but R or K.20-26 Because individual CDPK isoforms can, in general, recognize all four motifs, it appears that CDPKs may have a series of

PAGE 22

5 overlapping but non-identical polypeptide bi nding grooves that can accommodate the different sequences. Studies to determine these CDPK motifs have largely been performed by using synthetic peptides. However, although thes e peptides have represented powerful investigative tools, their small size and random conforma tion significantly limit their abilities to mimic the proteins they are intended to model. Therefore, using proteins to identify substrates for new phosphorylati on sequences could help define primary structural determinants of protein kinase specificity. Protein extracts are usually in the dena tured form, making it more difficult to interpret physiological relevanc e of results. Bylund and Krebs19 showed that phosphorylation may increase with the unfol ding of the protein substrate. Native lysozyme (which was not a substrate for th e cyclic AMP-dependent protein kinase of rabbit skeletal muscle) became susceptible to phosphorylation by the enzyme once the protein was denatured by heating.19 Therefore, many proteins may contain sites that can be phosphorylated once they have been expos ed. Careful interpretation is therefore needed for protein phosphorylation reactions observed in vitro since denatured proteins may become protein kinase substrates even though they were not substrates in their native state. Additionally, the mixing of proteins and kinase s from different subcellular compartments may lead to phosphorylati on of proteins that would not occur in vivo Despite these complications, the identi cation of phosphorylation motifs is of fundamental importance and may be useful for functional genomics and prediction of phosphoproteins. There are at least 34 CDPK isoforms in Arabidopsis, some of which have been implicated in drought stress, pat hogen response pathways, and the regulation

PAGE 23

6 of metabolic enzymes, transpor t proteins, and cell structure.7 Consequently, it is of fundamental importance to understand how this important group of Ser/Thr-kinases targets their substrate proteins. So ev en though these results might not signify physiological occurrences, they are important in terms of fundamental information, as the identification of new motifs may aid in th e understanding of how these kinases target their substrates. 14-3-3s Another family of proteins associated w ith phosphorylation is th e 14-3-3 proteins. These proteins were first identified as abunda nt brain proteins that were isolated as soluble, cytosolic, and acidic proteins.27, 28 Naming of these proteins was given according to their particular migration pattern on two-dimensional diethylaminoethyl cellulose (DEAE-cellulose) chromatography and starch ge l electrophoresis. The proteins were then named by Greek letters according to their re spective elution positions on HPLC. Further studies showed that these pr oteins were also present in all eukaryotic organisms examined to date, existing as protein families that contain highly conserved, but individually distinct isoforms.29-31 Of the organisms characterized, Arabidopsis has the largest family (10 distinct 14-3-3 pr oteins of 248 to 268 amino acids: GF14 GF14 GF14 GFl4 GF14 GF14 GF14 GF14 GF14, and GF14 ).28 Members of the 14-3-3 family are homoand heterodimers, whose L-shaped monomers come together to form a broad cen tral groove that cont ains two binding sites for target proteins. Phosphoryl ation of these binding part ners (and possibly also the 14-3-3 proteins themselves) may be impor tant in regulating these interactions.32, 33 To date, two different binding motifs have b een identified in n early all known 14-3-3 binding proteins, RSXpSXP and RXY/FXpSXP (where X denotes any amino acid

PAGE 24

7 residue, R represents a basic residue, and pS denotes phosphorylated serine); and other binding motifs have been discovered, in cluding unphosphorylated sites on a few proteins.32, 34, 35 Binding of these phosphoserine-containing pr oteins with 14-3-3 proteins implicated 14-3-3s as proteins that mediate inte raction among divers e components of many biological activities. In mammals, most of the known 14-3-3 binding proteins are components of intracellular signa lling pathways. In contrast 14-3-3s in plants have emerged as important regulators of phosphoryl ated enzymes of biosynthetic metabolism, ion channels and regulat ors of plant growth.36 While many proteins have been identified to bind with 14-3-3s, little is known about how many targets exist in plants. Also, since there are only two curre ntly known phosphoserine-cont aining binding motifs among target proteins, it would be of interest to discover whether there are more 14-3-3 binding motifs, and to determine if some of thes e targets are also ta rgets of CDPKs. As mentioned earlier, the common use of denatured proteins would actually be of benefit for these studies since crysta l structures of a 14-3-3 :phosphopeptide complex showed that the phosphopeptides bind in an extended confor mation, thus resulting in a greater chance of finding novel binding motifs.32 Detection and Analysis of Protein Phosphorylation To understand processes regulated by phos phorylation on the molecular level, one must first determine which proteins are phosphorylated, and second identify the exact sites of phosphorylation. However, as Mann et al.37 mentions, analysis of phosphoproteins is not straightforward for seve ral reasons. First, the stoichiometry of phosphorylation is generally relatively low; that is, only a small fraction of the available intracellular pool of a protein is phosphorylated at any given time as a result of a given

PAGE 25

8 stimulus. Second, the phosphorylated sites on proteins might vary, implying that any given phosphoprotein is hetero geneous; that is, it exis ts in several different phosphorylated forms. Third, many of the signaling molecules are present at low abundance within cells and in these cases, en richment is a prerequisite. Fourth, most analytical techniques used for studying pr otein phosphorylation have a limited dynamic range; which means that, although major phosphor ylation sites might be located easily, minor sites might be difficu lt to identify. Fifth, phosphatases present in cell lysates could dephosphorylate residues unless pr ecautions are taken to in hibit their activity during preparation and purification steps. Finall y, proteins can also be phosphorylated by kinases which are already pres ent during extraction and purification, so this needs to be prevented. The analysis of phosphorylation usually pr oceeds in a certain order, with the detection of the phosphorylated protein coming first. This may be done by using either radiolabeling, antibodies, or fluorescent labeling; and analysis is done by using chromatographic methods [high-performance liquid chromatography (HPLC), thin-layer chromatography TLC)], electrophoresis, wester n blotting, autoradiography, scintillation counting, or mass spectrometry (MS).1 Recently, a method for selectively staining phosphorylated proteins separa ted by polyacrylamide gel electrophoresis was developed. This proprietary fluorescent stain (Pro-Q Diamond Phosphoprot ein Gel Stain) developed by Molecular Probes allows dir ect, in-gel detectio n of phosphate groups attached to serine, threonine, and tyrosine residues, without the need for radiolabelling or antibodies.

PAGE 26

9 Another way to detect phosphoproteins is to screen for the pr esence of individual phosphopeptides. This can be done by first part ially hydrolyzing the la beled or unlabelled protein by partial acid, enzymatic, or alkali ne hydrolysis of the amide bonds of the protein to release the phosphopeptides. The resulting phosphopeptides then must be separated from other peptides by using e ither thin-layer chromatography (TLC), electrophoresis (one or two dimensional), or high-performance liquid chromatography (HPLC). Retention time, mass spectrome try, antibody recognition, or amino acid sequencing (Edman, MS/MS) can th en be used for identification. Recently, mass spectrometry-based methods have become increasingly popular for analyzing phosphopeptides resulting from prot eolytic digests of phosphorylated proteins because of their increased sensitivity and sp eed, and because they remove the need for protein radiolabelling. Mass Spectrometric-based Methods Mass spectrometry is becoming the me thod of choice for analyzing complex protein mixtures. Routine analysis of biomol ecules, particularly pr oteins and peptides, was made possible by the advent of two ma ss spectrometry ionization tools: matrixassisted laser desorption/ioni zation (MALDI) and electrospr ay ionization (ESI). Since 1988 when MALDI and ESI mass spectrometry we re first proven useful for analyzing peptides, proteins, carbohydrates, and oli gonucleotides, they have become the MS methods of choice for biopolymer analysis. Our study used three mass spectrometers: a matrix-assisted laser desorption/ionizati on time-of-flight mass spectrometer (MALDITOF-MS), a nanospray quadrupole ion trap mass spectrometer (nanoESI-QIT-MS), and a tandem quadrupole time-of-flight mass spectrom eter with interchangeable MALDI and nanospray sources (MALDI or ESI-QqTOF-MS).

PAGE 27

10 Ionization techniques Matrix-assisted laser desorption/ionization (MALDI), a soft ioni zation technique, was first described in the late 1980s as a technique used with mass spectrometry for analyzing large, polar, nonvolatile molecules.38 In this technique, a solid organic matrix compound that is strongly UV-absorbing at the designated wavelength, is dissolved in an appropriate solvent and mixed with the solution of the sample of inte rest (the analyte). A 0.5-3 L aliquot of this solution is then placed on a stainless-stee l target plate and allowed to dry. On drying, the analyte is co-crystallized with a large (104) molar excess of the solid matrix material. Once in the ma ss spectrometer (typically time-of-flight), the sample is then irradiated with a pulsed laser beam [usually a nitrogen laser in the ultraviolet range (337 nm)] for desorption and ionization of the analyte molecules. Even though the mechanism by which MALDI operates is still unclear, it is agreed that the matrix is critical and fills several roles. First, using the large excess of matrix helps to isolate analyte molecules from each other, thereby reducing intramolecular interactions. Second, the matrix absorbs large amounts of energy from the incoming photons of the pulsed laser beam, resulting in an explosive breakdown of the matrixanalyte lattice, sending both matrix and analyt e molecules into the gas phase. Third, the matrix is necessary for ionization of the anal yte because it transfers protons to the analyte via gas-phase reactions in the dense cloud that forms.39 The desorption/ionization process is shown in Figure 1-1.39 Matrix choice depends on the irradiance wavelength and the type of sample being analyzed. Comm on matrices that are used with N2 lasers operating in the UV at 337 nm are 3,5-dimethoxy-4-hydroxytrans -cinnamic acid (sinapinic acid),

PAGE 28

11 2,5-dihydroxy-benzoic acid (DHB), and -cyano-4-hydroxy-trans-cinnamic acid ( cyano) (Figure 1-2),40, 41 with the latter two being best for peptides. Electrospray ionization, another soft ionizat ion technique, was first introduced in the late 1980s. In this technique, ions are fo rmed from peptides and proteins by spraying a dilute solution of the analyte (typically di ssolved in a mixture of water, an organic modifier such as acetonitrile, and a few per cent by volume of a volatile acid) from a fine tip at atmospheric pressure. Generally, a hi gh electric field is cr eated by applying a high voltage to either the spray tip or the counter electrode, resultin g in a fine mist of droplets that are highly charged. This generated el ectric field (E) between the spray tip and counter electrode is e xpressed by the equation: E = (2V/r) ln(4d/r) where V is the voltage applied, r is the radius of the needle, and d is the distance between the spray tip and counter electrode. This imposed electric field will also penetrate to the liquid flowing through the needle, causing ions in solution to move toward the liquid surface. Accumulation of these charges at the surface then leads to de stabilization of the surface because the ions at th e surface are drawn to the counter electrode yet cant escape, resulting in the formation of a Tayl or cone (Figure 1-3). Due to the cones instability (influenced by the surface tension of the fluid ), charged droplets are emitted. The onset voltage (Von) required to initiate charged-dropl et emission is related to surface tension by the equation: Von = 2 x 105 ( r)0.5 n (4d/r) The radius (R) of the emitted droplet s will depend on the fluid density ( ), flow rate (Vf), and surface tension ( ), given by the relationship:

PAGE 29

12 R ( Vf 2 )1/3 Thus, the higher the flow rate (Vf), the larger the initial dropl et size which leads to lower ionization efficiency because the droplets ar e not so close in size to the Rayleigh limit (will be discussed later). On the same note, th e lower the flow rate (as in nanospray), the smaller the droplet size, the higher the su rface-to-volume ratios l eading to a larger amount of the analyte available for ionizat ion, thus a higher ionization efficiency. Once these small droplets are formed and are accelerated toward the counterelectrode, solvent rapidly eva porates and the analytes (peptides or proteins) in the droplets pick up one, two, or more protons fr om the solvent to form singly or more frequently, multiply charged ions (for example, [M+H]+1, [M+2H]+2, [M+3H]+3, etc.). As the solvent evaporates and the droplet shrinks the charge density on the surface increases to the point where Coulombic charge repulsi on overcomes the forces holding the droplet and solvated ions together (R ayleigh Limit), leading to disi ntegration of the droplet into smaller droplets (Figure 1-4). This limit wh ere the Coulombic explosions begin is given by: Q2 = 64 2 0 R3 where 0 is the permittivity of vacuum. Once the ions are emitted or evaporated from the droplet surface, the ions are then sample d into the high-vacuum region of the mass spectrometer for mass analysis and detec tion, most often using a quadrupole mass analyzer. Overall, the widespread acceptance of ESI-MS and MALDI-MS are accountable for several reasons. Both methods are usually combined with relatively low-cost, easily operated mass spectrometers. They both offe r high sensitivity (picomole-femtomolar

PAGE 30

13 range), accuracy ( 0.01%), and capability of analyzi ng molecules with a wide molecular weight range.39 Even though both techniques have many common capabilities, they do have their own unique capabilities that are of practical significance. Although both techniques work best with clean (salta nd detergent-free) samples, MALDI is more tolerant of many of the comm on biological buffers, that is, information can be obtained by MALDI directly from a relatively dirt y sample. Also, even though both methods can provide molecular weight information for la rge proteins, MALDI-MS is more sensitive and provides the information more easily. Howe ver, in the lower molecular weight range, ESI-MS usually provides more accurate molecular weight measurements as well as better mass resolution. Additionally, since MALDI produces predominantly singly charged molecular ions from peptides and prot eins, analysis of the resulting MALDI-MS spectrum is very straightforward. Finally, si nce ions in electrosp ray are produced at atmospheric pressure from flowing liquid st reams, ESI is ideally suited for on-line coupling to high-performance liquid chro matography, making it possible to analyze mixtures of peptides and proteins. With all of these considerations in mind, it can be said that MALDI and ESI are methods that can be used to complement each other. Mass analysis With the advent of MALDI and ESI, peptid e and protein analysis have been a large focus of efforts in mass spectrometry over th e last 17 years, bringing many types of mass analyzers into use in this area of research, namely time-of-flight, qua drupole ion trap, and triple quadrupole mass spectrometers, primarily because of their cost and ease of use. Time-of-flight mass analyzers are among the simplest of the mass analyzers.42 The principle of time-of-flight mass spectrometry involves measuring the time required for an

PAGE 31

14 ion to travel from the ioni zation source to the detector.43 For a simple TOF-MS, there are three components, an ionizati on source (typically MALDI), a field-free drift region, and a detector (Figure 1-5). Upon ionization, ions are accelerated out of the source under the influence of a strong electric field. Even though all the ions recei ve the same kinetic energy during acceleration at the ionization source, as they tr averse the field-free region they separate into groups according to ve locity because they may have different m/z values. These ions then sequentially stri ke the detector in order of increasing m/z value (lighter ions arriving first), upon which the time-of-flight analyzer converts the time-offlight of the ions to a mass-to -charge ratio using the equations: E = mv2 t = L/v = L [m / 2zeV]1/2 therefore, m/z = 2t2eV/L2 where t = time of flight (sec onds), L = length of flight tube (m), v = velocity (m/s), m = mass (kg), and z = charge. Simple linear mass spectrometers as desc ribed above are somewhat limited due to their rather low resolution. This low resolution is partly due to the initial kinetic energy spread of individual ion populat ions, that is, various member s of the same ion population will arrive at the detector at slightly differe nt times. An effective way to correct for this energy spread is through the use of a reflect ron (ion mirror) which acts as an energyfocusing device. When an ion reaches the re flectron it is slowed down until it is stopped by a voltage that is applied to the back end of the reflectron, the ion is turned around, and then reaccelerated out to a second detector at a slightly different angle so the flight path of the reflected ions does not cross with the ions entering the reflect ron (Figure 1-6). In

PAGE 32

15 the case of an ion population with a spread of slightly different kinetic energies, ions with a slightly lower energy will not penetrate th e reflectron as deeply, therefore turning around more quickly and catching up to those ions with full kinetic energy. While ions with slightly greater kinetic energies w ill penetrate more deeply, turn around more slowly, and have their flight times retard ed, allowing the other ions to catch up. The addition of this reflectron will then cause ions of a given m/z to be spacially focused into packets with flight times that are closer t ogether. The addition of the reflectron also increases the flight path for an ion without increasing the si ze of the flight tube, also resulting in an improvement in resolution by enhancing the time di spersion of ions of different m/z Another improvement that has been ma de for achieving better resolution on MALDI-TOF instruments is time-lag focusing or delayed extraction in the ionization source. In this source, ions are created in a field-free region and allowed to spread out before extraction and acceler ation voltages are applied. Quadrupole ion trap mass spectrometers are mass analyzers that operate by trapping ions in a three-dimensional electric field consisting of two end-cap electrodes and a ring electrode, each having a h yperbolic geometry (Figure 1-7).42, 44, 45 In the normal mode of use, the end-cap electrodes have an auxiliary oscillat ing potential of low amplitude applied while the ring electrode has an RF oscillating drive potential of 1 MHz, resulting in the creation of a potential well (quadrupolar trappi ng field). This field can be described as having a saddle shape that is constantly spinning whereby the field at any particular point in time will possess this saddle shape. For an ideal quadrupole field to be generated, the mathematical relationship presented below has to be fulfilled: r0 2 = 2z0 2

PAGE 33

16 where r0 is the radius of the ring electrode in the centr al horizontal plane and z0 is the separation of the two end-cap electrodes m easured along the axis of the ion trap. Typically, once the magnitude of r0 is given, the sizes of all the electrodes and electrode spacings are fixed. It should be noted that th e majority of commercial ion traps in use today have r0 at either 1.00 or 0.707 cm. An ions stability in this quadrupolar trapping field is dependent upon its m/z the potentials applied to the electrodes, and the inte rnal dimensions of the ion trap electrodes. An ion that is stable in this field will po ssess a trajectory that has the appearance of a Lissajous figure, allowing it be trapped within the specific el ectric field of the ion trap (Figure 1-8). Unstable ions will have trajector ies that increase in magnitude as they near the ring of the endcaps, resulting in their co llision with the electrodes. Determination of whether the trajectory of an i on will be stable or unstable under defined conditions of the electric field may be calculated with the Mathieu equations: az = -2ar = -16eU (1) m 2(r0 2 + 2z0 2) qz = -2qr = 8eV (2) m 2(r0 2 + 2z0 2) where az and qz are two reduced parameters, r sy mbolizes the radial direction, z symbolizes the axial direction, e is the charge of an ion, U is the DC amplitude applied to the ring electrode, V is the RF am plitude, m is the mass of an ion, is the angular drive frequency (2 frf), r0 is the radius, and z0 is distance from the center to the end-cap. It should be noted that an ion has to be stable in both the r and z directions to be confined within the trap, thus, ar and qr parameters also have to be considered: ar = 8eU (3) m 2(r0 2 + 2z0 2)

PAGE 34

17 qr = -4eV (4) m 2(r0 2 + 2z0 2) The resulting stable trajectories that are calculated from the operating parameters ar,z and qr,z can be displayed graphically as the Mathieu stability di agram (Figure 1-9), whereby the region of stability is defined by the boundaries at z=0, z=1, r=0, and r=1. This means that if an ion has an az and qz within this region, it will be stable in both the r and z directions and will be trapped in the ion trap. Typically, ions generated by electrospray ionization (external source) are focused into the ion trap using electrostatic lenses. Once in the ion trap, collisions with helium buffer gas at a pressure of 1 mTorr dampen th e kinetic energy of th e ions and contract trajectories toward the center of the trap, where a range of m/z values are held in stable orbits by the RF potential. As the RF potential on the ring electrode is increased, the ions become more energetic and de velop unstable trajectories alon g the axis of symmetry (the z-axis), then in order of in creasing m/z value, ions exit th e trap through holes in the endcap electrodes to a detector. As the RF amp litude is ramped and ions are ejected to the detector one at a time, a mass spectrum is generated; usually several such spectra (microscans) are obtained in succession a nd are then summed prior to display and recorded as a macroscan. A detailed desc ription of quadrupole ion traps has been discussed by March if additional information is necessary.44, 45 Along with measuring the m/z values of i ons introduced to the mass spectrometer, quadrupole ion trap mass spectrometers can also be used to obtain detailed structural information from these ions. This information is obtained by performing multiple massselective operations, one afte r another, that is, tandem mass spectrometry (MS/MS). The first mass-selective operation is used for the isolation of the ion species of interest

PAGE 35

18 (designated as the parent ion) and the second is used to de termine the mass/charge ratios of the fragment ions (product) formed by collision-induced disso cation (CID) of the isolated ion of interest. CI D refers to the process whereby the kinetic energy of the selected ion population is incr eased by applying a voltage res onant with the frequency of the precursor ion, causing more energetic collis ions with the He bath gas. Subjecting the ions to many hundreds of low-energy collisi ons will ultimately increase the internal energy of the ion until fragmentation occurs. Another mass spectrometer also possessing th e ability to perform MS/MS analysis that was available for this project wa s a tandem quadrupole time-of-flight mass spectrometer (QqTOF-MS), where Q refers to a mass-resolving quadrupole and q refers to an r.f.-only quadrupole or hexapole collisi on cell. This configuration can be regarded as either the addition of a mass-resolving quadrupole and co llision cell to an ESI-TOF, or the replacement of the third quadrupole (Q3) in a triple quadrupole by a TOF mass spectrometer,46 with the latter being the simplest de scription for the purpose of describing the basic principles. A thorough review of triple quadrupole instruments has been described by Yost and Boyd.47 A typical QqTOF configuration consists of three quadrupoles, Q0, Q1 and Q2, followed by a reflecting TOF mass analyzer w ith orthogonal injection of ions (Figure 110). Q0 provides collisional damping, Q1 acts as a mass filter, and Q2 is a collision cell. In the case when ions are provided by a highpressure electrospray source, Q0 acts as an ion guide with collisional cool ing and focusing of the ions as they enter the instrument. When single MS (or TOF-MS) measurements are required, the mass filter Q1 is operated in the r.f.-only mode serving as only a transmission element while the TOF analyzer

PAGE 36

19 records spectra. For MS/MS, Q1 is operated in the mass filter mode to only transmit the parent ion of interest which then gets accelerated to an energy of between 20 and 200 eV before it enters the collision cell Q2, wher e it is subjected to CID and subsequently collisionally cooled and focused befo re analysis by the TOF mass analyzer. Mass analysis of peptides and proteins In a generic mass spectrometry-based experiment, the typical proteomics experiment consists of 5 stages. In stage 1, proteins are isolated from cell lysates or tissues with gel electrophores is typically used as a met hod for biochemical fractionation. Since MS of whole proteins are less sensitiv e than MS of peptides and the mass of the intact protein is usually insufficient for prot ein identification, protei ns from stage 1 are typically enzymatically digested with trypsin for the formation of C-terminally protonated amino acid peptides (stage 2). These peptides are then separated by highperformance liquid chromatography followe d by ESI (stage 3) whereby ions are introduced to the mass spectometer for mass analysis, producing an MS spectrum (stage 4). The computer then generate s a prioritized list of these peptides for fragmentation and a series of MS/MS experiments ensues (stage 4). These MS and MS/MS spectra are then used for matching against a known protein databa se for the identifica tion of the proteins (stage 5).48 Stage 1, the sample extraction and prepara tion step, may be considered the most critical step in any proteomics study.49 In this regard, proteomic analysis of plant tissues is even more problematic than other orga nisms due to the invo lvement of several challenges. These include the fact that pl ant tissues typically have low amounts of proteins, they are often rich in proteases a nd materials that may severely interfere with downstream protein separation and anal ysis, including cell wall and storage

PAGE 37

20 polysaccharides, lipids, phenolic com pounds and a broad array of secondary metabolites.49 These contaminants pose a serious problem for one of the most commonly used separation techniques in proteomics 2-dimensional gel electrophoresis. Their presence can result in horizontal and vert ical streaking, smearing, and a reduction in the number of distinctly resolved protein spots. In order to alleviate this problem, several protein extraction techniques for plant tissu es have been compared by Saravanan and Rose, whereby the quantitative and qualitative ch aracteristics of the protein extracts were examined.49 From this study, it was demonstrated that the phenol-based method gave the greatest protein yield and the least contaminants. Once protein extraction from the plant tissu e has been performed, the proteins are typically separated by 1or 2-dimensional gel electrophoresis (GE), which are techniques that separate proteins by the application of an electric field. SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis) is a 1D-GE technique that separates proteins according to their molecular weights. 2D-GE involves the separation of proteins according to their isoelectric point (the pH at which a protein carries no net electric charge) in the first dimension, and separati on according to MW in the second dimension. Once the proteins are immobilized in the gel, they are t ypically visualized by a gel staining method.50, 51 Some gel stains are visible in visible light, others which are fluorescent stains that require visualizati on by UV light. Typical fluorescent imaging devices are CCD camera-based sy stems or laser scanner systems.52, 53 Once the proteins are visualized and the prot eins of interest are determined, the gel bands/spots are excised for enzymatic digestion.54 Prior to digestion, protein stains have to be removed to prevent later interference of enzyme activity or mass spectrometric

PAGE 38

21 analysis. Also, the presence of sulfhydryl-containing amino aci ds in a protein may result in the formation of disulfides bonds along the protein backbone resulting in the formation of a three-dimensional shape. As a result, i nner portions of the 3D protein structure may be inaccessible to the enzyme, thus unraveling of this structure is necessary for complete enzyme cleavage of the protein. This is achieved by the reducti on of the sulfhydryl groups with a reducing agent su ch as dithiothreitol (DTT), and subsequent alkylation to prevent reforming of the disulfide bond, resul ting in the linearizati on of the protein. At this point, an enzyme is added for protei n cleavage. Trypsin, a commonly used enzyme, cuts the protein at the C-terminal end of ly sine and arginine residues, resulting in the formation of C-terminally protonated amino acid peptides for mass spectrometric analysis. As mentioned earlier, the routine analysis of proteins and peptides has become possible due to the introduc tion of MALDI and ESI techni ques. Typically, MALDI-MS is used for the measurement of peptide masses in an enzymatic digest. ESI-MS is typically coupled with on-line reversed-p hase liquid chromatography sy stems for the separation of the peptides prior to intr oduction to the mass spectrome ter, upon which the peptide masses are measured and tandem MS of the peptides is performed for structural information. The tandem MS utility available on two of the mass spectrometers described above, the QIT-MS and QqTOF-MS, enables immense structural information to be obtained on proteins and peptides. Advancing this utility even more is the incorporation of decisionmaking algorithms that can automatically pe rform MS/MS experiments on precursor ions from a previously acquired full scan, enabling the instrument to make real-time decisions

PAGE 39

22 concerning the experiment at hand. The use of this algorithm is called data-dependent analysis. A typical example of its use would be for peptides eluting from an HPLC separation. As peptides enter the ion trap ma ss spectrometer, a full scan is obtained. Once an ion is detected above a preset threshold, the mass spectrometer will automatically switch from full scan mode to MS/MS mode on that ion. If there are coeluting peptides, the mass spectrometer will perform MS/MS on th e most intense ion, this ion would then be placed on an exclusion list, and the sec ond most intense ion from the full scan would then be subjected to MS/MS. So as each i on gets subjected to MS /MS analysis it is placed on an exclusion list and will not be removed from this lis t until after a userdefined length of time. As peptides undergo fragmentation by lo w-energy gas phase collisions during MS/MS, they undergo cleavage at the amide bonds (-CO-NH-) that join pairs of amino acid residues, generating a ladder of sequence ions.55 If the charge is retained on the Nterminal end after cleavage of the amide bond, b-type ions are formed. However, if the charge is retained on the C-terminal end, y-type ions are formed. The most commonly observed fragment ions and their nomencl ature are shown in Figure 1-11A. The determination of the amino acid sequence becomes possible if a complete series of either one or both ion types are present for the subt raction of the masses of adjacent sequence ions. If a phosphate group is present, as with phosphorylated peptides, fragmentation occurs differently, that is, very little cl eavage occurs along the peptide backbone (amide bonds) as would be expected. Instead, -elimination (removal of the phosphate moiety) occurs as the primary fragmentation with th e phosphate moiety being lost as either

PAGE 40

23 H3PO4, HPO3, H2PO4 -, HPO4 2-, PO4 3-, or PO3 -, depending on the ionization mode (positive or negative) being used (Fi gure 1-11B). These signature losses from phosphopeptides can be used as an indicat or to determine whether a peptide is phosphorylated. On the QIT-MS, a reconstruc ted chromatogram can be created by the software after MS/MS analysis to locate all peptides in a chromatogram that lost a neutral fragment (H3PO4 or HPO3) (a reconstructed neutral fragment chromatogram). However, sequencing of the peptide is generally di fficult due to insufficient fragmentation. A similar feature available on the QqTOF-MS is precursor ion scanni ng, whereby peptides losing the precursor ion (H2PO4 -, HPO4 2-, PO4 3-, or PO3 -) are recorded by the TOF-MS in the negative ion mode. The draw back of this method is that the experiment has to be repeated in the positive ion mode so that peptide sequencing may be obtained because fragments created in the negative ion m ode are very difficult to interpret. Data interpretation Once mass spectra are obtained, protein iden tification is performed according to the method of analysis. For MALDI-MS, peptide mapping, often referred to as peptide-mass mapping or peptide-mass fingerprinting is performed (Figure 1-12). In this method, proteins are identified by ma tching the list of experimental peptide masses to a list of predicted peptide masses that w ould occur after digestion with a specific enzyme of all entries in the protein database.42, 48 A match is generally found if a sufficient number of peptide ions are matched and there are not more than two proteins present. Typically, as the size of the database being se arched is increased, the level of uncertainty also increases due to the possibility of more proteins bei ng present that could generate peptides with similar masses.

PAGE 41

24 If tandem MS was performed, identificati on of the protein is more clear-cut because not only is the mass of the peptide known, but the peptide sequence information is also available from the peak pattern of the CID spectrum (Figure 1-13). The CID spectra are then scanned against a comprehens ive protein sequence database using one of a number of different algorithms, each with its own strengths and weaknesses.48, 56-60 Two of the most commonly used methods ar e cross-correlation and probability-based matching. In the cross-correlation method used in the SEQUEST search engine, theoretical tandem mass spectra are constructed for all peptide sequences in the protein database matching the mass of the experimental peptide and the overlap or cross-correlation of the predicted spectra with the experimental spect ra is used to determine the best match.42, 56-58 The quality of the match between the seque nce and the spectrum is indicated by the magnitude of the cross-correlation value and th e quality of the match versus all the other top ranking sequences in the database is s hown by the difference between the normalized cross-correlation score to the second ranke d sequence. In the probability-based matching method used in the MASCOT search engine, calculated fragments of peptides in the database are compared with observed peaks an d a score is calculated that reflects the statistical significance of the match between the spectrum and the sequences contained in a database. There are three advantages to this approach:60 1) A simple rule can be used to judge whether a result is signi ficant or not, guarding against fa lse positives. 2) Scores can be compared with those from other types of searches. 3) Search parameters can be readily optimized by iteration. For each of these met hods, the identified peptides are compiled into a protein hit list with co rresponding scores or statistics.

PAGE 42

25 The introduction of these search engine s have provided a means for handling the enormous amounts of CID spectra that can be produced from data-dependent scans, resulting in high-throughput proteomics. Howeve r, because protein id entifications rely on the matches with sequence databases, highthroughput proteomics is limited to those species having a comprehensive sequence database available. Unfortunately, identification and mappi ng of phosphorylated peptides by tandem mass spectrometry is not as straightforwar d as described above for several reasons.61 First, cleavage of the protein by trypsin can be inhibited due to the negatively charged modifications of the phosphate group, resu lting in incomplete peptide coverage. Secondly, phosphorylation is often sub-stoich iometric, resulting in the phosphopeptide being present in much lower abundance than the other unphosphorylated peptides. This will result in suppression of the phos phopeptide relative to its unphosphorylated counterpart during the mass sp ectrometric analysis. Suppression of the phosphopeptide is even more pronounced in the presence of ma ny other unphosphorylated peptides such as that found in the protein digest. Reducing the number of unphosphorylated peptides present in the sample by enriching for th e phosphopeptides will th erefore enhance mass spectrometric mapping of the phosphorylation site. Finally, performing tandem MS on the phosphopeptide to determine the location of the phosphate group can be difficult due to the instability of the phosphate moiety whereby the phosphate moiety may be eliminated before the peptide can even undergo fragmentation, making it difficult to locate the phosphorylation s ite. Also, since fragmentation of the phosphopeptide may result in only the elimination of the phosphate moiety and very little fragmentation along the backbone, database searching wi ll not identify the phosphopeptide.

PAGE 43

26 Isolation and enrichment techniques Isolation and enrichment techniques ha ve been developed to enhance mass spectrometric analysis of phosphopeptides ev en further, eliminating the need for radiolabeling, antibodies or fluorescent labe ling. The most common enrichment methods are immobilized metal-ion affinity chromatography (IMAC),62-67 and chemical modification methods.68-73 Once isolated, the resulting phosphopeptides can then be identified by mass spectrometry, however, this ca n be very difficult as will be discussed later. Immobilized metal-ion affinity chromatography Immobilized metal-ion affinity chro matography, originally recognized by Andersson and Porath, is a method used to se lectively isolate and enrich phosphopeptides from a peptide mixture via the interaction of the phosphate group on the peptide with the free coordination sites of metal ions immob ilized (chelated) to a stationary phase74 (Figure 1-14). Two of the more frequently used chelating ligands for IMAC are iminodiacetic acid (IDA) and nitrilo-triacetic acid (NTA), however, the majority of the published applications of IMAC for phosphopr otein and phosphopeptide characterization have used IDA. Townsend and coworkers reported that NTA sepharose was superior to IDA-sepharose for phosphopeptide purification by Fe(III)-IMAC because of its higher selectivity for phosphopeptides.75 However, studies performed by other groups have not confirmed this.67 Since the first report utiliz ing Fe(III) by Andersson and Porath, the use of several other metal ions have been reported, incl uding Al(III), Sc(III), Lu(III), Th(III),74, 76 and Ga(III).66 Of these metals, Fe(III) and Ga(III) seem to give best results for phosphopeptide isolation, with Ga(III) show ing less overall suppression effect and the

PAGE 44

27 ability to isolate multiply phosphorylated pep tides, while Fe(III) show s better selectivity for monophosphorylated peptides.66 Although the IMAC methodology is for th e isolation and enrichment of phosphopeptides, nonspecific binding could also occur for peptides possessing multiple carboxylic acid groups.62 These peptides could suppress the signal from trace-level phosphopeptides in the mixture. One way to prevent this problem is to convert the peptides in the mixture to the correspondi ng peptide methyl esters by replacing the carboxylic acid groups. This is especially n eeded when acidic amino acids (aspartate, glutamate, and S-carboxymethylated cystei ne) are present. By performing this methylation step prior to th e IMAC isolation, it is exp ected that nonspecific binding through carboxyl groups will be prevented, re sulting in the selective isolation of phosphopeptides only. The methylation step, how ever, may be a problem if there is moisture present due to the sensitivity of methanolic-HCl to water. This means that measures have to be taken to keep the r eaction mixtures extremely dry, which can make the sample preparation prior to mass sp ectrometry more labor intensive taking approximately three and a half hours compared to an hour without the methylation step. Another area of concern for the IMAC me thod is that the phosphate moiety may be lost during ionization or frag mentation due to the lability of the phosphate group. This means that once the phosphate group is remove d, localization of th e phosphorylation site will be difficult. Also, little sequence inform ation may be obtained because most of the fragmentation energy is used to remove the phosphate moiety. One way of alleviating this problem would be to replace the phosphate group with a more stable label.

PAGE 45

28 Chemical derivatization The problem of the lability of the phos phate group has been addressed by several proposed chemical modification methods whereby H3PO4 is removed by -elimination at high pH, a linker containing a thiol group is added by Michael addition, followed by the addition of a biotin-containing compound which act s as an affinity tag for purification of the phosphopeptides as well as a tag for phosphorylation site mapping67-73, 77 (Figure 115). Several linkers have been used fo r modification, includ ing ethanedithiol,69, 71, 72, 77 and various lengths of alkanethiols.73 Of these linkers, ethanedithiol seems to be the most popular due to the presence of a thiol group on both ends of the molecule, thus increasing the chances for it to be bound. Several types of biotinylated chemicals have been used, including iodoacetyl-PEO-biotin,72, 77 biotin-HPDP,71 and (+)-biotinyl,3maleimidopropionamidyl-3,4-dioxoctanediamine.69 The disadvantages of this technique are that it is labor-intensive, time-consum ing (several hours), and the many sample handling steps involved could le ad to significant sample loss. These problems can be eliminated if the label is synthesized and st ored prior to sample preparation, however, not very many publications report doing this.77 Qiagen phosphoprotein purification kit Within the past 2 years, the development of an enrichment method specifically for phosphorylated proteins from complex cel l lysates was introduced by Qiagen Incorporated. Purification of the phosphoproteins is performed by a proprietary affinity chromatography method. The little information th at is known about this column is that the phosphate groups on phosphoproteins are bound to the column with high specificity, while proteins without the phos phate groups will not bind to the column and therefore be found in the column flow-through. The bound phos phoproteins are then washed with a

PAGE 46

29 phosphate-buffered saline buffer (PBS) and st ored. The free phosphate in the elution buffer serves two functions: 1) displaces th e phosphoproteins from the column, and 2) inhibits the phosphate activity in the cell lysate, therefor e stabilizing the phosphorylation status of the proteins during dow nstream processing and storage.

PAGE 47

30 Figure 1-1. The MALDI process

PAGE 48

31 Figure 1-2. Commonly used MALDI matri ces for peptide and protein analysis Figure 1-3. Taylor cone formation

PAGE 49

32 Figure 1-4. The electros pray ionization process Figure 1-5. A linear time-of -flight mass spectrometer

PAGE 50

33 Figure 1-6. A reflectron timeof-flight mass spectrometer Figure 1-7. Three-dimensional idea l ion trap showing the asymptotes and the dimensions r0 and z0. Modified from March, R.E. International Journal of Mass Spectrometry 2000 200 285-312. Figure 1, page 287

PAGE 51

34 Figure 1-8. Trajectory of a trappe d ion. Modified from March, R.E. Journal of Mass Spectrometry 1997 32 351-369. Figure 8, page 356.

PAGE 52

35 Figure 1-9. Mathieu stability diagram in (az, qz) space for the region of simultaneous stability in both the r and z -directions near the origin for the threedimensional quadrupole ion trap. The qz-axis intersects the z=1 boundary at qz=0.908, which corresponds to qmax in the mass-selective instability mode. Modified from March, R.E. Journal of Mass Spectrometry 1997 32 351-369. Figure 7, page 356.

