The interaction networks of the budding yeast and human DNA replication-initiation proteins

ABSTRACT

DNA replication is a stringently regulated cellular process. In proliferating cells, DNA replication-initiation proteins (RIPs) are sequentially loaded onto replication origins during the M-to-G₁ transition to form the pre-replicative complex (pre-RC), a process known as replication licensing. Subsequently, additional RIPs are recruited to form the pre-initiation complex (pre-IC). RIPs and their regulators ensure that chromosomal DNA is replicated exactly once per cell cycle. Origin recognition complex (ORC) binds to, and marks replication origins throughout the cell cycle and recruits other RIPs including Noc3p, Ipi1-3p, Cdt1p, Cdc6p and Mcm2-7p to form the pre-RC. The detailed mechanisms and regulation of the pre-RC and its exact architecture still remain unclear. In this study, pairwise protein-protein interactions among 23 budding yeast and 16 human RIPs were systematically and comprehensively examined by yeast two-hybrid analysis. This study tested 470 pairs of yeast and 196 pairs of human RIPs, from which 113 and 96 positive interactions, respectively, were identified. While many of these interactions were previously reported, some were novel, including various ORC and MCM subunit interactions, ORC self-interactions, and the interactions of IPI3 and NOC3 with several pre-RC and pre-IC proteins. Ten of the novel interactions were further confirmed by co-immunoprecipitation assays. Furthermore, we identified the conserved interaction networks between the yeast and human RIPs. This study provides a foundation and framework for further understanding the architectures, interactions and functions of the yeast and human pre-RC and pre-IC.

Introduction

The initiation of eukaryotic DNA replication occurs under the strict control of cell cycle regulatory mechanisms to ensure that genomic DNA is faithfully duplicated once per cell cycle [1–5]. One of the most critical control mechanisms of DNA replication is the sequential loading of replication-initiation proteins (RIPs) onto replication origins for pre-replication complex (pre-RC) assembly and replication licensing [1–6]. Current models suggest that origin recognition complex (ORC; composed of Orc1-6p) associates with replication origins in a sequence specific manner throughout the cell cycle, marking them as the start sites of DNA replication [6–9]. During the M-to-G₁ transition, additional RIPs, including Noc3p, Ipi3p, Cdc6p and Cdt1p, facilitate the loading of minichromosome maintenance proteins (Mcm2-7p), resulting in the formation of the pre-RC and the completion of replication licensing [6–15].

Following pre-RC formation, Cdc45p and other RIPs associate with replication origins in late G₁ phase to form the pre-initiation complex (pre-IC) [16–21]. As a part of pre-IC, the GINS complex is a core component of the CMG (Cdc45-MCM-GINS) replicative helicase complex responsible for DNA duplex unwinding in S phase [14,20]. The pre-RC and pre-IC are activated by cyclin-dependent kinases (CDKs) and Dbf4p-dependent kinase (DDK) [22–25]. The phosphorylation of pre-RC and pre-IC proteins occurs as the cell passes the restriction point (START in yeast) and at the G₁-to-S transition [24,26,27]. In late M phase, Cdc14p phosphatase dephosphorylates ORC, Cdc6p and MCM proteins to reset their competency for replication licensing for the next cell cycle [4].

In this study, we have performed yeast two-hybrid analysis to systematically examine the pairwise protein-protein interactions among the budding yeast and human RIPs, and compare the interactions between the two organisms. Through our comprehensive study and literature surveys, we also demonstrate that several protein-protein interactions in the RIP networks are highly conserved between the lower and higher eukaryotic organisms. Critically, this study has revealed the existence of several previously unknown physical interactions and suggests that they may have functional roles within the interaction networks. The interactions between ORC and MCM subunits suggest that these two complexes contact at their interface. The self-interactions of ORC subunits are also of biological significance (unpublished). The interactions of Noc3p and Ipi3 with pre-RC and pre-IC components further suggest that these proteins are involved in DNA replication as well as ribosome biogenesis [9,13,28]. In addition, the biological significance of the interactions among GINS, ORC, Cdt1 and Ipi3 proteins are worth further study.

Materials and methods

Plasmid construction and strains

Competent Escherichia coli (E. coli) DH5α cells (Clontech) were used for plasmid propagation. S. cerevisiae AH109 strain of the Matchmaker GAL4 Two-hybrid System 3 (Clontech) was used as the host strain for yeast two-hybrid analysis. The genotype of AH109 is MATa, trp1-901, leu2-3, 112, ura3–52, his3-200, gal4Δ, gal80Δ, LYS2::GAL1, UAS-GAL1TATA-HIS3, GAL2UAS-GAL2TATA-ADE2, URA3::MEL1, UAS-MEL1TATA-lacZ, MEL1.

For yeast two-hybrid analysis, total human RNA was extracted from normal human blood using TRIzol reagent (Invitrogen). First-strand cDNA was synthesized using a cDNA kit (Fermentas). Coding sequences of the human genes were amplified by PCR with KOD polymerase (Novagen) using the first-strand cDNA as templates. The cDNA fragment encoding each human gene was cloned into the multiple cloning site of pGBKT7 (the “bait” vector) and pGADT7 (the “prey” vector) for the yeast two-hybrid assays (Clontech). The ORFs of the S. cerevisiae replication-initiation proteins were generated by PCR and cloned in the pGADT7 and pGBKT7 vectors as well.

For co-immunoprecipitation assays, the ORFs of the S. cerevisiae and human replication-initiation proteins were generated by PCR. The yCdt1 coding sequence was ligated into the pESC-HIS vector (containing a sequence for the FLAG epitope, downstream of the GAL10 promoter; Agilent Technologies) to generate the pESC-Cdt1-FLAG plasmid (HIS) [29]. The second multiple cloning site (MCS) in the pESC-Cdt1-FLAG construct (containing a sequence for c-myc epitope, downstream of the GAL1 promoter) was used to generate the pESC-His3-Psf1-Myc-Cdt1-FLAG, pESC-His3-Psf3-Myc-Cdt1-FLAG and pESC-His3-Sld5-Myc-Cdt1-FLAG plasmids. W303-1A (wild-type) yeast cells were transformed with the respective plasmids (Supplementary Figure S12F, G and H), while the W303-1A GALS-3HA-MCM6::natNT MCM4-GFP::KanMX strain (YL1223) expressing HA-Mcm6 was used to examine co-immunoprecipitation with endogenous ORC proteins (Supplementary Figure S12B, C and D). The un-transformed W303-1A strain was used as a control.

To generate HA-hORC5 plasmid, the hORC5 coding sequence was ligated into the pCMV-HA mammilian expression vector (Clontech). To generate FLAG-hORC5 plasmid, a FLAG-tag was inserted at the Hind III restriction site of the pcDNA3.0 mammilian expression vector (Invitrogen), and resulting vector was inserted with the hORC5 coding sequence. To generate HA-hNOC3 plasmid, the HA affinity tag was inserted between the start codon and the second codon of the hNOC3 coding sequence in the mammalian expression vector pcDNATM3.1/Zeo (+) (Life Technologies). HEK-293T cells were transfected with the tagged-protein expressing plasmid(s) or empty vector as stated (Supplementary Figure S12A and E).

