Isolation of the cDNA of human MCM10 homolog
A database search identified a human EST clone (AA312197) as a potential homolog of the S.cerevisiae MCM10
. This cDNA fragment was obtained by RT–PCR and used for screening a HeLa cDNA library. The longest human cDNA clone obtained (3721 bp) encoded a predicted protein of 874 amino acids with a calculated molecular mass of 98 kDa (Fig. a). The initiator methionine codon was preceded by an untranslated leader sequence that contained a stop codon in the same reading frame. The sequence surrounding the initiation codon matched with the Kozak consensus for eukaryotic translation initiation (41
). The cDNA contained an Alu
I repetitive sequence as well as poly(A)+
signal in the 3′ non-coding region. Human Mcm10 (HsMcm10) was 28% identical and 42% similar to D.melanogaster
Mcm10, 19% identical and 33% similar to C.elegans
Mcm10, 12% identical and 23% similar to S.pombe
Cdc23, and 11% identical and 21% similar to S.cerevisiae
Mcm10. The central region of HsMcm10 (from amino acids 236 to 435) showed a high level of identity with the D.melanogaster
homolog (42%), C.elegans
homolog (29%), S.pombe
homolog (24%) and S.cerevisiae
homolog (24%) (Fig. b). This central homologous region, which contains the sequence resembling a zinc-finger motif (CX9-10
H), previously noted in both S.cerevisiae MCM10
and S.pombe cdc23
gene products (27
), appears conserved among all five species. In S.cerevisiae
, the substitution of the second cysteine residue by proline [in the mcm10-43
) mutant] suggests an important role of this motif in the biological function of Mcm10 (26
Figure 1 Comparison of Mcm10 proteins of different species. (a) Schematic diagram of the similarities among human (Homo sapiens), D.melanogaster, C.elegans, S.pombe and S.cerevisiae Mcm10 proteins. Striped boxes indicate the conserved central domain. Black boxes (more ...) Schizosaccharomyces pombe cdc23
was able to complement both S.cerevisiae mcm10-43
) mutant (27
) and mcm10-1
mutant (data not shown). As an initial characterization of HsMcm10, we tested whether HsMcm10 could complement a temperature-sensitive allele of Mcm10. Full-length HsMcm10 was subcloned into an expression vector and introduced into the temperature-sensitive mcm10-1, mcm10-43
) and cdc23–M36
strains. HsMcm10 did not functionally replace mutant Mcm10 either in budding yeast or in fission yeast (data not shown).
HsMcm10 mRNA expression during the cell cycle
In both HeLa and NB1–RGB cells, northern blot analysis revealed the presence of two HsMcm10 transcripts of 5.0 and 3.7 kb in a ratio of 2:1. Genomic Southern blot analysis demonstrated that the human genome contained a single copy of the HsMcm10 gene (data not shown). The different mRNA species may result from an alternative splicing event in the 3′ non-coding region. We were not successful in isolating the cDNA clone which contained the entire 3′ non-coding region. However, nine of the 13 isolated cDNA clones, which contained the entire open reading frame, did not contain any alternatively spliced form, although the probe used for library screening recognized two kinds of mRNA species. We also made five probes of ~200 bp which hybridized to the coding region at every 500 bp. All five probes detected two mRNA species by northern blot analysis (data not shown). Therefore, it is unlikely that the two different mRNA species result from alternative splicing within the coding region.
We examined whether mRNA expression of HsMcm10 was growth dependent. The mRNA levels of most proteins involved in DNA replication increase at the G1/S boundary after quiescent mammalian cells are stimulated by serum. NB1–RGB cells were arrested at the G0 phase by serum starvation and stimulated to proliferate by serum addition. The mRNA level was measured by northern blot hybridization at the indicated times after serum addition (Fig. ). The mRNA level increased from 8 h after serum stimulation, reached maximum at 16 h and subsequently decreased. The ratio of the two mRNA species did not change throughout the cell cycle. This fluctuation is typical for DNA replication proteins, in accord with the notion that Mcm10 is necessary for DNA replication. As a positive control for the effect of serum stimulation, the mRNA level of PCNA, an auxiliary protein of DNA polymerase δ, was also examined and shown to increase before the synthesis of DNA.
