Introduction
Biochemical analysis of crude extracts of Skp2-/- cells revealed the presence in the cytosolic fraction of a Skp2-independent E3 activity that mediates the ubiquitination of p27Kip1 (ref. 7). This ubiquitination was not dependent on the phosphorylation of p27Kip1 on Thr 187, which is a prerequisite for Skp2-mediated ubiquitination. To isolate the cytoplasmic E3 responsible for p27Kip1 degradation, we purified the activity that mediates polyubiquitination of p27Kip1—in conjunction with a ubiquitin-activating enzyme (E1, Uba1) and a ubiquitin-conjugating enzyme (E2, UbcH5A)—from a rabbit reticulocyte lysate by chromatography on six different columns (Fig. 1a). The in vitro ubiquitination assay detected substantial E3 activity in fractions 14 to 18 of a Superose 6 gel-filtration column (Fig. 1b). SDS–polyacrylamide gel electrophoresis (PAGE) and Coomassie-blue staining of these fractions revealed two proteins that appeared to copurify with the activity. These molecules—of relative molecular mass (Mr) 140,000 (140K) and 50,000 (50K)—were designated KPC1 (Kip1 ubiquitination-promoting complex 1) and KPC2, respectively. Subsequent chromatography on a Mini Q anion exchange column yielded marked E3 activity in fractions 18 to 27; again, both KPC1 and KPC2 copurified with the activity (Fig. 1c). We determined the amino-acid sequences of KPC1 and KPC2 by mass spectrometric analysis (Fig. 1d). A database search revealed that the amino-acid sequences of the rabbit KPC1-derived peptides matched only those of human expressed sequence tags (ESTs), whereas rabbit KPC2 seems to be the orthologue of human glioblastoma cell differentiation-related 1 (GBDR1)12, whose function has been unknown.
Figure 1: Purification of KPC.
(a) Protocol for the purification of KPC. (b, c) Fractions from Superose 6 gel-filtration chromatography (b) and Mini Q anion exchange chromatography (c) were assayed for their ability to mediate p27Kip1 polyubiquitination, and also subjected to SDS–PAGE and Coomassie-blue staining. (d) Amino-acid sequences of peptides derived from rabbit KPC1 and KPC2. (e) Schematic representation of the structural organization of KPC1 and KPC2. Ub, ubiquitin.
Full size image (86 KB)The deduced amino-acid sequence of human KPC1 (see Supplementary Information, Fig. S1) suggests that this protein, which contains a RING-finger domain near its carboxy terminus, functions as a catalytic subunit (Fig. 1e). KPC1 also contains a SPRY domain, which was originally identified as a motif of unknown function that is present in three copies in the mammalian ryanodine receptor13. KPC2 contains an amino-terminal ubiquitin-like (UBL) domain and two C-terminal ubiquitin-associated (UBA) domains (Fig. 1e); its overall structure is similar to that of Rad23, which both binds to multi-ubiquitin chains attached to proteins as well as interacts with the proteasome14, 15. KPC2 also interacts with the proteasome and polyubiquitinated proteins through its UBL and UBA domains, respectively (T.H., T.K. & K.I.N., unpublished observations). These structural characteristics suggest that KPC is an authentic RING-finger type E3 complex.
