The 20–40 kDa monomeric small GTP-binding proteins (SMGs) that constitute the ras superfamily are present exclusively in eukaryotes (Takai et al., 2001). Those SMGs that have been identified (currently >100) have homology to ras, and belong to one of the five subfamilies namely ras, rho, rab, sar1/arf and ran (Takai et al., 2001). These ras-like proteins shuttle between inactive GDP- and active GTP-bound states. Ras subfamily proteins, including ras and rap1, are key players in receptor-linked signaling pathways that control cell growth and differentiation (Bokoch, 1993; Wagner and Williams, 1994). In fact, ras is one of the most frequently mutated and activated oncogenes in human and animal model systems. Ras activation has been observed in >90% of pancreatic carcinomas and 50% of colon carcinomas (Barbacid, 1987; Grunicke and Maly, 1993). Mutant ras is also found in high proportions of carcinogen-induced cancers including lung and bladder cancers (Mirvish, 1995). Although ras mutations have been reported in 35% of oral cancers in India and Southeast Asia where betel nut chewing and reverse smoking are prevalent, ras mutations in oral cancers are quite rare in patients in the Western Hemisphere (Clark et al., 1993; Saranath et al., 1991; Jordan and Daley, 1997). It is thus likely that activation of other ras-like proteins may play a role in oral carcinogenesis because SMGs are integral members of the growth factor signaling pathways (Bokoch, 1993). Since oropharyngeal squamous cell carcinomas (SCCs) have a low frequency of ras mutations and because our recent work in normal and immortalized oral keratinocytes demonstrated that rap1B is linked to growth cessation in normal oral keratinocytes (D'Silva et al., In Press, Journal of Cellular Physiology), we were prompted to investigate rap1 expression and regulation in malignant cells.
Rap1 is a ubiquitously expressed member of the ras subfamily but its function is poorly understood. Rap1 was initially identified as a ras-tumor-suppressor gene in ras-transformed NIH 3T3 fibroblasts (Kitayama et al., 1989). However, subsequent work has shown that rap1 is not a single protein. Rap1 has two isoforms, rap1A and rap1B. These isoforms are products of two different genes located on chromosomes 1 and 12, respectively (Noda, 1993). Rap1A, which is also referred to as smg p21 or Krev1, is 95% homologous to rap1B, differing by only nine amino acids (of a total of 184), six of which are located between amino acids 171 and 184 at the C-terminus (Bokoch, 1993).
The rap1 protein may have different cellular functions depending on the isoform and its subcellular localization, as well as the proliferative capacity and differentiation state of the particular cell type. For example, in rat salivary gland (D'Silva et al., 1997,1998), neutrophils (Maridonneau-Parini and de Gunzburg, 1992) and platelets (Lapetina et al., 1989; Berger et al., 1994), rap1 has been localized to secretory granules and is thought to regulate secretory granule formation or exocytosis (Bos et al., 2001). Rap1 proteins have also been localized on the Golgi of rat fibroblasts (Beranger et al., 1991), the zymogen granules, plasma membrane and microsomal membrane of pancreatic acini (Schnefel et al., 1992), and in the endocytic compartment of fibroblasts and skeletal muscle cells (Pizon et al., 1994,1996). More recently, rap1 was identified in the perinuclear region of COS-1 cells (Mochizuki et al., 2001). However, until the present study, rap1 has not been identified in the nucleus. In fact, of >100 SMGs expressed in eukaryotes, ran/TC4 is the only SMG hitherto identified in the nucleus (Moore and Blobel, 1993), where it plays a critical role in cell proliferation by regulating nuclear import and export (Moore, 2001; Takai et al., 2001).
In this report, we show high expression of rap1 in oropharyngeal SCCs and cell lines, and the first evidence of its localization inside the nucleus as well as in the cytoplasm. Furthermore, we show that the active GTP-bound form is localized in the nucleus, whereas the inactive GDP-bound form is trapped in the cytoplasm.
Rap1 expression in human oropharyngeal SCC cell lines
We have recently demonstrated that rap1B inhibits proliferation in normal human oral keratinocytes (D'Silva et al., In Press). Hence, in initial experiments, rap1 expression was assessed in oropharyngeal SCC cell lines, in which growth control mechanisms are disrupted. In all, 10 human oropharyngeal SCC cell lines: UM-SCC-11A, UM-SCC-11B, UM-SCC-14A, UM-SCC-14B, UM-SCC-22A, UM-SCC-22B, UM-SCC-17B, UM-SCC-74A, UM-SCC-81B and OSCC3, were screened (Figure 1a). Rap1 protein is most strongly expressed in UM-SCC-17B followed by UM-SCC-22B and UM-SCC-81B, and shows a lower level of expression in most of the other cell lines including OSCC3, a poorly differentiated oral SCC cell line (Figure 1a). As a result of reports in the literature linking rap1 to differentiation (York et al., 1998; Bos et al., 2001) and our own previous observation of increased rap1 expression during keratinocytes differentiation and growth cessation, we assessed the relationship between rap1 and transglutaminase expression in the SCC cell lines (Figure 1c). This marker of keratinocyte differentiation was variably expressed in the cell lines and showed no correlation with rap1 expression (Figure 1a, c). The highest expression of transglutaminase was noted in UM-SCC-11B, UM-SCC-17B and UM-SCC-74A, whereas UM-SCC-11A, UM-SCC-14B, UM-SCC-81B and OSCC3 showed slightly lower expression. UM-SCC-22A shows only a faint signal (Figure 1c) and transglutaminase expression in UM-SCC-14A and UM-SCC-22B was detected only on very prolonged exposures on duplicate membranes (data not shown). Equivalency of loading was assayed by blotting the membranes with mouse anti-GAPDH antibody (Figure 1b, d, respectively). Thus, rap1 is consistently but variably expressed in oropharyngeal SCC cell lines.
