© 2014 Nature America, Inc. All rights reserved. tional essential component: Rmi1 in yeast in Rmi1 component: essential tional substrates linked topologically (TopIII Top3 and (BLM) these are Top3 and Sgs1 helicase Bloom’sdisorder predisposition of syndrome. The yeast counterparts BLM with acts Institute Institute of Molecular Cancer Research, University of Zürich, Zürich, should Switzerland. Correspondence be addressed to N.H.T. ( Copenhagen, Copenhagen, Denmark. 1 products noncrossover generates of exclusively in kingdoms life, all conserved topoisomerase. 1A is which type , and a helicases RecQ between interplay by The unlinking DNA-strand to coupled helicase specific a of activity duplex-unwinding the through plished linked. The disjunction of the entangled can be accom outcome. noncrossover a toward events repair double-strand-breakage HR-mediated of 2.5% only to uting transformation the risk of loss of (LOH), a heterozygosity known of driver oncogenic carry however, crossovers, Mitotic products. noncrossover or over Holliday to junction an yield number equal the of symmetrical cross cleave to able are nucleases Several products. recombinant produce fashion typically noncrossover a or crossover a which either in processed, are produced, chromatids is sister two intermediate interconnects dHJ a ligation, and synthesis DNA after HR, canonical In strand. DNA damaged the of repair the for template a as forks sequence DNA replication homologous a uses stalled HR of restarting and maintenance the for and is required gaps and single-stranded breaks DNA double-strand Homologous recombination (HR) is a central pathway for the repair of TopIII thereby attaches to than obtained enzymatic Repair Richard D Bunker Nicolas Bocquet dissolution by Topoisomerase III Structural and mechanistic insight into Holliday-junction nature structural & molecular biology molecular & structural nature Received 15 March 2013; accepted 17 January 2014; published online 9 February 2014; Friedrich Friedrich Miescher Institute for Biomedical Research, Basel, Switzerland.

The DNA molecules within the dHJ intermediate are topologically are The the dHJ DNAwithin topologically intermediate molecules the

as

1 a first

0 of -gate

. Alternative pathways must . to exist therefore bias Alternative HR events an influencing 1

to 1

once DNA via a In dHJ process termed dissolution. humans, TopIII

enzyme

oligonucleotide-binding separation. the

opening 8

1

dHJs , 1 edge double-strand 9 a eQ eiae ht s uae i te cancer- the in mutated is that helicase RecQ a , . In mitotic cells, LOH appears to be rare, contrib rare, be to appears LOH cells, mitotic In . 1

, Anna H Bizard that

the

have 1 of

, Stephen C Kowalczykowski to

Topoisomerase

the relaxes

dynamics enable

converged.

gate 15 3

α breaks Department Department of Microbiology and Molecular Genetics, University of California, Davis, Davis, California, USA. , 1

) are required to process various various process to required are )

DNA 6 dissolution in , they act in concert with an addi with concert in act , they

of TopIII

3 domain

,

its

4 topology, via 2 How

. These dHJ intermediates intermediates dHJ These . III , Wassim Abdulrahman

opening 17

a a homologous , 1

through (TopIII 8

binding

and and the RMI1–RMI2 RMI1–RMI2 the and (OB

advance online publication online advance

is illustrate

and fold)

unknown. a 12–

) which of

together closing. 1

3

4 RMI recombination of , Petr Cejka . Although

RMI how

DNA 1 1 5– 2

Center Center for Healthy Aging, Department of Cellular and Molecular Medicine, University of to Here,

7

1 Our 1 with to to

binding , . TopIII α 2

- - - ­ TopIII . passes.

1

results , Nicolai B Larsen

we RMI activity but ssDNA-decatenase plasmids, modest do they possess supercoiled and TopIII intermediate, with the T strand being passed through the nick process giving rise to transesterification a 5 torus the of ing intertwined of gate a by possible made in-and-out is strand) The (T strand single DNA occurs. a of movement then relaxation duplex, linked nemically or dissolution dHJ substrate, HJ an dHJ formation. Should of the T and case C strands belong to the the same, plecto in and strand moved) one is only if hemicatenation alternatively, (or, catenation or and decatenation T is outcome When the duplexes, strand). separate to T belong strands or C strand (transfer nick the through ibly a allowing thus second strand), single-stranded DNA C (ssDNA) strandor to be transferred strand revers (cut strand DNA one in nick single-stranded transient a introduce and structure toroidal a form DNA duplexes separate two into intermediate hemicatenane the unhooks that step decatenation a by junctions it and followed is a into hemicatenane, HJs two the collapses migrating process the between supercoiling positive (TopIII topoisomerase a and other each toward Sgs1) or (BLM helicase a requires in results complex this of component instability genomic any of mutation dissolvasome. dHJ or minimal Loss the form proteins (RMI1) Rmi1 and complex in humans 3 , a

4 TopIII that step branch-migration a first involves dissolution dHJ The

a can , Ian D Hickson present

partners influences

1 1

RMI assumes do

disentangles

provide α produce i:10.1038/nsmb.277 3 0 α 1 1 . In yeast, as in humans in as yeast, In .

is a member of the type 1A family of topoisomerases, which α and RMI1

projects the , in isolation, are moderately effective in relaxing negatively

modulate

a

a it

2 crystal

toroidal

double 8

mechanistic . In this context, the C strand is cut in a reversible reversible a in cut is strand C the context, this In . to 18 19–

2 the a β behave , Mahamadou Faty ,

22 2 -sheets allowing reversible opening and clos and opening reversible allowing -sheets 2 23-residue & Nicolas H Thomä

1 , structure

2 . . Together, the RecQ helicase, Top3 (TopIII

final topoisomerase

Holliday 3 5 type .

as

hemicatenane 1

rationale

a A 25

hemicatenane

of 2 topoisomerase loop , junctions 2 2 1 4 6 , efficient decatenation requires requires decatenation efficient ,

to make the junctions migrate migrate junctions the make to . human [email protected]

into

function. for ′ -phosphotyrosine ssDNA-phosphotyrosine 1

, Simone Cavadini TopIII the

how (dHJs) intermediate 1 α

TopIII

or Top3) to relieve Top3) or relieve to dissolvase,

RMI s e l c i t r a fold. a 4

Present Present address: that

complexed 1 1 a

RMI ). stabilizes

gate, require 25–

2 2

1 rather 7 9 . This This . . . Top3

α 1

 , ­ - - )

© 2014 Nature America, Inc. All rights reserved. citrate), which belonged to belonged which citrate), group space Mg of presence the in (OB OB N-terminal + (DUF1767 RMI1 human with complex in 754–1001) residues of deletion after 1–753 TopIII TopIII the of structure the determined we questions, these address to effort an In by function activity. promoting decatenase topoisomerase modulates nases have diverged from their prokaryotic counterparts and how RMI1 tion, it is unclear to what extent the eukaryotic TopIII counterparts Top3 prokaryotic ( their from them distinguish that TopIII Human Construct RESULTS gate. topoisomerase the lining loop a through action topoisomerase the of approaches. biochemical We the dynamics show that modulates RMI1 and structural TopIII modulates elusive. remained TopIII by has dissolution dHJ that allows functionality decatenation passage efficient bestows strand and (Top3) of gating tural by basis modulates RMI1 which (Rmi1) gate topoisomerase the of tion conforma open the stabilizes Rmi1 which in model a suggested and Top3 of behavior complex open so-called increases the of Rmi1 level regulatory steady-state that the discovery this the to led of analysis activity relaxation the inhibit to serves also which Rmi1, of presence the position. in loop insertion RMI1 the TopIII and E119 RMI1 between interaction ionic while K529 and H530 E254, TopIII binding network a hydrogen-bonding of center the at is Q113 RMI1 network. bridge salt- and hydrogen-bonding corresponding the TopIII with stacks Y100 RMI1 neighboring F94; TopIII with stacking hydrophobic in in shown point TopIII the of patch hydrophobic the and segment loop ( network. hydrogen-bonding corresponding the depicting and OB from originating residues RMI1 between zipper ( sphere. a red by represented is metal catalytic The red. in shown is RMI1 (gray). domain CAP IV, and blue); noncatalytic (dark cap 5Y catalytic III, (green); domains topo-fold two the with domain gate II, (cyan); domain toprim I, domains showing complex ( TopIII human the of ( interface. binding 1 Figure s e l c i t r a  ( unit asymmetric crystallographic the in b Supplementary Figs. 1 Supplementary ) Overall architecture of the crystallized crystallized the of architecture ) Overall We set out to define how RMI1 (Rmi1) (Rmi1) RMI1 how define to out set We N and residues originating from the the from originating residues and β 10– α α α F262. ( F262. (helix gate

–RMI1 complex after 4–5 d for human TopIII human for d 4–5 after complex –RMI1 Overall architecture and TopIII and architecture Overall α e 12 gate of TopIIIof gate 12 ) Interaction between the RMI1- the between ) Interaction

design α Fig. Fig. f –RMI1 complex. We obtained crystals of the human human the of crystals Weobtained complex. –RMI1 ) Similar to to ) Similar α and yeast Top3 carry distinct sequence signatures signatures sequence distinct carry Top3 yeast and α α a

5 α of ) Domain organization organization ) Domain 19) (near proposed pivot pivot proposed (near 19) (Top3)by using function ). RMI1 P98 is engaged engaged is P98 RMI1 ). –RMI1 complex. complex. –RMI1

the 2+ (200 mM MgCl mM (200 and

TopIII α e . . ( , highlighting , highlighting c 2 ) Hydrophobic ) Hydrophobic d α 2 ) As in in ) As ). ). In informa of the absence structural 5 R338 locks locks R338 2 . Biochemical . Biochemical a 5 α . The struc The . –RMI residue residue

α c α 12 12 but but –RMI1 –RMI1 1 α

P 2 N

complex 4 ) or Ca or ) ), residues 1–219) ( 1–219) residues ), 1 2 α upeetr Fg 3 Fig. Supplementary - - 1

