Portal de Periódicos da CAPES

The Interaction of the Human MutL Homologues in Hereditary Nonpolyposis Colon Cancer

1999; Elsevier BV; Volume: 274; Issue: 10 Linguagem: Inglês

10.1074/jbc.274.10.6336

1083-351X

Shawn Guerrette, Samir Acharya, Richard Fishel,

Colorectal Cancer Treatments and Studies

Germline mutations in two human mismatch repair (MMR) genes, hMSH2 and hMLH1, appear to account for approximately 70% of the common cancer susceptibility syndrome hereditary nonpolyposis colorectal cancer (HNPCC). Although the hMLH1 protein has been found to copurify with another MMR protein hPMS2 as a heterodimer, their function in MMR is unknown. In this study, we have identified the physical interaction regions of both hMLH1 with hPMS2. We then examined the effects of hMLH1 missense alterations found in HNPCC kindreds for their interaction with hPMS2. Four of these missense alterations (L574P, K616Δ, R659P, and A681T) displayed >95% reduction in binding to hPMS2. Two additional missense alterations (K618A and K618T) displayed a >85% reduction in binding to hPMS2, whereas three missense alterations (S44F, V506A, and E578G) displayed 25–65% reduction in binding to hPMS2. Interestingly, two HNPCC missense alterations (Q542L and L582V) contained within the consensus interaction region displayed no effect on interaction with hPMS2, suggesting that they may affect other functions of hMLH1. These data confirm that functional deficiencies in the interaction of hMLH1 with hPMS2 are associated with HNPCC as well as suggest that other unknown functional alteration of the human MutL homologues may lead to tumorigenesis in HNPCC kindreds. Germline mutations in two human mismatch repair (MMR) genes, hMSH2 and hMLH1, appear to account for approximately 70% of the common cancer susceptibility syndrome hereditary nonpolyposis colorectal cancer (HNPCC). Although the hMLH1 protein has been found to copurify with another MMR protein hPMS2 as a heterodimer, their function in MMR is unknown. In this study, we have identified the physical interaction regions of both hMLH1 with hPMS2. We then examined the effects of hMLH1 missense alterations found in HNPCC kindreds for their interaction with hPMS2. Four of these missense alterations (L574P, K616Δ, R659P, and A681T) displayed >95% reduction in binding to hPMS2. Two additional missense alterations (K618A and K618T) displayed a >85% reduction in binding to hPMS2, whereas three missense alterations (S44F, V506A, and E578G) displayed 25–65% reduction in binding to hPMS2. Interestingly, two HNPCC missense alterations (Q542L and L582V) contained within the consensus interaction region displayed no effect on interaction with hPMS2, suggesting that they may affect other functions of hMLH1. These data confirm that functional deficiencies in the interaction of hMLH1 with hPMS2 are associated with HNPCC as well as suggest that other unknown functional alteration of the human MutL homologues may lead to tumorigenesis in HNPCC kindreds. Mismatch repair is characterized by the recognition and repair of mispaired nucleotides that may arise from misincorporation during DNA replication, recombination between DNA lacking perfect homology, and physical damage to DNA such as deamination of a methylated cytosine (1Friedberg E.C. Walker G.C. Siede W. DNA Repair and Mutagenesis. American Society for Microbiology, Washington, D. C.1995Google Scholar). The best studied MMR 1The abbreviations used are: MMR, mismatch repair; h-, human; HNPCC, hereditary nonpolyposis colorectal cancer; IVTT, in vitro transcription and translation; DAM, DNA adenine methylation; MSH, MutS homologue; MLH, MutL homologue; PMS, post meioticsegregant; k assoc, relative association constant; Intrel, relative interaction; IRm, mutant interaction ratio; IRwt, wild type interaction ratio; GST, glutathione S-transferase; mutH , mutL, mutS, MutHLS; PCR, polymerase chain reaction; PAGE, polyacrylamide gel electrophoresis.1The abbreviations used are: MMR, mismatch repair; h-, human; HNPCC, hereditary nonpolyposis colorectal cancer; IVTT, in vitro transcription and translation; DAM, DNA adenine methylation; MSH, MutS homologue; MLH, MutL homologue; PMS, post meioticsegregant; k assoc, relative association constant; Intrel, relative interaction; IRm, mutant interaction ratio; IRwt, wild type interaction ratio; GST, glutathione S-transferase; mutH , mutL, mutS, MutHLS; PCR, polymerase chain reaction; PAGE, polyacrylamide gel electrophoresis.system is the MutHLS or Dam-instructed pathway in Escherichia coli where mismatch repair is initiated at sites of transient undermethylation of DNA adeninemethylation (Dam) GATC sequences (for review, see Ref. 2Modrich P. Annu. Rev. Genet. 