PAGE 53

36 Figure 1-10. Tandem QqTOF mass spectromete r. Adapted from Chernushevich, I.V.; Loboda, A.V.; Thomson, B.A. Journal of Mass Spectrometry 2001 36, 849865. Figure 1, page 860.

PAGE 54

37 Figure 1-11. Peptide fragme ntation. A) Typical low-ener gy CID fragmentation of a peptide forming mainly b and y -ions. B) Fragmentation of a phosphopeptide resulting in mainly -elimination of the phosphate moiety as phosphoric acid.

PAGE 55

38 Figure 1-12. Peptide mass mapping. A protein se quence can be verified by site-specific digestion and measurement of the peptide ions for correlation with those predicted by the sequence. Conversely, if the identity of the protein is not known the peptide mass map can be used to search the protein database to find the sequence that best fits the mass map. Adapted from Yates, J.R. Journal of Mass Spectrometry 1998 33, 1-19. Figure 3, page 8.

PAGE 56

39 Figure 1-13. Tandem mass spectrometric se quencing. The ladder of fragment ions represents the amino acid sequence of the peptide. By subtracting the m/z values for adjacent ions of the same type the sequence can be elucidated. Conversely, the fragmentati on pattern can be used to search the protein or nucleotide database to find the amino aci d sequence that best fits the tandem mass spectrum. Adapted from Yates, J.R. Journal of Mass Spectrometry 1998 33, 1-19. Figure 4, page 10.

PAGE 57

40 Figure 1-14. Two of the most frequently used chelating ligands for IMAC: iminodiacetic acid and nitr ilo-triacetic acid.

PAGE 58

41 Figure 1-15. Chemical derivatization me thod for the affinity purification of a phosphorylated peptide. Adapted from Go she, M.; Conrads, T.; Panisko, E.; Angell, N.; Veenstra, T.; Smith, R. Analytical Chemistry 2001 73, 25782586. Figure 1, page 2581.

PAGE 59

42 CHAPTER 2 PHOSPHORYLATION DETECTION AND ENRICHMENT Reversible phosphorylation on serine, thre onine and tyrosine residues on a protein is one of the most common and important pos t-translational modifications involved in virtually all cellular processes. Key to the molecular understanding of these signals, is the identification of kinases, their substrates, and the sp ecific site of phosphorylation. However, the process of studying phosphorylatio n can be a very difficult and tedious task as mentioned in Chapter 1. A major contribut or to the difficulty of this process is suppression by unphosphorylated peptides in a complex mixture, making phosphopeptide isolation and enrichme nt a necessary step. For the development of a method for phosphopr otein analysis of a complex mixture (for example Arabidopsis thaliana protein extract), one of the first requirements is the determination of a means of detecting phosphopr oteins along with the limits of detection of the chosen method. Second, since isola tion and enrichment of the phosphorylated protein is necessary if using a MALD I-TOF-MS or QIT-MS, examination and comparisons of available enrichment methods would be advantageous for determining the most appropriate method. MALDI-TOF-MS can provide a means for ra pid sample preparation and analysis for method development, however, optimal c onditions for MALDI has to be determined prior to any other method development strate gies. As mentioned earlier, selection of matrix and solvent conditions is critical. This is even more important for the analysis of phosphorylated peptides because of their lo w response to mass spectrometry in positive

PAGE 60

43 ion mode due to the presence of the nega tively charged phosphate group. This negative charge interference in detection is even more pronounced when multiple phosphate groups are present in the peptide. Asara and Allison determined that one way to alleviate this problem is to reduce the negative charge interference by the a ddition of a positively charged species to the matrix spot.78 Typically, this may be done by adding a solution of ammonium citrate to the already co-cryst allized analyte and matrix spot. Hence, optimization of matrix conditions pr ior to this step is necessary. Once matrix conditions for phosphopeptide analysis have been optimized, comparison of enrichment techniques may be performed. As mentioned in Chapter 1, two published methods for phosphopeptide isolation and enrichment are IMAC and chemical derivatization methods. During the initial stages of this project, there were two commercially available IMAC columns on the market, the Pierce Phosphopeptide Isolation Kit and Millipores ZipTipMC, each having a potential advantage over the other. The isolation kit from Pierce is a small minispin column with Ga(III) chelated to an IDA resin, thus reducing the workload for th e user since the metal ions are already chelated and the minispin column format makes it possible to enrich several samples at once. On the other hand, Millipores ZipTipMC comes as a ziptip with the IDA resin only, that is, the user binds his or her meta l of choice during the preparation, thus giving the user the opportunity to te st various metal ligands for optimization. At the time, several methods had been reported for isol ation and enrichment of phosphopeptides by chemical derivatization methods, with Goshes77 publication being the most recent. Presented are the method development step s performed for determining the best isolation and enrichment techniques for th e purposes of this project. Beginning the

PAGE 61

44 method development was the optimization of matrix conditions for MALDI, followed by optimization of conditions for both available IMAC products as well as the chemical derivatization met hod published by Goshe.77 Once optimized, both enrichment techniques (IMAC and chemical modification) were compared for determination of the best method for the goal of this project. It shoul d be noted that at the later stages of this project, a proprietary Phosphoprotein Purifi cation Kit was developed by Qiagen Inc. which will not be discussed until Chapter 4. Experimental Methods Materials and Instruments The Pro-Q Diamond Phosphoprotein Gel Stai n (Pro-Q) and SYPRO Ruby Gel Stain (SYPRO) were obtained from Molecula r Probes, Inc (Eugene, OR). 10% Bis-Tris Novex NuPAGE polyacrylamide gels were from Invitrogen (Carlsbad, CA). Protein standards ( -casein, ovalbumin, bovine serum albu min, and carbonic anhydrase) as well as the evaluated MALDI matrices ( -cyano-4-hydroxycinnamic acid and dihydroxybenzoic acid) were purchased from Sigma Aldrich (St. Louis, MO). Sequencing grade trifluoroacetic acid was obtained from Applied Biosystems (Foster City, CA). The Phosphopeptide Purification Ki t (IMAC mini-spin co lumns), (+)-biotinyliodoacetamidyl-3,6-dioxaoctaned iamine (iodoacetyl-PEObiotin) and tris(2-carboxyethyl) phosphine hydrochloride (TCEP-HCl) were from Pierce (Rockf ord, IL). Reversed-phase C18 ZipTips and ZipTipMC were obtained fr om Millipore (Billerica, MA). Sequencinggrade modified trypsin was purchased fr om Promega (Madison, WI). Acetonitirile, ethanol, methanol, formic acid, glacial ace tic acid, barium hydroxide, sodium hydroxide, ammonium bicarbonate, and Coomassie Brilli ant Blue R250 (Coomassie) were obtained

PAGE 62

45 from Fisher Scientific (Fairlawn, NJ). HP LC grade water was purchased from Burdick & Jackson (Pleasant Prairie, WI). Fluorescent imaging of gels was acquire d with a Typhoon Scanner (GE Healthcare, Piscataway, NJ). Mass spectrometric measur ements were made using a Voyager-DE Pro Biospectrometry Workstation (Applie d Biosystems, Foster City, CA). Phosphoprotein Detection 1.86 g and 0.93 g -casein, ovalbumin, bovine serum albumin, and carbonic anhydrase were separated by 1-dimensional gel electrophoresis using a 10% Bis-Tris Novex NuPAGE polyacrylamide gel with MO PS running buffer at 200V for 1 hr. The gel was stained with the Pro-Q Diamond and imaged according to the manufacturers protocol. The gel was then counter-stain ed with SYPRO Ruby, imaged with the Typhoon, and counter-stained again with C oomassie. A serial dilution (250, 100, 50, 20, 10 and 1 ng) of a mixture of equal amounts of -casein, ovalbumin, and bovine serum albumin was then separated on a gel and stai ned as described above for detection limit determination. Matrix Optimization Saturated solutions of the MALDI matrices were prepared in the selected solvent systems in Table 2-1 and mixed with a synthetic peptide (PLARTLpSVAGLPGK). Unless otherwise noted, all MALDI analyses were performed using the dried-droplet method by mixing a 1:1 ratio of analyte to ma trix and allowing the spot to dry on the MALDI plate. Two spots of each matrix-analy te solution were spotted and once dried, 25 mM ammonium citrate was spotted on top of one of the two spots and allowed to dry.

PAGE 63

46 Table 2-1. Matrix solvent systems for evaluation. Matrix Solvent combinations Additive after spotting -cyano-4hydroxycinnamic acid 50% ACN, 0.3% TFA 50% ACN, 0.3% TFA 25 mM ammonium citrate 50% ACN, 3% FA 50% ACN, 3% FA 25 mM ammonium citrate 2,5-dihydroxybenzoic acid 50% ACN, 0.3% TFA 50% ACN, 0.3% TFA 25 mM ammonium citrate 50% ACN, 3% FA 50% ACN, 3% FA 25 mM ammonium citrate ACN acetonitrile, TFA trifluoroacet ic acid, FA formic acid Mass spectra were obtained with a MALD I-TOF MS instrument equipped with a 337 nm nitrogen laser and reflectron optics. All spectra were acquired in positive ionization mode under delayed extraction cond itions in reflectron mode. Spectra were obtained with an acceleration voltage of 20 kV and 100 laser shots at a laser repetition rate of 3.0 Hz. Laser intensity, extraction delay time, grid voltage and guide wire were all adjusted to obtain the best spectrum fo r each sample. An external calibration was performed before each spectrum was obtained with a calibration mixture consisting of 1.0 pmol Des-Arg1-Bradykinin, 1.3 pmol Angiotensin I, 1.3 pmol Glu1-fibrinopeptide B, 2.0 pmol ACTH (1-17 clip), and 1.5 pmol ACTH (18-39 clip). Phosphopeptide Enrichment A stock solution of -casein digest was prepared by incubating 1 mg of protein with 10 g trypsin in 100 mM ammonium bicarbonate at 37 0C overnight. Immobilized metal-ion affinity chromatography Both the Phosphopeptide Puri fication Kit and ZipTipMC were used according to the manufacturers protocol. Briefly, this i nvolves binding the samp le at low pH (0.1% acetic acid), followed by wash steps with low organic (0.1% acetic acid with 10% acetonitrile, and water) for the removal of unphosphorylated peptides, and then elution of

PAGE 64

47 the phosphopeptides with higher pH (0.3N ammonium hydroxide solution). For the ZipTipMC, the metal ion of choice is bound to the resin. Fe3+, Ga3+, Cu2+, and Ni3+ were used as the metal ions of choice. Isolated peptides were analyzed by MALDI analysis as described above. Chemical derivatization Chemical derivatization of the synthetic phosphopeptide was performed according to the protocol published by Goshe.77 Reaction conditions such as time and temperature were investigated for optimal yield. Results and Discussion Phosphoprotein Detection Analysis of the four protein standard s (including the phosphorylated proteins, casein and ovalbumin) with the Pro-Q st ain showed selective staining of the phosphorylated proteins (Figur e 2-1). Counter-staining with both SYPRO and Coomassie proved that all four proteins were indeed on the gel and the Pro-Q stain was selective for only the phosphorylated proteins Counter-staining also indi cated that the Pro-Q stain could be as sensitive as th e SYPRO stain (as low as 1 ng)53 compared to the Coomassie stain (as low as 100 ng),53 according to the intensity of the gel bands. Subsequent detection limit experiments determined that the Pro-Q stain could detect phosphoproteins at levels as low as 20 ng (Figure 2-2). Ho wever, it was noticed that large amounts (several g) of unphosphorylated proteins were detected by th e stain but appeared more as smears rather than very distinct bands. This problem could be alleviated by always including an unphosphorylated protein as a st andard among the samples and subtracting the corresponding signal for removal of b ackground staining. However, if there is a

PAGE 65

48 greater amount of an unphosphorylated protein pres ent in the sample of interest than the standard, background subtraction may not co mpletely correct for this problem. Matrix Optimization Initial studies utilized se veral matrix compounds, with -cyano giving the best results for peptide analysis. This matrix ( -cyano matrix containing 50% ACN/0.3%TFA) was then optimized for phos phopeptide enhancement by the addition of ammonium citrate. Figure 2-3 compares spectra of a phosphopeptide of 1459 Da obtained with and without the addition of ammonium citrate. The increase in intensity of the [M+H]+ upon addition of ammonium citrate show s that ammonium citrate does enhance the phosphopeptide signal and can be used with the matrix for optimal conditions for phosphopeptides. Phosphopeptide Enrichment Figure 2-4 shows a MALDI spectrum of a -casein digest. Expected unphosphorylated peptides can be seen, however, the phosphopeptides at 2062 (1 phosphorylation site) and 3121 (4 phosphorylation sites) m/z are not visible, proving that isolation and enrichment of phosphopeptides is necessary. Isolation and enrichment of the -casein phosphopeptides was then performed wi th the Millipore ZipTipMC. A variety of metal ions were tested to identify which woul d give the best isolation and enrichment of the phosphopeptides, as well as the least nonsp ecific binding. Figure 2-5 shows spectra of the isolated phosphopeptides obtained using Fe3+, Ga3+, Cu2+, and Ni3+ proving that the ZipTipMC does isolate and en rich for phosphopeptides. Of th e four metal ions used, Ga3+ and Fe3+ seem to give the best spectra with re spect to good signal being obtained for both the monophosphorylated and multiply phosphor ylated peptides. The monophosphorylated peptide could not be found in the spectrum obtained using Ni3+, however, the signal

PAGE 66

49 intensity of the multiply phosphorylated peptid e was higher than those obtained for the other spectra. The Pierce Phosphopeptide Isolation Kit was then used to isolate and enrich the casein phosphopeptides, giving similar re sults as the ZipTipMC with Ga3+ (results not shown). These results were expected since the Pierce Kit also uses Ga3+. The sensitivity of the column was then tested by isola ting and enriching for phosphopeptides from 1 picomole of -casein digest (Figure 2-6). At firs t glance, it seemed as if only the monophosphorylated peptide was isolated (Figur e 2-6A), however, af ter the addition of ammonium citrate to the spot, significant enhancement of the multiply phosphorylated peptide (3121 m/z) was seen, however, th e monophosphorylated peptide could not be seen possibly due to ionization suppression by the multiply phosphorylated peptide. This shows that this IMAC column does isolat e and enrich phosphopeptides at very low amounts (1 pmol) and also that ammonium c itrate can be added to the MALDI spot for the enhancement of the multiply phosphorylated peptides. Both IMAC methods were compared to decide which would be best for this project. In regards to levels of sensitivity for phosphopeptides, both gave similar results (femtomole range), leaving th e decision up to ease of use and cost. Of the two IMAC products mentioned, the Pierce product was th e least labor intensive, however, it was more costly. Due to the amount of samples th at would possibly have to be enriched, the Pierce Phosphopeptide Isolation Kit was c hosen as the better method of the two. For a comparison of the chemical deriva tization and IMAC methods, the protocol published by Goshe77 was tested for its utility as an enrichment technique. A synthesized phosphopeptide PLARTLpSVAGLPGK was used fo r optimization of each reaction step

PAGE 67

50 ( -elimination, Michael additi on, and biotinylation). For -elimination, optimal reaction conditions of temperature and time were examined. Once established, parameters for the addition of the linker were examined, that is, type of linker used (ethaneditiol (EDT) or 2mercaptoethylamine), addition of a reducing agent (TCEP-HCl) to prevent dimerization of the linker molecule, length of time of the reaction, and al so whether the reaction could occur simultaneously with -elimination. At this point, remo val of the EDT prior to the addition of biotin was critical for the preven tion of competition for the biotin between the EDT or 2-mercaptoethylamine-linked peptides and the free EDT or 2mercaptoethylamine molecules. If complete removal of the linkers is not achieved, the reaction yield will be minimal, which is not acceptable in real samples where phosphorylation can be at very low levels. Several methods such as extraction with diethyl ether or size-exclusion chromatography were utilized for the efficient removal of the linker molecules. Following the removal step, the addition of biotin was performed and optimized. Figure 2-7 shows the chemical derivatiz ation of a synthetic phosphopeptide. A spectrum of the phosphopeptide was firs t obtained (Figure 2-7A), followed by elimination at determined optimal conditions (55 0C for 1 hour in 0.5 M sodium hydroxide (Figure 2-7B). Figure 2-7C show s the addition of the EDT linker at the determined optimal conditions, that is, the -elimination and EDT addition steps are performed concurrently (55 0C for 1 hour in 0.5 M sodium hydroxide, 3.6 mM TCEPHCl and a 5 molar excess of EDT), followe d by removal of unreacted EDT with a Sephadex G-10 spin column. It should be mentioned that the addition of 2mercaptoethylamine as a linker was also at tempted but was unsuccessful. Biotinylation

PAGE 68

51 was then performed for 90 minutes in the da rk with constant stirring and the reaction mixture desalted with a Sephadex G-10 spin column (Figure 2-7D). This experiment proves that the chemical derivatization steps are possible, however, the isolation of the biotinylated peptide with an avidin colu mn was not done due to the low product yield which would not be acceptable for the purposes of this project because of low amounts of starting material that would be available. This procedure proved to be very labo r-intensive and time-consuming so an alternative method was attempte d. If a label could be synthesized and stored, sample handling and derivatization time would be re duced. With this in mind, the above label (EDT and biotin) was first synthesized, purif ied and then added to the phosphopeptide, unfortunately, no reaction seemed to have occurred when monitored by MALDI-TOFMS. The synthesis of a novel phosphorylation la bel was also attempted by the reaction of H-Cys(Trt)-NH2 and biotin-LC-OSu. The expected product (702 m/z) was not found in the MALDI spectrum of the re action mixture, however, the sodiated and potassiated forms could be seen at 724 and 740 m/z. The reaction mixture was then purified on an HPLC system and the fraction of interest collected and a MALDI spectrum obtained as shown in Figure 2-8A. In order to validate th at this was indeed the product of interest, several other matrices were investigated to see if the expected product could be seen without the adducts. Figure 2-8B shows a spectrum of th e expected product with 3hydroxypicolinic acid matrix, proving that the expected product was formed. Removal of the tert-butyl protecting group from the pr oduct was then performed and the expected loss of 57 m/z is seen in Figure 2-8C, resul ting in the deprotected label at 645 m/z, as

PAGE 69

52 well as the product with a sodium adduct as well as a potassium adduct. Even though the experiments were successful in synthesi zing a possible phosphorylation label, the reaction yield was very low and reaction w ith the phosphopeptide did not seem to occur after several attempts. Of the above methods attempted for isolation and enrichment of phosphopeptides, the IMAC columns seem to be more favor able when time and labor are considered. Samples can be prepared for analysis by IMAC in less than an hour, compared to 1 day by the chemical derivatization method, that is, providing that the isola tion with the avidin column works effectively. Also, the chemi cal derivatization met hod requires a larger quantity of sample and there is a much highe r potential of sample loss due to the many steps involved. With a ll of this in mind, the IMAC columns seemed to be the best method for isolation and enrichment of phosphopeptides.

PAGE 70

53 Figure 2-1. One-Dimensional gel electrophoresis of four protein standards. A) Selective phosphoprotein detection with the ProQ Diamond Phosphoprotein Gel Stain. B) Total protein staining with SYPR O Ruby protein stain. C) Total protein staining with Coomassie R-250 protein stain.

PAGE 71

54 Figure 2-2. Pro-Q stained gel of varying concentrations of pr otein standards. The lowest detection level of the two phosphoprotein standards appears to be 20 ng.

PAGE 72

55 Figure 2-3. Comparison of matrix conditi ons for phosphopeptide sp ectra. A) Spectrum obtained with -cyano matrix. B) Spectrum obtained with -cyano matrix and the addition of ammonium citrate. E nhancement of the phosphopeptide at m/z 1460 is observed.

PAGE 73

56 Figure 2-4. -casein digest spectrum. Unphosphorylat ed peptides are observed, however, the two known phosphopeptides at m/z 2062 and 3123 are not evident.

PAGE 74

57 Figure 2-5. Comparison of di fferent metal ions on the ZipTipMC for isolation of casein phosphopeptides. A) Fe3+. B) Ga3+. C) Cu2+. D) Ni3+.

PAGE 75

58 Figure 2-6. Spectra of -casein phosphopeptides isolated by the Pierce Phosphopeptide Isolation Kit. A) Isolated phosphopeptides spotted with -cyano matrix. B) Isolated phosphopeptides spotted with -cyano matrix and the addition of ammonium citrate.

PAGE 76

59 Figure 2-7. Chemical derivatization of phosphopeptide. A) Underivatized phosphorylated peptide. B) -elimination of peptide. C) Michael addition of EDT to the -eliminated peptide. D) Biotinylation of peptide.

PAGE 77

60 Figure 2-8. Spectrum of chemical derivatiza tion label. A) Spectrum of HPLC purified label using -cyano. B) Spectrum of HPLC purified label using 3hydroxypicolinic acid matrix. C) Spectrum of the deprotected label using cyano.

PAGE 78

61 CHAPTER 3 INVESTIGATING THE AUTOPHOSPHOR YLATION SITES OF A CALCIUMDEPENDENT PROTEIN KINASE Since the regulatory properties of au tophosphorylation of CDPKs still remain unclear, one of the goals of this project was to develop methods for the identification of phosphorylation sites of phosphoryl ated proteins and subseque ntly apply these methods for autophosphorylation site iden tification of CDPKs. With the availability of three different types of mass spectrometers each po ssessing a unique feature, the utilization of all three instruments was considered be neficial for complete coverage of autophosphorylation site mapping. As mentioned in Chapter 1, both the QITMS and QqTOF-MS come equipped with similar features for phosphoryla tion detection, that is, the re constructed neutral fragment chromatogram and precursor ion scanning, re spectively, both taking advantage of the lability of the phosphate group. Also, in-line re versed phase chromatography may be used with these mass spectrometers for separati on of peptides prior to MS and MS/MS analysis. Additionally, phosphorylation site ma pping is possible with the MS/MS feature. The MALDI-TOF-MS on the other hand, does not have the capa bility of MS/MS analysis; however, it has the advantage of ra pid sample preparation and analysis as well as affordability for smaller laboratories inte rested in proteomics. With this in mind, the development of a novel, simple, cost eff ective method for preliminary phosphorylation identification was proposed. The idea behind this was to take advantage of several difficulties associated with phosphopeptide an alysis, including hydrophilicity, ionization

PAGE 79

62 efficiency and suppression effects by unphosphoryl ated peptides. Often, protein digests are desalted with a ziptip which can result in the loss of phosphopeptides due to their hydrophilicity. Phosphopeptides may not be seen in a MALDI-TOF spectrum due to either loss of the phosphopeptide in the desalti ng step or to suppressi on effects. However, if these peptides were first dephosphorylated by -elimination prior to desalting and spotting, the peptides should be visible in the spectrum. The appearance of these peptides in the spectrum when compared to the spectrum of the untreated digest would then be an indication of a previously phosphorylated peptide that has been dephosphorylated. Targeted MS/MS analysis can then be perf ormed on these peptides for phosphorylation verification. Presented is the development of a novel, cost effective method for preliminary phosphorylation site identification on a MALDI-TOF-MS. Also presented is the comparison of various available methods for autophosphorylation site mapping of a calcium-dependent protein kinase. Experimental Methods Materials and Instruments NuPAGE 10% Bis-Tris SDS-PA GE gels were obtained from Invitrogen (Carlsbad, CA). Sequencing-grade modified trypsin wa s purchased from Promega (Madison, WI). The Phosphopeptide Purification Kit (IMAC mini-spin columns) was from Pierce (Rockford, IL). Reversed-phase C18 ZipTips were obtained from Millipore (Billerica, MA). The MALDI calibration mixture Sequ azyme Peptide Mass Standards Kit was purchased from PE Biosystems (Foster City, CA). Mass spectrometric measurements were made using either an LCQ Deca ion trap (ThermoFinnigan, San Jose, CA) equipped with a PicoView electrospray ionization so urce (New Objective, Ringoes, NJ) and an

PAGE 80

63 ABI 140D Solvent Delivery System (Perkin Elmer, Wellesley, MA) or a Voyager-DE Pro Biospectrometry Workstation (Applied Bi osystems, Foster City, CA) or a QSTAR (Applied Biosystems, Foster City, CA) e quipped with an LC Packings Ultimate nanoHPLC system (LC Packings, Sunnyvale, CA). Kinase Assay and Protein Preparation Recombinant calcium-dependent protein kinases 4 and 5 (CPK4 and CPK5) from Arabidopsis thaliana were a gift from Estelle Hraba k, University of New Hampshire. Both kinases were autophosphorylated by inc ubation for one hour at room temperature in 50 mM Tris, pH 7.5, 10 mM MgCl2, 1 mM EGTA, and 1.2 mM CaCl2, and either 1 mM [ -P32] ATP or 1 mM unlabeled ATP. P32-labelled kinases as well as the untreated kinases were resolved on an SDS gel fo r autoradiography. In preparation for mass spectral analysis GST-CPK5 autophosphorylated with cold ATP was resolved on an SDS gel along with several standards ( -casein, -casein, BSA and ovalbumin). Protein bands visualized by staining with Coomassie Brilli ant Blue R250 were excised from the gel for in-gel tryptic digestion 54. An in-solution tryptic digestion of -casein was also performed, however, reduction and alkylation of the protein was omitted. Samples were then dried in a centrifugal vacuum sy stem (SpeedVac) to near dryness. Phosphopeptide Enrichment Phosphopeptides from the autophosphorylat ed CPK5 digest were isolated and enriched with the Pierce Phosphopeptide Pu rification Kit (gallium IMAC mini-spin columns) according to protocol. The eluted phosphopeptides were dried and reconstituted in 0.1% TFA. The IMAC-enriched sample as we ll as a sample of the digest (unenriched sample) were desalted with a reversed-phase C18 ZipTip according to protocol and the

PAGE 81

64 eluted peptides were dried and reconstitu ted in 5% ACN/0.5% acetic acid for mass spectrometric analysis on the LCQ Deca and QSTAR. -Elimination -elimination of the in-solution and in-gel tryptic digests of CPK5 was performed according to Knights procedure68 in which the peptides we re dissolved in a 4:3:1 solution of H2O/DMSO/ethanol (50 L) follo wed by 23 L saturated Ba(OH)2 and 1 L 5 M NaOH. Samples were incubated at 37 0C for 2 hours then neutralized with HCl and dried with a SpeedVac. The samples were reconstituted in 0.1% TFA and desalted with a reversed-phase C18 ZipTip for analysis with the MALDI-TOF mass spectrometer. It should be noted that the in-solution digest of -casein was first used to test this method followed by the in-gel digests of all the standards ( -casein, -casein, ovalbumin, and BSA). Data-dependent LC/MS/MS on the QIT-MS Samples were introduced to the ion trap mass spectrometer via an on-line reversedphase capillary HPLC (50 m i.d. x 5 cm C 18 produced in-house) with an isocratic solvent delivery at 200 nL/min with 0% Solvent A (5% ACN/95% water/0.5% acetic acid) for 5 min, and a linear gradient was performed for 20-30 min to 60% Solvent B (95% ACN/5% water/0.5% acetic acid). The tryptic peptides were detected using datadependent acquisition whereby a full scan between m/z 300.0-2000.0 was first obtained followed by a CID spectrum of the top 4 precu rsor ions (collision energy = 35%). ESI conditions were as follows: capillary temperature, 200 0C; sheath gas flow, 0 L/min; auxiliary gas flow, 0 L/min; ESI voltage, 1.20 kV; capillary voltage, 7.00 V; tube lens offset, -5.00 V. The CID mass is olation window was set to 2.00 m/z units. The subsequent information was input into the pr otein database searching programs Sequest

PAGE 82

65 (ThermoQuest, San Jose, CA, USA) or MA SCOT (Matrix Science Inc, Boston, MA, USA). Additional phosphopeptides were identif ied by investigating th e neutral fragment chromatograms at m/z 98, 49 or 32.6 and manually in terpreting the MS/MS data. MALDI-TOF-MS Analysis of the -eliminated Digests The peptide samples of the untreated and -eliminated digests of -casein, -casein, ovalbumin, BSA and CPK5 were prepared us ing a matrix solution consisting of 53 mM HCCA in 50% acetonitrile/0.1% TFA in a 1:1 ra tio, that is, 1 L sample to 1 L matrix. The samples were then air dried at room te mperature on a stainless steel plate. Mass spectra were obtained with a MALDI-TOF MS instrument equipped with a 337 nm nitrogen laser and reflectron optics. All sp ectra were acquired in positive ionization mode. The instrument was operated under dela yed extraction conditions in reflectron mode, mirror voltage ratio of 1.12, a delay time of 150 ns and grid voltage 70% of full acceleration voltage (20 kV). Spectra were obtained with 100 laser shots at a laser intensity of 2625 and laser repetition rate of 3.0 Hz. An external calibration was performed before each spectrum was obtained with a calibration mixture consisting of 1.0 pmol Des-Arg1-Bradykinin, 1.3 pmol Angiotensin I, 1.3 pmol Glu1-fibrinopeptide B, 2.0 pmol ACTH (1-17 clip), and 1.5 pmol ACTH (18-39 clip). Once the spectra of the untreated and -eliminated digests were obtained, they were superimposed and peaks present in the -eliminated digest that were not vi sible in the untreated sample were investigated further as po ssible phosphopeptides that had been dephosphorylated. Precursor Scanning Tryptic digests were loaded into an EconoTip (New Objective, Woburn, MA) emitter and interfaced with the QSTAR instrument operated in precursor ion scanning mode for the PO3 fragment ion (-79 m/z). The needle voltage was maintained between -

PAGE 83

66 700 V and -800 V while the declustering and fo cusing potentials were -70 V and -225 V, respectively. Mass spectra were acquired usi ng a stepsize of 0.25 Da and a dwell time of 40 ms per m/z. The collision energy was adjusted to higher values (70-80 eV) in order to optimize the production of the phosphate-derived low-mass fragment ions using nitrogen as the collision gas. Data-dependent LC/MS/MS on the QqTOF-MS Capillary rpHPLC separation of protein digests was performed on a 15 cm x 75 um i.d. PepMap C18 column (LC Packings, San Francisco, CA) in combination with an Ultimate Capillary HPLC System (LC Packi ngs, San Francisco, CA) operated at a flow rate of 200 nL/min. Inline mass spectrometric analysis of the column eluate was accomplished by a hybrid quadrupole time-of-f light instrument (QSTAR, Applied Biosystems, Foster City, CA) equipped with a nanoelectrospray source. A two-point mass calibration was performed in MS/MS mode of operation using the known fragment ion masses of [Glu]-Fibrinopeptide ( m/z 175.119 and m/z 1056.475). Results and Discussion Kinase Autophosphorylation Figure 3-1 shows autoradi ography results of the P32-labelled kinases demonstrating that that the kinases were indeed autophosphorylated under autophosphorylation conditions. Ion Trap: Data-dependent LC/MS/MS of the Autophosphorylated CPK5 Digest Analysis by Sequest software of data -dependent LC/MS/MS spectra of the unenriched sample identifi ed the phosphopeptide NSLNI pS MR. An additional phosphopeptide DIY pT LSRK was revealed by manual an alysis of the reconstructed neutral fragment chromatogram of 98 m/z and the corresponding MS/MS spectrum of

PAGE 84

67 947.39 m/z (Figure 3-2). Analysis of the IMAC-e nriched digest by manual inspection of the neutral fragment chromatograms at 98 m/z and 32.6 m/z revealed a triply charged phosphopeptide G pS FKDKLDEGDNNKPEDYSK at 789.73 m/z as well as the previously identified peptide DIY pT LSRK (Figure 3-3). This s hows that wit hout the use of the IMAC columns for enrichment, the former peptide would have gone unnoticed possibly due to suppression effects from th e other unphosphorylated peptides present in the digest. Overall, three aut ophosphorylation sites we re identified by this data-dependent LC/MS/MS method and IMAC enrichment. Analysis of data from a similar experi ment by MASCOT software resulted in the identification of two pho sphorylation sites (DIY pT LSR and TMRNSLNI pS MR) previously identified by using both Sequest and neutral fragment scans. Neutral fragment scans from this run revealed MS/MS data at m/z values matching m/z values of the possible phosphopeptides TPNIRDIY pT L pS R and EMFQAMDT DNS GAIT FDELK (doubly phosphorylated but ambiguous as to which two sites). However, due to insufficient fragmentation, confirmation of these phosphorylation sites was not possible. It should also be noted that the phosphopeptide G pS FKDKLDEGDNNKPEDYSK identified previously by IMAC enrichment wa s not identified in this analysis possibly due to suppression effects. Overall, utilizing these various methods with the ion trap mass spectrometer resulted in the id entification of six possible phosphorylation sites of CPK5. MALDI-TOFMS Several controls ( -casein, -casein, ovalbumin, and BSA) were first used to test the MALDI-TOFMS -elimination method as a first pass method for predicting peptide phosphorylation. Figure 3-4 shows the result s of the superimposed spectra of the untreated and -eliminated -casein digests. Peaks th at were present in the -eliminated

PAGE 85

68 spectrum and absent in the untreat ed digest were matched with m/z values of theoretically dephosphorylated phosphopeptides. That is, peaks representing potentially dephosphorylated peptides were compared to a list of all po ssible phosphorylated peptides of -casein that was generated by the MS -Digest function of the databasesearching program Protein Pros pector. Both known phosphopeptides FQ pS EEQQQTEDELQDK and ELEELNVPGEIVE pS L pSpSpS EESITR were detected by this method, however, the monophosphorylated peptide was also indicated to be doubly phosphorylated and the tetraphosphorylated peptide also indicated to have six phosphorylation sites. Although these possibly false sites were detected, both phosphopeptides were the only peptides identified as be ing phosphorylated. Similar resu lts (not shown) were also attained for the other standards (phosphoryl ated and unphosphorylated proteins) whereby the known phosphopeptides were identified, howev er, several false positives were also identified. Although there were false positives, actual phosphopeptides were identified so it was decided that this appro ach could be used as a first pass method to find possible phosphopeptide targets for MS/MS verification. This method was applied to the autophosphor ylated CPK5 digest and three possible phosphopeptides were found, RTMRNSLNI pS MR also found with the LC/MS/MS Sequest and MASCOT analys is shown previously, L pT AHEVLRHPWICENGVAPDR and IIQRGHYS ERKAAELT K (one phosphorylation site bu t exact site is uncertain) (Figure 3-5). QqTOF-MS analysis Precursor ion scanning of the autophosphoryl ated digest indicat ed the presence of at least 12 phosphorylation sites. This was done by manually performing peptide mass

PAGE 86

69 mapping whereby all m/z values with the loss of PO3 obtained from the precursor ion scan were compared to a list of theore tically possible phosphopeptides generated by Protein Prospector MS Dige st, however, sequence verifications have proven to be difficult. Figure 3-6 shows the precursor ion scan from 400-500 m/z Among these were the three phosphopeptides previously iden tified by the ion trap experiments. Data-dependent LC/MS/MS analysis fo llowed by MASCOT database searching resulted in the identification of six phosphor ylation sites, including previously identified sites DIYTL pS R, NSLNI pS MRDA and G pS FKDKLDEGDNNKPEDYSK, as well as additional sites N pS LNI pS MR, ML pS SKPAER, and TSTT NLSSNSDHSPNAADIIAQEFSK (1 site) (Figure 3-7). It should be noted that the phosphopeptide fragmentation on this instrument was significantly better than that of the ion trap as many of the b and y ions were generated on the QSTAR as compared with the ion trap which showed mainly the dephosphor ylated peptide afte r collision-induced dissociation. The development of a simple, cost effective method for phosphorylation identification using -elimination and a MALDI-TOF MS has been shown to provide targets for sequence verification. This will be useful for laborator ies with only a MALDITOF instrument, however, sequences will have to be verified once the targets are found due to possible false positives. Comparison of results obtained by these th ree instruments gave significant overlap showing that these methods are indeed compli mentary (Figure 3-8). Overall, at least 17 possible phosphorylation sites were observed (Fig ure 3-9). Of these sites, 5 sites are seen in at least two of these methods and 7 have been verified by MS/MS analysis. Combining

PAGE 87

70 the three methods gave an increased iden tification of autophosphorylation sites that would not have been possible by any one me thod. Having these complimentary results gave an increase in the confidence that th ese sites are indeed valid, however, future experiments need to be performed to verify the sites, after which we will attempt to interpret their significance. Because the autophosphorylation properti es of CDPKs are not fully understood, localization of the autop hosphorylation sites of an Arabidopsis thaliana member of this family by various complimentary methods are presented as a contri bution towards their understanding. Compilation of these autophosphor ylation sites with those of previously published sites such as that of the tomato ( Le CPK1),79 ice plant (McCPK1)9 or arabidopsis thaliana (CPK1)80 will aid in their understand ing by possibly identifying a common sequence motif. Figure 3-1. Autoradiography of SDS-PAGE separated autophosphor ylated CPDKs. A) 2 hour exposure. B) 2 day exposure.

PAGE 88

71 Figure 3-2. Autophosphorylated CPK5 digest. A) Base peak chromatogram. B) Reconstructed neutral fragment 98 chromatogram. C) MS/MS of the phosphopeptide DIY pT LSRK (947.39 m/z) located by the neutral fragment chromatogram at the retention time 17.77 min.

PAGE 89

72 Figure 3-3. IMAC enriched phosphopeptides from autophosphorylated CPK5 digest. A) Base peak chromatogram. B) MS/MS of the phosphopeptide DIY pT LSRK located by the neutral fragment 98 chromatogram at 12.06 min. C) MS/MS of the phosphopeptide G pS FKDKLDEGDNNKPEDYSK located by the neutral fragment 32.6 chromatogram at 11.80 min.

PAGE 90

73 Figure 3-4. Spectra of -casein digest. A) Un treated digest. B) -eliminated digest. Dephosphorylated phosphopeptides in the -eliminated digest are indicated with

PAGE 91

74 Figure 3-5. Spectra of CPK5 di gest. A) Untreated digest. B) -eliminated digest. Dephosphorylated phosphopeptides in the -eliminated digest are indicated with

PAGE 92

75 Figure 3-6. Precursor ion scan (400 500 m/z ) of autophosphorylated CPK5 digest. m/z values were matched with predic ted phosphorylated protein digests.

PAGE 93

76 Figure 3-7. MS/MS spectrum and corres ponding MASCOT search result of the phosphopeptide TSTT NLSSNSDHSPNAADIIAQEFSK ( m/z 939).

