Glyceraldehyde 3-phosphate dehydrogenase is a SET ... - Nature

Oncogene (2006) 25, 4033–4042

& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

Glyceraldehyde 3-phosphate dehydrogenase is a SET-binding protein and regulates cyclin B-cdk1 activity S Carujo, JM Estanyol, A Ejarque, N Agell, O Bachs and MJ Pujol Departament de Biologia Cel lular i Anatomia Patolo`gica, Facultat de Medicina, Universitat de Barcelona, Barcelona, Spain

We report here that glyceraldehyde-3-phosphate dehydrogenase (GAPDH) interacts in vitro and in vivo with the protein SET. This interaction is performed through the acidic domain of SET located at the carboxy terminal region. On analysing the functional relevance of SETGAPDH interaction, we observed that GAPDH reverses in a dose-dependent manner, the inhibition of cyclin Bcdk1 activity produced by SET. Similarly to SET, GAPDH associates with cyclin B, suggesting that the regulation of cyclin B-cdk1 activity might be mediated not only by the interaction of GAPDH with SET but also with cyclin B. To analyse the putative role of GAPDH on cell cycle progression, HCT116 cells were transfected with a GAPDH expression vector. Results indicate that overexpression of GAPDH does not affect the timing of DNA replication but induces an increase in the number of mitosis, an advancement of the peak of cyclin B-cdk1 activity and an acceleration of cell cycle progression. All these results suggest that GAPDH might be involved in cell cycle regulation by modulating cyclin B-cdk1 activity. Oncogene (2006) 25, 4033–4042. doi:10.1038/sj.onc.1209433; published online 13 February 2006 Keywords: SET; GAPDH; cyclin B; cdk1; cell cycle

Introduction The p21Cip1 protein belongs to the CIP/KIP family of cyclin-dependent kinase inhibitors (CKIs) that are involved in cell cycle regulation. It interacts with cyclins and cdks by two regions located at the NH2-domain of the protein, and as a consequence, inhibits the activity of cyclin-cdk complexes (Chen et al., 1995; Luo et al., 1995; Nakanishi et al., 1995). In addition to the association with cyclin-cdks, p21Cip1 also interacts with a growing number of proteins, indicating that it is involved in the regulation of many other Correspondence: Dr O Bachs, Departament de Biologia Cel lular i Anatomia Patolo`gica, Facultat de Medicina, Universitat de Barcelona, Casanova 143, 08036 Barcelona, Spain. E-mail: [email protected] Received 20 April 2005; revised 7 December 2005; accepted 4 January 2006; published online 13 February 2006

cellular functions (Coqueret, 2003). We have previously reported that p21Cip1 interacts with the SET protein, which was firstly identified as a CAN-fusion gene in an acute undifferentiated leukemia (von Lindern et al., 1992; Adachi et al., 1994). SET belongs to a family of proteins, which includes the yeast nucleosome assembly protein NAP-I (Kawase et al., 1996), MB20 (Shen et al., 2001), Cell Division Antigen-1 (Chai et al., 2001), nucleolar transforming growth factor-b1 target (Ozbun et al., 2001), supressor of presenilin-2 (Wen et al., 2000) and testis-specific protein Y-encoded (Vogel et al., 1998). SET participates in a diversity of cellular functions. It has been identified as a potent inhibitor of protein phosphatase 2A (Li et al., 1996; Adler et al., 1997; Pandey et al., 2003) and as a regulator of DNA replication, chromatin remodeling and gene transcription (Matsumoto et al., 1993; Okuwaki and Nagata, 1998; Shikama et al., 2000; Seo et al., 2001; Cervoni et al., 2002; Kutney et al., 2004). The protein SET has been found associated with pp32, a protein considered to be a tumor suppressor, in two types of protein complexes, the INHAT complex (inhibitor of histone acetyltransferases) and the SET complex (Beresford et al., 2001; Seo et al., 2001; Fan et al., 2002; Kutney et al., 2004). Whereas the INHAT complex participates in chromatin remodelation by blocking the acetylation sites of histones, the SET complex has been proposed to be involved in the repair response to oxidative stress (Lieberman and Fan, 2003). SET has been also involved in cell cycle progression by regulating the activity of several cdks. It reverses the p21Cip1 inhibitory effect on cyclin E-cdk2 activity (Estanyol et al., 1999), inhibits cyclin B-cdk1 activity (Kellogg et al., 1995; Canela et al., 2003) and enhances the activity of p35nck5a –cdk5 (Qu et al., 2002). Thus, SET participates in a significant number of cellular processes related to the regulation of cell proliferation and oncogenic transformation. In this work we have identified glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as a SET binding protein. We report here that similarly to SET, GAPDH also interacts with cyclin B and reverses the inhibition of cyclin B-cdk1 produced by SET. Our results also reveal that overexpression of GAPDH stimulates cell proliferation, suggesting that GAPDH might be a positive regulator of cell cycle progression.

GAPDH and SET interaction S Carujo et al

4034

Results In vitro interaction between glyceraldehyde-3-phosphate dehydrogenase and SET protein With the aim of identifying new p21Cip1-binding proteins, we generated affinity chromatography columns of GSTp21Cip1-Sepharose 4B or GST-Sepharose 4B as described under Materials and methods. The columns were loaded with Molt-4 cell lysates, washed and the p21Cip1-binding proteins were eluted with a buffer containing 1 M KCl. The collected proteins were then analysed by gel electrophoresis. As shown in Figure 1a, approximately 10 major proteins were detected in the eluates from the GSTp21Cip1 column, whereas in contrast, no proteins were bound to the GST column. As an additional control, we performed Western blot analysis of two well-known p21Cip1-binding proteins (SET and PCNA) on the eluates of the GSTp21Cip1 column. As showed in the Figure 1b, both proteins were present in the eluates confirming that the GSTp21Cip1 column was working properly. To identify the p21Cip1-binding proteins, the bands were sliced from the gel, trypsinized and the resulting peptides were processed by mass spectrometry. A band of 38 kDa (Figure 1a, arrowhead) was identified in that way as GAPDH. Figure 1c shows the five peptides of

