The FHA domain protein SNIP1 is a regulator of the cell cycle and cyclin D1 expression

Abstract

Smad nuclear interacting protein 1 (SNIP1) is an evolutionarily conserved protein containing a forkhead-associated (FHA) domain that regulates gene expression through interactions with multiple transcriptional regulators. Here, we have used short interfering RNAs (siRNAs) to knockdown SNIP1 expression in human cell lines. Surprisingly, we found that reduction in SNIP1 levels resulted in significantly reduced cell proliferation and accumulation of cells in the G1 phase of the cell cycle. Consistent with this result, we observed that cyclin D1 protein and mRNA levels were reduced. Moreover, SNIP1 depletion results in inhibition of cyclin D1 promoter activity in a manner dependent upon a previously characterized binding site for the AP-1 transcription factor family. SNIP1 itself is induced upon serum stimulation immediately prior to cyclin D1 expression. These effects were independent of the tumour suppressors p53 and retinoblastoma (Rb), but were consistent with an interaction with BRG1, a component of the ATP-dependent chromatin remodelling complex, Swi/Snf. These results define both a new function for SNIP1 and identify a previously unrecognized regulator of the cell cycle and cyclin D1 expression.

Introduction

Smad nuclear interacting protein 1 (SNIP1) was originally identified in a yeast two-hybrid screen with members of the Smad transcription factor family, which mediate changes in gene expression in response to transforming growth factor β (TGFβ) (Kim et al., 2000). Endogenous SNIP1 coimmunoprecipitates with Smad4 in extracts from the mouse mammary gland epithelial cell line, NMuMg, and overexpressed SNIP1 represses TGFβ-induced transcription in reporter gene assays (Kim et al., 2000). Subsequent studies revealed that SNIP1 also interacts with the C/H1 domain of the transcriptional coactivators p300 and CREB binding protein (CBP) (Kim et al., 2000), which functions as a binding site for many transcriptional regulators (Goodman and Smolik, 2000). Through this interaction, SNIP1 overexpression was shown to inhibit the activity of the RelA(p65) nuclear factor-κB (NF-κB) subunit and displace binding of Smad4 to this domain (Kim et al., 2000, 2001). Therefore, SNIP1 has the potential to function as a regulator of many transcription factors. The majority of these studies have relied on overexpression of SNIP1, however, and the cellular function of the endogenous protein remains obscure.

SNIP1 itself is a 396 amino-acid protein that contains an N-terminal bipartite nuclear localization signal (NLS) and a putative forkhead-associated (FHA) domain in its C-terminus. It is conserved through evolution and homologues have been identified in organisms from Homo sapiens to Caenorhabditis elegans. Murine SNIP1 is 86% homologous to its human equivalent, while C. elegans SNIP1 (gene name C32E8.5) shows 47% overall homology (www.wormbase.org). The strongest homology is found within the C-terminal FHA domain. SNIP1 is ubiquitously expressed, and studies from Xenopus laevis show that ectopic expression of SNIP1 RNA in vivo inhibits dorsal mesoderm formation (Kim et al., 2000). Significantly, a C. elegans genome-wide RNA interference (RNAi) analysis has examined the effect of knocking down expression of the C. elegans SNIP1 protein. This study demonstrated that loss of SNIP1 resulted in embryonic lethality. Furthermore, knockdown of SNIP1 in adult worms resulted in growth defects and sterility (http://www.wormbase.org/db/gene/gene?name=C32E8.5).

A role in growth regulation would be consistent with the observation that SNIP1 contains an FHA domain. The FHA domain is a protein : protein interaction domain thought to specifically mediate binding to motifs containing phosphorylated, particularly phospho-threonine, residues (Durocher et al., 2000). FHA domains were originally identified as a sequence profile of 75 amino acids found in a variety of proteins including a small number of forkhead-type transcription factors (Pierrou et al., 1994). The FHA domain is one of the relatively rare protein modules to be found in both prokaryotes and eukaryotes (Hofmann and Bucher, 1995). In eukaryotes, proteins containing an FHA domain are found to be almost exclusively nuclear and seem to be prevalent among proteins which function in the cell cycle and DNA damage response (Zhou, 2000). Proteins containing FHA domains include checkpoint kinase 2 (CHK2) and mediator of DNA damage checkpoint protein 1 (MDC1) (Lee and Chung, 2001; Goldberg et al., 2003). The function of the SNIP1 FHA domain is unknown, however.

In this study, we have used RNAi to investigate the function of endogenous human SNIP1 to determine if it also has a role as a regulator of cell growth.

Results

SNIP1 is required for cell proliferation

To investigate the function of endogenous SNIP1 protein, small interfering RNAs (siRNAs) were designed and tested. Two siRNAs, to distinct regions of the SNIP1 coding sequence, were found to efficiently inhibit SNIP1 expression levels in human osteosarcoma U-2 OS cells (Figure 1a). Following siRNA transfection, it could be immediately observed that in those dishes where SNIP1 was depleted, fewer cells were observed (data not shown). Using an alamarBlue proliferation assay (Rocha et al., 2000), this effect was quantitated. SNIP1 knockdown resulted in significantly decreased proliferation of U-2 OS cells 48 h after the siRNA treatment (Figure 1b). This effect persisted until day 5 after siRNA treatment, when proliferation was again observed. This recovery could represent the fact that RNA interference is a temporary effect in the mammalian cell (Wilson et al., 2003) or possible outgrowth of untransfected cells. Importantly, both SNIP1 siRNA oligonucleotides gave the same result, demonstrating that the cellular effect observed was specific to SNIP1 and not a secondary affect of a particular sequence used.

