Enhancement of EGF- and PMA-mediated MAP kinase ... - Nature

Oncogene (1997) 15, 1437 ± 1444  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Enhancement of EGF- and PMA-mediated MAP kinase activation in cells expressing the human papillomavirus type 16 E5 protein Kerstin Crusius, Eeva Auvinen and Angel Alonso Deutsches Krebsforschungszentrum, Im Neuenheimer Feld-242, 69120 Heidelberg, Germany

In this report we demonstrate that cells expressing the human papillomavirus type 16 E5 open reading frame (HPV16-E5) show a greatly enhanced transcription of the immediate early genes after EGF or PMA treatment compared to control cells. This enhancement is due to ampli®cation of the signal transduction pathways in response to growth factors or phorbol esters. Upon shorttime EGF treatment of the E5-expressing cells we observed an increase in the activation of EGF receptors, resulting in a stronger activation of MAP kinases ERK1/2 compared to control-transfected cells. We also observed that in E5-expressing cells, treatment with PMA results in an increase in membrane-associated PKC activity, and a superactivation of the ERK1/2 MAP kinases. This superactivation is PKC-dependent, since pretreatment of the cells with the PKC inhibitor Ro 31-8220 inhibits MAP kinase activation and early gene transcription almost completely. Furthermore, treatment with genistein strongly reduces the PMAmediated superactivation of ERK1/2 kinases, demonstrating a PKC-mediated, tyrosine kinase-dependent pathway in the superinduction of MAP kinase activation. Thus, HPV16-E5 eects superactivation of MAP kinases over at least two dierent pathways, a PKCmediated, and another, receptor tyrosine-kinase mediated, PKC-independent one. Keywords: EGFR; TGF-a; PMA; TPA; fos; raf

Introduction Protein tyrosine kinase growth factor receptors are believed to function by initiating a phosphorylation cascade which activates target proteins ®nally responsible for transcriptional activation and mitogenic signals (Malarkey et al., 1995; Marshall, 1995). Epidermal growth factor receptor (EGFR) is known to possess an intrinsic tyrosine kinase activity. The receptor is activated upon ligand binding, and after binding to adaptor proteins it is able to activate ras (Mulcahy et al., 1985; Satoh et al., 1990). This activation is followed by a recruitment of raf to the plasma membrane which in turn activates MEK kinases, which ®nally phosphorylate MAP kinases (Gardner et al., 1993; Ahn et al., 1991; Crews et al., 1992, Allessi et al., 1994). These kinases migrate into the nucleus and phosphorylate their transcription factors like c-fos or members of the elk-1 family, which upon binding to the corresponding DNA Correspondence: A Alonso Received 12 September 1996; revised 28 May 1997; accepted 28 May 1997

sequences activate transcription of target genes (Cobb and Goldsmith, 1995). This signal is very rapid and lasts only a few minutes, after which the activation of the receptor is downregulated probably by the action of speci®c phosphatases (Stone and Dixon, 1994; Tomic et al., 1995; Misra-Press et al., 1995). Similar to growth factors, the phorbol ester PMA is able to induce transcription of early genes in ®broblasts (Gillmann, 1988; Harvat and Wharton, 1995; Hill and Treisman, 1995). Since PMA activates protein kinase C (PKC), and this activation seems to result in an inhibition of the EGFR activity (Carpenter and Wahl, 1990), it can be concluded that at least in the ®rst steps of signalling, PMA and EGF use dierent pathways. Furthermore, it has been demonstrated that PKC is able to activate c-raf in vitro and in vivo (Kolch et al., 1993), thus it is possible that both signals converge at c-raf and activate MAP kinases through MEK. In addition, PMA may also activate MAP kinases by other mechanisms independent of the activation of c-raf (Seger et al., 1995). In both cases, activation of elk-1 and immediate early gene (IE gene) transcription is the ®nal result. Human papillomaviruses (HPVs) have been associated with epithelial cancer development (zur Hausen, 1991; zur Hausen and Schneider, 1987, Durst et al., 1983). Whereas e.g. types 6 and 11 are mainly found in benign neoplasias (low-risk types), types 16, 18 and 33 (high-risk types) have been found associated with malignant growth (zur Hausen and Schneider, 1987). Two virus-encoded proteins of the high-risk HPV type 16 (HPV16), E6 and E7, are able to transform ®broblasts and keratinocytes in vitro. Another HPV16 open reading frame, E5, with coding capacity for a strongly hydrophobic protein, seems to be only weakly oncogenic (Leptak et al., 1991; Leechanachai et al., 1992; Pim et al., 1992; Bouvard et al., 1994). This contrasts with the situation observed in bovine papillomavirus type 1 (BPV1), where E5 is the main oncogene and is alone able to transform ®broblasts and keratinocytes (Schiller et al., 1986; DiMaio, 1991). The HPV16-E5 gene product has been mainly found associated with cellular membranes and its expression in immortalized keratinocytes results in a strong impairment of the gap-junction mediated cell-cell communication (Burkardt et al., 1989; Conrad et al., 1993; Oelze et al., 1995 and our unpublished results). Contrary to the BPV1-E5 protein, which has been found to bind with the EGFR, there seems to be no direct association between EGFR and the HPV16-E5 protein (Conrad et al., 1994), although contradictory results have been published (Hwang et al., 1995). Using human keratinocytes transfected with HPV16-E5, Straight et al. (1993) could demonstrate an inhibition of the downregulation of internalized EGF receptors

