MCM double hexamer loading visualized with human proteins

Protein sequences

Human cDNAs were cloned from RPE-1 and U2OS cells by total RNA isolation, oligo-dT reverse transcription, followed by PCR with gene-specific oligonucleotides and cloning into expression vectors. Sanger sequencing showed that the cloned cDNA coded for the reference protein sequence for the following proteins: ORC1, ORC2, ORC3 (NCBI: iso2, Uniprot: canonical iso), ORC4, ORC5, ORC6, MCM2, MCM3, MCM5, MCM6, MCM7, CDC6 and GMNN. For MCM4 and CDT1, the cloned cDNA coded for a natural variant that differed from NCBI reference protein sequences, but according to the Genome Aggregation Database (https://gnomad.broadinstitute.org) represented the majority allele (allele frequency > 0.5): MCM4_L650M (allele frequency: 0.862), CDT1_C234R (allele frequency: 0.999). These alleles were used and considered wild type throughout this study.

Expression and purification of human MCM2–7

MCM2–7 was expressed in insect cells using the biGBac baculovirus expression vector system¹⁰. Human cDNAs of MCM2, MCM3, MCM4, MCM5, MCM6 and MCM7 were cloned into pLIB. MCM3 was subcloned to contain an N-terminal TEV protease-cleavable Flag tag. pLIB-derived expression cassettes of MCM4, MCM5, MCM6, MCM7 were subcloned into pBIG1a (pBIG1a:MCM4, MCM5, MCM6, MCM7) and expression cassettes of MCM2 and Flag–MCM3 were subcloned into pBIG1b (pBIG1b:MCM2, Flag–MCM3). Expression cassettes of these two vectors were subcloned into pBIG2ab (pBIG2ab:MCM4, MCM5, MCM6, MCM7, MCM2, Flag–MCM3). Baculovirus was generated and amplified in Sf9 cells (Thermo Fisher, 12659017) using the EMBacY baculoviral genome⁴¹. For protein expression, Sf9 cells were infected and collected 52 h after infection, flash-frozen and stored at −80 °C.

Cell pellets were thawed on ice in MCM buffer (50 mM HEPES/KOH pH 7.6, 100 mM potassium glutamate, 5 mM magnesium acetate, 10% glycerol, 0.02% NP-40, 1 mM dithiothreitol (DTT)) + protease inhibitors (1 tablet per 50 ml Roche Complete Ultra EDTA-free, 10 µg ml⁻¹ leupeptin, 10 µg ml⁻¹ pepstatin A, 1 mM AEBSF, 1 µg ml⁻¹ aprotinin, 2 mM benzamidine). Cells were resuspended and lysed by Dounce homogenization using a tight-fitting pestle. The lysate was cleared by centrifugation (158,000g, 4 °C, 1 h) and incubated with anti-Flag M2 affinity gel (Sigma-Aldrich, A2220) for 1 h at 4 °C with rotation. The beads were transferred to a Econo-Pac gravity flow column (Bio-Rad, 732-1010) and washed with 2× 20 column volumes MCM buffer, 1× 20 column volumes MCM buffer + 5 mM ATP (Sigma-Aldrich, A2383). For dephosphorylation, the beads were resuspended in MCM buffer + 0.2 mg ml⁻¹ lambda protein phosphatase, 1 mM manganese (ii) chloride and incubated for 1 h at 4 °C. The beads were washed with 2× 20 column volumes MCM buffer. For elution and proteolytic tag removal the beads were resuspended in 5 column volumes MCM buffer + 80 µg ml⁻¹ TEV protease, 0.1 mg ml⁻¹ 3×Flag peptide for 2 h at 4 °C. The eluate was concentrated using an Amicon Ultra-15 concentrator (Merck, UFC903024) and further purified by gel filtration using a HiLoad 16/600 Superdex 200 pg column equilibrated in MCM buffer. Fractions containing stoichiometric MCM2–7 were concentrated, flash-frozen in liquid nitrogen, and stored at −80 °C.

Expression and purification of human ORC1–5

ORC1–5 was expressed in baculovirus-infected insect cells. The coding sequences of human ORC1, ORC2, ORC3, ORC4 and ORC5 were codon-optimized for Spodoptera frugiperda, synthesized (GeneArt, Thermo Fisher Scientific) and subcloned into modified pBIG1 vectors that contain a pLIB-derived polyhedrin expression cassette. ORC1 was subcloned into pBIG1a with a TEV protease-cleavable N-terminal 3× Flag tag. ORC2 was subcloned into pBIG1b, ORC3 into pBIG1c, ORC4 into pBIG1d, ORC5 into pBIG1e. Expression cassettes from these five vectors were subcloned into pBIG2abcde (pBIG2abcde:Flag-ORC1,2,3,4,5). Baculoviruses were generated using EMBacY and Sf9 cells. To express ORC1–5, Sf9 cells were infected, collected 52 h after infection, flash-frozen and stored at −80 °C.

Cell pellets were thawed in ORC1–5 lysis buffer (50 mM HEPES/KOH 7.6, 650 mM potassium chloride, 5 mM magnesium acetate, 1 mM ATP, 10% glycerol, 0.02% NP-40, 1 mM DTT, 2 mM benzamidine) + protease inhibitors (1 tablet per 50 ml Roche Complete Ultra EDTA-free, 10 µg ml⁻¹ leupeptin, 10 µg ml⁻¹ pepstatin A, 1 mM AEBSF, 1 µg ml⁻¹ aprotinin). Cells were lysed by Dounce homogenization, and the lysate cleared by centrifugation (158,000g, 4 °C, 1 h). ORC1–5 was bound to anti-Flag M2 affinity gel for 1 h at 4 °C. The column was washed with 2× 20 column volumes ORC1–5 lysis buffer + 4 mM ATP. For dephosphorylation, the beads were resuspended in 1 column volumes ORC1–5 lysis buffer + 0.2 mg ml⁻¹ lambda protein phosphatase, 1 mM manganese (ii) chloride, potassium chloride adjusted to 650 mM, and incubated at 4 °C for 1 h. The column was washed with ORC1–5 lysis buffer. For elution and proteolytic tag removal, the beads were resuspended in 5 column volumes ORC1–5 lysis buffer + 80 µg ml⁻¹ TEV protease, 0.1 mg ml⁻¹ 3× Flag peptide, potassium chloride adjusted to 650 mM and were incubated at 4 °C for 2 h. To remove TEV protease, the eluate was supplemented with 35 mM imidazole pH 8.0 and incubated with Ni-NTA Agarose (Invitrogen, R90115) for 1 h at 4 °C. The flowthrough was concentrated and further purified by gel filtration using a HiLoad 16/600 Superdex 200 pg column (Cytiva) equilibrated in ORC1–5 SEC buffer (50 mM HEPES/KOH pH 7.6, 650 mM potassium chloride, 5 mM magnesium acetate, 10% glycerol, 0.02% NP-40, 1 mM DTT). Fractions containing stoichiometric ORC1–5 were concentrated, flash-frozen, and stored at −80 °C.

