Evolution of Origin Sequence and Recognition for Licensing of Eukaryotic DNA Replication

17 The large size of eukaryotic chromosomes necessitates that the initiation of DNA replication 18 occurs at numerous origins of DNA replication. In S. cerevisiae, origins are highly DNA sequence-19 specific and are recognized by the Origin Recognition Complex (ORC) . In contrast, most 20 eukaryotes have lost features in ORC subunits that contribute to DNA sequence -specific 21 recognition, raising the question of how origins are identified. An analysis of origins in the genome 22 of the distantly related budding yeast Yarrowia lipolytica identified considerable variability in origin 23 sequence and structure . High-resolution structures demonstrate that Y. lipolytica origins are 24 recognized by a combination of ORC and Cdc6 in a manner different from S. cerevisiae. The 25 structure of Yarrowia ORC-Cdc6 bound to different origins shows considerable plasticity in 26 protein-DNA interactions. We compare these yeast structures to the structure of human ORC -27 CDC6 bound to DNA. These studies reveal information on the evolution of origin s and origin 28 recognition. 29 30 Nomenclature note: There is a different nomenclature for proteins in yeast and human cells. For 31 example, Cdc6 in yeasts is CDC6 in human cells. 32

33 The genome in eukaryotic cells is distributed over multiple large chromosomes that each contain 34 numerous origins of DNA replication to ensure that all of the DNA is duplicated precisely once per 35 cell division cycle 1–4. The location of origins in the genome is marked by the assembly of pre-36 Replicative Complexes (pre-RCs) prior to the initiation of actual DNA synthesis from each origin. 37 Pre-RCs are assembled on all potential origins, usually following exit from the previous mitosis or 38 during G1 -phase 5–11. The best characterized system for understanding the biochemistry of 39 complete DNA replication, including pre-RC assembly, derives from studies of the budding yeast 40 S. cerevisiae 5,7,12–17. 41 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 2 In S. cerevisiae, pre-RCs are assembled by the binding of the Origin Recognition Complex (ORC) 42 to specific DNA sequences called Autonomously Replicating Sequences (ARSs) that determine 43 the location of ~500 origins in the 13Mb genome 7,18,19. ORC, a six-subunit ATPase, binds to and 44 bends the origin DNA and then recruits the Cdc6 ATPase. Together, these proteins load two 45 copies of the Mcm2-7 hexamer that are chaperoned by the Cdt1 protein to form the MCM double 46 hexamer ( DH) 5,6,16,17,20. The MCM DH is destined to separate into two divergent replicative 47 helicases called the CMG (Cdc45-Mcm2-7-GINS), which encompasses Cdc45, the Mcm-2-7 48 hexamer, and the four-subunit GINS complex 21. The assembly of the helicase and subsequent 49 replication of DNA occurs following activation of the pre -RC by the S -phase Cyclin-Dependent 50 Kinases (CDKs) and the Cdc7-Dbf4 kinase (DDK) 16. 51 The origins of DNA replication in S. cerevisiae consist of multiple essential or important DNA 52 elements. The A and B1 DNA elements are recognized by ORC, whereas the B2 element is a 53 weak ORC binding site that is in an inverted orientation and of variable distance from the A and 54 B1 elements 12,22–26. Depending on this distance, two modes of assembly of pre-RCs can occur, 55 one requiring only a single ORC and the other involving two separate ORCs 12,17,26,27. Since the 56 genome of S. cerevisiae is relatively compact with little repeat sequences and has very short 57 intergenic DNA regions, origins of DNA replication have most likely evolved to be highly DNA 58 sequence-specific and located in non-transcribed regions of the genome so that the initiation of 59 DNA replication does not conflict with gene transcription 28. As a consequence, most S. cerevisiae 60 origins are located within the short intergenic DNA sequences. 18,29 Origin specificity in S. 61 cerevisiae occurs in part by the interaction of an a-helix in the Orc4 subunit that inserts into a 62 major groove in the origin DNA, a loop in the Orc2 subunit that inserts into a minor groove in the 63 origin DNA, and a lysine-rich region in the intrinsically disordered domain of Orc1 that also binds 64 a DNA minor groove 28,30–32. 65 A small clade of budding yeasts that are evolutionarily related to S. cerevisiae , including 66 Kluyveromyces lactis and Lachancea kluyveri have ARSs and origins that are related in sequence 67 to the S. cerevisiae origins 33,34. The Orc2 loop and Orc4 a-helix in these species are conserved 68 35. In contrast, all other eukaryotes, including other budding yeasts and fungi, and all animals and 69 plants have either lost completely or truncated these origins recognition elements 1,35. In some 70 budding yeasts such as Candida albicans and Pichia pastoris , ARS sequences have been 71 characterized and are very different from the S. cerevisiae clade of ARS sequences, 36,37 but the 72 manner in which the proteins interact with them has not been addressed. Other yeasts, such as 73 the fission yeast S. pombe, have gained an unusual A/T-rich hook domain in Orc4 that binds to 74 the A.T-rich origins of DNA replication, but this mode of origin recognition is not common. Similar 75 to pre -RC assembly using purified S. cerevisiae proteins, 38,39 pre-RC assembly has been 76 reconstituted with purified human proteins, demonstrating both a one -ORC and a two-ORC 77 mechanism of MCM DH loading onto non-specific DNA 17,40–42. While this may suggest that ORC 78 can determine the location of origins of DNA replication in human cells, a meta-analysis of multiple 79 studies that mapped ORC and MCM binding sites in the human genome showed a very poor 80 correlation with the location of origins of DNA replication 43. This may be due to technical reasons, 81 but there remains the matter of how origin recognition, and hence the specification of origin 82 location in most eukaryotes occurs. 83 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 3 In animal cells, such as C. elegans, Drosophila and mammalian cells , including human cell s, 84 origins of DNA replication have been mapped and they cor relate with genomic features such as 85 histone modifications, higher-order chromosome structure, and in many cases transcription start 86 sites 1–3,44,45. For example, in C. elegans, the efficiency of origins of DNA replication is associated 87 with histone H3-lysine-4-dimetylation (H3K4me2) and histone H3-lysine27-acetylation (H3K27Ac) 88 46. In Drosophila and human cells, DNA topology and certain chromatin features mark replication 89 origins, and they are commonly associated with regions that contain nearby predicted G4-quartet 90 DNA structures 47–51. In human cells, origins of DNA replication are located both at specific loci 91 such as open chromatin regions, but initiation of DNA replication can also occur in a distributed 92 fashion, where stochastic origin firing takes place in chromosome replication initiation domains 93 44,45,52–57. How these specific and distributed origins are specified is not known, but speculation 94 about epigenetic marking of the initiation of DNA replication is common 4,44,58,59. 95 In a study of the mechanism of origin specificity in S. cerevisiae, we noticed that the Orc4 a-helix 96 and Orc2 loop that provided DNA sequence-specific interactions with origins were only conserved 97 in the small clade of S. cerevisiae-related budding yeasts, whereas many other budding yeasts 98 and all other eukaryotes , including plants and animals, including human ORC, lacked these 99 conserved features 28. In this report, we first determined the structure of human ORC -CDC6 100 bound to a G/C rich DNA. Though we observed DNA bending seen in all ODC complexes to date 101 as well as a surprising minor groove contact, we reasoned that a stronger evolutionary 102 perspective would aid in understanding origin specification. We therefore began studies on 103 Yarrowia lipolytica, which lacked the origin-recognition features seen in S. cerevisiae. Y. lipolytica 104 is a non-conventional, oleaginous yeast that is widely used in biotechnology whose last common 105 ancestor with S. cerevisiae existed ~300 million years ago 60. Unlike S. cerevisiae, Yarrowia is 106 heterothallic, having two separate mating types, MatA and MatB. Previous studies identified a few 107 origins that are located near centromeres , probably because , unlike S. cerevisiae ARSs, 108 propagation of extra-chromosomal plasmids in Yarrowia requires both a centromere and an origin 109 sequence on the plasmid 61–64. To study DNA replication in Yarrowia more thoroughly, we mapped 110 the location of origins in all six chromosomes, demonstrating a genome organization of replication 111 timing domains reminiscent of those in the genomes of animal cells , including human cells. 112 Genetic analysis of two of these origins, one a centromere -associated origin and the other an 113 origin on a chromosome arm , uncovered a short ~30 bp essential region , and massive parallel 114 mutational analysis revealed that Y. lipolytica origins of DNA replication are heterogeneous. 115 Structural studies of ORC and Cdc6 bound to the two different origin DNA sequences 116 demonstrated that, unlike S. cerevisiae, Y. lipolytica origin recognition required both ORC and 117 Cdc6 for base -specific interactions and hence origin recognition, with some protein -DNA 118 interactions varied between the two origins. The results show a surprising plasticity in origin 119 sequences, structure, and recognition in different eukaryotes. We discuss the evolution of origin 120 recognition and specificity. 121 122

123 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 4 Cryo-EM structure of the human ORC–DNA–CDC6 complex 124 125 The human ORC and CDC6 bound to DNA (HsODC) was reconstituted by combining HsORC1–126 5, HsCDC6, and DNA in vitro. Since the sequence specificity of HsORC remains unknown, a 127 defined DNA 60 base -pair fragment with 70% G /C content was selected for complex assembly 128 after confirming ORC binding through biochemical assays. The 2.6 Å -resolution cryoEM map 129 (Figures 1A, S1A and S1B ) enabled the unambiguous placement of all five ORC subunits, 130 HsCDC6, and the bound DNA. The N-terminal regions of HsORC1 (amino acids 1-465), HsORC2 131 (aa 1-165), and HsCDC6 (aa 1 -151) proteins, which consist of intrinsically disordered regions 132 (IDRs) are disordered and therefore not visible. Of the 60 -bp DNA used , 29-bp were built with 133 confidence, and density corresponding to four ATP analogs was clearly resolved at the conserved 134 nucleotide-binding sites within ORC. Local resolution analysis demonstrated that the AAA+ 135 (ATPases Associated with diverse Activities) core of the complex, comprising RecA-like domains 136 from each ORC subunit and HsCDC6, was better resolved than the central DNA, which exhibited 137 flexibility and correspondingly lower resolution. 138 139 Overall architecture . HsODC adopts a closed -ring conformation, in which all six protein 140 components encircle a centrally located DNA duplex. The complex has a two -tiered layered 141 appearance, with the AAA+ domains forming one layer and the winged -helix domains (WHD) 142 forming a second layer. Each WHD from one subunit sits atop the AAA+ domain of a neighboring 143 subunit (Figure 1B, Suppl. Video 1). The HsODC structure is very similar to HsORC1 -5 bound 144 to DNA that copurified from the expression host cells (PDB ID: 7JPS) 65, with HsCDC6 closing the 145 ring around the DNA. The overall RMSD between the two structures is 4.0 Å over backbone 146 atoms, without the DNA and HsCDC6, though the RMSD between individual subunits is 147 considerably lower (between 0.6 and 1.5 Å). The HsCDC6 AAA+ domain is nestled between the 148 AAA+ domains of HsORC1 and HsORC2, and its WHD sits on top of the AAA+ domain of 149 HsORC1. The higher resolution of this structure , compared to ORC -DNA alone 65,66, brings 150 additional features into view. The HsORC2 WHD was not visible in the HsORC structure but there 151 is clear density for this domain in the HsODC structure, which is situated above the RecA domain 152 of HsCDC6. There is a slight widening of ~4.2Å at the interface between HsORC1 and HsORC2, 153 creating sufficient space to accommodate HsCDC6. This local broadening allows the RecA 154 domain of HsCDC6 to insert into the gap, where it establishes contacts with both HsORC1 and 155 HsORC2. An additional ATP -binding site is formed between HsCDC6 and HsORC1, as is the 156 case for ScODC and DmODC 67–69 . 157 158 DNA bending. Previous studies in S. cerevisiae have shown that DNA bending mediated by 159 ORC is important for origin licensing and subsequent MCM loading 17,25,31,70 . In S. cerevisiae, 160 replication origins contain an A/T -rich ARS consensus sequence (ACS) and ORC bends DNA 161 downstream of the ACS site by 40–55° relative to the axis of ORC’s central DNA-binding channel. 162 This bending is primarily driven by a basic amino acid patch within the Orc5 subunit (ORC5-BP), 163 which contains a long loop (AA 350 - 370) enriched with glycine and alanine residues, which 164 confer flexibility, along with basic amino acids that extend into the DNA minor groove. In our 165 HsODC structure, we observe clear density for only two HsORC5 arginine residues that make 166 limited contacts with the DNA backbone and are at the very beginning of this highly flexible loop 167 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 5 (Figure 1C ). The rest of the loop is disordered ( Figure S1C). These interactions might be 168 insufficient to achieve the degree of bending seen in S. cerevisiae ODC, resulting in a more 169 modest bend of ~20° relative to the axis of the central DNA -binding channel. This difference 170 suggests that the mechanism of DNA engagement and bending by human ODC may either be 171 inherently less pronounced than in yeast, potentially reflecting species -specific adaptations in 172 origin recognition, or that reduced bending may reflect the fact that the DNA used is not a bona 173 fide origin. 174 175 ORC-CDC6 engages DNA through backbone interactions with all protein subunits. Detailed 176 examination of the cryo-EM structure revealed that the five subunits of human ORC ( HsORC1–177 5) and HsCDC6 engage directly with the DNA duplex (Figure 1C and S1D). The majority of these 178 interactions are mediated through the RecA-like domains of the ORC subunits. Notably, only the 179 WHDs of HsORC3 and HsORC5 make direct contacts with the DNA, rather than all of them in the 180 case of ScODC, suggesting a specialized role for these WHDs in stabilizing DNA binding within 181 the closed -ring architecture during the assembly of the pre -Replicative Complex. The high -182 resolution of the cryo-EM map enabled detailed mapping of the protein–DNA interface within the 183 ODC. HsORC1 engages the DNA through residues T593 and H596, positioned to interact with 184 the phosphate backbone near the DNA entry point of the complex. HsORC2 contacts DNA via 185 T365 and R367. HsORC3 contributes a cluster of residues—R641, K697, and Q698 that interact 186 with the DNA backbone. Like all other ORC subunits, HsORC4 also exhibits a DNA-binding 187 interface, with residues K127, F129, S131, and T391 forming a broad contact surface that likely 188 plays a role in anchoring the DNA and stabilizing the complex (Figure 1C). 