PROTEOMIC ANALYSIS OF PP2A MUTANTS AND
PP2A-RELATED TUMOR VIRUS PROTEINS
by
Yiwang Zhou
A thesis submitted in conformity with the requirements for the degree of Master of Science Graduate Department of Molecular Genetics University of Toronto
© Copyright by Yiwang Zhou 2015
PROTEOMIC ANALYSIS OF PP2A MUTANTS AND
PP2A-RELATED TUMOR VIRUS PROTEINS
Yiwang Zhou Master of Science Graduate Department of Molecular Genetics University of Toronto 2015
ABSTRACT
Protein phosphatase 2A (PP2A) is a tumor suppressor, whose function and activity is influenced by mutations of PP2A subunits and PP2A-related tumor virus proteins. In this work, protein-protein interaction changes in the normal PP2A interaction network caused by mutations of PP2A scaffolding A (PPP2R1A) and catalytic C (PPP2CA) subunits have been characterized by AP-SWATH. The results show that the normal function and activity of PP2A holoenzyme may be significantly altered and influenced by these mutations. This study also characterizes the interactomes of several PP2A-related tumor viral oncoproteins, including E4orf4. It is revealed for the first time that ASPP-PP1 complex subunits are among the major interactors of E4orf4, suggesting the involvement of E4orf4 in the regulation of Hippo signaling pathway. The results gained in my M.Sc. work provide a deeper understanding of the consequences of PP2A subunits mutations and the PP2A-dependent and PP2A-independent functions of the PP2A-related tumor virus proteins.
! ii! ACKNOWLEDGEMENT
First and foremost, I would like to thank my supervisor, Dr. Anne-Claude Gingras for taking me as a master student, and for all the time and patience spent in providing me guidance throughout this study. Before joining Dr. Gingras' lab, I had no idea what real scientific research is and how to carry out an independent research project. These two years' experience develops my ability to work and think like a researcher, which is extremely worthwhile for my future career.
I would also like to express my appreciation to the members of Dr. Gingras' lab: Jean-Philippe Lambert for answering my numerous questions and for helping me analyze the SWATH data; Zhen-yuan Lin for teaching me affinity purification and for always being there when I had trouble dealing with the mass spectrometers; James Knight for helping me with data processing and visualization; Frank Liu for solving all the problems I had in ProHits; Wade Dunham and Marilyn Goudreault for helping me find everything I needed in the lab; and the rest of the lab members for assistance through the years. For smoothly carrying out this study, I also want to thank my committee members, Dr. Philip Kim, Dr. Stéphane Angers and Dr. Vuk Stambolic for their guidance, as well as Dr. Philip Branton and Dr. Egon Ogris for their collaboration.
In addition, I am indebted to my parents, who are always standing by me and offering me support and encouragement to pursue my dream. I am also grateful to have Huayun Hou as my roommate. Thanks for accompanying for these two years. Finally and most especially, I would thank to my boyfriend, Jiyuan Yang, for his understanding, support, patience and love during the past two years, no matter how far away we are separated.
! iii! TABLE OF CONTENTS
ABSTRACT...... ii ACKNOWLEDGEMENT...... iii LIST OF TABLES...... vi LIST OF FIGURES...... vii LIST OF ABBREVIATIONS...... ix
CHAPTER 1 INTRODUCTION 1.1 PP2A holoenzyme...... 1 1.2 The function of PP2A as a tumor suppressor...... 6 1.3 Structure-function of the PP2A subunits under study 1.3.1 Scaffolding A subunit, PPP2R1A...... 8 1.3.2 Catalytic C subunit, PPP2CA...... 9 1.3.3 Regulatory B subunit, PPP2R2A...... 10 1.4 PP2A-related tumor virus proteins 1.4.1 SV40 small T antigen...... 11 1.4.2 Polyomavirus middle T antigen...... 12 1.4.3 Adenovirus E4orf4 protein...... 13 1.5 Affinity purification and mass spectrometry...... 14 1.6 Proximity-dependent biotin identification with BioID...... 19 1.7 Quantitative mass spectrometry with the Data Independent Acquisition method SWATH...... 22 1.8 Objectives of this project...... 24 CHAPTER 2 MATERIALS AND METHODS 2.1 Plasmids...... 26 2.2 Cell lines, culture, transfection and collection...... 29 2.3 Affinity purification 2.3.1 Anti-FLAG affinity purification...... 30 2.3.2 Streptavidin affinity purification...... 32
