Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
Zurich Open Repository andArchiveUniversity of ZurichMain LibraryStrickhofstrasse 39CH-8057 Zurichwww.zora.uzh.ch
Year: 2020
Regulation of the MLH1-MLH3 endonuclease in meiosis
Cannavo, Elda ; Sanchez, Aurore ; Anand, Roopesh ; Ranjha, Lepakshi ; Hugener, Jannik ; Adam,Céline ; Acharya, Ananya ; Weyland, Nicolas ; Aran-Guiu, Xavier ; Charbonnier, Jean-Baptiste ;
Hoffmann, Eva R ; Borde, Valérie ; Matos, Joao ; Cejka, Petr
Abstract: During prophase of the first meiotic division, cells deliberately break their DNA1. These DNAbreaks are repaired by homologous recombination, which facilitates proper chromosome segregation andenables the reciprocal exchange of DNA segments between homologous chromosomes2. A pathway thatdepends on the MLH1–MLH3 (MutL) nuclease has been implicated in the biased processing of meioticrecombination intermediates into crossovers by an unknown mechanism3,4,5,6,7. Here we have biochem-ically reconstituted key elements of this pro-crossover pathway. We show that human MSH4–MSH5(MutS), which supports crossing over8, binds branched recombination intermediates and associates withMutL, stabilizing the ensemble at joint molecule structures and adjacent double-stranded DNA. MutSdirectly stimulates DNA cleavage by the MutL endonuclease. MutL activity is further stimulated byEXO1, but only when MutS is present. Replication factor C (RFC) and the proliferating cell nuclearantigen (PCNA) are additional components of the nuclease ensemble, thereby triggering crossing-over.Saccharomyces cerevisiae strains in which MutL cannot interact with PCNA present defects in formingcrossovers. Finally, the MutL–MutS–EXO1–RFC–PCNA nuclease ensemble preferentially cleaves DNAwith Holliday junctions, but shows no canonical resolvase activity. Instead, it probably processes mei-otic recombination intermediates by nicking double-stranded DNA adjacent to the junction points9. AsDNA nicking by MutL depends on its co-factors, the asymmetric distribution of MutS and RFC–PCNAon meiotic recombination intermediates may drive biased DNA cleavage. This mode of MutL nucleaseactivation might explain crossover-specific processing of Holliday junctions or their precursors in meioticchromosomes4.
DOI: https://doi.org/10.1038/s41586-020-2592-2
Posted at the Zurich Open Repository and Archive, University of ZurichZORA URL: https://doi.org/10.5167/uzh-199557Journal ArticlePublished Version
Originally published at:Cannavo, Elda; Sanchez, Aurore; Anand, Roopesh; Ranjha, Lepakshi; Hugener, Jannik; Adam, Céline;Acharya, Ananya; Weyland, Nicolas; Aran-Guiu, Xavier; Charbonnier, Jean-Baptiste; Hoffmann, Eva R;Borde, Valérie; Matos, Joao; Cejka, Petr (2020). Regulation of the MLH1-MLH3 endonuclease in meiosis.Nature, 586(7830):618-622.DOI: https://doi.org/10.1038/s41586-020-2592-2
1
RegulationoftheMLH1-MLH3endonucleaseinmeiosis1
2
EldaCannavo1#,AuroreSanchez1#,RoopeshAnand1#,LepakshiRanjha1,Jan-3
nikHugener2,CélineAdam3,4,AnanyaAcharya1,2,NicolasWeyland5,Xavier4
Aran-Guiu6, Jean-Baptiste Charbonnier7,8, Eva R. Hoffmann6,9, Valérie5
Borde3,4,JoaoMatos2andPetrCejka1,26
7
Affiliations:8
1Institute forResearch inBiomedicine,Universitàdella Svizzera italiana (USI),9
FacultyofBiomedicalSciences,Switzerland10
2Department of Biology, Institute of Biochemistry, Eidgenössische Technische11
Hochschule(ETH),Zürich,Switzerland12
3InstitutCurie,PSLResearchUniversity,CNRSUMR3244,Paris,France13
4ParisSorbonneUniversité,Paris,France14
5InstituteofMolecularCancerResearch,UniversityofZürich,Zürich,Switzerland15
6GenomeDamageandStabilityCentre,SchoolofLifeSciences,UniversityofSus-16
sex,Brighton,UK17
7I2BC, iBiTec-S, CEA, CNRS UMR 9198, Université Paris-Sud, Gif-sur-Yvette,18
France;19
8UniversitéParisSud,Orsay,France20
9DNRFCenterforChromosomeStability,DepartmentofCellularandMolecular21
Medicine,FacultyofHealthandMedicalSciences,UniversityofCopenhagen,Co-22
penhagen,Denmark23
24
#Theseauthorscontributedequally.25
26
Materials&Correspondence:PetrCejka, Institute forResearch inBiomedicine,27
ViaVincenzoVela6,6500Bellinzona,Switzerland;28
E-mail:[email protected]
30
2
SUMMARY31
32
Duringprophaseofthefirstmeioticdivision,cellsdeliberatelybreaktheirDNA1.33
TheseDNAbreaksarerepairedbyhomologousrecombination,whichfacilitates34
proper chromosome segregation andenables reciprocal exchangeofDNA seg-35
mentsbetweenhomologouschromosomes,promotingthusgeneticdiversityin36
theprogeny2.Geneticandcellulardataimplicatedapathwaydependentonthe37
MLH1-MLH3(MutLγ)nucleaseinthebiasedprocessingofmeioticrecombination38
intermediatesintocrossovers,butmechanismsthatleaditsactivationwereun-39
clear3-7.Here,webiochemicallyreconstitutedkeyelementsofthispro-crossover40
pathway.First,weshowthathumanMSH4-MSH5(MutSγ),whichwasknownto41
supportcrossingover8,bindsbranchedrecombinationintermediatesandphysi-42
callyassociateswithMutLγ.Thishelpsstabilizetheensembleatjointmolecule43
structuresandadjacentdsDNA.Second,weshowthatMutSγdirectlystimulates44
DNAcleavagebytheMutLγendonuclease,whichdemonstratesanunexpecteddi-45
rectfunctionforMutSγintriggeringcrossing-over.Third,wefindthatMutLγac-46
tivityisfurtherstimulatedbyEXO1,butonlywhenMutSγispresent.Fourth,we47
alsoidentifythereplicationfactorC(RFC)andtheproliferatingcellnuclearanti-48
gen(PCNA)asadditionalcomponentsofthenucleaseensemble,andshowthatS.49
cerevisiae strains expressing PCNA-interacting peptide (PIP) box-likemutated50
MutLγpresentstrikingdefectsinformingcrossovers.Finally,weshowthatthe51
MutLγ-MutSγ-EXO1-RFC-PCNA nuclease ensemble preferentially cleaves DNA52
withHolliday junctions,butshowsnocanonicalresolvaseactivity. Instead, the53
multilayerednucleaseensemblelikelyprocessesmeioticrecombinationinterme-54
diatesbynickingdsDNAadjacenttothejunctionpoints9.SinceDNAnickingby55
MutLγisdependentonitsco-factors,theasymmetricdistributionofMutSγand56
RFC-PCNAonmeioticrecombinationintermediatesmaydrivebiasedDNAcleav-57
age.ThisuniquemodeofMutLγnucleaseactivationmightexplaincrossover-spe-58
cificprocessingofHollidayjunctionswithinthemeioticchromosomalcontext4.59
60
61
3
MAINTEXT62
63
MutLγisanATP-stimulatedendonuclease64
TostudyhumanMutLγ(MLH1-MLH3),weexpressedandpurifiedtheheterodi-65
merfrominsectcells(ExtendedDataFig.1a,b).Similarlytothemismatchrepair66
(MMR)-specific MutLα (MLH1-PMS2)10 and yeast Mlh1-Mlh311,12, the human67
MLH1-MLH3complexnon-specificallynickeddouble-strandedsupercoiledDNA68
(scDNA)withmanganeseasametalco-factor(Fig.1a,ExtendedDataFig.1c-e)13.69
MutationsintheconservedmetalbindingDQHA(X)2E(X)4EmotifofMLH3abol-70
ishedtheendonuclease,indicatingthattheDNAcleavageactivitywasintrinsicto71
theMutLγheterodimer(Fig.1a).ATPpromotedthenucleaseactivity>2-fold(Fig.72
1a,ExtendedDataFig.1f).Almostnonucleaseactivitywasobservedwithmagne-73
sium(ExtendedDataFig.1g),whichisbelievedtobethespecificmetalco-factor10.74
ExperimentswithvariousATPanalogsrevealedthatATPhydrolysisbyMLH1-75
MLH3wasrequiredforthemaximalstimulationofDNAcleavage(ExtendedData76
Fig. 1h,i). TheN-termini of bothMLH1 andMLH3 proteins contain conserved77
WalkermotifsimplicatedinATPbindingandhydrolysis14.Todefinewhetherthe78
ATPaseofMLH1,MLH3orbothsubunitsoftheheterodimerpromotesitsnucleo-79
lyticactivity,wepreparedtherespectivehMutLγvariantswithmutationsinthe80
ATPasemotifs of either subunit individually or combined (ExtendedData Fig.81
1j,k)14.WeobservedthattheintegrityoftheATPasedomainofMLH1,andtoa82
muchlesserdegreeofMLH3,promotedthenucleaseactivityofMLH1-MLH3(Ex-83
tended Data Fig. 1l,m). Without ATP, the ATPase-deficient variants of MLH1-84
MLH3boundandcleavedDNAsimilarlyasthewildtypecomplex(ExtendedData85
Fig.1l-n).TheMutLγcomplexdidnotcleaveoligonucleotide-basedHollidayjunc-86
tions(HJ)DNA(ExtendedDataFig.1o).87
88
MutLγandMutSγbindtoDNAjunctions89
Yeast and human MutLγ complexes bind DNA with a preference towards90
branched structures suchasHolliday junctions11,15. Similarly, recombinanthu-91
manMutSγwasshowntobindHJs8.WefoundthatthehumanandyeastMutSγ92
complexesboundevenbetterHJprecursorssuchasD-loops(ExtendedDataFig.93
4
2a-f).ThisbindingpreferenceagreeswiththeproposedearlyfunctionofMutSγ94
tostabilizenascentstrandinvasionintermediatesthatmatureintosingle-endin-95
vasions, which helps ensure their crossover designation16. In contrast, single-96
strandedDNA(ssDNA)ordsDNAwasnotboundbyMutSγ(ExtendedDataFig.97
2b-f). Electrophoretic mobility shift assays demonstrated that the MutSγ and98
MutLγ complexesmoderately stabilized each other at the DNA junctions (Ex-99
tendedData Fig. 3a-e). Accordingly, the respectivehumanor yeastMutSγ and100
MutLγcomplexesdirectlyphysicallyinteract(Fig.1bandExtendedDataFig.3f-101
h)17. The very slowmigrationof theprotein-DNA complexeswas indicativeof102
multiple units of the heterocomplexes bound to the DNA substrate (Extended103
DataFig.3i,j),asshownpreviouslyforyeastMutLγ11.Wenotethatthepresence104
ofDNAjunctionswasessentialforstableDNAbinding(ExtendedDataFig.3a,e),105
whichsupportsamodelwherethebranchedDNAstructureservesasanucleation106
pointforaMutSγ-MutLγfilamentthatthenextendstotheadjacentdsDNAarms18.107
108
MutSγpromotestheMutLγnuclease109
PreviousinvivoexperimentsimplicatedMutSγinthestabilizationofnascentDNA110
jointmolecules early in themeiotic pro-crossover pathway8,17,19, but whether111
MutSγisdirectlyinvolvedlaterinnucleolyticprocessingwasnotclear.Usingour112
reconstitutedsystem,weobserved~3-foldstimulationoftheMLH1-MLH3endo-113
nucleasebyMSH4-MSH5(Fig.1c,ExtendedDataFig.4a-c),whichwasdependent114
ontheMLH3metalbindingmotif(ExtendedDataFig.4b).ATPpromotedDNA115
cleavagebytheMutSγ-MutLγensemble,andasinreactionswithMutLγalone,the116
maximalnucleaseactivitywasobservedwhenATPhydrolysiswaspossible(Ex-117
tendedDataFig.4d).AsMLH1andMLH3,alsoMSH4andMSH5proteinscontain118
conservedATPasedomains14,16.TheATPbinding/hydrolysismotifsinMLH1and119
MSH5werebothcrucial,whilethemotifinMSH4waslessimportantandinMLH3120
appeareddispensable(Fig.1d,e,ExtendedDataFig.4e,f).TheATPasemotifmu-121
tationsinMSH4orMSH5insteaddidnotaffectthecapacityofthetwosubunitsto122
formacomplexorbindDNA(ExtendedDataFig.4g,h).Thestimulatoryeffectwas123
likely facilitated by direct physical interactions between the cognate heterodi-124
mers,asyeastMsh4-Msh5didnotpromotethenucleaseactivityofhumanMutLγ125
5
(ExtendedDataFig.4i,j).MutSγ-MutLγexhibitednodetectablestructure-specific126
nucleaseorresolvaseactivity(ExtendedDataFig.4k,l).127
To assess how the other human MutS homologue complexes, compared to128
MutSγ,promotetheactivityofMutLγ,wesupplementedtheMLH1-MLH3nucle-129
ase reactions with the MMR factors MutSa (MSH2-MSH6) or MutSb (MSH2-130
MSH3).MutSb,butnothMutSa, couldstimulatetheMLH1-MLH3nucleasetoa131
similarlevelasMutSγ(ExtendedDataFig.5a-c)13.Thisagreeswithpreviousex-132
perimentsshowingthatyeastMutLγcouldpartiallysubstituteMutLαinthere-133
pairofinsertion/deletionmismatchesinMMR20.Thesedataalsounderpinthein-134
volvementofMutLγinthemetabolismoftrinucleotiderepeatslinkedtoseveral135
neurodegenerativediseases,aswellas rareMLH3mutations found inpatients136
withhereditarynonpolyposiscolorectalcancer(HNPCC)/Lynchsyndromechar-137
acterizedbymicrosatelliteinstability13,21,22.138
139
EXO1promotestheMutSγ-MutLγnuclease140
Genetic experimentswith budding yeast revealed a structural (nuclease-inde-141
pendent)functionofExo1intheMlh1-Mlh3pro-crossoverpathway5.Theeffect142
wasdependentonitsdirectinteractionwiththeMlh1subunitofMutLγ5,23,butit143
wasunclearwhether the interplaydirectlyaffects theMlh3endonuclease,and144
whetherthisfunctionisconservedinhighereukaryotes.Totestfortheeffectof145
EXO1 on the nuclease activity ofMLH1-MLH3,we used the nuclease-deficient146
EXO1(DA)varianttopreventdegradationoftheresultingnickedDNA(Extended147
Data Fig. 6a). We observed no stimulation of the MLH1-MLH3 nuclease by148
EXO1(DA)alone,butEXO1(DA)promotedDNAcleavage~2-3-foldwhenMSH4-149
MSH5waspresent(Fig.1fandExtendedDataFig.6b,c).Morethan40%DNAwas150
cleavedusingonly20nMconcentrationofthemulti-proteinensemble(Fig.1f).151
IncontrasttoMSH4-MSH5thatmoderatelystabilizedMLH1-MLH3onDNA,we152
detectednosuchcapacityofEXO1(DA)(ExtendedDataFig.6d).YeastExo1(DA)153
couldnotsubstitutehumanEXO1(DA)inthenucleaseassays(ExtendedDataFig.154
6e), inagreementwithadirectphysical interactionbetweenhumanEXO1(DA)155
andMLH1-MLH3(Fig.1g).Finally,EXO1(DA)didnotpromotethenucleaseactiv-156
ityofMLH1-MLH3withMutSb(ExtendedDataFig.6f),indicatingthatEXO1likely157
6
specificallystimulatestheendonucleaseactivityofMutSγ-MutLγinvolvedinmei-158
oticrecombination.159
160
PCNApromotestheMLH3nucleaseensemble161
WenextsetouttotestwhetherMLH1-MLH3withitsco-factorscancatalyzeDNA162
cleavageunderphysiologicalconditionsinmagnesium.Whilealmostnonuclease163
activityofMLH1-MLH3aloneinmagnesiumwasobserved,weaknickingwasseen164
upon adding MSH4-MSH5, and the reactions were further stimulated by165
EXO1(DA)(Fig.2a).AsRFC-PCNAareknowntodirecttheMutLaendonucleasein166
mismatchrepair10,wetestedfortheireffectonMutLγ.Notably,weobservedad-167
ditional~2-foldstimulationofDNAcleavagewhentheRFC-PCNAcomplexwas168
included(Fig.2a,ExtendedDataFig.7a-e).Nucleolyticcleavagewasdependent169
ontheintegrityoftheMLH3metal-bindingmotif(ExtendedDataFig.7b).Yeast170