Signal transduction and HPV16-E5 K Crusius et al

EGF-mediated superinduction of early gene transcription is tyrosine kinase dependent Since both EGF and PMA aect signal transduction through de®ned pathways, we decided to analyse

60 min PMA

45 min PMA

30 min PMA

60 min PMA

15 min PMA

45 min PMA

30 min PMA

Control 15 min PMA

45 min EGF

30 min EGF

Control

45 min EGF

A31-Neo

GAPDH

60 min PMA

45 min PMA

15 min PMA

A31-Neo 60 min PMA

45 min PMA

30 min PMA

15 min PMA

A31-E5

30 min PMA

b

Egr-1

GAPDH

100 nM Vanadate, 30 min EGF



30 min EGF

A31-Neo 100 nM Vanadate, 30 min EGF

A31-E5 50 nM Vanadate, 30 min EGF

A31-Neo


A31-E5

30 min EGF

c

30 min EGF 50 µM Genistein, 30 min EGF 100 µM Genistein, 30 min EGF 300 µM Genistein, 30 min EGF

In a series of experiments we analysed the eect of growth factors and phorbol esters on transcription of the immediate early genes in E5-expressing and control cells. E5- and vector-transfected cells were treated with EGF, and after 30 min total RNA was isolated and Northern blots were prepared. As shown in Figure 1a, transcription of the c-fos gene was greatly enhanced in E5-expressing cells upon 30 min treatment with EGF. A similar eect was seen after TGF-a treatment (results not shown). An enhancement was also found for c-myc transcription, indicating that the early genes were aected in general. These results con®rm and extend those of Leechanachai et al. (1992), who also showed an enhanced c-fos transcription in E5-expressing cells upon treatment with EGF. Furthermore, treatment of the cells with the phorbol ester PMA produced an increase in c-fos (Figure 1a) or Egr-1 (Figure 1b) gene transcription in E5-expressing cells compared to control cells. Quantitation of the hybridization signal by phosphoimaging demonstrated that c-fos transcription in E5-expressing cells was six to seven times more ecient than in the vector-transfectd cells. Thus, enhanced transcription of early genes in response to treatment with EGF, TGF-a, and PMA could be observed in E5-expressing cells.

A31-E5

A31-Neo

c-fos

Results E5 enhances immediate early gene transcription after EGF, TGF-a and PMA treatment

30 min EGF

A31-E5

10 min EGF

a 10 min EGF

with subsequent recycling to the plasma membrane. This mechanism would be mediated by the failure of the membrane-bound proton-ATPase to acidify the endosomal milieu, probably due to a direct binding of the E5 protein to the 16 kD subunit of the enzyme (Straight et al., 1993, 1995). This would result in an increased number of receptors at the plasma membrane and in a longer-lasting activation of the receptor upon long term treatment with EGF. Nevertheless, this mechanism seems not to be sucient to explain the large increase in early gene transcription, and thus, other mechanisms must be involved in the enhancement observed (Straight et al., 1993). Here we report the eect of HPV16-E5 expression on signal transduction. We show that treatment of E5expressing cells with EGF results in an increased activation of the EGF receptors compared to controls and a concomitant MAP kinases ERK1/2 superactivation and increased early gene transcription. Further, we show that treatment of E5-expressing cells with the phorbol ester PMA also results in an increased ERK1/ 2 activation and enhanced early gene transcription. This increase is paralleled by an increased translocation of PKC from the cytosol to the particulate fraction of the cells. Thus, E5 is able to modulate signal transduction by at least two dierent mechanisms, one involving increased EGFR activation and a second dependent of an increased PKC activation.

