P27Kip1 serine 10 phosphorylation determines its metabolism and interaction with cyclin-dependent kinases

Abstract

p27Kip1 is a critical modulator of cell proliferation by controlling assembly, localization and activity of cyclin-dependent kinase (CDK). p27Kip1 also plays important roles in malignant transformation, modulating cell movement and interaction with the extracellular matrix. A critical p27Kip1 feature is the lack of a stable tertiary structure that enhances its "adaptability" to different interactors and explains the heterogeneity of its function. The absence of a well-defined folding underlines the importance of p27Kip1 post-translational modifications that might highly impact the protein functions. Here, we characterize the metabolism and CDK interaction of phosphoserine10-p27Kip1 (pS10- p27Kip1), the major phosphoisoform of p27Kip1. By an experimental strategy based on specific immunoprecipitation and bidimensional electrophoresis, we established that pS10-p27Kip1 is mainly bound to cyclin E/CDK2 rather than to cyclin A/CDK2. pS10- p27Kip1 is more stable than non-modified p27Kip1, since it is not (or scarcely) phosphorylated on T187, the post-translational modification required for p27Kip1 removal in the nucleus. pS10-p27Kip1 does not bind CDK1. The lack of this interaction might represent a mechanism for facilitating CDK1 activation and allowing mitosis completion. In conclusion, we suggest that nuclear p27Kip1 follows 2 almost independent pathways operating at different rates. One pathway involves threonine-187 and tyrosine phosphorylations and drives the protein toward its Skp2-dependent removal. The other involves serine-10 phosphorylation and results in the elongation of p27Kip1 half-life and specific CDK interactions. Thus, pS10-p27Kip1, due to its stability, might be thought as a major responsible for the p27Kip1-dependent arrest of cells in G1/G0 phase.

Keywords: ATRA, all-trans retinoic acid; CDK, cyclin-dependent kinase; CKI, CDK inhibitor; IUP, intrinsically unstructured protein; cyclin-depedent kinases regulation; p27Kip1; p27Kip1 metabolism; p27Kip1 modifications; pS10-p27Kip1, phosphoserine 10 p27Kip1; pT187-p27Kip1, phosphothreonine p27Kip1.

Figure 1.

Characterization and cell cycle distribution of p27Kip1 isoforms. (A) p27Kip1 isoforms. On the top: spot 0, unmodified wild type (WT); spot 2, monophosphorylated form of not modified type (1P), and spot 4, doubly phosphorylated derivative of not modified type (2P). On the bottom: spot 1, a not phosphorylated uncharacterized isoform; spot 3, monophosphorylated derivative of form 1. The immunoblotting (WB p27) was performed employing mAbs against p27Kip1. (B) Lan-5 cells were synchronized as reported under Materials and Methods. Then, cytosolic and nuclear extracts were prepared and analyzed (30 μg extract) for p27Kip1 (p27) by immunoblotting. PKM2 (pyruvate kinase muscle isoenzyme 2) and HDAC1 (histone deacetylase 1) were used as control for cellular localization and equal loading. Two different exposition times of films (1 and 5 minutes) were shown. (C) 1 mg nuclear extracts from G0, G1, S and G2/M phases were immunoprecipitated with polyclonal antibodies against p27Kip1, pS10-p27Kip1 and pT187-p27Kip1. An unrelated antiserum (NR) was employed as control. The immunoprecipitated material (IP) was then analyzed by monodimensional immunoblotting by employing mAbs against p27Kip1. The blot was exposed to film for 1 and 10 minutes.

Figure 2.

Bidimensional analysis of nuclear extracts and pS10-p27Kip1-depleted nuclear extracts in different phases cell division cycle. Lan-5 cells were synchronized as reported under Materials and Methods and nuclear extracts were prepared. Then, the extracts of each phase (G0, G1, S and G2/M) was pS10-p27Kip1-depleted by immunoprecipitation. Then, equal amounts of untreated nuclear extracts and nuclear pS10-p27Kip1-depleted extracts were analyzed by 2D/WB. (A–D) Report the results of analysis. Note that different amounts of extract of each phase were employed in the experiment due to the difference of p27Kip1 content. Particularly, 500 μg in the G0 and G1 analyses (A and B), 4 mg in S (C) and 2 mg in G2/M (D). (E) 4 mg of untreated nuclear extract were analyzed by 2D/WB (top). pT187-p27Kip1 IP materials from 4 mg of nuclear extract were prepared and also analyzed by 2D/WB (bottom).

