@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Pharmaceutical Sciences, Faculty of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Thomas, Dylan"@en ; dcterms:issued "2013-12-31T08:00:00Z"@en, "2013"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """Protein arginine N-methyltransferases (PRMTs) constitute a family of post-translational modifying enzymes that modulate protein-protein interactions via the addition of methyl groups to arginine residues in protein substrates (1). PRMTs have been demonstrated to homooligomerize via a dimerization arm that binds with the outer surface of the S-adenosyl-Lmethionine (AdoMet) binding domain (2-5). In this body of work, I have demonstrated and quantified in vitro the strength of homodimerization for PRMT1 and PRMT6 and demonstrated that saturating concentrations of S-adenosyl-L-methionine (AdoMet) or S-adenosyl-Lhomocysteine (AdoHcy) respectively strengthen or weaken this interaction. This finding supports an ordered bisubstrate mechanism in which AdoMet binding promotes formation of the complete peptide-substrate binding groove through dimerization, and AdoHcy generation promotes dissociation of the dimeric complex and turnover of substrate. A kinetic study using HIV Tat peptides revealed oligomerization-dependent kinetic patterns with these substrates. Kinetic experiments were initially performed on HIV Tat peptide with novel ωN-substitutions to probe their ability to inhibit PRMT1, 4 and 6. It was found that these Tat-peptides act as substrate inhibitors for both PRMT1 and PRMT6 and that this substrate inhibition was mitigated as the enzyme concentration increased. A model was proposed that represents activity as the sum of each ordered oligomer in solution, with the monomer being uniquely susceptible to substrate inhibition. Diverging from strictly oligomerization effects, R1 fibrillarin-like peptide containing a single arginine was substituted to alter the pKa of the terminal guanidino group to better probe the physicochemical properties that control methyltransfer. Surprisingly, hydroxyl substituted R1 peptide demonstrated an enhanced catalytic constant with PRMT1. MS and MS² experiments demonstrate that only monomethylation occurs on substituted arginines with PRMT1, and that this addition is asymmetric. PRMT1 D51N, a catalytically compromised mutant, revealed the kcat as rate limiting in the presence of D₂O, and electrostatic potential maps indicate that deprotonation of hydroxyl substituted arginine produces a strong nucleophile capable of enhanced methyltransfer. Altogether, these studies support water mediated, ordered bisubstrate mechanism in which oligomerization modulates activity. Substrate inhibition and active site chemistry were investigated using novel chemically substituted peptide probes that highlight trends beyond what site-directed mutagenesis can reveal alone."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/44623?expand=metadata"@en ; skos:note "OLIGOMERIZATION DEPENDENT ENZYME KINETICS AND MECHANISTIC CHARACTERIZATION OF TYPE I PROTEIN ARGININE N- METHYLTRANSFERASES by Dylan Thomas B.M.L.Sc., The University of British Columbia, 2006 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Pharmaceutical Sciences) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) June 2013  Dylan Thomas, 2013 ii Abstract Protein arginine N-methyltransferases (PRMTs) constitute a family of post-translational modifying enzymes that modulate protein-protein interactions via the addition of methyl groups to arginine residues in protein substrates (1). PRMTs have been demonstrated to homo- oligomerize via a dimerization arm that binds with the outer surface of the S-adenosyl-L- methionine (AdoMet) binding domain (2-5). In this body of work, I have demonstrated and quantified in vitro the strength of homodimerization for PRMT1 and PRMT6 and demonstrated that saturating concentrations of S-adenosyl-L-methionine (AdoMet) or S-adenosyl-L- homocysteine (AdoHcy) respectively strengthen or weaken this interaction. This finding supports an ordered bisubstrate mechanism in which AdoMet binding promotes formation of the complete peptide-substrate binding groove through dimerization, and AdoHcy generation promotes dissociation of the dimeric complex and turnover of substrate. A kinetic study using HIV Tat peptides revealed oligomerization-dependent kinetic patterns with these substrates. Kinetic experiments were initially performed on HIV Tat peptide with novel ωN-substitutions to probe their ability to inhibit PRMT1, 4 and 6. It was found that these Tat-peptides act as substrate inhibitors for both PRMT1 and PRMT6 and that this substrate inhibition was mitigated as the enzyme concentration increased. A model was proposed that represents activity as the sum of each ordered oligomer in solution, with the monomer being uniquely susceptible to substrate inhibition. Diverging from strictly oligomerization effects, R1 fibrillarin-like peptide containing a single arginine was substituted to alter the pKa of the terminal guanidino group to better probe the physicochemical properties that control methyltransfer. Surprisingly, hydroxyl substituted R1 peptide demonstrated an enhanced catalytic constant with PRMT1. MS and MS 2 experiments iii demonstrate that only monomethylation occurs on substituted arginines with PRMT1, and that this addition is asymmetric. PRMT1 D51N, a catalytically compromised mutant, revealed the kcat as rate limiting in the presence of D2O, and electrostatic potential maps indicate that deprotonation of hydroxyl substituted arginine produces a strong nucleophile capable of enhanced methyltransfer. Altogether, these studies support water mediated, ordered bisubstrate mechanism in which oligomerization modulates activity. Substrate inhibition and active site chemistry were investigated using novel chemically substituted peptide probes that highlight trends beyond what site-directed mutagenesis can reveal alone. iv Preface This dissertation contains work completed wholly in Dr. Adam Frankel‟s laboratory and includes the combined efforts of several lab members, including Dr. Ted M. Lakowski, Dr. Magnolia L. Pak, Mynol Islam Vhuiyan, Jenny J. Kim, and myself. This work would not have been possible without the generous collaboration of Nathaniel M. Martin and colleagues. In Chapter 2, I performed all fluorescence based assays and optical assays. In Chapter 3, I completed all mass spectrometric assays utilized in quantifying Tat peptide methylation; I modeled all kinetic parameters and proposed all novel mathematical models listed. In Chapter 4, I performed IC50 experiments utilizing PRMT4 and PRMT6, I performed all kinetic isotope effect experiments, modeled all kinetic parameters except for IC50 values with PRMT1, and performed all gel based methylation assays. Dr. Ted M. Lakowski performed kinetic analysis and product inhibition studies in Chapter 2. Dr. Lakowski also synthesized and quantified H3 and H4 peptide used in kinetic experiments. In Chapter 3, Dr. Lakowski separated methylated substrates by gel electrophoresis from reactions containing PRMT1, PRMT4 or PRMT6 and Tat peptides, and performed densitometry on those methylated Tat peptides. He also quantified stock solutions of all Tat peptides utilizing multiple reaction monitoring. In Chapter 4, Dr. Lakowski performed MS 2 experiments to locate asymmetric monomethylation on amino substituted R1 peptide. HIV-1 Tat peptides used in Chapter 3 were synthesized and provided by Peter t‟Hart, Randy Van Ommeren and Dr. Nathaniel I. Martin. Fibrillarin-like R1 substituted peptides utilized in Chapter 4 were synthesized and provided by Timo Koopmans, Helmi Kreinin and Dr. Nathaniel I. Martin. Magnolia L. Pak expressed and purified mCit-PRMT1 and mCer-PRMT1 for Chapter 2. Mynol Islam Vhuiyan designed primers for and prepared plasmid DNA of the D51N PRMT1 variant in chapter 4. Jenny J. Kim helped subclone several fluorescently tagged PRMTs used in Chapter 2. Dr. Jennifer Bui performed Gaussian modeling of electrostatic potential for substituted R1-peptides in Chapter 4. Chapter 2 is reproduced with permission from Thomas, D., Lakowski, T. M., Pak, M. L., Kim, J. J., and Frankel, A. (2010) Forster resonance energy transfer measurements of cofactor- dependent effects on protein arginine N-methyltransferase homodimerization, Protein Sci 19, 2141-2151. Copyright  2010 The Protein Society, Wiley-Blackwell. Chapter 3 is reproduced with permission from Peter 't Hart*, Dylan Thomas*, Randy van Ommeren , Ted M. Lakowski , Adam Frankel and Nathaniel I. Martin. (2012) Analogues of the HIV-Tat peptide containing Nη-modified arginines as potent inhibitors of protein arginine N- methyltransferases. Med. Chem. Commun., 3, 1235-1244. Copyright  2011 Royal Society of Chemistry. Please see footnotes in the first pages of these chapters for similar information. v Table of contents Abstract ........................................................................................................................................... ii Preface............................................................................................................................................ iv Table of contents ............................................................................................................................. v List of tables ................................................................................................................................... vi List of figures ................................................................................................................................ vii List of abbreviations .................................................................................................................... viii Acknowledgements ........................................................................................................................ ix 1 Introduction ............................................................................................................................. 1 1.1 Protein arginine N-methyltransferases structure and activity ......................................... 1 1.2 Mechanisms of arginine methylation ............................................................................ 11 1.3 Sequence-based control of PRMT activity and common peptide substrates. ............... 16 1.4 Hypothesis and research objectives .............................................................................. 20 2 Förster resonance energy transfer measurements of cofactor-dependant effects on protein arginine N-methyltransferase homodimerization , ......................................................................... 21 2.1 Introduction ................................................................................................................... 21 2.2 Methods......................................................................................................................... 23 2.3 Results ........................................................................................................................... 28 2.4 Discussion ..................................................................................................................... 40 3 Analogues of the HIV-Tat peptide containing Nη-modified arginines as potent inhibitors of protein arginine N-methyltransferases .......................................................................................... 45 3.1 Introduction ................................................................................................................... 45 3.2 Methods......................................................................................................................... 47 3.3 Results ........................................................................................................................... 50 3.4 Discussion ..................................................................................................................... 60 4 N-substituted arginyl peptide molecular probes reveal an activated substrate for protein arginine N-methyltransferase activity ........................................................................................... 66 4.1 Introduction ................................................................................................................... 66 4.2 Methods......................................................................................................................... 68 4.3 Results ........................................................................................................................... 71 4.4 Discussion ..................................................................................................................... 83 5 Conclusions ........................................................................................................................... 92 References ..................................................................................................................................... 98 Appendix ..................................................................................................................................... 105 Appendix figures ..................................................................................................................... 105 vi List of tables Table 2.1. Apparent kinetic parameters for fluorescent and non-fluorescent PRMTs. ............... 31 Table 2.2. Dissociation constants for fluorescent PRMT1 and PRMT6 with and without cofactor. ............................................................................................................................................... 38 Table 3.1. Apparent kinetic constants for PRMT1 and PRMT6 using a substrate inhibition model..................................................................................................................................... 55 Table 4.1. Physicochemical properties of arginine and its derivatives. ....................................... 72 Table 4.2. Apparent kinetic constants for PRMT1, PRMT6 and PRMT1 D51N using R1- based peptides. ................................................................................................................................ 79 vii List of figures Figure 1.1. Activity classifications and conserved sequences in PRMTs. ..................................... 2 Figure 1.2. Major domains and a selection of conserved active site residues for PRMT1, 4 and 6. ................................................................................................................................................. 4 Figure 1.3. Conserved residues within the hPRMT1 active site. ................................................... 6 Figure 1.4. Contacts between the dimerization arm and αXY-helix shield AdoHcy from solvent. ................................................................................................................................................. 7 Figure 1.5. Cleland notation of bisubstrate mechanisms. ............................................................ 12 Figure 2.1. Formation of PRMT FRET pairs. .............................................................................. 29 Figure 2.2. Normalized absorption and emission spectra for mCerulean and mCitrine-PRMTs. 30 Figure 2.3. Activities of PRMT1 and PRMT6 with and without mCerulean or mCitrine and double-reciprocal plot of PRMT1 product inhibitor analysis. .............................................. 31 Figure 2.4. mCerulean and mCitrine-PRMTs produce FRET. .................................................... 36 Figure 2.5. Steady-state FRET binding for fluorescent PRMTs. ................................................. 38 Figure 2.6. FRET efficiency subunit contribution. ...................................................................... 39 Figure 2.7. Non-fluorescent PRMT competition with FRET pairs. ............................................ 39 Figure 3.1. Chemical substitutions on the terminal omega nitrogen of arginine. ........................ 51 Figure 3.2. Methylation of modified Tat peptides. ...................................................................... 53 Figure 3.3. Tat peptide methylation diverges from a hyperbolic fit and reveals substrate inhibition. .............................................................................................................................. 56 Figure 3.4. PRMT concentration affects substrate inhibition. ..................................................... 59 Figure 3.5. Substrate inhibition mechanisms for a sequential ordered bi-substrate reaction. ..... 63 Figure 4.1. ωN-substituted R1 peptides. ...................................................................................... 71 Figure 4.2. Coomassie staining and phosphor imaging of radiolabeled R1-peptides. ................. 73 Figure 4.3. Interogation of monomethylated R1-NH2 and R1-OH,using MS. ............................ 76 Figure 4.4. Michaelis-Menten curves for PRMT1, PRMT4 and PRMT6 in H2O and D2O. ........ 78 Figure 4.5. The 4-parameter sigmoidal fits for PRMT1, PRMT4 and PRMT6........................... 82 Figure 4.6. Electrostatic potential surrounding deprotonated Nε-substituted arginine fragments. ............................................................................................................................................... 86 Figure 4.7. Superimposition of PRMTs with water and arginine. ............................................... 89 Figure 4.8. A proposed water-based mechanism for activation of arginine in PRMT1. ............. 91 Figure A.1. Equilibration of FRET signal for PRMT1 and PRMT6. ........................................ 105 Figure A.2. Initial velocity of PRMT6 E58K with varying AdoMet. ....................................... 106 viii List of abbreviations aDMA ω-NG,NG-dimethylarginine AdoHcy S-adenosyl-L-homocysteine AdoMet S-adenosyl-L-methionine CARM1 Coactivator-associated arginine methyltransferase 1 HIV Human immunodeficiency virus hnRNPA1 Heterogeneous ribonucleoprotein A1 LBD Ligand-binding domain LC-MS/MS Liquid chromatography tandem mass spectrometry MLL Mixed lineage leukemia MMA ω-NG-monomethylarginine NMWCO Nominal molecular weight cut-off PABPN1 Nuclear poly(A) binding protein 1 PRMT1-9 Protein arginine N-methyltransferase 1-9 Sam68 Src substrate associated in mitosis (68kDa) sDMA ω-NG,NG-dimethylarginine SH3 SRC homology domain 3 TAR Transactivation response region Tat Transactivator of transcription TDRD3 Tudor domain containing protein 3 TPR Tetratricopeptide repeat TOF Time of flight ix Acknowledgements Firstly, I would like to acknowledge and thank my supervisor, Dr. Adam Frankel, for his continuous support and allowing me to conduct this research in his laboratory. Without his insightful contributions and genuine care for his students, this work and dissertation would not have come to be. On countless occasions he has helped me solve both professional and personal problems, and for that I am very grateful. Additionally, I would like to thank my committee members, Dr. Leann Howe, Dr. Judy Wong, Dr. David Grierson and Dr. Peter Soja, all of whom have provided ideas and solved several technical problems pertaining to my work. My committee both challenged and worked with me to help so I could obtain my professional goals. I would also like to express my gratitude to Dr. Ted M. Lakowski who not only trained me in nearly all techniques I perform in the laboratory, but provided knowledge and insight about the academic process and general philosophy on life, which has helped guide me through my PhD program. I also extend thanks to my colleagues in the laboratory, Dr. Magnolia L. Pak, Mynol I. Vhuiyan, and Jenny J. Kim all of whom contributed to the work found in this dissertation and assisted me in overcoming technical hurdles. I would also like to acknowledge my mom (Diana Lepine-Thomas), dad (David L. Thomas), and brothers (Phillip Thomas and Quinn Boeker-Thomas) for providing essentially boundless emotional support and guidance over the last six years. Without their help, I would not have completed this work. Last (but certainly not least) I would like to thank James Roland Cummins for providing emotional support, advice and helping me keep my life in perspective as I delved deeper into the rabbit hole. 1 1 Introduction 1.1 Protein arginine N-methyltransferases structure and activity Post-translational modifications provide an added level of control over the dizzying array of cellular processes maintained in equilibrium that regulate the function of mammalian cells. These modifications add additional layers of chemical diversity, modifying the intermolecular forces that modulate interactions between proteins and other biochemicals. Although several different types of post-translational modifications exist (6), this introduction will focus on arginine methylation. Arginine methylation replaces hydrogens on the terminal omega nitrogens of the arginine guanidino with methyl groups, reducing the number of available hydrogen bond donors without significantly altering sterics or electrostatics (7). Overexpression or abberent splicing of PRMTs has been implicated in breast (8), prostate (9), colorectal (10), lung (11), bladder cancer and leukemia in addition to other pathologies, such as hypertension (12) and heart disease (13). Thus, understanding how PRMT function is regulated will lead to improved design and discovery of potential inhibitors for these enzymes. Methylation of arginine is catalyzed by the Protein Arginine N-Methyltransferase (PRMT) family of enzymes in humans. To date this family is comprised of nine members for which eight have characterized activity (14). Methylation of arginine utilizes the cellular methyl donor, S-Adenosyl-L-Methionine (AdoMet) per catalytic cycle, transferring a single methyl group onto arginine to produce MMA and S-Adenosyl-L-Homocysteine (AdoHcy). After the initial methylation, a second methylation may occur either asymmetrically or symmetrically. These two types of activity are classified as type I and type II, respectively (Figure 1.1). The presence of only monomethylation has been putatively classified as type III (Figure 1.1). PRMT5 is currently the only type II human PRMT, and PRMT7 is the only type III (15, 16). All 2 remaining PRMTs are classified as type I, producing monomethylarginine and asymmetric dimethylarginine with the exception of PRMT9, for which no activity has been characterized (17-21). Figure 1.1. Activity classifications and conserved sequences in PRMTs. Type I, type II and type III activities that generate ADMA, SDMA or MMA only, respectively, are displayed. A diagram representative of the relative lengths of the 9 PRMT family members are aligned to compare the position of conserved sequence motifs involved in cofactor binding and catalysis. PRMT2, PRMT3, and PRMT8 have N- terminal SH3, zinc-finger and myristolization sites shown in the above figure. PRMT9 contains double N-terminal TPR repeat and C-terminal zinc-finger domains. This figure is adapted from Bedford‟s arginine methylation review (14). 