30 min EGF 50 µM Genistein, 30 min EGF 100 µM Genistein, 30 min EGF 300 µM Genistein, 30 min EGF

1438

Egr-1

GAPDH

Figure 1 Eect of EGF and PMA on early gene transcription in E5-expressing cells. (a) Cells were cultured for 24 h without FCS and then EGF or PMA were added for the time periods indicated. RNA was isolated, separated on agarose gels, blotted and hybridized with a radioactively labelled c-fos DNA probe. After exposure, the blot was rehybridized with a DNA probe speci®c for GAPDH (glyceraldehyde-3-phosphate dehydrogenase) to quantitate for RNA loading dierences. Control, untreated cells. (b) Cells were treated with PMA for 15, 30, 45 and 60 min and then RNA was isolated. Blots were prepared as indicated above and hybridized to a radioactive DNA probe for the transcription factor Egr-1. (c) After culturing the cells for 24 h in the absence of FCS, inhibitors at dierent concentrations were added for 30 min. EGF was added at the concentration described in Materials and methods and after further 30 min RNA was isolated, separated on agarose, blotted onto membranes and hybridized with a radioactive probe for Egr-1


whether enhancement of transcription could be inhibited by sodium orthovanadate or genistein, which are a protein phosphatase inhibitor and a protein tyrosine kinase inhibitor, respectively. Cells were grown to 80% con¯uency and then cultured for 24 h in the absence of serum. Inhibitors were added for 30 min, EGF was added, and the cells were cultured for further 30 min. RNA was isolated, blotted, and hybridized with radioactively labelled probes for c-fos and Egr-1, a zinc-®nger protein known to be a transcriptional activator (Cao et al., 1990). As shown in Figure 1c for Egr-1, genistein produced a clearly dose-dependent inhibition of Egr-1 mRNA transcription, indicating that the eect of HPV16-E5 protein on early gene transcription is mediated by protein tyrosine kinases. On the contrary, orthovanadate has no inhibitory eect on the enhancement of RNA transcription observed in E5-expressing cells. In fact it seems to produce a slight increase in Egr-1 mRNA transcription in both E5- and control-transfected cells, suggesting that phosphatases might be involved in the control of EGF-mediated IE (immediate early) gene transcription (Figure 1c). These results show that EGF-mediated superinduction of early gene transcription in E5-expressing cells occurs by a tyrosine kinase-dependent pathway, and cannot be inhibited by treatment with orthovanadate. EGF receptors are superactivated in E5-expressing cells Since EGF was able to enhance transcription of several early genes, we decided to analyse whether activation of the EGF receptors is modi®ed in E5-expressing cells. E5- and control-transfected cells were treated with EGF, and after dierent time periods total cell extracts were prepared. Immunoprecipitations were performed with antibodies speci®c for phosphotyrosine, and the resulting precipitates were separated on polyacrylamide gels, blotted, and the receptors detected using antibodies speci®cally recognizing the EGF receptor. As shown in Figure 2, after 5 min EGF treatment a clear increase in EGF receptor phosphorylation is seen in E5-expressing cells while no or only a slight increase is observed in control cells. These results demonstrate that treatment with EGF results in an enhanced tyrosine phosphorylation of the EGF receptor in E5expressing cells compared to control cells.

