The Mlh1-Pms1 endonuclease uses ATP to preserve DNA discontinuities as strand discrimination signals to facilitate mismatch repair

In eukaryotic post-replicative mismatch repair, MutS homologs (MSH) detect mismatches and recruit MLH complexes to nick the newly replicated DNA strand upon activation by the replication processivity clamp, PCNA. This incision enables mismatch removal and DNA repair. Biasing MLH endonuclease activity to the newly replicated DNA strand is crucial for repair. In reconstituted in vitro assays, PCNA is loaded at pre-existing discontinuities and orients the major MLH endonuclease Mlh1-Pms1/MLH1-PMS2 (yeast/human) to nick the discontinuous strand. In vivo, newly replicated DNA transiently contains discontinuities which are critical for efficient mismatch repair. How these discontinuities are preserved as strand discrimination signals during the window of time where mismatch repair occurs is unknown. Here, we demonstrate that yeast Mlh1-Pms1 uses ATP binding to recognize DNA discontinuities. This complex does not efficiently interact with PCNA, which partially suppresses ATPase activity, and prevents dissociation from the discontinuity. These data suggest that in addition to initiating mismatch repair by nicking newly replicated DNA, Mlh1-Pms1 protects strand discrimination signals, aiding in maintaining its own strand discrimination signposts. Our findings also highlight the significance of Mlh1-Pms1’s ATPase activity for inducing DNA dissociation, as mutant proteins deficient in this function become immobilized on DNA post-incision, explaining in vivo phenotypes.


38-mer complement to CMO169 used in discontinuous substrate
Table S1.Oligonucleotides used in this study.Primers CMO175-178 used to generate mutations to abrogate ATPase activity in Mlh1 and Pms1 subunits.Q5 mutagenesis (NEB) was performed using pMH1 and pMH8 as templates (see Materials and Methods) (31).CMO168-CMO170 and CMO184 were used to construct oligonucleotide substrates used to measure Mlh1-Pms1 DNA binding in Figure S2.

Figure S1 .
Figure S1.Plasmid-based substrates used in this study.(A-E) Location of sites for pre-existing nicks on circular plasmids (2.7 kb) used in experiments where circular substrates with breaks are used.Substrates contain between 1 and 5 pre-existing discontinuities introduced using restriction nicking endonucleases as described in the Materials and Methods.Distances and spacing between nicks are indicated on the maps in number of bases (b).(F-H) Location of sites for pre-existing nicks on linearized plasmids (2.7 kb in F, 4.3 kb in G-H) used in experiments where DNA was in linear form to facilitate PCNA self-loading.Distances and spacing between nicks are indicated on the maps in number of kilobases (kb).
Figure S2.Mlh1-Pms1 has modestly higher affinity for model oligonucleotide substrates containing a discontinuity than for intact duplex DNA.(A) Quantification of electrophoretic mobility shift assays performed on intact (continuous) and discontinuous oligonucleotide substrates shown as percent DNA bound as described in the Materials and Methods.All reactions contained 0, 20, 50, 100, and 200 nM of Mlh1-Pms1 in 120 mM NaCl and 0.5 mM ATP when indicated.For all data collected, n = 3.Data were fit to a sigmoidal function describing cooperative ligand binding.(B-E) Representative native TBE gels to demonstrate the gel shift of Mlh1-Pms1 on radiolabeled oligonucleotide substrate mimicking intact DNA or DNA with a pre-existing discontinuity.Where indicated, ATP was added to a final concentration of 0.5 mM.(B) Mlh1-Pms1 affinity for intact (continuous) DNA duplex without ATP.Lanes 1-6 are in increasing order of concentration 0, 20, 50, 100, and 200 nM.(C) Mlh1-Pms1 affinity for intact (continuous) DNA duplexes with 0.5 mM ATP.Lanes 1-6 are in order of decreasing concentrations of 200, 100, 50, 20, and 0 nM Mlh1-Pms1.(D) Mlh1-Pms1 affinity for discontinuous duplexes without ATP.Lanes 1-6 are in order of decreasing concentrations of 200, 100, 50, 20, and 0 nM Mlh1-Pms1.(E) Mlh1-Pms1 affinity for discontinuous DNA duplexes with 0.5 mM ATP.Lanes 1-5 are in increasing concentration order from 0-200 nM Mlh1-Pms1, skipping lane 6, and 300 mM Mlh1-Pms1 in lane 7. Gels were quantified using ImageJ software.Due to unstable binding at near physiological ionic strength, the amount of DNA bound was quantified as loss of DNA in the substrate band relative to negative controls.

Figure S5 .
Figure S5.Optimization of T7 exonuclease activity for Mlh1-Pms1 nick protection assays.(A) T7 exonuclease was titrated in lanes 3-5 (2.0, 4.0, and 8.0 units) on a 2.7 kb DNA with a single pre-existing nick and analyzed by a native agarose gel to identify an amount of T7 that that partially degrades nicked DNA in the absence of factors.Top band is intact double stranded circular DNA while lower band in gel is single stranded circular DNA strand after excision.(B) Quantification of data in A fit to a hyperbolic function.2.0 units of T7 exonuclease was used for protection experiments in Figure 5.
Figure S6.The ATPase sites in Mlh1-Pms1 are conserved and biologically relevant.(A) Clustal Omega Multiple Sequence Alignment of the first 179 amino acids from yeast Mlh1, yeast Pms1, E. coli MutL, human MLH1, and mouse MLH1.Boxed residues are the four ATPase motifs that are conserved among the GyrB, Hsp90, histidine kinase, MutL (GHKL) family (23, 26, 48) of ATPases.Residues within these motifs are color-coded and labeled for chemical properties to further demonstrate conservation.Red arrow indicates the position of the asparagine residue that was mutated in yeast Mlh1 and yeast Pms1 in this study to abolish ATPase activity.Blue triangles indicate a subset of missense mutations that have been identified in Lynch syndrome patients (43, 44) that cluster around ATPase motifs, including the conserved asparagine that is critical for ATPase activity.(B) Predicted structure of the amino-terminal domain of yeast Mlh1.AlphaFold2 was used predict the structure of the amino-terminal domain of yeast Mlh1.This was then aligned to crystal structure of the amino-terminal domain of E. coli MutL complexed with ADPNP (PDB: 1B63) (26) and to the crystal structure of the amino-terminal domain of human MLH1 complexed with ADP