Figure 3.

Interaction of p27Kip1 isoforms with different CDKs. (A) 500 μg nuclear and cytosolic extract aliquots were independently immunoprecipitated with a mAb against CDK1, pAbs against CDK2, CDK4, CDK6, an unrelated mAb (mAb NR) and an unrelated pAB (pAb NR). The IP materials were analyzed by a mAb against p27Kip1 (p27). The lanes of mAbs IPs (either against CDK1 and NR) gave non specific smears since mouse mAbs were employed either in the IP or in the immunodetection. 30 μg of nuclear and cytosolic extracts were analyzed in the input lanes. (B) Left: 500 μg Lan-5 nuclear extract aliquots were independently immunoprecipitated with pAbs against total p27Kip1 (IP 27, second lane); pS10-p27Kip1 (IP pS10–27, third lane), and an unrelated pAB (IP NR, fourth lane). The IP materials were analyzed by mAbs against p27Kip1 (p27), CDK4, CDK5 and a pAb against CDK7. 50 μg of nuclear extract was analyzed in the input lane. Right: exactly as in the experiment reported on the left except that the IP materials were analyzed by mABs against CDK1 and cyclin B1. (C) 2 mg of nuclear extract were immunoprecipitated with a pAB against CDK2. Then, the IP materials were separated by 2D/WB and analyzed by mABs against p27Kip1 (p27, top) and pABs against pS10-p27Kip1 (pS10-p27, bottom). (D) Two distinct 2 mg aliquot nuclear extracts of Lan-5 cells were immunoprecipitated as follows: top, one aliquot was immunoprecipitated with a non related antiserum and the supernatant was immunoprecipitated with antibodies against CDK2 (IP CDK2); bottom, the other aliquot was firstly immunoprecipitated with an antiserum against pS10-p27Kip1 and the supernatant was immunoprecipitated with antibodies against CDK2 (IP CDK2 pS10p27-DEPLETED). Then, the IP materials were analyzed by 2D/WB using a mAbs against p27Kip1. (E) As in panel D, except that the second IP was performed employing antibodies pABs against CDK4. (F) pT187-p27Kip1 IPs (2 mg extracts from G1, S and G2/M) were separated by 2D/WB and investigated for p27Kip1, CDK1 and CDK2 content by specific mAbs.

Figure 4.

Metabolism of pS10-p27Kip1 and interactions with cyclins. (A) Lan-5 cells were cultured in the presence of retinoic acid (ATRA, 5 μM) or MG132 (1 μM) or Epoxomicin (Epox, 2.5 μM) for 8 hours. Then, cells were collected and nuclear extracts prepared. Equal amount of protein (50 μg protein) were analyzed by protein gel blotting employing antibodies against p27Kip1. Note that the treatment with ATRA for 8 hours induces a low increase of p27Kip1. (B) 1 mg extract aliquots of Lan-5 cells treated as in panel A (i.e.s control, ATRA, MG132 and Epoxomicin) were immunoprecipitated with a serum against pT187-p27Kip1 or with unrelated pABs. The immunoprecipitated materials were then analyzed by 1D/WB with mABs anti-p27Kip1 (p27). IgG-H signals correspond to heavy chain of IgG. (C) 1 mg extract aliquots of Lan-5 cells treated as in panel A were immunoprecipitated with a serum against pS10-p27Kip1 and analyzed by 1D/WB with mABs anti-p27Kip1 (p27). (D) 500 μg protein of a MG132-treated Lan-5 cells were divided into 2 aliquots. One aliquot was depleted of pT187-p27Kip1 by IP (MG132 Treated pT187p27 depleted). The other aliquot was immunoprecipitated with a not related antiserum (MG132 treated). The supernatants of both the aliquots were analyzed by 2D/WB employing a mAbs against p27Kip1. (E) The same experiment described in the panel D except that the IP material was analyzed for pS10p27Kip1 content.