this protein that were sequenced by mass spectrometry. The binding of GAPDH to the GSTp21Cip1 columns was confirmed by Western blotting using specific antiGAPDH antibodies (Figure 1d). To analyse whether the interaction between GAPDH and p21Cip1 was direct or not, affinity chromatography assays were performed. Thus, purified GAPDH was loaded on a GSTp21Cip1-Sepharose 4B column and the binding analysed by Western blotting using antiGAPDH antibodies. Results showed that GAPDH did not directly bind to the GSTp21Cip1 column (Figure 2a). Thus, it is likely that the association of GAPDH with p21Cip1, observed when cell extracts were loaded in the GSTp21Cip1 column, could be mediated by an intermediate protein. To identify this protein, we loaded a GAPDH-sepharose 4B column with the eluates obtained from the GSTp21Cip1 column (Figure 2b). After washing, the proteins bound to the GAPDH-column were eluted with a buffer containing 1 M KCl. Then, the proteins were visualized by gel electrophoresis. Results showed that four major proteins were bound to the GAPDH-Sepharose 4B column, one of 39 kDa and three of 13–18 kDa (Figure 2c). By mass spectrometry, we identified the 39 kDa as the protein SET and the 13–18 kDa proteins as histones (data not shown). The presence of SET in the eluates was confirmed by

a

b Sepharose 4B-GSTp21 PC FT W

Sepharose 4B-GST

E

PC

Sepharose 4B GSTp21

FT E

E

WB

97 66 SET

45

PCNA 31

c 1 51 101 151 201 251 301

d

GKVKVGVDGF THGKFHGTVK FTTMEKAGAH CTTCLAPLAK AQNLIPASTG AKYDDIKKVV LNDTFVKLVS

GRIGRLVTRA AEDGKLVIDG LKGGAKRIVI VIHDHFGIVE AAKAVGKVIP KEASEGPLKG WYDNEFGYSE

AFNSGKVDIV KAITIFQERD SAPSADAPMF GLMTTVHAIT ELDGKLTGMA ILGYTEDEVV RVVDLMAHMA

Sepharose 4B-GSTp21 PC

FT

AINDPFIDLH PENIKWGDAT VMGVNHFKYA ATQKTVDSPS FRVPTANVSV SDDFNGSNHS SKE

YMVYMFQYDS AYVVESTGVF NSLKIISNAS GKLWRGGRGA LDLTCRLEKP SIFDAGAGIE

Sepharose 4B-GST

E

PC

FT

E

GAPDH

Figure 1 Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) interacts with p21Cip1. (a) Molt-4 cell extracts were loaded onto a GSTp21Cip1-sepharose 4B column or on a GST-sepharose 4B column. Samples from cell extracts (PC), flow-through (FT), wash (W) and eluates (E) were electrophoresed and visualized with Coomassie blue staining. One of the proteins present in the eluate from the GSTp21Cip1-sepharose 4B column (arrowhead) was sliced from the gel and micro-sequenced by mass spectrometry. (b) As a control, the eluates from the GSTp21Cip1-sepharose 4B column were analysed for the presence of two well-known p21Cip1-binding proteins, SET and PCNA. This analysis was performed by Western blotting (WB) using anti-SET and anti-PCNA antibodies. (c) From the mass spectrometry analysis of the sliced band five peptides (bold letters) corresponding to GAPDH were identified. (d) Eluates from the affinity chromatography columns were subjected to Western blotting using an anti-GAPDH antibody. Oncogene


4035 Sepharose 4B-GSTp21

a

Std

PC

FT

b

Sepharose 4B GSTp21

E

Sepharose 4B GAPDH

45

c

MOLT-4

GAPDH

31

Sepharose 4B GAPDH Std E2 97 66

FT

d

31

Sepharose 4B-GAPDH PC

21

E2

E1

45

FT

W

E2 WB-SET

14

Sepharose 4B-GSTp21

e

PC

FT

W

E WB-SET

Extract

IMD

Sepharose 4B-GSTp21 Extract FT W E WB-GAPDH

WB-SET

Figure 2 The protein SET mediates the interaction between p21Cip1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). (a) Purified GAPDH was loaded onto a GSTp21Cip1-sepharose 4B column. Samples from cell extracts (PC), flow-through (FT) and eluate (E) were electrophoresed and visualized with Coomassie blue staining. (b) Molt-4 cell extracts were loaded onto a GSTp21Cip1sepharose 4B column. The eluate from this column (E1) was subsequently loaded onto a GAPDH-sepharose 4B column and the bound proteins were then eluted (E2). (c) The proteins present in the eluate E2 were separated by electrophoresis and visualized with Coomassie blue staining. (d) Eluate E2 was subjected to Western blotting using an anti-SET antibody. (e) Purified SET was loaded onto a GSTp21Cip1-sepharose 4B column. Samples from purified SET (PC), flow-through (FT), wash (W) and eluate (E) were electrophoresed and visualized by Western blot with anti-SET antibodies (upper panel). HCT116 cell extracts were immunodepleted by using anti-SET antibodies. Total cell extracts and immunodepleted extracts (IMD) were subjected to Western blotting with antiGAPDH or anti-SET antibodies (bottom panel, left). Immunodepleted extracts were loaded onto a GSTp21Cip1-sepharose 4B column. Samples from the total extract, from the flow-through (FT), wash (W) and eluate (E) were electrophoresed and visualized by Western blot with anti-GAPDH and anti-SET antibodies (bottom panel, right).