Variable rap1 expression in human oropharyngeal SCC cell lines; no correlation with differentiation. Protein bands in whole cell lysates were detected with antibodies to rap1 (a) or transglutaminase (c) followed by immunochemiluminescence. Equivalency of loading was evaluated by blotting the same membranes with antibodies to GAPDH (b and d, respectively). The immunoblot data are representative of two independent experimentsFull figure and legend (100K)
Nuclear and cytoplasmic localization of rap1
To verify the immunoblot data, immunohistochemical studies of rap1 expression were performed using cytospins of UM-SCC-17B (Figure 2) and UM-SCC-22B (Figure 3). In UM-SCC-17B, rap1 expression was observed in the cytoplasm, as expected. However, we also noted variable nuclear staining (Figure 2a, red arrow). In UM-SCC-22B an even greater intensity of nuclear staining could be appreciated (Figure 3a, red arrow). Since the affinity-purified polyclonal antibody binds to rap1 as well as other less and more rapidly migrating peptides on immunoblots (Figures 2e and 3e, arrowheads), we investigated the possibility that the immunohistochemical staining with the polyclonal antibody might be nonspecific. Therefore, using a rap1-specific monoclonal antibody, we tested rap1 expression in UM-SCC-17B and UM-SCC-22B on cytospins (Figures 2b and 3b), cells cultured on Lab-tek slides (insets in Figures 2b and 3b) and immunoblots (Figures 2f and 3f). This antibody detected no cross-reactive peptides by immunoblot analysis (Figures 2f and 3f). The relatively greater intensity of nuclear distribution was especially evident in cells stained with the monoclonal antibody (Figure 3b). Punctate nuclear staining was observed with both UM-SCC-17B (Figure 2b) and UM-SCC-22B (Figure 3b). Interestingly, the immunohistochemical staining with the monoclonal antibody appeared to be more punctate than that observed with the anti-rap1 polyclonal antibody, which stained the nucleus more diffusely (Figures 2a and 3a, red arrows). Bluish hematoxylin-stained nuclei are masked with brown DAB precipitate (red arrows), whereas nuclei that were not stained with the rap1 antibody are blue (blue arrows). Thus, immunohistochemical localization in the nuclear fraction was more prominent with the monoclonal antibodies. The rabbit IgG (Figures 2c and 3c) and mouse IgG (Figures 2d and 3d), used as controls for nonspecific binding of the affinity-purified polyclonal and monoclonal primary antibodies, respectively, were appropriately negative. Immunoblot analysis confirmed that the majority of the rap1 signal is in the nuclear fraction (NE) rather than in the cytoplasmic extract (CE) in both UM-SCC-17B (Figure 2e, f) and UM-SCC-22B (Figure 3e, f). In Figure 3e, f, unprocessed recombinant rap1 (rR) (synthesized in E. coli, a generous gift from Dr T Fisher, University of North Carolina), which migrates slightly slower than the processed endogenous cellular form, was run to confirm antibody specificity. Equivalency of protein loading was assessed by India ink staining of the nitrocellulose membranes (Figures 2g and 3g).
Rap1 is present in the nuclear fraction of UM-SCC-17B cell line; immunohistochemical and immunoblot comparison of rap1 detection by polyclonal and monoclonal antibodies. Cytospins of UM-SCC-17B were incubated with rabbit anti-rap1 affinity-purified polyclonal antibody (a) or rabbit IgG (c) (2 g/ ml each) or mouse anti-rap1 monoclonal antibody (b) or mouse IgG (d) (10 g/ml each). The insets in (b) and (d) are UM-SCC-STAL-17B grown on tissue culture slides. Brown DAB reaction product with hematoxylin counterstain (bar=27 m, inset bar=30 m). Red arrows indicate nuclear staining, blue arrows are unstained nuclei. Protein bands on samples of NP40 whole cell lysate (CL), cytoplasmic and nuclear extracts (CE and NE respectively), (25 g each) were detected with monoclonal (f) antibodies to rap1. The membrane was subsequently washed and re-exposed to the rap1 polyclonal antibody (e) and exposed to film. The same membrane was subsequently stained with India ink for detection of equivalency of protein loading (g). The particulate (Pt) and cytosolic (CY) fractions were prepared as described, electrophoresed and blotted for rap1 (h)Full figure and legend (367K)
Rap1 is present in the nuclear fraction of UM-SCC-22B cell line; immunohistochemical and immunoblot comparison of rap1 detection by polyclonal and monoclonal antibodies. Cytospins of UM-SCC-22B were incubated with rabbit anti-rap1 affinity-purified polyclonal antibody (a) or rabbit IgG (c) (2 g/ ml each) or mouse anti-rap1 monoclonal antibody (b) or mouse IgG (d) (10 g/ml each). The insets in (b) and (d) are UM-SCC-22B grown on tissue culture slides. Brown DAB reaction product with hematoxylin counterstain (bar=27 m, inset bar=30 m). Protein bands on samples of unprocessed recombinant rap1 (rR, 10 ng), NP40 whole cell lysate (CL), cytoplasmic and nuclear extracts (CE and NE, respectively) (25 g each) were detected with monoclonal (f) antibodies to rap1. The membrane was subsequently washed and re-exposed to the rap1 polyclonal antibody (e) and exposed to film. At the end of the experiment the membrane was stained with India ink for detection of equivalency of protein loading (g). The particulate (Pt) and cytosolic (CY) fractions were prepared as described, electrophoresed and blotted for rap1 (h)Full figure and legend (269K)
Rap1 may function by shuttling between membrane-bound and soluble subcellular fractions. This prompted an investigation of rap1 localization in these two fractions. When cell extracts were fractionated into the particulate (Pt) and soluble cytosolic (CY) fractions, rap1 was mainly isolated in the particulate fractions of both cell lines (Figures 2h and 3h) and only a faint signal was identified in the cytosolic fraction of UM-SCC-17B (Figure 2h). No cytosolic rap1 signal was detected in UM-SCC-22B (Figure 3h). This suggests that rap1 exists primarily in a membrane-bound form in SCC cells. It may be only transiently expressed in the cytosolic fraction and otherwise shuttles between membrane-bound organelles. Such a possibility is consistent with this protein being geranylgeranylated (Takai et al., 2001).