1991; 25: 229-253Crossref PubMed Scopus (770) Google Scholar). Dam-instructed mismatch repair has been shown to require themutH, mutL, mutS, anduvrD(mutU) genes (3Rydberg B. Mutat. Res. 1978; 52: 11-24Crossref PubMed Scopus (77) Google Scholar, 4Glickman B.W. Radman M. Proc. Natl. Acad. Sci. U. S. A. 1980; 77: 1063-1067Crossref PubMed Scopus (270) Google Scholar). The mutL gene was first identified in the enteric bacteria Salmonella typhimuriumLT7 as a mutator (mutator LT7 or MutL) (5Demerec M. Lahr E.L. Miyake T. Galehran E. Balbinder E. Baric S. Hashimoto K. Glanville E.V. Gross J.D. Carnegie Inst. Wash. Year Book. 1957; 370: 390-406Google Scholar, 6Miyake T. Genetics. 1960; 45: 755-762Crossref PubMed Google Scholar) and later assigned as a member of the post-replication mismatch repair machinery (3Rydberg B. Mutat. Res. 1978; 52: 11-24Crossref PubMed Scopus (77) Google Scholar, 4Glickman B.W. Radman M. Proc. Natl. Acad. Sci. U. S. A. 1980; 77: 1063-1067Crossref PubMed Scopus (270) Google Scholar). The MutL protein was purified based on its ability to complement E. coli mutL extracts for mismatch repair (7Grilley M. Welsh K.M. Su S.S. Modrich P. J. Biol. Chem. 1989; 264: 1000-1004Abstract Full Text PDF PubMed Google Scholar). Although a specific biochemical function has not yet been identified for MutL, it does appear to interact with the mismatch recognition protein MutS (7Grilley M. Welsh K.M. Su S.S. Modrich P. J. Biol. Chem. 1989; 264: 1000-1004Abstract Full Text PDF PubMed Google Scholar), leading to subsequent activation of the endonucleolytic activity of MutH (8Welsh K.M. Lu A.L. Clark S. Modrich P. J. Biol. Chem. 1987; 262: 15624-15629Abstract Full Text PDF PubMed Google Scholar), which initiates excision repair by introducing a strand scission on the unmodified strand of hemimethylated GATC sequences (9Lahue R.S. Au K.G. Modrich P. Science. 1989; 245: 160-164Crossref PubMed Scopus (443) Google Scholar). This mechanism of methyl-directed mismatch repair appears to be unique to Gram-negative bacteria, because homologues of MutH have not been found in either Gram-positive bacteria or eukaryotes. The gene coding for MutL has been highly conserved throughout evolution (for review, see Refs. 10Fishel R. Wilson T. Curr. Opin. Genet. Dev. 1997; 7: 105-113Crossref PubMed Scopus (147) Google Scholar and 11Kolodner R. Genes Dev. 1996; 10: 1433-1442Crossref PubMed Scopus (539) Google Scholar). Homologous genes have been found in a number of organisms including the yeast Saccharomyces cerevisiae where there are at least fourMutLhomologues (MLH): MLH1, MLH2, MLH3, and PMS1 (Post MeioticSegregation) (Ref. 11Kolodner R. Genes Dev. 1996; 10: 1433-1442Crossref PubMed Scopus (539) Google Scholar; Stanford yeast sequence data base). Similar homologues of MutL have also been found in humans (12Bronner C.E. Baker S.M. Morrison P.T. Warren G. Smith L.G. Lescoe M.K. Kane M. Earabino C. Lipford J. Lindblom A. Tannergard P. Bollag R.J. Godwin A.R. Ward D.C. Nordenskjeld M. Fishel R. Kolodner R. Liskay R. Nature. 1994; 368: 258-261Crossref PubMed Scopus (1913) Google Scholar, 13Nicolaides N.C. Papadopoulos N. Liu B. Wei Y.F. Carter K.C. Ruben S.M. Rosen C.A. Haseltine W.A. Fleischmann R.D. Fraser C.M. Adams M.D. Venter J.C. Dunlop M.G. Hamilton S.R. Peterson G.M. de la Chapelle A. Vogelstein B. Kinzler K.W. Nature. 1994; 371: 75-80Crossref PubMed Scopus (1429) Google Scholar). Unfortunately the nomenclature has become complicated. Although human hMLH1 is most closely related to yeast MLH1, the human hPMS1 appears most closely related to yeast MLH2 and/or MLH3, and human hPMS2 is most closely related to yeast PMS1. Both genetic and biochemical evidence have suggested that the S. cerevisiae MLH1 and PMS1 form a heterodimer, and both genes have been shown to be essential for MMR (14Prolla T.A. Christie D.-M. Liskay R.M. Mol. Cell. Biol. 1994; 14: 407-415Crossref PubMed Google Scholar, 15Prolla T.A. Pang Q. Alani E. Kolodner R.D. Liskay R.M. Science. 1994; 265: 1091-1093Crossref PubMed Scopus (279) Google Scholar). Copurification of the human hMLH1-hPMS2, based on its ability to complement MMR in extracts derived from a cell line that was genetically deficient for hMLH1, has confirmed the evolutionary conservation of this heterodimer (16Li G.-M. Modrich P. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 1950-1954Crossref PubMed Scopus (349) Google Scholar). The function and interaction of hPMS1 as well as the yeast homologues MLH2 and MLH3 are unknown. Defects in the human MMR pathway have been strongly implicated in the etiology of hereditary nonpolyposis colorectal cancer (HNPCC) (12Bronner C.E. Baker S.M. Morrison P.T. Warren G. Smith L.G. Lescoe M.K. Kane M. Earabino C. Lipford J. Lindblom A. Tannergard P. Bollag R.J. Godwin A.R. Ward D.C. Nordenskjeld M. Fishel R. Kolodner R. Liskay R. Nature. 