! iv! 2.4 Immunoprecipitation-Western Blot...... 33 2.5 Mass spectrometry 2.5.1 MS/MS...... 34 2.5.2 SWATH...... 35 2.6 Data analysis and visualization 2.6.1 DDA data search...... 36 2.6.2 SAINT analysis...... 37 2.6.3 SWATH data analysis...... 38 CHAPTER 3 RESULTS 3.1 Protein-protein interaction changes imparted by PPP2R1A mutations...... 42 3.2 Validation of selected interaction changes for PPP2R1A mutations...... 48 3.3 Protein-protein interaction changes imparted by PPP2CA mutations...... 51 3.4 Validation of the interaction between ANKLE2 and PPP2CA mutants...... 59 3.5 Interacting proteins of PPP2R2A...... 62 3.6 Interacting proteins of PP2A-related tumor virus proteins 3.6.1 SV40 small T antigen interacting proteins...... 67 3.6.2 Polyomavirus middle T antigen interacting proteins...... 72 3.6.3 E4orf4 interacting proteins...... 72 CHAPTER 4 FUTURE DIRECTIONS AND DISCUSSION 4.1 Significance of the work...... 82 4.2 Discussion...... 83 4.3 Future directions...... 90 APPENDIX...... 92 REFERENCES...... 102
! v! LIST OF TABLES
Table 1. PP2A subunits...... 5 Table 2. Plasmids used in this study...... 27 Table 3. Mutations of PPP2R1A...... 43 Table 4. Mutations of PPP2CA...... 52 Table 5. Schematic representation of the various C subunit mutations...... 52 Table 6. Mutations of PPP2R2A...... 63 Table 7. Mutations of E4orf4...... 75
! vi! LIST OF FIGURES
Figure 1. Crystal structure of PP2A and the formation of PP2A holoenzyme...... 3 Figure 2. Workflow of affinity purification coupled with mass spectrometry (AP-MS)...... 16 Figure 3. Model for application of BioID method...... 20 Figure 4. AP-SWATH data analysis pipeline...... 39 Figure 5. Dot-plot representation of protein-protein interaction changes for eight PPP2R1A mutations relative to wild-type PPP2R1A...... 45 Figure 6. Validation of selected protein-protein interaction changes imparted by mutations of PPP2R1A...... 49 Figure 7. Dot-plots representation of protein-protein interaction changes for eleven PPP2CA mutations relative to wild-type PPP2CA...... 54 Figure 8. Validation of interaction changes for ANKLE2 imparted by mutations of PPP2CA...... 60 Figure 9. Characterization of the interacting proteins of wild-type PPP2R2A and PPP2R2A mutants...... 65 Figure 10. Characterization of the interacting proteins of wild-type PPP2R2A and PPP2R2A mutants using BioID...... 68 Figure 11. Characterization of the interacting proteins of wild-type SV40 ST and ST mutant...... 70 Figure 12. Characterization of the interacting proteins of PyMT...... 73 Figure 13. Characterization of the interacting proteins of wild-type E4orf4 and the class I E4orf4 mutant...... 77 Figure 14. Characterization of the interacting proteins of wild-type E4orf4, class I mutant and class II mutants...... 79 Figure 15. Immunofluorescence of Hela cells expressing wild-type PPP2CA and PPP2CA mutants...... 86 figure 1. Expression of FLAG tagged wild-type PPP2R1A and PPP2R1A mutants in HEK293 T-REx cells...... 92
! vii! figure 2. Expression of FLAG tagged wild-type PPP2CA and PPP2CA mutants in HEK293 T-REx cells...... 94 figure 3. Characterization of the interacting proteins of FLAG tagged wild-type PPP2CA and PPP2CA mutants...... 96 figure 4. Expression of FLAG tagged wild-type PPP2R2A and PPP2R2A mutants in HEK293 T-REx cells...... 98 figure 5. Expression of BirA*-FLAG tagged wild-type PPP2R2A and PPP2R2A mutants in HEK293 T-REx cells...... 100
! viii! LIST OF ABBREVIATIONS
AP Affinity purification AvgP Averaged probability CID Collision induced dissociation COSMIC Catalogue of Somatic Mutations in Cancer ctrl control DDA Data-dependent acquisition DIA Data-independent acquisition E4orf4 Early transcription region 4 open reading frame 4 ER Early region FDR False discovery rate HEAT Huntington-Elongation-A subunit-TOR HPLC High Performance Liquid Chromatography IP Immunoprecipitation LC Liquid chromatography LR Late region LT Large T antigen LUMIER Luminescence-based mammalian interactome mapping MLR most likely ratio MS Mass spectrometry MT Middle T antigen PP1 Protein phosphatase 1 PP2A Protein phosphatase 2A PPP Phosphoprotein phosphatase PyMT Polyomavirus middle T antigen S/MRM Selected/Multiple Reaction Monitoring SAINT Significance Analysis of INTeractome SILAC Stable Isotope Labeling by Amino acids in Cell culture ST Small T antigen
! ix! SV40 Simian virus 40 SWATH Sequential Window Acquisition of all THeoretical spectra TPP Trans-Proteomic Pipeline WB Western blot XIC extracted ion chromatography Y2H Yeast-two-hybrid
°C Degrees Celsius µg microgram µl microliter µM micromolar µm micrometer amu atomic mass units cm centimeter Da Daltons eV electron-volt m/z mass to charge ratio min minute ml milliliter mM millimolar ms microsecond ng nanogram nl nanoliter ppm parts permillion rpm revolutions per minute
! x! ! 1!