Rap1 expression in human oropharyngeal SCC tissues
To investigate the expression of rap1 in human oral SCC tissues, immunohistochemical studies were performed on 5 m sections of formalin-fixed, paraffin-embedded tissues using the rap1-specific polyclonal antibody. The rap1 epitope identified by the monoclonal anti-rap1 antibody was not accessible on the paraffin-embedded, formalin-fixed tissues. All 15 SCCs examined expressed rap1 (Figure 4, representative figure). Predominantly, cytoplasmic expression of rap1 was noted in large, differentiated malignant keratinocytes (Figure 4a). Some of the staining appeared to be perinuclear in distribution. Both nuclear (red arrows) and cytoplasmic localization were identified in cells that were generally smaller and morphologically less differentiated (Figure 4b). In cancer cells showing nuclear staining, the bluish hematoxylin-stained nuclei were masked by the brown DAB precipitate (Figure 4b, red arrows). The blue arrows show nuclei that do not stain with rap1 (Figure 4b). Human salivary gland also expresses rap1 strongly (D'Silva et al., 1997), predominantly in the cytoplasm of acinar cells (Figure 4c) although an occasional cell also shows nuclear expression (red arrows). Negative controls with rabbit IgG instead of the primary antibody were run for each specimen examined. These controls were appropriately negative (Figure 4d).
Rap1 is present in the nucleus and cytoplasm of human SCC. 5 m tissue sections of human SCC tissue were incubated with the rap1 polyclonal antibody (a, b, c) or rabbit IgG (d) at 2 g/ml. Brown DAB reaction product with hematoxylin counterstain. (a), (b) and (d) squamous cell carcinoma; (c) salivary gland (white bar for (a), (b) and (c)=62.5 m; black bar for (d)=62.5 m), (a), (b), (c) and (d) are from the same tissue section. Red arrows indicate nuclear staining. The data are representative of 15 tissue specimens with similar resultsFull figure and legend (505K)
Rap1 isoforms in oropharyngeal SCC cell lines
To investigate which rap1 isoform is expressed in different SCC cell lines, we performed real-time reverse transcriptase polymerase chain reaction (RT–PCR) with isoform-specific primer sets. Both rap1A and rap1B mRNAs were expressed in all SCC cell lines (Figure 5a). Rap1A and rap1B mRNAs are expressed most strongly in UM-SCC-17B. Consistent with the Western blot data, UM-SCC-11B exhibited the lowest expression of rap1B and rap1A mRNA. UM-SCC-14A and UM-SCC-14B also expressed very low levels of rap1A mRNA. The ratio of rap1B to rap1A expression varied between cell lines, with all eight-cell lines expressing either more rap1B or equivalent amounts of the two isoforms.
Quantification of rap1A and rap1B mRNAs in different SCC cell lines by Q-RT–PCR and Northern blots. Q-RT–PCR of RNA from SCC cell lines (a). The data are the means.d. of 4 replicates. Northern blot analysis with radiolabeled rap1A (b) or rap1B (c), with GAPDH as internal controlFull figure and legend (108K)
To verify that the real-time RT–PCR data were not a function of normalization to -actin, a cytoskeletal protein that may be variably overexpressed in malignant cells, Northern blot analysis of RNA from the different cell lines was performed. Rap1A RNA was detected in 30 g of RNA (Figure 5b), whereas rap1B RNA was detected in 20 g of RNA (Figure 5c). In all cell lines examined, rap1B (Figure 5c) was more strongly expressed than rap1A (Figure 5b). UM-SCC-17B showed the strongest rap1B signal (Figure 5c), when normalized to GAPDH, which corresponds to the Q-RT–PCR findings (Figure 5a). UM-SCC-22B, which has almost exclusively nuclear rap1 staining (Figure 3b), expresses both isoforms strongly (Figure 5b, c).
Rap1 isoforms in the nucleus
To investigate whether one or both rap1 isoforms are localized in the nucleus, wild-type rap1A (pTAR-rap1A) and rap1B (pTAR-rap1B) or the empty vector (pTARGET) were transfected into 3T3 fibroblasts. Endogenous cytoplasmic and nuclear expression was low but detectable in the empty vector (pTARGET)-transfected 3T3 control cells (6E). Both rap1A (Figure 6a) and rap1B (Figure 6c) were strongly expressed in transfected 3T3 fibroblasts. Many cells showed cytoplasmic localization of these two proteins and occasional cells showed both nuclear and cytoplasmic localization of both isoforms on immunohistochemical localization studies (Figure 6a, c, arrows). Interestingly, many of the transfected 3T3 cells exhibit strong perinuclear envelope staining. The corresponding rabbit IgG immunohistochemistry controls were appropriately negative (Figure 6b,d, f). Increased expression of rap1A or rap1B in transfected cells, observed in immunohistochemical studies, was corroborated with immunoblot data (Figure 6g). Rap1A and rap1B appear as doublets in 3T3 transfected cells (6G, arrowheads), which may be because of unprocessed and mature forms of the protein although this observation is also consistent with previous reports that rap1 exists in both phosphorylated and unphosphorylated forms (Lapetina et al., 1989). Note that the endogenous form (Figure 6g lanes pT and C) migrates rapidly and corresponds to the lower of the two bands in the transfected cells.
Both rap1 isoforms localize in the nucleus. 3T3 fibroblasts were transfected with pTAR-rap1A (a and b) or pTAR-rap1B (c and d) or pTARGET empty vector (e and f). Cytospins of transfected cells were fixed and incubated with rabbit anti-rap1 affinity-purified polyclonal antibody (a, c and e) or rabbit IgG (b, d and f) (2 g/ml each). Brown DAB reaction product with hematoxylin counterstain (bar=20 m). Arrows indicate nuclear staining (a and c). The corresponding NP40 cell lysates (40 g protein, each) were electrophoresed and blotted with rabbit anti-rap1 polyclonal antibodies (g) and immunochemiluminescent detectionFull figure and legend (273K)
Nuclear and cytoplasmic localization of rap1A and rap1B were confirmed by immunofluorescence double labeling visualized by confocal microscopy (Figure 7). Rap1A (Figure 7a, arrows) and rap1B (Figure 7c, arrows), detected by FITC-labeled mouse anti-rap1 monoclonal antibody, colocalized with propidium iodide-stained nuclei. The pTARGET vector control also showed a basal level of rap1 in the nucleus (Figure 7e). In immunofluorescence both the transfected and endogenous signal for rap1 appears to be inside the nucleus (Figure 7a, c, e) rather than perinuclear as it appeared in IHC (Figure 6a, c, e). The corresponding mouse IgG immunofluorescence controls were appropriately negative (Figure 7b, d, f). On immunoblots, rap1 in the CE appeared as a doublet with the higher Mr signal showing a greater intensity than the lower Mr signal (Figure 7g). The NE shows only the lower Mr signal. It is possible that the higher molecular mass form in the cytosol represents newly synthesized, unprocessed protein and/or that post-translational modifications such as phosphorylation, or lipid modification such as geranyl geranylation accompany movement of this protein between the cytoplasm and the nucleus. Equivalency of protein loading in CE fractions was confirmed by GAPDH staining (Figure 7g).