1994; 368: 258-261Crossref PubMed Scopus (1913) Google Scholar, 13Nicolaides N.C. Papadopoulos N. Liu B. Wei Y.F. Carter K.C. Ruben S.M. Rosen C.A. Haseltine W.A. Fleischmann R.D. Fraser C.M. Adams M.D. Venter J.C. Dunlop M.G. Hamilton S.R. Peterson G.M. de la Chapelle A. Vogelstein B. Kinzler K.W. Nature. 1994; 371: 75-80Crossref PubMed Scopus (1429) Google Scholar,17Fishel R. Lescoe M.K. Rao M.R. Copeland N.G. Jenkins N.A. Garber J. Kane M. Kolodner R. Cell. 1993; 75: 1027-1038Abstract Full Text PDF PubMed Scopus (2578) Google Scholar, 18Akiyama Y. Sato H. Yamada T. Nagasaki H. Tsuchiya A. Abe R. Yuasa Y. Cancer Res. 1997; 57: 3920-3923PubMed Google Scholar, 19Miyaki M. Konishi M. Tanaka K. Kikuchi Yanoshita R. Muraoka M. Yasuno M. Igari T. Koike M. Chiba M. Mori T. Nat. Genet. 1997; 17: 271-272Crossref PubMed Scopus (585) Google Scholar). Germline mutations in hMSH2 (humanMutS homologue) and hMLH1 account for approximately 70% of the HNPCC kindreds (20Peltomaki P. Vasen H.F. Gastroenterology. 1997; 113: 1146-1158Abstract Full Text Full Text PDF PubMed Scopus (676) Google Scholar). Mutations in the other MMR genes hPMS1, hPMS2, and hMSH6 appear rare; however, they have been reported in a few atypical families (13Nicolaides N.C. Papadopoulos N. Liu B. Wei Y.F. Carter K.C. Ruben S.M. Rosen C.A. Haseltine W.A. Fleischmann R.D. Fraser C.M. Adams M.D. Venter J.C. Dunlop M.G. Hamilton S.R. Peterson G.M. de la Chapelle A. Vogelstein B. Kinzler K.W. Nature. 1994; 371: 75-80Crossref PubMed Scopus (1429) Google Scholar, 18Akiyama Y. Sato H. Yamada T. Nagasaki H. Tsuchiya A. Abe R. Yuasa Y. Cancer Res. 1997; 57: 3920-3923PubMed Google Scholar, 19Miyaki M. Konishi M. Tanaka K. Kikuchi Yanoshita R. Muraoka M. Yasuno M. Igari T. Koike M. Chiba M. Mori T. Nat. Genet. 1997; 17: 271-272Crossref PubMed Scopus (585) Google Scholar). Biochemical characterization and structure/function analysis of the MMR proteins has contributed to our understanding of their contribution(s) to hereditary and sporadic carcinogenesis. In S. cerevisiae, an interaction region between MLH1 and PMS1 was localized to the carboxyl termini of both proteins using a two-hybrid in vivoassay system (21Pang Q. Prolla T.A. Liskay R.M. Mol. Cell. Biol. 1997; 17: 4465-4473Crossref PubMed Scopus (115) Google Scholar). In this study, we have localized the biochemical interaction region of hMLH1 and hPMS2 to similar, but not identical, carboxyl-terminal regions of both proteins. We found that the interaction region of hMLH1 lies between amino acids 506 and 675, whereas the interaction region of hPMS2 lies between amino acids 675 and 850. To evaluate the biochemical interactions of hMLH1 mutations found in HNPCC in vitro, we constructed several derivatives of the hMLH1 protein that contained mutations previously reported to cosegregate with cancer susceptibility. We found that the L574P, K616Δ, K618A, K618T, R659P, and A681T missense mutations of hMLH1 displayed a >80% loss of interaction with hPMS2, whereas the S44F, V506A, and E578G missense mutations displayed a 20–80% reduction in interaction with hPMS2. Two missense alterations (Q542L and L582V) showed no reduction in hPMS2 interaction, and one of those (Q542L) was also found to have no effect in the dominant mutator assay (22Shimodaira H. Filosi N. Shibata H. Suzuki T. Radice P. Kanamaru R. Friend S.H. Kolodner R.D. Ishioka C. Nat. Genet. 1998; 19: 384-389Crossref PubMed Scopus (121) Google Scholar). These data indicate that loss or reduced interaction between hMLH1 and hPMS2 may play a causative role in the development of HNPCC. However, other as yet undefined functional alterations also appear to contribute to cancer susceptibility. These data are largely consistent with anin vivo dominant mutator assay developed in yeast that was used to examine mutations of hMLH1 found in HNPCC (22Shimodaira H. Filosi N. Shibata H. Suzuki T. Radice P. Kanamaru R. Friend S.H. Kolodner R.D. Ishioka C. Nat. Genet. 1998; 19: 384-389Crossref PubMed Scopus (121) Google Scholar). Restriction endonucleases and T4 DNA ligase were purchased from New England Biolabs (Beverly, MA). PCR reaction were performed using the High Fidelity PCR kit from Boehringer Mannheim. Oligonucleotides were synthesized on an Applied Biosystems (Foster City, CA) 3948 nucleic acid synthesis and purification system. DNA plasmid constructs were purified using Qiagen (Hilden, Germany) DNA purification kits. In vitro transcription and translation (IVTT) reactions were performed using the Promega (Madison, WI) TNT-coupled rabbit reticulocyte lysate system. Radiolabeled [35S]methionine used to label proteins was obtained from NEN Life Science Products. Glutathione-linked agarose beads were purchased from Sigma. Tris-HCl, NaCl, Tween 20, and dithiothreitol were purchased form Amresco (Solon, OH). All other reagents used were of the highest quality commercially available. The cloning of hMLH1 and hPMS2 has been previously described (12Bronner C.E. Baker S.M. Morrison P.T. Warren G. Smith L.G. Lescoe M.K. Kane M. Earabino C. Lipford J. Lindblom A. Tannergard P. Bollag R.J. Godwin A.R. Ward D.C. Nordenskjeld M. Fishel R. Kolodner R. Liskay R. Nature. 