189 190 A particularly notable feature is the DNA -binding mode of HsORC5, which utilizes its WHD to 191 engage the DNA through a cluster of basic residues —R322, R326, and Y432. These residues 192 form a distinct basic patch that establishes strong electrostatic interactions with the phosphate 193 backbone of the DNA. This interaction likely contributes to the bending of the DNA toward the 194 protein surface. Such localized bending may play a critical role in the structural remodeling of the 195 origin DNA, enabling the recruitment and loading of downstream replication factors, such as the 196 MCM2–7 helicase, as shown for ScODC 70. HsCDC6 also participates in DNA binding, 197 contributing contacts via residues T236, T238, and Q240 to the phosphate backbone of the DNA. 198 These protein -DNA interactions across all six subunits establish a high -affinity DNA -binding 199 surface that facilitates origin recognition and pre-Replicative Complex assembly. 200 201 A direct interaction with a DNA base. Since sequence-specific origins have yet to be identified 202 in metazoans, we initially expected to observe only non -specific interactions with the DNA 203 backbone. However, residue R367 from HsORC2 appears to extend into the minor groove of the 204 DNA and interact with a nitrogen base edge (Figure 1D and S1E). Although the cryo-EM density 205 for the full guanidinium group of the arginine side chain is incomplete, the visible portion is 206 sufficient to model its orientation and infer that it is positioned to form hydrogen bonds with the 207 adenine and adjacent thymine base (Figure S1E). This type of contact is noteworthy, as minor 208 groove base interactions can contribute to sequence-preferential recognition, even in proteins or 209 DNA that are not strictly sequence -specific. The neighboring residue, T365, binds the DNA 210 phosphate backbone and likely stabilizes the orientation of R367, effectively “locking” it into 211 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 6 position. Although the DNA used in this study is not derived from a bona fide origin sequence, 212 the structural arrangement of T365 and R367 suggests a potential mechanism by which HsORC2 213 could engage in limited DNA sequence-dependent recognition. 214 215 Nucleotide binding sites . The HsODC complex was assembled in the presence of the non -216 hydrolyzable ATP analog AMPPNP. As in most AAA+ ATPases, including ORC, nucleotide 217 binding occurs at the interface between adjacent subunits, where conserved motifs from 218 neighboring proteins contribute to the formation of the nucleotide-binding pocket 71 . Upon addition 219 of CDC6 to the human ORC complex, we observed additional density corresponding to a 220 nucleotide at the interface between CDC6 and ORC1, with HsORC1 R670 serving as the arginine 221 finger, and HsCDC6 residues R388 and K208 coordinate the β- and γ-phosphates of ATP. The 222 cryo-EM densities for all four -nucleotide binding regions were sufficiently well -resolved to allow 223 modeling of a magnesium ion coordinated near the ATP analog (Figure S1F). 224 Genome-wide identification of origins of DNA replication in Yarrowia lipolytica 225 To investigate the evolution of origin recognition , we identified origins of DNA replication in 226 Yarrowia lipolytica (Figure 2A). A strain of Y. lipolytica was constructed that expressed the Herpes 227 Simplex Virus thymidine kinase (HSVTK) and the human Equilibrative Nucleoside Transporter 1 228 (ENT1) proteins to enable incorporation of the thymidine analog 5-ethynyl-2′-deoxyuridine (EdU). 229 The temporal dynamics of DNA replication in Y. lipolytica was investigated by EdU -labeling 230 following synchronization of cells by nutrient starvation and release into the cell division cycle by 231 re-feeding. EdU -positive cells were visualized using fluorescence microscopy and quantified 232 alongside budding index as an indicator of S-phase entry (Figure S2A and S2B). The peak of S 233 phase under these conditions was 60-75 minutes post release, however, when the DNA synthesis 234 inhibitor hydroxyurea (HU) was added, replication progression was slower, but more synchronous 235 (Figures 2A and S2C). 236 The whole genome landscape of DNA replication in Y. lipolytica was determined by continuous 237 labeling of DNA replication with EdU in the presence and absence of HU and harvesting the cells 238 at different times post release. Compared to S. cerevisiae , where 200mM HU inhibits DNA 239 replication and cell viability, Y. lipolytica is very sensitive to HU . While in the presence of 5mM 240 HU, cells can still progress through the cell cycle, higher levels of HU inhibit cell proliferation and 241 cell viability. Labeled DNA was detected by sequencing and mapped to a high-quality genome 242 assembly 72 . A total of 634 replication origin peaks were identified with much sharper peaks when 243 HU was added due to checkpoint inhibition of replication fork progression , allowing better 244 definition of the temporal activation of origins during S phase (Figure 2A). Importantly, HU did 245 not alter origin location. Early replicating origins were first detected at 30 minutes post release 246 and the latest active origins appeared at 120 minutes. Of the 634 replication origins identified 247 genome-wide, 289 were classified as early -firing under HU treatment. Most origins were in 248 intergeneic regions of the genome, with some at transcription start sites (Figure 2B). 249 Importantly, the spatial distribution of timing of activation of replication origins across the 250 chromosomes was not distributed uniformly, but instead formed discrete 150–300 kb clusters of 251 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 7 origins that were activated at the same time (Figure 2C), closely resembling the replication timing 252 domains described in the chromosomes of higher eukaryotes 73,74. 253 Genetic analysis of two Yarrowia origins of DNA replication 254 The 634 origins were annotated by referring to the chromosome (A through F) and counting each 255 origin from the left telomer e to the right telomere. For example, OriA-006 is a newly discovered 256 origin that is the sixth from the left telomere o n chromosome A. To genetically characterize the 257 DNA sequences under the EdU -seq peaks, we tested two origins using ARS assays: the 258 previously characterized, centromere-associated OriC-061 (previously called ARS18 or Ori3018) 259 61,62 and OriA-006. Unlike ARSs in S. cerevisiae in which an origin can support high frequency 260 transformation (HFT) and plasmid stability, Y. lipolytica requires both a centromere and an origin 261 of DNA replication be present on the mini-chromosome (Figure S3A) 61. Both OriA-006 and OriC-262 061 supported robust plasmid replication when cloned into an Ori⁻/CEN⁺ plasmid backbone , 263 including equivalent ARS activity (both HFT and plasmid stability) when placed in either 264 orientation or at variable distances from the centromere (Figure S3B). In contrast, two 600 bp 265 fragments not associated with EdU-seq peaks, one derived from a coding region on Chromosome 266 D and the other from a non -coding intergenic region on Chromosome E, could not support ARS 267 activity (Figure S3C). These findings demonstrate that not all genomic sequences can function 268 as replication origins, highlighting the requirement for specific DNA elements. Furthermore, they 269 reinforce the reliability of EdU-seq in identifying biologically active replication origins and provide 270 a foundation for dissecting the sequence and structural features critical for origin function in Y. 271 lipolytica. 272 Linker scan mutagenesis 75 overcomes concerns that deletion mutations can alter the spacing of 273 essential DNA sequences and was previously employed to dissect the structure of S. cerevisiae 274 origins 22–24. We therefore used linker scanning to analyze the sequence requirements of Y. 275 lipolytica origins within the 600 bp OriC-061 and OriA-006 fragments (Figure 3A and 3B). Each 276 mutant was screened for high-frequency transformation (HFT) and plasmid stability. In OriC-061, 277 45 linker mutants were tested and substitutions at positions 2 –6 severely reduced or abolished 278 HFT and plasmid stability, identifying this region as critical for replication initiation (Figure 3A). 279 Mutations outside this core retained wild -type–like stability (~41%), suggesting they are 280 dispensable for origin function. For OriA-006, linker insertions between positions 25–29 reduced 281 or eliminated both transformation efficiency and plasmid maintenance (Figure 3B). Together, 282 these results demonstrated that replication origin activity in Y. lipolytica is dependent on a short 283 ~30 bp sequence of DNA. These sequences are sufficient for origin activity since for each origin, 284 a 50 bp fragment supported ARS activity in the presence of a CEN (Figure S3D). 285 Structures of Yarrowia ORC-Cdc6 Bound to Different Origin DNAs 286 Origin DNA sequences are required for Cdc6 DNA binding. A Cdc6 DNA-binding assay using 287 size exclusion chromatography (SEC) was developed to test whether specific DNA sequences 288 were required for Cdc6 to co -elute with ORC. A 54 bp fragment of OriC-061 or a scrambled 289 version of the DNA with the same G/C content was used. Wild-type OriC-061 DNA and some 290 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 8 Cdc6 co-eluted with ORC (Figure 4A top, fractions 3 and 4 ), however, when the scrambled 291 DNA was used, ORC binding to DNA was greatly reduced and Cdc6 no longer co -eluted with 292 ORC (Figure 4A, bottom). This was validated by mass photometry (Figure S4A). 293 294 Structure of YlORC-DNAOriC-061-YlCdc6 suggests sequence-specific binding. Cryo-electron 295 microscopy (cryo-EM) was used to gain insight into the mechanism of YlORC origin binding. Using 296 the 54bp DNA fragment of OriC-061 in complex with YlORC and YlCdc6, a 2.7 Å resolution 297 structure of the YlORC-DNA54bpOriC-061-YlCdc6 (ODC) complex was obtained (Figure 4B, Suppl. 298 Video 2 and Video 3 , Figure S4B ). The winged-helix (WHD) and AAA+( -like) domains of all 299 YlOrc1-5 and YlCdc6 proteins were visible in addition to the second TFIIB domain and C-terminal 300 α-helix of YlOrc6 (Figure S4C). A somewhat weaker and less defined density for Orc2 -WHD 301 indicates flexibility of the domain while bound to DNA and Cdc6. In addition, neither the Orc1 302 bromo-adjacent homology (BAH) domain nor the N-terminal TFIIB domain of Orc6 were visible. 303 304 Akin to many AAA+ protein complexes and all ORC -Cdc6 structures determined to date 67–69, 305 Orc1-5 and Cdc6 form a two -tiered hexameric ring with ATP -binding sites between the AAA+ 306 domains of Cdc6/Orc1, Orc1/Orc4, Orc4/Orc5, and Orc3/Orc5, along with the contacts between 307 the AAA+ domains of the other subunits forming the first tier (Figure S4D). ATP and Mg2+ were 308 observed in the ATPase binding sites between Orc1/Orc4 and Cdc6/Orc1, while the Orc3/Orc5 309 and Orc4/5 sites contained ADP in the YlORC -DNA54bpOriC-061-YlCdc6 maps generated from gel 310 filtration-derived samples (Figure S4D). The winged-helix domain of each subunit sits atop the 311 adjacent subunit’s AAA+ domain, forming a domain -swapped second tier. Similar to the S. 312 cerevisiae ODC (ScODC) structures 67,69, the C-terminal TFIIB domain (TFIIB-B) and C-terminal 313 α-helix of Orc6 are visible. Like ScOrc6, YlOrc6 makes limited contacts with DNA and binds to 314 the complex in multiple places: the TFIIB -B domain contacts part of the Orc2 N -terminal coil 315 (residues 106-165), the WH domain of Orc3, and a small portion of the Orc5 basic patch (Orc5 -316 BP, residues 348-364), and to the Orc3 protrusion with Orc6’s C-terminal α-helix. Differing from 317 previous ODC structures , two small segments of the Cdc6 N -terminal IDR are bound to Orc1: 318 Cdc6[1-13] binds to the exterior surface of the complex between Orc1-AAA and Orc4-AAA, while 319 Cdc6[13-20] binds near the interface of Orc1 -AAA and Cdc6 -WHD (Figure S4E). Due to the 320 partial occupancy of both segments within the density map, we suggest that this region can bind 321 to Orc1 in either of the two conformations. 322 323 The sharpness of the DNA in the cryo -EM map was immediately evident, with purine and 324 pyrimidine densities at each respective position easily discernible, and base identities apparent 325 at most positions, indicating that the DNA is positioned in a discrete manner relative to the protein, 326 implying that YlORC -Cdc6 binds to Ori sequences in a DNA sequence-dependent manner 327 (Figure S4 F). The pattern of identifiable bases was u sed to define its register. The DNA is 328 significantly bent, similar to the DNA in ScODC structures 67,69, with a 40° bend occurring near the 329 interface between the WHD and AAA+ domains of the complex. 330 331 YlORC/YlCdc6 binds Ori DNA specifically . From the structural analysis, the interactions 332 between ORC/Cdc6 and DNA can be grouped into three sequence elements: A central region 333 with major/minor groove and phosphate backbone contacts comprised predominantly of Orc4 and 334 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 9 Cdc6 side chains we coined the Orc4/Cdc6 -interacting element, the AT element consisting of 335 minor groove and backbone interactions involving multiple subunits and a water -mediated 336 hydrogen bonding network at one end of the origin, and the Orc5 basic patch (Orc5-BP) element 337 on the opposite side of the origin, with minor groove, backbone, and water -mediated hydrogen 338 bonding interactions carried out by Orc2 and Orc5 (Figure 4C). 339 340 The Orc4/Cdc6 -interacting element is proximal to the large bend in the DNA and a site of 341 significant minor groove compression. K465 of YlOrc4 makes sequence-specific contacts with the 342 carbonyls of G31 and G32 in the major groove of the Y strand DNA (Figures 4C and 4D, bottom 343 right). This lysine emanates from an a-helix in the insertion loop, similar in location to the α-helix 344 of S. cerevisiae Orc4 (ScOrc4) in the ScODC complex that inserts itself into the major groove of 345 the DNA for sequence -specific binding. The YlOrc4 insertion helix is considerably smaller than 346 that of ScOrc4 (Figure S4G). It is angled so that a stretched K465 side chain reaches into the 347 DNA major groove , and salt-bridge interactions between K462, R466, and D474 stabilize the 348 helix. YlOrc4 K465 appears to form the only apparent sequence -specific interaction in the Orc4 349 insertion helix. Unlike the ScOrc4 insertion helix, which is highly conserved among fungi that are 350 predicted to bind origins in a similar manner to S. cerevisiae, this lysine is not conserved (Figure 351 S4G). 352 353 Near the YlOrc4 insertion helix resides a unique extended loop region of the Cdc6 WHD (Figure 354 4D). Sequence alignments to other eukaryotic Cdc6 proteins show that a large portion of the loop 355 comes from an insertion that is conserved within the family Dipodascales, although a similar 356 insertion may have separately evolved in more distant fungi such as Neurospora crassa (Figure 357 S4H). The density of this extended loop region is sharp and shows the loop interacting with the 358 DNA over an entire turn of the double helix, making major/minor groove and backbone contacts 359 throughout. Starting from the N -terminus of the loop, the sidechain of K548 is inserted into the 360 minor groove and hydrogen bonds with the O2 carbonyl of Y-T29. R557 reaches into the major 361 groove of the DNA and interacts with Y-G34 (Figure 4D). The DNA-binding loop appears to be 362 held in place by several electrostatic interactions with the DNA extending from the Orc4/Cdc6 -363 interacting element into the AT element. From these interactions, a structure -based preliminary 364 binding motif was constructed centered around the Orc4/Cdc6-interacting element motif, deemed 365 to be 5’-CNCCNRH-3’, where N denotes any nucleotide, R denotes purines, and H is not G. 366 367 Located upstream of the Orc4/Cdc6 -binding element, the AT element lacks major groove 368 interactions and is recognized by a network of backbone, minor groove, and water -mediated 369 hydrogen bond interactions involving all ODC subunits except Orc6 ( Figure 4 C). The most 370 prominent sequence-dependent interaction in this element is between the Orc3 R220 sidechain 371 and N3 of A15 via a coordinated water, and is further stabilized by an aspartate (D218) (Figure 372 S4I). The positioning of Orc3 R220 likely precludes a C -G base pair at the adjacent position 16 373 due to potential steric clashes between R220 and the minor groove amine of a guanine. The 374 sidechain of Orc4 K168 enters the minor groove near position 14 and may interact with the N3 375 nitrogen of A14, but the sidechain density is weak (Figure S4I). Adjacent to the AT element, the 376 Orc1[300-305] basic patch is visible, with a mix of backbone and minor groove interactions. In 377 addition, Orc2, Orc5, and Cdc6 all make contact with the phosphate backbone in this region. 