Figure 1
figure1

SNIP1 is required for U-2 OS cell proliferation. (a) A single transfection of SNIP1 siRNA oligonucleotide efficiently knocks down SNIP1 protein levels in U-2 OS cells. U-2 OS cells were transfected with the indicated SNIP1 and control siRNA oligonucleotides. At 48 h after transfection, whole-cell lysates were prepared, resolved by SDS–PAGE and immunoblotted with anti-SNIP1 or anti-β-actin antibodies. (b) SNIP1 depletion inhibits U-2 OS cell proliferation. U-2 OS cells were transfected with the indicated SNIP1 and control siRNA oligonucleotides. Cell proliferation was measured every 24 h for 5 days by alamarBlue assay. (c) SNIP1 depletion results in G1 cell cycle arrest. U-2 OS cells were transfected with SNIP1 and control siRNA oligonucleotides and harvested 48 h post transfection. A mock transfection, containing no siRNA oligonucleotide, was also performed. Cells were stained with propidium iodide and analysed by FACS analysis. The data shown were obtained from three independent experiments before calculating means and standard deviations, except the mock transfection, where data shown are representative of two independent experiments

To determine whether knockdown of SNIP1 resulted in apoptotic cell death and/or an effect on the cell cycle, siRNA transfected U-2 OS cells were analysed using fluorescence activated cell sorting (FACS). At 48 h post-transfection, cells were harvested, stained with propidium iodide and FACS analysis was performed. The results from this experiment demonstrated that loss of SNIP1 did not result in any apparent cell death at this or later time points, as evidenced by the absence of an increase in sub-G1 cells (Figure1c and data not shown). These cells did accumulate in the G1 phase of the cell cycle, concomitant with a decrease in both S and G2/M phases (Figure 1c). Again, both oligonucleotides targeting SNIP1 mRNA gave similar results (data not shown). Together, these observations indicated that in U-2 OS cells, loss of SNIP1 protein results in a G1 cell cycle arrest.

SNIP1's effects on the cell cycle are p53 independent

One of the principle functions of the p53 tumour suppressor is to induce G1 cell cycle arrest (Chehab et al., 2000; Vogelstein et al., 2000) and since U-2 OS cells possess wild-type p53 (Rocha et al., 2003b), it was possible that the effects of SNIP1 knockdown were indirect and resulted from its activation. Induction of p53 protein levels occurs principally due to post-translational mechanisms. Upon cellular stimulation, p53 becomes rapidly stabilized, with its half-life dramatically increasing from approximately 30 min to several hours. However, following knockdown of SNIP1, no increase in endogenous p53 protein or mRNA levels was observed (Figure 2a and b). In addition, Hdm2 mRNA expression levels were unaffected (Figure 2b). Hdm2 is a target gene of p53 that also functions as its inhibitor in a negative feedback loop (Grossman et al., 1998). Confirming that p53 transcriptional activity is not induced following SNIP1 knockdown, cotransfection of SNIP1 siRNA expression plasmids did not induce expression from the p53 responsive p21WAF1/CIP1 promoter (Figure 2c) or affect p21 protein levels (Figure 2d). In addition to these results, we were unable to detect any interaction between endogenous SNIP1 and endogenous p53 in U-2 OS cells by immunoprecipitation (data not shown).

Figure 2
figure2

SNIP1 depletion does not lead to activation of p53. (a, b) p53 protein and mRNA levels do not increase following SNIP1 knockdown in U-2 OS cells. U-2 OS cells were transfected with the indicated SNIP1 and control siRNA oligonucleotides. At 48 h after transfection, whole-cell protein lysates or RNA were prepared, and either resolved by SDS–PAGE and immunoblotted with anti-SNIP1, anti-p53 or anti-β-actin antibodies or subjected to semiquantitative PCR analysis for SNIP1, p53, Hdm2 and GAPDH expression. (c) U-2 OS cells were transfected with 1.5 μg of p21 luciferase reporter plasmid and either 3 μg of a control siRNA expression plasmid or 1 μg each of three SNIP1 siRNA plasmids (3 μg total). Cells were harvested 48 h after transfection and the luciferase activity was measured. (d) U-2 OS cells were transfected with the indicated SNIP1 and control siRNA oligonucleotides. At 48 h after transfection, whole-cell lysates were prepared, immunoblotted with anti-SNIP1, anti-p21 and anti-β-actin antibodies

To confirm the p53 independence of the effects of SNIP1 depletion as well as to investigate whether these effects could be seen in other cell types, we next examined the affect of SNIP1 siRNA treatment on p53-null H1299 non-small-cell lung carcinoma cells. Similar to the results seen with U-2 OS cells, depletion of SNIP1 significantly inhibited cell proliferation measured using the alamarBlue assay (Figure 3a). An increase in cells in G1 phase was also observed (Figure 3b). This latter effect was less dramatic than that seen in U-2 OS cells; however, this reflects the fact that a much higher proportion of H1299 cells are in G1 to begin with. Control experiments demonstrated that this cell cycle distribution was an intrinsic attribute of the cells that we were using and did not result from the transfection reagent or the control siRNA (data not shown). These results confirmed that the effects of SNIP1 depletion on cellular proliferation are p53 independent and are not restricted to one cell type.