2 2 with one heterodimer

2+ a b (200 mM calcium calcium mM (200 N terminu RMI1 TopIII α (Top3) decate 322–447 Domain III 21–194 Domain α loop RMI1 insertion α 1 262–321 &448–518 Domain II s (residues (residues advance online publication online advance 20 I Fig. Fig. Toprim 1 ). ). - - )

is absent in yeast, is responsible for RMI2 binding in the human human the in binding RMI2 complex for responsible is yeast, in absent is (OB α cerevisiae full-length of those to equivalent are crystallization TopIII human the for boundaries and The resolution, 3.25-Å domain it statistics. has validation similar Ca of presence the in grown crystal a of structure the Wedetermined also statistics. validation excellent and ( data diffraction Å. the with 2.85 agreement of good has resolution model The maximal a to it refined and replacement lar Mg TopIIIof structure the solved We dissolution. for sufficient ( dissolution dHJ in active remained BLM, with together tested when TopIII The α Supplementary Fig. 4a Fig. Supplementary 13 -helix bundle followed by an N-terminal (OB N-terminal an by followed bundle -helix Mg TopIII 90 194 α 2+ C 19 ° β ) oligonucleotide-binding domain oligonucleotide-binding ) orCa β OB1 9 α β α 3 β 2 in complex with RMI1 (TopIII RMI1 with complex in 10 31 β β Top3–Rmi1. Human RMI1 encompasses a DUF1767 three- β IV 1 5a α 2+ RMI1 insertionloop , 15 3 I α 2 α . Deletion of human RMI1 OB RMI1 human of Deletion . 1–753 12 α β c α 261 4 11 195–261 Domain IV and RMI1 and 1 DUF1767

II DUF1767 nature structural & molecular biology molecular & structural nature 321 C terminus Insertion loop 59 ), thus suggesting that these boundaries are are boundaries these that suggesting thus ), C terminus III 94 1–219 447 134 RMI1 N terminu OB II 90 constructs used in crystallization, crystallization, in used constructs α ° N 518 –RMI1 (OB –RMI1 s IV R338 TopIII α 637 17 α c f d e E119 TopIII 2+ α L290 L297 216 13 –Mg A295 13 , 2 α , TopIII , 0 12 α 655 A118 ( α loop insertion RMI C α A118 12 loop insertion RMI Fig. 1a Fig. C4 Zn has no adverse effect effect adverse no has A521 β Linker V298 2+ β E305 15 V116 α V116 15 N N L299 –RMI1) by molecu by –RMI1) M134 D522 19 R287 L302 ) complex used for for used complex ) ) and a C-terminal C-terminal a and ) 694 α F94 Y151 Y151 E265 A525 –Ca , α K529 V95 b L177 Y100 R176 Saccharomyces Saccharomyces 19 Y100 P98 C terminus L526 ). OB ). K78 2+ α RMI1 M136 11 H530 α OB –RMI1, at –RMI1, M149 2+ L181 I R255 F262 Q113 C Table 1 Table α C -bound -bound RMI1 E254 , which which , I211 11 V263 α β β β β α β β β V179 5A 3 1 I 4 1,001 c 625 1 2 4 - )

© 2014 Nature America, Inc. All rights reserved. in gray and and gray in (PDB (PDB state apo the in structures: topoisomerase IA type available the from taken ( shell. hydration the Mg contacting directly residue sole the as serves E41 model. the into inclusion their before molecules 1.5 at 2 composite annealing Simulated spheres. gray as depicted molecules water corresponding Mg the and residues 2 Figure ( date to solved members family 1A type all in found 3 Fig. Supplementary TopIII logical TopIII two interactions with distinct makes RMI1 crystal, the In base. its as acting ( doorknocker TopIII the of shape overall The Overview OB dissolution on dHJ coefficient asdefinedinref. a r.m.s. deviations B No. atoms R No. reflections Resolution (Å) Refinement Redundancy Completeness (%) I CC R R R Resolution (Å) Cell dimensions Space group Data collection Table 1 nature structural & molecular biology molecular & structural nature sphere. a red as or toprim (topoisomerase-primase), an acidic cluster (residues (residues cleavage C-strand cluster acidic an for required cofactor and D150 that D148, E152) binds the metal-ion (topoisomerase-primase), toprim or Members of this family contain five domains: characteristic domain I, Values inparenthesesareforhighest-resolutionshell. / factors (Å work pim meas merge Bond Bond angles (°) Bond lengths (Å) Water Ion Water Ion Protein number of pairs a σ Human TopIII 1/2 , , N I (%) b / outer shell; (%) ) is the minimal functional dissolvasome complex dissolvasome functional minimal the is ) , , (%) 1EC σ R c (gray). The omit difference map shows peaks (green) for water water for (green) peaks shows map difference omit The (gray). free (Å)

Data Data collection and refinement statistics Catalytic site of human TopIII human of site Catalytic L 2

3 Homo sapiens Homo ) of 3 a ) in magenta, Top1A from magenta, ) in

the α Fig. 1 Fig.

b RI ui b snl-atce M ( EM single-particle by unit –RMI1 α α

TopIII 28 displays the canonical toroidal topoisomerase shape oeue. e diinly ofre te bio the confirmed additionally We molecules. 2+ b in in vitro , 3 ) Structural alignment of different catalytic centers centers catalytic different of alignment ) Structural 1D6 -binding site. Mg site. -binding 5 mF 6 b 19.7 19.7 / 23.1 (24.0 / 27.6) 62.78–2.85 62.78–2.85 (2.97–2.85) 3 ; domain II, which creates the ~26-Å-diameter ~26-Å-diameter the creates ; which II, domain 94.00, 94.00, 94.00, 331.72 ). ), with TopIII with ), . TOPIII TopIII o a M – 0.814; 0.814; 41,328 41,328 (4,527) –RMI 3 DF ; thus, TopIII ; thus, 5 14.1 14.1 (14.7) 16.0 16.0 (175) 14.9 (163) 16.7 16.7 (2.2) 4.2 (44.8) 100 100 (100) 2+ ) in cyan, cyan, ) in α P 2.85 2.85 Å c 0.009 6,557 α , and D150 bridges a water molecule from from molecule a water bridges D150 , and –Mg 1.04 59.0 62.3 80.3 4 omit electron density map contoured contoured map density electron omit 45 (PDB (PDB n 1 1 1 α = 4,527 2 1 complex 2+ –RMI1 complex resembles a brass brass a resembles complex –RMI1 2 –RMI1 α 2+ T. maritima 4CG α E. coli E. . . ( is shown as a red sphere, with with sphere, a red as shown is forming the ring and RMI1 RMI1 and ring the forming a b α ) ) TopIII Random half-set(Pearson)correlation Y –BLM–RMI1 (DUF1767 + (DUF1767 –BLM–RMI1 ) in green. Mg green. ) in Top1A in the apo state state Top1A apo the in 19.7 19.7 / 22.5 (25.1 / 27.8) 65.05–3.25 65.05–3.25 (3.45-3.25) (PDB (PDB α

93.38, 93.38, 93.38, 331.62 catalytic-site catalytic-site TOPIII advance online publication online advance 0.727; 0.727; 27,765 27,765 (4,404) 99.8 99.8 (99.9) 19.7 (96.2) 16.1 (78.1) 8.4 8.4 (41.8) 2GA 8.2 8.2 (2.0) 5.3(5.3) P α E. coli E. 3.25 3.25 Å 0.010 6,632 2+ 3 Fig. 1 Fig. 1.06 78.9 81.8 83.5 4 –Ca i. 1 Fig. 2 n 1 5 1 . is shown shown is I = 4,328 2 3 2+ 1 4 2 ) Top3 –RMI1

b b ) 33– and and

3 5 - .

RMI1–TopIII TopIII in involved be to and disordered partially be to shown been has loop this RMI1, isolated ( between located or sugars oligonucleotides ligands, protein binding OB when other by proteins used topoisomerase fold–containing frequently the interface of an top using gate, the II) to (domain attaches ssDNA- RMI1 toprim the from away topoisomerase. the projecting to terminus N RMI1 opposite the with site, binding diametrically site a at domain Zn detectable no We observed model. the from omitted been have and of domain V (residues 638–753) could not be unambiguously assigned to IV a are well defined in our structure, density for includes the N-terminal part which I domains Although 655–694). 655–1001), (residues motif zinc-finger predicted (residues region C-terminal the V in domain and III; domain with fold winged-helix same the shares 5 the of formation and cleavage ssDNA for needed (Y362) tyrosine lytic folds forms winged-helix which III, domain topoisomerase gate formed with two antiparallel topo-fold domains to the (−1) position (−1) to the ssDNAcut anchoring in is implicated transition-state which and H414, stabilization, phosphate pentavalent in involved is which R364, include IA enzymes type prokaryotic and eukaryotic in roles parable com serving residues Key glutamine. a by substituted is it which in in cluster acidic identified viously ( the hydration shell ( provides additional hydrogen bonds to bridge water within molecules and ecules are well mol to located form potentially hydrogen bonds. D150 water the toward point E352 of atoms oxygen carboxylic and Y362 of group hydroxyl The configuration. octahedral an in E41 of ion is by coordinated and five water oxygen one molecules carboxylic TopIII in the cluster,acidic we the assigned to density of the proximity and the observed chemistry coordination Mg the Given crystallization. catalysis support and tallization TopIII ion in to one calcium corresponding at location same the peak signal ( cluster acidic-residue toprim the at sity TopIII Mechanism ( gate topoisomerase of the center at the contacts, by crystal stabilized not Fig. 2 Fig. i. 1b Fig. a RMI1 binds the toroidal TopIII toroidal the binds RMI1 2+ anomalous signal. anomalous TopIII ′ Supplementary Fig. 4c Fig. Supplementary popoyoie sN intermediate ssDNA -phosphotyrosine α α b showed a strong positive a showed strong positive –Ca and and , α e α

Mg Ca , –Mg f and and 2+

2+ 2+ of Supplementary Figs. 1 1 Figs. Supplementary α –RMI1. We note that Mg that We note –RMI1. RMI1