RFCcouldpartiallysubstituteforhumanRFC(ExtendedDataFig.7c),inaccord171
withthecapacityofyeastRFCtoloadhumanPCNAonscDNA10,24.Incontrast,no172
stimulationwasdetectedwhenRFCwasomitted fromthereactionmixtureor173
whenusing relaxedDNA (ExtendedData Fig. 7b,f), indicating that PCNAmust174
likelybeactively loadedonto scDNAbyRFC10,24.Nostimulationwasobserved175
whenusingyeastinsteadofhumanPCNA(ExtendedDataFig.7b),orinreactions176
withmanganese(ExtendedDataFig.7g).Thus,whileMutLγonitsownisapoor177
nuclease that requires manganese, MutSγ, EXO1 and RFC-PCNA activate it to178
cleaveefficientlyinabuffercontainingphysiologicalmagnesium,andthereaction179
isnolongerstimulatedbyadditionalmanganese(Fig.2b).180
RFC-PCNAcouldalsopromotethenucleaseactivityofMutLγalone,although181
toalesserextent(ExtendedDataFig.7h),suggestingthatMutSγandEXO1(DA)182
arenotstrictlyrequiredtomediatethestimulatoryeffectofRFC-PCNA.Inaccord,183
wefoundthatMutLγdirectlyphysicallyinteractswithPCNA(Fig.2c).ATPwas184
necessary for thenucleaseactivityof theensembleand forPCNA loading, and185
couldnotbereplacedbyADPorAMP-PNP,showingthatATPhydrolysiswasre-186
quired(Fig.2d,ExtendedDataFig.7i).Incontrasttothereactionsinmanganese,187
the integrity of the ATPasemotifs of all four MutSγ andMutLγ subunits was188
7
requiredformaximalcleavageactivityinmagnesium(Fig.2e,ExtendedDataFig.189
7j).190
Notably, the nuclease ensemble preferentially cleaved plasmid-length DNA191
withpalindromicrepeatsformingaHJ-likestructure(cruciformDNA,Fig.2f),in192
agreementwiththebindingpreferenceoftheMutSγandMutLγheterodimersto193
theserecombinationintermediates(ExtendedDataFig.2b-f)8,11,15.However,the194
activityofthecomplexonthecruciformDNAprimarilyyieldednickedproducts195
(Fig.2f),unlike canonicalHJ resolvases thatgive rise to linearDNAuponcon-196
certedcleavageofbothDNAstrandsatthejunctions25.Wealsonotethatwedid197
notobserveanycleavageofmodelHJorD-loopoligonucleotide-basedsubstrates198
(ExtendedDataFig.7k).OurdatasuggestthatthehMutLγensembleprocesses199
recombination intermediates by resolution-independent nicking, in agreement200
withresultsobtainedfromsequencingofheteroduplexDNAarisingduringmeio-201
sisinyeastcells,whichindicatedcleavagebynickingsomedistanceawayfrom202
theDNAjunctionpoints9.203
Interactionswith PCNA are oftenmediated by a PIP-boxmotif.We supple-204
mentedtheMutLγensemblenucleaseassayswithaPIP-boxpeptidederivedfrom205
p2126,oracontrolpeptidewithkeyresiduesmutated.ThecompetingPIP-box206
peptideeliminatedthestimulatoryeffectofRFC-PCNA,whilethecontrolpeptide207
hadnoeffect(Fig.3a),demonstratingthatthePCNAfunctioninstimulatingthe208
nucleaseactivityoftheMutLγensembleisdependentonaninteractionviaaPIP-209
boxlikemotif.WenextanalyzedseveralmutantsofconservedPIP-box-likese-210
quencesinMLH1,MLH3andEXO1(DA)27-29(Fig.3b).Therespectivemutations211
didnotnotablyaffectthenucleasereactionspersewithoutPCNAorthecapacity212
tobindDNA,but themutantsbecamepartly refractory to stimulationbyRFC-213
PCNA, in particularwhen the PIP-box-likemutations ofmultiple factorswere214
combined(Fig.3c,ExtendedDataFig.8a-e).Furthermore,thecorrespondingmu-215
tationsintheyeasthomologuesofMlh1andMlh3(Fig.3b)resultedinmeiotic216
defects,as indicatedbyadecrease inthe frequencyofcrossoversat theCEN8-217
THR1intervalleadingtochromosomenon-disjunctionandreducedsporeviabil-218
ity(Fig.3d,e,ExtendedDataFig.9a).WenotedthatthestabilityofMlh1Pwasre-219
ducedinvivo,whichmaycontributetotheveryseverephenotype,butthelevels220
ofMlh3Pwerecomparabletowildtype(ExtendedDataFig.9b,c).Wealsofound221
8
yeastMlh1andMlh3asapartofacomplexwithRfc1inmeioticcellsatthetime222
ofjointmoleculeresolution(ExtendedDataFig.9d).Finally,usingchromatinim-223
munoprecipitationandsynchronousmeioticyeastcultures,weobservedanen-224
richmentofRfc1atbothnaturalandengineeredDSBhotspotsinlatemeioticpro-225
phaseatthetimewhenjointmoleculesareresolvedintocrossovers(Fig.3f).This226
coincideswiththeaccumulationofMlh3atthesamehotspots30.Theaccumula-227
tionofRfc1atsitesofrecombinationwasindependentofMlh3,suggestingthatit228
mayberetainedfromanearlierstepofDNAsynthesis(ExtendedDataFig.9e).229
WedemonstratedherethatMutSγ,EXO1andRFC-PCNAarerequiredtoacti-230
vatethenucleaseactivityofMutLγunderphysiologicalconditionswithmagne-231
sium,butwefailedtodetectanycanonicalHJresolvaseactivity.Rather,ourdata232
suggestthatthenucleaseensembleprocessesmeioticjointmoleculeintermedi-233
ates by biased resolution-independent nicking of dsDNA in the vicinity ofHJs.234
SinceHJsaresymmetricandtheirresolutioncanyieldbothcrossoversandnon-235
crossovers25, how is the crossover bias established?AsMutSγ likely stabilizes236
asymmetricHJprecursors, itmaybeasymmetricallydistributedat themature237
jointmoleculescontainingHJs(ExtendedDataFig.10).Similarly,PCNAislikely238
loadedasymmetricallyatjointmoleculestofacilitateDNAsynthesisbypolymer-239
ased,oratstranddiscontinuitiesbeforetheligationofdoubleHJstakesplace.We240
propose that the asymmetricpresenceof theMutLγnuclease co-factorsmight241
providethesignaltoguaranteebiased,crossover-specificprocessingofmeiotic242
jointmoleculeintermediates(ExtendedDataFig.10).InthismodeltheMMRre-243
actionthatlimitsmutagenesisduringDNAreplication11isrepurposedinmeiotic244
recombinationto insteadpromotediversityamongtherecombinedDNAmole-245
cules.246
247
248
REFERENCES249
250
1 Keeney, S., Giroux, C. N. & Kleckner, N. Meiosis-specific DNA double-251
strand breaks are catalyzed by Spo11, a member of a widely conserved 252
protein family. Cell 88, 375-384 (1997). 253
2 Hunter, N. Meiotic Recombination: The Essence of Heredity. Cold Spring 254
Harb Perspect Biol 7, doi:10.1101/cshperspect.a016618 (2015). 255
9
3 Nishant, K. T., Plys, A. J. & Alani, E. A mutation in the putative MLH3 256
endonuclease domain confers a defect in both mismatch repair and meiosis in 257
Saccharomyces cerevisiae. Genetics 179, 747-755, 258
doi:10.1534/genetics.108.086645 (2008). 259
4 Zakharyevich, K., Tang, S., Ma, Y. & Hunter, N. Delineation of joint 260
molecule resolution pathways in meiosis identifies a crossover-specific 261
resolvase. Cell 149, 334-347, doi:10.1016/j.cell.2012.03.023 (2012). 262
5 Zakharyevich, K. et al. Temporally and biochemically distinct activities of 263
Exo1 during meiosis: double-strand break resection and resolution of double 264
Holliday junctions. Mol Cell 40, 1001-1015, 265
doi:10.1016/j.molcel.2010.11.032 (2010). 266
6 De Muyt, A. et al. BLM helicase ortholog Sgs1 is a central regulator of 267
meiotic recombination intermediate metabolism. Mol Cell 46, 43-53, 268
doi:10.1016/j.molcel.2012.02.020 (2012). 269
7 Svetlanov, A., Baudat, F., Cohen, P. E. & de Massy, B. Distinct functions of 270
MLH3 at recombination hot spots in the mouse. Genetics 178, 1937-1945, 271
doi:10.1534/genetics.107.084798 (2008). 272
8 Snowden, T., Acharya, S., Butz, C., Berardini, M. & Fishel, R. hMSH4-273
hMSH5 recognizes Holliday Junctions and forms a meiosis-specific sliding 274
clamp that embraces homologous chromosomes. Mol Cell 15, 437-451, 275
doi:10.1016/j.molcel.2004.06.040 (2004). 276
9 Marsolier-Kergoat, M. C., Khan, M. M., Schott, J., Zhu, X. & Llorente, B. 277
Mechanistic View and Genetic Control of DNA Recombination during 278
Meiosis. Mol Cell 70, 9-20 e26, doi:10.1016/j.molcel.2018.02.032 (2018). 279
10 Kadyrov, F. A., Dzantiev, L., Constantin, N. & Modrich, P. Endonucleolytic 280
function of MutLalpha in human mismatch repair. Cell 126, 297-308, 281
doi:10.1016/j.cell.2006.05.039 (2006). 282
11 Ranjha, L., Anand, R. & Cejka, P. The Saccharomyces cerevisiae Mlh1-Mlh3 283
heterodimer is an endonuclease that preferentially binds to Holliday junctions. 284
J Biol Chem 289, 5674-5686, doi:10.1074/jbc.M113.533810 (2014). 285
12 Rogacheva, M. V. et al. Mlh1-Mlh3, a meiotic crossover and DNA mismatch 286
repair factor, is a Msh2-Msh3-stimulated endonuclease. J Biol Chem 289, 287
5664-5673, doi:10.1074/jbc.M113.534644 (2014). 288
13 Kadyrova, L. Y., Gujar, V., Burdett, V., Modrich, P. L. & Kadyrov, F. A. 289
Human MutLgamma, the MLH1-MLH3 heterodimer, is an endonuclease that 290
promotes DNA expansion. Proc Natl Acad Sci U S A 117, 3535-3542, 291
doi:10.1073/pnas.1914718117 (2020). 292
14 Sonntag Brown, M., Lim, E., Chen, C., Nishant, K. T. & Alani, E. Genetic 293
analysis of mlh3 mutations reveals interactions between crossover promoting 294
factors during meiosis in baker's yeast. G3 (Bethesda) 3, 9-22, 295
doi:10.1534/g3.112.004622 (2013). 296
15 Claeys Bouuaert, C. & Keeney, S. Distinct DNA-binding surfaces in the 297
ATPase and linker domains of MutLgamma determine its substrate 298
specificities and exert separable functions in meiotic recombination and 299
mismatch repair. PLoS Genet 13, e1006722, 300
doi:10.1371/journal.pgen.1006722 (2017). 301
16 Nishant, K. T., Chen, C., Shinohara, M., Shinohara, A. & Alani, E. Genetic 302
analysis of baker's yeast Msh4-Msh5 reveals a threshold crossover level for 303
meiotic viability. PLoS Genet 6, doi:10.1371/journal.pgen.1001083 (2010). 304
10
17 Santucci-Darmanin, S. et al. The DNA mismatch-repair MLH3 protein 305
interacts with MSH4 in meiotic cells, supporting a role for this MutL 306
homolog in mammalian meiotic recombination. Hum Mol Genet 11, 1697-307
1706 (2002). 308
18 Manhart, C. M. et al. The mismatch repair and meiotic recombination 309
endonuclease Mlh1-Mlh3 is activated by polymer formation and can cleave 310
DNA substrates in trans. PLoS Biol 15, e2001164, 311
doi:10.1371/journal.pbio.2001164 (2017). 312
19 Kneitz, B. et al. MutS homolog 4 localization to meiotic chromosomes is 313
required for chromosome pairing during meiosis in male and female mice. 314
Genes Dev 14, 1085-1097 (2000). 315
20 Flores-Rozas, H. & Kolodner, R. D. The Saccharomyces cerevisiae MLH3 316
gene functions in MSH3-dependent suppression of frameshift mutations. Proc 317
Natl Acad Sci U S A 95, 12404-12409 (1998). 318
21 Lipkin, S. M. et al. MLH3: a DNA mismatch repair gene associated with 319
mammalian microsatellite instability. Nat Genet 24, 27-35, 320
doi:10.1038/71643 (2000). 321
22 Wu, Y. et al. A role for MLH3 in hereditary nonpolyposis colorectal cancer. 322
Nat Genet 29, 137-138, doi:10.1038/ng1001-137 (2001). 323
23 Dherin, C. et al. Characterization of a highly conserved binding site of Mlh1 324
required for exonuclease I-dependent mismatch repair. Mol Cell Biol 29, 907-325
918, doi:10.1128/MCB.00945-08 (2009). 326
24 Pluciennik, A. et al. PCNA function in the activation and strand direction of 327
MutLalpha endonuclease in mismatch repair. Proc Natl Acad Sci U S A 107, 328
16066-16071, doi:10.1073/pnas.1010662107 (2010). 329
25 Rass, U. et al. Mechanism of Holliday junction resolution by the human 330
GEN1 protein. Genes Dev 24, 1559-1569, doi:10.1101/gad.585310 (2010). 331
26 Bruning, J. B. & Shamoo, Y. Structural and thermodynamic analysis of 332
human PCNA with peptides derived from DNA polymerase-delta p66 subunit 333
and flap endonuclease-1. Structure 12, 2209-2219, 334
doi:10.1016/j.str.2004.09.018 (2004). 335
27 Lee, S. D. & Alani, E. Analysis of interactions between mismatch repair 336
initiation factors and the replication processivity factor PCNA. J Mol Biol 337
355, 175-184, doi:10.1016/j.jmb.2005.10.059 (2006). 338
28 Liberti, S. E. et al. Bi-directional routing of DNA mismatch repair protein 339
human exonuclease 1 to replication foci and DNA double strand breaks. DNA 340
Repair (Amst) 10, 73-86, doi:10.1016/j.dnarep.2010.09.023 (2011). 341
29 Genschel, J. et al. Interaction of proliferating cell nuclear antigen with PMS2 342
is required for MutLalpha activation and function in mismatch repair. Proc 343
Natl Acad Sci U S A 114, 4930-4935, doi:10.1073/pnas.1702561114 (2017). 344
30 Sanchez, A. A., C; Rauh, F; Duroc, Y; Ranjha, L; Lombard, B; Mu, X; Loew, 345
D; Keeney, S; Cejka, P; Guérois R; Klein F; Charbonnier, JB; Borde, V. 346
Mechanism of in vivo activation of the MutLγ-Exo1 complex for meiotic 347
crossover formation. bioRxiv 2019.12.16.876623, 348
doi:https://doi.org/10.1101/2019.12.16.876623 (2019). 349
350
351
11
FIGURELEGENDS352
353
Figure1.MSH4-MSH5andEXO1(DA)interactwithandpromotetheMLH1-354
MLH3endonuclease.a,Top,alignmentofthemetalbindingmotifinMLH3.Con-355
servedresiduesarehighlightedinred.Substitutionmutationsusedinthisstudy356
areinitalics.Bottom,nucleaseassayswithMLH1-MLH3and2.7kbp-longscDNA,357
with5mMmanganeseacetate,withoutorwith0.5mMATP,at37°C.b,Protein358
interaction assays with immobilized MSH4-MSH5 (bait, 220 nM) and MLH1-359
MLH3(prey).The10%polyacrylamidegelwasstainedwithsilver.c,Nuclease360
assayswithMutSγandMutLγ(5mMmanganeseacetate,30°C,0.5mMATP).Av-361
eragesshown;n=3;errorbars,SEM.d,Top,alignmentofMSH5andMSH4ATPase362
domains.Conservedresiduesarehighlightedinred.Alaninesubstitutionsusedin363
thisstudyareinitalics.Bottom,nucleaseassayswithMutSγATPasevariants(5364
mMmanganeseacetate,30°C,0.5mMATP).Averagesshown;errorbars,SEM;365
n=3.e,Top,alignmentofMLH1andMLH3ATPasedomains.Conservedresidues366
arehighlightedinred.Alaninesubstitutionsusedinthisstudyareinitalics.Bot-367
tom,nucleaseassayswithMutLγATPasevariants(5mMmanganeseacetate,30368
°C,0.5mMATP).Averagesshown;errorbars,SEM;n=3.f,Quantitationofkinetic369
nuclease assays with MutSγ and MutLγ, without or with nuclease-deficient370