Figure 3
figure3

SNIP1 knockdown inhibits cell proliferation in p53-null H1299 cells. (a, b) Cell proliferation assays and FACS analysis were conducted in H1299 cells as described for U-2 OS cells in Figure 1

SNIP1 does not interact with p300 in all cell types

SNIP1 has previously been reported to bind the C/H1 domain of the p300 transcriptional coactivator (Kim et al., 2000). p300-null cells exhibit proliferative defects (Yao et al., 1998) and it was possible, therefore, that the effects that we observed might result from an effect on p300 function. Previously, the interaction of endogenous SNIP1 with p300 was identified in a murine mammary gland epithelial cell line, NMuMg (Kim et al., 2000). However, it was unknown if SNIP1 associates with p300 in other cell types. Confirming the previously published observations, SNIP1 was seen to coimmunoprecipitate with p300 in extracts prepared from NMuMg cells (Figure 4a). In contrast, no interaction was observed between the two proteins in U-2 OS or HeLa cells (Figure 4a). Furthermore, gel filtration analysis using HeLa nuclear extract demonstrated that SNIP1 did not coelute with p300, confirming that these two proteins do not interact in this cell line (Figure 4b). p300 was found to exist in a relatively large complex (1 MDa) as previously reported (Bradney et al., 2003), but SNIP1 is a component of an even larger, unknown, complex. SNIP1 was not detected in any other fraction of the analysis, suggesting that all cellular SNIP1 is present in this high molecular weight complex (data not shown). SNIP1 was also found in a high molecular weight complex in extracts prepared from U-2 OS cells (data not shown).

Figure 4
figure4

SNIP1 associates with BRG1 and not p300 in U-2 OS and HeLa cells. (a) Association of SNIP1 with p300 is cell type specific. Endogenous SNIP1 was immunoprecipitated from 200 μg of NMuMg cells, U-2 OS cell whole-cell extracts or 1 mg of HeLa cell nuclear protein extract. The immunoprecipitated complex was resolved by SDS–PAGE and immunoblotted for SNIP1 and p300. Input lanes represent 10, 5 and 1% of total protein from NMuMg, U-2 OS and HeLa extract used in each immunoprecipitation, respectively. Preimmune serum from the rabbit used to raise the SNIP1 antibody was used as a negative control. (b) SNIP1 is found in a high molecular weight complex. HeLa nuclear protein extract was resolved by Superose 6 gel filtration. Fractions (50 μl) were collected and resolved by SDS–PAGE followed by immunoblotting for SNIP1, p300 and BRG1. Input lanes represent 5% of the HeLa nuclear extract used in the gel filtration analysis. SNIP1 was not detected in any fraction containing lower molecular weight complexes. The positions of molecular weight markers used during column calibration are shown. (c) SNIP1 coimmunoprecipitates with BRG1. Endogenous SNIP1 was immunoprecipitated from 200 μg of HeLa cell nuclear protein extract. The immunoprecipitated complex was resolved by SDS–PAGE and immunoblotted for SNIP1 and BRG1. Input lanes represent 10% of total protein from HeLa extract used. Preimmune serum from the rabbit used to raise the SNIP1 antibody was used as a negative control. (d) Colocalization of transfected SNIP1 and BRG1. U-2 OS cells were transfected with YFP-SNIP1 and GFP-BRG1 expression plasmids. At 48 h post transfection, cells were analysed and images were acquired using a DeltaVision microscope. Images were deconvoluted using SoftWorx (Applied Precision). DAPI stain was used to stain the cellular DNA. The yellow colour in the merged image indicates colocalization of YFP-SNIP1 with GFP-BRG1. (e) Characterization of the 2B12 SNIP1 monoclonal antibody. SNIP1 monoclonal antibody (clone 2B12) was tested for specificity by immunoblotting 100 μg of Hela nuclear extract. (f) Endogenous SNIP1 partially colocalizes with BRG1 and BRM. U-2 OS cells were permeabilized and stained with SNIP1 (2B12) (red), BRG1 (green) and BRM (green) antibodies. DAPI stain was used to stain cellular DNA. Image analysis was performed as in (d). Since these are digitalized images, both BRG1 and BRM signals are set to green even though they are from the same cell, using separate secondary antibodies. The yellow colour in the merged image indicates colocalization

Interestingly, SNIP1 partially coeluted with BRG1, a component of the SWI/SNF chromatin remodelling complex (Figure 4b) (Peterson and Workman, 2000). Furthermore, a low level of BRG1 was found to coimmunoprecipitate with SNIP1 (Figure 4c). Therefore, we next determined whether SNIP1 colocalizes with BRG1 within cells. Transfection of GFP and YFP tagged BRG1 and SNIP1 expression plasmids into U-2 OS cells revealed a dramatic colocalization between the two proteins in large discrete nuclear bodies (Figure 4d). To investigate if the endogenous proteins colocalized in U-2 OS cells, a monoclonal antibody to SNIP1 was used (Figure 4e). This antibody confirmed that SNIP1 is a predominantly nuclear protein and furthermore, consistent with the results of the gel filtration analysis of HeLa cell nuclear extracts, a partial colocalization with BRG1 was observed (Figure 4f). Partial colocalization was also seen with BRM, a BRG1-related protein that can also be a component of Swi/Snf chromatin remodelling complexes (Peterson and Workman, 2000). The endogenous proteins exhibited a quite distinct pattern of subnuclear localization to the fusion proteins, being present in many more, smaller foci, suggesting that overexpression of SNIP1 and BRG1 can result in nuclear protein aggregates. These experiments indicate that while SNIP1 might exert its effects by interacting with p300 in some cell lines, this does not appear to be the case in U-2 OS cells, where this interaction is absent. SNIP1 does interact, either directly or indirectly, with another transcriptional regulator, BRG1. SNIP1 does not appear to be a component of Swi/Snf itself as it only associates with a small fraction of the cellular BRG1 and BRM. It might, however, exert a specific regulatory function.

