A meiosis-specific AAA+ assembly reveals repurposing of ORC during budding yeast gametogenesis

ORC (Orc1-6) is an AAA+ complex that loads the AAA+ MCM helicase to replication origins. Orc1, a subunit of ORC, functionally interacts with budding yeast Pch2, a meiosis-specific AAA+ protein. Pch2 regulates several chromosomal events of gametogenesis, but mechanisms that dictate Pch2 function remain poorly understood. We demonstrate that ORC directly interacts with an AAA+ Pch2 hexamer. The ORC-Pch2 assembly is established without Cdc6, a factor crucial for ORC-MCM binding. Biochemical analysis suggests that Pch2 utilizes ORC’s Cdc6-binding interface and employs its non-enzymatic NH2-terminal domain and AAA+ core to engage ORC. In contrast to phenotypes observed upon Orc1 impairment, nuclear depletion of other subunits of ORC does not lead to Pch2-like phenotypes, indicating that ORC integrity per se is not required to support Pch2 function. We thus reveal functional interplay between Pch2 and ORC, and uncover the repurposing of ORC to establish a non-canonical and meiosis-specific AAA+ assembly.


Flow cytometry 137
Flow cytometry was used to assess synchronous passage through the meiotic program (as judged by 138 duplication of the genomic content) and was performed as described (14). For analysis of rapamycin-139 induced phenotype, mitotic cultures were grown to saturation and diluted to OD600 1.0, and 140 rapamycin was added. Samples for flow cytometry were taken at the indicated time points. 141 142

Western blot analysis 143
For Western blot analysis, protein lysates from yeast meiotic cultures were prepared using 144 trichloroacetic acid (TCA)-precipitation and run on 10% SDS-gels (unless otherwise indicated), 145 transferred for 90 minutes at 300 mA and blot with the selected primary antibody/secondary antibody, 146 as described (14). 147 148

Southern blot analysis 149
For Southern blot assay, DNA from meiotic samples was prepared as described (14). DNA was 150 digested with HindIII (to detect DSBs at the control YCR047C hotspot) or ApaLI (to monitor DSBs in 151 the region of interest: right rDNA flank; YLR164W), followed by gel electrophoresis, blotting of the 152 membranes and radioactive (32P) hybridization using probes specific for YCR047C (chromosome III;153 209,030) or YLR164W (chromosome XII;493,932) (for detection of DSBs in 154 hotspot control region or rDNA, respectively) (14). DSBs signals were monitored by exposing an X-155 ray film to the membranes and further developed using a Typhoon Trio scanner (GE Healthcare) after 156 one week of exposure. 157 158

In vivo co-immunoprecipitation 159
For immunoprecipitation assays, 100 mL meiotic cultures at OD600 1.9 were grown, harvested after 160 4.5 hours (unless otherwise indicated), washed with cold H 2 O and snap frozen. Acid-washed glass 161 beads were then added, together with 300 μL of ice-cold IP buffer (50 mM Tris-HCl pH 7.5, 150 mM 162 NaCl, 1% Triton X-100, 1 mM EDTA pH 8.0, with protease inhibitors) and the cells broken with a 163 Fastprep disruptor (FastPrep®-24, MP Biomedicals) by two 45 sec cycles on speed 6. The lysate was 164 subsequently spun 3 min at 3000 rpm and the supernatant transferred to a falcon tube. The lysate was 165 next sonicated by 25 cycles (30 sec on/ 30 sec off), high power range, using a Bioruptor  Plus sonication device, Diagenode) and then spun down 20 min at 15000 rpm. Supernatant was 167 transferred to a new eppendorf tube, and 50 μL of input was taken. For α-Flag/ HA/TAP-based IPs, 1 168 μL of antibody (α-Flag-M2 antibody, Sigma-Aldrich / α-HA, Biolegend/ α-TAP, Thermo Fisher 169 Scientfic) was added to the lysate and rotated for 3 hours. After the incubation step, 30 μL of 170 Dynabeads protein G (Invitrogen, Thermo Fisher Scientfic) were added and rotated overnight at 4°C. 171 For α-Orc2-based IPs, lysate was precleared with 10 μL of Dynabeads protein G for 1 hour at 4°C. 172 Lysate was then incubated with 2 μL of α-V5 (IgG isotype control; Invitrogen) or 11 μL of α-Orc2 173 (Santa Cruz Biotechnology) for 3 h at 4°C, followed by 3 h incubation with 25 μL of Dynabeads 174 protein G. 175 The reactions were washed 4 times with ~ 500 μL of ice-cold IP buffer. In the last wash, beads were 176 transferred to a new eppendorf tube. 55 μL of loading buffer was added, boiled at 95°C and run in a 177 SDS gel. The inputs followed a TCA precipitation step. Briefly, 10% TCA was added and incubated 178 during 30 min on ice. Pellet was then washed with ice-cold acetone, spun and dried on ice, and further 179 resuspended in TCA resuspension buffer (50 mM Tris-HCl 7.5, 6 M Urea). After incubating for 30 180 min on ice, pellet was dissolved by pipetting and vortexing. Finally, 10 μL of loading buffer was added 181 and samples were boiled at 95°C and run in a SDS gel together with the IP samples. Note that for the 182 experiments shown in Figure 4D, 50 mL of sporulation culture, instead of 100 mL, were collected to 183 perform the IP protocol. 184 185

Expression and purification of recombinant proteins in bacteria 246