PAGE 94

77 Figure 3-8. Overlap of phosphopeptides identif ied by all methods. Blue squares indicate peptides identified on the MALDI-TOF -MS, green squares indicate those identified on the QIT-MS, and ye llow indicate those identified on the QSTAR. Sequences in orange are se quences identified by precursor ion scanning. Figure 3-9. Phosphorylated peptid es identified by all methods.

PAGE 95

78 CHAPTER 4 IDENTIFICATION OF IN VITRO SUBS TRATES OF A CALCIUM-DEPENDENT PROTEIN KINASE In order to gain an understanding of th e specificity and function of calciumdependent protein kinases, subs trates of these kinases need to be determined. Several approaches may be applied to achieve this goal: 1) mass spectrometric identification of substrates phosphorylated by CDPKs in vitro 2) tandem affinity purification and identification of in vivo substrates, and 3) CDPK substr ate traps with yeast two-hybrid systems. For the purposes of this project, the mass spectrometric identification of in vitro substrates approach was determined as the method of choice due to the availability of mass spectrometry facilities. Several problems are associated with this approach. First, if cell lysates are used as the source of substrate proteins, protein phos phatases and kinases al ready present in the cell lysate need to be inhibited. Second, all phosphates al ready present in the substrate proteins need to be removed prior to phosphoryl ation by the kinase in order to be able to determine the source of prot ein phosphorylation. Finally, sinc e many important substrates may be present at low abundance, isolation and enrichment of these proteins is necessary. Several commercial products were available to solve these problems. First, denaturation of proteins resu lts in the inhibition of phospha tases and kinases. Several denaturing extraction methods have been compared by Saravanan and Rose49 with the phenol-based method giving the highest protei n yield and best resolution and spot intensity with gel electrophores is. Invitrogens TRIzol Reag ent, a ready-to-use reagent

PAGE 96

79 consisting of a mono-phasic solu tion of phenol and guanidine isothiocyanate, can be used for easy extraction of proteins from plant tissue. This occurs by the disruption of cells and dissolving of cell components by the reagent, resulting in an over all purified protein extract after the removal of RNA and DNA with chloroform and ethanol, respectively. For dephosphorylation of pr oteins in the extracts, several phosphatases were available: calf intestinal alkaline phosphatase (CIP) (i n-solution or immobilized on agarose), Antarctic phosphatase, and bi otinylated phosphatase. Each of these phosphatases has its advantage. CIP in soluti on is the cheapest of the four and since derivatizations such as imm obilization are not performed on the protein, active sites should be easily accessible for dephosphorylatio n of proteins in the extract. However, inhibition of the phosphatase after treatment a nd prior to kinase treatment is necessary. Agarose-immobilized calf intestinal phosphata se and biotinylated phosphatase have the advantage that they can be removed easily from the protein extract; however, as mentioned above, immobilization can possibly inhibit the activity of a percentage of the phosphatase due to blocking of the active si tes. Antarctic phosphatase, which can be completely deactivated by a short heat trea tment, also has the advantage of facile inhibition prior to kinase treatment. Since several types of phosphatases were available, determination of the most a ppropriate phosphatase for the pur poses of this project was deemed necessary. Finally, the development of a propr ietary product from Qiagen (The PhosphoProtein Purification Kit) designed for the specific purificati on of phosphorylated proteins from complex cell lysates has ma de enrichment of low abundance proteins possible. The principle of this method is th at proteins that carr y a phosphate group on any

PAGE 97

80 amino acid are bound with high specificity to a PhosphoProtein Purification Column, while proteins without phosphate groups do not bind to the column and can therefore be found in the column flow-through fraction. Bind ing of phosphorylated proteins occurs by flowing the lysate (~0.1 mg/mL) through the co lumn at a flow rate of about 0.5 mL/min. Low lysate concentration and flow rate are used to ensure that all phosphate groups are easily accessible and are not hidden within prot ein complexes, and that complete binding of phosphorylated proteins occurs. Since 7-15% of proteins from ce lls are expected to carry one or more phosphate groups, the expect ed yield from one of these columns for 2.5 mg of protein from a cell lysate is 175 g of phosphorylated protein. The maximum binding capacity of one of these co lumns is 500 g of phosphorylated protein. This chapter demonstrates the development of a method for the identification of substrates phosphorylated by kinases in vitro A comparison of the performance of available phosphatases for dephosphorylating plant extract is shown as well as the comparison of various phosphatase inhibiti on methods. Application of the optimized dephosphorylation and inhibition steps followed by enrichment of the resulting in vitro phosphorylated substrates is shown. Experimental Methods Materials and Instruments TRIzol Reagent and NuPAGE 10% Bis-Tris SD S-PAGE gels were obtained from Invitrogen (Carlsbad, CA). Calf Intestinal Alkaline Phosphatase, Antarctic Phosphatase, Biotinylated Phosphatase and Streptavidin Magnetic Beads were purchased from New England Biolabs, Inc. (Beverly, MA). Immob ilized Calf Intestinal Alkaline Phosphatase was purchased from Sigma (St. Louis, MO ). The PhosphoProtein Purification Kit was from Qiagen Inc. (Valencia, CA). Pro-Q Diamond Phosphoprotei n Gel Stain was from

PAGE 98

81 Molecular Probes (Eugene, OR). Sequencing-gr ade modified trypsin was purchased from Promega (Madison, WI). CPK4 was a gift from Estelle Hrabak.81 Mass spectrometric measurements were made using an LCQ Deca ion trap (ThermoFinnigan, San Jose, CA) equipped w ith a PicoView electrospray ionization source (New Objective, Ringoes, NJ) and an ABI 140D Solvent Delivery System (Perkin Elmer, Wellesley, MA). Method Development Protein extract preparation Proteins were extracted from mature Arabidopsis thaliana leaves with TRIZOL Reagent by grinding the leaves with a pre-chilled mortar a nd pestle and liquid nitrogen. Ground leaves were then transf erred to a chilled Corex centrifuge tube and 5 mL of TRIZOL Reagent was added for every 500 mg plant leaves. Ground leaves were allowed to sit in the TRIZOL Reagent for 5 minutes at room temperature after which the sample was centrifuged at 10,900 x g for 5 minutes at 40C. The supernatant was then transferred to a fresh Corex tube, centrif uged for another 5 minutes, and the resulting supernatant transferred to an Oakridge tube. One mL of chloroform was then added and the tube shaken vigorously by hand for 15 s econds then allowed to stand at room temperature for 2 minutes. The sample was centrifuged at 10,900 x g for 15 minutes and the upper (aqueous) layer completely removed. 1.5 mL of absolute ethanol was then added to the lower phase (phenol-chlorofor m phase) and the sample mixed by inversion followed by incubation for 3 minutes at room temperature. Centrifugation of the sample at 483 x g was followed by the transfer of the supernatant to a fresh Corex tube. Three volumes of ice cold acetone was added to th e supernatant followed by centrifugation for 2 minutes at 2,860 x g The supernatant was then decanted, the acetone wash was

PAGE 99

82 repeated, and the pellet was air dried. The prot ein pellet was solubili zed with 1% SDS in 50 mM Tris-HCl, pH 7.5 for 30 minutes at 50 0C. The sample was diluted to obtain a final SDS concentration of 0.1%, and it was dial yzed against three changes of 1L of 50 mM Tris-HCl, pH 7.5. Dialyzed extract was con centrated by ultrafiltration in an Amicon concentrator containing a 10,000 molecula r weight cut-off membrane. Protein quantification was performed by Bradford Protein Assay. Dephosphorylation of the protein extract Several phosphatases were used for dephos phorylation optimization of the protein extract: Calf Intestinal Alkaline Phosphata se, Antarctic Phosphatase, Biotinylated Phosphatase, and Immobilized Calf Intest inal Alkaline Phospha tase. All phosphatases were used according to the manufacturers protocols and the resulting samples were separated on an SDS gel and stained with the Pro-Q Diamond Phosphoprotein Gel Stain according to the manufacturers protocol. CI P was determined as the most suitable phosphatase for these experiments. Phosphatase inhibition Since the CIP was in solution, several methods of phosphatase inhibition were tested: the addition of sodium orthovanadate, EDTA or phosphoserine. The model protein used for these studies was a fusion protein in which maltose binding protein was linked to the N-terminus of soybean serine acetyl tr ansferase (MBP-SAT), which had previously been phosphorylated by CDPK4. Varying con centrations and times for each inhibitor were used to optimize the i nhibition step. Also, kinase ac tion after addition of these inhibitors was observed to ensure that the kinase was not being i nhibited. This was done by adding CDPK4 to the inhibited sample w ith the needed phosphor ylation buffer, 1mM

PAGE 100

83 ATP, 10 mM MgCl2, 1 mM EGTA, and 1.2 mM CaCl2. The Pro-Q Diamond Phosphoprotein Gel Stain was used to monitor all of these reactions. In vitro phosphorylation of Arabidopsis thaliana extract with CDPK4 Proteins were extracted from 2 g of mature Arabidopsis thaliana according to the method described previously. Removal of phosphate groups added to the proteins by in vivo phosphorylation was then performed by addi ng 6,750 units of CIP to the extract in a buffer containing 100 mM NaCl, 50 mM Tris-HCl, 10 mM MgCl2, and 1 mM dithiothreitol, pH 7.9 at 37 0C for 1 hour. Sodium orthovanadate was then added to a final concentration of 10 mM and th e sample incubated at room temperature for 5 minutes. Buffers and excess vanadate was removed from the extract by dialysis in 50 mM Tris buffer, pH 7.5. The dephosphorylated protei n extract was then split into two equal aliquots for a control and an in vitro phosphorylated sample. Phosphorylation of the dephosphorylated extract was performed by incubation with CDPK4 in a buffer containing 1 mM ATP in 50 mM Tris, pH 7.5, 10 mM MgCl2, 1 mM EGTA, and 1.2 mM CaCl2 in the cold overnight, followed by incubati on at room temperature for one hour. To the control sample, the same conditions we re applied except for the addition of the kinase. Phosphoprotein enrichment Excess ATP was removed from the extract by dialysis in 50 mM Tris buffer, pH 7.5 followed by phosphoprotein enrichment with Qiagens PhosphoProtein Purification columns according to manufacturers protoc ol. Eluates and flow-throughs from both samples were then resolved on an SDS ge l and visualized by staining with Pro-Q Diamond Phosphoprotein Gel Stain according to protocol and Coomassie Brilliant Blue

PAGE 101

84 R250. Gel bands were excised fo r in-gel tryptic digestion54 and subsequently dried in a centrifugal vacuum system (S peedVac) to near dryness. Data-dependent LC/MS/MS on the ion trap Samples were introduced to the ion trap mass spectrometer via an on-line reversedphase capillary HPLC (75 m x 5 cm C18 Ne w Objective) with an isocratic solvent delivery at 200 nL/min with 3% Solvent A (5% ACN/95% water/0.5% acetic acid) for 5 min, and a linear gradient was performed fo r 85 min to 60% Solvent B (95% ACN/5% water/0.5% acetic acid). The tryptic peptides were dete cted using data-dependent acquisition whereby a full scan between m/z 300.0-2000.0 was first obtained followed by a CID spectrum of the top 4 precursor ions (collision energy = 38.0%). ESI conditions were as follows: capillary temperature, 190 0C; sheath gas flow, 0 L/min; auxiliary gas flow, 0 L/min; ESI voltage, 1.30 kV; capillary voltage, 36.0 V; tube lens offset, 10.00 V. The CID mass isolation window was set to 2.50 m/z units. The subsequent information was input into the protein database search ing programs Turbo Sequest (ThermoQuest, San Jose, CA, USA) or MASCOT (Matri x Science Inc, Boston, MA, USA) and comparison of results was performed with the DTA Select program (ThermoQuest, San Jose, CA, USA). Results and Discussion Four phosphatases were investigated for dephosphorylation of the protein extract: Calf Intestinal Alkaline P hosphatase, Antarctic Phosphatase Biotinylated Phosphatase, and Immobilized Calf Intestinal Alkaline P hosphatase (Figure 4-1) Dephosphorylation of the extracts was monitored by the Pro-Q stain. According to th e results obtained from the stain, CIP gave the best results with regards to dephosphorylation of the proteins (Figure 4-1A). The immobilized CIP was expect ed to be the best choice because of the

PAGE 102

85 easy removal of the phosphatase. However, the dephosphorylation efficiency was lower than expected (Figure 4-1B) and there was also leakage of the phosphatase from the agarose. The biotinylated and Antarctic phosphatases also did not dephosphorylate the extract very well (Figure 4-1C and D). Although the biotinylated phosphatase was supposed to have the advantage of easy remova l with streptavidin magnetic beads, all of the phosphatase was not removed by the b eads. Heat inactivation of Antarctic Phosphatase by heating at 65 0C for 5 minutes worked very well with regards to denaturing the phosphatase; however, most of the proteins in the extract were precipitated by the heat treatment. Overall, the soluble ca lf intestinal alkaline phosphatase was the best choice since it gave th e best dephosphorylation of th e extract. The disadvantage of this phosphatase is that it had to be in hibited prior to the phosphorylation step. Once the CIP was chosen as the phosphata se to be used for dephosphorylating the extract, a method of phosphatase inhibition th at would not interfer e with subsequent phosphorylation of the proteins by CDPK4 had to be determined. This was necessary to ensure that the phosphatase was not dephosphor ylating as the kinase was phosphorylating the proteins. Several known methods of phosphata se inhibition were te sted: inhibition of the active site of the phosphatase with vana date, binding the require d phosphatase cations with EDTA, and competitive inhibition with phosphoserine (Figure 4-2). Testing was performed by comparing the intensity of th e Pro-Q stained band of MBP-SAT with the phosphatase followed by the inhibitor then the kinase with that of a control (MBP-SAT phosphorylated by CDPK4). Inhibition was monitored by observing whether the MBPSAT was phosphorylated by the CDPK4 afte r the phosphatase in the sample was inhibited. Of the three methods tested, inhibitio n with vanadate gave the desired results of

PAGE 103

86 both inhibiting the phosphatase as well as not in terfering with the ki nase action (Figure 42B), that is, the MBP-SAT was phosphorylated by the kina se after inhibition of the phosphatase by vanadate. EDTA and phos phoserine gave the same results, autophosphorylation of the CIP (formation of ph osphoserine in the active site as a result of CIPs catalytic mechanism) and only a small degree of phosphor ylation of the MBPSAT. After comparing these three methods, it was determined that vanadate efficiently inhibited CIP and was also compatible with the CDPK4 phosphorylation step. Once these individual steps were optimized, they could be combined for the overall schematic for identification of substrates of CDPK4 (Figure 4-3) The resulting 13.8 mg protein extracted from 2.0 g of leaves was used according to the sample preparation method in Figure 4-3. Resulting SDS-PAGE separation of the two samples showed more intense bands in the CDPK4 treated sa mple when visualized by both Pro-Q and Coomassie staining (Figure 4-4). Unseparated pr otein bands were a resu lt of the presence of salts and detergents from the elution buffer of the Qiagen kit. Twenty-seven equivalent gel slices were excised and digested from both samples for data-dependent LC/MS/MS (an example may be seen in Figure 4-5) a nd the resulting spectr a analyzed by Turbo Sequest and MASCOT for protein identification. Identified prot eins were then input into the DTA Select program for comp arison of results. Prot eins that were present in only the CDPK4-treated sample were identified as possi ble substrates. Table 1 shows a list of 29 possible substrates of CDPK4, each of which wa s identified by at leas t two peptides with significant database scores. Identified along with these 29 proteins was CDPK4 which is known to be autophosphorylated on a single seri ne residue. The identification of CDPK4 in this experiment shows that the method doe s work since CDPK4 is a positive control in

PAGE 104

87 this sample. An additional 100 proteins were also identified from only the CDPK4 treated sample, however, only one peptide was matche d by the database searching program, and the probability of false identifica tion in this set of proteins is very high. Since this set of experiments were executed only once, repetition will be necessary to determine reproducibility. Ideally, identification of the phosphorylation sites of these proteins would be further proof that these prot eins are indeed substrates of the kinase. Unfortunately, phosphorylation site identification was very problematic due to the lability of the phosphate moiety. Reconstructed neutral frag ment chromatograms showed many possible phosphopeptides, however, due to insufficient fragmentation of the peptides, sequence information for these peptides was not obtained. Because only the m/z for these peptides was known, and there were many possible peptide matches for each m/z identification by database searching was not feasible. Als o, manual interpretation of the corresponding MS/MS spectra was difficult, because severa l proteins were identified for each run, which makes identifying the corresp onding protein difficult and mapping the phosphorylation site almost impossible. We have developed a method for the identifi cation of substrates of a kinase from a complex system. This was accomplished by the incorporation of several methods including the utilizati on of several newer technologies that have become available for phosphorylation analysis, namely, the Pro-Q Diamond Phosphoprotein Gel Stain for visualization of the optimization steps, a nd the Qiagen PhosphoProtein Purification Kit for simplification of the sample, both of wh ich eliminated the need for radio-labeling.

PAGE 105

88 Also shown is the comparison of various phos phatases available as well as a means of phosphatase inhibition. The obtained results of at least 29 potentia l substrates of CDPK4 may be used for future studies towards the unde rstanding of the function of CDPK4 as well as the CDPK family. This may be accomplished once substr ates of these kinases are identified and compared with the obtained results. Also, phosphorylation site mapping of the substrates may be more efficient with another type of mass spectrometer th at will produce better fragmentation for database search matching. Thes e proteins could also be used for future verification experiments where by the purified proteins may be used for testing their interaction with the CDPK4 or other related kinases.

PAGE 106

89 Figure 4-1. Pro-Q Diamond Phosphoprotein Gel Stain of Arabidopsis thaliana extract dephosphorylation with phosphatases. A) Ca lf intestinal alkaline phosphatase (in solution). B) Immobilized calf intestinal alkaline phosphatase. C) Biotinylated phosphatase. D) Antarctic phosphatase. Phosphatase present in the extract is denoted by *.

PAGE 107

90 Figure 4-2. Inhibiting Calf Intestinal Alkalin e Phosphatase. A) ProQ stained gel showing EDTA inhibition. B) ProQ stained ge l showing vanadate and phosphoserine inhibition.

PAGE 108

91 Figure 4-3. Steps for identifying in vitro CDPK4 substrates from Arabidopsis thaliana. A and B represent the control and CDPK4 treated samples, respectively.

PAGE 109

92 Figure 4-4. SDS gel electrophoresis separation of the control and CDPK4 treated Qiagen samples. A) CPK4 treated flow-through. B) CPK4 treated eluate. C) Control flow-through. D) Control eluate.

PAGE 110

93 Figure 4-5. Base peak chromatogram and fu ll MS spectrum of tryp tic digest of band 20,21.

PAGE 111

94 Table 4-1. List of proteins identif ied in only the CDPK4 treated sample. Function or location Provisional name Gel band # Photosynthesis /chloroplast chlorophyll A-B binding protein 15, b1 Photosystem I reaction centre subunit N, chloroplast precursor 16, 17,18 Photosystem II 10 kDa polypeptide, chloroplast precursor 20,21 Photosystem II 22 kDa protein, chloroplast precursor 20,21, b7 Oxygen-evolving enhancer protein 2-2, chloroplast precursor 3, 16, 20 Splice isoform 1 of chlorophyll a-b binding protein CP29.2, chloroplast precursor 7,8,9, 14, 19 Cytochrome b559 alpha subunit 17,18, b4, b11 Fructose-bisphosphate aldolase 15, b8 Fructose-bisphosphate aldolase 15 Ribosomal 50S Ribosomal protein L4, chloroplast precursor 17,18, b10 60S ribosomal protein L7-3 17,18, 19 40S ribosomal protein S18 20,21 Nuclear Nucleosome assembly protein 16, b9 Protein synthesis, processing, trafficking Nascent polypeptide-associated complex (NAC) domain 17,18 Similar to nascent polypeptide associated complex alpha chain 17,18 Peptidyl-prolyl cis-tran s isomerase, chloroplast precursor b1, b2, b8 Elongation factor 1-beta 1 16 Metabolism ATP synthase B chain 15, 20,21 ATP synthase alpha chain 3, 7,8,9, 10 Putative H+-transporting ATP synthase 10, 17,18 Glutathione S-transferase 11 20,21 Glutathione S-transferase like protein 20,21 Uridylyltransferase-related 17,18 Serine hydroxymethyltransferase, mitochondrial precursor 11 Glyceraldehyde-3-phosphate dehydrogenase B, chloroplast precursor 11, 14, 15, b8 Phosphoribulokinase, chloroplast precursor b8 Splice Isoform 2 of Ribulose bisphosphate carboxylase/oxygenase activase, chloroplast precursor 3, 4,5, 7,8,9, 10, 11, b2, b7 Signaling/Stress Calcium-dependent protei n kinase, isoform 4 2, 3, 4,5, b1, b3 Heat shock cognate 70 kDa protein 2 3, 7,8,9 Other myrosinase-associated protein 3, 11, 13, 14

PAGE 112

95 CHAPTER 5 14-3-3 INTERACTORS FROM ARABIDIOPSIS THALIANA As mentioned previously, 14-3-3 protei ns regulate many ce llular processes by binding to phosphorylated sites in diverse target proteins. In plants, they have emerged as important regulators of phosphorylated en zymes of biosynthetic metabolism, ion channels, and regulators of plant growth.36 Despite all that is known about these proteins, wide-scale studies of all interactors from a plant have not been performed to our knowledge. One means of performing this task is to utilize protein-protein affinity chromatography which is a purif ication method that exploits the unique interaction of one molecule with a second, complementary bindi ng molecule (ligand). The basic procedure for affinity chromatography is shown in Figure 5-1. First, the ligand is covalently coupled to an insoluble matrix such as agarose, and the resulting slurry poured into a column. An impure mixture containing the protein to be is olated is then applied to the column for interaction of the protein of interest with th e ligand. That is, the prot ein of interest will bind specifically to the immobilized ligand wh ile all other proteins will not. The unbound proteins will then wash off the column wh ile the bound interacting protein will remain on the column. Interacting proteins can then be removed from the column by using an elution buffer that will disrupt the proteins inte raction with the ligand. The application of affinity chromatogra phy may be used for the purposes of this project for the identification of interactor proteins from Arabidopsis thaliana Utilization of an entire leaf extract will pr ovide a means for identifying all interactors present, that is,

PAGE 113

96 not only proteins of a certain class. Additionally, th e use of denatured proteins will also aid in determining binding motifs for thes e proteins after th e investigation of phosphorylation motifs of identifie d interactors. Presented is the application of 14-3-3 affinity chromatography for isolating Arabidopsis thaliana interactor proteins. Experimental Methods Materials and Instruments TRIzol Reagent and NuPAGE 10% Bis-Tris SD S-PAGE gels were obtained from Invitrogen (Carlsbad, CA). Microcon YM10 concentrators were from Millipore (Billerica, MA). Pro-Q Diamond Phosphoprotein Gel Stai n was from Molecular Probes (Eugene, OR). Sequencing-grade modified trypsin was purchased from Promega (Madison, WI). A FAMOS autosampler from LC Packings (Sunnyvale, CA) was used for automated sample loading. Capillary rpHP LC separation of protein digests was performed on a 15 cm x 75 um i.d. PepMap C18 column from LC Packings (San Francisco, CA) in combination with an Ultim ate Capillary HPLC System (LC Packings, San Francisco, CA). Inline mass spectrometric analysis was accomplished by a hybrid quadrupole time-of-flight instrument (QSTAR Applied Biosystems, Foster City, CA) equipped with a nanoe lectrospray source. Protein Extract Preparation Proteins were extracted from mature Arabidopsis thaliana leaves with TRIZOL Reagent by grinding the leaves with a pre-chilled mortar a nd pestle and liquid nitrogen. Ground leaves were then transf erred to a chilled Corex centrifuge tube and 5 mL of TRIZOL Reagent was added for every 500 mg plant leaves. Ground leaves were allowed to sit in the TRIZOL Reagent for 5 minutes at room temperature after which the

PAGE 114

97 sample was centrifuged at 10,900 x g for 5 minutes at 40C. The supernatant was then transferred to a fresh Corex tube, centrif uged for another 5 minutes, and the resulting supernatant transferred to an Oakridge tube. One mL of chloroform was then added and the tube shaken vigorously by hand for 15 s econds then allowed to stand at room temperature for 2 minutes. The sample was centrifuged at 10,900 x g for 15 minutes and the upper (aqueous) layer completely removed. 1.5 mL of absolute ethanol was then added to the lower phase (phenol-chlorofor m phase) and the sample mixed by inversion followed by incubation for 3 minutes at room temperature. Centrifugation of the sample at 483 x g was followed by the transfer of the supernatant to a fresh Corex tube. Three volumes of ice cold acetone was added to the supernatant followed by centrifugation for 2 minutes at 2,860 x g The supernatant was then decanted, the acetone wash was repeated, and the pellet was air dried. The prot ein pellet was solubili zed with 1% SDS in 50 mM Tris-HCl, pH 7.5 for 30 minutes at 50 0C. The sample was diluted to obtain a final SDS concentration of 0.1%, and it was dial yzed against three changes of 1L of 50 mM Tris-HCl, pH 7.5. The sample was washed with binding buffer (50 mM Tris-HCl, 250 mM NaCl, and 5 mM MgCl2, pH 7.5) prior to concentra tion. Protein quantitation was then performed by Bradford Protein Assay. 14-3-3 Affinity Purification 14-3-3 isoforms immobilized on agarose we re provided by Dr. Paul Sehnke and Dr. Robert Ferl, Dept. of Horticultural Sciences, University of Florida. The 14-3-3 resin was washed with 10 mL of binding buffer ( 50 mM Tris-HCl, 250 mM NaCl, and 5 mM MgCl2, pH 7.5). Sample volume was brought up to 10 mL with binding buffer and incubated with the resin for 2 days in a co ld room with constant rotation, followed by incubation at room temperature for 1 hour. Th e protein extract was then flowed off the

PAGE 115

98 column and the column washed with 6 mL of binding buffer. Interacting proteins were then eluted from the resin by washing w ith 10 aliquots of 500 L elution buffer (1M acetic acid, 500 mM NaCl, pH 2). Once elute d, samples were neutralized with 1M NaOH and concentrated with microcon concentr ators (10,000 molecular weight cut-off) according to the manufacturers instructions. Eluted proteins were separated on an SDS gel and proteins were visualized by st aining and imaging with the Pro-Q Diamond Phosphoprotein stain and Coomassie stain. Gel bands were excised and digested with trypsin overnight. The digested samples were then collected and placed in the FAMOS autosampler for mass spectrometric analysis. Amino Acid Sequencing by nanoESI QqTOF MS Analysis MS/MS experiments for peptide sequencing were performed by loading 10 L of each sample at 10 L/min for 5 minutes with a FAMOS autosampler onto the C18 precolumn for desalting and concentration of the sample. The switching-valve position was then changed and the trapped peptides were back flushed a nd separated on the C18 nano column operated at a flow rate of 200 nL/min with a gradient of 5% to 60% acetonitrile over 30 minutes. MS/MS data we re acquired by Information Dependent Acquisition (IDA) mode. Protein Identification Fragment ion data generated by Informa tion Dependent Acquisition (IDA) via the QSTAR were searched against the NCBI nr sequence database usi ng the Mascot (Matrix Science, Boston, MA) database search engine. Probability-based MOWSE scores above the default significant value were considered for protein identification in addition to validation by manual interpretation of the MS/MS data.

PAGE 116

99 Results and Discussion Protein extraction from 2.0 g of Arabidopsis thaliana leaves resulted in 6.4 mg of protein extract for incubation with the 14-3-3 re sin. Visualization of the untreated protein extract, 14-3-3 flow-through and eluate by Pr o-Q and Coomassie staining can be seen in Figure 5-2. As can be seen, many protein band s were intensely stai ned with the Pro-Q Diamond Phosphoprotein stain, indi cating that phosphorylated proteins were affinity purified by the 14-3-3 resin. However, once st ained by Coomassie, several bands were only faintly stained indicating that these proteins were pr esent at only low amounts. Mass spectrometric analysis of the tryptic digests of these proteins resulted in the identification of 263 proteins (Table 5-1), several of whic h were identified as being phosphorylated by the MASCOT search engine. Figures 5-3 th rough 5-25 show the MASCOT search results for the identified phosphorylated proteins. It should be noted that some of these results show insufficient fragmentation resulting in overall low scores for MASCOT, however, according to the peptide mapping it seems as though most of these phosphorylation sites are real. Additionally, low fragmentation for phosphopeptides is typical for MS/MS analysis. The results presented here show the 14-33 affinity purification of 263 putative phosphorylated proteins from Arabidopsis thaliana Interpretation of the biological significance of these results is yet to be determined. However, additional experiments need to be performed to ensure that these pr oteins interacted with the 14-3-3s because of phosphorylation. One way of performing this would be to dephosphorylate the extract prior to affinity purification with the 14-3-3s. If interacti on of the protein does not occur after dephosphorylation of the pr otein then it may be concluded that the interaction was primarily because of phosphorylation.

PAGE 117

100 Figure 5-1. General procedure for affinity chromatography.

PAGE 118

101 Figure 5-2. Gel images of 14-3-3 affinity purified Arabidopsis thaliana proteins. A) ProQ Diamond Phosphoprotein Stain. B) Coom assie Stain. Lanes 1, 2, and 3 are TRIzol protein extract, 14-3-3 flow -through, and 14-3-3 eluate, respectively.

PAGE 119

102 Figure 5-3. MASCOT search results of the phosphopeptide NAGpSRLVVR (m/z 1050.5895) from photosystem I subunit PSI-E-like protein (gi|7269730) identified in band 2. Figure 5-4. MASCOT search results of the phosphopeptide QERFSQILpTPR (m/z 1453.7791) from Nuf2 family protein (gi|15219846) identif ied in band 3.

PAGE 120

103 Figure 5-5. MASCOT search results of the phosphopep tide SRLSSAAAKPSVpTA (m/z 1424.7812) from ribosomal protein S6 (gi|2662469) identif ied in band 4. Figure 5-6. MASCOT search results of the phosphopeptide AMAVpSGAVLSGIGSSFLpTGGKR (m/z 2225.2202) from Lhcb6 protein (gi|4741960) identified in band 4.

PAGE 121

104 Figure 5-7. MASCOT search results of the phosphopeptide pTWEKLQMAAR (m/z 1312.7833) from laminin receptor homologue (gi|16380) identified in band 6. Figure 5-8. MASCOT search results of the phosphopeptide LEAIEpTAK (m/z 953.5758) from cysteine synthase (gi|1488519) identified in band 6.

PAGE 122

105 Figure 5-9. MASCOT search results of the phosphopeptid e SRLpSSAAAKPSVTA (m/z 1424.7800) from ribosomal protein S6 (gi|2662469) identif ied in band 7. Figure 5-10. MASCOT sear ch results of the phosphopep tide LpSSAAAKPSVTA (m/z 1181.6309) from ribosomal protein S6-lik e (gi|7270073) identified in band 8.

PAGE 123

106 Figure 5-11. MASCOT s earch results of the phos phopeptide SRLpSSAAAKPSVTA (m/z 1424.7857) from ribosomal protei n S6-like (gi|7270073 ) identified in band 8. Figure 5-12. MASCOT sear ch results of the phosphope ptide LpSSAPAKPVAA (m/z 1090.6041) from ribosomal protein S6 (gi|2224751) identif ied in band 8.

PAGE 124

107 Figure 5-13. MASCOT sear ch results of the phosphope ptide pSRLSSAPAKPVAA (m/z 1333.7495) from ribosomal protein S6 (gi|2224751) identif ied in band 8. Figure 5-14. MASCOT sear ch results of the phosphopep tide SLGGSRPGLPpTGR (m/z 1333.7944) from unknown (gi|21592536) identified in band 8.

PAGE 125

108 Figure 5-15. MASCOT s earch results of the phos phopeptide IKLPSGpSK (m/z 908.6211) from 60S ribosomal protein L2 (gi|22135870) identified in band 10. Figure 5-16. MASCOT s earch results of the phos phopeptide IKLPSGpSK (m/z 908.6211) from putative ribosomal protei n L8 (gi|7270565) identified in band 10.

PAGE 126

109 Figure 5-17. MASCOT search result s of the phosphopeptide AESLNPLNFpSSSKPK (m/z 1697.9166) from ATP-dependent Cl p protease proteolytic subunit ClpR4, putative (gi|21593086) identified in band 11. Figure 5-18. MASCOT search results of the phosphopeptide ALVTLIEKGVAFEpTIPVDLMK (m/z 2366.3912) from Glutathione Stransferase (gi|27363352) identified in band 12.

PAGE 127

110 Figure 5-19. MASCOT sear ch results of the phosphope ptide KVEMLDGVpTIVR (m/z 1454.8631) from putative ribosomal prot ein L9 (gi|12642868) identified in band 12. Figure 5-20. MASCOT s earch results of the phos phopeptide LApTGEPLR (m/z 935.5762) from putative protein (g i|7573368) identified in band 12.

PAGE 128

111 Figure 5-21. MASCOT sear ch results of the phosphope ptide SFGLDSpSQAR (m/z 1146.6510) from putative protein 1 photosystem II oxygen-evolving complex (gi|4835233) identified in band 13. Figure 5-22. MASCOT search results of the phosphopeptide IGTADVLAFFLPGVVpSQVFK (m/z 2187.3049) from unknown protein (gi|3152582) identified in band 14.

PAGE 129

112 Figure 5-23. MASCOT sear ch results of the phosphopep tide SAGSVGKSAGpSEK (m/z 1243.7179) from putative TNP1-like transposon protein (gi|4734013) identified in band 14. Figure 5-24. MASCOT sear ch results of the phosphopep tide SpSGIALpSSRLHYASPIK (m/z 1946.0432) from peptidylprolyl isom erase ROC4 (gi|6899901) identified in band 15.

PAGE 130

113 Figure 5-25. MASCOT sear ch results of the phosphope ptide IDCEpSACVAR (m/z 1259.5482) from GAST1-like protein (g i|21618022) identifie d in band 19.