EXO1(DA)(20nM).Theassayswerecarriedoutat30°Cwith5mMmanganese371
acetateand2mMATP.Averagesshown;errorbars,SEM;n=3.g,Proteininterac-372
tionassayswithimmobilizedMutLγ(bait)andEXO1(prey).The10%polyacryla-373
midegelwasstainedwithsilver.374
375
Figure2.RFC-PCNApromoteDNAcleavagebytheMutLγ-MutSγ-EXO1(DA)376
ensemble.a,Nucleaseassayswith5.6kbp-longscDNAandindicatedhumanpro-377
teinswascarriedoutwith5mMmagnesiumacetateand2mMATPat37°C.A378
representative experiment is shown at the bottom, a quantitation (averages379
shown;n=7;errorbars,SEM)atthetop.b,Representativenucleaseassayscarried380
outwith5mMmagnesiumacetate,5mMmanganeseacetateor5/1mMmagne-381
sium/manganeseacetate,respectively,asindicated,withindicatedproteins,con-382
taining 2 mM ATP and incubated at 37 °C. c, Protein interaction assays with383
12
immobilizedMLH1-MLH3(bait)andPCNA(prey).The10%polyacrylamidegel384
wasstainedwithsilver.d,Nucleasereactionsasinpanela,lane5,buteitherwith-385
outATP,withATPorwithnon-hydrolysableAMP-PNP(2mM).Averagesshown;386
errorbars, SEM;n=4.e,Nuclease reactionswithMLH1-MLH3(50nM),MSH4-387
MSH5(50nM),EXO1(DA)(50nM)andyRFC-hPCNA(50-100nM,respectively),388
lane2.Lanes3-7containinsteadMLH1-MLH3orMSH4-MSH5variants(50nM)389
deficientinATPhydrolysis,asindicated.SeeschemesinFig.1d,eforthespecific390
mutations.Reactionswerecarriedoutwith5mMmagnesiumacetateand2mM391
ATPat37°C.Averagesshown;errorbars,SEM;n=4.f,Representativenuclease392
reactionswithMLH1-MLH3,MSH4-MSH5,EXO1(DA)andyRFC-hPCNA,asindi-393
cated,with3.5kbp-longdsDNAeithercontaining(left)ornot(right)DNArepeat394
formingHJ-likecruciformDNA.Averagesshown;errorbars,SEM;n=9.395
396
Figure3.ThestimulationoftheMLH3nucleaseensemblebyPCNArequires397
aPIPboxmotifandisconservedinevolution.a,NucleaseassayswithMLH1-398
MLH3,MSH4-MSH5,EXO1(DA),andyRFC-hPCNA,asindicated,with5mMmag-399
nesiumacetateand2mMATPat37°C.Thereactionsweresupplementedwitha400
p21PIP-boxwildtypeormutatedcontrolpeptide,whereindicated(670nM,~5-401
foldoverKdofwildtypepeptideforPCNA)26.Averagesshown;errorbars,SEM;402
n=4.b,AlignmentofPIP-box-likemotifsfromvarioushumanorS.cerevisiaepro-403
teins.Residuesmorelikelytobeconservedarehighlightedinred.Wildtypehu-404
manandyeastEXO1,MLH3/Mlh3andMLH1/Mlh1weremutatedtocreatere-405
spective(P)variantswithindicatedresiduesubstitutions(Ainitalics).c,Nuclease406
assays with MLH1-MLH3, MSH4-MSH5, EXO1(DA), and yRFC-hPCNA, as indi-407
cated,with5mMmagnesiumacetateand2mMATPat37°C.Whereindicated,408
wild type MLH1 was replaced with MLH1P (Q562A-I565A-F568A), wild type409
MLH3 with MLH3P (Q341A-V344A-F347A), and EXO1(DA), with EXO1(DA)P410
(D173A-Q788A-L791A).Averagesshown;errorbars,SEM,n=5.d,Recombination411
frequency,expressedasamapdistanceincentimorgans,wasassayedinthewild412
typeS.cerevisiaestrain,mlh1Dandmlh3D,andinstrainscomplementedwithan413
untaggedconstructexpressingwildtypeMlh1,Mlh1P(Q572A-L575A-F578A)or414
Mlh3P(Q293A-V296A-F300A).Allvariantswereexpressedfromtheendogenous415
gene locus. Averages shown; error bars, SD; n≥900 from 3 independent416
13
experiments.e,FrequencyofchromosomeVIIInon-disjunctioninstrainsasde-417
scribedinpaneld.Averagesshown;errorbars,SD;n≥900from3independent418
experiments.f,Rfc1-TAPlevelsattheindicatedmeioticDSBhotspotsrelativetoa419
negativecontrolsite(NFT1)wereassessedbyChIPandqPCRduringameiotic420
time-course (synchronized pCUP1-IME1 cells).Averagesshown;errorbars,SD;421
n=2.ThecartoonillustratesthepositionofsitesanalyzedbyqPCRrelativetothe422
meioticchromosomestructure.423
424
425
14
METHODS426
427
Preparationofexpressionvectors428
To prepare the MSH4-MSH5 expression vector, the MSH4-STREP and MSH5-429
8xHISconstructswerecodon-optimizedforexpressioninSpodopterafrugiperda430
Sf9cellsandsynthesized(GenScript).ThegeneswereamplifiedbyPCRusingM13431
forwardandreverseprimers(seeSupplementaryDataTable1forsequencesof432
alloligonucleotides)anddigestedwithSalIandHindIII(forMSH4)orSmaIand433
KpnI(forMSH5)restrictionendonucleases(NewEnglandBiolabs).Digestedfrag-434
ments were ligated into corresponding sites in pFBDM (Addgene) to obtain435
pFBDM-hMSH4co-STREPandpFBDM-hMSH5co-HIS,respectively.Bothplasmids436
werethendigestedwithBamHIandHindIII(NewEnglandBiolabs)andligatedto437
generate pFBDM-hMSH4co-STREP-hMSH5co-HIS. To prepare expression con-438
structscodingfortheATPasevariantsofhMSH4andhMSH5,therespectivecon-439
servedresiduesintheWalkerAmotifs(seeref16)weremutatedbyQuikChange440
IIsite-directedmutagenesiskit(AgilentTechnologies).ToprepareMSH4G685A,441
the pFB-hMSH4co-STREP-hMSH5co-HIS vector was mutated with primers442
HMSH4G685A_FO and HMSH4G685A_RE. This created pFB-hMSH4coG685A-443
STREP-hMSH5co-HIS. To prepare MSH5G597A, HMSH5G597A_FO and444
HMSH5G597A_RE primers were used to create pFB-hMSH4co-STREP-445
hMSH5coG597A-HIS.Wealsoprepareda construct combiningbothmutations,446
buttheresultingmutantcomplexwasnotstableandcouldnotbepurified.447
TopreparetheMLH1andMLH3expressionvectors,bothgeneswereamplified448
byPCRfrompFL-his-MLH3co-MLH1cocontainingbothMLH1andMLH3genes,449
whichwerecodon-optimizedforinsectcellexpression.ToamplifyMLH1,FLAG-450
hMLH1co_FO and hMLH1co_RE primers were used. The PCR product was di-451
gested withNheI and XbaI (New England Biolabs) and inserted in pFB-MBP-452
MLH3-his11 creatingpFB-FLAG-hMLH1co(thesequenceofMLH3wasremoved453
during this step). Similarly, MLH3 was amplified using MLH3co_FO and454
MLH3co_RE.ThePCRproductwasdigestedwithNheIandXmaI(NewEngland455
Biolabs)andinsertedintopFB-MBP-MLH3-HIS,generatingpFB-MBP-hMLH3co.456
Thesequenceofnon-optimizedMLH3wasremovedduringthisstep.Theconsen-457
sus metal-binding motif in MLH3 is DQHAADE (conserved residues458
15
underlined)3,10.Topreparethenuclease-deadvariant,thesequenceofwildtype459
MLH3inpFB-MBP-MLH3co-HISwasmutatedusingprimersHMLH33ND_FOand460
HMLH33ND_RE. This created a sequence with 3 point mutations including461
D1223N,Q1224KandE1229K(NKHAADK,mutatedresiduesinitalics),andthe462
resulting vector was pFB-MBP-MLH3co3ND-HIS. To disrupt the ATPase of463
hMLH114,thepFB-FLAG-hMLH1cowasmutatedusingprimersHMLH1E34A_FO464
andHMLH1E34A_RE.ThiscreatedpFB-FLAG-MLH1E34A.Tomutatethecorre-465
sponding conserved residue inMLH3, thepFB-HIS-MBP-MLH3cowasmutated466
usingprimersHMLH3E28A_FOandHMLH3E28A_RE.ThiscreatedpFB-HIS-MBP-467
MLH3E28A.TopreparetheMLH1Pvariant,thepFB-FLAG-hMLH1coplasmidwas468
mutatedusingHMLH1_PIP1_3AFOandHMLH1_PIP1_3AREprimers.Toprepare469
the MLH3P variant, the pFB-FLAG-hMLH3co plasmid was mutated using470
HMLH3_PIP_3AFOandHMLH3_PIP_3AREprimers.471
TopreparepFB-hEXO1-FLAG,thesequencecodingforwildtypehumanEXO1472
(orEXO1[DA],containingtheD173Amutationinactivatingitsnuclease)wasam-473
plifiedbyPCRusingprimersHEXO1_FOandHEXO1_RE,andrespectivevectors(a474
kindgiftfromStefanoFerrari,UniversityofZurich)31astemplates.ThePCRprod-475
uctsweredigestedbyBamHIandXmaI(NewEnglandBiolabs),andclonedinto476
correspondingsitesinpFB-MBP-Sae2-HIS32(thesequenceofMBP-Sae2wasre-477
movedduringtheprocess,FLAG-tagwasaddedtotheC-terminusandaHIS-tag478
fromtheoriginalconstructwasnottranslatedduetoaStopcodon).479
TopreparepFB-hMSH2-FLAG, thesequencecoding forMSH2wasamplified480
from pFB-hMSH233 using primers HMSH2FLAG_FO and HMSH2FLAG_RE. The481
PCRproductwasdigestedbyBamHIandXhoI(NewEnglandBiolabs),andcloned482
into corresponding sites inpFB-MBP-Sae2 (the sequenceofMBP-Sae2was re-483
movedduringtheprocess,FLAG-tagwasaddedtotheC-terminusofMSH2anda484
HIS-tagfromtheoriginalconstructwasnottranslatedduetoaStopcodon).485
To prepare pFB-HIS-yMLH1, pFB-GST-yMLH111 was digested using BamHI486
(NewEnglandBiolabs)toremovetheGSTtag.Thisprocedureleftbehindasingle487
BamHIsite.TwocomplementaryoligonucleotidesHis-ForandHis-Revwerean-488
nealedtoeachother,andclonedintotheBamHIsite.Thisintroducedasequence489
codingfor8xHIStagbeforetheyeastMLH1genecreatingpFB-HIS-yMLH1.490
16
TopreparepFB-MBP-yMLH3,a terminationcodonwas introducedafter the491
yMLH3gene inpFB-MBP-yMLH3-HIS11sothattheHIS-tagwouldnotbetrans-492
latedwiththeyMlh3protein.Thiswascarriedoutbysite-directedmutagenesis493
usingforwardprimer329andreverseprimer330.494
TopreparetheexpressionvectorforyeastMsh4,theyeastMSH4genewasam-495
plifiedfromthegenomicDNAoftheS.cerevisiaeSK1strainusingforwardprimer496
258andreverseprimer259b.Thereverseprimerintroducedthesequencefor497
the C-terminal STREP affinity tag. The amplified product was digested with498
BamHIandHindIII(NewEnglandBiolabs)andclonedintocorrespondingsitesof499
pFB-GST-MLH111tocreatepFB-yMSH4-STREP.TheyeastMSH5genewasampli-500
fiedfromthegenomicDNAoftheS.cerevisiaeW303strainusingforwardprimer501
265andreverseprimer266.TheMSH5genewas thencloned intoBamHIand502
XhoIrestrictionsitesofpFB-MBP-MLH3-HIS11tocreatepFB-yMSH5-HIS.503
504
PurificationofhumanMLH1-MLH3505
The bacmids and baculoviruses were prepared individually using pFB-FLAG-506
hMLH1co andpFB-HIS-MBP-hMLH3co vectors according tomanufacturer’s in-507
structions(Bac-to-Bacsystem,LifeTechnologies).Spodopterafrugiperda9(Sf9)508
cellswereseededat500,000cellsperml16hbefore infection.Thecellswere509
thenco-infectedwithbothbaculovirusesand incubated for52hat27 °Cwith510
constantagitation.Thecellswerethenharvested(500xg,10min)andwashed511
oncewithPBS(137mMNaCl,2.7mMKCl,10mMNa2HPO4,1.8mMKH2PO4).The512
pelletsweresnap-frozeninliquidnitrogenandstoredat-80°C.Allsubsequent513
stepswerecarriedoutoniceorat4°C.Thepelletswereresuspendedin3vol-514
umesof lysisbuffer[50mMTris-HClpH7.5,1mMdithiothreitol(DTT),1mM515
ethylenediaminetetraacetic acid (EDTA), 1 mM phenylmethylsulfonyl fluoride516
(PMSF),1:400(volume/volume)proteaseinhibitorcocktail(Sigma,P8340),30517
μg/ml leupeptin (Merck)] and incubated for 20min with continuous stirring.518
Next,1/2volumeof50%glycerolwasadded,followedby6.5%volumeof5M519
NaCl(finalconcentration305mM).Thesuspensionwasfurtherincubatedfor30520
minwithcontinuousstirring.Thecellsuspensionwascentrifugedfor30minat521
48,000xg toobtainsolubleextract.Thesupernatantwas transferred to tubes522
containingpre-equilibratedAmyloseresin(NewEnglandBiolabs,4mlper1lof523
17
Sf9culture)andincubatedfor1hwithcontinuousagitation.Theresinwascol-524
lectedbyspinningat2,000xgfor2minandwashedextensivelybatchwiseand525
onadisposablecolumn(10ml,ThermoFisher)withAmylosewashbuffer [50526
mMTris-HClpH7.5,1mMβ-mercaptoethanol(β-ME),1mMPMSF,10%glycerol,527
300mMNaCl].ProteinwaselutedwithAmyloseelutionbuffer[50mMTris-HCl528
pH 7.5, 0.5 β-ME, 1 mM PMSF, 10% glycerol, 300 mM NaCl, 10 mMmaltose529
(Sigma)]andthetotalproteinconcentrationwasestimatedbyBradfordassay.To530
cleaveoffthemaltosebindingtag(MBP),1/6(weight/weight)ofPreScissionpro-531
tease(PP)34,withrespecttototalproteinconcentrationintheeluate,wasadded532
andincubatedfor1h.Next,thecleavedamyloseeluatewasdilutedbyadding1/2533
volumeofFLAGdilutionbuffer(50mMTris-HClpH7.5,1mMPMSF,10%glyc-534
erol,300mMNaCl)tolowertheconcentrationofβ-ME.Thedilutedeluatewas535
then incubated batchwise for 1 hwith pre-equilibrated anti-FLAGM2 affinity536
resin(Sigma,A2220,0.8ml).TheresinwaswashedextensivelywithFLAGwash537
buffer(50mMTris-HClpH7.5,0.5mMβ-ME,1mMPMSF,10%glycerol,150mM538
NaCl).ProteinwaselutedwithFLAGwashbuffercontaining150ng/μl3xFLAG539
peptide(Sigma),aliquoted,frozeninliquidnitrogenandstoredat-80°C.Thefinal540
constructcontainedaFLAGtagattheN-terminusofMLH1.Theyieldfrom1lcul-541
turewas~0.5mgandtheconcentration~2μM.AllMLH1-MLH3mutantswere542
expressedandpurifiedusingthesameprocedure.543
544
PurificationofhumanMSH4-MSH5545
The human MSH4-MSH5 complex was expressed from a dual pFB-hMSH4co-546
STREP-hMSH5co-HISvectorinSf9cellsusingtheBac-to-Bacsystemasdescribed547
above.Allpurificationstepswerecarriedouton iceorat4°C.Thecellpellets548
wereresuspendedin3volumesofnickel-nitriloaceticacid(NiNTA)lysisbuffer549
[50mMTris-HClpH7.5,2mMβ-ME,1mMPMSF,1:400(volume/volume)prote-550
aseinhibitorcocktail(Sigma,P8340),30μg/mlleupeptin(Merck),20mMimid-551
azole]andincubatedfor20minwithcontinuousstirring.Next1/2volumeof50%552
glycerolwasadded, followedby6.5%volumeof5MNaCl(finalconcentration553
305mM),andthesuspensionwasfurtherincubatedfor30minwithcontinuous554
stirring.Toobtainsolubleextract,thesuspensionwascentrifugedat48,000xg555
for 30 min. The soluble extract was transferred to a tube containing pre-556
18
equilibratedNiNTAresin(Qiagen,4mlper1lSf9cells)andincubatedfor1hwith557
continuousmixing.TheNiNTAresinwascollectedbycentrifugationat2,000xg558
for2min.Theresinwaswashedextensivelybatchwiseandonadisposablecol-559
umnwithNiNTAwashbuffer(50mMTris-HClpH7.5,2mMβ-ME,300mMNaCl,560
1mMPMSF,10%glycerol,20mM imidazole).ProteinwaselutedwithNiNTA561
washbuffercontaining250mMimidazole.Theelutedsamplewasincubatedwith562
pre-equilibratedStrep-TactinSuperflowresin(Qiagen,0.7ml) for90minwith563
continuousmixing.Theresinwastransferredtoadisposablecolumnandwashed564
extensivelywithStrepwashbuffer(50mMTris-HClpH7.5,2mMβ-ME,300mM565
NaCl,1mMPMSF,10%glycerol).ProteinwaselutedwithStrepwashbuffercon-566
taining2.5mMd-Desthiobiotin(Sigma,D1411)andstoredat-80°Caftersnap567
freezing in liquidnitrogen.The finalconstructcontainedaSTREPtagat theC-568
terminusofMSH4andaHIS-tagattheC-terminusofMSH5.Thevariantsofthe569
MSH4-MSH5complexwerepurifiedusingthesameprocedure.Wenotethatthe570