Hop1 was purified from bacterial cells. Briefly, the coding sequence of Hop1 was sub-cloned into a 247 pET28a vector for expression of recombinant NH 2 -terminally polyhistidine-tagged Hop1 (6xHis-248 Hop1). For protein expression, BL21 RIPL cells were transformed with the resulting vector and further 249 used to inoculate 11 L of LB media, supplemented with kanamycin and chloramphenicol. Cultures 250 were grown at 37°C with vigorous shaking until OD600 ~ 0.6-0.8. Protein expression was induced by 251 addition of 0.25 mM IPTG overnight at 18°C. Cells were harvested by centrifugation at 4500 rpm for 252 15 min and the pellet washed with PBS and immediately snap frozen. For protein purification, cell 253 pellets were resuspended in buffer A (50 mM Hepes, pH 7.5, 300 mM NaCl, 5 mM Imidazole, 10% 254 Glycerol, 0.05% Tween-20, 5 mM β-mercaptoethanol) supplemented with benzonase and protease 255 inhibitors (1 mM PMSF and Serva protease inhibitor mix). Cells were lysed using a microfluidizer 256 (Microfluidizer M-110S, Microfluidics Corporation), centrifuged at 30000 rpm, 4°C for 1h and the 257 lysate filtered. The clear lysate was firstly passed through a 5 mL TALON column (GE Healthcare). 258 After extensive washing, protein was eluted with an imidazole gradient between buffer A and buffer B 259 (buffer A supplemented with 400 mM imidazole). Eluate was pooled, diluted 2:1 in buffer A without 260 NaCl and imidazole, and subsequently loaded into a Heparin column (HiTrap Heparin 16/10, GE 261 Healthcare), previously equilibrated with buffer C (20 mM Hepes,pH 7.5,150 mM NaCl,5 mM 262 MgCl2, 10% Glycerol, 5 mM β-mercaptoethanol). Protein was further eluted in a gradient between 263 buffer C and D (buffer C with 1 M NaCl), and fractions pooled and concentrated using a 30K Amicon-264 Ultra-15 centrifugal filter. Concentrated protein was spun down 15 min in a bench-top centrifuge (4°C) 265 and immediately loaded onto a HiLoad 16/600 Superdex 200 column (GE Healthcare), pre-266 equilibrated in gel filtration buffer consisting of 20 mM HEPES pH 7.5, 300 mM NaCl, 5 mM MgCl2, 267 5% glycerol and 2 mM β-mercaptoethanol. Fractions were analyzed by SDS-PAGE and those fractions 268 containing 6xHis-Hop1 were concentrated with an Amicon-Ultra-15 concentrator (MWCO 30 kDa), 269 snap frozen and kept at -80°C until further use. 270

Cross-linking Mass-Spectrometry 320
Cross-linking Mass-Spectrometry (XL-MS) was performed as described (20). Briefly, 0.75 µM of His-321 MBP-Pch2 was mixed with 1.5 µM of His-ORC complex in 200 µL of buffer (30 mM HEPES pH 7.5, 322 150 mM NaCl, 2 mM TCEP) and incubated at 4°C for 90 minutes. DSBU (disuccinimidyl dibutyric 323 urea -also known as BuUrBu-, Alinda Chemical Limited) was added to a final concentration of 3 mM 324 and incubated at 25°C for 1 hour. The reaction was stopped by adding Tris-HCl pH 8.0 to a final 325 concentration of 100 mM and incubated at 25°C for an additional 30 min. 10 µL of protein sample was 326 taken before and after adding the cross-linker for analysis by SDS-PAGE. SDS-PAGE gel was stained 327 with CBB. Cross-linked protein complexes were precipitated by adding 4 volumes of cold acetone (-328 20°C overnight), centrifuged 5 min at 15000 rpm and the pellet was dried at room temperature. 329 Protein pellets were denatured in denaturation-reduction solution (8 M urea, 1mM DTT) for 30 min at 330 25°C. Cysteine residues were alkylated by adding 5.5 mM chloroacetamide and incubating for 20 min 331 at 25°C. ABC buffer (20 mM ammonium bicarbonate pH 8.0) was added to reduce the final 332 concentration of urea to 4M. Sample was digested by Lys-C (2 µg) at 25°C for 3h, followed by 333 overnight Trypsin (1 µg) digestion in buffer containing 100 mM Tris-HCl pH 8.5, 1 mM CaCl2 at 334 25°C. The digestion was stopped by adding trifluoroacetic acid (TFA) to a final concentration of 0.2%. 335 Resulting peptides after digestion were run in three independent Size-Exclusion Chromatography 336 (SEC) runs on a Superdex Peptide 3.2/ 300 column (GE Healthcare) connected to an ÄKTAmicro 337 FPLC system (GE Healthcare). SEC runs were performed at a flow rate of 0.1 mL/min in buffer 338 containing 30% acetonitrile and 0.1% formic acid. 100 µL fractions were collected and the same 339 fractions from the three SEC runs were pooled, dried and submitted to LC-MS/MS analysis. 340 LC-MS/MS analysis was performed as previously reported using an Ultimate 3000 RSLC nano system 341 and a Q-Exactive Plus mass spectrometer (Thermo Fisher Scientific) (20). Peptides were dissolved in 342 water containing 0.1% TFA and were separated on the Ultimate 3000 RSLC nano system (precolumn: 343 C18, Acclaim PepMap, 300 μm × 5 mm, 5 μm, 100 Å, separation column: C18, Acclaim PepMap, 75 344 μm × 500 mm, 2 μm, 100 Å, Thermo Fisher Scientific). After loading the sample on the precolumn, a 345 multistep gradient from 5−40% B (90 min), 40−60% B (5 min), and 60−95% B (5 min) was used with 346 a flow rate of 300 nL/min; solvent A: water + 0.1% formic acid; solvent B: acetonitrile + 0. Raw data from the Q-Exactive Plus mass spectrometer were converted to Mascot generic files (MGF) 355 format. Program MeroX (version 1.6.6.6) was used for cross-link identification (21). Combined MS 356 data in MGF format and the protein sequences in FASTA format were loaded on the program and MS 357 spectra matching cross-linked