To confirm that KPC1 and KPC2 are sufficient for polyubiquitination of p27Kip1, we generated the corresponding recombinant hexahistidine (His6)-tagged human proteins in insect cells and purified them with a Ni2+-based resin (Fig. 2a). Immunoprecipitation and immunoblot analysis of lysates of insect cells expressing one or both human proteins revealed that KPC1 formed a complex with KPC2 (Fig. 2b). KPC1(
R), a deletion mutant of KPC1 lacking the RING-finger domain, also interacted with KPC2, suggesting that this domain is dispensable for binding to KPC2. The recombinant KPC1 and KPC2 proteins exhibited pronounced E3 activity for p27Kip1, but not p21Cip1 or p57Kip2, in the presence of E1 (Uba1) and E2 (UbcH5A, 100 ng 10 
l-1) enzymes (Fig. 2c), suggesting that KPC1 and KPC2 are sufficient for E3 activity towards p27Kip1. At a lower concentration of UbcH5A (30 ng 10 l-1), monoubiquitinated forms of p27Kip1 were virtually undetectable in the absence of KPC1–KPC2 (Fig. 2d); the addition of KPC then markedly increased the production of both mono- and polyubiquitinated forms of p27Kip1. Replacement of the glutathione S-transferase (GST)–ubiquitin fusion protein in the reaction mixture with ubiquitin resulted in a shift in the position of bands of high molecular mass (Fig. 2e), suggesting that the bands with lower electrophoretic mobilities corresponded to polyubiquitinated p27Kip1, not to aggregates of p27Kip1 or to nonspecific proteins that cross-reacted with the antibodies to (anti-) p27Kip1. Use of the K48R mutant of ubiquitin in the assay resulted in a substantial decrease in the generation of polyubiquitinated forms of p27Kip1.
Figure 2: Polyubiquitination of p27Kip1 by KPC.
(a) Purified recombinant human His6–Flag–KPC1 and His6–HSV–KPC2 expressed in insect cells were subjected to SDS–PAGE and staining with Coomassie blue. (b) Interaction between KPC1 and KPC2 in insect cells. Lysates of cells expressing the indicated combinations of KPC2 and either wild-type (WT) KPC1 or the KPC1(R) mutant were subjected to immunoprecipitation (IP) with anti-Flag, and the resulting precipitates were subjected to immunoblot analysis (IB) with anti-Flag or anti-HSV. (c–e) Polyubiquitination of p27Kip1 mediated by recombinant KPC. The recombinant KPC1–KPC2 complex was assayed for the ability to mediate polyubiquitination of p27Kip1, p57Kip2, or p21Cip1 in the presence of the indicated combinations of Uba1, 100 ng (c) or 30 ng (d, e) of UbcH5A, GST–ubiquitin (c, d) or ubiquitin (e), and ATP. (f) Requirement for the RING-finger domain in KPC1 function. KPC1 derivatives (WT or R) were assayed for the ability to mediate polyubiquitination of p27Kip1 in the absence or presence of KPC2. (g) E2 preference of KPC. The indicated E2 enzymes were tested for their ability to support the polyubiquitination of p27Kip1 by recombinant KPC1–KPC2.
KPC1 was sufficient for polyubiquitination of p27Kip1 in the absence of KPC2; indeed, the extent of polyubiquitination was greater in the absence of KPC2 than in its presence (Fig. 2f). A similar effect has been described for Rad23, which inhibits multi-ubiquitin chain assembly15. KPC1(R) did not support the polyubiquitination of p27Kip1, indicating that the RING-finger domain is indispensable for E3 activity. We examined the preference of KPC for seven different E2 enzymes in the presence of an E1 (Fig. 2g). Polyubiquitination activity of KPC was apparent with only Ubc4 or UbcH5A.
To investigate whether KPC contributes to the degradation of p27Kip1in vivo, we examined the potential physiological interaction between the endogenous proteins in NIH 3T3 cells. Cells stably expressing a short hairpin RNA (shRNA) specific for KPC1 mRNA, or an shRNA specific for enhanced green fluorescent protein (EGFP) mRNA as a control, were synchronized in G0 phase by serum deprivation, stimulated to re-enter G1 phase of the cell cycle by incubation with complete medium, and then incubated further in the presence of the proteasome inhibitor MG132 to facilitate the potential interaction of p27Kip1 with KPC. Cell lysates were then subjected to immunoprecipitation with anti-KPC1 or anti-KPC2 or with control immunoglobulin G (IgG). Immunoblot analysis of the resulting precipitates revealed that p27Kip1 and KPC2 were coprecipitated by the anti-KPC1, and that p27Kip1 and KPC1 were coprecipitated by the anti-KPC2 (Fig. 3a). Depletion of KPC1 by RNA interference (RNAi) resulted in a decrease in the amount of p27Kip1 that was coprecipitated by either anti-KPC1 or anti-KPC2, and this decrease was proportional to the reduction in the cellular abundance of KPC1. Direct interaction between recombinant KPC and p27Kip1in vitro was also demonstrated by immunoprecipitation and immunoblot analysis (Fig. 3b).