doublemutantMSH4G685A-MSH5G597Aheterodimerwasnotstableandcould571
notbepurified.572
573
PurificationofhumanEXO1(DA)574
ThepFB-EXO1(D173A)-FLAGvectorwasusedtopreparerecombinantbaculovi-575
rusandtheproteinwasexpressedinSf9cellsasdescribedabove.Frozencellpel-576
letwasthawedandresuspendedin3pelletvolumesoflysisbuffer[50mMTris-577
HClpH7.5,0.5mMβ-ME,1mMEDTA,1:400(volume/volume)proteaseinhibitor578
cocktail(Sigma,P8340),0.5mMPMSF,20µg/mlleupeptin].Thecellsuspension579
wasincubatedwithgentlestirringfor10min.1/2volumeof50%glyceroland580
6.5%volumeof5MNaCl(finalconcentration305mM)wereadded.Thesuspen-581
sionwasincubatedfor30minwithstirring.Theextractwasthencentrifugedat582
48,000xgfor30min.Thesolubleextractwasaddedtopre-equilibratedM2anti583
FLAGaffinityresin(Sigma,A2220,2mlresinforpurificationfrom1lSf9cellcul-584
ture)andincubatedbatchwisefor45min.Thesuspensionwasthencentrifuged585
(2,000xg,5min),thesupernatant(FLAGflowthrough)removed,andtheresin586
wastransferredtoadisposablechromatographycolumn.Theresinwaswashed587
with50resinvolumesofTBSbuffer(20mMTris-HClpH7.5,150mMNaCl,0.5588
mMβ-ME,0.5mMPMSF,10%glycerol)supplementedwith0.1%NP40.Thiswas589
19
followedbywashingwith10resinvolumesofTBSbufferwithoutNP40.EXO1-590
FLAGwaselutedwithTBSbuffersupplementedwith150ng/μl3xFLAGpeptide591
(Sigma, F4799). Fractions containing detectable protein (as estimated by the592
Bradfordmethod)werepooled,appliedonadisposablecolumnwith1mlpre-593
equilibratedBiorex70resin(Bio-Rad),andflow-throughwascollected.Thesam-594
plewasthendilutedbyadding1volumeofdilutionbuffer(50mMTris-HClpH595
7.5,5mMβ-ME,0.5mMPMSF,10%glycerol).DilutedFLAG-EXO1wasappliedon596
1mlHiTrapSPHPcolumn(GEHealthcare)pre-equilibratedwithSbufferA(50597
mMTris-HClpH7.5,75mMNaCl,5mMb-ME,10%glycerol)at0.8ml/min.The598
columnwaswashedwith20mlSbufferA,andelutedwith8mllinearsaltgradi-599
entinSbufferA(75mMto1MNaCl).Peakfractionswerepooled,aliquoted,fro-600
zeninliquidnitrogenandstoredat-80°C.Theprocedureyieldedaround~0.15601
mgofproteinfrom1lofSf9culture,withanapproximateconcentrationof~1μM.602
603
PurificationofhumanMSH2-MSH6andMSH2-MSH3heterodimers604
TopreparetheMSH2-MSH6heterodimer,theSf9cellswereco-infectedwithre-605
combinant baculoviruses prepared from pFB-hMSH2-FLAG and pFB-hMSH6-606
HIS33vectors.Thepurificationwascarriedoutat4°Coronice.Thecellpellets607
wereresuspendedin3volumesoflysisbuffer[50mMTris-HClpH7.5,1:400[vol-608
ume/volume]proteaseinhibitorcocktail(Sigma,P8340),1mMPMSF,60μg/ml609
leupeptin,0.5mMβ-ME,20mMimidazole].Thesamplewasincubatedwhilestir-610
ringfor20min.1/2volumeof50%glycerolwasadded,followedby6.5%volume611
5MNaCl(finalconcentration305mM).Thecellsuspensionwasincubatedfor30612
minwithstirring.Toobtainsolubleextract,thesuspensionwascentrifuged(30613
min,48,000xg).Thesupernatantwasmixedwithpre-equilibrated2mlNiNTA614
resin(purificationfrom800mlSf9cells)andincubatedbatchwisefor1h.The615
resinwasthenwashedbatchwiseandoncolumnwithwashbuffer[30mMTris-616
HClpH7.5,1:1,000(volume/volume)proteaseinhibitorcocktail(Sigma,P8340),617
15μg/mlleupeptin,0.5mMβ-ME,0.5mMPMSF,20mMimidazole,300mMNaCl,618
10%glycerol].Boundproteinwaselutedwithelutionbuffer[30mMTris-HClpH619
7.5, 1:1,000 (volume/volume) protease inhibitor cocktail (Sigma, P8340), 15620
μg/mlleupeptin,0.5mMβ-ME,0.5mMPMSF,300mMimidazole,150mMNaCl,621
20
10%glycerol].Thepooledfractionsweredilutedwith7volumesofdilutionbuffer622
(30mMTris-HClpH7.5,15μg/mlleupeptin,0.5mMβ-ME,0.5mMPMSF,150623
mMNaCl,10%glycerol),andmixedwith0.7mlpre-equilibratedanti-FLAGM2624
affinity gel (Sigma). The suspensionwas incubated batchwise for 60min. The625
samplewascentrifuged(5min,1,000g)andresinwastransferredtoadisposable626
chromatographycolumn.The resinwas thenwashedextensivelywithdilution627
buffer.Theheterodimerwaselutedwithdilutionbuffersupplementedwith200628
μg/ml 3x FLAG peptide (Sigma). Eluates containing protein were pooled, ali-629
quoted,frozeninliquidnitrogenandstoredat-80°C.TheMSH2-MSH3heterodi-630
merwaspreparedusingthesameprocedure,usingpFB-hMSH3-HIS35.631
632
PurificationofyeastMsh4-Msh5633
BaculovirusesexpressingMsh4andMsh5werepreparedindividuallyusingthe634
Bac-to-BacsystemandpFB-yMSH4-STREPandpFB-yMSH5-HISvectors.Sf9cells635
wereco-infectedwithoptimizedratiosofbothvirusestoexpressbothproteins636
togetherasaheterodimer.Thecellswereharvested52hafterinfection,washed637
withPBS,andthepelletswerefrozeninliquidnitrogenandstoredat-80°Cuntil638
use.Thesubsequentstepswerecarriedoutoniceorat4°C.Thecellpelletwas639
resuspended in lysisbuffer [50mMTris-HClpH7.5,2mMβ-ME,1mMEDTA,640
1:400(volume/volume)proteaseinhibitorcocktail(Sigma,P8340),1mMPMSF,641
30μg/mlleupeptin,20mMimidazole]for20min.Then,50%glycerolwasadded642
toafinalconcentrationof16%,followedby5MNaCltoafinalconcentrationof643
305mM.Thesuspensionwasincubatedforfurther30minwithgentleagitation.644
Thetotalcellextractwascentrifugedat48,000xgfor30mintoobtainsoluble645
extract.TheextractwasthenboundtoNiNTAresin(Qiagen)for60minbatchwise646
followedbyextensivewashingwithNiNTAwashbuffer(50mMTris-HClpH7.5,647
2mMβ-ME,300mMNaCl,10%glycerol,1mMPMSF,10μg/mlleupeptin,20mM648
imidazole) both batchwise and on a column. The heterodimer was eluted by649
NiNTAelutionbuffer(NiNTAwashbuffercontaining250mMimidazole).Theelu-650
atewasfurtherincubatedwithpre-equilibratedStrep-TactinSuperflowresin(Qi-651
agen)for60minbatchwise.Theprotein-boundresinwasthenwashedintwose-652
quentialsteps;firstwithSTREPwashbufferI(50mMTrispH7.5,2mMβ-ME,10653
%glycerol,1mMPMSFand300mMNaCl)andthenwithSTREPwashbufferII654
21
(50mMTrispH7.5,2mMβ-ME,10%glycerol,1mMPMSFand50mMNaCl).655
TheheterodimerwaselutedwithSTREPwashbufferIIcontaining2.5mMd-Des-656
thiobiotin(Sigma).Theeluatewasthenappliedonapre-equilibratedHiTrapQ657
HPcolumn(GEHealthcare).ThecolumnwaswashedwithSTREPwashbufferII658
andproteinwaselutedwithalineargradientofNaCl(50to600mM)inSTREP659
wash buffer II. Collected fractionswere analyzed on SDS-PAGE, peak samples660
were pooled, aliquoted and stored at -80 °C. The final construct contained a661
STREPtagattheC-terminusofyMsh4andaHIS-tagattheC-terminusofyMsh5.662
Theprocedureyielded~0.15mgofproteinfrom4lofSf9culture,withanap-663
proximateconcentrationof~1μM.664
665
PurificationofyeastMlh1-Mlh3666
TheyMlh1-yMlh3heterodimerwasexpressedusingpFB-HIS-yMLH1andpFB-667
MBP-yMLH3and theBac-to-Bac systemandpurifiedusingaffinity chromatog-668
raphy11.Briefly,thecellswereresuspendedinlysisbuffercontaining50mMTris-669
HClpH7.5,1mMDTT,1mMEDTA,1:400(volume/volume),proteaseinhibitor670
cocktail(Sigma,P8340),1mMPMSF,30μg/mlleupeptinandincubatedfor20671
min.Subsequentlyglycerol[finalconcentration16%(volume/volume)]andNaCl672
(finalconcentration305mM)wereadded.Uponfurther incubationfor30min673
andcentrifugation(48,000xg,30min),theclearedextractwasthensubjectedto674
affinitychromatographywithAmyloseresin(NewEnglandBiolabs),theMBPtag675
wascleavedwithPreScissionproteaseandtheheterodimerwasfurtherpurified676
onNi-NTAagarose(Qiagen)11.Thefinaleluatewasdialyzedinto50mMTris-HCl677
pH7.5,5mMβ-ME,10%glycerol,0.5mMPMSFand300mMNaCl.Aliquotswere678
flashfrozenandstoredat-80°Cuntiluse.Thepurificationyielded~1mgprotein679
from2.4lcultureandtheconcentrationwas5.9μM.680
681
PurificationofyeastandhumanRFC,PCNAandtheKuheterodimer682
HumanPCNAwasexpressedinE.colicells(1l)frompET23C-his-hPCNAvector683
(akindgiftfromUlrichHuebscher,UniversityofZurich).Transformedcellswere684
grown to OD 0.5, and induced with 0.5 mM isopropyl β-D-1-thiogalactopyra-685
noside(IPTG)for3.5hat37°C.Cellswerelysedbysonicationinlysisbuffer(20686
22
mMTris-HClpH7.5,250mMNaCl,2mMβ-ME,5mMimidazole,1mMPMSF,687
1:250SigmaproteaseinhibitorcocktailP8340).Thelysatewasclearedbycen-688
trifugation(48,000xg,30min)andboundto2mlNiNTAresin(Qiagen)for1h689
batchwise.Resinwaswashedwithwashbuffer(20mMTris-HClpH7.5,250mM690
NaCl,2mMβ-ME,30mMimidazole,1mMPMSF),andPCNAwaselutedwithelu-691
tionbuffer(washbuffersupplementedwith400mMimidazole).Thesamplewas692
dilutedtoconductivitycorrespondingto100mMNaCl,andloadedonHiTrapQ693
column.Thecolumnwasdevelopedbyasaltgradient(100mMto1MNaCl)in20694
mMTris-HClpH7.5,2mMβ-MEand10%glycerol.ThefractionscontainingPCNA695
werepooled, aliquoted and stored at -80 °C. Yeast PCNAwasprepared asde-696
scribed36.697
ToexpresshumanRFC,theSf9cells(1.4l)wereinfectedwithrecombinant698
baculovirus prepared with vector pFBDM-MBP-RFC1-RFC2-3-4-His-5 (a kind699
giftfromJosefJiricny,ETHZurich,RFC1subunitMBP-taggedandRFC5subunit700
his-tagged).Thepurificationwascarriedoutat4°Coron ice.Thecellpellets701
wereresuspendedin3volumeslysisbuffer[50mMTris-HClpH7.5,2mMβ-ME,702
1:300(volume/volume)proteaseinhibitorcocktail(Sigma,P8340),1mMPMSF,703
30μg/mlleupeptin,15mMimidazole).Thecellswereletswellingonicefor20704
minandmixedoccasionally.Afterwards,1/2volumeof50%glycerolwasadded,705
followedby6.5%volume5MNaCl(finalconcentration305mM),andthesus-706
pensionwasincubatedwhilestirringfor30min.Thecellsuspensionwascentri-707
fugedat55,000xg for30mintoobtainsolubleextract.Thesupernatantwas708
mixedwithpre-equilibrated2mlNiNTAresin(Qiagen)andbatchincubatedwith709
gentleagitationfor1h.Theresinwaswashed3timesbatchwise,andwith15710
resinvolumesoncolumnwithwashbufferI(50mMTris-HClpH7.5,2mMb-ME,711
0.25MNaCl,10%Glycerol,1mMPMSF,20mMimidazole).TheRFCcomplexwas712
elutedwithNiNTAelutionbuffer(50mMTris-HClpH7.5,2mMb-ME,250mM713
NaCl,10%glycerol,1mMPMSF,300mMimidazole).Theeluatewasdirectlyap-714
plied in flowon1.5mlamyloseresin(NewEnglandBiolabs)pre-equilibrated715
withwashbuffer2(50mMTrispH7.5,2mMb-ME,0.1MNaCl,10%Glycerol,1716
mMPMSF),andwashedwithwashbuffer2.TheRFCcomplexwaselutedwith717
washbuffer2supplementedwith10mMmaltose.Theproteinconcentrationwas718
estimated by the Bradfordmethod, and the samplewas incubatedwith 20%719
23
(w/w) Prescission protease for 2 h at 4 °C. The samplewas then applied on720
HiTrapQcolumn(0.5ml/min),pre-equilibratedinQbufferA(50mMTrispH721
7.5,5mMb-ME,0.1MNaCl,10%Glycerol),thecolumnwasthenwashedwithQ722
bufferA,andelutedinthesamebufferwithasaltgradient(0.1Mto1M)in6723
columnvolumes.ThefractionscontainingRFCwerepooled,aliquoted,frozenin724
liquidnitrogenandstoredat-80°C.Thepurificationyielded~1mlof8.5μM725
RFC.726
YeastRFCwasexpressedinE.colicells(4l)transformedwithpEAO271(akind727
giftfromE.Alani,CornellUniversity).CellsweregrowntoOD0.5,andinduced728
with0.5mMIPTGfor3hat37°C.Cellswereresuspendedinlysisbuffer(60mM729
HEPES-NaOHpH7.5,250mMNaCl,2mMβ-ME,0.5mMEDTA,1:250Sigmapro-730
teaseinhibitorcocktailP8340,1mMPMSF,10%glycerol)anddisruptedbyson-731
ication.Theclearedextractwasloadedon5mlHiTrapSPcolumn,washedwith732
bufferSPA(30mMHEPES-NaOHpH7.5,300mMNaCl,2mMβ-ME,0.5mMEDTA,733
1mMPMSF,10%glycerol)andelutedwithasaltgradient(300mMto600mM734
NaCl). Eluted fractions were analyzed by polyacrylamide gel electrophoresis,735
pooledanddilutedtoconductivitycorrespondingto110mMNaCl.Thediluted736
samplewasappliedonHiTrapQcolumn,andelutedin110to600mMNaClgra-737
dientin30mMHEPES-NaOHpH7.5,2mMβ-ME,1mMPMSFand10%glycerol.738
Theeluatewasaliquotedandstoredat-80°C.ThepreparationoftheyeastKu739
heterodimerwasdescribedpreviously37.740
741
Nucleaseassays742
Thereactions(15µl)werecarriedoutin25mMTris-acetatepH7.5,1mMDTT,743
0.1mg/mlbovineserumalbumin(BSA,NewEnglandBiolabs),andasindicated744
manganeseormagnesiumacetate(5mM),ATP(concentrationsasindicated,GE745
Healthcare,27-1006-01)andplasmid-basedDNAsubstrate[100ngperreaction,746
either2.7kbp-longpUC19(Fig.1andrelatedExtendedDataFigures),5.6kbp-747
longpFB-RPA2(Figures2and3,andrelatedExtendedDataFigures),10.3kbp-748
longpFB-HIS-MBP-hMLH3co(ExtendedDataFig.4j),orpAG25(Addgene)orcru-749
ciformpIRbke8mut25(Fig.2f,ExtendedDataFig.4k)].Inexperimentswith32P-750
labeled oligonucleotide-based DNA the substrate concentration was 1 nM, in751
24
molecules.Whereindicated,ADP(AlfaAesar,J60672),AMP-PNP(TorontoChem-752
ical,A634303)orATP-γ-S(Cayman,14957)wereusedinsteadofATP.Wherein-753
dicated,thereactionsweresupplementedwithPIPboxpeptidederivedfromp21754
(GRKRRQTSMTDFYHSKRRLIFS) or control peptidewith key residuesmutated755
(underlined,GRKRRATSATDFYHSKRRLIFS)(Genecust).Thereactionbufferwas756
assembledonice,andtherecombinantproteinswerethenaddedonice(MLH1-757
MLH3proteinwasalwaysaddedlast).Thereactions,unlessindicatedotherwise,758
wereincubatedfor60minat30°Cor37°C.Thereactionsweresupplemented759
withproteinstorageordilutionbuffertocompensateforcomponentsintroduced760
with recombinantproteins ineachparticularexperiment, this resulted in final761
NaClconcentrations~30mM.Thereactionswereterminatedwith5µlSTOPso-762
lution(150mMEDTA,2%SDS,30%glycerol,0.01%bromophenolblue),1µlpro-763
teinaseK(Roche,03115828001,18mg/ml)andfurtherincubatedfor60minat764
50°C.Thereactionproductswerethenseparatedbyelectrophoresisin1%aga-765
rose(Sigma,A9539)containingGelRed(Biotium)inTAEbuffer.UsingBio-Rad766
SubCellGTsystem(gellength26cm),theseparationwascarriedoutfor90min767
at120V.Thegelswerethenimaged(InGenius3,GeneSys).Theresultswerequan-768
titatedusingImageJandexpressedas%ofnickedDNAversusthetotalDNAin769
eachparticular lane;anynickedDNApresent incontrol (noprotein)reactions770
wasremovedasabackground.Forgelsourcedata,seeSupplementaryFigure1.771
772
Electrophoreticmobilityshiftassays773
TheDNAbindingreactionswerecarriedout in15μlvolume inbindingbuffer774
