Importin-β Mediates Cdc7 Nuclear Import by Binding to the Kinase Insert II Domain, Which Can Be Antagonized by Importin-α*

We investigated the nuclear import mechanism of Cdc7, which is essential for the initiation of DNA replication. Here we report that importin-β binds directly to Cdc7 via the Kinase Insert II domain, promoting its nuclear import. Although both importin-α and -β bind to Cdc7 via the Kinase Insert II domain in a mutually independent manner, the binding affinity of Cdc7 for importin-β is ∼10 times higher than for importin-α at low protein concentrations of an equimolar ratio. Immunodepletion of importin-β, but not importin-α, abrogates Cdc7 nuclear import, and the addition of importin-β to the importin-depleted cytosol restores Cdc7 nuclear import. Furthermore, transduction of anti-importin-β, but not anti-importin-α antibodies, into live cells inhibits Cdc7 nuclear import. Unexpectedly, we found that Cdc7 nuclear import is inhibited by competitive binding of importin-α to Cdc7. Further studies by site-directed mutagenesis suggest that Lys306 and Lys309 within the Kinase Insert II domain are critical for Cdc7 nuclear localization.

The Cdc7 protein consists of 11 putative kinase domains that are highly conserved in all known Cdc7-related mammalian proteins (12,15,16). In addition, there are two kinase inserts: the Kinase Insert II spans from amino acids 203 to 370, and the Kinase Insert III from 440 to 538 (12,(15)(16)(17). A potential short Kinase Insert I also exists at the amino acid residues 75-88. The amino acid sequences of the kinase inserts are the most diverse regions of the entire Cdc7 protein, and thus the inserts are thought to be involved in species-specific regulation and/or interactions with other proteins such as Dbf4 (4).
We examined the regulatory mechanism of human Cdc7 (huCdc7) nuclear transportation using in vitro and in vivo assays. We found that huCdc7 is directly bound and translocated into the nucleus by importin-␤. The binding site is mapped to the Cdc7 Kinase Insert II, and the Lys 306 and Lys 309 residues within this domain are critical for Cdc7 nuclear localization. Most interestingly, importin-␣ can competitively bind to the Cdc7 Kinase Insert II, and can thus effectively inhibit importin-␤-mediated Cdc7 nuclear transportation. Our data also raises the possibility that the binding of Cdc7 to importin-␣ could be involved in the activation or maintenance of the replication checkpoint in response to cell damaging agents such as irradiation and anticancer agents.

EXPERIMENTAL PROCEDURES
Cell Culture and DNA Transfection-HeLa and HEK293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum (HyClone, Logan, UT) and a combination of penicillin (50 units/ml) and streptomycin (50 g/ml). Chinese hamster ovary (CHO) cells were grown in minimal essential medium supplemented with 10% Fetal Clone II (HyClone). Cells grown on a coverglass were transfected with plasmids for 12 h using Lipofectamine PLUS TM reagent as suggested by the supplier (Invitrogen) and as described previously (28).
Protein Expression and Purification-Recombinant GST, GST⅐Cdc7, GST⅐importin-␤, and GST⅐importin-␣2 proteins were expressed in Escherichia coli (BL21), and were purified by affinity chromatography on glutathione-Sepharose 4B as suggested by the manufacturer (Amersham Biosciences). GST pull-down assays were carried out using beads conjugated with either GST (negative control) or GST⅐Cdc7 recombinant proteins. Recombinant His-tagged proteins were purified using BD Talon metal affinity resins according to the manufacturers instruction (BD Biosciences) and as described previously (30,31). Prior to carrying out nuclear import assays, all purified proteins were extensively dialyzed against Nuclear Transport Buffer (NTB: 20 mM HEPES (pH 7.3), 110 mM potassium acetate, 5 mM sodium acetate, 2 mM magnesium acetate, 1 mM EGTA, 2 mM dithiothreitol, and protease inhibitor mixture).
Protein-Protein Interactions in Solution ("Solution Binding Assay")-Purified recombinant His-importin-␣2 or His-importin-␤ proteins were incubated for 30 min at room temperature with the glutathione-Sepharose 4B beads pretreated with GST⅐Cdc7.