PAGE 131

114 Table 5-1. Proteins identified from 14-3-3 affinity chromatography of Arabidopsis thaliana proteins. Function or location gi number Provisional name Protein synthesis, processing, trafficking gi|4733972 14-3-3 protein (grf15), putative gi|21618266 14-3-3 protein GF14omega (grf2) gi|21553476 14-3-3-like protein GF14 iota (general regulatory factor 12) gi|2129595 14-3-3 protein homolog GF14 upsilon chain gi|166717 GF14 psi chain gi|1531629 GF14 mu gi|1256534 GF14 chi chain gi|12247993 putative 14-3-3 protein GF14epsilon gi|15222146 eukaryotic translation initiation factor, putative (EIF4B5) gi|295789 elongation factor 1-alpha gi|23397095 putative chloroplast transla tion elongation factor EF-Tu precursor gi|21592448 cytosolic cyclophilin ROC3 gi|6899901 peptidylprolyl isomerase ROC4* gi|38454142 peptide methionine sulfoxide reductase-like protein gi|804950 cysteine synthase* gi|6850835 cysteine synthase AtcysC1 gi|16305 glycine rich protein gi|16301 glycine rich protein gi|6899935 ubiquitin extension protein (UBQ5) gi|6572081 non-specific lipid transfer protein gi|26451353 putative non-specific lipid transfer protein nLTP gi|4056469 Strong similarity to gb|M95166 ADP-ribosylation factor gi|23198346 adenylate translocator gi|16160 adenosine nucleotide translocator gi|21280861 putative cytosolic factor protein gi|2058282 atranbp1a Signaling/stress gi|23197658 apospory-associated protein C-like protein gi|20148361 possible apospory-associated like protein gi|30794013 putative calcium-binding protein, calreticulin gi|1429207 annexin gi|7572922 copper homeostasis factor gi|7269278 hsp 70-like protein gi|7269269 HSP90-like protein gi|1906826 heat shock protein gi|8953699 4-nitrophenylphosphatase-like gi|34485583 extracellular calcium sensing receptor gi|7939542 AIG2 protein-like (avirulence induced gene protein) gi|21593920 signal recognition particle 19 kDa protein subunit, putative gi|21436247 putative pathogenesis-related PR-1 protein gi|8778701 Probable phosphop antothenoylcysteine decarboxylase gi|2765081 g5bf *Identified as being phosphorylated by MASCOT search engine

PAGE 132

115 Table 5-1. Continued. Function or location gi number Provisional name Metabolism gi|497788 glutathione S-transferase gi|23197738 glutathione S-transferase gi|2266412 glutathione S-transferase gi|6730003 Chain A, glutathione S-tran sferase in complex with herbicide gi|27363352 glutathione S-transferase* gi|7263568 sedoheptulose-bisphosphatase precursor gi|166702 glyceraldehyde 3-phosphate dehydrogenase A subunit gi|6721173 glyceraldehyde-3-phosphate dehydrogenase C subunit (GapC) gi|336390 glyceraldehyde 3-phosphate dehydrogenase B subunit gi|19699140 putative glyceraldehyde-3-phosphate dehydrogenase gi|23198084 phosphoglycerate kinase, putative gi|21536853 phosphoglycerate kinase, putative gi|12644295 Phosphoglycerate kinase, chloroplast precursor gi|1022805 phosphoglycerate kinase gi|30102502 aminomethyltransferase-like precursor protein gi|21689621 p-nitrophenylphosphatase-like protein gi|21360517 myrosinase-associated protein like gi|11242 fructose-bisphosphatase gi|16226653 fructose-bisphosphate aldolase like protein gi|7529717 fructose bisphosphate aldolase-like protein gi|14334740 putative fructose bisphosphate aldolase gi|7267660 H+-transporting ATP synthase-like protein gi|5708095 ATP synthase gamma chain, chloroplast precursor gi|5881679 ATPase alpha subunit gi|6522554 peroxidase gi|405611 peroxidase gi|16173 L-ascorbate peroxidase gi|9279611 peroxiredoxin Q-like protein gi|7529720 peroxiredoxin-like protein gi|11994212 glycolate oxidase gi|22531128 glycolate oxidase gi|9755746 formate dehydrogenase (FDH) gi|8809648 ripening-related protein-like; contains similarity to pectinesterase gi|8778996 Contains similarity to ferredoxin-NADP+ reductase and contains an oxidoreductase FAD/NAD-binding PF|00175 domain gi|871992 thioglucosidase gi|871990 thioglucosidase gi|7671423 microbody NAD-dependent malate dehydrogenase gi|3929649 mitochondrial NAD-dependent malate dehydrogenase *Identified as being phosphorylated by MASCOT search engine

PAGE 133

116 Table 5-1. Continued. Function or location gi number Provisional name Metabolism gi|7529712 monodehydroascorbate reductase (NADH)-like protein gi|21536569 NADH:ubiquinone oxidoreductase-like protein gi|7076787 cytosolic tr iosephosphate isomerase gi|21593477 putative triosephosphate isomerase gi|6735305 beta-1, 3-glucanase 2 (BG2) gi|4741197 aldose 1epimerase-like protein gi|23197622 phosphoribulokinase precursor gi|22327649 glycosyl hydrolase family 1 protein gi|21593796 glutamate-ammonia ligase (EC 6.3.1.2) precursor, chloroplast gi|19171469 isocitrate dehydrogenase gi|11761812 glutathione dependent dehydroascorbate reductase precursor gi|887939 Gibberellin-regulated protein GAST1 protein homolog gi|21618022 Gibberellin-regulated protein GAST1-like protein* Photosynthesis/chloroplast proteins gi|295792 ribulose bisphosphate carboxylase gi|1944432 ribulosebisphosphate carboxylase gi|16194 ribulose bisphosphate carboxylase gi|23505787 ribulose bisphosphate carboxylase small chain 1b precursor (RuBisCO small subunit 1b) gi|13926229 ribulose bisphosphate carboxylase, small subunit protein gi|5881702 large subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase gi|15450379 Rubisco activase gi|7267731 chlorophyll a/b-binding protein-like gi|6522530 chlorophyll a-b binding protein 4 precursor homolog gi|16374 chlorophyll a/b binding protein (LHCP AB 180) gi|13265501 chlorophyll a/b-binding protein gi|6403491 putative chlorophyll a/b-binding protein gi|7327812 chlorophyll a/b-binding protein CP29 gi|12642854 putative photosystem II type I chlorophyll a/b binding protein gi|430947 PSI type III chlorophyll a/b-binding protein gi|23296426 putative photosystem II type I chlorophyll a/b binding protein gi|10176952 photosystem I reaction centre subunit psaN precursor gi|7269762 putative photosystem I reaction center subunit II precursor gi|7269730 photosystem I subunit PSI-E-like protein* gi|5732201 photosystem I subunit II precursor gi|5732205 photosystem I subunit IV precursor gi|12642864 putative photosystem I subunit V precursor gi|2924280 PSI 9kD protein *Identified as being phosphorylated by MASCOT search engine

PAGE 134

117 Table 5-1. Continued. Function or location gi number Provisional name Photosynthesis/chloroplast proteins gi|9759370 photosystem II stability/assembly factor HCF136 gi|7268925 photosystem II oxygen-evolving complex protein 3like gi|1076373 photosystem II oxygen-evolving complex protein 2 (fragment) gi|21592906 23 kDa polypeptide of oxygen-evolving comlex (OEC) gi|22571 33 kDa oxygen-evolving protein gi|10177538 33 kDa polypeptide of oxygen-evolving complex gi|4835233 putative protein 1 photosystem II oxygen-evolving complex* gi|6175179 putative ribose 5-phosphate isomerase gi|99698 glutamate-ammonia ligase (EC 6.3.1.2), cytosolic (clone lambdaAtgskb6) gi|6016709 carbonic anhydrase, chloroplast precursor gi|28973241 putative thioredoxin-m gi|4741960 Lhcb6 protein* gi|4741944 Lhcb2 protein gi|20465661 putative ferredoxin-NADP+ reductase gi|7270198 putative component of cytochrome B6-F complex gi|5881707 cytochrome f gi|20465723 putative cytoch rome c oxidase subunit gi|7267158 putative fibrillin gi|20148241 putative fibrillin Ribosomal gi|2244857 ribosomal protein gi|166858 ribosomal protein gi|5881760 ribosomal protein L2 gi|2791998 ribosomal protein L4 gi|7267628 putative L5 ribosomal protein gi|21537351 putative ribosomal protein L7 gi|7270565 putative ribosomal protein L8* gi|6562271 ribosomal protein L8 homolog gi|12642868 putative ribosomal protein L9* gi|468771 ribosomal protein L12 gi|16497 Plastid ribosomal protein CL15 gi|550544 ribosomal protein L16 gi|12484209 putative ribosomal protein L18 gi|33589762 ribosomal protein like (L19) gi|3482935 Putative ribosomal protein L21 gi|2654122 ribosomal protein L23a gi|21554529 ribosomal protein L27, putative gi|19310757 putative ribosomal protein L28 gi|10177580 ribosomal protein L32 gi|23296539 putative ribosomal protein S1 gi|6598334 putative ribosomal protein S4 gi|2662469 ribosomal protein S6* gi|2224751 ribosomal protein S6* *Identified as being phosphorylated by MASCOT search engine

PAGE 135

118 Table 5-1. Continued. Function or location gi number Provisional name Ribosomal gi|166867 ribosomal protein S11 (probable start codon at bp 67) gi|7270073 ribosomal protein S6-like* gi|7269981 ribosomal protein S11-like gi|7267097 putative ribosomal protein S13 gi|5881753 ibosomal protein S15 gi|17978783 S18.A ribosomal protein gi|21592469 30S ribosomal protein S20 gi|6706420 40S ribosomal protein S2 homolog gi|7671467 40S ibosomal protein S6 gi|21592577 40S ribosomal protein S8-like gi|6682246 putative 40S ribosomal protein S23 gi|7572932 40S ribosomal protein S26 homolog gi|21436345 putative 50S ribosomal protein L27 gi|21593299 chloroplast 50S ribosomal protein L31, putative gi|7529726 60S ribosomal protein-like gi|7413634 60S ribosomal protein-like gi|23397084 putative 60S ribosomal protein gi|22135870 60S ribosomal protein L2* gi|9795602 Putative 60S ribosomal protein L6 gi|21592925 putative 60S ribosomal protein L6 gi|23308199 60S ribosomal protein L7A gi|7362767 60S ribosomal protein L7A protein gi|10176857 60S ribosomal protein L13 gi|6522565 60S ribosomal protein L13, BBC1 protein gi|21593767 60S ribosomal protein L13a gi|6642654 putative 60S ribosomal protein L13A gi|21594923 putative 60S ribosomal protein L17 gi|30102898 60S ribosomal protein L18A gi|17104643 putative 60S ribosomal protein L35 gi|21592412 putative 60S ribosomal protein L35 gi|21593754 60S ribosomal protein L37a gi|23507773 60S ribosomal protein L38-like protein Nuclear gi|20466610 putative nucl eosome assembly protein gi|99684 DNA-binding protein gi|7573443 mRNA binding protein precursor-like gi|12325357 RNA-binding protein, putative gi|16320 Histone H1-2 gi|16314 histone H1-1 gi|1617013 histone H2B like protein gi|4914322 putative transcription factor gi|19548045 putative SET protein, phospatase 2A inhibitor gi|18377868 fibrillin precursor-like protein gi|9663025 DIP2 protein gi|9663023 DIP1 protein gi|6730705 Putative phosphatase 2A inhibitor gi|15219846 Nuf2 family protein (Myosin-like protein)* Protease gi|1354272 aspartic proteinase *Identified as being phosphorylated by MASCOT search engine

PAGE 136

119 Table 5-1. Continued. Function or location gi number Provisional name Nuclear gi|21593086 ATP-dependent Clp protease proteolytic subunit ClpR4, putative* gi|27754300 putative cysteine proteinase inhibitor Protease gi|23397070 putative cysteine proteinase AALP Kinase gi|7378615 protein kinase-like gi|7268618 receptor serine/th reonine kinase-like protein Membrane gi|7268821 endomembrane-associated protein gi|9759532 outer membrane lipoprotein-like gi|1143394 V-type proton-ATPase gi|16380 laminin receptor homologue* gi|16323452 putative thaumatin protein Putative proteins gi|7635455 putative protein gi|7573368 putative protein* gi|7362762 putative protein gi|7287993 putative protein gi|7269995 putative protein gi|7269839 putative protein gi|7269521 putative protein gi|7269388 putative protein gi|7267543 putative protein gi|7019666 putative protein gi|6562320 putative protein gi|5541681 putative protein gi|4886277 putative protein gi|10045563 putative protein Unnamed/ hypothetical/ unknown Proteins gi|9758664 unnamed protein product gi|8809633 unnamed protein product gi|9755652 hypothetical protein gi|6017109 hypothetical protein gi|4185132 hypothetical protein gi|20198164 hypothetical protein gi|7485430 hypothetical protein At2g37660 [imported] gi|3513730 hypothetical protein gi|6566279 hypothetical protein RF12 gi|9795608 Unknown protein gi|9755448 Unknown protein gi|7658343 unknown protein gi|6041839 unknown protein gi|28973215 unknown protein gi|28393989 unknown protein gi|22136800 unknown protein gi|21450872 unknown protein gi|21436055 unknown protein gi|15294226 unknown protein gi|14334418 unknown protein gi|3193285 unknown protein *Identified as being phosphorylated by MASCOT search engine

PAGE 137

120 Table 5-1. Continued. Function or location gi number Provisional name Unnamed/ hypothetical/ unknown Proteins gi|3152582 unknown protein* gi|30023784 unknown protein gi|23505947 unknown protein gi|23505937 unknown protein gi|13265523 unknown protein gi|21592865 unknown gi|21592536 unknown* gi|21536786 unknown gi|4454459 expressed protein Other gi|5734756 Similar to SOUL Protein gi|4734013 putative TNP1-like transposon protein* gi|20259257 putative 33 kDa secretory protein gi|17682 Wilm's tumor suppressor homologue gi|15222811 leucine-rich repeat family protein *Identified as being phosphorylated by MASCOT search engine

PAGE 138

121 CHAPTER 6 HIGH-THROUGHPUT PHOSPHOPROTEOM ICS OF ARABIDOPSIS THALIANA Two-dimensional gel electrophoresis (2-DE) has been exploited over the past 20 or more years for the large-scale analysis of proteins. Several reviews have provided a detailed overview of this separation t echnique along with associated sample preparation.82-84 Briefly, 2-DE involves the separation of proteins by displacement in two dimensions oriented at right angles to one another. In th e first dimension, proteins are separated by isoelectric focusing (separati on by charge, pI) and in the second dimension by sodium dodecyl sulfate-polyacrylamide ge l electrophoresis (SDS-PAGE) (separation by molecular weight). Although this technique was developed in the mid-1970s,85-87 its real expansion as a useful technique had to wait for the deve lopment of microanalytical techniques that were able to identify protei ns at the small amounts availa ble in 2-DE. Initially, Edman sequencing was the microanalytical technique used for these identifications, however, mass spectrometry has recently become the meth od of choice due to its higher sensitivity as compared with Edman sequencing and its capability to characterize almost any posttranslational modification.84 This combination of 2-DE with mass spectrometry resulted in an increase in popularity of this appro ach in the early 1990s, coining the word proteomics, the large-scale study of proteins. Despite its popularity, 2-DE is associated with numerous technical difficulties and inadequacies. First, the reproducibility of th e same sample from gel to gel may vary. Second, although the resolution of 2-DE seems impr essive, it is still not sufficient for the

PAGE 139

122 enormous amounts of proteins present in the sample which can resu lt in comigration of several proteins in the same spot. Third, due to the enormous chemical diversity of proteins and their very divergent expressi on in cells and tissues protein samples may consist mainly of very abundant proteins, resulting in poor repr esentation of low abundance proteins. Fourth, r ecovery of hydrophobic proteins and large molecular weight proteins tend to be difficult due to solubi lity problems. Finally, the 2-DE technique requires a degree of technical proficiency and hi gh quality protein sample preparation is a necessity due to dramatic interference in sepa ration if contaminants are present. 2-DE is also notoriously difficult to auto mate resulting in limited throughput. Regardless of these problems, 2-DE has pr oven to be a very useful tool with regards to separating complex protein mixtures for comparative studies as well as for the identification of post-translational modificat ions such as phosphorylation. The basis of recognition of modified proteins in 2-DE is simple. If a protein becomes phosphorylated, one of the separation parameters (pI or MW ) must be altered in the modified form. Traditionally, identification of a phosphor ylation modification on a protein was determined by a shift in a proteins pI. Howe ver, the recent developm ent of the sensitive fluorescent Pro-Q Diamond Phosphoprotein Gel St ain has provided a means of selectively staining phosphorylat ed proteins in polyacrylamide gels. The detected proteins may then be excised, digested with trypsin, and analyzed by mass spectrometry for protein identification a nd phosphorylation site mapping. As in any approach, prepar ation of large quantities of samples can be a major bottleneck. Alleviating this problem has become possible with the development of robotics for post-electrophoretic preparati ons prior to mass spectrometric analysis.

PAGE 140

123 Included in these preparations are gel spot excision and digestion. Among the various robotic instrumentation availa ble for spot picking and di gestion are the InvestigatorTM ProPicTM and the InvestigatorTM ProGestTM. The InvestigatorTM ProPicTM comes integrated with three key processes: highresolution gel imaging, choices in image analys is with HT PC Analyzer Software and HT Analyzer 2-D Evolution, and protein spot cuttin g. Gel imaging is executed in a fully light tight enclosure with an interchangeable UV light source for imaging of fluorescent stained gels via CCD-based imaging t echnology. Once imaged, the HT Analyzer Software is then used for selecting sp ots for excision by generating a picklist automatically or by manual point and c lick methods. Gel plugs of 1.8 mm are then excised by using gel hydration, cutting and gent le vacuum extraction, and transferred to 96-well plates. Utilizing the ProPics fully enclosed processing environment minimizes sample handling and the risk of keratin cont amination. Also, with the automation of the instrument during picking and transferring of th e spots, the users time is freed to attend to other tasks. Once the spot picking process is completed, the 96-well plate can then be directly transferred to the InvestigatorTM ProGestTM for in-gel digestion of all samples simultaneously. The InvestigatorTM ProGestTM comes equipped with automated protocols and temperature-controlled reactions that save time and increase reproducibility. Customization of the protocols is also avai lable. The automation of the tedious manual procedures such as wash steps significantly decreases the introduction of human errors, increasing reproducibility. Additionally, samp le processing is performed by using a concentric dual needle design that delivers liquid reagents and pressurized nitrogen to

PAGE 141

124 each well. Waste removal and sample recovery without aspiration steps is possible due to the use of pierced reaction plates. This om ission of aspiration steps during sample processing eliminates the risk of sample cr oss-contamination or gel plug carry-over. Overall, the use of the ProGest system with its fully enclosed processing environment minimizes sample handling, the risk of ke ratin contamination, cross-contamination among samples, and increases throughput. The results presented here demonstrate th e utility of the phosphoprotein stain with the automation of robotics as a means for high-throughput phosphoprot eomic analysis of Arabidopsis thaliana Experimental Methods Materials and Instruments TRIzol Reagent was obtained from Invitrogen (Carlsbad, CA). The 2-D Cleanup Kit and PlusOne 2-D Quant Kit and immobilized linear gradient strips (pH 3-11) were purchased from GE Healthcare (Piscataway, NJ). 8-16% Tris-glycine polyacrylamide gels were from Bio-Rad (Hercu les, CA). Equilibration Buff ers I and II were purchased from Genomic Solutions (Ann Arbor, MI). Pro-Q Diamond Phosphoprotein Gel Stain, PeppermintStickTM Phosphoprotein Molecular Weight Ma rkers, and Colloidal Blue Gel Stain were from Invitrogen. Sequencing-gr ade modified trypsin was purchased from Promega (Madison, WI). All other solvents were obtained from Fisher Scientific (Fairlawn, NJ). Isoelectric focusing was performed on an IPGphor system (GE Healthcare, Piscataway NJ). SDS-PAGE was performe d on a Protean 2 (BioRad, Hercules, CA). Fluorescent imaging of gels was acquired with a Typhoon 8600 variable mode scanner (GE Healthcare, Piscataway, NJ). Automated s pot excision and digestion were carried out

PAGE 142

125 with the InvestigatorTM ProPicTM and the InvestigatorTM ProGestTM from Genomic Solutions. A FAMOS autosampler from LC Packings (Sunnyvale, CA) was used for automated sample loading. Capillary rpHP LC separation of protein digests was performed on a 15 cm x 75 um i.d. PepMap C18 column from LC Packings (San Francisco, CA) in combination with an Ultim ate Capillary HPLC System (LC Packings, San Francisco, CA). Inline mass spectrometric analysis was accomplished by a hybrid quadrupole time-of-flight instrument (QSTAR Applied Biosystems, Foster City, CA) equipped with a nanoe lectrospray source. Sample Preparation Proteins were extracted from mature Arabidopsis thaliana leaves with TRIZOL Reagent by grinding the leaves with a pre-chilled mortar a nd pestle and liquid nitrogen. Ground leaves were then transf erred to a chilled Corex centrifuge tube and 5 mL of TRIZOL Reagent was added for every 500 mg plant leaves. Ground leaves were allowed to sit in the TRIZOL Reagent for 5 minutes at room temperature after which the sample was centrifuged at 10,900 x g for 5 minutes at 40C. The supernatant was then transferred to a fresh Corex tube, centrif uged for another 5 minutes, and the resulting supernatant transferred to an Oakridge tube. One mL of chloroform was then added and the tube shaken vigorously by hand for 15 s econds then allowed to stand at room temperature for 2 minutes. The sample was centrifuged at 10,900 x g for 15 minutes and the upper (aqueous) layer completely removed. 1.5 mL of absolute ethanol was then added to the lower phase (phenol-chlorofor m phase) and the sample mixed by inversion followed by incubation for 3 minutes at room temperature. Centrifugation of the sample at 483 x g was followed by the transfer of the supernatant to a fresh Corex tube. Three volumes of ice cold acetone was added to the supernatant followed by centrifugation for

PAGE 143

126 2 minutes at 2,860 x g The supernatant was then decanted, the acetone wash was repeated, and the pellet was air dried. The prot ein pellet was then so lubilized with an 8M UTCEB rehydration buffer containing 8M ur ea, 2M thiourea, 4% CHAPS, 0.2% SDS, 100 mM DTT, and 0.5% pH 3-11 IPG buffer. Protein quantitation was then performed with the PlusOne 2-D Quant Kit accord ing to the manufacturers protocol. 2-Dimensional Gel Electrophoresis For IEF in the first dimension, 400 uL (1.432 mg) of the protein sample was combined with 1.2 L pH 3-11 IPG buffer and 2L Orange G dye. The solution was then applied on an immobilized linear gradient strip (18 cm, pH 3-11) for rehydration overnight. IEF was then performed using th e following three steps: 500 V for 6 hours, 1000 V for 8 hours, and 8000 V for 5 hours. The strip was then removed for equilibration (reduction and alkylation) prior to separation in the s econd dimension. First, incubation of the strip with 15 mL Equilibration Buff er I (6M urea, 130 mM DTT, 30% glycerol, 45 mM Tris base, 1.6% SDS, 0.002% Bromophe nol Blue, pH 7) for 30 minutes was executed for reduction of the proteins. Th e buffer was then decanted and 15 mL Equilibration Buffer II (6M urea, 135 mM iodoacetamide, 30% glycerol, 45 mM Tris base, 1.6% SDS, 1.6% Bromophenol Blue, pH 7) was then added for another 30 minutes of incubation for alkylation of the proteins The second dimension was then performed by running on an 18x18 cm 8-13% Tris-Glycine-SDS -PAGE gel. The gel was run at 10 mA for 1 hour followed by 24 mA for 5 hours at 10 0C. Protein Visualization and Analysis Pro-Q staining was performed according to the manufacturers protocol for visualization of phosphorylated proteins. Th e fluorescent stained gel was then imaged with the Typhoon scanner at 532/560 nm excitation/emission. Following this, the gel was

PAGE 144

127 then counter-stained with Colloidal Blue Gel Stain for total protein staining and the image captured with an office scanner. Both the Pro-Q and Colloidal images were then superimposed for identification of the protei n spots for analysis, that is, protein spots from the Pro-Q image were identified in the Colloidal image for analysis. Automated Spot Picking and Digestion The Colloidal stained gel was then imaged with the InvestigatorTM ProPicTM and the HT Analyzer Software was then used fo r selecting spots for excision by generating a picklist by manual point and cl ick methods. Gel plugs of 1.8 mm were then excised and transferred to 96-well plates for digestion with the InvestigatorTM ProGestTM according to one of the automated protocols. The digested samples were then collected and placed in the FAMOS autosampler for mass spectrometric analysis. Amino Acid Sequencing by nanoESI QqTOF MS Analysis MS/MS experiments for peptide sequencing were performed by loading 10 L of each sample at 10 L/min for 5 minutes with a FAMOS autosampler onto the C18 precolumn for desalting and concentration of the sample. The switching-valve position was then changed and the trapped peptides were back flushed a nd separated on the C18 nano column operated at a flow rate of 200 nL/min with a gradient of 5% to 60% acetonitrile over 30 minutes. MS/MS data we re acquired by Information Dependent Acquisition (IDA) mode. Protein Identification Fragment ion data generated by Informa tion Dependent Acquisition (IDA) via the QSTAR were searched against the NCBI nr sequence database usi ng the Mascot (Matrix Science, Boston, MA) database search engine. Probability-based MOWSE scores above

PAGE 145

128 the default significant value were considered for protein identification in addition to validation by manual interpretation of the MS/MS data. Results and Discussion Protein Visualization and Analysis Pro-Q staining of the 2-D gel revealed at l east 31 very intense protein spots (Figure 6-1) and a large number of weakly stai ned spots after background subtraction of unphosphorylated proteins in the standard mark ers that were co-electrophoresced on the gel. Following this, the gel was counter-staine d with Colloidal Protei n Stain (Figure 6-2) for easy visualization of protei n spots, however, several spot s were very faintly stained. Some spots that were intensel y stained by the Pro-Q were faintly stained by the Colloidal Blue, suggesting that these are highly phosphoryl ated proteins. Protei n spots were then chosen for analysis by superimposing the Pro-Q and Colloidal images of the gel. Automated Spot Picking and Digestion The gel was then imaged with the spot picker and a pick list was created for excision of protein spots (Figure 63). Operation of the InvestigatorTM ProPicTM resulted in the excision of only 18 spots. The remain ing 13 spots were cut by the robot but not removed from the gel, therefore, manual excision was performed for these spots. Automated tryptic digest ion by the InvestigatorTM ProGestTM resulted in the digestion of all samples, however, the final sample volumes varied, 2 of the 31 samples were placed in the wrong wells, and 3 samples had some of the sample placed on top of the plate beside the well instead of in the well. Protein Identification LC/MS/MS analysis of the tryptic digests of the individual spot s resulted in the identification of 2 or more proteins per spot due to co-migration of proteins on the gel.

PAGE 146

129 Causes of co-migration are overloading of the gel and inadequate separation. Several solutions to the latter problem have become av ailable, such as frac tionation of the sample prior to 2-DE or using zoom gels which sepa rate the proteins over smaller pI ranges. It should also be noted that many proteins were found in multiple spots. In 2-D analysis, multiple spots are presumably due to post-translational modifications, degradation of proteins in-vivo or in vitro or expression of differential isoforms derived from different genes.88 Overall, LC/MS/MS analysis of the tryptic peptides obtained from the excised gel spots revealed the identities of at least 138 proteins that we re identified by the MASCOT search engine with at least 2 peptides with significant scores (Table 6-1). Identified proteins were grouped according to thei r biological function or location. These 138 proteins can be considered as putative phosphoproteins of Arabidopsis thaliana however, verification by phosphorylation si te mapping proved to be difficult for several possible reasons. First, identification of the site could have been missed by database searching due to insufficient fragmentation during sequencing. Second, proteins at the acidic end of the gel were not separated very well resulting in th e identification of several proteins in one gel spot, which could result in insufficient amounts of the phosphor ylated protein for phosphorylation site mapping. Third, since spots of interest were manually selected by superimposing the Pro-Q stained image and Colloidal image, it is possible that the excised spot was not on the correct target since gel plugs of only 1.8 mm were excised. Also, since the Pro-Q stain is more sensitive than the Colloidal stain, several spots were of very low intensity making spot picking somewhat difficult. This problem could be alleviated if spots could be excised directly from the Pro-Q imaged gel, however, this

PAGE 147

130 capability was not available. Figures 6-4 through 6-12 show MS/MS spectra of possible phosphopeptides identified by the MASCOT search engine. Of these 138 proteins identified, it is expect ed that only some of these proteins are truly phosphorylated since some spots were found to contain several proteins, that is, only one of the several proteins c ould be phosphorylated. However, since phosphorylation site mapping did not give the identification of th e phosphorylation sites for most of the identified pr oteins it is difficult to say which of these 138 proteins are being detected by the Pro-Q stain. Despite the fact that the phosphorylation site was not identified it is still strongly believed that many of these proteins are indeed phosphorylated due to the fact that they were detected by the Pro-Q stain and also that many of the proteins were found in several spots running side by side which is a key indicator of a protein that has b een post-translationally modified. Overall, the results show that with th e use of robotics for spot excision and proteolytic digestion in combination with selective staining of phosphorylated proteins, phosphoproteome analysis can become high-throughput, however, improvements need to be made with the robotic instrumentati on to prevent loss of samples. Also, the development of a filter for imaging Pro-Q stained gels with the spot picker for direct excision would also increase throughput and decrease possible picking errors.

PAGE 148

131 Figure 6-1. Pro-Q Diamond Phosphoprotein Gel Stai n image indicating potential phosphorylated proteins from a twodimensional gel electrophoresis separation of Arabidopsis thaliana protein extract.

PAGE 149

132 Figure 6-2. Colloidal Gel Stain image indicatin g all proteins from a two-dimensional gel electrophoresis separation of Arabidopsis thaliana protein extract.

PAGE 150

133 Figure 6-3. InvestigatorTM ProPicTM image of Colloidal stai ned gel and corresponding spots for excision.

PAGE 151

134 Figure 6-4. MS/MS spectrum and corres ponding MASCOT search results of the phosphopeptide QTGpSLYpSDWDLLPAK ( m/z 1852.9480) from the unknown protein (gi|30725696) id entified in spot 1.

PAGE 152

135 Figure 6-5. MS/MS spectrum and corres ponding MASCOT search results of the phosphopeptide QpTGSLYpSDWDLLPAK ( m/z 1852.9516) from the unknown protein (gi|30725696) id entified in spot 2.

PAGE 153

136 Figure 6-6. MS/MS spectrum and corres ponding MASCOT search results of the phosphopeptide SGpSGDDEEGSYGR ( m/z 1394.4946) from the unknown protein (gi|23308191) id entified in spot 3.

PAGE 154

137 Figure 6-7. MS/MS spectrum and corres ponding MASCOT search results of the phosphopeptide NRSGpSGDDEEGSYGR ( m/z 1665.6520) from the unknown protein (gi|23308191) id entified in spot 3.

PAGE 155

138 Figure 6-8. MS/MS spectrum and corres ponding MASCOT search results of the phosphopeptide pTTGEEEKK ( m/z 1000.5043) from the low temperatureinduced protein (gi|509262) identified in spot 8.

PAGE 156

139 Figure 6-9. MS/MS spectrum and corres ponding MASCOT search results of the phosphopeptide SFGLDSpSQAR ( m/z 1146.6510) from putative protein 1 photosystem II oxygen-evolving complex (g i|4835233) identified in spot 13.

PAGE 157

140 Figure 6-10. MS/MS spectrum and corres ponding MASCOT search results of the phosphopeptide GTGTANQCPpTI DGGSETFSFKPGKYAGK ( m/z 2955.3193) from the 33 kDa polypeptide of oxygen-evolving complex (gi|10177538) identified in spot 14.

PAGE 158

141 Figure 6-11. MS/MS spectrum and corres ponding MASCOT search results of the phosphopeptide SPASDpTYVIFGEAK ( m/z 1563.7827) from the unknown protein (gi|48310641) id entified in spot 15.

PAGE 159

142 Figure 6-12. MS/MS spectrum and corres ponding MASCOT search results of the phosphopeptide RSPpSPPPAR ( m/z 1043.5207) from the RSZp22 protein (gi|2582645) identified in spot 19.

PAGE 160

143 Table 6-1. List of proteins identified from 2-DE spots. Function or location gi number Provisional name Spot # Photosynthesis/ chloroplast gi|5734518 photosystem I subunit III precursor 24 gi|4835233 putative protein 1 photosystem II oxygenevolving complex* 12,13 gi|23198314 putative chloroplast RNA binding protein precursor 10 gi|22571 33 kDa oxygen-evolving protein 15,16 gi|10177538 33 kDa polypeptide of oxygen-evolving complex* 12,13,14 gi|21592906 23 kDa polypeptide of oxygen-evolving complex (OEC) 18 gi|16374 chlorophyll a/b binding protein (LHCP AB 180) 10,13,17,18 gi|12642854 putative photosystem II type I chlorophyll a/b binding protein 17 gi|6539610 thioredoxin m2 29,30 gi|28973241 putative thioredoxin-m 28,29,30,31 gi|9759188 thylakoid lumenal 17.4 kD protein, chloroplast precursor 25 Ribosomal gi|23296539 putative ribosomal protein S1 7 gi|21553851 40S ribosomal protein S2 14 gi|6566279 RF12 15 gi|6598334 putative ribosomal protein S4 16 gi|21592577 40S ribosomal protein S8-like 16 gi|166867 ribosomal protein S11 (probable start codon at bp 67) 23 gi|21537373 40S ribosomal protein S12-2 20,22,23 gi|8567795 40S ribosomal protein S17, putative 24 gi|5881716 ribosomal protein S18 27 gi|4567232 40S ribosomal protein S25 28,29 gi|7270903 ribosomal protein S25 28,29 gi|21553790 50S ribosomal protein L29 24 gi|10177580 ribosomal protein L32 24 gi|7268562 ribosomal protein L32-like protein 24,26 gi|6478915 putative 60S acidic ribosomal protein P0 10 gi|5923684 putative 60S acidic ribosomal protein, 5' partial 15 gi|23397084 putative 60S ribosomal protein 23 gi|21617903 60S acidic ribosomal protein P3 24 gi|23308199 60S ribosomal protein L7A 16 gi|16497 Plastid ribosomal protein CL15 19 Metabolism gi|6056409 glutathione S-transferase 19 gi|3461818 putative glutathione S-transferase 19 gi|2266412 glutathione S-transferase 19 gi|20197312 glutathione S-transferase (GST6) 19 gi|8885622 N-glyceraldehyde-2-phosphotransferase-like 14 gi|7572929 alpha-galact osidase-like protein 6 gi|7263568 sedoheptulose-bisphosphatase precursor 8,20 *Identified as being phosphorylated by MASCOT search engine

PAGE 161

144 Table 6-1. Continued. Function or location gi number Provisional name Spot # Metabolism gi|38603934 S-adenosylmethionine:2demethylmenaquinone methyltransferaselike 21,23 gi|6728985 putative S-adenosylmethionine:2demethylmenaquinone methyltransferase 23 gi|6175179 putative ribose 5-phosphate isomerase 17 gi|5834508 thiosulfat e sulfurtransferase 10 gi|2895510 putative pectin methylesterase 19 gi|21592462 glycine decarboxylase complex H-protein 24 gi|21554426 biotin carboxyl carrier protein precursor-like protein 12 gi|1066348 acetyl-CoA carboxylas e biotin-containing subunit 12 gi|23505901 diaminopimelate epimerase like protein 10 gi|21689621 p-nitrophenylphosphatase-like protein 12,13 gi|1944432 ribulosebisphosphate carboxylase, large subunit 1,2,3 gi|16194 ribulose bisphosphate carboxylase, small subunit 27,28,29,30,31 gi|13926229 ribulose bisphosphate carboxylase, small subunit protein 27,28,29,30,31 Protein synthesis, processing, trafficking gi|6686821 elongation factor 1B alpha-subunit 11,15,16 gi|461073 elongation factor 1 beta, EF-1 beta {internal fragments} 15 gi|398606 eEF-1beta 15 gi|7270846 multiubiquitin chain binding protein (MBP1) 2,6 gi|6899901 peptidylprolyl isomerase ROC4 20,22 gi|11762200 cyclophilin/peptidylprolyl isomerase ROC4 20 gi|18398710 nascent polypeptide-associated complex (NAC) domain-containing protein 15 gi|4115918 similar to nascent polypeptide associated complex alpha chain 11,14 gi|6561948 alpha NAC-like protein 11 gi|38454142 peptide methionine sulfoxide reductase-like protein 12,13 gi|2582645 RSZp22 protein (splicing factor) 19 gi|487791 GF14omega isoform 14 gi|1256534 GF14 chi chain 14,15 gi|12247993 putative 14-3-3 protein GF14epsilon 16 gi|16974503 eukaryotic initiation factor 5A (eIF-5A) like protein 23 gi|9295717 initiation factor 5A-4, putative (eIF-5A) 20,22,23 gi|12083338 putative glycine cleavag e system H protein precursor 24 *Identified as being phosphorylated by MASCOT search engine

PAGE 162

145 Table 6-1. Continued. Function or location gi number Provisional name Spot # Signaling/stress gi|21553555 dehydration stress-induced protein 24,30 gi|6957717 putative RAD23 5 gi|388259 cor47 (low temperature-induced, ABA, severe water stress) 1,2,4 gi|509262 lti45 (low temperature-induced)* 8 gi|21593067 HSP associated protein like 5 gi|20258909 putative protein phosphatase 2C 7 gi|16223 calmodulin 27,28 gi|15217459 calreticulin 2 (CRT2) 6,7 gi|12643243 Calreticulin 2 precursor 1 gi|3212877 similar to late embryogenesis abundant proteins 9 gi|23197658 apospory-associated protein C-like protein 12,13 gi|21555216 submergence induced protein 2A 18 gi|11761812 glutathione dependent dehydroascorbate reductase precursor 17 gi|7269215 putative major latex protein 26 Nuclear gi|99684 DNA-binding protein 13,14,16 gi|549975 nucleosome assembly protein I-like protein; similar to mouse nap I 1 gi|12325357 RNA-binding protein, putative; 35994-37391 18 gi|30023780 putative RNA-binding protein 16 gi|20334750 histone H2A like protein 24 Protease gi|26452816 putative carboxyl-terminal proteinase 12 gi|10177410 HCF106 16 gi|23397070 putative cysteine proteinase AALP 17,18 gi|19548039 Cysteine proteinase RD21a precursor 11,12,15,16 gi|21595063 putative aspartyl protease 10 gi|1354272 aspartic proteinase 13 gi|10177282 20S proteasome subunit PAF1 10 gi|4887543 ATP-dependent Clp protease subunit ClpP 19 Membrane gi|7268821 endomembrane-associated protein 10 Cytoskeleton gi|7268885 tubulin beta-9 chain 5 gi|16323374 putative tubulin beta-4 chain 5 gi|20334778 beta tubulin 1, putative 5 Putative Proteins gi|9955526 putative protein 1 gi|7594543 putative protein 19 gi|7340724 putative protein 23 gi|7287993 putative protein 12,13 gi|7270957 putative protein 24 gi|7269995 putative protein 17 gi|7269580 putative protein 19 gi|7269388 putative protein 21,23 gi|7267907 putative protein 20 Putative Proteins gi|7267473 putative protein (fragment) 3 gi|6759449 putative protein 6 *Identified as being phosphorylated by MASCOT search engine

PAGE 163

146 Table 6-1. Continued. Function or location gi number Provisional name Spot # Unnamed/hypothetical/ unknown proteins gi|9758428 unnamed protein product 6 gi|16471 unnamed protein product 7 gi|15795158 unnamed protein product 16 gi|7268071 hypothetical protein 7 gi|25406857 hypothetical protein [imported] 23 gi|6466955 unknown protein 10 gi|3128209 unknown protein 6 gi|29824163 unknown protein 14 gi|28393989 unknown protein 17 gi|24417462 unknown 6 gi|24417358 unknown 20 gi|24417296 unknown 6 gi|23296326 unknown protein 8 gi|21555690 unknown 20,22,23 gi|21555497 unknown 20,21,22 gi|21554910 unknown 28,29 gi|21386997 unknown protein 29,30 gi|20258893 unknown protein 16 gi|18377530 unknown protein 12 gi|17065622 unknown protein 5 gi|13751875 unknown protein 7 gi|48310641 unknown protein* 15,16 gi|30725696 unknown protein* 1,2 gi|23308191 unknown protein* 2,3,4 gi|19699303 unknown protein 14,15 gi|19699114 unknown protein 16 gi|13265523 unknown protein 9,12,13 gi|3193303 unknown protein 11,14,15 *Identified as being phosphorylated by MASCOT search engine

PAGE 164

147 CHAPTER 7 RESEARCH OVERVIEW Mass spectrometry has emerged as th e method of choice for phosphoprotein analysis. Presented was the application of several mass spectrometric techniques for phosphorylation analysis. This research was begun by optimizing matrix conditions for MALDI-TOFMS analysis of phosphopeptides an d investigation of various enrichment techniques. Once optimized, these enrichment techniques were then applied to various phosphoproteomic projects relating to Arabidopsis thaliana Project 1, the development of methods fo r identifying phosphorylated sites of a protein, demonstrated the development of a simple, cost effective method for phosphorylation identification using -elimination and a MALDI-TOF MS. This method will be useful for laboratories with on ly a MALDI-TOF-MS instrument, however, sequences will have to be verified once the targets are found due to possible false positives. Application of the developed met hods resulted in the identification of numerous phosphorylation sites of a calcium-d ependent protein kinase. Attainment of this goal was possible with the use of th ree different types of mass spectrometers (MALDI-TOF-MS, QIT-MS and QqTOF-MS). Comparison of results obtained by these three instruments gave significant overlap showing that these methods are indeed complimentary. Although numerous autophosphory lation sites were id entified, precursor ion scanning indicated the pr esence of other sites that co uld be phosphorylated, however these sites remained elusive with the types of analyses attempted. Since the autophosphorylation properties of CDPKs are not fully understood, this work is presented

PAGE 165

148 as a contribution towards their understandi ng. Compilation of these sites along with identified sites of other CDPKs will aid in their understanding by possibly identifying a common sequence motif among these kinases. Project 2 demonstrates the implementation of a method for the identification of 28 possible substrates of a CDPK from a complex system. Achievement of this goal was accomplished by the utilization of severa l newer technologies for phosphorylation analysis that became available during the ti me of this project. Included were various separation techniques as well as mass spectrometr ic analysis. The information attained in this project may be used towards the understa nding of CDPKs, incl uding their functions based on target substrates as well as thei r target specificity based on binding motifs. However, additional work has to be performed to achieve this goal because of the lack of phosphorylation site mapping due to insuffici ent fragmentation of the phosphopeptides after MS/MS analysis. Despite this setback, th e information attained may be useful for designing future experiments, that is, the proteins identified ma y be tested individually as substrates of the kinase for verification. The application of affinity chromatography and mass spectrometery was then performed for the identification of 14-3-3 interactors from Arabidopsis thaliana protein extracts for Project 3. Illustrated in Chapte r 5 was the identifica tion of 263 interacting proteins, included among these were severa l verified phosphorylated proteins. It is expected that a much larger fraction of these identified interactors are phosphorylated, however, phosphorylation sites of these protei ns are needed. Further interpretation of these results may lead to an increase in th e knowledge of the binding specificity of these interacting proteins with the 14-3-3s whic h may lead to the discovery of additional

PAGE 166

149 binding motifs for this family of proteins. Since 14-3-3 proteins are known to interact with phosphorylated proteins, th e goal of this project was to correlate these identified proteins with substrates of the kinase to see if there was any overlap between the two families of proteins. Table 7-1 shows the protei ns that were identified in both sets of analyses. At this time, the biological si gnificance of this information has not been determined, however, this information may be used in the future for understanding the function of the 14-3-3 proteins as well as their overlap with CDPKs. Finally, Project 4 involved the developm ent of a high-throughput approach for phosphoproteomic analyses. Illustrated was the a pplication of robotic instrumentation for high-throughput phosphoproteomi c analysis of a complex Arabidopsis protein extract resulting in the identification of 138 proteins, many of which are expected phosphorylated proteins. Howeve r, verification of the phosphor ylation sites of many of these proteins was not possible. Comparison of the identified poten tial phosphorylated proteins with that of the 143-3 interactors resulted in th e identification of 39 proteins (Table 7-2), several of which were verifi ed phosphorylated proteins, increasing our confidence that these proteins are likely to be phosphorylated. The incorporation of robotics in this analysis resu lted in less time required by th e user for sample preparation, however, improvements still need to be made to the instrumentation due to occasional errors. Also, the addition of a filter specific for the Pro-Q stain will also result in an increase in throughput for this meth od. This high-throughput approach for phosphorylation analysis may be applied to othe r projects, freeing the time of the user for other important tasks.