Expression and purification of human CDC6

CDC6 was expressed in insect cells. Human cDNA of CDC6 was cloned into pLIB to contain an N-terminal TEV protease-cleavable Flag tag. The baculovirus was generated in Sf9 cells using the EMBacY genome. For expression, Sf9 cells were infected and the culture collected 52 h after infection. The cell pellet was flash-frozen in liquid nitrogen and stored at −80 °C.

The cell pellet was thawed on ice in CDC6 lysis buffer (50 mM HEPES/KOH pH 7.6, 650 mM potassium chloride, 5 mM magnesium acetate, 4 mM benzamidine, 1 mM ATP, 10% glycerol, 0.02% NP-40, 1 mM DTT) + protease inhibitors (1 tablet per 50 ml Roche Complete Ultra EDTA-free, 10 µg ml⁻¹ leupeptin, 10 µg ml⁻¹ pepstatin A, 1 mM AEBSF, 1 µg ml⁻¹ aprotinin). Cells were lysed with a Dounce homogenizer and the lysates centrifuged (158,000g, 4 °C, 1 h). The cleared lysate was incubated with anti-Flag M2 affinity gel for 1 h at 4 °C. The beads were washed with 2× 20 column volumes CDC6 lysis buffer + 4 mM ATP. For dephosphorylation, the beads were resuspended in CDC6 lysis buffer + 0.2 mg ml⁻¹ lambda protein phosphatase, 1 mM manganese (ii) chloride and incubated for 1 h at 4 °C. The column was washed with CDC6 lysis buffer, followed by CDC6 HTP-wash buffer (50 mM potassium phosphate pH 7.6, 75 mM potassium acetate, 5 mM magnesium acetate, 0.1% Triton X-100, 1 mM DTT, 2 mM ATP). For elution, the beads were resuspended in CDC6 HTP-wash buffer + 80 µg ml⁻¹ TEV protease, 0.1 mg ml⁻¹ 3× Flag peptide and incubated for 2 h at 4 °C. To remove TEV protease, the eluate was supplemented with 30 mM imidazole pH 8.0 and incubated with Ni-NTA agarose for 1 h at 4 °C. The CDC6-containing flowthrough was collected. A hydroxyapatite column was prepared by resuspending 2 g Bio-Gel HTP Hydroxyapatite (Bio-Rad, 130-0420) in 12 ml CDC6 HTP-wash buffer. To remove fine particles, the beads were allowed to settle for 2 min and the supernatant was removed. Two more times, the beads were resuspended in 10 ml CDC6 HTP-wash buffer, allowed to settle for 2 min, and the supernatant removed. Then, 4 ml of a 50% slurry were incubated with CDC6 for 15 min at 4 °C. The beads were transferred to a gravity flow column and washed with 2 ml CDC6 HTP-wash buffer, followed by 5 ml CDC6 HTP-rinse buffer (50 mM potassium phosphate pH 7.6, 150 mM potassium acetate, 5 mM magnesium acetate, 0.1% Triton X-100, 15% glycerol, 1 mM DTT). CDC6 was eluted by applying 10 ml CDC6 HTP-elution buffer (50 mM potassium phosphate pH 7.6, 400 mM potassium acetate, 5 mM magnesium acetate, 0.1% Triton X-100, 15% glycerol, 1 mM DTT). The eluate was dialysed 2×1 h at 4 °C against CDC6 dialysis buffer (50 mM HEPES/KOH pH 7.6, 650 mM potassium chloride, 5 mM magnesium acetate, 10% glycerol, 0.02% NP-40, 1 mM DTT). CDC6 was concentrated, flash-frozen, and stored at −80 °C.

Expression and purification of human CDT1

CDT1 was expressed in insect cells. The coding sequence of human CDT1 was codon-optimized for S. frugiperda, synthesized (GeneArt, Thermo Fisher Scientific) and subcloned into pLIB as a fusion protein with an N-terminal Flag–His–SumoEu1 fusion⁴². The baculovirus was generated in Sf9 cells using the EMBacY genome. Expression cultures were collected 52 h after infection, snap-frozen in liquid nitrogen and stored at −80 °C.

The cell pellet was thawed in CDT1 buffer (50 mM HEPES pH 7.6, 650 mM potassium chloride, 5 mM magnesium acetate, 10% glycerol, 0.02% NP-40, 1 mM DTT) + protease inhibitors (1 tablet per 50 ml Roche Complete Ultra EDTA-free, 10 µg ml⁻¹ leupeptin, 10 µg ml⁻¹ pepstatin A, 1 mM AEBSF, 1 µg ml⁻¹ aprotinin). Cells were lysed using a Dounce homogenizer. The lysate was cleared by centrifugation (158,000g, 4 °C, 1 h) and incubated with anti-Flag M2 affinity gel for 1 h at 4 °C. The beads were transferred to a gravity flow column and washed twice with CDT1 buffer + 5 mM ATP. The beads were resuspended in CDT1 buffer + 0.2 mg ml⁻¹ lambda protein phosphatase, 1 mM manganese (ii) chloride and incubated for 1 h at 4 °C. The beads were washed with CDT1 buffer. For proteolytic elution the beads were resuspended in CDT1 buffer + 80 µg ml⁻¹ His-SENP1_EuH protease⁴² and incubated for 2 h at 4 °C. To remove His-SENP1_EuH protease, 35 mM imidazole pH 8.0 was added and the eluate incubated with Ni-NTA agarose for 1 h at 4 °C. The CDT1-containing flowthrough was concentrated and further purified by gel filtration on a Superdex 200 Increase 10/300GL column (Cytiva) using CDT1 buffer. CDT1-containing fractions were concentrated, snap-frozen, and stored at −80 °C.

Expression and purification of human ORC6

ORC6 was expressed in Escherichia coli. The coding sequence of human ORC6 was codon-optimized for E. coli, synthesized (GeneArt, Thermo Fisher Scientific) and subcloned into pK27Sumo to encode the fusion protein His–Sumo–ORC6. The protein was expressed using the strain T7express lysY (NEB, C3010I). Expression was induced with 0.4 mM IPTG at OD600 ~ 0.6 at 16 °C. The culture was collected after 16 h, flash-frozen, and stored at −80 °C.

