Illllllllllllllllllllllllllllllllllllllllllllllllllllllllllll 3 1293 017704 LIBRARY Michigan State University This is to certify that the thesis entitled Use a fie teak! flue-l) [)1 If 6*€m 7’0 /:/ent]‘0L/S/ fcros+a<'/cno/’ rr ens/7a ,«rox 'Je ”flu flSSDQop‘fe flmfe n5 presented by Z. (21;! I7 61" has been accepted towards fulfillment of the requirements for KLJ degree 1n Wafimg/ Wm Major professor D‘ate Zl/l I/fij 07639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN Box to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 1M WW“ USE OF THE YEAST TWO-HYBRID SYSTEM TO IDENTIFY PROSTAGLANDIN ENDOPEROXIDE H SYNTHASE ASSOCIATED PROTEINS By Liqun On A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the Degree of MASTER OF SCIENCE Department of Microbiology 1 999 ABSTRACT USE OF THE YEAST TWO-HYBRID SYSTEM To IDENTIFY PROSTAGLANDIN ENDOPEROXIDE H SYNTHASE ASSOCIATED PROTEINS By Liqun Gu PGHS-l and PGHS-Z, the two central isozymes in the pathway for conversion of arachidonic acid to the biologically active prostaglandins and thromboxanes, are very similar in structure and show nearly identical catalytic preperties toward arachidonate metabolism. However the two isozymes are not redundant, they each play separate and specialized roles in cellular signaling and developmental processes. To test whether PGHS-l and PGHS-Z form complexes with other proteins, thereby conferring isozyme differences in apparent substrate affinity, Signal transduction, or catalytic regulation, we used the yeast two-hybrid system to identify candidate genes of proteins that interacted with PGHS. Two candidate proteins ATl 5 and AT18 were shown to interact with PGHS-2 carboxyl terminus in the yeast two-hybrid system. However after serial testing both in vitro and in viva, neither ATlS nor AT18 could be demonstrated to associated with PGHS-Z. To my family iii ACKNOWLEDGMENTS I am grateful to my advisor Dr. David L. DeWitt for his guidance and encouragement during my graduate studies at Michigan State University. He not only helped me learn more about science, but also about people. I also thank my committee members: Dr. Walter J .Esselman, Dr. John C.Fyfe, Dr. Richard C.Schwartz and Dr. William L.Smith for their advice. and support through out the course of my study. I would like to thank my past and present coworkers in the lab, members of Dr. Smith’s and Dr. Garavito’s lab for friendship and kind help. Finally, I want to express my gratitude to my family in China. Without their love, encouragement and sacrifice, it would not have been possible for me to get my education and pursue my dreams. iv TABLE OF CONTENTS List of Tables ......................................................................................... vi List of Figures ....................................................................................... vii List of Abbreviations .............................................................................. viii Introduction ............................................................................................. 1 Literature Review ............................................................................. 1 Research Project Goals ...................................................................... 4 Background ................................................................................... 5 Materials and Methods .............................................................................. 11 Construction of bait Plasmids ............................................................ 11 Library Plasmids ........................................................................... l6 Yeast Two-Hybrid Screening and Testing of Positive Clones ....................... 16 Expression of ATlS and AT18 proteins in Bacteria ................................... 18 In vitro Elution Assay ..................................................................... 20 In vitro Slot Blotting ....................................................................... 20 In vitro Affigel Assay ..................................................................... 21 Additional Two-Hybrid Testing in Yeast to Verify Interaction ...................... 21 Results ................................................................................................. 23 Cloning of ATl 5 and AT18 by Yeast Two-Hybrid Screening ....................... 23 In vitro Testing of PGHS-Z-ATI 5 and PGHS-2-AT18 Interaction .................. 27 In vivo Testing in Yeast to Verify Interaction .......................................... 32 Discussion ............................................................................................ 36 References ............................................................................................ 39 LIST OF TABLES 1. Description of the Plasmids ..................................................................... 13 II. Genotypes of the Yeast Host Strain ........................................................... 15 III. Yeast Two-Hybrid Screen Result ............................................................ 24 LIST OF FIGURES 1. Comparison of the primary structure of PGHS-2 and PGHS-1 ............................ 2 2. The Principle of Yeast Two-Hybrid System .................................................. 6 3. PGHS protein sequences used for bait to screen the yeast two-hybrid system .......... 9 4. Location in the PGHS-l and PGHS-2 proteins of the amino acid sequences used to construct the yeast two-hybrid bait plarnids (standard one-letter symbol). . 1 O 5. Specific primers used in PCR reactions to generate PGHS-1 -EGF, PGHS-1 carboxyl terminal, PGHS-2-EGF and PGHS-2 carboxyl terminal cDNA sequences for construction of the yeast two-hybrid bait plasmid ........................ 12 6. Screening an GAL4-Activation Domain (AD) fusion library for proteins that interact with a bait protein ..................................................................... 17 7. Protocol used to eliminate false positives that arise in a two-hybrid library screening ......................................................................................... l9 8. Two of the candidate clones isolated by the yeast two-hybrid screening: ATM and AT15 ................................................................................. 26 9. SDS-PAGE analysis of the expression of His-tag-ATI 8 and ATI 5 protein. . . . . .......28 10. In vitro Binding Assay ......................................................................... 29 11. Slot blotting ...................................................................................... 3O 12. Immunoblot of PGHS-2 retained on Affigel ................................................ 31 13. Western blots of candidate proteins and bait protein using the GAL4 DNA-BD, AD mAbs and Dinah (Ab against PGHSZ carboxy terminus) ............................ 33 14. Interaction between AT15, AT18 protein with PGHS-2 C-tenninus as determined by serniquantitative yeast two-hybrid assays based on the degree of induction of the two reporter genes H183 and B-gal ...................................................... 