Western blot using an anti-SET antibody (Figure 2d). To further confirm that SET was required for GAPDH binding to p21Cip1, SET immunodepleted lysates were loaded onto the GSTp21Cip1 column and then the association of GAPDH to the column was analysed by Western blotting. Results revealed that when SET was absent from the lysates GAPDH was not able to associate with p21Cip1 (Figure 2e). To confirm that SET directly binds to GAPDH, pull down experiments were performed with purified SET

and GAPDH-sepharose 4B beads. Results revealed that GAPDH associates directly with SET (Figure 3a). Similar results were obtained by affinity chromatography using purified SET and GAPDH-Sepharose 4B columns (Figure 3b). Histones did not bind to GAPDH (data not shown), indicating that the association of histones with the GAPDH column might be indirect probably due to their ability to interact with SET (Seo et al., 2001). We subsequently aimed to characterize the interaction among p21Cip1, SET and GAPDH. Thus, we Oncogene


4036

a

b

Sepharose 4B-GAPDH Pull down B

Sepharose 4B-GAPDH

NB 45

48

PC

FT

E WB-SET

31

SET

36 29

45

Sepharose 4B PC FT

WB-SET

31

c

E

Sepharose 4B-GSTp21 PC FT

FT W Wn

E

WB-SET

WB-GAPDH

Figure 3 Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) directly interacts with SET. (a) Purified SET was pulled-down with GAPDH-sepharose 4B beads. Bound (B) and not bound (NB) fractions were analysed by gel electrophoresis and coomassie blue staining. (b) Purified SET was loaded onto a GAPDH-sepharose 4B column or a sepharose 4B column as a control. Samples from the input (PC), flow-through (FT) and eluate (E) were electrophoresed and visualized by Western blot using anti-SET antibodies. (c) Purified SET was loaded onto a GSTp21Cip1-sepharose 4B column to generate p21Cip1-SET complexes. After the elimination of the not bound SET, the column was loaded with purified GAPDH. Samples from input (PC), from the flow-through obtained after SET loading (FTSET), the flow-through obtained after GAPDH loading (FTG), the first wash (W1), the last wash (Wn) and eluate (E) were electrophoresed and visualized by Western blot with anti-GAPDH and anti-SET antibodies.

loaded the GSTp21Cip1 column with SET in order to generate GSTp21Cip1-SET complexes in the column. Then, after the elimination of the not-bound SET, we added purified GAPDH onto the column and analysed its binding to the complexes. Interestingly, when the column was cleaned with the washing buffer, both SET and GAPDH were removed from the column (Figure 3c). These results indicated that the binding of GAPDH to SET disrupts the interaction between p21Cip1 and SET. In vivo association between SET and glyceraldehyde-3phosphate dehydrogenase To investigate whether SET and GAPDH may interact in vivo, we first examined the intracellular distribution of GAPDH and SET proteins in HCT116 human colon cancer cells. Thus, immunocytochemical experiments using antibodies against GAPDH or SET proteins were performed. As it can be seen in Figure 4a, cells showed the SET protein mainly located in the nucleus, being the nucleoli excluded, whereas GAPDH was distributed in both nuclear and cytoplasmatic compartments. Results revealed a co-localization of both proteins in the nucleus of a significant number of cells. We subsequently aimed to study the co-localization of both proteins along the cell cycle. Thus, HCT116 cells were synchronized by serum starvation and the colocalization of both proteins was analysed by immunoOncogene

Figure 4 Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) interacts with the acidic carboxy-terminal domain of SET. (a) The intracellular localization of SET and GAPDH was analysed by immunocytochemistry in HCT116 cells. (b) Immunocytochemical analysis using anti-SET and anti-GAPDH antibodies were performed in synchronized HCT116 cells at 12, 15 and 18 h after re-entry into the cell cycle. Co-localization of both proteins was quantified and results represented in the graph as percentage of the cells showing co-localization respect to the total number of cells. (c) HCT116 cell extracts were immunoprecipitated with anti-SET antibodies and the immunoprecipitates subjected to Western blotting using anti-SET and anti-GAPDH antibodies. Immunoprecipitation with a nonrelated antibody was used as a control (d) To identify the SET domain that binds to GAPDH, three SET fragments were generated, SET 1 (aa 1–80), SET 2 (aa 81–180) and SET 3 (aa 181–277). (e) The SET fragments were loaded onto a GAPDH-sepharose 4B column or onto a sepharose 4B column. Then, samples from cell extracts (PC), flow-through (FT), wash (W) and eluate (E) were electrophoresed and visualized with Coomassie blue staining.

cytochemistry at different times after re-entry into the cell cycle. Results revealed that the number of cells showing co-localization between SET and GAPDH varies during the cell cycle (Figure 4b). Thus, at 12 h (S phase) an 18% of cells showed co-localization of both