Figure 2 Expression of human Mcm10 mRNA after the growth stimulation. After cultured in DMEM containing 0.4% fetal calf serum for 72 h, NB1–RGB cells were released by changing the medium to DMEM containing 15% fetal calf serum. Two micrograms (more ...) HsMCM10
gene was found in the human genome database (accession number AL138764) and mapped on chromosome 10. The upstream sequence of HsMcm10 contained several potential E2F-binding motifs. The transcription of HsMcm10 may be regulated by E2F transcription factors at the G1
/S boundary as shown for many other DNA replication proteins including dihydrofolate reductase, DNA polymerase α, Orc1, Mcm4 and Cdc6 (42
Subcellular localization of the HsMcm10 in mammalian cells
To determine the subcellular localization of HsMcm10, we made the expression vector for HsMcm10. The entire coding region except for the first methionine codon was subcloned into the multiple cloning sites of pSRhisA, which is a mammalian expression vector carrying the SRα promoter as well as six-His-tag in the upstream region of the multicloning sites (33
). COS-1 cells were transiently transfected with pSRhisA/HsMcm10, and 24 h after transfection, cells were fixed and stained with monoclonal antibody specific to the six-His-tag and FITC-conjugated secondary antibody. Ectopically expressed HsMcm10 protein was localized in the nuclei and formed a fine array of foci which were excluded from the nucleoli (Fig. a). These foci were resistant to detergent extraction as shown in Figure b. HsMcm10 was also subcloned into ptetHA, which contains the tetracycline-regulated promoter as well as HA epitope-tag upstream of the multicloning sites, and co-transfected with a tTA-expressing plasmid. HsMcm10 expressed by the ptetHA vector gave the same results (data not shown).
Figure 3 Localization of ectopically expressed HsMcm10 in COS-1 cells. (a) COS-1 cells were transiently transfected with the expression vector encoding the six-His-tagged HsMcm10 protein. Cells were fixed at 24 h after transfection, and the expression was detected (more ...)
The resistance of immunostaining against detergent extraction suggests that HsMcm10 may be associated with chromatin. For further analysis of subcellular localization of HsMcm10, we raised a polyclonal rabbit antibody against the recombinant fragment of HsMcm10. The major protein band of 105 kDa, which was slightly larger than the predicted molecular mass, was detected by immunoblotting of HeLa total cell lysate (Fig. a, lane 1, indicated by an arrow). To confirm the specificity of the antibody, HA-tagged HsMcm10 was ectopically expressed by transient transfection in HeLa cells and detected by immunoblotting. Addition of HA-epitope to HsMcm10 produced a 120-kDa protein which was detected by the anti-HsMcm10 antibody (Fig. a, lanes 2 and 3, indicated by an arrowhead). The 120-kDa protein was also detected by the anti-HA antibody (data not shown). Therefore, we concluded that the antibody specifically detected HsMcm10 in human cell extract. The minor protein band of 86 kDa which was detected in HeLa total cell lysate (Fig. a, lane 1, indicated by an asterisk) seems to be a degradation product of HsMcm10. We observed similar degradation product of overexpressed HA-tagged HsMcm10 as shown in Figure a, lane 3 (indicated by a dot).
Figure 4 Subcellular distribution of endogenous HsMcm10 in HeLa cells. (a) Detection of HsMcm10 protein with a polyclonal antibody. Whole cell lysates from 2 × 104 HeLa cells were subjected to immunoblotting (lane 1). The arrow indicates the endogenous (more ...)