378 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 10 These interactions extend th e binding motif in the 5’ direction of OriC-061 to 5’ -379 ATNNNNCNCCNRH-3’. 380 381 On the opposite side of the Orc4/Cdc6 -binding element are the sites of two clusters of protein -382 DNA interactions: one encompassing X-strand positions 27 and 30 to 33 mediated predominantly 383 by Orc3, and another mediated by Orc2 and the Orc5 basic patch (Orc5-BP) from positions X-35-384 41 (Figure S4J). Apart from backbone interactions, Orc3 R672 forms a minor groove H-bond with 385 T25 O2 carbonyl on the Y strand. Orc5 R357 forms a hydrogen bond with the O2 carbonyl of T16, 386 while R362 forms a hydrogen bond to the carbonyl of T38 (Figure S4J). The other significant 387 interaction in the Orc5-BP element is an H-bond between Orc2 R150 and G35 in the major groove. 388 With the addition of the Orc5-BP element, the inferred binding motif of YlORC and YlCdc6 to OriC-389 061 extends in the 3’ direction to 5’-ATNNNNCNCCNRHNNNNNNNGNNYR-3’, where Y denotes 390 a pyrimidine. 391 392 Comparison of YlODC with two different origin sequences provide a consensus 393 recognition sequence. To determine whether the interactions between YlORC/YlCdc6 and DNA 394 are consistent between different origin sequences, a 2.6 Å cryo-EM structure was determined 395 using a 60-bp fragment from OriA-006, the other origin validated by the linker scan assay (Figure 396 S4K, S4L). Many of the critical interactions remained the same (Figure S4M), with a Ca RMSD 397 of 0.60 Å between the two structures, however, an excess of ATP was used for this cryo -EM 398 sample and ATP is now visible in each ATP binding site. 399 400 In the AT element, the sidechain of Orc4 K168 appears to have stronger density overall compared 401 to YlODC54bpOriC-061, but still appears to have multiple conformations. Unlike YlODC54bpOriC-061, the 402 lysine sidechain could hydrogen bond to a water between K168 and the T51 O2 carbonyl on the 403 Y strand, the O2 carbonyl of T11, or a water between the lysine amine and the N3 nitrogen of 404 A50. The most significant change in the Orc5-BP element is the loss of Orc2 R150 density in the 405 map with the change of G35 to a cytosine, consistent with the loss of this interaction. 406 407 The overall conformations of the Orc4 insertion helix and Cdc6 DNA-binding loop are very similar 408 between the two origins. However, the positioning of the Cdc6 R557 sidechain is different: with, 409 the arginine H-bonding with G34 on the Y strand in the major groove in OriC-061, while with OriA-410 006, R557 interacts with G44 on the Y strand, the equivalent of one base pair away (Figure S4N). 411 This would then alter the Orc4/Cdc6 element motif from CNCCNRH in OriC-061 to CNNCCNRH 412 in OriA-006, pointing towards added flexibility for Orc4/Cdc6 binding requirements. With the 413 observed changes at other elements included, we suggest the structure -based YlORC - and 414 YlCdc6-binding sequence 5’-ATNNNXXNCCNRHNNNNNNNNNNYR-3’, where at least one X is 415 a cytosine. 416 417 Mutation of origin sequences disrupt ODC assembly 418 419 The importance of the sequence -specific interactions was determined by measuring the effects 420 of origin mutations on in vivo genetic assays, including colony formation and plasmid stability 421 (Figure 5A) and a biochemical assay for DNA/Cdc6 binding to ORC (Figure 5B). The first series 422 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 11 of mutant ori fragments tested were in the Orc4/Cdc6-interacting element of OriA-006, as the site 423 involves multiple major groove interactions which could confer base specificity . Changing 424 individual cytosines (C17A or C21A) along with the complementary base in the motif does not 425 affect the phenotype in vivo, but mutation of two cytosines [C20A, C21A] leads to smaller colonies, 426 a 50% reduction in plasmid stability , and reduced Cdc6 loading (Figure 5A, 5B and S5A). All 427 triple mutants (CA, CG, or CT at positions 17, 20, and 21) are inviable in vivo (Figure 5A 428 and S5A) and mutation of all three cytosines to guanines greatly reduced Cdc6 binding to ORC-429 DNA (Figure 5B and S5A).The lack of sensitivity to the X-strand position 17 mutant alone could 430 be caused by the flexible nature of the Cdc6 R557 sidechain coordinating it, as it can potentially 431 shift to the adjacent position as seen in OriC-061 while still affecting binding specificity (Figure 432 S4N). Changes to the Ori sequence at position 23, the site of the Cdc6 K548 interaction, led to a 433 large decrease in the number of colonies, a 90% decrease in plasmid stability once returned to 434 non-selective media (Figure 5A), and a sharp reduction of Cdc6 loading in vitro (Figure 5B), 435 indicating the importance of this minor groove interaction for proper origin licensing. Mutations in 436 other regions of the Ori sequence had similar effects. Both A12G and T13C mutations to the AT 437 element of OriA-006 (Figure 5B) result in a decrease in in vitro Cdc6 loading (Figure 5B). 438 439 To study the structural effects of Ori sequence mutations on the binding of the ODC, a 2. 6 Å 440 resolution cryo-EM structure was determined for the double cytosine to guanine mutant [C20G, 441 C21G] of OriA-006 (YlODC60bpOriA-006-CNNGGNR) (Figure S5D,E). The structure is nearly identical to 442 YlODC60bpOriA-006-WT, with the largest change occurring at the Orc4 α-helix (Figure 5C ). The 443 secondary structure of this helix unravels and becomes a structured loop. Orc4 K465, which 444 contacts consecutive guanines in the WT OriA-006, points away from the DNA, and Orc4 D474 445 is unable to form a stabilizing salt bridge, as Orc4 R466 now forms multiple H -bonds with the 446 mutant guanine, now located on the opposite strand. The Cdc6 DNA -binding loop appears to 447 have no noticeable changes to its structure, suggesting that Cdc6 loading and specificity might 448 be independent of Orc4 sequence-specific binding. 449 450 DNA deformability and bendability are critical for DNA replication 451 452 Compared to the strict sequence requirements for origin licensing in S. cerevisiae, Y. lipolytica 453 has fewer essential base-specific contacts in the Orc4/Cdc6-binding element to allow for origin 454 licensing. Are there other factors that could play a role in origin licensing specificity? Noticeable 455 in the YlODC structure was a significant compression of the minor groove to allow the DNA to 456 bend. If the bendability of the Ori has a significant role in origin licensing specificity, replacing 457 segments of the bent region of the Ori DNA with a rigid dA tract should inhibit origin licensing 76,77. 458 The structural properties of DNA have been suggested previously to play a part in Drosophila 459 ORC binding preference, and more recent studies have shown that bending of DNA via the Orc5-460 BP was critical for DNA replication in S. cerevisiae 25,70,78 . To test this, we examined the effect on 461 DNA binding, Cdc6 loading, and the in vivo effects of incorporating a 6 nt-long dA tract around 462 the region of OriA-006 that is bent in the YlODC structure, starting at position 23 (abbreviated as 463 A[23-28]) and moving it downstream 1 bp at a time (Figure S5B). 464 465 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 12 There were discrepancies between the Cdc6 loading assay and the in vivo phenotypic effects of 466 the 6A mutants tested (Figure S5B and S5C). At A[24-29] a decrease in Cdc6 loading was seen, 467 but no deleterious effects were observed in vivo. Mutant A[25-30] had >85% of WT OriA-006 468 activity, Cdc6 loading efficiency increased, whereas with A[26-31] and A[27 -32] Cdc6 loading 469 assay results had near WT Cdc6 loading efficacy, but either an extreme reduction or a complete 470 lack of growth was observed in vivo. Insertions from A[28-33] through A[31-36], showed complete 471 inhibition of Cdc6 loading and no growth in vivo, as the 6A tract enters the Orc5 -BP element. 472 Decreases in the 260/280 ratio, indicating reduced DNA binding, were seen in the A[28 -33] 473 through A[31-36] mutants compared to WT. 474 Large-scale mutational analysis of Yarrowia Origins of DNA Replication 475 To identify sequence preferences required for origin function, a Massively Parallel Origin 476 Selection (MPOS) assay was performed using mutagenized libraries (~15% mutation density 35) 477 spanning 90 bp regions of OriC-061 and OriA-006. These libraries were cloned into Ori⁻/CEN⁺ 478 plasmids and introduced into Y. lipolytica for growth-based competitive selection. Deep 479 sequencing of plasmids from pre-selection and post-selection time points was performed, and a 480 quantitative model was trained using MAVE-NN to predict post-selection enrichment as a function 481 of DNA sequence 79. 482 In OriA-006, the most striking changes occurred between positions 10 and 37, where specific 483 nucleotides showed strong selection signatures, highlighting this region as a functionally critical 484 core of the origin, which aligns with the linker scan sensitive area (Figure 6A, the numbers refer 485 to the numbering of base pairs in Figure S4M ). A similar pattern was observed for OriC-061, 486 though with slightly lower resolution due to reduced transformation efficiency and higher 487

signal (Figure S6A), confirming that both sequence preference and positional 488 sensitivity are conserved features in this origin. 489 Bases critical for ORC–CDC6 binding showed enrichment when unmutated and depletion when 490 altered, reflecting their essential role in origin establishment. Conversely, mutations that allowed 491 initial transformation but led to reduced plasmid stability highlighted bases important for 492 maintenance. This dual-phase insight was central to parsing the functional logic of ARS activity 493 in Y. lipolytica. 494 A conserved sequence motif, 5’-YATRNNNNNNCNAWTTNNNNNNYNYAA-3’, emerged from 495 MPOS assay analysis as a central feature of functional origins. Targeted mutagenesis within the 496 YAT, YNYA, and flanking regions revealed key nucleotide requirements for Ori function (Figure 497 6B). Mutation in a critical position such as A 12-to-C within the YAT region abolished colony 498 formation, indicating a loss of Ori activity. Multiple substitutions in the YNYA region , including 499 mutations at posit ions 33, 34, 35 and 36 resulted in either reduced plasmid instability or no 500 transformation, suggesting that these positions are critical for origin activity. These findings align 501 with MPOS results, which showed selection against guanine (G) at T33, reinforcing a preference 502 for thymine (T) or cytosine (C). This preference likely reflects the structural compatibility of 503 pyrimidines with ORC binding, particularly through minor groove interactions involving ORC3, 504 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 13 ORC2, and ORC5. For example, the YA motif appears critical for minor groove engagement by 505 ORC5, and substitutions that disrupt this local DNA shape abolish origin function. 506 Cross-Species Validation of a Replication Origin Motif Using ROC Analysis . Next we 507 evaluated the predictive power of MPOS-identified motifs by evaluating their ability to distinguish 508 functional origins from genomic background. The motif derived from the Y. lipolytica MPOS assay 509 exhibited a remarkable ability to distinguish EDU -seq peaks from randomly sampled DNA from 510 the Y. lipolytica genome, yielding an AUROC of 0.80 (Figure 6C). By comparison, an AUROC of 511 0.91 was obtained when an analogous motif was inferred from S. cerevisiae MPOS data of 28 and 512 used to distinguish sequences in OriDB 18 from random DNA in the S. cerevisiae genome (Figure 513 6C). The predictive power of the Y. lipolytic a motif is considerably less than that of the S. 514 cerevisiae motif. This difference may be due to reduced motif accuracy arising from the lower 515 transformation rate of Y. lipolytica relative to S. cerevisiae, or from the reduced number of variant 516 sequences used in the MPOS assay (7,000 synthetic sequences in Y. lipolytica versus millions of 517 randomly mutagenized sequences in S. cerevisiae). Alternatively, Y. lipolytica origins may be 518 more variable than S. cerevisiae origins, so much so that some origins do not contain a match to 519 the motif. 520 An analysis of nucleotide frequencies across 54 Y. lipolytica early origins of replication compared 521 to the frequencies across 5351 close matches to the motif in non-origin sequences showed a 522 skew for T/A sequences 5’ to the core consensus sequence and a skew for A/T base pairs 3’ to 523 the core consensus sequence only in early origins , but not in non -origins (Figure S6B). This 524 suggests that the arrangement facilitates initiation of DNA replication in the generally G/C rich 525 genome since the average %G /C in 160 bp around motif in early origins is 44%, whereas the 526 Yarrowia genome is on average 48.9% G/C 72 . 527