4037

proteins, whereas a decrease was observed at 15 h (9% of cells). Then, at the beginning of mitosis (18 h), colocalization increased again (19%). To finally determine whether SET and GAPDH may interact in vivo, immunoprecipitation experiments were carried out. HCT116 cell extracts were immunoprecipitated using an anti-SET antibody and the presence of GAPDH in the immunoprecipitates was analysed by Western blotting, using anti-GAPDH antibodies. Results showed in Figure 4c revealed that endogenous SET and GAPDH co-immunoprecipitated, indicating that both proteins might be associated in the cells. In order to identify the domains of the SET protein that interact with GAPDH, we generated three SET fragments we named SET1 (aa 1–80), SET2 (aa 81–180) and SET3 (aa 181–277) (Figure 4d). Then, affinity chromatography experiments using these three fragments and GAPDH-sepharose 4B columns were performed. Results revealed that SET3 but not SET1 interact with GAPDH. A slight binding of GAPDH to SET2 fragment was also observed (Figure 4e). These results were confirmed by pull down experiments (data not shown). Glyceraldehyde-3-phosphate dehydrogenase reverses the inhibitory effect of SET on cyclin B-cdk1 activity We further aimed to study the functional relevance of SET-GAPDH interaction. As SET protein is a cyclin Bcdk1 inhibitor (Canela et al., 2003), we analysed whether GAPDH could affect this inhibitory activity of SET. Thus, we first checked whether GAPDH alone could affect cyclin B-cdk1 activity. HCT116 cell extracts were immunoprecipitated with anti-cyclin B antibodies and then cyclin B-cdk1 activity was analysed in the immunoprecipitates in the absence or in the presence of different amounts of purified GAPDH. As it can be observed in Figure 5a, growing amounts of GAPDH did not affect cyclin B-cdk1 activity. In contrast, as expected by previous results (Canela et al., 2003), SET inhibited cyclin B-cdk1 activity (Figure 5b). Finally, cyclin B-cdk1 activity was analysed in the presence of a fixed amount of SET (5 mM) and growing levels of GAPDH. Results indicate that the inhibition of cyclin B-cdk1 activity produced by SET is reverted in a dose-dependent manner by the addition of GAPDH (Figure 5c). Cyclin B interacts with both glyceraldehyde-3-phosphate dehydrogenase and SET As it has been previously reported that SET interacts in vitro with cyclin B (Kellogg et al., 1995) and we report here that GAPDH is able to block the inhibition of cyclin B-cdk1 activity induced by SET, we aimed to study whether GAPDH could interact with cyclin B. Thus, affinity chromatography experiments using a GAPDH-sepharose 4B column and purified GSTcyclin B were carried out. Purified GSTcyclin B or GST (as a control) was loaded onto a GAPDH-sepharose 4B column. Results revealed that GSTcyclin B but not GST associated with the GAPDH column, indicating that GAPDH and cyclin B interact in vitro (Figure 6a,

Figure 5 Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) reverses the inhibition of cyclin B-Cdk1 activity induced by SET. (a) HCT116 cell extracts were immunoprecipitated with an anticyclin B antibody. Then, cyclin B-Cdk1 activity was determined in the immunoprecipitates in the absence or in the presence of growing concentrations of exogenous GAPDH. An immunoprecipitation with normal rabbit serum (NRS) was used as a control. Histone H1 was used as a substrate (upper panel, left). A Western blot with anti-cyclin B antibodies was performed as a control of the immunoprecipitation (bottom pane, left). Cyclin B-cdk1 activities were quantified and represented in the graph as a percentage of the maximal activity (right panel). (b) Cyclin B-Cdk1 activity was determined in the immunoprecipitates in the absence or in the presence of 5 mM SET. An immunoprecipitation with normal rabbit serum (NRS) was used as a control. A Western blot with anti-cyclin B antibodies was performed as a control of the immunoprecipitation. (c) Cyclin B-cdk1 activity was determined in the immunoprecipitates in the absence of any protein, in the presence of 5 mM SET, or in the presence of 5 mM SET and growing amounts of GAPDH. The activity was quantified and represented as the mean value7s.d. of four independent experiments.

upper panels). An additional control was performed by loading purified GSTcyclin B onto a sepharose 4B column. Results indicate that cyclin B does not interact with the beads (Figure 6a, bottom panel). We also aimed to study the intracellular distribution of both proteins. Thus, immunocytochemical experiments were performed in HCT116 cells using anti-GAPDH and anticyclin B antibodies. As observed in Figure 6b, both cyclin B and GAPDH are mostly cytoplasmatic and in a significant number of cells, a co-localization of both proteins in this cellular compartment was observed. However, also a nuclear co-localization of both proteins was observed in a limited number of cells. To finally determine whether cyclin B and GAPDH could interact in vivo, immunoprecipitation experiments Oncogene


4038

a

0h 12h

24h 36h 48h

67-

WB GFP-GAPDH

31GFP

n° cells × 105 / dish

b 28 24 20 16 12 8 4 0

GFP GFP-GAPDH 0

12

24 Hours

36

48

Figure 7 Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is a positive regulator of cell proliferation. HCT116 cells were transfected with expression vectors harboring GFP-GAPDH or GFP and were subsequently cultured from 24 h. Then, cells were collected at different times of growing. (a) HCT116 cell extracts were subjected to Western blotting with anti GFP antibodies. (b) The number of cells was counted at different times of growing. Results are expressed as the mean value7s.d. of four independent experiments.

showed that both proteins co-immunoprecipitated (Figure 6d). Figure 6 Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and SET interact with cyclin B. (a) Fifteen micrograms of purified GSTcyclin B1 or purified GST were loaded onto a GAPDHsepharose 4B column. Samples from the input (PC), flow-through (FT), wash (W) and eluate (E) were electrophoresed and visualized with coomassie blue staining. An additional control was performed by loading purified GSTcyclin B1 on a sepharose 4B column. (b) The intracellular localization of GAPDH and cyclin B was analysed by immunocytochemistry in HCT116 cells. (c) HCT116 cell extracts were immunoprecipitated with anti-GAPDH antibodies and the presence of GAPDH and cyclin B in the immunoprecipitates was analysed by Western blotting using antiGAPDH and anti-cyclin B antibodies. Immunoprecipitation with a nonrelated antibody was used as a control (C ) (d) HCT116 cell extracts were immunoprecipitated with anti-SET antibodies and the presence of SET and cyclin B in the immunoprecipitates was analysed by Western blotting using anti-SET and anti-cyclin B antibodies. Immunoprecipitation with a nonrelated antibody was used as a control (C ).

were carried out. HCT116 cell extracts were immunoprecipitated with an anti-GAPDH antibody and the presence of cyclin B in the immunoprecipitates was analysed by Western blotting, using an anti-cyclin B antibody. Results showed in Figure 6c revealed that endogenous cyclin B and GAPDH co-immunoprecipitated, indicating that both proteins might be associated in the cells. We also studied the in vivo interaction between cyclin B and SET by immunoprecipitation experiments. Thus, HCT116 cell extracts were immunoprecipitated with an anti-SET antibody. Then, the presence of cyclin B in the immunoprecipitates was determined by Western blotting using an anti-cyclin B antibody. Results clearly Oncogene