peptides were identified. In the settings of MeroX, the precursor 358 precision and the fragment ion precision were changed to 10.0 and 20.0 ppm, respectively. RISE mode 359 was used and the maximum missing ions was set to 1. MeroX estimates the false discovery rate (FDR) 360 by comparison of the distribution of the cross-link candidates found using provided protein sequences 361 and the distribution of the candidates found from decoy search using shuffled sequences. A 2% FDR 362 was used as the cut-off to exclude the candidates with lower MeroX scores. The results of cross-link 363 data were exported in comma-separated values (CSV) format. Cross-link network maps were 364 generated using the xVis web site (https://xvis.genzentrum.lmu.de) (22). Validation of the datasets was 365 performed by identifying 13 intra-MBP crosslinks and using a published crystal structure of MBP 366 (PDB 1FQB, (23)) to map Cα-Cα distances between identified crosslinked amino acids. The average 367 Cα-Cα was 14.41 Å, which is in good agreement with the Cα-Cα distance (12 Å) which the cross-368 linked state of DSBU is able to facilitate (Supplementary Table 3). 369

Pch2 interacts with the ORC complex in meiotic G2/prophase 373
We previously showed that Pch2 functionally interacts with Orc1 (14), but the biochemical 374 basis of this interaction remains poorly understood. To start to define the interaction between Pch2 and 375 Orc1, we investigated how this interaction depended on Pch2 hexamer formation and ATP hydrolysis 376 activity in vivo. We employed an ATP hydrolysis mutant within the Walker B domain of Pch2 (pch2-377 E399Q) ( Figure 1A), which is unable to support rDNA-associated DSB protection (14). In other 378 AAA+ enzymes, mutating this critical residue in the Walker B domain prevents efficient ATP 379 hydrolysis and stalls the stereotypical catalytic cycle of AAA+ enzymes. This often leads to stabilized 380 interactions between AAA+ proteins and their clients and/or adaptors. Equivalent mutants in other 381 AAA+ enzymes have been used to trap enzyme:client and/or enzyme:adaptor interactions (9,24). We 382 detected an increased interaction between Pch2 and Orc1 in cells expressing Pch2-E399Q as compared 383 to cells expressing wild type Pch2 ( Figure 1A and B). We next investigated a different mutant Pch2 384 allele, which carried a mutation within the Walker A motif (K320R). Mutations in residues located 385 within this motif have been shown to reduce ATP binding (9). When we probed the interaction 386 between Pch2 and Orc1, Orc1-TAP failed to co-immunoprecipitate Pch2-K320R ( Figure 1A and C). 387 Considering that mutations in the Walker A motif lead to monomerization of Pch2 in vivo (25), our 388 data suggest that the efficient interaction between Pch2 and Orc1 relies on ATP binding and Pch2 389 hexamer formation. As a whole, these experiments indicate that Pch2 interacts with Orc1 in a manner 390 that is consistent with a stereotypical AAA+: client and/or adaptor interaction. 391 Many, if not all functions ascribed to Orc1 involve its assembly into the six-component Origin 392 Recognition Complex (ORC; consisting of Orc1-6) (15)). We therefore tested whether in addition to 393 Orc1, other subunits of ORC also interacted with Pch2. We employed the pch2-E399Q allele to 394 stabilize in vivo interactions. Our co-immunoprecipitation (co-IP) assays revealed an interaction 395 between TAP-tagged versions of Orc2 and Orc5 and Pch2 during meiotic G2/prophase ( Figure 1D-F). 396 Similarly, we detected this interaction between 3xFLAG-tagged Pch2-E399Q and Orc2 by Co-IP using 397 a α-Orc2 antibody ( Figure 1G). Furthermore, an unbiased mass-spectrometric analysis of the Pch2-398 E399Q interactome identified Orc5 in addition to Orc1, indicating that Pch2 interacts with multiple 399 ORC subunits (VBR and GV, unpublished observations). As a whole, we conclude that Pch2 interacts 400 with ORC during meiotic G2/prophase. 401 To enable the chromosomal loading of the MCM AAA+ replicative helicase at origins of DNA 402 replication, ORC (i.e. Orc1-6) associates with Cdc6, an additional AAA+ protein (15)( Figure 1D). 403 Pch2 is expressed during meiotic S-phase and G2/prophase, whereas Cdc6 availability is restricted to 404 G1 phase (26), also in the meiotic program (27). This suggests that the interaction between Pch2 and 405 ORC occurs independently of Cdc6. We employed a meiosis-specific null allele of CDC6 (cdc6-mn) 406 (28) which interferes with pre-meiotic DNA replication ( Figure 1H), to investigate if absence of Cdc6 407 influenced Pch2-ORC binding. (Note that in the cdc6-mn background, despite a failure to undergo bulk 408 DNA replication, meiotic progression is unaffected and cells initiate DSB formation in a meiotic 409 G2/prophase-like state (28,29)). The interaction between Pch2 and Orc1 in the cdc6-mn background 410 was similar to the binding that was observed in CDC6 cells ( Figure 1I), indicating that ORC-Pch2 411 assembly occurs independently of Cdc6. 412 We have previously shown that Pch2 protects ribosomal (r)DNA array borders (i.e. the ~1-10 413 outermost rDNA repeats and ~50 kb of single copy flanking sequences) against meiotic DSB 414 formation ( (14), and Figure 1J). In agreement with our observation that the interaction between Pch2 415 and ORC does not depend on Cdc6, we observed that Cdc6 depletion (via cdc6-mn) did not trigger a 416 Pch2-like phenotype at rDNA borders, as judged by the analysis of meiotic DSB formation at the right 417 rDNA flank (YLR164W) ( Figure 1K). In addition, pch2Δcdc6-mn efficiently formed DSBs within the 418 right rDNA flank ( Figure 1K), demonstrating that bulk (Cdc6-dependent) DNA replication is not 419 required for DSB formation in these regions in cells lacking Pch2. Thus, these data show that Pch2 and 420 ORC functionally interact during meiotic G2/prophase, and that this interaction does not require Cdc6. 421 422