1994; 368: 258-261Crossref PubMed Scopus (1913) Google Scholar). Both hMLH1 and hPMS2 were subcloned into pET-29a using NdeI and BamHI. Glutathione S-transferase (GST) fusion proteins were made using the pGEX system (Amersham Pharmacia Biotech). For ease of subcloning, pGEX-4T-2 was modified as follows. 1) The vector DNA was digested with EcoRI and BamHI and gel purified, and 2) the following linker was introduced by ligation (top strand, 5′-GAT CCG AGA ACC TGT ACT TCC AGG GAC ATA TGG CCA TGG GTA CCG-3′; bottom strand, 5′-AAT TCG GTA CCC ATG GCC ATA TGT CCC TGG AAG TAC AGG TTC TCG-3′); this vector is referred to as pGEX-SG1 and allows for subcloning using NdeI and NcoI restriction endonuclease sites in which the ATG initiation codon is in-frame with the GST moiety. This vector also contains a TEV protease site just upstream of the NdeI and NcoI sites. AnNdeI restriction site (CATATG) was introduced into both hMLH1 and hPMS2 at the initiator ATG by a method that was similar to the introduction of an NcoI site in hMSH2 (23Fishel R. Ewel A. Lescoe M.K. Cancer Res. 1994; 54: 5539-5542PubMed Google Scholar, 24Kallal L. Kurjan J. Mol. Cell. Biol. 1997; 17: 2897-2907Crossref PubMed Google Scholar). pGEX-hMLH1 and pGEX-hPMS2 were constructed by digesting the pGEX-SG1 and the pET29a constructs of hMLH1 and hPMS2 with NdeI andNotI and ligating the genes into the vector. The hMLH1 deletion mutants were constructed using PCR truncation mutagenesis. Forward primers were designed by using a codon that had a guanine in the first position and adding the next 17 nucleotides to the following sequence; 5′-GCG GAT CCC ATG G-3′. The reverse primers were designed by adding the first 18 nucleotides of the complementary strand to the following sequence; 5′-GGC ATA CTC GAG CTA-3′. Using a forward and reverse primer, a PCR reaction was performed using pET-hMLH1 as template. The PCR product and pET-24d were digested withNcoI and XhoI, gel-purified, and ligated together. The hPMS2 deletion mutants were constructed in the same manner. The forward primers were designed by using a codon with a guanine in the first position and adding the next 18 nucleotides to the following sequence: 5′-GGC TAC GGT CTC CAT ATG-3′. Reverse primers were designed by adding the first 18 nucleotides of the complementary strand to the following sequence: 5′-CGC GGA GGA TCC CTA-3′. The PCR product and pET-29a were digested with NdeI and BamHI, gel purified, and ligated together. All of these constructs were sequenced in their entirety (ABI377) to eliminate possible PCR mutagenesis. The HNPCC mutants of hMLH1 were constructed by overlapping PCR using primers containing the required mutation (24Kallal L. Kurjan J. Mol. Cell. Biol. 1997; 17: 2897-2907Crossref PubMed Google Scholar). The PCR products were gel-purified and ligated into bluescript-hMLH1. These constructs were sequenced to verify mutagenesis and then subcloned into pET-29a. The pET constructs were then sequenced in their entirety to eliminate the possibility of second site mutations. An overnight culture of pGEX-hPMS2 or pGEX-hMLH1 was grown in LB medium with 50 μg/ml ampicillin. 50 ml of LB with ampicillin was inoculated with 1 ml overnight culture and grown to an absorbance at 600 nm of 0.5. Isopropyl-1-thio-b-d-galactopyranoside was added to a final concentration of 0.1 mm and placed in a shaker at 30 °C for 2 h. Induced cells were pelleted and resuspended in 800 μl of phosphate-buffered saline (Boehringer Mannheim) plus protease inhibitors (0.5 mm phenylmethylsulfonyl fluoride, 0.8 mg/ml leupeptin, 0.8 mg/ml pepstatin, and 0.1 mm EDTA). Lysozyme was added to a concentration of 1 mg/ml and left on ice for 30 min. Triton X-100 and dithiothreitol were then added to final concentrations of 0.2% and 2 mm, respectively, and the lysate was frozen and thawed two times to completely lyse the cells. DNase I (Boehringer Mannheim) was added to a final concentration of 20 μg/ml, and the lysate was incubated on ice for 20 min. Cell debris was cleared by centrifugation at 14,000 rpm in a refrigerated Eppendorf (Model 5402) centrifuge for 30 min. and the supernatant