SNIP1 regulates cyclin D1 expression

Since SNIP1 depletion resulted in a G1 arrest in U-2 OS cells, we next investigated effects on known regulators of this stage of the cell cycle. Consistent with this observation, siRNA-mediated knockdown of SNIP1 resulted in a reduction in both cyclin D1 and cyclin E protein and mRNA levels (Figure 5a–d). Similar effects were seen with both SNIP1 siRNAs (data not shown) and with two distinct cyclin D1 primer sets over a range of RNA concentrations (Figure 5c). Consistent with a lack of effect on cell cycle distribution (Figure 1c), no difference was observed with SNIP1 or cyclin D1 levels between a mock transfection lacking any siRNA oligonucleotide and our control siRNA (Figure 5b). In contrast, and in keeping with our earlier observation of an absence of apoptosis in these cells (Figure 1c), caspase 3 protein levels (Figure 5a) and Bcl-2 and Bcl-xL mRNA levels (Figure 5d) were not affected by loss of SNIP1. Similarly, no differences in expression of the RelA NF-κB, IκBα subunit or our β-actin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) controls were seen (Figure 5a and d). Therefore, the effects of SNIP1 would appear to be specific and do not represent a general repression of transcription.

Figure 5
figure5

SNIP1 is required for cyclin D1 expression. (ag) SNIP1 depletion results in downregulation of cyclin D1 mRNA and protein. U-2 OS (a–d), H1299 (e and f) or Saos-2 (g) cells were transfected with either a mock transfection (b) or the indicated SNIP1 and control siRNA oligonucleotides. At 48 h after transfection, whole-cell lysates or RNA were prepared, and either resolved by SDS–PAGE and immunoblotted with the indicated antibodies or subjected to semiquantitative PCR analysis with primers to the genes shown. In (c), to confirm the downregulation of cyclin D1 message, an additional cyclin D1 primer set was used (primer set B) and a titration of RNA levels confirmed the quantitative nature of the assay. (h) Rb depletion does not affect SNIP1-dependent inhibition of cyclin D1 expression. U-2 OS cells were transfected with the indicated SNIP1, Rb and control siRNA oligonucleotides. At 48 h after transfection, whole-cell lysates were prepared, resolved by SDS–PAGE and immunoblotted with anti-SNIP1, Rb, cyclin D1 or β-actin antibodies

Cyclin D1 protein and mRNA levels were also reduced in H1299 cells following SNIP1 knockdown (Figure 5e and f). Similarly, SNIP1 depletion in Saos-2 osteosarcoma cells, which are null for both p53 and Rb, also resulted in reduced cyclin D1 levels (Figure 5g). General toxicity of the transfection reagent prevented alamarBlue analysis of proliferation in these cells (data not shown). These observations confirm the p53 independence of SNIP1's regulation of cyclin D1. It also demonstrates that this effect is independent of Rb function. This latter observation was confirmed by using an Rb siRNA SMARTPool in U-2 OS cells. Downregulation of Rb did not interfere with the loss of cyclin D1 observed upon SNIP1 depletion (Figure 5h).

Interestingly, cyclin E expression was not inhibited upon loss of SNIP1 in either H1299 or Saos-2 cells (Figure 5f and g). This observation suggested that cyclin D1 is the primary target of SNIP1, and that repression of cyclin E could be a secondary consequence, possibly requiring the activities of p53 or Rb.

Cell cycle regulation of SNIP1 protein levels

These results all suggested that SNIP1 itself might be a cell cycle-regulated protein. To investigate this hypothesis, SNIP1 levels following serum starvation and subsequent serum readdition were investigated. Serum starvation of U-2 OS cells results in a G1 synchronization of the cell cycle as these cells are dependent on serum growth factors for proliferation (Beadling et al., 2001). Addition of serum then releases the cells from this arrest, which then proceed through the G1 and S phases of the cell cycle in a synchronized manner (Beadling et al., 2001). Following serum starvation, SNIP1 and cyclin D1 protein levels were very low (Figure 6a). Following the addition of serum, a striking induction of SNIP1 protein was observed, preceding and then overlapping the induction of cyclin D1 (Figure 6a). By contrast, induction of c-Jun, a subunit of the AP-1 transcription factor complex and a known regulator of cyclin D1 expression (Albanese et al., 1995; Watanabe et al., 1996), immediately preceded induction of SNIP1. SNIP1 levels peaked at the 6-h time point and then subsequently declined. Serum induction of SNIP1 was confirmed by immunofluorescence using the 2B12 anti-SNIP1 monoclonal antibody (Figure 6b). The appearance of c-Jun protein just prior to SNIP1 induction raised the possibility that c-Jun might regulate SNIP1 expression. Consistent with this hypothesis, transfection of a c-jun expression plasmid into U-2 OS cells significantly induced endogenous SNIP1 protein levels in a dose-dependent manner (Figure 6c). Together, these results indicate that SNIP1 is a cell cycle-regulated protein and possibly a transcriptional target of the c-jun proto-oncogene.

