10.1 0 Mismatch Repair The DNA mismatch repair system removes mismatches and short insertions or deletions that are present in DNA. Mismatch repair corrects rare base pair mismatches and short dele- tions or insertions that appear in DNA following replication. DNA polymerases introduce about one mispaired nucleotide per 10° nucleo- tides. The 3°95" proofreading exonuclease increases replication fidelity by 100-fold by removing mispaired nucleotides before the growing chain can be extended further Although an error frequency of 1 nu- cleotide in 10” may seem extremely low, it would result in a high mu- tation rate. The second type of error that arises during replication, short insertions and deletions, result from the fact that repeated-se- quence motifs such as [CA], or [A],, in microsatellites (see Chapter 7) sometimes dissociate and then re-anneal incorrectly, As a result of this slippage, the newly synthesized strand will have a different num- ber of repears than the template strand. Introduction of an insertion or deletion into the newly synthesized DNA is likely to produce a mutation if it affects a region that codes for a protein or an essential RNA molecule. Cells with a nonfunctional mismatch repair system have a high rate of mutation due to their inability to efficiently repair base pair mismatches, short insertions, or short deletions that arise during replication. We begin our examination of mismatch repair by considering the i patch repair system because this system has been the most y studied. Although this system provides valuable lessons for studying mismatch repair in other organisms, it differs from the mismatch repair systems used by gram-positive bacteria and eukary- otes in one important respect. The E. coli mismatch repair system can distinguish a newly synthesized strand from a parental strand because only the latter has methyl groups attached to sites with the sequence GATC. ___E, coli has a deoxyadenosine methylase that transfers methyl groups from S-adenosylmethionine molecules te deoxyadenosines in GATC sequences. However, the timing of methylation by deoxyadenosine methylase lags behind that of nucleotide addition at the replication fork by about two minutes, so the newly synthesized strand is tran- siently unmethylated. The E. coli mismatch repair system exploits this period of transient unmethylation to identify and cut GATC sites in a newly synthesized strand with a mismatch, Genetic and biochemical studies have demonstrated that three E. coli proteins, MutS, MutL, and MutH, are dedicated to mismatch re- pair. Although these proteins are essential for mismatch repair, they are not sufficient, Several additional enzymes and protein factors also make important contributions. Among these enzymes and protein. factors are: DNA helicase II (UvrD), single-stcanded DNA binding protein (SSB), 5’-3" exonucleases (ExoVII or Rec]), 3’-+5’ exonucle- ases (Exol, ExoVIl, or ExoX), DNA polymerase III holoenzyme, DMNA ligase, and deoxyadenosine methylase. The E. colf mismatchrepair system has been reconstituted im vitro from purilied proteins, permitting us to determine how each enzyme and protein factor con- tibutes ro the process. The assay system is based on the recovery of a restriction endonuclease site that is lost because of the mismatch but recovered by the action of the mismatch repair system. The re- sults of the f vitro studies are summarized in FIGURE 10.40. CH, CH, Jw Mut, MutH, oe z 3 s a Mull | Helicase ll, 558, ATP | Helicase Il, SSB, ATP CHs Hy a a o —Ft Exol, ExoWll or Exox. Shy - — Cty P| 3 FIGURE 10.40 E. coli mismatch repair system. The newly synthesized ONA strand (light blue) with a mismatch (orange) is transiently unmethylated at GATC sites. The parent strand (dark blue} has methylated GATC sites The mismatch-activated MutS+MutL+ ATP complex stimulates the MutH endonuclease to incise the nearest unmethylated GATC sequence {either 5' or 3' from the mismatch), Helicase I! (UvrD) unwinds the DNA and SSB binds to the single strands. If the incision is 5° to the mismatch (left), then ExoVl or RecJ endonu- cleases hydrolyze the nicked strand in a §'3' direction. If the incision is 3° to the mismatch (right), then Exol, Exell, or Ecox exonucleases hydrolyze the nicked strand ina 3'5' direction. DNA polymerase Ill holoenzyme fills the gap with new DNA (shown in red), DNA ligase seals the remaining nick and deoxyadenosine methylase adds 4 methyl group to the