Dot-blot Analysis-Purified recombinant proteins were dot blotted onto a polyvinylidene difluoride membrane (Amersham Biosciences) using a CONVERTIBLE TM filtration manifold system (Invitrogen), and the membrane was then air-dried for 40 min at room temperature. The membrane was "blocked" for 2 h with PBS buffer containing 0.2% Tween 20 and 5% nonfat dry milk, followed by incubation with recom-binant importin-␣2 (1 mg/ml) or importin-␤ (1 mg/ml) proteins at 4°C overnight. The membrane was washed three times with PBST for 10 min each, and then proteins bound to the membrane were detected by Western blot analysis. Densitometry was carried out using the Quanti-Scan version 3.0 (Biosoft, Cambridge, United Kingdom).
Immunofluorescence Analysis-Cells grown on a coverglass were fixed with 4% paraformaldehyde for 10 min at room temperature, and were then permeabilized with 0.1% Triton X-100 in PBS for 3 min. After being treated with PBS containing 1% bovine serum albumin for 2 h at room temperature, cells were incubated with primary antibodies for 2 h at room temperature. The cells were washed three times with PBS (5 min each wash), and then incubated with FITC-or rhodamine-conjugated rabbit anti-mouse IgG for 45 min at room temperature. Subsequently, cells were washed with PBS, mounted on a slide glass, and then visualized by fluorescence microscopy (Axiovert 100, Carl Zeiss). Hoechst 33258 (Sigma) was used to visualize the nuclei.
In Vitro Nuclear Import Assay-Nuclear import assays in digitoninpermeabilized cells were performed as described previously (32) with minor modifications. Briefly, cells grown on a coverglass were permeabilized with cold NTB containing 40 g/ml digitonin (Roche) for 5 min. The cells were washed once with cold NTB, and excessive buffer was removed. Subsequently, the cells on a coverslip were treated with a small amount of nuclear transport mixture and incubated at room temperature for 30 min. A complete nuclear transport mixture contained HeLa cell cytosol (8 mg/ml), recombinant GST⅐Cdc7 proteins, and an energy generating system (1 mM ATP, 5 mM creatine phosphate, 20 units of creating phosphokinase/ml) and protease inhibitor mixture) (32). Nuclear import was determined by immunofluorescence using the FITC-conjugated anti-GST antibodies (Santa Cruz).
To deplete importin-␣2 and/or importin-␤, HeLa cell cytosol was incubated with 10 g of anti-importin-␣2 and/or 10 g of anti-importin-␤ antibodies at 4°C for 4 h, which was further incubated for another 1 h with protein-agarose A/G (Santa Cruz). The antibody-importin complexes bound to agarose beads were then removed by centrifugation at 1,000 ϫ g for 2 min. This procedure was repeated three times to completely remove importin-␣2 and/or -␤ proteins. For reconstitution experiments, purified recombinant importin-␣2 (0.125 mg/ml) and/or importin-␤ (0.125 mg/ml) were added to the cytosol depleted of the importins. For the docking assay, recombinant importin-␤ was used as a source of import factor instead of HeLa cell cytosol, as previously described (30,33,34).
Chariot-based Antibody Transduction-HeLa and HEK293 cells grown on a coverglass were transduced with anti-importin-␤ or -importin-␣2 antibodies using a Chariot TM transfection kit (Active Motif, Carlsbad, CA) according to the manufacturer's instruction. Normal goat IgG was used as negative control. At 12 h post-transfection, endogenous Cdc7 was detected by indirect immunostaining with an anti-Cdc7 antibody.

RESULTS
Both Importin-␣2 and -␤ Interact with HuCdc7-We wanted to confirm the nuclear localization of tagged Cdc7 proteins at the onset of this investigation because several tagged recombinant Cdc7 proteins were to be used for this work. As expected, GFP⅐Cdc7 protein was localized in the nucleus (Fig. 1A). To learn how huCdc7 is transported into the nucleus, we analyzed the proteins associated with Cdc7 by a "pulldown" assay using glutathione-Sepharose beads coupled with recombinant GST or GST⅐Cdc7 as described under "Experimental Procedures." PAGE and subsequent Western blot analysis of the proteins bound to the beads showed that importin-␣2, -␤, and MCM2 proteins interacted with Cdc7, whereas PCNA, 14-3-3␤ and importin-␣1 did not ( Fig. 1B; data not shown). Because huCdc7 does not contain a cNLS, the binding of huCdc7 with importin-␣2 was unexpected. The data shown in Fig.  1B, however, could not distinguish whether: 1) importin-␣2 and -␤ bind to Cdc7 in a mutually exclusive manner; or 2) a classical cargo (i.e. Cdc7)-importin-␣/-␤ complex forms. Subsequent studies showed that the former is true (below).