Rap1A and rap1B in the nucleus in transfected cells, detection by confocal microscopy. 3T3 fibroblasts were transfected with pTAR -rap1A (a and b) or pTAR-rap1B (c and d) or pTARGET empty vector (e and f), as described. Cells were collected at 24 h post-transfection and cytospins were prepared. The cytospins were fixed and incubated with mouse anti-rap1 monoclonal antibody (a, c and e) or mouse IgG (b, d and f) (10 g/ ml each) (bar=25 m). Arrows indicate transfected cells with nuclear staining (a and c). The two panels in (a) and (c) are two different areas in the same slide. Rap1 in the corresponding nuclear and cytoplasmic extracts (40 g protein, each) was detected by immunoblot analysis with mouse anti-rap1 monoclonal antibody (g). The membrane was subsequently washed and re-exposed to GAPDHFull figure and legend (250K)
Active GTP-bound rap1 localizes in the nucleus
Based on previous observations that ranGTP and ranGDP are primarily in the nucleus and cytoplasm respectively (Yoneda, 2000), we postulated that rap1BGTP is localized in the nucleus whereas rap1BGDP is localized in the cytoplasm. The cDNAs from dominant negative (pTAR-rap1B-N17) and constitutively active rap1B (pTAR-rap1B-V12) were ligated to enhanced GFP (EGFP) in a pEGFP-C1 vector so that exogenously expressed rap1 could be distinguished from the endogenously expressed protein in Western blots or fluorescence. Dominant negative (EGFP-rap1B-N17), constitutively active (EGFP-rap1B-V12) and wild-type (EGFP-rap1B-wt) rap1B were strongly expressed in 3T3 fibroblasts as detected by the GFP antibody (Figure 8a) and by the rap1 antibody (Figure 8b). The EGFP-rap1B fusion protein migrates more slowly than EGFP alone, consistent with the higher molecular mass of the fusion protein (Figure 8a). Endogenous rap1 and EGFP-rap1B fusion proteins are both detected with the rap1 antibody (Figure 8b), except in the EGFP vector control lane where only endogenous rap1 is detected. Endogenous rap1 (Figure 8b) and GAPDH (Figure 8c) show equivalency of loading. Functional activity of the dominant negative, constitutively active and wild-type rap1B protein was assayed with ralGDS (ral guanine nucleotide dissociation stimulator), as described previously (Franke et al., 1997). The rap-binding domain (RBD) of ralGDS binds only rap1GTP but not the inactive GDP-bound form (Franke et al., 1997). As predicted, a stronger rap1 signal was obtained with the active mutant than with the wild-type construct (Figure 8d, V12 and wt, respectively) at equivalent protein loading, whereas no signal was obtained with the dominant negative construct (Figure 8d, N17).
Expression and functional analysis of dominant negative, constitutively active and wild-type rap1B constructs. Dominant negative and constitutively active rap1B constructs were generated by site-directed mutagenesis, as described in Materials and methods. Lysates of 3T3 fibroblasts transfected with dominant negative (N17), constitutively active (V12), or wild type (wt) EGFP-rap1B or empty vector (EGFP) (20 g each) were prepared and immunoblotted with rabbit anti-GFP (a), mouse anti-rap1 (b) or mouse anti-GAPDH antibodies (c). The active form of rap1 in each of these cell groups (500 g each) was retrieved by ral-GDS (d)Full figure and legend (78K)
After confirming the functional status of the constructs, 3T3 fibroblasts were transfected and the localization of exogenously expressed rap1BGTP and rap1BGDP was assessed by confocal microscopy. As shown in Figure 9a, constitutively active rap1B (EGFP-rap1B-V12; middle panel) is expressed in the nucleus and cytoplasm, whereas dominant negative rap1B (EGFP-rap1B-N17; left panel) is expressed primarily in the cytoplasm of 3T3 fibroblasts, including the perinuclear region. As expected, because of its small molecular mass EGFP alone diffuses passively into the nucleus (Zhang et al., 2002), and is present in both the nucleus and cytoplasm (Figure 9a, right panel). Isolation of nuclear and cytoplasmic extracts (NE and CE) from transfected cells shows that endogenous rap1 is present in the nuclear and cytoplasmic fractions (Figure 9b) as shown earlier (Figure 7). The exogenously expressed EGFP-rap1 fusion protein was identified by its larger mass (upward shift in Mr) when blotted with either anti-rap1 (Figure 9b) or anti-GFP antibodies (Figure 9c). Heterogenous nuclear ribonucleoprotein (hnRNP), a nuclear protein (Habelhah et al., 2001), and GAPDH staining were used to assess equivalency of loading of the NE and CE fractions, respectively. Figure 9d shows enrichment of hnRNP, in all the samples designated NE. Thus, we conclude that activation of rap1 by GTP-nucleotide binding promotes nuclear translocation of rap1.