Glyceraldehyde-3-phosphate dehydrogenase behaves as a positive regulator of cell proliferation As our results indicate that GAPDH is able to regulate cyclin B-cdk1 activity in vitro, we aimed to study whether overexpression of GAPDH might affect cell proliferation. To analyse this aspect, HCT116 cells were transfected with an expression vector harboring GFPGAPDH or GFP (that was used as a control). Cells were then cultured during 24 h after transfection and the number of cells was subsequently measured at different times of growing. Under these experimental conditions, the ectopic expression of GAPDH or GFP in the cells was maintained at least until 36 h after the start of the experiment (Figure 7a). As shown in Figure 7b, cells overexpressing GAPDH proliferate faster than control cells, indicating that GAPDH positively regulates cell proliferation. In order to understand which phase of the cell cycle was affected by the overexpression of GAPDH, we analysed whether GAPDH could regulate DNA replication. To this end, (3H)-Thymidine incorporation into DNA was determined in synchronized GFP-transfected cells (control) or in cells overexpressing GFP-GAPDH. Results indicate that overexpression of GAPDH did not modify the timing of DNA replication during the cell cycle (Figure 8a). We then studied whether overexpression of GAPDH could affect mitosis. Thus, immunocytochemical experiments, using antibodies against phosphorylated histone H3 (a mitotic marker) were performed. Results revealed that the number of mitosis was significantly higher (a 50% increase) in HCT116


4039

cells overexpressing GAPDH (Figure 8c). Specifically, the peak of cyclin B-cdk1 activity occurred earlier in cells overexpressing GAPDH than in control cells.

Thymidineincorporation (cpm)

a 200000 150000 GFP-GAPDH GFP

100000 50000 6

8

b

10 12 Hours

H3P

2

14

16

GFP GFP-GAPDH

1

0 12 Hours post-transfection 120

c % of activity

100 GAPDH GFP

80 60 40 20 0 12

14

16 18 Hours

20

22

Figure 8 Overexpression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) advances the peak of cyclin B-cdk1 activity. (a) The incorporation of (3H)-Thymidine into DNA was determined in synchronized HCT116 cells at different times of the cell cycle. Results are represented as the mean values in cpm7s.d. of four independent experiments (b) The number of mitosis was determined in cells transfected with GFP-GAPDH or GFP at 12 h after transfection by immunocytochemistry using specific anti-phosphohistone H3 as described under Materials and methods. The number of cells in mitosis is represented in arbitrary units. (c) Extracts from synchronized HCT116 cells were immunoprecipitated with an anticyclin B antibody and cyclin B-Cdk1 activity was determined in the immunoprecipitates. The activity was measured in synchronized cells at different times after cell cycle resumption. Results are represented as the percentage of the maximal activity (mean value7s.d. of three independent experiments).

cells overexpressing GAPDH than in control cells (Figure 8b). We were subsequently interested to investigate whether the increase in the number of cells in mitosis produced by GAPDH overexpression could be related to a modification in cyclin B-cdk1 activity. Thus, cyclin B-cdk1 activity was determined in synchronized HCT116 cells at different times after re-entry of the cells into the cell cycle. Cells extracts collected at different times were immunoprecipitated with anticyclin B antibodies and the kinase activity was determined in the immunoprecipitates. Results indicated that the timing of cyclin B-cdk1 activity was modified in

Discussion We report here the observation that GAPDH directly interacts with SET, a previously identified p21Cip1binding protein (Estanyol et al., 1999). Thus, we further aimed to analyse the functional consequences of this interaction. Glyceraldehyde-3-phosphate dehydrogenase is a classical metabolic enzyme involved in energy production. However, numerous studies have established that it is a multifunctional protein in mammalian cells (Sirover, 1999). Interestingly, it is clear now that GAPDH displays a distinct membrane, cytosolic and nuclear localization. In those subcellular sites, it is involved in a number of cellular functions, including endocytosis and membrane fusion (Glaser and Gross, 1995; Robbins et al., 1995), vesicular secretory transport (Tisdale, 2001), translational control (Yi et al., 2000), nuclear tRNA transport (Singh and Green, 1993), DNA replication (Zheng et al., 2003) and DNA repair (Krynetski et al., 2001). Moreover, during the last years, new nuclear functions for GAPDH have emerged. These include the transcriptional regulation of histone gene expression (Zheng et al., 2003), the role of GAPDH in nuclear membrane fusion (Nakagawa et al., 2003), in the modulation of telomere structure (Sundararaj et al., 2004), in the recognition of fraudulent nucleotides in DNA arising from cancer chemotherapy (Krynetski et al., 2003) and in triggering apoptosis (Chuang et al., 2005; Hara et al., 2005). Results reported here indicate that SET associates with GAPDH through the carboxy terminal domain of the protein (aa 181–277). This region of SET contains a large acidic domain that has been described to be involved in a number of functions as for instance, DNA replication of adenovirus genome (Nagata et al., 1995) or cyclin B-cdk1 inhibition (Canela et al., 2003). As SET inhibits cyclin B-Cdk1 activity, we first focused our efforts to explore whether GAPDH could affect the inhibition of cyclin B-Cdk1 activity produced by SET (Canela et al., 2003). Our results indicate that in vitro, GAPDH reverses the inhibitory effect of SET on cyclin B-cdk1 complexes. SET has been reported to be a cyclin B-binding protein (Kellogg et al., 1995; Canela et al., 2003). Interestingly, the SET domain involved in cyclin B binding but also in the inhibition of cyclin B-Cdk1 activity is the acidic carboxy terminal region of the protein (aa 181–277) (Canela et al., 2003). This domain is the same that we report here to be responsible for the association of SET with GAPDH. These results suggest that both GAPDH and cyclin B could compete for the binding to SET. It is thus likely that the effect of GAPDH on reversing the inhibition of cyclin B-cdk1 activity produced by SET could rely on this competiOncogene