Figure 3: Interaction between KPC and p27Kip1 and subcellular localization of KPC.
(a) Interaction between endogenous p27Kip1 and KPC1 or KPC2 in NIH 3T3 cells. Lysates of cells expressing KPC1-1 or EGFP shRNAs were subjected to immunoprecipitation (IP), followed by immunoblot analysis (IB) with anti-KPC1, anti-KPC2, or anti-p27Kip1. (b) Association of p27Kip1 with the KPC1–KPC2 complex in vitro. Recombinant p27Kip1 and KPC1–KPC2 complex were mixed, immunoprecipitated with anti-p27Kip1, and subjected to immunoblot analysis. (c) Immunofluorescence analysis of the subcellular localization of KPC1 and KPC2. NIH 3T3 cells expressing His6–Flag–KPC1 and HA–KPC2 were stained with anti-Flag (red), anti-HA (green) and Hoechst 33258 (blue). Scale bars, 25 m. (d) Subcellular fractionation of endogenous KPC1 and KPC2. Cytosolic (C) or nuclear (N) fractions prepared from NIH 3T3 cells either in asynchronous culture (AS) or synchronized at G0, G1, or S phase were subjected to immunoblot analysis with anti-KPC1, anti-KPC2, anti-
-tubulin (cytosolic marker), or anti-lamin B1 (nuclear marker). Asterisk, non-specific bands.
The subcellular localization of KPC was examined by immunofluorescence analysis of NIH 3T3 cells infected with retroviral vectors encoding Flag epitope-tagged KPC1 and haemagglutinin epitope (HA)-tagged KPC2. The two proteins exhibited almost identical distributions in the cytoplasm and were absent from the nucleus (Fig. 3c). We also examined the subcellular localization of endogenous KPC by immunoblot analysis of cytosolic and nuclear fractions prepared from NIH 3T3 cells. Both KPC1 and KPC2 were detected only in the cytosolic fraction irrespective of the cell-cycle stage (Fig. 3d). The cytoplasmic localization of KPC is consistent with the facts that p27Kip1 is exported from the nucleus to the cytoplasm at the G0–G1 transition9, 11, and that polyubiquitination activity for p27Kip1 was shown to be present in the cytosolic fraction of crude extracts of Skp2-deficient cells7. These results suggest that, after its translocation from the nucleus to the cytoplasm, p27Kip1 is targeted by KPC and ubiquitinated.
We next determined the effect of overexpression of KPC on p27Kip1 degradation at the G0–G1 transition in NIH 3T3 cells. Expression of wild-type KPC1 and KPC2 markedly promoted the degradation of p27Kip1 compared with that apparent in control cells, whereas coexpression of the KPC1(R) mutant and KPC2 delayed p27Kip1 degradation (Fig. 4a), suggesting that KPC1(R) functions as a dominant-negative mutant. Given that the nuclear export of p27Kip1 mediated by CRM1 seems to be required for its KPC-mediated degradation in the cytoplasm, we examined the effect of treating cells with leptomycin B, a specific inhibitor of CRM1-dependent nuclear export, at a concentration (5 ng ml-1) sufficient to inhibit p27Kip1 translocation from the nucleus to the cytoplasm9, 16. Neither the promotion nor the inhibition of p27Kip1 degradation by wild-type and mutant KPC1, respectively, was evident in cells exposed to leptomycin B (Fig. 4b). The nuclear export of p27Kip1 by CRM1 thus indeed seems to be necessary for KPC-mediated proteolysis, although we cannot eliminate the possibility that another protein that regulates KPC is affected by leptomycin B.