Active, GTP-bound rap1 is localized in the nucleus in 3T3 fibroblasts and SCC cells; detection by confocal microscopy and immunoblot analysis. 3T3 fibroblasts were transfected with EGFP-rap1B-N17 or EGFP-rap1B-V12 or EGFP empty vector (a), as described. Cells were fixed at 24 h post-transfection and EGFP fluorescence was visualized with a confocal microscope (a) (bar=25 m). Nuclear (NE) and cytoplasmic (CE) extracts were prepared from duplicate samples using the NE-PER nuclear extraction kit from Pierce, according to the manufacturer's instructions. These extracts (40 g) were electrophoresed, transferred and blotted with mouse anti-rap1 monoclonal (b), rabbit anti-GFP polyclonal (c), goat anti-hnRNP polyclonal (d) and mouse anti-GAPDH monoclonal (e) antibodies. NE and CE extracts were also prepared from UM-SCC-(14A and 17B). Active GTP-bound (f, upper panel) and total rap1 (f, lower panel) were evaluated in these extractsFull figure and legend (103K)
Based on these observations that exogenously expressed active GTP-bound rap1 localizes in the nucleus in NIH 3T3 fibroblasts, we investigated the distribution of endogenous active rap1 in the nuclear (NE) and cytoplasmic (CE) fractions of SCC cell lines (Figure 9f). Consistent with the preferential localization of rap1-GTP in the nucleus, both UM-SCC-17B, which exhibits equivalent distribution of total rap1 between the nuclear and cytoplasmic fractions, and UM-SCC-14A, which exhibits a stronger total rap1 signal in the nucleus than in the cytoplasm (Figure 9f, lower panel), have a marked enrichment of active GTP-bound rap1 in the nucleus (Figure 9f, upper panel). UM-SCC-14A was further characterized with respect to the effects of serum starvation and feeding on rap1 translocation (Figure 10). In the presence of serum, rap1 translocates to the nucleus whereas under serum-free conditions rap1 is localized almost exclusively in the cytoplasm (Figure 10a). Equivalency of cytoplasmic protein loading was evaluated by GAPDH. Nuclear enrichment of the NE fractions and equivalency of loading was observed by blotting with antihistone antibody (Figure 10a). As shown in Figure 10b, equivalent expression of rap1 is present in whole cell lysates of serum-treated and serum-starved cells, indicating that the change in localization is not secondary to changes in total rap1. The immunoblot data are supported by immunohistochemical studies (Figure 10c–f). In serum-treated cells, rap1 expression is prominent in the nucleus masking the hematoxylin staining of the nuclei (Figure 10d), whereas in the serum-starved cells the rap1 staining is in the cytoplasm (Figure 10c). Mouse IgG controls were appropriately negative (Figure 10e, f). These data are consistent with a role for growth factors in rap1 activation and translocation.
c Serum induces rap1 translocation to the nucleus in SCC cells. Nuclear (NE, 24 g) and cytoplasmic (CE, 30 g) extracts of serum-treated or serum-starved UM-SCC-14A were electrophoresed and blotted with rap1, histone or GAPDH antibodies (a). Whole cell lysates (24 g) from similarly treated cells were also electrophoresed and blotted with rap1 or GAPDH antibodies (b). Cytospins from these treatment groups were fixed and incubated with rap1 monoclonal antibody (c and d) or mouse IgG (e and f, bar=30 m). Arrows indicate nuclear (red arrows) or cytoplasmic (blue arrows) staining in cellsFull figure and legend (229K)
Ras is a small G-protein that is frequently activated by mutations in human and animal cancers (Grunicke and Maly, 1993; Barbacid, 1987). Activation of ras bypasses the upstream signaling pathways from growth factors at the cell surface. Furthermore, activated ras is capable of transforming immortalized cells and is one of the factors necessary to convert normal cells to malignant (Lundberg et al., 2002). In the United States and other Western countries squamous cell carcinomas of the head and neck rarely have activating ras mutations (Clark et al., 1993; Saranath et al., 1991; Jordan and Daley, 1997), which is in sharp contrast to most other carcinomas, including other tobacco carcinogen-induced cancers such as those of the bladder and lung (Mirvish, 1995). This has led to the speculation that some other small G-protein might have a role in the transformation of oral squamous epithelial cells. We have been interested in the expression and function of rap1 in normal keratinocyte biology. Rap1 has a role in differentiation in epidermal keratinocytes (Schmitt and Stork, 2001) and in neuronal cells (Vossler et al., 1997; York et al., 1998). Extending these observations, we recently showed that rap1 inhibits proliferation in normal keratinocytes (manuscript In press). Based on these observations we hypothesized that aberrantly regulated rap1 could play a role in malignant transformation of epithelial cells. Therefore, we set out to examine the expression and functional activities of rap1 in malignant keratinocytes. The work in the present paper reveals that rap1 is expressed to varying degrees in both tissue biopsies of squamous cell carcinomas derived from the upper aerodigestive tract and in cell lines derived from this tumor type. We also report the rather surprising discovery that in these tumor cells, rap1 is frequently expressed predominantly in the nucleus. This makes rap1 only the second of >100 known small G-proteins to be expressed in the nucleus. Furthermore, we show by several methods that it is the GTP-bound, activated form of rap1 that translocates to the nucleus and we show that restricting or supplying growth factors, such as those that are present in serum, can modulate this activation and translocation. We suggest that altered growth factor pathways in squamous tumors are linked to aberrant activation of rap1 and that unscheduled nuclear translocation in response to activated growth factor signaling overrides the usual growth inhibitory effect of this SMG.
Localization of active (GTP-bound) rap1 primarily in the nucleus and inactive (GDP-bound) rap1 primarily in the cytoplasm of SCC cell lines suggests that this protein shuttles between these two compartments. In this regard, ran, the only other SMG that has been identified in the nucleus, regulates nucleocytoplasmic transport by translocating between the nucleus and cytosol by switching between an active and an inactive state (Moore, 2001; Takai et al., 2001). Similarly, in SCC cell lines, rap1GTP is primarily localized in the nucleus, whereas rap1GDP is primarily localized in the cytoplasm. In fact, in UM-SCC-14A nearly all rap1 is localized to the nucleus and is in the GTP-bound active form. In contrast, in transfection studies with normal fibroblasts, we observed rap1BGTP in both the nucleus and cytoplasm, but the presence of active rap1 in the cytoplasm could represent overloading of the cell. Consistent with our observations of rap1 activation and nuclear localization, Ichiba et al. (1999) identified a novel rap1 exchange factor, GFR (guanine nucleotide exchange factor for rap1), in the nuclei of HeLa cells. More recently, Epac, another exchange factor for rap1, was shown to localize to the nucleus during metaphase (Qiao et al., 2002). Hence, the presence of the rap1 exchange factors, GFR or Epac in the nucleus, is consistent with nuclear localization and nucleotide exchange to active rap1GTP.