4040

tion. This possibility is compatible with the observation that GAPDH is also able to directly interact with cyclin B and that this interaction does not affect cyclin B-cdk1 activity. These results opened the possibility that GAPDH could be involved in the regulation of cell proliferation. Thus, we subsequently aimed to analyse the effect of GAPDH overexpression on cell cycle progression. Our observations revealed that GAPDH overexpression results in an advanced peak of cyclin B-cdk1 activity in synchronized cells. Moreover, overexpression of GAPDH also induces an acceleration of cell cycle progression. Interestingly, in these experimental conditions, the timing of DNA replication was not affected. Thus, these data are compatible with the possibility that GAPDH might be regulating G2/M transition. It has been recently reported that the nuclear accumulation of GAPDH is a cell cycle dependent process. Specifically, in A549 cells, nuclear GAPDH is primarily detectable in S phase, it remains high at G2/M and it strongly decreases by the end of mitosis. At G0/G1, GAPDH is almost completely excluded from the nucleus (Sundararaj et al., 2004). These results indicate that at G2/M transition, cells may contain significant levels of GAPDH in the nucleus that might be able to regulate cyclin B-cdk1 activity. Cyclin B-cdk1 activity is regulated at different levels, that is, by phosphorylation of cdk1 at thr 160, which activates the complex; by phosphorylations at thr 14 and tyr 15 of cdk1, which inactivates the complex (these phosphorylated residues might in turn be dephosphorylated by the protein phosphatase CDC25C and being the complex activated in that way), and by the association of CKIs to the complex (Morgan, 1997). The identification of GAPDH as a cyclin B-binding protein and its ability to reverse the inhibition produced by SET but not that of p21Cip1 on cyclin B-cdk1 complexes reveal a new regulatory mechanism for the activity of these complexes and indicates that the regulation of cyclin BCdk1 activity is a much more complicated process that it was previously described. In addition to cell cycle regulation, SET also participates in a diversity of other functions as regulation of PP2A activity, chromatin remodeling and transcription, histone acetylation and apoptosis (Li et al., 1996; Okuwaki and Nagata, 1998; Seo et al., 2001; Cervoni et al., 2002; Fan et al., 2003). Thus, the association of GAPDH with SET could also be a new mechanism involved in the regulation of these cellular functions that merits to be explored in the next future.

Materials and methods Plasmids Human GAPDH cDNA in a pChug 20.2 plasmid was a Gift from Dr Sirover. It was subsequently inserted in a pGEX-5X-3 vector by BamH1–EcoR1 sites and in a pGFP-C2 vector by BglII and EcoR1 sites. The three SET fragments (SET1, aa 1–80; SET2, aa 81–180; and SET3, aa 181–277) were in a pGEX-4T-2 vector as previously described (Canela et al., Oncogene

2003). The human cyclin B1 and p21cip1 cDNAs were inserted into a pGEX-KG vector by NcoI and XhoI and NdeI–HindIII sites, respectively. Cell culture and transfections The lymphoblastoid human Molt-4 cell line was cultured in RPMI-1640 containing 10% of fetal calf serum (FCS). The colon carcinoma human HCT116 cell line was cultured in Dubelcco’s modified Eagle’s–Ham’s F12 medium (1:1) (Biological Industries, Kibbutz beit Haemek, Israel) containing 10% FCS. Cells were maintained al 371C in humidified atmosphere in a 5% of CO2. Transfection of HCT116 cells with GFP or GFP-GAPDH expression vectors was performed with Lipofectamet (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions. To synchronize HCT116 cells, they were cultured in 0.5% FCS for 24 h at confluence. Then, they were trypsinized and cultured at low confluence in a medium containing 10% FCS. Antibodies and purified proteins Anti-GAPDH antibody (MBA374) was from Chemicon (Temecula, CA, USA), anti-SET antibody (ab-1183) from Abcam (Cambridge, UK) and anti-cyclin B1 (05–373) and anti-phospho histone H3 (06–570) antibodies were from Upstate Biotechnology (Lake placid, NY, USA). All the others antibodies, anti-PCNA (SC-56), anti-cdk4 (SC-260 g) and anti-p21Cip1 (SC-397) were obtained from Santa Cruz Biotehnology (Santa Cruz, CA, USA). Purified human erythrocyte GAPDH was from Sigma (St Louis, MO, USA). Protein expression and purification Glutathione S-transferase (GST) and GST-fusion proteins were expressed in Escherichia coli and subsequently purified by absorption to glutathione-Sepharose beads (Amersham Biosciences, Uppsala, Sweden) as previously described (SanchezPiris et al., 2002). In several cases, GST was separated from the GST fusion proteins by digestion with thrombin protease according to the manufacturer (Sigma). Affinity chromatography GSTp21Cip1-sepharose 4B columns were prepared by coupling 5–7 mg of purified GSTp21Cip1 to 3 ml of CNBr-activated Sepharose 4B (Amersham Pharmacia Biotech, Uppsala, Sweden). GST-sepharose 4B columns were also built in the same way. Molt-4 cells were lysed in a buffer containing 50 mM HEPES, pH 7.6, 150 mM KCl, 1 mM MgCl2, 1 mM EGTA, 1 mM DTT, 1 mM PMSF, 10 mg/ml leupeptin and 1 mg/ml aprotinin (lysis buffer). Cell extracts (70 mg) were loaded onto the GSTp21Cip1-sepharose 4B or GST-sepharose 4B columns. After washing in 50 vol of lysis buffer, the bound proteins were eluted with the same buffer but containing 1 M KCl. GAPDHsepharose 4B columns were prepared by coupling 3.5 mg of purified GAPDH to 1.75 ml of CNBr-activated Sepharose 4B as mentioned above. Eluates from the GSTp21Cip1 sepharose 4B column were dyalized against a buffer containing 50 mM HEPES, pH 7.6, 50 mM KCl, 1 mM EGTA, 1 mM MgCl2, 1 mM PMSF, 10 mg/ml leupeptin and 1 mg/ml aprotinin before being loaded onto the GAPDH-sepharose 4B column. Elution of proteins bound to the GAPDH-sepharose 4B column was performed using lysis buffer containing 1 M KCl. Purified GST-cyclin B1 (15 mg) was loaded on the GAPDH-sepharose 4B column and its binding analysed as mentioned above. The fractions were then analysed by SDS–polyacrylamide gel electrophoresis (SDS–PAGE).