was transferred to a new microcentrifuge tube with rehydrated GST beads such that approximately 10–50 ng of protein were bound to each 25 μl of beads (see below for quantification of GST fusion protein levels). The lysate/GST beads were incubated at 4 °C on a rocking platform. After rocking at 4 °C for 1–2 h, the lysate/GST beads were spun at 1,000 rpm in an Eppendorf microcentrifuge for 30 s, the supernatant was removed, and the beads were gently resuspended in 500 μl of binding buffer (20 mm Tris, pH 7.5, 10% glycerol, 150 mm NaCl, 5 mm EDTA, 1 mmdithiothreitol, 0.1% Tween 20, 0.75 mg/ml bovine serum albumin, 0.5 mm phenylmethylsulfonyl fluoride, 0.8 mg/ml leupeptin, and 0.8 mg/ml pepstatin). The centrifugation/resuspension was repeated three times to wash the beads free of most nonspecific lysate proteins. The slurry was then added to a 14-ml sterile polypropylene tube, diluted with binding buffer to approximately 50 μl of packed glutathione beads/ml and incubated at 4 °C on a rocking platform for 30 min to allow bovine serum albumin to coat the beads. 500 μl (10–50 ng of bound GST fusion protein) of these coated GST fusion protein-associated glutathione beads was then aliquoted into 1.5-ml Microcentrifuge tubes. GST fusion protein expression levels were determined by binding the lysate to glutathione beads (as above), followed by 3× wash and quantification of protein on Coomassie-stained SDS-PAGE gels using bovine serum albumin as a standard (14Prolla T.A. Christie D.-M. Liskay R.M. Mol. Cell. Biol. 1994; 14: 407-415Crossref PubMed Google Scholar). IVTT (Promega) Reactions with [35S]methionine were performed with pET-hPMS2 or pET-hMLH1 using purified DNA (Qiagen) according to the manufacturer's recommendations. IVTT reactions were prerun to determine the relative molar concentration of each construct. This was calculated using the specific activity of the [35S]methionine, correcting for the number of methionines in each IVTT construct and using SDS-PAGE and a Molecular Dynamics PhosphorImager with ImageQuant software (Sunnyvale, CA) to quantitate labeled peptide. Up to 10 μl of the IVTT protein was added to each tube such that each sample had relative equimolar concentrations of IVTT protein. An IVTT reaction that used pET24d as the vector was added to normalize the total amount of IVTT buffer/reaction mixture in each tube. They were incubated for at least 1 h at 4 °C on a rocker. The beads were washed three times with the binding buffer and then resuspended in 50 μl of SDS loading buffer (0.25 Tris, pH 6.8, 5% sucrose, 2% SDS, 5% 2-mercaptoethanol, and 0.005% bromphenol blue). The samples were resolved on an SDS-PAGE and then imaged using a Molecular Dynamics PhosphorImager. The GST-IVTT interaction assay system is not absolutely quantitative and is likely to depend on the relative association constant (k assoc) of the individual hMLH1 mutant proteins with hPMS2. Thus, subtle changes in the relative concentration of interacting peptides may influence the ultimate measure of interaction. To provide modest control between experiments for such concentration-dependent processes, we determined the approximate molar concentrations of the GST fusion protein and the IVTT protein (see "Materials and Methods" above), and each interaction experiment was designed to contain a nearly identical ratio of the test peptides. Relative interaction (Intrel) with hPMS2 by each missense mutant hMLH1 protein was determined as the fraction of the mutant interaction ratio (IRm) divided by the wild type interaction ratio (IRwt). Intrel = IRm/IRwt. The mutant interaction ratio IRm was determined by quantitating the amount of interacting peptide (Molecular Dynamic PhosphorImager with ImageQuant software) and dividing this number by the quantitated amount of pre-experiment IVTT-expressed protein (Molecular Dynamic PhosphorImager with ImageQuant software). This quantitation was determined (as above) for each experiment which a priori contains all of the mutant proteins to be tested as well as the wild type protein control on a single SDS-PAGE gel. The wild type interaction ratio (IRwt) was calculated similarly by quantifying the amount of wild type IVTT protein precipitated in an interaction experiment and dividing it by the IVTT expression control. In this way, each interaction precipitation was normalized to the amount of protein expressed by IVTT and introduced into the interaction experiment. Results are presented as the mean and S.D. of four separate experiments. These statistical methods would appear make the GST-IVTT assay system semiquantitative in nature. The hMLH1 and hPMS2 proteins were purified as a stable heterodimer based on complementation of MMR in a human hMLH1-deficient cell extract (16Li G.