PAGE 167

150 The results of this research are pr esented as a contri bution towards the understanding of the physiological roles of calcium-dependent protein kinases and 14-3-3 proteins. Application of the presented methods may be appl ied to other kinases in the CDPK superfamily for identification of signaling networks in which each kinase participates, imparting insight into the physiological roles of these kinases. Since these kinases are found in vascular and nonvascul ar plants, a better understanding of the regulation of key aspects such as cellular function, metabolism, and response to external signals will be gained for a wide range of plants. These findings will have a great impact on biotechnology which will in turn affect im portant areas such as agriculture. The developed methods may also be applied to other fields of study, for example, studying kinases in disease research. Table 7-1. List of proteins identified as both CDPK substr ates and 14-3-3 interactors. Function or location Provisional name Photosynthesis/chloroplast Fructose-bisphosphate aldolase Nuclear Nucleosome assembly protein Protein synthesis, processing, trafficking Peptidyl-prolyl cis-trans isom erase, chloroplast precursor Metabolism ATP synthase alpha chain Putative H+-transporting ATP synthase Glutathione S-transferase like protein Glyceraldehyde-3-phosphate dehydrogenase B, chloroplast precursor Phosphoribulokinase, chloroplast precursor Splice Isoform 2 of Ribulose bisphosphate carboxylase/oxygenase activase, Signaling/stress Heat shock cognate 70 kDa protein 2

PAGE 168

151 Table 7-2. Correlating proteins identif ied in 2-DE and 14-3-3 experiments. Function or location gi number Provisional name Photosynthesis/ chloroplast gi|16374 chlorophyll a/b binding protein (LHCP AB 180) gi|12642854 putative photosystem II type I chlorophyll a/b binding protein gi|21592906 23 kDa polypeptide of oxygen-evolving comlex (OEC) gi|22571 33 kDa oxygen-evolving protein gi|10177538 33 kDa polypeptide of oxygen-evolving complex* gi|4835233 putative protein 1 photosystem II oxygen-evolving complex* gi|28973241 putative thioredoxin-m Ribosomal gi|23296539 putative ribosomal protein S1 gi|16497 Plastid ribosomal protein CL15 gi|10177580 ribosomal protein L32 gi|6598334 putative ribosomal protein S4 gi|166867 ribosomal protein S11 (probable start codon at bp 67) gi|21592577 40S ribosomal protein S8-like gi|23397084 putative 60S ribosomal protein gi|23308199 60S ribosomal protein L7A gi|6566279 RF12 Metabolism gi|2266412 glutathione S-transferase gi|7263568 sedoheptulose-bisphosphatase precursor gi|21689621 p-nitrophenylphosph atase-like protein gi|6175179 putative ribose 5-phosphate isomerase gi|1944432 ribulosebisphosphate carboxylase gi|16194 ribulose bisphosphate carboxylase gi|13926229 ribulose bisphosphate carboxylase, small subunit protein Protein synthesis, processing, trafficking gi|1256534 GF14 chi chain gi|12247993 putative 14-3-3 protein GF14epsilon gi|6899901 peptidylprolyl isomerase ROC4* gi|38454142 peptide methionine sulfoxide reductase-like protein Signaling/stress gi|11761812 glutathione dependent dehydroascorbate reductase precursor gi|23197658 apospory-associated protein C-like protein Nuclear gi|12325357 RNA-binding protein, putative; 35994-37391 gi|99684 DNA-binding protein Protease gi|1354272 aspartic proteinase gi|23397070 putative cysteine proteinase AALP Membrane gi|7268821 endomembrane-associated protein Putative proteins gi|7269388 putative protein gi|7269995 putative protein gi|7287993 putative protein Unknown proteins gi|13265523 unknown protein gi|28393989 unknown protein *Identified as being phosphorylated by MASCOT search engine

PAGE 169

152 LIST OF REFERENCES 1. Yan, J. X.; Packer, N. H.; Gooley, A. A.; Williams, K. L. Journal of Chromatography A 1998 808 23-41. 2. Steen, H.; Kuster, B.; Mann, M. Journal of Mass Spectrometry 2001 36 782-790. 3. Bennett, K.; Stensballe, A.; Podtelej nikov, A.; Moniatte, M.; Jensen, O. Journal of Mass Spectrometry 2002 37 179-190. 4. Wang, D.; Harper, J. F.; Gribskov, M. Plant Physiology 2003 132 2152-2165. 5. Harmon, A. C.; Gribskov, M.; Gubrium, E.; Harper, J. F. New Phytologist 2001 151 175-183. 6. Harper, J. F.; Bretton, G.; Harmon, A. C. Annual Review of Plant Biology 2004 55 263-288. 7. Harmon, A. C.; Gribskov, M.; Harper, J. F. Trends in Plant Science 2000 5 154159. 8. Cheng, S. H.; Willmann, M. R.; Chen, H. C.; Sheen, J. Plant Physiology 2002 129 469-485. 9. Chehab, E. W.; Patharkar, O. R.; Hegeman, A. D.; Taybi, T.; Cushman, J. C. Plant Physiology 2004 135 1430-1446. 10. Chaudhuri, S.; Seal, A.; DasGupta, M. Plant Physiology 1999 120 859-866. 11. Saha, P.; Singh, M. Biochemistry Journal 1995 305 ( Pt 1), 205-210. 12. Anil, V. S.; Rao, K. S. Phytochemistry 2001 58 203-212. 13. Harper, J.; Harmon, A. Nature Reviews. Molecular Cell Biology 2005 In Press. 14. Dammann, C.; Ichida, A.; Hong, B. M.; Romanowsky, S. M.; Hrabak, E. M.; Harmon, A. C.; Pickard, B. G.; Harper, J. F. Plant Physiology 2003 132 18401848. 15. Lu, S. X.; Hrabak, E. M. Plant Physiology 2002 128 1008-1021. 16. Anil, V. S.; Harmon, A. C.; Rao, K. S. Plant and Cell Physiology 2003 44 367376.

PAGE 170

153 17. Pical, C.; Fredlund, K. M.; Petit, P. X.; Sommarin, M.; Moller, I. M. FEBS Letters 1993 336 347-351. 18. Kennelly, P. J.; Krebs, E. G. Journal of Biological Chemistry 1991 266 1555515558. 19. Bylund, D. B.; Krebs, E. G. Journal of Biological Chemistry 1975 250 6355-6361. 20. Huang, J. Z.; Hardin, S. C.; Huber, S. C. Archives of Biochemistry and Biophysics 2001 393 61-66. 21. Hernandez Sebastia, C.; Hardin, S. C.; Clous e, S. D.; Kieber, J. J.; Huber, S. C. Archives of Bioche mistry and Biophysics 2004 428 81-91. 22. Roberts, D. M.; Harmon, A. C. Annual Review of Plant Physiology and Molecular Biology 1992 43 375-414. 23. Neumann, G. M.; Thomas, I.; Polya, G. M. Plant Science 1996 114 45-51. 24. Huang, J. Z.; Huber, S. C. Plant and Cell Physiology 2001 42 1079-1087. 25. Loog, M.; Toomik, R.; Sak, K.; Muszynska, G.; Jarv, J.; Ek, P. European Journal of Biochemistry 2000 267 337-343. 26. Sebastia, C. H.; Hardin, S. C.; Clouse, S. D.; Kieber, J. J.; Huber, S. C. Archives of Biochemistry and Biophysics 2004 428 81-89. 27. Moore, B. W.; Perez, V. J. Physiological and Biochemical Aspects of Nervous Integration 1976 343-359. 28. Wu, K.; Rooney, M. F.; Ferl, R. J. Plant Physiology 1997 114 1421-1431. 29. Ferl, R. J.; Lu, G.; Bowen, B. W. Genetica 1994 92 129-138. 30. Ferl, R. J. Annual Review of Plant Physiolo gy and Plant Molecular Biology 1996 47 49-73. 31. Aitken, A. Trends in Cell Biology 1996 6 341-347. 32. Yaffe, M. B.; Rittinger, K.; Volinia, S. ; Caron, P. R.; Aitken, A.; Leffers, H.; Gamblin, S. J.; Smerdon, S. J.; Cantley, L. C. Cell 1997 91 961-971. 33. Muslin, A. J.; Tanner, J. W.; Allen, P. M.; Shaw, A. S. Cell 1996 84 889-897. 34. Wurtele, M.; Jelich-Ottmann, C. ; Wittinghofer, A.; Oecking, C. Embo Journal 2003 22 987-994. 35. Fuglsang, A. T.; Borch, J.; Bych, K.; Jahn, T. P.; Roepstorff, P.; Palmgren, M. G. Journal of Biological Chemistry 2003 278 42266-42272.

PAGE 171

154 36. Rubio, M. P.; Geraghty, K. M.; Wong, B. H.; Wood, N. T.; Campbell, D. G.; Morrice, N.; Mackintosh, C. Biochemistry Journal 2004 379 395-408. 37. Mann, M.; Ong, S.; Gronborg, M.; Steen, H.; Jensen, O.; Pandey, A. Trends in Biotechnology 2002 20 261-268. 38. Carr, S. A.; Hemling, M. E.; Bean, M. F.; Roberts, G. D. Analytical Chemistry 1991 63 2802-2824. 39. Siuzdak, G. Proceedings of the National Academy of Science of the United States of America 1994 91 11290-11297. 40. Hillenkamp, F.; Karas, M.; Beavis, R. C.; Chait, B. T. Analytical Chemistry 1991 63 1193A-1203A. 41. Beavis, R. C.; Chaudhary, T.; Chait, B. T. Organic Mass Spectrometry 1992 27 156-158. 42. Yates, J. R., 3rd. Journal of Mass Spectrometry 1998 33 1-19. 43. Guilhaus. Journal of Mass Spectrometry 1995 30 1519-1532. 44. March, R. E. Journal of Mass Spectrometry 1997 32 351-369. 45. March, R. E. Rapid Communications in Mass Spectrometry 1998 12 1534-1554. 46. Chernushevich, I. V.; Loboda, A. V.; Thomson, B. A. Journal of Mass Spectrometry 2001 36 849-865. 47. Yost, R. A.; Boyd, R. K. Methods of Enzymology 1990 193 154-200. 48. Aebersold, R.; Mann, M. Nature 2003 422 198-207. 49. Saravanan, R. S.; Rose, J. K. Proteomics 2004 4 2522-2532. 50. Lauber, W. M.; Carroll, J. A.; Dufield, D. R.; Kiesel, J. R.; Radabaugh, M. R.; Malone, J. P. Electrophoresis 2001 22 906-918. 51. Yan, J. X.; Harry, R. A.; Spibey, C.; Dunn, M. J. Electrophoresis 2000 21 36573665. 52. Miller, M. D., Jr.; Acey, R. A.; Lee, L. Y.; Edwards, A. J. Electrophoresis 2001 22 791-800. 53. Patton, W. F. Electrophoresis 2000 21 1123-1144. 54. Wilm, M.; Shevchenko, A.; Houthaeve, T.; Br eit, S.; Schweigerer, L.; Fotsis, T.; Mann, M. Nature 1996 379 466-469.

PAGE 172

155 55. Hunt, D. F.; Yates, J. R., 3rd; Shabanowitz, J.; Winston, S.; Hauer, C. R. Proceedings of the National Academy of Sc ience of the United States of America 1986 83 6233-6237. 56. Yates, J. R., 3rd; Eng, J. K.; McCormack, A. L.; Schieltz, D. Analytical Chemistry 1995 67 1426-1436. 57. Yates, J. R., 3rd; Eng, J. K.; McCormack, A. L. Analytical Chemistry 1995 67 3202-3210. 58. Yates, J. R., 3rd; Morgan, S. F.; Gatlin, C. L.; Griffin, P. R.; Eng, J. K. Analytical Chemistry 1998 70 3557-3565. 59. Mann, M.; Wilm, M. Analytical Chemistry 1994 66 4390-4399. 60. Perkins, D. N.; Pappin, D. J.; Creasy, D. M.; Cottrell, J. S. Electrophoresis 1999 20 3551-3567. 61. McLachlin, D.; Chait, B. Current Opinion in Chemical Biology 2001 5 591-602. 62. Ficarro, S.; McCleland, M.; Stukenberg, P.; Burke, D.; Ross, M.; Shabanowitz, J.; Hunt, D.; White, F. Nature Biotechnology 2002 20 301-305. 63. Raska, C.; Parker, C.; Dominski, Z.; Marzluff, W.; Glish, G.; Pope, R.; Borchers, C. Analytical Chemistry 2002 74 3429-3433. 64. Stensballe, A.; Andersen, S.; Jensen, O. Proteomics 2001 1 207-222. 65. Cao, P.; Stults, J. Rapid Communications in Mass Spectrometry 2000 14 16001606. 66. Zhou, W.; Merrick, A.; Khaledi, M.; Tomer, K. Journal of the American Society for Mass Spectrometry 2000 11 273-282. 67. Posewitz, M.; Tempst, P. Analytical Chemistry 1999 71 2883-2892. 68. Knight, Z.; Schilling, B.; Row, R.; Kenski, D.; Gibson, B.; Shokat, K. Nature Biotechnology 2003 21 1047-1054. 69. Oda, Y.; Nagasu, T.; Chait, B. Nature Biotechnology 2001 19 379-382. 70. McLachlin, D.; Chait, B. Analytical Chemistry 2003 75 6826-6836. 71. Adamczyk, M.; Gebler, J.; Wu, J. Rapid Communications in Mass Spectrometry 2001 15 1481-1488. 72. Goshe, M.; Veenstra, T.; Panisko, E.; Conrads, T.; Angell, N.; Smith, R. Analytical Chemistry 2002 74 607-616.

PAGE 173

156 73. Molloy, M.; Andrews, P. Analytical Chemistry 2001 73 5387-5394. 74. Andersson, L.; Porath, Journal of Analytical Biochemistry 1986 154 250-254. 75. Neville, D. C.; Rozanas, C. R.; Price, E. M.; Gruis, D. B.; Verkman, A. S.; Townsend, R. R. Protein Science 1997 6 2436-2445. 76. Andersson, L. Journal of Chromatography A 1991 539 327-334. 77. Goshe, M.; Conrads, T.; Panisko, E.; Angell, N.; Veenstra, T.; Smith, R. Analytical Chemistry 2001 73 2578-2586. 78. Asara, J. M.; Allison, J. Journal of the American Soc iety for Mass Spectrometry 1999 10 35-44. 79. Rutschmann, F.; Stalder, U.; Piotro wski, M.; Oecking, C.; Schaller, A. Plant Physiology 2002 129 156-168. 80. Hegeman, A. D.; Harms, A. C.; Sussman, M. R.; Bunner, A. E.; Harper, J. F. Journal of the American Societry for Mass Spectrometry 2004 15 647-653. 81. Hakes, D.; Dixon, J. Analytical Biochemistry 1992 202 293-298. 82. Gorg, A.; Obermaier, C.; Boguth, G.; Hard er, A.; Scheibe, B.; Wildgruber, R.; Weiss, W. Electrophoresis 2000 21 1037-1053. 83. Lilley, K. S.; Razzaq, A.; Dupree, P. Current Opinion in Chemical Biology 2002 6 46-50. 84. Rabilloud, T. Proteomics 2002 2 3-10. 85. Klose, J. Humangenetik 1975 26 231-243. 86. MacGillivray, A. J.; Rickwood, D. European Journal of Biochemistry 1974 41 181-190. 87. O'Farrell, P. H. Journal of Biological Chemistry 1975 250 4007-4021. 88. Nam, M. H.; Heo, E. J.; Kim, J. Y.; Kim, S. I.; Kwon, K. H.; Seo, J. B.; Kwon, O.; Yoo, J. S.; Park, Y. M. Proteomics 2003 3 2351-2367.

PAGE 174

157 BIOGRAPHICAL SKETCH Camille Strachan was born in St. Andrew, Jamaica on March 17, 1978. She is the daughter of Marjorie and J ohn Strachan, who also parent ed her 3 brothers, Wayne, Kenroy and Stephen. Throughout her years at Immaculate Conception High School, she aspired to be become a doctor and conseque ntly took 4 years of chemistry, biology and mathematics in order to achieve this dream Encouraging her along this path was her chemistry teacher, Ms. Bowen who always trie d to convince Camille that electrons and atoms do exist even though they cant be s een! Despite her not be ing convinced at the time, she enjoyed the excitement of chemis try labs when she could see the wonderful colors of compounds formed in Inorganic La b or the changes in color of the Bunsen burner flame when testing various metals. Also sharing words of encouragement was Sister Helen Rose who taught her Plant Biol ogy. Even though she enjoyed learning about plants, Camille never thought she would ever use this information again. Little did she know how useful this information would be to her future studies! During this time she also enjoyed extra-curricular activities such as modern dancing and playing the piano, violin and clarinet. At age 18, she decided to broaden her horizons by attending school in another country so she left the sunny island of Jama ica to attend Florida Atlantic University (FAU) in Boca Raton, FL as an honors student in the 5-year masters program in Physical Therapy. At the beginning of her second semester she began tutoring Chemistry and Algebra in the Office of Multicultural Affair s and remained with th is position throughout

PAGE 175

158 her studies at FAU. Unfortuna tely, during her sophomore year of studies she experienced the sudden loss of her beloved father. During her time at FAU she took several psychology, biology, chemistry, and anatomy and physiology classes. At the end of her junior year, her passion for chemistry t ook over and Camille decided to become a chemistry major. In May 2000, she graduated CUME LAUDE from FAU with a Bachelor of Arts in Chemistry and minor in Psychology. In Fall 2000, she enrolled in the Analyti cal Division of the Chemistry Department at the University of Florida as an Alumni Fellow. During this time she was introduced to mass spectrometry by Dr. Richard Yost and decide d to continue her research in this field. Working under the supervision of Dr. James Winefordner, she entered a collaboration project funded by NSF with Dr. Alice Ha rmon from the Botany Department and Dr. Nancy Denslow from the Protein Core Facili ty. With the help of both Drs. Harmon and Denslow, she was able to learn many sample handling techniques as well as some of the biology behind the projects, fulfilling her de sire to do both chemistry and biology. During her last year of research she was funded by a Dow Chemical Fellowship as well as a supplemental grant funded by NSF. Upon comp letion of her projects, she received her Doctor of Philosophy degree in chemistry in August 2005. Camille has been awarded a two year postdoctoral fellow position with Centocor, a Johnson & Johnson company.

PAGE 176

159 55. Hunt, D. F.; Yates, J. R., 3rd; Shabanowitz, J.; Winston, S.; Hauer, C. R. Proceedings of the National Academy of Sc ience of the United States of America 1986 83 6233-6237. 56. Yates, J. R., 3rd; Eng, J. K.; McCormack, A. L.; Schieltz, D. Analytical Chemistry 1995 67 1426-1436. 57. Yates, J. R., 3rd; Eng, J. K.; McCormack, A. L. Analytical Chemistry 1995 67 3202-3210. 58. Yates, J. R., 3rd; Morgan, S. F.; Gatlin, C. L.; Griffin, P. R.; Eng, J. K. Analytical Chemistry 1998 70 3557-3565. 59. Mann, M.; Wilm, M. Analytical Chemistry 1994 66 4390-4399. 60. Perkins, D. N.; Pappin, D. J.; Creasy, D. M.; Cottrell, J. S. Electrophoresis 1999 20 3551-3567. 61. McLachlin, D.; Chait, B. Current Opinion in Chemical Biology 2001 5 591-602. 62. Ficarro, S.; McCleland, M.; Stukenberg, P.; Burke, D.; Ross, M.; Shabanowitz, J.; Hunt, D.; White, F. Nature Biotechnology 2002 20 301-305. 63. Raska, C.; Parker, C.; Dominski, Z.; Marzluff, W.; Glish, G.; Pope, R.; Borchers, C. Analytical Chemistry 2002 74 3429-3433. 64. Stensballe, A.; Andersen, S.; Jensen, O. Proteomics 2001 1 207-222. 65. Cao, P.; Stults, J. Rapid Communications in Mass Spectrometry 2000 14 16001606. 66. Zhou, W.; Merrick, A.; Khaledi, M.; Tomer, K. Journal of the American Society for Mass Spectrometry 2000 11 273-282. 67. Posewitz, M.; Tempst, P. Analytical Chemistry 1999 71 2883-2892. 68. Knight, Z.; Schilling, B.; Row, R.; Kenski, D.; Gibson, B.; Shokat, K. Nature Biotechnology 2003 21 1047-1054. 69. Oda, Y.; Nagasu, T.; Chait, B. Nature Biotechnology 2001 19 379-382. 70. McLachlin, D.; Chait, B. Analytical Chemistry 2003 75 6826-6836. 71. Adamczyk, M.; Gebler, J.; Wu, J. Rapid Communications in Mass Spectrometry 2001 15 1481-1488. 72. Goshe, M.; Veenstra, T.; Panisko, E.; Conrads, T.; Angell, N.; Smith, R. Analytical Chemistry 2002 74 607-616.

PAGE 177

160 73. Molloy, M.; Andrews, P. Analytical Chemistry 2001 73 5387-5394. 74. Andersson, L.; Porath, Journal of Analytical Biochemistry 1986 154 250-254. 75. Neville, D. C.; Rozanas, C. R.; Price, E. M.; Gruis, D. B.; Verkman, A. S.; Townsend, R. R. Protein Science 1997 6 2436-2445. 76. Andersson, L. Journal of Chromatography A 1991 539 327-334. 77. Goshe, M.; Conrads, T.; Panisko, E.; Angell, N.; Veenstra, T.; Smith, R. Analytical Chemistry 2001 73 2578-2586. 78. Asara, J. M.; Allison, J. Journal of the American Soc iety for Mass Spectrometry 1999 10 35-44. 79. Rutschmann, F.; Stalder, U.; Piotro wski, M.; Oecking, C.; Schaller, A. Plant Physiology 2002 129 156-168. 80. Hegeman, A. D.; Harms, A. C.; Sussman, M. R.; Bunner, A. E.; Harper, J. F. Journal of the American Societry for Mass Spectrometry 2004 15 647-653. 81. Hakes, D.; Dixon, J. Analytical Biochemistry 1992 202 293-298. 82. Gorg, A.; Obermaier, C.; Boguth, G.; Hard er, A.; Scheibe, B.; Wildgruber, R.; Weiss, W. Electrophoresis 2000 21 1037-1053. 83. Lilley, K. S.; Razzaq, A.; Dupree, P. Current Opinion in Chemical Biology 2002 6 46-50. 84. Rabilloud, T. Proteomics 2002 2 3-10. 85. Klose, J. Humangenetik 1975 26 231-243. 86. MacGillivray, A. J.; Rickwood, D. European Journal of Biochemistry 1974 41 181-190. 87. O'Farrell, P. H. Journal of Biological Chemistry 1975 250 4007-4021. 88. Nam, M. H.; Heo, E. J.; Kim, J. Y.; Kim, S. I.; Kwon, K. H.; Seo, J. B.; Kwon, O.; Yoo, J. S.; Park, Y. M. Proteomics 2003 3 2351-2367.

PAGE 178

161 BIOGRAPHICAL SKETCH Camille Strachan was born in St. Andrew, Jamaica on March 17, 1978. She is the daughter of Marjorie and J ohn Strachan, who also parent ed her 3 brothers, Wayne, Kenroy and Stephen. Throughout her years at Immaculate Conception High School, she aspired to be become a doctor and conseque ntly took 4 years of chemistry, biology and mathematics in order to achieve this dream Encouraging her along this path was her chemistry teacher, Ms. Bowen who always trie d to convince Camille that electrons and atoms do exist even though they cant be s een! Despite her not be ing convinced at the time, she enjoyed the excitement of chemis try labs when she could see the wonderful colors of compounds formed in Inorganic La b or the changes in color of the Bunsen burner flame when testing various metals. Also sharing words of encouragement was Sister Helen Rose who taught her Plant Biol ogy. Even though she enjoyed learning about plants, Camille never thought she would ever use this information again. Little did she know how useful this information would be to her future studies! During this time she also enjoyed extra-curricular activities such as modern dancing and playing the piano, violin and clarinet. At age 18, she decided to broaden her horizons by attending school in another country so she left the sunny island of Jama ica to attend Florida Atlantic University (FAU) in Boca Raton, FL as an honors student in the 5-year masters program in Physical Therapy. At the beginning of her second semester she began tutoring Chemistry and Algebra in the Office of Multicultural Affair s and remained with th is position throughout

PAGE 179

162 her studies at FAU. Unfortuna tely, during her sophomore year of studies she experienced the sudden loss of her beloved father. During her time at FAU she took several psychology, biology, chemistry, and anatomy and physiology classes. At the end of her junior year, her passion for chemistry t ook over and Camille decided to become a chemistry major. In May 2000, she graduated CUME LAUDE from FAU with a Bachelor of Arts in Chemistry and minor in Psychology. In Fall 2000, she enrolled in the Analyti cal Division of the Chemistry Department at the University of Florida as an Alumni Fellow. During this time she was introduced to mass spectrometry by Dr. Richard Yost and decide d to continue her research in this field. Working under the supervision of Dr. James Winefordner, she entered a collaboration project funded by NSF with Dr. Alice Ha rmon from the Botany Department and Dr. Nancy Denslow from the Protein Core Facili ty. With the help of both Drs. Harmon and Denslow, she was able to learn many sample handling techniques as well as some of the biology behind the projects, fulfilling her de sire to do both chemistry and biology. During her last year of research she was funded by a Dow Chemical Fellowship as well as a supplemental grant funded by NSF. Upon comp letion of her projects, she received her Doctor of Philosophy degree in chemistry in August 2005. Camille has been awarded a two year postdoctoral fellow position with Centocor, a Johnson & Johnson company.


Permanent Link: http://ufdc.ufl.edu/UFE0011590/00001

Material Information

Title: Phosphoproteomics of Arabadopsis thaliana
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0011590:00001

Permanent Link: http://ufdc.ufl.edu/UFE0011590/00001

Material Information

Title: Phosphoproteomics of Arabadopsis thaliana
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0011590:00001


This item has the following downloads:


Full Text












PHOSPHOPROTEOMICS OF Arabidopsis thaliana


By

CAMILLE STRACHAN

















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Camille Nicola Strachan

































This is dedicated to my parents, John and Marjorie Strachan, who always gave me their
loving support and encouragement; to my second parents, Leon and Daphney Strachan,
for their love and support throughout the years; and to my brothers, Wayne, Ken and
Stephen, for their love and friendship.















ACKNOWLEDGMENTS

This work would not have been possible without the help of many individuals,

namely, my three supervisory committee members: Dr. James Winefordner, Dr. Nancy

Denslow, and Dr. Alice Harmon. I thank Dr. Winefordner for welcoming me into his

group and allowing me to do a collaboration project even though it was outside of his

field. Despite the fact that this project was not in his area, he always made the effort to

gain an understanding of proteomics. Being a part of his group truly meant a lot to me

and it was an inspiration to see his determination to keep his research group family by

planning group trips.

I thank Dr. Denslow for taking the time to meet with me as I tried to find my ideal

project (application of mass spectrometry to a biological project) which at the time

seemed impossible to find. She discussed the various projects available and welcomed

me into her lab to begin what seemed an impossible but exciting task. I am truly grateful

for her invaluable guidance as I entered the world of research, an area with which I had

very limited experience. Her never-ending enthusiasm and encouragement always gave

me hope that anything is possible; you just have to figure out the right approach. How

many approaches do I have to try, to figure out the right one? Her unending enthusiasm

showed me that in research you have to keep trying until you find it; and once you do,

you'll be guaranteed one of her famous handshakes in congratulations! Her passion for

science was contagious and I will always try to emulate it. The knowledge she imparted

to me in our weekly group meetings was invaluable and for that I am also truly grateful. I









am also thankful to her for starting me on the collaboration project with Dr. Harmon, to

whom I am eternally grateful.

Dr. Harmon's unending patience with me (as I tried to grasp the biological

concepts behind the projects) and with my never-ending questions about biological

sample preparation was greatly appreciated. During my initial stages of working in her

lab, she actually took the time to perform several experiments with me. I am eternally

grateful to her for those learning experiences and the privilege of performing experiments

with her. Working in her lab was a tremendous pleasure as I watched her enthusiasm for

research and her patience with her students as she explained various approaches and also

performed various experiments with them. She always made time to sit with me and

discuss any results or approaches that I had questions about. She was an advisor and also

a mentor, always lending a listening ear and giving advice as I vented my frustrations

with research or daily life. She was a true inspiration to me and has encouraged me to

continue my career in research and development. I am eternally indebted to her.

I also owe my gratitude to members of the three groups that I am associated with. I

thank the Winefordner group for their friendship and for always making me feel

welcomed even though I was not based in the Winefordner lab. I am grateful to Dr. Ben

Smith and Dr. Nico Omenetto for their input and interest in group meetings. I thank the

Protein Core Facility for the use of its mass spectrometers and for always making me feel

welcomed in the lab. Most especially I thank Dr. Stanley Stevens and Marjorie Chow. I

thank Dr. Stevens for sharing his knowledge of mass spectrometry and database

searching, as well as his renditions of catchy tunes during a frustrating day of lab work. I

am truly grateful to Marjorie Chow for her tremendous help and support with sample-









preparation techniques, and for always giving me a listening ear as I vented about one

thing or another. I thank Scott McMillen for MALDI-TOF-MS training. I extend my

gratitude to Alfred Chung for giving me insight into synthesis and for always

encouraging me to choose the best career path. I am truly grateful to Scott McLung for

always ensuring that the QITMS was running optimally. Our friendly banter in the lab

always brought a smile to my face, especially with the constant chocolate supply that

Scott always provided. I also thank Margaret Joyner from the Harmon lab for her

constant encouragement and reminiscing conversations, and for always ensuring that I

had anything I needed in the lab to achieve my goals.

Last but not least, I would like to extend my gratitude to my family and friends for

their unending love, support, encouragement, and patience. Each day I give thanks for my

good fortune of having such a wonderful family. My mother's unending strength and

determination are only a few of the qualities that I try to emulate. My father's

organizational qualities and attention to detail I hope to one day be able to achieve. I wish

he could be here to see what I have achieved. I know that he would be proud. The

presence of newly made friends and colleagues made my time here pleasurable. Finally, I

thank my Heavenly Father for giving me the ability to achieve all that I have achieved so

far. Without Him I would never have gotten this far. He has given me the strength and

perseverance to strive for excellence even when times were rough. If there is anyone I

have left out, I thank them.
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES .............. .................. ............. ................... ..... ....