containing25mMHEPESpH7.8,5mMmagnesiumchloride,5%(volume/vol-775
ume)glycerol,1mMDTT,50μg/mlBSA,6.6ng/μldsDNA(inassayswithyeast776
proteins)or3.3ng/μldsDNA(inassayswithhumanproteins)ascompetitor(50777
bp-long),0.5nMDNAsubstrate(32P-labelled,inmolecules)andrespectivecon-778
centrationsofrecombinantproteins(yeastorhumanMSH4-MSH5complexand779
theirvariants,MLH1-MLH3andvariants).Theoligonucleotide-basedDNAsub-780
stratesweressDNA(labelledoligonucleotidePC1253),dsDNA(labelledPC1253781
andPC1253C),Y-structure(labelledPC1254andPC1253),HJ(labelledPC1253782
andPC1254,PC1255andPC1256)andD-Loop(labelledBB,andBT, INVaand783
INVb).MgCl2wasreplacedby3mMEDTAwhereindicated.Thereactionswere784
25
assembledandincubatedonicefor15min,followedbytheadditionof5μlEMSA785
loadingdye(50%glycerol,0.01%bromophenolblue).Theproductsweresepa-786
ratedon6%nativepolyacrylamidegel(19:1acrylamide-bisacrylamide,BioRad)787
onice.Thegelsweredriedon17CHRpaper(Whatman),exposedtostoragephos-788
phor screens (GEHealthcare), and scannedbyPhosphorimager (TyphoonFLA789
9500,GEHealthcare).ThequantitationwascarriedoutbyImageQuantsoftware790
(GE Healthcare) and graphs were plotted using Prism software (Prism 8,791
Graphpad).Forgelsourcedata,seeSupplementaryFigure1.792
Forthe"super-shift"assayscomprisingyMlh1-yMlh3andyMsh4-yMsh5,there-793
actionswerecarriedoutasmentionedabove(withmagnesiumorEDTA,asindi-794
cated),exceptthattheproductswereseparatedon0.6%agarosegelinTAEbuffer795
at4°C(1h,100V).ThegelsweredriedonDE81paper(Whatman)andscanned796
asabove.Inthesuper-shiftassayswithMLH1-MLH3,MSH4-MSH5andEXO1,the797
reactionbufferadditionallycontained75mMNaCland10μMATP.TheDNAbind-798
ingassayswithyKu70-80werecarriedoutsimilarly,withoutsaltandATP,and799
wereincubatedfor30minat30°C.800
801
Proteininteractionassays802
Totestforprotein-proteininteractions,recombinant“bait”proteinwasimmobi-803
lizedonbeadscoupledtoaspecificantibodyandincubatedwiththe“prey”pro-804
tein.Afterremovalofunboundproteinbybeadswashing,proteinswereeither805
detectedbysilverstainingorbywesternblot.Forgelsourcedata,seeSupplemen-806
taryFigure1.807
TotestfortheinteractionbetweenMLH1-MLH3andMSH4-MSH5,0.7µganti-808
MLH1 antibody (Abcam, ab92312)was captured on 15µl Protein Gmagnetic809
beads (Dynabeads, Invitrogen) by incubating in 50µl PBS-T (PBSwith 0.02%810
Tween-20) for60minwithgentlemixingat regular intervals.Thebeadswere811
washed3timesonmagneticrackswith150µlPBS-Ttoremoveunboundanti-812
bodies.Thebeadswerethenmixedwith165nMrecombinantMLH1-MLH3and813
220nMMSH4-MSH5in50µlbindingbufferI(25mMHEPESpH7.8,3mMEDTA,814
1mMDTT,50µg/mlBSA,80mMNaCl)and incubatedon ice for45minwith815
gentleagitationatregularintervals.Beadswerethenwashed3timeswith150µl816
26
washbufferI(25mMHEPESpH7.8,3mMEDTA,1mMDTT,0.02%Tween-20,817
80mMNaCl)andproteinswereelutedbyboilingthebeadsinSDSbuffer(50mM818
Tris-HCl pH 6.8, 1.6% sodium dodecyl sulphate, 100 mM DTT, 10% glycerol,819
0.01%bromophenolblue)for3minat95°C.Theeluatewasseparatedona10%820
SDS-PAGEgelandproteinsweredetectedbysilverstaining.Toperformtheex-821
perimentreciprocally,5µganti-HISantibody(Genscript,A00186)wascaptured822
onProteinGbeads(Dynabeads,Invitrogen)asdescribedabove.Therecombinant823
protein complexes, as above,were then addedand incubated in50µl binding824
bufferII(25mMHEPESpH7.8,3mMEDTA,1mMDTT,50µg/mlBSA,80mM825
NaCl) for 45 min with gentle agitation at regular intervals. Beads were then826
washed3timeswithwashbufferII(25mMHEPESpH7.8,3mMEDTA,1mM827
DTT,80mMNaCl,0.1%TritonX-100).Thesubsequentstepswerecarriedoutas828
describedabove.Totestforspecies-specificinteractionsasshowninExtended829
DataFig.3g,thesameprocedurewasfollowedexcept100nMofeitherhuman830
MSH4-MSH5oryeastMsh4-Msh5wasincubatedwith400nMMLH1-MLH3.To831
testfortheinteractionbetweenyeastMlh1-Mlh3andMsh4-Msh5,10µlProtein832
G beads were used to capture 1 µg anti-STREP antibody (Biorad, MCA2489).833
yMsh4-yMsh5(120nM)wasincubatedwiththebeadsin60µlbindingbufferIII834
(25mMTris-HClpH7.5,3mMEDTA,1mMDTT,20mg/mlBSA,60mMNaCl)for835
60minwithcontinuousmixing.Next,thebeadswerewashed3timeswith150µl836
washbufferIII(25mMTris-HClpH7.5,3mMEDTA,1mMDTT,120mMNaCl,837
0.05%TritonX-100).300nMyMlh1-yMlh3wasthenaddedtotheresuspended838
beadsin60µlbindingbufferIII,andincubatedforadditional60minwithcontin-839
uousmixing.Beadswerewashed3timeswith150µlwashbufferIIIandboiled840
afterwardsfor3minat95°CinSDSbuffertoelutetheproteins.Theproteincom-841
plexesweredetectedbywesternblotwithanti-HISantibody(Genscript,A00186).842
TotestfortheinteractionbetweenMLH1-MLH3andEXO1,0.33µganti-MLH1843
antibody (Abcamab223844)was capturedon10µl proteinGmagneticbeads844
(Dynabeads,Invitrogen)byincubatingin50µlPBS-T(PBSwith0.1%Tween-20)845
for2hat4°Cwithgentlemixingatregularintervals.Thebeadswerewashed4846
timesonmagnetic rackswith150µlPBS-T to removeunboundantibody.The847
beadswerethenmixedwith1µgrecombinantMLH1-MLH3and0.5µgEXO1in848
27
200µlbindingbufferI(25mMTris-HClpH7.5,3mMEDTA,1mMDTT,20µg/ml849
BSA,300mMNaCl)andincubatedonicefor2hwithgentleagitationatregular850
intervals.Beadswerethenwashed4timeswith300µlwashbufferI(50mMTris-851
HClpH7.5,3mMEDTA,1mMDTT,300mMNaCl,0.05%TritonX-100)andpro-852
teinswereelutedbyboilingthebeadsinSDSbuffer(50mMTris-HClpH6.8,1.6%853
sodiumdodecylsulphate,100mMDTT,10%glycerol,0.01%bromophenolblue)854
for3minat95°C.Theeluatewasseparatedona10%SDS-PAGEgelandproteins855
weredetectedbysilverstaining.856
TotestfortheinteractionbetweenhumanMLH1-MLH3andhumanPCNAor857
EXO1,1µganti-MLH1antibody(Abcamab223844)wascapturedon15µlprotein858
Gmagneticbeads(Dynabeads,Invitrogen)byincubatingin50µlPBS-T(PBSwith859
0.1%Tween-20)for1hatroomtemperaturewithgentlemixingatregularinter-860
vals.Thebeadswerewashed3timesonmagneticrackswith150µlPBS-Ttore-861
moveunboundantibody.Thebeadswerethenmixedwith1.5µgeachrecombi-862
nantMLH1-MLH3andPCNAorEXO1,in60µlbindingbufferI(25mMTris-HCl863
pH7.5,3mMEDTA,1mMDTT,20µg/mlBSA,60mMNaCl)andincubatedonice864
for1hwithgentleagitationatregularintervals.Beadswerethenwashed4times865
with150µlwashbufferI(50mMTris-HClpH7.5,3mMEDTA,1mMDTT,120866
mMNaCl,0.05%TritonX-100)andproteinswereelutedbyboilingthebeadsin867
SDSbuffer(50mMTris-HClpH6.8,1.6%sodiumdodecylsulphate,100mMDTT,868
10%glycerol,0.01%bromophenolblue)for3minat95°C.Avidin(Sigma,A9275,869
110ng/µl)wasaddedtotheeluateasastabilizer.Theeluatewasseparatedona870
10%SDS-PAGEgelandproteinsweredetectedbysilverstaining.871
872
Yeastmanipulations873
AllyeaststrainsarederivativesoftheSK1backgroundandarelistedinSupple-874
mentaryDataTable2.Yeaststrainswereobtainedbydirecttransformationor875
crossing to obtain the desired genotype. The following alleles have been de-876
scribed previously: mlh1Δ, mlh3Δ as well as spore-autonomous fluorescent877
markerforthelivecellrecombinationassays38,39.878
YIplac211 plasmid derivatives carryingMLH1 (pYIplac211-MLH1, pML535) or879
MLH3(pYIplac211-MLH3,pML536),aswellastherespectivepromoter(~500bp880
28
upstreamofATG)andterminator(~200bpdownstreamofSTOP)regionswere881
used tocomplementmlh1∆ormlh3∆mutantstrains, respectively.pYIplac211-882
MLH1andpYIplac211-MLH3werelinearizedandintegratedinthepromoterre-883
gion of the respective genomic loci. pYIplac211-MLH1Q572A-L575A-F578A(pML538),884
encodingMlh1P,andpYIplac211-MLH3Q293A-V296A-F300A(pML540),encodingMlh3P,885
weregeneratedbyrestrictiondigest-mediatedinsertionofasyntheticfragment886
carrying the respectivemutations into pYIplac211-MLH1or pYIplac211-MLH3.887
PCR-basedC-terminaltaggingofMLH1andMLH3wasperformedusingstandard888
procedures40.889
Rfc1wasC-terminally taggedwithTAPtag.TheMlh1-HAandthe internally890
FLAG-Myc-taggedMlh3constructsweredescribedpreviously30,41.Transformants891
wereconfirmedusingPCRdiscriminatingbetweencorrectandincorrectintegra-892
tionsandsequencing.Allexperimentswereperformedat30°C.Twodifferentap-893
proacheswereusedformeiosisinduction.Inthefirstone,cellsweregrowninSPS894
presporulationmediumandtransferred insporulationmediumasdescribed42.895
Forhighlysynchronouscopper-induciblemeiosis,theprocedurewasdescribed43.896
Briefly, cellswere grown in YPD to exponential phase. Exponentially growing897
yeastwereinoculatedatOD600=0.05intoreducedglucoseYPD(1%yeastextract,898
2%peptone,1%glucose)andgrowntoanOD600=11-12for16-18h.Cellswere899
washed, resuspended in sporulation medium (1.0% [w/v] potassium acetate,900
0.02%[w/v]raffinose,0.001%polypropyleneglycol)atOD600=2.5.After2h,901
copper(II)sulfate(50µM)wasaddedtoinduceIME1expressionfromtheCUP1902
promoter.903
904
Analysisofrecombinationusingspore-autonomousfluorescence905
Thespore-autonomousfluorescenceanalysisofrecombinationwasperformedas906
described39,withminormodifications.Diploidyeastcellcolonieswerestreaked907
onYP2%glycerolplates,grownfor48h,andsinglecolonieswereexpandedtwicein908
YPDplatesat30°Cfor24h.Cellswerethentransferredtosporulationmedium909
plates(SPM,2%KAc)andincubatedat30°Cfor48h.Sporeswereresuspended910
inSPM,brieflysonicatedandtransferredontoPoly-L-Lysinecoatedmicroscopy911
slides.ImageswerecapturedinfourchannelsusingaWide-fieldDeltaVisionmul-912
tiplexedmicroscopewitha60x1.4NADICOilPlanApoNobjectiveandapeco.edge913
29
5.5cameraunderthecontrolofSoftworx(AppliedPrecision).Imageswerepro-914
cessed in Fiji and the pattern of spore fluorescence in tetrads was manually915
scored.Onlytetradswitheachfluorescentmarkeroccurringintwosporeswere916
includedinthefinalassay.Recombinationfrequency,expressedasmapdistance917
incentimorganswascalculatedusingtheStahllabonlinetools(https://elizabeth-918
housworth.com/StahlLabOnlineTools/)44. Three replicates using independent919
cloneswereanalyzed.≥900tetradswerescoredforeachgenotype.920
921
Analysisofsporeviability922
Sporeviabilitywasdeterminedbymicrodissectionof≥156sporesfromatleast923
twoindependentexperimentsafterinductionofmeiosisonSPMplatesat30°C924
for24h.925
926
Proteinstabilityanalysesbywesternblotting927
Proteinextractsfromyeastwereperformedusingthetrichloroaceticacid(TCA)928
method45.Briefly,exponentiallygrowingcultureswereharvestedanddisrupted929
usingglassbeadsin10%TCA.Precipitateswerecollectedbycentrifugation,re-930
suspendedin2xNuPAGEsamplebuffer(Invitrogen),andneutralizedwith1M931
Tris.Sampleswerethenboiledat95°Cfor5min,clearedbycentrifugation,and932
separatedinNuPAGE3-8%Tris-Acetategels(Invitrogen).Aftergelelectrophore-933
sis,proteinsweretransferredontoPVDFmembranes(GEHealthcare).Antibodies934
targeting the following tags or proteins were used: mouse anti-FLAG HRP-935
conjugated (1:15,000, A8592-1MG, Sigma), rabbit anti-FLAG (1:2000, F7425-936
.2MGSigma),rabbitanti-Crm1(1:5000,agiftfromK.Weis,ETHZurich),swine937
anti-rabbitHRP-conjugated(1:5000,P0399,Dako).Forgelsourcedata,seeSup-938
plementaryFigure1.939
940
Co-immunoprecipitationandWesternblotanalysis941
1.2x109cellswereharvested,washedoncewithPBS,andlysedin3mllysisbuffer942
[20mMHEPES-KOHpH7.5,150mMNaCl,0.5%TritonX-100,10%glycerol,1943
mMMgCl2,2mMEDTA;1mMPMSF;1xCompleteMiniEDTA-Free(Roche);1X944
PhosSTOP(Roche);125U/mlbenzonase]withglassbeadsthreetimesfor30sin945
aFastprepinstrument(MPBiomedicals,SantaAna,CA).Thelysatewasincubated946
30
1h at 4 °C. 100μl of PanMouse IgGmagnetic beads (ThermoScientific)were947
washedwith100μllysisbuffer,preincubatedin100μg/mlBSAinlysisbufferfor948
2hat4°Candthenwashedtwicewith100μllysisbuffer.Thelysatewascleared949
bycentrifugationat13,000xg for5minand incubatedovernightat4°Cwith950
washedPanMouse IgGmagneticbeads.Themagneticbeadswerewashed four951
timeswith1mlwashbuffer[20mMHEPES-KOHpH7.5,150mMNaCl,0.5%Tri-952
tonX-100,5%Glycerol,1mMMgCl2,2mMEDTA,1mMPMSF,1xCompleteMini953
EDTA-Free(Roche)].Thebeadswereresuspendedin30μlTEV-Cbuffer(20mM954
Tris-HCl pH 8, 0.5mM EDTA, 150mMNaCl, 0.1%NP-40, 5% glycerol, 1mM955
MgCl2,1mMDTT)with3μlTEVprotease(1mg/ml)andincubatedfor2hat23956
°Cunderagitation.Theeluatewastransferredtoanewtube.Beadseluatewas957
heatedat95°Cfor10minandloadedonpolyacrylamidegel[4-12%Bis-Trisgel958
(Invitrogen)]andruninMOPSSDSRunningBuffer(LifeTechnologies).Proteins959
werethentransferredtoPVDFmembraneusingTrans-Blot®Turbo™Transfer960
System(Biorad)at1Aconstant,upto25Vfor45min.Proteinsweredetected961
usingc-Mycmousemonoclonalantibody(9E10,SantaCruz,1:500),HA.11mouse962
monoclonalantibody(16B12,Biolegend,1:750)orTAPrabbitmonoclonalanti-963
body(Invitrogen,CAB1001,1:4,000).TheTAPantibodystilldetectstheCBP(Cal-964
modulinBindingProtein)moietyafterTEVcleavageoftheTAPtag.Signalwas965
detectedusingtheSuperSignalWestPicoorFemtoChemiluminescentSubstrate966
(ThermoFisher).ImageswereacquiredwithaChemidocsystem(Biorad).Forgel967
sourcedata,seeSupplementaryFigure1.968
969
Chromatinimmunoprecipitationandreal-timequantitativePCR970
Foreachmeiotictimepoint,2x108cellswereprocessedasdescribed46,withthe971
followingmodifications:lysiswasperformedinlysisbufferplus1mMPMSF,50972
μg/mlaprotininand1xCompleteMiniEDTA-Free(Roche),using0.5mmzirco-973
nium/silicabeads(BiospecProducts,Bartlesville,OK).Thelysatewasdirectlyap-974
pliedon50μlPanMouseIgGmagneticbeads.Beforeuse,magneticbeadswere975
blockedwith5μg/μlBSAfor4hat4°C.976
QuantitativePCRwasperformedfromtheimmunoprecipitatedDNAorthewhole977
cellextractusingaQuantStudio5Real-TimePCRSystemandSYBRGreenPCR978
mastermix(AppliedBiosystems,ThermoScientific)asdescribed46.Resultswere979
31
expressedas%ofDNAinthetotalinputpresentintheimmunoprecipitatedsam-980