The binding of MCM2 to Cdc7 is consistent with the published data that MCM2 protein is the major physiological substrate of Cdc7 (16,35,36) (Fig. 1B, MCM2). PCNA, a nuclear protein, and 14-3-3␤, a ubiquitous protein, did not co-precipitate with Cdc7, suggesting that these proteins may not interact with Cdc7 at all (Fig. 1B).
HuCdc7 Protein Showed Much Higher Binding Affinity for Importin-␤ than for Importin-␣2 at Low Protein Concentrations-To determine whether the classical importin-␣/-␤ pathway mediates Cdc7 nuclear import, we investigated the interactions between Cdc7 and importin-␣2 or -␤ using an in vitro pull-down assay. Immobilized GST, GST⅐importin-␣2, or GST⅐importin-␤ on beads was incubated with cell extracts prepared from CHO cells expressing GFP, GFP⅐Cdc7, Cdc7, GFP⅐PCNA, or GFP⅐ NLS⅐PCNA. As shown in Fig. 2A, Western blot analysis using anti-GFP or -Cdc7 antibodies suggests that both Cdc7 and GFP⅐Cdc7 similarly interact with GST⅐importin-␣2 and GST⅐importin-␤, which is consistent with the data in Fig. 1B.
GST⅐importin-␣2 specifically interacted with GFP⅐NLS⅐PCNA containing a cNLS (PKKKRKV) derived from SV40 T antigen, but not with GFP⅐PCNA without a cNLS ( Fig. 2A, lane 3). This suggests that the GST and GFP tags of these proteins do not affect the interactions between these proteins. Similarly, Sekimoto et al. (37) previously showed that GST⅐importin-␣ or GST⅐importin-␤ fusion proteins did not affect the normal function of the importin proteins. An auxiliary observation of this experiment is that PCNA, which does not contain a cNLS, may be imported into the nucleus by directly interacting with importin-␤. However, we did not pursue the PCNA nuclear transportation mechanism further.
To analyze the nature of interactions between Cdc7 and importin-␣2/-␤, we incubated purified recombinant GST⅐Cdc7 with purified recombinant His-importin-␣2 or -␤. As shown in Fig. 2C, Cdc7 directly FIGURE 1. Both importin-␣2 and -␤ co-precipitation with huCdc7. A, GFP⅐ huCdc7 is localized in the nucleus. HeLa cells transfected with pEGFP-huCdc7 were fixed with paraformaldehyde at 12 h post-transfection, washed, mounted, and visualized under a fluorescent microscope (GFP panels). Alternatively, HeLa cells fixed as above were permeabilized with Triton X-100, incubated with an anti-Cdc7 antibody, and then further incubated with a secondary antibody conjugated with FITC (FITC panels). B, glutathione-Sepharose beads coupled with GST or GST⅐Cdc7 were incubated with HeLa cell lysates. Proteins bound to the columns were separated by SDS-PAGE, followed by Western blotting with anti-importin-␤, -importin-␣2, -MCM2, -PCNA, and -14-3-3 ␤ antibodies.  4) were immobilized on glutathione-Sepharose 4B beads, which were then incubated with purified recombinant importin-␣2 (lanes 1 and 3) or -␤ (lanes 2 and 4) tagged with His 6 . After washing, the protein bound to the beads were analyzed by SDS-PAGE and Western blotting using an anti-His antibody. Input proteins are shown in lanes 5-8. D, 8 g of purified recombinant GST, GST⅐Cdc7, importin-␣2 or -␤ proteins were dot blotted onto a polyvinylidene difluoride membrane as indicated in each panel. The membrane was then incubated with purified recombinant importin-␤ (left panels) or -␣2 (right panels). The importin-␤ and -␣2 proteins bound were detected by Western blot analysis using anti-importin-␤ (left panels) or -importin-␣2 antibodies (right panels), respectively. E, different amounts of purified recombinant importin-␤ or -␣2 proteins were dot blotted onto a polyvinylidene difluoride membrane, which was then incubated with purified recombinant GST⅐Cdc7 (1 mg/ml). The GST⅐Cdc7 proteins bound were detected by Western blot with an anti-GST antibody. Ratio, the molar ratio of importin-␣2 and -␤ applied onto the membrane, which was 2:1. F, densitometric values of the bands in panel E were determined using QuantiScan version 3.0 (Biosoft).