35 vii PGHS COX SP EGF MBD NSAID EPA FLAP DNA-BD 3-AT X-gal ER LIST OF ABBREVIATIONS Prostaglandin endoperoxide H synthases Cyclooxygenase Signal peptide Epidermal growth factor Membrane binding domain Nonsteroidal anti-inflammatory drug 5,8,1 1,14,17-eicosapentaenoic acid 5-lipoxygenase-activating protein DNA binding domain Activation domain 1 ,2,4-3 -aminotriazole 5-bromo-4~chloro—3-indolyl-B-D-galactoside Isopropyl-B-D—thiogalactopyranoside Horseradish peroxidase monoclonal antibody endoplasmic reticulum viii INTRODUCTION Literature Review The prostaglandin endoperoxide H synthases (PGHS) are the central enzymes in the pathway for conversion of arachidonic acid to the biologically active prostaglandins and thromboxanes. There are two PGHS isozymes, PGHS-1 and PGHS-2, also known as cyclooxygenase-l and -2 (COX-1 and COX-2). PGHS-l is constitutively expressed in most tissues, and is thought to mediate “housekeeping” functions, including cytoprotection of the gastric mucosa, regulation of renal blood flow, and platelet aggregation. In contrast, PGHS-2 is usually undetectable in most tissues, but it can be induced to express at high levels in migratory and other responding cells by pro- inflammatory and mitogenic stimuli. PGHS-2 is generally considered to be a mediator of inflammation (14). PGHS-1 and PGHS-2 are very similar in structure and Show nearly identical catalytic properties toward arachidonate metabolism (5). The two enzymes are about 60% identical in amino acid sequence and both have four main functional domains (Figure 1): a) a signal peptide (SP); b) a Short amino-terminal epidermal growth factor homology domain (BGF); c) a putative membrane binding domain (MBD); and d) a large catalytic domain. The catalytic domains are highly conserved. However, the amino terminus of PGHS-2 is slightly truncated compared with that of PGHS-l , and the carboxyl terminus of PGHS-2 possesses an l8-residue cassette that is absent in PGHS-l . The biological Significance of these subtle differences in the primary amino acid sequence still needs to be determined. Domain EG SP F MBD Catalytic 18 AA 14 59 38 70 o % Similarity with PGHS-l PGHS-2 Figure 1. Comparison of the primary structure of PGHS-2 and PGHS-1. Knockout mice for PGHS-1 and PGHS-2 display distinct phenotypes suggesting that these two enzymes are not redundant and that each play separate and specialized roles in cellular signaling and developmental processes (6,7). The inducible PGHS-2 functions more than simply to augment constitutive prostaglandin synthesis by PGHS- l , because induction of PGHS-2 often results in only a nominal increase in cellular PGH2 biosynthetic activity (8). In addition, prostanoid synthesis catalyzed by PGHS-l or PGHS-2 is coupled to different extracellular Stimuli and derives from different arachidonate substrate pools, possibly via coupling to different phospholipase systems. In some cells, prostaglandin synthesis proceeds via PGHS-2, even though significant PGHS- ] enzyme is present (9). Taken together, these observations imply the existence of two partially independent PGHS-1 and PGHS-2 prostanoid biosynthetic pathways (10), which may result from different microenvironments in which the enzymes are located, or due to specific interactions of each enzyme with other proteins. Besides the cellular traits that distinguish the two isozymes from one another, PGHS isozymes differ subtly in their substrate specificity and nonsteroidal anti- inflammatory drug (N SAID) affinities (11,12). While both enzymes have Similar Km’s and catalytic efficiencies for arachidonic acid, PGHS-2 has been shown to use alternate fatty acid substrates, such as a-linolenic acid and 5,8,11,14,17-eicosapentaenoic acid (EPA), more efficiently than PGHS-l. NSAIDS have been identified that selectively inhibit either PGHS-l or PGHS-2. The larger and more accommodating active site of PGHS-2 compared with that of PGHS-1 might be responsible for some of these differences. However, other factors might account for their differential binding activities. Another important enzyme for eicosanoid metabolism, 5-lipoxygenase, depends on FLAP (an 18kDa subsidiary protein known as S-lipoxygenase-activating protein) for arachidonate delivery (13,14). This precedent opens the possibility that regulation of prostaglandin synthesis by PGHS-l or PGHS-2 could also depend on protein-protein interactions, thus modifying the activity or subcelllular location of PGHS-1 or PGHS-2. Another argument supporting separate biological roles for PGHS-l and PGHS-2 has come from recent research on the mechanisms whereby chronic aspirin users have reduced incidence of colon cancer (15). Normal colon epithelial cells are found to constitutively express PGHS-l, but not PGHS-2, while most colon carcinoma cells express normal levels of PGHS-1 and high levels of PGHS-2 (16-18). Furthermore, NSAIDS which specifically inhibit PGHS-2, but not PGHS-l, reduce intestinal polyp formation in normal mice. Genetic ablation of PGHS-2 in knockout mice also reduces polyp formation (19). These results suggest that PGHS-2 plays a specific signaling role in colon epithelial cells. All the above data raise the question: Do PGHS-l and PGHS-2 form complexes with other proteins, thereby conferring isozyme differences in apparent substrate affinity, Signal transduction, or catalytic regulation? Research Project Goals The primary hypothesis that we would like to test is that the unique biological properties of PGHS-1 and PGHS-2 result from specific interaction of the enzyme with other proteins that may modify the activity and/or subcellular location of PGHS. Since there are obvious structural differences between PGHS-1 and PGHS-l that could possibly mediate differential protein-protein interactions (i.e. the 18—residue cassette), we plan: Aim 1: To identify candidate genes of proteins that interact with PGHS, using the yeast two-hybrid system. Aim 2: To confirm the Specific association of the candidate protein and PGHS in vitro. Aim 3: To confirm the association of PGHS and the candidate protein in vivo. Backgron The yeast two-hybrid assay is based on the fact that many eukaryotic transcriptional activators are composed of two physically separable, functionally independent domains. The yeast GAL4 transcriptional activator, for example, contains a DNA binding domain (DNA-ED) and a transcriptional activation domain (AD). The DNA-BD recognizes and binds to a sequence (U AS) in the upstream regions of GAL4- responsive genes, while the AD interacts with other components of the transcription machinery needed to initiate transcription. When both domains are part of the same protein, they are able to activate gene transcription. If physically separated by recombinant DNA technology and expressed in the same host cell, however, the GAL4 DNA-BD and AD peptides do not directly interact with each other and cannot activate the responsive genes. In the yeast two-hybrid systems, two different cloning vectors are used to generate separate fusion proteins of the two GAL4 domains. The recombinant hybrid proteins are co-expressed in yeast where they are targeted to the yeast nucleus. If the non- GAL4 portions of the two hybrid proteins interact with each other, the DNA-BD and transcriptional AD are re-united and can again activate transcription (Figure 2). Thus, as a result of a two-hybrid interaction, the GAL4 transcriptional activator will be functionally reconstituted and will activate transcription of reporter genes (lacZ and HIS3) having upstream GAL4 binding sites. This makes the protein interaction l. GAL4 transcriptional activator (wild type) transcription 2. GALA DNA-binding domain and transcriptional activation domain (physically separated by reeomvbinant DNA technology) AD 3. In two-hybrid system, two different cloning vectors are used to generate separate fusions of these GAL4 doamins to genes encoding proteins that potentially interact with each other Figure 2. The principle