GATC site (not shown in figure}. Adapted from Lyer, A. R., et al., Chem. Rev. 106 (2008): 302-323.The process begins when MutS as either a homodimer or homo- tetramer binds to the mismatch, For simplicity, Figure 10.40 shows a homedimer bur this is not firmly established. MutS recruits MutL (a homodimer) in an ATP-dependent fashion, Then the MutS*MutL complex activates MutH, which makes an incision at the nearest un- methylated GATC site, either 5° or 3’ to the mismatch, in the newly synthesized strand, Helicase II (UvrD) unwinds the DNA and 555 binds to the resulting single strands. When the incision is 5’ to the mismatch, Exe VII or Rec] exonucleases hydrolyze the nicked strand in a 5°3* direction. When the incision is 3’ to the mismatch, Exol, ExoVII, or EcoX exonucleases hydrolyze the nicked strand ina 3° ta ¥ direction. DNA polymerase II] holoenzyme fills the gap with new DNA (shown in red). DNA ligase seals the remaining nick and de- oxyadenosine methylase adds a methyl group to the GATC site. All organisms that have a mismatch repair system have MutS and MutL. homologs. However, MutH is present only in E. coli and some other gram-negative bacteria, Organisms that lack MutH require some means other than the recognition of the unmethylated GATC site to distinguish between newly synthesized DNA strands and parental strands. The E. coli mismatch repair system suggests a possible mech- anism. MutH is not required if the DNA molecule already has a nick on either the 5’ or 3’ side of the mismatch. Under these conditions, the E. coli system will remove the mismatch from the nicked strand. Titia K. Sixma and coworkers have solved the crystal structure for a C-terminal truncated E. cofi MutS homodimer in complex with a 30 bp DNA oligomer with a G/T mismatch (FIGURE 10.41a). The deletion of the last $3 amino acid residues in the truncated MutS changes the protein from a tetramer to a dimer in solution, perhaps explaining why the truncated MutS does not support mismatch re- pair. The DNA oligomer passes through the larger of two channels formed by the MutS monemers, The function of the other channel, which is large enough to accommodate another double-stranded DNA segment, is not known. The Mur$ dimer also has two ATPase domains, which are located at the far end of the structure with re- spect to the DNA channel. Although the MurS dimer is made of two identical polypeptides, the MutS*DNA oligomer complex is asym- metric. Only one MutS subunit makes contact with the mismatch base pair, The Phe residue ina highly conserved Phe-X-Glu (where X is any amino acid residue) motif of one of the subunits inserts into the minor groove of the DNA helix at the mismatch site, causing a 60° bend in the DNA (FIGURE 10.416). The subunit that makes contact with the mismatch has an ADP bound to its ATPase site. The ATPase site in the other subunit appears to be empty, Although ATP is not re- quired for the MutS to bind DNA, it is required for specificity, MutL exists as. a homodimer in solution. However, its crystal struc= ture has not yet been solved. MutL interacts with MutS but itis not yet known whether this interaction takes place at the mismatch site or af- ter MutS has moved from this site. In fact, there is conflicting evidence about whether MutS actually moves from the mismatch site or remains bound to this site. In the former case, MutS, with or without MutL, could use the energy provided by ATP to slide to the GATC site. In thetay FIGURE 10.41 Structure of a C-terminal truncated E. coli MutS homodimer in complex with a 30 bp DNA oligomer with a G/T mismatch at position 9. fa) The MutS dimer forms a clamp containing two channels. The DNA oligomer (white) passes through the larger channel at the top of the dimer. Only one subunit (green) makes contact with the G/T mismatch. The same subunit also binds ADP (orange spacefill structure) at a site that is distal to the DNA channel. The other subunit (blue) does not have a nucleotide in its ATPase site. (b) The phenylalanine (Phe-36) in the highly conserved Phe-X-Glu motif stacks with a mispaired base after interca into the helix through the minor groove. Thy and Gua indicate thymine and guanine, respectively. Part a Protein Data Bank ID: IE3M. Lamers, M. H., et al., Mature 407 (2000); 711-717. Part b photo courtesy of T. X. Sixma, The Netherlands Cancer Institute, and Niels de Wind, Leiden University Medical Center. latter case, some type of DNA bending would