EGF, 30 min

EGF, 15 min

EGF, 5 min

control

A31-Neo EGF, 30 min

EGF, 15 min

EGF, 5 min

kD 213 —

control

A31-E5

EGFR

Figure 2 Activation of EGF receptors in E5-expressing and vector-transfected cells. Cells at 70% con¯uency were cultured in the absence of FCS for 24 h and then EGF was added for the time indicated in the Figure. Cell lysates in RIPA buer were prepared from 46107 cells. Proteins were immunoprecipitated with a phosphotyrosine antibody and Western blotted with an antibody recognizing the mouse EGF receptor. Detection was performed by chemiluminescence. Control, without EGF treatment

1439

EGF- and PMA-treatment of E5-expressing cells lead to superactivation of MAP kinases In ®broblasts, activation of c-fos transcription occurs through binding of the phosphorylated elk-1 protein to the serum response element (SRE) forming a ternary complex (Hill and Treisman, 1995). elk-1 has been shown to be a substrate for activated MAP kinase (Marais et al., 1993; Davis, 1993). Consequently, we wanted to analyse whether the enhanced c-fos/Egr-1 transcription observed in EGF- or PMA-treated E5expressing cells correlates with an enhanced activation of MAP kinases. We therefore studied the activation of ERK1/2 in E5-expressing and vector-transfected cells treated with EGF for dierent periods of time. As shown in Figure 3, treatment with EGF results in an increase of ERK1/2 phosphorylation already after 5 min. An increased phosphorylation is still observed 15 min after addition of the growth factor when no more activation of the MAP kinases can be detected in vector-transfected cells (Figure 3). Thus, E5-expressing cells contain more phosphorylated/activated ERK1/2 protein, and this phosphorylation lasts longer than in the control cells. Moreover, treatment of the cells with PMA results in a longer-lasting activation of ERK1/2 in E5-expressing cells compared to EGF treatment of the cells (Figure 3), suggesting again that EGF and PMA mediate enhanced MAP kinase activation by dierent mechanisms. To further substantiate the results presented above, we used in situ kinase assay to de®ne ERK1/2 superactivation in E5-expressing cells upon treatment with EGF or PMA. Proteins from E5- and controltransfectd cells, treated as described above, were separated on polyacrylamide gels containing 0.25 mg/ ml myelin basic protein (MBP) and in situ kinase assays were performed as described (Leevers and Marshall, 1992). As shown in Figure 4, after 3 min treatment with EGF a strong activation of the kinase activity of proteins migrating at 42/44 kD in E5expressing cells can be observed, whereas only a very weak activation is observed in vector-transfected cells. A similar activation is observed when the cells were treated with PMA (Figure 5). An interesting observation was that the maximum MAP kinase activation was observed 15 min after PMA treatment but only 3 ± 5 min after addition of EGF (compare Figures 4 and 5). Quanti®cation by phosphoimaging revealed a 4 to 5-fold increase in ERK1/2 activity in E5-expressing cells compared to control cells. Our results demonstrate that treatment with EGF or PMA increases the kinase activity of the MAP kinases ERK1/2 in E5-expressing cells compared to control cells. PKC activity is enhanced in membranes of E5transfected cells Whereas the signalling pathway after EGF stimulation is well characterized and comprises activation of the phosphorylation cascade ras-raf-MEK-MAP kinase, the downstream targets in PMA signalling through PKC are less investigated. Since PMA activates PKC and the results presented above demonstrated an activation of c-raf after PMA treatment, we hypothesized that the superactivation of ERK1/2 by PMA might result from a superactivation of PKC and


further signalling downstream by the c-raf-MEK-MAP kinase pathway. To test this hypothesis, we treated E5expressing and control cells with EGF or PMA and measured the amount of activatable PKC bound to the

cellular membranes. As shown in Figure 6, treatment with PMA increases activatable membrane-bound PKC in the E5-expressing cells 3.5-fold using a pseudosubstrate peptide as substrate, whereas in vector-trans-

ERK 2-P-protein

30 min PMA

15 min PMA

5 min PMA

15 min EGF

5 min EGF

Control

30 min PMA

15 min PMA

A31-Neo 5 min PMA

30 min EGF

15 min EGF

5 min EGF

Control

A31-E5

30 min EGF

a

kD 50.6 —

ERK 1-P ERK 2-P

35.5 —

Erk2 protein

15 min PMA

5 min PMA

Control

15 min PMA

5 min PMA

15 min EGF

5 min EGF

Control

15 min EGF

A31-Neo

A31-E5

5 min EGF

b

Figure 3 (a) ERK1/2 activation in E5-expressing cells after EGF/PMA treatment. Cells were cultured as indicated in the legend to Figure 1 and treated with EGF or PMA for 5, 15 or 30 min. Cellular extracts were prepared, and the proteins separated on polyacrylamide gel electrophoresis. Western blotting was performed using antibodies speci®cally recognizing the phosphorylated forms of ERK1/2 MAP kinases. (b) Blots were stripped and reacted with a polyclonal antibody (kindly provided by Dr J Bos) recognizing both the phosphorylated and unphosphorylated forms of ERK2. A phosphorylated form of the ERK2 protein (ERK 2P) was used as control