3 Conserved structural features and sequence motifs of PRMT enzymes. Crystal structures for PRMT1, PRMT3, PRMT4 or coactivator-associated arginine methyltranferase I (CARM1), PRMT5 and more recently PRMT6 reveal a number of conserved sequence motifs, in addition to a general conserved structure consisting of an N-terminal helix bundle, followed by the AdoMet binding domain, a beta barrel domain and finally a dimerization arm (Figure 1.2) (2- 5, 22-24). Although the Rossmann fold based cofactor pocket is common to many methyltransferases (25), the beta barrel domain is unique to PRMTs. PRMTs have been demonstrated to form homo-oligomers and in some cases hetero-oligomers via x-ray crystallography, dynamic light scattering, bimolecular fluorescence complementation (BiFC), and Co-immunoprecipitation (IP) (2, 4, 20, 26-28). Homodimerization has been shown to occur through hydrophobic interactions with the dimerization arm and outer surface of the AdoMet binding domain with and without cofactor. This forms a complex with C2-symmetry and an inner anionic channel ideal for electrostatically holding positively charged arginines adjacent to the cofactor-binding pocket. Crystal structures of PRMT1 illustrate the presence of several peptide binding grooves that circumnavigate the beta barrel domain, with a final groove inserting within the anionic channel to position the arginine containing protein substrate next to the cofactor binding site (2). Crystal structures reveal that yeast HMT1 forms hexamers while human PRMT1 forms homodimers (5). However, dynamic light scattering has demonstrated that human PRMT1 is capable of forming much higher molecular weight structures than homodimers (2). 4 Figure 1.2. Major domains and a selection of conserved active site residues for PRMT1, 4 and 6. Panel A displays an overlay of PRMT1 (red), PRMT4 (green) and PRMT6 (blue) highlighting the structural conservation among type I PRMTs. Residues in panel B represent numbering for PRMT1, and the αXY-helix has been removed to better visualize the active site residues within. Structures were compiled in UCSF Chimera 1.6.2. Alignments of primary PRMT sequences reveal segments involved in substrate stabilization or catalysis. Standard classification of these sequence motifs is divided into motif I (VLD/EVGxGxG), post motif I (V/IxG/AxD/E), motif II (F/I/VDI/L/K), III (LR/KxxG) and the THW loop(14). All motifs are found within the AdoMet binding domain with the exception of 5 the THW loop, which originates in the C-terminal beta barrel domain and folds into the active site. The threonine of the THW loop is purported to form Van der Waal interactions with AdoMet, however, it lies in tandem with a conserved aspartic acid present on an N-terminal αX- helix that may function in a His-Asp proton relay involved in activation of arginine, priming it to receive a methyl group (3). The remainder of the motifs operate via hydrophobic and Van Der Waals interactions, conforming the cofactor pocket to promote AdoMet binding. Key residues regulate cofactor stabilization in type I enzymes. Conserved residues and sequence motifs have been implicated in the stabilization of overall tertiary and quaternary structure of PRMTs, as well as stabilization of the apo- and holoenzyme. The PRMT1 crystal structure has revealed three main stabilizing factors holding AdoHcy within the cofactor binding pocket: 1) A glycine-rich loop (G78 and G80) that stabilizes the homocysteine portion of AdoHcy via Van der Waals forces, 2) E100 that forms hydrogen bonds stabilizing the position of the ribose component of AdoHcy, and 3) E129 that forms hydrogen bonds with the amino component of adenine (2) (Figure 1.3). Related to these observations, the crystal structure of the conserved PRMT3 core (residues 208-528) stabilizes AdoHcy through similar interactions with different residues. Again, amino acids interact with the 3 major components of AdoHcy comprised of the homocysteine, adenine and the ribose component. R236 and D258 stabilize the amino group of methionine of AdoHcy, D282 hydrogen bonds with dual hydroxyls on the ribose sugar, and E311, the main chain nitrogen of I310 and a water molecule hydrogen bond with the nitrogens of adenine (3). Similar stabilizing factors are present with PRMT4 using a triple strategy of interacting with the three major components of AdoHcy. R169 present on the αZ- helix interacts with the carboxyl of methionine in AdoHcy, E215 hydrogen bonds with ribose 6 hydroxyls and E244, S272 and the main chain nitrogen of V243 interact with N6, N1 and N7 adenine nitrogens respectively (4). Figure 1.3. Conserved residues within the hPRMT1 active site. Conserved residues involved in stabilizing AdoHcy (R54, G78, G80, E100, E129, E144) formation of the active site (D51, R54, H293) or in some cases, specifying PRMT type (M48, M155) are shown above, with the relative positions of the THW-loop and αYZ-helix turn. Blue lines represent theoreticl hydrogen bonding sites within 2 angstroms in length. Coordinates are adapted from PDB 1ORI. The N-terminal helix bundle (αX-, αY- and αZ-helices) play a major role in both stabilizing AdoHcy (via PRMT3 R236 and PRMT4 R169) and forming a pivotal peptide binding groove within central PRMT-homodimer anionic channel (Figure 1.2A and Figure 1.4) (3, 29). This helix bundle has since been shown to be present in not only PRMT3 and 4 structures, but PRMT6 (PDB 4HC4) as well. Structures of PRMTs with AdoHcy display the αXY-helix as ordered, sitting over top of the cofactor binding pocket restricting free diffusion of AdoHcy. Interestingly, apoenzyme structures reveal a more disordered orientation for this helix bundle, supporting an induced fit model in which cofactor binding triggers encapsulation by these N- 7 terminal helices, forming the necessary peptide-binding groove to properly position protein containing arginines within the active site (Figure 1.4). Figure 1.4. Contacts between the dimerization arm and αXY-helix shield AdoHcy from solvent. A homodimer of PRMT6 is shown with magnification into the AdoMet binding pocket. Within the magnified area, the surface map represents the dimer arm of the opposite monomer (A) within the homodimeric complex, and makes contacts with the αYZ-helix (B and C respectively) to help stabilize the methyldepleted cofactor, AdoHcy (D). Coordinates are adapted from PDB 4HC4. Key residues stabilize methyltransfer and alter the generation of asymmetric or symmetric dimethylarginine. As mentioned earlier, several conserved amino acids stabilize the cofactor, however, there are also several key residues that are postulated to interact with arginine, making it a better nucleophile for methyltransfer. Residue numbers for PRMT1 will be used for the purpose of discussion although these residues are present in other PRMTs. Within the AdoMet binding domain there is a double-E hairpin loop containing two conserved glutamic acid residues. E144Q and E153Q in PRMT1 have both been shown via site-directed mutagenesis to be inactive (2, 26, 30). E144 interacts with R54, which in turn forms a salt bridge 8 with the carboxy terminus of methionine within AdoMet, stabilizing the cofactor position. However, it is also noteworthy that E144 is in proper orientation within the active site to hydrogen bond with a single water molecule that bridges the space between E144 and H293 of the THW loop. H293 has been proposed to operate in a proton relay that may deprotonate a single water molecule, which subsequently deprotonates the guanidino group of arginine. E144D and E144Q mutations both compromise activity and change PRMT1 oligomer size (2), suggesting that E144 also has a role in controlling PRMT1 structural changes. E153 is also important and the E153 to Q replacement has been demonstrated to completely abolish all activity without modifying PRMT1‟s oligomerization state. E153 hydrogen bonds with both the omega and delta nitrogens of the guanidino group, stabilizing the position of arginine. E153 is also hypothesized to draw electronegativity through the guanidino group, making the methyl accepting nitrogen a better nucleophile. M155 and M48 are both key conserved residues within the AdoMet binding domain and their function has been investigated. M155A and M48L replacements have both been demonstrated to compromise activity (31) (Figure 1.3). It had also been suggested that M155 may control the generation of asymmetric or symmetric dimethylarginine since its R-group points within the active site, sterically hindering the possible formation of SDMA. However, the M155A variant, which allows enough space for PRMT1 to generate SDMA, fail to produce any (31). M48 is in position to interact via Van der Waals forces with AdoMet, and stabilizes the position of the cofactor. M48L and M48A variants both compromise activity implicating it in proper enzyme function (31). Additionally, the M48L mutation was found to prefer N-terminal arginines when methylating R3 peptide (ac-GGRGGFGGRGGFGGRGGFGG) and produce largely MMA as opposed to wild type PRMT1, which produced mostly ADMA. Curiously, 9 F379 in C.elegans PRMT5, which is in a similar structural position to M48 for PRMT1, has been shown to partially modulate the formation of SDMA and ADMA (24). The F379M substitution was shown to not only enhance PRMT5‟s activity, but allow the enzyme to produce a combination of both ADMA and SDMA. Sequence alignments between cePRMT5 and hPRMT1 reveal that F379 and M48 are present in the same location within the αY-helix, further implicating this N-terminal αXY-helix position in proper enzyme activity. The THW loop is a highly conserved structure but its function is a point of much deliberation. The THW loop is found on the C-terminal region of PRMT1 and is present within the beta barrel domain. Bridging the gap between β10 and β11 strands, the THW loop bends inward towards the active site and has been proposed to form part of a His-Asp relay (Figure 1.2B and 1.3, D51 and H293)(4). H293A mutation in PRMT1 has compromised activity and D2O solvent isotope effect experiments have identified catalysis as the rate limiting step for this variant (30). It was proposed that the H293A mutation may be disrupting the overall structure of PRMT1, however, the active site must be partially intact as some activity is still present in the variant. Related to the THW loop, D51, the aspartic acid involved in this proposed His-Asp relay, has been mutated to arginine and been shown to be inactive (32). PRMT oligomerization and its effect on activity. Prior analysis of PRMT oligomerization and its effect on activity has been largely based upon deletion of the dimerization arm (2, 5). Crystal structures of human PRMT1 and yeast HMT1 both demonstrated that upon removal or mutation of the dimerization arm, monomers could no longer associate and enzyme activity was lost. Although these experiments were a good starting point, it‟s unknown whether complete deletion of the dimerization arm triggered misfolding of other domains. In addition to compromising activity, the D51R replacement in PRMT1 also reduced 10 homodimerization (32). In 2007, Higashimoto et al performed a study showing that Serine-229 phosphorylation negatively regulated CARM1 activity. An isoelectric change, S229E mimicking phosphorylation was created, and this variant had decreased activity, compromised AdoMet binding, reduced estrogen receptor transactivation and importantly, compromised oligomerization (33). S229, located on the outer face of the AdoMet binding domain, contributes to form the surface upon which the dimerization arm of the sister PRMT binds to form homodimer. The presence of a large, charged functional group adjacent to the hydrophobic, dimerization arm binding surface likely destabilizes this interaction. Compromised AdoMet binding also suggested that dimerization may be required to fully stabilize the cofactor within the AdoMet binding pocket(33). Feng et al, while studying transient kinetics of PRMT1, observed that as PRMT1 concentration was increased there was a concomitant increase in kcat, as well as the appearance of oligomers via PRMT1 cross-linking and subsequent polyacrylamide gel electrophoresis (34). PRMT1 formed homodimers and higher order oligomers as its concentration was increased. Soon afterwards, Pak et al. reported that increased PRMT2 concentration in the presence of PRMT1 synergistically enhanced the kcat of the complex similar to higher concentrations of PRMT1, implying that the formation of hetero- and homo-oligomers may enhance activity. PRMT1 was shown to be the catalytic unit in this complex by repeating the experiment with inactive variants PRMT1 E153Q and PRMT2 E220Q (26). Together, these studies suggest that PRMT oligomerization modulates enzyme activity. 11 1.2 Mechanisms of arginine methylation For type I PRMTs, Two prominent views exist within the field and the pertinent studies for both will be presented here. There are several studies demonstrating that type I PRMTs operate through a distributive, ordered bisubstrate mechanism in which AdoMet binds first, followed by the protein substrate. Following methyltransfer, the methylated peptide dissociates first and the cofactor product of methyl transfer, AdoHcy, dissociates from the complex last (35, 36) (Figure 1.5). Multiple reaction monitoring of monomethyl and asymmetric dimethylarginine generation has shown that the Km for monomethylarginine is lower than that of unmethylated arginine containing peptide, providing evidence for a system in which methylation is apparently processive as the enzyme will preferentially methylate monomethylated peptide when selecting from a homogenous, free diffusion system (35). Alternatively, a partially processive, random bisubstrate mechanism has also been proposed as a combination of work from one group analyzing PRMT1 and PRMT6 (37, 38) (Figure 1.5). Biochemical evidence for random and ordered sequential bisubstrate mechanisms. The method of determining the mechanism of bisubstrate reactions involves performing product inhibition studies in which kinetic progress curves are performed in the presence of different concentrations of each product produced in the reaction, referred to as product inhibition (39). The results of these experiments are patterns of intersecting lines on a Lineweaver-Burke plot unique to the different types of mechanisms possible for the enzyme. In an ordered bisubstrate system, all lines should intersect in the upper left quadrant of the Lineweaver-Burke plot, which is diagnostic for mixed inhibition with the exception of when product Q (the last product to dissociate) is varied in the presence of substrate A (the first substrate to bind). In this particular case, all curves will intersect on the x-axis, indicating competitive inhibition (39). This single 12 case produces competitive inhibition because product Q is a modified form of substrate A, and thus can bind and compete for initial docking in the enzyme. Alternatively, a random bisubstrate system will show compentitive patterns with products provided that binding is in rapid equilibrium and catalysis is rate-limiting (Figure 1.5). More complex patterns will be obtained if dead-end complexes form. Figure 1.5. Cleland notation of bisubstrate mechanisms. Panel A illustrates an ordered bisubstrate mechanism in which AdoMet (A) binds first and peptide (B) binds second followed by methylated peptide dissociation (P) and finally AdoHcy dissociation (Q). Panel B displays a random bisubstrate system in which AdoMet (A) or peptide (B) can bind in any order and methylated peptide (P) and AdoHcy (Q) can dissociate in any order. Obiayno et al produced evidence of a partially processive mechanism in which methyltransfer occurs in a random bisubstrate sequence. Using radioisotopic labeling with [methyl- 14 C]-AdoMet and subsequent densitometric analysis of the radioactive substrates, product inhibition Lineweaver-Burke plots utilizing PRMT6 produced a pattern of intersecting lines that crossed on the x-axis, consistent with competitive inhibition (37, 38). Partial processivity has been proposed as a result of a recent study in which MMA and ADMA formation was shown to vary with the type of substrate incubated with PRMT1(40). In a processive mechanism, repeated rounds of catalysis occur without full dissociation of one or more substrates from the enzyme active site. Alternatively, in a distributive mechanism, all 13 substrates fully dissociate from the enzyme before repeated rounds of catalysis may occur (39). Human ribonuclear protein A1 and K (hnRNPA1/K), Src substrate associated in mitosis (Sam68) and fibrillarin-like peptides were analyzed with varying position and numbers of arginines to both assess whether preferential arginines were methylated within the sequence. Double turnover studies were also performed to determine the level of processivity present for PRMT1 with a variety of peptide substrates. PRMT1 was prepared at 20 µM with 40 µM AdoMet and saturating peptide substrate. The rationale of this experiment was to provide a ratio of 2:1 cofactor to enzyme to restrict the system to two catalytic cycles. Varying peptide substrates produced different MMA:ADMA ratios thus the system was assigned as “partially processive” in which the propensity to produce ADMA and MMA varied depending on the substrate being methylated (40). Partial processivity implies that the PRMT can alternate between distributive and processive mechanisms dependant upon substrate sequence. Our group and others have generated data that support an ordered bisubstrate mechanism for PRMT1, PRMT3 and PRMT6 in which methylation is distributive (35, 36). Multiple reaction monitoring, in which unique fragments of methylated arginine are counted as they are isolated by mass, was performed to directly quantify MMA and ADMA formation using a fibrillarin-like R1 peptide with a single arginine with or without monomethylation (35). Product inhibition studies were performed with these two substrates and the intersecting pattern of the Lineweaver-Burke plot demonstrated mixed inhibition in all cases except for when AdoMet (A) was varied in the presence of AdoHcy (Q). This resulted in competitive inhibition, consistent with an ordered bisubstrate system. Additionally, the Km of monomethylated R1 peptide was shown to be 3x lower in concentration than unmethylated R1(35), supporting a system in which formation of ADMA was „apparently processive‟ as MMA-R1 would be converted into ADMA- 14 R1 shortly after production due to its superior affinity constant. Kölbel took a three-pronged approach, analyzing whether PRMT1 and PRMT3 were distributive or processive. Firstly, it was determined that in RxR containing peptides, mono- or asymmetric dimethylation of adjacent arginines did not affect the kinetic constants dictating the methylation neighboring arginine residues. Secondly, TOF analysis of multiple arginine containing peptides revealed production of MMA:ADMA up to a 5:1 ratio. In a processive system, involving homodimers, only ADMA should be produced per catalytic cycle, as there are only two cofactor-loaded active sites per substrate turnover event. Thirdly, tryptophan fluorescence quenching experiments were performed with RxR-2 peptide (containing 2 tryptophan residues), revealing that the dissociation kinetics were similar to the catalytic constant of total methyl transfer regardless of the presence of AdoMet or AdoHcy. This supported a system in which processitivity was unlikely as substrate dissociation occurred on a timescale proportional to transfer of a single methyl group (36). An additional study has demonstrated that cofactor binding triggers multiple structural changes that improve the ability to recruit peptide substrates to the active site. Transient kinetic analysis of PRMT1 performed by Feng et al showed that association of fluorescently labeled acetylated histone H4 tail peptide increased when saturating levels of AdoHcy were present in in solution (34). Additionally, in the presence of AdoHcy dissociation of these peptides was shown to decrease. It was suggested that the microenvironment of PRMT1 may change in the presence of cofactor to better facilitate peptide binding. It was observed that this binding event best fits to a double exponential plot, indicating that peptide binding may also be comprised of multiple steps with unique kinetic constants. Association and dissociation rate constants for peptide binding were more rapid than those associated with catalysis, so it was concluded that catalysis is 15 the rate-limiting step in this system (34). Together these results support system in which cofactor binding triggers several, slower structural conformations to occur, fully forming the active site and preparing the enzyme to recruit a protein or peptide substrate into the active site. Structural data supports an ordered bisubstrate mechanism. The most compelling evidence for an ordered bisubstrate system originates for human PRMT crystal structures themselves, and interestingly, from experiments performed to generate PRMT specific inhibitors. Upon binding of AdoHcy in the cofactor binding pocket, a conformational change has been shown to occur for PRMT4 in which the N-terminal αXY helix reorients to occlude the AdoHcy from solvent and helps to form an electronegative peptide binding groove that runs along the inner channel of the PRMT4 homodimer (4, 29). Sack et al. demonstrated that binding of a pyrazole based PRMT4 specific inhibitor only occurred in the presence of AdoHcy as evidenced by isothermal calorimetry (29). Structural analysis of the PRMT4:AdoHcy:inhibitor ternary complex demonstrated that the inhibitor bound within the peptide binding groove adjacent to the cofactor binding pocket, docking in similar proximity to where arginine would be situated in a substrate peptide. Previously stated, crystal structures of human PRMT3, PRMT4 and human PRMT6 possess the αXY helix and in both cases solvent exposed electronegative residues present on the helices help form the electronegative peptide binding groove that passes through the inner channel of the homodimer. This further justifies why pyrazole inhibitor binding could only be modeled in the presence of AdoHcy. 