528 The Yarrowia lipolytica genome has 634 origins of DNA replication that are distributed into large 529 replication timing domains of early and late replicating regions, much like the A and B replication 530 timing domains of vertebrate chromosome replication 80,81. Interestingly, both centromeres and 531 telomeres replicate early in Y. lipolytica. The early and late re plicating domains in vertebrates, 532 including in human cells, correspond to topologically associated domains (TADs) and 533 euchromatin and heterochromatin, respectively. What determines the structure of the Y. lipolytica 534 replication timing domains remains to be determined, but the global temporal pattern of replication 535 in this yeast is very different from the replication timing in S. cerevisiae, which falls into two classes 536 based on temporal control by the S-phase cyclin-dependent protein kinases 82. 537 Origins of DNA replication in S. cerevisiae are specified by ORC recognizing the ARS consensus 538 sequence 5’ -WTTTAYRTTTW-3’and bending the DNA 17,31,83. After interacting with ORC, t he 539 Cdc6 initiator-specific motif (ISM) and WH domains bind to the DNA phosphate backbone (but do 540 not form base-specific interactions), thereby contributing to origin DNA binding 67,69. Thus, in S. 541 cerevisiae, the location of origins in the genome is primarily due to ORC. The ScOrc4 and ScOrc2 542 base-specific contacts are major contributors to DNA-sequence-specific interactions, so much so 543 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 14 that it was predicted that the presence of the Orc4 α-helix correlated with DNA sequence-specific 544 binding 28,32. But S. cerevisiae ORC also has minor groove and backbone interactions that 545 contribute to ORC -DNA binding. Clearly, however, analysis of origin recognition in Y. lipolytica 546 suggests that the re are alternative mec hanisms for base-specific origin recognition because in 547 this yeast both YlORC and YlCdc6 are required . The Y. lipolytica consensus sequence 5’-548 ATNNNXXNCCNRHNNNNNNNNNNYR-3’, supported by both structural data and high -549 throughput mutagenesis data, is substantially different from the S. cerevisiae sequence, both in 550 length and flexible base composition. The YlOrc4 α-helix is substantially reduced and only a single 551 lysine (K465) contacts adjacent G/C base pairs in the 5’-XXNCCNRH-3’ Orc4/Cdc6-interaction 552 motif. This α-helix is completely missing in the human ORC4 subunit, suggesting evolution of the 553 Orc4 protein toward increasing sequence-specific origin recognition in the yeasts (Figure 7 ). 554 Mutation of single YlOrc4 interacting bases did not affect Y. lipolytica origin activity (Figure 5A), 555 in contrast to single base changes in the ScORC binding site significantly compromising origin 556 activity 84. When YlOrc4 two adjacent interacting base pair s are mutated, the Orc4 α-helix 557 collapses and the base-interacting K465 residue, switches with R466, which now makes base-558 specific interactions, albeit with guanines now located on the opposite strand (Figure 5C). Thus, 559 there is considerable plasticity in ORC binding to origins in Y. lipolytica since the different origins 560 display alternative protein-DNA interactions. 561 Unlike the situation in S. cerevisiae, the YlORC bound weakly to origin DNA and YlCdc6 increased 562 the specificity and affinity for origin DNA by making several contacts, including a base-specific 563 interaction via R557 to the Orc4/Cdc6-interacting element, adjacent to the Orc4 α-helix interaction 564 site (Figure 4D). Interestingly, a single base change in the base in the major grove of either OriA-565 006 or OriC-061 that interacts with YlCdc6 R557 does not affect origin function. This may be due 566 to the flexibility of the YlCdc6 R557 ability to interact with neighboring base-pairs in the major 567 groove, depending on the origin (Figure S4N). In contrast, in both origins, mutation in the minor 568 groove base pair that interacts with YlCdc6 K548 severely compromises origin activity, 569 underscoring the importance of YlCdc6 in origin recognition. 570 The extensive DNA interactions in the AT and Orc5-BP elements in the consensus sequence 571 contribute to origin function, since single-point mutations in both elements eliminate origin activity. 572 These include base pairs that interact with Orc3 in the AT element and Orc 2 and Orc 5 in the 573 Orc5-BP element. An Orc1 loop ( residues 300-305) interacts with the minor gro ove non-574 specifically in addition to other residues binding the phosphate backbone. A conserved β-loop in 575 this region binds to a minor gro ove in the Drosophila ORC-CDC6-DNA structure, but does not 576 make nucleotide -specific contacts 68. Indeed, unlike Y. lipolytica ORC-Cdc6 interaction s with 577 DNA, all DNA contacts in the Drosophila ORC-CDC6-DNA structure on a 60 bp AT-rich DNA lack 578 base pair specificity. One common feature in all structures, however, is the interaction between 579 Orc5/ORC5 with the DNA to stabilize the DNA bend (Figure 7). As noted previously 25,85–87, it is 580 possible that DNA sequences that have the propensity to bend upon ORC or ORC-CDC6 binding 581 is an essential feature of all origins of DNA replication. 582 It was surprising that, in the structure of the human ORC-CDC6-DNA, we observed that HsORC2 583 residue R367 appears to interact in a minor grove with the edge of a nucleotide, suggesting for 584 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 15 the first time that human ORC-CDC6 might have direct DNA sequence dependent interactions or 585 origin recognition . In an analogous way, the S. cerevisiae Orc2 W396 residue makes p-586 interactions in a minor gro ove of ARS1 origin DNA, contributing to origin specificity. In contrast, 587 in Yarrowia, residues in the non-conserved Orc2 loop interact with the phosphate backbone of 588 DNA, likely contributing to affinity but not specificity (Figure 7). The DNA in the human ORC-589 CDC6-DNA structure is a G/C rich sequence that is not an origin of DNA replication or a known 590 ORC-CDC6 binding site in the human genome, so it is possible that ORC-CDC6 interaction with 591 an authentic origin or binding site might reveal additional base-specific interactions. 592 It is remarkable that for such a fundamental process as specification of origins of DNA replication 593 that there have been many solutions to defining the location of origin sequences in the eukaryotic 594 genome. This variation may be a result of differences in genome organization and gene density, 595 as well as other features such as genome size, G/C content and 3D structure. The Yarrowia and 596 human ORC -Cdc6-DNA structures presented here, together with previous analyses of S. 597 cerevisiae and Drosophila structures have highlighted some key aspects of the evolution of origin 598 recognition and specification. 599

of the study 600 We have only characterized in detail two Y. lipolytica origins and yet there are some sequences 601 under EdU peaks that lack the consensus sequence identified here. Mutagenesis and structural 602 studies with ORC and Cdc6 using these origins may reveal even greater flexibility in origin 603 specification uncovered in this study. Furthermore, continued analysis of the HsODC using 604 different DNA sequences, including known ORC-CDC6 DNA binding sites, may reveal additional 605 base-specific origins recognition in human cells. 606

607 This research was supported by grants from the National Institutes of Health ( GM045436, 608 GM133777, HG011787), the Howard Hughes Medical Institute and the Goldring Family 609 Foundation. Core DNA sequencing was facilitated by the Cold Spring Harbor Laboratory (CSHL) 610 DNA sequencing and analysis shared resource, supported by t he Cancer Center grant 611 (CA13106). We thank Dennis Thomas for managing the CSHL Cryogenic Microscopy Core 612 shared resource, and members of the Stillman and Joshua-Tor laboratories for suggestions and 613 advice. L.J. is an Investigator of the Howard Hughes Medical Institute. 614 Author contributions 615 B.S. and L. J. conceived the study. B.S., L. J., and J.B.K. designed the study. J.B, N.Z., O.P.C. 616 and K.L. performed the experiments. J.B., N.Z., O.E.D, O.P.C., K.L., J.B.K., L.J., and B.S. 617 analyzed the data. N.Z, J.B, O.P .C, J.B.K, L.J. and B.S. wrote the paper with input from all 618 authors. B.S., J.B.K. and L. J. provided funding and oversaw the project. 619 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 16 Declaration of interests 620 The authors declare no conflicts of interest. 621 622 Figure Legends 623 624 Figure 1 Structure of the HsORC–Cdc6–DNA complex. (A) Cryo-EM 3D map of the HsODC 625 complex bound to a 60 -bp DNA duplex, with each protein subunit shown in a distinct color. (B) 626 Cartoon representation of the atomic model fitted into the cryo -EM map, highlighting all protein 627 subunits of HsORC, Cdc6, and the DNA molecule. This view emphasizes the arrangement of the 628 proteins around the central DNA-binding channel. (C) Interactions between HsORC subunits and 629 DNA, shown at the amino acid level. Interacting residues are colored as above, providing a clear 630 visualization of how each subunit contributes to DNA engagement. (D) Close-up view of residue 631 R367 from the HsORC2 subunit, which establishes three distinct contacts with DNA nucleotides, 632 specifically thymine from chain H and guanine and adenine from chain I. 633 634 Figure 2: Genome-wide profiling and genomic context of replication origins in Yarrowia 635 lipolytica. (A) Temporal mapping of replication origin activity throughout Yarrowia 636 lipolytica genome using EdU-seq. EdU-seq signal tracks across a 2.3 Mb region of Chromosome 637 A following release from starvation into S phase in the presence (blue) or absence (red) of 5 mM 638 hydroxyurea (HU). Samples were collected at the indicated time points. HU -treated cells show 639 temporally resolved activation of replication origins, while untreated cells exhibit more extensive 640 EdU incorporation. (B) Enrichment analysis of EdU -seq peaks relative to genomic annotations. 641 Intergenic and promoter/TSS regions are enriched for replication origins, while exonic regions are 642 depleted. Log2(observed/expected) values indicate the degree of enrichment or depletion across 643 different genomic features. (C) Genome-wide origin firing maps across all six Y. 644 lipolytica chromosomes at 30 and 120 minutes post -release in the presence of HU. Each track 645 shows EdU-seq signal along individual chromosomes (A–F), with peaks corresponding to active 646 replication origins. Regions highlighted in black boxes represent early -firing origins (30’), while 647 red boxes highlight later -firing origins (120’). Centromeres (black boxes), rDNA clusters (red 648 boxes), and telomeric regions (orange diamonds) are annotated for reference. 649 Figure 3. Linker -scanning mutagenesis reveals essential regions for origin activity in Y. 650 lipolytica OriC-061 and OriA-006. (A) Mitotic stability assay of OriC-061 mutants carrying XhoI 651 linker substitutions (CTCGAG) at positions 2–45. Images show colony formation following plasmid 652 transformation and selection and percentage (% URA+ retention) for different linker mutations. 653 Wild-type (WT) OriC-061 shows ~41% mitotic stability, while substitutions in linkers 2 to 6 654 drastically reduce origin activity to ≤3%. A zoomed-in view highlights the six critical linker mutants 655 with corresponding colonies on the plate after the initial. transformation, with the percentage 656 plasmid stability values shown. (B) Mitotic stability assay of OriA-006 mutants with linker 657 substitutions at positions 1–44. WT OriA-006 exhibits ~45% mitotic stability. A cluster of mutations 658 (linkers 25 –29) results in complete loss of origin activity (0 –5% stability), identifying a key 659 functional region. Sequences and representative colony phenotypes for these five mutants are 660 shown below the graph. 661 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 17 Figure 4. YlORC and YlCdc6 coordinate to bind origins specifically. (A) SDS-PAGE and 662 agarose gel electrophoresis of gel filtration fractions from samples containing YlOrc1 -6[SST-663 YlOrc1], YlCdc6, and a 54 -bp DNA oligonucleotide derived from OriC-061 (top panel) and a 664 scrambled sequence thereof (bottom pane l). The ag arose gel in the top panel was spliced 665 together as shown in the black boxes , but represents DNA from the same experiment. (B) (left) 666 The 2.7 Å resolution unsharpened map of the YlORC-DNA54bpOriC-061-YlCdc6 complex and (right) 667 the ribbon representation of the derived structure. (C) A diagram of the visible protein -DNA 668 interactions seen in the YlORC -DNA54bpOriC-061-YlCdc6 structure. N indicates any nucleotide; W 669 denotes an A or T; XX denotes a C in either of these positions; R indicates a purine; H indicates 670 A, C, or T; Y indicates a pyrimidine. (D) A ribbon representation of YlOrc4 ( blue) and YlCdc6 671 (pink) near the Orc4/Cdc6 element of OriC-061 in the YlORC -DNA54bpOriC-061-YlCdc6 structure, 672 with insets of the YlOrc4 insertion helix (bottom right) and the YlCdc6 DNA binding loop (top right, 673 bottom left). (E) A top-down representation of the complex near the AT element shows all proteins 674 except YlOrc6 forming electrostatic interactions with OriC-061 in the region. 675 676 Figure 5. Effects of mutations on origin recognition and function. (A) Mutational analysis of 677 the Orc4/Cdc6-interacting element motif in OriA-006 (top) and OriC-061 (bottom) and the effect 678 on origin activity. Images show colony formation following plasmid transformation and selection 679 and percentage (% URA+ retention) for different mutations. Base changes are highlighted in red, 680 and their colony formation is shown in the adjacent images. Boxes represent the linker scan 681 mutations (see Figure 3). (B) SDS-PAGE results of the peak fractions from the Cdc6 loading 682 assay utilizing mutant OriA-006 sequences, indicating the differences in Cdc6 co -elution 683 (numbers refer to the X -strand). (C) Comparison of the structure of the YlOrc4 insertion loop 684 between YlORC -DNA60bpOriA-006-YlCdc6 ( left) and YlORC -DNA60bpOriA-CNNGGNR-YlCdc6 ( right) 685 structures. The yellow bases represent the bases that were mutated in OriA-006. 686 687 Figure 6. Massively parallel origin selection assay and quantitative modeling. (A) 688 Quantitative model for Y. lipolytica origin specificity derived from a massively parallel origin 689 selection (MPOS) assay carried out on 90 bp sequences containing OriA-006 mutagenized at 690 15% per bp. Logo illustrates an additive model trained using MAVE-NN 79 to distinguish selected 691 variants from input variants. Sequence coordinates match those in Fig. S4M. The endogenous 692 OriA-006 sequence is shown above, and linker positions 25-29 from Fig. 3B are boxed. Fig S6A 693 provides a similar analysis for OriC-061. (B) Functional origin assay of single-nucleotide mutants 694 within the AT and the Orc5 -BP motifs . Images show c olony formation following plasmid 695 transformation and selection and percentage (% URA+ retention) for different mutations. (C) 696 Sensitivity and specificity of core motifs in MPOS-derived models. (i) Core motif of the Y. lipolytica 697 MPOS model and corresponding z -score distributions for EdU peaks and random Y. lipolytica 698 genomic regions. (ii) Core motif from a model trained on S. cerevisiae MPOS data 28 , together 699 with z-score distributions of this model on origins from OriDB 18 and on random S. cerevisiae 700 genomic regions. (iii) ROC curves for the two core motifs on their respective positive and negative 701 genomic regions. 702 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 18 Figure 7. Origin recognition throughout evolution . A comparison of analogous structural 703 features within the ODC complex of metazoans (Homo sapiens), Saccharomyces cerevisiae, and 704 Yarrowia lipolytica. 705 706 Supplemental Video 1 An animation of the 2.6 Å-resolution unsharpened map of the HsORC-707 CDC6 complex rotated along two axes. 708 709 Supplemental Video 2 An animation of the 2.7 Å resolution unsharpened map of the YlORC -710 DNA54bpOriC-061-YlCdc6 complex rotated along two axes. 711 712 Supplemental Video 3 An animation of the refined molecular structure of the YlORC-DNA54bpOriC-713 061-YlCdc6 complex rotated along two axes. 714