pleandnormalizedbythenegativecontrolsiteinthemiddleofNFT1,a3.5kb981
longgene.Forthemeiotictime-courseinFigure3f,thedatawerefurthernormal-982
izedbythevalueatthe2htime-point(timeofmeiosisinductionbycopperaddi-983
tion).PrimersforGAT1,BUD23,HIS4LEU2,AxisandNFT1havebeendescribed30.984
985
ADDITIONALREFERENCES986
987
31 El-Shemerly, M., Hess, D., Pyakurel, A. K., Moselhy, S. & Ferrari, S. ATR-988
dependent pathways control hEXO1 stability in response to stalled forks. 989
Nucleic Acids Res 36, 511-519, doi:10.1093/nar/gkm1052 (2008). 990
32 Cannavo, E. & Cejka, P. Sae2 promotes dsDNA endonuclease activity within 991
Mre11-Rad50-Xrs2 to resect DNA breaks. Nature 514, 122-125, 992
doi:10.1038/nature13771 (2014). 993
33 Iaccarino, I., Marra, G., Palombo, F. & Jiricny, J. hMSH2 and hMSH6 play 994
distinct roles in mismatch binding and contribute differently to the ATPase 995
activity of hMutSalpha. EMBO J 17, 2677-2686, 996
doi:10.1093/emboj/17.9.2677 (1998). 997
34 Anand, R., Pinto, C. & Cejka, P. Methods to Study DNA End Resection I: 998
Recombinant Protein Purification. Methods Enzymol 600, 25-66, 999
doi:10.1016/bs.mie.2017.11.008 (2018). 1000
35 Palombo, F. et al. hMutSbeta, a heterodimer of hMSH2 and hMSH3, binds to 1001
insertion/deletion loops in DNA. Curr Biol 6, 1181-1184, doi:10.1016/s0960-1002
9822(02)70685-4 (1996). 1003
36 Biswas, E. E., Chen, P. H. & Biswas, S. B. Overexpression and rapid 1004
purification of biologically active yeast proliferating cell nuclear antigen. 1005
Protein Expr Purif 6, 763-770, doi:10.1006/prep.1995.0007 (1995). 1006
37 Reginato, G., Cannavo, E. & Cejka, P. Physiological protein blocks direct the 1007
Mre11-Rad50-Xrs2 and Sae2 nuclease complex to initiate DNA end 1008
resection. Genes Dev, doi:10.1101/gad.308254.117 (2018). 1009
38 Arter, M. et al. Regulated Crossing-Over Requires Inactivation of 1010
Yen1/GEN1 Resolvase during Meiotic Prophase I. Dev Cell 45, 785-800 1011
e786, doi:10.1016/j.devcel.2018.05.020 (2018). 1012
39 Thacker, D., Lam, I., Knop, M. & Keeney, S. Exploiting spore-autonomous 1013
fluorescent protein expression to quantify meiotic chromosome behaviors in 1014
Saccharomyces cerevisiae. Genetics 189, 423-439, 1015
doi:10.1534/genetics.111.131326 (2011). 1016
40 Wild, P. et al. Network Rewiring of Homologous Recombination Enzymes 1017
during Mitotic Proliferation and Meiosis. Mol Cell 75, 859-874 e854, 1018
doi:10.1016/j.molcel.2019.06.022 (2019). 1019
41 Duroc, Y. et al. Concerted action of the MutLbeta heterodimer and Mer3 1020
helicase regulates the global extent of meiotic gene conversion. Elife 6, 1021
doi:10.7554/eLife.21900 (2017). 1022
42 Murakami, H., Borde, V., Nicolas, A. & Keeney, S. Gel electrophoresis 1023
assays for analyzing DNA double-strand breaks in Saccharomyces cerevisiae 1024
32
at various spatial resolutions. Methods Mol Biol 557, 117-142, 1025
doi:10.1007/978-1-59745-527-5_9 (2009). 1026
43 Chia, M. & van Werven, F. J. Temporal Expression of a Master Regulator 1027
Drives Synchronous Sporulation in Budding Yeast. G3 (Bethesda) 6, 3553-1028
3560, doi:10.1534/g3.116.034983 (2016). 1029
44 Stahl, F. W. & Lande, R. Estimating interference and linkage map distance 1030
from two-factor tetrad data. Genetics 139, 1449-1454 (1995). 1031
45 Matos, J. et al. Dbf4-dependent CDC7 kinase links DNA replication to the 1032
segregation of homologous chromosomes in meiosis I. Cell 135, 662-678, 1033
doi:10.1016/j.cell.2008.10.026 (2008). 1034
46 Borde, V. et al. Histone H3 lysine 4 trimethylation marks meiotic 1035
recombination initiation sites. EMBO J 28, 99-111, 1036
doi:10.1038/emboj.2008.257 (2009).1037
1038
MAINTEXTSTATEMENTS1039
1040
Acknowledgements1041
ThisworkwassupportedbygrantsfromtheSwissNationalScienceFoundation1042
(31003A_17544)andERC(681-630)toP.C.,InstitutCurieandCNRStoV.B.,by1043
AgenceNationaledelaRecherche(ANR-15-CE11-0011)toV.B.andJ.-B.C.,bythe1044
NovoNordiskFoundation(NNF15OC0016662)andERC(724718)toE.R.H,and1045
theSwissNationalScienceFoundation(155823and176108)toJ.M.Wethank1046
JosefJiricny(ETHZurich)andmembersoftheCejkalaboratoryforhelpfulcom-1047
mentsonthemanuscriptandNeilHunterforcommunicatingresultspriorto1048
publication.1049
1050
Conflictofinterest1051
Theauthorsdeclarenoconflictofinterest.1052
1053
Authorcontributions1054
E.C.,A.S.,R.A.andP.C.planned,performedandanalyzedthemajorityoftheexper-1055
imentsandwrotethepaper.L.R.andA.A.performedmostoftheexperimentswith1056
yeastrecombinantproteinsandelectrophoreticmobilityshiftassays.N.W.per-1057
formed experiments to define simultaneousDNAbinding byMLH1-MLH3 and1058
MSH4-MSH5.J.H.performedexperimentswithyeastmlh1andmlh3variantsmu-1059
tated inPIP-box-likesequences,andthedatawereanalyzedtogetherwith J.M.1060
ChipexperimentsandRfc1-Mlh1andRfc1-Mlh3pulldownassayswerecarried1061
33
outbyC.A.,thedatawereanalyzedtogetherwithV.B.J-B.C.helpedpreparethe1062
MLH1-MLH3expressionconstruct anddesignedexperimentswith thePIP-box1063
peptide.X.A-G.andE.R.H.preparedtheMSH4-MSH5expressionconstruct.Allau-1064
thorscontributedtopreparethefinalversionofthemanuscript.1065
1066
DataAvailabilityStatement1067
Allrelevantdatageneratedoranalyzedduringthisstudyareincludedinthispub-1068
lishedarticleanditssupplementaryinformationfile.1069
1070
EXTENDEDDATAFIGURELEGENDS1071
1072
Extended Data Figure 1. ATP hydrolysis promotes MLH1-MLH3 to nick1073
scDNA.a,AschemeofMLH1andMLH3constructs.Themaltose-bindingprotein1074
(MBP)onMLH3wascleavedduringproteinpurification.b,RecombinantMLH1-1075
MLH3usedinthisstudy.The4-15%gradientpolyacrylamidegelwasstainedwith1076
CoomassieBlue.c,NucleaseassaywithMLH1-MLH3and2.7kbp-longsupercoiled1077
DNA(scDNA)asasubstrate.Thereactionwith5mMmanganeseacetatewasin-1078
cubatedwithoutATPat37 °C.d,Quantitationofassayssuchas inc.Averages1079
shown;errorbars,SEM;n=3.e,NucleaseassaywithMLH1-MLH3(300nM)and1080
2.7kbp-longscDNA.LinearDNAwasusedasamarker.Theassaywascarriedout1081
at37°Candcontained5mMmanganeseacetateandATP(0.5mM).TheMLH1-1082
MLH3nucleaseintroducesnicksindsDNAbutdoesnotlinearizedsDNA.f,Quan-1083
titationofnucleaseassayswithMLH1-MLH3withoutorwithATP(0.5mM),inthe1084
presenceofmanganese(5mM).Averagesshown;errorbars,SEM;n=4.g,Nucle-1085
aseassaywithMLH1-MLH3and5mMmagnesiumacetate.Thereactionbuffer1086
containedATP(0.5mM).Theassaywascarriedoutat37°C.Theheterodimer1087
exhibitsbarelydetectablenucleaseactivityinmagnesium.h,Nucleaseassaywith1088
MLH1-MLH3andvariousnucleotidecofactors(ADP,ATPandnon-hydrolysable1089
ATPanalogsATP-γ-SandAMP-PNP,all0.5mM).Theassaywascarriedoutat371090
°Cwith5mMmanganeseacetate.Thepanelshowsarepresentativeexperiment.1091
i,Quantitationofnucleaseassayssuchasinpanelh,supplementedwithvarious1092
nucleotide co-factorsand their analogs (0.5mM).Averages shown;errorbars,1093
34
SEM; n=4. j, Purified MLH1-MLH3 variants used in this study. MLH1(EA),1094
MLH1(E34A); MLH3(EA), MLH3(E28A); MLH3(3ND), MLH3(D1223N-Q1224K-1095
E1229K).The4-15%gradientpolyacrylamidegelwas stainedwithCoomassie1096
Blue. k,AlignmentofMLH1andMLH3ATPasemotifs.Conservedresiduesare1097
highlightedinred.AlaninesubstitutionsinMLH3andMLH1mutantsusedinthis1098
studyareinitalics.l,NucleaseassaywithwildtypeMLH1-MLH3andindicated1099
variantsdeficientinATPhydrolysis,withoutorwithATP(0.5mM).Theassaywas1100
carriedoutat37°C,with5mMmanganeseacetate.Thepanelshowsarepresenta-1101
tiveexperiment.m,Quantitationofnucleaseassaysasshowninpanell,without1102
orwithATP(0.5mM),witheitherwildtypeorMLH1-MLH3variantsmutatedin1103
conservedATPasedomain residues.Averages shown; errorbars, SEM;n≥4.n,1104
ElectrophoreticmobilityshiftassaywithindicatedMLH1-MLH3variants,oligo-1105
nucleotide-basedHJasthesubstrate, intheabsenceofATPandnomagnesium1106
(with3mMEDTA).Asterisk(*)indicatesthepositionoftheradioactivelabel.A1107
representative experiment is shown at the bottom, a quantitation (averages1108
shown,n=3;errorbars,SEM)atthetop.o,NucleaseassayswithwildtypeMLH1-1109
MLH3onoligonucleotide-basedDNAsubstrates(Hollidayjunction,HJandnicked1110
Hollidayjunction,nickedHJ).Theasteriskindicatesthepositionoftheradioactive1111
label.Theassaywascarriedoutat37°C,with5mMmanganeseormagnesium1112
acetate,asindicated,withATP(1mM).Theproductswereanalyzedby10%na-1113
tivepolyacrylamidegelelectrophoresis.1114
1115
ExtendedDataFigure2.HumanandyeastMutSγcomplexespreferentially1116
bindbranchedDNAintermediates.a,RecombinanthumanMSH4-MSH5used1117
in this study.b,Electrophoreticmobility shiftassayswithhumanMSH4-MSH51118
andindicatedDNAsubstrates.Asterisk(*)indicatesthepositionoftheradioac-1119
tivelabel.Theassayswerecarriedoutinabuffercontaining2mMmagnesium1120
acetatewithout ATP. c, Quantitation of DNA binding assays such as shown in1121
panelb.Averagesshown;errorbars,SEM;n=3.d,Electrophoreticmobilityshift1122
assayswith yeastMsh4-Msh5and indicatedDNA substrates.Asterisk (*) indi-1123
catesthepositionoftheradioactivelabel.Theassayswerecarriedoutinabuffer1124
containing2mMmagnesiumacetatewithoutATP.e,Quantitationofexperiments1125
such as shown in panel d. Averages shown; error bars, range; n=2. f,1126
35
QuantificationofelectrophoreticmobilityshiftassayswithyeastMsh4-Msh5and1127
indicated DNA substrates, without magnesium (with 3 mM EDTA). Averages1128
shown;errorbars,range;n=2.1129
1130
ExtendedDataFigure3.MutSγandMutLγphysically interactandmoder-1131
atelystabilizeeachotheratDNAjunctions.a,Toinvestigatetheinterplayof1132
MutLγandMutSγatDNAjunctions,weperformedelectrophoreticmobilityshift1133
assayswitheitherorbothcomplexesundermorestringentconditions(75mM1134
NaCl,2mMmagnesiumacetate), separatedon0.6%agarosegels.Under these1135
conditions,MSH4-MSH5lostthecapacitytostablybindHJs/D-Loops,butcould1136
helpstabilizetheMutSγ-MutLγcomplex.ThebindingofMutLγalonewasnotsta-1137
ble,asevidencedbyaweakprotein-DNAbandandthepresenceofsmearinthe1138
lanesindicativeofcomplexesthatdissociatedduringelectrophoresis.Theaddi-1139
tionofMutSγresultedinamoderatestabilizationoftheprotein-DNAcomplex,1140
and aminor super-shift in electrophoreticmobility of the stable protein-DNA1141
band(indicatedbytheredandbluearrows).Shownarerepresentativeexperi-1142
ments.b,Electrophoreticmobilityshiftassaysasinpanela,butwithoutmagne-1143
sium(with3mMEDTA).c,Quantitationofassayssuchasshowninpanelb.The1144
Yaxis indicatesrelativeprotein-DNAcomplexstability,obtainedupondividing1145
theprotein-DNAbandintensity(seeblueorredarrowsinpanelb)bytheinten-1146
sityoftheradioactivesignalinthelaneabovethefreesubstrateband,butbelow1147
theprotein-DNAband.Averagesshown;errorbars,SEM;n=5.d,Assaysasina,1148
with humanMutLγ and either human or yeastMutSγ. The supershiftwas ob-1149
servedonlywhenthecognatehumancomplexeswerecombined.e,Electropho-1150
reticmobilityshiftassaysasina,butwithyeastMutLγandMutSγcomplexes.f,1151
ProteininteractionassayswithimmobilizedMLH1-MLH3(bait)andMSH4-MSH51152
(prey).The10%polyacrylamidegelwasstainedwithsilver.g,Proteininteraction1153
assayswithimmobilizedhumanMSH4-MSH5oryeastMsh4-Msh5thatwereused1154
asbaits,andhumanMLH1-MLH3(prey).Theelutedproteinswereanalyzedby1155
silver staining. Although interaction between yeast Msh4-Msh5 and human1156
MLH1-MLH3wasstilldetected,itwasweakerthantheinteractionbetweenthe1157
cognateMSH4-MSH5andMLH1-MLH3complexes.h, Protein interactionassay1158
with immobilized yeast Msh4-Msh5 (bait) and yeast Mlh1-Mlh3 (prey). The1159
36
eluted proteinswere analyzed bywestern blotting. i, Electrophoreticmobility1160
shift assayswithMLH1-MLH3andMSH4-MSH5, as indicated, andoligonucleo-1161
tide-based HJ DNA substrate. 32P-labeled λDNA/HindIII digest was used as a1162
marker.TheDNA-boundMLH1-MLH3andMSH4-MSH5speciesmigratehighup1163
ontheagarosegelwheretheresolutioncapacityislimited.j,Electrophoreticmo-1164
bilityshiftassaywithyeastKu70-Ku80heterodimerandHJDNAsubstrate.Ku1165
bound thedsDNAends of the fourHJ arms, resulting in up to4 heterodimers1166
bound to the DNA substrate (lanes 5-7). Comparisonwith λ DNA/HindIII and1167
panelirevealedthattheKu-DNAcomplexmigratesmuchfasterthanDNA-bound1168
MLH1-MLH3andMSH4-MSH5.ThissuggeststhatmultipleunitsofMLH1-MLH31169
andMSH4-MSH5bindDNA.1170
1171
Extended Data Figure 4. MSH4-MSH5 promotes DNA cleavage by MLH1-1172
MLH3,butthecomplexdoesnotexhibitresolvaseactivity.a,Quantitationof1173
kineticnucleaseassayswithMLH1-MLH3(50nM)withoutorwithMSH4-MSH51174
(50nM)using5.6kbp-longscDNA.Theassayswerecarriedoutat30°Cinthe1175
presenceof5mMmanganeseacetateand2mMATP.Averagesshown;errorbars,1176
SEM;n=3.b,NucleaseassayswithMSH4-MSH5andeitherwildtypeMLH1-MLH31177
or nuclease-dead MLH1-MLH3 (D1223N-Q1224K-E1229K, 3ND). The assays1178
werecarriedoutat30°Cinthepresenceof5mMmanganeseacetateand0.5mM1179
ATP. c, Quantitation of nuclease assays with variousMLH1-MLH3 andMSH4-1180
MSH5concentrations,asindicated.Theassayswerecarriedoutat30°Cinthe1181
presenceof5mMmanganeseacetateand0.5mMATP.Averagesshown;error1182
bars,SEM,n=3.Theefficiencyofnucleasecleavagewasgenerallydependenton1183
theconcentrationsused.Whenusing50nMMLH1-MLH3,themaximalcleavage1184
efficiencywasachieved togetherwith50nMMSH4-MSH5,no further increase1185
whenusing100nMMSH4-MSH5wasobserved.Thissuggeststhatbothheterodi-1186
mersmayformastoichiometriccomplex.Viceversa,whenusing50nMMSH4-1187