CHAPTER 1 INTRODUCTION
1.1 PP2A holoenzyme Reversible protein phosphorylation has been widely recognized as one of the most important regulatory mechanisms in eukaryotic cells. Proteins are phosphorylated by kinases and dephosphorylated by phosphatases, which results in altering the properties of those proteins, including their activities, localization, etc.[1]. Phosphorylation of serine and threonine residues is mediated by ~350 kinases[2], but a more limited number of serine/threonine phosphatases antagonize them[3]. How a limited number of catalytic subunits for phosphatases can still target specific substrates has been the topic of many studies, and it is now widely accepted that several members of the evolutionary conserved "PPP" (phosphoprotein phosphatase) subfamily of serine/threonine phosphatases (13 members in human) acquire this specificity in part by their association with non-catalytic subunits that provide localization or substrate recognition clues. In the best-characterized example, the protein phosphatase 1 (PP1) associates with at least 200 validated proteins in mammalian cells as dimers or trimers, by recognizing mutually exclusive short linear motifs on these interactors[4]; these interactors help bringing the phosphatase in the vicinity of the substrates, dock to the substrate and sometimes block substrate-binding channels, functioning in this last case as inhibitors.
Together with PP1, protein phosphatase 2A (PP2A) is a widely expressed Ser/Thr phosphatase, accounting for a large proportion of the total Ser/Thr phosphatase activity in the cell. Through regulation of multiple signaling pathways, PP2A is involved in controlling many essential aspects of biology, such the cell cycle, cell growth, apoptosis, etc.[5-7].
Like PP1, PP2A does not function in isolation in cells: it is most often present as a heterotrimer. The trimeric holoenzyme PP2A is composed of a catalytic C subunit (PP2A C; two genes, PPP2CA and PPP2CB, exist in mammals), a scaffolding A
! ! 2! subunit (PP2A A; two genes, PPP2R1A and PPP2R1B are expressed in humans), and one of many regulatory B subunits (PP2A B) (Fig. 1a). The PP2A B subunits can be classified into four structurally unrelated families: B, B', B'' and B'''[5, 6] (subunit classification is shown in Table 1). Multiple families, isoforms, and splice variants of the PP2A B subunits allow (together with the two catalytic and two scaffolding subunits) the combinatorial formation of more than 60 different PP2A holoenzymes[8]. The PP2A trimeric holoenzyme is thought to be assembled in two steps. First, one of the two isoforms of PP2A A and C subunits associate to generate the PP2A core dimer. In order to gain full activity towards specific PP2A substrates, this core dimer then binds to one PP2A B subunit to yield the trimeric holoenzyme[5-7, 9] (Fig. 1b). Isoforms for either PP2A A and C subunits have high sequence similarity[10-13]. Although PP2A B subunits within a same family (B, B’, B’’, B’’’) also exhibit sequence similarity, PP2A B subunits from different families have no sequence homology[5, 6]. Based on such variety, PP2A B subunits virtually determine the substrate specificity and localization of PP2A holoenzymes[9].
The catalytic subunits of PP1 and PP2 have remained highly constant during the course of evolution and may be the most conserved among all known enzymes[14]. Studies of PP1 across different species have shown that the amino sequence of PP1 C subunits from mammals and Drosophila shared more than 90% overall identity. The corresponding PP1 C subunits in yeast and Aspergillus are over 80% identical to the mammalian PP1 C. PP2A is also well conserved[15]. For example, the α-isoform of the PP2A C subunit from human and rabbit are 100% identical and the β-isoform shares 99.7% identity between these species[16]. The remarkable conservation of the catalytic subunits of protein phosphatases may reflect the fact that they play an essential role in regulating cellular functions[14]. The A subunits of PP2A are also highly conserved during evolution and the α-isoform of PP2A A subunit has a similar changing rate with histone H4, one of the most conserved protein known so far. Although PP2A B subunits have diverse structures and substrate specificity, B subunits of the same family from different species exhibit some identity. Structural
! ! 3!