Figure 4: Effect of KPC overexpression or depletion on p27Kip1 degradation.
(a) NIH 3T3 cells expressing KPC2 and either wild-type (WT) or mutant (R) KPC1 were synchronized in G0 phase and then stimulated for the indicated times. Cell lysates were subjected to immunoblot analysis (IB) with anti-p27Kip1 or anti-GSK-3
. (b) NIH 3T3 cells were treated and analysed as in a in the presence of leptomycin B. (c) Depletion of KPC1 by RNAi. Lysates prepared from NIH 3T3 cells expressing KPC1 or EGFP shRNAs were subjected to immunoblot analysis with anti-KPC1 or anti-KPC2. Asterisk, non-specific bands. (d) NIH 3T3 cells expressing KPC1 or EGFP shRNAs were treated and analysed as in a. (e) NIH 3T3 cells were infected with retroviruses encoding KPC1-1 or EGFP shRNAs for 4 h and were then treated and analysed as in a.
We applied RNAi to determine the effects of depletion of endogenous KPC1 from NIH 3T3 cells on p27Kip1 degradation and cell-cycle progression. Cells infected with a retroviral vector encoding either KPC1-1 or KPC1-2 shRNAs specific for KPC1 mRNA exhibited a >80% decrease in the abundance of KPC1 compared with that apparent in cells infected with a control vector for EGFP shRNA (Fig. 4c). Infection of these shRNA vectors did not induce the interferon response (see Supplementary Information, Fig. S2a). Expression of EGFP or KPC1-1-M1 (mismatched) shRNA did not affect KPC1 expression, p27Kip1 degradation or cell-cycle progression (see Supplementary Information, Fig. S2b–d). The reduction in the amount of p27Kip1 apparent in response to serum restimulation of serum-deprived cells was markedly inhibited by expression of KPC1-1 or KPC1-2 shRNAs (Fig. 4d). To eliminate concerns associated with down-regulation of KPC1 prior to serum deprivation, we infected NIH 3T3 cells with retroviruses encoding KPC1-1 or EGFP shRNAs for 4 h and then synchronized the cells in G0 phase by serum withdrawal. Under these conditions, p27Kip1 degradation was again inhibited by expression of KPC1-1 shRNA (Fig. 4e), indicating that KPC1 mediates p27Kip1 degradation only at the G0–G1 transition.
We next examined the effects on p27Kip1 abundance or cell-cycle progression of depletion of KPC1 in wild-type, Skp2-/-, or Skp2-/-p27-/- mouse embryonic fibroblasts (MEFs). Wild-type, Skp2-/-, or Skp2-/-p27-/- MEFs infected with a retroviral vector encoding KPC1-1 shRNA similarly exhibited a >80% decrease in the abundance of KPC1 compared with that apparent in cells infected with a control vector (Fig. 5a). As in NIH 3T3 cells (Fig. 4d), depletion of KPC1 in wild-type MEFs resulted in inhibition of p27Kip1 degradation at the G0–G1 transition but did not substantially affect progression of the cell cycle from G0 to S phase, as monitored by immunoblot analysis of cyclin A or by flow cytometry (Fig. 5b). Deficiency of Skp2 alone also did not result in a delay in cell-cycle progression from G0 to S phase (Fig. 5c). However, depletion of KPC1 in Skp2-/- MEFs resulted in both a pronounced impairment of p27Kip1 degradation and a delay in cell-cycle progression. This delay seemed to be caused by the accumulation of p27Kip1 that resulted from the deficiency of both KPC1 and Skp2, given that cell-cycle progression was not delayed by depletion of KPC1 in Skp2-/-p27-/- MEFs (Fig. 5d). These data suggest that, even in the absence of KPC, p27Kip1 is exported from the nucleus and is unable to inhibit cyclin–CDK complexes in the nucleus during G1 phase, and that Skp2 is then necessary to degrade accumulating p27Kip1 that is shuttled back into the nucleus from the cytoplasm during S phase.