Figure 6
figure6

SNIP1 induction following serum stimulation. (a) Analysis of serum-induced proteins in U-2 OS cells. U-2 OS cells were serum starved for 96 h and then released back into the cell cycle by readdition of serum. Whole-cell lysates were harvested at the indicated times post serum addition, resolved by SDS–PAGE and then immunoblotted using anti-SNIP1, cyclin D1, c-Jun and β-actin antibodies. (b) Serum stimulation induces SNIP1 expression. U-2 OS cells were serum starved as outlined in (a). Cells were permeabilized and stained with the 2B12 SNIP1 monoclonal antibody. DAPI stain was used to stain the cellular DNA. Immunofluorescence analysis was performed as in Figure 4. (c) c-Jun induces SNIP1 expression. U-2 OS cells were transfected with indicated levels of RSV c-jun expression plasmid. RSV controls were included as appropriate such that all transfections had equivalent levels of expression plasmid. Whole-cell lysates were harvested 48 h after transfection, resolved by SDS–PAGE and then immunoblotted using anti-SNIP1, c-Jun and β-actin antibodies. ns indicates the position of a nonspecific band seen in some experiments

SNIP1 regulates the cyclin D1 promoter

We next investigated whether SNIP1 depletion affected the activity of the cyclin D1 and cyclin E promoters. SNIP1 siRNA expression plasmids were cotransfected with cyclin D1 and cyclin E promoter luciferase reporter constructs. Consistent with the effects seen with the endogenous genes and proteins, knockdown of SNIP1 resulted in reduced levels of activity of both the cyclin D1 and cyclin E promoters in U-2 OS cells (Figure 7a). Earlier results using the p21 promoter had demonstrated that these effects did not result from nonspecific repression of transcription (Figure 2c). Consistent with this and also the effects seen on endogenous gene expression (Figure 5d), knockdown of SNIP1 also had a minimal effect on the NF-κB-regulated Bcl-xL promoter and an NF-κB responsive, 3 × κB luciferase reporter (Figure7a and data not shown). In contrast, and consistent with the effects on the endogenous proteins, repression of the cyclin D1 but not the cyclin E promoter activity was seen in Saos-2 cells (Figure 7b). These results indicate that, at least in part, SNIP1 regulation of cyclin D1 expression occurs through regulation of its promoter activity. Analysis of cyclin D1 promoter mutants (Figure 8a) revealed that repression resulting from loss of SNIP1 was totally dependent upon a previously characterized AP-1 site present in the promoter (Albanese et al., 1995; Watanabe et al., 1996) (Figure 8b). No dependence on previously characterized NF-κB-binding sites was seen, in contrast to our previous studies investigating the mechanism of p53-mediated repression of the cyclin D1 promoter (Rocha et al., 2003b). Electrophoretic mobility shift analysis (EMSA) using the cyclin D1 promoter AP-1 site revealed that binding to this site consisted predominantly of c-Jun-containing protein complexes (Figure 8c). Furthermore, SNIP1 knockdown did not inhibit binding of c-Jun to this site. Rather, a slight increase in binding was observed (Figure 8c).

Figure 7
figure7

SNIP1 regulates cyclin D1 promoter activity. U-2 OS (a) and Saos-2 (b) cells were transfected with 1.5 μg of cyclin D1, cyclin E and Bcl-xL luciferase reporter plasmids as indicated and either 3 μg of a control siRNA expression plasmid or 1 μg each of three SNIP1 siRNA plasmids (3 μg total). Cells were harvested 48 h after transfection and the luciferase activity was measured

Figure 8
figure8

SNIP1 regulation of the cyclin D1 promoter requires an intact AP-1-binding site. (a) Schematic diagram of cyclin D1 promoter and luciferase reporter plasmids used showing AP-1- and NF-κB-binding sites. (b) Analysis of SNIP1-dependent regulation of the cyclin D1 promoter. U-2 OS cells were transfected with 1.5 μg of the indicated cyclin D1 promoter reporter plasmids and either 3 μg of a control siRNA expression plasmid or 1 μg each of three SNIP1 siRNA plasmids (3 μg total). Cells were harvested 48 h after transfection and the luciferase activity was measured. (c) EMSA analysis of AP-1 DNA-binding activity following SNIP1 knockdown. Nuclear protein extracts were prepared from U-2 OS cells treated with the indicated siRNA oligonucleotides and EMSA analysis was performed using 32P-labeled cyclin D1 AP-1 oligonucleotide probe and 2 μg of protein. In total, 0.25 μg of anti-c-Jun antibody or IgG control was included as indicated. (d) Analysis of BRG1-dependent regulation of the cyclin D1 promoter. U-2 OS cells were transfected with 1.5 μg of the indicated cyclin D1 promoter reporter plasmids and either 3 μg of a control siRNA expression plasmid or 1.5 μg each of two BRG1 siRNA plasmids (3 μg total). Cells were harvested 48 h after transfection and the luciferase activity was measured

SNIP1 did not coimmunoprecipitate with c-Jun (data not shown), suggesting an indirect mechanism of regulation, possibly by interacting with a coactivator or corepressor. Since our earlier data indicated that SNIP1 does not interact with p300, in U-2 OS cells (Figure 4), we investigated whether BRG1, with which it does interact (Figure 4), can also regulate cyclin D1 promoter activity. Strikingly, knockdown of BRG1 also repressed the cyclin D1 promoter in an AP-1-binding site-dependent manner (Figure 8d). Although further experimentation will be required to determine the exact mechanism through which SNIP1 regulates cyclin D1 promoter activity this result, together with those of Figure 4, suggests that it might do so by influencing the activity and function of the Swi/Snf chromatin remodelling complex.