MSH5,afurtherincreaseofDNAcleavagewasobservedwhenMLH1-MLH3con-1188
centrationsexceeded50nM,whichisinagreementwiththecapacityofMLH1-1189
MLH3tocleaveDNAonitsown.d,QuantitationofnucleaseassayswithMLH1-1190
MLH3andMSH4-MSH5,as indicated, in thepresenceofvariousnucleotideco-1191
factors or their analogs (2mM). The assays were carried out at 30 °C in the1192
37
presenceof5mMmanganeseacetate.Averagesshown;errorbars,SEM;n≥4.e,1193
RepresentativenucleaseassayswithMSH4-MSH5andvariantsofMLH1-MLH31194
deficientinATPhydrolysis,asindicated.Theassayswerecarriedoutat30°Cin1195
thepresenceof5mMmanganeseacetateand0.5mMATP.f,Representativenu-1196
cleaseassayswithMLH1-MLH3andvariantsofMSH4-MSH5deficientinATPhy-1197
drolysis,asindicated.Theassayswerecarriedoutat30°Cinthepresenceof51198
mMmanganeseacetateand0.5mMATP.g,RecombinantMSH4-MSH5anditsvar-1199
iants used in this study.MSH4(G685A),MSH4(GA);MSH5(G597A),MSH5(GA).1200
The 4-15% gradient polyacrylamide gel was stained with Coomassie Blue. h,1201
Quantitation of electrophoreticmobility shift assayswithMSH4-MSH5 and its1202
ATPasemotifmutant variants. Oligonucleotide-basedHJwas used as the sub-1203
strate.Asterisk(*)indicatesthepositionoftheradioactivelabel.ATPwasnotin-1204
cludedinthebindingbuffer.ThemutationsdidnotaffectthecapacityofMSH4-1205
MSH5tobindDNA.Averagesshown;errorbars,SEM;n=3.i,Nucleasereactions1206
werecarriedoutwithyeastorhumanMutSgandMutLgcomplexes,asindicated1207
(50nM),with2.7kbp-longscDNAsubstrate.WhilehumanMutSgpromotedDNA1208
cleavagebyhumanMutLg(comparelanes2and3),yeastMutSgdidnotnotably1209
promoteDNAcleavagebyhumanMutLg(comparelanes2and5),andrecipro-1210
cally,humanMutSgdidnotpromoteDNAcleavagebyyeastMutLg(comparelanes1211
7 and 8). j, Quantitation of nuclease assayswith human and yeastMutSg and1212
MutLgcomplexesasinpaneli,butwith10.3kbp-longscDNAsubstrate.k,Cleav-1213
ageofpIRbke8mutcruciformDNA(invertedrepeatsfoldingbacktoformaHol-1214
lidayjunctionstructure)byMutSgandMutLgcomplexes.Thequantitationbelow1215
thelanesrepresentsanaveragefromtwoindependentexperiments.Simultane-1216
ouscleavageofbothstrandsatthejunctionpointwouldleadtolinearDNA.No1217
linearDNAwasobservedwithMutSgandMutLg, indicatinga lackofcanonical1218
resolvaseactivity.l,Representativenucleaseassayswithindicatedproteinsand1219
oligonucleotide-basedHJDNA.Asterisk(*)indicatesthepositionoftheradioac-1220
tivelabel.NoDNAcleavagewasobserved,indicatingalackofstructure-specific1221
DNAcleavageactivityontheoligonucleotide-basedsubstrate.Theproductswere1222
analyzedby15%denaturingpolyacrylamidegelelectrophoresis.1223
1224
38
ExtendedDataFigure5.MutSbbutnotMutSastimulatesMutLgtoasimilar1225
extentasMutSg.a,RecombinantMutSb(MSH2-MSH3)usedinthisstudy.b,Re-1226
combinant MutSa (MSH2-MSH6) used in this study. c, Nuclease assays with1227
MLH1-MLH3,MSH4-MSH5,andMSH2-MSH3orMSH2-MSH6,as indicated.The1228
assayswerecarriedoutat30°Cinthepresenceof5mMmanganeseacetateand1229
0.5mMATP.Arepresentativeexperimentisshownatthebottom,aquantitation1230
(averagesshown;n=3;errorbars,SEM)atthetop.1231
1232
ExtendedDataFigure6.StimulationofthenucleaseactivityofMutSg-MutLg1233
byEXO1(D173A).a,RecombinantEXO1(D173A),usedinthisstudy.The4-15%1234
gradientpolyacrylamidegelwasstainedwithCoomassieBlue.b,Nucleaseassays1235
with MLH1-MLH3 and MSH4-MSH5, as indicated, without (left) or with1236
EXO1(D173A)(right).Theassayswerecarriedoutat30°Cinthepresenceof51237
mMmanganeseacetateand0.5mMATP.Arepresentativeexperimentisshown1238
atthebottom,aquantitation(averagesshown;n=3;errorbars,SEM)atthetop.1239
c,NucleaseassayswithMLH1-MLH3and/orEXO1(D173A),asindicated.Theas-1240
sayswerecarriedoutat30°Cinthepresenceof5mMmanganeseacetateand0.51241
mMATP.Arepresentativeexperimentisshownatthebottom,aquantitation(av-1242
eragesshown;n=4;errorbars,SEM)atthetop.EXO1(DA)doesnotpromotethe1243
nucleaseofMLH1-MLH3alone.ThelimitedDNAcleavageinlane3likelyresults1244
fromresidualnucleaseactivityofEXO1(D173A)thatbecomesapparentathigh1245
proteinconcentrations(100nM)inthepresenceofmanganese.d,Quantitationof1246
electrophoretic mobility shift assays with MLH1-MLH3, MSH4-MSH5 and1247
EXO1(D173A),asindicated.Theprotein-DNAspecieswereresolvedin1%aga-1248
rosegels.Averagesshown;errorbars,SEM;n=5.EXO1(D173A)didnotnotably1249
affect DNA binding of MLH1-MLH3 andMSH4-MSH5. e, Nuclease assays with1250
MLH1-MLH3, MSH4-MSH5 with either human EXO1(D173A) or yeast1251
Exo1(D173A),asindicated.Theassayswerecarriedoutat30°Cinthepresence1252
of 5mMmanganese acetate and 0.5mMATP. A representative experiment is1253
shownatthebottom,aquantitation(averagesshown;n=5;errorbars,SEM)at1254
thetop.f,NucleaseassayswithMLH1-MLH3,MSH2-MSH3andEXO1(D173A),as1255
indicated. The assays were carried out at 30 °C in the presence of 5 mM1256
39
manganeseacetateand0.5mMATP.Arepresentativeexperimentisshownatthe1257
bottom,aquantitation(averagesshown;n=3;errorbars,SEM)atthetop.1258
1259
Extended Data Figure 7. RFC-PCNA promote the nuclease activity of the1260
MutSg-MutLg-EXO1(DA)ensemble.a,RecombinanthumanandyeastRFCand1261
PCNAusedinthisstudy.The4-15%gradientpolyacrylamidegelwasstainedwith1262
CoomassieBlue.b,NucleaseassayswithscDNAandindicatedproteins(all50nM,1263
excepthumanPCNA,100nM)werecarriedoutwith5mMmagnesiumacetate1264
and2mMATPat37°C.Arepresentativeexperimentisshownatthebottom,a1265
quantitation(averagesshown;n≥4;errorbars,SEM)atthetop.c,Experimentsas1266
inpanelb,comparingtheefficacyofhumanandyeastRFCasapartoftheMLH31267
nucleaseensemble.Averagesshown;n=4;errorbars,SEM.d,Nucleasereactions1268
containingMLH1-MLH3(50nM),MSH4-MSH5(50nM),EXO1(D173A)(50nM)1269
and yRFC-hPCNA (50-100 nM, respectively) (column 1), withoutMSH4-MSH51270
(column2)orwithouthEXO1(D173A) (column3).Reactionswere carriedout1271
with5mMmagnesiumacetateand2mMATPat37°C.Averagesshown;error1272
bars, SEM; n≥4. e, Kinetic nuclease assays withMLH1-MLH3 (50 nM), MSH4-1273
MSH5 (50 nM), EXO1(D173A) (50 nM) and yRFC-hPCNA (50-100 nM, respec-1274
tively),asindicated.Reactionswerecarriedoutwith5mMmagnesiumacetate1275
and2mMATPat37°C.Averagesshown;errorbars,SEM;n≥5.f,Nucleaseassays1276
withMLH1-MLH3 (50 nM),MSH4-MSH5 (50 nM), EXO1(D173A) (50 nM) and1277
hRFC-hPCNA(50-100nM,respectively),as indicated,withsupercoiled(left)or1278
relaxedDNA(right).Reactionswerecarriedoutwith5mMmagnesiumacetate1279
and2mMATPat37°C.Shownisarepresentativeexperiment.RFC-PCNAdonot1280
stimulate the cleavage of relaxed DNA. g, Nuclease assays with MLH1-MLH3,1281
MSH4-hMSH5,EXO1(D173A)withoutorwithyRFC-hPCNA,asindicated.Theas-1282
sayswerecarriedoutat37°Cinthepresenceof5mMmanganeseacetateand21283
mMATP.Arepresentativeexperimentisshownatthebottom,aquantitation(av-1284
eragesshown;n=3;errorbars,SEM)atthetop.Withoutmagnesium,nostimula-1285
tion of DNA cleavage by RFC-PCNAwas observed.h, Nuclease reactions with1286
MLH1-MLH3(50nM),MSH4-MSH5(50nM),EXO1(D173A)(50nM)andyRFC-1287
hPCNA(50-100nM,respectively),asindicated.Reactionswerecarriedoutwith1288
5mMmagnesiumacetateand2mMATPat37°C.Averagesshown;errorbars,1289
40
SEM;n≥5. i,NucleaseassayswithMLH1-MLH3(50nM),MSH4-MSH5(50nM),1290
EXO1(D173A) (50nM) and yRFC-hPCNA (50-100nM, respectively) and5mM1291
magnesiumacetate,eitherwithnonucleotideco-factor(lane2),withATP(2mM,1292
lane3)orADP(2mM,lane4).ATPisstrictlyrequiredforDNAcleavagebythe1293
nucleaseensemble.j,RepresentativenucleaseassayswithMLH1-MLH3(50nM),1294
MSH4-MSH5(50nM),EXO1(D173A)(50nM)andyRFC-hPCNA(50-100nM,re-1295
spectively),lane2.Lanes3-7containinsteadMLH1-MLH3orMSH4-MSH5vari-1296
antsdeficientinATPhydrolysis,asindicated.SeeFig.1d,eforthespecificmuta-1297
tions.Reactionswerecarriedoutwith5mMmagnesiumacetateand2mMATP1298
at37°C.Averagesshown;errorbars,SEM;n=4.k,Nucleaseassayswithindicated1299
oligonucleotide-basedsubstratescarriedoutat37 °C in thepresenceof5mM1300
manganeseacetateand2mMATP.Allproteins30nM,asindicated.Asterisk(*)1301
indicatesthepositionoftheradioactivelabel.Thereactionproductswereana-1302
lyzedona15%denaturingpolyacrylamidegel.NoDNAcleavagewasobserved.1303
1304
ExtendedDataFigure8.PIPbox-likemotifsinEXO1,MLH3andMLH1facil-1305
itatethestimulatoryeffectofRFC-PCNAonthehMLH3nucleaseensemble.1306
a,TheMLH1P-MLH3Pvariant(seeFig.3b)isnotimpairedinHJ-binding.Electro-1307
phoreticmobilityshiftassaywascarriedoutwith5ng/reactiondsDNAcompeti-1308
torand3mMEDTA(nomagnesium).Asterisk(*) indicates thepositionof the1309
radioactive label. b, TheMLH1PandMLH3P variant combinations are not im-1310
pairedinnucleaseactivitywithoutorwithMSH4-MSH5andEXO1(D173A)inthe1311
absenceofRFC-PCNA.Thenucleaseassayswereperformedwith5mMmanga-1312
neseacetateand2mMATPat37°C.Averagesshown;errorbars,SEM,n=3.c,1313
Nuclease assayswithMSH4-MSH5 (50nM), EXO1(D173A) (50nM) andyRFC-1314
hPCNA(50-100nM),andarespectiveMLH1-MLH3variant,asindicated(seeFig.1315
3b).Mutations in thePIP-box likemotifreducethestimulationof thenuclease1316
ensemblebyRFC-PCNA.Theassayswerecarriedoutwith5mMmagnesiumace-1317
tate and 2 mM ATP at 37 °C. Averages shown; error bars, SEM, n=5. d, The1318
EXO1P(D173A)variantwithmutatedPIP-boxmotif(seeFig.3b)isnotaffectedin1319
its ability to promote the nuclease of MLH1-MLH3 andMSH4-MSH5 (without1320
RFC-PCNA).Theassayswerecarriedoutwith5mMmanganeseacetateand2mM1321
ATPat37°C.Averagesshown;errorbars,SEM,n=4.e,TheEXO1P(D173A)variant1322
41
with mutated PIP-box motif (see Fig. 3b), in complex with MLH1-MLH3 and1323
MSH4-MSH5impairsthestimulatoryfunctionofyRFC-hPCNA(50-100nM).The1324
assayswerecarriedoutwith5mMmagnesiumacetateand2mMATPat37°C.1325
Averagesshown;errorbars,SEM,n=5.1326
1327
ExtendedDataFigure9.RFC-PCNApromotemeioticrecombinationinyeast1328
cells.a, Sporeviabilityupon tetradmicrodissection, analyzed in thewild type1329
strain,mlh1Dandmlh3D,andinstrainscomplementedwithaconstructexpress-1330
inguntaggedMlh1P(Q572A-L575A-F578A)orMlh3P(Q293A-V296A-F300A)at1331
theendogenouschromosomallocus.Atleast156sporesfrom2biologicalrepli-1332
cateswereanalyzedforeachgenotype.b,WesternblotanalysisofMlh1Pexpres-1333
sioninyeast.TCAextractswerepreparedfromexponentiallyproliferatingSK11334
strainsexpressingMLH1,MLH1-FLAGorMLH1P-FLAGfromtheendogenousgene1335
locus. ThePIP-box-likemutationaffects thestabilityof theFLAG-taggedMlh11336
protein.Blotswereprobedwithanti-FLAGantibody(Sigma,F7425).Crm1isa1337
proteinnormalizationcontrol.Asteriskdenotesacross-reactingband.c,Western1338
blotanalysisofMlh3Pexpressioninyeast.Asinb,butwithMLH3,MLH3-FLAGor1339
MLH3P-FLAG constructs. Blots were probed with anti-FLAG antibodies: Sigma1340
F7425(leftpanel);A8592(rightpanel).Crm1isaproteinnormalizationcontrol.1341
Mlh1-FLAGandMlh3P-FLAGshowedcomparableexpressionlevels.Asterisksde-1342
notecross-reactingbands.d,ApulldownofTAP-taggedyeastRfc1-5andassoci-1343
atedproteinsfrommeioticcellextractsfrompCUP1-IME1cells5h30minafter1344
theinductionofmeiosis.ThepresenceofMlh1-HAandMlh3-MycintheTEVelu-1345
atewasanalyzedbyWesternblotting.e,Rfc1-TAPlevelsat thethree indicated1346
meioticDSBhotspotsrelativetoanegativecontrolsite(NFT1)wereassessedby1347
ChIPandqPCRinndt80∆-arrestedcellsafter7hinmeiosis.Mlh3isnotrequired1348
fortherecruitmentofRFCtothemeioticDSBhotspots.MLH3:VBD2136;mlh3∆:1349
VBD2137.Averagesshow;errorbars,SD,n=2.1350
1351
ExtendedDataFigure10.Apossiblemodelforbiasedresolutionofrecom-1352
bination intermediates by the MLH3 nuclease ensemble. Meiotic dsDNA1353
breaks (a) are resected (1) and the resultingDNAoverhang invadesmatching1354
42
DNAonahomologouschromosome(2).TheunstableD-Loopintermediates(b)1355
are stabilized by MSH4-MSH5 (3), DNA synthesis by RFC-PCNA-Pold (4) and1356
branchmigration(5),leadingtomorestablestructurestermedsingle-endinva-1357
sions(c).Thisisfollowedbyasecondendcapture(6),andmoreDNAsynthesis1358
(7)leadingtoprecursorsofdoubleHollidayjunctions(d)andlatermatureddou-1359
bleHollidayjunctions(e).Asaresultoftheprevioussteps,MSH4-MSH5andRFC-1360
PCNAmaybepresentasymmetricallyatthe(d)or(e)intermediatesatthejunc-1361
tionspointsortheirvicinity.Theasymmetricpresenceoftheco-factorsthendi-1362
rectsandstimulatesthebiasedDNAcleavage(9)of(d)or(e)structuresbyMLH1-1363
MLH3-EXO1.Uponfinalprocessing(10)andligation(11),theultimateresultisa1364
DNAcrossovercharacterizedbyreciprocalexchangeoftheDNAarmsofthere-1365
combiningchromosomes.1366
1367
1368
Figure 1
a
1 2 3 4 5 6 Lane
No
pro
tein
No ATP
MLH1-MLH3 (300 nM)
No
pro
tein
MLH1-MLH3(3ND) (300 nM)
+ ATP
++ +
+––
–
scDNA
nicked
DNA
–
Mn2+
+ + + + – 80 220 400 400 – MLH1-MLH3 (nM)
1 2 3 4 5 6 7 8 9 Lane
kDa Ma
rke
r
116
97
66
45
MLH3
MSH4MSH5
MLH1
anti-HIS
MSH4-MSH5– – –+
b Pull down by anti-HIS
200
MSH4-MSH5
MLH1-MLH3
+++
–
scDNA
nicked DNA
–+
(both 50 nM)
No
pro
tein
c
1 2 3 4 Lane
0
10
20
30
40
50
scD
NA
cle
avag
e (
%)
Mn2+
MSH4-MSH5 (50 nM)
MLH1-MLH3
MLH1(EA)-MLH3
MLH1-MLH3(EA)
MLH1(EA)-MLH3(EA)
––––
++
++
– –
–
–– ––
–––
––
All (50 nM)No
pro
tein
e
Mn2+
0
10
20
30
40
scD
NA
cle
avag
e (
%)MLH1-MLH3 (50 nM)
MSH4-MSH5
MSH4(GA)-MSH5
MSH4-MSH5(GA)
–––
–––
++
+––
–
– ––
All (50 nM)No
pro
tein
d
Mn2+
0
10
20
30
40
scD
NA
cle
avag
e (
%)
+–
Input
250150
100
75
50
kDa
Ma
rke
r
37
anti-MLH1
EXO1
MLH3
MLH1
1 2 3 4 Lane
MLH1-MLH3
+ +–
anti-MLH1
EXO1
++ +
Pulldown
+––
MLH1-MLH3 (20 nM)
MSH4-MSH5 (20 nM)
Mn2+
0 20 40 600
10
20
30
40
50
Time (min)
scD
NA
cle
avag
e (
%)
EXO1(DA)
No EXO1(DA)
f
g
Figure 2
0
5
10
15
20
scD
NA
cle
avag
e (