30 min PMA

15 min PMA

5 min PMA

Control

A31-Neo 30 min PMA

15 min PMA

Control

15 min EGF

5 min EGF

3 min EGF

Control

60 min EGF

A31-Neo 15 min EGF

5 min EGF

3 min EGF

A31-E5

5 min PMA

A31-E5

Control

1440

kD 121 — 84 —

p80

53.2 —

p55 ERK 1/2

Figure 4 EGF-mediated activation of ERK1/2 determined by in situ kinase assay. Cells were cultured as before, treated with EGF for the times indicated and extracts prepared. The proteins were separated on polyacrylamide gels containing 0.25 mg/ml of MBP and assayed for the presence of kinase activity. The low panel shows a histogram with the radioactivity values as quantitated by phosphoimaging, Kontrolle = Control

Figure 5 PMA-mediated activation of ERK1/2 determined by in situ kinase assay. Culture conditions, preparation of extracts, gel electrophoresis, and in situ kinase assay were performed as described in the legend to Figure 4. As in Figure 4, the lower part of the Figure shows the radioactivity values for the ERK1/2 band quantitated by phosphoimaging, Kontrolle = Control


Howsley, 1991). Further, the results presented above suggest that the enhanced activation of MAP kinases in E5-expressing cells is mediated by an increase of activatable membrane-bound PKC. To test whether superinduction of MAP kinases after PMA treatment was completely PKC-dependent, we treated E5expressing and control cells with dierent concentrations of the PKC inhibitor Ro 31-8220 for 30 min, added PMA, and then measured the ERK1/2 activation at dierent times as described in Materials and methods. As shown in Figure 7, treatment with 1 mM Ro 31-8220 abolished almost completely ERK1/2 activation in both E5-expressing and control cells 5 min after addition of the phorbol ester. This inhibition was complete when MAP kinases were assayed 15 min after addition of PMA. Similar results were obtained when the inhibitor concentration was raised up to 3 mM. These results indicate that PMAmediated superinduction of MAP kinase activation in E5-expressing cells is a PKC-dependent process. We wanted to analyse whether EGF-mediated superactivation of ERK1/2 in E5-expressing cells was mediated by protein kinase C. For this, we treated the cells with 1 or 3 mM PKC inhibitor Ro 31-8220 for 30 min and then added EGF for 5 or 15 min. We observed that 1 mM of the inhibitor was unable to block activation of ERK1/2 after 5 min, indicating that superactivation of MAP kinases by EGF in E5expressing cells was not mediated by PKC (Figure 8). A moderate inhibition of ERK1/2 phosphorylation was observed using a 3 mM concentration. Since it has been shown that PKC may activate MAP kinases probably using dierent pathways (Leevers and Marshall, 1992; Robbins et al., 1992), we wanted to establish whether MAP kinase activation after PMA treatment is mediated by tyrosine kinases. E5-expressing and control cells were treated with genistein for 30 min, PMA was then added and MAP-kinase activation was measured. In both E5expressing and control cells activation of ERK1/2 was

fected cells only a twofold activation can be observed. Similar results were obtained using a MARCKS peptide as substrate (results not shown). Thus, in E5expressing cells more PKC is translocated into the membranes, indicating an increased PKC activity. This superinduction would probably cause an enhanced downstream signalling, resulting in increased activation of ERK1/2 MAP kinases. Finally, it should be pointed out that the observed increased PKC activity is not due to increased synthesis of PKC, since no appreciable dierences in the total amount of PKC a and b isotypes were found between E5-expressing and control-transfected cells (results not shown). MAP kinase superactivation by PMA is PKC- and tyrosine kinase-dependent It has been shown that PMA eects signalling through a PKC-mediated pathway (Gilmore and Martin, 1983;

Figure 6 PKC translocation into membranes in EGF- and PMA-treated cells. Cells were cultured for 24 h in the absence of FCS, treated with EGF or PMA for 10 min and membranes prepared as described in Materials and methods. PKC activity was measured using a pseudosubstrate peptide as substrate. Abscissa, radioactivity in c.p.m./mg protein. Black columns, E5expressing cells. Empty columns, vector-transfected cells. The experiment was repeated three times with similar results, Eth = ethanol