16 1.3 Sequence-based control of PRMT activity and common peptide substrates. The PRMT RGG methylation motif. Early comparisons of PRMT substrate sequences such as those found in fibrillarin, hnRNPA1, bFGF, nucleolin and others, revealed a distinct motif (41). ADMA was present on (G/F)GGRGG(G/F) independent of where this consensus sequence appeared in protein substrates. This sequence was classified as the „RGG‟ repeat and the term RGG was commonly used to identify the location of ADMA. The RGG repeat was highly clustered within RNA binding proteins and PRMT activity has now been demonstrated to modulate RNA-protein interactions (42). Despite this compelling consensus sequence, many good PRMT substrates are asymmetrically dimethylated on arginines that are not nestled within this consensus sequence. Histones, one of the first discovered cellular substrates for PRMTs, contain N-terminal arginines that are not nestled within an RGG motif (43). Wooderchak et al. embarked on a study to better characterize the types of sequences that could be methylated by PRMT1. Peptide libraries containing variations on the RGG motifs were purchased and incubated with PRMT1. The levels of MMA and ADMA formation were measured and it was discovered that several different sequences could be methylated when the second amino acid in the RGG motif was modified (RxG) (44). Although the RGG motif was still favoured, this result (in combination with the presence of ADMA on non-RGG substrates such as histones) suggested that other sequences could be candidates for arginine methylation and should be considered beyond simply screening for RGG repeats (45). Interestingly, this study also noted that the ratio of MMA:ADMA initially consisted of mostly MMA and was gradually dominated by ADMA. It was concluded based on this observation that the mechanism of methyl transfer was distributive (44). It is worth noting that the R3 peptide (ac- 17 GGRGGFGGRGGFGGRGGFGG) and R1 fibrillarin like peptide (WGGYSRGGYGGW) are based upon the RGG repeat found in fibrillarin, which is heavily methylated by PRMT1 (46). Conformational peptide changes alter methylation activity. Aside from the RGG motif, the RxR motif has been identified as a common consensus sequence for PRMT activity (41). This sequence motif was analyzed by Kölbel et al. in the context of peptides based upon nuclear poly(A) binding protein 1 (PABPN1) which contains several RxR repeats that results in up to 13 unique methylation sites. It was found that R289 was preferentially converted to ADMA in vitro and the presence of ADMA at other arginines did not greatly change the methylation of R289. Peptides based on PABPN1 were derived and arginine 10 (R10 - corresponding to R289) was also discovered to be the primary site of ADMA formation. By substituting randomized flanking amino acids, it was shown that proline 9 (P9) was essential to this preferential methylation of R10. Far UV circular dichroism spectroscopy revealed that the presence of P9 in RxR13 peptide induced a reverse turn conformation, which consequently improved methylation of R10 (47). In addition to this study, Gui et al. performed mass spectrometric analysis on an R2 peptide (GGRGGFGGKGGFGGRGGFG) and demonstrated that PRMT1 methylation of this substrate was non-stochastic and preferentially targeted the N-terminal arginine first (40). Methylation of the second arginine in R2 did not occur until the first arginine was asymmetrically dimethylated. The key observation in preferential arginine methylation in RxR13 peptide and R2 peptide is that solvent exposed, N-terminal arginines are preferentially methylated by PRMT1. Although with the R2 peptide both arginines were claimed to be identical in local environment, the N-terminal arginine is more solvent exposed than the nestled secondary arginine. Thus, both backbone conformation as well as N-terminal proximity affects 18 solvent exposure and consequently methylation priority of internal arginines within peptide substrates. This observation is validated by histone methylation, which occurs largely on N- terminal extensions rather than residues buried within the globular core of histones. Solvent exposure combined with protein-protein recruitment to the nucleosome likely facilitates efficient methylation of individual histones despite lack of an RGG motif. Histones and arginine methylation. Several studies presented within this dissertation utilize histone-based peptides and whole histones as substrates to demonstrate arginine methylation. When arginine methylation was first discovered by purification from calf thymus, it was surmised that histones comprised the first group of important PRMT substrates (48). Since then, several important arginine methylation marks on histones have been characterized, as well as the individual PRMT family members that catalyze these post-translational modifications. PRMT1 catalyzes H4R3me2a, which has been shown to be a mark of gene expression through the recruitment of transcription factors such as TDRD3 (49). Although not a type I enzyme, PRMT5 generates H4R3me2s, which is a mark of gene repression, suggesting that PRMT1 and PRMT5 compete for the same methylation location to determine the activation state of genes (50). In vitro, PRMT1 and PRMT6 can methylate calf thymus histone H2A, H2B, H3 and H4 (28). PRMT4 and PRMT6 both target histone H3 in situ. Recruitment of PRMT4 to promoters triggers the generation of H3R17me2a and H3R26me2a (51-53). Interestingly, cleavage of PRMT4‟s C-terminal tail eliminates its ability to methylate histone H3, suggesting that it may play a role in substrate recruitment (4). H3R17me2a and H2R26me2a have been linked in concert with lysine acetylation to disassociate the nucleosome remodeling and deacetylase complex (NuRD) and TIF1 family co-repressors from the histone complex (54). Additionally, it 19 has been noted that K23 acetylation recruits CARM1 binding to histone H3 (55). The best characterized histone methylation site for PRMT6 is the H3R2me2a (56). This mark is mutually exclusive with the H3K4me3 activation mark, initially assigning PRMT6 as a repressor of gene expression (57). However, PRMT6 has also been shown to produce H4R3me2a and H2AR4me2a marks which are both associated with gene expression (58). PRMT6 is also capable of methylation and subsequently improved activity of DNA polymerase-β (59) in addition to coactivating hormone receptors (60). Thus PRMT6 may possess a more flexible function, both repressing and activating genes in a context-dependent manner. Of the remaining type I enzymes, PRMT2 is capable of methylating whole histone H4 and it is suggested that PRMT3 does not play a major role in epigenetic pathways as it is wholly located within the cytosol (17, 61). HIV Tat and PRMT6. Chapter 3 presents a series of substrate inhibitors based upon HIV Tat peptide, which is a methylation substrate specifically of PRMT6 in HIV infected cells (62). Tat protein is a transactivator of gene expression that is produced early in the HIV life cycle. This viral protein interacts with the Tat transactivation response region (TAR) found 5‟ of HIV RNA, and improves the activity of RNA polymerase II by recruiting cyclin T1 and cyclin dependent kinase 9 (CDK9). This complex phosphorylates RNA polymerase II resulting in enhanced activity (63, 64). HIV-1 Tat is heavy modified by several post-translational enzymes. It possesses a number of acetylated cysteine and lysine residues, it is ubiquitinated at K71 and is arginine methylated within an arginine-rich sequence found at residues 48-58 (65). To investigate which PRMT was responsible arginine methylation, HIV-1 Tat protein was expressed, purified and incubated with PRMT4, PRMT6 and PRMT7. Of those enzymes, only PRMT6 was able to 20 significantly methylate Tat (66). It was surmised that HIV Tat must be a substrate for PRMT6. Co-IP experiments verified an interaction between HIV Tat and PRMT6, and overexpression of PRMT6 caused a reduction in Tat transactivation measured by an HIV promoter based luciferase reporter. Additionally, PRMT6 knockdown with siRNA increased levels of viral protein as measured by a Tat sensitive LTR-lacZ CD4+ HeLa cell line (62). The same group went on to demonstrate that HIV-1 Tat was primarily methylated at R52 (66). In conjunction with PRMT6‟s specific activity when using Tat as a substrate, HIV-1 Tat-based peptides became an attractive template due to its ability to traverse cell membranes (67). Methylation of HIV-1 Tat by PRMT6 ultimately decreases its affinity for the 5‟ TAR, attenuating production of viral RNA and promoting cellular repression of the virus (66). 1.4 Hypothesis and research objectives The work within this dissertation was aimed at elucidating whether PRMT homomeric and/or heteromeric interactions are able to modulate enzyme activity and/or substrate specificity. Additionally, given the conflicting data surrounding the type I PRMT mechanism of catalysis, unique molecular tools using HIV-1 Tat and fibrillarin-like R1 peptide were developed to probe PRMT activity beyond what was capable from site-directed mutagenesis alone. The specific research hypothesis is as stated, “Type I PRMT activity and substrate specificity is dependent upon and can be modified by homo- and hetero-oligomerization.” The studies found herein utilize a number of in vitro assays, such as resonance energy transfer, liquid chromatography coupled mass spectrometry, radiomethylation assays and kinetic solvent effects in conjunction with unique ωN-substituted R1 peptide and HIV-1 Tat peptide substrates to not only quantify oligomerization, but how the fundamental chemistry of type I PRMTs responds to these novel substrates. 21 2 Förster resonance energy transfer measurements of cofactor- dependant effects on protein arginine N-methyltransferase homodimerization 1,2 2.1 Introduction Structures of PRMTs reveal a common mode of dimerization between catalytic subunits (2-5, 53). Each subunit contains a dimerization helix-turn-helix that protrudes from the C-terminal β- barrel and rests upon the N-terminal AdoMet binding domain of the other subunit, forming a central anionic cavity with two opposing active sites. Removing the dimerization helix-turn-helix from Rmt1p and PRMT1 has been shown to eliminate homodimerization, AdoMet binding, and methyltransferase activity (2, 5). More recently, Higashimoto et al. (2007) have shown that CARM1 is phosphorylated on S229 on the dimerization helix-turn-helix, and a phosphoserine mimic S229E mutation significantly reduced AdoMet binding, enzyme activity in vitro, homodimerization, and CARM1-mediated transactivation of estrogen receptor-dependent transcription (33). Taken together these results underscore the important relationship between PRMT homodimerization and methyltransferase activity. Although PRMT1 and PRMT6 possess a high degree of sequence identity in the dimer arm implying a similar structure-function relationship, the physiological role of homodimerization has not been demonstrated for PRMT6. 1 A version of chapter 2 has been published. Thomas, D., Lakowski, T. M., Pak, M. L., Kim, J. J., and Frankel, A. (2010) Forster resonance energy transfer measurements of cofactor-dependent effects on protein arginine N- methyltransferase homodimerization, Protein Sci 19, 2141-2151. Dylan Thomas performed all resonance energy transfer and fluorescence based experiments. Dylan Thomas also prepared several of the fluorescently tagged enzymes used in this study. Dr. Ted Lakowski performed product inhibition and enzyme kinetics. Dr. Lakowski also wrote portions of the manuscript pertaining to enzyme kinetics. Dr. Magnolia L. Pak produced mCer and mCit-tagged PRMT1. Jenny J. Kim helped subclone several templates from which mCer and mCit-PRMTs were produced. 2 DNA constructs for mRFP1, eGFP and mCyan were generous gifts from Dr. Judy Wong and Dr. Luois Lefebvre from the University of British Columbia. 22 In this study, the dissociation constant of homodimerization is measured for PRMT1 and PRMT6 in the presence or absence of AdoMet and AdoHcy. N-terminal mCerulean- and mCitrine-PRMT1 and 6 were expressed and purified to model oligomerization affinity utilizing resonance energy transfer. A parabolic pattern of energy transfer was generated by titrating mCitrine conjugated PRMTs into their associated mCerulean binding partner. By performing titrations in the presence or absence of saturating AdoMet or AdoHcy, it was found that the dissociation constant decreased in the presence of AdoMet for PRMT1, and increased in the presence of AdoHcy for PRMT6. This pattern supports a mechanism in which AdoMet induces stabilization of homodierization and AdoHcy promotes dissociation, implying PRMTs must form free monomers for cofactor turnover to take place. Steady-state enzyme kinetics and linear fluorescence quenching experiments were performed that demonstrate N-terminal fluorescent conjugation with eGFP variants does not effect PRMT activity or substrate specificity, and resonance energy transfer occurs between dimers and not higher order oligomers. . 23 2.2 Methods Expression plasmids. Plasmids harboring enhanced green fluorescent protein (eGFP) and enhanced cyan fluorescent protein (eCFP) were generously donated by Drs. Judy Wong (Faculty of Pharmaceutical Sciences, The University of British Columbia) and Louis Lefebvre (Department of Medical Genetics, The University of British Columbia), respectively. The expression vector for human PRMT6 in pET28a(+) was previously described (19). Rat PRMT1 in pGEX-2T (17) was humanized through a H161Y mutation to make PRMT1v1 (68, 69), and then sub-cloned into pET28a(+) using BamHI and XhoI restriction sites. PRMT1v1 (isoform 1) is used in all fluorescent constructs with PRMT1. Plasmids pET28a(+)-eGFP-PRMT6 and pET28a(+)-eCFP-PRMT6 were generated by sub-cloning eGFP and eCFP sequences into an NdeI site 5‟ to the PRMT6 gene within the pET28a(+) vector using the primers 5‟ - GGA ATT CCA TAT GGT GAG CAA GGG CGA GGA GC - 3‟ and 5‟ - GGA ATT CCA TAT GCT TGT ACA GCT CGT CCA TGC CGA G - 3‟. Expression plasmids pmCer-PRMT6 and pmCit- PRMT6, which code for mCer- and mCit-PRMT6 fusions, were generated through multiple rounds of site-directed mutagenesis on pET28a(+)-eGFP-PRMT6 and pET28a(+)-eCFP-PRMT6 templates. To generate pmCer-PRMT1 and pmCit-PRMT1 expression vectors, sequences coding for mCer and mCit were PCR-amplified and sub-cloned into the NdeI site 5‟ to PRMT1v1 within the pET28a(+)-PRMT1 vector using identical primers to those used to amplify eGFP above since the 5‟ and 3‟ sequences are identical. All of the fluorescent fusions code for the protein sequence AMSTGGQQMGR as a linker between the N-terminal fluorescent protein and the PRMT. All eGFP variants contained an A206K mutation to disrupt any intrinsic oligomerization. Protein expression and isolation. Expression of both fluorescent (mCer or mCit attached) and non-fluorescent PRMT1 and PRMT6 are induced with 1.0 µM isopropyl--D- 24 thiogalactopyranoside at 30˚C in BL21(DE3) pLysS gold cells (Stratagene) overnight in LB medium (Fisher Scientific) containing an additional 1.0% glucose, 50 µg/ml kanamycin, and 35 µg/ml chloramphenicol. The cells are harvested via centrifugation in a Beckman model J2-21 centrifuge at 10,000 x g for 15 min and are immediately frozen at -80˚C until purification. After resuspension in lysis buffer (50 mM HEPES-KOH, pH 7.6, 1.0 M NH4Cl, 10 mM MgCl2, 0.1% lysozyme, 25 U/ml DNase I, 0.2 µM Triton X-100, 7.0 mM -mercaptoethanol, 1.0 mM phenylmethanesulphonylfluoride, and complete EDTA-free protease inhibitor cocktail tablets (Roche product code 04693132001) according to volume requirements) at 2 mL per gram wet weight of cells, cells are sonicated using a Branson Sonifier 450 on ice for eight 30-s pulses at 50% duty cycle with 30-s pauses in between. Each protein is first purified via a 1.0-ml HisTrap FF affinity column (GE Healthcare) per 2.0 L bacterial culture using an established method (17). The eluent from the first step is purified using a HiLoad 26/60 Superdex 200 pg column (GE Healthcare). Each sample is collected and exchanged into a storage buffer (100 mM HEPES- KOH, pH 8.0, 200 mM NaCl, 1 mM DTT, 10% glycerol, and 2 mM EDTA) using Amicon Ultra ultracentrifugal filters with a 10-kDa molecular weight cut-off (Millipore), frozen in liquid nitrogen, and stored at -80˚C (17). Protein quantification and spectral characteristics. The concentration of fluorescent fusion proteins is measured using the extinction coefficients for mCit ( nm = 77,000 M -1 cm -1 ) or mCer (434 nm = 43,000 M -1 cm -1 ) (70, 71). The concentrations of unconjugated PRMT1 and PRMT6 are determined by separation of purified proteins on SDS-PAGE and subsequent densitometry of Coomassie blue-stained bands as described previously (17). The emission spectra of mCit and mCer fusion proteins, excited at 434 nm, are recorded in methylation buffer consisting of 50 mM HEPES-KOH, pH 8.0, 10 mM NaCl, and 1.0 mM DTT 25 in 3-mL polystyrene cuvettes (Sarstedt) on a Cary Eclipse Fluorescence Spectrophotometer (Varian) using medium gain, medium scan speed, a 5-nm excitation slit width, and 10-nm emission slit width. PRMT activity assays. Non-fluorescent PRMT1, mCer-PRMT1 or mCit-PRMT1 at a concentration of 400 nM are incubated at 37 C for 1 h with increasing histone H4 tail peptide (SGRGKGGKGLGKGGAKRHRKVW) (17) and a constant saturating concentration of AdoMet (250 µM) in methylation buffer in a final volume of 80 µL. The H4 tail peptide is used at concentrations of 1.0, 2.0, 5.0, 10, 20, 40, and 100 M. Similar reactions are also carried out with 400 nM non-fluorescent PRMT6, mCer-PRMT6 or mCit-PRMT6 using the histone H3 tail peptide (ARTKQTARKSTGGKAPRKQLATKAAW) . Reactions are stopped by heating at 80 C for 5 min and the reaction samples are dried in a vacuum centrifuge, acid hydrolyzed in the vapor phase with 6.0 M HCl at 110 C in vacuo as described previously (35). Samples are reconstituted in 0.1% aqueous formic acid and 0.05% trifluoroacetic acid, and the amount of MMA, aDMA and the total methylation are measured according to a previously described UPLC-MS/MS assay (17). The resulting data is fit to the Michaelis–Menten equation using Sigma Plot 8 (SYSTAT) to generate apparent KM values for the above peptides. In order to establish an AdoHcy KI value for PRMT1, methylation assays with 400 nM PRMT1 and the product inhibitor AdoHcy at concentrations of 0, 0.5, 1.0, 5.0 and 20 µM are performed with constant 80 µM H4 tail peptide and variable concentrations of 1.0, 2.0, 5.0, 10 and 25 µM AdoMet. The samples are treated as described above, and data are analyzed according to previously described methods (39). FRET measurements. All steady-state FRET measurements are performed in a microplate format on a Synergy Mx Monochromator-based Multi-mode Microplate Reader (Biotek). 26 Measurements are taken using excitation and emission slit widths of 9 nm. Sample wells are filled to an 80-µL final volume in a 384-well black polystyrene non-binding surface microplate (Corning #3575). The mCer donor is excited at 434 nm and fluorescence emitted from the mCit acceptor is measured at 529 nm in absolute fluorescence. Sensitivity is adjusted with an 8-mm height correction from the upper plane of the sample wells. For KD measurements each plate contain five rows comprised of sixteen different fluorescent protein concentrations (four rows are used for samples with fluorescent PRMT6 and AdoMet), and two additional rows for mCit and mCer background signal. Sample wells are prepared in quintuplicate and contain a 0.5-µM mCer-PRMT1 or 0.5-µM mCer-PRMT6 solution to which varying concentrations of mCit-PRMT1 or mCit-PRMT6 are added to saturate FRET signal. Plate readings are acquired after 60 min incubation at 37 ˚C. Background signals are determined by measuring fluorescence of individual fluorescent proteins at concentrations corresponding to experimental samples, and these background signals are subtracted from the experimental FRET signals to generate background-corrected data, which is fit using Caligator set to a 1:1 ratio for protein binding to calculate KD values (72). For efficiency measurements used to assess subunit contributions to FRET, 11 solutions are premixed containing either 1.0 µM mCer-PRMT1 or 1.0 µM mCer-PRMT6, as well as varying concentrations up to 1.0 µM mCit-PRMT1 or 1.0 µM mCit-PRMT6, respectively. These solutions are then pre-incubated at 37 ˚C for 1 h and transferred in 80-µL aliquots into 384-well plates in triplicate. Efficiency measurements are performed with 434-nm excitation and 475-nm emission wavelengths. PRMT1 and PRMT6 FRET specificity assays are performed by premixing mCer-PRMT1 with mCit-PRMT1 or mCer-PRMT6 with mCit-PRMT6 at 1.0-µM of each fluorescent protein 27 along with varying concentrations up to 5.0 µM of non-fluorescent PRMT1, PRMT6, or methylation buffer as a control. The solutions are incubated for 60 min at 37 ˚C prior to exciting samples at 434 nm and measuring fluorescence at both 475 nm and 529 nm. 28 2.3 Results This study focuses on the quantification of PRMT homodimerization using FRET. We create fusion proteins of PRMT1 and PRMT6 bearing either mCer or mCit on their N-termini to measure FRET interactions between subunits using the absorption maximum at 434 nm for mCer and the emission maximum at 529 nm for mCit (22), thus providing a simple method for quantifying homodimer binding affinity. The FRET response is produced when a mCer-PRMT dimerizes with a mCit-PRMT [Figure 2.1]. In the absence of this pair forming, all of the radiative energy from mCer emits at its own emission maximum at 475 nm, and FRET will not occur. Implicit in Figure 2.1, mCer-PRMT and mCit-PRMT homodimers must dissociate into monomers and then re-associate into a mCer/mCit-PRMT dimer in order to observe a FRET signal. 29 Figure 2.1. Formation of PRMT FRET pairs. Upon addition of mCer-PRMT (1) and mCit-PRMT (2), the populations of homodimers will dissociate into monomers and recombine into mixed homodimers, bringing mCer and mCit in close proximity as an active FRET pair (3). When excited with 434-nm light, only (3) will produce a FRET signal at 529 nm. 30 Fluorescent PRMT spectral properties and enzymatic activities. Prior to performing FRET experiments, all fluorescent fusion proteins are investigated for appropriate spectral characteristics and enzymatic activity. The absorption and emission spectra for fluorescent PRMTs are recorded from 450 to 650 nm using a 434-nm excitation wavelength, and normalized to a value of 1.0 for comparison [Figure 2.2]. The absorption and emission wavelengths for all conjugated proteins are consistent with those of unconjugated mCer and mCit (70, 73). These data demonstrate that the fluorescent components of the conjugated PRMTs are properly folded and possess functional fluorophores. Using a UPLC tandem mass spectrometry assay (described in Materials and Methods) the activities of mCer- and mCit-PRMTs are compared to those of their respective unconjugated proteins [Figure 2.3]. The apparent Vmax and KM values (Table 2.1) are similar for respective fluorescent and non-fluorescent PRMTs, suggesting that the PRMT component of each fluorescent fusion protein is properly folded, and that the fluorescent attachment has little or no effect on the activity of the conjugated PRMT. Figure 2.2. Normalized absorption and emission spectra for mCerulean and mCitrine-PRMTs. Normalized spectra are overlaid for mCer-PRMT1 absorption (1) and emission (2), as well as mCit-PRMT1 absorption (3) and emission (4). The emission spectra are collected for both mCer-PRMTs and mCit-PRMTs using a 434-nm excitation wavelength. Fluorescent PRMT6 produces identical spectra, whereas PRMT1 and PRMT6 without attached fluorophores do not possess intrinsic fluorescence within this range using 434-nm excitation (data not shown). 31 Figure 2.3. Activities of PRMT1 and PRMT6 with and without mCerulean or mCitrine and double- reciprocal plot of PRMT1 product inhibitor analysis. The initial velocity of reactions is determined as described in the Materials and Methods section. (A) Methylation assays using PRMT1 (●), mCit-PRMT1 (○), or mCer-PRMT1 (▼) with increasing concentrations of the H4 tail peptide are shown. (B) Methylation assays using PRMT6 (●), mCit-PRMT6 (○), or mCer-PRMT6 (▼) with increasing concentrations of the H3 tail peptide are shown. The initial rate is calculated as pmol per min per nanomole of enzyme to accommodate the differences in molecular weights between fluorescent and non-fluorescent PRMTs, and kinetic parameters are listed in Table 2.1. (C) In the presence of a constant 80-µM H4 tail peptide concentration, variable concentrations of 1.0, 2.0, 5.0, 10 and 25 µM AdoMet are incubated in methylation buffer. These reactions are repeated in the presence of fixed AdoHcy concentrations of 0 µM (□), 0.5 µM (▲), 1.0 µM (○ ), 5.0 µM (■) and 20 µM (●). The pattern of intersecting lines on the y-axis is indicative of competitive inhibition for the product inhibitor AdoHcy. Table 2.1. Apparent kinetic parameters for fluorescent and non-fluorescent PRMTs. Enzyme Substrate Vmax (pmol/minnmol)* KM (µM) PRMT1 H4 tail 274.5 (5.9) 6.9 (0.2) mCit-PRMT1 H4 tail 302 (26) 10.3 (1.9) mCer-PRMT1 H4 tail 294.1 (6.2) 5.8 (0.1) PRMT6 H3 tail 13.2 (0.5) 0.8 (0.1) mCit-PRMT6 H3 tail 12.8 (0.4) 0.5 (0.1) mCer-PRMT6 H3 tail 9.7 (0.4) 0.9 (0.3) *Vmax is calculated as pmol per min per nmol of enzyme to account for mass differences between fluorescent and non-fluorescent PRMTs. Numbers in parentheses represent standard deviations. 