In vitro reconstitution demonstrates a direct interaction between Pch2 and ORC 423
To gain understanding of the biochemical basis underlying ORC-Pch2 binding, we sought to in 424 vitro reconstitute this complex. For this, we expressed and purified budding yeast Pch2 (carrying a 425 NH 2 -terminal His-MBP tag) through a baculovirus-based protein expression system. As judged by 426 size exclusion chromatography (SEC), purified Pch2 assembled into an apparent hexamer (predicted 427 size ~636 kDa), with a minor fraction that appears to be monomeric (size of ~106 kDa for His-MBP-428 Pch2) (Figure 2A). We confirmed functionality of our affinity purified Pch2 by demonstrating a direct 429 interaction with Hop1, a confirmed substrate of Pch2, as previously described (10) (Supplementary 430 Figure 1A). We next tested whether Pch2 directly interacted with ORC, by using ORC (i.e. Orc1-6; 431 with Orc1 carrying a His-tag, total size ~414 kDa) purified from insect cells (see Supplementary 432 Figure 1B). Solid phase pulldown experiments revealed that Pch2 is able to interact with the entire 433 ORC (i.e. Orc1-6) ( Figure 2B and C). This demonstrates that these AAA+ proteins indeed interact 434 directly. Next, we asked whether this interaction could also be reconstituted in solution. Size Exclusion 435 Chromatography (SEC) analysis confirmed that ORC and Pch2 form a complex in solution, as judged 436 by a reduced retention volume (which is indicative of a larger and/or more elongated complex) when 437 combined, as compared to the elution profiles of Pch2 or ORC individually ( Figure 2D). We suggest 438 that ORC and Pch2 interact with each other in an ORC (Orc1-6 hexamer) to Pch2 (hexamer) fashion, 439 yielding what would be a complex of ~ 1 MDa. Taken together, these experiments demonstrate that 440 ORC directly interacts with Pch2 to establish a meiosis-specific AAA+ to AAA+ assembly. Since this 441 interaction does not require Cdc6, this assembly represents an interaction of ORC with an AAA+ 442 protein which is biochemically distinct from the interaction of ORC with the MCM AAA+ complex.  Table 2). We used 457 these non-redundant crosslinks to generate crosslink network maps for the ORC-Pch2 assembly by 458 using xVis (https://xvis.genzentrum.lmu.de). These 313 crosslinks consist of 121 intermolecular 459 crosslinks (i.e. crosslinks between peptides originating from two different proteins) and 192 460 intramolecular crosslinks (i.e. crosslinks between peptides originating from a single protein). We 461 identified 96 Pch2-Pch2 crosslinks ( Figure 3C, red lines and Supplementary Table 3). Since Pch2 462 forms a homo-hexamer, we cannot distinguish whether Pch2-Pch2-crosslinked peptides originate from 463 intra-or intermolecular crosslinked peptides. We observed 77 crosslinks between ORC subunits (i.e.  Table 3). When comparing crosslink abundance between individual ORC subunits with 466 a published crystal structure of ORC to model the position of each subunit ( Figure 3F; based on 467 structure PBD 5v8f; (30)), we noted that neighboring subunits often displayed the most abundant 468 crosslinks (for example Orc1/Orc2, Orc2/Orc3 and Orc3/Orc5; see Supplementary Table 3). However, 469 several observed crosslinks span considerable distance when based on the ORC structure we used for 470 analysis (PBD 5v8f; (30)), arguing for significant levels of flexibility within our ORC preparation. Of 471 note, our ORC complex is devoid of Cdc6, and also not bound to MCM-Cdt1, contrary to the reported 472 structure (30), which conceivably could affect complex topology. Furthermore, we cannot exclude that 473 Pch2 leads to structural rearrangements within ORC upon binding. 474 We next focused on the 96 Pch2-Pch2 crosslinks ( Figure  of note when considering these crosslinks. First, we find crosslinks that contain Pch2 peptides from 489 both its enzymatic AAA+ core (12 out of 21; 57%) and its non-catalytic NTD (9 out of 21; 43%) (see 490 also Supplementary Figure 2). We interpret this to indicate that Pch2 makes extensive contacts with 491 the ORC complex, whereby both its enzymatic core and its NTD are involved. Many AAA+ ATPases 492 (including TRIP13, the mammalian homolog of Pch2 (12) (13)) engage clients/adaptors via an initial 493 engagement using their NTDs, and subsequently show interactions mediated through AAA+ 494 core:client binding (9). The observation that both Pch2's AAA+ core and NTD are involved in ORC 495 binding, is consistent with a scenario in which Pch2 binds to ORC in a AAA+:client and/or adaptor-496 type engagement. It is conceivable that Pch2 uses its NTD for the initial recognition of ORC, whereas 497 subsequent AAA+ mediated interactions stabilize this complex formation. Second, a large fraction of 498 the total Pch2-ORC crosslinks is established between Pch2 and Orc1/Orc2 (10 out of 21; 48 %). 