of yeast two-hybrid system. phenotypically detectable. The same principle applies to screening a yeast two-hybrid library . The gene encoding a target protein is cloned into the DNA-BD vector to generate a fusion protein, referred to as a “bait”. Likewise, a cDNA library is constructed in the AD vector to generate chimeras of various proteins encoded by the library cDNAs fused to the GAL4 AD. The two types of hybrid plasmids are then cotransfonned into a yeast host strain, which is auxotrophic for Trp and Len, for library screening. The transformants are plated on minimal medium lacking Leu, Trp, and His to select for those that contain both types of plasmids (i.e., Leu+, Trp? and that also express interacting hybrid proteins (HisI). Primary His+ transformants are tested for expression of a second reporter gene (lacZ) using a sensitive and rapid filter assay for B-galactosidase activity (20). The yeast two-hybrid system has previously been used successfully in identifying one protein that interacts with the catalytic regions of both PGHS-1 and PGHS-2 (residues 381- 498, using the translation start site in human PGHS-1 as residue 1) (21). However, these early experiments determined that, when large regions of PGHS (300- 400 residues) or regions of PGHS containing the membrane-binding domains (MBD) were used as baits, no interacting proteins were detected. These results could have been obtained because no other interacting proteins exist, or because these baits were poorly expressed or were not translocated to the nucleus because of the MBD in the bait. Short carboxyl terminal PGHS-1 baits were found to be unsuitable for library screening as well, because they intrinsically transactivated the reporter genes. To avoid the same problems and to identify proteins that interact specifically with PGHS-1 or PGHS-2, we decided to choose three PGHS short sequences as baits in our yeast two-hybrid screening. The baits we employed were the human PGHS-l-EGF domain, residues 24-84, the human PGHS-Z-EGF domain, residues 18-70, (using the translation start site in human PGHS-2 as residue 1) and the human PGHS-2 carboxyl terminal sequence, residues 555-604 (Figure 3 and 4). We chose the PGHS-2 carboxyl terminal sequence because this sequence contains an 18-amino acid cassette that is absent in PGHS-1, which represents the most pronounced amino acid sequence difference between PGHS-1 and PGHS-2. The PGHS-EGF domains were chosen because the EGF domains in other proteins are often responsible for protein-protein interaction. Although the MBDS represent the second most dissimilar region between PGHS-l and PGHS-2, these sequences are too hydrophobic to translocate into the nucleus, and cannot be used in yeast two-hybrid system. EGF SP MBD Catalytic PGHS-1 PGHS'I PGHS-1 EGF Bait C-ter Bait EGF ‘ SP MBD Catalytic 18 AA m PGHS-2 PGHS'Z PGHS-2 EGF Bait C-ter Bait Figure 3. PGHS protein sequences used for bait to screen the yeast two-hybrid system. The four PGHS sequences depicted were cloned into pAS2-l vector. The resulting plasmids expressed PGHS-GAL4-BD fusion proteins.The PGHS-l constructs contained the following amino acids: PGHS-l-EGF, residues 24-84; and PGHS-l-COOH, residues 556-599. The PGHS-2 constructs contained the following amino acids: PGHS-Z-EGF, residues 18-70; and PGHS-Z-COOH, residues 555-604 . HUMAN-2 PGHS-l-IFG Bait MSRSLLL RFLLFLLLLP PLPVLHEDPG APTPVNPCCY YPCQHQGICV RFGLDRYQC .............. MLA RALLLCAVLA LSHnANpccs HPCQNRGVCM SVGFDQYKC aaas-z-sar sate CTIPGLWTWL HUMAN-1 HUMAN-2 HUMAN-1 HUMAN-2 HUMAN-1 HUMAN-2 HUMAN-1 HUMAN-2 HUMAN-1 HUMAN-2 HUMAN-l HUMAN-2 HUMAN-l HUMAN-2 TVRSNLIPSP TSRSHLIDSP ARRFLLRRKF VEKLLLRRKF LBRQYQLRLF LARQRKLRLF LMLYATLWLR LMMYATIWLR QLKFDPELLF KLKFDPELLF GVEALVDAFS GITQFVESFT FQELVGEKBM FEELTGEKEM PICSPEYWKP VICSPAYWKP CSTPEFLTRI PTYNSAHDYI PTYNADYGYK IPDPQGTNLM IPDPQGSNMM KDGKLKYQVL KDGKMKYQII BHNRVCDLLK EHNRVCDVLK GVQFQYRNRI NKQFQYQNRI RQIAGRIGGG RQIAGRVAGG AAELEELYGD SAELEALYGD PGHS-l-C Btit SWESFSNVSY SWEAFSNLSY FAFFAQHFTH FAFFAQHFTH DGEMYPPSVE DGEMYPPTVK AEHPTWGDEQ QEHPEWGDEQ AMEFNHLYHW AAEFNTLYHW RNMDHHILHV RNVPPAVQKV IDALEFYPGL IDAVELYPAL 57 43 THFLLTHGRW VHYILTHFKG YTRILPSVPK YTRALPPVPD QFFKTSGKMG QFFKTDHKRG EAPVLMHYPR DTQAEMIYPP LFQTTRLILI LFQTSRLILI HPLMPDSFKV HPLLPDTFQI AVDVIRESRE SQASTDQSRQ LLEKCHPNSI LVEKPRPDAI FWEFVNA.TF FWNVVNNIPF DCPTPMGTKG DCPTPLGVKG PGFTKALGHG PAFTNGLGHG GIPPQSQMAV QVPEHLRFAV GETIKIVIEE GETIKIVIED GSQEYSYEQF HDQKYNYQQF MRLQPFNEYR MKYQSFNEYR FGESMIEIGA FGETMVEVGA OBT-PGEB-l IREMLMRLVL LRNAIMSYVL KKQLPDAQLL KKQLPDSNEI VDLGHIYGDN VDLNHIYGET GQEVFGLLPG GQEVFGLVPG YVQQLSGYFL YVQHLSGYHF LFNTSMLVDY IYNNSILLEH KRFGMKPYTS KRFMLKPYES PFSLKGLLGN PFSLKGLMGN STFGGEVGHN IVKTATLKKL VCLNTKTCPY VSFRVPDASQ STPGGEVGFQ IINTASIQSL IqNNVKGCPr Tsrsvpnpéi IKTVTINASS S83 oer-m4 STE L" 599 604 DDGPAVE... PGHS-Z-C Bait Figure 4. Location in the PGHS-1 and PGHS-2 proteins of the amino acid sequences used to construct the yeast two-hybrid bait plamids (standard one-letter symbols). The numbers on the far right refer to the amino acid position. Those sequences used as baits in yeast two-hybrid screening are boxed (residues 24-84 and 116 103 176 163 236 223 296 283 356 343 416 403 476 463 536 523 593 556-599 of PGHS-l, residues 18-70 and 555-604 of PGHS-2). Sequences used in GST fusion proteins are shaded (residues 575-599 of PGHS-1 , residues 565-604 of PGHS-2). 10 MATERIALS AND METHODS Construction of bait Plasmids For yeast two-hybrid screening, four pairs of upper and lower primers (Figure 5) complementary to the human PGHS-l-EGF domain, human PGHS-Z-EGF domain, human PGHS-1 carboxyl terminal sequence, and human PGHS-2 carboxyl terminal sequence were synthesized and used in PCR reactions to generate four corresponding PGHS cDNA fragments. The PCR products were first cloned into the plasmid pCR2.l TA (Invitrogen), then subcloned into GAIA-DNA-binding-domain vector pASZ-l (which carries a selectable marker T RPl) (Table 1). Structures of the four pAS2-1 constructs were confirmed by restriction endonuclease analysis and DNA sequencing. The three pAS2-1 constructs, which encoded the human PGHS-l-EGF domain (residues 24-84), the PGHS-Z-EGF domain (residues 18-70), or the PGHS-2 carboxyl terminal sequence (residues 555-604) fusion proteins were designated as pASlEF G, pAS2EGF and pASZC respectively, and were used as baits in the yeast two-hybrid system (Table I). The fourth pAS2-1 construct, which encoded the human PGHS-l carboxyl terminal sequence (residues 556-599) fusion protein, was designated as pASlC and was used as a control for clones that reacted with pASZC. To ensure that these PGHS baits by themselves did not activate transcription, they were used to transform the yeast strain Y190 (Table II) separately. The Y190 transformants were selected on minimal media lacking tryptophan and were tested for autonomous reporter gene (lacZ) activation by the B-galactosidase assay. The transformants expressing the GAL4 fusion proteins were confirmed by western blot using a monoclonal antibody (mAb) against the GAL4-DNA-binding- ll Human PGHS-l-EGF Bait 5’ Primer, 29mer, nucleotides coding for the peptide sequence Ala-ASp-Pro-Gly-Ala-Pro-Thr (residues 24-30): 5’ CCC ATA TGG CGG ACC CAG GGG CGC CCA CG 3’ Ndel 3’ Primer, 31mer, complementary to nucleotides coding for the peptide sequence Arg-Asn-Ser-Leu-Arg-Pro-Ser-AMB (residues 78-84): 5’ CCG ACG TCC TAG CTG GGC CGC AGT GAA 'I'I‘C C 3’ Human PGHS-l C’ Bait 5’ Primer, 28mer, nucleotides coding for the peptide sequence Asn-Ile-Val-Lys-Thr-Ala-Thr (residues 556-562): 5’ TCC