32 FRET from fluorescent PRMT homodimerization. To test the feasibility of using FRET to measure PRMT homodimerization as outlined in Figure 2.1, up to 3.5 µM mCit-PRMT1 is titrated into a solution of 1.0 µM mCer-PRMT1 while scanning the fluorescence emission wavelengths from 450 to 650 nm [Figure 2.4(A)]. The initial spectrum of mCer-PRMT1 is consistent with the spectrum of mCer alone (70). As mCit-PRMT1 is titrated into mCer-PRMT1, a peak appears at the emission maximum for mCit-PRMT1 (529 nm). For each increase in mCit- PRMT1 concentration, a corresponding drop is observed in mCer-PRMT1 emission at 475 nm greater than that which can be accounted for by dilution alone. Energy transfer from 434 nm to 529 nm is demonstrated by an increase in 529-nm fluorescence with a concomitant decrease in 475-nm fluorescence (i.e., donor emission). These spectral changes are a direct demonstration of the FRET phenomenon. Similar results are observed for fluorescent PRMT6 proteins (data not shown). When excited using 434 nm light, both mCer- and mCit-PRMTs are able to produce 529 nm emissions not attributable to FRET. To ensure the signal produced from protein mixing is due to FRET and not background fluorescence, a multi-well plate assay is performed with the inclusion of background controls [Figure 2.4(B)]. The sum of mCer- and mCit-PRMT6 fluorescence emissions (i.e., total background signal) is less than the fluorescence observed when the two fluorescent PRMTs are combined, thus demonstrating that additional fluorescence at 529 nm is produced from FRET as a result of PRMT6 homodimerization. These background controls are also employed in FRET experiments with fluorescent PRMT1 proteins (data not shown). PRMT dissociation constants. We determine the KD values for PRMT dimerization by varying the mCit-PRMT concentration with a fixed mCer-PRMT concentration. A broad range of mCit-PRMT concentrations are initially used to estimate a KD value for each set of 33 experimental conditions. Each experiment is then repeated with an appropriate range of mCit- PRMT concentrations that produce data points above and below each estimated KD value to best fit the data in subsequent binding curves [Figure 2.5]. We note that using the same mCit concentration range for all experimental conditions (e.g., with or without cofactor) would result in inaccurate KD value estimation. Fluorescence readings are acquired using multi-well plate format and the data fit to equation 2.1, producing a hyperbolic fit for equimolar binding (72). The background-corrected FRET signal is proportional to the ratio of the fluorescent PRMT FRET pair concentration ([Dimer]) to the total PRMT concentration ([PRMT]total), which can also be expressed in terms of the concentrations of PRMTs conjugated to mCer ([mCer]) and mCit ([mCit]), as well as the KD value. Based on the kinetic data presented above we make the assumption that the dimerization KD values are the same irrespective of the fluorescent attachment. It is important to note that the KD value is not calculated as ½ of the FRET maximal signal, but rather as a function of best fit using equation 2.1, which takes into account both monomeric and dimeric PRMT populations. Equation 2.1 ][2 ][][ ][ ][ ]][[4)][]([ 2 mCer KmCitmCer PRMT Dimer FRET mCitmCerKmCitmCerD total D   The calculated KD values for all multi-well FRET assays are listed in Table 2.2. The presence of AdoMet decreases the KD of dimerization for PRMT1 by 4-fold compared to PRMT1 alone or with AdoHcy. In contrast, the presence of AdoHcy increases the KD of dimerization for PRMT6 by 6-fold compared to PRMT6 alone or with AdoMet. Interestingly, the KD values for both PRMT1 and PRMT6 are greater in the presence of AdoHcy than in the presence of AdoMet. These results demonstrate that the presence of cofactors can differentially affect PRMT dimerization. 34 PRMT6 dimerization appears to be more sensitive to the presence of AdoHcy than PRMT1. Thus we compared AdoHcy dissociation constant (KI) values for PRMT1 and PRMT6 in order to expose a possible regulatory mechanism for PRMT-selective inhibition. We used the mass spectrometry-based assay to determine the KI value for PRMT1 since we have previously established the AdoHcy KI value for PRMT6 (35). As expected the double-reciprocal plot of the inhibition data reveal a series of lines increasing in slope with increasing AdoHcy concentrations that intersect on the y-axis [Figure 2.3(C)], indicating that the inhibition is competitive. The AdoHcy KI = 5.8  0.5 µM for PRMT1, which is 4-fold higher than the KI value previously calculated for PRMT6 (35). Therefore, not only is PRMT6 dimerization more sensitive to AdoHcy concentration than PRMT1 dimerization, but the enzyme activity is more sensitive as well. PRMT dimer contribution to FRET. PRMT1 has been shown to form high order oligomers under purification and crystallographic conditions (2, 5). To investigate whether FRET signals from fluorescent PRMT1 and PRMT6 proteins are attributed to complexes larger than dimers under our experimental conditions, we measure the efficiency of energy transfer between fluorophores using excitation at 434 nm and emission at 475 nm to capture the quenching of mCer-PRMT fluorescence caused by mCit-PRMT, thus allowing us to assess the oligomeric contribution to FRET. Here, efficiency is defined as the magnitude of energy transfer from donor to acceptor (equation 2.2) where DA is the 475-nm emission of the donor/acceptor pair and D is the 475-nm emission of the donor alone (74). Equation 2.2 35 As the concentration of mCit-PRMT increases, more donor/acceptor pairs form and the efficiency increases. Efficiency increases are linearly related to the mole fraction of the FRET acceptor when dimeric complexes are formed, which is observed for PRMT1 and PRMT6 as shown in Figure 2.6 where efficiency data fit linearly with R 2 values of 0.91 and 0.96, respectively. Unlike the case for dimers, the contribution to FRET for larger oligomeric complexes shows a hyperbolic curve when efficiency is plotted against the mole fraction of FRET acceptor. Efficiency curves for various oligomers are plotted [Figure 2.6] using a simplified binomial model (equation 2.3) where %Q is the quenching from FRET, α is an efficiency constant unique to each FRET system, PA is the mole fraction of acceptor, and n is the number of oligomers (75, 76). Our PRMT1 and PRMT6 efficiency data support the formation of dimer FRET complexes. Equation 2.3 Quenching is extrapolated to the mole fraction of one for PRMT1 and PRMT6 to estimate maximum efficiency of 16% and 25%, respectively [Figure 2.6]. Extrapolated maximal efficiencies for PRMT1 and PRMT6 demonstrate a higher efficiency energy transfer for PRMT6. It is possible that the mCer/mCit portions of the fusion proteins are held in closer proximity for PRMT6 than for PRMT1. 36 Figure 2.4. mCerulean and mCitrine-PRMTs produce FRET. (A) A 1.0-µM solution of mCer-PRMT1 is added to a cuvette to a final volume of 1.5 mL. mCit PRMT1 is then titrated for sixteen 10-µL additions into the sample, covering a concentration range of 0 to 3.5 µM. After each addition the solution is allowed to stir for 2 min prior to scanning for wavelength emission between 450 to 650 nm using a Varian benchtop fluorometer as described in the Materials and Methods section. (B) The emission at 529 nm is measured using a Biotek micro-plate reader as described in the Materials and Methods section. The background fluorescence from 0.5 µM mCer-PRMT6 alone (▲) remains constant, and the background fluorescence contributions from 0 to 2.08 µM mCit-PRMT6 alone (●) increases linearly with increasing protein. The combination of a fixed concentration of mCer-PRMT6 (0.5 µM) with varying concentrations of mCit-PRMT6 (■) shows greater fluorescence intensity than the sum of both background signals (♦). 37 Dimer subunit specificity. In order to demonstrate that FRET pairs are established through the binding of two PRMT subunits, non-fluorescent PRMTs are used to disrupt FRET from mCer/mCit-PRMTs. In these experiments, increasing concentrations of non-fluorescent PRMT1 and PRMT6 are mixed with FRET pairs and fluorescence is measured at emission wavelengths of 475 nm and 529 nm to capture the change in energy transfer. As shown in Figure 2.7(A), the presence of non-fluorescent PRMT1 results in a concentration-dependent decrease in energy transfer between mCer/mCit-PRMT1, whereas the addition of buffer has no effect. Interestingly, non-fluorescent PRMT6 also disrupts the mCer/mCit-PRMT1 FRET pair, but to a lesser extent. When non-fluorescent PRMT6 is added to mCer/mCit-PRMT6 [Figure 2.7(B)], the energy transfer between the FRET pair is decreased in a concentration-dependent manner. The addition of buffer or non-fluorescent PRMT1 does not disrupt the mCer/mCit-PRMT6 FRET. We can conclude from these experiments that PRMT1 and PRMT6 can compete with their own FRET pairs, demonstrating specificity of the FRET signals. Given that PRMT6 can weakly disrupt PRMT1 FRET pairs, we proceeded to test for PRMT1/PRMT6 heterodimerization. Varying concentrations of mCit-PRMT6 with a fixed mCer-PRMT1 concentration are used to detect FRET consistent with dimerization. Although a weak FRET signal is detected above background [data not shown], the protein concentrations required to reach a saturation point adequate to fit a dissociation curve and calculate a KD value are not achievable under our assay conditions. PRMT1/6 heterodimers are not likely to compose an appreciably large population in vivo given the relatively tight association between their respective homodimers [Table 2.2]. 38 Figure 2.5. Steady-state FRET binding for fluorescent PRMTs. FRET measurements are performed as described in the Materials and Methods section between mCer- and mCit- PRMTs. Protein binding curves are shown for (A) PRMT1 (■), PRMT1 with 500 µM AdoMet (●), PRMT1 with 20 µM AdoHcy (▲), (B) PRMT6 (■), PRMT6 with 500 µM AdoMet (●) and PRMT6 with 20 µM AdoHcy (▲). All experimental groups contain 0.5 µM mCer-PRMT1 or mCer-PRMT6. The dissociation constants derived from protein binding curves for fluorescent PRMT1 and PRMT6 are shown with their standard deviations. Table 2.2. Dissociation constants for fluorescent PRMT1 and PRMT6 with and without cofactor. Enzyme Cofactor KD (nM)* PRMT1 – 110 (26) PRMT1 AdoMet 30 (14) PRMT1 AdoHcy 110 (38) PRMT6 – 210 (34) PRMT6 AdoMet 180 (104) PRMT6 AdoHcy 1100 (67) *Numbers in parentheses represent standard deviations. 39 Figure 2.6. FRET efficiency subunit contribution. Up to 1.0-µM mCit-PRMT1 or -PRMT6 is mixed with 1.0-μM mCer-PRMT1 or -PRMT6 and 475-nm emissions were collected. (A) Efficiency measurements for the mCer/mCit-PRMT1 FRET pair indicate a linear relationship when plotted against mole fraction of mCit-PRMT1. (B) Similar results are shown for the mCer/mCit-PRMT6 FRET pair. Both PRMT1 and PRMT6 theoretical efficiencies are obtained by extrapolating to a mole fraction of one. Percent efficiencies (%Q) for trimer, tetramer, pentamer and hexamer are modeled from right to left for both enzymes and plotted in gray. Figure 2.7. Non-fluorescent PRMT competition with FRET pairs. (A) The 529 nm/475 nm ratio of the mCer/mCit-PRMT1 FRET pair is plotted with addition of buffer (○), increasing non-fluorescent PRMT6 (■) and increasing non-fluorescent PRMT1 (♦) concentrations. (B) The 529 nm/475 nm ratio of the mCer/mCit-PRMT6 FRET pair is also plotted with addition of buffer (○), increasing non- fluorescent PRMT1 (■) and increasing non-fluorescent PRMT6 (♦) concentrations. Error bars represent standard deviation. 40 2.4 Discussion FRET to measure PRMT homodimerization. Spectroscopic techniques utilizing FRET provide a useful and accurate means of quantifying protein-protein interactions (71). Until now this technique has not been applied to the measurement of PRMT homodimers. In this work we show that PRMT dimerization can indeed be measured under various conditions using FRET. The major advantage of this technique is its compatibility with a multi-well plate format so that uniform sample equilibration can be achieved over multiple PRMT concentrations, avoiding sources of time-dependent fluctuations in fluorescence. This assay is made possible by attaching mCer and mCit to the N-termini of PRMT1 and PRMT6. Even though different human PRMT1 splice variants, differing in N-terminal length and sequence have been shown to exhibit differential enzyme activity and substrate specificity (69), we do not observe any differences in kinetic constants or substrate specificity between fluorescent and non-fluorescent PRMTs [Table 2.1]. Our study shows that the attachment of additional sequence on the N-termini of PRMT1 and PRMT6 does not affect their enzyme functions, suggesting that dimerization is also not affected. For the purpose of fitting FRET data we have made the assumption that the purported PRMT interaction is a 1:1 dimer unaffected by the presence of either N-terminal fluorescent protein mCer or mCit. This necessary assumption implies that during FRET experiments, the pool of monomeric PRMT is the same regardless of the accessory fluorescent protein. Without this assumption it would be necessary to attribute the FRET signal to two separate dissociation constants to determine individual monomer and dimer concentrations. Under the parameters used 41 in this study, the concentration of all three homodimerized species in Figure 2.1 are the same at the equivalence point for mCer- and mCit-PRMTs. PRMT oligomerization. The structure of yeast Rmt1p has been shown to form a trimer of dimers (i.e., hexamer) within its crystal lattice, but in solution it exists mostly as a dimer and its propensity to oligomerize occurs mostly at higher concentrations (0.1 to 4.0 mg/mL) (5), well above concentrations used in this study. Its mammalian homolog PRMT1 exists as a dimer within its crystal structure lattice (2). Dynamic light scattering and size exclusion analyses have estimated the PRMT1 molecular weight to be nearly 6-fold greater than the molecular weight of a dimer and 9-fold greater in the presence of AdoHcy (2). These molecular weights are not consistent with dimeric or hexameric structures, but are more likely caused by high molecular weight aggregates that form as a result of the high concentrations needed for native size determination. In addition, the mobile phase used to perform these experiments contained 5% glycerol, which can reduce PRMT1 activity [data not shown]. In this study, glycerol concentrations are kept below 1% (final concentration) and [PRMT]total does not exceed 2.1 µM for PRMT1. Aside from PRMT1, no evidence exists currently to suggest that PRMT6 is capable of forming high order oligomers beyond dimers. Efficiency data [Fig. 6] provide evidence that FRET occurs between two PRMT subunits. It is important to note that the relationship derived by Adair and Engelman (1994) applies to relatively small oligomeric complexes and assumes that each subunit can interact with all surrounding subunits (75). We cannot rule out the possibility that FRET between dimers occurs within a higher order oligomer, yet the spectral data from which we derive PRMT dissociation constants is generated from a 1:1 binding interaction between fluorescent PRMTs. 42 Regulatory implications for PRMT dissociation constants. Previous kinetic investigation of PRMT6 has demonstrated that it uses a Bi-Bi sequential ordered enzyme mechanism in which AdoMet associates first and AdoHcy dissociates last from the enzyme during a catalytic cycle (35). This mechanism is largely supported by crystal structures of PRMT1, 3, and 4 in complex with AdoHcy that show the cofactor buried underneath N-terminal α-helices (αX and/or αY) (2- 4, 53). Once positioned over the cofactor these α-helices serve as an upper ridge along one side of an acidic groove into which a methyl-accepting polypeptide can dock, and αY also establishes a portion of the contact surface for PRMT dimerization believed to be critical for enzyme activity. We find that PRMT1 and PRMT6 subunits discriminate between AdoMet and AdoHcy in the formation of homodimers consistent with facilitating enzyme turnover. The presence of AdoMet favors PRMT1 dimerization 4-fold and PRMT6 dimerization 6-fold over the presence of AdoHcy [Table 2.2 and Figure 2.5(B)], suggesting that the PRMT in complex with AdoMet facilitates dimer association in preparation for additional reaction steps to proceed, and the PRMT in complex with AdoHcy triggers dimer dissociation so that the product inhibitor can be released. The results of this study also point to some differences between PRMT1 and PRMT6 in response to AdoMet or AdoHcy. While the PRMT1 affinities towards AdoMet and AdoHcy are similar (dissociation constants K A S = 3.5 µM for AdoMet (17) and KI = 5.8  0.5 µM for AdoHcy), the PRMT6 affinity towards AdoHcy (KI = 1.4 µM) is approximately 10-fold higher than its affinity towards AdoMet (K A S = 16.5 µM) (35). These affinity differences suggest that PRMT6 activity can be more susceptible to the feedback inhibition of AdoHcy than PRMT1 activity. Relative intracellular levels of AdoMet and AdoHcy can also impact enzyme activity. The cellular [AdoMet]/[AdoHcy] ratio, also referred to as methylation potential, has been shown 43 to vary in different human cell lines. For example, this ratio was measured at 53.4 in liver cancer HepG2 cells, 21.1 in liver cancer SK-HEP-1 cells, 14.4 in breast cancer MCF-7 cells, 7.1 in embryonic kidney HEK293 cells, and 6.6 in cervical cancer HeLa cells (77). If we consider the ratio of dissociation equilibrium constants K A S and KI, then the expression rearranges to yield equation 2.4, where the concentration of PRMT bound to AdoMet is [PRMT•AdoMet], the concentration of PRMT bound to AdoHcy is [PRMT•AdoHcy], and the methylation potential is MP. Using equation 4 we calculate that the ratio of PRMT6-bound AdoMet to AdoHcy is 0.56 in HeLa cells (i.e., more PRMT6 is bound to AdoHcy than AdoMet), whereas the same ratio for PRMT1-bound cofactors is 11, thus demonstrating that in cells with lower methylation potential PRMT6 is susceptible to inhibition as a result of its higher affinity for AdoHcy over AdoMet. Equation 2.4          A S I K K MP AdoHcyPRMT AdoMetPRMT ]•[ ]•[ Alterations in the methylation potential can also affect protein-protein interactions as demonstrated by Herrmann et al. (2009), who reported recently that GFP fusion proteins of PRMT1 and PRMT6 expressed in HEK293T cells exhibited diffusion characteristics consistent with high molecular weight complexes in fluorescence recovery after photobleaching experiments. In the presence of adenosine dialdehyde, which is an AdoHcy hydrolase inhibitor that causes intracellular AdoHcy accumulation and subsequent inhibition of AdoMet-dependent methylation, a portion of GFP-PRMT1 became immobilized in the nucleus, whereas diffusion of nuclear GFP-PRMT6 increased (78). The authors propose that PRMTs respond differently to the accumulation of unmethylated substrates, yet our results add another possibility that PRMTs respond differently to increased intracellular AdoHcy. The dimerization KD values for PRMT6 in the presence of either AdoMet or AdoHcy are respectively 6- and 10-fold higher for the 44 corresponding values for PRMT1 [Table 2.2]. As the major methyltransferase in cells (21, 79), PRMT1 may require a tight subunit interaction as a means to withstand changes in cellular methylation potential, whereas other PRMTs (e.g., PRMT6) may be more sensitive to different cofactor concentrations for regulatory purposes. 45 3 Analogues of the HIV-Tat peptide containing Nη -modified arginines as potent inhibitors of protein arginine N- methyltransferases 3 3.1 Introduction Divergent from the previous analysis in which the affinity of PRMT homodimerization was assessed, we embarked on designing new inhibitors and substrates for PRMT family members. To generate peptidomimetic inhibitors capable of specifically targeting PRMT6 (relative to other PRMTs and lysine methyltransferases) we chose to modify a fragment of the HIV-Tat protein known to be a PRMT6-specific peptide substrate. Working together, the groups of Wainberg and Richard recently identified the HIV-1 transactivator protein (Tat) as a unique substrate for PRMT6 (62). In follow-up studies using smaller Tat-derived peptides, the same authors further demonstrated that methylation occurs predominantly within the arginine rich motif (ARM) of Tat at R52 and to a lesser extent at R53. Single and double point mutations within the Tat-ARM sequence resulted in decreases in PRMT6-dependent HIV-1 repression, with R52K producing the largest impact (66). Strikingly, a comparative analysis of the methylation of full-length Tat by PRMTs 1, 3, 4, 5, 6, and 7 revealed an exquisite specificity for methylation by PRMT6. This specificity, coupled with the Tat peptide‟s intrinsic cell penetrating 3 A versioin of Chapter 3 has been published. Peter 't Hart*, Dylan Thomas*, Randy van Ommeren , Ted M. Lakowski , Adam Frankel and Nathaniel I. Martin. (2011) Analogues of the HIV-Tat peptide containing Nη- modified arginines as potent inhibitors of protein arginine N-methyltransferases. Med. Chem. Commun., 2012,3, 1235-1244. *These authors contributed equally to this work. Peter „t Hart and Randy van Ommeren of Dr. Nathaniel N. Martin‟s laboratory synthesized all substituted Tat peptides. Dr. Lakowski quantified final working stocks of Tat peptides and initial gel based activity assays. Dylan Thomas performed all substrate inhibition and oligomerization assays as well as data modeling. Dylan Thomas wrote parts of results and all of the discussion section. 46 ability(67), suggests that the HIV-Tat peptide itself may serve as a template for the design of new cell permeable, PRMT6-selective peptidomimetic inhibitors. HIV Tat based peptides were synthetically modified on a terminal nitrogen of R52 to probe their function as selective PRMT6 inhibitors. Although in vivo, R52 is the preferred methylation site, these peptides acted as substrates despite the modification due to the presence of surrounding arginines that become preferred post-synthesis. Despite this, steady-state kinetics revealed substrate inhibition for both PRMT1 and PRMT6 when using the modified Tat peptides as substrates. Apparent Vmax, Km and Ki values were generated revealing subtle differences dependant upon the type of nitrogen substitution. Using histone H3 tail peptide as a control, it was noted that small levels of substrate inhibition were present which had not been observed before in previous studies. This initiated a series of experiments comparing levels of substrate inhibition as a function of enzyme concentration. Kinetic constants should not change as a function of enzyme concentration as they are normalized by total enzyme present, however, it was shown that apparent Vmax increased as a function of enzyme concentration and that substrate inhibition was mitigated as enzyme concentration rose. This led to the rational conclusion that PRMT oligomerization was altering their kinetic constants and a model was proposed in which enzyme activity is a sum of activity produced by each individual oligomeric state involved in the reaction. 