2+ ssDNA complex places the RMI1 insertion loop, which is is which loop, insertion RMI1 the places complex β 4 –RMI1 (Ca –RMI1 0 1 1 and upeetr Fg 4b Fig. Supplementary ( 42 Fig. Fig. 2 Fig. 1c Fig. , 4 3

( cleavage β D148 Fig. 2 Fig. 2, 2, into the gate central of the topoisomerase a E41 ). E352, conserved ). in E352, orthologs conserved eukaryotic , d Y362 2+ ). RMI1 projects a large insertion loop, loop, insertion a large projects RMI1 ). H 2+ α ). H b 2 in TopIII in 2 binding O dependence of DNA relaxation, the the relaxation, DNA of dependence O ). 4 mF mF 1 , but neither Zn neither but , α Escherichia coli Escherichia structure via its OB its via structure o o D150 – H – and H H 2 2 O 2 2+ DF O DF 3 O α Fig. 2 Fig. 2 2a and Ca and . The arrangement of the the of arrangement The . –Ca c c b 38 , electron density electron as Mg difference electron den electron difference c ) is not part of the pre the of part not is ) ). In the structure of of structure the In ). , 3 2 2+ 9 a 9 H. sapiens T. maritim E. coli E. coli K42 and carries the cata the and carries ; domain IV, which which IV, domain ; ) and an anomalous anomalous an and ) D148 –RMI1). The metal metal The –RMI1). 2+ s e l c i t r a E41 2+ topoisomerases, topoisomerases, H414 Top3inapostate permitted crys permitted Top1A inapostate or Cs or Y362 a D150 Top1A TopIII N +

E352 α allowed allowed β E152 -barrel -barrel R364 3 2+ 7  ------;

© 2014 Nature America, Inc. All rights reserved. with TopIII with and strand ( Y151 ( L181 L177, RMI1 from chains alkyl with nature, in hydrophobic mainly is face Å 1,565 of area surface total a occupying interface, ing OB RMI1 The decatenases. 1A type prokaryotic of structural characteristic obvious features display not does and relaxase a to resemblance decatenation loop) the as to referred henceforth is (and Top3 family prokaryotic the in topoi the of ( cavity somerase central the lining both IV), (domain 502–519 to by (encoded TopIII between differences ( Å) of 2.7 deviation (r.m.s. pronounced and C 480 over Å 1.6 of deviation from relaxases Top1A the TopIII TopIII previously proposed as family, topoisomerase 1A type the within mechanism catalytic one-metal a favor therefore, findings, Our ion. metal second the to role equivalent functionally a serving thus tion, state transition during arising transesterifica pentavalent-phosphate in ion metal second TopII the of that to equivalent site the at situated is ( 1A family type topoisomerase the within conserved (K42), residue lysine a Instead, sites. binding further of indication no with Mg one find We cleavage. ssDNA metal-mediated 1A type into insight provides now structure RMI1 TopIII the ions, metal catalytic bound of signs show not did ases pombe dubliniensis Candida graminicola clavigera Grossmannia japonicus obchizosaccharomyces rouxii Zygosaccharomyces for Sequences Methods). (Online construct rlRmi1 the in cerevisiae S. the within box black-outlined in the Regions RMI1. human and strains yeast 1 ( at contoured loop, decatenation the RMI1 of modeling before calculated envelope 2 unbiased an TopIII final ( (502–519). loop decatenation the and 255 and 241 residues between loop coli E. TopIII human from diverges where sites at motifs inserts TopIII human and ( 3 Figure s e l c i t r a  the on c a

) Comparison between between Comparison ) ) Sequence alignments between different different between alignments Sequence ) Whereas previous structures of prokaryotic type 1A topoisomer 1A type prokaryotic of structures previous Whereas E. coli E. 2a

RMI 4 Top3 loop segments shown are the the are shown segments loop Top3 a and and α 4 ), whereas the similarity with the ), the whereas similarity β β . K42 is ideally positioned to stabilize the negatively charged charged negatively the stabilize to positioned ideally is K42 .

3–

domains I, III and IV share close structural similarity with with similarity structural close share IV and III I, domains resembles 10– residues 241–255 ( 241–255 residues 1 The RMI1 decatenation loop. loop. decatenation RMI1 The N -TopIII β α Fig. 3 Fig. attaches to the TopIII the to attaches , , H. sapiens H. β α 4 loop) and methionines M134, M136 and M149 ( M149 and M136 M134, methionines and 4 loop) –Ca sequence have been randomized randomized been have sequence α β Colletotrichum higginsanum Colletotrichum topB 3 strands, respectively) forming a hydrophobic zipper zipper hydrophobic a forming respectively) strands, 3 L302, L297, V298, L299 and A295 on on A295 and L299 V298, L297, L302, mF 12 loop ( loop 12 4– S. cerevisiae S. 2+ β o α a a ), however, are the absence of two loops equivalent equivalent loops of two absence however, the are ), – – –RMI1 model overlaid with with overlaid model –RMI1 5A loop), V179 ( V179 loop), 5A ). The latter has been implicated in decatenation decatenation in implicated been has latter The ). (green). RMI1 (red) (red) RMI1 (green). 35

interface a DF , , , 4

are shown. are 5 prokaryotic , , Schizosaccharomyces Schizosaccharomyces . . Thus TopIII c Colletotrichum Colletotrichum Fig. 1 Fig. electron density density electron E. coli E. , , Pichia pastoris Pichia α , E. coli E. α E. coli E. . The two unique unique two The .

E. coli E. and the prokaryotic Top3 decatenases Top3decatenases prokaryotic the and c Top3 (gray) (gray) Top3 ). These residues form a hydrophobic hydrophobic a form residues These ). α , positions; positions;

and and α

Top3 Top3 Top3 numbering; domain II) and and II) domain Top3 numbering; b type 2+ β gate and acts as the major bind major the as acts and gate α ) The The ) 5A strand), the aromatic ring of ring aromatic the strand), 5A or Ca or , , at the structural level, has close

Thermatoga maritima Thermatoga σ 1 E. E. coli , Fig. 3 Fig. . A ,

relaxase 2+ Supplementary Figs. 1 Figs. Supplementary Top3 is less decatenase ion in our structures, structures, our in ion a ). The most-apparent most-apparent The ). a c E. col E. coli α 12 loop 512–519 α i H. sapien S. pombe C. dubliniensis C. higginsanum C. graminicola G. clavigera S. japonicus P. pastori Z. rouxii S. cerevisiae Conservation H. sapien S. pombe C. dubliniensis C. higginsanum C. graminicola G. clavigera S. japonicus P. pastori Z. rouxii S. cerevisiae Conservation 241–255loop decatenation α 12, and L290 L290 and 12, 2 18 . This inter This . β α advance online publication online advance 10 19 s s s s Fig. 2 Fig. (r.m.s. (r.m.s. α I, I, 11 12 56 12 12 13 12 85 13 11 91 86 16 88 88 98 84 43 88 67 α 4 β b 6 2 4 4 4 0 2 7 4 – 2 - - - - - ) . within the TopIII the within loop insertion RMI1 the of contribution the dissect Tofunctionally Decatenation motifs. structural by RMI1 provided be principally could ability decatenation whether ( the note that in regions where the TopIII ( loop insertion RMI1 the by occupied when closed effectively is gate topoisomerase into moved be I domains to and the an between III, through opening cavity the central has ssDNA TopIII the the Because in Å. 13 opening to Å 25 central from the gate of size the restricts loop tion opening gate for the to ( V263 and TopIII against stack Y100 and P98 RMI1 prominently, TopIII the with interactions stacking hydrophobic lated 1 TopIII from emanates loop The insertion RMI1 discussion. for basis the as used be subsequently will complex TopIII insertion loop can RMI1 be fully traced in the the electron density 131, maps obtained with and the 120 residues around density electron ( in TopIII ordered less being although slightly: differs 3 strands RMI1 between inserted is which 94–134, residues loop, RMI1 The The TopIII ( R176 and K78 residues TopIII accommodates that groove Fig. 3a Supplementary Fig. 4c Fig. Supplementary b b When examining the structure of the TopIII of the structure the examining When inser RMI1 the in crystal, the trapped state In conformational the ). In the structures of the two TopIII two the of structures In the ). ). Residues from the insertion loop of RMI1 form tight interca tight form RMI1 of loop insertion the from Residues ). E. coli E.

RMI decatenation loop RMI1 α α α E. coli E. , b β –Ca E305 and further stabilize binding ( binding stabilize further and E305 gate, creating two layers of aligned and parallel rings ( rings parallel and aligned of layers two creating gate, E. coli Human TopIII RMI1 1 ). We therefore focused on the RMI1 insertion loop, examining 1 and and 1 Top3 decatenase, RMI1 donates structural motifs motifs structural donates RMI1 Top3 decatenase, insertion Fig. 1e Fig. 2+ Top3 topoisomerase 1, is expected to serve as a pivot point point pivot a as serve to expected is 1, topoisomerase –RMI1 –RMI1 data (

and β Fig. 1 Fig. 4 b 2, lines the gate of the topoisomerase ( topoisomerase the of gate the lines 2, 6 α α . ,

–RMI1 complex, we switched to yeast as a model model a as yeast to switched we complex, –RMI1

f RMI1 dHJ ). This region in TopIII in region This ). nature structural & molecular biology molecular & structural nature loop b ).

dissolution β –

4– TopIII inserts e Fig. Fig. 3 ) with an apparent discontinuity of the the of discontinuity apparent an with ) β 5A) form a bifurcated salt bridge with with bridge salt bifurcated a form 5A) α b

into ). ). Hence, the TopIII α

α β helix helix in vitro 1 (D94) and protrudes into the 1 and protrudes (D94) structure diverges structure from that of

the α K1 –RMI1 complexes, the loop loop the complexes, –RMI1 09 N1

TopIII α α 10 12. RMI1 OB RMI1 12. Fig. 1 Fig. , by structural analogy analogy structural by , T1 α D1 11 –RMI1 complex, we complex, –RMI1 14 a K1 T1 N1 P1 d