Yeast two-hybrid assays

The yeast two-hybrid analysis was performed as described [30,31]. Pairs of the bait (pGBKT7) and prey (pGADT7) plasmids were used to co-transform AH109, the host strain of the Matchmaker GAL4 Two-hybrid System 3 (Clontech), and the cells were plated on SCM-Trp-Leu (Synthetic Complete Medium lacking tryptophan and leucine; SCM-2) plates and allowed to grow at 30ºC for 3 d. Transformants were then streaked and/or patched on SCM-2, SCM-Trp-Leu-His (SCM-3), and SCM-Trp-Leu-His-Ade (SCM-4) plates. Negative controls included combinations of each prey plasmid with empty BD vector and of each bait plasmid with the empty AD vector, while the AD-T and BD-p53 combination were used as the positive control. The YL187 yeast strain was used as a mating partner for the AH109 strain to identify potential false positive results from autonomous transcriptional activation of the reporter gene by a single fusion protein in the absence of an interacting partner.

Immunoblotting analysis

Expression of the fusion-protein in the transformed AH109 strain was tested by immunoblotting. Protein samples were mixed with equal volumes of 2X Laemmli’s buffer, boiled at 100ºC for 3-5 min, and then loaded onto 10–12.5% SDS-PAGE gels. The proteins were then transferred onto nitrocellulose membrane. The membrane was blocked for 30 min in TBST containing 5% dry milk. The HA- and Myc-tagged fusion proteins were probed with anti-c-Myc (Roche, 9E10, 1:5,000) and anti-HA (Roche, 12CA5, 1:10,000) antibodies which were then recognized with the secondary antibodies (HRP-conjugated anti-mouse, 1:20,000; Thermo Fisher Scientific Inc.). The signals were visualized using the SuperSignal™ reagent (Pierce).

Human and yeast protein co-immunoprecipitation assay

Co-IP assays with yeast cell extracts were performed as previously described [9,29]. Co-IP assays with human HEK-293T cell extracts were performed as previously described [32]. Yeast and human cell extracts were DNase I digested to avoid DNA-mediated protein interactions prior to antibody incubation (see Supplementary Figure S12 legend for further details). Co-IP was performed with the following antibodies; anti-c-Myc (Roche, 9E10), anti-HA (Roche, 12CA5), anti-FLAG M2 (Sigma), anti-yOrc2 (SB67), anti-yOrc3 (SB3) and anti-yOrc5 (SB5) (SB-No. antibodies were gifts from Dr. Bruce Stillman, CSHL). Immunoblotting was performed as previously described [9,29], with the following antibodies; mouse anti-HA (Roche, 12CA5, 1:10,000), mouse anti-c-Myc (Roche, 9E10, 1:5,000) mouse anti-FLAG M2 (Sigma, 1:5000), mouse anti-yOrc2 (SB67, 1:1,000), mouse anti-yOrc3 (SB3, 1:20,000), mouse anti-yOrc5 (SB5, 1:1,000), mouse anti-BM28 (hMCM2, BD Bioscience, 1:1000), goat Anti-hMCM3 (Santa Cruz, 1:500) and mouse anti-hMCM7 (Santa Cruz, 1:500) antibodies, which were then recognized with the secondary antibody (HRP-conjugated anti-mouse LC, 1:10,000; Thermo Fisher Scientific Inc).

Results

Examining protein expression and transcriptional activation of the reporter genes by individual yeast two-hybrid fusion proteins

Target proteins to be analyzed were expressed in the host yeast cells as both the DNA binding domain (BD) and activation domain (AD) fusion proteins. The protein expression of all fusion proteins in the subsequent yeast two-hybrid experiments were examined by immunoblotting. The results confirmed the expression of all fusion proteins being tested with a few exceptions as noted (Supplementary Figure S1). We also performed tests in the yeast two-hybrid system to exclude the possibility of transcriptional activation of the reporter genes by each single fusion protein expressed from the plasmids. The results confirmed the absence of reporter gene activation by individual fusion proteins with a few exceptions as noted (Supplementary Figure S2). The fusion proteins that had expression and did not activate reporter gene expression without an interaction partner were used in the subsequent yeast two-hybrid analysis.

Pairwise interactions of the yeast and human ORC subunits

As summarized by diagrams in Figures 1-10, positive interactions were detected as being in one direction (e.g. BD-fusion protein 1 interacted with AD-fusion protein 2, but BD-fusion protein 2 did not interact with AD-fusion protein 1; represented by single-directional arrows), or two directions (having reciprocal interactions between the BD-fusion protein and AD-fusion protein; represented by double arrows). In general, one-directional interactions were relatively weak interactions, while two-directional interactions were relatively strong [9,30].

Figure 1.

Budding yeast and human ORC inter-subunit interactions examined by yeast two-hybrid analysis. Yeast host cells expressing yeast or human AD-fusion proteins and BD-fusion proteins were patched or streaked onto SCM-Trp-Leu (SCM-2), SCM-Trp-Leu-His (SCM-3) and SCM-Trp-Leu-His-Ade (SCM-4) plates and then incubated at 30°C for three days to test cell growth. Cells expressing AD and individual BD-fusion proteins and cells expressing BD and individual AD-fusion proteins were used as the negative controls, while cells co-expressing AD-T (T, SV40 large T-antigen) and BD-p53 were used as the positive control. (a, b) Summary of the budding yeast protein-protein interactions (see data in Supplementary Figure S3 and Table S1). O1-O6, ORC1-ORC6. One-directional arrows represent protein-protein interactions that were observed in a single direction (BD-fusion protein 1 interacted with AD-fusion protein 2, but BD-fusion protein 2 did not interact with AD-fusion protein 1). Double arrows represent interactions detected in both directions (BD-fusion protein 1 interacted with AD-fusion protein 2, and BD-fusion protein 2 also interacted with AD-fusion protein 1). The thickness of the arrows indicates the observed relative strength of the interactions, with thin arrows representing the weakest interaction (+) and the thickest representing the strongest (+++). Self-interactions of te same protein were indicated by dashed circles. (c) Human protein-protein interactions indicated by cell growth on SCM-2, SCM-3 and SCM-4 plates (see Supplementary Figure S4 and S5 for additional data). Circles indicate the positive control. (d-f) Summary of the human protein-protein interactions.

Figure 10.

A summary diagram of human pre-RC protein-protein interactions. Host yeast cells expressing human AD-fusion proteins and BD-fusion proteins were streaked onto SCM-2, SCM-3 and SCM-4 plates and incubated at 30°C for three days to examine the relative strength of the interactions indicated by cell growth relative to the positive and negative controls as indicated (see Supplementary Figures S4-S11 for data). Positive interactions are depicted in the diagram. Refer to Figure 1 for the meaning of the symbols.

To test the pairwise physical interactions among yeast and human ORC subunits, we performed yeast two-hybrid analysis with ORC subunits expressed in the host yeast cells as both the BD) and activation domain (AD) fusion proteins.

We were able to observe many positive physical interactions for the yeast and human ORC subunits (Figure 1(a,b), Supplementary Figure S3, Tables 1 and 2, and Supplementary Table S1), some of which were previously reported (summarized and referenced in Table 1) and some of which were previously unknown (summarized and referenced in Table 2). These novel interactions will be described in the forthcoming sections. Importantly, many of the positive interactions observed in this study were consistent with previously reported data from co-IP with yeast and human cell extracts, and/or from other physical and genetic experiments, as summarized and referenced in Tables 1 and 2, and described in the proceeding sections. Some of these novel interactions were further confirmed by co-IP in this study (Table 3).

Table 1.

Physical interactions identified by yeast two-hybrid analysis in this study that have previously been reported.