ERK 2-P-protein

3 µM Ro 31-8220, 15 min PMA

3 µM Ro 31-8220, 5 min PMA

1 µM Ro 31-8220, 15 min PMA

1 µM Ro 31-8220, 5 min PMA

5 min PMA

3 µM Ro 31-8220

1µM Ro 31-8220

Control

3 µM Ro 31-8220, 15 min PMA

3 µM Ro 31-8220, 5 min PMA

A31-Neo 1 µM Ro 31-8220, 15 min PMA

1 µM Ro 31-8220, 5 min PMA

5 min PMA

3 µM Ro 31-8220

1 µM Ro 31-8220

Control

A31-E5

kD 50.6 —

35.5 —

ERK 1-P ERK 2-P

Figure 7 Eect of the PKC inhibitor Ro 31-8220 on the PMA-mediated MAP kinase activation. E5-expressing or controltransfected cells were treated with 1 or 3 mM Ro 31-8220 for 30 min and then with PMA for 5 or 15 min. Total protein extracts were prepared, separated on polyacrylamide gels and Western blotted as described in the legend to Figure 3

1441


1442

ERK 2-P-protein

3 µM Ro 31-8220, 15 min EGF

3 µM Ro 31-8220, 5 min EGF

1 µM Ro 31-8220, 15 min EGF

1 µM Ro 31-8220, 5 min EGF

5 min EGF

3 µM Ro 31-8220

1mM Ro 31-8220

Control

3 µM Ro 31-8220, 15 min EGF

3 µM Ro 31-8220, 5 min EGF

A31-Neo 1 µM Ro 31-8220, 15 min EGF

1 µM Ro 31-8220, 5 min EGF

5 min EGF

3 µM Ro 31-8220

1 µM Ro 31-8220

Control

A31-E5

kD 50.6 —

35.5 —

ERK 1-P ERK 2-P

Figure 8 Eect of Ro 31-8220 on the EGF-mediated MAP kinase activation. Cells were treated as described before with Ro 318220 for 30 min, EGF was added for 5 or 15 min, and extracts were prepared and processed as indicated in the legend to Figure 3

strongly decreased. This inhibition was already observed 5 min after PMA addition and did not change after 15 min treatment (results not shown). These results demonstrate that genistein-sensitive tyrosine kinases were involved in PMA-mediated signalling. Discussion The results presented in this report demonstrate that EGF and PMA produce a strong activation of MAP kinase and immediate early gene transcription in E5expressing cells but not in vector-transfected cells. We also found an increased EGFR activation upon ligand treatment, suggesting a direct relationship between E5mediated EGF receptor activation, MAP kinase activation, and enhanced early gene transcription. Furthermore, the fact that genistein was able to block transcription almost completely favours a mechanism coupled to tyrosine kinases. This agrees with the inability of indomethacin to downregulate enhanced transcription of early genes in E5-expressing cells, indicating that cyclooxygenase metabolites are not involved in the E5-mediated upregulation of transcription (our unpublished results). This contrasts again with the situation found in 3T3 ®broblasts, in which transcription of Egr-1 and c-fos mRNA after arachidonic acid treatment is cyclooxygenase-dependent and can be inhibited by pretreatment of the cells with indomethacin (Danesch et al., 1994). An increase in receptor phosphorylation after long term EGF treatment has been described in E5transfected keratinocytes (Straight et al., 1993). These authors concluded that the increase was the result of increased recycling of the receptor to the plasma membrane, due to the inability of the proton-ATPase in the endosomal membrane to acidify the endosomal milieu (Straight et al., 1995). This inability would be mediated by binding of the 16 kD subunit of the