30 12 13 gate 31 ). L1 α α A1 V1 15 –Mg –Ca 16 29 Figs. 1b 1b Figs. α α P1 Q1 T1 gate. Most gate. F94, F262 F262 F94, E1 W1 24 N V1 17 21 2+ 2+ 28 -domain -domain 27 22 in trans in A1 –RMI1 –RMI1 –RMI1 –RMI1 P1 18 E1 K1 T1 A1 26 Fig. Fig. and 19 23 25 20 α - -

© 2014 Nature America, Inc. All rights reserved. tyrosine of domain III domainof tyrosine open form, with the scissile phosphate covalently attachednicked speciesto thecorresponds catalytic to Top3 being stabilized in an ssDNA-boundtype Rmi1 added. Previous work demonstrated that this accumulationaccumulation ofof nicked DNA as a function of the concentration of quantificationwild- ( separated from fully relaxed DNA. This analysis ( ofEtBr, however, allows nicked circular DNA to electrophoreticallybe ( gels(EtBr)-free bromide ethidium on detectableDNA) relaxed tially (scDNA) in the presence of Mg experiments. biochemical subsequent for basis Top3yeast ( binding in proficient is (rlRmi1) Rmi1 randomized-loop designated struct, ( loop RMI1 human the from loop insertion taken Rmi1 residues scrambled randomly 23 yeast with V145) to D87 acid (residues 58–amino the replaced then We Top3 binding. defective in and solubility limited with resulted derivatives Rmi1 tested virus constructs etch All sites. tobacco recognition flanking (TEV) using protease by excised or deleted be could relative to the 23 residues present in the human RMI1 ( residues 39 by extended segment insertion-loop corresponding the OB the overall domain organization of the metazoan orthologs (DUF1767 has+ been biochemically dissected in detail system because the yeast dissolution reaction mediated by Top3–Rmi1 data in as sticks. ( between yeast and human proteins are represented and L ( (uncropped gel image in L of WT RMI1 (lanes 3–5), L Lanes 6–9 were as in lane 2, but after addition presence (lane 2) of both human TopIII dHJ substrate incubated in the absence (lane 1) or in product of dHJ dissolution (uncropped gel image (lane 8) is a migration marker indicating the (lanes 4–6). A dHJ substrate digested with RsaI (lane 3) or increasing concentrations of GST-rlRmi1,in lane 2, but with addition of GST-WT Rmi1 (lane 2) of both Top3 and Sgs1. Lanes 3–6 are as incubated in the absence (lane 1) or presence yeast Top3, Sgs1 and Rmi1. dHJ substrate was assay of a short synthetic dHJ junction, using the Supplementary Fig. 6b overnight incubation (uncropped gel image in substrate with the different complexes after as shown in (black) and rlRmi1 (gray) on EtBr-stained gels, the nicked-DNA band intensity for WT Rmi1 Fig. 6a are shown (uncropped gel image in of WT Rmi1 (lanes 4–7) and rlRmi1 (lanes 9–12) Rmi1) versus rlRmi1. Increasing concentrations to assess the effect of yeast Rmi1 wild-type (WT ( the open Top3 and stimulates dHJ dissolution. Figure 4 nature structural & molecular biology molecular & structural nature observe any increase in nicked DNA ( f a Fig. 4 Fig. 2 ) Scheme representation of the human L ) Yeast Top3 relaxation assay using pUC19 DNA RMI1 (lanes 9–11) at increasing concentrations Supplementary Fig. 6c Isolatedwild-typeTop3 ablerelaxistonegatively supercoiled DNA the whether tested Wefirst N isrin op ( loop) insertion + 2 RMI1 constructs. Residues conserved ). Sc, supercoiled. ( e a

. , comparison of lanes 2 and 3). Running gels in the presence the in gelsRunning 3). and 2 lanes ofcomparison , The RMI1 decatenation loop stabilizes g ) Stimulation-factor quantitation of a . ( c ) Digest of single-stranded M13 Fig. 4 ). ( Supplementary Fig. Fig. 5a Supplementary b Supplementary Fig. 6d ) showed that there was a six-fold increase in the ). ( 25 d b 1 ) dHJ dissolution , RMI1 (lanes 6–8) or ) Quantitation of 2 e 6 i. 3 Fig. ) As in . When using rlRmi1, however, we did not however,did rlRmi1, weusing When . 2+ Fig. 3 Fig. Supplementary S. cerevisiae S. , producing multiple topoisomers (par c d and and , with the α c and BLM. ). We found that this Rmi1 con Rmi1 We ). this that found Fig. 4 1 RMI1 upeetr Fg 2 Fig. Supplementary

25 , ). 2 a 6 Rmi1 decatenation loop loop decatenation Rmi1 , lanes 9–12, and . S. cerevisiae ); therefore, it the forms ); therefore,

Fig. 4 advance online publication online advance f Top3 c WT Rmi1 rlRmi1 Top3 a

a Digested

, lanes 4 to 7) and M13 Relaxation Relaxation

Fig. Fig. 3 –EtBr +EtBr 2 1 Rmi1 retains M13 + + 1

TopIII Nicked Fig. 4 + 4 3 c WT Rmi1 7 6 5 4 3 2 – – – ). ), with with ), rlRmi α P98 + + + – – – –L 5

b WT Rmi1 P131 1 – – – ). 6 RMI1 - - rlRmi1 25

50 – tion and decatenation and tion catena plasmid a in defective be to construct rlRmi1 the found we Similarly, diminished. greatly was efficiency dissolution the rlRmi1, previously observed as complex, Top3–Rmi1 the by dHJs stimulated of dissolution strongly the Rmi1 Wild-type intermediate. final the of hemicatenane decatenation the to Rmi1 of contribution the examine ( dis dHJs in synthetic Sgs1 a of and solution Top3 of action concerted the examined next We The Top3 the of gate. closing and of opening dynamics the influences directly probably hence in Rmi1 Wecleavage. to single-strand inhibit loop that conclude the insertion of ability with the Rmi1 not principal while interfering conformation Top3 open a in stabilizing in nicked, is involved loop insertion Rmi1 that the demonstrate experiments These and shortening. scrambling ( rlRmi1 conclude that the effects on Top3 catalysis in observed the presence of Wetherefore Top3 stability. for deleterious not thus is itself rlRmi1 ( overnight incubation after ( h 1 after rlRmi1, or type wild of ence pres the in cleavage ssDNA M13 of levels comparable Wefragments. observed smaller into ssDNA cleave effectively to able was alone ( assays in relaxation used concentration Rmi1 highest ssDNA at viral M13 with the nM, 400 Top3–rlRmi1 bated complexes incu we this, For cleavage. ssDNA also Top3-mediated with rlRmi1 interferes whether examined next we aggregation) through ple, 100 + rlRmi1 WT Rmi1 Top3–Sgs1 To rule out nonspecific inactivation of Top3 by rlRmi1 (for exam Top3of (for by rlRmi1 inactivation To nonspecific out rule d + + –

200 insertion + – – 9 8 + – Fig. 4 Fig.

26 Dissolution 25

6 5 4 3 2 1 – – 10 – – – , TopIII 2 50 7 + – – – Hwvr we wl-ye m1 a sbtttd by substituted was Rmi1 wild-type when However, . + + 11

a loop + 100 ) are probably a direct consequence of loop-sequence of loop-sequence ) are consequence a probably direct α 20 – 12 + – –L 200 L115 20 P131 + – 2 RMI1 100 is + – Sc Topoisomers Linear Nicked Sc, relaxed Linear Nicked

200 key 8 7 2

6 Linear assay using a related biochemical reaction reaction biochemical related a using assay RsaI

9 for dHJ Fig. 4 Fig.

Fig. 4 Fig. dHJ g WT RMI1 BLM–TopIII

c e b dissolution L L ). As judged by ssDNA cleavage, cleavage, ssDNA by judged As ). 2 1 d Stimulation factor Appearance of nicked RMI1 RMI1 4 5 1 2 3 ). This substrate allowed us to to us allowed substrate This ). DNA (fold) 2 0 1 2 3 4 5 6 7 Supplementary Fig. 5b Fig. Supplementary Rmi1 concentration(nM 2 0 RMI1 concentration(nM 1 – – – – 0 rlRmi1 WT Rmi1 + 8 7 6 5 4 3 2 – – –

in vitro + 5 5 – –

40 5 + – – s e l c i t r a WT RMI1 L L 25 2 1 + – – 0 RMI1 RMI1 100 + – – 60 Fig. 4 Fig.

5 100 + – 25 – + – – 80 100 200 + 1 9 – – c ) )

). ). Top3 5 + – – 1 0 25 100 + – – ) or or ) 1 100  - - - - - © 2014 Nature America, Inc. All rights reserved. nick, and subsequent C-strand religation and DNA release must must release DNA and religation C-strand subsequent and nick, C-strand the through moved be to has T strand the completed, be to for suggested been has what to analogously width, in Å 30 to up of opening an create may ( conformation open presumed a for of decatenation, on based known and structures simulations available cycle, reaction we built a TopIII somerase. To illustrate what we propose happens in the TopIIIdonated is which ment, seg decatenation-loop RMI1 the on dependent is junction Holliday synthetic a of Dissolution topoisomerases. 1A type prokaryotic and e ae eemnd h srcue f h hmn TopIII human the (OB of structure the determined have We DISCUSSION complexes. yeast required is loop dissol decatenation for RMI1 the that and process decate the nation to contribute 99–116 residues RMI1-loop human that in (quantified RMI1 of tions and the activity could not be rescued by addition of higher concentra stimulation, wild-type to relative 50% approximately of impairment near wild-type ability to stimulate dHJ dissolution, L ( proteins human and yeast between of loop portion the decatenation the latter most being conserved also OB RMI1 nation loop. In contrast, L OB RMI1 the between interaction only the at loop retaining its point, narrow most Gly linker inserted between L115 and P131 ( residues P98 and P131, and L deletion constructs: L TopIII for dHJ dissolution, we reverted to the human required loop decatenation Rmi1 the within loop. decatenation (residues Rmi1 the as 94–134) loop insertion Rmi1 the to refer dissolution dHJ of Top3, is also required for decatenation and loop, with its ability to stabilize the open state 5c Fig. ssDNA– cognate and Top3protein), binding ( (the protein RPA replace can yeast which (SSB), protein binding Sgs1, on dependent completed. is process decatenation the and (5), released is substrate the (4), closure gate After gate. the into duplex DNA separate a second, from originating strand) (T strand a DNA move to time sufficient allowing 3, step stabilizes (red) RMI1 (3). gate the of opening after pass can T strand the which through gap a generating thus (2), C strand the in a nick introduces and (1) substrate DNA the onto TopIII example. an as taken process, dissolution dHJ the of product final the a hemicatenane, of decatenation the with TopIII the depicts model TopIIIby catalyzed dissolution 5 Figure s e l c i t r a 

In an effort to further define the region region the define further to effort an In E. coli E. N ) ) complex and have found TopIII strong between resemblance α ). These data suggest that the Rmi1 Rmi1 the that suggest data These ).