Interaction [*, bi-directional interaction; h, human; y, yeast (S. cerevisiae)]	Literature reference and Method(s) (a, b, c, d, e, f, g, h, i)	Interaction (*, bi-directional interaction; h, human; y, S. cerevisiae)	Literature reference and Method(s) (a, b, c, d, e, f, g, h, i)
*hORC1 + hORC2	[34 d, 35 a, 40 b]	*yOrc2 + yMcm5	[44 a, 51 h]
*hORC2 + hORC3	[34 d, 40 b, 46 c, 47 b]	BD-yOrc5 + AD-yMcm2	[14 c, 55 i, 60 a, 67 b]
*hORC2 + hORC4	[34 d, 35 a, 40 b, 47 b, 48 g]	BD-yOrc6 + AD-yMcm2	[44 a, 60 a]
*hORC2 + hORC5	[34 d, 35 a, 40 b, 48 g, 50 c]	BD-yMcm3 + AD-yOrc2	[44 a]
BD-hORC2 + AD-hORC6	[40 b, 47 b, 52 d]	BD-yMcm7 + AD-yOrc2	[41 j, 44 a, 51 h]
BD-hORC4 + AD-hORC3	[35 a, 40 b, 47 b, 48 g]	*hCDC6 + hORC1	[35 a]
BD-hORC4 + AD-hORC5	[35 a, 48 g]	*hCDC6 + hORC2	[35 a]
BD-hORC4 + AD-hORC6	[35 a, 47 b]	*hCDC6 + hMCM2	[35 a]
BD-hORC5 + AD-hORC1	[33 d, 35 a]	*hCDC6 + hMCM3	[35 a, 69 b]
BD-hORC5 + AD-hORC3	[40 b, 48 g, 50 c]	*hCDC6 + hMCM7	[35 a, 69 b]
*yOrc1 + yOrc1	[36 d, 37 d]	BD-yCdc6 + AD-yOrc3	[42 a]
*yOrc1 + yOrc2	[9 e, 36 b g, 37 d, 38 c, 39 g, 41 j]	BD-yMcm3 + AD-yCdc6	[41 j, 44 a]
*yOrc2 + yOrc2	[37 d]	*hCDT1 + hORC1	[35 a]
*yOrc2 + yOrc3	[9 e, 37 d, 42 a, 43 e, 44 a]	*hCDT1 + hORC2	[35 a]
*yOrc2 + yOrc5	[9 e, 37 d, 42 a, 43 e, 44 a, 51 h]	BD-yCdt1 + AD-yOrc2	[84 e]
*yOrc2 + yOrc6	[9 e, 37 d, 43 e, 44 a, 53 a]	BD-yCdt1 + AD-yOrc3	[41 j, 44 a]
BD-yOrc3 + AD-yOrc1	[9 e, 36 b d, 37 d, 38 c, 39 g]	BD-yCdt1 + AD-yMcm5	[83 a]
*yOrc4 + yOrc5	[9 e, 37 d, 42 a, 43 e, 44 a]	*hIPI3 + hORC1	[28 e]
BD-yOrc4 + AD-yOrc1	[9 e, 36 d, 37 d, 38 c, 39 g, 42 a, 43 e, 44 a]	*hIPI3 + hORC2	[28 e]
BD-yOrc4 + AD-yOrc2	[37 d, 44 a, 49 b]	*hIPI3 + hORC3	[28 e]
*yOrc5 + yOrc5	[37 d, 43 e]	*hIPI3 + hMCM2	[28 e]
BD-yOrc5 + AD-yOrc3	[37 d, 42 a, 43 e, 44 a]	*hIPI3 + hMCM3	[28 e]
*yOrc6 + yOrc6	[37 d, 43 e]	*hIPI3 + hMCM4	[28 e]
BD-yOrc6 + AD-yOrc3	[9 e, 37 d, 43 e]	*hIPI3 + hMCM7	[28 e, b]
*yMcm2 + yMcm2	[9 e, 78 b]	BD-hORC4 + AD-hIPI3	[28 e]
*yMcm2 + yMcm4	[9 e, 29 b, 39 g, 54 b, 55 i, 56 d, 57 b, 58 c]	BD-hIPI3 + AD-hORC5	[28 e]
*yMcm2 + yMcm6	[9 e, 29 b, 44 a, 54 b, 55 i, 58 c, 59 a, 60 a]	*yNoc3 + yNoc3	[71 d]
BD-yMcm2 + AD-yMcm7	[9 e, 55 i, 60 a]	BD-yCdt1 + AD-yNoc3	[9 e]
*yMcm3 + yMcm5	[9 e, 44 a, 54 b, 55 i, 59 a, 61 e, 62 d]	BD-yIpi3 + AD-yCdt1	[9 e]
*yMcm3 + yMcm7	[9 e, 44 a, 54 b, 55 i, 59 a, 61 e]	*ySld5 + ySld5	[64 b, 56 d]
BD-yMcm3 + AD-yMcm4	[9 e, 29 b, 39 g, 54 b, 55 i, 56 d, 57 b, 63 b, 64 b]	*ySld5 + yPsf1	[54 b, 56 d, 57 b, 63 b, 64 b, 72 e]
*yMcm4 + yMcm6	[9 e, 29 b, 39 g, 44 a, 54 b, 55 i, 56 d, 57 b, 59 a, 64 b]	*ySld5 + yPsf2	[56 d, 63 b, 64 b, 71 d, 72 e, 88 e]
*yMcm4 + yMcm7	[9 e, 39 g, 54 b, 55 i, 56 d, 57 b, 59 a, 63 b]	BD-ySld5 + AD-yPsf3	[54 b, 56 d, 63 b, 64 b, 71 d, 72 e,]
*yMcm5 + yMcm5	[9 e, 62 d]	*yPsf1 + yPsf2	[56 d, 71 d, 72 e]
BD-yMcm5 + AD-yMcm4	[9 e, 39 g, 54 b, 55 i, 56 d, 57 b, 62 d, 63 b, 64 b]	BD-yPsf1 + AD-yPsf3	[56 d, 71 d, 72 e]
BD-yMcm5 + AD-yMcm6	[9 e, 44 a, 55 i, 62 d]	BD-yPsf2 + AD-yPsf3	[56 d, 71 d, 72 e]
BD-yMcm5 + AD-yMcm7	[9 e, 55 I, 61 e, 62 d]	*yPol30 + yPol30	[76 i]
*yMcm6 + yMcm6	[9 e]	BD-yMcm10 + AD-yPol30	[74 e]
BD-yMcm6 + AD-yMcm7	[9 e, 44 a, 55 i, 59 a]	BD-yCdc45 + AD-yOrc2	[21 h]
*yMcm7 + ycmM7	[9 e, 61 e]	*yMcm10 + yMcm2	[57 b]
*hORC2 + hMCM2	[35 a]	BD-yMcm10 + AD-yMcm3	[75 e]
*hORC2 + hMCM5	[46 c,]	BD-yMcm10 + AD-yMcm7	[75 e]
*hORC3 + hMCM7	[35 a]	BD-yMcm4 + AD-yMcm10	[57 b]
*hORC4 + hMCM3	[35 a]	BD-yMcm6 + AD-yMcm10	[57 b]
*hORC5 + hMCM2	[35 a]	BD-yMcm6 + AD-ySld5	[54 b, 56 d, 63 b, 73 a]
*hORC5 + hMCM3	[35 a]	BD-yMcm6 + AD-yPsf1	[73 a]
*hORC6 + hMCM4	[35 a]	BD-yMcm6 + AD-yPsf2	[73 a]
BD-hORC6 + AD-hMCM2	[35 a]	BD-yMcm6 + AD-yPsf3	[73 a]
*yOrc2 + yMcm2	[60 a]

Table 1 summarizes the physical interactions we identified in this study that have been reported by various methods in previous publications. Some of the interactions identified by yeast two-hybrid in this study have also been identified previously using the same method by either us or other groups. Interactions discovered by genetic screenings have also been used as a reference for some previously reported physical interactions. (a, Reconstituted Complex; b, Affinity Capture-Western; c, Co-fractionation; d, Affinity Capture-MS; e, Yeast two-hybrid; f, Co-localization; g, Co-purification; h, Synthetic Lethality; I, Dosage Growth Defect; j, Negative Genetic) (a-g, Physical interactions; h-j, Genetic interactions).