interacts with either importin-␣2 or -␤ independent of the other importin. The dot-blot analysis using purified proteins also showed a similar result (Fig. 2D). Because the data in Fig. 2C appears to suggest that Cdc7 has higher affinity for importin-␤ than for -␣ (compared lanes 3 and 4), we carried out a more detailed dot blot experiment by applying different amounts and molar ratios of purified recombinant His-importin-␤ and His-importin-␣2 onto each dot on a membrane. The membrane was then incubated with GST⅐Cdc7, followed by Western blot analysis with an anti-GST antibody. As shown in Fig. 2, E and F, the binding affinity of Cdc7 for both importin-␣2 and -␤ is dose-dependent. However, the affinity of Cdc7 for importin-␤ is significantly higher than for importin-␣2 at low protein concentrations. For example, the relative signal intensity of Cdc7 bound to importin-␤ at the concentration of 0.05 g/dot was 83 (arbitrary) units, whereas that of importin-␣2 at the same concentration was 17 units (Fig. 2, E and F). Furthermore, 0.25 g/dot of importin-␣2 was required to reach the relative intensity of 75 units (Fig.  2F). Considering the fact that the same amount of importin-␣2 protein contains twice as many molecules than importin-␤ (i.e. molar ratio is 2:1 at the same amount of importin-␣2 and -␤ proteins), the binding affinity of Cdc7 for importin-␤ is ϳ10-fold higher than for importin-␣2 at low protein concentrations (0.05-0.25 g/blot). Taken together, the data shown in Figs. 1 and 2 suggest that Cdc7 has high binding affinity for importin-␤ in the absence or presence of a low concentration of importin-␣2, raising the possibility that importin-␤ is responsible for Cdc7 nuclear import.
Importin-␤, but Not Importin-␣, Mediates Cdc7 Nuclear Import-To test the hypothesis that importin-␤ is directly responsible for Cdc7 nuclear import, we carried out a series of in vitro nuclear import assays in the presence or absence of importin-␤ as described below. As shown in Fig. 3A, HeLa cell cytosol can effectively transport Cdc7 in the digitonin-permeabilized cells, confirming that the HeLa cell cytosol contains all the necessary factors for Cdc7 nuclear import. To identify cytosolic factors responsible for Cdc7 nuclear transportation, we carried out an in vitro nuclear import assay using HeLa cytosol depleted of both importin-␣2 and -␤. This depleted cytosol did not support Cdc7 nuclear import (Fig. 3B, panels I and II). As expected, Cdc7 nuclear localization was restored when both importin-␣2 and -␤ recombinant proteins were added back to the depleted cytosol (Fig. 3B, panels III and IV). Supple- FIGURE 3. In vitro analysis suggests that importin-␤ (imp), but not importin-␣2, mediates Cdc7 nuclear import. A, purified recombinant GST⅐Cdc7 (0.1 mg/ml) proteins were used as nuclear import substrates in the presence of HeLa cytosol and an energy generating system (Cytosol) or in the presence of transport buffer (Buffer). Merged, the image of Cdc7 was merged with that of the nucleus (Hoechst). B, the Cdc7 nuclear import assay was carried out using HeLa cytosol depleted of both importin-␣2 and -␤ (I and II). The depleted cytosol was supplemented with both importin-␣2 and -␤ (III and IV), importin-␣2 only (V and VI), or importin-␤ only (VII and VIII). C, purified recombinant GST⅐Cdc7 (0.1 mg/ml) was used as a nuclear import substrate in the presence of HeLa cytosol depleted of either importin-␣2 (III and IV) or importin-␤ (VII and VIII). Depleted cytosol was supplemented with purified recombinant importin-␣2 (V and VI) or importin-␤ (IX and X). HeLa cytosol incubated with normal goat IgG was used as a negative control (I and II). D, depletion of importin proteins from HeLa cytosol was confirmed by Western blot analysis. mentation of purified recombinant importin-␣2 to the depleted HeLa cytosol did not restore Cdc7 nuclear import (Fig. 3B, panels V and VI). Importantly, Cdc7 nuclear import was restored when recombinant importin-␤ alone was added back to the depleted cytosol (Fig. 3B, panels VII and VIII), suggesting importin-␤ is directly responsible for the Cdc7 nuclear transportation. To confirm this finding, we depleted either importin-␣2 or -␤ alone from the "complete" HeLa cytosol. Consistent with the data shown in Fig. 3B, the absence or presence of importin-␣2 did not affect Cdc7 nuclear import (Fig. 3C, III-VI). In contrast, the cytosol lacking importin-␤ could not support Cdc7 nuclear transportation, which was restored when purified recombinant importin-␤ was added (Fig. 3C, VII-X). This result further confirms that importin-␤, but not importin-␣2, is responsible for Cdc7 nuclear import, although both can bind to huCdc7. Note that the depletion of the proteins was confirmed by Western blot analysis (Fig. 3D).