499 Although these two subunits are the largest polypeptides of the Orc1-6 complex (which might affect 500 the distribution of the observed crosslinks), we note that Orc1/Orc2 are neighboring the position that is 501 occupied by Cdc6 when it interacts with ORC. In our preparations, Cdc6 is not present, leaving this 502 space unoccupied. We thus speculate that Pch2 utilizes this "vacated" Cdc6-binding position to 503 interact with ORC. In agreement with this is our earlier finding that, in vivo, Pch2 binding with ORC 504 occurs independently of Cdc6 ( Figure 1I). We attempted to map the identified 21 inter-ORC-Pch2 505 crosslinks onto an ORC structure (PBD 5v8f (30), Figure 3F; crosslinked residues are marked by a 506 black dot). Due to the absence of regions of ORC within the used crystal structure, we were unable to 507 map several of the ORC-Pch2 crosslinks (i.e. crosslinks with Orc2, Orc5 and Orc6). Mapping of 508 observed crosslinks showed a distribution of crosslinked residues across a large region of ORC, 509 suggesting that Pch2 establishes extensive contacts with the ORC complex. Interestingly, when we 510 analyzed the position of these residues in a structure containing Cdc6, we noted that three crosslinked 511 residues within Orc1 (K612, T614 and S615) were located in a position that is shielded by Cdc6,512 according to the ORC-Cdc6-Cdt1-MCM complex structure (PBD 5v8f;(30) formation. In contrast, our XL-MS analysis suggests that the NTD of Pch2 is involved in mediating 520 binding to ORC. We aimed to establish whether indeed the NTD was involved in Pch2-ORC 521 assembly. For this, we first employed yeast two-hybrid (Y2H) analysis, to show that Pch2 lacking its 522 NTD (amino acids 2-242) was unable to interact with Orc1 ( Figure 4A and B). We next investigated 523 the interaction between Pch2 and Orc1 in meiotic G2/prophase, by expressing an identical truncated 524 version of Pch2 (3xFLAG-Pch2-243-564). This truncated version of Pch2 was impaired in its ability to 525 interact with Orc1 ( Figure 4C and D). The residual interaction of Pch2-243-564 with Orc1 might 526 indicate that, in meiotic cells, Pch2 lacking its NTD retains a certain degree of affinity towards ORC 527 ( Figure 4D). We next purified Pch2 lacking the NTD (His-MBP-Pch2-243-564) from insect cells. By 528 SEC, we observed that this Pch2 protein eluted at an apparent size that indicated a more extended 529 shape or less organized assembly as compared to full length Pch2 (data not shown). We have observed 530 a similar behavior when purifying Pch2 proteins that harbor specific amino acid mutations within the 531 NTD (unpublished observations, MAVF and GV). These findings imply a role for the NTD in 532 stabilizing and/or maintaining Pch2 into a stable, well-ordered hexamer (see also above). Importantly, 533 the ability of purified Pch2 243-564 to interact with ORC was abolished, further demonstrating an 534 important contribution of the NTD of Pch2 in directing interaction with ORC ( Figure 4E). 535 We next asked whether the NTD of Pch2 was sufficient for ORC binding. Based on Pch2 sequence 536 conservation and secondary structure predictions, we performed Y2H analyses using a series of 537 COOH-truncated fragments of Pch2. These analyses revealed that the NTD of Pch2 (consisting of 538 amino acids 2-242) is sufficient to establish the interaction with Orc1 ( Figure 4F). Further truncations 539 of the NTD identified a minimal fragment of Pch2 (containing amino acids 2-144) sufficient for the 540 interaction between Pch2 and Orc1. In agreement with these observations, our XL-MS analysis 541 identified several crosslinks between Pch2 and ORC-subunits that consisted of Pch2-peptides that are 542 located within this region of the NTD (K88, K18, K43; Figure 3D and E), underscoring the importance 543 of this domain in mediating the interaction between Pch2 and ORC. We attempted to express 544 corresponding Pch2-NTD fragments in meiosis, but observed that often these fragments were poorly 545 expressed (unpublished observations, MAVF and GV). This precluded us from performing in vivo 546 interaction studies. To further test a role of the NTD of Pch2 in mediating interaction with ORC, we 547 expressed recombinant NTD fragments. We noted that, similarly to our in vivo observations, many 548 recombinantly-produced fragments were poorly expressed or aggregated under purifying conditions 549 (unpublished observations, MAVF and GV). We managed to express and purify the minimal NH 2 -550 terminal fragment of Pch2 (His-MBP-Pch2-2-144) that was sufficient for Orc1 interaction in our Y2H 551 analysis. SEC analysis suggested that this fragment exists as a monomer (expected size ~59 kDa), 552 which is in agreement with the crucial role AAA+ domains play in mediating hexamerization of 553 AAA+ complexes ( Figure 4G). This fragment was capable of interacting with ORC, albeit to 554 significantly lesser extent than full length Pch2 ( Figure 4H and I). This could indicate additional 555 binding interfaces between Pch2 and ORC that lie outside of this domain (as suggested by the 556 observation of additional crosslinks containing peptides from regions outside of the NTD of Pch2, and 557 by the residual in vivo interaction we observed between Pch2-ΔNTD and Orc1; see above). 558 Alternatively, hexamer formation of Pch2 (driven by AAA+ to AAA+ interactions) increases the local 559 effective concentration of the NTD, and this could contribute to efficient binding between Pch2 and 560