HeLa whole cell extracts were fractionated into Triton X-100-soluble fractions, chromatin-bound fractions solubilized with DNaseI digestion, and DNaseI-unextractable fractions, and subcellular localization of HsMcm10 was investigated by immunoblotting (Fig. b). When extracts were prepared from asynchronously growing HeLa cells, HsMcm10 proteins were mainly localized in DNaseI-resistant fraction. About 10% of HsMcm10 was localized in Triton X-100-extractable fraction. To confirm that fractionation was performed properly, localization of Mcm7 and lamin B1 were investigated. As demonstrated previously, Mcm7 resolved into a doublet on SDS–PAGE. More than half of the Mcm7 proteins were free in soluble fractions in HeLa cells, while the rest of Mcm7 was bound to chromatin, which is consistent with the previous report by Fujita et al.
). On the other hand, lamin B1 was localized in DNaseI-resistant structure.
Next we fractionated the lysates of HeLa cells arrested at the G1/S boundary or at M phase. Most HsMcm10 proteins were localized in DNaseI-resistant fraction at G1/S boundary, whereas HsMcm10 dissociated from DNaseI-resistant structure and became soluble at M phase (Fig. b). To determine whether HsMcm10 dissociates from nuclear structures during the progression of S phase in the way that Mcm2–7 complex does, HeLa cells were arrested at the G1/S boundary by aphidicolin treatment following mitotic shake-off, and then released in aphidicolin-free medium. HsMcm10 remained in DNaseI-resistant structure during S phase, whereas Mcm7 dissociated from chromatin during S phase progression (Fig. c and d). HsMcm10 started to dissociate from DNaseI-resistant structure 6 h after release. At 6 h after release, G2 phase cells, which had intact lamina structure and no BrdU incorporation, began to increase. M phase cells began to appear at 9 h after release. Therefore, HsMcm10 seems to dissociate from DNaseI-resistant nuclear structure in G2 phase.
HsMcm10 associated with other replication factors
We speculated that Mcm10 might interact with other initiation factors, since the previous report demonstrated that Mcm10 is involved in initiating DNA replication in budding yeast (25
). Therefore, the cDNAs for DNA replication factors were subcloned into the expression vector described above and co-transfected with pSRhisA/HsMcm10 or ptetHA/HsMcm10. When HsMcm10 was co-expressed with human Orc2, HsMcm10 was found to co-localize with not all, but some Orc2 in the foci (Fig. a), whereas HsMcm10 was not co-localized with PCNA, DNA polymerase α or p21 (data not shown). To confirm that the co-localization between HsMcm10 and Orc2 did not result from the non-specific aggregation from overexpression, the expression vector for GFP-tagged human lamin A mutant, which lacks the C-terminal CaaX sequence, was also introduced with ptetHA/HsMcm10 as a negative control. The CaaX motif is necessary for the efficient integration of lamin A protein into the nuclear lamina, whereas lamin A without the CaaX motif accumulates within nuclei as multiple foci when overexpressed (30
). HsMcm10 was not co-localized with GFP-tagged lamin A mutant as shown in Figure b, suggesting that co-localization of HsMcm10 and Orc2 did not result from the non-specific aggregation caused by overexpression. Co-localization of Mcm10 and Orc was also observed in budding yeast (44
Figure 5 Co-localization of ectopically expressed HsMcm10 and Orc2 in COS-1 cells. The expression vector encoding the His-tagged HsMcm10 protein was co-transfected with either the vector expressing the HA-tagged Orc2 (a) or that expressing the GFP-tagged lamin (more ...)
As an independent method for detection of the physical interaction between Mcm10 and Orc2, we used the two-hybrid system to examine whether similar interactions could occur in yeast cells. Full-length cDNAs encoding the HsMcm10 and human Orc2 genes were fused to the B42 transactivation domain (AD) and the N-terminal LexA DNA-binding domain (BD), respectively. These constructs were introduced in the same cell and β-galactosidase activity was measured. As shown in Figure , positive interaction was detected in cells co-expressing the HsMcm10 and Orc2 proteins. The β-galactosidase activity caused by HsMcm10 and Orc2 interaction was strong and was ~70% of the activity observed with control cells harboring simian virus 40 large T antigen in the AD vector and p53 protein in the BD vector. When the HsMcm10 and Orc2 cDNAs were switched between BD vector and AD vector, a similar strong signal was obtained (data not shown).