Table 2.

Novel physical interactions identified by yeast two-hybrid analysis in this study.

Novel physical interaction (*, bi-directional interaction; h, human; y, S. cerevisiae)	Physical interaction Absent/present in human (h) or yeast (y) homologs. [Literature reference and Method(s) (a, b, c, d, e, f, g, h, i)]	Novel physical interaction (*,bi-directional interaction; h, human; y = S. cerevisiae)	Physical interaction Absent/present in human (h) or yeast (y) proteins. [Literature reference and Method(s) (a, b, c, d, e, f, g, h, i)]
*hORC1 + hORC4	[This work, 9 e, 36 b, 37 d, 38 c, 39 g, 42 a, 43 e, 44 a, 71 d]	*hNOC3 + hMCM2	[13 b]
*hORC2 + hORC2	[This work, 37 d]	*hNOC3 + hMCM3	Absent
*hORC4 + hORC4	[37 d]	*hNOC3 + hMCM7	Absent
*hORC5 + hORC5	[This work, 37 d, 43 e]	*hNOC3 + hIPI3	Absent
*hORC1 + hMCM2	[60 a, 83 a]	*hIPI3 + hCDC6	Absent
*hORC1 + hMCM4	[83 a]	*hIPI3 + hCDT1	[This work, 9 e]
*hORC1 + hMCM6	[83 a]	*hNOC3 + hNOC3	[This work, 71 d]
*hORC1 + hMCM7	[44 a, 83 a]	*yPsf1 + yPsf1	Absent
*hORC2 + hMCM4	[51 h]	*ySld5 + yOrc2	Absent [41 j]
*hORC2 + hMCM6	[This work, 41 j]	*yPsf1 + yOrc2	Absent [41 j]
*hORC2 + hMCM7	[This work, 41 j, 44 a, 51 h]	*ySld5 + yOrc5	Absent
*hORC3 + hMCM4	Absent [41 j]	*yPsf2 + yOrc2	Absent
BD-hORC3 + AD-hMCM5	[This work, 41 j]	BD-yMcm10 + AD-yOrc2	Absent
*hORC4 + hMCM2	Absent [41 j]	BD-yMcm10 + AD-yOrc3	Absent [41 j]
*hORC4 + hMCM4	Absent	BD-yMcm10 + AD-yOrc6	Absent [41 j]
*hORC4 + hMCM6	Absent [41 j]	BD-yMcm10 + AD-yCdt1	Absent
*hORC5 + hMCM4	[51 h]	BD-yMcm10 + AD-yCdc14	Absent
*hORC5 + hMCM7	[51 h]	BD-yNoc3 + AD-yCdc14	Absent
*hORC6 + hMCM5	Absent [41 j]	BD-yIpi3 + AD-yPsf1	Absent
BD-hMCM3 + AD-hORC1	[83 a]	BD-yIpi3 + AD-yPsf2	Absent
BD-hMCM6 + AD-hORC5	[This work]	BD-yIpi3 + AD-yCdc45	Absent
BD-yMCM6 + AD-yORC2	Absent [41 j]	BD-yIpi3 + AD-yMcm10	Absent
BD-yMCM5 + AD-yORC3	Absent [41 j]	BD-yIpi3 + AD-yCdc14	Absent
BD-yMcm6 + AD-yORC3	[This work]	BD-yOrc2 + AD-yPsf3	Absent
BD-yMcm6 + AD-yORC5	[This work]	BD-yCdt1 + AD-ySdl5	Absent
*hCDC6 + hORC4	[41 j, 44 a]	BD-yCdt1 + AD-yPsf1	Absent [41 j]
*hNOC3 + hORC2	[9 e]	BD-yCdt1 + AD-yPsf2	Absent
*hNOC3 + hORC4	Absent	BD-yCdt1 + AD-yPsf3	Absent
*hNOC3 + hORC5	[9 e]

Table 2 summarizes the novel physical interactions we identified in this study. Some of the novel interactions we identify have been identified in the homolog’s of other organisms. Some of the novel physical interactions we identified in this study have been reported by genetic methods as referenced (a, Reconstituted Complex; b, Affinity Capture-Western; c, Co-fractionation; d, Affinity Capture-MS; e, Yeast two-hybrid; f, Co-localization; g, Co-purification; h, Synthetic Lethality; I, Dosage Growth Defect; j, Negative Genetic) (a-g, Physical interactions; h-j, Genetic interactions).

Table 3.

Novel physical interactions verified by co-immunoprecipitation assay.

Novel yeast two-hybrid interactions verified by co-IP (*, bi-directional interaction; h, human; y, S. cerevisiae)
*hORC5 + hORC5	BD-yMcm6 + AD-yOrc3
*hNOC3 + hMCM2	BD-yMcm6 + AD-yOrc5
*hNOC3 + hMCM3	BD-yCdt1 + AD-ySdl5
*hNOC3 + hMCM7	BD-yCdt1 + AD-yPsf1
BD-yMcm6 + AD-yOrc2	BD-yCdt1 + AD-yPsf3

Table 3 summarizes the novel physical interactions identified by the yeast two-hybrid assay, that were further confirmed by co-IP.

The observed positive interactions among the yeast ORC subunits are summarized in Figure 1(a,b), Supplementary Figure S3, Tables 1 and 2, and Supplementary Table S1. With six yeast ORC subunits in both the BD- and AD-fusions, all 36 possible pairs of combinations were tested, of which 19 pairs of positive interactions were observed in either one direction (e.g. BD-yOrc4 interacted with AD-yOrc1, but BD-yOrc1 and AD-yOrc4 had no interaction) or two directions (e.g. BD-yOrc1 interacted with AD-yOrc2, and conversely, BD-yOrc2 also interacted with AD-yOrc1). Among the positive interactions in either one or both directions, the following have not previously been detected by yeast two-hybrid analysis, but have been observed using other methods in previous studies (Table 1): yOrc1 with yOrc1, yOrc2 and yOrc3; yOrc2 with yOrc2 and yOrc5; yOrc5 with yOrc4 and yOrc5; and yOrc6 with yOrc6.

Similar to yeast ORC, all 36 possible pairs of combinations among the human ORC subunits were tested, of which 19 pairs of interactions were positive in either one or two directions (Figure 1(c-f), Supplementary Figures S4, S5, and Tables 1 and 2). Among the positive interactions, the following were previously unknown: hORC1 with hORC4; hORC2 with hORC2; hORC4 with hORC4; and hORC5 with hORC5. We further confirmed the hORC5 self-interaction by co-IP (Supplementary Figure S12A and Table 3). HA-hORC5 and FLAG-hORC5 could be co-immunoprecipitated by either anti-FLAG antibody or anti-HA antibody, but not the control IgG, from the DNase I treated extract of HEK-293T cells co-expressing the two proteins. As negative controls, HA-hORC5 was not co-immunoprecipitated by anti-FLAG antibody from the extracts without hORC5-FLAG, and hORC5-FLAG was not co-immunoprecipitated by anti-HA antibody from extracts without HA-ORC5.