ATA TGA ACA TTG TCA AGA CGG CCA C 3’ Ndel 3 ’ Primer, 29mer, complementary to nucleotides coding for the peptide sequence Arg-Pro-Ser-Thr-Glu-Leu-OPA (residues 594-599): 5’ CGG AAT TCT CAG AGC TCT GTG GAT GGT CG 3’ 13le Human PGHS-Z-EGF Bait 5’ Primer, 29mer, nucleotides coding for the peptide sequence Ala-Asn-Pro-CyS-Cys-Ser-His (residues 18-25): 5’ CCC ATA TGG CAA ATC CTC CTT GCT GTT CC 3’ Ndel 3’ Primer, 33mer, complementary to nucleotides coding for the peptide sequence Lys-Leu-Phe-Leu-Lys-Pro-Thr-AMB (residues 64-70): 5’ GGC CTA GGC TAA GTG GGT TTC AGA AAT AAT TTT 3’ Human PGHS-2 C’ Bait 5’ Primer, 28mer, nucleotides coding for the peptide sequence Asn-Asn-Val-LyS-Gly-Cys-Pro (residues 555-560): 5’ CCC ATA TGA ATA ACG TGA AGG GCT GTC C 3’ Ndel 3’ Primer, 29mer, complementary to nucleotides coding for the peptide sequence Glu-Arg-Ser-Thr-Glu-Leu-AMB (residues 598-604): 5’ GGC TTA AGC TAC AGT TCA GTC GAA CGT TC 3’ Figure 5. Specific primers used in PCR reactions to generate PGHS-l-EGF, PGHS-l carboxyl terminal, PGHS-Z-EGF and PGHS-2 carboxyl terminal cDNA sequences for construction of the yeast two-hybrid bait plasmid. (OPA: opal stop codon; AMB: amber stop codon) 12 TABLE I DESCRIPTION OF THE PLASMIDS Vector Mption Size Clonming vectors pAsz-l GAL4(1_147) DNA-ED, mm, 3.41:}, amp’, C ”F 2, pACTZ GAL4(768—88I) AD, LEU2,, 8.1kb amp’, HA epitope tag Recombinant plasmids . p AS lEGF PGHS-l-EGF (24-84) in pAS2-l 8,5kb IRPI, arnp’ p ASZEGF PGHS-Z-EGF (18-70) in pASZ-l 8.5kb IRPI, amp’ pASlC PGHS-l-COOH (556-599) in pASZ-l 8.5kb m1, amp’ pASZC PGHS-Z-COOH (SSS-604) in pASZ-l 8.5kb m1, amp’ pASnIS Arts in pASZ-l TRPI, ampr 9-1kb pASnl8 AT18 in pASZ-l mm, mi 96% pACT2C PGHS-Z-COOI-I (SSS-604) in pACT2 8.3kb LEU2,, amp” PM"15 AT15 in pACT2 LEU2,, amp’ 8-8kb PAT” AT18 in pACT2 1.1502,, mp7 93% l3 TABLE I DESCRIPTION OF THE PLASMIDS (continued) Vector Description Size Clonming vectors pCR2.l kan', amp’ 3.9kb pET-28a kan’ , 6 X Histidine-tag 5.4kb pGEX-4T amp’, GST-tag 5.0kb Recombinant plasmids pET-AT15 AT15 in pET-28a 6.21m kan’, 6 XHistidine—tag pET-AT18 AT18 in pET-28a 6.7kb kan’, 6 X Histidine-tag l4 TABLE II GENOTYPES OF THE YEAST HOST STRAINS Strain Genotype Reporters Auxotrophy Y190 CG1945 MA Ta, ura3-52, his3-200, lysZ-SOI, [-1153, [ac Z adeZ-l 01, "pl-901, leu2-3, I 12. gal4-del. gal80-del, cyhrZ. LYSZJJGALI UAS'HIS3TA TA'HIS3' URS3.'.’GAL I UAS'GAL I TA TA-lacZ MA Ta, ura3-52, his3-200, IysZ-801, H153, [ac Z adeZ-I 01. trpl -90 I . Ieu2-3. I 12. gaI4-del, gnu-542, gal80-538.cyhr2, LYSZ::GALI (MS—GAL] TA TA-HIS3, URS3::GAL4 I 7.":ers(x3)-CyCl TA TA- lacZ trpI ,Ieu2, cyh' 2 trp1 ,leu2, cyhr 2 15 domain (Clontech). Library Plasmids The human brain cDNA library was constructed in the GAIA-activation-domain vector pACT2 (Table I), which carries a selectable marker LE U2. It was purchased from Clontech. Yeast Two-Hybrid Screening and Testing of Positive Clones The yeast two-hybrid library was screened following transformation into the Y190 yeast strain, which harbors the reporter genes HIS3 and IacZ under the control of upstream GAL4 transcription regulatory element (Clontech, Matchmaker Two-Hybrid System 2). Briefly, Y190 was co-transformed with a PGHS bait plasmid and cDNA library constructed in the GAL4-transcriptional-activation-domain vector pACT2. An estimated 3,500,000 recombinants from the human brain cDNA library were screened. Yeast co-transfonnation was done by a procedure using lithium acetate, single stranded DNA and polyethylene glycol (22). Double-transformants containing plasmids encoding PGHS bait and interacting proteins were selected for on minimal media deficient in tryptophan, histidine and leucine, and containing 25 mM 1,2,4-3-aminotriazole (3-AT) (Sigma). Yeast colonies surviving the medium selection were subsequently assayed for B- galactosidase activity in the presence of 5-bromo-4-chloro-3-indolyl-B-D-galactoside (X- gal) by filter lifi assays (Figure 6). Plasmids containing the library genes that interacted with PGHS baits were isolated fiom B-galactosidase-positive colonies and transformed into E. coli strain DH5a for amplification. These plasmids were then co-transfonned into yeast strain CD 1 945 (Table II) along with their PGHS bait partners. The double transformants were subjected 16 GAL4 DNA-BD bait plasmid GAL4 AD fusion library Marker: TRPl Marker: LEU2 Cotransform yeast reporter strain Y190 Plate culture on SDl-I-lisl-LeuI-Trp Only His+ colonies grow: these are candidates for a two-hybrid interaction Perform B-galactosidase assay Blue color: B-galactosidase + White color: B-galactosidase - indicates putative positive colony one class of false positives, (interaction between the likely due to nonspecific two-hybrid proteins) activation of the HIS3 reporter only Figure 6. Screening an GAL4-Activation Domain (AD) fusion library for proteins that interact with a bait protein. Double-transforrnant colonies, which survive the His’ selective medium and have B-galactosidase activity, contain potential-bait-interacting hybrid proteins. to a second round of testing to verify the ability of the library plasmids to induce transactivation of the HIS3 and IacZ reporter genes. All library plasmids that passed the above tests were transformed alone into CG1945 yeast. If their encoded proteins possessed intrinsic transactivating activities toward the reporter gene, they were discarded. The plasmids that passed those verification procedures were again transformed into CG1945 with a plasmid encoding lamin C fused to GAL4-ED or the fourth pAS2-l bait construct, which encoded the human PGHS-l carboxyl terminal sequence (residues 556-599) fused to GAL4-ED (23). These pairings tested for non-specific activation by the pACT2 encoded proteins via a mechanism other than interacting with GAL4-BD. The clones encoding proteins that interacted with lamin C or PGHS-1 carboxyl terminal sequence were discarded. Plasmids that passed all the above tests, were demonstrated to encode proteins that interacted specifically only with the PGHS baits (Figure 7). Sequences of the candidate library cDNA clones were compared to GenBank database with the Blasta and Tfasta sequence comparison programs. Expression of AT15 and AT18 proteins in Bacteria To express the two proteins AT15 and AT18, found to interact with pASZC, their respective cDNA insert was digested out of the pACT2 library plasmid, and subcloned, in-frarne, into the prokaryotic expression vector pET-28a (N ovagen), giving rise to pET- ATl 5 and pET-AT18. The identity and orientation of the constructs were confirmed by DNA sequencing. The pET-ATl 5 or pET-AT18 vector expressed the AT15 or AT18 protein as a fusion protein containing 6 consecutive histidine residues (His-tag) at the NH; terminus. The His-tag was used for purification of the recombinant AT15 or AT18 protein. The plasmids were transformed into E. coli strain BL21(DE3), a Ion mutant l8 Isolate AD-library DNA from putative positive colony i Transform E. coli i Isolate AD-library DNA from E. coli (fl m g.) k_J k_2 Cotransformation (l) AD-Iibmy DNA (2) AD-library DNA (3) AD-library DNA +Bait plasmid + L i C plasmid B-galactosidase assay (1) Blue colonies (2) White colonies (3) White colonies if true positive if true positive if true positive Figure 7. Protocol used to eliminate false positives that arise in a two-hybrid library screening. Plasmids, which didn't intrinsically activate transcription, which encoded proteins interacting specifically only with the PGHS baits, not with nonspecific protein, e.g. Lamin C, were selected as true postives. l9 strain containing the T7 polymerase under the control of lac UV5 promoter. Addition of isopropyl-B-D-thiogalactopyranoside ([PTG) induces the expression of the fusion proteins in this system. The His-tag-ATIS and His-tag-AT18 were purified from crude bacterial lysates by Ni-NTA Resin (Qiagen) according to the manufacturer’s instruction. In vitro Elution Assay 50 ug of purified ovine PGHS-1 or mouse PGHS-2 were incubated together with 50 pg of either His-tag-ATlS or His-tag-AT18 in 1 ml PBS (50 mM sodium phosphate, 300 mM NaCl, pH8.0) at 4 °C for l h to allow dimerization. Ovine PGHS-1 and mouse PGHS-2 were provided by Mike Malkowski. 