Discussion

In this report, we demonstrate a new function for the SNIP1 protein as a regulator of the cell cycle and cyclin D1 expression, in multiple cell lines, independent of p53 and Rb status. That this is a primary effect of SNIP1 is suggested by its induction immediately prior to cyclin D1 expression, following serum stimulation (Figure 6). Furthermore, serum induction of SNIP1 expression immediately follows induction of the proto-oncogene c-Jun and overexpression of c-Jun increases SNIP1 protein levels (Figure 6). That SNIP1 is also required for transcriptional activation through the AP-1 site of the cyclin D1 promoter (Figure 8) suggests a feedback loop between these proteins. Together, these results demonstrate that SNIP1 is an important cellular regulator of cell proliferation and the cell cycle, functionally linked to known oncogenes.

SNIP1 had previously been identified as a p300 coactivator interacting protein (Kim et al., 2000, 2001) and our studies were originally initiated to further investigate the functional significance of this observation. Although we verified the interaction of SNIP1 with p300 in NMuMg cells, we did not observe this interaction in the U-2 OS cells used in many of our experiments (Figure 4). Instead, we find that SNIP1 is present in a high molecular weight complex that does not contain p300. We cannot rule out that a more transient association between SNIP1 and p300/CBP does occur in these cells, but it would appear not to be one of its principle interaction partners. These results do suggest that SNIP1 association with p300/CBP might be regulated by post-translational modifications and that its function will therefore vary in different cell types or according to the cellular environment.

We did obtain evidence suggesting an interaction between SNIP1 and another transcriptional regulator, BRG1 (Figure 4), a component of the Swi/Snf ATP-dependent chromatin remodelling complex (Peterson and Workman, 2000). Endogenous SNIP1 only interacts with a small fraction of cellular BRG1 and its homologue BRM, however, indicating it is not a component of the Swi/Snf complex itself (Figure 4). Consistent with SNIP1 indirectly regulating AP-1 activity on the cyclin D1 promoter, we do not find that SNIP1 directly binds c-Jun (data not shown) and depletion of SNIP1 does not affect AP-1 binding by EMSA analysis (Figure 8). Furthermore, siRNAs targeting BRG1 mimic the effects of SNIP1 siRNAs on the cyclin D1 promoter: both repress the promoter in an AP-1 site-dependent manner (Figure 8). It is interesting to note that Swi/Snf has been previously demonstrated to interact with and regulate an AP-1 complex consisting of Fos/Jun dimers (Ito et al., 2001). An attractive hypothesis, therefore, would involve SNIP1 being required for recruitment of BRG1 to the cyclin D1 promoter. Alternatively, however, SNIP1 might affect the balance between BRG1 and BRM complex recruitment to the promoter or even regulate Swi/Snf chromatin remodelling activity in some way. Determination of how SNIP1 regulates BRG1 and Swi/Snf function will require further analysis and will be the subject of future studies. In particular, since a role for BRG1 as a regulator of the mammalian cell cycle has been previously demonstrated (Muchardt and Yaniv, 2001), it will be interesting to examine how widespread the effects of SNIP1 on Swi/Snf function are. Although we have focused on cyclin D1 expression here, it is perfectly possible that the SNIP1/BRG1(BRM) complex regulates other genes associated with the cell cycle. Indeed in U-2 OS cells, we do observe downregulation of cyclin E expression (Figure 5). Although this might be a secondary consequence of loss of cyclin D1, it is also consistent with the previously reported interaction of Swi/Snf with Rb (Muchardt and Yaniv, 2001) and it is interesting to note that we do not see repression of cyclin E in Rb-null Saos2 cells.

Depletion of SNIP1 and BRG1 results in an inhibition of cyclin D1 promoter activity that is abolished upon mutation of the AP-1-binding site (Figure 8). However, mutation of this site does not mimic the repressive effects of SNIP1/BRG1 depletion, and promoter activity is essentially unchanged. This implies a balance between activator and repressor proteins functioning through this promoter element: removal of the activator complex leads to dominance by the repressor complex, but mutation of the site removes both. This is perhaps not surprising given that these transfections are into an asynchronous cell population where both activation and inhibition of cyclin D1 promoter activity will be occurring.

We have also found that overexpression of SNIP1 in U-2 OS cells results in a relatively nonspecific repression of reporter gene activity (data not shown). Consistent with this, transiently expressed YFP-SNIP1 fusion protein accumulates in large nuclear bodies with a very different appearance to the pattern seen with endogenous SNIP1 protein (Figure 4). Although this data confirmed the association of SNIP1 with BRG1, which also forms these large nuclear bodies when overexpressed, it underlines the problems that can be presented with transient transfection of some cellular regulators. For this reason, we have not performed functional experiments in which SNIP1 has been transiently expressed since interpretation of these effects is hard. The presence of SNIP1 in a high molecular weight complex could provide an explanation for the ineffectiveness of this approach. Overexpressed SNIP1 will disturb the stoichiometry of this complex, resulting in dominant negative and nonspecific effects.