4041 SET immunodepletion For the immunodepletion of SET, HCT116 cell extracts (150 mg) were incubated with anti-SET antibodies (15 ml) (ab1183, abcam), overnight at 41C. Then, A/G protein was added for 1 h 30 min at 41C. After centrifugation, the supernatant was checked for the levels of SET and then it was used for affinity chromatography analysis. Pull-down experiments Pull-down experiments were performed as previously described (Canela et al., 2003) using GAPDH-sepharose 4B beads. Full-length SET and its different fragments (2–5 mg) were incubated with immobilized GAPDH in binding buffer (50 mM Tris-HCl, pH 7.2, 150 mM NaCl, 1% Triton X-100 and 1 mM DTT) for 1 h at room temperature. Then, the beads were extensively washed with binding buffer. After washing, the bound fraction was extracted from the beads with Laemmli sample buffer. The unbound fraction corresponds to the supernatant of the first wash. The fractions were then analysed by SDS–PAGE. Immunoprecipitation, gel electrophoresis and Western blotting HCT116 cells were lysed in buffer A (50 mM Tris-HCl, pH 7.4, 5 mM EDTA, 250 mM NaCl, and 0.1% Triton X-100) containing 50 mM NaF, 0.1 mM Na3VO4, 1 mM PMSF, 10 mg/ml leupeptin and 0.5 mg/ml aprotinin, on ice for 30 min. Samples were then clarified by centrifugation at 10 000 g for 10 min at 41C. Supernatants were subsequently incubated overnight at 41C with the specific antibodies, followed by incubation with protein A- (Pierce) or protein G- (Sigma) agarose beads for 1 h at 41C. After washing three times with buffer A, the immunoprecipitates were then electrophoresed in 12% SDS– PAGE gels and subsequently subjected to Western blotting as described (Canela et al., 2003). Cyclin B-cdk1 kinase activity To determine the activity of cyclin B-cdk1 complexes, 500 mg of lysed HCT116 cells were immunoprecipitated with a specific anti-cyclin B1 antibody and the kinase activity measured in the immunoprecipitates as previously described (Canela et al., 2003). In the kinase assays, different amounts of SET or GAPDH were added as mentioned in the text. Mass spectrometry analysis The bands from SDS–PAGE gels were manually sliced from the gel and digested with trypsin following conventional procedures (Shevchenko et al., 2000). Sequences from the resulting peptides were obtained by using a LCQ DECA XP ion trap mass spectrometer (ThermoQuest Finnigan MAT, San Jose, CA, USA), as described (Carrascal et al., 2002). (3H)-Thymidine incorporation HCT116 cells (4 105) were transfected with GFP or GFPGAPDH expression vectors. At 24 h after transfection, cells were synchronized by serum starvation (0.5% FCS) for 24 h. Then, (3H)-Thymidine (25 Ci/mmol, Amersham) was added to

the plates at different times as mentioned in the text at a final concentration of 8 mCi/ml. Cells were then harvested after 1 h incubation at 371C. They were then washed three times in PBS. Cells were subsequently lysed in 50 mM Tris-HCl, pH 7.5, 100 mM DTT and 0.2% SDS. DNA was subsequently precipitated using 5% TCA/1% sodium pyrophosphate, for 1 h at 41C. After 5 min of centrifugation at 12 000 r.p.m. in a microfuge, pellets were washed twice with 5% TCA/1% sodium pyrophosphate. Ethanol was then added to the pellets and incubated for 2 h at 41C. After centrifugation for 10 min at 12 000 r.p.m., the pellets were dried and then incubated with 500 ml of 0.3 M NaOH for 1 h at 41C. Finally, 100 ml of 1.5 M HCl and 5 ml of scintillation liquid were added to each sample. Radioactivity was finally quantified in a scintillation counter. Cell proliferation assay HCT116 cells were transfected with GFP or GFP-GAPDH expression vectors as mentioned above. At 24 h after transfection, cells were placed in p35 dishes (2 105 cells/dish). Then, at 12, 24, 36 at 48 h of growth, cells were trypsinized and counted in a Neubauer chamber. Determination of the mitotic index To determine the number of cells in mitosis, immunofluorescence experiments were performed using anti-phosphorylated histone H3 (06-570, UBI). Three independent experiments were performed. In each experiment a total of 300 cells were counted. Immunofluorescence Wild-type HCT116 cells or HCT116 transfected cells were seeded in culture dishes containing glass coverslips and allowed to grow for 24 h. Cells were then fixed with 4% paraformaldehyde for 30 min and incubated with primary antibodies for 1 h at room temperature. After washing three times in PBS, cells were incubated with secondary antibodies conjugated with FITC- or TRIC-conjugated antibodies for 45 min at room temperature. After washing in PBS, coverslips were mounted on slides with Mowiol (Calbiochem, Darmstad, Germany). Immunofluorescence was measured using a confocal laser fluorescence microscope. Acknowledgements We are grateful to Dr Sonia Brun for technical support and to Dr Nu´ria Canela and Jorge Vera for their collaboration. Mass spectrometry analysis were performed at the ‘Unitat de proteo`mica dels Serveis cientific-te`cnics de la Universitat de Barcelona, Facultat de Medicina’. This work was supported by Grants SAF2002-00452, SAF 2003-08329, GEN2003-20243C08 from the Ministerio de Educacioń y Ciencia of Spain. Financial support was also obtained from the networks of cooperative research on Cancer (C03/10) from the Instituto Carlos III.