ORC. The latter interpretation is in agreement with our observation that the in vivo interaction between 561
Pch2 and Orc1 is severely diminished in cells expressing a Pch2 Walker A domain mutant, which is 562 expected to disrupt ATP binding and hexamerization (25). We conclude that the NTD of Pch2 563 provides a crucial contribution to ORC-Pch2 complex formation (Figure 4). 564 565

In vivo analysis of the functional connection between Pch2 and ORC 566
We previously demonstrated that Pch2 is required to prevent rDNA-associated meiotic DSB 567 formation (14). Inactivating Orc1 (via a temperature-sensitive allele of ORC1, orc1-161) triggers a 568 similar rDNA-associated phenotype as observed in cells lacking Pch2, which shows that Orc1 and 569 Pch2 collaborate to protect the rDNA against DSB formation and instability in meiosis (14). Since our 570 biochemical analysis demonstrates that Pch2 binds to ORC, we aimed to address whether ORC is 571 required for Pch2 function at rDNA borders during meiotic DSB formation and recombination. ORC 572 subunits are essential for cell viability, and we thus employed the "anchor away" method (31), which 573 has been used to efficiently deplete chromosomal factors in budding yeast meiosis (32-34), to 574 inactivate selected ORC subunits ( Figure 5A). Mitotically proliferating diploid cells that carry FRB-575 tagged versions of ORC2 or ORC5 (orc2-FRB and orc5-FRB) exhibited a strong growth defect when 576 grown in the presence of rapamycin ( Figure 5B), demonstrating efficient nuclear depletion of Orc2 and 577 Orc5. To investigate the efficacy and timing of this functional depletion, we used flow cytometry to 578 query DNA replication in logarithmically growing cultures after treatment with rapamycin. In the 579 orc2-FRB or orc5-FRB backgrounds, addition of rapamycin induced DNA replication to cease (as 580 judged by an accumulation of 2N-containing cells) within 180 minutes of treatment, with the first 581 effects detectable after 90 minutes ( Figure 5C). These experiments indicate a rapid and efficient 582 functional depletion of Orc2 or Orc5. We used these ORC alleles to investigate rDNA-associated DSB 583 formation (by probing meiotic DSB formation at the right rDNA flank; YLR164W (14)). Surprisingly, 584 rapamycin-induced depletion of Orc2 or Orc5 did not trigger an increase in rDNA-associated DSB 585 formation, in contrast to what is observed in cells lacking Pch2 or in cells expressing a temperature-586 sensitive allele of ORC1 (14) ( Figure 5D). Meiotic progression seemed normal under these conditions, 587 since meiotic DSB formation at a control locus (YCR047C; chromosome III) occurred normally 588 ( Figure 5D), and pre-meiotic DNA replication timing appeared unaffected under this treatment 589 regimen (data not shown). MCM association with origins of replication (the critical ORC-dependent 590 step during DNA replication) occurs prior to induction into the meiotic program (and thus rapamycin 591 exposure in our experimental setup) (27), and therefore nuclear depletion of ORC in this regimen is 592 not expected to interfere with efficient pre-meiotic DNA replication. We cannot currently exclude that 593 incomplete depletion of Orc2/5 precludes us to expose a role for Orc2/5 in controlling Pch2's rDNA-594 associated phenotype. However, based on the viability effects ( Figure 5B), and on the timing of the 595 observed effects of Orc2/5-depletion during vegetative growth (i.e. within 90-180 minutes; Figure 5C) 596 as compared to the duration of rapamycin treatment in our meiotic experiments (up to a maximum of 8 597 hours), we favor the interpretation that in vivo, Orc2 and Orc5 are not strictly required for rDNA-598 associated Pch2 function. In agreement with this conclusion are experiments in which cells were 599 exposed to longer periods of rapamycin treatment by adding the drug in pre-meiotic cultures (i.e. 3 600 hours prior to initiation of meiotic cultures). Under these conditions we equally failed to see an effect 601 of Orc2/Orc5 depletion on rDNA-associated DNA break formation, despite the appearance of (mild) 602 pre-meiotic DNA replication defects (unpublished observations, MAVF and GV). Based on these 603 results, we conclude that the rDNA-associated function of Pch2 does not strictly depend on Orc2/Orc5 604 function. 