Figure 6 Two-hybrid interactions between human Mcm10 (HsMCM10), human Orc2 (HsORC), and mouse Mcm2–7 (mMCM2–7). Interactions between pairs of fusion proteins containing either an N-terminal LexA DNA-binding domain (BD) or a B42 transactivation (more ...)
The interaction between Mcm10 and five of the six Mcm2–7 proteins (Mcm2–4, 6 and 7) has been reported in budding yeast (26
). The cDNAs for mouse Mcm2–7 were subcloned into the BD vector and cotransfected with HsMcm10 in the AD vector. Mcm2–7 proteins are well conserved between human and mouse, and their proteins are functionally equivalent and interchangeable when the Mcm2–7 complex was reconstituted in vitro
(Z.You, personal communication). As shown in Figure , interactions between HsMcm10 and mouse Mcm2 and Mcm6 were observed. A similar interaction was observed when the cDNAs were switched between BD and AD vectors (data not shown). None of the other Mcm proteins interacted with HsMcm10. The β-galactosidase activity observed from the interaction between HsMcm10 and Mcm6 was strong and almost identical to the activity caused by HsMcm10 and Orc2. The β-galactosidase activity observed from the interaction between HsMcm10 and Mcm2 was only 10% of the activity detected in cells with simian virus 40 large T antigen and p53 protein (as the positive control), suggesting that the interaction was unstable in yeast. We did not detect an interaction between HsMcm10 and Mcm3, 4 or 7 in the two-hybrid system that was reported to occur in budding yeast (25
). However, the inability to detect interactions in the two-hybrid system does not exclude the possibility that these proteins do not interact in vivo.
The interaction between HsMcm10 and Mcm2, 3, 4 and 7 may require post-translational modification, for example.
To obtain further evidence of the association of HsMcm10 with replication proteins in vivo, immunoprecipitation was carried out. The plasmids expressing His–HsMcm10 and HA–Orc2 were transiently transfected into COS-1 cells, and HsMcm10 or Orc2 proteins were precipitated. Since HsMcm10 in the nuclease-resistant structure was insoluble and not suitable for immunoprecipitation, HsMcm10 in the Triton-extractable fraction was used for immunoprecipitation at first. However, both His–HsMcm10 and HA–Orc2 in the soluble fraction non-specifically bound to protein G–Sepharose beads, even in the absence of antibodies. Therefore, we used a minor fraction of His–HsMcm10 and HA–Orc2 which was released from DNaseI treatment. HA–Orc2 was co-precipitated with His–HsMcm10 when anti-HsMcm10 antibody was used (Fig. , lane 7). Because anti-HsMcm10 antibody reacted with human, mouse and monkey Mcm10 proteins (data not shown), a small amount of HA–Orc2 was precipitated by anti-HsMcm10 antibody via endogenous monkey MCM10 (Fig. , lane 6). Furthermore, His–HsMcm10 was co-precipitated with HA–Orc2 when anti-HA antibody was used (Fig. , lane 11), confirming that HsMcm10 and Orc2 associated each other in mammalian cells. On the other hand, the Mcm2–7 complex was not present in the precipitates in both cases, although there was a large amount of Mcm2–7 complex in the DNaseI-extractable fraction (data not shown). However, this result does not exclude the possibility that HsMcm10 interacts with the Mcm2–7 complex in vivo. The interaction with the Mcm2–7 complex may be unstable and/or transient. It is also possible that only a small fraction of Mcm2–7, which are close to the replication origins, for example, interacts with HsMcm10.
Figure 7 Immunoprecipitation assay with His–HsMcm10 and HA–Orc2. Fifty micrograms of the chromatin-bound fraction containing vector control (lanes 5 and 9), HA–Orc2 (lanes 6 and 10), coexpressed His–HsMcm10 and HA–Orc2 (more ...)