be required to ensure that the MutS*MutL complex bound to the mismatch site was also able to interact with the GATC site. In either case, the Muts* Mul complex would need to activate MutH 50 thar it could nick the un- methylated GATC site. Eukaryotes have proteins that are homologous to Mut$ and Mutl bur lack homologs to MutH, Three human MutS homologs, desig MSH2, MSH3, and MSH6, participate in mismatch repair. MSH2 and MSH6 combine to form a heterodimer called MutSa, and MSH2 and MSH3 combine co form a second heterodimer called MutSB. MutSo ini- tiares mismatch repair at single mismatches and small insertion/del loops, whereas MutSP only initiates mismateh repair at insertion/dele- tion loops of various sizes, The structures of Mutha and MuthB are thought to be similar to the Mut homodimer in bacteria. The MSH6 subunit, but not the MSH2 subunit, in Mute: is the source of the Phe-X- Glu motif that makes contact with the mismatch. Mammalian ho- mologs of the bacterial MutL. protein that participate in misma are designated MLH1 and PMS2 and the heterodimer containing these ovo subunits is called MutLa Paul Modrich and coworkers have recently reconstituted the hu man mismatch repair system in vitro. A strand break on either side of the mismatch is sufficient ro direct repair. Purified human MutSa, nated tion h repairMurLa, Exol (a 5’3° exonuclease), RPA, sliding clamp PCNA, clamp loader RFC, and DNA polymerase 6 are required for bidirec- tional repair. MutSa, Exol, and RPA suffice to excise a mismatch when the nick is on the 5” side of the mismatch but MutLa, RPC, and PCNA also are required when the nick is on the 3’-side of the mis- match. The observation that the mismatch repair system can degrade newly synthesized strands with a nick on the 3’-side of the mismatch was very puzzling because Exol, the only exonuclease added to the system, degrades DNA in a 5’—3' direction. Modrich and coworkers solved the puzzle by demonstrating that MutLa is a latent endonucle- ase that is activated in a mismatch-, MutSa-, RFC-, PCNA-, and ATP-dependent fashion. Once activared, MutL« preferentially makes incisions in the strand that already has a nick, that is, the discontinu- ous strand during replication. The endonuclease activity appears to require an amino acid motif present in the PMS2 but not rhe MLH1 subunit. This motif is also present in MutL homologs from other or- ganisms, including many bacteria. However, the motif is not present in MutL proteins from bacteria such as E. colf with mismatch repair systems that depend on GATC methylation. FIGURE 10.42 presents a F-directed 5'-diractad MutSa, PCNA 5 x 5 Mi “te uihoe Mutha pts a -@ -e oes ee ee Se a ee en F gs r 5 5° to 3" hydrolysis ae ¥ io 3 hydrolysis ate MulSo, MutLe, MutSa, MutLer, Exol, RPA AOR? -@ Exel, RPA, ADE) +@ — —z — — ¥ ——OpeSessssss—: J——QSESSASSOOS— 5 RPA FIGURE 10.42 Incision of the discontinuous heteroduplex strand in human mismateh repair. MutSa, PCNA, and REC activate a latent MutLa endonuclease, which nicks the discontinuous strand of a DNA with a nick on the 5’ or 3 side of the mismatch in an ATP-dependent reaction (red arrows}. For simplicity, protein components are shown only in the top structures. For a nick on the 3 side, this produces a new 5° terminus.on the distal side of tho mismatch that serves as an entry site for MutSce-activated Exol, which removes the mismatch in a5" to 3’ hydrolytic reaction con- trolled by RPA. Adapted from Kadyrov, F.A., Cell 126 (2006); 297-308. 7schematic diagram that shows 3°. and 3’-directed human mismatch repair pathways. MutSa, PCNA, and RFC cooperate to activate the latent MutLe endonuclease, which nicks the discontinuous strand of a DNA strand on both the 5’ and 3’ sides of the mismatch. When the original nick is on the 3*-side of the mismatch, MutL« incisions pro- duces a new 5” terminus on the far side of the mismatch that can serve as an entry site for MutSa-activated Exol, allowing it to remove the mismatch using its §’3" exonuclease activity. RPA stinwlates Exol activity in the presence of MutSa as long as the mismatch is present. Once the mismatch has been removed, RPA inhibits the exonuclease, probably by displacing Exol from the DNA. Individuals with a non-functional mismatch repair system due to a defective MutSe or MutLe suffer from non-polyposis colon cancer (HNPCC), an autosomal recessive syndrome that greatly increases their predis- position to develop intestinal cancer.