–RMI1 system and engineered two two engineered and system –RMI1 Model for the final steps of dHJ dHJ of steps final the for Model Top1A N u and the N-terminal half of the RMI1 loop ( loop RMI1 of the half N-terminal the and in eitd y TopIII by mediated tion N and TopIIIthe 46 n vitro in , 4 7 1 . This model depicts domain III and the gate in in gate the and III domain depicts model This . RMI1, with a Ser-Gly-Ser-Gly insertion between . coli E. α E. coli E. –RMI1 catalytic cycle cycle catalytic –RMI1 in trans in α . We subsequently subsequently We . 2 RMI1 RMI1 retains the TopIII (in green) loads loads green) (in α 2 Supplementary Supplementary single-strand– RMI1, with an Ala-Gly-Gly-Ser-Gly-Ser- –RMI1. The The –RMI1. TopIA α α gate and hence removing the decate the removing gate and hence i. 4e Fig. model, representing different stages into the central gate of the topoi the of gate central the into Fig. Fig. 4 Fig. 4 Fig. 6 α . For the topoisomerase cycle cycle topoisomerase the For . 5 , g –RMI1 in the human and and human the in –RMI1 ) and suggests that the gate gate the that suggests and ) ). We conclude, therefore, therefore, conclude, We ). Fig. 4e f ). Whereas L Whereas ).

, f ). L α 1 RMI1 showed an interaction with interaction 1 RMI1 links the Fig. 1e Fig. T strand 1 2 RMI1 had had RMI1 advance online publication online advance α α –RMI1 –RMI1 –RMI1 , f ), the the ), C strand α - - - - ­

C-strand binding o efcie eaeain ( decatenation effective for ( Experimentally we P98–K109 find that around human RMI1 residuesresidues 96–116 areinvolving required loop decatenation RMI1 the of beginning the to proximity in is gate the of region pivot ( form open the of model the In The Top3 enzymes prokaryotic in complex Top3–Rmi1 the decatenation experiments by accommodated is DNA stranded (approximately ten-fold less efficient than for ssDNA) by which TopIII double-the within DNA double-stranded of accommodation disfavor to expected are ( gate the within points ment OB RMI1 the of presence the region, this in loop the of plasticity necessary the Despite DNA. with crowding) interactions steric (through repulsive nor residues) hydrophobic or charged (through attractive crucial DNA,with neither contact exerts direct in be to likely is which loop, the of part this that conclude we gave to rise only minor dHJ defects very dissolution tion or substitution of the tip of the loop in the human L ( site this at exists variability conformation siae the within 39 residues the additional where is the locus also presumed ssDNA T-strand–binding site This (residues 116–131). site the Mg Ca Mg the comparing when observed loop decatenation the of plasticity the with line in is This Å). 7 approximately being ssDNA of diameter the (with point widest its at Å 11 around only measures gate and loop the topoisomerase the decatenation between the cavity is thus required to accommodate the ssDNA upon gate closure, because occur ( 2 2+ Hemicatenane Rmi1 loop are inserted, thus further supporting the notion that that notion the supporting further thus inserted, are loop Rmi1

RMI -bound crystal forms ( crystal -bound substrate T Fig. 2+ -bound structure, the loop segment is disordered around this

decatenation 5 C ). A certain degree of plasticity within the decatenation loop Final dissolutio 1 α 2

gate. This is in agreement with the low efficiency efficiency low the with agreement in is This gate. by WTRMI1 the openstat Promotion of nature structural & molecular biology molecular & structural nature 5

loop 2 strand release Gate openingand 3 5 rlRMI 4 Fig. Fig. 4 and contrasts with the behavior seen with 8 n 5 i. 4e Fig. .

e modulates Fig. 2 Fig. 1 3 Fig. Fig. b and T-strand passage Gate openingan , b g 5 ) impart steric constraints that that constraints steric impart ) ). In addition, RMI1 residues residues RMI1 addition, In ). , step 3), we observed that the the that observed we 3), step , Supplementary Fig. 4d Supplementary

C-strand religatio Gate closureand the 3 Fig. 3 Fig. N

domain and its attach its and domain gate d c

). Because trunca Because ). dynamic in in vitro n 2 4 Rmi1 Rmi1 mutant ( S. S. cerevi Fig. 1 Fig. Fig. Fig. 4 2+ , - and and - e ). ). In e e ). ). ), ), - - -

© 2014 Nature America, Inc. All rights reserved. of these macromolecular machines macromolecular these of put when roles in context the and regulatory catalytic crucial assume nonetheless can isolation, in unstructured are which loop, RMI1 the toposomes as to referred (also assembles large their within function topoisomerase modulate proteins how binding TopIII of structure The complexes. romolecular mac of context the in function proteins of majority the cell, the In Conclusion dissolvasome. RecQ-containing the within are integrates turally studies Further TopIII relaxation. how understand to required than rather hemicatenane dHJ a of decatenation the allowing specificity, and dynamics its modulates but instead action of change mode the topoisomerase fundamentally not does RMI1 that find We now role. scaffolding noncatalytic and In RecQ helicase. both a yeast, the serves catalytic Sgs1 RecQ helicase TopIII of minimally, consisting, some of on a the presence dissolva is complete dependent dHJ dissolution TopIII times. gate-opening extended require not does hence and distances shorter over occurs probably strands of movement DNAprocess), in an C (working intramolecular strand the does as duplex same the to belongs strand T the which in mode, process’) ‘intermolecular an (as tion ment of these strands in and out of the gate, thereby driving decatena move the permits This dissolution). dHJ in example, (for duplexes different from originating ssDNA T distant strands to accommodate time sufficient with topoisomerase the provide will state open lized stabi its in gate the structural which in model a The support data biochemical and enzymes). IA type prokaryotic in found loops lent equiva the functionally from (or the loop RMI1 possibly originating elements structural through stabilized is open, once gate, the RMI1, somerase gate opens. Although topoi C-strand cleavage the is not strand dependent onC single-stranded the of scission and binding the resolution. dHJ and decatenation effective for required is RMI1 in loop. loop this decatenation Moreover, prokaryotic the of loci two the to segment, proximity loop in a inserts RMI1 that find now We decatenation. loop-mediated against argument an as used been Top3(TopIII being also activity impaired relaxation with albeit activity, decatenation of in loop this of Deletion decatenases in prokaryotic the is exclusively found gate, the which lining loop decatenation of the or absence striking presence the is most difference The differences. structural minimal only with cores IA type example, (for nation Prokaryotic itself. topoisomerases implicated topoisomerase in relaxation and those involved in decate the of manipulation addition RMI1 allowing direct by probing of the effect of RMI1 on decatenation without stimulated is decatenation that TopIII the of Analysis Toward TopIII the of TopIII and loop RMI1 the between among conserved reasonably ( also orthologs are K109 to P96 from ranging nature structural & molecular biology molecular & structural nature We propose a model based We on ( our data new based a model propose a

–RMI a 4

5 unified Fig. Fig. 3 . The structural equivalent of this loop in the eukaryotic eukaryotic the in loop this of equivalent structural The . α α 1 opening-closing equilibrium. opening-closing ) decatenases is absent, a feature that has previously previously has that feature a absent, is decatenases ) function c

). ). We that therefore this region contact hypothesize mechanism E. coli E. α E. coli E.

in –RMI1 system has the great advantage advantage great the has system –RMI1

Top3) share almost identical enzymatic enzymatic identical almost Top3)share the

Top3 results in almost complete loss loss complete almost in Top3 results for

dissolvasome

decatenation 5 α α 0 2 . –RMI1 functionally and struc and functionally –RMI1 contributes to the modulation modulation the to contributes 5 α . In contrast, in its relaxation relaxation its in contrast, In . (Top3), RMI1 (Rmi1) and a and (Rmi1) (Top3),RMI1 Fig. Fig.

4 in trans in 9 α ). Epitopes such as as such Epitopes ). advance online publication online advance –RMI1 illustrates illustrates –RMI1 5 ), whereby after after ), whereby n trans in , positioned positioned , , thus thus , 3 5 ------.