Figure 5: Effects of KPC1 depletion on p27Kip1 degradation and cell-cycle progression.
(a) Lysates of wild-type (WT), Skp2-/-, or Skp2-/-p27-/- MEFs expressing KPC1-1 or EGFP shRNAs were subjected to immunoblot analysis (IB) with anti-KPC1 or anti-KPC2. Asterisk, non-specific bands. (b–d) Wild-type (b), Skp2-/- (c), or Skp2-/-p27-/- (d) MEFs expressing KPC1-1 or EGFP shRNAs were analysed as in Fig. 4a. The percentage of cells in S phase is shown. (e) Wild-type or Skp2-/- MEFs were synchronized in G0 phase. Eight hours (G1 phase) or 17 h (S phase) after release, the cells were labelled with [35S]methionine and [35S]cysteine for 1 h. After culture in complete medium for the indicated chase periods, the cells were lysed and subjected to immunoprecipitation with anti-p27Kip1, and the resulting precipitates were analysed by SDS–PAGE and autoradiography. The intensities of the p27Kip1 bands were quantified by scanning densitometry.
Full size image (76 KB)We next investigated the effects of a deficiency of KPC1, Skp2, or both KPC1 and Skp2 on p27Kip1 stability in G1 phase or S phase by pulse-chase analysis (Fig. 5e). At G1 phase, the half-life of p27Kip1 was similarly short (<4 h) in wild-type and Skp2-/- MEFs expressing EGFP shRNA, whereas that of p27Kip1 in wild-type or Skp2-/- MEFs expressing the KPC1-1 shRNA was markedly increased. In contrast, at S phase, p27Kip1 in wild-type MEFs expressing either KPC1-1 or EGFP shRNAs was unstable (half-life, <2 h), whereas p27Kip1 in Skp2-/- MEFs expressing KPC1-1 or EGFP shRNAs was stable. These results suggest that p27Kip1 degradation in G1 phase is distinct from that in S phase, and that these two events are mediated by KPC1 and Skp2, respectively.
During the course of this study, we observed that both the rate of increase in cell number (see Supplementary Information, Fig. S3a) and the proportion of cells in S phase (data not shown) were lower for NIH 3T3 cells expressing the KPC1-1 or KPC1-2 shRNAs than in control cells. These findings suggested that KPC controls not only the progression from G0 to S phase, but also the growth of continuously cycling cells during long-term culture. However, the growth rate as well as the proportion of cells in S phase for p27-/- MEFs expressing the KPC1-1 shRNA did not differ from those apparent for p27-/- MEFs expressing EGFP shRNA (see Supplementary Information, Fig. S3b, c; data not shown). The reduced growth rate of KPC1-deficient cells may thus be attributable to the defect in the degradation of p27Kip1. Although the reason for the discrepancy between the results obtained from cells in short-term culture (within 24 h of release from G0 arrest) (Fig. 5b) and in long-term culture (>48 h) (see Supplementary Information, Fig. S3a, c) remains to be determined, we propose that p27Kip1 gradually accumulates in cells expressing KPC1 shRNA and eventually exceeds the threshold required for inhibition of cyclin–CDK complexes during long-term culture.