%)
scDNA
nicked DNAlinear DNA
Plasmid
with HJ
Plasmid
without HJ
No
pro
tein
No
pro
tein
RFC-PCNA–– –
––+ +
+++–+
1 2 3 4 5 6 7 Lane
Mg2+
d
MLH1-MLH3 +
250
150
100
75
50
kDa
37
anti-MLH1
PCNA
MLH3
MLH1
++ –
anti-MLH1
PCNA
+
+
–
–
–
–
–
–
1 2 3 Lane
0
20
40
60
scD
NA
cle
avag
e (
%)
MLH1-MLH3
No
pro
tein
MSH4-MSH5
EXO1 (DA)
RFC-PCNA++++
++
+
++
+
–––
–––
––––
a
scDNA
nicked DNA
All proteins
(50 nM) except
PCNA
(100 nM)
Mg2+
All
proteins (50 nM)
except
PCNA (100 nM)
b
No
pro
tein
No
pro
tein
scDNA
nicked DNA
Mg
2+
Mn
2+
Mn
2+
Mg
2+
MLH1-MLH3
(300 nM)
Mn
2+
Mg
2+
Mn
2+
Mg
2+
MLH1-MLH3
MSH4-MSH5
EXO1(DA)
RFC-PCNA
1 2 3 4 5 6 7 8 Lane
Mg2+
scDNA
nicked DNA
No
pro
tein
ATP
AMP-PNP
++
––
–– –
–
1 2 3 4 Lane
0
20
40
60
scD
NA
cle
avag
e (
%)
MLH1-MLH3
MSH4-MSH5EXO1(DA)
RFC-PCNA
MLH1-MLH3
MSH4-MSH5EXO1(DA)
0
10
20
30
40
50
scD
NA
cle
avag
e (
%)
Pulldown
–
+–
No
pro
tein
MLH1-MLH3
MSH4-MSH5EXO1(DA)
RFC-PCNA
ML
H1
(EA
)
No
su
bstitu
tio
n
ML
H3
(EA
)
ML
H1
(EA
)-M
LH
3(E
A)
MS
H4
(GA
)
MS
H5
(GA
)
Su
bstitu
tio
ns
All
proteins (50 nM)
except
PCNA (100 nM)
scDNA
nicked DNA
c
ef
Mg2+
scD
NA
cle
avag
e (
%)
scD
NA
cle
avag
e (
%)
All
proteins (50 nM)
except
PCNA (100 nM)
1 2 3 4 5 Lane
scDNA
nicked DNA
0
10
20
30
40
scD
NA
cle
avag
e (
%)
MLH1-MLH3
++++–+++–– RFC-PCNA
No
pro
tein
PIP
PIP
mu
t
No
ne
Mg2+
++++–MSH4-MSH5
1 2 3 4 5 Lane
Competing peptide
a b
c
e
wild
typ
e
mlh1∆ mlh3∆
Complementation
Mlh
1
Mlh
1P
Mlh
3
Mlh
3P
d
0
5
10
15
20
Gen
eti
c d
ista
nce (
cM
)
No
ne
No
ne
wild
typ
e
mlh1∆ mlh3∆
Complementation
Mlh
1
Mlh
1P
Mlh
3
Mlh
3P
No
ne
No
ne
0
2
4
6
8
Ch
rom
oso
me
no
n-d
isju
ncti
on
(%
)
MSH4-MSH5
scDNA
nicked DNA
RFC-PCNA
MLH1P-MLH3P
EXO1
(DA)
EXO1
(DA)P
No
pro
tein
––
+ + + + + +– + + +– –
Mg2+
MLH1-MLH3
EXO1
(DA)
1 2 3 4 5 6 7 Lane
0
5
10
15
20
25
scD
NA
cle
avag
e (
%)
f
0
2
4
6
8
2 4 6 8
neg. control
HIS4LEU2
BUD23
GAT1
axis
Time in meiosis (h)
Re
lati
ve
en
ric
hm
en
t o
f R
fc1
++++–EXO1 (DA)
Figure 3
All
proteins (50 nM)
except
PCNA
(100 nM)
All
proteins (30 nM)
except
RFC(50 nM)
PCNA (100 nM)
Extended Data Figure 1
1 2 3 4 5 6 7 8 9 10 Lane
No
pro
tein
nicked DNA
scDNA
No
pro
tein
No ATP + ATP All proteins (300 nM)
MLH1-MLH3
MLH1(E34A)-MLH3
MLH1-MLH3(E28A)
MLH1(E34A)-MLH3(E28A)
++
++
–– – –
– –––––
–– –++
++
– –– – –
– –––––
a
50 100 200 400 600 800 (nM)
Mg2+
scDNA
nicked DNA
MLH1-MLH3
No
pro
tein
f
200
11697
66
45
31
22
14
kDa ML
H1
-ML
H3
ML
H1
(EA
)-M
LH
3
MLH3
(or variant)MLH1
(or variant)
ML
H1
-ML
H3
(EA
)
ML
H1
(EA
)-M
LH
3(E
A)
ML
H1
-ML
H3
(3N
D)
1 2 3 4 5 6 Lane
Ma
rke
r
j
1 2 3 Lane
scD
NA
lin
ea
rD
NA
linear DNA
scDNA
scD
NA
+ M
LH
1-M
LH
3
nicked DNA
b
m
Mn2+
Nicked
HJ
0 20 40 75 150
MLH1-MLH3
HJ
(nM)
Substrate
Free label Free label
Substrate
o
Mg2+Mg2+
g
1 2 3 4 5 6 7 Lane
Mn2+
1 2 3 4 5 6 Lane
AT
P
AD
P
AM
P-P
NP
No
co
facto
r
MLH1-MLH3
(200 nM)
scDNA
nicked DNA
Mn2+No
pro
tein
AT
P-γ
-S
200
11697
66
45
31
MLH3
MLH1
Ma
rke
r
1 2 Lane
hM
utL
γ
0 100 200 300 4000
10
20
30
40
hMLH1-hMLH3 (nM)
scD
NA
cle
avag
e (
%)
No ATP
No
pro
tein
50 100 200 400 (nM)
d
No ATP
scDNA
nicked DNA
MLH1-MLH3 Mn2+
Mn2+
1 2 3 4 5 Lane
c
MLH1-MLH3 (nM)
e
MLH1-MLH3
(300 nM)Mn2+
–ATP +ATP0
20
40
60
80
scD
NA
cle
avag
e (
%)
iMLH1-MLH3
(200 nM)
Mn2+
No
co
fac
tor
+A
TP
+A
DP
+A
MP
-PN
P
+A
TP
-γ-S
0
10
20
30
40
scD
NA
cle
avag
e (
%)
h
–
No ATP + ATP
MLH1-MLH3
MLH1(EA)-MLH3
MLH1-MLH3(EA)
MLH1(EA)-
MLH3(EA)
++
++
– –– – –
– –––––
–++
++
– –– – –
– –––––
All
(300 nM)
Mn2+
0
20
40
60
scD
NA
cle
avag
e (
%)
0 20 40 75 150
MLH1-MLH3
(nM)
1 2 3 4 5 Lane
Mn2+
Substrate
Free label
HJ
0 20 40 75 150
MLH1-MLH3
(nM)
k
l
1 2 3 4 5 Lane1 2 3 4 5 Lane
0
20
40
60
80
100
DN
A b
ind
ing
(%
)
ML
H1
-ML
H3
ML
H1
(E3
4A
)-M
LH
3
No
pro
tein
ML
H1
(E3
4A
)-M
LH
3(E
28
A)
ML
H1
-ML
H3
(3N
D)
ML
H1
-ML
H3
(E2
8A
)Wells
All p
rote
ins (
25
0 n
M)
1 2 3 4 5 6HJ
No Mg2+No ATP
Protein-
bound DNA
n
Extended Data Figure 2
MSH5MSH4
a
Ma
rke
r
200
116
97
66
45
hM
utS
γ
31
kDa
No
pro
tein
25 50 75 100
MSH4-MSH5 (nM)
Free DNA
Protein-bound
DNA
Free DNA
Protein-bound
DNA
Free DNA
Protein-bound
DNA
Wells
Wells
Wells
dsDNA
HJ
D-Loop
Mg2+
b
ssDNA
Free DNA
Protein-bound
DNA
Wells
0 25 50 75 1000
20
40
60
80
100
hMSH4-hMSH5 (nM)
DN
A b
ind
ing
(%
)
dsDNA
HJ
D-Loop
ssDNA
Mg2+
0 50 100 150 2000
20
40
60
80
yMsh4-yMsh5 (nM)
DN
A b
ind
ing
(%
)
HJ
D-Loop
ssDNA
dsDNA
Y-Structure
e
0 50 100 150 2000
20
40
60
80
100HJ
D-Loop
ssDNAdsDNA
Y-Structure
yMsh4-yMsh5 (nM)
DN
A b
ind
ing
(%
)
fNo Mg2+
Mg2+
Mg2+
25 50 100 200
yMsh4-yMsh5
(nM)
No
pro
tein
ssDNA
Y-StructureFree DNA
Protein-bound
DNA
Wells
Free DNA
Wells
Free DNA
Wells
Free DNA
Wells
Free DNA
Wells
D-Loop
HJ
dsDNA
Protein-bound
DNA
Protein-bound
DNA
Protein-bound
DNA
1 2 3 4 5 Lane
Protein-bound
DNA
c
d
1 2 Lane
MSH4-MSH5 (nM)
1 2 3 4 5 Lane
a
75 0 75 100 0 100 hSγ (nM)
No Mg2+
0 75 75 0 100 100 hLγ (nM)
No
pro
tein
1 2 3 4 5 6 7 Lane
Free DNA
DNA-Lγ
DNA-Sγ-Lγ
Unstable
DNA-proteinspecies
HJ
b
No Mg2+
23.19.4 6.6 4.4
2.3 + 2.0
No
pro
tein
1 2 3 4 Lane
kbp
λ D
NA
/Hin
dII
I
Both (100 nM)
No
pro
tein
1 2 3 4 5 6 7 8 Lane
23.19.4 6.6 4.4
2.3 + 2.0λ D
NA
/Hin
dII
I
2 4 8 15 30 60
yKu70-yKu80 (nM)
kbp
Mg2+
0.6
ij
1 2 3 Lane
MLH1-MLH3 (400 nM)
MSH4-MSH5 (100 nM)
yMsh4-yMsh5 (100 nM)– +
+ + ++
–– –
MLH3
MSH4, MSH5
MLH1
anti-HIS
yMsh4, yMsh5
f
Extended Data Figure 3
yMsh4-yMsh5
yMlh1-yMlh3
1 2 3 Lane
yMsh5-his
yMlh1-hisWB:
anti-HIS
+ +–+ + –
ySγ (nM)
yLγ (nM)No
pro
tein
75 100 0 0 75 100
0 0 75 100 75 100
Free DNA
Free DNA
DNA-Lγ or
DNA-Sγ
DNA-Sγ-Lγ
HJ
dsDNA
e
HJ
75 0 75 100 0 100 hSγ (nM)
0 75 75 0 100 100 hLγ (nM)
No
pro
tein Mg2+
Free DNA
DNA-Lγ
DNA-Sγ-Lγ
Unstable
DNA-proteinspecies
Free DNA
DNA-LγDNA-Sγ-Lγ
Unstable
DNA-protein
speciesD-Loop
1 2 3 4 5 6 7 Lane
Free DNA
dsDNA
d
hSγNo
pro
tein
1 2 3 4 5 6 Lane
hLγ
ySγ
+
+ +++
++
–
– – –– –
– –
Free DNA
DNA-Lγ or
DNA-Sγ
DNA-Sγ-Lγ
HJ
All proteins
(100 nM)
Mg2+
1 2 3 4 5 6 7 Lane
MLH1-MLH3
MSH4-MSH5
200
Pull down by
anti-MLH1
116
97
66
45
kDa Ma
rke
r
MLH3
MSH4MSH5
MLH1
anti-hMLH1
+ + +– – –– + + – +–
g
h
Mg2+
0
1
2
3
DN
A b
ind
ing
sta
bilit
y (
art
ific
ial u
nit
s)
hLγ
hSγ–
Protein
concentration
(nM)75 75 100 100
+ + + +–+ +
c
Human MLH1-MLH3 = hLγ
Human MSH4-MSH5 = hSγ
Abbreviations
MSH4-MSH5
1 2 3 4 5 6 7 Lane
Input
MLH1-MLH3
–+ +
+
Free
DNA
Free
DNA
Protein-bound
DNAProtein-
bound
DNA
MSH4(G685A)-MSH5
MSH4-MSH5(G597A)
scDNA
nicked DNA
No
pro
tein
MLH1-MLH3
(50 nM)
MSH4-MSH5–––
++
+– ––
– ––
All (50 nM)
1 2 3 4 5 Lane
No
pro
tein
MLH1-MLH3
MLH1(E34A)-MLH3
MLH1-MLH3(E28A)
MLH1(E34A)-MLH3(E28A)
–++
++
++
++
––
–– – –
–– –
– – –
––––
–
–
–––
––
scDNA
nicked DNA
No
MSH4-MSH5
50 nM
MSH4-MSH5 All (50 nM)
1 2 3 4 5 6 7 8 9 Lane
a
fe
i
b
MS
H4
-MS
H5
MS
H4
(GA
)-M
SH
5
MS
H4
-MS
H5
(GA
)
Ma
rke
r
200
11697
66
45
31
22
14
kDa
MSH4 (or variant)
1 2 3 4 Lane
MSH5 (or variant)
g
0
20
40
60
80
100
DN
A b
ind
ing
(%
)
h
MSH4-MSH5
HJ
MSH4(G685A)-MSH5
MSH4-MSH5(G597A)
++
+
– –––
––
0
20
40
60
80
scD
NA
cle
avag
e (
%)
MLH1-MLH3 (nM)
MSH4-MSH5 (nM)
00
05
0
50
50
50
50
50
50
50
50
25
50
10
0
20
0 00 0
25
50
10
0
MLH1-hMLH3
titrationMSH4-MSH5
titration
Mn2+
Mn2+
Mn2+
1 2 3 4 5 6 7 8 Lane
No
pro
tein
MLH1-MLH3
MSH4-MSH5
yMlh1-yMlh3
yMsh4-yMsh5
+ + +– – – –– – ––+ + +– – – – – + +– – – + + – –
scDNA
nicked DNA
Mn2+
Extended Data Figure 4
0 10 20 30 40 50 60
0
10
20
30
40
50
60
Time (min)
scD
NA
cle
avag
e (
%)
hMLH1-hMLH3
hMLH1-hMLH3 hMSH4-hMSH5
Mn2+
c
MLH1-MLH3 (50 nM)
MSH4-MSH5 (50 nM)
Substrate
Anticipated
resolvase
product
Free label
50
100
nt
10
20
30
40
––
+++
–
1 2 3 4 Lane
l
50 nt
50 nt50 nt
50 ntMSH4-MSH5 (100 nM)
1 2 3 4 Lane
MLH1-MLH3 (100 nM)
No
pro
tein
Cruciform
DNA
+ ++ +
––
scDNA
nicked DNA
0 23 40 0 Cleavage (%)
Ma
rke
r
linear DNA
k
MLH1-MLH3, MSH4-MSH5
Mn2+Mn2+
1 2 3 Lane
No
pro
tein
MSH4-MSH5++
scDNA
nicked DNA
All proteins
(50 nM)
Mn2+
Wild
typ
e M
LH
1-M
LH
3
ML
H1
-ML
H3
(3
ND
)
j
Mn2+
0
20
40
60
80
scD
NA
cle
avag
e (
%)
yMsh4-yMsh5
yMlh1-yMlh3
MSH4-MSH5
MLH1-MLH3
–
No
pro
tein
––– + + +
+ +
– –
–––
––+
+–– +
–
–
–
MLH1-MLH3
Both
(50 nM)0
20
40
60
scD
NA
cle
avag
e (
%)
MLH1-MLH3
MSH4-MSH5
– ++– – +
– ++– – +
– ++– – +
– ++– – +
– ++– – +
No c
ofa
ctor
ATP
AD
P
AM
P-P
MP
ATP
-γ-S
d
Mn2+
c
200
1169766
45
31
22
kDa Ma
rke
r
hM
utSb
MSH3MSH2
a
Ma
rke
r
hM
utSa
200
1169766
45
31
22
kDa
MSH6
MSH2
b
scDNA
nicked DNA
MSH2-MSH6 (hMutSa)
MSH2-MSH3 (hMutSb)
MSH4-MSH5 (hMutSγ)–––
–– –
––
––
–
–
++
+
– ++
+–
––
––
MLH1-MLH3
(50 nM)
No
pro
tein
All (50 nM)
1 2 3 4 5 6 7 8 Lane
Mn2+
0
20
40
60
80
scD
NA
cle
avag
e (
%)
Extended Data Figure 5
1 2 Lane
1 2 Lane
No
pro
tein
0
20
40
60scD
NA
cle
avag
e (
%)
+0
20
40
60
80
DN
A b
ind
ing
(%
)
No Mg2+
With Mg2+, with ATP
EXO1(D173A)
MLH1-MLH3
MSH4-MSH5
All (75 nM)
+ ++ +
+ +
+ ++ ++ +
––––––
d
Extended Data Figure 6
0
10
20
30
40
50
scD
NA
cle
avag
e (
%)
EXO1(D173A)
MLH1-MLH3
Both (100 nM)
+
No
pro
tein
+++
––
––
1 2 3 4 Lane
c
scDNA
nicked DNA
Mn2+
scDNA
nicked DNA
No
EXO1(DA)
EXO1(D173A)
(50 nM)
MLH1-MLH3
MSH4-MSH4––
+ + + ++ +––
––
+––
Both
(30 nM)
1 2 3 4 5 6 7 8 Lane
200
116
97
66
45
31
22
14
kDa Ma
rke
r
EX
O1
(D1
73
A)
a b
1 2 Lane
Mn2+
+ EXO1(D173A)
yExo1(D173A)+–
––––
–
scDNA
nicked DNA
MLH1-MLH3 MSH4-MSH5
Both (30 nM)
No
pro
tein
1 2 3 4 Lane
Both
(50 nM)
Mn2+
0
10
20
30
40
50
scD
NA
cle
avag
e (
%)
e
scDNA
nicked DNA
MLH1-MLH3
(20 nM)
MSH2-MSH3 (20 nM)
EXO1(D173A) (50 nM)+ +
+–
––
0
10
20
30
40
50
scD
NA
cle
avag
e (
%)
No
pro
tein
––
1 2 3 4 Lane
f
Mn2+
0
10
20
30
40
50
scD
NA
cle
avag
e (
%)
EXO1(D173A) (50 nM)
MSH4-MSH5
(50 nM)
No
pro
tein
MLH1-MLH3 (20 nM)
RFC (50 nM)
PCNA (100 nM)
–––
––
+––
+––
++–
+–+
+++
+++
–+
+++
+ –
1 2 3 4 5 6 7 8 9 10 Lane
scDNA
nicked DNA
Extended Data Figure 7
Mn2+
Mg2+
No
pro
tein
MLH1-MLH3 (50 nM)
MSH4-MSH5 (50 nM)
EXO1(D173A) (50 nM)
RFC (50 nM)
PCNA (100 nM)
ATP
ADP––
––+
+
1 2 3 4 Lane
scDNA
nicked DNA
Mg2+
a
kDa
yRfc1
250
Ma
rke
r
Ye
ast
RF
C
Hu
ma
n R
FC
Hu
ma
n P
CN
A
150
100
70
50
40
30
20
PCNA
RFC1
RFC2-5
1 2 3 4 LaneMLH1-MLH3
MSH4-MSH5
EXO1(D173A)
Yeast RFC
Yeast PCNA
Human PCNA
No
pro
tein
–
–––––
+
+
+
++
+
++
+
++
ML
H1
-ML
H3
+ MSH4-MSH5
+ EXO1(DA)
+ RFC-PCNA
MLH1-MLH3(3ND)+++ + +
––––––
–
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Lane
–––
–––
–––
–––
–––
–––
–––
–
+
+ +–
––
– ––
– –––– +
+++
++ +
+
+ + ++ + +
– – – – – –
–––––
+– + +– –
––
–+–+ +
+–
scDNA
nicked DNA
nicked DNA
scDNA
Mg2+
All proteins (50 nM)
except
PCNA (100 nM)
0
10
20
30
40
50
60
scD
NA
cle
avag
e (
%)
0
10
20
30
40
50
60
scD
NA
cle
avag
e (
%)
MLH1-MLH3
MSH4-MSH5
EXO1(D173A)
Human PCNA
Yeast RFC
Human RFC
No
pro
tein
+ + ++ + ++ + ++ + +
++
––––
––
––
––
All proteins
(50 nM)
except
PCNA
(100 nM)
Mg2+
scDNA
nicked DNA
1 2 3 4 Lane
b c
MLH1-MLH3, MSH4-MSH5
EXO1(D173A), RFC-PCNA
No
MS
H4
-MS
H5
No
EX
O1
(DA
)
0
10
20
30
40
50
scD
NA
cle
avag
e (
%)
Mg2+
d
e
f
g
MLH1-MLH3
MSH4-MSH5
EXO1(D173A)
RFC-PCNA+++– – –
+ + + + + ++ +
+ +– –– –
– – – –
Mg2+
0
10
20
30
scD
NA
cle
avag
e (
%)
h i
No
pro
tein
No
su
bstitu
tio
n
ML
H1
(EA
)-M
LH
3
ML
H1
-ML
H3
(EA
)
ML
H1
(EA
)-M
LH
3(E
A)
MS
H4
(GA
)-M
SH
5
MS
H4
-MS
H5
(GA
)
Mu
tan
t va
ria
nts
scDNA
nicked DNA
WT MLH1-MLH3
WT MSH4-MSH5
EXO1(D173A)
RFC-PCNA
––––
++++
– – –– –+ +
+++
+ +++
++
++
++
k
No
pro
tein
No
pro
tein MLH1-MLH3
–– ++ ++ +
+
1 2 3 4 5 6 Lane
EXO1(D173A)
Mg2+
+ ++ +
+ ++ +
MSH4-MSH5
RFC-PCNA
Mg2+
EXO1(D173A)
RFC-PCNA
MSH4-MSH5
MLH1-MLH3
j
MLH1-MLH3
MSH4-MSH5
EXO1(D173A)
RFC-PCNA
––––
+ ++ ++ +
+–
––––
+ ++ ++ +
+–
Supercoiled
DNA
Relaxed
DNA
sc/rel DNA
nicked DNA
Mg2+
0 20 40 600
10
20
30
40
50
60
Time (min)
scD
NA
cle
avag
e (
%)
MLH1-MLH3
MLH1-MLH3, MSH4-MSH5
MLH1-MLH3, MSH4-MSH5, EXO1(DA)
MLH1-MLH3, MSH4-MSH5, EXO1(DA), RFC-PCNA
HJ
D-Loop
0
20
40
60
scD
NA
cle
avag
e (
%)
b
Mn2+
No
pro
tein
EXO1(DA)
EXO1(DA)P
––
++–
– ––
(30 nM)
MLH1-MLH3
MSH4-MSH5
Both (30 nM)
1 2 3 4 Lane
0
5
10
15
20
25
scD
NA
cle
avag
e (
%)
scDNA
nicked DNA
d
scD
NA
cle
avag
e (
%)
MLH1P-
MLH3
MLH1-
MLH3P
MSH4-MSH5
RFC-PCNA
– + + + + + +– – + + +– –
scDNA
nicked DNA
Mg2+
MLH1-
MLH3
1 2 3 4 5 6 7 Lane
c
scD
NA
cle
avag
e (
%)
Extended Data Figure 8a
MLH1-MLH3 (30 nM)
MLH1P-MLH3P (30 nM)No
pro
tein
++–
–
HJ
Free DNA
Protein-bound
DNA
MSH4-MSH50
10
20
30
40
scD
NA
cle
avag
e (
%)
scDNA
nicked DNA
No
pro
tein
No
pro
tein
ML
H1
P-M
LH
3
ML
H1
-ML
H3
P
ML
H1
-ML
H3
ML
H1
P-M
LH
3
ML
H1
-ML
H3
P
ML
H1
-ML
H3
ML
H1
P-M
LH
3P
(200 nM) (30 nM)
+ + + +– – – –
MLH1-MLH3 variant
1 2 3 4 5 6 7 8 9 Lane
Mn2+
1.8-fold 1.5-fold 1.4-fold
0
10
20
30
40
scD
NA
cle
avag
e (
%)
scDNA
nicked DNA
No
pro
tein
MLH1-MLH3
MSH4-MSH5Both (50 nM)
RFC-PCNA– – – ++ –
EXO1
(DA)
EXO1
(DA)P
1 2 3 4 5 6 Lane
e
(30 nM)
Both
(30 nM)
scD
NA
cle
avag
e (
%)
scD
NA
cle
avag
e (
%)
scD
NA
cle
avag
e (
%)
scD
NA
cle
avag
e (
%)
+ + + +– – – –EXO1(DA)
– + + + + + + EXO1(DA)
No
pro
tein
Mg2+
All
proteins (50 nM)
except
PCNA (100 nM)
––
Re
lati
ve
en
ric
hm
en
t o
f R
fc1
0
1
2
3
4
5
MLH3 mlh3∆
neg. control
HIS4LEU2
BUD23
GAT1
mlh1∆ mlh3∆
Complementation
wild
typ
e
No
ne
Mlh
1
Mlh
1P
No
ne
Mlh
3
Mlh
3P
Extended Data Figure 9
a
e
0
20
40
60
80
100
Sp
ore
via
bilit
y (
%)
Rfc1-TAP
Input
Rfc1-TAP
Mlh1-HA
Rfc1
Mlh1-HA
–++ +
–+++
TEV
eluate
Mlh3-Myc
Input
Rfc1-TAP
Mlh3-Myc
–++ +
–+++
TEV
eluate
Rfc1
Rfc1-TAP
ML
H1
ML
H1
-FL
AG
mlh
1P-F
LA
G
ML
H3
ML
H3
-FL
AG
mlh
3P-F
LA
G
Mlh1-FLAG
Crm1
ML
H3
ML
H3
-FL
AG
mlh
3P-F
LA
G
Mlh3-FLAG
Crm1
***
Mlh3-FLAG*
Crm1
b c
d
Extended Data Figure 10