To examine whether importin-␤ also mediates Cdc7 nuclear import in vivo, we transduced normal IgG, anti-importin-␣2, or anti-importin-␤ antibodies into HeLa cells using a Chariot TM transduction system (Active Motif). The transduction of nonspecific IgG and anti-importin-␣2 antibodies into HeLa cells did not affect the nuclear localization of endogenous Cdc7 (Fig. 4, I-VIII). In contrast, the transduction of anti-importin-␤ antibodies into HeLa cells effectively inhibited Cdc7 nuclear import (Fig. 4, IX-XII), confirming that importin-␤ is responsible for Cdc7 nuclear import in vivo. We also carried out a similar experiment using HEK293 cells, and found exactly the same result as shown in Fig. 4 (data not shown). Based on the data obtained from in vitro and in vivo assays, we concluded that importin-␤, but not importin-␣, directly mediates Cdc7 nuclear import. This conclusion is further supported by the data obtained by "docking" assay (below).

FIGURE 5. The Cdc7 Kinase Insert II domain is required for interaction of Cdc7 with importin-␤ and its nuclear localization.
A, a schematic representation of wild-type huCdc7 and several deletion mutants used in this study. B, purified recombinant His-importin-␤ (1 g) or His-importin-␣2 (0.5 g) proteins were incubated with glutathione-Sepharose beads coupled with wild-type GST⅐Cdc7 or mutant GST⅐Cdc7 proteins (10 g). Subsequently, proteins bound to the beads were analyzed by PAGE-Western blot analysis using an anti-His antibody (upper panels). Total input wild-type GST⅐Cdc7 and mutant GST⅐Cdc7 detected by anti-GST antibodies are shown in the lower panels. C, the binding ability of Cdc7 for importin-␤ is directly correlated to the docking of Cdc7 to the NPC (arrows). Note that a blown-up image is present as supplemental Fig. 1 to show the difference between proteins in the cytoplasm and docked at the nuclear membrane. D, the GFP⅐huCdc7 K306A,K309A double mutant is localized in cytoplasm. The GFP⅐Cdc7 K306A,K309A or K306A,K309A double mutant was transfected into CHO cells and the subcellular locations of these Cdc7 mutants were analyzed at 15 h post-transfection by fluorescence microscopy. E, the Cdc7 K306A,K309A mutant has significantly lower binding affinity for importin-␤ than wild type. The experiment was carried out similarly as described in Fig. 2A. Imp-␤, immobilized recombinant GST⅐importin-␤; Total, total proteins extracts from CHO cells transfected with GFP⅐Cdc7 wild-type or GFP⅐Cdc7 K306A,K309A mutant; Bound, GFP⅐Cdc7 wild-type or GFP⅐Cdc7 K306A,K309A proteins bound to the immobilized recombinant GST⅐importin-␤.
To determine whether the in vitro binding of Cdc7 to importin-␤ is directly relevant to Cdc7 nuclear transportation, we carried out in vitro docking assays. Digitonin-permeabilized HeLa cells were incubated with GST⅐Cdc7 or GST-tagged Cdc7 deletion mutants in the presence of purified recombinant His-importin-␤ and an energy generating system (note that cytosol was not used in this experiment). Because Ran protein and GTP were not added, the cargo protein (Cdc7) was not expected to be in the nucleus but "docked" at the NPC by this assay system (18,20). Consistent with in vitro binding experiments, wild-type Cdc7 and Cdc7(⌬1-146) efficiently docked at the NPC in the presence of importin-␤, but Cdc7(⌬1-431) and Cdc7(⌬196 -372) did not. The deletion mutants GST⅐Cdc7(⌬279 -574) and GST⅐Cdc7(⌬1-279) also docked at the NPC, albeit at lower levels (Fig. 5C). This result demonstrates that in vitro binding of importin-␤ to Cdc7 is directly relevant to Cdc7 nuclear transportation in an importin-␤-dependent manner.