Finally, we compared the stabilities in proliferating NIH 3T3 cells of wild-type p27Kip1 and a form of the protein with a mutated nuclear localization sequence (NLS), the latter of which localizes to the cytoplasm (see Supplementary Information, Fig. S3d). Pulse-chase analysis revealed that wild-type p27Kip1 was unstable (half-life, <2 h) in the cycling cells regardless of the absence or presence of KPC1 (see Supplementary Information, Fig. S3e), similar to the results showing that KPC did not affect p27Kip1 stability in S phase (Fig. 5e). The NLS mutant of p27Kip1 was unexpectedly stable and not affected by RNAi-mediated depletion of KPC1, suggesting that prior localization of p27Kip1 in the nucleus is required for its degradation mediated by KPC in the cytoplasm. Overexpression of KPC1 and KPC2 partially rescued this phenotype. These data thus suggest that KPC regulates cell growth by mediating the cytoplasmic degradation of p27Kip1 that has been exported from the nucleus.
KPC seems to be constitutively active throughout the cell cycle, including the G0 phase7. The nuclear export of p27Kip1 seems important for the rapid clearance of this protein that accompanies the entry of cells into the cell cycle from the resting stage. We therefore propose that translocation and proteolysis of p27Kip1 are coupled in order to remove this CKI promptly and irreversibly. Given that p27Kip1 contains an NLS, p27Kip1 molecules that have exited the nucleus may shuttle back into this organelle. The function of KPC might thus be to terminate the export–import cycle of p27Kip1 (see Supplementary Information, Fig. S4). This cycle also seems to be broken by signalling by means of the phosphatidylinositol 3-kinase and Akt pathway. The NLS of human p27Kip1 contains a consensus site for phosphorylation by Akt at Thr 157, and p27Kip1 phosphorylation by Akt impairs its nuclear import17, 18, 19. The biological significance of phosphorylation of this residue remains to be determined, however, given that, despite the high overall sequence homology between human and mouse p27Kip1, Thr 157 is not conserved in the mouse protein.
The degradation of p27Kip1 is regulated by two distinct mechanisms: translocation-coupled cytoplasmic ubiquitination by KPC at G1 phase and nuclear ubiquitination by Skp2 at S and G2 phases. KPC-dependent degradation of p27Kip1 seems to occur transiently, even though KPC activity is present in the cytoplasm throughout the cell cycle. These observations suggest that the nuclear export of p27Kip1 takes place only at G1 phase. This system may therefore ensure the rapid and efficient elimination of p27Kip1 at G1 phase. Given that mice deficient in Skp2 complete embryonic development6, KPC and Skp2 may be functionally redundant. We have recently shown that the CKIs p21Cip1 and p57Kip2 are also targets of Skp2-mediated ubiquitination20, 21. In contrast, our in vitro ubiquitination data suggest that KPC is specific for p27Kip1. However, p57Kip2 is also rapidly degraded at the G0–G1 transition and this event does not require Skp221. The possible role of KPC in p57Kip2 degradation thus warrants further investigation.
Methods
In vitro assay of ubiquitination activity.
For monitoring of the purification of the E3 for p27Kip1, portions of column fractions were mixed with 50 ng of Uba1, 100 ng (unless indicated as 30 ng) of UbcH5A, 3 g of GST–ubiquitin, and 50 ng of recombinant p27Kip1 in a final volume of 10 l containing 40 mM Hepes-NaOH (pH 7.9), 60 mM potassium acetate, 2 mM dithiothreitol, 5 mM MgCl2, 0.5 mM EDTA, 10% glycerol and 1.5 mM ATP. Reaction mixtures were incubated for 30 min at 26 °C.
Purification of KPC1 and KPC2 from rabbit reticulocyte lysate.