11. 10. 9. 8. 7. 6. 5. 4. 3. 2. 1. prior to publication. Part of this work was performed at PXII beamline of the Swiss thank A. andCosta colleagues at the London Research Institute for sharing results H. Gut and M. Renatus for help anddiscussion. Wefruitful would also like to K. Shimada, J. Keusch, E. Fischer, A. D.Scrima, Hess, D.K. Boehm, Klein, for Cellular Imaging and NanoAnalytics (c-CINA). We thank S. Gasser, U. Rass, We are grateful to H. Stahlberg and team for use of the EM facilities at the Center the US National Institutes of Health (GM-41347 and GM-62653) (to S.C.K.). Nordea Foundation (to I.D.H.), The Villum Kann Rasmussen Fund (to I.D.H.) and 693-2009) (to N.B.) and European Research Council (to N.H.T. and I.D.H.), the European Molecular OrganizationBiology long-term Fellowship ALTF(EMBO Marie Curie Fellowship 253555-TOPO) (FP7-PEOPLE-2009-IEF (to N.B.), grants from the Krebsforschung Schweiz (to (KFS-2986-08-2012) N.H.T.), This work was supported by the Novartis Research Foundation (to N.H.T.), version of the pape Note: Any Supplementary Information and Source Data files are available in the TOPIII the reprints/index.htm at online available is information permissions and Reprints The authors declare no competing interests.financial other coauthors. N.H.T. the project. supervised help of P.C., S.C.K. and I.D.H.; N.B. and N.H.T. wrote the paper with the help of the constructions and drop-assay tests. N.B. and N.H.T. the designed project with the three-dimensional model reconstruction. M.F. and N.B.L. performed yeast-strain and Sgs1 proteins as well as SSB protein. S.C. out carried EM single-particle and human RMI1-mutant proteins for dissolution assays. P.C. provided yeast Top3 assay. N.B. performed the construct design, expression and purification of contributions by W.A. and N.B., and dissolution assays. W.A. did the M13-digest A.H.B. performed the relaxation, catenation and decatenation assays, with N.H.T. outcarried stages final of model refinement and map improvements. structure determination crystallization, and model refinement. R.D.B. and N.B. outcarried molecular biology, protein expression and purification, Light Source at the Paul Scherrer Institute, Villigen, Switzerland. ited in the under accession codes codes accession under Bank Data Protein for the TOPIII the in ited codes. Accession the pape the of in version available are references associated any and Methods M A C AUTH ckn O

ethods

MP u L & iko, .. h Boms ydoe eiae upess rsig over crossing suppresses helicase syndrome Bloom’s The I.D. Hickson, & L. Wu, J.R. LaRocque, S. Thiagalingam, W.K. Cavenee, formation. crossover and resolution junction Holliday Recombination: W.D.Heyer, control Regulatory S.C. West, & J.M. Skehel, S., Maslen, M.G., Blanco, J., Matos, eukaryotic the in junctions Holliday J.A. Solinger, & K.T. Ehmsen, W.D., Heyer, double-strand- F.W. The Stahl, & R.J. Rothstein, T.L., Orr-Weaver, J.W., Szostak, fungi. in conversion for mechanism A Holliday,R. genomic maintains recombination homologous Mitotic M. restart. Jasin, & M.E. Moynahan, fork replication mammalian of Pathways T. Helleday, & E. Petermann, during homologous recombination. homologous during USA Sci. cells. Acad. Natl. mammalian (BLM)-deficient helicase syndrome Bloom and wild-type cancers. colorectal retinoblastoma. in Biol. Curr. Cell mitosis. and during intermediates recombination DNA of resolution the of sight? in resolution nucleus: recombination. for model repair break (2007). (2010). tumorigenesis. suppresses and stability Biol. Cell Mol. Rev. Nat. o O E

147 TIN wl R R C

, 158–172 (2011). 158–172 , e O G

α d 14 NTRIBUTI FINANCIAL INT FINANCIAL –RMI1 structure bound to a Ca a to bound structure –RMI1 g , R56–R58 (2004). R56–R58 , α m r l et al. et . . –RMI1 structure bound to bound a Mg structure –RMI1 t al. et e et al. et nts Nature oriae o te oes ae en depos been have models the of Coordinates Proc. Natl. Acad. Sci. USA Sci. Acad. Natl. Proc. r Expression of recessive alleles by chromosomal mechanisms chromosomal by alleles recessive of Expression .

108 nehmlg eobnto ad os f eeoyoiy in heterozygosity of loss and recombination Interhomolog Mechanisms underlying losses of heterozygosity in human in heterozygosity of losses underlying Mechanisms O

11

NS , 11971–11976 (2011). 11971–11976 , 305 , 683–687 (2010). 683–687 , Trends Biochem. Sci. TrendsBiochem. , 779–784 (1983). 779–784 , E R E Nature STS Cell a. e. o. el Biol. Cell Mol. Rev. Nat.

426 33 , 25–35 (1983). 25–35 ,

, 870–874 (2003). 870–874 , 98 , 2698–2702 (2001). 2698–2702 ,

28 2+ , 548–557 (2003). 548–557 , ion. Genet. Res. Genet. 2+ http://www.nature.c s e l c i t r a ion and

11 89 4CH 196–207 , , 285–307 ,

onlin 4CG

online T

Proc. for om/

Y  e -

© 2014 Nature America, Inc. All rights reserved. 29. 28. 27. 26. 25. 24. 23. 22. 21. 20. 19. 18. 17. 16. 15. 14. 13. 12. s e l c i t r a 

Tse, Y.C., Kirkegaard, K. & Wang, J.C. Covalent bonds between protein and DNA: and protein between bonds Covalent J.C. Wang, & K. Y.C.,Kirkegaard, Tse, evolutionary and mechanisms, molecular Structure, J.M. Berger, & K.D. Corbett, the by III topoisomerase DNA of activation and Binding S.J. Brill, & C.F. Chen, Rmi1 S.C. Kowalczykowski, & I.D. Hickson, C.Z., Bachrati, J.L., P.,Plank, Cejka, of Decatenation S.C. Kowalczykowski, & C.C. Dombrowski, J.L., Plank, P., Cejka, Mankouri, H.W. & Hickson, I.D. The RecQ helicase-topoisomerase III-Rmi1 complex: M. Chang, Sgs1-Top3 the of member a Rmi1, T. Enomoto, & A. Ui, M., Seki, M.S., Lai, Wu, L. dissolvasome junction Holliday double A P. Sung, & W. Bussen, S., Raynard, T.R.Singh, Rmi1/Nce4 Yeast S.J. Brill, & F.S.,C.E. Y.Q.,Nallaseth, Slagle, Lan, J.R., Mullen, D. Xu, full-length The S.C. Kowalczykowski, & P. Cejka, Harmon, F.G., DiGate, R.J. & Kowalczykowski, S.C. III RecQ helicase and topoisomerase topoisomerase DNA the Mapping S.J. Brill, & V. Kaliraman, W.M., Fricke, sgs1 yeast between Interaction J.C. Wang, & M.F. Noirot-Gros, R.J., Bennett, yeast The R. Rothstein, & L. Arthur, C., Bendixen, J.P., McDonald, S., Gangloff, omto o popoyoie ikg bten eti DA oosmrss and topoisomerases DNA certain DNA. between linkage phosphotyrosine of formation (2004). 95–118 topoisomerases. DNA in relationships by subunit. Rmi1 dissolution during junctions Holliday double Sgs1–Top3. of decatenation stimulates chromosomes. disentangling the by DNA (2007). ‘dissolvasome’? structure-specific DNA a complex. III helicase/Topo RecQ (2005). the of member 8 III topoisomerase BLM, comprising dissolvasome. Dev. junction Holliday helicase-double Bloom the of component 25 complex. Sgs1-Top3 the of subunit a as stability genome controls stability. genome maintain Chem. Biol. J. junctions. Holliday unwinds preferentially that helicase DNA vigorous a is protein ope i bdig es, otiue t sse crmtd cohesion. chromatid sister to contributes yeast, budding in complex (2006). intermediates. recombination (2006). 13861–13864 for mechanism conserved a function: recombination. DNA of passage control strand DNA novel a comprise III helicase. DNA (2001). Sgs1 the of domain binding III. topoisomerase (2000). DNA and helicase gyrase. reverse eukaryotic type I topoisomerase Top3 interacts with Sgs1, a DNA helicase homolog: a potential , 685–690 (2007). 685–690 , , 4476–4487 (2005). 4476–4487 ,

22 J. Biol. Chem. Biol. J. et al. et al. et , 2856–2868 (2008). 2856–2868 , et al. et BLAP75/RMI1 promotes the BLM-dependent dissolution of homologous RMI, a new OB-fold complex essential for to protein syndrome Bloom for essential complex OB-fold new a RMI, t al. et Nat. Struct. Mol. Biol. Mol. Struct. Nat. . cerevisiae S.

J. Biol. Chem. Biol. J. 285 BLAP18/RMI2, a novel OB-fold-containing protein, is an essential an is protein, OB-fold-containing novel a BLAP18/RMI2, M1NE, sprso o gnm isaiiy ecds a encodes instability, genome of suppressor a RMI1/NCE4, , 8290–8301 (2010). 8290–8301 ,

255 , 5560–5565 (1980). 5560–5565 , g1Tp-m1 n RA ope: mcaim for mechanism a complex: RPA and Sgs1-Top3-Rmi1 Mol. Cell. Biol. Cell. Mol. Dev. Genes Mol. Cell Mol.

282 Mol. Cell Mol. rc Nt. cd Si USA Sci. Acad. Natl. Proc. , 28971–28979 (2007). 28971–28979 ,

17

α

, 1377–1382 (2010). 1377–1382 , 47 22 ad BLAP75. and , nu Rv Bohs Boo. Struct. Biomol. Biophys. Rev. Annu.