Further deletion analysis suggested that the Cdc7 segment spanning amino acids 306 -313 is important for Cdc7 nuclear localization (data not shown). Therefore, we generated several single and double point mutations within this region, and then examined the nuclear localization of each mutant in transfected CHO cells. As shown in Fig. 5D, Cdc7 K306A and K309A single mutants were localized to the nucleus. However, the Cdc7 K306A,K309A double point mutant was localized in the cytoplasm, suggesting that the Lys 306 and Lys 309 residues are critical for  ). B, purified recombinant GST⅐Cdc7, His-importin-␤, and His-importin-␣2 (20 g each) were mixed in PBS and loaded onto a gel-filtration column. Fractions collected were subject to SDS-PAGE and Western blot analysis using anti-Cdc7 or -His antibodies. C, signal intensity of each band in panel B was quantified using QuantiScan version 3.0, and the relative values are expressed in a diagram format. D, a schematic diagram of protein complexes in fractions 2-5 and 6 -10 are shown. ␣ and ␤ denote importin-␣2 and -␤, respectively. N and C denote the NH 2 and COOH termini of the protein, respectively. E, importin-␣2 competitively inhibits the importin-␤-mediated Cdc7 nuclear import. The importin-␤-mediated docking of purified recombinant GST⅐Cdc7 protein (0.1 mg/ml) at the NPC was analyzed in the presence of incrementally increased amounts of importin-␣2. huCdc7 nuclear localization. Importantly, the binding affinity of importin-␤ for the Cdc7 K306A,K309A mutant is approximately three times lower than for wild-type Cdc7 (Fig. 5E), again demonstrating that importin-␤ directly bind to Cdc7 and transport it into the nucleus.
The differential binding of Cdc7 to importin-␣2 and -␤ raises the possibility that importin-␣2 may inhibit the importin-␤-mediated Cdc7 nuclear transportation, in contrast to its classical role of promoting transport of the cargo⅐importin-␣/-␤ complex into the nucleus. The interference of Cdc7⅐importin-␤ complex formation by importin-␣2 could be because of a direct interaction of importin-␣2 with Cdc7 through the Kinase Insert II domain as predicted by the data shown in Fig. 5. Alternatively, importin-␣2 could indirectly disrupt (or interfere) the Cdc7⅐importin-␤ complex (formation) by binding to importin-␤. To determine which of these possibilities is correct, we carried out a binding assay using importin-␣2 lacking the IBB domain (⌬IBB; ⌬1-51). Like full-length importin-␣2, a high level of His-importin-␣2⌬IBB (1:10) disrupted the association of GST⅐Cdc7 with His-importin-␤, suggesting that the inhibition of Cdc7 binding to importin-␤ is due to the direct interactions of importin-␣2 to Cdc7 through the Kinase Insert II domain (Figs. 5B and 6A, lane 5).
To further characterize the binding of Cdc7 to importin-␣2 and -␤, a gel-filtration assay was carried out using a Superdex 200 column (Amersham Biosciences). The mixture of purified recombinant proteins (i.e. His-importin-␣2, His-importin-␤, and GST⅐Cdc7) in PBS buffer was loaded onto a column and fractionated by chromatography (Fig. 6B). The protein complexes collected in each fraction were analyzed by Western blotting using either anti-Cdc7 or -His antibodies. Consistent with the data shown in Fig. 6A, the profile of GST⅐Cdc7 and His-importin-␤ proteins co-eluted with fractions 2-5, suggesting that Cdc7 primarily forms a complex with importin-␤, but not with importin-␣2 (Fig.  6, B and C). Fractions 6 -8 contain all three proteins. This raises the possibility that certain complexes may contain all three proteins, Cdc7, importin-␣2, and -␤. Alternatively, these fractions may contain a mixture of complexes, Cdc7⅐importin-␣2 and Cdc7⅐importin-␤. Because importin-␣⌬IBB can effectively compete with importin-␤ for Cdc7, we prefer the latter possibility. This conclusion is also consistent with the expectation that the total molecular weight of protein complexes collected in the fractions would be doublets but not triplets. Fractions 9 and 10 mainly contain Cdc7 and importin-␣2, suggesting that Cdc7 can form a complex with importin-␣2, which is consistent with the data shown in Fig. 6A (lane 6).