Figure 1. Crystal structure of PP2A and the formation of PP2A holoenzyme. (a) PP2A is composed of a scaffolding A subunit (red), a catalytic C subunit (blue) and one of many regulatory B subunits (yellow). The PP2A B subunit shown here is
PPP2R5A. Structure solved by Xu, et al.; Cell (2006) 127:1239-1251 (image from PDB
Japan). (b) PP2A C subunit first binds to A subunit to form the core dimer of PP2A. Then one of the B subunits interacts with this core dimer through its association with PP2A A at the other end, making the PP2A holoenzyme.
! ! 4!
! ! ! a
C subunit
A subunit
B subunit
! b
C C B C
A A A B
PP2A core dimer PP2A holoenzyme ! ! ! ! ! ! ! ! ! ! ! !
! ! 5!
Table 1. PP2A subunits Subunit function Family Protein Isoform Scaffolding A subunit PPP2R1A Aα PPP2R1B Aβ Catalytic C subunit PPP2CA Cα PPP2CB Cβ Regulatory B subunit B (B55) PPP2R2A Bα PPP2R2B Bβ PPP2R2C Bγ PPP2R2D Bδ B' (B56) PPP2R5A B'α PPP2R5B B'β PPP2R5C B'γ PPP2R5D B'δ PPP2R5E B'ε B'' PPP2R3A B''α PPP2R3B B''β PPP2R3C B''γ B''' STRN STRN3 STRN4
! ! 6! studies of human B55α subunit and the yeast ortholog Cdc55 reveal that they have a high degree of similarity, with approximately 56% identity in their crystal structures[17].
1.2 The function of PP2A as a tumor suppressor Dysregulation of the balance between protein phosphorylation and dephosphorylation plays an important role in cancer initiation and maintenance. Abundant evidence has supported that kinases are often mutated to active, oncogenic forms. In addition to kinases, recent findings have also pointed out the importance of protein phosphatases in malignant transformation. As a major Ser/Thr phosphatase, PP2A has been demonstrated to function as a tumor suppressor[8, 18-23]. Convincing evidence has been provided to show that suppression of PP2A activity can cooperate with other oncogenic changes to cause transformation of various cell types[8].
The first line of evidence illustrating the function of PP2A as a tumor suppressor came from the studies of chemical and protein inhibitors of PP2A, including okadaic acid (OA), CIP2A, I1PP2A/I2PP2A, etc.[8, 19, 23]. OA is a tumor promoter that can stimulate premature meiosis and mitosis[5], and increased expression of several proto-oncogenes such as c-fos and c-jun[24], both of which will contribute to the promotion of carcinogenesis[25, 26]. As an inhibitor of PP2A, the experiments with OA suggested that PP2A plays a negative role in regulating cell growth [27]; however, it needs to be mentioned that OA, like all other chemical PP2A inhibitors identified to date, inhibit not only PP2A (all holoenzymes), but also its closest relatives PP4 and PP6 and in many instances other enzymes of the PPP family, including PP1[28]. Structural analyses of phosphatase catalytic subunits help explain this lack of specificity: all these compounds interact at the phosphatase active site, which is the most conserved area of these phosphatases. Results solely based on these chemical inhibitors therefore need to be interpreted with caution.
By contrast to the low intrinsic specificity of chemical inhibitors to the catalytic
! ! 7! subunits of PP2A, protein inhibitors have the potential to bind to surfaces specific to one but not another enzyme. For example, CIP2A (cancerous inhibitor of PP2A) selectively targets PP2A to inhibit its phosphatase activity towards c-Myc, resulting in the increased phosphorylation of c-Myc as well as the reduction in its degradation[29]. CIP2A expression is upregulated in transformed cell lines and cancer tissue samples[8], which correlates with the enhanced stability and overexpression of c-Myc in a wide variety of human cancers, suggesting one more link between PP2A and cancer[30]. I1PP2A and I2PP2A are two other protein inhibitors of PP2A activity toward several phosphorylated substrates[31]. Although the function of I1PP2A and I2PP2A remain incompletely understood, it has been proposed that the inhibition of PP2A by I2PP2A stimulates the activity of the MEK-ERK-MAPK pathway and c-Jun phosphorylation[32, 33]. Defects in the MEK-ERK-MAPK pathway have been identified in many human cancers, and lead to uncontrolled growth of the tumor cells[34, 35]. Activation of c-Jun by phosphorylation is required during tumor initiation and progression[36]. The identification of these PP2A inhibitors in contributing to cancer development contributed bolstering the role of PP2A as a tumor suppressor.