3

, 886–896 (2012). 886–896 , , 2843–2855 (2008). 2843–2855 , 14 . il Chem. Biol. J. , 611–620 (1999). 611–620 , rns ice. Sci. Biochem. Trends , 8391–8398 (1994). 8391–8398 , . il Chem. Biol. J. acaoye cerevisiae Saccharomyces MO J. EMBO

. il Chem. Biol. J. 275

103 276 26898–26905 , 24

Mol. Cell. Biol. Cell. Mol. 4068–4073 , 2024–2033 , 32 8848–8855 , advance online publication online advance MO Rep. EMBO 538–546 ,

Genes Sgs1 281

33 , ,

42. 41. 40. 39. 38. 37. 36. 35. 34. 33. 32. 31. 30. 50. 49. 48. 47. 46. 45. 44. 43.

hnea A, iae RJ & odao, . rsa srcue f cmlx f a of complex a of structure Crystal A. Mondragon, & R.J. DiGate, A., Changela, H. Goulaouic, a proteins: fold Oligonucleotide/oligosaccharide-binding L. Zou, & R.L. Flynn, motif. binding DNA helix-turn-helix The B.W. Matthews, & R.G. Brennan, complex: CAP-DNA a T.A.of Steitz, structure & Crystal G.C. Shields, S.C., Schultz, Duguet, M., Serre, M.C. & Bouthier de La Tour, C. A universal type IA topoisomerase in domain catalytic conserved a Toprim: E.V. Koonin, & D.D. Leipe, L., Aravind, of structure The R. DiGate, & A. Mondragón, Crystal P. Reinemer, & D. Schomburg, B., Wieland, A., Harrenga, G., Hansen, 67K the of structure Three-dimensional A. Mondragon, & J.C. Wang, C.D., Lima, Wang, F. Hoadley,K.A. G.W.Brown, topoisomerase & Human I.D. Hickson, J., Ou, C.Z., Yang,Bachrati, J., oao-aca VM e a. pta ad eprl raiain f multi-protein of organization temporal and Spatial al. et V.M. Bolanos-Garcia, puzzle? simple a topoisomerases: IA Type Tour, C.B. la de & T. Viard, topoisomerase- IA type of mechanism The R.J. DiGate, & A. Mondragon, Z., Li, B. Xiong, in changes Conformational A. Mondragon, & C.D. Lima, H., Feinberg, a of Identification R.J. DiGate, & K.J. Marians, H., Hiasa, A., Mondragon, Z., Li, novel A J.M. Berger, & N. Osheroff, J.E., of Deweese, A.B., Burgin, B.H., studies Schmidt, Structural A. Mondragon, & R.J. DiGate, A., Changela, growing family of genome guardians. genome of family growing Chem. degrees. 90 by bent is DNA the fold. and nucleases family OLD primases, proteins. RecR DnaG-type topoisomerases, II and IA type III. 358 from I topoisomerase length full of structure (1994). of fragment N-terminal complex. BLM the of dissolvasome. syndrome Bloom Philos. Trans. A Math. Phys. Eng. Sci. Eng. Phys. Math. Trans.A Philos. systems. cell-regulatory information-rich in control sensitive achieving assemblies: 89 ye A N tpioeae ih snl-tadd N molecule. DNA (2001). single-stranded 1077–1081 a with topoisomerase DNA IA type III (2010). Chem. Biol. J. III mediated DNA topological transformations. topological DNA mediated simulation. dynamics molecular and I. topoisomerase intermediates. replication of decatenation for essential domain unique Nature and unified two-metal mechanism for topoisomerase-DNA DNA cleavage by type II and IA IA topoisomerases. type novel a intermediate. reveal conformational complexes III-DNA topoisomerase α α , 456–467 (2007). 456–467 , . RMI1. and BLM by stimulated is that decatenase DNA single-stranded a is Structure , 1328–1340 (2006). 1328–1340 , Nucleic Acids Res. Acids Nucleic J. Mol. Biol. Mol. J.

264 465 et al. et al. et , 1903–1906 (1989). 1903–1906 , , 641–644 (2010). 641–644 ,

Crystal structures of RMI1 and RMI2, two OB-fold regulatory subunits 7 et al. et et al. et

The type IA topoisomerase catalytic cycle: A normal mode analysis mode normal A cycle: catalytic topoisomerase IA type The , 1373–1383 (1999). 1373–1383 , 285 Nucleic Acids Res. Acids Nucleic Nat. Struct. Biol. Struct. Nat.

359 Purification and characterization of human DNA topoisomerase DNA human of characterization and Purification , 21426–21436 (2010). 21426–21436 , Structure and cellular roles of the RMI core complex from the from complex core RMI the of roles cellular and Structure

Structure nature structural & molecular biology molecular & structural nature , 805–812 (2006). 805–812 ,

27 . coli E. , 2443–2450 (1999). 2443–2450 , J. Mol. Biol. Mol. J. shrci coli Escherichia

Structure 18 N tpioeae I. topoisomerase DNA Science

, 1159–1170 (2010). 1159–1170 , 6 26 , 918–922 (1999). 918–922 , Proteins Crit. Rev. Biochem. Mol. Biol. Mol. Biochem. Rev. Crit. , 4205–4213 (1998). 4205–4213 ,

370

253

18 Mol. Microbiol. Mol. 368 , 3023–3039 (2012). 3023–3039 , Mol. Cell Mol. , 1149–1158 (2010). 1149–1158 ,

, 1001–1007 (1991). 1001–1007 , 71 Escherichia coli Escherichia hrooa maritima Thermotoga , 105–118 (2007). 105–118 , N tpioeae III-catalysed topoisomerase DNA , 1984–1994 (2008). 1984–1994 ,

7 , 301–307 (2001). 301–307 ,

35 Nature , 888–895 (2000). 888–895 , DNA topoisomerase DNA

367

. 45 . o. Biol. Mol. J. Nature E. coli E. 138–146 , , 266–275 , Biochimie . coli E. J. Biol. J.

DNA 411 ,

© 2014 Nature America, Inc. All rights reserved. at high resolution. Final data-processing statistics are given in in given are statistics data-processing Final resolution. high at signal useful substantial is there that suggests This 0.4. than greaterremained CC the cutoff anisotropic the by defined range resolution the However,over sets. along the a ( difference the ellipsoid, directions for anisotropy in tetragonal When crystals. represented as a thermal the along Å 3.16 is P factor amplitudes by TRUNCATE AIMLESS with merging for space-group prediction and format conversion before additional scaling and XDS withprocessed wereData collected.were Mg of presence the in grown crystals from sets Data Å. 1.0 wavelengthof a Switzerland)with Pilatus 6M detector (Dectris) at beamline PXII of theprocessing. and Swiss collection Data Light Source (Villigen, glycol before flash cooling in liquid nitrogen. was performed by supplementation of the growth condition with 35% ethylene rods growing to maximal dimensions within 4–5 d. Cryopreservation of crystals 200 mM calcium citrate. Crystals in either Mg grown in the presence of Ca presencethe ofMg PEG 2000, 100 mM Tris-HCl, pH 7.0, and 200 mM MgCl Complex crystallization. pooled and subjected to crystallization experiments. NaCl, and 0.2 mM TCEP). Fractions containing the heterodimeric complex were Superdexona columnHealthcare)mM 200(GE 300 7.4, HEPES,pH mM (50 a final step, the purified protein was subjected to size-exclusion chromatography by 50 mM Tris, pH 8.0, 0.2 mM TCEP) to remove traces of endogenous step.DNA. This As was followed by POROSQ anion-exchange chromatography (buffered trifugation. Streptactin affinity purification was performed as ultracen a by firstclarified subsequentlywere lysates andsonication,strep-affinity by performed β mM 1 NaCl, mM 200 8.0, Tris-HCl,pH mM 50 in Lysis infection. after h 48 10 × 4 ofconcentration a at infected were Cells insect crystallization. cells with a pFastBac dual transfer vector to (Bac-to-Bac system, Invitrogen). prior (residues 1–753)–RMI1 (residues 1–219) complex was coexpressed expression in High Five Protein ONLINE doi: Buster ( produced after iterative model building with COOT and refinement with auto rection of several side chain rotamers. The final TopIII bythe Mg revealed of the refined models and the interpretability of the election density maps. This high-resolutionofweak the datarefinementin markedly improved qualitythe the CC (2.97–2.85 Å) shell in resolutionthe directionoutermost the of resolution,the this At found. was Å 2.85 of limit strategy of Karplus and Diederichs initial 3.3-Å model stalled, the data were reassessed with the ‘paired-refinement’ the outermost resolution shell and refined with PHENIX.REFINE CHAINSAW and a search model based on hRMI1 (PDB withan was solved by molecular replacement (MR) in space Structuregroup determination and refinement. outliers, respectively. TOPIII and 96.0% of residues were in the favored regions of the Ramachandran plot for -mercaptoethanol, 1 mM PMSF and protease inhibitor cocktail (Sigma) was (Sigma) inhibitorcocktailprotease and PMSF mM -mercaptoethanol,1 4 ∆ 1 The initial TopIII group, space true the in Analyzed anisotropic. strongly are sets data Both B = 24.0 Å 24.0 = B 2 10.1038/nsmb.2775 1 2, the resolution at which which at resolution the 2, 1/2 α http://www.gl E. coliE. –Mg m was 0.97, with an overall CC c 2+ axis. The anisotropy resulted in high merging

F METHODS ion and for an additional 39 water molecules and enabled the cor o – DF– 2+ Top1A 2 1/2 and resolution limits of 3.64 Å along the the along Å 3.64 of limits resolution and –RMI1 and TOPIII 2+ creain ofiin bten adm af aa sets data half random between coefficient (correlation (TOPIII c 2+ difference-map for peaks waterthe molecules coordinated α ab orCa 3 –Mg obalphasing.com/buster 3 plane and 2.63 Å along the the along Å 2.63 and plane search modelproposed by BALBES and prepared with 5 The complex (8 mg/ml) was crystallized in 8–12% (w/v) ∆ 2 2+ . Averaged intensities were converted to structure- to converted were intensitiesAveraged . B) is 26.6 Å 26.6 is B) α 2+ 2+ –RMI1 model was iteratively rebuilt with COOT –Mg R . The conditionsThe. of crystallizationfor crystals the were Tris, pH 7.0, 8% (w/v) PEG 5000 MME, and merge 5 3 〈 2+ X-ray diffraction data were recorded with a with recorded were data diffraction X-ray with autoBuster, and an optimal high-resolution 5 α I/ 5 2 was less than 50%. After efforts to improve the 6 –RMI1) and Ca and –RMI1) ab . -Ca σ or REFMAC 〉 plane CC = 2 for the TopIII the for 2 = 1/2 2 2+ . The TOPIII The . of 0.81 and a –RMI1, respectively, with 0% and 0.1% The TopIII 2+ / or Ca ) with data to 2.85-Å resolution, 5 1/2 1 , and POINTLESSand , 6 5 cells/ml and harvested after harvested and cells/ml 7 was 0.30, and along the at 3.3-Å resolution, at which 2+ 2+ ua Strep-TopIII Human α α c appeared as square-based α (TOPIII axis, the two principal principal two the axis, ∆ –Mg –Ca –Mg 2 P B of 25.9 Å 3 R . Crystals grew only in 4 N α factors for both data ab 1 2+ –Mg 2 B 2 2+ –RMI1 model was plane and 2.88 Å 2.88 and plane 1 I –RMI1 data have data –RMI1 2 with PHASER ). –RMI1 structure α Table 2+ –Ca –RMI1 data –RMI1 5 2 2 . Inclusion 2+ was used used was 〈 1 I/ –RMI1) –RMI1) . 96.7% 96.7% . σ c 〉 axis = 2 = 5 5 3 5 α 5 4 - - - )