Rabbit reticulocyte lysate was prepared as described22. The lysate (
20 g of protein) was incubated for 45 min at 4 °C with 700 ml of DE52 resin (Whatman, Maidstone, UK) that had been equilibrated with buffer A (50 mM Tris-HCl (pH 7.4), 0.1 mM dithiothreitol, 10% glycerol). The slurry was filtered at a rate of 400 ml h-1 into a 10-cm-diameter column and then washed at the same flow rate with buffer A. Proteins were eluted from the column stepwise at a rate of 700 ml h-1 with buffer A containing 100 or 300 mM KCl, and 140-ml fractions were collected. Fractions containing E3 activity for p27Kip1 were concentrated by (NH4)2SO4 (60%) precipitation, resuspended in 10 ml of buffer A, and dialysed against buffer A until the conductivity of the dialysate was similar to that of buffer A containing 150 mM KCl. The dialysate was centrifuged for 15 min at 12,000g, and the resulting supernatant was applied at a rate of 10 ml h-1 to a Superdex 200 gel-filtration column (2.6 by 60 cm; Amersham pharmacia biotech, Piscataway, NJ) that had been equilibrated with buffer A containing 150 mM KCl. Proteins were eluted at 10 ml h-1, and 10-ml fractions were collected. Fractions containing E3 activity for p27Kip1, which eluted in a discrete peak corresponding to an apparent native molecular size of between 70K and 440K, were pooled, adjusted to 5 mM potassium phosphate, and centrifuged for 20 min at 60,000g. The resulting supernatant was applied to a ceramic hydroxyapatite type 1 column (1 by 10 cm; Bio-Rad, Hercules, CA) that had been equilibrated with buffer P (5 mM potassium phosphate (pH 7.4), 50 mM NaCl, 0.1 mM dithiothreitol, 10% glycerol). Elution was performed at a rate of 2 ml min-1 with a 60-ml linear gradient of 5 to 300 mM potassium phosphate in buffer P, and 2-ml fractions were collected. Fractions containing E3 activity for p27Kip1, which eluted between 20 and 50 mM potassium phosphate, were pooled, diluted with buffer A to a conductivity similar to that of buffer A containing 40 mM KCl, and centrifuged for 20 min at 60,000g. The resulting supernatant was applied to a Mono Q column (1 by 10 cm; Amersham pharmacia biotech) that had been equilibrated with buffer A containing 40 mM KCl. Proteins were eluted at 1 ml min-1 with a 30-ml linear gradient of 40 to 350 mM KCl in buffer A, and 1-ml fractions were collected. Those containing E3 activity for p27Kip1, which eluted between 100 and 160 mM KCl, were pooled and applied to a Superose 6 gel-filtration column (1 by 30 cm; Amersham pharmacia biotech) that had been equilibrated with buffer A containing 150 mM KCl. Elution was performed at 0.5 ml min-1, and 1-ml fractions were collected. Fractions containing E3 activity for p27Kip1, which eluted as a discrete species with an apparent native molecular mass of between 70K and 440K, were pooled, diluted with buffer A to a conductivity similar to that of buffer A containing 80 mM KCl, and centrifuged for 20 min at 60,000g. The resulting supernatant was applied to a Mini Q column (0.46 by 5 cm; Amersham pharmacia biotech) that had been equilibrated with buffer A containing 80 mM KCl. Bound material was eluted at 150 l min-1 with a 4.5-ml linear gradient of 80 to 350 mM KCl in buffer A, and 150-l fractions were collected. E3 activity for p27Kip1 eluted between 130 and 180 mM KCl.
Cloning of KPC1 and KPC2 cDNAs.
Purified rabbit KPC1 and KPC2 were separated by SDS–PAGE on a 10% gel and stained with Coomassie blue. The stained bands were excised from the gel and the proteins therein were subjected to in-gel reduction, S-carboxyamidomethylation and digestion with trypsin. The resulting peptides were either analysed by LCQ ion-trap mass spectrometry or subjected to chromatography on a RPC C2/C18 column (2.1 by 100 mm; Amersham pharmacia biotech). Peptides isolated from the latter column were then subjected to sequencing by automated Edman degradation. IMAGE Consortium cDNA clones (GenBank accession numbers BE885419 and BE885914) encoding human KPC1 were obtained from Research Genetics. The cDNA encoding human KPC2, previously identified as GBDR1 (ref. 12), was amplified by the polymerase chain reaction from human liver cDNA (Clontech, Palo Alto, CA).