Mochizuki et al. (2001) have recently shown that variability in EGF or NGF-induced activation of rap1 in the perinuclear region of COS-1 and PC12 cells, respectively, may regulate differentiation or survival responses. These investigators suggested that the differences in rap1 cellular functions may be related to the intracellular site at which the protein is localized and activated. In the nucleus, RanGTP regulates release of cargo proteins from carrier and adapter proteins (Yoneda, 2000; Takai et al., 2001). RanGTP in the nucleus controls the cell cycle by delaying the initiation of mitosis (Moore, 2001). If growth factor activation of rap1 regulates cell survival, as suggested by Mochizuki et al. (2001), this could explain the higher expression of rap1 in squamous carcinoma cells.
In the cytoplasm, rap1 appears to be primarily in the particulate fraction of the cells rather than the cytosol. This suggests that rap1 is only transiently expressed in the cytosolic fraction and must shuttle between or back to membrane-bound organelles, consistent with this protein being geranylgeranylated (Takai et al., 2001). It remains to be determined whether nuclear or cytoplasmic localization requires geranylgeranylation.
In summary, this is the first report of an SMG, other than members of the ran subfamily, being localized in the nucleus. Moreover, it is active rap1GTP that is present in the nucleus whereas rap1GDP is retained in the cytoplasm. Furthermore, growth factors stimulate nuclear translocation and activation of rap1 in tumor cells. These factors are consistent with rap1 playing a role in the uncontrolled growth of human squamous cell carcinomas.
Materials and methods
Oropharyngeal SCC cell lines (Carey, 1994) were grown to 60–80% confluence in Dulbecco's modified Eagle's medium (DMEM, GIBCO, Grand Island, NY, USA) containing 10% fetal bovine serum, penicillin 100 /ml, streptomycin 100 g/ml and 50 g/ml L-glutamine. For the nuclear and cytoplasmic localization studies, the cells were plated in this medium at 2.5 105 cells/100 mm dish for 48 h and then grown for 7 days in DMEM/F12, penicillin, streptomycin and L-glutamine (Zou et al., 1999) in the presence or absence of 5% fetal bovine serum.
Western blot analysis
For the preparation of lysates, UM-SCC cell lines were grown to 60–80% confluence, then washed once with ice-cold PBS and lysed in 1% Nonidet P-40 (NP40) lysis buffer (50 mM Tris-HCl, pH 7.4, 200 mM NaCl, 2 mM MgCl2, with protease inhibitors and 10% glycerol) on ice for 10 min. Particulate material in the lysates was pelleted by centrifugation at 15,700 g for 10 min at 4°C. The supernatant (NP-40 extract) was collected and protein content was measured by the Bio-Rad protein assay (Bio-Rad, Richmond, CA, USA). Equal amounts of protein (20–60 g) were electrophoresed on SDS/12% PAGE gels and transferred to nitrocellulose membranes (Schleicher and Schuell). Prior to blocking, the membrane was stained with Ponceau S for verification of efficient transfer and protein loading. The membranes were incubated in Tris-buffered saline (TBS) containing 5% nonfat dry milk and 0.1% Tween-20 (Bio-Rad) to block nonspecific binding. Membranes were placed in the primary antibody for 1 h at RT or overnight at 4°C. Primary antibody concentrations were: rabbit anti-rap1 affinity-purified polyclonal antibody (1 : 200, Santa Cruz, CA, USA); mouse anti-rap1 monoclonal antibody (1 : 150; Transduction Laboratories), mouse anti-transglutaminase monoclonal antibody (1 : 1000, Neomarkers, Fremont, CA, USA); mouse anti-GFP (green fluorescent protein; Santa Cruz); or mouse anti-glyceraldehyde-3-phosphate dehydrogenase monoclonal antibody (GAPDH, 1 : 5000; Chemicon International, Temecula, CA, USA). Membranes were washed in TBS containing 0.1% Tween-20 (TBS-T). Affinity-purified horseradish peroxidase-linked donkey anti-rabbit or goat anti-mouse IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) as secondary antibodies were used to detect binding of primary antibodies. Visualization of the immunoreactive proteins was accomplished by the chemiluminescence system (Pierce, Rockford, IL) and exposure to X-ray film.
The nuclear and cytoplasmic subcellular fractions of cells were prepared as follows. PBS-washed cells were resuspended in cytoplasmic extraction buffer (CE buffer) (10 mM HEPES, pH 7.6, 60 mM KCl, 1 mM EDTA, 1 mM dithiothreitol and protease inhibitors) to which NP40 was added to a final concentration of 0.2%. After incubation for 3 min on ice, the suspension was centrifuged at 370 g for 5 min to sediment nuclei. The supernatant was further clarified by centrifuging at 15 700 g for 10 min at 4°C and the cytoplasmic extract (CE) was collected. The nuclear pellet was washed gently with CE buffer without NP40 and centrifuged at 370 g for 5 min. Nuclear extract (NE) was obtained by suspending the nuclear pellet in 1% NP40-containing lysis buffer for 15 min on ice with occasional vortexing. Nuclear extract was clarified by centrifuging at 15 700 g for 10 min at 4°C and collecting the supernatant.
The cytosolic and particulate fractions of cells were prepared as follows: Cells were homogenized in homogenizing buffer (10 mM each of HEPES, pH 6.2, NaCl and MgCl2, 0.05 mM EDTA and protease inhibitors) and centrifuged at 375 g. The supernatant was centrifuged at 100 000 g for 1 h. The resultant supernatant was collected as the cytosolic fraction and the pellet as the particulate fraction. The particulate fraction was resuspended in lysis buffer (10 mM Tris-HCl, pH 7.4, 50 mM NaCl, 10 mM MgCl2, 0.05 mM EDTA and protease inhibitors).