715 Yeast Strain Construction 716 Yeast strains for EdU -sequencing and ARS assays were derived from Yarrowia lipolytica PO1f 717 (MATA, leu2 -270, ura3 -302, xpr2 -322, axp1 -2). Strains used for genome annotation were 718 provided by Dr. Richard Rachubinski (University of Alberta). To generate the TK+ strain 719 (YlB0002), a BrdU -Inc cassette with a URA3 marker was integrated at the IntE1 locus on 720 chromosome 5. The cassette, under TEF and GPD promoters, expressed Herpes Simplex Virus 721 thymidine kinase (HSV-TK) and the human Equilibrative Nucleoside Transporter 1 (hENT1) 722 genes. Correct integration was confirmed via selective growth on -URA and FOA plates, colony 723 PCR, and vector verification before transformation. 724 Plasmid Construction for BrdU-Inc Cassette 725 The BrdU-Inc cassette was cloned into the EasyClone vector pCfB6677 88 obtained from Addgene 726 and targeted to the IntE1 locus. The cassette was flanked by loxP sites for URA3 marker removal 727 via Cre recombinase. HSV-TK and hENT1 were placed under the TEF and GPD promoters, 728 respectively, and amplified from pNC1164 and pCfB8742. USER cloning enabled precise 729 assembly using uracil -containing primers and enzymatic treatment to generate overhangs for 730 directional ligation. The final construct was validated by PCR, restriction mapping, and 731 sequencing. 732 ARS Plasmid Construction 733 The pSCARS1 plasmid 89 was modified to create pYl001 by removing SC-Trp and ORI1068, and 734 adding KpnI and BglII sites. GFP remained under TEF promoter control. To construct pYl002 - 735 pYl011, replication origins were amplified from PO1f genomic DNA and ligated into pYl001 at KpnI 736 or BglII sites. Constructs were verified by PCR and restriction digestion for downstream ARS 737 assays and stability tests. 738 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 19 Yeast Transformation 739 PO1f cells (5 × 10⁷) were grown overnight and transformed according to Dahlin et al 2021 88 . For 740 genomic integrations, 500 ng of the linearized vector was used; for ARS assays, 15 µg of the 741 circular plasmid. Cells were heat -shocked at 39°C, recovered in YPD, and plated on -URA. 742 Integration was confirmed using PCR. 743 Synchronization of YlB0002 (TK+) 744 To synchronize cells in G0/G1, YlB0002 was grown in YPD for 72 hours when they reached 745 stationary phase and then diluted 1:10 into fresh medium to re-enter the cell cycle. EdU (500 µM 746 for sequencing, 100 µM for imaging) (Thermo Fisher, E10187) and HU (5 mM) (Sigma, H8627) 747 were added as needed. Samples were collected at multiple time points for EdU imaging, flow 748 cytometry, and sequencing. 749 EdU Imaging in YlB0002 750 Synchronized or log-phase cells were labeled with 100 µM EdU (Thermo Fisher, E10187). After 751 fixation (3.7% PFA), cells were permeabilized with Triton X -100 and subjected to Click -iT 752 chemistry using Alexa Fluor 488 (Thermo Fisher, C10387). DNA was stained with Hoechst 33342, 753 and cells were mounted in anti-fade solution for imaging with 63X or 100X oil objectives. 754 Flow Cytometry 755 To assess cell cycle progression in Y. lipolytica, cultures were grown at 30°C to mid-log phase or 756 harvested at specific time points. Cells were pelleted and washed twice with sterile water, then 757 fixed in 70% ethanol and incubated overnight at 4°C. Following fixation, cells were pelleted, 758 washed twice with sterile water, and resuspended in FC buffer (50 mM sodium citrate, pH 7.0, 759 0.1% sodium azide). For RNA and protein degradation, samples were sequentially treated with 760 RNase A (0.1 mg/ml) and proteinase K (0.2 mg/ml) for 1 hour each at 55°C. Cells were then 761 stained with SYTOX Green Nucleic Acid stain (Thermo Fisher S7020). Samples were sonicated 762 and diluted before flow cytometry or FACS analysis to assess DNA content and cell cycle 763 distribution. 764 EdU-Seq of Synchronous Yarrowia Cells ± HU 765 To track DNA replication dynamics, Y. lipolytica cells were synchronized in G0/G1 by 72 -hour 766 culture in YPD. Cells were then released into fresh YPD containing 500 µM EdU (± 5 mM HU), 767 and samples were collected over a 210 -minute time course. Flow cytometry confirmed 768 synchronization. 769 For mapping sites of EdU incorporation , DNA from the time-course samples was fragmented, 770 EdU-labeled DNA was captured via biotin -azide Click-iT reaction (Thermo Fisher, C10365) and 771 Streptavidin T1 beads (Thermo Fisher 65602). Libraries were prepared using the Illumina TruSeq 772 Kit (Illumina IP -202-1012). Sequencing identified newly replicated regions, providing high -773 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 20 resolution replication timing profiles. Sequencing data revealed replication origin firing across S 774 phase 35. 775 EdU Incorporation and DNA Preparation 776 To identify early - and late-firing origins, stationary phase cells were transferred into fresh YPD 777 containing either EdU+HU or EdU alone. A mock (no EdU) control was included. At each time 778 point, replication was halted with sodium azide (0.1%), and cells were collected for flow cytometry 779 and DNA extraction. DNA was purified using a Qiagen Genomic-tip after Zymolyase-20T (Sunrise 780 Science Products, N0766391) digestion and lysis. Purified genomic DNA was fragmented using 781 the Bioruptor Pico and checked on a Bioanalyzer (100–550 bp fragments). 782 Click Labeling and DNA Enrichment 783 EdU-labeled DNA was tagged with biotin via Click -iT chemistry and purified using magnetic 784 Streptavidin beads. Bound DNA was eluted and further cleaned using MinElute columns (Qiagen). 785 Library Prep and Sequencing 786 Biotinylated DNA was ligated to adapters and PCR -amplified using the Illumina TruSeq kit. The 787 resulting libraries were sequenced to profile replication timing at high resolution. 788 EdU-seq Pre-processing and genome alignment 789 The paired -end reads were trimmed using Fastp v.0.23.2 90 using the default settings and 790 automatic adapter detection, and quality control was performed using FastQC (v0.12.1; available 791 online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). The trimmed reads were 792 aligned to Yarrowia Lipolytica strain E122 (MATA) using Bowtie2 v.2.4.4 91 with the default 793 options. The output SAM alignment files were converted to BAM format, sorted and indexed 794 using SAMtools Samtools v.1.19.2 92 To produce the illustrated genome coverage tracks and for 795 visualization purposes, we used bamCoverage from Deeptools v.3.5.1 93 was used to generate 796 the coverage tracks with normalization option of bins per million mapped reads (BPM). 797 A total of 75 EdU -seq libraries were prepared across 11 time points (15 –210 minutes), with 798 biological and technical replicates ensuring data reproducibility (Supplement Table 1). Fifty high-799 quality libraries were selected for downstream analysis. Temporal replication patterns were 800 consistent across replicates, demonstrating the robustness of the protocol. 801 Supplemental Table 1 802 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 21 803 The libraries used in the analysis are summarized in the table below. 804 805 Name of the library Duration of EdU label Treatment 9-30min-eh5_S3_trm_matAfinal.sorted.bw 30 min 500 mM EdU + 5mM HU S07_S7_trm_matAfinal.sorted-45minold.bw 45 min 500 mM EdU + 5mM HU 72-45-eh-5_S3_trm_matAfinal.sorted.bw 45 min 500 mM EdU + 5mM HU S03_S3_trm_matAfinal.sorted-new45eh.bw 45 min 500 mM EdU + 5mM HU S04_S4_trm_matAfinal.sorted-new60eh.bw 60 min 500 mM EdU + 5mM HU S09_S9_trm_matAfinal.sorted-75min.bw 75 min 500 mM EdU + 5mM HU S08_S8_trm_matAfinal.sorted-new90eh.bw 90 min 500 mM EdU + 5mM HU S07_S7_trm_matAfinal.sorted-90eh.bw 90 min 500 mM EdU + 5mM HU S14_S14_trm_matAfinal.sorted-120min-old.bw 120 min 500 mM EdU + 5mM HU S11_S11_trm_matAfinal.sorted-new180- 642peakcount.bw 180 min 500 mM EdU + 5mM HU S12_S12_trm_matAfinal.sorted-new210.bw 210 min 500 mM EdU + 5mM HU 30min_S4_trm_matAfinal.sorted.bw 30 min 500 mM EdU .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 22 45min_S5_trm_matAfinal.sorted.bw 45 min 500 mM EdU 45min_S6_trm_matAfinal.sorted.bw 45 min 500 mM EdU S04_S4_trm_matAfinal.sorted.bw 60 min 500 mM EdU S06_S6_trm_matAfinal.sorted.bw 60 min 500 mM EdU 806 EdU-seq Peak calling and annotation 807 For peak calling, we employed MACS2 v2.2.6 94 using a p -value filter of 0.01 and a minimum 808 length of 300 base pairs. These peaks were annotated using annotatePeaks.pl from HOMER 809 v.4.11 http://homer.ucsd.edu/homer 95 , which also provides the annotation class (Promoter -810 TSS/TTS/Gene/Intergenic) enrichment analysis. 811 Stability Assays 812 For plasmid stability, transformants were grown in selective media, then diluted into YPD and 813 regrown. After 30 hours, cells were plated on selective and non -selective plates. Colony counts 814 on both plates were used to estimate plasmid loss , expressed as percentage . In GFP-based 815 assays, cells harboring plasmids expressing Green Fluorescent Protein (GFP) under the control 816 of the TEF promoter with the CYC1 terminator were analyzed via flow cytometry according to 817 Lopez et al. 89 . GFP+ percentages were calculated using BD FACSC, with wild -type cells as a 818 negative control. 819 XhoI Linker Scanning via In-Fusion Cloning 820 XhoI linker scanning mutants of OriA-006 and OriC-061 were generated using In-Fusion cloning 821 (TaKaRa 638945). Primers with XhoI sites and 15 -bp overlaps were used in high -fidelity PCR 822 (PrimeSTAR Max,TaKaRa R045A). Mutant plasmids were circularized via inverse PCR and In -823 Fusion assembled, purified, and transformed into Y. lipolytica for downstream analysis. 824 Massively parallel origin selection (MPOS) assay 825 MPOS assays were carried out as in 28 with minor modifications. Plasmid libraries for OriA-006 826 and for OriC-061 were constructed starting with corresponding wild -type plasmids used in the 827 linker scanning experiments . 90 bp regions centered on the essential sites identified by linker 828 scanning were then mutagenized at 15% per bp. Mutagenesis was carried out using 829 computationally designed oligo pools, each containing 7,000 variants, synthesized by Agilent. To 830 increase the percentage of correctly cloned plasmids, variants were cloned using a ccdB cassette 831 replacement strategy based on that of Kinney et al. (2010) 96 , but with Gibson cloning instead of 832 Golden Gate cloning. Each plasmid library was then transformed into Y. lipolytica and subjected 833 to selection via growth on SC-URA plates, after which bulk DNA was extracted. Amplicons 834 containing variant sequences flanked by barcodes and primers for Illumina sequencing were then 835 prepared using template DNA (plasmid DNA from the initial libraries or bulk DNA extract ed from 836 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 23 cells) amplified by PCR using custom primers. Amplicons were subjected to PE150 sequencing 837 on an Illumina NextSeq 2000 instrument using a P1 flow cell and XLEAP-SBS chemistry. 838 Quantitative modeling of MPOS data 839 For each of the two loci, OriA-006 and OriC-061, MAVE-NN 79 was used to train an additive model 840 that distinguishes MPOS-selected origins from those in the initial library. The OriA-006 model was 841 trained using 2,911,260 pre-selection reads and 2,082,266 post-selection reads. The OriC-061 842 model was trained using 8,820,054 pre-selection reads and 7,264,362 post-selection reads. 843 Sequence logos illustrating these models are shown in Figure 6A (OriA-006) and Figure S6A 844 (OriC-061). Logos were created using Logomaker 97 . The additive model for origin specificity in 845 S. cerevisiae was previously reported by Hu et al. 35 and computed in a similar manner using an 846 early version of MAVE-NN. 847 ROC analysis of MPOS-derived motifs 848 To carry out the ROC analysis in Figure 6C a 32 bp core motif was extracted from the Y. lipolytica 849 OriA-006 additive model. This motif was used to scan 1,000 bp regions of Y. lipolytica genomic 850 DNA centered on either the EdU peaks identified above (positive set; 623 regions) or on randomly 851 chosen genomic locations (control set; 623 regions). The maximum motif score observed in each 852 genomic region was recorded. The scores for positive and negative regions were then converted 853 to z-scores (Figure 6Ci). A similar analysis was performed using a n 18 bp core motif extracted 854 from the S. cerevisiae ARS1 MPOS model (Figure 6Cii). This motif was used to scan 1 ,000 bp 855 regions of S. cerevisiae genomic DNA centered either on origins in OriDB 18 (positive set; 410 856 regions) or on randomly chosen genomic DNA (control set ; 410 regions). Figure 6Ciii shows 857 ROC curves corresponding to these core motifs and their respective positive and negative control 858 sets. 859 Expression and Purification of Human ORC Subunits (HsORC1–5) 860 861 Codon-optimized HsORC1 (NP_004144.2), N-terminally fused with twin Strep and SUMO tags, 862 was cloned into the pFL vector for expression in insect cells. The remaining synthetic human ORC 863 genes—HsORC2 (NP_006181.1), HsORC3 (NP_862820.1), HsORC4 (NP_859525.1), and 864 HsORC5 (NP_002544.1)—were cloned into the MultiBac baculovirus expression system 98 . A 865 twin StrepTag followed by a TEV cleavage site was also introduced at the N-terminus of HsORC3 866 to facilitate affinity purification. All HsORC proteins are full length. Recombinant expression of 867 HsORC1 and separately of the HsORC2-5 complex were performed in Sf9 insect cells infected 868 with baculovirus and cultured in CCM3 medium (GE Healthcare Life Sciences, Pittsburgh, PA) for 869 48 hours. Cell pellets for both HsORC1 and HsORC2-5 were resuspended separately in lysis 870 buffer containing 50 mM HEPES -NaOH (pH 7.5), 300 mM KCl, 30 mM potassium glutamate, 5 871 mM magnesium acetate, 5 mM dithiothreitol (DTT), and 2 mM ATP. HsORC1-expressing cells 872 were lysed by sonication, and lysates were clarified by centrifugation at 143,000 × g for 45 873 minutes. The supernatant was loaded onto a 5 mL StrepTactin agarose beads onto a gravity flow 874 column. After washing, bound HsORC1 protein was eluted with lysis buffer supplemented with 5 875 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 24 mM desthiobiotin. The HsORC2-5 complex was purified in parallel using an identical Strep -tag 876 affinity protocol. Purified HsORC1 and HsORC2-5 protein fractions were combined, and TEV 877 protease was added for tag cleavage, followed by incubation at 4 °C for 12 hours. The mixture 878 was subsequently diluted to 150 mM KCl and subjected to ion exchange chromatography using 879 a HiTrap SP column with a linear gradient from 150 to 1000 mM KCl. Protein-containing fractions 880 were analyzed by SDS -PAGE. Fractions containing all five ORC subunits were pooled, 881 concentrated, and further purified by size exclusion chromatography using a Superose 6 Increase 882 10/300 GL column (GE Healthcare) equilibrated with minimal buffer (25 mM HEPES -NaOH (pH 883 7.5), 100 mM KCl, 2 mM DTT). Final protein purity was assessed by SDS -PAGE, and pure 884 fractions were concentrated using an Amicon® Ultra centrifugal filter (50 kDa MWCO). Aliquots 885 of ~3–5 μM were snap-frozen in liquid nitrogen and stored at –80 °C. 886 887 Expression and Purification of Human CDC6 Protein 888 889 The human CDC6 (HsCDC6) gene was cloned into the pET28a vector to allow for IPTG inducible 890 expression in Escherichia coli. The construct encoded full-length HsCDC6 with an N-terminal His-891 SUMO tag. The resulting plasmid was transformed into E. coli Rosetta (DE3) cells and cultured 892 in 4 liters of Terrific Broth (TB) supplemented with kanamycin. Cells were grown at 37°C until 893 reaching an optical density (OD ₆₀₀) of 0.8–1.0, at which point expression was induced with 0.5 894 mM IPTG. Protein expression was carried out for 16 hours at 16°C, and cells were harvested by 895 centrifugation at 3,500 × g for 15 minutes . Cell pellets were resuspended in lysis buffer (50 mM 896 HEPES, pH 7.0, 300 mM NaCl, 10 mM imidazole, 2 mM β -mercaptoethanol) and lysed by 897 sonication on ice. The lysate was clarified by centrifugation and loaded onto a gravity-flow column 898 packed with Ni-NTA agarose beads and pre-equilibrated with lysis buffer. The column was 899 washed with 10 column volumes (CV) of lysis buffer , followed by 5 CV of high -salt buffer (lysis 900 buffer with 500 mM NaCl). Bound protein was eluted using lysis buffer containing 400 mM 901 imidazole. Eluted protein was incubated with TEV protease overnight at 4°C to cleave the His -902 SUMO-TEV tag. The cleaved sample was diluted and applied to a HiTrap SP column (GE 903 Healthcare) equilibrated in buffer (50 mM HEPES-NaOH, pH 7.0, 150 mM NaCl, 1 mM DTT) for 904 ion exchange chromatography . Protein was eluted with a linear gradient of Buffer B (50 mM 905 HEPES-NaOH, pH 7.0, 1 M NaCl, 1 mM DTT). Protein peak fractions were pooled and subjected 906 to size exclusion chromatography (SEC) using a Superdex 200 Increase 10/300 GL column , 907 equilibrated with SEC buffer (25 mM HEPES-NaOH, pH 7.5, 100 mM KCl, 2 mM DTT). Protein 908 purity was assessed by SDS-PAGE, and pure fractions were concentrated using an Amicon® 909 Ultra centrifugal filter (30 kDa MWCO). Final protein aliquots at concentrations of approximately 910 10–15 μM were snap-frozen in liquid nitrogen and stored at –80°C. 911 912 Protein purification and preparation of YlODC 913 914 Synthetic, full -length genes of Yarrowia lipolytica ORC1 (YlORC1) (RefSeq: XP_502645.1), 915 YlORC2 (XP_503147.3), YlORC3 (XP_505428.1), YlORC4 (XP_504002.3), YlORC5 916 (XP_500387.1), and YlORC6 (XP_506105.1) were codon optimized and cloned into either pFL 917 (pH promoter YlORC1, p10 promoter YlORC6), pSPL (pH promoter YlORC2, p10 promoter 918 YlORC5), or pUCDM (pH promoter YlORC4, p10 promoter YlORC3) plasmids for expression via 919 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 25 the MultiBac baculovirus expression system 98 . To increase protein expression and solubility, in 920 addition to providing a tag for affinity chromatography, an N -terminal Twin Strep-SumoStar-TEV 921 tag was added to YlORC1. The tagged version of YlOrc1 was utilized for the Cdc6 loading assay 922 and EM studies. YlCDC6 (XP_501295.1) was codon optimized and cloned into a pET28b bacterial 923 expression cassette (Novagen) containing an N-terminal 8xHis-TEV tag for affinity purification. 