LIST OF FIGURES ......... ......................... ...... ........ ............ xi

ABSTRACT ........ .............. ............. ...... ...................... xvi

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

Calcium -D dependent Protein K inases ...................................... ........... ..................1
14-3-3s .............. .............. ....... ................... ...............
Detection and Analysis of Protein Phosphorylation................................ ...............
Mass Spectrometric-based Methods.......................... ........... ..............9
Ionization techniques...................................... .......... .........................10
M ass an aly sis.................... ........................... ................ 13
Mass analysis of peptides and proteins ............ .............................. 19
D ata interpretation ................................. .......................... .............. 23
Isolation and enrichment techniques ................................. ................. 26
Immobilized metal-ion affinity chromatography ......................................26
Chemical derivatization.................................................. 28
Qiagen phosphoprotein purification kit............................................. 28

2 PHOSPHORYLATION DETECTION AND ENRICHMENT ...............................42

E x p erim mental M eth od s..................................................................... .....................44
M materials and Instrum ents ............................................................................44
Phosphoprotein D election ............................................................................45
M atrix O ptim ization ........................................ .............................................4 5
Phosphopeptide E nrichm ent .......................................... .......... ....................46
Immobilized metal-ion affinity chromatography ......................................46
Chemical derivatization..................... ...... ........................... 47
R results and D iscu ssion ................................. .............. ................ ..... .......... 47
Phosphoprotein D election ............................................................................47
M atrix O ptim ization ........................................ .............................................4 8
Phosphopeptide Enrichm ent................................................... ............... ... 48









3 INVESTIGATING THE AUTOPHOSPHORYLATION SITES OF A
CALCIUM-DEPENDENT PROTEIN KINASE ............... .................. ............61

E x p erim mental M eth od s..................................................................... .....................62
M materials and Instrum ents ............................................................................62
Kinase A ssay and Protein Preparation ..................................... .................63
Phosphopeptide E nrichm ent.......................................................... ............... 63
P -E lim in atio n ............................ ................................................. 6 4
Data-dependent LC/MS/MS on the QIT-MS ............................................... 64
MALDI-TOF-MS Analysis of the P-eliminated Digests................................65
Precursor Scanning ................. ........ ......... .................... 65
Data-dependent LC/MS/MS on the QqTOF-MS ..........................................66
R esu lts an d D iscu ssion ..................................................................... ................ .. 6 6
Kinase Autophosphorylation ........... ......... ... ..... ............ ...... ..........66
Ion Trap: Data-dependent LC/MS/MS of the Autophosphorylated CPK5
D ig e st ....................................................................................................6 6
M A L D I-T O FM S ......................... ............................ .. ........ .... ...... ...... 67
Q qT O F -M S analysis............ ......................................................... ... .... ....... 68

4 IDENTIFICATION OF IN VITRO SUBSTRATES OF A CALCIUM-
DEPENDENT PROTEIN KINASE ........................................ ....................... 78

E xperim ental M ethods........................................................................ .................. 80
M materials and Instrum ents ............................................................................80
M ethod D ev elopm ent ............................................................... ..................... 8 1
Protein extract preparation ........................................ ........ ............... 81
Dephosphorylation of the protein extract...............................................82
Phosphatase inhibition.............................. .................... .... ................... 82
In vitro phosphorylation of Arabidopsis thaliana extract with CDPK4 ......83
Phosphoprotein enrichment.......... ..............................................83
Data-dependent LC/MS/MS on the ion trap ..........................................84
R esu lts an d D iscu ssion ..................................................................... ................ .. 84

5 14-3-3 INTERACTORS FROM ARABIDIOPSIS THALIANA..............................95

E x p erim mental M eth od s..................................................................... .....................96
M materials and Instrum ents ............................................................................96
Protein Extract Preparation........................................... ........................... 96
14-3-3 Affinity Purification.............................. .... ................. 97
Amino Acid Sequencing by nanoESI QqTOF MS Analysis............................ 98
Protein Identification ......... ...... .................... ........ ... .... ................. 98
R results and D iscu ssion ................................. .............. ................ ..... .......... 99









6 HIGH-THROUGHPUT PHOSPHOPROTEOMICS OF ARABIDOPSIS
T H A L IA N A ........................................................................................................ 12 1

E xperim mental M ethods................................................................... ..................... 124
M materials and Instrum ents ...........................................................................124
Sam ple Preparation ......... .......................... .......... .. .... ............125
2-Dimensional Gel Electrophoresis.................... .... ...................... 126
Protein Visualization and Analysis ....................................... ............... 126
Autom ated Spot Picking and Digestion ....................................................... 127
Amino Acid Sequencing by nanoESI QqTOF MS Analysis...........................127
P protein Identification ....... ............. .... .......... .............. ....................127
Results and Discussion ........... ......... ..... ....... .......... ..... ........... 128
Protein Visualization and Analysis ....................................... ............... 128
Autom ated Spot Picking and Digestion .................................... ............... 128
Protein Identification .................................. ......... ....... ................128

7 RESEARCH OVERVIEW ............ ..... ................................. 147

LIST OF REFEREN CES .................................................................. ............... 152

B IO G R A PH ICA L SK ETCH ............ .................................................... .....................161
















LIST OF TABLES


Table page

2-1 Matrix solvent systems for evaluation. ....................................... ............... 46

4-1 List of proteins identified in only the CDPK4 treated sample...............................94

5-1 Proteins identified from 14-3-3 affinity chromatography ofArabidopsis thaliana
p protein s. ................................................. .............. .........................1 14

6-1 List of proteins identified from 2-DE spots. ................................ .....................143

7-1 List of proteins identified as both CDPK substrates and 14-3-3 interactors..........150

7-2 Correlating proteins identified in 2-DE and 14-3-3 experiments ...........................151
















LIST OF FIGURES

Figure page

1-1 T he M A L D I process......................................................................... .................. 30

1-2 Commonly used MALDI matrices for peptide and protein analysis .......................31

1-3 Taylor cone form action ........................................................... ........31

1-4 The electrospray ionization process ........................................ ...... ............... 32

1-5 A linear time-of-flight mass spectrometer .................................... ............... 32

1-6 A reflectron time-of-flight mass spectrometer .......................... ............... 33

1-7 Three-dimensional ideal ion trap showing the asymptotes and the dimensions
ro a n d z o .............................................................................. .. 3 3

1-8 Trajectory of a trapped ion ............. ..... .. ................. ............... ............... 34

1-9 M athieu stability diagram ............................................... .............................. 35

1-10 Tandem QqTOF mass spectrometer............................ .. .......................36

1-11 P eptide fragm entation ..................................................................... ...................37

1-12 P eptide m ass m apping ...................... .... ......... ............................. ............... 38

1-13 Tandem m ass spectrom etric sequencing ............................................. ..................39

1-14 Two of the most frequently used chelating ligands for IMAC ..............................40

1-15 Chemical derivatization method for the affinity purification of a phosphorylated
p ep tid e ..............................................................................4 1

2-1 One-Dimensional gel electrophoresis of four protein standards.............................53

2-2 Pro-Q stained gel of varying concentrations of protein standards ...........................54

2-3 Comparison of matrix conditions for phosphopeptide spectra..............................55

2-4 -casein digest spectrum ............................................................... .....................56









2-5 Comparison of different metal ions on the ZipTipMC for isolation of P-casein
phosphopeptides. A) Fe3+. B) Ga3+. C) Cu2+. D) Ni3+ ..............................57

2-6 Spectra of P-casein phosphopeptides isolated by the Pierce Phosphopeptide
Isolation K it......................................................................................... 58

2-7 Chemical derivatization of phosphopeptide .........................................................59

2-8 Spectrum of chem ical derivatization label ........................................ ...................60

3-1 Autoradiography of SDS-PAGE separated autophosphorylated CPDKs ................70

3-2 Autophosphorylated CPK 5 digest ................................. .....................71

3-3 IMAC enriched phosphopeptides from autophosphorylated CPK5 digest .............72

3-4 Spectra of 0-casein digest.............................................................. .....................73

3-5 Spectra of C PK 5 digest ................................................. ............................... 74

3-6 Precursor ion scan (400 500 m/z) of autophosphorylated CPK5 digest.................75

3-7 MS/MS spectrum and corresponding MASCOT search result of the
phosphopeptide TSTTNLSSNSDHSPNAADIIAQEFSK (m/z 939).......................76

3-8 Overlap of phosphopeptides identified by all methods............................................77

4-1 Pro-Q Diamond Phosphoprotein Gel Stain ofArabidopsis thaliana extract
dephosphorylation with phosphatases........................ ..........................89

4-2 Inhibiting Calf Intestinal Alkaline Phosphatase ...... ..... ..................................... 90

4-3 Steps for identifying in vitro CDPK4 substrates from Arabidopsis thaliana ..........91

4-4 SDS gel electrophoresis separation of the control and CDPK4 treated Qiagen
sam ples ..................................... .................. .............. .......... 92

4-5 Base peak chromatogram and full MS spectrum of tryptic digest of band 20,21....93

5-1 General procedure for affinity chromatography...............................................100

5-2 Gel images of 14-3-3 affinity purified Arabidopsis thaliana proteins.................101

5-3 MASCOT search results of the phosphopeptide NAGpSRLVVR (m/z
1050.5895) from photosystem I subunit PSI-E-like protein (gi|7269730)
identified in band 2 .......... .. .... .............................. ..... .... .. ............ 102

5-4 MASCOT search results of the phosphopeptide QERFSQILpTPR (m/z
1453.7791) from Nuf2 family protein (gi|15219846) identified in band 3............102









5-5 MASCOT search results of the phosphopeptide SRLSSAAAKPSVpTA (m/z
1424.7812) from ribosomal protein S6 (gil2662469) identified in band 4. ...........103

5-6 MASCOT search results of the phosphopeptide
AMAVpSGAVLSGIGSSFLpTGGKR (m/z 2225.2202) from Lhcb6 protein
(gi|4741960) identified in band 4. ............. .................................. ......... ........ 103

5-7 MASCOT search results of the phosphopeptide pTWEKLQMAAR (m/z
1312.7833) from laminin receptor homologue (gi116380) identified in band 6.....104

5-8 MASCOT search results of the phosphopeptide LEAIEpTAK (m/z 953.5758)
from cysteine synthase (gi11488519) identified in band 6 ............. .................. 104

5-9 MASCOT search results of the phosphopeptide SRLpSSAAAKPSVTA (m/z
1424.7800) from ribosomal protein S6 (gi|2662469) identified in band 7. ...........105

5-10 MASCOT search results of the phosphopeptide LpSSAAAKPSVTA (m/z
1181.6309) from ribosomal protein S6-like (gi17270073) identified in band 8.....105

5-11 MASCOT search results of the phosphopeptide SRLpSSAAAKPSVTA (m/z
1424.7857) from ribosomal protein S6-like (gi17270073) identified in band 8.....106

5-12 MASCOT search results of the phosphopeptide LpSSAPAKPVAA (m/z
1090.6041) from ribosomal protein S6 (gi12224751) identified in band 8. ...........106

5-13 MASCOT search results of the phosphopeptide pSRLSSAPAKPVAA (m/z
1333.7495) from ribosomal protein S6 (gi12224751) identified in band 8 ...........107

5-14 MASCOT search results of the phosphopeptide SLGGSRPGLPpTGR (m/z
1333.7944) from unknown (gi|21592536) identified in band 8 ...........................107

5-15 MASCOT search results of the phosphopeptide IKLPSGpSK (m/z 908.6211)
from 60S ribosomal protein L2 (gi|22135870) identified in band 10. .................108

5-16 MASCOT search results of the phosphopeptide IKLPSGpSK (m/z 908.6211)
from putative ribosomal protein L8 (gil7270565) identified in band 10..............108

5-17 MASCOT search results of the phosphopeptide AESLNPLNFpSSSKPK (m/z
1697.9166) from ATP-dependent Clp protease proteolytic subunit ClpR4,
putative (gi|21593086) identified in band 11. ................ ................. ........109

5-18 MASCOT search results of the phosphopeptide
ALVTLIEKGVAFEpTIPVDLMK (m/z 2366.3912) from Glutathione S-
transferase (gi|27363352) identified in band 12 ....................................................109

5-19 MASCOT search results of the phosphopeptide KVEMLDGVpTIVR (m/z
1454.8631) from putative ribosomal protein L9 (gi112642868) identified in band
12 ..... ........ ........................................... .........110









5-20 MASCOT search results of the phosphopeptide LApTGEPLR (m/z 935.5762)
from putative protein (gil7573368) identified in band 12. ............. ..................1..10

5-21 MASCOT search results of the phosphopeptide SFGLDSpSQAR (m/z
1146.6510) from putative protein 1 photosystem II oxygen-evolving complex
(gil4835233) identified in band 13 ..... ................ ....................... ..... ............... 111

5-22 MASCOT search results of the phosphopeptide
IGTADVLAFFLPGVVpSQVFK (m/z 2187.3049) from unknown protein
(gi|3152582) identified in band 14. ........................ ................ ...............111

5-23 MASCOT search results of the phosphopeptide SAGSVGKSAGpSEK (m/z
1243.7179) from putative TNP -like transposon protein (gi 4734013) identified
in band 14. ...................................................................... ..........112

5-24 MASCOT search results of the phosphopeptide SpSGIALpSSRLHYASPIK
(m/z 1946.0432) from peptidylprolyl isomerase ROC4 (gil6899901) identified in
band 15 ........... ..... .... ........... ... ............. 112

5-25 MASCOT search results of the phosphopeptide IDCEpSACVAR (m/z
1259.5482) from GAST1-like protein (gi121618022) identified in band 19.........113

6-1 Pro-Q Diamond Phosphoprotein Gel Stain image indicating potential
phosphorylated proteins from a two-dimensional gel electrophoresis separation
of Arabidopsis thaliana protein extract ....... ........ ........................................ 131

6-2 Colloidal Gel Stain image indicating all proteins from a two-dimensional gel
electrophoresis separation of Arabidopsis thaliana protein extract.....................132

6-3 InvestigatorT ProPicT image of Colloidal stained gel and corresponding spots
for excision................................... .................................. ........ 133

6-4 MS/MS spectrum and corresponding MASCOT search results of the
phosphopeptide QTGpSLYpSDWDLLPAK (m/z 1852.9480) from the unknown
protein (gi|30725696) identified in spot 1....................................... ................... 134

6-5 MS/MS spectrum and corresponding MASCOT search results of the
phosphopeptide QpTGSLYpSDWDLLPAK (m/z 1852.9516) from the unknown
protein (gi|30725696) identified in spot 2 ............................ ................... ....... 135

6-6 MS/MS spectrum and corresponding MASCOT search results of the
phosphopeptide SGpSGDDEEGSYGR (m/z 1394.4946) from the unknown
protein (gil23308191) identified in spot 3........................................................ 136

6-7 MS/MS spectrum and corresponding MASCOT search results of the
phosphopeptide NRSGpSGDDEEGSYGR (m/z 1665.6520) from the unknown
protein (gil23308191) identified in spot 3........................................................ 137









6-8 MS/MS spectrum and corresponding MASCOT search results of the
phosphopeptide pTTGEEEKK (m/z 1000.5043) from the low temperature-
induced protein (gi1509262) identified in spot 8. ..... ................... ............... 138

6-9 MS/MS spectrum and corresponding MASCOT search results of the
phosphopeptide SFGLDSpSQAR (m/z 1146.6510) from putative protein 1
photosystem II oxygen-evolving complex (gil4835233) identified in spot 13. .....139

6-10 MS/MS spectrum and corresponding MASCOT search results of the
phosphopeptide GTGTANQCPpTIDGGSETFSFKPGKYAGK (m/z 2955.3193)
from the 33 kDa polypeptide of oxygen-evolving complex (gi 10177538)
identified in spot 14.............. .. ........ ................... ............ ....140

6-11 MS/MS spectrum and corresponding MASCOT search results of the
phosphopeptide SPASDpTYVIFGEAK (m/z 1563.7827) from the unknown
protein (gi 48310641) identified in spot 15 ............ ................ .......................... 141

6-12 MS/MS spectrum and corresponding MASCOT search results of the
phosphopeptide RSPpSPPPAR (m/z 1043.5207) from the RSZp22 protein
(gi|2582645) identified in spot 19. .............................................. ............... 142















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

PHOSPHOPROTEOMICS OF Arabidopsis thaliana

By

Camille Nicola Strachan

August 2005

Chair: James D. Winefordner
Cochair: Alice C. Harmon
Major Department: Chemistry

Reversible protein phosphorylation on serine, threonine and tyrosine residues is one

of the most common and important regulatory modifications of intracellular proteins,

playing a role in many biological and biomedical phenomena such as cellular signal

transduction, cell growth, cell differentiation, cell division, metabolism and cancer. Mass

spectrometry has emerged as the method of choice for identifying phosphorylation sites

in phosphopeptides because of its advantages over previous methods (high performance

liquid chromatography separation of radiolabelled proteins with 32P or 33P followed by

Edman degradation) including its increased sensitivity and speed, and because it

eliminates the need for protein radiolabelling.

This research focused on the application of mass spectrometry to

phosphoproteomic analyses of Arabidopsis thaliana. We demonstrated the application of

mass spectrometry to four phosphoproteomic projects. Before working on these projects,

we examined the development of methods for phosphorylation enrichment and analysis.









These methods were then applied to the various projects. Project 1, identifying

autophosphorylation sites of a calcium-dependent protein kinase, demonstrated the use of

several complementary methods for identifying numerous autophosphorylation sites of

the protein. Project 2, identifying substrates of a calcium-dependent protein kinase from

Arabidopsis thaliana, demonstrated the application of several newer technologies for

identifying numerous substrates of the kinase. Project 3, identifying 14-3-3 interactors

from Arabidopsis thaliana, examined the identification of numerous protein interactors,

several of which were proven to be phosphorylated. These interactors were then shown to

overlap with the substrates identified for the kinase, possibly indicating interaction

between the two families of proteins. Finally, project 4, the application of robotic

instrumentation was demonstrated as a means for high-throughput phosphoproteomic

analysis, which resulted in identifying several phosphorylated proteins.














CHAPTER 1
INTRODUCTION

Reversible protein phosphorylation on serine, threonine and tyrosine residues is one

of the most common and important regulatory modifications of intracellular proteins,

playing a role in many biological and biomedical phenomena such as cellular signal

transduction, cell growth, cell differentiation, cell division, metabolism, and cancer.1 2

Highlighting its importance is the fact that up to one-third of all proteins in a cell are

phosphorylated at any given time, and as much as 5% of all the genes in a vertebrate

genome code for enzymes involved in phosphorylation kinasess) or dephosphorylation

(phosphatases).2 Due to its importance, research has been initiated in many areas of

biomedical research towards the understanding of the regulatory properties of protein

phosphorylation. Included in these studies are the investigation of the function of protein

phosphorylation in cell cycle regulation, enzyme activation/deactivation, and protein-

protein association.

Calcium-Dependent Protein Kinases

In eukaryotes, protein kinases regulate key aspects of cellular function (such as

metabolism) and responses to external signals by catalyzing the transfer of the terminal

group of ATP to seryl or threonyl residues in a variety of protein substrates.3 Recent

mapping of the Arabidopsis genome provides the first opportunity to identify all the

protein kinases present in a plant model and to begin to understand their physiological

roles. The Arabidopsis genome encodes 1085 typical protein kinases, which is about 4%

of the predicted 25,500 genes.4 These plant kinases differ from animal kinases, since in









plants they phosphorylate only serine and threonine residues; while in animals, tyrosine is

the predominant residue that is phosphorylated. Moreover, a number of kinase families in

plants are either not found in animals or yeast, or are highly divergent. Some of these are

calcium-dependent protein kinases (CDPKs) found in vascular and nonvascular plants,

green algae, and certain protozoa ciliatess and apicomplexans).5 These enzymes are

proposed to be involved in all aspects of plant development and physiology, and

participate in the coupling of cellular responses to environmental and developmental

signals.6

Regulation of the CDPK kinase activity depends on calcium signaling and possibly

autophosphorylation on Ser/Thr residues of the kinase itself; however, the regulatory

effects of autophosphorylation still remain unclear.7-9 Autophosphorylation of a CDPK

from groundnut was suggested to be a prerequisite for its activation,10 while inhibition of

activity was seen after autophosphorylation of a winged bean.1l On the other hand,

preautophosphorylation of a CDPK from sandalwood had no effect on kinase activity.12

Also, conflicting results were seen for a CDPK from ice plant whereby mutation of either

one of the two autophosphorylation sites increased activity, but mutations at both

residues dramatically decreased activity.9 This regulatory process needs further study to

learn if common autophosphorylation sites exist among this family of kinases that will

eventually lead to a greater understanding in the role of this mechanism.

Another aspect of CDPKs that is not understood and requires further investigation

is their recognition of substrate proteins. It is expected that CDPKs have access to

hundreds of potential substrates in the cytosol and nucleus since they are found as both

soluble and membrane-anchored isoforms.13 The expectation is that most isoforms will be









found associated with membranes, such as the plasma membrane,14 peroxisomes,14

endoplasmic reticulum,15 seed oil bodies,16 and mitochondria.17 In addition to their

widespread subcellular distribution, there is evidence that some CDPKs can change

locations in response to a stress treatment.9 This was seen when an isoform McCPK1

from ice plant (I /le,'i///ll,,,//h /////crystallinum) was tagged with a fluorescent protein

and transiently expressed in leaves. The tagged McCPK1 showed a pronounced shift in

localization from the plasma membrane to the nucleus in response to a salt or dehydration

stress, indicating that CDPK targeting is dynamic.

Again, very few substrates are known. Knowledge of the mechanism that these

enzymes use to recognize their diverse substrate proteins is even more limited. Typically,

the sites phosphorylated by a particular protein kinase share a set of common sequence

elements (its consensus sequence) whose existence is necessary and sufficient for

recognition by that enzyme.18 These common sequence elements refer to the sequence

elements immediately surrounding the site(s) phosphorylated by the kinase, generally

taking the form of a short linear sequence of amino acids. According to Kennelly and

Krebs,18 several assumptions are implied in the formulation of a consensus sequence:

1) The existence of a consensus sequence on a protein is essential and adequate for its

recognition as a substrate by a particular protein kinase. 2) The specificity-determining

feature of the phosphorylation site is contained in a neighboring sequence of amino acids

around the phosphoacceptor, not including elements from different polypeptide chains or

from widely scattered portions of a single polypeptide chain. 3) Not all sequence

positions surrounding the phosphoacceptor group carry equal weight in determining the

recognition code.18









Summarizing the complexities of the substrate-recognition process as a set of short

recognition sequences has its usefulness in its simplicity, which has facilitating the

evaluation and application of a large body of observations. However, this sequence can

be an oversimplification that can lead one to think that the primary sequence alone

controls recognition; when in fact, factors such as secondary/tertiary structure or distant

secondary recognition sites play a significant role in substrate recognition.18, 19 The

secondary/tertiary structure of the protein may actually determine substrate specificity by

denying access to potential phosphoacceptor groups. This means that the existence of

intricate secondary/tertiary structures could be an important key to substrate recognition.

That is, the more complex the determinants, the more discriminating the kinase. The

presence of a consensus sequence does not assure that a protein is a substrate of the

kinase, but instead functions as a guide whose implications must be confirmed.

In early studies of CDPK substrate specificity, two simple phosphorylation motifs

were recognized; Basic-3-x-x-[S/T]o and So-x-Basic+2. To date, in-depth analyses have

resulted in the reporting of four consensus sequences for CDPKs with some differences

apparent among the isoforms: 1) (-5-x-Basic-3-x-x-So-x-x-x-(p+4 (minimal) or Basic-6-p-5-

x-Basic-3-x-x-So-x-x-x-(p+4-Basic 6 (optimal), 2) [Basic-9-Basic-8-x-7-Basic-6] -(-5-x-x-x-x-

[S/T]o-x-Basic+2, where the exact ordering of residues within brackets is not specified,

3) (p-l-[ST]o-(p+l-X-Basic+3-Basic+4, and 4) [AL]-5-X-4-R-3-X-2-X-i-So-X+1-R+2-Z+3-R+4,

where (p is a hydrophobic residue, x is any amino acid, and Basic is a basic amino acid

residue (K or R), and Z is any residue but R or K.20-26 Because individual CDPK isoforms

can, in general, recognize all four motifs, it appears that CDPKs may have a series of









overlapping but non-identical polypeptide binding grooves that can accommodate the

different sequences.

Studies to determine these CDPK motifs have largely been performed by using

synthetic peptides. However, although these peptides have represented powerful

investigative tools, their small size and random conformation significantly limit their

abilities to mimic the proteins they are intended to model. Therefore, using proteins to

identify substrates for new phosphorylation sequences could help define primary

structural determinants of protein kinase specificity.

Protein extracts are usually in the denatured form, making it more difficult to

interpret physiological relevance of results. Bylund and Krebs19 showed that

phosphorylation may increase with the unfolding of the protein substrate. Native

lysozyme (which was not a substrate for the cyclic AMP-dependent protein kinase of

rabbit skeletal muscle) became susceptible to phosphorylation by the enzyme once the

protein was denatured by heating.19 Therefore, many proteins may contain sites that can

be phosphorylated once they have been exposed. Careful interpretation is therefore

needed for protein phosphorylation reactions observed in vitro, since denatured proteins

may become protein kinase substrates even though they were not substrates in their

native state. Additionally, the mixing of proteins and kinases from different subcellular

compartments may lead to phosphorylation of proteins that would not occur in vivo.

Despite these complications, the identification of phosphorylation motifs is of

fundamental importance and may be useful for functional genomics and prediction of

phosphoproteins. There are at least 34 CDPK isoforms in Arabidopsis, some of which

have been implicated in drought stress, pathogen response pathways, and the regulation









of metabolic enzymes, transport proteins, and cell structure.7 Consequently, it is of

fundamental importance to understand how this important group of Ser/Thr-kinases

targets their substrate proteins. So even though these results might not signify

physiological occurrences, they are important in terms of fundamental information, as the

identification of new motifs may aid in the understanding of how these kinases target

their substrates.

14-3-3s

Another family of proteins associated with phosphorylation is the 14-3-3 proteins.

These proteins were first identified as abundant brain proteins that were isolated as

soluble, cytosolic, and acidic proteins.27 28 Naming of these proteins was given according

to their particular migration pattern on two-dimensional diethylaminoethyl cellulose

(DEAE-cellulose) chromatography and starch gel electrophoresis. The proteins were then

named by Greek letters according to their respective elution positions on HPLC. Further

studies showed that these proteins were also present in all eukaryotic organisms

examined to date, existing as protein families that contain highly conserved, but

individually distinct isoforms.29-31 Of the organisms characterized, Arabidopsis has the

largest family (10 distinct 14-3-3 proteins of 248 to 268 amino acids: GF14W, GF14X,

GF14(p GF14co, GF14), GF14r, GF14s, GF14K, GF14i, and GF14v).28

Members of the 14-3-3 family are homo- and heterodimers, whose L-shaped

monomers come together to form a broad central groove that contains two binding sites

for target proteins. Phosphorylation of these binding partners (and possibly also the

14-3-3 proteins themselves) may be important in regulating these interactions.32 33 To

date, two different binding motifs have been identified in nearly all known 14-3-3

binding proteins, RSXpSXP and RXY/FXpSXP (where 'X' denotes 'any amino acid









residue', R represents a basic residue, and pS denotes phosphorylated serine); and other

binding motifs have been discovered, including unphosphorylated sites on a few

proteins.32,34,35

Binding of these phosphoserine-containing proteins with 14-3-3 proteins implicated

14-3-3s as proteins that mediate interaction among diverse components of many

biological activities. In mammals, most of the known 14-3-3 binding proteins are

components of intracellular signalling pathways. In contrast, 14-3-3s in plants have

emerged as important regulators of phosphorylated enzymes of biosynthetic metabolism,

ion channels and regulators of plant growth.36 While many proteins have been identified

to bind with 14-3-3s, little is known about how many targets exist in plants. Also, since

there are only two currently known phosphoserine-containing binding motifs among

target proteins, it would be of interest to discover whether there are more 14-3-3 binding

motifs, and to determine if some of these targets are also targets of CDPKs. As

mentioned earlier, the common use of denatured proteins would actually be of benefit for

these studies since crystal structures of a 14-3-3 :phosphopeptide complex showed that

the phosphopeptides bind in an extended conformation, thus resulting in a greater chance

of finding novel binding motifs.32

Detection and Analysis of Protein Phosphorylation

To understand processes regulated by phosphorylation on the molecular level, one

must first determine which proteins are phosphorylated, and second identify the exact

sites of phosphorylation. However, as Mann et al.37 mentions, analysis of

phosphoproteins is not straightforward for several reasons. First, the stoichiometry of

phosphorylation is generally relatively low; that is, only a small fraction of the available

intracellular pool of a protein is phosphorylated at any given time as a result of a given









stimulus. Second, the phosphorylated sites on proteins might vary, implying that any

given phosphoprotein is heterogeneous; that is, it exists in several different

phosphorylated forms. Third, many of the signaling molecules are present at low

abundance within cells and in these cases, enrichment is a prerequisite. Fourth, most

analytical techniques used for studying protein phosphorylation have a limited dynamic

range; which means that, although major phosphorylation sites might be located easily,

minor sites might be difficult to identify. Fifth, phosphatases present in cell lysates could

dephosphorylate residues unless precautions are taken to inhibit their activity during

preparation and purification steps. Finally, proteins can also be phosphorylated by

kinases which are already present during extraction and purification, so this needs to be

prevented.

The analysis of phosphorylation usually proceeds in a certain order, with the

detection of the phosphorylated protein coming first. This may be done by using either

radiolabeling, antibodies, or fluorescent labeling; and analysis is done by using

chromatographic methods [high-performance liquid chromatography (HPLC), thin-layer

chromatography TLC)], electrophoresis, western blotting, autoradiography, scintillation

counting, or mass spectrometry (MS).1 Recently, a method for selectively staining

phosphorylated proteins separated by polyacrylamide gel electrophoresis was developed.

This proprietary fluorescent stain (Pro-Q Diamond Phosphoprotein Gel Stain)

developed by Molecular Probes allows direct, in-gel detection of phosphate groups

attached to serine, threonine, and tyrosine residues, without the need for radiolabelling or

antibodies.









Another way to detect phosphoproteins is to screen for the presence of individual

phosphopeptides. This can be done by first partially hydrolyzing the labeled or unlabelled

protein by partial acid, enzymatic, or alkaline hydrolysis of the amide bonds of the

protein to release the phosphopeptides. The resulting phosphopeptides then must be

separated from other peptides by using either thin-layer chromatography (TLC),

electrophoresis (one or two dimensional), or high-performance liquid chromatography

(HPLC). Retention time, mass spectrometry, antibody recognition, or amino acid

sequencing (Edman, MS/MS) can then be used for identification.

Recently, mass spectrometry-based methods have become increasingly popular for

analyzing phosphopeptides resulting from proteolytic digests of phosphorylated proteins

because of their increased sensitivity and speed, and because they remove the need for

protein radiolabelling.

Mass Spectrometric-based Methods

Mass spectrometry is becoming the method of choice for analyzing complex

protein mixtures. Routine analysis of biomolecules, particularly proteins and peptides,

was made possible by the advent of two mass spectrometry ionization tools: matrix-

assisted laser desorption/ionization (MALDI) and electrospray ionization (ESI). Since

1988 when MALDI and ESI mass spectrometry were first proven useful for analyzing

peptides, proteins, carbohydrates, and oligonucleotides, they have become the MS

methods of choice for biopolymer analysis. Our study used three mass spectrometers: a

matrix-assisted laser desorption/ionization time-of-flight mass spectrometer (MALDI-

TOF-MS), a nanospray quadrupole ion trap mass spectrometer (nanoESI-QIT-MS), and a

tandem quadrupole time-of-flight mass spectrometer with interchangeable MALDI and

nanospray sources (MALDI or ESI-QqTOF-MS).









Ionization techniques

Matrix-assisted laser desorption/ionization (MALDI), a "soft ionization" technique,

was first described in the late 1980s as a technique used with mass spectrometry for

analyzing large, polar, nonvolatile molecules.38 In this technique, a solid organic matrix

compound that is strongly UV-absorbing at the designated wavelength, is dissolved in an

appropriate solvent and mixed with the solution of the sample of interest (the analyte). A

0.5-3 ptL aliquot of this solution is then placed on a stainless-steel target plate and

allowed to dry. On drying, the analyte is co-crystallized with a large (104) molar excess

of the solid matrix material. Once in the mass spectrometer (typically time-of-flight), the

sample is then irradiated with a pulsed laser beam [usually a nitrogen laser in the

ultraviolet range (337 nm)] for desorption and ionization of the analyte molecules.

Even though the mechanism by which MALDI operates is still unclear, it is agreed

that the matrix is critical and fills several roles. First, using the large excess of matrix

helps to isolate analyte molecules from each other, thereby reducing intramolecular

interactions. Second, the matrix absorbs large amounts of energy from the incoming

photons of the pulsed laser beam, resulting in an explosive breakdown of the matrix-

analyte lattice, sending both matrix and analyte molecules into the gas phase. Third, the

matrix is necessary for ionization of the analyte because it transfers protons to the analyte

via gas-phase reactions in the dense cloud that forms.39 The desorption/ionization process

is shown in Figure 1-1.39 Matrix choice depends on the irradiance wavelength and the

type of sample being analyzed. Common matrices that are used with N2 lasers operating

in the UV at 337 nm are 3,5-dimethoxy-4-hydroxy-trans-cinnamic acid (sinapinic acid),









2,5-dihydroxy-benzoic acid (DHB), and a-cyano-4-hydroxy-trans-cinnamic acid (a-

cyano) (Figure 1-2),40 41 with the latter two being best for peptides.

Electrospray ionization, another "soft ionization" technique, was first introduced in

the late 1980s. In this technique, ions are formed from peptides and proteins by spraying

a dilute solution of the analyte (typically dissolved in a mixture of water, an organic

modifier such as acetonitrile, and a few percent by volume of a volatile acid) from a fine

tip at atmospheric pressure. Generally, a high electric field is created by applying a high

voltage to either the spray tip or the counter electrode, resulting in a fine mist of droplets

that are highly charged. This generated electric field (E) between the spray tip and

counter electrode is expressed by the equation:

E = (2V/r) ln(4d/r)

where V is the voltage applied, r is the radius of the needle, and d is the distance between

the spray tip and counter electrode. This imposed electric field will also penetrate to the

liquid flowing through the needle, causing ions in solution to move toward the liquid

surface. Accumulation of these charges at the surface then leads to destabilization of the

surface because the ions at the surface are drawn to the counter electrode yet can't

escape, resulting in the formation of a Taylor cone (Figure 1-3). Due to the cone's

instability (influenced by the surface tension of the fluid y), charged droplets are emitted.

The onset voltage (Von) required to initiate charged-droplet emission is related to surface

tension by the equation:

Von = 2 x 105 (y r).5 n (4d/r)

The radius (R) of the emitted droplets will depend on the fluid density (p), flow rate

(Vf), and surface tension (y), given by the relationship:









R oc (p Vf 2)1/3

Thus, the higher the flow rate (Vf), the larger the initial droplet size which leads to lower

ionization efficiency because the droplets are not so close in size to the Rayleigh limit

(will be discussed later). On the same note, the lower the flow rate (as in nanospray), the

smaller the droplet size, the higher the surface-to-volume ratios leading to a larger

amount of the analyte available for ionization, thus a higher ionization efficiency.

Once these small droplets are formed and are accelerated toward the counter-

electrode, solvent rapidly evaporates and the analytes (peptides or proteins) in the

droplets pick up one, two, or more protons from the solvent to form singly or more

frequently, multiply charged ions (for example, [M+H]+1, [M+2H]+2, [M+3H]+3, etc.). As

the solvent evaporates and the droplet shrinks, the charge density on the surface increases

to the point where Coulombic charge repulsion overcomes the forces holding the droplet

and solvated ions together (Rayleigh Limit), leading to disintegration of the droplet into

smaller droplets (Figure 1-4). This limit where the Coulombic explosions begin is given

by:

Q2= 642 0 y R3

where S0 is the permittivity of vacuum. Once the ions are "emitted" or "evaporated" from

the droplet surface, the ions are then sampled into the high-vacuum region of the mass

spectrometer for mass analysis and detection, most often using a quadrupole mass

analyzer.

Overall, the widespread acceptance of ESI-MS and MALDI-MS are accountable

for several reasons. Both methods are usually combined with relatively low-cost, easily

operated mass spectrometers. They both offer high sensitivity (picomole-femtomolar









range), accuracy (+ 0.01%), and capability of analyzing molecules with a wide molecular

weight range.39 Even though both techniques have many common capabilities, they do

have their own unique capabilities that are of practical significance. Although both

techniques work best with clean (salt- and detergent-free) samples, MALDI is more

tolerant of many of the common biological buffers, that is, information can be obtained

by MALDI directly from a relatively dirty sample. Also, even though both methods can

provide molecular weight information for large proteins, MALDI-MS is more sensitive

and provides the information more easily. However, in the lower molecular weight range,

ESI-MS usually provides more accurate molecular weight measurements as well as better

mass resolution. Additionally, since MALDI produces predominantly singly charged

molecular ions from peptides and proteins, analysis of the resulting MALDI-MS

spectrum is very straightforward. Finally, since ions in electrospray are produced at

atmospheric pressure from flowing liquid streams, ESI is ideally suited for on-line

coupling to high-performance liquid chromatography, making it possible to analyze

mixtures of peptides and proteins. With all of these considerations in mind, it can be said

that MALDI and ESI are methods that can be used to complement each other.

Mass analysis

With the advent of MALDI and ESI, peptide and protein analysis have been a large

focus of efforts in mass spectrometry over the last 17 years, bringing many types of mass

analyzers into use in this area of research, namely time-of-flight, quadrupole ion trap, and

triple quadrupole mass spectrometers, primarily because of their cost and ease of use.

Time-of-flight mass analyzers are among the simplest of the mass analyzers.42 The

principle of time-of-flight mass spectrometry involves measuring the time required for an









ion to travel from the ionization source to the detector.43 For a simple TOF-MS, there are

three components, an ionization source (typically MALDI), a field-free drift region, and a

detector (Figure 1-5). Upon ionization, ions are accelerated out of the source under the

influence of a strong electric field. Even though all the ions receive the same kinetic

energy during acceleration at the ionization source, as they traverse the field-free region

they separate into groups according to velocity because they may have different m/z

values. These ions then sequentially strike the detector in order of increasing m/z value

(lighter ions arriving first), upon which the time-of-flight analyzer converts the time-of-

flight of the ions to a mass-to-charge ratio using the equations:

E = 2 mv2

t = L/v = L [m / 2zeV]1/2

therefore, m/z = 2t2eV/L2

where t = time of flight (seconds), L = length of flight tube (m), v = velocity (m/s),

m = mass (kg), and z = charge.

Simple linear mass spectrometers as described above are somewhat limited due to

their rather low resolution. This low resolution is partly due to the initial kinetic energy

spread of individual ion populations, that is, various members of the same ion population

will arrive at the detector at slightly different times. An effective way to correct for this

energy spread is through the use of a reflectron (ion mirror) which acts as an energy-

focusing device. When an ion reaches the reflectron it is slowed down until it is stopped

by a voltage that is applied to the back end of the reflectron, the ion is turned around, and

then reaccelerated out to a second detector at a slightly different angle so the flight path

of the reflected ions does not cross with the ions entering the reflectron (Figure 1-6). In









the case of an ion population with a spread of slightly different kinetic energies, ions with

a slightly lower energy will not penetrate the reflectron as deeply, therefore turning

around more quickly and catching up to those ions with full kinetic energy. While ions

with slightly greater kinetic energies will penetrate more deeply, turn around more

slowly, and have their flight times retarded, allowing the other ions to catch up. The

addition of this reflectron will then cause ions of a given m/z to be specially focused into

packets with flight times that are closer together. The addition of the reflectron also

increases the flight path for an ion without increasing the size of the flight tube, also

resulting in an improvement in resolution by enhancing the time dispersion of ions of

different m/z. Another improvement that has been made for achieving better resolution on

MALDI-TOF instruments is time-lag focusing or "delayed extraction" in the ionization

source. In this source, ions are created in a field-free region and allowed to spread out

before extraction and acceleration voltages are applied.

Quadrupole ion trap mass spectrometers are mass analyzers that operate by

trapping ions in a three-dimensional electric field consisting of two end-cap electrodes

and a ring electrode, each having a hyperbolic geometry (Figure 1-7).42, 44, 45 In the

normal mode of use, the end-cap electrodes have an auxiliary oscillating potential of low

amplitude applied while the ring electrode has an RF oscillating drive potential of 1

MHz, resulting in the creation of a potential well (quadrupolar trapping field). This field

can be described as having a saddle shape that is constantly spinning whereby the field at

any particular point in time will possess this saddle shape. For an ideal quadrupole field

to be generated, the mathematical relationship presented below has to be fulfilled:

ro2 = 2zo2









where ro is the radius of the ring electrode in the central horizontal plane and zo is the

separation of the two end-cap electrodes measured along the axis of the ion trap.

Typically, once the magnitude of ro is given, the sizes of all the electrodes and electrode

spacings are fixed. It should be noted that the majority of commercial ion traps in use

today have ro at either 1.00 or 0.707 cm.

An ion's stability in this quadrupolar trapping field is dependent upon its m/z, the

potentials applied to the electrodes, and the internal dimensions of the ion trap electrodes.

An ion that is stable in this field will possess a trajectory that has the appearance of a

Lissajous figure, allowing it be trapped within the specific electric field of the ion trap

(Figure 1-8). Unstable ions will have trajectories that increase in magnitude as they near

the ring of the endcaps, resulting in their collision with the electrodes. Determination of

whether the trajectory of an ion will be stable or unstable under defined conditions of the

electric field may be calculated with the Mathieu equations:

az = -2ar = -16eU (1)
mQ2(ro2 + 2zo2)
qz = -2qr = 8eV (2)
mQ2(ro2 + 2zo2)

where az and qz are two reduced parameters, r symbolizes the radial direction, z

symbolizes the axial direction, e is the charge of an ion, U is the DC amplitude applied to

the ring electrode, V is the RF amplitude, m is the mass of an ion, Q is the angular drive

frequency (27frf), ro is the radius, and zo is distance from the center to the end-cap. It

should be noted that an ion has to be stable in both the r and z directions to be confined

within the trap, thus, ar and qr parameters also have to be considered:

ar= 8eU (3)
mQ2(r02 + 2zo2)









qr = -4eV (4)
mQ2(r02 + 2z02)

The resulting stable trajectories that are calculated from the operating parameters ar,z and

qr,z can be displayed graphically as the Mathieu stability diagram (Figure 1-9), whereby

the region of stability is defined by the boundaries at Pz=0, Pz=l, Pr=0, and Pr=l. This

means that if an ion has an az and qz within this region, it will be stable in both the r and z

directions and will be trapped in the ion trap.

Typically, ions generated by electrospray ionization (external source) are focused

into the ion trap using electrostatic lenses. Once in the ion trap, collisions with helium

buffer gas at a pressure of 1 mTorr dampen the kinetic energy of the ions and contract

trajectories toward the center of the trap, where a range of m/z values are held in stable

orbits by the RF potential. As the RF potential on the ring electrode is increased, the ions

become more energetic and develop unstable trajectories along the axis of symmetry (the

z-axis), then in order of increasing m/z value, ions exit the trap through holes in the end-

cap electrodes to a detector. As the RF amplitude is ramped and ions are ejected to the

detector one at a time, a mass spectrum is generated; usually several such spectra

(microscans) are obtained in succession and are then summed prior to display and

recorded as a macroscan. A detailed description of quadrupole ion traps has been

discussed by March if additional information is necessary.44'45

Along with measuring the m/z values of ions introduced to the mass spectrometer,

quadrupole ion trap mass spectrometers can also be used to obtain detailed structural

information from these ions. This information is obtained by performing multiple mass-

selective operations, one after another, that is, tandem mass spectrometry (MS/MS). The

first mass-selective operation is used for the isolation of the ion species of interest









(designated as the parent ion), and the second is used to determine the mass/charge ratios

of the fragment ions (product) formed by collision-induced dissocation (CID) of the

isolated ion of interest. CID refers to the process whereby the kinetic energy of the

selected ion population is increased by applying a voltage resonant with the frequency of

the precursor ion, causing more energetic collisions with the He bath gas. Subjecting the

ions to many hundreds of low-energy collisions will ultimately increase the internal

energy of the ion until fragmentation occurs.