47 3.2 Methods Quantification of Tat peptides. Tat peptide samples corresponding to approximately 1.0 μM, initially estimated by weight, were dried in 300-μL (Waters, WAT094170) inserts and the tubes hydrolyzed in 6 N HCl at 110 ˚C for 24 h in vacuo and reconstituted in 0.5% acetic acid and 0.01% trifluoroacetic acid (TFA). Modified Tat peptides were quantified by measurement of total lysine using an Agilent Technologies 1290 Infinity HPLC with a (2.1 x 100 mm) Waters Acquity BEH C18 column connected to an AB SCIEX QTRAP 5500 mass spectrometer running at a flow rate of 0.15 mL/min at 45 ˚C. The mobile 0.5% acetic acid and 0.01% TFA was used isocratically for 2min and switched to 0.5% acetic acid, 0.01% TFA, and 30% methanol for an additional 1 min. Lysine was quantitated by multiple reaction monitoring for the precursor ion [M+H] 146.9m/z, and the product ion 84.1 m/z. Lysine standards were used between 400 and 5000 nM. Initial methylation of Tat peptides. Each Tat peptide at 250 μM was incubated at 37 ˚C overnight (16 h) with 150 μM [methyl-14C] AdoMet and 2.0 μM PRMT1, PRMT4 or PRMT6 in methylation buffer (50 mM HEPES 10 mM NaCl, 1.0 mM DTT, pH 8) in a 20 μL final volume. The reactions were terminated by addition of 5X tricine sample dilution buffer and the methylated peptides were separated by 17% tricine gel electrophoresis according to previously described methods (80). The gels were fixed with 5% glutaraldehyde and stained according to previous protocols (81). Similar to our experiences with other peptides we found that failure to fix the gels in this way resulted in leakage of peptides from the gel during fixing, staining, destaining and drying (82). Dried gels were exposed to storage phosphor screens (GE Healthcare) for 16 h and scanned on a Typhoon 9400 imager (GE Healthcare). The above reactions were repeated except that the source of methyl groups was unlabeled AdoMet. These 48 reactions were passed through a 30-kDa molecular weight cut off filter to remove the enzyme from the peptides, dried, and hydrolysed (see above). The amounts of aDMA and MMA were measured according to methods described below. Detection of methylation at substituted R52. We utilized MS to determine the potential for PRMT-mediated methylation at the modified R52 residue in the Tat peptide series with the same reaction samples of PRMT1, PRMT4, or PRMT6 with the Tat peptides described above. Using a series of product ion scans we selected masses corresponding to [M+H], [M+CH3+H] and [M+2CH3+H] for Tat peptide analogues 1, 3, 4, 5, 6, and 7 by scanning over the appropriate m/z range. Enzyme assays. All enzymes were expressed and isolated using previously described methods (17, 83). For substrate inhibition assays, mixes of enzyme (50-800 nM) and Tat peptide analogues 1-8 (0.25-200 μM) and 200 μM AdoMet (Sigma) were incubated at 37 ˚C in 1x reaction buffer (17). To keep all reactions within the linear range 400- and 800-nM enzyme reactions were incubated for 60 min and 50- and 100-nM enzyme reactions were incubated for 120 min. Samples were flash frozen in liquid nitrogen to stop catalysis, thawed, and spin-filtered at 12,000 x g at 4 ˚C using 30-kDa molecular weight cut-off filters (VWR 82031-354) for 15 min to separate the enzyme from the peptide substrate. Sample eluates were transferred into 300-μL glass inserts and dried using a Thermo Savant SC110A speed vacuum. The dried reactions were hydrolysed with 200 μL 6N HCl at 110 ˚C for 24 h in vacuo and reconstituted in 0.5% acetic acid and 0.01% trifluoroacetic acid (TFA). Reactions containing Tat analogue 4 were reconstituted in 0.5% acetic acid and 0.05% TFA. MMA and aDMA were separated and quantified using the same LC-MS/MS instrumentation described above. Liquid chromatography was performed for 5.5 min at 45 ˚C at 49 0.150 mL/min. For all Tat peptides except 4, buffer A contained 0.5% acetic acid and 0.01% TFA, and buffer B contained 30% methanol, 0.5% acetic acid and 0.01% TFA. Buffer A and B for 4 were identical except 0.05% TFA was used to help chromatographically separate N-ethyl- L-arginine from aDMA. Ions were acquired using a 30-V cone voltage at 400 ˚C. Fragments were generated using 20-meV collision energy. Precursor ions 203.1 and 189.2 m/z were selected corresponding to aDMA and MMA, respectively, and were quantified via multiple reaction monitoring using their generated 46.1 and 74.2 m/z product ions as previously described (17, 82). 50 3.3 Results Chemically modified Tat peptides are substrates for PRMT1 and PRMT6. To initially determine if the modified Tat peptides (Figure 3.1) were substrates for PRMT1, PRMT4 or PRMT6, these enzymes were used in radioactive methylation reactions with each peptide using [methyl- 14 C]-AdoMet as a source of methyl groups. The methylated peptides were separated using tricine gel electrophoresis and exposed to storage phosphor screens. The developed gels in Figure 3.2 A demonstrate that all enzymes exhibit methylation above background for all peptides; however, PRMT1 and PRMT6 show much higher levels of methylation relative to PRMT4. These results were corroborated in reactions with unlabeled AdoMet analysed using mass spectrometry (MS) to measure enzymatically-produced aDMA and MMA (Figure 3.2 B). The no-enzyme control groups produced no quantifiable methylarginine species by MS (data not shown), consistent with gel-based results in Figure 3.2 A (bottom gel). By MS we determined that PRMT1 and PRMT6 exhibited highest activity towards Tat-peptide analogues 2, 4, 7 and 8 in overnight methylation reactions, whereas PRMT4 was at least 10-fold less active than either PRMT1 or PRMT6. Generally, PRMT1 and PRMT6 produced more aDMA than MMA. In contrast, PRMT4 produced more MMA than aDMA for all peptides. 51 Figure 3.1. Chemical substitutions on the terminal omega nitrogen of arginine. Having determined that modified Tat peptides are robust substrates for PRMT1 and PRMT6, we examined if the N-modified R52 residues could be methylated. Using MS of hydrolysed methylation reactions of Tat peptides with and without PRMT1 or PRMT6, we could not detect any masses consistent with methylation of any of the substituted arginines. We could, however, detect the parent masses of the unmethylated, substituted arginines in all cases except for Tat-peptide analogues 6 and 7 (data not shown). These results are in contrast to our previous observation that PRMT1 can methylate an ethyl-substituted arginine residue within a peptide devoid of other arginine residues (82). The same substituted arginine residue in analogue 4 was not methylated by PRMTs in this study, suggesting that the presence of flanking, unmodified arginine residues presents more favourable targets for PRMTs. Chemically modified Tat peptides are substrate inhibitors for PRMT1 and PRMT6. Initial velocity reaction kinetics for PRMT1 and PRMT6 revealed a pattern consistent with substrate inhibition (39) within 0-to-40 μM concentrations of Tat peptide (Figure 3.3). This pattern was generated with 100 nM of PRMT1 or PRMT6 incubated with Tat peptides 1-8 in the presence of AdoMet. The curves were fit to a substrate inhibition model (Equation 3.1) and the derived apparent maximum velocities (Vmax), Michaelis-Menten (Km) and inhibition (Ki) constants for Tat peptides are displayed in Table 3.1. 52 Equation 3.1 The above expression is simplified and based upon substrate inhibition of an ordered bi- substrate system (84). Apparent Vmax, Km and Ki values generally varied within an order of magnitude for both PRMTs (the Ki for Tat peptide analogue 6 for PRMT6 is an exception). PRMT6 exhibited universally lower Vmax values compared to PRMT1 (Table 3.1), but interestingly both PRMT1 and PRMT6 demonstrated their respective highest activities towards Tat peptide analogue 4. 53 Figure 3.2. Methylation of modified Tat peptides. (A) Tat peptides 1-8 were methylated by PRMT1, PRMT4, and PRMT6 using [methyl- 14 C]AdoMet and detected on a storage phosphor screen after separating peptides from proteins via 17% tricine gel electrophoresis. A no-enzyme control was included. (B) The same experiment in (A) was repeated with unlabeled AdoMet. The peptides were hydrolysed and aDMA and MMA were detected using MS. Total methyl groups transferred were calculated as (2[aDMA]+[MMA]). 54 Inhibition patterns are biphasic at high substrate concentrations. Non-linear least square fits for substrate inhibition were derived using the 0-to-40 μM substrate range because at Tat peptide concentrations exceeding 40 μM, PRMT6 (and PRMT1 to a lesser extent) deviated from the model. For PRMT1 and PRMT6, the 100-µM point differed from the predicted value for all Tat peptides analysed and was consistently higher than a best fit would predict. Tat peptide analogue 8 in particular exhibited activity at a concentration of 100 µM for both PRMT1 and PRMT6 that matched or exceeded the methylation maximum at a 5.0-μM substrate concentration (Figure 3.3). The decision to fit in this range was made so that all experimental data could be compared. PRMT1 can be reasonably fit to substrate inhibition over the entire substrate range and result in slightly larger Ki values. Within the narrower substrate range of 0 to 40 µM, PRMT6 exhibited consistently higher apparent Ki values for all Tat peptides than did PRMT1, often differing by an order of magnitude for each respective Tat peptide (Table 3.1). Tat peptide analogue 6 was fit using substrate inhibition so that the data set could be compared to all other Tat peptides for PRMT1 and PRMT6, but it should be noted that the full substrate concentration range can also model well to a Michaelis-Menten-Henri hyperbolic fit. Both methods of fitting produce similar Vmax and Km values for analogue 6 with PRMT6 (Table 3.1 and data not shown), and the substrate inhibition model shows a large difference between Ki and Km values by at least two orders of magnitude. It is apparent from Equation 3.1 that when Ki is large, as in the case for analogue 6 with PRMT6, the relationship reduces to a standard Michaelis-Menten-Henri model, thus both equations yield similar Vmax and Km values. 55 Table 3.1. Apparent kinetic constants for PRMT1 and PRMT6 using a substrate inhibition model. PRMT1 PRMT6 Substrate Vmax a (pmol/min nmol) Km (μM) Ki (μM) Vmax a (pmol/min nmol) Km (μM) Ki (μM) Tat 1 28.1 (5.4) 12.2 (2.1) 2.68 (0.53) 2.85 (0.40) 3.32 (0.58) 19.9 (5.2) Tat 2 31.3 (0.2) 10.7 (0.1) 4.79 (0.20) 4.29 (0.23) 2.18 (0.09) 35.1 (3.0) Tat 3 36.9 (5.8) 25.6 (5.0) 2.98 (0.50) 1.86 (0.10) 2.07 (0.14) 43.3 (5.1) Tat 4 91.2 (7.3) 5.23 (0.59) 4.29 (0.45) 11.9 (0.6) 1.76 (0.25) 31.3 (6.2) Tat 5 28.3 (7.2) 6.59 (1.58) 7.62 (2.85) 1.39 (0.02) 0.838 (0.017) 77.3 (2.9) Tat 6 38.3 (26.2) 21.4 (13.3) 2.90 (2.12) 1.31 (0.05) 1.13 (0.13) >100 Tat 7 49.1 (6.3) 18.7 (2.8) 2.68 (0.47) 2.64 (0.02) 1.76 (0.07) 63.0 (1.8) Tat 8 11.2 (0.9) 3.92 (0.20) 7.60 (0.72) 1.63 (0.07) 0.751 (0.067) 35.9 (3.0) a Vmax values are normalized per nmol of enzyme to account for velocity differences as a result of changes in PRMT concentration. *Bracketed numbers indicate standard deviations. 56 Figure 3.3. Tat peptide methylation diverges from a hyperbolic fit and reveals substrate inhibition. Initial rate of total methyl groups transferred (pmol methyl group per min per nmol enzyme) is shown for 100 nM PRMT1 (A) and PRMT6 (B) using Tat peptide analogues 1-8 at 0, 0.25, 0.5, 0.75, 1.0, 5.0,15, 40 and 100 μM. Error bars represent standard deviations (n = 2). Non-linear least square fits were derived using data points from 0 to 40 µM as indicated by a solid line. To highlight deviation from substrate inhibition at a 100-µM substrate concentration, the fit was extrapolated as indicated by a dotted grey line. 57 PRMT concentration affects substrate inhibition. In contrast to methylation of Tat peptides, methylation of the histone H3 tail peptide (ARTKQTARKSTGGKAPRKQLAT- KAAW) with 100 nM PRMT6 revealed weak substrate inhibition within the concentration range used. Previous methylation of the H3 tail peptide modelled closely to a Michaelis-Menten-Henri hyperbolic fit rather than substrate inhibition. The only difference between the two experiments was enzyme concentration. To resolve this discrepancy, assays were performed with 400 nM PRMT6 (matching the original published methylation experiment of histone H3 tail peptide) with 0 to 100 μM histone H3 tail peptide or Tat peptide analogue 2. Analogue 2 was selected because it contains the native Tat sequence and modelled most closely to substrate inhibition for PRMT6 over the full concentration range of 0 to 100 μM. Despite being normalized by enzyme concentration, 400 nM PRMT6 showed an increase in Vmax compared to 100 µM PRMT6 and a trend towards a standard hyperbolic fit rather than substrate inhibition (Figure 3.4 A). This result supports the notion that enzyme concentration may be a key variable in modifying kinetic constants for PRMT6. To further investigate the effect of enzyme concentration on PRMT1 and PRMT6, enzyme activity assays were performed for Tat peptide 2 at 0-to-200 μM substrate concentrations with 50 to 800 nM PRMT1 or PRMT6 and 200 μM AdoMet. As the concentration of PRMT1 was increased from 50 to 800 nM enzyme, the apparent Ki fluctuated within an order of magnitude, however, the Km steadily increased, exceeding an order of magnitude difference between the 50 and 800-nM enzyme concentrations (data not shown). The apparent Vmax underwent an initial doubling from 50 to 100 nM enzyme, and then remained nearly constant at higher concentrations (Figure 3.4 B). Qualitatively, the curve at 50 nM PRMT1 most closely matched substrate inhibition up to 200 μM. As the enzyme concentration was increased to 800 58 nM PRMT1, the 100- and 200-μM substrate data points increasingly deviated from the fit, producing higher levels of methylation than substrate inhibition would predict. The effect of enzyme concentration on PRMT6 resulted in drastic changes in apparent Vmax, Km and Ki values. In doubling the enzyme concentrations from 50 to 100 nM PRMT6, a 50% decrease in apparent Ki was observed and remained relatively constant at higher PRMT6 concentrations (data not shown). Increasing PRMT6 concentration from 50 to 800 nM was met with over an 87-fold increase in the apparent Km and a 9-fold increase in the apparent Vmax (Figure 3.4 C). The greatest increase in apparent Vmax occurred between 100- and 400-nM enzyme concentrations. Substrate inhibition models offer little value in predicting activity for PRMT6 at Tat peptide concentrations exceeding 40 μM. Indeed, enzyme activities at 100- and 200-μM substrate concentrations were far in excess of the predicted values, and this effect was most prominent at 400 and 800 nM PRMT6. The data points at the higher PRMT6 concentrations would produce an excellent fit using a standard Michaelis-Menten-Henri relationship. 59 Figure 3.4. PRMT concentration affects substrate inhibition. (A) Initial rate of total methylation for histone H3 tail peptide (left) and Tat B peptide (right) is shown with 100 nM (♦) and 400 nM (■) PRMT6. In both cases, increasing enzyme concentration demonstrates a loss of substrate inhibition. Tat B was methylated with 50 nM (♦), 100 nM (■), 400 nM (▲) and 800 nM (●) PRMT1 (B) or PRMT6 (C). Graphed to the right are the apparent Vmax values for each corresponding enzyme concentration. All curves are modelled using a substrate inhibition model from 0 to 40 µM as indicated by a solid line and extrapolated to 200 µM to highlight deviation from the fit as indicated by a dotted grey line. 60 3.4 Discussion Tat-ARM derived peptides as PRMT inhibitors. Inspired by published reports describing HIV-Tat as a selective substrate for PRMT6 (62, 66), we designed and synthesised a number of truncated Tat peptide analogues bearing N-substitutions at R52, the arginine residue known to be methylated. These analogues were generated with the aim of developing PRMT6 selective peptide inhibitors. The results obtained are intriguing as each of the peptides was found to be readily methylated by both PRMT1 and PRMT6 at other arginine residues present in the peptide. Furthermore, at the appropriate concentration, these peptides were also shown to behave as substrate-inhibitors of both PRMT1 and PRMT6 (discussed in greater detail in following section). The observation that truncated Tat-ARM peptide analogues are readily methylated at arginine residues other than R52 by both PRMT1 and PRMT6 was somewhat unexpected considering the highly specific N-methylation of R52 by PRMT6 (66). These results suggest that in order to achieve PRMT6 selectivity, the full-length HIV-Tat peptide may be required. As can be seen in the solution structure of HIV-Tat (85), the side chain of R52 is solvent exposed and likely resides in a unique steric and electronic environment, possibly contributing to PRMT6 selective methylation. By comparison, the arginine residue corresponding to R52 in the linear 13-mer Tat-ARM analogues used in the present study are not likely to possess the same unique properties relative to the other arginines present. To address this issue, one approach that may be of value in future investigations could involve examining the use of cyclized peptidomimetics of the Tat-ARM. Cyclization of the Tat-ARM is expected to restrict its conformation and may place the R52 residue in an environment that more accurately mimics that found in full-length Tat. Such cyclic Tat-ARM analogues have been previously described in the literature and are known 61 to maintain the functional characteristics and cell permeability of full length HIV-tat.32 It may therefore be of value to evaluate the methylation behaviour of the various PRMTs towards cyclized mimetics of the Tat-ARM. PRMT mechanism for Tat peptide substrate inhibition. The previously characterized enzyme mechanisms for PRMT1 and PRMT6 provide a rational explanation for how substrate inhibition may occur. Presented in Figure 3.5 is the Cleland notation for an ordered bi-substrate reaction with peptide substrate-generated inhibition. Two abortive complexes are possible as a result of high levels of peptide substrate. The first abortive complex that can form is EBB in which one peptide (B) invades the AdoMet-binding pocket and prevents cofactor (A) from binding while another peptide binds within the peptide-binding groove (39). Typically, EBB complexes occur when structural similarities exist between substrates A and B, as well as between the sites in which they occupy (39). In the case of PRMTs, the two substrates have dramatically different structures, with the peptide being far too large to fit appropriately within the AdoMet-binding pocket (2, 25). The second abortive complex is EBQ in which a peptide (or methylated peptide) docks within the peptide binding groove before AdoHcy (Q) release can occur and prevents the enzyme from entering another catalytic cycle (39). Inhibition can also occur as a result of the peptide docking within the binding groove in an atypical conformation so that the arginine residue is not properly positioned to accept a methyl group. Structural evidence supports a model in which EBQ substrate inhibition can occur by locking AdoHcy within its binding pocket. The N-terminal X and Y helices present in structures of PRMT3, PRMT4, and PRMT5 fold over and trap AdoHcy within its binding pocket, as well as form an adjacent peptide binding groove (3, 4, 24, 29). These structural studies suggest that it is possible for a methylated peptide product to dissociate from the enzyme while AdoHcy continues to occupy 62 the active site and accept another peptide substrate within the binding groove to produce an abortive EBQ complex (Figure 3.5). In support of this model, binding of a pyrazole-based inhibitor to PRMT4 within this peptide-binding groove required the presence of AdoHcy, suggesting that occupancy of the AdoMet-binding pocket is a prerequisite to peptide binding (17, 29). Additionally, an ordered bi-substrate mechanism has been proposed for PRMT1, PRMT3 and PRMT6 via direct quantification of MMA and ADMA using mass spectrometry (35, 36). Under this model, tight association of a methylated peptide may prevent dissociation of AdoHcy since the peptide must dissociate before AdoHcy can be exchanged for AdoMet. However, a random bi-substrate mechanism has also been proposed for PRMT1 and PRMT6 using densitometry of radioactive products on gels (37, 38). Under this system, substrate inhibition may still occur through docking of an incompatible peptide conformation within the active site. Continued efforts in defining the structural and mechanistic features of PRMTs will help to establish a route for designing methyltransferase-selective inhibitors. The Tat-based peptide sequence is highly electropositive with multiple arginine and lysine residues, and this feature is expected to be ideally suited for binding to electronegative grooves such as those found on the PRMT1 enzyme surface (2). Interestingly, most of the chemical modifications to R52 in the Tat peptide did not affect its ability to act as a substrate inhibitor for PRMT1 and PRMT6, which is reflected in the consistencies among Ki values (Table 3.1). This behaviour may suggest an alternate method of substrate inhibition where the Tat peptide binds to one of the electronegative grooves, resulting in a conformational change or partial occupancy of the active site that negatively impacts enzyme activity. A notable exception to this trend was Tat analogue 6 for PRMT6, which showed a hyperbolic fit rather than substrate inhibition. The cyclopropyl substituent on 6, which adopts a constrained and bent conformation, 63 disrupted the ability of the peptide to form an EBQ complex with PRMT6, but did not prevent the formation of an EAB complex that led to substrate methylation. This unique kinetic behaviour for 6 may point to a structural difference between PRMT6 and PRMT1. Figure 3.5. Substrate inhibition mechanisms for a sequential ordered bi-substrate reaction. The PRMT reaction mechanism is shown using standard Cleland notation where representations of substrates and reactants are as follows: A is AdoMet, B is peptide, P is methylated peptide, and Q is AdoHcy. Substrate inhibition can occur through the formation of EBB or EBQ abortive species. EBB lacks cofactor required for substrate methylation, and EBQ cannot take part in another round of methylation without exchanging Q for A. Both species are considered abortive complexes as denoted by asterisks. Oligomerisation as a regulator of activity. The changes in kinetic constants as a function of enzyme concentration that we observed (Figure 3.4) may be facilitated by the generation of multiple enzyme species through oligomerisation. Feng and co-workers demonstrated that when using histone H4 as a methylation substrate, the apparent kcat increased as PRMT1 concentration increased (34). Additionally, our group showed an increase in the apparent Vmax of PRMT1 by increasing the concentration of PRMT2 or catalytically inactive PRMT1 (26). In the current study, the apparent Vmax values for PRMT1 and PRMT6 showed maximal change (Figure 3.4 B-C) when transitioning through the dissociation constants (Kd values) previously measured at 110 and 210 nM, respectively (28). These results imply that PRMT kinetic behaviour is, in part, dictated by the oligomerisation state of the enzyme. 