Antibodies.
Polyclonal anti-KPC1 or anti-KPC2 were generated in rabbits by standard procedures with fragments of recombinant mouse KPC1 (KPC1-N, residues 1–302; KPC1-C, residues 762–1073) or KPC2 (residues 1–409) as antigens. Monoclonal anti-p27Kip1, to GSK-3 and to HSP90 were obtained from BD Biosciences, San Jose, CA. A monoclonal antibody (M2) to Flag was from Sigma, St Louis, MO. A monoclonal antibody to HSV was from Novagen, Madison, WI. A monoclonal antibody (9E10) to Myc was from Roche Molecular Biochemicals (Basel, Switzerland). Monoclonal anti-Skp2 (2B12), to lamin B1 (L-5) and to -tubulin (TU01), were from Zymed, San Francisco, CA. Rabbit polyclonal anti-HA (Y11), to cyclin A (H432) and to p27Kip1 (C19), were from Santa Cruz Biotechnology, Santa Cruz, CA.
Baculovirus expression system.
Complementary DNAs encoding human wild-type KPC1 or the mutant KPC1(R), tagged at their N termini with His6 and Flag, and a cDNA for human KPC2 containing N-terminal His6 and HSV tags were subcloned into pBacPAK9. Recombinant baculoviruses were generated with the BacPAK baculovirus expression system (Clontech). Recombinant proteins were expressed in Sf21 cells and purified as described23, 24.
Retrovirus expression system.
Complementary DNAs encoding human wild-type KPC1 or KPC1(R) containing His6 and Flag tags at their N termini, a cDNA for human KPC2 containing a C-terminal HA tag, and cDNAs for human wild-type p27Kip1 or the NLS mutant thereof (which contains the mutations R152A, K153A and R154A) with N-terminal Myc tags were subcloned into pMX-puro (provided by T. Kitamura, University of Tokyo)25, and the resulting vectors were used to transfect Plat E cells and thereby generate recombinant retroviruses. NIH 3T3 cells were infected with the recombinant retroviruses and selected in medium containing puromycin (10 g ml-1). Cells stably expressing the recombinant proteins were pooled for experiments.
Isolation of MEFs.
Primary MEFs were isolated from 13.5-day-postcoitum wild-type, Skp2-/-, p27-/-, or Skp2-/-p27-/- embryos and cultured as described previously8, 26. Only nonsenescent MEFs (passages 1 to 3) were used for experiments.
RNAi.
The pMX-puro II vector was constructed by deletion of the U3 portion of the 3' long terminal repeat of pMX-puro. The mouse U6 gene promoter followed by DNA corresponding to an shRNA sequence was subcloned into the NotI and XhoI sites of pMX-puro II, yielding pMX-puro II-U6/siRNA. The DNA for the shRNA encoded a 21-nucleotide hairpin sequence specific to the mRNA target, with a loop sequence (-TTCAAGAGA-) separating the two complementary domains, and contained a tract of five T nucleotides to terminate transcription. The hairpin sequences specific for mouse KPC1 (KPC1-1, KPC1-2) and for EGFP (Clontech) mRNAs corresponded to nucleotides 1852 to 1872 (KPC1-1), 402 to 422 (KPC1-2) and 126 to 146 (EGFP) of the respective coding regions. A mismatch shRNA KPC1-1-M1 contains two point mutations at positions 1860 (T
C) and 1862 (CA). Recombinant retroviruses were generated and used to infect NIH 3T3 cells or MEFs as described above. After selection in medium containing puromycin (10 g ml-1), cells stably expressing the shRNAs were pooled for experiments.
Accession numbers.
Homo sapiens and Mus musculus KPC1 accession numbers are AY744152 and AY744153, respectively.
BIND identifiers.
Three BIND identifiers (www.bind.ca) are associated with this manuscript: 182416, 182417 and 182418.
Note: Supplementary Information is available on the Nature Cell Biology website.