924 925 For YlORC, Sf9 insect cells were incubated with baculovirus for 72 hr in Hyclone CCM3 media 926 (GE Healthcare Life Sciences, Pittsburg, PA). For YlCdc6 expression, BL21 (DE3) strain E. coli 927 were transformed and then selected for with a kanamycin -supplemented LB starter culture, 928 followed by inoculation of kanamycin-supplemented Terrific Broth (TB). Expression was induced 929 at OD 1 with the addition of 0.5 mM IPTG and incubated at 17ºC overnight. 930 931 Unless otherwise noted, each purification step was carried out at 4ºC. For YlORC purification, 932 insect cell pellets were thawed in a 30ºC water bath, resuspended in lysis buffer (50 mM HEPES-933 NaOH (pH 7.5), 150 mM potassium acetate (KOAc) (pH 7.5), 50 mM potassium glutamate (K -934 Glu), 50 mM arginine hydrochloride, 10 mM Mg(OAc) 2, 6.5 mM dithiothreitol (DTT), 1.65 mM 935 adenosine triphosphate (ATP), 10% glycerol), supplemented with a protease inhibitor cocktail (1 936 mM PMSF, 2 µM pepstatin, 2 µM leupeptin, 1 mM benzamidine, 1:1725 dilution of Millipore-Sigma 937 aprotinin (A6279)) in addition to 1X cOmplete EDTA-free protease inhibitor cocktail (Roche) and 938 sonicated. Lysate was then centrifuged at 38,000g for 1 hour, after which the supernatant was 939 collected and 3 mL of either StrepTactin Superflow resin or StrepTactin 4Flow resin was added 940 to the supernatant and incubated for 90 minutes. The resin was washed and YlORC was eluted 941 using lysis buffer supplemented with 5 mM desthiobiotin. YlORC-containing fractions were pooled 942 together, and l-phosphatase was added at a ~2:1 YlORC:phosphatase molar ratio, along with 1 943 mM manganese chloride, and incubated at 4ºC for 36 -48 hours. In preparations used for 944 biochemical assays, YlORC was simultaneously treated with TEV protease at an YlORC:TEV 945 mass ratio of 15:1. The phosphatase -treated elution then underwent anion exchange 946 chromatography (HiTrap Q HP 5 mL, Cytiva) followed by size exclusion chromatography 947 (Superose 6 increase 10/300 GL, Cytiva) equilibrated in minimal buffer (25 mM HEPES -NaOH 948 (pH 7.5), 100 mM KOAc (pH 7.5), 50 mM K -Glu, 5 mM Mg(OAc) 2, 1 mM DTT, 5% glycerol). 949 Aliquots were made with YlORC concentrated to »2.4 mg/mL. 950 951 For YlCdc6 purification, cell pellets were thawed in similar conditions except that the lysis buffer 952 contained an additional 10 mM imidazole. For YlCdc6 affinity purification, 5 mL of Ni-NTA agarose 953 was added to clarified lysate, with resin washes done with 25 mM imidazole supplemented lysis 954 buffer, and eluted with 50 mM, 100 mM, 250 mM, and 500 mM imidazole supplemented lysis 955 buffer. TEV protease was added and incubated overnight at 4º. YlCdc6 was further purified using 956 cation exchange chromatography (HiTrap SP HP 5 mL, Cytiva) followed by size exclusion 957 chromatography (Superdex 200 increase 10/300 GL, Cytiva) into minimal buffer (25 mM HEPES-958 NaOH (pH 7.5), 100 mM NaCl, 1 mM DTT). Aliquots were made by supplementation with glycerol 959 to 5%, and YlCdc6 was concentrated to 5 mg/mL. 960 961 Synthetic ori oligonucleotides for structural analysis 962 963 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 26 Oligonucleotides were ordered from IDT and were then annealed to their complementary 964 oligonucleotides by incubating them together at 95ºC for 5 minutes, followed by a temperature 965 decrease of 2ºC/min until 25ºC was reached. The ordered oligonucleotides are displayed in the 966 table below: 967 968 OriA-006ori(60bp) 5’-CTCCACCCAATATGCCCCTCCAATCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGATTGGAGGGGCATATTGGGTGGAG -3’ OriA-006 CNNGGNR (60bp) 5’-CTCCACCCAATATGCCCCTGGAATCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGATTCCAGGGGCATATTGGGTGGAG -3’ OriA-006GNNGGNR (60bp) 5’-CTCCACCCAATATGCCGCTGGAATCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGATTCCAGCGGCATATTGGGTGGAG -3’ OriA-006GNNCCNR (60bp) 5’-CTCCACCCAATATGCCGCTCCAATCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGATTGGAGCGGCATATTGGGTGGAG -3’ OriA-006GNNCCNT (60bp) 5’-CTCCACCCAATATGCCGCTCCATTCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGAATGGAGCGGCATATTGGGTGGAG -3’ OriA-006CNNAANR (60bp) 5’-CTCCACCCAATATGCCCCTAAAATCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGATTTTAGGGGCATATTGGGTGGAG -3’ OriA-006ANNAANR (60bp) 5’-CTCCACCCAATATGCCACTAAAATCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGATTTTAGTGGCATATTGGGTGGAG -3’ OriA-006CNNCCNT (60bp) 5’-CTCCACCCAATATGCCCCTCCATTCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGAATGGAGGGGCATATTGGGTGGAG -3’ OriA-006CNNCCNC (60bp) 5’-CTCCACCCAATATGCCCCTCCA CTCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGA GTGGAGGGGCATATTGGGTGGAG -3’ oriA-006 R+0 6A 5’-CTCCACCCAATATGCCCCTCCAAAAAAACTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGTTTTTTTGGAGGGGCATATTGGGTGGAG -3’ oriA-006 R+1 6A 5’-CTCCACCCAATATGCCCCTCCAAAAAAAATCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGATTTTTTTTGGAGGGGCATATTGGGTGGAG -3’ OriA-006 R+2 6A 5’-CTCCACCCAATATGCCCCTCCAATAAAAAACCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGTTTTTTATTGGAGGGGCATATTGGGTGGAG -3’ OriA-006 R+3 6A 5’-CTCCACCCAATATGCCCCTCCAATCAAAAAACTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGTTTTTTGATTGGAGGGGCATATTGGGTGGAG -3’ OriA-006 R+4 6A 5’-CTCCACCCAATATGCCCCTCCAATCCAAAAAATACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTATTTTTTGGATTGGAGGGGCATATTGGGTGGAG -3’ OriA-006 R+5 6A 5’-CTCCACCCAATATGCCCCTCCAATCCAAAAAAAACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTTTTTTTTGGATTGGAGGGGCATATTGGGTGGAG -3’ OriA-006 R+6 6A 5’-CTCCACCCAATATGCCCCTCCAATCCAGAAAAAACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTTTTTTCTGGATTGGAGGGGCATATTGGGTGGAG -3’ OriA-006 R+7 6A 5’-CTCCACCCAATATGCCCCTCCAATCCAGCAAAAAAAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTTTTTTTGCTGGATTGGAGGGGCATATTGGGTGGAG -3’ OriA-006 R+8 6A 5’-CTCCACCCAATATGCCCCTCCAATCCAGCTAAAAAAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTTTTTTAGCTGGATTGGAGGGGCATATTGGGTGGAG -3’ OriA-006 R+5 XhoI 5’-CTCCACCCAATATGCCCCTCCAATCCACTCGAGACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTCTCGAGTGGATTGGAGGGGCATATTGGGTGGAG -3’ OriC-061 (45bp) 5’-CCCAATATTACACCCAAGTAGCATGCATAAGCTAAAAGTAACTCG -3’ 5’-CGAGTTACTTTTAGCTTATGCATGCTACTTGGGTGTAATATTGGG -3’ OriC-061 scramble (45bp) 5’-CACAGAAACTAATAAGACAAACACGCATCTGCTTATTGCGCACTA -3’ 5’-TAGTGCGCAATAAGCAGATGCGTGTTTGTCTTATTAGTTTCTGTG -3’ OriC-061 (54bp) 5’-TGGTACCGATCCCAATATTACACCCAAGTAGCATGCATAAGCTAAAAGTAACTC -3’ 5’-GAGTTACTTTTAGCTTATGCATGCTACTTGGGTGTAATATTGGGATCGGTACCA -3’ OriC-061 (60bp) 5’-CGATGGTACCGATCCCAATATTACACCCAAGTAGCATGCATAAGCTAAAAGTAACTCGCA -3’ 5’-TGCGAGTTACTTTTAGCTTATGCATGCTACTTGGGTGTAATATTGGGATCGGTACCATCG -3’ OriA-006 YGTR (60bp) 5’-CTCCACCCAATGTGCCCCTCCAATCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGATTGGAGGGGCACATTGGGTGGAG -3’ OriA-006 YACR (60bp) 5’-CTCCACCCAATACGCCCCTCCAATCCAGCTCCTACAAGTCGGGGTTGAGACTGCACCAAA -3’ 5’-TTTGGTGCAGTCTCAACCCCGACTTGTAGGAGCTGGATTGGAGGGGCGTATTGGGTGGAG -3’ .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 27 969 Cdc6 loading assay 970 971 To test the dependence of YlORC’s ability to bind DNA and load YlCdc6 on the origin, annealed 972 oligonucleotides representing ori variants and 5X reaction buffer (750 mM KOAc (pH 7.5), 250 973 mM HEPES-NaOH (pH 7.5), 50 mM Mg(OAc)2, 5 mM DTT, 5 mM ATP, 50% glycerol) were added 974 to purified YlORC1 -6, with an ORC:DNA molar ratio of 1:1.5, and allowed to incubate at room 975 temperature for 10 minutes. For YlCdc6-containing assays, purified YlCdc6 was added following 976 the ORC -DNA incubation at an ORC:Cdc6 molar ratio of 1:4 and was incubated at room 977 temperature for 10 minutes. Samples were then loaded onto a Superose 6 increase 3.2/300 978 microkit column (Cytiva) equilibrated in SEC buffer (25 mM HEPES -NaOH (pH 7.5), 100 mM 979 NaCl, 1 mM DTT) and fraction samples were run on SDS-PAGE and visualized using ReadyBlue 980 Protein Gel Stain (Sigma). For experiments determining the total quantity of DNA in each fraction, 981 SDS-PAGE gels were stained with SYBR Gold and imaged prior to ReadyBlue staining. 982 260/280nm UV absorbance ratios obtained from chromatograms were also used for qualitative 983 analysis of DNA and Cdc6 binding to ORC. 984 985 Gel images were quantified using the ImageJ -based software package Fiji 99 . For calculating 986 relative loading efficiency, the intensity ratio of the YlCdc6 band to each ORC subunit band was 987 normalized to the intensity ratios observed while running the assay with a 60bp fragment of the 988 WT OriA-006 and averaged. 989 990 Cryo-EM sample preparation 991 992 HsODC: Purified HsORC1–5, HsCDC6, DNA and AMP-PNP were mixed at final concentrations 993 of 2.5 μM, 5.0 μM, 7.5 μM and 10 μM, respectively. To reduce preferred particle orientations and 994 promote the formation of thin ice layers over grid holes, lauryl maltose neopentyl glycol (LMNG; 995 Anatrace, Maumee, OH) was added to a final concentration of 0.05% (w/v). For cryo-electron 996 microscopy, 4 μL of the protein –DNA complex was applied to a glow-discharged Quantifoil R 997 0.6/1, 300 mesh copper grid. The grid was incubated for 10 seconds at 25°C and 90% humidity, 998 blotted for 3.0 seconds, and then rapidly plunge-frozen into liquid ethane using a Leica EM GP2 999 automatic plunge freezer (Leica Microsystems, Buffalo Grove, IL). 1000 1001 1002 YlODC60bpOri-A006-WT: a similar protocol was used with key differences. A 60 bp OriA-006 fragment 1003 was used instead of the 45 bp OriC-061 fragment, and the final protein concentration was 1 1004 mg/mL. The sample was applied to a glow -discharged lacey carbon grid and blotted for 3.0 1005 seconds. 1006 1007 YlODC54bpOriC-061: purified YlORC1 -6 was mixed with glycerol -free 5X loading assay reaction 1008 buffer, a 54 bp OriC-061 fragment, and YlCdc6 at an ORC:DNA:Cdc6 molar ratio of 1:1.5:4 in a 1009 stepwise fashion, followed by gel filtration using a Superose 6 increase 3.2/300 microkit column 1010 (Cytiva), identical to the Cdc6 loading assay. Fractions containing the YlORC-DNA-Cdc6 complex 1011 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 28 were then concentrated with a 0.5 mL Amicon Ultra 50kDa MWCO centrifugal filter to a final 1012 protein concentration of 1 -1.25 mg/mL without additional ATP supplementation. Following 1013 concentration, lauryl maltose neopentyl glycol (LMNG) was added to 0.05% to reduce a preferred 1014 orientation issue. 4 µL of sample were applied to a non-glow discharged Quantifoil R 1.2/1.3 300 1015 mesh copper grid (previously washed with ethyl acetate), incubated for 10 seconds at 25ºC and 1016 95% humidity, blotted for 2.9 seconds, and plunged into liquid ethane using a Leica Automatic 1017 Plunge Freezer EM GP2. 1018 1019 YlODC60bpOri-A006-CNNGGNR: sample preparation followed the protocol for the YlODC45bpOriC-061 1020 sample, with key differences. A 60 bp OriA-006 fragment containing the CNNGGNR mutation 1021 (see oligo table for sequence) was substituted, and samples had a final protein concentration of 1022 1.6 mg/mL. Additionally, a non -glow discharged Quantifoil R 1.2/1.3 300 mesh copper grid 1023 (previously washed with ethyl acetate) was used and was blotted for 2.7 seconds. 1024 1025 Cryo-EM data acquisition 1026 1027 HsODC: Cryo-EM data were collected on a Titan Krios transmission electron microscope 1028 (ThermoFisher Scientific) operating at 300 keV. Data were acquired using EPU software 1029 (v2.10.0.5, ThermoFisher Scientific), and dose-fractionated movies were recorded on a K3 direct 1030 electron detector (Gatan) in electron counting mode. HsODC samples were applied to Quantifoil 1031 R 0.6/1 grids, and 30 -frame movies were collected at an exposure rate of 1.44 e ⁻/Ų/frame, 1032 yielding a cumulative dose of 43.2 e ⁻/Ų. A total of 7088 micrographs were acquired at 81,000× 1033 nominal magnification, with a defocus range of 0.6–2.2 μm. 1034 1035 YlODC: Cryo-electron microscopy data were collected using an FEI/ThermoFisher Titan Krios 1036 TEM operating at 300 keV. A Gatan K3 direct electron detector equipped with a BioQuantum 1037 energy filter was utilized to semi -automatically collect dose -fractionated movies with 1038 ThermoFisher EPU data collection software. For the YlODC45bpOriC-061 maps, two collections on 1039 consecutive days accrued 8978 and 9310 exposures, respectively, with movies collected with 30 1040 frames at a dose rate of 1.98 e/Å2 per frame, resulting in a cumulative dose of 59.4 e/Å2. For the 1041 YlODC54bpOriC-061 data collection, 30-frame movies were collected over three consecutive days, 1042 resulting in 9309, 8758, and 2274 exposures taken, respectively, at a dose rate of 1.44 e/Å2 per 1043 frame, totaling a cumulative dose of 43.2 e/Å2. For the YlODC60bpOri-A006-WT data collection, a single 1044 session was used to collect 8340 exposures, with movies containing 40 frames at a dose rate of 1045 1.97 e/Å2 per frame, totaling 78.8 e/Å2 in cumulative dose. YlODC60bpOri-A006-CNNGGNR data collection 1046 included 8428 exposures from a single session, with 40 frames per movie, a dose rate of 1.37 1047 e/Å2 per frame, and a cumulative dose of 54.8 e/Å2. 1048 1049 Cryo-EM data processing 1050 1051 HsODC: Real -time preprocessing, including motion correction, CTF estimation, and particle 1052 picking, was performed in WARP (v1.0.9). Particle picking used the BoxNet pretrained neural 1053 network implemented in TensorFlow, with a particle diameter of 180 Å and a threshold score of 1054 0.6, resulting in 898,455 coordinates. Subsequent image processing was carried out in 1055 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 29 cryoSPARC v3.2. Particles were extracted and subjected to multiple rounds of 2D classification, 1056 and well-resolved subsets were selected for ab initio 3D reconstruction. Separation into 3 –5 ab 1057 initio classes proved critical for improving map quality. These models were used for 3D 1058 heterogeneous refinement against the full dataset, yielding 443,190 selected particles for HsODC. 1059 This subset was further classified into three classes, and the best class was refined. 1060 Homogeneous and non-uniform refinements for the best 3D class (130,819 particles) produced a 1061 cryo-EM map at 2.6 Å resolution, as determined by the gold -standard FSC (GSFSC) criterion. 1062 The final sharpened map was used for model building and visualization (Figure S1B). 