64 An interesting property of the data sets is the consistently high enzyme activity at Tat peptide concentrations of 100 and 200 μM following substrate inhibition. Such a phenomenon could occur through formation of two or more separate enzyme species in solution that possess different kinetic constants for which methyltransferase activity would be a result of the additive contributions of all catalytically active species in solution. Thus we propose a mechanism in which one species (monomer) is subject to substrate inhibition as evidenced by such a pattern at low enzyme concentrations. At high enzyme concentrations another species (oligomer) may predominate with its own set of kinetic constants. A mathematical model for such a system is shown below: Equation 3.2 This model sums the activities of all species based on their relative proportion, with the assumption that only the monomer is affected by substrate inhibition. The initial term from Equation 3.1 for substrate inhibition is multiplied by the fraction of enzyme in a monomeric state. The second term is the sum of activities for all oligomers formed, multiplied by their fractional presence as dictated by their KD values. E represents the total concentration of enzyme in solution. While the model expressed in Equation 3.2 represents one possibility and can accurately fit the data sets, a number of issues must be considered. First, different values for the kinetic constants can result in an excellent fit, but may not accurately represent the system. Second, it may not be feasible to isolate an enzyme species at a specific concentration, solve for dissociation and kinetic constants, and then use those values to solve a missing part of the equation under different experimental conditions. Third, dimerization has been shown to play a 65 role in activity (2, 26, 32-34) so the question remains as to whether formation of an inactive EBQ complex can alter the ability of an enzyme to oligomerise. Such a dynamic system likely has many interdependent variables, including more than those described in this paper, such as the kinetic aspects of oligomerization. Lastly, we have assumed only the monomer is affected by substrate inhibition because this phenomenon is most prevalent at low enzyme concentrations. The interplay between substrate and enzyme may be fluid and related to exchange between monomer and oligomer through mechanisms listed above. 66 4 N-substituted arginyl peptide molecular probes reveal an activated substrate for protein arginine N-methyltransferase activity 4 4.1 Introduction In Chapter 3 it was demonstrated that HIV Tat peptide containing a terminally substituted R52 residue acted as substrate inhibitors for PRMT1 and PRMT6. Additionally, apparent Km and Vmax for PRMT1 and PRMT6 changed as a function of enzyme concentration when methylating these substrates, producing evidence that oligomerization may not only mitigate substrate inhibition but each enzyme oligomer has its own unique kinetic constants. A mathematical relationship was postulated that states that the total methyltransfer observed is the sum of the kinetic contributions of monomer and all higher order oligomers in solution, and that only the monomer is susceptible to substrate inhibitions. Deviating from Tat peptide, I decided to undertake a new study analyzing the effect of substrate pKa on methyltransfer for PRMT1, 4 and 6. The mechanism of methyl transfer by PRMTs has recently been under scrutiny, and although work has been done attempting to characterize the chemistry involved, a consensus on the mechanism has not yet been reached. A number of arginine substitutions were made to probe how pKa differences modulate PRMT activity to better understand limitations of the enzyme‟s chemistry. Hydroxy substituted arginine revealed an unusually high apparent kcat for PRMT1, positioning it as a better substrate for PRMT1 than monomethyl or unsubstituted arginine. Kinetic isotope effect experiments using 4 All ωN-substituted R1 peptides were synthesized in Dr. Nathaniel N. Martin‟s lab by Timo Koopmans and Helmi Kreinin. Dr. Ted M. Lakowski performed MS 2 analysis on proteinase K treated R1-peptides and IC50 analysis on PRMT1. Mynol Islam Vhuiyan generated the PRMT D51N expression plasmid. Dr. Jennifer Bui performed Gaussian electrostatic potential maps. Dylan Thomas expressed and purified PRMT1 D51N, PRMT6, performed all kinetic isotope effect experiments, gel assays, infusion based mass spectrometry and IC50 analysis for PRMT4 and 6. 67 various substituted arginine peptides revealed different rate limiting steps for PRMT1 and PRMT6, implicating water in the catalytic reaction. Methylation of hydroxyarginine by normally inactive variants of PRMT1, E153Q and D51N, solidify the hydroxy substitution as an activated substrate analogue that reveals water as an important and rational player in the previously proposed His-Asp proton relay mechanism of arginine methylation. Additionally, nitroarginine was identified as a promising arginine substitution for general PRMT inhibition. 68 4.2 Methods Peptide quantification. All R1 peptides were quantified spectrophotometrically using an extinction coefficient of 11,000 M -1 cm -1 at a 280-nm wavelength. The extinction coefficient of the peptide was calculated using the SIB Swiss Institution of Bioinformatics ExPASy ProtParam tool based upon the theoretical values for tryptophan absorbance. This method has been previously used to quantify R1 based peptides (82). 14 C Methylation and tricine gel separation of the R1 peptide series. Each R1 peptide at 250 µM was incubated in methylation buffer (50 mM HEPES pH 8.0, 10 mM NaCl, 1 mM dithiothreotol (DTT)) (17) along with 100 µM S-adenosyl-L-[methyl- 14 C]methionine and 2 µM of PRMT1, 4, 6 or PRMT1 D51N for 16 h at 37 ˚C. These enzymes were isolated as previously described (17, 28). Reactions were stopped by introduction of 5x tricine gel loading buffer and separated electrophoretically on 17% tricine gels (80). Gels were fixed using 5% glutaraldehyde (Sigma-Aldrich) according to previous methods to prevent peptide diffusion out of the gels (86). The gels were then stained with Coomassie Blue, dried and exposed to a storage phosphor screen (GE healthcare) for 24-72 h before development on a Typhoon 9400 imager (GE Healthcare) at 100-µm resolution. Enzyme kinetics & kinetic isotope effects. Enzyme concentrations of 200 nM PRMT1, 400 nM PRMT4, 400 nM PRMT6, and 200 nM PRMT1 D51N were incubated at 37 ˚C with 200 µM AdoMet and 0, 1, 2, 5, 10, 20, 40, 100, and 200 µM R1-OH, R1-NH2 and R1-CH3 peptide. Reactions were performed for 60 min (PRMT1), 90 min (PRMT4 and PRMT6), and 120 min (D51N) in methylation buffer with H2O or D2O (final solvent composition of >95% D2O) for KIE experiments prior to flash freezing. After thawing samples, AdoHcy and peptide substrates were spin filtered through 12,000-Da molecular weight cut-off flat bottom filters (VWR 82031- 69 354) at 4 ˚C for 10 min. The filtrate was loaded into 300-µL glass inserts (Waters, WAT094170) and samples were analyzed via LC-MS/MS on an ABSciex QT5500 mass spectrometer. Separation and quantification of AdoHcy was achieved utilizing a Waters Acquity UPLC BEH C18 reverse phase column (2.1 x 100 mm) with buffer conditions previously described (87). Peptide methylation analyses. In order to determine methylation of peptides bearing heteroatom substituents, 250 µM R1-OH or R1-NH2 was incubated with 100 µM AdoMet and 2 µM PRMT1 in methylation buffer for 16 h at 37 ˚C. After filtering samples as described above, a 1/100 dilution of each filtrate was made in 0.01% TFA and 0.5% acetic acid and infused into an ABSciex QT5500 mass spectrometer using a Supelco 2 mL HPLC syringe. Q1 scans in positive ion scanning mode were performed from 600 to 700 m/z range to acquire doubly charged peptide masses. To determine the position of methylation, 300 µM of R1-NH2 or R1-OH and 250 µM AdoMet were incubated in methylation buffer with and without 150 nM PRMT1 for 24 h at 37 ˚C. After incubation, proteinase K (1.0 mg/mL) was added to a final concentration of 3.3 µg/mL and further incubated for an additional 24 h at 37 ˚C to digest the peptide into its amino acids. Protease was removed by passage of the reaction mixtures through 30,000 NMWCO centrifugal filters and amino acids were recovered. MS 2 spectra were recorded selecting precursor ions 190.12, 191.11 and 204.1, 205.13 m/z, corresponding respectively to unmodified N- aminoarginine and N-hydroxyarginine as well as their methylated forms, using a declustering potential of 55 V and collision energy of 18 eV scanning from 30 to 220 m/z. IC50 progress curves. PRMT1, 4, and 6 were pre-incubated at a concentration of 200 µM with 0, 1, 2, 5, 10, 25, 50, 100, and 300 µM R1-NO2 or R1-canavanine (R1-can) in methylation buffer for 1 h at 37 ˚C. Reactions were initiated by introducing 15 µM H4 tail 70 peptide (SGRGKGGKGLGKGGAKRHRKVW) and 10 µM AdoMet for PRMT1, 20 µM H3 tail peptide (ARTKQTARKSTGGKAPRKQLATKAAW) and 20 µM AdoMet for PRMT4, and 5 µM H3 tail peptide with 20 µM AdoMet for PRMT6; concentrations were varied around their respective Km values for each enzyme (83). After incubation for 1 h at 37 ˚C, reactions were processed as described above and analyzed via LC-MS/MS using previously described methods for AdoHcy quantification (87). Since PRMT4 and PRMT6 reactions could not be accurately modeled with AdoHcy quantification due to background automethylation, these samples were processed to quantify methylated arginines. The samples were dried and hydrolyzed in the vapor phase using 6 N HCl for 24 h at 110 ˚C in vacuo. Hydrolyzed samples were resuspended in 0.01% trifluoroacetic acid and 0.5% acetic acid, and MMA and aDMA were quantified via LC- MS/MS using previously described buffer conditions and multiple reaction monitoring protocols (17). 71 4.3 Results Methylation of the R1 series of peptides. Peptides based on the previously described R1 sequence were synthesized in which the single arginine residue was modified at the guanidino group with the aim of altering its pKa and size (Figure 4.1 and Table 4.1) (35). Initially, we tested the propensity of these peptides to act as substrates for PRMT1, 4, and 6 in methylation reactions containing [methyl- 14 C]-AdoMet. As shown in Figure 4.2, all wild-type enzymes methylated R1-OH, R1-NH2, unmodified R1, and R1-CH3, whereas methylation above background was detected for R1-can. R1-ethyl was weakly methylated by PRMT1 only, consistent with our previous findings (82). Lastly, methylation of R1-NO2 was not detected. These results indicate that steric and pKa differences among arginine analogues account for the ability of these peptides to be methylated. Figure 4.1. ωN-substituted R1 peptides. The sequence of the R1 peptide is indicated along with the seven R1-peptide variants used in the present study: (1) R1-nitro (R1-NO2), (2) R1-canavanine (R1-can), (3) R1-hydroxy (R1-OH), (4) R1-amino (R1-NH2), (5) R1-native (R1), (6) R1-monomethyl (R1-CH3), and (7) R1-ethyl. The associated pKa values for each respective substituted arginine can be found in Table 4.1. 72 Table 4.1. Physicochemical properties of arginine and its derivatives. Amino Acid pKa a Volume difference (Å3) b Arginine 13.88 ± 0.70 (12.48) (88) 0 N-Nitroarginine 3.83 ± 0.50 +24.448 Canavanine 8.01 ± 0.70 (7.01) (89) -7.817 N-Hydroxyarginine 8.68 ± 0.69 +6.432 N-Aminoarginine 11.09 ± 0.70 +9.703 N-Monomethylarginine 14.06 ± 0.70 +17.675 N-Ethylarginine --- +34.477 a The listed pKa values were obtained using Advanced Chemistry Development Software v11.02 calculated for the following CAS registry numbers: 74- 79-3, 149967-90-8, 282727-72-4, 133374-43-3, 57444-72-1, and 17035-90-4 unless otherwise indicated. b Volume differences using arginine as the standard were calculated using Molinspiration property engine v2011.04 In addition to wild type enzymes, D51N and E153Q variants of PRMT1 were tested to assess whether methylation of the R1 peptide series was affected by amino acid residues in the active site supposedly involved in arginine binding or catalysis (2). D51 is purported to play a role in a catalytic His-Asp dyad for deprotonation of the substrate arginine guanidino group, and E153 is positioned to form stabilizing hydrogen bonds to the arginine guanidino group (2, 3). Previous studies have shown that D51R and E153Q substitution variants abolish enzyme activity (2, 26, 30, 32, 34). We found that D51N PRMT1 demonstrated compromised methylation of all peptides compared to wild type PRMT1. However, reactions containing R1-OH and R1-CH3 exhibited the highest levels of methylation for this variant. PRMT1 E153Q showed above background levels of methylation utilizing R1-OH as a substrate, highlighting the hydroxyl modification as a possible activated species or transition state analogue capable of partially mitigating the loss of function in the E153Q substitution variant (Figure 4.2). 73 Figure 4.2. Coomassie staining and phosphor imaging of radiolabeled R1-peptides. Radioisotopic labeling of R1-nitro, -canavinine, -hydroxy, -amino, -native, -monomethyl and -ethyl (lanes 1-7) with [methyl- 14 C]-AdoMet via PRMT1 (A), PRMT4 (B), PRMT6 (C), PRMT1 D51N (D), PRMT1 E153Q (E), and no enzyme control (F). Shown below in each panel is the Coomassie Blue-stained gel corresponding to the phosphor image directly above. 74 Monomethylation of R1 peptides at heteroatom-substituted N atom. We have previously shown that PRMT1 transfers a single methyl group to the substituted N atom on the R1-ethyl peptide (82). In order to determine the site of PRMT1-catalyzed methylation for R1-OH and R1-NH2 peptides, we performed methylation reactions with and without enzyme to analyze the methylated and unmethylated peptides by mass spectrometry (Figure 4.3, panel A and B). Single MS electrospray ionization (Q1) scans from 600 to 700 m/z identified doubly charged base-peaks at 659.8 m/z [M+2H] for unmethylated R1-OH, and 666.9 m/z [M+2H] for methylated R1-OH (the +7 amu difference is attributed to a single methyl addition for the doubly charged peptide). Similar results were obtained for R1-NH2. In order to investigate which N atom on R1-OH and R1-NH2 peptides PRMT1 monomethylates, MS 2 experiments were performed on both the unmethylated and methylated peptides. In order to confirm the position of methylation in both the R1-OH and R1-NH2 peptides, we incubated each with AdoMet in the presence or absence of PRMT1, and digested the peptides into component amino acids using proteinase K. The resulting preparations produced unmethylated N-hydroxyarginine and N-aminoarginine in the no-enzyme control group and monomethylated forms of N-hydroxyarginine and N-aminoarginine in the PRMT1 group. Despite applying several fragmentation conditions to assess the methylated N-hydroxyarginine sample, we were unable to identify the specific location of methylation within the N- hydroxyarginine moiety. In contrast, the MS 2 data obtained when examining the digests of the R1-NH2 peptide showed product ions (M+H) in the enzyme treatment group consistent with the formation of methyl-aminoarginine (Figure 4.3, panel C and D). Several peaks in the control spectrum increase by 14 m/z in the enzyme treatment group, indicating the addition of a single methyl group to the aminoarginine upon incubation with PRMT1. The peak at 74.9 m/z 75 corresponds to the amino-guanidino group in the control spectrum, which is replaced by an 89.1 m/z peak that is larger by 14 m/z, confirming the presence of a methyl group on the amino- guanidino region of aminoarginine. Previous studies have shown that the guanidine unit of substituted arginines is a commonly observed fragment in MS 2 spectra (17, 82). The presence of a peak at 47 m/z in the spectrum of the PRMT1 treatment group suggests the formation of methyl-hydrazine, indicating that the methyl group resides on the N atom or on the N-amino substituent itself. 76 Figure 4.3. Interogation of monomethylated R1-NH2 and R1-OH,using MS. Mass spectrometry of the methylated R1-OH (A) and R1-NH2 (B) peptides are shown in blue. No-enzyme controls are displayed in red in the background for comparison. The MS 2 spectra of unmodified aminoarginine (C) and methyl-aminoarginine (D) derived from enzymatic digestion with proteinase K of R1-NH2 in the presence and absence of PRMT1 are shown. The structures of selected product ions are illustrated above or adjacent to the corresponding peaks with the transferred methyl group in bold. The precursor ion mass-to-charge ratios are listed above the corresponding spectra. Spectral intensity is scaled to percent intensity. 77 Enzyme kinetics and kinetic isotope effect analysis. From our initial evaluation of the R1 peptide series, we found that PRMTs were active toward both natural and unnatural substrates. These findings provided us with the opportunity to perform structure-activity relationship studies and gain mechanistic insight into PRMT-catalyzed methylation by comparing kinetic constants under steady-state conditions for PRMT1, PRMT4, and PRMT6. The natural substrate R1-CH3 was used along with heteroatom-modified R1-OH and R1-NH2 isosteres in order to compare kinetic constants for single methylation events. Data were generated in parallel under H2O and D2O solvent conditions to determine KIEs for the above enzyme-substrate combinations (Figure 4.4). A summary of all derived apparent kinetic constants can be found in Table 4.2. 78 Figure 4.4. Michaelis-Menten curves for PRMT1, PRMT4 and PRMT6 in H2O and D2O. Steady-state enzyme kinetic curves are shown for PRMT1 (A), PRMT6 (B), PRMT4 (C) and PRMT1 D51N (D) with 0 to 200 µM R1-OH in H2O (●) and D2O (■), R1-NH2 in H2O (♦) and D2O (▲), and R1-CH3 in H2O (▼) and D2O (X). Dashed red lines indicated best fits of experiments performed in D2O. 79 Table 4.2. Apparent kinetic constants for PRMT1, PRMT6 and PRMT1 D51N using R1- based peptides. Solvent Enzyme Substrate Km (µM) kcat (min -1) kcat/Km ∙ 10 3 (min-1∙M-1) KIE° (kcat/Km)H/ (kcat/Km)D KIE° (kcat)H/(kcat)D Rate-limiting step a H2O PRMT1 R1-OH 93.1 (7.5) b 1.49 (0.5) 16.1 (8.1) 0.935 (11.5) 2.74 (7.9) product dissociation D2O R1-OH 31.7 (16.1) 0.543 (7.9) 17.2 (8.1) H2O R1-NH2 41.1 (32.6) 0.512 (8.6) 13.0 (24.6) 2.23 (26.5) 1.78 (14.1) catalysis D2O R1-NH2 49.6 (21.0) 0.287 (11.2) 5.84 (9.9) H2O R1-CH3 13.9 (12.2) 0.220 (3.1) 15.9 (8.8) 2.88 (9.1) 1.15 (3.2) substrate association D2O R1-CH3 34.5 (1.7) 0.191 (0.6) 5.54 (2.3) H2O D51N R1-OH 157 (21.0) 0.00511 (15.5) 0.0330 (36.4) 0.687 (40.0) 0.373 (16.5) catalysis D2O R1-OH 245 (19.6) 0.0137 (5.8) 0.0481 (16.6) H2O PRMT6 R1-OH 462 (21.0) 0.0162 (2.0) 0.0361 (19.4) 0.200 (19.5) 0.506 (2.2) catalysis D2O R1-OH 177 (2.6) 0.0320 (1.0) 0.180 (1.7) H2O R1-NH2 100 (52.0) 0.00475 (24.0) 0.0516 (31.0) 0.465 (38.8) 0.910 (24.1) substrate association D2O R1-NH2 48.3 (22.2) 0.00522 (1.7) 0.111 (23.4) H2O R1-CH3 306 (5.6) 0.0433 (6.6) 0.142 (1.4) 0.356 (9.1) 0.421 (7.0) catalysis D2O R1-CH3 260 (11.1) 0.103 (2.3) 0.398 (9.0) a Analysis to determine rate limiting step is drawn from Northrop, 1975 (90). b Numbers in parentheses show coefficients of variation. For enzyme kinetics of PRMT1, PRMT4, and PRMT6 performed in H2O, each enzyme demonstrated varying behavior in methylating the modified R1 peptides. No apparent kinetic constants could be generated for PRMT4 as none of the reactions could be saturated within 0 to 200 µM peptide concentrations. In the concentration range analyzed, R1-CH3 demonstrated the highest levels of methyl transfer for PRMT4. PRMT1 displayed the highest kcat and Km for R1- OH, resulting in a specificity constant (kcat/Km) that was similar for R1-CH3. PRMT6 demonstrated a clear preference for R1-CH3 with a higher kcat and specificity constant compared to R1-NH2 and R1-OH. R1-CH3 remained an excellent substrate for all tested PRMTs, yet R1- OH demonstrated a much larger kcat for PRMT1 than with any other enzyme-substrate pair (Table 4.2). 80 Kinetic constants derived in D2O differed greatly from those found using H2O as a solvent. In D2O, the kcat and Km for R1-OH with PRMT1 both decreased in concert, resulting in a largely unchanged specificity constant. In contrast, R1-CH3 suffered an increase in Km while retaining a similar kcat, yielding a lower specificity constant. Using PRMT6 in D2O, all substrates produced an increased kcat with a concomitant decrease in Km relative to activity in H2O, resulting in increased specificity constants and inverse KIEs in all cases. The efficiency of R1-OH methylation by PRMT1 was unaffected by D2O, whereas PRMT6-mediated methylation efficiency of all R1-peptides improved in D2O (Table 4.2). Unlike the E153Q variant, PRMT1 D51N had sufficient activity to be quantified using our mass spectrometry assay to derive apparent kinetic constants. Although a saturating concentration of R1-CH3 for the D51N variant could not be attained within a concentration range of 0 to 200 µM, apparent kinetic constants could be calculated for R1-OH in both H2O and D2O. In these experiments an unexpected change was observed within the concentrations tested wherein R1-CH3 demonstrated the highest levels of methyl transfer instead of R1-OH as seen for wild type PRMT1. Additionally, methyl transfer using R1-OH as a substrate improved in the presence of D2O as demonstrated by an inverse KIE. This data underscores that D51 is more important for methylation of R1-OH than R1-CH3. The results above highlight some key differences in methyl transfer when using modified R1 peptides as substrates. PRMT1, PRMT6, and PRMT1 D51N have varying rate-limiting steps evidenced by the presence of KIEs in different kinetic rate constants. Additionally, R1-OH produced the highest levels of methyl transfer only for wild type PRMT1. PRMT1 D51N and PRMT6 exhibited the highest activity when utilizing R1-CH3 as substrate. Together, these results reveal that water is likely to be involved in methyl transfer. 81 N-substituted peptides as type I PRMT inhibitors. No methyl transfer was observed when incubating PRMT1, PRMT4, or PRMT6 with R1-NO2 and only minimal methylation was seen with R1-can. Given these results, both peptides were thus assessed as competitive inhibitors of these PRMTs. IC50 curves were generated using a range of 0 to 300 µM R1-NO2 or R1-can peptides in reactions containing histone H4 tail peptide (for PRMT1) or H3 tail peptide (for PRMT4 and 6) and AdoMet. All reactions were initially tested for AdoHcy production to model IC50 values; however, PRMT4 and PRMT6 produced unusually high levels of automethylation. To circumvent this problem, MMA and aDMA were measured as described in the Materials and Methods section. R1-can demonstrated inhibition of PRMT1 (IC50 = 30.4 µM), and was a poor inhibitor of PRMT4 and PRMT6 (data not shown). This is not surprising as PRMT1 demonstrated weak methylation of R1-can, validating its ability to appropriately dock within the active site. However, R1-NO2 demonstrated inhibition for PRMT1, PRMT4, and PRMT6 with IC50 values of 21.2, 47.2, and 44.9 µM, respectively (Figure 4.5). Given these results, incorporation of N-nitroarginine in place of arginine in PRMT substrate peptides may represent a general approach for designing PRMT inhibitors. Inhibition of all three PRMTs by R1-NO2, whose N-nitro substitution is relatively larger in size and bears the lowest guanidino group pKa than other arginine analogues (Table 4.1), was surprising since it lacks the positive charge characteristic of PRMT-binding peptides (2). Also of interest is the finding that R1- NO2 inhibits PRMT4 as most R1 peptide-based inhibitors prepared to date have demonstrated poor PRMT4 inhibition (82, 83). It should also be noted that in addition to the low guanidino group pKa, the sterics of the N-nitro substitution may also play a role in preventing methylation of R1-NO2 by PRMTs. It can be concluded from these results 82 that inhibition of all three PRMTs by R1-NO2 demonstrates that positive charge on the substrate arginine is not a prerequisite for docking to the enzyme. Figure 4.5. The 4-parameter sigmoidal fits for PRMT1, PRMT4 and PRMT6. Methylation activity plots with 4-parameter sigmoidal fits used to generated IC50 values were overlayed with the averaged data for PRMT1 inhibited by 0 to 300 µM R1-NO2 (▲) and R1-can (X), PRMT4 inhibited by 0 to 300 µM R1-NO2 (■), and PRMT6 inhibited by 0 to 300 µM R1-NO2 (●). Plots are displayed as a percent of maximum activity from AdoHcy production (PRMT1) or total methyl transfer (PRMT4 and PRMT6). 