A comparison of the pairwise interactions among the yeast and human ORC subunits observed in this study and by others, as cited below, indicate that the following positive interactions were common between the budding yeast and human proteins: ORC1 interacts with ORC2 [9,33–41] and ORC4 [9,36–39,42–44]; ORC2 interacts with ORC2 [37], ORC3 [9,34,35,37,40,42–48], ORC4 [34,35,37,40,44,47–49], ORC5 [9,34,35,37,40,42–44,48,50,51] and ORC6 [37,40,43,44,47,52,53]; ORC3 interacts with ORC5 [37,40,42–44,48,50]; ORC4 interacts with ORC5 [9,30,35,37,43,44,48]; and ORC5 interacts with ORC5 [37,43]. These common ORC subunit interactions demonstrate a high degree of conservation in the ORC subunit interactions between the two species.

Inter-subunit interactions of the yeast and human MCM proteins

We have previously reported the studies of pairwise interactions among all six human MCM subunits and some of the yeast MCM subunits using the yeast two-hybrid system combined with co-IP and other methods [9,30]. In this study we examined all 36 possible pairs of combinations of the six yeast MCM subunits in both the BD- and AD-fusions, of which 22 pairs of interactions were positive in either one or two directions (Figure 2(a,b), Supplementary Figure S3, Tables 1 and 2, and Supplememtary Table S1). Although no novel interactions were identified, this study further confirmed our previously findings on budding yeast MCM proteins [9] and revealed the relative strengths of the interactions based on the host cell growth on SCM-4 plates (1+, weak; 2+; intermediate; 3+, strong; Figure 2, Supplementary Figure S3, Tables 1 and 2, and Supplementary Table S1).

Figure 2.

Budding yeast MCM intra-subunit interactions examined by yeast two-hybrid analysis. Host yeast cells expressing yeast AD-fusion proteins and BD-fusion proteins were streaked onto SCM-2, SCM-3 and SCM-4 plates and incubated at 30°C for three days to examine the relative strength of the interactions indicated by cell growth. The positive and negative controls were the same as those in Figure 1. (a) Summary of results from the yeast two-hybrid analysis (see Supplementary Figure S3 and S1 Table for data). (b) Summary diagram of the yeast two-hybrid interactions of MCM subunits. Refer to Figure 1 for the meaning of the symbols.

A comparison of the yeast and human MCM two-hybrid interaction data from this study and others (cited below) indicated that the following interactions were conserved between the budding yeast and human proteins: MCM2 interacts with MCM4 [9,29,30,39,54–58], MCM6 [9,29,30,44,54,55,58–60], and MCM7 [9,27,48,52]; MCM3 interacts with MCM5 [9,30,44,54,55,59,61,62] and MCM7 [9,30,44,54,55,59,61]; and MCM4 interacts with MCM5 [9,29,30,39,54–57,62–64], MCM6 [9,30,39,44,54–57,59,64–66], and MCM7 [9,30,39,54–57,59,63,65,66]. Furthermore, both yeast and human MCM5, −6 and −7 self-interact [9,30,61,62]. These data indicate a high degree of conservation in the MCM subunit interactions between yeast and humans.

Interactions of the yeast and human ORC subunits with MCM subunits

We also examined the interactions of the yeast and human ORC subunits with MCM subunits. We detected extensive interactions between ORC and MCM proteins from yeast (Figure 3(a), Supplementary Figure S3, Tables 1 and 2, and Supplementary Table S1) and humans (Figure (3b-d), Supplementary Figure S6, S7; and Tables 1 and 2).

Figure 3.

Budding yeast and human ORC and MCM, inter-subunit interactions examined by yeast two-hybrid analysis. (a) Summary diagram of the S. cerevisiae ORC and MCM inter-subunit interactions (see data in Supplementary Figure S3 and Table S1). (b) Host yeast cells expressing human AD-fusion proteins and BD-fusion proteins were patched onto SCM-2, SCM-3 and SCM-4 plates and incubated at 30°C for three days to test cell growth (see Supplementary Figures S6 and S7 for additional data). The positive and negative controls were the same as those in Figure 1. (c, d) Summary diagrams of the human ORC and MCM inter-subunit interactions indicated by cell growth on the SCM-3 (c) and SCM-4 (d) plates. Refer to Figure 1 for the meaning of the symbols.

With six yeast ORC subunits and six yeast MCM subunits in both the BD- and AD-fusions, all 72 possible pairs of yeast ORC-MCM combinations were tested, of which 12 pairs of positive interactions were observed in either one or two directions. Of the positive interactions, the following were previously unknown: yMcm5 with yOrc3; yMcm6 with yOrc2; yMcm6 with yOrc3; and yMcm6 with yOrc5. The interactions of yMcm6 with yOrc2, yOrc3 and yOrc5 were further confirmed by co-IP (Supplementary Figure S12B, C, D and Supplementary Table S3). DNase I treated extracts from yeast expressing HA-yMcm6 were incubated with anti-HA antibody, anti-yOrc2, anti-yOrc3 or anti-yOrc5 antibody, to immunoprecipitate HA-yMcm6, and endogenous Orc2, yOrc3 and yOrc5, respectively. HA-yMcm6 and yOrc2 were reciprocally co-immunoprecipitated by either anti-Myc or anti-yOrc2 antibody, but not by the control IgG (Supplementary Figure S12B). Similar reciprocal co-IP results were observed for the interactions of HA-yMcm6 with yOrc3 (Supplementary Figure S12C) and of HA-yMcm6 with yOrc5 (Supplementary Figure S12C).

Similarly, all 72 possible pairs of human ORC-MCM combinations were tested, of which 46 pairs of interactions were positive in either one or two directions (Figure 3(b,c), Supplementary Figures S6, S7, and Tables 1 and 2). The following positive human ORC-MCM interactions have not been previously reported: hORC1 with hMCM2, hMCM3, hMCM4, hMCM6 and hMCM7; hORC2 with hMCM4, hMCM6 and hMCM7; hORC3 with hMCM4, and hMCM5; hORC4 with hMCM2, hMCM4 and hMCM6; hORC5 with hMCM4, hORC5 and hMCM7; and hORC6 with hMCM5.

A comparison of yeast and human ORC-MCM interaction data from this study and others (cited below) indicated that the following interactions were conserved between budding yeast and humans: ORC2 interacts with MCM2 [35,60], MCM5 [44,46,51], MCM6 [41], and MCM7 [41,44,51]; ORC3 interacts with MCM5 [41] and MCM6. ORC5 interacts with MCM2 [14,35,55,60,67] and MCM6. These data strongly suggest that both yeast and human ORC and MCM proteins interact with each other in a conserved manner.

We can also see from the above analyses that the human ORC-MCM interactions observed in the yeast two-hybrid system were more abundant than the corresponding yeast proteins. Interestingly, yeast two-hybrid studies of mouse ORC-MCM interactions [68] indicated striking similarities to the human ORC-MCM interactions we observed.

Interactions of yeast and human CDC6 with ORC and MCM subunits

We have previously reported that yCdc6p interacts with yOrc1p and yOrc2p [9]. In this study, all 24 possible pairs of combinations ofCDC6 with each of the six ORC subunits and six MCM subunits in two directions were tested for both the yeast and human proteins. We found four pairs of positive interactions for the yeast proteins (Figure 4(a). Supplementary Figure S3, Tables 1 and 2 and Supplementary Table S1) and 12 pairs for the human proteins (Figure 4(b-d), Tables 1 and 2, and Supplementary Figure S8) in either one or two directions. These interactions are consistent with the role of CDC6 as a licensing factor in yeast and human DNA replication.