1063 1064 YlODC: WARP was utilized for motion correction, CTF estimation, and particle picking (via 1065 BoxNet neural network trained on manually picked micrographs) from the collected micrographs 1066 for all datasets 100 . For particle picking, a particle diameter of either 180 Å or 200 Å and a 1067 thresholding score of either 0.3, 0.4, or 0.5 were used, yielding 1,822,720 particles for 1068 YlODC45bpOriC-061, 3,805,196 particles for YlODC54bpOriC-061, and 793,509 particles for YlODC60bpOri-1069 A006-WT. While WARP pre-processing and picking was carried out for the YlODC60bpOri-A006-CNNGGNR 1070 exposures originally, cryoSPARC’s pre-processing and particle picking tools were used instead, 1071 described below. All downstream processing was carried out using cryoSPARC v4 101–103. All 2D 1072 classifications underwent an extra final iteration, all heterogeneous refinements listed used 1073 particles binned to 128 pixels, and all non -uniform refinements used “minimize over per -particle 1074 scale” and underwent one extra final pass. 1075 1076 YlODC45bpOriC-061: 1,822,720 particles were picked and extracted by WARP/BoxNet unbinned with 1077 a box size of 480 pixels and imported into cryoSPARC for processing. Each day of collection 1078 underwent its own set of 2D classification, resulting in two 2D classification jobs, one with 970,731 1079 starting particles and 194,059 particles selected for further processing, and another with 851,989 1080 particles with 457,613 selected. The 194,059 particles were used for ab-initio model generation 1081 of three classes. The generated ab-initio structures were then used in a heterogeneous 1082 refinement that included the entire particle dataset (1,822,720 particles), resulting in 802,503 1083 particles constituting the best class. This class then underwent homogenous refinement followed 1084 by non-uniform refinement, resulting in a 2.9 Å resolution map. Particles from this map underwent 1085 another round of 2D classification, selecting for 533,196 particles. Another heterogeneous 1086 refinement was carried out, using the ab-initio models as templates, resulting in a class with 1087 302,465 particles. This class was then subjected to non -uniform refinement with simultaneous 1088 per-particle scale minimization, per-particle defocus refinement, and CTF refinement of per-group 1089 CTF parameters, spherical aberration, and tetrafoil, producing a 2.7 Å map. Finally, 3D 1090 classification was carried out, and the class containing the strongest Orc2 -WHD was selected 1091 and underwent a final non-uniform refinement, resulting in a final map with a resolution of 2.7 Å 1092 from 84,914 particles 1093 1094 YlODC54bpOriC-061: 3,805,196 particles were picked and extracted by WARP/BoxNet unbinned with 1095 a box size of 440 pixels and imported into cryoSPARC for processing. Following a check for 1096 corrupt particles, each day’s full particle set was used in ab-initio model generation to generate 4 1097 classes. The best class of each set was used in non -uniform refinement with per -particle scale 1098 minimization, per-particle defocus refinement, and CTF refinement of per-group CTF parameters, 1099 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 30 spherical aberration, and tetrafoil to produce 3 maps (1,512,226 particles  538,187 particles, 1100 2.6 Å resolution map; 509,844 particles  210,193 particles, 2.6 Å resolution map; 1,377,300 1101 particles  488,199 particles, 2.5 Å resolution map) to validate pixel size and spherical aberration 1102 parameters. 2D classification with 250 classes was then carried out on the entire dataset, from 1103 which most (3,705,630) particles were selected for ab -initio model generation commenced from 1104 3,705,630 particles sorted into eight classes, the best of which was then flipped for proper 1105 handedness. Heterogeneous refinement was then carried out, producing a 2.3 Å resolution map 1106 from 1,453,157 particles that appeared heterogeneous in density. 3D classification was carried 1107 out with 10 classes at a filter resolution of 8 Å, which were then used to produce 6 maps via non-1108 uniform refinement as the result of combining some 3D classification output classes into one 1109 structure. One of these structures underwent a further 3D classification into 2 classes at a filter 1110 resolution of 10 Å, selecting for the class that contained the Orc1 -AAA domain while bound to 1111 DNA. Following this, a large -scale heterogeneous refinement of the entire dataset was done 1112 again, this time split into 14 classes: the aforementioned Orc1-AAA containing map, three maps 1113 derived from the 3D classification, the YlODC45bpOriC -061 map, and a map of YlORC -DNA 1114 produced from a previous collection. Following this large -scale classification/refinement, a 1115 YlODC54bpOriC-061 map at 2.5 Å resolution was produced with weak Cdc6 density from 1116 558,557 particles. This class then underwent 3 cycles of 3D classification and non -uniform 1117 refinement with per-particle scale minimization, per-particle defocus refinement, CTF refinement 1118 of per -group CTF parameters, spherical aberration, and tetrafoil refinement. Each 3D 1119 classification consisted of only 2 classes and used a filter resolution of 10 Å. After the 3 rounds of 1120 refinement and re-classification, a 2.7 Å map from 70,712 particles was produced. Following this, 1121 the Subset Particles job was used to select particles by per-particle scale, leaving 51,599 particles 1122 for use in another non -uniform refinement, producing a 2.7 Å resolution map to be used as the 1123 final YlODC54bpOriC-061 map. A schematic summary of data processing is shown in Figure 1124 S4B. 1125 1126 YlODC60bpOri-A006-WT: 793,509 particles were picked and extracted by WARP/BoxNet unbinned 1127 with a box size of 440 pixels and imported into cryoSPARC for processing. Following a check for 1128 corrupt particles, 2D classification into 200 classes was done, with 416,680 particles selected for 1129 ab-initio model generation of 6 maps. The YlODC54bpOriC-061 map with weak Cdc6 density was 1130 imported into the project and heterogeneous refinement with the six ab-initio maps and the 1131 imported YlODC54bpOriC-061 map was carried out on the entire particle dataset. The best 1132 heterogeneous refinement volume/particles underwent non -uniform refinement with the same 1133 scale, defocus, and CTF refinement corrections done in previously mentioned non -uniform 1134 refinements, and a 2.8 Å map from 255,191 particles was produced. Particles used in this map 1135 were then repicked from motion- and CTF-corrected micrographs (carried out in cryoSPARC) to 1136 generate a 2.9 Å resolution map, which then underwent Reference -Based Motion Correction to 1137 produce a 2.7 Å resolution map from 246,739 particles. 3D classification at a filter resolution of 1138 10 Å into 3 classes were performed, and the best two classes were combined for another round 1139 of non-uniform refinement, followed by a Subset Particles job (by per -particle scale) and a final 1140 non-uniform refinement to produce a 2.6 Å resolution map from 125,267 particles, which was then 1141 used for model building of YlODC60bpOri-A006-WT. A schematic summary of data processing is shown 1142 in Figure S4K. 1143 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 31 1144 YlODC60bpOri-A006-CNNGGNR: 8,428 movies were imported into cryoSPARC and underwent patch 1145 motion correction and patch CTF corrections to generate corrected micrographs. Template 1146 picking of micrographs commenced using representative 2D class averages of the particles used 1147 in the final YlODC60bpOri-A006-WT refinement, resulting in 4,708,382 particles being picked. 1148 Micrographs and respective particles were then analyzed using the Micrograph Junk Detector job, 1149 and after exposure and particle curation resulted in 2,710,398 particles from 7,330 micrographs. 1150 Particles were then extracted with a box size of 432 px and Fourier cropped to 128 px. Initial 2D 1151 classification utilized 200 classes, of which 2,4,69,383 particles from 184 classes were selected 1152 for use in further processing. Ab-initio model generation of 8 maps from 400,000 particles failed 1153 to produce a YlODC structure, so the YlODC60bpOri-A006-WT map was imported into the project, and 1154 heterogeneous refinement with the eight ab-initio maps along with the imported YlODC60bpOri-A006-1155 WT map was performed on the particle dataset. The best heterogeneous refinement volume/class 1156 (481,746 particles) underwent non-uniform refinement and further 2D classification to produce a 1157 cleaned stack of particles (305,326 particles) for particle re -extraction without Fourier cropping. 1158 Non-uniform refinement of the extracted particles produced a 2.9 Å resolution map, which was 1159 used as the reference volume for Reference-Based Motion Correction and re-refined to produce 1160 a 2.6 Å resolution map. Particles were then subset by per-particle scale and re-refined, generating 1161 a 2.5 Å resolution map of the complex with variable Cdc6 density. 3D classification into three 1162 classes generated a class from 51,222 particles containing the full YlODC complex, which was 1163 used for a final non-uniform refinement to produce the final 2.56 Å resolution map of YlODC60bpOri-1164 A006-CNNGGNR used for model building. All non-uniform refinements of the unbinned particles utilized 1165 per-particle scale minimization, per-particle defocus refinement, and CTF refinement of per-group 1166 CTF parameters, spherical aberration, and tetrafoil. A schematic summary of data processing is 1167 shown in Figure S5D. 1168 1169 Model building and validation 1170 1171 HsODC: The atomic model of HsORC (PDB ID: 7JPS) was used as the starting model for HsODC 1172 and rigid-body fitted into the cryo -EM density using ChimeraX. Regions of HsORC that were 1173 missing or did not fit well into the density were rebuilt manually in Coot. Iterative model building 1174 and refinement were performed in PHENIX (v1.20.1 –4487-000), with secondary -structure 1175 restraints applied throughout. Model validation was carried out using MolProbity and PHENIX 1176 validation tools. The final model showed good stereochemistry, with >95% of residues in favored 1177 regions of the Ramachandran plot, <0.5% outliers, and all bond length and bond angle deviations 1178 within acceptable limits. Structural figures were generated using ChimeraX and PyMOL (v2.5.5, 1179 Schrödinger, LLC). 1180 1181 YlODC: For the YlODC45bpOriC-061 structure, AlphaFold 2 models for each subunit were docked 1182 into the density individually using the “fit to map” functionality in ChimeraX 104 , then refined using 1183 the Coot software package 105 . The density for the DNA was sharp enough to allow us to discern 1184 purines and pyrimidines, allowing us to produce a generic DNA -B form model of the respective 1185 DNA sequence and manually rebuild it in Coot. This structure was then used as the basis of the 1186 other ODC structures. Structures were then further refined in Coot using “Real Space Refinement” 1187 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 32 function and ligands, ions, and waters were manually built. The Phenix software package was 1188 then used to further refine and finalize the structures, as well as provide validation metrics, via its 1189 “Real Space Refine” functionality 106 . Figures using these structures, along with comparisons to 1190 previously published ORC/ODC structures, were generated using ChimeraX. The preliminary 1191 YlODC45bpOriC-061 structure was then used as the basis for the YlODC54bpOriC-061 structure, which 1192 itself was used as the starting point for the YlODC60bpOri-A006-WT structure. Additionally, the 1193 YlODC60bpOri-A006-WT structure was used as a reference for building the YlODC60bpOri-A006-CNNGGNR 1194 structure. 1195 1196 1197

1198 1199 1. Hu, Y. & Stillman, B. Origins of DNA replication in eukaryotes. Mol. Cell 83, 352–372 1200 (2023). 1201 2. Hyrien, O., Guilbaud, G. & Krude, T. The double life of mammalian DNA replication origins. 1202 Genes Dev. 39, 304–324 (2025). 1203 3. Ekundayo, B. & Bleichert, F. Origins of DNA replication. PLoS genetics 15, e1008320 1204 (2019). 1205 4. Lee, C. S. K., Weiβ, M. & Hamperl, S. Where and when to start: Regulating DNA replication 1206 origin activity in eukaryotic genomes. Nucleus 14, 2229642 (2023). 1207 5. Costa, A. & Diffley, J. F. X. The Initiation of Eukaryotic DNA Replication. Annu. Rev. 1208 Biochem. 91, 107–131 (2022). 1209 6. Bleichert, F., Botchan, M. R. & Berger, J. M. Mechanisms for initiating cellular DNA 1210 replication. Science 355, eaah6317 (2017). 1211 7. Bell, S. P. & Labib, K. Chromosome Duplication in Saccharomyces cerevisiae. Genetics 203, 1212 1027–1067 (2016). 1213 8. Musiałek, M. W. & Rybaczek, D. Behavior of replication origins in Eukaryota - spatio-1214 temporal dynamics of licensing and firing. Cell cycle (Georgetown, Tex) 0 (2015) 1215 doi:10.1080/15384101.2015.1056421. 1216 9. Nishitani, H. & Lygerou, Z. Control of DNA replication licensing in a cell cycle. Genes Cells 1217 7, 523–534 (2002). 1218 10. Rhind, N. DNA replication timing: Biochemical mechanisms and biological significance. 1219 Bioessays 2200097 (2022) doi:10.1002/bies.202200097. 1220 11. Limas, J. C. & Cook, J. G. Preparation for DNA replication: the key to a successful S phase. 1221 FEBS letters 593, 2853–2867 (2019). 1222 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 33 12. Speck, C. & Reuter, L. M. Compact Origins and Where to Find Them: ORC’s Guide to 1223 Genome‐Wide Licensing. BioEssays e70018 (2025) doi:10.1002/bies.70018. 1224 13. Yeeles, J. T. P., Deegan, T. D., Janska, A., Early, A. & Diffley, J. F. X. Regulated eukaryotic 1225 DNA replication origin firing with purified proteins. Nature 519, 431–435 (2015). 1226 14. Kurat, C. F., Yeeles, J. T. P., Patel, H., Early, A. & Diffley, J. F. X. Chromatin Controls 1227 DNA Replication Origin Selection, Lagging-Strand Synthesis, and Replication Fork Rates. Mol. 1228 Cell 65, 117–130 (2017). 1229 15. Devbhandari, S., Jiang, J., Kumar, C., Whitehouse, I. & Remus, D. Chromatin Constrains the 1230 Initiation and Elongation of DNA Replication. Mol. Cell 65, 131–141 (2017). 1231 16. Zhang, Q., Lam, W. H. & Zhai, Y. Assembly and activation of replicative helicases at origin 1232 DNA for replication initiation. Curr. Opin. Struct. Biol. 88, 102876 (2024). 1233 17. Stillman, B., Diffley, J. F. X. & Iwasa, J. H. Mechanisms for licensing origins of DNA 1234 replication in eukaryotic cells. Nat. Struct. Mol. Biol. 1–11 (2025) doi:10.1038/s41594-025-1235 01587-5. 1236 18. Siow, C. C., Nieduszynska, S. R., Müller, C. A. & Nieduszynski, C. A. OriDB, the DNA 1237 replication origin database updated and extended. Nucleic acids research 40, D682-6 (2012). 1238 19. Brewer, B. J. & Fangman, W. L. The localization of replication origins on ARS plasmids in 1239 S. cerevisiae. Cell 51, 463–471 (1987). 1240 20. Attali, I., Botchan, M. R. & Berger, J. M. Structural Mechanisms for Replicating DNA in 1241 Eukaryotes. Annu. Rev. Biochem. 90, 1–30 (2021). 1242 21. Moyer, S. E., Lewis, P. W. & Botchan, M. R. Isolation of the Cdc45/Mcm2-7/GINS (CMG) 1243 complex, a candidate for the eukaryotic DNA replication fork helicase. Proceedings of the 1244 National Academy of Sciences of the United States of America 103, 10236–10241 (2006). 1245 22. Marahrens, Y. & Stillman, B. A Yeast Chromosomal Origin of DNA Replication Defined by 1246 Multiple Functional Elements. Science 255, 817–823 (1992). 1247 23. Theis, J. F. & Newlon, C. S. Domain B of ARS307 contains two functional elements and 1248 contributes to chromosomal replication origin function. Mol Cell Biol 14, 7652–7659 (1994). 1249 24. Rao, H., Marahrens, Y. & Stillman, B. Functional Conservation of Multiple Elements in 1250 Yeast Chromosomal Replicators. Mol. Cell. Biol. 14, 7643–7651 (1994). 1251 25. Reuter, L. M. et al. MCM2-7 loading-dependent ORC release ensures genome-wide origin 1252 licensing. Nat. Commun. 