Figure 5.

p27Kip1 and S10-p27Kip1 interactions with cyclins. (A) Equal amounts (1 mg) of asynchronous nuclear Lan-5 cell extract and nuclear extract obtained from G1, S and G2/M phases were immunoprecipitated with polyclonal antibodies against p27Kip1, pS10-p27Kip1 and pT187-p27Kip1. An unrelated antiserum (NR) was employed as control. The filter was probed with mAbs against p27Kip1 and mAbs against Cyclin E. (B) Top: Cyclin E IP from G1 nuclear extract (500 μg) was analyzed by 2D/WB employed a monoclonal antibody against p27Kip1. Bottom: Cyclin E IP from S nuclear extract (4 mg) was analyzed by 2D/WB employed a mAbs against p27Kip1. (C) Cyclin A IPs were prepared from 2 different extracts and analyzed by 2D/WB employed a mAbs against p27Kip1. Top: the extract employed (from S + G2/M phases) was previously depleted of CDK1 by immunoprecipitation. Thus, the blot represents p27Kip1 phosphoisoforms interacting cyclin A/CDK2. Bottom: the extract employed (from S + G2/M phases) was previously depleted of CDK2 by immunoprecipitation. Thus, the blot represents p27Kip1 isoforms interacting witj cyclin A/CDK1. (D) Cyclin B IPs (from G2/M phase) was analyzed by 2D/WB using a mAbs against p27Kip1. (E) 200 μg of nuclear Lan-5 cells S phase extract was immunoprecipitated with pABs against p27Kip1. Analogously, 400 μg of nuclear Lan-5 cells S phase extract was immunoprecipitated with pABs against pS10-p27Kip1. The immunoprecipitated materials were analyzed for p27Kip1, CDK2, cyclin A and cyclin E contents.

Figure 6.

Metabolism of recombinant p27Kip1 mutants and characterization of pT198-p27Kip1 isoforms. (A) Equal amounts of recombinant wild-type p27Kip1(WT), p27Kip1-T187A (T187A), p27Kip1-S10E (S10E), and p27Kip1-S10A (S10A) were incubated with recombinant active CDK2/cyclin E (see Materials and Methods for additional details). After the reaction, the assay mixtures were analyzed by monodimensional western blotting using antibodies against CDK2, p27Kip1, and pT187p27Kip1. (B) Equal amounts of recombinant wild-type p27Kip1 (WT p27), p27Kip1-S10E (p27S10E), and p27Kip1-Y88A (p27Y88A) were incubated with recombinant active Abl (rAbl). After the reaction, the assay mixtures were analyzed by monodimensional protein gel blotting by using antibodies against p27Kip1 and phosphotyrosine (PY). (C) Lan-5 cells were transfected with 3 distinct expression vectors, namely pcDNA3-p27Kip1-S10E (p27S10E), pcDNA3-p27Kip1-S10E/T187A (p27S10E,T187A), and pcDNA3-p27Kip1-S10E/T187A/T198V (p27S10E,T187A,T198V). After 24 hours, cells were collected and the nuclear extracts were prepared. 200 μg protein extracts were then analyzed by 2D/WB employing antibodies against p27Kip1, (D) Nuclear (N) and cytosolic (C) extracts of asynchronous Lan-5 cells were prepared. Each extract was immunoprecipitated with polyclonal antibodies against pT198–p27Kip1 (IP pT198p27). Identical IP were performed employing a non related antiserum. The input (50 μg protein) and the different IP were analyzed by 1D/WB employing antibodies against p27Kip1. The blots were exposed to the films for different time intervals. (E) Nuclear (Nucleus) and cytosolic (Cytosol) extracts of growing Lan-5 were prepared. Then, 1 mg of each cellular fraction were immunoprecipitated with polyclonal antibodies against pT198–p27Kip1 (IP pT198p27) and analyzed by 2D/WB employing antibodies against p27Kip1. (F) 2 mg of nuclear extract of Lan-5 was prepared divided into 2 aliquots. One aliquot was immunoprecipitated with non-related antiserum and the supernatant recovered. The other aliquot was immunoprecipitated using antibodies against pS10p27Kip1 and the supernatant recovered (pS10p27 depleted). Then, the 2 supernatants were immunoprecipitated with a polyclonal antiserum against pT198p27Kip1 and the IP materials were analyzed by 2D/WB employing antibodies against p27Kip1.