The cell pellet was thawed in ORC6 lysis buffer (50 mM HEPES pH 7.6, 500 mM potassium chloride, 5 mM magnesium acetate, 10% glycerol, 0.02% NP-40, 1 mM DTT, 35 mM imidazole) + protease inhibitors (1 tablet per 50 ml Roche Complete EDTA-free, 1 mM AEBSF), lysozyme. The cells were resuspended and lysed by sonication. The lysate was centrifuged (158,000g, 4 °C, 1 h) and the cleared lysate incubated with Ni-NTA agarose (Invitrogen) for 1 h at 4 °C. The beads were transferred to a gravity flow column and washed twice with ORC6 lysis buffer + 5 mM ATP and twice with ORC6 lysis buffer. The protein was eluted with ORC6 elution buffer (50 mM HEPES/KOH pH 7.6, 200 mM potassium chloride, 5 mM magnesium acetate, 10% glycerol, 0.02% NP-40, 1 mM DTT, 300 mM imidazole pH 8.0). The Sumo-specific protease His-Ulp1 was added at 0.08 mg ml⁻¹ and the protein was dialysed overnight against ORC6 SEC buffer (50 mM HEPES/KOH pH 7.6, 200 mM potassium chloride, 5 mM magnesium acetate, 10% glycerol, 0.02% NP-40, 1 mM DTT). To remove His-Ulp1 protease and His–Sumo, the imidazole concentration was adjusted to 35 mM and the protein incubated with Ni-NTA agarose (Invitrogen) for 1 h at 4 °C. The ORC6 containing flowthrough was concentrated and further purified by gel filtration using a HiLoad 16/600 Superdex 200 pg column (Cytiva) equilibrated in ORC6 SEC buffer. ORC6 containing fractions were pooled, snap-frozen, and stored at −80 °C.

Expression and purification of truncated loading factors

The ORC1–5 complex with ORC1(∆N) (ORC1–5(1∆N)), and CDC6(∆N) and CDT1(∆N) with N-terminal IDR truncations were expressed in insect cells using a Flag–His–SumoEu1 fusion system⁴² and sequences that were codon-optimized for S. frugiperda and synthesized (GeneArt, Thermo Fisher Scientific). The SumoEu1 fusions were only partially stable in Sf9 cells giving relatively low yields. Flag–His–SumoEu1–ORC1(391–861) was cloned into pBIG1a containing a pLIB-derived polyhedrin expression cassette. The expression cassette was subcloned together with ORC2–5 expression cassettes into pBIG2abcde (pBIG2abcde:Flag–His–SumoEu1–ORC1(∆N), ORC2–5). Flag–His–SumoEu1–CDC6(∆N) (CDC6 residues 143–560) and Flag–His–SumoEu1–CDT1(∆N) (CDT1 residues 167–546) were cloned into pLIB. Baculoviruses were generated in Sf9 cells using EMBacY. Sf9 expression cultures were collected 52 h after infection, flash-frozen and stored at −80 °C.

The three proteins were purified with the same purification protocol with only the gel filtration column differing. Cell pellets were thawed in Wash-300 buffer (50 mM HEPES/KOH pH 7.6, 300 mM potassium chloride, 5 mM magnesium acetate, 1 mM ATP, 10% glycerol, 0.02% NP-40, 1 mM DTT) + protease inhibitors (1 tablet per 50 ml Roche Complete Ultra EDTA-free, 10 µg ml⁻¹ leupeptin, 10 µg ml⁻¹ pepstatin A, 1 mM AEBSF, 1 µg ml⁻¹ aprotinin, 4 mM benzamidine) and lysed using a Dounce homogenizer. The lysate was centrifuged (158,000g, 4 °C, 1 h) and the cleared lysate incubated with anti-Flag M2 affinity gel for 2 h at 4 °C. The beads were washed twice with Wash-300 buffer + 4 mM ATP, and then resuspended in Wash-300 buffer + 0.2 mg ml⁻¹ lambda protein phosphatase, 1 mM manganese (ii) chloride and incubated for 1 h at 4 °C. The beads were washed with Wash-300 buffer. For proteolytic elution, the beads were resuspended in Wash-300 buffer + 80 µg ml⁻¹ His-SENP1_EuH protease⁴² and incubated for 2 h at 4 °C. To remove His-SENP1_EuH protease, the eluate was supplemented with 35 mM imidazole pH 8.0 and incubated with Ni-NTA agarose (Invitrogen) for 1 h at 4 °C. The flowthrough was concentrated and further purified by gel filtration using SEC-∆N buffer (50 mM HEPES/KOH pH 7.6, 500 mM potassium glutamate, 5 mM magnesium acetate, 10% glycerol, 0.02% NP-40, 1 mM DTT). ORC1–5(1∆N) was purified using a HiLoad 16/600 Superdex 200 pg column (Cytiva). CDC6(∆N) and CDT1(∆N) were purified using a Superdex 200 Increase 10/300GL column (Cytiva). Fractions containing the respective protein were concentrated, snap-frozen, and stored at −80 °C.

Expression and purification of human geminin

Geminin was expressed in insect cells. Human cDNA of geminin was cloned into pLIB with an N-terminal Flag tag. The baculovirus was generated in Sf9 cells using EMBacY. Geminin was expressed in Sf9 cells and the culture collected 52 h after infection. The cell pellet was stored at −80 °C.

The cell pellet was thawed in GMNN lysis buffer (50 mM HEPES/KOH pH 7.6, 300 mM potassium chloride, 5 mM magnesium acetate, 10% glycerol, 0.02% NP-40, 1 mM DTT) + protease inhibitors (1 tablet per 50 ml Roche Complete Ultra EDTA-free, 10 µg ml⁻¹ leupeptin, 10 µg ml⁻¹ pepstatin A, 1 mM AEBSF, 1 µg ml⁻¹ aprotinin) and lysed using a Dounce homogenizer. The lysate was cleared by centrifugation (158,000g, 60 min, 4 °C) and incubated with anti-Flag M2 affinity gel for 1 h. The beads were washed with GMNN lysis buffer + 5 mM ATP, followed by GMNN lysis buffer. Protein was eluted in GMNN lysis buffer + 0.1 mg ml⁻¹ 3×Flag peptide. The protein was concentrated and further purified by gel filtration on a Superdex 200 Increase 10/300GL column equilibrated in GMNN SEC buffer (50 mM HEPES/KOH pH 7.6, 200 mM potassium chloride, 5 mM magnesium acetate, 10% glycerol, 0.02% NP-40, 1 mM DTT). Geminin containing fractions were concentrated and the protein concentration determined for the homodimer. The protein was flash-frozen, and stored at −80 °C.

Mass spectrometry

Protein preparations of human MCM2–7, ORC1–5 (FL), ORC1–5(1∆N), CDC6 (FL), CDC6(∆N), CDT1 (FL) and CDT1(∆N) were subjected to mass spectrometry. Insect cell ORC6 and CDC6 were not detected in any purification.

Nuclease footprinting assay with human proteins

MCM loading reactions were performed in assay buffer (25 mM HEPES/KOH pH 7.6, 100 mM sodium acetate, 10 mM magnesium acetate, 1 mM DTT). MCM was typically loaded onto a ARS1-containing 10.6 kb plasmid that was purified by caesium chloride density gradient centrifugation (pJY22⁴³).

A 20 µl reaction with truncated proteins typically contained 4 nM (27.5 ng µl⁻¹) plasmid DNA (10.6 kb), 2 mM ATP, 60 nM MCM2–7, 150 nM ORC6, 150 nM CDC6(∆N), 150 nM CDT1(∆N), and 90 nM ORC1–5(1∆N). Stocks of MCM2–7, CDC6(∆N), CDT1(∆N), and ORC1–5(1∆N) were diluted to a 10× working concentration in assay buffer. ORC6 was diluted to a 20× working concentration in assay buffer + 200 mM sodium chloride. Reactions were started by adding MCM2–7, ORC6, CDC6(∆N), CDT1(∆N) and ORC1–5(1∆N) to a mix of DNA and ATP in assay buffer.