605 The lack of a role for Orc2/5 in mediating Pch2-dependent suppression of rDNA instability is in stark 606 contrast with the role of Orc1 (14), suggesting that Orc1 could be a central mediator of the interaction 607 between ORC-Pch2. Several observations support this hypothesis. First, when comparing Pch2 Co-IP 608 efficiencies of Orc1, Orc2 and Orc5, we consistently find the strongest interaction with Orc1 ( Figure  609 1F), arguing that Orc1 is a central interactor of Pch2. Second, we observed several intermolecular 610 crosslinks containing peptides from the MBP-moiety (that is NH 2 -terminally fused to Pch2 in His-611 MBP-Pch2). In addition to 17 intermolecular crosslinks between MBP and Pch2 (which are expected 612 since these two polypeptides are covalently linked), we observed 6 MBP-Orc1 intermolecular 613 crosslinks ( Figure 5E, and Supplementary Table 3). MBP-derived crosslinks with Orc1 were unique: 614 there were no crosslinks observed between MBP and other ORC subunits. Since efficient crosslinking 615 depends on proximity of ~ 12Å between Cα's of crosslinked amino acids, these data argue that Orc1 is 616 in close vicinity of MBP (and, by extension, Pch2). Third, by analyzing the interaction between 617 individual ORC subunits (Orc1-4, and Orc6) and Pch2 using Y2H analysis, we observed an interaction 618 between Orc1 and Pch2, as reported earlier (14) but did not detect an interaction between Pch2 and 619 other individual ORC subunits ( Figure 5F). This result strengthens the conclusion that, within ORC, 620 Orc1 is a major interaction partner of Pch2. To test the premise that Orc1 is a crucial mediator of 621 ORC-Pch2 assembly, we probed the interaction between Pch2 and Orc2/Orc5 in the presence of a 622 temperature-sensitive allele of ORC1 (orc1-161). In this situation, Orc2 and Orc5 showed a decreased 623 ability to immunoprecipitate Pch2, further strengthening the premise that Orc1 is crucial in mediating 624 the interaction between ORC and Pch2 ( Figure 5G). 625 Altogether, our data suggest that in vivo, Pch2 interacts with the entire ORC, with Orc1 being an 626 important mediator of this interaction. Functionally, Orc1 is a crucial binding partner for Pch2. Thus, 627 we conclude that during meiotic G2/prophase, ORC is repurposed to interact together with Pch2, in a 628 biochemical and functional manner that is uniquely distinct from its well-documented role in the 629 chromosomal loading of the AAA+ MCM helicase assembly. 630 631 DISCUSSION 632 The hexameric AAA+ ORC complex is an essential regulator of eukaryotic DNA replication. It 633 forms the loading platform for the chromosomal association of the replicative helicase MCM, a 634 hexameric AAA+ complex (15,16). Here we show that, during the meiotic program of budding yeast, 635 ORC interacts with another AAA+ protein: Pch2. Our data reveal several interesting biochemical 636 characteristics about the ORC-Pch2 assembly. First, we show that the ORC-Pch2 assembly does not 637 require Cdc6 (or any other accessory factors). This is in stark contrast to the highly regulated 638 interaction between ORC and MCM (15,16). Expression of Pch2 is induced during S-phase and peaks 639 during G2/prophase, when Pch2 is involved in many processes controlling meiotic DSB formation and 640 recombination. During this time of the cell cycle, ORC is not complexed with Cdc6 (26,27) and, as 641 such, would be available for association with Pch2. In line with such a temporal separation of Pch2-642 and Cdc6-bound ORC, we found evidence from in vitro reconstitution that Pch2 might (partially) use 643 the binding pocket that in ORC-Cdc6 is occupied by Cdc6. In future experiments, our biochemical 644 reconstitution should allow us to test whether Cdc6 and Pch2 binding to ORC is mutually exclusive. 645 Binding of a monomer of the Cdc6 AAA+ protein to the five other AAA+ like ORC-proteins (Orc1-5) 646 establishes the functional ring-shaped ORC hexamer (i.e. a Cdc6-Orc1-5 hexamer), which, in this 647 composition, is proficient in loading the MCM AAA+ hexamer. (Note that Orc6 is a non-AAA+ 648 domain-containing component of ORC that does not directly contribute to Cdc6-ORC AAA+ hexamer 649 assembly (15,16)). An intriguing possibility was that a monomer of Pch2 AAA+ protein could, in lieu 650 of Cdc6, establish a complex with Orc1-5 (i.e. a 1:5 Pch2:Orc1-5 hexamer). However, we do not find 651 evidence supporting such a binding mode. First, when we reconstituted the ORC-Pch2 complex, we 652 observed that the pool of Pch2 that elutes at the expected size of a Pch2 hexamer interacts with ORC 653 (as judged by SEC analysis; Figure 2D). Second, our combined XL-MS and biochemical analyses 654 indicate that the non-AAA+ domain of Pch2 (i.e. its NTD) provides a key contribution to the efficient 655 binding of Pch2 to ORC (Figure 3 and 4). This kind of behavior would not be expected if a 1:5 656 Pch2:ORC (Orc1-5) would be established via binding principles that are similar to Cdc6-ORC, 657 wherein AAA+ to AAA+ interactions are the main driver of complex formation. Third, a Walker A 658 domain Pch2 mutant that is expected to monomerize (25) (Figure 1), fails to interact with ORC in vivo. 659 Although our current in vitro reconstitutions cannot formally exclude the establishment of a 1:6 660 Pch2:ORC complex that then is bound to an hexamer of Pch2 (in a manner analogous to a 1:6 Cdc6-661 ORC (Orc1-6): hexameric MCM assembly), we interpret our experiments to indicate