geometry and appropriate bond lengths (Mg model. The metal sites in both structures were restrained to maintain octahedral structure similaritylocal restraints TopIII imposedonfinal the TopIIIfor as way same the in refined and CC the of CC a Å), (3.45–3.25 shell resolution outer the (in resolution, selected with similar CC TopIII the of fitting rigid-body by determined TopIII The chain. per group TLS one TCEP immediately beforemM 0.2 itand NaCl wasmM 100 applied7.4, pH HEPES,to mM to 25 glow-dischargedin nM 200 of centration holey carbon grids Electron microscopy. are given in models final the for statistics Refinement model. the fromomitted were built tional residues of the TopIII sidechains were truncated at C with one hydrated Mg loop to be traced. The final refined model contains RMI1 Ca percentile, among the best for this resolution. TopIIIthe quality,places model analyze to reflections was omitted from refinement for cross-validation. MolProbity 2.42Å, both with variants were generated by a site-directed mutagenesis strategy, cloned in pGEX junctions.Hollidaydouble ofDissolution of EtBr.presence the in gel agarose 1% a on loaded were kDNA of (Topogen). ng Products 100 of presence the in but were, assays catenation the as performed were assays previoulsy of in the presence as 100 ng described pUC19, were alone (Sigma) or incubated with 200 nM GST–WT 1 Rmi1 or GST-rlRmi1 and Sgs1 nM 60 Top3, nM 100 assays, catenation the For Technologies). (Life II Green SYBR with stained were gels The °C. 37 at h 1 for K Proteinase and EDTA with incubation by stopped were reactions The 100 7.4, pH were performed for 1 h, 3 h or overnight in a reaction buffer of 25 mM HEPES, presence or absence of EtBr. and incubation for 1 h at 37 °C. The samples were run on 1% agarose gel in the MgOAc). The reactions were stopped by the addition of EDTA and Proteinase100 7, K pH (200 ng per reaction) and incubated 1 h at 42different constructs)Rmi1were incubatedpresencethe in ofsubstrate °C PUC19 in reaction (Top3and bufferproteinsassay, relaxation (25 the For mM buffer). HEPES, (final glycerol 10% and Proteins were stored in 50 mMaccording to Tris,a previously described protocolpH 7.0, 150 mM NaCl,cells (High Five) as GST N-terminal fusion proteins1 with C-terminal His mM ref. in outlined proteins. Relaxation, M13 digest assays and catenation-decatenation assays with yeast the model further. improvenot incrementsdid Smaller 10°. to set orientationforgenerationwas the final model ( refinementconvergence,of cycles several After EMAN. implementedin as py TopIIIof structure crystal the from estimated dimensions with blob Gaussian featureless a was refinementsangular for used model initial the bias,To modelEMAN2. avoid of3.7totalÅ.A of 4,000 particles were picked and subsequently analyzed with micrographs CCD at 130,000×nominal magnification, resulting pixela in size microscope equipped with a LaB6 filament electron and operated at 80 kV.CM10 Philips Images were a recorded as with examined was grid The solution. acetateuranyl (m/w) 2% with negativelystained and s 45 for grids EM the to (Cu 400-mesh grids, Quantifoil Micro Tool GmbH). The protein was adsorbed The RMI1 decatenation loop (residues 116–130) was ordered in the TopIII M13 (200 ng per reaction, M13 mp18 phage ssDNA from Clontech) digestions 2+ 1/2 –RMI1 crystal as opposed to that of TopIII of 0.73 and a ab plane and a CC a and plane Yeast Top3, Sgs1 and bacterial SSB were expressed and purified as purified and expressed were SSB bacterial and YeastTop3, Sgs1 µ Table g/ml, 100 g/ml, µ g/ml BSA, 32% glycerol, 5 mM NaOAc, and 2 mM MgOAc. MgOAc. mM 2 and NaOAc, mM 5 glycerol, 32% BSA, g/ml 2 Supplementary Fig. 3 3 σ 1 . Yeast Rmi1 WT and rl constructs were expressed in insect insect in expressedwere constructs rl andYeast . WT Rmi1 α ∆ . =0.02 Å), and the same randomly assigned of set 5% of all –RMI1, generated with the command makeinitialmodel. command the with generated –RMI1, B of 35.5 Å 2+ The sample used for crystal growth was diluted to a con µ /Ca nature structural & molecular biology molecular & structural nature g/ml BSA, 32% glycerol, 5 mM NaOAc, and 0.5 mM 0.5 NaOAc,and mM 5 glycerol, 32% BSA, g/ml 1/2 2+ α 0.97 in the direction of the of direction the in 0.97 ion bound to Glu41 in monodentate mode. Several N and C termini that could not be unequivocally β 2 ,owing to a lack of electron density, and addi ). The TopIII 1/2 cutoffs derived for TopIII ) was obtained. The angular spacing used α –Ca α α 2 –Mg The humanL The 2+ 3 –Ca α to avoid nuclease contamination. α 2+ -OH –Ca –Mg α –RMI1 crystal structure was was structure crystal –RMI1 2+ 2+ –Mg –RMI1 with the addition of addition the with –RMI1 –RMI1 model in the 100th 100th the in model –RMI1 2+ 2 1/2 2+ = 2.09 Å and Ca –RMI1 model was rebuilt of 0.24 in the direction the in 0.24 of –RMI1, thus allowing the 2+ 2–216 –RMI1 data at 3.25-Å at data –RMI1 c axis, with an overall an with axis, 2 1 5 and L and . The decatenation . The decatenation β and TopIII -mercaptoethanol, µ α α M M –Mg –Mg 2 RMI1-loop E. coli E. 2+ 2+ 2+ -OH 5 –RMI1 –RMI1 α 8 6 , used 20–637 tags, SSB SSB 2 α = – - -

© 2014 Nature America, Inc. All rights reserved. absence of RMI1. absence in the of observed by product percentage the of divided RMI1 in presence the in (represented factor screen, stimulation The Typhoonscanner. a with recorded phosphorimager was image a the and to overnight exposed were gels After the °C. dried, 4 at being run electrophoresis gel acrylamide native 8% with separated ref. in described as were dialyzed overnight against final buffer. around 140 mM NaCl) from RMI1 (elution around 200 mM(elution tag NaCl).GST ofseparation the Healthcare)allowingThe GE resin, proteinsQ (POROS protease). RMI1 protein was further purified by anion-exchange was performed.chromatography This was followed by overnight TEV cleavage (4 °C, 2% of TEV sonication. After ultracentrifugation clarification, a GST-Sepharose affinity step 20 °C after induction with 0.2 mM IPTG). The cells were harvested and lysed by in produced and plasmids nature structural & molecular biology molecular & structural nature The dHJ substrate preparation and dissolution reactions were carried out out carried were reactions dissolution and preparation substrate dHJ The Fig. 4 Fig. 5 9 g with the exception that the reaction products were were products reaction the that exception the with ) was calculated as the percentage of product formed formed product of percentage the as calculated was ) E. coli E. ROSETTA strain (overnight expression at expression (overnight strain ROSETTA

59. 58. 57. 56. 55. 54. 53. 52. 51.

ahai CZ & iko, .. islto o dul Hlia jntos y the by junctions Holliday double of Dissolution I.D. Hickson, & C.Z. Bachrati, V.B. Chen, macromolecular of Refinement E.J. Dodson, & A.A. Vagin, G.N., Murshudov, Adams, P.D. of development and Features K. Cowtan, & W.G. Scott, B., P.,Lohkamp, Emsley, A.J. McCoy, quality. data and model crystallographic Linking K. Diederichs, & P.A. Karplus, M.D. Winn, Kabsch, W. Automatic processing of rotation diffraction data from crystals of initially (2009). III topoisomerase and BLM of action concerted crystallography. 53 method. maximum-likelihood the by structures solution. structure Coot. (2007). 658–674 Science Crystallogr. Biol. D Crystallogr. unknown symmetry and cell constants. cell and symmetry unknown , 240–255 (1997). 240–255 , Acta Crystallogr. D Biol. Crystallogr. Biol. D Crystallogr. Acta

336 et al. t al. et , 1030–1033 (2012). 1030–1033 , t al. et t al. et Acta Crystallogr. D Biol. Crystallogr. Biol. D Crystallogr. Acta PHENIX: a comprehensive Python-based system for macromolecular oPoiy alao srcue aiain o macromolecular for validation structure all-atom MolProbity: vriw f h CP sie n cret developments. current and suite CCP4 the of Overview Acta Crystallogr. D Biol. Crystallogr. Biol. D Crystallogr. Acta hsr rsalgahc software. crystallographic Phaser

67 , 235–242 (2011). 235–242 , J. Appl. Crystallogr. Appl. J.

66 , 486–501 (2010). 486–501 , α Acta Crystallogr. D Biol. Crystallogr.Biol. Crystallogr.D Acta . Methods Mol. Biol. Mol. Methods

66

, 12–21 (2010). 12–21 , 66 . pl Crystallogr. Appl. J. doi: , 213–221 (2010). 213–221 ,

26 10.1038/nsmb.2775 , 795–800 (1993). 795–800 ,

582 , 91–102 ,

Acta 40 ,