A 20 µl reaction with full-length proteins typically contained 4 nM (27.5 ng µl⁻¹) plasmid DNA (10.6 kb), 2 mM ATP, 60 nM MCM2–7, 150 nM ORC6, 120 nM ORC1–5, 150 nM CDC6, and 150 nM CDT1. MCM2–7 was diluted to 10× working concentration in assay buffer. ORC6 was diluted to 20× working concentration in assay buffer + 200 mM sodium chloride. A 20× OCC mix of ORC1–5, CDC6 and CDT1 was prepared in assay buffer + 650 mM sodium chloride. Reactions were started by adding MCM2–7, ORC6, and the OCC mix to a mix of DNA and ATP in assay buffer.

In experiments that contained truncated proteins as well as their full-length counterparts, ORC1–5 and ORC1–5(1∆N) were both used at 120 nM and the 20× OCC mix (650 mM sodium chloride) method was used. When proteins were omitted, a salt-containing buffer was added instead to maintain the same salt concentration throughout the experiment.

Reactions were started in 2-min intervals and incubated at 37 °C in a thermomixer with shaking at 1,250 rpm. After 30 min of MCM loading, 2 µl of benzonase nuclease (Sigma-Aldrich, E1014) was added and the mix incubated at 37 °C with shaking for either 0.5 min or 20 min depending on the experiment. Then, 20 µl of the mix was transferred to a tube containing 10 µl 3× Stop buffer (assay buffer + 100 mM EDTA, 500 µM proteinase K (Sigma-Aldrich, 107393), 1% SDS) and the mix incubated at 37 °C for 20 min with shaking.

The sample was diluted with TE (10 mM Tris/HCl pH 8.0, 1 mM EDTA) to 200 µl and an equal volume of phenol/chloroform/isoamyl alcohol (25/24/1; Invitrogen UltraPure, 15593031) was added. The sample was vortexed for 1 min, transferred to a 5PRIME Phase Lock Gel Heavy spin column (VWR, 733-2478), and spun for 5 min at 20,000g. The aqueous phase was transferred to a new tube, 20 µl of 3 M sodium acetate pH 5.2, 1 µl of 20 mg ml⁻¹ glycogen (Thermo Scientific, R0561), and 550 µl ethanol were added and DNA precipitated overnight at −20 °C. The DNA was pelleted (20,000g, 4 °C, 40 min), washed with 80% ethanol (20,000g, 4 °C, 10 min), and air-dried. The pellet was resuspended in 3 µl TE. Then, 1.5 µl 20% Ficoll 400 (Sigma-Aldrich, F2637) was added, and samples were loaded on a pre-run 4–20% Novex TBE gel (Invitrogen, EC62252BOX). The gel was run using TBE running buffer (Invitrogen, LC6675) at 150 V for 50 min, stained with SYBR Gold (Invitrogen, S11494) for 30 min, and imaged using an Amersham ImageQuant 800 imager.

Salt stability experiments

For the salt stability experiment in Fig. 1j, the nuclease footprinting assay was performed with the following modifications. MCM loading reactions (20 µl) were set up in assay buffer and were incubated (37 °C, 30 min, 1,250 rpm). Then, 10 µl of 3× sodium chloride in 1× assay buffer solutions were added to achieve the indicated sodium chloride concentrations, and the reaction was incubated (37 °C, 15 min, 1,250 rpm). The reactions were diluted to a volume of 200 µl with solutions that adjusted all reaction buffers to 25 mM HEPES/KOH pH 7.6, 20 mM sodium acetate, 150 mM sodium chloride, 10 mM magnesium acetate, 1 mM DTT. Immediately, 10 µl of benzonase nuclease (Sigma-Aldrich, E1014) was added and the mix incubated (37 °C, 0.5 min, 1,250 rpm). Then, 200 µl were transferred into 100 µl 3× Stop buffer and the mix incubated (37 °C, 20 min, 1,250 rpm). The samples were extracted with an equal volume of phenol–chloroform–isoamyl alcohol. The aqueous phase was diluted 3-fold with TE, and DNA precipitated with 50% isopropanol, 200 mM sodium chloride, 20 µg glycogen, followed by an 80% ethanol wash. The salt stability experiments in Extended Data Fig. 1e,f were performed in a similar way with the following modifications. MCM was loaded for 30 min in a single large reaction and split into two tubes before adding assay buffer or a 3× sodium chloride in 1× assay buffer solution to achieve the indicated salt concentration. Samples were taken at the indicated timepoints, diluted, treated with benzonase, and transferred into Stop buffer. For the salt stability experiments in Fig. 3g and Extended Data Fig. 1g, MCM was loaded using the indicated reaction conditions for 30 min. Then, assay buffer or NaCl-containing assay buffer was added to achieve the indicated salt concentration. After 15 min incubation, samples were diluted, treated with benzonase for 0.5 min and transferred into Stop buffer.

Geminin inhibition experiments

For the geminin inhibition experiment shown in Fig. 1i, geminin and CDT1 were mixed on ice prior to setting up the MCM loading reactions. Geminin/CDT1 mixes were prepared at 20× concentration in assay buffer + 200 mM sodium chloride. Reactions were then started by adding MCM (10×, assay buffer), ORC6 (20×, 200 mM NaCl), geminin-CDT1 mix (20×, 200 mM NaCl), and ORC1–5/CDC6 mix (40×, 650 mM NaCl) to a mix of DNA and ATP in assay buffer. For the ATPγS timecourse experiment in Extended Data Fig. 1m geminin was added at minute 9 at 300 nM (twofold excess over 150 nM CDT1(∆N)).

Nuclease footprinting assay with yeast proteins

Nuclease footprinting experiments with yeast proteins were performed at 30 °C instead of 37 °C. Reactions with yeast proteins contained 4 nM (27.5 ng µl⁻¹) pJY22 plasmid DNA (10.6 kb, caesium chloride purified), 2 mM ATP, 100 nM Mcm2–7/Cdt1, 40 nM yCdc6 and 40 nM Orc1–6. In experiments containing human and yeast reactions, both were carried out at 30 °C.