We then determined whether importin-␣2 could interfere with importin-␤-mediated Cdc7 docking at the NPC. As shown in Fig. 6E, importin-␣2 can significantly inhibit importin-␤-mediated Cdc7 translocation to the NPC at a molar ratio of 1:5 (␤:␣2) and completely blocked the docking at the 1:10 molar ratio (␤:␣2). This data are consistent with the idea that the interference of the Cdc7⅐importin-␤ complex by importin-␣2 is directly relevant to the importin-␣2-mediated down-regulation of Cdc7 nuclear import.
To further examine the inhibitory effect of importin-␣2 on Cdc7 nuclear transportation, we carried out nuclear import assays using HeLa cell cytosol. GST⅐Cdc7 is imported into the nucleus in the presence of the HeLa cytosol and an energy generating system in the digitoninpermeabilized HeLa (Fig. 7A) and CHO (Fig. 7B) cells. When a low dose of importin-␣2 (0.125 mg/ml) was added to the HeLa cytosol, importin-␣2 did not notably inhibit Cdc7 nuclear import (Fig. 7A, V and VI). When the amount of importin-␣2 was increased to 0.25 mg/ml, Cdc7 nuclear import was almost completely inhibited (Fig. 7A, VII and VIII). Interestingly, the inhibitory effect of importin-␣2 was more pronounced in the CHO cells (compared panels V-VIII in Fig. 7, A and B). The reasons for this difference between HeLa and CHO cells are currently unknown.

DISCUSSION
We were surprised by the finding that huCdc7 could bind to both importin-␣ and -␤, because huCdc7 does not contain any known cNLS. However, further studies by pull-down and dot-blot analyses confirmed our initial finding. Subsequently, in vitro binding assays carried out using different molar ratios of importin-␣ and -␤ led us to conclude that huCdc7, although bound by both importins, has a much higher affinity for importin-␤ than for importin-␣ at low protein concentrations. This observation suggests that a similar preferential binding of Cdc7 to importin-␤ may occur in the cytoplasm, because the low protein concentrations used for our experiments are more relevant to physiological conditions. This may be the reason why Cdc7 can be readily transported into the nucleus despite significant amounts of importin-␣ in the cyto- plasm under normal cellular conditions. Consistent with this hypothesis, we found that the binding of Cdc7 by importin-␤ is directly relevant to Cdc7 nuclear import. We first demonstrated that the HeLa cytosol depleted of both importin-␣ and -␤ did not support Cdc7 nuclear transportation, which was restored by adding purified recombinant importin-␣ and -␤ (Fig. 3B). Importantly, supplementation of importin-␤ alone to the cytosol depleted of both importin-␣ and -␤ could restore Cdc7 nuclear import. In contrast, supplementation of purified importin-␣ to the depleted cytosol did not restore Cdc7 nuclear import. This data strongly suggests that importin-␤, but not -␣, is responsible for Cdc7 nuclear import. We then repeated the experiment using the HeLa cytosol-depleted importin-␣ or -␤ only (Fig. 3C). Depletion of importin-␤, but not importin-␣, inhibited Cdc7 nuclear localization, and supplementing the depleted cytosol with purified recombinant importin-␤ restored Cdc7 nuclear import. This data further confirms that import-␤ is the sole factor required for the initiation of huCdc7 nuclear import, which was also consistent with the results of in vivo antibody transduction and in vitro docking assays (Figs. 4 and 5).
In vitro binding assays using Cdc7 mutants suggest that the Cdc7 Kinase Insert II (amino acids 203-370) contains binding site(s) for both importin-␣ and -␤ (Fig. 5). The mutant Cdc7(⌬1-146) and wild-type Cdc7 showed a comparable binding affinity for importin-␤, whereas the mutant Cdc7(⌬196 -372) did not bind to importin-␤ at all, suggesting that the Kinase Insert II domain is necessary and sufficient for stable binding to importin-␣ and -␤. Interestingly, both Cdc7(⌬1-197) and Cdc7(⌬297-574) could bind to importin-␤, albeit at lower affinity, but not to importin-␣ (Fig. 5). This data raises the following possibilities: 1) there are two separate subregions for importin-␤ binding, but only one site for importin-␣ within the Kinase Insert II domain; or 2) there is only one binding site for both importin-␣ and -␤ within the entire Kinase Insert II region. If the former is true, one of the two importin-␤ binding sites would be at amino acids 196 -279 and the other at 280 -372. Judging from the signal intensities of protein bands in Fig. 5B, importin-␤ may bind to these two subregions with almost equal probability. If there is only one binding site within the Kinase Insert II domain, it may span both up-and downstream regions of the amino acid residue 279/280 junction. In this case, however, each of the 196 -279 and 280 -372 protein segments alone may have sufficient binding affinity for importin-␤, although both protein segments may be required for more stable binding to importin-␤. If the one-binding site scenario is correct, more than one binding motif may be present within the Kinase Insert II domain. At this point, we cannot rule out either possibility, although we prefer the one-binding site model because a double point mutation within a small protein segment could inhibit Cdc7 nuclear localization as discussed below.