that ORC (Orc1-662 6) is complexed with an hexamer of Pch2. Our results also suggest that Pch2 employs a stereotypical 663 AAA+ to client/adaptor binding mode towards ORC: i) binding is increased in a mutant that stalls ATP 664 hydrolysis ( Figure 1B), ii) hexamerization is required for efficient interaction ( Figure 1C), and iii) the 665 non-enzymatic NTD of Pch2 plays a crucial role in mediating the interaction between Pch2 and ORC 666 ( Figure 4). If the binding of Pch2 with ORC is in line with an AAA+ to client/adaptor interaction, can 667 ORC then be considered a client or an adaptor of Pch2? Together with our earlier observations, which 668 revealed that Orc1 is required for the nucleolar localization and function of Pch2 (14), our current 669 analysis is in agreement with an adaptor-like role for Orc1 (i.e. by aiding in proper subcellular 670 localization of Pch2). Based on these experiments, we favor a model in which Orc1 (and ORC) acts as 671 a localized chromosomal recruiter of Pch2, in line with an adaptor-like role for ORC in facilitating 672 Pch2 function. Nonetheless, we cannot currently exclude that ORC function/composition is also 673 influenced by Pch2 activity in an AAA+ to client relationship, and our in vitro reconstitution 674 experiments have the promise of addressing this intriguing possibility. Pch2 uses its enzymatic activity 675 to influence the chromosomal association of its clients, chromosomal HORMA domain-containing 676 proteins (10,11). Since Pch2-mediated removal of HORMA proteins has been associated with local 677 control of DSB activity and meiotic recombination, also within the rDNA (3,14,35), an interesting 678 question remains whether, and if so, how the interaction between Pch2 and ORC plays a direct role in 679 Pch2 activity-driven Hop1 removal from specific chromosomal regions. 680 A surprising aspect of our work is the finding that depleting subunits of ORC other than Orc1 (i.e. 681 Orc2 or Orc5) did not lead to a Pch2-associated phenotype at the rDNA locus ( Figure 5). Although we 682 cannot exclude that our depletion strategy for these subunits is incomplete, based on our data we favor 683 the conclusion that these subunits are not strictly required for Pch2 function. In combination with the 684 fact that Orc1 is required for Pch2 function at the rDNA (14), and appears to act as a major interacting 685 partner for Pch2, we envision two possible (not mutually exclusive) molecular explanations. First, 686 since inactivating Orc2/Orc5 is expected to lead to diminished origin binding of ORC, we suggest that 687 the role of Pch2-ORC at the rDNA could be executed away from origins of replication. Second, it is 688 possible that in vivo, Orc1 exists in two pools: one where it is complexed with Orc2-6 (i.e. ORC), and 689 one where it exists as a monomer. Conceivably, Pch2 could interact with both pools. If Orc1 is the 690 protein that provides the needed functionality to Pch2 (whether complexed with ORC or not), 691 inactivating other ORC components (like Orc2/Orc5) would not per se trigger Pch2-like phenotypes. 692 In either case, our findings point to a non-canonical role for Orc1/ORC in mediating the activity of 693 Pch2 during meiotic G2/prophase. The recruitment of Pch2 to the nucleolus is diminished in an orc1-694 161 mutant background (14) and Orc1 should thus contain a chromosome-binding activity that is 695 required for nucleolar recruitment of Pch2. Interestingly, Orc1 contains a nucleosome binding module 696 (a Bromo-Adjacent Homology (BAH) domain) (36,37), and we previously showed that this domain is 697 required for the rDNA-associated role of Pch2 (14). Future work should be focused on understanding 698 how the BAH domain of Orc1 biochemically and functionally contributes to Pch2 function in relation 699 to nucleosome/chromatin association. 700 In conclusion, we have used a combination of in vivo and in vitro analyses to reveal the establishment 701 of a meiosis-specific AAA+ assembly between ORC and Pch2. By establishing an in vitro 702 reconstituted assembly of Pch2 and ORC combined with in vivo analysis, we have shed light on an 703 interaction between Pch2 and an AAA+ adaptor-like protein complex, which is important for localized 704 chromosomal recruitment of Pch2. Our experiments reveal interesting characteristics of this assembly 705 and highlight a certain plasticity in the ability of ORC to interact with distinct AAA+ proteins. 706 Understanding the biochemical, structural and functional connections between these two ATPases in 707 more detail will be an important avenue for future research. crosslinks. Blue= inter-ORC, red= intra-ORC and intra-Pch2, black= inter-ORC-Pch2. D. Table  878 showing inter-ORC-Pch2 crosslinks. Indicated are residues in Pch2, and ORC subunits, domain of 879 Pch2 involved (NTD: . N indicates how often crosslinks were identified. 880 MeroX score is indicated. A. Schematic of ORC assembly and of rapamycin-based anchor away method. B. 10-fold serial 906 dilution spotting assay for anchor-away strains (untagged,. Strains are grown 30 on YP-Dextrose (YPD) or YPD + rapamycin (1μg/mL). C. Flow cytometry analysis of efficiency of 908   Orc6 Orc5 His-MBP-Pch2 His-Orc1