Experiments with human origin sequences

A 2,398 bp fragment of the origin at the human lamin B2 (LMNB2) locus, and a 2,398 bp fragment of the origin at the human MYC locus were cloned into the vector pBIG1c. The LMNB2 origin fragment was amplified from human genomic DNA by PCR using oligonucleotides hOri-LamB2-2.4_for (AACGCTCTATGGTCTAAAGATTTACTCAGCAGCCCGGTG) and hOri-LamB2-2.4_rev (AACCCCGATTGAGATATAGATTTTGAGAATTGAGTCTTTGGAAACACTAAG). The MYC origin fragment was amplified using oligonucleotides hOri-Myc_for (AACGCTCTATGGTCTAAAGATTTAAGCTTGTTTGGCCGTTTTAGGG) and hOri-Myc_rev (AACCCCGATTGAGATATAGATTTCTCGAGGCAGGAGGGGAG). The fragments were inserted into pBIG1c that was linearized with the restriction enzyme SwaI using Gibson assembly, and the constructs were sequence verified. Doubly biotinylated DNA fragments of 2,398 bp size were generated by PCR using 5′-biotinylated oligonucleotides (Integrated DNA Technologies): LamB2 ori using Bio-LamB2-2.4_for ([5′-biotin]ACTCAGCAGCCCGGTG) and Bio-LamB2-2.4_rev ([5′-biotin]TGAGAATTGAGTCTTTGGAAACACTAAG), Myc ori using Bio-Myc-2.4_for ([5′-biotin]AAGCTTGTTTGGCCGTTTTAGGG) and Bio-Myc-2.4_rev ([5′-biotin]CTCGAGGCAGGAGGGGAG), yeast Ars1 using Bio-Ars1-2.4_for ([5′-biotin]GGTGGAGATATTCCTTATGGCATG) and Bio-Ars1-2.4_rev ([5′-biotin]GTAATTCGACCATTCCGACACAG) on pJY22, no origin (pET21a backbone) using Bio-pET21a-2.4_for ([5′-biotin]CCACAGGTGCGGTTGC) and Bio-pET21a-2.4_rev ([5′-biotin]TTCACCGTCATCACCGAAAC) on pET21a. The PCR products were column purified (QIAquick PCR purification kit, Qiagen) and purity confirmed by agarose gel electrophoresis. For the experiments shown in Extended Data Fig. 1h,i nuclease footprinting assays were performed using 4 nM doubly biotinylated PCR products and 400 nM streptavidin (Thermo Fisher, 434301).

DNA templates for electron microscopy experiments

The DNA substrate used for electron microscopy imaging of wild-type MCM loading was modified from the pGC209 (ref. ⁴⁴) construct, containing two inverted ACS sequences spaced by 70 bp, by adding Widom601 and Widom603 strong positioning sequences at both ends. The Widom sequences map 7 and 5 base pairs away from the inverted ACS sites, making the nucleosome-free region 148 base pairs long. For the experiment shown in Extended Data Fig. 1a, the pGC211 (ref. ⁴⁴) construct, which has two inverted ACS sequences spaced by 90 bp, was modified to include Widom601 and Widom603 strong positioning sequences on both ends. The nucleosome-free region is 168 base pairs long in this construct. For MCM5(AG) mutant, Widom sequences were swapped for a suicide substrate for covalent M.HpaII methyltransferase binding (equally efficient at blocking double hexamer sliding²⁸). The plasmids were synthesized by Eurofins and used for PCR amplification with the primer pairs NCP F/NCP R (for nucleosome reconstitution) or Gid70-MTRB F/Gid70-MTRB R (for methyltransferase capping).

Reconstitution of nucleosomes with yeast histones and preparation of HpaII-flanked origins was carried out as described²⁸. In brief, amplified templates were purified by anion exchange chromatography on a 1 ml RESOURCE Q column (Cytiva), followed by ethanol precipitation. DNA pellets were resuspended in TE buffer, mixed with purified yeast histone octamers and subjected to dialysis with decreasing NaCl concentration to reconstitute nucleosomes^28,45. The chromatinized construct was purified by size exclusion chromatography using a Superose 6 Increase 3.2/300 column (Cytiva). For the methyltransferase construct, DNA was incubated with M.HpaII in 1:6 molar ratio at 30 °C overnight in buffer 1 (50 mM potassium acetate, 25 mM Tris pH 7.5, 10 mM magnesium acetate, 1 mg ml⁻¹ bovine serum albumin (BSA), 150 μM S-adenosyl-methionine (NEB)). The M.HpaII–70bp–M.HpaII construct was isolated by anion exchange chromatography using a 1 ml RESOURCE Q column (Cytiva).

NCP F, 5′-(Des)-CGATAGAACTCGGGCCGCCCTGGAGAATCGCGGTGCCG-3′; NCP R, 5′-CCTGCACCCCAGGGACTTGAAGTAATAAGGAC-3′; Gid70-MTRB F, atatatCC*GGcctgtATCTCGATTTTTTTATGTTTAGTTTCGC; Gid70-MTRB R -TGGGCGCC*GGAACTGGGTGCTGTaTTTTTATGTTTAGTTCG; (Des), Desthiobiotin TEG; C*, 5-fluoro-2′-deoxycytosine.

Human MCM loading for cryo-EM

For the hDH loading reaction, 45 nM of chromatinized DNA (nucleosome–Gid70–nucleosome) were incubated with 120 nM ORC1–5(1∆N), 120 nM ORC6, 150 nM CDC6(∆N), 150 nM CDT1(∆N) and 60 nM MCM2–7 in EM buffer (25 mM HEPES-KOH pH 7.6, 100 mM potassium glutamate,10 mM magnesium acetate, 1 mM DTT, 2 mM ATP) resulting in a final volume of 35 µl. Incubation was carried out for 30 min at 37 °C and 1,250 rpm constant mixing.

The MCM recruitment reaction was established by substituting ATP with ATPγS and a chromatinized origin concentration of 70 nM.

Negative-stain electron microscopy reactions and imaging

The MCM loading reaction for the negative-stain experiment shown in Extended Data Fig. 1a was performed similarly as described above, but using 20 nM ORC1–5(1∆N), 20 nM ORC6, 20 nM CDC6(∆N), 40 nM MCM2–7, 40 nM CDT1(∆N), 7.5 nM of chromatinized DNA (nucleosome–Gid90–nucleosome) in a total volume of 20 µl. Dropout experiments shown in Supplementary Fig. 1 were carried out with the same sample concentrations as described for cryo-EM, but using a final volume of 20 µl per reaction, and omitting one factor at a time. Reactions were diluted twofold (Extended Data Fig. 1a) and 4-fold (Supplementary Fig. 1) in EM buffer where nucleotide was omitted. 300-mesh copper grids coated with a layer of continuous carbon (EM Resolutions, C300Cu100) were glow-discharged at 25 mA for 1 min using a GloQube Plus Glow Discharge System (Quorum), before applying 4 µl of the sample for 2 min. Grids were stained with two successive applications of 4 µl of 2% (w/v) uranyl acetate solution. Excess stain was blotted after 40 s using filter paper. Micrographs were collected using a FEI Tecnai G2 Spirit transmission electron microscope operated at 120 keV, equipped with a 2 K x 2 K GATAN UltraScan 1000 CCD camera. Data collection was carried out at a nominal magnification of 30,000×, yielding a pixel size of 3.45 Å at the specimen level, and a defocus range of −0.6 to −1.4 µm.