200 ill of 50% slurry of Ni-NTA agarose was then added to each mixture, and the mixture was incubated at 4 °C for l h, followed by centrifugation at 1000 rpm for 5 min at 4 °C. The supematants were aspirated, and the pellets were washed 3 times with 1 ml PBS and twice with PBS containing 10% glycerol (pH6.0) by repeated centrifugation. Bound proteins were eluted with 1 ml PBS containing 10% glycerol and 300mM irnidazole. Aliquots (20 ul) of the eluted proteins were separated by electrophoresis on a 15% polyacrylamide gel using Tris glycine/SDS buffer (Qiagen)- In vitro Slot Blotting The carboxyl terminal sequences of PGHS-l (residues 575-599) and PGHS-2 (residues 566-604) were generated by PCR, subcloned in-frame into pGEX-4T (Phannacia) and transformed into E. coli to generate GST-PGHS-l-COOH and GST- PGHS-Z-COOH fusions (Figure 4). The fusion proteins were purified from crude bacterial lysates by Mike Malkowski under non-denaturing conditions by chromatography on glutathione-Sepharose (Phannacia). 20 Purified GST-PGHS-l-COOH, GST-PGHS-Z-COOH, ovine PGHS-l, and mouse PGHS-2 were blotted onto nitrocellulose, and incubated with 0.5 11ng His-tag-ATl 5 or His-tag-AT18 in PBS at 4 °C for l h. After washing the nitrocellulose filters with PBS (3 times), bound His-tag-ATl 5 and His-tag-AT18 were visualized on X-ray film using Ni- NTA conjugated horseradish peroxidase (HRP) and a chemiluminescence reagent. BSA and His-tag-ATl 5, His-tag-AT18 were also blotted onto the nitrocellulose as controls. In vitro Afl'igel Assay AT15 and AT18 affinity agarose were prepared by irreversibly cross-linking His- tag-ATl 5 or AT18 protein to Affigel agarose beads (1 mg/ml bed volume, Bio-Rad Laboratories). 40 ul of beads were then incubated in PBS with 4 ug purified human PGHS-2 for 1 h at 4 °C, and washed three times with PBS. Bound PGHS-2 was eluted by boiling in l x SDS-PAGE loading buffer and the samples were analyzed by western blotting using polyclonal antibody against PGHS-2 (24). Additional Two-Hybrid Testing in Yeast to Verify Interaction AT15 and AT18 cDNA inserts were released by digestion with restriction enzymes from the yeast library plasmids, GAL4-activation—domain vector pACT 2 (Table I), and subcloned in-frame into a GAIA-DNA-binding-domain vector, pAS2-l , to construct pASnl 5 and pASn18. The two pASZ-l constructs were confirmed by restriction endonuclease analysis and DNA sequencing. These plasmids were transformed into CG1945 (Table II) separately. The CGl945 transformants were selected on minimal media lacking tryptophan and were tested for autonomous reporter gene (lacZ) activation by a B-galactosidase assay. The transfonnants were also analyzed for expression of the firsion proteins by western blot using the monoclonal antibody (mAb) against the GAL4- 21 DNA-binding-domain (Clontech). The PGHS-2 carboxyl terminal sequence (residues 555-604) was excised from the GAL4-DNA-binding-domain vector pAS2- l , and subcloned in-fi'ame into the GAL4- activation-domain vector pACT2 to construct pACT2C. The structure of the pACT2C construct was confirmed by restriction endonuclease analysis and DNA sequencing. CG1945 transformant of pACT2C was selected on minimal media lacking leucine and was tested for autonomous reporter gene (lacZ) activation by a B-galactosidase assay. The transforrnant was also analyzed by western blot using the monoclonal antibody (mAb) against the GAIA—activation-domain (Clontech). The pASnl 5 and pASn18 plasmids were then transformed into CG1945 either alone or with pACT2C or with the empty GAL4-activation-domain vector pACT2. Single and double transformants were tested for autonomous reporter gene lacZ activation by a B-galactosidase assay and autonomous reporter gene HIS activation by re-streaking on minimal media lacking histidine. 22 RESULTS Cloning of ATIS and AT18 by Yeast Two-Hybrid Screening Four cDNA sequences, corresponding to the PGHS-l-EGF domain (residues 24- 84), the PGHS-Z-EGF domain (residues 18-70), PGHS-2 carboxyl terminal sequence (residues 555-604), and PGHS-l carboxyl terminal sequence (residues 556-599) (Figure 3 and 4), were cloned into the GAIA-DNA-binding-domain vector pASZ-l, to generate four PGHS-bait plasmids: pASlEGF, pASZEGF, pAS2C and pASlC. Three of them: pASlEGF, pASZEGF, pAS2C were used as baits for yeast two-hybrid screening to identify proteins that might interact with the PGHS isozymes. All four of them were used separately to transform the yeast strain Y190 to check for intrinsic transactivation activity toward the host reporter genes. The transfonnants were subjected to a B-galactosidase assay and were found to express no lacZ activity, showing that the bait plasmids were suitable for library screening. Western blot analysis, using the transforrnant lysates and monoclonal antibody (mAb) against the GAL4-DNA-binding-domain, showed that all four baits were properly expressed and fused with the GAIA-DNA-binding-domain in the correct orientation and reading frame. An estimated 3,500,000 human brain library transformants were screened using pASlEGF, pAS2EGF and pASZC. We used this library, because PGHS-2 is constitutively expressed in the brain. Among these, approximately 940,000 were screened with pAS2C, 300,000 with pAS2EGF, and 2,200,000 with pASlEGF (Table III). From these combined screens, a total of 58 colonies survived selection on minimal media lacking Trp, His and Len, but only 21 of these contained detectable B-galactosidase 23 38. .oz - 0989.82 usage 95. _ 3.2 @382 .oz 532.. e855 aseaeaez BS. 33“ «OZ .30 05 .o mmvm seen 33...: _ 2. z 85. m S 8 8°82 name.— 38. 82 058%.82 use; .5. «a: 38. 82 £295 e855 328.252 2.5. n 888 new u e. e a 3:? M 3: s e ees s awe-“Wm. a E... 2.: s e Es s 883 5-0 38% NA: m _.._.< m e um NimmOm 2.: s e E»: s 389 NE 3? 8.5.8 e 3:023 2.38.5 REESE Bee—U newsflaeuanabeo 25.8.. 