-M. Modrich P. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 1950-1954Crossref PubMed Scopus (349) Google Scholar). To map the interaction region(s) between these proteins, we developed an assay that relies on the use of a GST fusion protein expressed inE. coli as a "bait" and in vitro transcribed and translated (IVTT) protein as "prey." This assay proved to be effective for all of the combinations that would be necessary for this study; GST-hPMS2::IVTT-hMLH1 (Fig.1 A) and GST-hMLH1::IVTT-hPMS2 (Fig. 1 B). The interaction for each of these IVTT full-length peptides was specific for the GST fusion proteins because we observed nearly undetectable nonspecific background binding as demonstrated by incubation and centrifugal precipitation of the IVTT-expressed protein with the glutathione beads alone, E. coli lysate + glutathione beads, and pGEX (the GST moiety alone) + glutathione beads as controls (Fig. 1, 2nd and 4th lanes). Furthermore, densitometric comparison of the pGEX (only)lane with the GST fusion (either hMLH1 or hPMS2)lane demonstrated that the signal to background ratio in this assay approaches 200-fold. These results suggested that this bait-prey system was sufficient to map the interaction region(s) of the hMLH1-hPMS2 heterodimer. Using the GST-IVTT assay, we mapped the interaction domain of hMLH1 with hPMS2 (Fig. 2). In each experiment, a comparison between pairs of signals obtained when IVTT-hMLH1 peptide is added to the GST vector alone versus the GST-hPMS2 fusion is required to determine relevant interaction. Amino acid residues 1–500 of hMLH1 did not appear to significantly interact with GST-hPMS2 (Fig. 2, pairs 2, 4, and 5). However, significant interaction was identified in the carboxyl terminus of the MLH1 protein (amino acids 253–756: Fig. 2, pairs 3 and6). This apparent interaction region was further resolved by truncating this carboxyl-terminal interaction fragment in 25 amino acid increments. We found that amino acid residues 506–756 of hMLH1 interacted with GST-hPMS2, whereas hMLH1 amino acid residues 531–756 showed no detectable interaction (Fig. 2; pairs 6 and7, respectively). Furthermore, hMLH1 amino acid residues 410–675 displayed interaction with GST-hPMS2, whereas hMLH1 amino acid residues 410–650 did not (Fig. 2; pairs 8 and 9, respectively). These results suggest that the most significant hPMS2 interaction region of hMLH1 is located between amino acid residues 506 and 675. It is worth noting that there is a significantly reduced interaction with hPMS2 when one compares hMLH1 amino acid residues 506–756 with hMLH1 amino acid residues 410–675. These results suggest that the interaction region of hMLH1 with hPMS2 may extend the carboxyl terminus from amino acid residue 675. We also determined the interaction domain of hPMS2 with hMLH1 (Fig.3). As suggested above, a comparison between pairs of signals obtained when IVTT-hPMS2 peptide was added to the GST vector alone versus the GST-hMLH1 fusion is required to determine relevant interaction. The first 600-amino acid residues of hPMS2 did not display significant binding to GST-hMLH1 (Fig. 3; pairs2, 4, and 5). In contrast, amino acid residues 602–862 of hPMS2 appeared sufficient to bind GST-hMLH1 (Fig.3; pair 6). Using the truncation strategy described above as a mechanism to resolve the interaction region, we found that amino acid residues 675–862 of hPMS2 bound to GST-hMLH1 (Fig. 3; pair7) and that the interaction was lost with truncation of 25 amino acid residues (amino acids 700–862: Fig. 3; pair 8). Similarly, amino acid residues 675–850 of hPMS2 interacted with GST-hMLH1 (Fig. 3; pair 9), and the interaction was lost with truncation of 25 amino acid residues (amino acids 675–825: Fig.3, pair 10). These results suggest that the hMLH1 interaction region of hPMS2 is located between amino acid residues 675 and 850. Several missense mutations of hMLH1 have been reported to cosegregate with HNPCC and are found to be within the region of hMLH1 that we had determined to interact with hPMS2 (20Peltomaki P. Vasen H.F. Gastroenterology. 