Figure 4.

yCdc6 and hCDC6 interact with ORC and MCM subunits in the yeast two-hybrid system. (a) A summary diagram of the S. cerevisiae CDC6 interactions with yORC and yMCM subunits (see data in Supplementary Figure S3 and Table S1). (b, c) Host yeast cells expressing human AD-fusion proteins and BD-fusion proteins were patched onto SCM-2 and SCM-3 plates and incubated at 30°C for three days to test cell growth (see Supplementary Figure S8 for additional data). The positive and negative controls were the same as those in Figure 1. Positive interactions are marked by grey shades in the tables. (d) A summary diagram of the human CDC6 interactions with hORC and hMCM subunit. Refer to Figure 1 for the meaning of the symbols.

The following positive interactions have not previously been reported: hCDC6 with hORC1, −2, −4 and with hMCM2, −3 and −7 (Figure 4(b-d), Tables 1 and 2, and Supplementary Figure S8). In addition, the interaction between yCdc6 and yOrc3 detected here has not been detected by yeast two-hybrid assay before, although it was detected by cryo-EM [42]. Similarly, the interaction between yCdc6 and yMcm3 detected here has not been detected by yeast two-hybrid assay before, although it was detected by negative genetic screening [41] and reconstituted complex [44] methods previously, as shown in Table 1.

Comparing the yeast and human protein data from this study and others (cited below), CDC6 interacted with ORC1 [9,35,41,42], ORC2 [9,35,41,42,44,51,67] and MCM3 [9,35,41,44,69] in both yeast and humans. Collectively, these data indicate the conserved interactions of CDC6 with ORC and MCM proteins.

Interactions of human and yeast CDT1 with ORC and MCM subunits

We have previously reported that yCdt1p interacts with yMcm6p and yOrc6p and that hCDT1 interacts with hMCM2p [9,29,32]. In this study, all 24 possible pairs of combinations of CDT1 with each of the six ORC subunits and six MCM subunits in two directions were tested for both the yeast and human proteins. We found six pairs of positive interactions for the yeast proteins (Figure 5(a), Figure S3 and Supplementary Table S1) and six pairs for the human proteins (Figure 5(b-d) and Figure S9) in either one or two directions. These interactions are consistent with the role of CDT1 as a licensing factor in yeast and human DNA replication.

Figure 5.

yCdt1 and hCDT1 interact with ORC and MCM subunits in the yeast two-hybrid system. (a) A summary diagram of the S. cerevisiae CDT1 interactions with yORC and yMCM subunits (see data in Supplementary Figure S3 and Table S1). (b, c) Host yeast cells expressing human AD-fusion proteins and BD-fusion proteins were patched onto SCM-2 and SCM-3 plates and incubated at 30°C for three days to test cell growth (see Supplementary Figure S9 for additional data). The positive and negative controls were the same as those in Figure 1. Positive interactions are marked by grey shades in the tables. (d) A summary diagram of the human CDT1 interactions with hORC and hMCM subunits. Refer to Figure 1 for the meaning of the symbols.

The following positive interaction has not been previously detected by yeast two-hybrid analysis: yCdt1 with yOrc3 (Figure 5(a), Figure S3 and Table S1). Comparatively, both yeast and human CDT1 interacted with MCM6 [this study and refs [29,70].].

Interactions of yeast and human NOC3 and IPI3 with ORC and MCM subunits

We have previously reported that yNoc3p interacts with yOrc2p, yOrc3p, yOrc5p, yOrc6p, yMcm6p and yCdt1p [9,13] (Supplementary Figure S3 and Supplementary Table S1). In this study, our yeast two-hybrid analysis showed that hNOC3 (FAD24) had interactions with hORC2, hORC4, hORC5, hMCM2, hMCM3 and hMCM7 (Supplementary Figure S10). Furthermore, we also confirmed the interactions of hNOC3 with hMCM2, hMCM3 and hMCM7 by co-IP (Supplementary Figure S12E and Table 3). hMCM2, hMCM3 and hMCM7 were found to be co-immunoprecipitated with HA-hNOC3 by the anti-HA antibody, but not the control IgG, from DNase I treated extracts of HEK-293T cells expressing HA-hNOC3. As another negative control, hMCM2, hMCM3 or hMCM7 were not co-immunoprecipitated by anti-HA antibody from extracts without HA-hNOC3.

These interactions are consistent with the role of NOC3 as a pre-RC component in yeast and human cells. Comparatively, both yeast and human NOC3 interacted with ORC2 and ORC5 [this study and ref [9].].

We have also previously reported that yIpi3p interacts with yOrc2p, yCdc6p, yCdt1p and yMcm4 [9] (Supplementary Figure S3 and Supplementary Table S1). We also reported that hIPI3 (WDR18) interacts with hORC1, hORC2, hORC3, hORC4, hORC5, hMCM2, hMCM3, hMCM4 and hMCM7 [28].

Comparatively, both yeast and human IPI3 interacted with ORC2 and MCM4 [9,28]. The collective evidence is consistent with human IPI3, being a licensing factor in pre-RC formation in both yeast and human cells.

Interactions among yeast and human CDC6, CDT1, NOC3 and IPI3 proteins

We have previously reported that yIpi3 interacts with yCdt1 and yCdc6, while yNoc3 interacts with itself and yCdt1 [9]. In this study, a total of 12 pairs of BD- and AD-fusion yeast proteins and 16 pairs of BD- and AD-fusion human proteins were tested to examine the interactions among Cdc6p, Cdt1p, Noc3p and Ipi3p. Four other combinations of the yeast proteins could not be tested because AD-yIpi3 did not express (Supplementary Figure S1). We found 4 pairs of positive interactions for the yeast proteins (Figure 6(a), Supplementary Figure S3 and Supplementary Table S1) and 7 pairs for the human proteins (Figure 6(b,c) and Supplementary Figure S11) in either one or two directions. The following positive interactions have not been previously detected: yIpi3 with yNoc3; hNOC3 with hIPI3 and hNOC3; hIPI3 with hCDC6 and hCDT1 (Figure 6(b,c) and Supplementary Figure S11). These interactions further support the roles of NOC3 and IPI3 in pre-RC formation. Comparatively, NOC3 interacted with itself in both yeast and humans [this study and refs [9,71].]. Additionally, CDC6 interacts with IPI3 in both yeast and humans [this study and ref [9].].

Figure 6.

Inter-subunit interactions of budding yeast and human CDC6, CDT1, NOC3 and IPI3 proteins in the yeast two-hybrid system. (a) A summary diagram and table of the S. cerevisiae MCM-loading protein interactions (see data in Supplementary Figure S3 and Table S1). (b) Host yeast cells expressing human AD-fusion proteins and BD-fusion proteins were patched onto SCM-2 and SCM-3 plates and incubated at 30°C for three days to test cell growth (see Supplementary Figure S11 for additional data). The positive and negative controls were the same as those in Figure 1. (c) A summary table and diagram of the human MCM-loading proteins’ interactions. Refer to Figure 1 for the meaning of the symbols.