8.: :5 85.58335 .1: 95: been 8:29 328m 3 anger—am BADmm—u— ZHHm—Um Guy—EOE hm<fl> :— HAM—4Q. 24 activity and were studied further. When transformed alone, 4 of the 21 clones were eliminated because they intrinsically transactivated the reporter genes. Another 9 were eliminated because they showed nonspecific binding activity to heterologous lamin C protein. Eight clones survived all tests. These clones were sequenced. It was determined that two clones which interacted with the carboxyl terminal PGHS-2 bait (AT14 and AT15), contained identical cDNA fragments. The open reading frame of this cDNA encoded a 260 amino acid fiagment showing 70% amino acid identity to Lin-7 (C. elegans), 41% identity to PSD-95 and 38% identity to thg-l/-2 over its length. An alignment of these cDNAs revealed the homology of AT14 with AT15 corresponded to the shared PDZ repeat domains in these proteins (Figure 8), which are protein modules that bind to the carboxyl terminal ends of target proteins (25-27). Another clone AT18 that also interacted with the carboxyl terminal PGHS-2 bait, encoded for a fragment of 566 amino acids with 69% amino acid identity to a cortactin- binding protein. Two of the clones that interacted with the PGHS-l-EGF domain, were ATP synthase: a protein commonly obtained as an artifact in the yeast two-hybrid screening. They were not examined further. Three other clones identified by the interaction with the PGHS- l -EGF domain, were novel proteins not present in the genetic databases, and were also not examined further (Table III). To verify that the proteins encoded by the pAT14/15 and pAT18 interacted specifically with PGHS-2, these clones were transformed into the CG1945 yeast host strain in the presence or absence of plasmid pASZ-l which encoded the GAL4 DNA- binding-domain, or pAS2C which encoded the PGHS-2 carboxyl terminal sequence, or pASlC which encoded the PGHS-l carboxyl terminal sequence. Both pAT14/15 and 25 I PDZ 260 Candidate Clone ATl4/AT15 21:2 1. Lin-7 (70% Id) PDZ 1 199 no 295 —- 2. PSD-95 (41% Id) . PDzr PDZZ run 8113 GK I ‘5 151160 2“ 313 393 435 ‘95 534 712 724 3' hm“ (38% 1") run run run sns GK I 162/194 224 310 319 403 4“ 546 588/668 669/702 73‘ 914 926 Figure 8. Two of the candidate clones isolated by the yeast two-hybrid screening: AT14 and AT15. Showing are the alignment between the candidate clone and the PDZ domains of Lin-7, PSD-95 repeat 3 and thg repeat 3. Percent identidies for the PDZ domains are shown (percent Id.). (Lin-7: C. elegans protein invoved in vulval development, PSD-95: post-synaptic density protein, thg: human DlgA homologous protein, DlgA: Drosophila discs-large tumor suppressor protein.) pAT18 clones were positive in the B-galactosidase activity assay only when co- transformed with the plasmid encoding the PGHS-2 carboxyl terminus (pAS2C). Therefore, we concluded that the isolated clones AT14/15 and AT18 represent proteins that interact specifically only with the carboxyl terminus of PGHS-2, and not with the carboxyl terminus of PGHS-l , or the GAL4-DNA-binding-domain alone. In vitro Testing of PGHS-2-AT15 and PGHS-2-AT18 Interaction To test the in vitro binding of AT15 and AT18 to PGHS-2, AT15 and AT18 were expressed as His-tag fusion proteins in bacteria and were purified by chromatography using Ni-NTA resin (Figure 9). Purified His-tag-ATI 5 or His-tag-AT18 was incubated with purified ovine PGHS-l or mouse PGHS-2 to allow the formation of AT] 5-PGHS-2 or AT18-PGHS-2 complexes, which were then precipitated by nickel chelated Ni-NTA agarose. As shown in Figure 10, His-tag fusion proteins AT] 5 and AT18 were precipitated with Ni-NTA agarose, but PGHS-2 was not co-precipitated with either AT15 or AT18 in this assay. A second approach to demonstrate an interaction with AT15 and AT18 was then tried in which purified GST-fusion PGHS-l carboxyl terminus, GST-fusion PGHS-2 carboxyl terminus, ovine PGHS-l, and mouse PGHS-2 were applied by slot-blotting to nitrocellulose filters. These filters were then incubated with His-tag-ATlS or His-tag- AT18, and bound His-tag proteins were detected with Ni-NTA conjugated HRP. Neither His-tag-ATI 5 nor His-tag-AT18 could be detected bound to either GST fusion PGHS-2 carboxyl terminus or the holo-PGHS-2 protein (Figure 1 l). Purified His-tag-ATl 5 or AT18 protein was next cross-linked to Affi gel agarose beads (1 mg/ml) and incubated with l/ 10 the amount of the purified human PGHS-2. 27 Figure 9. SDS-PAGE analysis of the expression of His-tag-AT18 and AT15 protein. AT18 and AT15 were subcloned into pET-28a to generate His-tag fusion proteins. E. coli strain BL21(DE3) was transformed with the indicated recombinant pET-28a plasmid, and protein expression was induced by addition of IPTG. The His-tag fusion proteins were purified by Ni-NTA resin. Homogenates (20 ug) of BL21(DE3) bacteria transformed with no plasmid (lane 2), pET-ATlS (lane 3), or pET-AT18 (lane 4) and purified His-tag-ATIS (lane 5) and His-tag-AT18 (lane 6). Molecular weight marker is in lane 1. 28 1 2 3 4 s 6 7 8 910 11 er10'3 ' -200 -116 -97 PGHS-l&-2- - ~' - 66 His-tag-ATI 8 - —45 His-tag-ATls- -3, Figure 10. In vitro Binding Assay. 50 pg purified His-tag-AT15 or -AT18 were incubated with 50 ug PGHS to allow heterodimerization. The His-tag-ATIS and His-tag-AT18 were then precipitated with Ni-NTA agarose. Following washing of the agarose beads, bound proteins were eluted with 0.3M imidizole, separated by 15% SDS-PAGE, and stained with Coumassi Blue. Lane 1. His-tag-AT18. Lane 2. His-tag—AT18 + ovine PGHS-l. Lane 3. His-tag- AT18 + mouse PGHS-2. Lane 4. His-tag-galactosidase (control). Lane 5. His-tag- galactosidase + ovine PGHS-l (control). Lane 6. His-tag-galactosidase + mouse PGHS-2 (control). Lane 7. His-tag-ATIS. Lane 8. His-tag-ATIS + mouse PGHS-2. Lane 9. ovine PGHS-l . Lane 10. mouse PGHS-2. Lane 1 l. Bio-Rad SDS-PAGE MW standard. 29 A Incubated w/ AT15 2 118 5 its 10 us PGHS-l-hu PGHS-Z-hu PGHS-l-ov . PGHS-Z-mu , BSA ,' AT15 W CI- .- B Incubated w/ AT18 2|.lg Sug 10 jig PGHS-l-hu PGHS-Z-hu PGHS-l-ov l PGHS-Z-mu BSA 1 .. _ " . j s.-. AT18 “II-'- ~ “ Figure 11. Slot Blotting. Three aliquots (2 jig, 5 jig, or 10 jig) of GST fusion human PGHS-1 carboxyl terminus, GST fusion human PGHS-2 carboxyl terminus, ovine PGHS-l, mouse PGHS-2, BSA, His-tag-ATlS or His-tag-AT18 were blotted onto nitrocellulose. The filters were probed with either 0.5 mg/ml His-tag-ATI 5 (Panel A) or His-tag-AT18 (Panel B), and bound His-tag-ATI 5 and His-tag-AT18 were detected with Ni-NTA conjugated HRP. 30 Total PGHS2_ AT15 AT18 BSA Sham 12345678910111213 -WA--~-- ‘ .na"*~+m ._or. ' ‘ Figure 12. Immunoblot of PGHS-2 retained on Affigel. Equal amounts of purified human PGHS-2 (lane 1) were incubated with Affigel beads (at a ratio of l jig PGHS-2/10 jil beads) containing either His-tag AT15 (lane 2, 3) or His-tag AT18 (lane 4-9) or BSA (lane 10, 11) or sham (lane 12, 13). The retained PGHS-2 were eluted with 1x SDS loading buffer and detected by Western blotting. No consistent reproducible or specific binding of PGHS-2 to either Affigel-His-tag-ATI 5 or -AT18 was observed. 31 Following washing of the Affigel, the Affigel was treated with l x SDS sample buffer and the sample buffer was analyzed. As shown in Figure 12, no consistent reproducible or specific binding of PGHS-2 to either Affigel-His-tag-ATl 5 or -AT18 could be detected by western blot analysis using a PGHS-2 polyclonal antibody (24). In viva Testing in Yeast to Verify Interaction Since the above in vitro binding assays could not demonstrate that AT15 and AT18 bind PGHS-2 in vitro, additional experiments were designed to detect the existence of AT15-PGHS-2 and AT18-PGHS-2 complexes in yeast. Cloning vectors were switched by moving the AT15 and AT18 insert from the GAIA-activation-domain to the GAL4- DNA-binding-domain vector, and PGHS-2 carboxyl cDNA sequence fi'om the DNA-BD to AD vector. Six constructs pASnl 5, pASn18, pAS2C (containing AT15, AT18 or PGHS-2 carboxyl terminal cDNA in the pASZ-l vector, respectively), pATl 5, pATl 8 and pACT2C (containing AT15, AT18 or PGHS-2 carboxyl terminal cDNA in the pACT2 vector, respectively) were used to transform CGl945 separately. The transformants were selected on minimal media lacking Trp or Leu. Western blotting, using the transfonnant lysates and monoclonal antibody (mAb) against the GAL4-DNA- binding-domain or GAL4-activation-domain, showed that all six insert proteins were properly expressed and fused with GAIA-DNA-binding-domain or GAL4-activation- domain in the correct orientation and reading frame (Figure 13). When the two-hybrid assay was repeated, yeast co-transformed with pASn15 (containing the AT15 cDNA in pASZ-l vector) and pACT2C (containing the PGHS-2 carboxyl terminal cDNA in pACT 2 vector) had no detectable B-galactosidase activity. These co-transforrnant colonies also couldn’t grow on histidine selected media when re- 32 A B C er10‘312345 12345 1234 “6'? Exit-’44,?