1997; 113: 1146-1158Abstract Full Text Full Text PDF PubMed Scopus (676) Google Scholar). As a first step in determining the functional consequences of these alterations, we have examined hMLH1 containing the V506A, Q542L, L574P, E578G, L582V, K616Δ, K618A, K618T, R659P, and A681T missense mutations for their interaction with hPMS2. We also tested the S44F mutation, because alteration of the homologous residue in S. cerevisiae has been reported to effect the interaction of yeast Mlh1p with Pms1p (21Pang Q. Prolla T.A. Liskay R.M. Mol. Cell. Biol. 1997; 17: 4465-4473Crossref PubMed Scopus (115) Google Scholar). In each of four separate experiments, we determined the relative IVTT protein expression (Fig. 4 A) and exposed equivalent molar quantities of IVTT protein to identical quantities of GST-hPMS1 for each interaction pair under study. To rule out selective degradation of mutant proteins, we assessed the stability of the missense mutant proteins in the glutathione agarose-GST-lysate by incubating the IVTT proteins for 5 h under interaction conditions (a normal interaction incubation is 1 h) and then resolving an aliquot by SDS-PAGE. We found no detectable degradation of the mutant peptides during this 5-fold-longer incubation period (data not shown). Interestingly, this lack of degradation could be demonstrated for the full-length protein as well as the numerous internal-start peptides that constitute the bulk of the background lower molecular weight material. Thus, it is unlikely that any reduction in interaction with hPMS2 is because of either altered expression levels or decreased stability of the mutated proteins. Initial experiments suggested that several of the hMLH1 missense mutant proteins might display subtle alterations in binding to hPMS2. To substantiate this comparison, we developed a semiquantitative measurement of hMLH1 interaction with hPMS2 (see "Materials and Methods"). A representative experiment is shown in Fig.4 B. Six of the hMLH1 missense mutant proteins (L574P, K616Δ, K618A, K618T, R659P, and A681T) showed >85% reduction in their interaction with hPMS2. The S44F, E578G, and V506A MLH1 missense mutant proteins displayed 25–65% of that observed with wild type interaction with hPMS2 (26.5 ± 7.1%, 54.9 ± 9.2%, and 65.7 ± 10.9%, respectively). Two hMLH1 missense mutant proteins, Q542L and L582V, appeared to have little effect on interaction with hPMS2 (102.5 ± 2.7% and 101.5 ± 30% of wild type interaction, respectively). It is interesting to note that two different hMLH1 missense mutations at residue 618 (K618A and K618T) had nearly identical effects on their interaction with hPMS2 (11.8 ± 2.5% and 5.3 ± 1.3% of wild type, respectively), thus reducing the likelihood that the altered interactions observed were artificial. Mutations in the human MutL homolog, hMLH1, appear to account for approximately 35% of HNPCC patients (20Peltomaki P. Vasen H.F. Gastroenterology. 1997; 113: 1146-1158Abstract Full Text Full Text PDF PubMed Scopus (676) Google Scholar). The hMLH1 protein was purified as a heterodimer with hPMS2 protein by biochemical complementation of MMR in an hMLH1-deficient cell extract (16Li G.-M. Modrich P. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 1950-1954Crossref PubMed Scopus (349) Google Scholar). Although the role of the hMLH1-hPMS2 heterodimer in MMR has not been elucidated, it is believed to be an integral component of the MMR machinery. These observations suggested that a disruption of the hMLH1-hPMS2 heterodimer might block the process of MMR and manifest itself as HNPCC. To test whether disruption of hMLH1 and hPMS2 interaction could be a cause of cancer susceptibility in HNPCC, we first mapped the interaction region(s) of hMLH1 and hPMS2. We found that this interaction was confined to a single region located in the carboxyl terminus of both proteins. This observation is similar to that reported for the respective homologues (MLH1 and PMS1) in S. cerevisiae. Comparison of the data presented here and similar studies with the yeast homologues suggest that these heterodimers interact through similar regions on each of these proteins. A summary of the hMLH1 and hPMS2 interaction regions is detailed in Fig.5. Interestingly, sequence comparison of the five known MLH1 homologues does not reveal significant conserved homology within the interaction region (15% identity; 38/250 amino acids). However, sequence comparison of the five PMS1 (PMS2) homologues (excluding hPMS1, which is largely nonhomologous) reveals significant homology within the consensus interaction region (35% identity; 61/175 amino acids). We found a strong correlation between HNPCC mutations and a loss of hMLH1 interaction with hPMS2. We tested 11 hMLH1 missense alterations reported in HNPCC kindreds and located in the consensus interaction domain of hMLH1. Nine of these missense alterations displayed reduced interaction with hPMS2. Four of these alterations (L574P, K616D, R659P, and A681T) displayed nearly undetectable interaction with hPMS2 (>99%), whereas two alterations (K618A and K618T) displayed a significant loss of interaction (>85%). Because gross alteration of protein structure can not be eliminated and is likely to be functionally significant for several of these missense mutations, our observations with the K618T