Intra-subunit interactions of yeast GINS complex proteins and inter-subunit interactions of yeast mcm10, Cdc14, Pol30 and Cdc45 proteins in the yeast two-hybrid system

The initiation of DNA replication in yeast requires the formation of the heterotetrameric GINS complex [20,72,73]. The GINS complex is composed of Sld5p (GINS4), Psf1p (GINS1), Psf2p (GINS2) and Psf3p (GINS3) [20,72,73]. In this study, a total of 12 pairs of the BD- and AD-fusions from yeast Sld5p, Psf1p, Psf2p and Psf3p were tested. Four other combinations could not be tested because BD-yPsf3 did not express (Supplementary Figure S1). We found 11 pairs of positive interactions (Figure 7(a,b), Supplementary Figure S3 and Supplementary Table S1) in either one or two directions. The following positive interactions have not been previously detected by yeast two-hybrid analysis: yPsf1 with yPsf1; ySld5 with yPsf3 (Figure 7(a,b), Supplementary Figure S3 and Supplementary Table S1).

Figure 7.

Intra-subunit interactions of yeast GINS complex proteins and inter-subunit interactions of yeast Mcm10, Cdc14, Pol30 and Cdc45 proteins in the yeast two-hybrid system. Host yeast cells expressing yeast AD-fusion proteins and BD-fusion proteins were streaked onto SCM-2 and SCM-3 plates (except for BD-ySld5 combined with AD-fusion proteins that were tested on SCM-4 plates) and incubated at 30°C for three days to examine the relative strength of the interactions indicated by cell growth (see Supplementary Figure S3 and S1 Table for data). The positive and negative controls were the same as those in Figure 1. (a, b) A summary table and diagram of the GINS intra-subunit interactions. (c, d) A summary table and diagram of the Mcm10, Cdc14, Pol30 and Cdc45 inter-subunit interactions. Refer to Figure 1 for the meaning of the symbols.

Mcm10p is involved in the origin association and stabilization of the MCM complex in DNA replication initiation [74,75]. Cdc14p is involved in mitotic exit and enabling replication licensing by dephophorylating CDK1 substrates including ORC, MCM and Cdc6p [4]. Pol30p functions as a clamp for DNA polymerase delta and creates a loading platform for other proteins that are involved in DNA replication and repair [74,76]. Cdc45p in complex with Mcm2-7p Cdc6p and Cdc18p is involved in DNA replication initiation [21]. In this study, 12 pairs of the BD- and AD-fusions of yeast Mcm10p, Cdc14p, Pol30p and Cdc45p were tested. Four other combinations could not be tested because BD-yCDC14 could self-activate reporter gene expression (Supplementary Figure S2). We found three pairs of positive interactions (Figure 7(c,d), Supplementary Figure S3 and Supplementary Table S1) in either one or two directions. The following positive interaction has not been previously detected by yeast two-hybrid analysis: yMcm10 with yCdc14 (Figure 7(c,d), Supplementary Figure S3 and Supplementary Table S1).

Additionally, we also examined whether or not individual yGINS protein complex components interact with yMcm10p, yCdc14p, yPol30p and yCdc45p. Twenty-four pairs of combinations were tested. We found that none of the tested pairs had interaction (Figure S3 and Supplementary Table S1).

Interactions of yeast Mcm10, Cdc14, pol30, CDC45 and GINS protein with ORC, MCM and MCM-loading proteins

In this study, a total of 218 pairs of the BD- and AD-fusions from yeast Mcm10p, Cdc14p, Pol30p, Cdc45p, Sld3p, Psf1p, Psf2p, Psf3p, Cdc6p, Cdt1p, Noc3p, Ipi3p, six ORC and six MCM subunits were tested. We found 33 pairs of positive interactions (Figures 8 and 9, Supplementary Figure S3 and Supplementary Table S1) in either one or two directions. The following positive interactions have not been previously detected by yeast two-hybrid analysis: ySld5 with yOrc2 and yOrc5; yPsf1 with yOrc2, yIpi3 and yCdt1; yPsf2 with yOrc2 and yCdt1; yPsf3 with yCdt1; yMcm10 with yOrc2 and yOrc3; yORC6 yIpi3 and yCdt1; yNoc3 with yCdc14; yIpi3 with yCdc45 and yCdc14; yORC2 with yPsf3; yCdt1 with ySdl5 (Figures 8 and 9, Supplementary Figure S3 and Supplementary Table S1). The interactions of ySld5 with yCdt1, yPsf2 with yCdt1, and yPsf3 with yCdt1were further confirmed by co-IP (Supplementary Figure S12F, G, H and Table 3). DNase I treated extracts of yeast cells co-expressing Myc-yPsf1 and yCdt1-FLAG were separately treated with anti-FLAG and anti-Myc antibodies to immunoprecipitate Myc-yPsf1 and yCdt1-FLAG, respectively. Myc-yPsf1 and yCdt1-FLAG were reciprocally co-immunoprecipitated by either anti-FLAG antibody or anti-Myc antibody, but not the control IgG. As another negative control, Cdt1-FLAG was not immunoprecipitated by anti-Myc antibody from cell extracts without expressing Myc-yPsf1 (Supplementary Figure S12F). Similar results were observed for the interactions of yCdt1 with yPsf3 (Supplementary Figure S12G) and yCdt1 with ySld5 (Supplementary Figure S12H).

Figure 8.

A summary diagram of the interactions of yeast GINS complex proteins with ORC, MCM and MCM-loading proteins. Host yeast cells expressing yeast AD-fusion proteins and BD-fusion proteins were streaked onto SCM-2 and SCM-3 plates (except for BD-ySld5 and BD-yMcm7 combined with AD-fusion proteins that were tested on SCM-4 plates) and incubated at 30°C for three days to examine the relative strength of the interactions indicated by cell growth (see Supplementary Figure S3 and S1 Table for data). The positive and negative controls were the same as those in Figure 1. Positive interactions are depicted in the diagram. Refer to Figure 1 for the meaning of the symbols.

Figure 9.

Interactions of yeast Mcm10, Cdc14, Pol30 and Cdc45 with ORC, MCM and MCM-loading proteins. Host yeast cells expressing yeast AD-fusion proteins and BD-fusion proteins were streaked onto SCM-2 and SCM-3 plates (except for BD-yMcm7 combined with AD-fusion proteins that were tested on SCM-4 plates) and incubated at 30°C for three days to examine the relative strength of the interactions indicated by cell growth (see Supplementary Figure S3 and S1 Table for data). The positive and negative controls were the same as those in Figure 1. Positive interactions are depicted in the diagram. Refer to Figure 1 for the meaning of the symbols.

Discussion

Previous studies on pre-RC and pre-IC components have focused on individual proteins, complexes or sub-complexes in an effort to determine the pre-RC inter-subunit interactions and architecture. We have for the first time conducted a comprehensive yeast two-hybrid study aimed at giving a holistic insight into the pairwise physical interactions of the yeast and human pre-RC and pre-IC components. We identified many previously unknown interactions as well as interactions reported in existing literature. We additionally demonstrated conservation of physical interactions between proteins in these two species. The evidence overwhelmingly supports similar parallels of interactions between yeast and human pre-RC components. Our results therefore, confirm and corroborate results from previous studies that utilize alternate techniques such as reconstituted complex, affinity capture-western, co-fractionation, affinity capture-MS, co-localization, co-fractionation and co-purification to demonstrate interaction or function (Table 1). Lastly, some key novel interactions identified by our yeast two-hybrid analysis were further confirmed by co-immunoprecipitation assay (Supplementary Figure S12 and Table 3).