15, 7306 (2024). 1253 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 34 26. Coster, G. & Diffley, J. F. X. Bidirectional eukaryotic DNA replication is established by 1254 quasi-symmetrical helicase loading. Science 357, 314–318 (2017). 1255 27. Gupta, S., Friedman, L. J., Gelles, J. & Bell, S. P. A helicase-tethered ORC flip enables 1256 bidirectional helicase loading. Elife 10, e74282 (2021). 1257 28. Hu, Y. et al. Evolution of DNA replication origin specification and gene silencing 1258 mechanisms. Nat. Commun. 11, 5175 (2020). 1259 29. Wang, D. & Gao, F. Comprehensive Analysis of Replication Origins in Saccharomyces 1260 cerevisiae Genomes. Frontiers in Microbiology 10, 2122 (2019). 1261 30. Yuan, Z. et al. Structural basis of Mcm2–7 replicative helicase loading by ORC–Cdc6 and 1262 Cdt1. Nat. Struct. Mol. Biol. 24, 316–324 (2017). 1263 31. Li, N. et al. Structure of the origin recognition complex bound to DNA replication origin. 1264 Nature 559, 217–222 (2018). 1265 32. Lee, C. S. K. et al. Humanizing the yeast origin recognition complex. Nat. Commun. 12, 33 1266 (2021). 1267 33. Liachko, I. et al. A comprehensive genome-wide map of autonomously replicating sequences 1268 in a naive genome. PLoS genetics 6, e1000946 (2010). 1269 34. Liachko, I. et al. Novel features of ARS selection in budding yeast Lachancea kluyveri. BMC 1270 Genom. 12, 633 (2011). 1271 35. Hu, Y. et al. Evolution of DNA replication origin specification and gene silencing 1272 mechanisms. Nat. Commun. 11, 5175 (2020). 1273 36. Tsai, H.-J. et al. Origin replication complex binding, nucleosome depletion patterns, and a 1274 primary sequence motif can predict origins of replication in a genome with epigenetic 1275 centromeres. Mbio 5, e01703-14 (2014). 1276 37. Liachko, I. et al. GC-rich DNA elements enable replication origin activity in the 1277 methylotrophic yeast Pichia pastoris. PLoS genetics 10, e1004169 (2014). 1278 38. Evrin, C. et al. A double-hexameric MCM2-7 complex is loaded onto origin DNA during 1279 licensing of eukaryotic DNA replication. Proc. Natl. Acad. Sci. 106, 20240–20245 (2009). 1280 39. Remus, D. et al. Concerted Loading of Mcm2–7 Double Hexamers around DNA during 1281 DNA Replication Origin Licensing. Cell 139, 719–730 (2009). 1282 40. Weissmann, F. et al. MCM double hexamer loading visualized with human proteins. Nature 1283 636, 499–508 (2024). 1284 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 35 41. Yang, R., Hunker, O., Wise, M. & Bleichert, F. Multiple mechanisms for licensing human 1285 replication origins. Nature 636, 488–498 (2024). 1286 42. Wells, J. N. et al. Reconstitution of human DNA licensing and the structural and functional 1287 analysis of key intermediates. Nat. Commun. 16, 478 (2025). 1288 43. Tian, M. et al. Integrative analysis of DNA replication origins and ORC-/MCM-binding sites 1289 in human cells reveals a lack of overlap. eLife 12, RP89548 (2024). 1290 44. Ganier, O., Prorok, P., Akerman, I. & Méchali, M. Metazoan DNA replication origins. 1291 Current opinion in cell biology 58, 134–141 (2019). 1292 45. Wang, W. et al. Genome-wide mapping of human DNA replication by optical replication 1293 mapping supports a stochastic model of eukaryotic replication. Mol Cell (2021) 1294 doi:10.1016/j.molcel.2021.05.024. 1295 46. Pourkarimi, E., Bellush, J. M. & Whitehouse, I. Spatiotemporal coupling and decoupling of 1296 gene transcription with DNA replication origins during embryogenesis in C. elegans. Elife 5, 1297 e21728 (2016). 1298 47. Comoglio, F. et al. High-Resolution Profiling of Drosophila Replication Start Sites Reveals a 1299 DNA Shape and Chromatin Signature of Metazoan Origins. Cell Reports 11, 821–834 (2015). 1300 48. Masai, H. & Tanaka, T. G-quadruplex DNA and RNA: Their roles in regulation of DNA 1301 replication and other biological functions. Biochem. Biophys. Res. Commun. 531, 25–38 (2020). 1302 49. Valton, A.-L. & Prioleau, M.-N. G-Quadruplexes in DNA Replication: A Problem or a 1303 Necessity? Trends in Genetics 32, 697–706 (2016). 1304 50. Valton, A. L. et al. G4 motifs affect origin positioning and efficiency in two vertebrate 1305 replicators. The EMBO journal 33, 732–746 (2014). 1306 51. Prorok, P. et al. Involvement of G-quadruplex regions in mammalian replication origin 1307 activity. Nat Commun 10, 3274 (2019). 1308 52. Bartlett, D. A., Dileep, V., Baslan, T. & Gilbert, D. M. Mapping Replication Timing in 1309 Single Mammalian Cells. Curr Protoc 2, e334 (2022). 1310 53. Carrington, J. T. et al. Most human DNA replication initiation is dispersed throughout the 1311 genome with only a minority within previously identified initiation zones. Genome Biol. 26, 122 1312 (2025). 1313 54. Akerman, I. et al. A predictable conserved DNA base composition signature defines human 1314 core DNA replication origins. Nat Commun 11, 4826 (2020). 1315 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 36 55. Guilbaud, G. et al. Determination of human DNA replication origin position and efficiency 1316 reveals principles of initiation zone organisation. Nucleic Acids Res (2022) 1317 doi:10.1093/nar/gkac555. 1318 56. Langley, A. R., Gräf, S., Smith, J. C. & Krude, T. Genome-wide identification and 1319 characterisation of human DNA replication origins by initiation site sequencing (ini-seq). 1320 Nucleic acids research 44, 10230–10247 (2016). 1321 57. Petryk, N. et al. Replication landscape of the human genome. Nature Communications 7, 1322 10208 (2016). 1323 58. Krude, T., Bi, J., Doran, R., Jones, R. A. & Smith, J. C. Human DNA replication initiation 1324 sites are specified epigenetically by oxidation of 5-methyl-deoxycytidine. Nucleic Acids Res. 53, 1325 gkaf362 (2025). 1326 59. Méchali, M., Yoshida, K., Coulombe, P. & Pasero, P. Genetic and epigenetic determinants of 1327 DNA replication origins, position and activation. Curr. Opin. Genet. Dev. 23, 124–131 (2013). 1328 60. Srikant, S., Gaudet, R. & Murray, A. W. Extending the reach of homology by using 1329 successive computational filters to find yeast pheromone genes. Curr. Biol. (2023) 1330 doi:10.1016/j.cub.2023.08.039. 1331 61. Vernis, L. et al. An Origin of Replication and a Centromere Are Both Needed To Establish a 1332 Replicative Plasmid in the Yeast Yarrowia lipolytica. Mol. Cell. Biol. 17, 1995–2004 (1997). 1333 62. Vernis, L. et al. Short DNA Fragments without Sequence Similarity Are Initiation Sites for 1334 Replication in the Chromosome of the Yeast Yarrowia lipolytica. Mol. Biol. Cell 10, 757–769 1335 (1999). 1336 63. Vernis, L. et al. Only Centromeres Can Supply the Partition System Required for ARS 1337 Function in the Yeast Yarrowia lipolytica. J. Mol. Biol. 305, 203–217 (2001). 1338 64. Lopez, C., Cao, M., Yao, Z. & Shao, Z. Investigating the role of noncoding regulatory DNA 1339 in plasmid development for Yarrowia lipolytica. doi:10.22541/au.160691063.38058320/v1. 1340 65. Jaremko, M. J., On, K. F., Thomas, D. R., Stillman, B. & Joshua-Tor, L. The dynamic nature 1341 of the human origin recognition complex revealed through five cryoEM structures. eLife 9, 1342 e58622 (2020). 1343 66. Tocilj, A. et al. Structure of the active form of human Origin Recognition Complex and its 1344 ATPase motor module. (2017) doi:10.2210/pdb5ujm/pdb. 1345 67. Schmidt, J. M. et al. A mechanism of origin licensing control through autoinhibition of S. 1346 cerevisiae ORC·DNA·Cdc6. Nat. Commun. 13, 1059 (2022). 1347 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 37 68. Schmidt, J. M. & Bleichert, F. Structural mechanism for replication origin binding and 1348 remodeling by a metazoan origin recognition complex and its co-loader Cdc6. Nat Commun 11, 1349 4263 (2020). 1350 69. Feng, X. et al. The structure of ORC–Cdc6 on an origin DNA reveals the mechanism of 1351 ORC activation by the replication initiator Cdc6. Nat. Commun. 12, 3883 (2021). 1352 70. Lam, W. H. et al. DNA bending mediated by ORC is essential for replication licensing in 1353 budding yeast. Proc. Natl. Acad. Sci. 122, e2502277122 (2025). 1354 71. Erzberger, J. P. & Berger, J. M. Evolutionary Relationships and Structural Mechanism of 1355 AAA+ Proteins. Annu. Rev. Biophys. Biomol. Struct. 35, 93–114 (2006). 1356 72. Zali, N. et al. Genome sequence assembly and annotation of MATA and MATB strains of 1357 Yarrowia lipolytica. NAR Genom. Bioinform. 7, lqaf175 (2025). 1358 73. Fu, H., Baris, A. & Aladjem, M. I. Replication timing and nuclear structure. Curr. Opin. Cell 1359 Biol. 52, 43–50 (2018). 1360 74. Marchal, C., Sima, J. & Gilbert, D. M. Control of DNA replication timing in the 3D genome. 1361 Nat. Rev. Mol. Cell Biol. 20, 721–737 (2019). 1362 75. McKnight, S. L. & Kingsbury, R. Transcriptional Control Signals of a Eukaryotic Protein-1363 Coding Gene. Science 217, 316–324 (1982). 1364 76. Nelson, H. C. M., Finch, J. T., Luisi, B. F. & Klug, A. The structure of an 1365 oligo(dA)·oligo(dT) tract and its biological implications. Nature 330, 221–226 (1987). 1366 77. Haran, T. E., Kahn, J. D. & Crothers, D. M. Sequence Elements Responsible for DNA 1367 Curvature. J. Mol. Biol. 244, 135–143 (1994). 1368 78. Remus, D., Beall, E. L. & Botchan, M. R. DNA topology, not DNA sequence, is a critical 1369 determinant for Drosophila ORC–DNA binding. EMBO J. 23, 897–907 (2004). 1370 79. Tareen, A. et al. MAVE-NN: learning genotype-phenotype maps from multiplex assays of 1371 variant effect. Genome Biol. 23, 98 (2022). 1372 80. Oji, A., Choubani, L., Miura, H. & Hiratani, I. Structure and dynamics of nuclear A/B 1373 compartments and subcompartments. Curr. Opin. Cell Biol. 90, 102406 (2024). 1374 81. Vouzas, A. E. & Gilbert, D. M. Replication timing and transcriptional control: beyond cause 1375 and effect — part IV. Curr Opin Genet Dev 79, 102031 (2023). 1376 82. McCune, H. J. et al. The Temporal Program of Chromosome Replication: Genomewide 1377 Replication in clb5Δ Saccharomyces cerevisiae. Genetics 180, 1833–1847 (2008). 1378 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 38 83. Chang, F. et al. High-resolution analysis of four efficient yeast replication origins reveals 1379 new insights into the ORC and putative MCM binding elements. Nucleic Acids Res. 39, 6523–1380 6535 (2011). 1381 84. Bell, S. P. & Stillman, B. ATP-dependent recognition of eukaryotic origins of DNA 1382 replication by a multiprotein complex. Nature 357, 128–134 (1992). 1383 85. Bleichert, F., Leitner, A., Aebersold, R., Botchan, M. R. & Berger, J. M. Conformational 1384 control and DNA-binding mechanism of the metazoan origin recognition complex. Proceedings 1385 of the National Academy of Sciences 115, E5906–E5915 (2018). 1386 86. Miller, T. C. R., Locke, J., Greiwe, J. F., Diffley, J. F. X. & Costa, A. Mechanism of head-to-1387 head MCM double-hexamer formation revealed by cryo-EM. Nature 575, 704–710 (2019). 1388 87. Yuan, Z. et al. Structural mechanism of helicase loading onto replication origin DNA by 1389 ORC-Cdc6. Proc. Natl. Acad. Sci. 117, 17747–17756 (2020). 1390 88. Dahlin, J., Holkenbrink, C. & Borodina, I. Yarrowia lipolytica, Methods and Protocols. in 1391 vol. 2307 41–68 (2021). 1392 89. Lopez, C., Cao, M., Yao, Z. & Shao, Z. Revisiting the unique structure of autonomously 1393 replicating sequences in Yarrowia lipolytica and its role in pathway engineering. Appl Microbiol 1394 Biot 1–14 (2021) doi:10.1007/s00253-021-11399-4. 1395 90. Chen, S., Zhou, Y., Chen, Y. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. 1396 Bioinformatics 34, i884–i890 (2018). 1397 91. Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 1398 357–359 (2012). 1399 92. Danecek, P. et al. Twelve years of SAMtools and BCFtools. GigaScience 10, giab008 1400 (2021). 1401 93. Ramírez, F. et al. deepTools2: a next generation web server for deep-sequencing data 1402 analysis. Nucleic Acids Res. 44, W160–W165 (2016). 1403 94. Zhang, Y. et al. Model-based Analysis of ChIP-Seq (MACS). Genome Biol. 9, R137 (2008). 1404 95. Heinz, S. et al. Simple Combinations of Lineage-Determining Transcription Factors Prime 1405 cis-Regulatory Elements Required for Macrophage and B Cell Identities. Mol. Cell 38, 576–589 1406 (2010). 1407 96. Kinney, J. B., Murugan, A., Callan, C. G. & Cox, E. C. Using deep sequencing to 1408 characterize the biophysical mechanism of a transcriptional regulatory sequence. Proceedings of 1409 the National Academy of Sciences 107, 9158–9163 (2010). 1410 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 39 97. Tareen, A. & Kinney, J. B. Logomaker: beautiful sequence logos in Python. Bioinformatics 1411 36, 2272–2274 (2020). 1412 98. Bieniossek, C., Richmond, T. J. & Berger, I. MultiBac: Multigene Baculovirus‐Based 1413 Eukaryotic Protein Complex Production. Curr. Protoc. Protein Sci. 51, 5.20.1-5.20.26 (2008). 1414 99. Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 1415 9, 676–682 (2012). 1416 100. Tegunov, D. & Cramer, P. Real-time cryo-electron microscopy data preprocessing with 1417 Warp. Nat. Methods 16, 1146–1152 (2019). 1418 101. Zivanov, J., Nakane, T. & Scheres, S. H. W. Estimation of high-order aberrations and 1419 anisotropic magnification from cryo-EM data sets in RELION-3.1. IUCrJ 7, 253–267 (2020). 1420 102. Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for 1421 rapid unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017). 1422 103. Punjani, A., Zhang, H. & Fleet, D. J. Non-uniform refinement: adaptive regularization 1423 improves single-particle cryo-EM reconstruction. Nat. Methods 17, 1214–1221 (2020). 1424 104. Goddard, T. D. et al. UCSF ChimeraX: Meeting modern challenges in visualization and 1425 analysis. Protein science : a publication of the Protein Society 27, 14–25 (2018). 1426 105. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. 1427 Acta Crystallogr. Sect. D 66, 486–501 (2010). 1428 106. Liebschner, D. et al. Macromolecular structure determination using X‐rays, neutrons and 1429 electrons: recent developments in Phenix. Acta Crystallogr. Sect. D 75, 861–877 (2019). 1430 1431 1432 1433 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint D BA C Figure 1 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint Figure 2 30’ 75’ 90’ 180’ 210’ 60’ 45’ 120’ 30’ 45’ 60’ +HU 160’ 2.3 Mb Chr. A -HU Log2 Ratio (obs/exp)Total size (bp)Number of peaksAnnotation 0.0574597943150TTS -0.99620384998Exon 0.7013196821163Intergenic 0.1796787639241Promoter/TSS F E D C B CEN Telomeric repeats rDNA clusters A 30’ 120’ A B C .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint 0 10 20 30 40 50 60 WT OriC-061 Linkers 2 to 45 AATATT ACACCC AAGTAG CATGCA TAAGCT CTCGAG CTCGAG CTCGAG –TC-AG CTC-AG 2 3 4 5 6 WT 3% 1.2% 41%0% 2% 0% Mitotic Stability (%) Figure 3 A B 25 26 27 28 29 0 10 20 30 40 50 60 WT Linkers 1 to 44 Boxed linkers 25 to 29 WT 5% 45%0% 0% 0% 0% OriA-006 Mitotic Stability (%) .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint A B C D E Figure 4 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint Figure 5 C WT oriA006 (CNNCCNR) oriA006 CNNGGNR mutant B A OriA-006 10 29 28 27 26 25 15 20 25 30 35 5- ATATGCCCCTCCAATCCAGCTCCTACAAGT -3 CDC6 ORC4 5- AATATTACACCCAAGTAGCATGCATAAGCT -3 3 4 5 6 7 15 20 25 30 35 40 OriC-061 CDC6 ORC4 C20,21 A C17,20,21G A 23 C 4% C17 A C21 A C21,23,24G A 26 C 4% C21 A 48% C23 A 48% C24 A 48% 0% 0% 48% 22% 47% C17,20,21T 0% 0% C T CDC6 CDC6 45% WT 21,23,34 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint Figure 6 A B C Wild type sequence OriA-006 5- ATATGCCCCTCCAATCCAGCTCCTACAAGT -3OriA-006 ORC3 CDC6 ORC3 ORC2 & ORC5CDC6 ORC4 T11 G A 12 C 12 29 28 27 26 25 17 22 27 32 37 T33 G A 34 C C 35 G A 36 C C35 G A 36 C T34 G C 35 A T 33 A A 34 T 25% 5% 0% 0% 0% 0% 0% .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint S. cerevisiae Y. lipolytica H. sapiens ORC2 loop ORC-CDC6-DNA bend ORC4 α-helix Figure 7 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 10, 2026. ; https://doi.org/10.64898/2026.03.10.710760doi: bioRxiv preprint