Alteration of PP2A function by mutations in specific PP2A subunits such as the scaffolding A subunit (PPP2R1A and PPP2R1B) has also been detected in several cancers. The earliest study has reported somatic mutations of PPP2R1B in 15% of primary lung tumors, in 6% of lung tumor-derived cell lines, and in 15% of colorectal carcinomas[37]. Further research also detected somatic alterations of PPP2R1B in colon and breast cancers[38-40]. Although at a low frequency, somatic mutations of PPP2R1A have also been identified in a variety of primary human tumors[38, 41, 42]. Biochemical studies showed that these cancer-associated PP2A scaffolding A subunit mutants were defective in binding to other PP2A subunits, including PP2A B' and C subunits, resulting in altered formation of the PP2A trimeric holoenzyme[43, 44]. Taken together, the mutations of PP2A A subunits cause the dysfunction of PP2A by preventing the formation of PP2A holoenzymes, which suggests that PPP2R1A and
! ! 8!
PPP2R1B act as tumor suppressor genes.
The last clue indicating the function of PP2A as a tumor suppressor came with the discovery that PP2A is the target of several viral oncoproteins, such as the small T antigen (ST) of two transforming DNA tumor viruses, simian virus 40 (SV40) and polyomavirus[18]. The ST disrupts PP2A by binding to the A subunit but not the C subunit, resulting in the displacement of specific PP2A B subunits[45-47]. The loss of B subunits further leads to the inhibition of PP2A activity towards certain substrates, resulting in enhanced phosphorylation of these proteins which are involved in the regulation of cell growth or cell proliferation[8]. For instance, the expression of SV40 ST can stimulate cell transformation through inhibiting the dephosphorylation of Akt and c-Myc by PP2A[48, 49]. The activation of Akt and c-Myc by hyperphosphorylation represent frequent alterations observed in human cancer cells[30, 50]. Other tumor virus proteins, including the polyomavirus middle T antigen (PyMT) and the adenovirus E4orf4 protein, also alter PP2A activity, and will be discussed later in Chapter 1.4.
1.3 Structure-function of the PP2A subunits under study 1.3.1 Scaffolding A subunit, PPP2R1A PPP2R1A is the α-isoform of the PP2A scaffolding A subunit. This isoform is ubiquitously expressed in all normal tissues and is typically 10-100 times more abundant than the β-isoform PPP2R1B[12, 51]. It contains 15 repeated HEAT (Huntington-Elongation-A subunit-TOR) motifs, each of which consisting of 39 amino acids[12, 52, 53]. These HEAT motifs generate a curved and hook-like helical structure for PPP2R1A[52, 54, 55], which acts as a structural assembly base to bind the C subunit and the B subunits[20]. The PP2A B and C subunits interact with PPP2R1A at different regions. The HEAT motifs 11-15 form strong hydrogen bonds and hydrophobic interactions with PP2A C subunits, while HEAT motifs 2-7 have loose interaction with different PP2A B subunits and no interaction with C subunit[9, 55-58].
! ! 9!
There are 107 mutations of PPP2R1A reported so far in COSMIC (Catalogue of Somatic Mutations in Cancer). Among these somatic mutations, 60 are located in the subunit B binding region of PPP2R1A, and 22 are within the subunit C binding region. These mutations may cause significant changes to the normal function of PP2A by influencing the interaction between PPP2R1A and PP2A B/C subunits.
1.3.2 Catalytic C subunit, PPP2CA PPP2CA is the α-isoform of the PP2A catalytic C subunit. It shares 97% sequence identity with the β-isoform PPP2CB and is ubiquitously expressed in almost every tissue[20]. PPP2CA consists of seven exons and six introns. Exons 2-6 are involved in substrate binding and catalysis. Exons 1 and 7, however, serve to regulate the interaction between PPP2CA and PP2A A/B subunits[59]. The C-terminal tail of PPP2CA ("T304-PDYF-L309") is uniquely conserved from yeast to human[60]. It harbors extensive post-translational modifications (Leu309 carboxymethylation; Thr304 and Tyr307 phosphorylation)[61-65]. Although as mentioned above in the structure of PPP2R1A, the PP2A B subunits bind to the C subunit through its interaction with the scaffolding A subunit, both biochemical and structural studies have highlighted the importance of the post-translational modifications of the PPP2CA C-terminal tail in regulating the dynamic exchange of the B subunits, which influences the specificity of PP2A holoenzymes[9, 56].