Further electron microscopy investigation of MCM loading with and without ORC6 was performed as follows. Reactions were set up by mixing 45 nM of a M.HpaII-Gid70-M.HpaII DNA template with 90 nM ORC1–5(1∆N), 150 nM CDC6(∆N), 150 nM CDT1(∆N), 60 nM MCM2–7 in EM buffer containing either 2 mM ATP or 2 mM ATPγS, with or without 150 nM ORC6. Samples were incubated at 37 °C for 30 min under agitation, diluted 1:4 in EM buffer and immediately used for negative staining as described above. Grids were imaged on a FEI Tecnai G2 Spirit microscope using a RIO16 camera at a pixel size of 3.1 Å per pixel. Particles were picked using crYOLO⁴⁶ and extracted with a box size of 144 pixels in Relion 4⁴⁷. After 2 rounds of 2D classification, well-averaging particles were counted, class populations were visualized in a 10 × 10 dot plot in Prism10.

Negative-stain image processing

Negative-stain images were processed using Relion 3.1⁴⁷. Particles were picked using Topaz v0.2.5⁴⁸. Contrast transfer function (CTF) parameters were estimated using Gctf v1.06⁴⁹. Extracted particles were then subjected to reference-free 2D classification.

Cryo-EM reactions and imaging with ORC6

UltrAuFoil R1.2/1.3 300-mesh grids (Quantifoil) were glow-discharged at 40 mA for 5 min using a GloQube Plus Glow Discharge System (Quorum), before applying 3 µl graphene oxide dispersion (10 ml graphene oxide flake dispersion (Sigma) diluted in 80 ml water; aggregates removed by centrifugation at 500g for 1 min). Incubation was carried out for three minutes, followed by blotting of excess liquid and three successive washes with 20 µl droplets of water. After 1–2 h drying at room temperature, 4 µl of the undiluted (ATP) reaction or the 3:1 diluted (ATPγS) reaction were applied to grids for 60 s at room temperature and 90% humidity in a Vitrobot Mark IV (Thermo Fisher). Grids were double-side plotted with force 0 for 5 s and immediately plunge frozen in liquid ethane. Micrographs were collected in counting mode using a pixel size of 1.08 Å on a Titan Krios transmission electron microscope with a K2 Summit direct electron detector and BioQuantum energy filter. A total electron dose of 49.28 e⁻ Å⁻² was used over 32 dose-fractioned movie frames and a total exposure time of 9.4 s. The defocus ranged from −1.0 to −2.5 µm. 3,589 movies were collected for the ATP reaction and a total of 31,569 movies for the ATPγS reaction.

Cryo-EM reactions and imaging without ORC6

Gid70 DNA template (45 nM), carrying a TwinStrep-tagged M.HpaII roadblock at each end, were mixed with 90 nM ORC1–5(1∆N), 150 nM CDC6(∆N), 150 nM CDT1(∆N) and 60 nM MCM2–7 in EM buffer to a total volume of 40 µl, and incubated for 30 min under agitation at 37 °C. Four microlitres of 50% diluted loading reaction were applied onto UltrAuFoil R1.2/1.3 300-mesh grids, which were coated with a graphene oxide layer as described above. After one minute of on-grid incubation in a Vitrobot Mark IV, the grids were blotted from both sides with Whatman Filter paper (blotting strength 0, blot time 3.5 s) and plunged into liquified ethane. Grids were clipped and stored in liquid nitrogen prior to data collection.

hDH loading reactions without CDC6 imaged

A hDH loading reaction was prepared by co-incubating 45 nM of a M.HpaII-Gid70-M.HpaII DNA with 90 nM ORC1–5(1∆N), 150 nM CDT1(∆N) and 60 nM MCM2–7 in assay buffer containing 60 mM sodium acetate (supplemented with 2 mM ATP) for 30 min at 37 °C under agitation. The sample was diluted 1:4 and subsequently negative-stained as described above. One hundred and fourteen micrographs were acquired using a RIO16 detector on a FEI TECNAI G2 Spirit Microscope at 3.1 Å per pixel. 69,785 particles were extracted with a 144-pixel box after crYOLO⁴⁶ picking and submitted to multiple rounds of 2D classification to identify MCM hDH particles in Relion 4⁴⁷.

Cryo-image processing of hDH loading

Image processing was performed using Relion 4.0b-GPU and cryoSPARC v3.3.2⁵⁰ at different stages of the processing pipeline as indicated in Extended Data Fig. 2. Beam-induced motion was accounted for by the Relion implementation with 5 × 5 patches and CTF parameters were estimated using CTFFIND v4.1.10⁵¹. Particle picking was carried out using Topaz v0.2.4⁴⁸, followed by particle extraction with a 440-pixel box and rescaling to 110 pixels in cryoSPARC. Three thousand, five hundred and eighty-nine micrographs with 970,326 particles were selected based on the CTF fit resolution of 2.57–4.50 Å, CTF fit cross-correlation of 0.07–0.27 and median pick score of 20.18–43.55. Three rounds of reference-free 2D classification, ab initio reconstruction and heterogeneous refinement were conducted to identify 49,485 hDH particles. After particle extraction without downscaling and using a 400-pixel box, the particle stack was further cleaned up by 2D classification, which yielded 19,049 particles used to compute high-resolution hDH 2D class averages. Ab initio reconstruction, followed by homogeneous, non-uniform and local refinement with C₂ symmetry resulted in a map with 3.1 Å resolution. The particle stack was re-extracted, re-grouped, and cleaned from duplicate particles using Relion. Fifteen thousand, eight hundred and seventy-four particles of the highest quality were isolated by 3D classification without alignment using a 320 Å mask and a regularization parameter T of 4. Particles were polished⁴⁷ subjected to 2D classification without alignment and 3D refinement imposing C₂ symmetry. CTF parameters were optimized⁴⁷ in three rounds (first, per-particle defocus, per-micrograph astigmatism; second, per-particle defocus, per-particle astigmatism, beamtilt; third, per-particle defocus, per-particle astigmatism, beamtilt, trefoil, 4th order aberrations) to yield a 3.3 Å resolution hDH structure. Homogeneous and non-uniform refinement with C₂ symmetry in cryoSPARC resulted in the final map at 3.1 Å resolution.

To determine the structure of hSH from the same dataset, a new Topaz model was trained. One hundred and seventy-eight thousand, eight hundred and fifty three particles were extracted with a 440-pixel box, downscaled to 110 pixels. Smaller particles and contaminations were removed by 2D classification in cryoSPARC. Initial volumes were obtained by ab initio reconstruction with six classes. Thirty-seven thousand, three hundred and ninety-six hSH particles were then isolated in two rounds of heterogeneous refinement (Extended Data Fig. 7a) and re-extracted with a 400-pixel box. Homogeneous, non-uniform and local refinement with a mask encompassing the entire hSH yielded a 3.4 Å resolution structure. The same particles were re-extracted and 3D refined in Relion. CTF refinement (per-particle defocus, per-micrograph astigmatism), Bayesian polishing and one additional round of CTF refinement (per-particle defocus, per-particle astigmatism, beamtilt) were carried out.  Twenty-five thousand and sixty-nine high-resolution hSH particles were isolated by 3D classification. Homogeneous, non-uniform and local refinement with a mask around the hSH yielded 3.2 Å resolution in cryoSPARC.