That SNIP1 is required for cyclin D1 expression and has such profound effects on cell proliferation has important implications for a possible role in cancer. Cyclin D1 is overexpressed in most human breast tumours (Weinstatsaslow et al., 1995). Deregulation of the factors that regulate its expression can also be expected in tumours where it is found at high levels. To date, SNIP1 has been the subject of relatively few publications, but our observations demonstrate that SNIP1 is an important regulator of the cell cycle that is worthy of much greater study. Understanding SNIP1 function will lead to greater comprehension of the mechanisms underlying the cell cycle and could also provide important insights into tumorigenesis.

Materials and methods

siRNA transfections

SNIP1 siRNA transfections were performed as described in Anderson and Perkins (2002), except that the cells were transfected once. All siRNA duplex oligonucleotides were synthesized by Dharmacon Research Inc. The Rb siRNA was purchased as SMARTPool, comprising a mixture of siRNA sequences. The sequences used for other siRNAs were as follows (sense strand only): (See Table 1)

Table 1 Table 1

SNIP1 siRNA's targeted regions of SNIP1 mRNA starting at position +268 and +879, respectively, relative to the start site of translation.

Antibodies

The SNIP1 polyclonal rabbit antibody was raised by Diagnostics Scotland using a purified, recombinant His-tagged SNIP1 protein fragment (amino acids 120–396). The SNIP1 monoclonal antibody 2B12 was raised using a purified, recombinant His-tagged SNIP1 fragment (amino acids 2–80) by Hybricore GmbH. Other antibodies used in this manuscript were anti-cyclin D1 (cat# 556470, Pharmingen) anti-p300 (cat# 554215, Pharmingen), anti-β-actin (cat# A5441 Sigma), anti-caspase 3 (sc-7148 Santa Cruz) anti-RelA (sc-372 Santa Cruz), anti-cyclin E (sc-248 Santa Cruz), anti-BRG1 (sc-10768 Santa Cruz), anti-BRM (sc-6450 Santa Cruz). The anti-p53 DO1 antibody was provided by Cancer Research UK and the anti-c-Jun antibody was provided by Simon Morton/Philip Cohen (University of Dundee).

Plasmids

Cyclin D1 and cyclin E promoter luciferase plasmids were obtained from Dr Richard Pestell (Georgetown University, Washington, DC, USA) and have been described previously (Albanese et al., 1995; Rocha et al., 2003b). The p21 promoter luciferase plasmid was obtained from Professor David Lane (University of Dundee). YFP-SNIP1 and GFP-BRG1 expression plasmids were generated by inserting their respective cDNAs into either pEYFP or pEGFP backbones (Clontech).

The siRNA expression plasmid has been described previously (Girdwood et al., 2003). siRNA oligonucleotides targeting SNIP1 and BRG1 were designed as described elsewhere (Brummelkamp et al., 2002). The sequences used were as follows (sense strand only): (See Table 2)

Table 2 Table 2

Gel filtration analysis

HeLa nuclear extract (250 μg) (Computer Cell Culture, Belgium) was applied to a 2.4 ml 3.2/30 Superose-6 gel filtration column (Amersham-Pharmacia Biotech). The column was run at a flow rate of 25 μl/min, samples were collected as 50 μl volumes and examined for the presence of SNIP1, BRG1 and p300 by Western blot analysis.

Cell proliferation assays

Cells were seeded at 5000 cells per well in a 96-well plate and following transfection, cell proliferation was assessed every 24 h for 5 days with the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide-like colorimetric alamarBlue assay that is based on the detection of metabolic activity (Biosource International, Camarillo, CA, USA). Absorbance was measured at 570 and 600 nm using a Rosys Anthos 2001 spectrophotometer (Rocha et al., 2000).

Primers for PCR analysis

Primers used for cyclin D1(A), GAPDH and Bcl-xL have been described previously (Rocha et al., 2003a, b). Other primers to SNIP1, cyclin E, Bcl-2, IκBα, Hdm2, cyclin D1(B) were as follows: (See Table 3)

Table 3 Table 3

Immunofluorescence microscopy

For immunofluorescence, cells grown on coverslips were fixed after washing once in PBS by incubation in 3.7% formaldehyde/PBS (pH 6.8) for 5 min. Cells were permeabilized in PBS-0.1% Triton X-100 for 15 min and then blocked in PBS-0.05% Tween supplemented with 1% normal donkey serum for 30 min. Anti-SNIP1 monoclonal antibody was used undiluted. Rabbit anti-BRG1 and Goat anti-BRM antibodies were used at 1 : 100 dilution. All second antibodies (labelled with either FITC, TRITC or Cy5) were purchased from Jackson Laboratories.

Other experimental procedures

Transient transfections, immunoprecipitations, protein extracts, EMSA analysis, RNA extractions, PCR analysis, Western blots, cell culture conditions and FACS analysis were all performed as described previously (Webster and Perkins, 1999; Anderson and Perkins, 2003; Rocha et al., 2003a, b). Luciferase assays were performed according to the manufacturer's instructions (Promega) and results were normalized for protein concentration with all experiments being performed a minimum of three times before calculating means and standard deviations as shown in figures. EMSA oligonucleotides corresponding to the cyclin D1 AP-1 site were as previously described (Albanese et al., 1995).

References

  1. Albanese C, Johnson J, Watanabe G, Eklund N, Vu D, Arnold A and Pestell RG . (1995). J. Biol. Chem., 270, 23589–23597.