Another mass spectrometer also possessing the ability to perform MS/MS analysis

that was available for this project was a tandem quadrupole time-of-flight mass

spectrometer (QqTOF-MS), where Q refers to a mass-resolving quadrupole and q refers

to an r.f.-only quadrupole or hexapole collision cell. This configuration can be regarded

as either the addition of a mass-resolving quadrupole and collision cell to an ESI-TOF, or

the replacement of the third quadrupole (Q3) in a triple quadrupole by a TOF mass

spectrometer,46 with the latter being the simplest description for the purpose of describing

the basic principles. A thorough review of triple quadrupole instruments has been

described by Yost and Boyd.47

A typical QqTOF configuration consists of three quadrupoles, QO, Q1 and Q2,

followed by a reflecting TOF mass analyzer with orthogonal injection of ions (Figure 1-

10). QO provides collisional damping, Q1 acts as a mass filter, and Q2 is a collision cell.

In the case when ions are provided by a high-pressure electrospray source, QO acts as an

ion guide with collisional cooling and focusing of the ions as they enter the instrument.

When single MS (or TOF-MS) measurements are required, the mass filter Q1 is operated

in the r.f.-only mode serving as only a transmission element while the TOF analyzer









records spectra. For MS/MS, Q1 is operated in the mass filter mode to only transmit the

parent ion of interest which then gets accelerated to an energy of between 20 and 200 eV

before it enters the collision cell Q2, where it is subjected to CID and subsequently

collisionally cooled and focused before analysis by the TOF mass analyzer.

Mass analysis of peptides and proteins

In a generic mass spectrometry-based experiment, the typical proteomics

experiment consists of 5 stages. In stage 1, proteins are isolated from cell lysates or

tissues with gel electrophoresis typically used as a method for biochemical fractionation.

Since MS of whole proteins are less sensitive than MS of peptides and the mass of the

intact protein is usually insufficient for protein identification, proteins from stage 1 are

typically enzymatically digested with trypsin for the formation of C-terminally

protonated amino acid peptides (stage 2). These peptides are then separated by high-

performance liquid chromatography followed by ESI (stage 3) whereby ions are

introduced to the mass spectometer for mass analysis, producing an MS spectrum (stage

4). The computer then generates a prioritized list of these peptides for fragmentation and

a series of MS/MS experiments ensues (stage 4). These MS and MS/MS spectra are then

used for matching against a known protein database for the identification of the proteins

(stage 5).48

Stage 1, the sample extraction and preparation step, may be considered the most

critical step in any proteomics study.49 In this regard, proteomic analysis of plant tissues

is even more problematic than other organisms due to the involvement of several

challenges. These include the fact that plant tissues typically have low amounts of

proteins, they are often rich in proteases and materials that may severely interfere with

downstream protein separation and analysis, including cell wall and storage









polysaccharides, lipids, phenolic compounds and a broad array of secondary

metabolites.49 These contaminants pose a serious problem for one of the most commonly

used separation techniques in proteomics, 2-dimensional gel electrophoresis. Their

presence can result in horizontal and vertical streaking, smearing, and a reduction in the

number of distinctly resolved protein spots. In order to alleviate this problem, several

protein extraction techniques for plant tissues have been compared by Saravanan and

Rose, whereby the quantitative and qualitative characteristics of the protein extracts were

examined.49 From this study, it was demonstrated that the phenol-based method gave the

greatest protein yield and the least contaminants.

Once protein extraction from the plant tissue has been performed, the proteins are

typically separated by 1- or 2-dimensional gel electrophoresis (GE), which are techniques

that separate proteins by the application of an electric field. SDS-PAGE (sodium dodecyl

sulfate-polyacrylamide gel electrophoresis) is a 1D-GE technique that separates proteins

according to their molecular weights. 2D-GE involves the separation of proteins

according to their isoelectric point (the pH at which a protein carries no net electric

charge) in the first dimension, and separation according to MW in the second dimension.

Once the proteins are immobilized in the gel, they are typically visualized by a gel

staining method.50' 51 Some gel stains are visible in visible light, others which are

fluorescent stains that require visualization by UV light. Typical fluorescent imaging

devices are CCD camera-based systems or laser scanner systems.52'53

Once the proteins are visualized and the proteins of interest are determined, the gel

bands/spots are excised for enzymatic digestion.54 Prior to digestion, protein stains have

to be removed to prevent later interference of enzyme activity or mass spectrometric









analysis. Also, the presence of sulfhydryl-containing amino acids in a protein may result

in the formation of disulfides bonds along the protein backbone resulting in the formation

of a three-dimensional shape. As a result, inner portions of the 3D protein structure may

be inaccessible to the enzyme, thus unraveling of this structure is necessary for complete

enzyme cleavage of the protein. This is achieved by the reduction of the sulfhydryl

groups with a reducing agent such as dithiothreitol (DTT), and subsequent alkylation to

prevent reforming of the disulfide bond, resulting in the linearization of the protein. At

this point, an enzyme is added for protein cleavage. Trypsin, a commonly used enzyme,

cuts the protein at the C-terminal end of lysine and arginine residues, resulting in the

formation of C-terminally protonated amino acid peptides for mass spectrometric

analysis.

As mentioned earlier, the routine analysis of proteins and peptides has become

possible due to the introduction of MALDI and ESI techniques. Typically, MALDI-MS is

used for the measurement of peptide masses in an enzymatic digest. ESI-MS is typically

coupled with on-line reversed-phase liquid chromatography systems for the separation of

the peptides prior to introduction to the mass spectrometer, upon which the peptide

masses are measured and tandem MS of the peptides is performed for structural

information.

The tandem MS utility available on two of the mass spectrometers described above,

the QIT-MS and QqTOF-MS, enables immense structural information to be obtained on

proteins and peptides. Advancing this utility even more is the incorporation of decision-

making algorithms that can automatically perform MS/MS experiments on precursor ions

from a previously acquired full scan, enabling the instrument to make real-time decisions









concerning the experiment at hand. The use of this algorithm is called data-dependent

analysis. A typical example of its use would be for peptides eluting from an HPLC

separation. As peptides enter the ion trap mass spectrometer, a full scan is obtained. Once

an ion is detected above a preset threshold, the mass spectrometer will automatically

switch from full scan mode to MS/MS mode on that ion. If there are coeluting peptides,

the mass spectrometer will perform MS/MS on the most intense ion, this ion would then

be placed on an exclusion list, and the second most intense ion from the full scan would

then be subjected to MS/MS. So as each ion gets subjected to MS/MS analysis it is

placed on an exclusion list and will not be removed from this list until after a user-

defined length of time.

As peptides undergo fragmentation by low-energy gas phase collisions during

MS/MS, they undergo cleavage at the amide bonds (-CO-NH-) that join pairs of amino

acid residues, generating a ladder of sequence ions.5 If the charge is retained on the N-

terminal end after cleavage of the amide bond, b-type ions are formed. However, if the

charge is retained on the C-terminal end, y-type ions are formed. The most commonly

observed fragment ions and their nomenclature are shown in Figure 1-11A. The

determination of the amino acid sequence becomes possible if a complete series of either

one or both ion types are present for the subtraction of the masses of adjacent sequence

ions.

If a phosphate group is present, as with phosphorylated peptides, fragmentation

occurs differently, that is, very little cleavage occurs along the peptide backbone amidee

bonds) as would be expected. Instead, p-elimination (removal of the phosphate moiety)

occurs as the primary fragmentation with the phosphate moiety being lost as either









H3PO4, HPO3, H2PO4, HPO42-, PO43-, or P03-, depending on the ionization mode

(positive or negative) being used (Figure 1-11B). These signature losses from

phosphopeptides can be used as an indicator to determine whether a peptide is

phosphorylated. On the QIT-MS, a reconstructed chromatogram can be created by the

software after MS/MS analysis to locate all peptides in a chromatogram that lost a neutral

fragment (H3PO4 or HPO3) (a reconstructed neutral fragment chromatogram). However,

sequencing of the peptide is generally difficult due to insufficient fragmentation. A

similar feature available on the QqTOF-MS is precursor ion scanning, whereby peptides

losing the precursor ion (H2PO4-, HPO42-, PO43-, or P03-) are recorded by the TOF-MS in

the negative ion mode. The drawback of this method is that the experiment has to be

repeated in the positive ion mode so that peptide sequencing may be obtained because

fragments created in the negative ion mode are very difficult to interpret.

Data interpretation

Once mass spectra are obtained, protein identification is performed according to the

method of analysis. For MALDI-MS, peptide mapping, often referred to as peptide-mass

mapping or peptide-mass fingerprinting is performed (Figure 1-12). In this method,

proteins are identified by matching the list of experimental peptide masses to a list of

predicted peptide masses that would occur after digestion with a specific enzyme of all

entries in the protein database.42 48 A match is generally found if a sufficient number of

peptide ions are matched and there are not more than two proteins present. Typically, as

the size of the database being searched is increased, the level of uncertainty also increases

due to the possibility of more proteins being present that could generate peptides with

similar masses.









If tandem MS was performed, identification of the protein is more clear-cut

because not only is the mass of the peptide known, but the peptide sequence information

is also available from the peak pattern of the CID spectrum (Figure 1-13). The CID

spectra are then scanned against a comprehensive protein sequence database using one of

a number of different algorithms, each with its own strengths and weaknesses.48 56-60 Two

of the most commonly used methods are cross-correlation and probability-based

matching.

In the cross-correlation method used in the SEQUEST search engine, theoretical

tandem mass spectra are constructed for all peptide sequences in the protein database

matching the mass of the experimental peptide and the overlap or cross-correlation of the

predicted spectra with the experimental spectra is used to determine the best match.42' 56-58

The quality of the match between the sequence and the spectrum is indicated by the

magnitude of the cross-correlation value and the quality of the match versus all the other

top ranking sequences in the database is shown by the difference between the normalized

cross-correlation score to the second ranked sequence. In the probability-based matching

method used in the MASCOT search engine, calculated fragments of peptides in the

database are compared with observed peaks and a score is calculated that reflects the

statistical significance of the match between the spectrum and the sequences contained in

a database. There are three advantages to this approach:60 1) A simple rule can be used to

judge whether a result is significant or not, guarding against false positives. 2) Scores can

be compared with those from other types of searches. 3) Search parameters can be readily

optimized by iteration. For each of these methods, the identified peptides are compiled

into a protein hit list with corresponding scores or statistics.









The introduction of these search engines have provided a means for handling the

enormous amounts of CID spectra that can be produced from data-dependent scans,

resulting in high-throughput proteomics. However, because protein identifications rely on

the matches with sequence databases, high-throughput proteomics is limited to those

species having a comprehensive sequence database available.

Unfortunately, identification and mapping of phosphorylated peptides by tandem

mass spectrometry is not as straightforward as described above for several reasons.61

First, cleavage of the protein by trypsin can be inhibited due to the negatively charged

modifications of the phosphate group, resulting in incomplete peptide coverage.

Secondly, phosphorylation is often sub-stoichiometric, resulting in the phosphopeptide

being present in much lower abundance than the other unphosphorylated peptides. This

will result in suppression of the phosphopeptide relative to its unphosphorylated

counterpart during the mass spectrometric analysis. Suppression of the phosphopeptide is

even more pronounced in the presence of many other unphosphorylated peptides such as

that found in the protein digest. Reducing the number of unphosphorylated peptides

present in the sample by enriching for the phosphopeptides will therefore enhance mass

spectrometric mapping of the phosphorylation site. Finally, performing tandem MS on

the phosphopeptide to determine the location of the phosphate group can be difficult due

to the instability of the phosphate moiety whereby the phosphate moiety may be

eliminated before the peptide can even undergo fragmentation, making it difficult to

locate the phosphorylation site. Also, since fragmentation of the phosphopeptide may

result in only the elimination of the phosphate moiety and very little fragmentation along

the backbone, database searching will not identify the phosphopeptide.









Isolation and enrichment techniques

Isolation and enrichment techniques have been developed to enhance mass

spectrometric analysis of phosphopeptides even further, eliminating the need for

radiolabeling, antibodies or fluorescent labeling. The most common enrichment methods

are immobilized metal-ion affinity chromatography (IMAC),62-67 and chemical

modification methods.6873 Once isolated, the resulting phosphopeptides can then be

identified by mass spectrometry, however, this can be very difficult as will be discussed

later.

Immobilized metal-ion affinity chromatography

Immobilized metal-ion affinity chromatography, originally recognized by

Andersson and Porath, is a method used to selectively isolate and enrich phosphopeptides

from a peptide mixture via the interaction of the phosphate group on the peptide with the

free coordination sites of metal ions immobilized (chelated) to a stationary phase74

(Figure 1-14). Two of the more frequently used chelating ligands for IMAC are imino-

diacetic acid (IDA) and nitrilo-triacetic acid (NTA), however, the majority of the

published applications of IMAC for phosphoprotein and phosphopeptide characterization

have used IDA. Townsend and coworkers reported that NTA sepharose was superior to

IDA-sepharose for phosphopeptide purification by Fe(III)-IMAC because of its higher

selectivity for phosphopeptides.75 However, studies performed by other groups have not

confirmed this.67

Since the first report utilizing Fe(III) by Andersson and Porath, the use of several

other metal ions have been reported, including Al(III), Sc(III), Lu(III), Th(III),74'76 and

Ga(III).66 Of these metals, Fe(III) and Ga(III) seem to give best results for

phosphopeptide isolation, with Ga(III) showing less overall suppression effect and the









ability to isolate multiply phosphorylated peptides, while Fe(III) shows better selectivity

for monophosphorylated peptides.66

Although the IMAC methodology is for the isolation and enrichment of

phosphopeptides, nonspecific binding could also occur for peptides possessing multiple

carboxylic acid groups.62 These peptides could suppress the signal from trace-level

phosphopeptides in the mixture. One way to prevent this problem is to convert the

peptides in the mixture to the corresponding peptide methyl esters by replacing the

carboxylic acid groups. This is especially needed when acidic amino acids (aspartate,

glutamate, and S-carboxymethylated cysteine) are present. By performing this

methylation step prior to the IMAC isolation, it is expected that nonspecific binding

through carboxyl groups will be prevented, resulting in the selective isolation of

phosphopeptides only. The methylation step, however, may be a problem if there is

moisture present due to the sensitivity of methanolic-HCl to water. This means that

measures have to be taken to keep the reaction mixtures extremely dry, which can make

the sample preparation prior to mass spectrometry more labor intensive taking

approximately three and a half hours compared to an hour without the methylation step.

Another area of concern for the IMAC method is that the phosphate moiety may be

lost during ionization or fragmentation due to the liability of the phosphate group. This

means that once the phosphate group is removed, localization of the phosphorylation site

will be difficult. Also, little sequence information may be obtained because most of the

fragmentation energy is used to remove the phosphate moiety. One way of alleviating this

problem would be to replace the phosphate group with a more stable label.









Chemical derivatization

The problem of the liability of the phosphate group has been addressed by several

proposed chemical modification methods whereby H3P04 is removed by 3-elimination at

high pH, a linker containing a thiol group is added by Michael addition, followed by the

addition of a biotin-containing compound which acts as an affinity tag for purification of

the phosphopeptides as well as a tag for phosphorylation site mapping67-73 77 (Figure 1-

15). Several linkers have been used for modification, including ethanedithiol,69' 71, 72, 77

and various lengths of alkanethiols.73 Of these linkers, ethanedithiol seems to be the most

popular due to the presence of a thiol group on both ends of the molecule, thus increasing

the chances for it to be bound. Several types of biotinylated chemicals have been used,

including iodoacetyl-PEO-biotin,72 77 biotin-HPDP,71 and (+)-biotinyl,3-

maleimidopropionamidyl-3,4-dioxoctanediamine.69 The disadvantages of this technique

are that it is labor-intensive, time-consuming (several hours), and the many sample

handling steps involved could lead to significant sample loss. These problems can be

eliminated if the label is synthesized and stored prior to sample preparation, however, not

very many publications report doing this.77

Qiagen phosphoprotein purification kit

Within the past 2 years, the development of an enrichment method specifically for

phosphorylated proteins from complex cell lysates was introduced by Qiagen

Incorporated. Purification of the phosphoproteins is performed by a proprietary affinity

chromatography method. The little information that is known about this column is that

the phosphate groups on phosphoproteins are bound to the column with high specificity,

while proteins without the phosphate groups will not bind to the column and therefore be

found in the column flow-through. The bound phosphoproteins are then washed with a






29


phosphate-buffered saline buffer (PBS) and stored. The free phosphate in the elution

buffer serves two functions: 1) displaces the phosphoproteins from the column, and 2)

inhibits the phosphate activity in the cell lysate, therefore stabilizing the phosphorylation

status of the proteins during downstream processing and storage.











Ptlsed laser
beam


\ Desorbed 'plume' of
maix and alyte
ions






' I


Sample plate 7


a S

00


9 w
0
B'

B
B


I


B



D* 0
mm


B
D I
1 0




pI

*


0

B


Analte

Matrix


Caion (eg. Na+ ox


a

em *

. . . .. . . . .
.......................m e ............
. . . . .


... . . . . . . . . . . . . . . . .' . . . .


Figure 1-1. The MALDI process












CH= C-H- CGOH


Sinapinic acid


I= C
CN


HO'
OH
a-cyano4-hydroxycbinarric acid 2,5-dihydroxybenzoic acid


Figure 1-2. Commonly used MALDI matrices for peptide and protein analysis














Taylor cone


Figure 1-3. Taylor cone formation










Desolvafion






tit Ai
I .


t+fe&


Mass SpDc-mdr


4^


uemitthe Curiombic
(GrIunnd) Epbsin

Figure 1-4. The electrospray ionization process


4k


Desolvad
Ihs


Figure 1-5. A linear time-of-flight mass spectrometer

Figure 1-5. A linear time-of-flight mass spectrometer

























Figure 1-6. A reflectron time-of-flight mass spectrometer


End-cap elechmue


Figure 1-7. Three-dimensional ideal ion trap showing the asymptotes and the dimensions
ro and zo. Modified from March, R.E. International Journal of Mass
Spectrometry. 2000, 200, 285-312. Figure 1, page 287.








Lissajous crve shape


5


3



N o


0
Y-AXJS fmm


Figure 1-8. Trajectory of a trapped ion. Modified from March, R.E. Journal of Mass
Spectrometry. 1997, 32, 351-369. Figure 8, page 356.















0-6,


1.0 O
X0.1


021













\0 ,4 0...
06










L2 04 06 US 1O 12 1_4 1B

Figure 1-9. Mathieu stability diagram in (az, qz) space for the region of simultaneous
stability in both the r- and z-directions near the origin for the three-
dimensional quadrupole ion trap. The q,-axis intersects the fl=l boundary at
q,=0.908, which corresponds to qmax in the mass-selective instability mode.
Modified from March, R.E. Journal of Mass Spectrometry. 1997, 32, 351-369.
Figure 7, page 356.











QI 01


10 mTmn


I*: i : :* *: *:


IL


MCP
Dekxb


i


Figure 1-10. Tandem QqTOF mass spectrometer. Adapted from Chemushevich, I.V.;
Loboda, A.V.; Thomson, B.A. Journal of Mass Spectrometry. 2001, 36, 849-
865. Figure 1, page 860.










R, / R2 4 R3

H N CH CO NH CH- CON-CH- COOH


b, ,


0

HO- P- OH
I
R, CH- R!,

I I I
H2N CH ~- CO- NH CH- CO- NH -CH- COOH
phosphaosedne

I


R, CH R,
H I IC
H N CH CO- NH C -CO -NH ~CH COOH


0
II
+ mD- P- OH


Figure 1-11. Peptide fragmentation. A) Typical low-energy CID fragmentation of a
peptide forming mainly b and y-ions. B) Fragmentation of a phosphopeptide
resulting in mainly 3-elimination of the phosphate moiety as phosphoric acid.










M)KS LVKMU%--AECA-DDMAAKAVEG-E I LSN -aI uSVAYFFNYGWHR.--
~ aaana aaaanara~ananaaaaawwwwUwws aananaana~


Pian keaimAm by Best Fi~ Dabase Seqprns

I3ISELVQKcAKAEGAERB DDMAAlKAVrEoG .SUF M EE .LSIVAVYFNYGWFRR.

Figure 1-12. Peptide mass mapping. A protein sequence can be verified by site-specific
digestion and measurement of the peptide ions for correlation with those
predicted by the sequence. Conversely, if the identity of the protein is not
known the peptide mass map can be used to search the protein database to find
the sequence that best fits the mass map. Adapted from Yates, J.R. Journal of
Mass Spectrometry. 1998, 33, 1-19. Figure 3, page 8.














S272 729 1295

171 |52 10152 1424

I I I I I I I I I I I I I I I I


2W0 400 O 12 o W 14W0























XmKSEVQAKLAEQBYFVDDMAAMKAVTOG.EQ S FMA I .SWVAYFVNYGWHR


SewpuMM ad PRna kMel d



Figure 1-13. Tandem mass spectrometric sequencing. The ladder of fragment ions
represents the amino acid sequence of the peptide. By subtracting the m/z
values for adjacent ions of the same type the sequence can be elucidated.
Conversely, the fragmentation pattern can be used to search the protein or
nucleotide database to find the amino acid sequence that best fits the tandem
mass spectrum. Adapted from Yates, J.R. Journal ofMass Spectrometry.
1998, 33, 1-19. Figure 4, page 10.









0




/ "a0
II
\ HW
ReSii N



II
0


N-----
/


Resin 0
HNc7


SHP


NIkel-DA (hin-DiAcetic acid) resin NhlN-HIA p1ao-TIrAe c add) Imse i

Figure 1-14. Two of the most frequently used chelating ligands for IMAC: imino-
diacetic acid and nitrilo-triacetic acid.



















0


or
o/d


SCL CLS


POO4


9tep 1: betaeliminstion


s





H 1 C H


0
NH NH
N -O N --
.O..r0-... N S
Smp3: 6wtlny1 niLon


7


Lis itherMor D


H





LCC
Sup 2: MichtnI Mion








>"

H ,CHX H

H NM


Figure 1-15. Chemical derivatization method for the affinity purification of a
phosphorylated peptide. Adapted from Goshe, M.; Conrads, T.; Panisko, E.;
Angell, N.; Veenstra, T.; Smith, R. Analytical Chemistry. 2001, 73, 2578-
2586. Figure 1, page 2581.


OH-


H


ptosp hot rvl x-
pItGphotf1nmly x-CH














CHAPTER 2
PHOSPHORYLATION DETECTION AND ENRICHMENT

Reversible phosphorylation on serine, threonine and tyrosine residues on a protein

is one of the most common and important post-translational modifications involved in

virtually all cellular processes. Key to the molecular understanding of these signals, is the

identification of kinases, their substrates, and the specific site of phosphorylation.

However, the process of studying phosphorylation can be a very difficult and tedious task

as mentioned in Chapter 1. A major contributor to the difficulty of this process is

suppression by unphosphorylated peptides in a complex mixture, making phosphopeptide

isolation and enrichment a necessary step.

For the development of a method for phosphoprotein analysis of a complex mixture

(for example Arabidopsis thaliana protein extract), one of the first requirements is the

determination of a means of detecting phosphoproteins along with the limits of detection

of the chosen method. Second, since isolation and enrichment of the phosphorylated

protein is necessary if using a MALDI-TOF-MS or QIT-MS, examination and

comparisons of available enrichment methods would be advantageous for determining the

most appropriate method.

MALDI-TOF-MS can provide a means for rapid sample preparation and analysis

for method development, however, optimal conditions for MALDI has to be determined

prior to any other method development strategies. As mentioned earlier, selection of

matrix and solvent conditions is critical. This is even more important for the analysis of

phosphorylated peptides because of their low response to mass spectrometry in positive









ion mode due to the presence of the negatively charged phosphate group. This negative

charge interference in detection is even more pronounced when multiple phosphate

groups are present in the peptide. Asara and Allison determined that one way to alleviate

this problem is to reduce the negative charge interference by the addition of a positively

charged species to the matrix spot.78 Typically, this may be done by adding a solution of

ammonium citrate to the already co-crystallized analyte and matrix spot. Hence,

optimization of matrix conditions prior to this step is necessary.

Once matrix conditions for phosphopeptide analysis have been optimized,

comparison of enrichment techniques may be performed. As mentioned in Chapter 1, two

published methods for phosphopeptide isolation and enrichment are IMAC and chemical

derivatization methods. During the initial stages of this project, there were two

commercially available IMAC columns on the market, the Pierce Phosphopeptide

Isolation Kit and Millipore's ZipTipMC, each having a potential advantage over the

other. The isolation kit from Pierce is a small minispin column with Ga(III) chelated to an

IDA resin, thus reducing the workload for the user since the metal ions are already

chelated and the minispin column format makes it possible to enrich several samples at

once. On the other hand, Millipore's ZipTipMC comes as a ziptip with the IDA resin

only, that is, the user binds his or her metal of choice during the preparation, thus giving

the user the opportunity to test various metal ligands for optimization. At the time,

several methods had been reported for isolation and enrichment of phosphopeptides by

chemical derivatization methods, with Goshe' s77 publication being the most recent.

Presented are the method development steps performed for determining the best

isolation and enrichment techniques for the purposes of this project. Beginning the









method development was the optimization of matrix conditions for MALDI, followed by

optimization of conditions for both available IMAC products as well as the chemical

derivatization method published by Goshe.77 Once optimized, both enrichment

techniques (IMAC and chemical modification) were compared for determination of the

best method for the goal of this project. It should be noted that at the later stages of this

project, a proprietary Phosphoprotein Purification Kit was developed by Qiagen Inc.

which will not be discussed until Chapter 4.

Experimental Methods

Materials and Instruments

The Pro-Q Diamond Phosphoprotein Gel Stain (Pro-Q) and SYPRO Ruby Gel

Stain (SYPRO) were obtained from Molecular Probes, Inc (Eugene, OR). 10% Bis-Tris

Novex NuPAGE polyacrylamide gels were from Invitrogen (Carlsbad, CA). Protein

standards (p-casein, ovalbumin, bovine serum albumin, and carbonic anhydrase) as well

as the evaluated MALDI matrices ( a-cyano-4-hydroxycinnamic acid and

dihydroxybenzoic acid) were purchased from Sigma Aldrich (St. Louis, MO).

Sequencing grade trifluoroacetic acid was obtained from Applied Biosystems (Foster

City, CA). The Phosphopeptide Purification Kit (IMAC mini-spin columns), (+)-biotinyl-

iodoacetamidyl-3,6-dioxaoctanediamine (iodoacetyl-PEObiotin) and tris(2-carboxyethyl)

phosphine hydrochloride (TCEP-HC1) were from Pierce (Rockford, IL). Reversed-phase

C18 ZipTips and ZipTipMC were obtained from Millipore (Billerica, MA). Sequencing-

grade modified trypsin was purchased from Promega (Madison, WI). Acetonitirile,

ethanol, methanol, formic acid, glacial acetic acid, barium hydroxide, sodium hydroxide,

ammonium bicarbonate, and Coomassie Brilliant Blue R250 (Coomassie) were obtained









from Fisher Scientific (Fairlawn, NJ). HPLC grade water was purchased from Burdick &

Jackson (Pleasant Prairie, WI).

Fluorescent imaging of gels was acquired with a Typhoon Scanner (GE Healthcare,

Piscataway, NJ). Mass spectrometric measurements were made using a Voyager-DE Pro

Biospectrometry Workstation (Applied Biosystems, Foster City, CA).

Phosphoprotein Detection

1.86 tg and 0.93 tg p-casein, ovalbumin, bovine serum albumin, and carbonic

anhydrase were separated by 1-dimensional gel electrophoresis using a 10% Bis-Tris

Novex NuPAGE polyacrylamide gel with MOPS running buffer at 200V for 1 hr. The gel

was stained with the Pro-Q Diamond and imaged according to the manufacturer's

protocol. The gel was then counter-stained with SYPRO Ruby, imaged with the

Typhoon, and counter-stained again with Coomassie. A serial dilution (250, 100, 50, 20,

10 and 1 ng) of a mixture of equal amounts of p-casein, ovalbumin, and bovine serum

albumin was then separated on a gel and stained as described above for detection limit

determination.

Matrix Optimization

Saturated solutions of the MALDI matrices were prepared in the selected solvent

systems in Table 2-1 and mixed with a synthetic peptide (PLARTLpSVAGLPGK).

Unless otherwise noted, all MALDI analyses were performed using the dried-droplet

method by mixing a 1:1 ratio of analyte to matrix and allowing the spot to dry on the

MALDI plate. Two spots of each matrix-analyte solution were spotted and once dried, 25

mM ammonium citrate was spotted on top of one of the two spots and allowed to dry.









Table 2-1. Matrix solvent systems for evaluation.
Matrix Solvent combinations Additive after spotting
a-cyano-4- 50% ACN, 0.3% TFA
hydroxycinnamic acid
50% ACN, 0.3% TFA 25 mM ammonium citrate
50% ACN, 3% FA
50% ACN, 3% FA 25 mM ammonium citrate
2,5-dihydroxybenzoic acid 50% ACN, 0.3% TFA
50% ACN, 0.3% TFA 25 mM ammonium citrate
50% ACN, 3% FA
50% ACN, 3% FA 25 mM ammonium citrate
ACN acetonitrile, TFA trifluoroacetic acid, FA formic acid

Mass spectra were obtained with a MALDI-TOF MS instrument equipped with a

337 nm nitrogen laser and reflectron optics. All spectra were acquired in positive

ionization mode under delayed extraction conditions in reflectron mode. Spectra were

obtained with an acceleration voltage of 20 kV and 100 laser shots at a laser repetition

rate of 3.0 Hz. Laser intensity, extraction delay time, grid voltage and guide wire were all

adjusted to obtain the best spectrum for each sample. An external calibration was

performed before each spectrum was obtained with a calibration mixture consisting of 1.0

pmol Des-Argl-Bradykinin, 1.3 pmol Angiotensin I, 1.3 pmol Glul-fibrinopeptide B, 2.0

pmol ACTH (1-17 clip), and 1.5 pmol ACTH (18-39 clip).

Phosphopeptide Enrichment

A stock solution of p-casein digest was prepared by incubating 1 mg of protein

with 10 pg trypsin in 100 mM ammonium bicarbonate at 37 OC overnight.

Immobilized metal-ion affinity chromatography

Both the Phosphopeptide Purification Kit and ZipTipMC were used according to

the manufacturer's protocol. Briefly, this involves binding the sample at low pH (0.1%

acetic acid), followed by wash steps with low organic (0.1% acetic acid with 10%

acetonitrile, and water) for the removal of unphosphorylated peptides, and then elution of









the phosphopeptides with higher pH (0.3N ammonium hydroxide solution). For the

ZipTipMC, the metal ion of choice is bound to the resin. Fe3+, Ga3+, Cu2+, and Ni3+ were

used as the metal ions of choice. Isolated peptides were analyzed by MALDI analysis as

described above.

Chemical derivatization

Chemical derivatization of the synthetic phosphopeptide was performed according

to the protocol published by Goshe.77 Reaction conditions such as time and temperature

were investigated for optimal yield.

Results and Discussion

Phosphoprotein Detection

Analysis of the four protein standards (including the phosphorylated proteins, P-

casein and ovalbumin) with the Pro-Q stain showed selective staining of the

phosphorylated proteins (Figure 2-1). Counter-staining with both SYPRO and Coomassie

proved that all four proteins were indeed on the gel and the Pro-Q stain was selective for

only the phosphorylated proteins. Counter-staining also indicated that the Pro-Q stain

could be as sensitive as the SYPRO stain (as low as 1 ng)53 compared to the Coomassie

stain (as low as 100 ng),53 according to the intensity of the gel bands. Subsequent

detection limit experiments determined that the Pro-Q stain could detect phosphoproteins

at levels as low as 20 ng (Figure 2-2). However, it was noticed that large amounts

(several Gg) of unphosphorylated proteins were detected by the stain but appeared more

as smears rather than very distinct bands. This problem could be alleviated by always

including an unphosphorylated protein as a standard among the samples and subtracting

the corresponding signal for removal of background staining. However, if there is a









greater amount of an unphosphorylated protein present in the sample of interest than the

standard, background subtraction may not completely correct for this problem.

Matrix Optimization

Initial studies utilized several matrix compounds, with a-cyano giving the best

results for peptide analysis. This matrix (a-cyano matrix containing 50%

ACN/0.3%TFA) was then optimized for phosphopeptide enhancement by the addition of

ammonium citrate. Figure 2-3 compares spectra of a phosphopeptide of 1459 Da obtained

with and without the addition of ammonium citrate. The increase in intensity of the

[M+H] upon addition of ammonium citrate shows that ammonium citrate does enhance

the phosphopeptide signal and can be used with the matrix for optimal conditions for

phosphopeptides.

Phosphopeptide Enrichment

Figure 2-4 shows a MALDI spectrum of a P-casein digest. Expected

unphosphorylated peptides can be seen, however, the phosphopeptides at 2062 (1

phosphorylation site) and 3121 (4 phosphorylation sites) m/z are not visible, proving that

isolation and enrichment of phosphopeptides is necessary. Isolation and enrichment of the

P-casein phosphopeptides was then performed with the Millipore ZipTipMC. A variety of

metal ions were tested to identify which would give the best isolation and enrichment of

the phosphopeptides, as well as the least nonspecific binding. Figure 2-5 shows spectra of

the isolated phosphopeptides obtained using Fe3+, Ga3+, Cu2+, and Ni3+ proving that the

ZipTipMC does isolate and enrich for phosphopeptides. Of the four metal ions used, Ga3

and Fe3+ seem to give the best spectra with respect to good signal being obtained for both

the monophosphorylated and multiply phosphorylated peptides. The monophosphorylated

peptide could not be found in the spectrum obtained using Ni3+, however, the signal









intensity of the multiply phosphorylated peptide was higher than those obtained for the

other spectra.

The Pierce Phosphopeptide Isolation Kit was then used to isolate and enrich the P-

casein phosphopeptides, giving similar results as the ZipTipMC with Ga3+ (results not

shown). These results were expected since the Pierce Kit also uses Ga3+. The sensitivity

of the column was then tested by isolating and enriching for phosphopeptides from 1

picomole of 0-casein digest (Figure 2-6). At first glance, it seemed as if only the

monophosphorylated peptide was isolated (Figure 2-6A), however, after the addition of

ammonium citrate to the spot, significant enhancement of the multiply phosphorylated

peptide (3121 m/z) was seen, however, the monophosphorylated peptide could not be

seen possibly due to ionization suppression by the multiply phosphorylated peptide. This

shows that this IMAC column does isolate and enrich phosphopeptides at very low

amounts (1 pmol) and also that ammonium citrate can be added to the MALDI spot for

the enhancement of the multiply phosphorylated peptides.

Both IMAC methods were compared to decide which would be best for this

project. In regards to levels of sensitivity for phosphopeptides, both gave similar results

(femtomole range), leaving the decision up to ease of use and cost. Of the two IMAC

products mentioned, the Pierce product was the least labor intensive, however, it was

more costly. Due to the amount of samples that would possibly have to be enriched, the

Pierce Phosphopeptide Isolation Kit was chosen as the better method of the two.

For a comparison of the chemical derivatization and IMAC methods, the protocol

published by Goshe77 was tested for its utility as an enrichment technique. A synthesized

phosphopeptide PLARTLpSVAGLPGK was used for optimization of each reaction step









(p-elimination, Michael addition, and biotinylation). For P-elimination, optimal reaction

conditions of temperature and time were examined. Once established, parameters for the

addition of the linker were examined, that is, type of linker used (ethaneditiol (EDT) or 2-

mercaptoethylamine), addition of a reducing agent (TCEP-HC1) to prevent dimerization

of the linker molecule, length of time of the reaction, and also whether the reaction could

occur simultaneously with p-elimination. At this point, removal of the EDT prior to the

addition of biotin was critical for the prevention of competition for the biotin between the

EDT or 2-mercaptoethylamine-linked peptides and the free EDT or 2-

mercaptoethylamine molecules. If complete removal of the linkers is not achieved, the

reaction yield will be minimal, which is not acceptable in real samples where

phosphorylation can be at very low levels. Several methods such as extraction with

diethyl ether or size-exclusion chromatography were utilized for the efficient removal of

the linker molecules. Following the removal step, the addition of biotin was performed

and optimized.

Figure 2-7 shows the chemical derivatization of a synthetic phosphopeptide. A

spectrum of the phosphopeptide was first obtained (Figure 2-7A), followed by P-

elimination at determined optimal conditions (55 OC for 1 hour in 0.5 M sodium

hydroxide (Figure 2-7B). Figure 2-7C shows the addition of the EDT linker at the

determined optimal conditions, that is, the P-elimination and EDT addition steps are

performed concurrently (55 OC for 1 hour in 0.5 M sodium hydroxide, 3.6 mM TCEP-

HC1 and a 5 molar excess of EDT), followed by removal of unreacted EDT with a

Sephadex G-10 spin column. It should be mentioned that the addition of 2-

mercaptoethylamine as a linker was also attempted but was unsuccessful. Biotinylation









was then performed for 90 minutes in the dark with constant stirring and the reaction

mixture desalted with a Sephadex G-10 spin column (Figure 2-7D). This experiment

proves that the chemical derivatization steps are possible, however, the isolation of the

biotinylated peptide with an avidin column was not done due to the low product yield

which would not be acceptable for the purposes of this project because of low amounts of

starting material that would be available.

This procedure proved to be very labor-intensive and time-consuming so an

alternative method was attempted. If a label could be synthesized and stored, sample

handling and derivatization time would be reduced. With this in mind, the above label

(EDT and biotin) was first synthesized, purified and then added to the phosphopeptide,

unfortunately, no reaction seemed to have occurred when monitored by MALDI-TOF-

MS.