Cultured UM-SCC cell lines were trypsinized and cytospins prepared at 8 104 cells per slide. Cytospins or Lab-tek slides were fixed in a mixture of methanol and acetone, and air-dried. The slides were stained for rap1 as follows. Slides were incubated with rabbit anti-rap1 affinity-purified polyclonal antibody or mouse anti-rap1 monoclonal antibody for 1 h at RT. Primary antibody binding was detected using biotinylated donkey anti-rabbit or goat anti-mouse IgG (Biocare Medical, Walnut Creek, CA, USA), followed by incubation with biotin–streptavidin conjugated to peroxidase (Biocare Medical) and the chromogen DAB500 (Biocare Medical) The immunodetection was followed by hematoxylin counterstaining.
After appropriate Human Subjects Institutional Review Board approval, 5 m sections from formalin-fixed, paraffin-embedded human intraoral SCC tissue were deparaffinized and rehydrated. Antigen retrieval was performed with 10 mM sodium citrate buffer, pH 6, at 95°C. The sections were stained with affinity-purified rabbit anti-rap1 polyclonal antibody (2 g/ml). Immunodetection was carried out as described above for cytospins. Under the conditions described here, the mouse anti-rap1 monoclonal antibody could not detect antigens on paraffin-embedded, formalin-fixed tissue.
For double immunofluorescence labeling, 3T3 fibroblasts transfected with pTAR-rap1A or pTAR-rap1B, or empty vector were trypsinized and cytospins prepared. The slides were initially treated with mouse anti-rap1 monoclonal antibody or mouse IgG followed by immunodetection using FITC (fluorescein isothiocyanate)- conjugated goat anti-mouse IgG (1 : 60; Jackson Immunoresearch). The washed slide was subsequently stained with propidium iodide (PI 1 g/ml, RNAse 100 g/ml in PBS containing 0.1% BSA). The cells were visualized on a laser scanning confocal microscope (MRC600, BIORAD). Slides were photographed with a 63X oil immersion lens, digitized and processed with Adobe Photoshop software (ver 5.0, San Jose, CA, USA).
Quantitative real-time Q-RT–PCR
Real-time RT–PCR was performed using the Perkin-Elmer TaqMan system (Heid et al., 1996). Briefly, two sets of isoform-specific primers were designed in exons that spanned the intron–exon boundaries in the region corresponding to the C-termini of rap1A and rap1B. The primers did not amplify genomic DNA and demonstrated isoform specificity when tested against plasmids containing either the rap1A or rap1B cDNA sequences (kind gift from Dr Veronique Pizon, Institut Andre Lwoff, France). The TaqMan probes, which are designed to bind to the middle of the regions being amplified, have reporter and quencher fluorescent dye molecules attached to each probe. During the extension cycle, the reporter dye is cleaved by the nuclease component of the polymerase. This releases the reporter from the quencher, thereby allowing the reporter to fluoresce. This fluorescence is quantified and during the logarithmic phase of the reaction is proportional to the amount of RNA being amplified.
Equivalent concentrations of total RNA from different cell lines were electrophoresed on a 1% formaldehyde–agarose gel, transferred to nylon membranes and immobilized by UV cross-linking. The RNA was hybridized with a 32P-labeled fragment of the rap1A or rap1B cDNA (kind gift of Dr. Veronique Pizon), in 50% formamide, 5 Denhardt's solution, 0.25% SDS, 100 g/ml single–stranded DNA, 5 SSC buffer (3 M sodium chloride, 0.3 M sodium citrate, pH 7.0) and 10% Dextran, overnight at 42°C. Rap1A or rap1B expression was visualized on autoradiographs.
The rap1A and rap1B cDNAs provided by Dr Pizon were inserted into pTARGET (Promega), an efficient mammalian expression vector with a CMV promoter upstream of the multiple cloning site. The newly synthesized plasmids, designated pTAR-rap1A and pTAR-rap1B respectively, were sequenced to verify the presence and orientation of the appropriate cDNA. Constitutively active (i.e. GTP-bound rap1B) and dominant negative rap1B were generated by substitution of Val for Gly at position 12 and Asn for Ser at position 17, respectively, a strategy similar to that described previously (Pizon et al., 1999). These mutations were introduced using the GeneEditor in vitro site-directed mutagenesis system (Promega) with the oligonucleotides 5'-GT-CGTTCTTGGCTCAGTTGGCGTTGGAAAGTG-3' and 5'-GGAGGCGTTGGAAAGAACGCTTTGACTGTAGAA-3', respectively. The mutations in the pTAR-rap1B-V12G and pTAR-rap1B-N17 plasmids were confirmed by sequencing.
Full-length PCR fragments of wild-type, constitutively active and dominant negative rap1B were inserted into pEGFP-C1 vector (Clontech) at Hind III and ECoR1 sites. Primers for PCR were: sense: 5'-CCCAAGCTTGCATCATG-CGTGAGTAT-3' and antisense: 5'-CCGGAATTCCATTTAGTATATTAAAG-3'. The activity of the constructs was validated with ralGDS as described previously (Franke et al., 1997). The rap-binding domain of ral-GDS binds only the active form of rap1 (Franke et al., 1997). The construct for ral-GDS was a generous gift from Dr. Johannes L. Bos (University Medical Centre Utrecht, The Netherlands).
3T3 cells (5 104) seeded in 60 mm dishes were transfected 24 h later with wild-type or constitutively active rap1B-containing plasmids using Superfect (Qiagen, Valencia, CA, USA). The empty vector was used as a control for transfection effects on endogenous gene expression. Cells were collected and tested 24 h after transfection.
- Barbacid M. (1987). Annu. Rev. Biochem., 56, 779–827. | Article | PubMed | ISI | ChemPort |
- Beranger F, Goud B, Tavitian A and de Gunzburg J. (1991). Proc. Natl. Acad. Sci. USA, 88, 1606–1610. | Article | PubMed | ChemPort |
- Berger G, Quarck R, Tenza D, Levy-Toledano S, de Gunzburg J and Cramer EM. (1994). Br. J. Haematol., 88, 372–382. | PubMed | ISI | ChemPort |
- Bokoch GM. (1993). Biochem. J., 289(Part 1), 17–24. | PubMed | ISI | ChemPort |
- Bos JL, de Rooij J and Reedquist KA. (2001). Nat. Rev. Mol. Cell Biol., 2, 369–377. | Article | PubMed | ISI | ChemPort |
- Carey TE. (1994). Head and Neck Tumor Cell Lines. in. Atlas of Human Tumor Cell Lines. R. Hay, A.G., J-G. Park (eds). Academic Press Inc., Harcourt Brace Jovanovich Publishers, pp 79–120.