Cryo-image processing of MCM recruitment

Movies were corrected for beam-induced motion using the Relion implementation with 5 × 5 patches in Relion 4.0b-GPU⁵² and CTF parameters were estimated using CTFFIND v4.1.13⁵¹. A Topaz model⁴⁸ was trained and 1,334,277 particles were picked from 31,569 micrographs. Particles were extracted with a 416-pixel box, rescaled to 104 pixels. Six hundred and twenty-nine thousand, two hundred and forty-one MCM-containing particles (that is, hOCCM, hSH and hMO*) were isolated using reference-free 2D classification. Initial volumes were generated using ab initio reconstruction in cryoSPARC. Low-pass filtered volumes were used for multi-reference 3D classification in Relion (Extended Data Fig. 5). Particles contributing to the different complexes (114,995 hSH, 170,792 hOCCM and 203,088 hMO* particles) were re-extracted using a 400-pixel box without rescaling. Homogeneous, non-uniform and local refinements with masks around the entire respective complex resulted in 3.6 Å resolution structure of the hSH, 4.0 Å hOCCM and 3.7 Å hMO* in cryoSPARC. Bayesian polishing was carried out for each particle stack individually in Relion⁴⁷.

The hOCCM was then subjected to three rounds of CTF refinement (first, per-particle defocus, per-micrograph astigmatism; second, anisotropic magnification; third, per-particle defocus, per-particle astigmatism, beamtilt) followed by another round of Bayesian polishing in Relion. A mask around CDC6 and portions of ORC1–5 was generated to carry out focused 3D classification without alignment. This yielded 100,567 particles with good density for ORC–CDC6. 3D refinement in Relion was followed by homogeneous refinement in cryoSPARC. Local refinement with a mask around ORC–CDC6 was used to improve alignment on this part of the complex. Thirty-four thousand, one hundred and sixteen hOCCM particles with well-resolved CDC6 density and 49,771 hOC₁M particles that lacked CDC6 were isolated by 3D classification without alignment in cryoSPARC. Both hOCCM and hOC₁M structures were locally refined to 3.8 and 4.1 Å, respectively, using a mask encompassing the entire complex.

Complexes containing single-loaded hexamers, hSH and hMO* particles were initially processed together (Extended Data Fig. 5). Three-dimensional refinement in Relion using a mask around the MCM yielded a 3.8 Å hSH structure. Two rounds of CTF refinement (first, per-particle defocus, per-micrograph astigmatism; second, anisotropic magnification) and another round of Bayesian polishing was carried out. One hundred and thirty-five thousand, seven hundred and forty-two hSH particles and 182,341 hMO* particles were subsequently separated by multi-reference 3D classification using the initial maps of hSH and hMO*, low-pass filtered to 30 Å. The hSH was refined in cryoSPARC (homogeneous, non-uniform, followed by local refinement) to 3.4 Å. Homogeneous and non-uniform refinement of the hMO* resulted in a consensus map solved to 3.6 Å resolution. The MCM portion of the complex was locally refined to 3.5 Å, while the ORC was refined to 4.0 Å.

Cryo-EM of hDH loading without ORC6

A total of 5,158 movies were collected in counting mode on a 200 kV Talos transmission electron microscope using a pixel size of 1.61 Å per pixel with a total dose of 48 e⁻ Å⁻² and a defocus range of −2 to −3.5 μM (step size 0.25 μM). Movies were corrected for beam-induced motion with the Relion implementation with 5 × 5 patches in Relion 5.0 and CTF parameters estimated with CTFFIND v4.1.13⁵¹. Subsequent micrograph curation reduced the number of micrographs to 2,168. Four hundred and fifty-six thousand, six hundred and eighty-six particles were picked using template matching in cryoSPARC v4.4.1⁵⁰ and extracted at a pixel size of 3.22 Å per pixel (2× binned) and subjected to two rounds of 2D classification. A total of 77,002 particles were used to generate an ab initio reconstruction in C₁. Particles were re-extracted at full resolution, refined and then symmetry expanded in C₂. Three-dimensional classification was performed using 4 classes. The class with the highest quality DNA density containing 74,880 particles was subject to a local refinement resulting in a 4.1 Å structure according to gold-standard FSC at 0.143 criterion.

MCM5(AG) DH loading for cryo-EM

A 50-µl MCM DH loading reaction was assembled using the protocol described in ‘Human MCM loading for cryo-EM’. In short, 45 nM M.HpaII–Gid70–M.HpaII DNA capped at each end with TwinStrep-tagged M.HpaII, was co-incubated with 90 nM ORC1–5(1∆N), 150 nM CDC6(∆N), 150 nM CDT1(∆N) and 60 nM MCM5(AG)–MCM2–7 in EM buffer at 37 °C under agitation. After 30 min, the loading reaction was diluted either by 40% or 75% with EM buffer and immediately used for cryo-EM.

UltrAuFoil R1.2/1.3 300-mesh grids were coated with a graphene oxide support as described above. Four microlitres of diluted loading reaction were applied onto each grid in a Vitrobot Mark IV set to 22 °C and 90% humidity. After one minute on-grid incubation, grids were double-side blotted for 4.5 s with blot force 0 and plunged into liquid ethane. Grids were subsequently clipped and stored in liquid nitrogen until data collection.

Cryo-EM of MCM5(AG) DH

In total, 13,203 movies were acquired from two grids at 92,000× magnification (1.58 Å per pixel) on a Glacios microscope equipped with a Falcon 3 direct electron detector operated in linear mode. A total dose of 50 e⁻ Å⁻² (exposure time 1.12 s) and a defocus range of −1 to −2.5 µm (0.3 µm step size) were applied. Movie frames were aligned using Relion 4.0 (ref. ⁵²) and CTF was estimated using GCTF v1.06 (ref. ⁴⁹). Two million, eight hundred and seventy-five thousand, eight hundred and twenty-five particles were picked using a pre-trained Topaz network⁴⁸, extracted with a box size of 64 pixels (4× binned to 6.32 Å per pixel) and imported in cryoSPARC v4.0⁵⁰ for 2 rounds of 2D classification. After 2D cleaning, two ab initio models were generated with C₁ symmetry using a subset of 50,000 particles. All 379,664 particles were subjected to one round of heterogenous refinement against the two ab initio reconstructions, followed by homogenous refinement in C₁. Particles were re-extracted in Relion with a box size of 300 pixels (unbinned at 1.58 Å per pixel) and 2D-classified in cryoSPARC. The resulting 322,548 particles were used to generate an unbinned ab initio reconstruction in C₁, yielding a 5.76 Å map after non-uniform refinement in C₁. Three-dimensional classification in Relion without alignment using a regularization parameter T of 200 distributed the particles into four classes with roughly equal populations, differing by presence of both, either or no MCM6 WH domain. The best-resolved class (showing density for both MCM6 WH domains) was non-uniform and locally refined in cryoSPARC with C₂ symmetry, yielding a final resolution of 5.6 Å according to gold-standard FSC and the 0.143 criterion. Local resolution estimation was carried out in cryoSPARC. Refinement statistics are reported in Extended Data Table 1.