The synthesis of a novel phosphorylation label was also attempted by the reaction

of H-Cys(Trt)-NH2 and biotin-LC-OSu. The expected product (702 m/z) was not found in

the MALDI spectrum of the reaction mixture, however, the sodiated and potassiated

forms could be seen at 724 and 740 m/z. The reaction mixture was then purified on an

HPLC system and the fraction of interest collected and a MALDI spectrum obtained as

shown in Figure 2-8A. In order to validate that this was indeed the product of interest,

several other matrices were investigated to see if the expected product could be seen

without the adducts. Figure 2-8B shows a spectrum of the expected product with 3-

hydroxypicolinic acid matrix, proving that the expected product was formed. Removal of

the tert-butyl protecting group from the product was then performed and the expected

loss of 57 m/z is seen in Figure 2-8C, resulting in the deprotected label at 645 m/z, as









well as the product with a sodium adduct as well as a potassium adduct. Even though the

experiments were successful in synthesizing a possible phosphorylation label, the

reaction yield was very low and reaction with the phosphopeptide did not seem to occur

after several attempts.

Of the above methods attempted for isolation and enrichment of phosphopeptides,

the IMAC columns seem to be more favorable when time and labor are considered.

Samples can be prepared for analysis by IMAC in less than an hour, compared to 1 day

by the chemical derivatization method, that is, providing that the isolation with the avidin

column works effectively. Also, the chemical derivatization method requires a larger

quantity of sample and there is a much higher potential of sample loss due to the many

steps involved. With all of this in mind, the IMAC columns seemed to be the best method

for isolation and enrichment of phosphopeptides.











A 123456789



1 6

tea


B 12345678 9





'U e-


r


lo-


1. Pcasi I Ou mlmmu I BA I CA (1-8 pg eadi)
2 pcasin (16 pig)
3a pasan (IU3 ig)
4. Ovabfmn (1.Ms]
5. Ovabmun
6B SA (186 pg)
7. B S(0.3 pg)
8. CA (16 jg)
9 CA (0-93 g)


Figure 2-1. One-Dimensional gel electrophoresis of four protein standards. A) Selective
phosphoprotein detection with the Pro-Q Diamond Phosphoprotein Gel Stain.
B) Total protein staining with SYPRO Ruby protein stain. C) Total protein
staining with Coomassie R-250 protein stain.


%4 S


C 12345678 9











1 2 3 4 5 6 7 8 9







H *PO

1$ -- --ca


-casem Co rng)
Ovatnfm (500 rig)
BS A (00o g)
-casem i OvaMnSm I SA (250 ng eah)
-caseM I/ Ounmi I/ OSA (100 ng eah)
-casen I Ovuaftrin I BSA (5 ng each)
p-ceasemn Ova~ I BSA (20 ng enad)
p-case I OwaRiuln I BSA (10 ng each)
p-caseM I Ovaftru I BSA (1 ng each)


Figure 2-2. Pro-Q stained gel of varying concentrations of protein standards. The lowest
detection level of the two phosphoprotein standards appears to be 20 ng.

































[II. IJ


1459L


J


14~0.1


J


122.4 145U.6
Mass (mu)


5.3E+4












5.3E+4 *


1740.8 2001.0


Figure 2-3. Comparison of matrix conditions for phosphopeptide spectra. A) Spectrum
obtained with a-cyano matrix. B) Spectrum obtained with a-cyano matrix and
the addition of ammonium citrate. Enhancement of the phosphopeptide at m/z
1460 is observed.


9 100









0


I 1i


FlAu.u bu..


I ~" ------- ---- --- ""~,~,~CC~b




















40 7.4


830_4


i .


dl


29112

/


3722/7

(


m lL ,L .J A.- -..,A.LihF t-- .. -. Lu


50-0


1400.2


2300.4


3200.6


41008


501S 0


Mass nfr
Figure 2-4. P-casein digest spectrum. Unphosphorylated peptides are observed, however,
the two known phosphopeptides at m/z 2062 and 3123 are not evident.


I
















2062-6



A I I


206110


Fe3


3120-8
/


--- -- -


ri-

it
r _I


3120_1

/


IkLj .


15594


2419-8 32802
Mass n.M


69m0


4140.6


n


,2-4E-4


I


50(


1-0


Figure 2-5. Comparison of different metal ions on the ZipTipMC for isolation of 3-
casein phosphopeptides. A) Fe3 B) Ga3+. C) Cu2+. D) Ni3+


1 C C"1* 3120-1 7!



r2161_1

I.0


r
ir
2 L m


31193


- LL If


m ..


Me ... -


M


6a-


.yJ









A







1 & .L i,1.,


2161_6
2-61
/






k.L .a.h ,. iL I ikIL- ..Ih, -t s.


- .. MI I 1 0 --An


-ii


155094


312110
/


24198 3280-2
Mass mna)


4140.6


Figure 2-6. Spectra of p-casein phosphopeptides isolated by the Pierce Phosphopeptide
Isolation Kit. A) Isolated phosphopeptides spotted with a-cyano matrix. B)
Isolated phosphopeptides spotted with a-cyano matrix and the addition of
ammonium citrate.


3236.7


100, B


1_7E-4


5001-0


.......... uY


^^*cNm.,h 11 II!I 1 M^ll* i i iM^^~


0 -
60-09


1 I ,. .











K+HT



L1


L. I


1280.G


1036-6 12772


1517-4


1754.2


15178 1758.4


[+HR -
I PO, \
II. I


i+H'- IPO,, + EDT
x


Ph ___________________________ ii


13188


15s52


1811.6


D D


1112


[M+Hi-11- PO4 + EDT

L, 1 \


1410


I[+Hr- INPO,+ EDT+
,it


1708


Figure 2-7. Chemical derivatization of phosphopeptide. A) Underivatized
phosphorylated peptide. B) p-elimination of peptide. C) Michael addition of
EDT to the 3-eliminated peptide. D) Biotinylation of peptide.


1L- W .. ..-- L .V ..,-I. -- -. I


07.0


104"38


100I C


10724


3-3E+4


~h-~-------- -- -- n


2354


I L


o100 A


5_8E+4





in


3E+4


1. I

















wI.. .. .... l


iwar
....


Is"


SSA


A SE-4



is


iiL S |is


r [+M +Kj


mA


740_9


K ,


QaM


IaSs


Figure 2-8. Spectrum of chemical derivatization label. A) Spectrum of HPLC purified
label using a-cyano. B) Spectrum of HPLC purified label using 3-
hydroxypicolinic acid matrix. C) Spectrum of the deprotected label using a-
cyano.


4 A


mSo


c I-E-


1 1


ALA


13E44




1M1J


ia.J


3L2


i2s


MI I 1 -L I L. I L II I I ..


IluPI~Cr c~ ~ -Rrr


. .. . .


R'


t ,t














CHAPTER 3
INVESTIGATING THE AUTOPHOSPHORYLATION SITES OF A CALCIUM-
DEPENDENT PROTEIN KINASE

Since the regulatory properties of autophosphorylation of CDPKs still remain

unclear, one of the goals of this project was to develop methods for the identification of

phosphorylation sites of phosphorylated proteins and subsequently apply these methods

for autophosphorylation site identification of CDPKs. With the availability of three

different types of mass spectrometers each possessing a unique feature, the utilization of

all three instruments was considered beneficial for complete coverage of

autophosphorylation site mapping.

As mentioned in Chapter 1, both the QIT-MS and QqTOF-MS come equipped with

similar features for phosphorylation detection, that is, the reconstructed neutral fragment

chromatogram and precursor ion scanning, respectively, both taking advantage of the

liability of the phosphate group. Also, in-line reversed phase chromatography may be used

with these mass spectrometers for separation of peptides prior to MS and MS/MS

analysis. Additionally, phosphorylation site mapping is possible with the MS/MS feature.

The MALDI-TOF-MS on the other hand, does not have the capability of MS/MS

analysis; however, it has the advantage of rapid sample preparation and analysis as well

as affordability for smaller laboratories interested in proteomics. With this in mind, the

development of a novel, simple, cost effective method for preliminary phosphorylation

identification was proposed. The idea behind this was to take advantage of several

difficulties associated with phosphopeptide analysis, including hydrophilicity, ionization









efficiency and suppression effects by unphosphorylated peptides. Often, protein digests

are desalted with a ziptip which can result in the loss of phosphopeptides due to their

hydrophilicity. Phosphopeptides may not be seen in a MALDI-TOF spectrum due to

either loss of the phosphopeptide in the desalting step or to suppression effects. However,

if these peptides were first dephosphorylated by p-elimination prior to desalting and

spotting, the peptides should be visible in the spectrum. The appearance of these peptides

in the spectrum when compared to the spectrum of the untreated digest would then be an

indication of a previously phosphorylated peptide that has been dephosphorylated.

Targeted MS/MS analysis can then be performed on these peptides for phosphorylation

verification.

Presented is the development of a novel, cost effective method for preliminary

phosphorylation site identification on a MALDI-TOF-MS. Also presented is the

comparison of various available methods for autophosphorylation site mapping of a

calcium-dependent protein kinase.

Experimental Methods

Materials and Instruments

NuPAGE 10% Bis-Tris SDS-PAGE gels were obtained from Invitrogen (Carlsbad,

CA). Sequencing-grade modified trypsin was purchased from Promega (Madison, WI).

The Phosphopeptide Purification Kit (IMAC mini-spin columns) was from Pierce

(Rockford, IL). Reversed-phase C18 ZipTips were obtained from Millipore (Billerica,

MA). The MALDI calibration mixture Sequazyme Peptide Mass Standards Kit was

purchased from PE Biosystems (Foster City, CA). Mass spectrometric measurements

were made using either an LCQ Deca ion trap (ThermoFinnigan, San Jose, CA) equipped

with a PicoView electrospray ionization source (New Objective, Ringoes, NJ) and an









ABI 140D Solvent Delivery System (Perkin Elmer, Wellesley, MA) or a Voyager-DE

Pro Biospectrometry Workstation (Applied Biosystems, Foster City, CA) or a QSTAR

(Applied Biosystems, Foster City, CA) equipped with an LC Packings Ultimate

nanoHPLC system (LC Packings, Sunnyvale, CA).

Kinase Assay and Protein Preparation

Recombinant calcium-dependent protein kinases 4 and 5 (CPK4 and CPK5) from

Arabidopsis thaliana were a gift from Estelle Hrabak, University of New Hampshire.

Both kinases were autophosphorylated by incubation for one hour at room temperature in

50 mM Tris, pH 7.5, 10 mM MgCl2, 1 mM EGTA, and 1.2 mM CaCl2, and either 1 mM

[y-P32] ATP or 1 mM unlabeled ATP. P32-labelled kinases as well as the untreated

kinases were resolved on an SDS gel for autoradiography. In preparation for mass

spectral analysis GST-CPK5 autophosphorylated with cold ATP was resolved on an SDS

gel along with several standards (a-casein, 0-casein, BSA and ovalbumin). Protein bands

visualized by staining with Coomassie Brilliant Blue R250 were excised from the gel for

in-gel tryptic digestion 54. An in-solution tryptic digestion of P-casein was also

performed, however, reduction and alkylation of the protein was omitted. Samples were

then dried in a centrifugal vacuum system (SpeedVac) to near dryness.

Phosphopeptide Enrichment

Phosphopeptides from the autophosphorylated CPK5 digest were isolated and

enriched with the Pierce Phosphopeptide Purification Kit (gallium IMAC mini-spin

columns) according to protocol. The eluted phosphopeptides were dried and reconstituted

in 0.1% TFA. The IMAC-enriched sample as well as a sample of the digest (unenriched

sample) were desalted with a reversed-phase C18 ZipTip according to protocol and the









eluted peptides were dried and reconstituted in 5% ACN/0.5% acetic acid for mass

spectrometric analysis on the LCQ Deca and QSTAR.

P-Elimination

p-elimination of the in-solution and in-gel tryptic digests of CPK5 was performed

according to Knight's procedure68 in which the peptides were dissolved in a 4:3:1

solution of H20/DMSO/ethanol (50 p)L) followed by 23 itL saturated Ba(OH)2 and 1 itL

5 M NaOH. Samples were incubated at 37 OC for 2 hours then neutralized with HC1 and

dried with a SpeedVac. The samples were reconstituted in 0.1% TFA and desalted with a

reversed-phase C18 ZipTip for analysis with the MALDI-TOF mass spectrometer. It

should be noted that the in-solution digest of p-casein was first used to test this method

followed by the in-gel digests of all the standards (p-casein, a-casein, ovalbumin, and

BSA).

Data-dependent LC/MS/MS on the QIT-MS

Samples were introduced to the ion trap mass spectrometer via an on-line reversed-

phase capillary HPLC (50 [tm i.d. x 5 cm C18 produced in-house) with an isocratic

solvent delivery at 200 nL/min with 0% Solvent A (5% ACN/95% water/0.5% acetic

acid) for 5 min, and a linear gradient was performed for 20-30 min to 60% Solvent B

(95% ACN/5% water/0.5% acetic acid). The tryptic peptides were detected using data-

dependent acquisition whereby a full scan between m/z 300.0-2000.0 was first obtained

followed by a CID spectrum of the top 4 precursor ions (collision energy = 35%). ESI

conditions were as follows: capillary temperature, 200 OC; sheath gas flow, 0 L/min;

auxiliary gas flow, 0 L/min; ESI voltage, 1.20 kV; capillary voltage, 7.00 V; tube lens

offset, -5.00 V. The CID mass isolation window was set to 2.00 m/z units. The

subsequent information was input into the protein database searching programs Sequest









(ThermoQuest, San Jose, CA, USA) or MASCOT (Matrix Science Inc, Boston, MA,

USA). Additional phosphopeptides were identified by investigating the neutral fragment

chromatograms at m/z 98, 49 or 32.6 and manually interpreting the MS/MS data.

MALDI-TOF-MS Analysis of the P-eliminated Digests

The peptide samples of the untreated and P-eliminated digests of p-casein, a-casein,

ovalbumin, BSA and CPK5 were prepared using a matrix solution consisting of 53 mM

HCCA in 50% acetonitrile/0.1% TFA in a 1:1 ratio, that is, 1 [iL sample to 1 [iL matrix.

The samples were then air dried at room temperature on a stainless steel plate. Mass

spectra were obtained with a MALDI-TOF MS instrument equipped with a 337 nm

nitrogen laser and reflectron optics. All spectra were acquired in positive ionization

mode. The instrument was operated under delayed extraction conditions in reflectron

mode, mirror voltage ratio of 1.12, a delay time of 150 ns and grid voltage 70% of full

acceleration voltage (20 kV). Spectra were obtained with 100 laser shots at a laser

intensity of 2625 and laser repetition rate of 3.0 Hz. An external calibration was

performed before each spectrum was obtained with a calibration mixture consisting of 1.0

pmol Des-Argl-Bradykinin, 1.3 pmol Angiotensin I, 1.3 pmol Glul-fibrinopeptide B, 2.0

pmol ACTH (1-17 clip), and 1.5 pmol ACTH (18-39 clip). Once the spectra of the

untreated and P-eliminated digests were obtained, they were superimposed and peaks

present in the P-eliminated digest that were not visible in the untreated sample were

investigated further as possible phosphopeptides that had been dephosphorylated.

Precursor Scanning

Tryptic digests were loaded into an EconoTip (New Objective, Woburn, MA)

emitter and interfaced with the QSTAR instrument operated in precursor ion scanning

mode for the PO3- fragment ion (-79 m/z). The needle voltage was maintained between -









700 V and -800 V while the declustering and focusing potentials were -70 V and -225 V,

respectively. Mass spectra were acquired using a stepsize of 0.25 Da and a dwell time of

40 ms per m/z. The collision energy was adjusted to higher values (70-80 eV) in order to

optimize the production of the phosphate-derived low-mass fragment ions using nitrogen

as the collision gas.

Data-dependent LC/MS/MS on the QqTOF-MS

Capillary rpHPLC separation of protein digests was performed on a 15 cm x 75 um

i.d. PepMap C18 column (LC Packings, San Francisco, CA) in combination with an

Ultimate Capillary HPLC System (LC Packings, San Francisco, CA) operated at a flow

rate of 200 nL/min. Inline mass spectrometric analysis of the column eluate was

accomplished by a hybrid quadrupole time-of-flight instrument (QSTAR, Applied

Biosystems, Foster City, CA) equipped with a nanoelectrospray source. A two-point mass

calibration was performed in MS/MS mode of operation using the known fragment ion

masses of [Glu]-Fibrinopeptide (m/z 175.119 and m/z 1056.475).

Results and Discussion

Kinase Autophosphorylation

Figure 3-1 shows autoradiography results of the P32-labelled kinases demonstrating

that that the kinases were indeed autophosphorylated under autophosphorylation

conditions.

Ion Trap: Data-dependent LC/MS/MS of the Autophosphorylated CPK5 Digest

Analysis by Sequest software of data-dependent LC/MS/MS spectra of the

unenriched sample identified the phosphopeptide NSLNIpSMR. An additional

phosphopeptide DIYpTLSRK was revealed by manual analysis of the reconstructed

neutral fragment chromatogram of 98 m/z and the corresponding MS/MS spectrum of









947.39 m/z (Figure 3-2). Analysis of the IMAC-enriched digest by manual inspection of

the neutral fragment chromatograms at 98 m/z and 32.6 m/z revealed a triply charged

phosphopeptide GpSFKDKLDEGDNNKPEDYSK at 789.73 m/z as well as the

previously identified peptide DIYpTLSRK (Figure 3-3). This shows that without the use

of the IMAC columns for enrichment, the former peptide would have gone unnoticed

possibly due to suppression effects from the other unphosphorylated peptides present in

the digest. Overall, three autophosphorylation sites were identified by this data-dependent

LC/MS/MS method and IMAC enrichment.

Analysis of data from a similar experiment by MASCOT software resulted in the

identification of two phosphorylation sites (DIYpTLSR and TMRNSLNIpSMR)

previously identified by using both Sequest and neutral fragment scans. Neutral fragment

scans from this run revealed MS/MS data at m/z values matching m/z values of the

possible phosphopeptides TPNIRDIYpTLpSR and EMFQAMDTDNSGAITFDELK

(doubly phosphorylated but ambiguous as to which two sites). However, due to

insufficient fragmentation, confirmation of these phosphorylation sites was not possible.

It should also be noted that the phosphopeptide GpSFKDKLDEGDNNKPEDYSK

identified previously by IMAC enrichment was not identified in this analysis possibly

due to suppression effects. Overall, utilizing these various methods with the ion trap mass

spectrometer resulted in the identification of six possible phosphorylation sites of CPK5.

MALDI-TOFMS

Several controls (p-casein, a-casein, ovalbumin, and BSA) were first used to test

the MALDI-TOFMS p-elimination method as a first pass method for predicting peptide

phosphorylation. Figure 3-4 shows the results of the superimposed spectra of the

untreated and P-eliminated P-casein digests. Peaks that were present in the P-eliminated









spectrum and absent in the untreated digest were matched with m/z values of theoretically

dephosphorylated phosphopeptides. That is, peaks representing potentially

dephosphorylated peptides were compared to a list of all possible phosphorylated

peptides of 0-casein that was generated by the MS-Digest function of the database-

searching program Protein Prospector. Both known phosphopeptides

FQpSEEQQQTEDELQDK and ELEELNVPGEIVEpSLpSpSpSEESITR were detected

by this method, however, the monophosphorylated peptide was also indicated to be

doubly phosphorylated and the tetraphosphorylated peptide also indicated to have six

phosphorylation sites.

Although these possibly false sites were detected, both phosphopeptides were the

only peptides identified as being phosphorylated. Similar results (not shown) were also

attained for the other standards (phosphorylated and unphosphorylated proteins) whereby

the known phosphopeptides were identified, however, several false positives were also

identified. Although there were false positives, actual phosphopeptides were identified so

it was decided that this approach could be used as a first pass method to find possible

phosphopeptide targets for MS/MS verification.

This method was applied to the autophosphorylated CPK5 digest and three possible

phosphopeptides were found, RTMRNSLNIpSMR also found with the LC/MS/MS

Sequest and MASCOT analysis shown previously, LpTAHEVLRHPWICENGVAPDR

and IIQRGHYSERKAAELTK (one phosphorylation site but exact site is uncertain)

(Figure 3-5).

QqTOF-MS analysis

Precursor ion scanning of the autophosphorylated digest indicated the presence of

at least 12 phosphorylation sites. This was done by manually performing peptide mass









mapping whereby all m/z values with the loss of PO3 obtained from the precursor ion

scan were compared to a list of theoretically possible phosphopeptides generated by

Protein Prospector MS Digest, however, sequence verifications have proven to be

difficult. Figure 3-6 shows the precursor ion scan from 400-500 m/z. Among these were

the three phosphopeptides previously identified by the ion trap experiments.

Data-dependent LC/MS/MS analysis followed by MASCOT database searching

resulted in the identification of six phosphorylation sites, including previously identified

sites DIYTLpSR, NSLNIpSMRDA and GpSFKDKLDEGDNNKPEDYSK, as well as

additional sites NpSLNIpSMR, MLpSSKPAER, and

TSTTNLSSNSDHSPNAADIIAQEFSK (1 site) (Figure 3-7). It should be noted that the

phosphopeptide fragmentation on this instrument was significantly better than that of the

ion trap as many of the b and y ions were generated on the QSTAR as compared with the

ion trap which showed mainly the dephosphorylated peptide after collision-induced

dissociation.

The development of a simple, cost effective method for phosphorylation

identification using p-elimination and a MALDI-TOF MS has been shown to provide

targets for sequence verification. This will be useful for laboratories with only a MALDI-

TOF instrument, however, sequences will have to be verified once the targets are found

due to possible false positives.

Comparison of results obtained by these three instruments gave significant overlap

showing that these methods are indeed complimentary (Figure 3-8). Overall, at least 17

possible phosphorylation sites were observed (Figure 3-9). Of these sites, 5 sites are seen

in at least two of these methods and 7 have been verified by MS/MS analysis. Combining









the three methods gave an increased identification of autophosphorylation sites that

would not have been possible by any one method. Having these complimentary results

gave an increase in the confidence that these sites are indeed valid, however, future

experiments need to be performed to verify the sites, after which we will attempt to

interpret their significance.

Because the autophosphorylation properties of CDPKs are not fully understood,

localization of the autophosphorylation sites of an Arabidopsis thaliana member of this

family by various complimentary methods are presented as a contribution towards their

understanding. Compilation of these autophosphorylation sites with those of previously

published sites such as that of the tomato (LeCPK1),79 ice plant (McCPK1)9 or

arabidopsis thaliana (CPK1)80 will aid in their understanding by possibly identifying a

common sequence motif.






















Figure 3-1. Autoradiography of SDS-PAGE separated autophosphorylated CPDKs. A) 2
hour exposure. B) 2 day exposure.










P&L- 43.9


Ljig


2 0 S 0 21 10 32 7 24 0 If


16 22II-rre
2 4 6 S 10 12 14 16 18 20 22 24 26 2S 30 F




M 17J7 M Rft -r-


I 2 4 86 10 12 14 16 1 20 22 24 26 2 30


3493


r


c i 543t15






,, faHL 73+1- 2


7w sO 1100 1300 195w 17w1 1s


Figure 3-2. Autophosphorylated CPK5 digest. A) Base peak chromatogram. B)
Reconstructed neutral fragment 98 chromatogram. C) MS/MS of the
phosphopeptide DIYpTLSRK (947.39 m/z) located by the neutral fragment
chromatogram at the retention time 17.77 min.


- Ii- ilk


m


in












M- 1.7


D 2 4 H H 1D 12 14 1D 18 2D


B


SIC
I PN,
rr


I Iii 1111 'II--- I I IIiii. i


mI IT DI i TI DIII
WoD 5w1 M WD


111DD
111)o


I_ ilMED


131D 15D) 17DD 19D1


757-1
IP*H-
../HKYW


h.7 hrH0 Y2


EL2t 5972 L 5L~


T 93OL5
B73M 9E"a
sas


M 1210 I





121D-5


liT '. '. '. '' I 1 .. .....


.' 5i . . i


131


1539
rsmll


Figure 3-3. IMAC enriched phosphopeptides from autophosphorylated CPK5 digest. A)
Base peak chromatogram. B) MS/MS of the phosphopeptide DIYpTLSRK
located by the neutral fragment 98 chromatogram at 12.06 min. C) MS/MS of
the phosphopeptide GpSFKDKLDEGDNNKPEDYSK located by the neutral
fragment 32.6 chromatogram at 11.80 min.


1I

e



I
^ a
e
U
a


r I


3 11HI


I1 111


S


" h^JFO 7" 1 MR2 ~
M75Ll rJ 7dC9l






73


9- Uhmbddist





A















II


























Figure 3-4. Spectra of 3-casein digest. A) Untreated digest. B) 3-eliminated digest.
Dephosphorylated phosphopeptides in the 3-eliminated digest are indicated
with *








74


t00, 1.1E4

Un9Leated digest
S-diminiated digest

80


70


60 A


50


40


30


20


10-


S102 ISA 2040.6 252 301.0 Mass (f



1 1*77""








Is








0+4











Figure 3-5. Spectra of CPK5 digest. A) Untreated digest. B) j-eliminated digest.
Dephosphorylated phosphopeptides in the j-eliminated digest are indicated
with *
Fiue35 pcr fCK iet )Unrae iet )peiiae iet












425.34B2
1.20e4



1.00e4 4079521
I 463.257





O- D
D
o 6000.0 447.2855




S4000.35.06






400 410 420 430 440 450 460 470 480 490 500

m&z, amu


Figure 3-6. Precursor ion scan (400 500 m/z) of autophosphorylated CPK5 digest. m/z
values were matched with predicted phosphorylated protein digests.


















2zz4391


10aI Al 14M 14 M14









800 100I 12W 140 160 1800


,
N
s: 3T


*r rt
s < v

4
.0 .. l
39


I i 4I
zoo 400


rs 4




If I
Kfi


rs
Y
u



i L


kr,
* C;

--
o -
3t I


3i -


-' I I M"W1T LL I "vw .-,Wpm"


600 800


10 0 I12 140 1I r I 6 I I


331 2a


200 40


Figure 3-7. MS/MS spectrum and corresponding MASCOT search result of the
phosphopeptide TSTTNLSSNSDHSPNAADIIAQEFSK (m/z 939).


I


7T1 5z2


. .J,














M(;NrCRwilV;S DKLDEGDNNK
D PALVI PLR EPIMRRNPDN
VDYACKSISK RKLISKEDVE
GELFDRI QR GHYSERKAIAI
IDFGLSVFFK PGQIFTDVVWG
'TQQGi DAVL KGYIDFESDP
APDRALDPAV LSRLKQFSAM
DELKAGLRKY GSTLKDTEIH
YFDKDGSGFI TIDELQQACV
GRRAaRSIS IBMRDA


PEDYSKTSTT
QAYYVLGHEK'
DVRREIQfIMH
I,TK i V;(VVKI
SPYYVAPEVL
WPV IDSLAKD
NKLKKMALKV
lj]MDAADVDN
EHGMADVFLE


NLSSNSDHSP
PN I HI I YT k.
BLGBGSIOVT
ACII: ,(;VMI' l
LKRYGPEADV
L RI S. .- 1
IAESLSEEEI
SGTIDYSEFI
DIIKEVDQNN


NAADIIAQEF
fi I.(.;Q QP T(;
IKGAYEDSLY
DLKPENFLLV
WTAGVILYTL
A i AGLREMFQAM
AATIHLNKLE
DGKIDYGEFV


SKDESUNHSB
TY [.CTL' I ASG
VH IVMELCAG
NKDDDFSLKA
LSGVPPFWAE
1IHPWICENGV
DTDNSGAITF
R ; EH LVAA FQ
EMMQKGNAGV


Figure 3-8. Overlap of phosphopeptides identified by all methods. Blue squares indicate
peptides identified on the MALDI-TOF-MS, green squares indicate those
identified on the QIT-MS, and yellow indicate those identified on the
QSTAR. Sequences in orange are sequences identified by precursor ion
scanning.


MGNSCRGSFKDKLDE
PNIRDIYTLSRKLGQ
IRDIYTLSRKLGQGQ
DLIRRMLSSKPAERL
GRRTMRNSLNISMRD
MRNSLNISMRDA
KPEDYSKTSTTNLSSNSD
NAGVGRRTMRNSLNI


> C ned by MS/MS




A


(1 phospho)


IIQRGHYSERKAAEL
ERKAAELTKIIVGVV
GVVEACHSLGVMHRD
SKPAERLTABEVLRH
EMFQAMDTDNSGAI T
QAMDTDNSGAITFDE
TDNSGAITFDELKAG
EFSKDNNSNNNSKDPALVI (1 phospho)
HILAGHGSIVTIKGAYED (1 phospho)


Figure 3-9. Phosphorylated peptides identified by all methods.


1
61
121
181
241
301
361
421
481
541














CHAPTER 4
IDENTIFICATION OF IN VITRO SUBSTRATES OF A CALCIUM-DEPENDENT
PROTEIN KINASE

In order to gain an understanding of the specificity and function of calcium-

dependent protein kinases, substrates of these kinases need to be determined. Several

approaches may be applied to achieve this goal: 1) mass spectrometric identification of

substrates phosphorylated by CDPKs in vitro, 2) tandem affinity purification and

identification of in vivo substrates, and 3) CDPK substrate traps with yeast two-hybrid

systems. For the purposes of this project, the mass spectrometric identification of in vitro

substrates approach was determined as the method of choice due to the availability of

mass spectrometry facilities.

Several problems are associated with this approach. First, if cell lysates are used as

the source of substrate proteins, protein phosphatases and kinases already present in the

cell lysate need to be inhibited. Second, all phosphates already present in the substrate

proteins need to be removed prior to phosphorylation by the kinase in order to be able to

determine the source of protein phosphorylation. Finally, since many important substrates

may be present at low abundance, isolation and enrichment of these proteins is necessary.

Several commercial products were available to solve these problems. First,

denaturation of proteins results in the inhibition of phosphatases and kinases. Several

denaturing extraction methods have been compared by Saravanan and Rose49 with the

phenol-based method giving the highest protein yield and best resolution and spot

intensity with gel electrophoresis. Invitrogen's TRIzol Reagent, a ready-to-use reagent









consisting of a mono-phasic solution of phenol and guanidine isothiocyanate, can be used

for easy extraction of proteins from plant tissue. This occurs by the disruption of cells and

dissolving of cell components by the reagent, resulting in an overall purified protein

extract after the removal of RNA and DNA with chloroform and ethanol, respectively.

For dephosphorylation of proteins in the extracts, several phosphatases were

available: calf intestinal alkaline phosphatase (CIP) (in-solution or immobilized on

agarose), Antarctic phosphatase, and biotinylated phosphatase. Each of these

phosphatases has its advantage. CIP in solution is the cheapest of the four and since

derivatizations such as immobilization are not performed on the protein, active sites

should be easily accessible for dephosphorylation of proteins in the extract. However,

inhibition of the phosphatase after treatment and prior to kinase treatment is necessary.

Agarose-immobilized calf intestinal phosphatase and biotinylated phosphatase have the

advantage that they can be removed easily from the protein extract; however, as

mentioned above, immobilization can possibly inhibit the activity of a percentage of the

phosphatase due to blocking of the active sites. Antarctic phosphatase, which can be

completely deactivated by a short heat treatment, also has the advantage of facile

inhibition prior to kinase treatment. Since several types of phosphatases were available,

determination of the most appropriate phosphatase for the purposes of this project was

deemed necessary.

Finally, the development of a proprietary product from Qiagen (The

PhosphoProtein Purification Kit) designed for the specific purification of phosphorylated

proteins from complex cell lysates has made enrichment of low abundance proteins

possible. The principle of this method is that proteins that carry a phosphate group on any









amino acid are bound with high specificity to a PhosphoProtein Purification Column,

while proteins without phosphate groups do not bind to the column and can therefore be

found in the column flow-through fraction. Binding of phosphorylated proteins occurs by

flowing the lysate (-0.1 mg/mL) through the column at a flow rate of about 0.5 mL/min.

Low lysate concentration and flow rate are used to ensure that all phosphate groups are

easily accessible and are not hidden within protein complexes, and that complete binding

of phosphorylated proteins occurs. Since 7-15% of proteins from cells are expected to

carry one or more phosphate groups, the expected yield from one of these columns for

2.5 mg of protein from a cell lysate is 175-500 tg of phosphorylated protein. The

maximum binding capacity of one of these columns is 500 [g of phosphorylated protein.

This chapter demonstrates the development of a method for the identification of

substrates phosphorylated by kinases in vitro. A comparison of the performance of

available phosphatases for dephosphorylating plant extract is shown as well as the

comparison of various phosphatase inhibition methods. Application of the optimized

dephosphorylation and inhibition steps followed by enrichment of the resulting in vitro

phosphorylated substrates is shown.

Experimental Methods

Materials and Instruments

TRIzol Reagent and NuPAGE 10% Bis-Tris SDS-PAGE gels were obtained from

Invitrogen (Carlsbad, CA). Calf Intestinal Alkaline Phosphatase, Antarctic Phosphatase,

Biotinylated Phosphatase and Streptavidin Magnetic Beads were purchased from New

England Biolabs, Inc. (Beverly, MA). Immobilized Calf Intestinal Alkaline Phosphatase

was purchased from Sigma (St. Louis, MO). The PhosphoProtein Purification Kit was

from Qiagen Inc. (Valencia, CA). Pro-Q Diamond Phosphoprotein Gel Stain was from









Molecular Probes (Eugene, OR). Sequencing-grade modified trypsin was purchased from

Promega (Madison, WI). CPK4 was a gift from Estelle Hrabak.81

Mass spectrometric measurements were made using an LCQ Deca ion trap

(ThermoFinnigan, San Jose, CA) equipped with a PicoView electrospray ionization

source (New Objective, Ringoes, NJ) and an ABI 140D Solvent Delivery System (Perkin

Elmer, Wellesley, MA).

Method Development

Protein extract preparation

Proteins were extracted from mature Arabidopsis thaliana leaves with TRIZOL

Reagent by grinding the leaves with a pre-chilled mortar and pestle and liquid nitrogen.

Ground leaves were then transferred to a chilled Corex centrifuge tube and 5 mL of

TRIZOL Reagent was added for every 500 mg plant leaves. Ground leaves were

allowed to sit in the TRIZOL Reagent for 5 minutes at room temperature after which the

sample was centrifuged at 10,900 xg for 5 minutes at 40C. The supernatant was then

transferred to a fresh Corex tube, centrifuged for another 5 minutes, and the resulting

supernatant transferred to an Oakridge tube. One mL of chloroform was then added and

the tube shaken vigorously by hand for 15 seconds then allowed to stand at room

temperature for 2 minutes. The sample was centrifuged at 10,900 xg for 15 minutes and

the upper (aqueous) layer completely removed. 1.5 mL of absolute ethanol was then

added to the lower phase (phenol-chloroform phase) and the sample mixed by inversion

followed by incubation for 3 minutes at room temperature. Centrifugation of the sample

at 483 xg was followed by the transfer of the supernatant to a fresh Corex tube. Three

volumes of ice cold acetone was added to the supernatant followed by centrifugation for

2 minutes at 2,860 xg. The supernatant was then decanted, the acetone wash was









repeated, and the pellet was air dried. The protein pellet was solubilized with 1% SDS in

50 mM Tris-HC1, pH 7.5 for 30 minutes at 50 C. The sample was diluted to obtain a

final SDS concentration of 0.1%, and it was dialyzed against three changes of 1L of 50

mM Tris-HC1, pH 7.5. Dialyzed extract was concentrated by ultrafiltration in an Amicon

concentrator containing a 10,000 molecular weight cut-off membrane. Protein

quantification was performed by Bradford Protein Assay.

Dephosphorylation of the protein extract

Several phosphatases were used for dephosphorylation optimization of the protein

extract: Calf Intestinal Alkaline Phosphatase, Antarctic Phosphatase, Biotinylated

Phosphatase, and Immobilized Calf Intestinal Alkaline Phosphatase. All phosphatases

were used according to the manufacturers' protocols and the resulting samples were

separated on an SDS gel and stained with the Pro-Q Diamond Phosphoprotein Gel Stain

according to the manufacturer's protocol. CIP was determined as the most suitable

phosphatase for these experiments.

Phosphatase inhibition

Since the CIP was in solution, several methods of phosphatase inhibition were

tested: the addition of sodium orthovanadate, EDTA or phosphoserine. The model protein

used for these studies was a fusion protein in which maltose binding protein was linked to

the N-terminus of soybean serine acetyl transferase (MBP-SAT), which had previously

been phosphorylated by CDPK4. Varying concentrations and times for each inhibitor

were used to optimize the inhibition step. Also, kinase action after addition of these

inhibitors was observed to ensure that the kinase was not being inhibited. This was done

by adding CDPK4 to the inhibited sample with the needed phosphorylation buffer, ImM









ATP, 10 mM MgC12, 1 mM EGTA, and 1.2 mM CaC12. The Pro-Q Diamond

Phosphoprotein Gel Stain was used to monitor all of these reactions.

In vitro phosphorylation of Arabidopsis thaliana extract with CDPK4

Proteins were extracted from 2 g of mature Arabidopsis thaliana according to the

method described previously. Removal of phosphate groups added to the proteins by in

vivo phosphorylation was then performed by adding 6,750 units of CIP to the extract in a

buffer containing 100 mM NaC1, 50 mM Tris-HC1, 10 mM MgCl2, and 1 mM

dithiothreitol, pH 7.9 at 37 OC for 1 hour. Sodium orthovanadate was then added to a final

concentration of 10 mM and the sample incubated at room temperature for 5 minutes.

Buffers and excess vanadate was removed from the extract by dialysis in 50 mM Tris

buffer, pH 7.5. The dephosphorylated protein extract was then split into two equal

aliquots for a control and an in vitro phosphorylated sample. Phosphorylation of the

dephosphorylated extract was performed by incubation with CDPK4 in a buffer

containing 1 mM ATP in 50 mM Tris, pH 7.5, 10 mM MgCl2, 1 mM EGTA, and 1.2 mM

CaC12 in the cold overnight, followed by incubation at room temperature for one hour. To

the control sample, the same conditions were applied except for the addition of the

kinase.

Phosphoprotein enrichment

Excess ATP was removed from the extract by dialysis in 50 mM Tris buffer, pH

7.5 followed by phosphoprotein enrichment with Qiagen's PhosphoProtein Purification

columns according to manufacturer's protocol. Eluates and flow-throughs from both

samples were then resolved on an SDS gel and visualized by staining with Pro-Q

Diamond Phosphoprotein Gel Stain according to protocol and Coomassie Brilliant Blue