“ i“ .:. 80—erl " ' Figure 13. Western blots of candidate proteins and bait protein using the GAL4 DNA-BB, AD mAbs and Dinah (Ab against PGHS2 carboxy terminus). Soluble protein extracts were prepared from yeast strain CG1945 transformed (separately) with the indicated vector. Samples equivalent to ~1-l .5 OD600 units of cells were resolved on a 15% polyacrylamide/SDS gel and electroblotted to a PVDF filter. The blots were probed with either the GAL4 DNA-BD mAb (0.5 jig/m1; Panel A) or the GAL4 AD mAb (0.4 jig/ml; Panel B) or Dinah (0.1 jig/ml; Panel C), followed by HRP-conjugated polyclon- al Goat Anti-Mouse IgG or HRP-conjugated polyclonal Goat Anti- Rabbit IgG. Signals were detected using a ECL detection assay and a l-5min exposure of X-ray film. Panel A. Lane 1: untransformed CGl945 control. Lane 2: pAS2 (a GAL4 DNA-BD vector). Lane 3: pAS2C. Lane 4: pASn15. Lane 5: pASn18. Panel B. Lane 1: untransformed CG1945 control. Lane 2: pACT2 (a GAL4 AD vector). Lane 3: pACT2C. Lane 4: pATlS. Lane 5: pATl8. Panel C. Lane 1. Human PGHS-2 standard. Lane 2. untransformed CG1945 control. Lane 3. pAS2C. Lane 4: pACT2C. 33 streaked (Figure 14). The fact that no interaction of these hybrid proteins could be observed after switching the vectors suggests that AT15 does not interact with the PGHS- 2 carboxyl terminus in a biologically relevant manner. Yeast co-transfonned with pASnl 8 (containing AT18 cDNA in pAS2-l vector) and pACT2C had strong B-galactosidase activity. However, yeast transformed with the pASn18 plasmid alone or co-transformed with pASn18 and the empty pACT2 vector, also had B-galactosidase activity, albeit less than when co-transformed with both pASnl 8 and pACT2C. All the transformant colonies grew on histidine selected media when re- streaked. This suggested that while AT18 might interact with the PGHS-2 carboxyl terminus, AT18 also had intrinsic transactivating activity toward the host reporter genes when expressed as a GAL4 DNA-BD chimeric protein. 34 Gal4 DB PGHS2c' pAS2CN F llllllllll C G814 AD HIS3 Beta-Gal pACT2 (Control) I 7 7 7 7| - - AT15 pATls ZZZ; + ++ AT18 pm 2222mm + +++ Ga14 DB AT15 pASnIS N r - C 01114 AD HIS3 Beta-Gal pACT2 (Control) I 7 7 7 7| - - PGHS2c' mm EZZZZEEIIEIIm - - Ga14 DB AT18 pASnls NI m0 5314 AD HIS3 Beta-Gal pACT2 (Control) r/7/A + + PGHSZc' pACT2C? IZZZZJIIHJJIHI + ++ 35 Figure 14. Interaction between AT15, AT18 protein with PGHS-2 C-terminus as determined by semiquantitative yeast two-hybrid assays based on the degree of induction of the two reporter genes HIS3 and B-gal. Yeast strain CG1945 was co-transfonned with the indicated plasmids. HIS3 activity was determined by the yeast colonies grown on medium lacking histidine: +, growth; -, no significant growth. B-gal activity was estimated from the time taken for colonies to turn blue in 5-bromo -4-chloro-3-indolyl b-D-galactoside filter lift assays at room temperature: +-H, <30 min -H-, 45-90 min; +, 90-480 min; -, no Significant activity. DISCUSSION The use of the yeast two-hybrid system to identify PGHS associated proteins in the endoplasmic reticulum (ER) has been found to be more complicated than searching for cytosolic proteins (21). The main difficulties result because PGHS contains a membrane-binding domain, and requires disulfide-bond formation and N-glycosylation for proper folding. Since the yeast system lacks the proper protein modification system, it should be advantageous to use small PGHS domains as baits in the yeast two-hybrid screening. This logic formed the partial basis for our choosing of the carboxyl terminal sequence of PGHS-2 as bait for these studies. Another reason that this sequence was chosen is that the carboxyl terminus of PGHS-2 contains sequences unique to this isozyme that we thought likely to have a specific function. Our choosing of the EGF domains of PGHS-l and PGHS-2 as baits was based on the fact that EGF domains commonly function as protein-protein interaction domains, and they have already been identified as dimerization domains in the PGHS homodirner (28). Five candidate clones were obtained with the PGHS-l-EGF bait, three of which are novel genes not known in the database, and were not characterized. Three candidate clones were obtained with the PGHS-2 carboxy] terminal bait, among which was a novel protein, AT15, possessing a PDZ domain, and AT18 a novel protein similar to cortactin binding protein. Since these clones interacted only with the PGHS-2 carboxyl terminus, which contains the 18 amino acid insert, and not with the PGHS-1 carboxyl terminus, we focused our characterization on these two candidate proteins: AT15 and AT18. The PDZ domains are protein modules that bind to the carboxy] terminal ends of 36 target proteins. They were originally identified in the post-synaptic density protein PSD- 95 as three repeats of about 90 residues containing the conserved motif Gly-Leu-Gly-Phe (GLGF) (26). PDZ domains have since been found in a diverse and growing set of proteins with different functions, e. g. p55, thg, Dsh, etc (29). Since PDZ domains appear to mediate direct protein-protein interactions by interacting with the carboxyl termini of the target proteins, it was particularly interesting for us to further characterize AT15’S binding specificity to PGHS-2. We expressed and purified His-tag fusion proteins of AT15 and AT18 in E. coli, and tried different approaches to detect the interaction between AT15 and PGHS-2, or AT18 and PGHS-2. However, none of the in vitro assays could prove that AT15 or AT18 bound PGHS-2. ' The lack of protein modification in E. coli suggested that perhaps only in yeast were AT15 and AT18 properly modified, or that there might be a third protein in yeast required for the interaction of AT] 5 and AT18 with the PGHS-2 carboxyl terminus. Therefore we re-tested the binding capacity of AT 1 5 and AT18 to PGHS-2 carboxyl terminus in yeast by switching vectors, moving the AT15 and AT18 insert from the activation-domain (AD) vector to the DNA-binding-domain (DNA-BD) vector, and PGHS-2 carboxyl cDNA sequence from the DNA-BD vector to the AD vector. When the two-hybrid assay was repeated for AT15 and PGHS-2 carboxyl terminus after switching vector, neither the B-galactosidase nor the histidine reporter gene’s transcription was transactivated. The lack of reporter gene expression suggests that AT15 does not interact ' with the PGHS-2 carboxyl terminus afier the vector switching. On the other hand, transcription of both B-galactosidase and histidine genes were 37 activated in yeast either transformed with AT18 in the GAL4-DNA-binding-domain (DNA-BD) vector alone, or co-transformed with AT18 and the empty activation-domain (AD) vector, or with AT18 and PGHS-2 carboxyl terminus in the AD vector. The intensity of B-galactosidase activity in yeast co-transformed with both AT18 and PGHS-2 carboxyl terminus was stronger than that with single AT18 or with AT18 and the empty AD vector. The expression of B-galactosidase and histidine might suggest that AT18 interacted with the PGHS-2 carboxyl terminus after the vector switching. However, AT18 also had intrinsic transactivating activity toward the host reporter genes when it was fiised with GAL4-DNA-binding-domain, suggests