- Clark LJ, Edington K, Swan IR, McLay KA, Newlands WJ, Wills LC, Young HA, Johnston PW, Mitchell R and Robertson G et al. (1993). Br. J. Cancer, 68, 617–620. | PubMed | ISI | ChemPort |
- D'Silva NJ, DiJulio DH, Belton CM, Jacobson KL and Watson EL. (1997). J. Histochem. Cytochem., 45, 965–973. | PubMed |
- D'Silva NJ, Jacobson KL, Ott SM and Watson EL. (1998). Am. J. Physiol., 274(6 Part 1), C1667–C1673. | PubMed | ChemPort |
- Franke B, Akkerman JW and Bos JL. (1997). EMBO J., 16, 252–259. | Article | PubMed | ISI | ChemPort |
- Grunicke HH and Maly K. (1993). Crit. Rev. Oncog., 4, 389–402. | PubMed | ISI | ChemPort |
- Habelhah H, Shah K, Huang L, Ostareck-Lederer A, Burlingame AL, Shokat KM, Hentze MW and Ronai Z. (2001). Nat. Cell Biol., 3, 325–330. | Article | PubMed | ISI | ChemPort |
- Heid CA, Stevens J, Livak KJ and Williams PM. (1996). Genome Res., 6, 986–994. | Article | PubMed | ISI | ChemPort |
- Ichiba T, Hoshi Y, Eto Y, Tajima N and Kuraishi Y. (1999). FEBS Lett., 457, 85–89. | Article | PubMed | ISI | ChemPort |
- Jordan RC and Daley T. (1997). J. Can. Dent. Assoc., 63, 517–518–521–515.
- Kitayama H, Sugimoto Y, Matsuzaki T, Ikawa Y and Noda M. (1989). Cell, 56, 77–84. | Article | PubMed | ISI | ChemPort |
- Lapetina EG, Lacal JC, Reep BR and Molina y Vedia L. (1989). Proc. Natl. Acad. Sci. USA, 86, 3131–3134. | PubMed | ChemPort |
- Lundberg AS, Randell SH, Stewart SA, Elenbaas B, Hartwell KA, Brooks MW, Fleming MD, Olsen JC, Miller SW, Weinberg RA and Hahn WC. (2002). Oncogene, 21, 4577–4586. | Article | PubMed | ISI | ChemPort |
- Maridonneau-Parini I and de Gunzburg J. (1992). J. Biol. Chem., 267, 6396–6402. | PubMed | ChemPort |
- Mirvish SS. (1995). Cancer Lett., 93, 17–48. | Article | PubMed | ISI | ChemPort |
- Mochizuki N, Yamashita S, Kurokawa K, Ohba Y, Nagai T, Miyawaki A and Matsuda M. (2001). Nature, 411, 1065–1068. | Article | PubMed | ISI | ChemPort |
- Moore JD. (2001). Bioessays, 23, 77–85. | Article | PubMed | ISI | ChemPort |
- Moore MS and Blobel G. (1993). Nature, 365, 661–663. | Article | PubMed | ISI | ChemPort |
- Noda M. (1993). Biochim. Biophys. Acta, 1155, 97–109. | Article | PubMed | ISI | ChemPort |
- Pizon V, Cifuentes-Diaz C, Mege RM, Baldacci G and Rieger F. (1996). Eur. J. Cell Biol., 69, 224–235. | PubMed | ISI | ChemPort |
- Pizon V, Desjardins M, Bucci C, Parton RG and Zerial M. (1994). J. Cell Sci., 107(Part 6), 1661–1670. | PubMed | ISI | ChemPort |
- Pizon V, Mechali F and Baldacci G. (1999). Exp. Cell Res., 246, 56–68. | PubMed |
- Qiao J, Mei FC, Popov VL, Vergara LA and Cheng X. (2002). J. Biol. Chem., 8, 8.
- Saranath D, Chang SE, Bhoite LT, Panchal RG, Kerr IB, Mehta AR, Johnson NW and Deo MG. (1991). Br. J. Cancer, 63, 573–578. | PubMed | ISI | ChemPort |
- Schmitt JM and Stork PJ. (2001). Mol. Cell Biol., 21, 3671–3683. | Article | PubMed | ISI | ChemPort |
- Schnefel S, Zimmerman P, Profrock A, Jahn R, Aktories K, Zeuzem S, Haase W and Schulz I. (1992). Cell Physiol. Biochem., 2, 77–89.
- Takai Y, Sasaki T and Matozaki T. (2001). Physiol. Rev., 81, 153–208. | PubMed | ISI | ChemPort |
- Vossler MR, Yao H, York RD, Pan MG, Rim CS and Stork PJ. (1997). Cell, 89, 73–82. | Article | PubMed | ISI | ChemPort |
- Wagner AC and Williams JA. (1994). Am. J. Physiol., 266(1 Pt 1), G1–14. | PubMed |
- Yoneda Y. (2000). Genes Cells, 5, 777–787. | Article | PubMed | ChemPort |
- York RD, Yao H, Dillon T, Ellig CL, Eckert SP, McCleskey EW and Stork PJ. (1998). Nature, 392(6676), 622–626. | Article | PubMed | ISI | ChemPort |
- Zhang YA, Okada A, Lew CH and McConnell SK. (2002). Mol. Cell Neurosci., 19, 430–446. | PubMed |
- Zou CP, Hong WK and Lotan R. (1999). Differentiation, 64, 123–132. | PubMed |
This work was supported in part by a grant from NIH-NIDCR DE00452-01 (NJD) and the University of Michigan's, Head and Neck SPORE grant (1 P50 CA97248).