Our ORC inter-subunit interaction data agree with previous reports concerning the EM structure of yeast ORC, its subunit interactions and spacial positions [42]. Moreover, our data also corroborate previous yeast two-hybrid studies on yORC complex [43]. Furthermore, four yORC subunits, yOrc2p, yOrc3p, yOrc5p and yOrc6p, associate to form the core of the yORC hexamer [42], whereas yOrc1p and yOrc4p, whose ATPase activity are essential for pre-RC formation, form the upper part of the hexamer [42,43]. Both our yeast and human ORC interaction data indicate that the core ORC sub-complex consists of ORC2, −3, −4 and −5 as these components possess the most inter-subunit interactions. Both human and yeast ORC1 and ORC6 have comparatively fewer interactions, which is consistent with previous reports [40,48,77].

Similarly, with respects to inter-subunit interactions of MCM complex, our data agree with previously published yeast and human MCM protein interaction studies [30,58,78,79]. Our data showing the self-interaction of the MCM subunits further support the findings that the MCM complex is loaded on replication origins as a head-to-head double hexamer [39,55,80].

Furthermore, the interactions we have reported between budding yeast and human ORC and MCM proteins suggest that the two complexes interact with each other at the interface as shown in previous studies on the mouse and budding yeast pre-RC proteins [51,68]. Comparatively, human and mouse interactions are strikingly more abundant as compared to the yeast, possibly owing to the different levels of organization and complexity between higher and lower order eukaryotes. Our data indicate that all six hORC subunits interact with hMCM4. Furthermore, hMCM3 interacts with hORC1, −4 and −5, while hMCM5 interacts with hORC2, −3 and −6.

Our current and published data clearly indicate that yeast and human NOC3 and IPI3 interact with several ORC, CDC6, CDT1 and MCM proteins, in both organisms. Interestingly, both NOC3 and IPI3 interact with the ORC subunits that form the core of the ORC hexamer, indicating these two proteins are integral components of the pre-RC [42,43]. Our data also show that CDC6 interacts with several ORC and MCM subunits, consistent with electron microscopy studies of the ORC-Cdc6p complex that has a ring shaped structure similar in dimensions to that of the MCM complex [81]. These data further match the requirement and the downstream action of ORC-CDC6 complex to load the MCM complex [69,82]. Furthermore, our data are also consistent with cryo-EM studies indicating the structure of the ORC-Cdc6 and Cdt1-MCM2-7 heteroheptamers [53]. Similarly, our data also shows that both human and yeast CDT1 interact with several ORC and MCM components, further confirming its role in pre-RC formation in eukaryotes [83]. The interactions of yeast Cdt1p with Orc2 and −6 reported in our study are also consistent with the findings of a previous study [84].

Interestingly, human NOC3, like its yeast homolog [71], has self-interactions. This finding may be of biological significance as we have identified several self-interacting components of both human and yeast ORC and MCM (Figures 1-3). Moreover, we found that both yeast and human NOC3 interact with IPI3 (Figure 6). It is interesting to note that both yNoc3p and yIpi3p have also been reported as being involved in ribosome biogenesis [9,13]. Furthermore, both yNoc3 and yIpi3 were found to interact with yCdt1, while this interaction was absent in the human proteins (Figure 6). On the other hand, hIPI3 does interact with hCDC6, while hCDC6 interacts with hCDT1 (Figure 6, Supplementary Figure S8). These data indicate that our yeast two-hybrid study indeed identified many physiologically relevant interactions which may be direct or indirect protein-protein interactions, as hinted by the hNOC3-hIPI3-hCDC6-hCDT1 interactions.

Inter-subunit interactions among the GINS complex proteins are consistent with previous reports in budding yeast and other model organisms [10,32,42,85,86]. Some pre-IC proteins were also found to interact with the pre-RC proteins (Figure 10). Our results, consistent with others [20,87,88], demonstrate that ySld5p, yPsf1p, yPsf2p and yPsf3p associate tightly to form the yGINS complex. Moreover, we discovered that the yGINS complex interacts with the yORC and yMCM complexes, yCdt1p and yIpi3p (Figures 7 and 9). It has been reported that GINS and the MCM complex, together with Cdc45p, form a large CMG (Cdc45-MCM-GINS) complex to perform the helicase function [16,20,78]. As our yeast data indicates that yIPI3 interacts with yORC, yMCM, yGINS and yCDC45, yIpi3, in addition to being a component of the pre-RC [9], may also form a larger previously unidentified complex in pre-IC.

Whilst large scale yeast two-hybrid screenings are usually not saturating and likely to miss some interactions, our comprehensive pairwise tests of individual proteins under optimal conditions could detect most, if not all, potential interactions. However, while we have identified numerous protein-protein interactions among the yeast and human pre-RC components, many of which are similar between the yeast and human proteins, we are cautious to note that certain expected interactions were not detected, while some unexpected interactions were observed. We therefore suggest that some interactions may be weak and/or transient during the association and dissociation of functional protein complexes. Alternatively, the use of full-length ORFs fused to the BD and AD domains in yeast two-hybrid assays may mask certain real interactions between proteins, especially the large ones. On the other hand, some detected interactions may not be physiologically relevant. For example, it is difficult to imagine how hORC2 interacts with five other hORC subunits in the ORC-CDC6 ring structure.

In conclusion, this study identified 113 positive interactions among 24 budding yeast DNA replication proteins and 96 positive interactions among 16 human DNA replication proteins, many of which were previously unknown. Of these novel interactions, six yeast and four human protein interactions were further confirmed by co-immunoprecipitation assays. We also identified several self-interactions among the pre-RC components, most notably NOC3, ORC and MCM proteins. We further demonstrate a link between pre-RC and pre-IC as yeast GINS complex proteins interact with yMCM-loading factors, yORC and yMCM proteins. Moreover, we identified yPsf1 and yPsf2 as having self-interactions. Furthermore, we show that yMcm10 interacts with yMCM (yMcm2, −3, −4, −6 and −7) as previously reported [75]. Additionally, we report that yMcm10 also interacts with yIpi3, yOrc2, −3, and −6, while yCdc45 interacts with yOrc2 and yIpi3. Interestingly, yCdc14 also interacts with yIpi3 and yNoc3. As yCdc14 has been reported to dephosphorylate targets of CDKs, which results in mitotic exit and enables pre-RC [4], yNoc3 and yIpi3 may also be CDK substrates. Lastly, we additionally found that yIpi3p interacts with yPsf1, −2, yCdc14p and yCdc45p. This study may serve to enhance our understanding of the overall pre-RC architecture and functions in eukaryotes. Importantly, the biological significance of several identified novel interactions has yet to be examined.

Funding Statement

This work was supported by the Center for Nasopharyngeal Carcinoma Research, Hong Kong [AoE/M-06/08]; Hong Kong Research Grants Council [GRF 661713]; Guangzhou Committee of Science and Information Innovation, China [201604020038]; Foshan Science and Technology Bureau, Guangdong, China [2015IT100132].

Acknowledgments

We gratefully acknowledge Kelvin K.L. Sou, Doris Y. Qin, Winnie W.Y Cheung, Fortune S.W. Chung and Jeffery Yeung for providing valuable technical assistance. We would also like to thank Dr. John F. Scott for his review of the manuscript. This work was supported by grants from Hong Kong Research Grants Council (GRF 661713), Center for Nasopharyngeal Carcinoma Research, Hong Kong (AoE/M-06/08), Foshan Science and Technology Bureau, Guangdong, China (2015IT100132), and Guangzhou Committee of Science and Information Innovation, China (201604020038).

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplemental material

Supplemental data for this article can be accessed here.