that AT18 may be a transcriptional activator by itself. The cumulative evidence led us to conclude that neither AT15 nor AT18 interacts with PGHS-2 in any biologically relevant manner. In general, the yeast two-hybrid system has been used extensively to detect protein-protein interactions (20, 30). However, detection of a Specific interaction between the bait protein and the library-encoded proteins in this heterologous assay system does not necessarily indicate that there is a corresponding interaction in the proteins’ native environment. For example, two members of a protein family (i.e., leucine zipper protein family) may produce a signal in the yeast two-hybrid assay, but they may never normally be present in the same cell type, cellular compartment, or during the same stage of the cell cycle. Similarly, the two-hybrid system may detect interactions between domains of proteins that are not present in the same subcellular location in a cell. Alternatively, if an interaction is mediated via a short sequence, that sequence may not be exposed on the native protein. Therefore, it is very important to verify the protein-protein interactions detected by the two-hybrid system by other biological or biochemical experiments. 38 REFERENCE Smith W.L., Marnett L.J.: Prostaglandin endoperoxide synthases, in: Sigel H., Sigel A., (eds): Metal Ions in Biological Systems. Vol. 30. New York: Marcel Dekker, 1994, pp. 163-199 Xie W., Chipman J .G., Robertson D.L., Erikson R.L., Simmons D.L.: Expression of a mitogen-responsive gene encoding prostaglandin synthase is regulated by mRNA splicing. Proc Natl Acad Sci USA 88:2692-2696, 1991 Kujubu D.A., Fletcher B.S., Vamum B.C., Lirn R.W., Herschman H.R.: TISlO, a phorbol ester tumor promoter inducible mRNA from Swiss 3T3 cells, encodes a novel prostaglandin synthase/cyclooxygenase homologue. J Biol Chem 266:12866-12872, 1991 O'Banion M.K., Sadowski H.B., Winn V., Young DA: A serum- and glucocorticoid-regulated 4 kb mRNA encodes a cyclooxygenase-related protein. J Biol Chem 266:23261-23267, 1991 Barnett J., Chow J., Ives D., Chiou M., Mackenzie R., Osen E., Nguyen B., Tsing S., Bach C., Freire J ., et a1: Purification, characterization and selective inhibition of human prostaglandin G/H synthase 1 and 2 expressed in the baculovirus system. Biochim Biophys Acta 1209: 1 30-139, 1994 Langenbach R., Morham S.G., Tiano H.F., Loftin C.D., Ghanayem B.I., Chulada P.C., Mahler J.F., Lee C.A., Goulding E.H., Kluckman K.D., et a1: Prostaglandin synthase 1 gene disruption in mice reduces arachidonic acid-induced inflammation and indomethacin-induced gastric ulceration. Cell 83:483-492, 1995 Morham S.G., Langenbach R., Loftin C.D., Tiano H.F., Vouloumanos N., Jennette J.C., Mahler J.F., Kluckman K.D., Ledford A., Lee C.A., et a1: Prostaglandin synthase 2 gene disruption causes severe renal pathology in the mouse. Cell 83:473-482, 1995 Murakami M., Matsumoto R., Austen K.F., Arm J .P.: Prostaglandin endoperoxide synthase-1 and -2 couple to different transmembrane stimuli to generate prostaglandin D2 in mouse bone marrow-derived mast cells. J Biol Chem 269:22269-22275, 1994 Reddy S.T., Herschman H.R.: Li gand-induced prostaglandin synthesis requires expression of the TISlO/PGS-2 prostaglandin synthase gene in murine fibroblasts and macrophages. J Biol Chem 269:15473-15480, 1994 39 10. ll. 12. 13. 14. 15. 16. 17. 18. 19. 20. Smith W.L., Garavito R.M., DeWitt D.L.: Prostaglandin endoperoxide H synthases (cyclooxygenaseS)-l and -2. J Biol Chem 271:33157-33160, 1996 Laneuville 0., Breuer D.K., Xu N., Huang Z.H., Gage D.A., Watson J .T., Lagarde M., DeWitt D.L., Smith W.L.: Fatty acid substrate Specificities of human prostaglandin-endoperoxide H synthase-1 and -2. Formation of lZ-hydroxy-(9Z, l3E/Z, 15Z)- octadecatrienoic acids fiom alpha-linolenic acid. J Biol Chem 270:19330-19336, 1995 Bhattacharyya D19, Lecomte M., Rieke C.J., Garavito M., Smith W.L.: Involvement of arginine 120, glutamate 524, and tyrosine 355 in the binding of arachidonate and 2-phenylpropionic acid inhibitors to the cyclooxygenase active Site of ovine prostaglandin endoperoxide H synthase-l. J Biol Chem 271:2179- 2184, 1996 Mancini J.A., Abramovitz M., Cox M.E., Wong E., Charleson S., Perrier H., Wang Z., Prasit P., Vickers P.J.: 5-lipoxygenase-activating protein is an arachidonate binding protein. FEBS Lett 318:277-281, 1993 Abramovitz M., Wong E., Cox M.E., Richardson C.D., Li C., Vickers P.J.: 5- lipoxygenase-activating protein stimulates the utilization of arachidonic acid by 5- lipoxygenase. Eur J Biochem 215:105-111, 1993 Giovannucci E., Egan K.M., Hunter D.J., Starnpfer M.J., Colditz G.A., Willett W.C., Speizer F.E.: Aspirin and the risk of colorectal cancer in women [see comments]. N Engl J Med 333:609-614, 1995 Eberhart C.E., Coffey R.J., Radhika A., Giardiello F.M., Ferrenbach S., DuBois R.N.: Up-regulation of cyclooxygenase 2 gene expression in human colorectal adenomas and adenocarcinomas. Gastroenterology 107:1183-1188, 1994 Sano H., Kawahito Y., Wilder R.L., Hashiramoto A., Mikai S., Asai K., Kimura S., Kata H., Kondo M., Hla T.: Expression of cyclooxygenase-1 and -2 in human colorectal cancer. Cancer Res 55:3785-3789, 1995 Kargman S.L., O'Neill G.P., Vicker P.J., Evans J.F., Mancini J.A., Jothy S.: Expression of Prostaglandin G/H synthase-l and -2 protein in hmnan colon cancer. Cancer Res 55:2556-2559, 1995 Oshirna M., Dinchuk J.E., Kargman S.L., Oshirna H., Hancock B., Kwong E., Trzaskos J .M., Evans J .F., Taketo M.M.: Supression of intestinal polyposis in ApcDelta7 16 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 87:803-809, 1996 Clontech Catalog # K1604-l: Matchmaker Two-Hybrid System 2 (PTlO30-1) 40 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Ballif B.A., Mincek N.V., Barratt J.T., Wilson M.L., Simmons D.L.: Interaction of cyclooxygenases with an apoptosis- and autoimmunity-associated protein. Proc Natl Acad Sci U S A 93:5544-5549, 1996 Gietz R.D., Schiestl R.H., Willems A.R., Woods R.A.: Studies on the transfonnation of intact yeast cells by the LiAc/SS-DNA/PEG procedure. Yeast 11:355-360, 1995 Bartel P, Chien CT, Stemglanz R, Fields S. Elimination of false positives that arise in using the two-hybrid system. Biotechniques l4(6):920-4, 1993 O'Sullivan M.G., Huggins Jr. E.M., Meade E.A., DeWitt D.L., McCall C.E.: Lipopolysaccharide induces prostaglandin H synthase-2 in alveolar macrophages. Biochem Biophys Res Commun 187:1123-1127, 1992 Sirnske J.S., Kaech S.M., Harp S.A., Kim S.K.: LET-23 receptor localization by the cell junction protein LIN-7 during C. elegans vulva] induction. Cell 85:195- 204, 1996 Cho K.D., Hunt C.A., Kennedy M.B.: The rat brain post-synaptic density fraction contains a homologue of the Drosophila discs-large tumor suppressor protein. Neuron 9:929-942, 1992 Lue R.A., Marfatia S.M., Branton D., Chishti A.H.: Cloning and characterization of hdlg: the human homologue of the Drosophila discs large tumor suppressor binds to protein 4.]. Proc Natl Acad Sci U S A 91:9818-9822, 1994 Picot D., 1.01] P.J., Garavito M.: The X-ray crystal structure of the membrane protein prostaglandin H2 synthase-l. Nature 367:243-249, 1994 Saras J ., Heldin C.H.: PDZ domains bind carboxy-terminal sequences of target proteins. Trends Biochem Sci 21:455-458, 1996 Wu T., Angus C.W., Yao X.L., Logun C., Shelhamer J.H.: P11, a unique member of the 8100 family of calcium-binding proteins, interacts with and inhibits the activity of the 85-kDa cytosolic phospholipase A2. J Biol Chem 272:17145- 17153, 1997 41 "‘111111111111111“