83 4.4 Discussion The kinetic isotope effect reveals differences in the rate limiting steps involving water. PRMT1, PRMT1 D51N, and PRMT6 all showed unique KIEs when using D2O as solvent. Although it may appear that these enzymes utilize water differently, a more plausible explanation is that the observed KIEs are reflected in different steps of the enzymatic reaction. Steady state analysis of KIEs gives useful information on the rate-limiting step of a reaction in which a bond- breaking event occurs with the isotope being utilized. In this case, water is proposed to be both a solvent and possibly a reactant, playing roles in both substrate association and product dissociation, as well as in the reaction chemistry. D2O has the additional capacity to exchange deuterium with arginyl protons under assay conditions used, creating the possibility of KIEs generated from direct perdeuteration of guanidino N atoms (91). In this system, the catalytic step may be hidden depending on whether D2O is involved in substrate association, enzyme catalysis, or product dissociation. Given that oligomerization and N-terminal αXY-helical motions are also critical components of the catalytic mechanism (4, 29), secondary KIEs in which D2O stabilizes high mobility components of the enzyme must also be considered (92). Despite such complicating factors, a method of dealing with “partial rate-limiting systems” was proposed by Northrop in which kinetic constants are grouped into the three major events: substrate binding (ka), catalysis (kb), and product dissociation (kc) (90). Although KIEs are an amalgam of overall effects on the entire enzyme mechanism encompassing all three kinetic terms, they can be compared to assess which steps are rate limiting and if a deuterium- oxygen cleavage or formation step is occurring. Since (Km)H/(Km)D, (kcat)H/(kcat)D, or (kcat/Km)H/(kcat/Km)D share rate constants, they may change in concert with one another, thus obscuring the steps affected by the KIE. A series of diagnostic criteria were proposed to solve 84 this issue in comparing kcat/Km KIEs to those of kcat alone. When kcat/Km experiences a KIE and kcat does not, this criteria indicate perturbations in substrate association rate constants unique to Km, and manifest as an apparent change in substrate binding. If both kcat/Km and kcat experience KIEs, then catalysis is concluded to be the rate-limiting step. If kcat/Km experiences no KIE while kcat does, then the effect is unique to product dissociation rate constants that are found within kcat and indicates rate-limiting product release (90). The guidelines described above are generally applicable when considering the PRMT1- mediated methylation of R1-CH3 and R1-OH, wherein substrate association and product dissociation steps display KIEs and thereby imply that they are rate limiting. By comparison, the R1-NH2 peptide demonstrated a KIE in catalysis, implicating D2O in a bond-breaking event. This observation is important as it provides some evidence against the possibility of a secondary KIE that stems from increased structural stabilization in D2O. Since PRMT1-catalyzed methylation is faster than for other PRMTs (17, 36, 38, 47, 86), it is perhaps not surprising that catalysis is efficient enough to be “invisible” to the KIE for R1-CH3 and R1-OH. An inverse KIE was observed with PRMT6 in kcat and kcat/Km for all substrates used; a drastically different behavior than what was observed for PRMT1. Catalysis appeared rate limiting for PRMT6 with R1-OH and R1-CH3, and methyl transfer universally improved in D2O. The presence of an inverse KIE suggests that the deuterium is stabilizing a high-energy bond during catalysis, resulting in a transfer that is energetically favorable. Spectroscopic studies investigating deuterium interactions with imidazole have provided evidence that the deuterium- nitrogen hydrogen bond between imidazole and D2O is shorter and more stable than a standard hydrogen bond (93). This effect could potentially manifest in energetically favorable deuterium transfer in the context of hydroxyl anion generation (i.e., water activation) for subsequent 85 guanidino deprotonation. Alternatively, D2O could stabilize the tertiary structure of high mobility components of PRMT6, which have been demonstrated to produce inverse kinetic isotope effects with other proteins (92). The PRMT1 D51N variant exhibited fascinating and instructive kinetics that implicates water in the reaction mechanism. Although initial velocity curves could not be saturated for R1- CH3 within the concentrations tested, D51N demonstrated higher levels of methyl transfer for the substrate R1-CH3 than for R1-OH, which is distinct from results produced by wild type PRMT1. The D51N variant reduced catalysis to the rate-limiting step as evidenced by inverse KIEs in both kcat and kcat/Km. In comparing wild type PRMT1 to its D51N variant in D2O when R1-OH was utilized as a substrate, the rate limiting step shifted from product dissociation to catalysis. This shift not only implies that D51 is critical for proper enzyme function, but that D2O and D51 interact either directly or indirectly within the reaction mechanism as suggested by a KIE on catalysis itself. The higher methylation activity using R1-CH3 as a substrate with PRMT1 D51N indicates that D51 is more involved in promoting the hypermethylation of R1-OH. Therefore, a deprotonation step involving D51 may be critically important for R1-OH activation and subsequent methyl transfer. 86 Figure 4.6. Electrostatic potential surrounding deprotonated Nε-substituted arginine fragments. The electrostatic potential maps of arginine analogue fragments bearing +1 charges (A), neutral charges (B), or -1 charges (C) were calculated using Gaussian software 09, revision C.01. The electrostatic potential scale is in units of kcal/mol. 87 Hydroxyarginine as an activated substrate. The N-hydroxyarginine intermediate plays a vital role in the catalytic mechanism of the nitric oxide synthase (eNOS, iNOS, nNOS) (94-96). The presence of this compound in nature suggests that hydroxyarginine may represent a unique species that is capable of overcoming energetic barriers to arginine metabolism and post- translational modification. In the case of PRMTs, the hydroxyl group in N-hydroxyarginine could be exerting its effect via oxime anion formation through deprotonation of the neutral hydroxyarginine tautomer via the His-Asp dyad (Figure 4.6). This is supported by the reduced level of methylation on R1-OH by the D51N variant compared to wild type PRMT1. Electrostatic potential maps of arginine analogues used in this study (Figure 4.6) demonstrate little difference on the substituted N atom for all +1 charge and neutral structures, whereas the oxime anion (i.e., -1 charge) form of N-hydroxyarginine stands out from its protonated forms. Deprotonation and subsequent oxime anion formation of this residue localizes negative charge on the substituted N atom, and may prime it for nucleophilic attack on the methylsulfonium group of AdoMet. This reaction is not without precedence as previous studies have shown that the nitrogen atom of an oxime anion can readily undergo alkylation (97). Curiously, only PRMT1 and its E153Q variant exhibited enhanced activity when using R1-OH as a substrate, implying that the PRMT1 active site has unique features that best tolerate N- hydroxyarginine as a substrate. Given that in PRMT1, E153 is positioned to redistribute electrons through the guanidino group toward the nucleophilic N in position for SN2 attack on AdoMet, it is possible that oximes, as in the case of deprotonated N-hydroxyarginine, are able to perform essentially the same function by acting as a source of electrons for the N atom, thereby producing activity in the normally inactive E153N variant. 88 A mechanism of water-mediated arginine activation. The data in this study suggest that water plays a role in facilitating substrate deprotonation. This deprotonation event allows for the subsequent “SN2-like” attack of the guanidine group at the positively charged methylsulfonium center in AdoMet. The first hint at a water-mediated mechanism for PRMTs can be found in the crystal structure of PRMT3 (3) in which the arrangement of two water molecules occupy the equivalent position of the substrate arginine, and a third water molecule resides between E326 (equivalent to E144 in PRMT1) and H476 (equivalent to H293 in PRMT1) of the His-Asp dyad (Figure 4.7). Therefore, we propose the potential mechanism of catalysis utilizing water detailed in Figure 4.8. KIEs in both PRMT1 D51N and PRMT6 demonstrate that catalysis is the rate- limiting step for these enzymes and may utilize solvent in their active site chemistry. However, the inverse KIE suggests that the reaction is stabilized in D2O. An inverse KIE may be observed when deuterium is transferred onto the highest frequency oscillator, resulting in a lower net system energy (98). Thus, the overall reaction is driven forward thermodynamically by lowering the energy of intermediate steps in the enzyme mechanism. 89 Figure 4.7. Superimposition of PRMTs with water and arginine. Superimposition of PRMT1, PRMT3, PRMT4 and PRMT6 demonstrate the putative position of arginine and water molecules within the active site. Water molecules are represented by magenta spheres. Distance values between the central water molecule and surrounding residues are measured in angstroms. It should be noted that while a water-mediated mechanism is rationally supported by both previously reported structural data and the biochemical evidence provided here, other explanations are feasible, including the presence of secondary KIEs. Stabilization of tertiary structure in D2O has been demonstrated and could facilitate better methyl transfer should the His-Asp dyad be involved in structurally bridging the THW loop with the αYZ-helices as suggested by Thompson and coworkers (30). Alternatively, given that the rate-limiting step is catalysis for wild type PRMT1 with R1-NH2, the active site is likely to accommodate a water 90 molecule poised between the His-Asp dyad and E144 in addition to an N-arginine substitution (Figure 4.7). In the case of R1-OH, D2O may promote direct deprotonation of the arginine oxime by simply providing kinetic barriers to deuterium water deprotonation. In both cases an inverse KIE could feasibly be observed, and diagnostic criteria proposed by Northrop would not apply, as they do not account for secondary KIEs (90, 93). Additionally, it has been previously demonstrated that deuteration of arginine occurs rapidly in both acidic and basic conditions (91). This phenomenon could produce KIEs through direct deprotonation of arginine by the His-Asp dyad, thus complicating the interpretation of data for R1-NH2 methylation by PRMT1 since it was the only combination to demonstrate a non-inverse KIE. Although this phenomenon should not create an inverse KIE, secondary KIEs that are rate limiting could mask other rapid steps that are occurring within the reaction mechanism. Our proposed mechanism for the methylation of R1-OH occurs through direct deprotonation of the hydroxyl group of N-hydroxyarginine to form an oxime anion (Figure 4.8). Such reactivity patterns have previously been described for analogous systems, leading to alkylation of the nitrogen center directly bound to the oxygen (97, 99). The precedence for water-mediated deprotonation of a basic residue has been established for the Set7/9 methyltransferase, which utilizes a water channel to ultimately deprotonate the terminal amine of lysine prior to methyl transfer (100). Structural evidence and the kinetic data outlined above provide a rational case for considering water as a plausible player in the PRMT catalytic mechanism. 91 Figure 4.8. A proposed water-based mechanism for activation of arginine in PRMT1. The following schematic demonstrates utilization of water in the PRMT1 reaction mechanism. The His293-Asp51 dyad deprotonates a single water molecule, which in turn deprotonates arginine, thus increasing its nucleophilicity and propensity to react with the methylsulfonium of AdoMet. Inserts show the proposed mechanism for deprotonation of the hydroxyl group of R1-OH to form an oxime anion intermediate prior to methyl transfer. 92 5 Conclusions The long-term goal of this research is to understand how PRMT catalytic activity and binding interactions are related, to paint a picture of how these enzymes integrate into signaling pathways, and in some instances, contribute to disease. The work performed here fundamentally addresses how PRMTs operate in catalysis and in large protein complexes. What sets this work apart from other efforts is the novelty of the tools and methods developed independently and collaboratively. The new peptide-based tools have added a new twist to structure activity relationship studies of PRMTs and will allow us to uncover the catalytic function of PRMT active site residues important for methylating the target N η atom that exceeds what can be accomplished by site-directed mutagenesis alone. We have also applied fluorescence-based methods to spearhead investigations into PRMT homo- and heterodimerization. By providing a clearer picture of how PRMTs work, we may help to define their contributions to biological pathways and the pathology of diseases like cancer and cardiovascular disease. In Chapter 2, evidence was presented that N-terminal fluorescently tagged PRMT1 and PRMT6 form stable homodimers, and that the presence of AdoMet or AdoHcy strengthens or weakens the binding affinity of homodimerization respectively. Utilizing mass spectrometry and phosphor imaging, it was demonstrated that the apparent kinetic constants and substrate specificity of fluorescently tagged PRMT1 and PRMT6 were similar to wild-type enzymes. Fluorescent quenching experiments demonstrated that resonance energy transfer occurs between dimers as opposed to higher order oligomers, and PRMT competition assays also demonstrated that a PRMT1-PRMT6 heterocomplex is able to form when PRMT6 is added to PRMT1. This data supports a mechanism in which AdoMet binding promotes homodimerization and subsequent methylation, followed by AdoHcy production and dissociation of the homodimer to 93 turn over the cofactor. It also suggests that a PRMT1-PRMT6 heterocomplex may be possible in vivo under appropriate conditions. In Chapter 3, HIV Tat-derived peptides with an ωN-substituted nitrogen were tested in pursuit of functional groups that may act as promising selective inhibitors for PRMT1, PRMT4 and PRMT6. These peptides acted as substrates for PRMT1 and PRMT6, but not for PRMT4. Enzyme kinetics of these novel substrates revealed substrate inhibition, which was demonstrated to be dependent on enzyme concentration and suggests that different enzyme oligomers may be present in solution with variable sensitivity to substrate inhibition. PRMT1 was most susceptible to substrate inhibition, and PRMT6 did not demonstrate substrate inhibition for the cyclopropyl- substituted Tat peptide, indicating that this structure interrupts dead-end complex formation. A mathematical model was proposed in which activity is the sum of the activity contributions of each unique oligomeric complex in solution, and that only the monomer was susceptible to substrate inhibition. Finally, to further expand our knowledge on PRMT catalytic mechanism, R1 fibrillarin- like peptide were synthesized bearing a single arginine with terminal substitutions that lower the pKa of the guanidino group. These substrates were used to assess whether improving nucleophilicity of the terminal guanidino nitrogens would reveal details about the conditions required for methyl transfer. These substituted R1 peptides were found to be good methyl acceptors with the exception of R1-can and R1-nitro, which displayed poor or no methylation respectively. Despite these results, R1-nitro produced useful inhibition of PRMT1, PRMT4 and PRMT6 while R1-can inhibited PRMT1 alone. Steady-state enzyme kinetics in the presence of either H2O or D2O as solvent revealed that the hydroxylated R1 peptide was an excellent substrate for PRMT1, with an enhanced kcat and Km compared to monomethylated R1. 94 Additionally, methylation of hydroxylated R1 peptide with the PRMT1 inactive variant E153Q produced above-background activity, partially circumventing the stabilization provided by that active site residue. In conjunction with the previous observations, kinetic isotope effects revealed that catalysis was rate-limiting in nearly all cases but PRMT1 with hydroxylated and aminated R1 peptides. Upon introducing the D51N mutation to PRMT1, catalysis became rate limiting, implicating D51 and water in methyl transfer. Electrostatic potential maps support a model in which hydroxylated R1 is deprotonated in place of water, producing an activated oxime with focused electronegativity on both the terminal oxygen and nitrogen atoms, improving nucleophilicity and improving rates of methyl transfer in PRMT1. Oligomerization in conjunction with structural data supports a distributive sequential ordered mechanism for PRMT1 and PRMT6. Cofactor effects on oligomerization of PRMTs is another piece of the overall picture revealed by biochemical data detailing a sequence of events in which substrate binding induces structural changes in an ordered fashion. These structural changes provide stabilization for both the cofactor and the substrate arginine- containing protein. Apoenzyme structures reveal conserved residues that are oriented in directions not conducive to catalysis. In the absence of AdoHcy, E144 in PRMT1 points into the AdoMet binding pocket, occluding possible cofactor binding, and R54 orients itself into the open pocket where the substrate guanidino group would sit for methylation (2-4, 29, 53). Upon AdoHcy binding, these residues reorient to stabilize the carboxy terminus of methionine in AdoMet and help to coordinate water in position to be deprotonated by the D51-H293 proton relay. In addition to this, cofactor binding must play a separate role, subtly altering dissociation strength of homodimers in the absence of peptide. As evidenced in this dissertation, AdoMet improves homodimer binding and AdoHcy weakens it. Dimerization has been previously shown 95 to be critical for activity (2, 5, 26, 33, 34), thus oligomerization should precede full formation of the active site. Crystal structures demonstrate that the N-terminal αXY-helix cluster is required to fully form the peptide-binding groove adjacent to the cofactor binding pocket and folds inwards to occlude the cofactor from solvent. In structures of PRMT homodimers, the dimerization arm makes several contacts with the outer face of the AdoMet binding domain and the αXY-helix bundle, forcing it to remain in position, fully forming the electronegative peptide binding groove that funnels down the inner channel of the PRMT homodimer (2, 3). Transient kinetic analysis experiments have verified a multistep conformational change that improves peptide binding in the presence of AdoHcy, further supporting that cofactor binding is critical for peptide recruitment (34). Post-methylation, AdoHcy production then facilitates dissociation of the PRMT homodimer to release the constrained αXY-helices, allowing both peptide and AdoHcy to exchange for AdoMet and unmethylated peptide. Additionally, it is sensible that the mechanism of PRMT methylation would be distributive because oligomerization traps AdoMet and AdoHcy within the cofactor binding pocket, making any processive methylation impossible beyond two sequential additions. It should be considered that a peptide containing arginine may be able to rotate between two cofactor-loaded active sites within the central channel of the homodimer. However, the presence of uneven ratios of MMA to ADMA utilizing a variety of substrates (36, 40, 44) supports a system in which at least some peptide fully dissociates as MMA before receiving a second methyl group. Additionally, it has been demonstrated that MMA can increase in concentration over time independent of and before detectable levels of ADMA are present, clearly supporting a distributive model (44). In conjunction with the improved Km of R1-CH3 for PRMT1 and PRMT6, it follows that an “apparently processive” 96 system may be observed in which dissociation and re-association kinetics are rapid post- monomethylation (35). HIV Tat and substrate inhibition in the context of an ordered system. HIV Tat- based peptides proved to be a useful tool in investigating PRMT kinetics. In the context of the aforementioned model in which substrate binding is ordered and sequential, it is logical that inhibition by these electropositive peptides could be facilitated by locking the αXY-helix in place via docking strongly within the electronegative peptide-binding groove, preventing cofactor turnover. Ki values were curiously close for all substitutions of Tat peptide within PRMT family members, indicating that inhibition is a result of a different physicochemical property other than those conferred by the individual substitutions. With two lysine residues and six arginine residues, the electropositive charge would interact tightly within the PRMT active site. The exception to this is cyclopropyl-Tat (Tat 6, Figure 3.1), which displayed no substrate inhibition with PRMT6. As previously mentioned, the cyclopropyl group must prevent formation of the inhibitory enzyme-substrate complex. The most interesting of the observations involving HIV Tat peptide is the mitigation of substrate inhibition with increasing enzyme concentration. Although this phenomenon was modeled as the sum of monomer activity susceptible to substrate inhibition plus oligomer activity impervious to substrate inhibition, it is difficult to experimentally validate the model beyond observations of this unique system. Equation 3.2 represents a simplified example of the dynamics possible in this enzymatic system. The function holds possible for all examples with the exception of Tat 8, which displays unique behavior that cannot be explained by the sum of both a substrate inhibition profile and hyperbolic fit, even if the hyperbola was purely within the linear range. Such examples underscore that PRMT oligomerization and subsequent activity are controlled by multiple conformational steps that can 97 be modified by cofactor and peptide binding, all of which have unique kinetic constants dictating this complex enzymatic equilibrium. The His-Asp proton relay is critical for activity of PRMT1 and operates through a water-mediated mechanism. Verifying and building upon work performed by Rust et al., the D51N variant of PRMT1 was validated in my work as being important for proper functioning of the enzyme. I found that activity decreased by two orders of magnitude compared to wildtype enzyme and substrate preference was altered (Figure 4.4D, Table 4.2). Kinetic isotope effect experiments demonstrated a change in rate-limiting step between wildtype PRMT1 (product dissociation) and PRMT1 D51N (catalysis) when utilizing D2O as a solvent and R1-OH as the substrate. It was also found for wildtype PRMT6 that catalysis was rate-limiting for R1-OH and R1-CH3, implicating water in the enzyme chemistry. Previously mentioned, a PRMT1 H293A mutation was made that was shown to compromise activity by disrupting catalysis using the kinetic isotope effect (30). Together this data underscores the importance of the His-Asp dyad composed of residues from the THW loop and αY-helix, two structures also important for cofactor binding and homodimerization, respectively. Future Work. The outer face of the AdoMet binding domain is a ripe location for mutational analysis to disrupt the hydrophobic interactions stabilizing dimer formation. A properly folded null-dimerzation variant would be valuable in characterizing a purely monomeric PRMT. Additionally, given the structural conservation between the dimerization arm and AdoMet binding domain of several PRMT family members (26), there are likely several other heterodimeric or hetero-oligomeric complexes that have yet to be characterized. The presence of canonical heterodimers would add several layers of exciting complexity and elegance to the field of arginine methylation and open up new paths to both recruit substrates and modulate activity. 