and K618A alterations provide significant foundation for the idea that protein structure is not the only contributor to heterodimer protein interaction. This notion is underlined by the observation that the K616Δ alteration (which is located in the first of three lysine residues that includes Lys-618) displays nearly undetectable interaction. Thus, it is appears possible that one or all three of these lysine residues form a charge-based contact between hMLH1 and hPMS2. Although the S44F alteration of hMLH1 was found to be outside of the consensus hPMS2 interaction region, we tested this missense alteration in our system because the homologous alteration in S. cerevisiae displayed reduced interaction with PMS1 (21Pang Q. Prolla T.A. Liskay R.M. Mol. Cell. Biol. 1997; 17: 4465-4473Crossref PubMed Scopus (115) Google Scholar). We found a reproducible reduction in the ability of S44F hMLH1 to interact with hPMS2 (26.5% of wild type). We consider three possibilities. 1) The GST-IVTT assay for mapping these regions may have failed to detect a possible interaction domain at the amino terminus of hMLH1, where the fusion is located, 2) the S44 amino acid may be in close spatial proximity to the consensus hPMS2 interaction region, or 3) substitution of Ser-44 with a phenylalanine gravely disrupts the protein structure such that the downstream interaction region is largely affected. The V506A and E578G hMLH1 missense alterations, both, appeared to display consistent reduced binding to hPMS2. Although it is still possible that these alterations do not alter interaction with hPMS2in vivo, it is interesting to note that a functional defect has been confirmed for these alterations using a yeast dominant mutator assay that would appear to depend on similar protein-protein interactions (22Shimodaira H. Filosi N. Shibata H. Suzuki T. Radice P. Kanamaru R. Friend S.H. Kolodner R.D. Ishioka C. Nat. Genet. 1998; 19: 384-389Crossref PubMed Scopus (121) Google Scholar). Two hMLH1 missense alterations, Q542L and L582V, did not appear to affect interaction with hPMS2. These alterations have been reportedly identified in patients meeting the Amsterdam criteria for HNPCC (25Vasen H.F. Mecklin J.P. Khan P.M. Lynch H.T. Dis. Colon Rectum. 1991; 34: 424-425Crossref PubMed Scopus (1726) Google Scholar,26Han H.J. Maruyama M. Baba S. Park J.G. Nakamura Y. Hum. Mol. Genet. 1995; 4: 237-242Crossref PubMed Scopus (167) Google Scholar). We can not rule out the possibility that an altered interaction by these amino acid changes is obscured by the amino-terminal GST fusion of these constructs. However, more provocative possibilities are that the Q542L and L582V may be either nonfunctional polymorphisms or that these alterations affect a function of the hMLH1 protein other than interaction with hPMS2. Interestingly, the Q542L missense alteration has no effect in an independent yeast dominant mutator assay system (22Shimodaira H. Filosi N. Shibata H. Suzuki T. Radice P. Kanamaru R. Friend S.H. Kolodner R.D. Ishioka C. Nat. Genet. 1998; 19: 384-389Crossref PubMed Scopus (121) Google Scholar). Because these alterations are reportedly from well defined HNPCC kindreds, these results appear to support the notion that there are other functional alterations of the human MutL homologues that may contribute to cancer susceptibility. The 11 hMLH1 proteins containing missense alterations appear to establish a strong link between HNPCC and loss of interaction with hPMS2. It is interesting to note that there have been no reports of single-amino acid substitutions in hPMS2, which are located in the hMLH1 interaction domain. This observation may suggest that there is another molecular partner for hMLH1 that has a redundant function with hPMS2 (similar to the findings with hMSH3 and hMSH6). hPMS1 could be a candidate for this role. We have previously identified the interaction domains of hMSH2 with hMSH3 and hMSH2 with hMSH6 (27Guerrette S. Wilson T. Gradia S. Fishel R. Mol. Cell. Biol. 1998; 18: 6616-6623Crossref PubMed Scopus (120) Google Scholar). However, we found no effect of missense alterations found in HNPCC kindreds on the interaction with either hMSH3 and hMSH6. This observation suggested that alteration of a static interaction between hMSH2 with hMSH3 or hMSH6 was unlikely to play a causative role in HNPCC. Here we confirm the correlation between altered interaction between hMLH1 and hPMS2 and susceptibility to HNPCC. We thank Hansjuerg Alder and the employees of the Sidney Kimmel Nucleic Acid Facility for nucleotide synthesis and sequencing, Christoph Schmutte for helping to prepare the figures for this manuscript, and Teresa Wilson, Scott Gradia, and Greg Tombline for helpful discussions.