Carboxymethylation of Leu309 is catalyzed by LCMT1 (leucine carboxyl methyltransferase 1)[62, 64] and is reversible through the function of phosphatase methylesterase PME-1[66, 67]. Leu309 methylation is specifically required in recruiting B subunits to PP2A core dimer. However, regulatory subunits from other subfamilies, including B', B'' and B''', have apparently no preference for the methylation at Leu309 or not[68-77]. Phosphorylation of Thr304 and Tyr307 was reported to be catalyzed by several kinases, and the dephosphorylation process can be accomplished by PP2A itself[61]. In contrast to Leu309 methylation, phosphorylation at these two sites was suggested to be associated with PP2A inactivation[60].
! ! 10!
However, the consequences of the phosphorylation events are still not clear. The phosphorylated sites also act in concert with another functional element in PP2A, the β12-β13 loop that includes an arginine at position 268. Expression of PPP2CA R268E in NIH3T3 cells induces a massive increase in cell volume and the formation of large, irregularly shaped nuclei. This phenotype is most likely dependent on the interaction between PPP2CA and B subunits since only the PPP2CA mutations (R268E and R268E/T304A/Y307F) that maintain the interaction with B subunits cause the severe phenotype[78]. All these observations suggest the important roles PPP2CA C-terminal tail and R268 play in keeping the normal function of PP2A holoenzymes.
1.3.3 Regulatory B subunit, PPP2R2A PPP2R2A is the α-isoform of the PP2A B55 regulatory subunits. Structural analysis of the four isoforms in the B55 subfamily revealed that PPP2R2A, PPP2R2B, PPP2R2C, and PPP2R2D only have minor differences in their structures[79]. Therefore, PPP2R2A serves as a good model to study the substrates binding sites in regulatory subunits from PP2A B55 family.
The resolved crystal structure of PPP2R2A shows that it is a seven-bladed β-propeller protein. Each of the blades consists of four anti-parallel β-strands. The three-dimensional structure of the PP2A holoenzyme containing PPP2R2A as the regulatory subunit indicates that the β-hairpin arm on the bottom face of PPP2R2A interacts with PP2A A subunit at the HEAT motifs 2-7, while the catalytic subunit binds to PP2A A at the other end[58, 80]. There is a putative substrate-binding groove on the top face of PPP2R2A[17].
It is generally believed that most binding of the PP2APPP2R2A (this will designate a specific PP2A trimer by the name of its specific B regulatory subunit) substrates appears to be solely with PPP2R2A (by opposition to binding through a A or C interaction)[80]. After binding to PPP2R2A, the substrates will be rapidly dephosphorylated by the catalytic C subunit and then released from the PP2A
! ! 11! holoenzyme, making it difficult to identify substrates for PPP2R2A because of the transient interaction. If there is a mutant of PPP2R2A that is still able to bind to the substrates but not to the PP2A core dimer, the interaction with the substrate could potentially be stabilized enough for it to get trapped, and therefore identifiable in a purification experiment. New PP2A substrates may be identified with the bound trapped substrates to these free PPP2R2A subunits.
1.4 PP2A-related tumor virus proteins 1.4.1 SV40 small T antigen Simian virus 40 (SV40) is the best characterized member of the Polyomaviridae family of small DNA tumor viruses[81]. Studies have reported that infection of SV40 in non-permissive host cells can contribute to cell transformation and tumor formation[82-85]. The SV40 genome is divided into two regions based on the timing of gene expression after SV40 infection; an Early Region (ER) and a Late Region (LR). In cells that are transformed by SV40 infection, only the ER is expressed and the expression of ER is sufficient to induce cell transformation. There are three proteins encoded by SV40 ER; the large T antigen (LT), small T antigen (ST) and 17kT antigen[86, 87]. SV40 LT functions to stimulate cell transformation through the inhibition of two tumor suppressor proteins, p53 and pRB[86-90]. SV40 ST, however, targets PP2A, which is the only cellular protein known to interact with SV40 ST except chaperones [45, 46].