hDH model building

AlphaFold-Multimer⁵³ was used to generate models of the ATPase tier (including the WH domains) of the hexameric human MCM2–7 assembly as well as the amino-terminal tier. Each model was rigid-body docked into one hexamer of the 3.1-Å resolution hDH map using UCSF ChimeraX v1.6.1 (ref. ⁵⁴). Each chain was refined in Coot v0.9.8.1 EL⁵⁵ and sections that could not be confidently built were deleted⁵⁴. The models of the ATPase and amino-terminal tiers were then combined. ATP, ADP, magnesium and zinc ions were added in the pertinent sites. Idealized B-form DNA was first docked into the density of the double helix and then manually modified to account for the stretch of underwound and melted DNA within the N-terminal dimerization interface. The model consisting of the MCM2–7 hexamer, ligands and DNA was adjusted using ISOLDE 1.6.0 (ref. ⁵⁶), imposing ligand, secondary and base pairing restraints. This was followed by an iterative process of real space refinement with restraints on geometry, secondary structure, metal coordination and nucleic acid planarity in Phenix v1.21 (ref. ⁵⁷) and manual adjustments in Coot and ISOLDE. To generate the hDH, a copy of the refined MCM hexamer was generated and rigid-body docked into the second hexamer of the hDH map. Clashes at the interface of the two hexamers were addressed using ISOLDE. The resulting model was subjected to real space refinement in Phenix. Refinement statistics are reported in Extended Data Table 1.

hSH model building

The atomic model of the hDH was rigid-body docked into the 3.2 Å cryo-EM map of the hSH using UCSF ChimeraX v1.6.1 (ref. ⁵⁸). The second hexamer was deleted, nucleotides were inspected and the DNA was replaced by an idealized B-form duplex DNA using Coot v0.9.8.1 EL⁵⁹. The model was refined with ISOLDE 1.6.0 (ref. ⁵⁶) and Phenix real space refinement with restraints on geometry, secondary structure, metal coordination and nucleic acid planarity⁵⁷. Refinement statistics are reported in Extended Data Table 1.

hMO* model building

To generate the model of the hMO* complex, the atomic structure of the hSH described above was docked into the map of the globally refined hMO* map using UCSF ChimeraX v1.6.1 (ref. ⁵⁴). The atomic model of H. sapiens ORC6 was retrieved from the AlphaFold Protein Structure Database (accession code AF-Q9Y5N6-F1). To guide the positioning of ORC6, the yMO (PDB entry 6RQC)²⁸ was aligned with the human MCM in the map. Based on the alignment with the S. cerevisiae Orc6, the N-terminal cyclin box domain (residues 1–94) of human ORC6 was positioned next to the N-terminal domains of MCM2 and MCM6. The domain was then rigid-body docked into the density. The second cyclin box domain (residues 95–190) was docked into the adjacent density of the map that is positioned between MCM6 and ORC5. Docking solutions with the highest cross-correlation scores were chosen. The starting model of ORC1–5 was the open conformation of the human ORC1–5 complex (PDB code 7JPR)⁶⁰. The model was docked into the locally refined map of the ORC at 4.0 Å resolution. The DNA model of the yMO complex (PDB entry 6RQC)²⁸ was fit into the DNA density in the globally refined hMO* map. The DNA model was refined in Coot v0.9.8.1 EL⁵⁹ applying all-molecule self-restraints 6.0 and DNA B-form restraints, before combining it with the DNA in the hSH. ORC6 was combined with the MCM and DNA into one PDB model. The two models of ORC6–MCM and ORC1–5 were iteratively refined with ISOLDE 1.6.0 (ref. ⁵⁶) and Phenix real space refinement with restraints on geometry, secondary structure, metal coordination and nucleic acid planarity⁵⁷ against the globally refined map of hMO* and the locally refined map of the ORC, respectively. For illustration purposes, a composite map was generated using Phenix and the refined ORC6–MCM and ORC1–5 models were combined into one PDB model. Refinement statistics are reported in Extended Data Table 2.

hOCCM model building

To assemble the atomic model of hOCCM, the human ORC1–5–DNA model (PDB entry 7JPS), the ORC2 WH domain of the ORC1–5 model (PDB entry 7JPR)⁶⁰ and the prediction of human CDC6 retrieved from AlphaFold Protein Structure Database (accession code AF-Q99741-F1) were rigid-body docked into the 3.8 Å cryo-EM map using UCSF ChimeraX v1.6.1 (ref. ⁵⁴). Unstructured parts of the models, which were not visible in the hOCCM map, were deleted. The hSH model was docked into the map and each subunit was split and fitted independently into the density. The N-terminal tiers of MCM7, MCM3 and MCM5 were deleted from the atomic coordinates file as the local quality of region of the map was deemed insufficient for model building. The positioning of the ATPase tiers of the same MCM subunits was guided by the position in the hSH and no further adjustment of atomic positions was carried out due to the limited local resolution. The atomic models of the MCM3 and MCM7 WH domains were overlayed with equivalent domains in yOCCM. The atomic model of CDT1 was retrieved from the AlphaFold database (accession code AF-Q9H211-F1) and the three structured domains (residues 167–387, 418–440 and 441–546) were extracted from the model. Each domain was overlayed with yeast Cdt1 in yOCCM (PDB entry 5V8F)²⁹. All models were combined into one PDB model and adjusted using Coot v0.9.8.1 EL⁵⁹ and ISOLDE 1.6.0 (ref. ⁵⁶). Models of ATPγS and magnesium were added in the ATPase active sites of MCM2–6, MCM6–4, ORC1–4, ORC4–5 and ORC5–3. A 39-mer idealized B-form duplex DNA was generated and fit into the density by applying all-molecule self-restraints 6.0 and DNA B-form restraints in Coot. The DNA coordinates were then merged with the hOCCM model before carrying out Phenix real space refinement with restraints on geometry, secondary structure, metal coordination and nucleic acid planarity⁵⁷. Refinement statistics are reported in Extended Data Table 3.

Analysis of protein–DNA contacts

Protein–DNA contacts were analysed using the DNAproDB web-based visualization tool^61,62.

Statistics and reproducibility

Proteins were independently purified at least twice (Fig. 1a). The experiments in Figs. 1d,f,g,j, 2e,g and 3a,f, Extended Data Fig. 1a,n and Supplementary Fig. 1a–d were performed three times. The experiments in Figs. 1c,e,h,i and 3g,i,k, Extended Data Figs. 1b–i,m, 3d and 4a and Supplementary Figs. 1e–h and 2b were performed twice. The micrographs shown in Extended Data Figs. 2a and 4b are representative micrographs of cryo-EM datasets. Similar experiments have been analysed multiple times by negative-stain electron microscopy.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.