  2. Anderson LA and Perkins ND . (2002). J. Biol. Chem., 277, 29550–29554.

  3. Anderson LA and Perkins ND . (2003). Mol. Cell. Biol., 23, 721–732.

  4. Beadling C, Cereseto A, Fan W, Naramura M and Smith KA . (2001). Oncogene, 20, 1771–1783.

  5. Bradney C, Hjelmeland M, Komatsu Y, Yoshida M, Yao TP and Zhuang Y . (2003). J. Biol. Chem., 278, 2370–2376.

  6. Brummelkamp TR, Bernards R and Agami R . (2002). Science, 296, 550–553.

  7. Chehab NH, Malikzay A, Appel M and Halazonetis TD . (2000). Genes Dev., 14, 278–288.

  8. Durocher D, Taylor IA, Sarbassova D, Haire LF, Westcott SL, Jackson SP, Smerdon SJ and Yaffe MB . (2000). Mol. Cell, 6, 1169–1182.

  9. Girdwood D, Bumpass D, Vaughan OA, Thain A, Anderson LA, Snowden AW, Garcia-Wilson E, Perkins ND and Hay RT . (2003). Mol. Cell, 11, 1043–1054.

  10. Goldberg M, Stucki M, Falck J, D'Amours D, Rahman D, Pappin D, Bartek J and Jackson SP . (2003). Nature, 421, 952–956.

  11. Goodman RH and Smolik S . (2000). Genes Dev., 14, 1553–1577.

  12. Grossman SR, Perez M, Kung AL, Joseph M, Mansur C, Xiao ZX, Kumar S, Howley PM and Livingston DM . (1998). Mol. Cell, 2, 405–415.

  13. Hofmann K and Bucher P . (1995). Trends Biochem. Sci., 20, 347–349.

  14. Ito T, Yamauchi M, Nishina M, Yamamichi N, Mizutani T, Ui M, Murakami M and Iba H . (2001). J. Biol. Chem., 276, 2852–2857.

  15. Kim RH, Flanders KC, Reffey SB, Anderson LA, Duckett CS, Perkins ND and Roberts AB . (2001). J. Biol. Chem., 276, 46297–46304.

  16. Kim RH, Wang D, Tsang M, Martin J, Huff C, de Caestecker MP, Parks WT, Meng XW, Lechleider RJ, Wang TW and Roberts AB . (2000). Genes Dev., 14, 1605–1616.

  17. Lee CH and Chung JH . (2001). J. Biol. Chem., 276, 30537–30541.

  18. Muchardt C and Yaniv M . (2001). Oncogene, 20, 3067–3075.

  19. Peterson CL and Workman JL . (2000). Curr. Opin. Genet. Dev., 10, 187–192.

  20. Pierrou S, Hellqvist M, Samuelsson L, Enerback S and Carlsson P . (1994). EMBO J., 13, 5002–5012.

  21. Rocha S, Campbell KJ and Perkins ND . (2003a). Mol. Cell, 12, 15–25.

  22. Rocha S, Martin AM, Meek DW and Perkins ND . (2003b). Mol. Cell. Biol., 23, 4713–4727.

  23. Rocha S, Soengas MS, Lowe SW, Glanzmann C, Fabbro D, Winterhalter K, Bodis S and Pruschy M . (2000). Cell Growth Differ., 11, 491–499.

  24. Vogelstein B, Lane D and Levine AJ . (2000). Nature, 408, 307–310.

  25. Watanabe G, Lee RJ, Albanese C, Rainey WE, Batlle D and Pestell RG . (1996). J. Biol. Chem., 271, 22570–22577.

  26. Webster GA and Perkins ND . (1999). Mol. Cell. Biol., 19, 3485–3495.

  27. Weinstatsaslow D, Merino MJ, Manrow RE, Lawrence JA, Bluth RF, Wittenbel KD, Simpson JF, Page DL and Steeg PS . (1995). Nat. Med., 1, 1257–1260.

  28. Wilson JA, Jayasena S, Khvorova A, Sabatinos S, Rodrigue-Gervais IG, Arya S, Sasrangi F, Harris-Brandts M, Beaulieu S and Richardson CD . (2003). Proc. Natl. Acad. Sci. USA, 100, 2783–2788.

  29. Yao TP, Oh SP, Fuchs M, Zhou ND, Ch'ng LE, Newsome D, Bronson RT, Li E, Livingston DM and Eckner R . (1998). Cell, 93, 361–372.

  30. Zhou MM . (2000). Nat. Struct. Biol., 7, 1085–1087.

Download references

Acknowledgements

We are particularly grateful to Dr Anita Roberts for help, assistance and encouragement to investigate SNIP1. We thank Maren Oehlmann for help with gel filtration analysis, Rosie Clarke for help with FACS analysis, all the members of the NDP laboratory and the Division of Gene Regulation and Expression at the University of Dundee for their help and assistance. NDP is funded by a Royal Society University Fellowship, KCR was formally a BBSRC PhD student but is now funded by the Association of International Cancer Research (AICR). TOH and NW are funded by a Wellcome Trust Senior Research Fellowship.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Neil D Perkins.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Roche, K., Wiechens, N., Owen-Hughes, T. et al. The FHA domain protein SNIP1 is a regulator of the cell cycle and cyclin D1 expression. Oncogene 23, 8185–8195 (2004). https://doi.org/10.1038/sj.onc.1208025

Download citation

Keywords

  • forkhead-associated domain
  • p300
  • AP-1
  • Swi/Snf
  • BRG1

Further reading