98 References 1. Bedford, M. T., and Clarke, S. G. (2009) Protein arginine methylation in mammals: who, what, and why, Mol Cell 33, 1-13. 2. Zhang, X., and Cheng, X. (2003) Structure of the predominant protein arginine methyltransferase PRMT1 and analysis of its binding to substrate peptides, Structure. 11, 509-520. 3. Zhang, X., Zhou, L., and Cheng, X. (2000) Crystal structure of the conserved core of protein arginine methyltransferase PRMT3, Embo J 19, 3509-3519. 4. Troffer-Charlier, N., Cura, V., Hassenboehler, P., Moras, D., and Cavarelli, J. (2007) Functional insights from structures of coactivator-associated arginine methyltransferase 1 domains, Embo J 26, 4391-4401. 5. Weiss, V. H., McBride, A. E., Soriano, M. A., Filman, D. J., Silver, P. A., and Hogle, J. M. (2000) The structure and oligomerization of the yeast arginine methyltransferase, Hmt1, Nat Struct Biol 7, 1165-1171. 6. Deribe, Y. L., Pawson, T., and Dikic, I. (2010) Post-translational modifications in signal integration, Nat Struct & Mol Biol 17, 666-672. 7. Calnan, B. J., Tidor, B., Biancalana, S., Hudson, D., and Frankel, A. D. (1991) Arginine- mediated RNA recognition: the arginine fork, Science 252, 1167-1171. 8. Zhong, J., Cao, R. X., Zu, X. Y., Hong, T., Yang, J., Liu, L., Xiao, X. H., Ding, W. J., Zhao, Q., Liu, J. H., and Wen, G. B. (2012) Identification and characterization of novel spliced variants of PRMT2 in breast carcinoma, FEBS J 279, 316-335. 9. Hong, H., Kao, C., Jeng, M. H., Eble, J. N., Koch, M. O., Gardner, T. A., Zhang, S., Li, L., Pan, C. X., Hu, Z., MacLennan, G. T., and Cheng, L. (2004) Aberrant expression of CARM1, a transcriptional coactivator of androgen receptor, in the development of prostate carcinoma and androgen-independent status, Cancer 101, 83-89. 10. Kim, Y. R., Lee, B. K., Park, R. Y., Nguyen, N. T., Bae, J. A., Kwon, D. D., and Jung, C. (2010) Differential CARM1 expression in prostate and colorectal cancers, BMC Cancer 10, 197. 11. Yoshimatsu, M., Toyokawa, G., Hayami, S., Unoki, M., Tsunoda, T., Field, H. I., Kelly, J. D., Neal, D. E., Maehara, Y., Ponder, B. A., Nakamura, Y., and Hamamoto, R. (2011) Dysregulation of PRMT1 and PRMT6, Type I arginine methyltransferases, is involved in various types of human cancers, Int J Cancer 128, 562-573. 12. Matsuguma, K., Ueda, S., Yamagishi, S., Matsumoto, Y., Kaneyuki, U., Shibata, R., Fujimura, T., Matsuoka, H., Kimoto, M., Kato, S., Imaizumi, T., and Okuda, S. (2006) Molecular mechanism for elevation of asymmetric dimethylarginine and its role for hypertension in chronic kidney disease, J Am Soc Nephrol 17, 2176-2183. 13. Chen, X., Niroomand, F., Liu, Z., Zankl, A., Katus, H. A., Jahn, L., and Tiefenbacher, C. P. (2006) Expression of nitric oxide related enzymes in coronary heart disease, Basic Res Cardiol 101, 346-353. 14. Bedford, M. T. (2007) Arginine methylation at a glance, J Cell Sci 120, 4243-4246. 15. Branscombe, T. L., Frankel, A., Lee, J. H., Cook, J. R., Yang, Z., Pestka, S., and Clarke, S. (2001) PRMT5 (Janus kinase-binding protein 1) catalyzes the formation of symmetric dimethylarginine residues in proteins, J Biol Chem 276, 32971-32976. 99 16. Zurita-Lopez, C. I., Sandberg, T., Kelly, R., and Clarke, S. G. (2012) Human protein arginine methyltransferase 7 (PRMT7) is a type III enzyme forming omega-NG- monomethylated arginine residues, J Biol Chem 287, 7859-7870. 17. Lakowski, T. M., and Frankel, A. (2009) Kinetic analysis of human protein arginine N- methyltransferase 2: formation of monomethyl- and asymmetric dimethyl-arginine residues on histone H4, Biochem J 421, 253-261. 18. Tang, J., Gary, J. D., Clarke, S., and Herschman, H. R. (1998) PRMT 3, a type I protein arginine N-methyltransferase that differs from PRMT1 in its oligomerization, subcellular localization, substrate specificity, and regulation, J Biol Chem 273, 16935-16945. 19. Frankel, A., Yadav, N., Lee, J., Branscombe, T. L., Clarke, S., and Bedford, M. T. (2002) The novel human protein arginine N-methyltransferase PRMT6 is a nuclear enzyme displaying unique substrate specificity, J Biol Chem 277, 3537-3543. 20. Lee, J., Sayegh, J., Daniel, J., Clarke, S., and Bedford, M. T. (2005) PRMT8, a new membrane-bound tissue-specific member of the protein arginine methyltransferase family, J Biol Chem 280, 32890-32896. 21. Tang, J., Frankel, A., Cook, R. J., Kim, S., Paik, W. K., Williams, K. R., Clarke, S., and Herschman, H. R. (2000) PRMT1 is the predominant type I protein arginine methyltransferase in mammalian cells, J Biol Chem. 275, 7723-7730. 22. Troffer-Charlier, N., Cura, V., Hassenboehler, P., Moras, D., and Cavarelli, J. (2007) Expression, purification, crystallization and preliminary crystallographic study of isolated modules of the mouse coactivator-associated arginine methyltransferase 1, Acta Crystallog 63, 330-333. 23. Antonysamy, S., Bonday, Z., Campbell, R. M., Doyle, B., Druzina, Z., Gheyi, T., Han, B., Jungheim, L. N., Qian, Y., Rauch, C., Russell, M., Sauder, J. M., Wasserman, S. R., Weichert, K., Willard, F. S., Zhang, A., and Emtage, S. (2012) Crystal structure of the human PRMT5:MEP50 complex, Proc Nat Acad Sci USA109, 17960-17965. 24. Sun, L., Wang, M., Lv, Z., Yang, N., Liu, Y., Bao, S., Gong, W., and Xu, R. M. Structural insights into protein arginine symmetric dimethylation by PRMT5, Proc Nat Acad Sci USA 108, 20538-20543. 25. Fauman, E. B., Blumenthal, R. M., and Cheng, X. (1999) Structure and evolution of AdoMet-dependent methyltransferases, In AdoMet-dependent methyltransferases: structures and functions (Cheng, X., and Blumenthal, R. M., Eds.), pp 1-38, World Scientific Publishing, New Jersey. 26. Pak, M. L., Lakowski, T. M., Thomas, D., Vhuiyan, M. I., Husecken, K., and Frankel, A. (2011) A protein arginine N-methyltransferase 1 (PRMT1) and 2 heteromeric interaction increases PRMT1 enzymatic activity, Biochemistry 50, 8226-8240. 27. Chen, D., Ma, H., Hong, H., Koh, S. S., Huang, S. M., Schurter, B. T., Aswad, D. W., and Stallcup, M. R. (1999) Regulation of transcription by a protein methyltransferase, Science 284, 2174-2177. 28. Thomas, D., Lakowski, T. M., Pak, M. L., Kim, J. J., and Frankel, A. (2010) Forster resonance energy transfer measurements of cofactor-dependent effects on protein arginine N-methyltransferase homodimerization, Protein Sci 19, 2141-2151. 29. Sack, J. S., Thieffine, S., Bandiera, T., Fasolini, M., Duke, G. J., Jayaraman, L., Kish, K. F., Klei, H. E., Purandare, A. V., Rosettani, P., Troiani, S., Xie, D., and Bertrand, J. A. Structural basis for CARM1 inhibition by indole and pyrazole inhibitors, Biochem J 436, 331-339. 100 30. Rust, H. L., Zurita-Lopez, C. I., Clarke, S., and Thompson, P. R. (2011) Mechanistic studies on transcriptional coactivator protein arginine methyltransferase 1, Biochemistry 50, 3332-3345. 31. Gui, S., Wooderchak, W. L., Daly, M. P., Porter, P. J., Johnson, S. J., and Hevel, J. M. (2011) Investigation of the molecular origins of protein-arginine methyltransferase I (PRMT1) product specificity reveals a role for two conserved methionine residues, J Biol Chem 286, 29118-29126. 32. Lee, D. Y., Ianculescu, I., Purcell, D., Zhang, X., Cheng, X., and Stallcup, M. R. (2007) Surface-scanning mutational analysis of protein arginine methyltransferase 1: roles of specific amino acids in methyltransferase substrate specificity, oligomerization, and coactivator function, Mol Endocrinol 21, 1381-1393. 33. Higashimoto, K., Kuhn, P., Desai, D., Cheng, X., and Xu, W. (2007) Phosphorylation- mediated inactivation of coactivator-associated arginine methyltransferase 1, Proc Nat Acad Sci USA 104, 12318-12323. 34. Feng, Y., Xie, N., Jin, M., Stahley, M. R., Stivers, J. T., and Zheng, Y. G. A transient kinetic analysis of PRMT1 catalysis, Biochemistry 50, 7033-7044. 35. Lakowski, T. M., and Frankel, A. (2008) A kinetic study of human protein arginine N- methyltransferase 6 reveals a distributive mechanism, J Biol Chem 283, 10015-10025. 36. Kolbel, K., Ihling, C., Bellmann-Sickert, K., Neundorf, I., Beck-Sickinger, A. G., Sinz, A., Kuhn, U., and Wahle, E. (2009) Type I Arginine Methyltransferases PRMT1 and PRMT-3 Act Distributively, J Biol Chem 284, 8274-8282. 37. Obianyo, O., Osborne, T. C., and Thompson, P. R. (2008) Kinetic mechanism of protein arginine methyltransferase 1, Biochemistry 47, 10420-10427. 38. Obianyo, O., and Thompson, P. R. Kinetic mechanism of protein arginine methyltransferase 6 (PRMT6), J Biol Chem 287, 6062-6071. 39. Segel, I. H. (1993) Enzyme kinetics : behavior and analysis of rapid equilibrium and steady-state enzyme systems, Wiley, New York. 40. Gui, S., Wooderchak-Donahue, W. L., Zang, T., Chen, D., Daly, M. P., Zhou, Z. S., and Hevel, J. M. (2013) Substrate-induced control of product formation by protein arginine methyltransferase 1, Biochemistry 52, 199-209. 41. Gary, J. D., and Clarke, S. (1998) RNA and protein interactions modulated by protein arginine methylation, Progress in nucleic acid research and molecular biology 61, 65- 131. 42. Boisvert, F. M., Cote, J., Boulanger, M. C., and Richard, S. (2003) A proteomic analysis of arginine-methylated protein complexes, Mol Cell Proteomics. 2, 1319-1330. 43. Paik, W. K., and Kim, S. (1967) Enzymatic methylation of protein fractions from calf thymus nuclei, Biochem Biophys Res Commun 29, 14-20. 44. Wooderchak, W. L., Zang, T., Zhou, Z. S., Acuna, M., Tahara, S. M., and Hevel, J. M. (2008) Substrate profiling of PRMT1 reveals amino acid sequences that extend beyond the \"RGG\" paradigm, Biochemistry 47, 9456-9466. 45. Di Lorenzo, A., and Bedford, M. T. (2011) Histone arginine methylation, FEBS letters 585, 2024-2031. 46. Lin, C. H., Huang, H. M., Hsieh, M., Pollard, K. M., and Li, C. (2002) Arginine methylation of recombinant murine fibrillarin by protein arginine methyltransferase, J Protein Chem 21, 447-453. 101 47. Kolbel, K., Ihling, C., Kuhn, U., Neundorf, I., Otto, S., Stichel, J., Robaa, D., Beck- Sickinger, A. G., Sinz, A., and Wahle, E. (2012) Peptide backbone conformation affects the substrate preference of protein arginine methyltransferase I, Biochemistry 51, 5463- 5475. 48. Rawal, N., Rajpurohit, R., Paik, W. K., and Kim, S. (1994) Purification and characterization of S-adenosylmethionine-protein-arginine N-methyltransferase from rat liver, Biochem J 300 ( Pt 2), 483-489. 49. Yang, Y., Lu, Y., Espejo, A., Wu, J., Xu, W., Liang, S., and Bedford, M. T. (2010) TDRD3 is an effector molecule for arginine-methylated histone marks, Mol Cell 40, 1016-1023. 50. Lacroix, M., El Messaoudi, S., Rodier, G., Le Cam, A., Sardet, C., and Fabbrizio, E. (2008) The histone-binding protein COPR5 is required for nuclear functions of the protein arginine methyltransferase PRMT5, EMBO reports 9, 452-458. 51. Bauer, U. M., Daujat, S., Nielsen, S. J., Nightingale, K., and Kouzarides, T. (2002) Methylation at arginine 17 of histone H3 is linked to gene activation, EMBO reports 3, 39-44. 52. Ma, H., Baumann, C. T., Li, H., Strahl, B. D., Rice, R., Jelinek, M. A., Aswad, D. W., Allis, C. D., Hager, G. L., and Stallcup, M. R. (2001) Hormone-dependent, CARM1- directed, arginine-specific methylation of histone H3 on a steroid-regulated promoter, Curr Biol 11, 1981-1985. 53. Yue, W. W., Hassler, M., Roe, S. M., Thompson-Vale, V., and Pearl, L. H. (2007) Insights into histone code syntax from structural and biochemical studies of CARM1 methyltransferase, Embo J 26, 4402-4412. 54. Wu, J., Cui, N., Wang, R., Li, J., and Wong, J. (2012) A role for CARM1-mediated histone H3 arginine methylation in protecting histone acetylation by releasing corepressors from chromatin, PloS one 7, e34692. 55. Daujat, S., Bauer, U. M., Shah, V., Turner, B., Berger, S., and Kouzarides, T. (2002) Crosstalk between CARM1 methylation and CBP acetylation on histone H3, Curr Biol 12, 2090-2097. 56. Iberg, A. N., Espejo, A., Cheng, D., Kim, D., Michaud-Levesque, J., Richard, S., and Bedford, M. T. (2008) Arginine methylation of the histone H3 tail impedes effector binding, J Biol Chem 283, 3006-3010. 57. Guccione, E., Bassi, C., Casadio, F., Martinato, F., Cesaroni, M., Schuchlautz, H., Luscher, B., and Amati, B. (2007) Methylation of histone H3R2 by PRMT6 and H3K4 by an MLL complex are mutually exclusive, Nature 449, 933-937. 58. Hyllus, D., Stein, C., Schnabel, K., Schiltz, E., Imhof, A., Dou, Y., Hsieh, J., and Bauer, U. M. (2007) PRMT6-mediated methylation of R2 in histone H3 antagonizes H3 K4 trimethylation, Genes Dev 21, 3369-3380. 59. El-Andaloussi, N., Valovka, T., Toueille, M., Steinacher, R., Focke, F., Gehrig, P., Covic, M., Hassa, P. O., Schar, P., Hubscher, U., and Hottiger, M. O. (2006) Arginine methylation regulates DNA polymerase beta, Mol Cell 22, 51-62. 60. Harrison, M. J., Tang, Y. H., and Dowhan, D. H. (2010) Protein arginine methyltransferase 6 regulates multiple aspects of gene expression, Nucleic Acids Res 38, 2201-2216. 102 61. Frankel, A., and Clarke, S. (2000) PRMT3 is a distinct member of the protein arginine N- methyltransferase family. Conferral of substrate specificity by a zinc-finger domain, J Biol Chem. 275, 32974-32982. 62. Boulanger, M. C., Liang, C., Russell, R. S., Lin, R., Bedford, M. T., Wainberg, M. A., and Richard, S. (2005) Methylation of Tat by PRMT6 regulates human immunodeficiency virus type 1 gene expression, J Virol. 79, 124-131. 63. Liang, C., and Wainberg, M. A. (2002) The role of Tat in HIV-1 replication: an activator and/or a suppressor?, AIDS reviews 4, 41-49. 64. Wei, P., Garber, M. E., Fang, S. M., Fischer, W. H., and Jones, K. A. (1998) A novel CDK9-associated C-type cyclin interacts directly with HIV-1 Tat and mediates its high- affinity, loop-specific binding to TAR RNA, Cell 92, 451-462. 65. Hetzer, C., Dormeyer, W., Schnolzer, M., and Ott, M. (2005) Decoding Tat: the biology of HIV Tat posttranslational modifications, Microbes Infect 7, 1364-1369. 66. Xie, B., Invernizzi, C. F., Richard, S., and Wainberg, M. A. (2007) Arginine methylation of the human immunodeficiency virus type 1 Tat protein by PRMT6 negatively affects Tat Interactions with both cyclin T1 and the Tat transactivation region, J Virol. 81, 4226- 4234. 67. Mishra, A., Lai, G. H., Schmidt, N. W., Sun, V. Z., Rodriguez, A. R., Tong, R., Tang, L., Cheng, J., Deming, T. J., Kamei, D. T., and Wong, G. C. (2011) Translocation of HIV TAT peptide and analogues induced by multiplexed membrane and cytoskeletal interactions, Proc Nat Acad Sci USA 108, 16883-16888. 68. Osborne, T. C., Obianyo, O., Zhang, X., Cheng, X., and Thompson, P. R. (2007) Protein Arginine Methyltransferase 1: Positively Charged Residues in Substrate Peptides Distal to the Site of Methylation Are Important for Substrate Binding and Catalysis, Biochemistry 46, 13370-13381. 69. Goulet, I., Gauvin, G., Boisvenue, S., and Cote, J. (2007) Alternative splicing yields protein arginine methyltransferase 1 isoforms with distinct activity, substrate specificity, and subcellular localization, J Biol Chem 282, 33009-33021. 70. Rizzo, M. A., Springer, G. H., Granada, B., and Piston, D. W. (2004) An improved cyan fluorescent protein variant useful for FRET, Nat Biotechnol 22, 445-449. 71. Shaner, N. C., Steinbach, P. A., and Tsien, R. Y. (2005) A guide to choosing fluorescent proteins, Nat Methods 2, 905-909. 72. Andre, I., and Linse, S. (2002) Measurement of Ca2+-binding constants of proteins and presentation of the CaLigator software, Anal Biochem 305, 195-205. 73. Heikal, A. A., Hess, S. T., Baird, G. S., Tsien, R. Y., and Webb, W. W. (2000) Molecular spectroscopy and dynamics of intrinsically fluorescent proteins: coral red (dsRed) and yellow (Citrine), Proc Nat Acad Sci USA 97, 11996-12001. 74. Andrews, D. L., and Demidov, A. A. (1999) Resonance energy transfer, Wiley, New York. 75. Adair, B. D., and Engelman, D. M. (1994) Glycophorin A helical transmembrane domains dimerize in phospholipid bilayers: a resonance energy transfer study, Biochemistry 33, 5539-5544. 76. Li, M., Reddy, L. G., Bennett, R., Silva, N. D., Jr., Jones, L. R., and Thomas, D. D. (1999) A fluorescence energy transfer method for analyzing protein oligomeric structure: application to phospholamban, Biophys J 76, 2587-2599. 103 77. Hermes, M., Osswald, H., Mattar, J., and Kloor, D. (2004) Influence of an altered methylation potential on mRNA methylation and gene expression in HepG2 cells, Exp Cell Res 294, 325-334. 78. Herrmann, F., Pably, P., Eckerich, C., Bedford, M. T., and Fackelmayer, F. O. (2009) Human protein arginine methyltransferases in vivo - distinct properties of eight canonical members of the PRMT family, J Cell Sci 122, 667-677. 79. Pawlak, M. R., Scherer, C. A., Chen, J., Roshon, M. J., and Ruley, H. E. (2000) Arginine N-methyltransferase 1 is required for early postimplantation mouse development, but cells deficient in the enzyme are viable, Mol Cell Biol 20, 4859-4869. 80. Schagger, H., and von Jagow, G. (1987) Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa, Anal Biochem 166, 368-379. 81. Wiltfang, J., Arold, N., and Neuhoff, V. (1991) A new multiphasic buffer system for sodium dodecyl sulfate-polyacrylamide gel electrophoresis of proteins and peptides with molecular masses 100,000-1000, and their detection with picomolar sensitivity, Electrophoresis 12, 352-366. 82. Lakowski, T. M., t Hart, P., Ahern, C. A., Martin, N. I., and Frankel, A. (2010) Neta- substituted arginyl peptide inhibitors of protein arginine N-methyltransferases, ACS Chem Biol 5, 1053-1063. 83. t Hart, P., Lakowski, T. M., Thomas, D., Frankel, A., and Martin, N. I. (2011) Peptidic partial bisubstrates as inhibitors of the protein arginine N-methyltransferases, Chembiochem 12, 1427-1432. 84. Tyapochkin, E., Cook, P. F., and Chen, G. (2009) para-Nitrophenyl sulfate activation of human sulfotransferase 1A1 is consistent with intercepting the E.PAP complex and reformation of E.PAPS, J Biol Chem 284, 29357-29364. 85. Bayer, P., Kraft, M., Ejchart, A., Westendorp, M., Frank, R., and Rosch, P. (1995) Structural studies of HIV-1 Tat protein, J Mol Biol 247, 529-535. 86. Hart‡, P. T., Thomas‡, D., Ommeren, R. v., Lakowski, T. M., Frankel*, A., and Martin*, N. I. (2012) Analogues of the HIV-Tat peptide containing N-modified arginines as potent inhibitors of protein arginine N-methyltransferases, MedChemComm 3, 1235-1244. 87. Lakowski, T. M., and Frankel, A. (2009) Sources of AdoHcy background in measuring PRMT activity using tandem mass spectrometry, Anal Biochem 396, 158-160. 88. Simms, H. S. (1930) The Arginine and Prearginine Groups in Edestin, The Journal of general physiology 14, 87-97. 89. Boyar, A., and Marsh, R. E. (1982) L-Canavanine, a Paradigm for the Structures of Substituted Guanidines, J Am Chem Soc 104, 1995-1998. 90. Northrop, D. B. (1975) Steady-state analysis of kinetic isotope effects in enzymic reactions, Biochemistry 14, 2644-2651. 91. Liepinsh, E., and Otting, G. (1996) Proton exchange rates from amino acid side chains-- implications for image contrast, Magn Reson Med 35, 30-42. 92. Makhatadze, G. I., Clore, G. M., and Gronenborn, A. M. (1995) Solvent isotope effect and protein stability, Nat Struct Biol 2, 852-855. 93. Yuan, C., Tu, S., Gelb, M. H., and Tsai, M. D. (2005) Unusual four-bond secondary H/D isotope effect supports a short-strong hydrogen bond between phospholipase A2 and a transition state analogue inhibitor, Biochemistry 44, 4748-4754. 104 94. Stoll, S., NejatyJahromy, Y., Woodward, J. J., Ozarowski, A., Marletta, M. A., and Britt, R. D. (2010) Nitric oxide synthase stabilizes the tetrahydrobiopterin cofactor radical by controlling its protonation state, J Am Chem Soc 132, 11812-11823. 95. Luzzi, S. D., and Marletta, M. A. (2005) L-arginine analogs as alternate substrates for nitric oxide synthase, Bioorg Med Chem Let 15, 3934-3941. 96. Martin, N. I., Woodward, J. J., Winter, M. B., Beeson, W. T., and Marletta, M. A. (2007) Design and Synthesis of C5 Methylated l-Arginine Analogues as Active Site Probes for Nitric Oxide Synthase, J Am Chem Soc 129, 12563-12570. 97. Osborn, H. M. I., Gemmell, N., and Harwood, L. M. (2002) 1,3-Dipolar cycloaddition reactions of carbohydrate derived nitrones and oximes, J Chem Soc, Perkin Trans 1, 2419-2438. 98. Lowry, T. H., and Richardson, K. S. (1987) Mechanism and theory in organic chemistry, 3rd ed., Harper & Row, New York. 99. Katritzky, A. R., Khashab, N. M., Bobrov, S., and Yoshioka, M. (2006) Synthesis of mono- and symmetrical di-N-hydroxy- and N-aminoguanidines, J Org Chem 71, 6753- 6758. 100. Zhang, X., and Bruice, T. C. (2008) Enzymatic mechanism and product specificity of SET-domain protein lysine methyltransferases, Proc Nat Acad Sci USA 105, 5728-5732. 105 Appendix Appendix figures Figure A.1. Equilibration of FRET signal for PRMT1 and PRMT6. Time course equilibration of fluorescently tagged mix of 0.5 µM each of mCit/mCer-PRMT1 alone (◊), with saturating AdoMet (□), and with saturating AdoHcy (▲) and 0.5 µM each of mCit/mCer-PRMT6 alone (x), with saturating AdoMet (■) and saturating AdoHcy (●). FRET was measured as described in chapter 2. PRMT1 and 6 must equilibrate for at least 60-100 minutes at 37 ○ C to obtain optimal fluorescent signal. 106 Figure A.2. Initial velocity of PRMT6 E58K with varying AdoMet. PRMT6 E58K (♦) and PRMT6 (▲) were incubated with 200 µM histone H3 tail peptide with increased concentrations of AdoMet as shown on the x-axis. MMA and ADMA generation were quantified using mass spectrometric methods described in chapter 3. PRMT6 E58K demonstrated an enhanced apparent Vmax and higher apparent Km than wild type PRMT6. E58K is a conserved residue on the outer surface of the AdoMet binding domain that lies in close proximity to R204 of within the dimerization arm of the opposite monomer in the homodimeric complex. This interaction may stabilize the position of the N-terminal αXY-helix, which could account for the variable kinetic characteristics seen above. "@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "2013-11"@en ; edm:isShownAt "10.14288/1.0073933"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Pharmaceutical Sciences"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "Attribution-NonCommercial-NoDerivatives 4.0 International"@en ; ns0:rightsURI "http://creativecommons.org/licenses/by-nc-nd/4.0/"@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Oligomerization dependent enzyme kinetics and mechanistic characterization of type I protein arginine N-methyltransferases"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/44623"@en .