The N-terminal region of SV40 ST shares sequence homology with the J domain of DnaJ from E. coli. The C-terminal region of SV40 ST consists of a Zinc-binding domain that interacts with the HEAT motifs 3-6 in PP2A A subunit, which overlap with the binding sites for PP2A B' subunits. After directly interacting with PP2A A subunit through the Zinc-binding domain, the J domain of SV40 ST mediates interaction with the PP2A C subunit[45]. It was reported that SV40 ST modulates the function of PP2A by two modes. Since the binding site of SV40 ST on the scaffolding A subunit overlaps with the B' subunits, binding of ST to the PP2A core dimer can
! ! 12! prevent the formation of PP2A holoenzymes with B' as the regulatory subunits, which alters the substrate specificity, localization and phosphatase activity of PP2A. In addition, the J domain of ST may directly interact with the substrate-binding region of the C subunit of PP2A, leading to the inhibition of the phosphatase activity of PP2A by competing with substrate for access to the active site[45]. Although SV40 ST affects the function of PP2A in these two ways, little is known about other possible functions of SV40 ST during infection.
1.4.2 Polyomavirus middle T antigen The Early Region of polyoma virus encodes three viral proteins: the small T antigen, middle T antigen (PyMT) and large T antigen, which have similar amino acid sequences. These three T antigens share a common amino-terminal sequence, and the small and middle T antigens share the central region. The carboxyl-terminal region of each antigen, however, is different[91]. In spite of the similar sequence, PyMT is the sole primary transforming protein of polyomavirus: Studies have reported that all of the polyoma virus non-transforming mutations alter the PyMT coding sequence. This demonstration was also confirmed by the observation that only the part of the early region that encoded PyMT was required for transformation[92-97]. This indicates that PyMT is tightly associated with the transforming properties of polyoma virus.
PyMT carries out its transforming function by associating with and modulating the activities of cellular proteins involved in control of cell proliferation[91]. These proteins include c-SRC, c-yes, PI 3-kinase, PP2A, etc.[47, 98-101]. Take c-SRC as an example. When expressed in host cells, PyMT interacts with c-SRC and leads to its dephosphorylation at Tyr52, which increases the tyrosine kinase activity of c-SRC, the cellular version of the v-src oncogene[102-105]. PP2A is another major interactor of PyMT. It has been revealed that the membrane-located PyMT binds to PP2A A and C subunits at its N terminus. There are strong indications that the association between PyMT and PP2A alters the substrate specificity and phosphatase activity of PP2A holoenzyme[91]. Although PyMT has been linked to the biology of PP2A for many
! ! 13! years, little is known about how PyMT regulates the function of PP2A.
1.4.3 Adenovirus E4orf4 protein E4orf4 is the product of human adenovirus early transcription region 4 open reading frame 4. It is a 114-residue protein that has no extensive sequence homology with any known proteins[106, 107]. E4orf4 is a multifunctional viral regulator, which is involved in the down regulation of virally modulated signal transduction, in the control of alternative splicing of late viral mRNAs, and in the induction of apoptosis in transformed cells[108-115]. General interest in E4orf4 emanates from the finding that when expressed in cells at high levels, E4orf4 exhibits tumor cell-specific, p53-independent toxicity in a variety of human cancer cell lines[106, 116-123]. This discovery of the tumor cell-specific killing ability of E4orf4 provides a new possibility for the design of anti-cancer drugs.
The major insight into the mechanism of how E4orf4 promotes cell killing comes from the finding that E4orf4 interacts with the tumor suppressor PP2A. E4orf4 binds to all members of the PP2A B55 subfamily of regulatory subunits but no other B subunits[111, 117]. Toxicity induced by E4orf4 is largely dependent on its ability to associate with the highly conserved PP2A B55 subunits. It is proposed that E4orf4 inhibits PP2A activity by binding to the putative substrate-binding groove of B55 and preventing access of substrates[17]. At high E4orf4 expression levels, this results in cell death through the failure to dephosphorylate substrates required for cell cycle progression[114, 124]. Two classes of E4orf4 mutants that are deficient in the process of cell death induction have been identified so far. Class I E4orf4 mutants fail to kill cancer cells and also loose the binding to B55, suggesting that interactions with PP2A are essential for the toxicity of E4orf4. Class II E4orf4 mutants, however, can still bind B55 but are deficient in cell killing, indicating that although PP2A binding is necessary for the induction of cell killing, it is not sufficient[117].
Although hyperphosphorylation of one of the PP2A substrates, p107 (a member of the
! ! 14! retinoblastoma gene family), has been identified following overexpression of E4orf4[17], which confirms the model for inhibition of PP2A activity by E4orf4, little is known about the influence of E4orf4 on other PP2A substrates. In addition, since the association with PP2A is not sufficient for E4orf4 to induce cell killing, it is necessary to detect other components required for E4orf4 toxicity.