References Adachi Y, Pavlakis GN, Copeland TD. (1994). J Biol Chem 269: 2258–2262. Adler HT, Nallaseth FS, Walter G, Tkachuk DC. (1997). J Biol Chem 272: 28407–28414. Beresford PJ, Zhang D, Oh DY, Fan Z, Greer EL, Russo ML et al. (2001). J Biol Chem 276: 43285–43293.

Canela N, Rodriguez-Vilarrupla A, Estanyol JM, Diaz C, Pujol MJ, Agell N et al. (2003). J Biol Chem 278: 1158–1164. Carrascal M, Carujo S, Bachs O, Abian J. (2002). Proteomics 2: 455–468. Cervoni N, Detich N, Seo SB, Chakravarti D, Szyf M. (2002). J Biol Chem 277: 25026–25031. Oncogene


4042 Chai Z, Sarcevic B, Mawson A, Toh BH. (2001). J Biol Chem 276: 33665–33674. Chen J, Jackson PK, Kirschner MW, Dutta A. (1995). Nature 374: 386–388. Chuang DM, Hough C, Senatorov VV. (2005). Annu Rev Pharmacol Toxicol 45: 269–290. Coqueret O. (2003). Trends Cell Biol 13: 65–70. Estanyol JM, Jaumot M, Casanovas O, Rodriguez-Vilarrupla A, Agell N, Bachs O. (1999). J Biol Chem 274: 33161–33165. Fan Z, Beresford PJ, Oh DY, Zhang D, Lieberman J. (2003). Cell 112: 659–672. Fan Z, Beresford PJ, Zhang D, Lieberman J. (2002). Mol Cell Biol 22: 2810–2820. Glaser PE, Gross RW. (1995). Biochemistry 34: 12193–12203. Hara MR, Agrawal N, Kim SF, Cascio MB, Fujimuro M, Ozeki Y et al. (2005). Nat Cell Biol 7: 665–674. Kawase H, Okuwaki M, Miyaji M, Ohba R, Handa H, Ishimi Y et al. (1996). Genes Cells 1: 1045–1056. Kellogg DR, Kikuchi A, Fujii-Nakata T, Turck CW, Murray AW. (1995). J Cell Biol 130: 661–673. Krynetski EY, Krynetskaia NF, Bianchi ME, Evans WE. (2003). Cancer Res 63: 100–106. Krynetski EY, Krynetskaia NF, Gallo AE, Murti KG, Evans WE. (2001). Mol Pharmacol 59: 367–374. Kutney SN, Hong R, Macfarlan T, Chakravarti D. (2004). J Biol Chem 279: 30850–30855. Li M, Makkinje A, Damuni Z. (1996). J Biol Chem 271: 11059–11062. Lieberman J, Fan Z. (2003). Curr Opin Immunol 15: 553–559. Luo Y, Hurwitz J, Massague J. (1995). Nature 375: 159–161. Matsumoto K, Nagata K, Ui M, Hanaoka F. (1993). J Biol Chem 268: 10582–10587. Morgan DO. (1997). Annu Rev Cell Dev Biol 13: 261–291. Nagata K, Kawase H, Handa H, Yano K, Yamasaki M, Ishimi Y et al. (1995). Proc Natl Acad Sci USA 92: 4279–4283.

Oncogene

Nakagawa T, Hirano Y, Inomata A, Yokota S, Miyachi K, Kaneda M et al. (2003). J Biol Chem 278: 20395–20404. Nakanishi M, Adami GR, Robetorye RS, Noda A, Venable SF, Dimitrov D et al. (1995). Proc Natl Acad Sci USA 92: 4352–4356. Okuwaki M, Nagata K. (1998). J Biol Chem 273: 34511–34518. Ozbun LL, You L, Kiang S, Angdisen J, Martinez A, Jakowlew SB. (2001). Genomics 73: 179–193. Pandey AV, Mellon SH, Miller WL. (2003). J Biol Chem 278: 2837–2844. Qu D, Li Q, Lim HY, Cheung NS, Li R, Wang JH et al. (2002). J Biol Chem 277: 7324–7332. Robbins AR, Ward RD, Oliver C. (1995). J Cell Biol 130: 1093–1104. Sanchez-Piris M, Posas F, Alemany V, Winge I, Hidalgo E, Bachs O et al. (2002). J Biol Chem 277: 17722–17727. Seo SB, McNamara P, Heo S, Turner A, Lane WS, Chakravarti D. (2001). Cell 104: 119–130. Shen HH, Huang AM, Hoheisel J, Tsai SF. (2001). Genomics 71: 21–33. Shevchenko A, Chernushevich I, Wilm M, Mann M. (2000). Methods Mol Biol 146: 1–16. Shikama N, Chan HM, Krstic-Demonacos M, Smith L, Lee CW, Cairns W et al. (2000). Mol Cell Biol 20: 8933–8943. Singh R, Green MR. (1993). Science 259: 365–368. Sirover MA. (1999). Biochim Biophys Acta 1432: 159–184. Sundararaj KP, Wood RE, Ponnusamy S, Salas AM, Szulc Z, Bielawska A et al. (2004). J Biol Chem 279: 6152–6162. Tisdale EJ. (2001). J Biol Chem 276: 2480–2486. Vogel T, Dittrich O, Mehraein Y, Dechend F, Schnieders F, Schmidtke J. (1998). Cytogenet Cell Genet 81: 265–270. von Lindern M, van Baal S, Wiegant J, Raap A, Hagemeijer A, Grosveld G. (1992). Mol Cell Biol 12: 3346–3355. Wen C, Levitan D, Li X, Greenwald I. (2000). Proc Natl Acad Sci USA 97: 14524–14529. Yi M, Schultz DE, Lemon SM. (2000). J Virol 74: 6459–6468. Zheng L, Roeder RG, Luo Y. (2003). Cell 114: 255–266.