Considering that the Kinase Insert II comprises ϳ170 amino acids, the necessity of this large stretch of protein segment for importin binding is quite different from the cNLS-mediated nuclear transportation. All known proteins transported into the nucleus by importin-␤ appear to require a large stretch of protein segment (23,27,38). These authors suggest that a large binding region is required for importin-␤-mediated protein nuclear transportation because the binding of importin-␤ to a cargo protein is by structural interactions. We, however, found that the Lys 306 and Lys 309 residues are essential for huCdc7 nuclear localization as the Cdc7 K306A,K309A double point mutant was not localized in the nucleus. This data suggest that the ternary structure of the Cdc7 K306A,K309A mutant has significantly altered so that it can no longer efficiently interact with importin-␤. Alternatively, importin-␤-mediated protein nuclear localization may also require a specific amino acid sequence (in addition to the structural requirement). In this regard, it may be worth further investigation of the potential role of XKXXKXX (which is from 305 VKLMKQA of huCdc7) in an importin-␤-mediated protein nuclear localization when more proteins directly transported by importin-␤ are identified and characterized.
The most surprising finding in this study is that importin-␣ can competitively impede the complex formation of Cdc7 with importin-␤, which is directly relevant to the inhibition of Cdc7 nuclear localization (Figs. 6 and 7). Because importin-␣ inhibits docking of Cdc7 at the NPC in vitro, the inhibition of the Cdc7 nuclear transportation by importin-␣ is likely at the protein-protein interaction step in the cytoplasm (as opposed to transportation through the NPC). Furthermore, this inhibition of Cdc7 nuclear import by importin-␣ is independent of importin-␣ binding to importin-␤ (Fig. 6).
Our data from dot blot and gel-filtration assays suggest that the affinity between Cdc7 and importin-␤ is strong when equimolar concentrations of importin-␣ and -␤ are present (a in Fig. 7C). Therefore, the formation of the Cdc7⅐importin-␤ complex may readily occur under normal cytoplasmic conditions, resulting in effective Cdc7 nuclear import. For Cdc7 nuclear localization, the two lysine residues at 306 and 309 are critically important (Fig. 5D), perhaps by promoting stable binding between Cdc7 and importin-␤ as the Cdc7 K306A,K309A mutant shows significantly lower binding affinity for importin-␤ (Fig. 5E). The binding between Cdc7 and importin-␣ (b in Fig. 7C) would not normally occur under the equimolar concentration conditions (Fig. 6A). Therefore, importin-␣ cannot impede Cdc7 nuclear transportation under the normal cytoplasmic conditions. This inability of importin-␣ binding to Cdc7 may also reduce the association potential of importin-␣ with importin-␤ (c of the Fig. 7C), because the IBB of importin-␣ has a selfinhibitory function (39,40). If, however, a cargo protein contains a cNLS, this binding dynamic can change dramatically. This is because importin-␣ can bind to the cargo protein, by which the binding potential of importin-␣ to -␤ can also increase through the release of the IBB domain from its cis-inhibitory conformation (39,40).
It is well known that cells can rapidly activate a replication checkpoint in response to cell damaging agents such as irradiation and anticancer agents (41,42). Interestingly, Miyamoto et al. (43) have recently found that cellular stress caused by irradiation can induce the nuclear accumulation of importin-␣ and inhibit a conventional nuclear import. In these contexts, our model predicts an interesting possibility that importin-␣ may be able to bind to Cdc7 in the nucleus (where no importin-␤ is present), and is thus involved in the down-regulation of replication initiation when a cell faces a crisis. Because an effective checkpoint operation can increase cell survival, the Cdc7 binding to importin-␣ in the nucleus in response to cell damage may have significant implication to decreases in the efficacy of anti-cancer therapies.