1.5 Affinity purification and mass spectrometry It is becoming increasingly clear that cellular organization implicates the association of proteins into functional units known as "protein complexes"[125]. Proteins generally interact with each other and form complexes in a time- and space-dependent manner, which is critical for cells to carry out and maintain their normal functions. In the case of signaling molecules such as PPP family phosphatases, a protein complex provides functional activity by interacting with its specific substrates. At the same time, the specialized function of a protein complex itself can also be dependent on the post-translational modifications or conformational changes caused by interaction with other neighboring proteins[126-128]. Because of the important roles protein complexes and protein-protein interaction networks play in maintaining the function and organization of the living cells, their analysis is of great importance in biological research.
Different approaches have been applied to characterize protein complexes and, more generally, protein-protein interaction networks. Affinity purification coupled with mass spectrometry (AP-MS) has recently been regarded as an indispensible tool in proteomic analysis. Compared to other methods, AP-MS possesses several advantages. First, in contrast to yeast-two-hybrid and related methods, AP-MS can be performed in a near physiological context, especially when expression levels of the protein of interest is controlled such that it is similar to the level of the endogenous protein [129, 130]. Second, AP-MS also allows the detection of post-translational modifications, including phosphorylation, methylation, ubiquitinylation, etc. In addition, AP-MS does not necessitate prior knowledge of the interacting proteins (though it can only
! ! 15! capture proteins expressed in the cell analyzed and detects interactions occurring only under tested conditions). Although less sensitive and rapid than other approaches such as Y2H[131], LUMIER[132], etc., only femtomole levels of peptides are required for identification using AP-MS, and sequencing of single peptide can be accomplished within hundreds of milliseconds. Based on these advantages, AP-MS has increasingly become the choice for the discovery of biologically relevant protein-protein interactions[130].
The generic workflow (Fig. 2) of AP-MS most often begins with tagging of the protein of interest with an epitope that can be used for purification, which can be achieved either by transfecting a tagged gene into cells or by targeting the endogenous gene using homologous recombination or other strategies[133]. The recombinant protein expression system varies (for example, I use here a tetracycline-inducible promoter), but expression levels should preferably be at near endogenous level in order to prevent the drawbacks resulting from overexpression. It is more likely for the overexpressed proteins to get misfolded, which will result in their association with molecular chaperones. The overexpressed proteins are also often mislocalized and dysregulated, inducing abnormal interactions or post-translational modifications. Moreover, the balance between the overexpressed proteins and their cellular binding partners are disrupted, making it more difficult to identify the bona fide protein-protein interactions[129]. Once cells expressing the recombinant proteins of interest are generated and treated as desired (e.g. with a growth factor or an inhibitor), the tagged protein is purified from the cell lysate together with its binding partners. The extracted proteins, including the bait itself, its interactors as well as the contaminants, are then degraded enzymatically into peptides, usually by trypsin, producing peptides with C-terminally protonated amino acids, which provides an advantage in subsequent peptide sequencing[133]. The prepared sample will then be analyzed by MS. In general, the MS analysis involves two steps, peptide separation by reversed-phase liquid chromatography followed by tandem MS (MS/MS). In the first MS scan, the mass/charge ratio (m/z) of the intact peptide is measured. The n most
! ! 16!
Figure 2. Workflow of affinity purification coupled with mass spectrometry (AP-MS). (a) A protein of interest (bait) is fused with a tag and expressed in cells. Affinity purification is performed using beads with antibody specifically targeting the tag. The protein, together with its binding partners and some unspecific interacting proteins, is pulled down and digested into peptides enzymatically. The peptide mixture is then analyzed by liquid chromatography (reverse-phase HPLC) and mass spectrometry. (b) Peptides are first separated by LC, ionized by electrospray and then enter MS. In tandem MS (MS/MS) analysis, the MS1 survey scan of the masses of individual precursor ions is performed firstly. A precursor ion (in red) is selected for fragmentation based on its abundance in MS1. The fragmented ions are detected and analyzed later to generate a MS/MS spectrum. This spectrum can be used to search the database in order to identify the peptides and finally characterize the proteins.
!
! ! 17!
! ! a Bait
tag
Bait tag LC-MS/MS
Bead Bait tag ! Bait tag LC-MS/MS ! ! Bead !
! LC MS1 MS/MS ! b ! LC MS1 MS/MS