proton pump ATPase to the E5 protein, leading to a reduction in the downregulation of the receptors. Binding of the 16 kD subunit to the BPV1 and HPV16-E5 proteins has been demonstrated in vitro and is probably responsible for the inhibition of the normal function of the proton-ATPase (Goldstein et al., 1992; Faulkner Valle and Banks, 1995). Nevertheless, other mechanisms must be involved in the enhanced activation of the receptor. From one side, we did not observe increased amounts of total receptor at the plasma membrane after EGF treatment (results not shown), and from the other side, the early time points at which we detected the increased phosphorylation argue against a role of recycling alone in increasing the receptor amount and consequently the activity. In this context, it has been recently reported that the HPV16-E5 protein may complex with the EGF receptor when overexpressed in COS cells (Hwang et al., 1995), although other authors have been unable to demonstrate such complexes (Conrad et al., 1994). Since the increased transcription of immediate early genes produced by E5 could be inhibited by genistein, tyrosine kinases must be involved in the upregulation of early gene transcription. It has been reported that speci®c phosphatases might be involved in downregulation of the activated receptors (Tomic et al., 1995 and references therein). Thus, one mechanism that can be envisaged is the inhibition of these phosphatases by an E5-mediated mechanism producing the observed increase in receptor activation. Our results on the eect of orthovanadate on EGFR activation support this hypothesis. Superactivation of the EGF receptors would produce an increased signalling, resulting in a superactivation of the MAP kinases ERK1/2 as demonstrated in our experiments. An interesting ®nding was that treatment of E5expressing cells with the phorbol ester PMA also leads to an increased ERK1/2 activation and early gene transcription compared to control cells. Since the target


of PMA is protein kinase C, and activation of PKC results in threonine phosphorylation and consequent inhibition of EGF receptor activity (Carpenter and Wahl, 1990), there must exist an additional E5mediated mechanism separate from the response to EGF, which is responsible for the increased signalling observed. Our results showed that PMA-treatment of E5-expressing cells clearly increases membrane-bound PKC activity. Translocation of PKC to the membranes has been demonstrated to be necessary for enzymatic activation (Newton, 1995 and references therein). Thus, PMA is able to augment the amount of activatable PKC in E5-expressing cells, resulting in an increased downstream signalling. It has been previously shown that PKC is able to activate c-raf in vitro and in vivo (Misra-Press et al., 1995). Since increase in early gene transcription after PMA treatment can be inhibited with genistein, protein tyrosine kinases must also be involved in the enhancement observed. Thus, one may explain enhanced ERK1/2 activation and early gene transcription after PMA treatment by an E5mediated relocation of PKC to the membranes, subsequent activation and increased phosphorylation of c-raf. At this point both signals coming from EGF and those from PMA converge, resulting in a process which is PKC- and tyrosine kinase-dependent, as observed in our experiments. Indeed, we could demonstrate that both EGF and PMA treatment of E5-expressing cells produced an enhanced activation of c-raf and MEK, resulting in the overactivation of the MAP kinases (results not shown). This overactivation would ®nally be responsible for the enhanced transcription of the c-fos and Egr-1 genes. Interestingly, MAP kinase activation seems to be speci®c for the E5 protein of the human papillomaviruses, since in ®brosarcoma cells transfected with HPV 16 E6 or E7 open reading frames no enhancement of ERK1/2 phosphorylation could be demonstrated (Gu and Matlashewski, 1995). An interesting point is that treatment of E5expressing cells with genistein inhibited PMA-induced MAP kinase activation almost completely. Since activation of ERK1/2 is accomplished by the dual speci®city kinase MEK, we might conclude that PKC is probably using this pathway and that other, MEKindependent, genistein-insensitive pathways for the activation of ERK1/2 are not involved. Such a MEK-independent pathway has been suggested for insulin-treated ®broblasts (Seger et al., 1995). In conclusion, HPV16-E5 is able to enhance MAP kinase activation and early gene transcription by at least two dierent mechanisms, one mediated by an upregulation of the EGF receptor activation upon ligand treatment, and another by an increase in membrane-bound activatable PKC. Further experiments are required to elucidate the molecular mechanisms responsible for these eects. Materials and methods Materials EGF, TGF-a, proteinase inhibitors and anti-phosphotyrosine antibodies were obtained from Boehringer Mannheim. Phorbol 12-myristate 13-acetate, sodium

orthovanadate, genistein, and phosphatidylserine were obtained from Sigma Munich. g-32P-ATP (5000 Ci/mmol) and the ECL Western blot detection system were from Amersham. Antibodies to the EGF receptor were from Santa Cruz Biotechnology (Heidelberg, Germany). The PKC inhibitor Ro 31-8220 was from Calbiochem (Bad Soden, Germany). Cells and cell culture A31-3T3 ®broblasts transfected with the HPV16-E5 open reading frame under the control of the MLV-LTR promoter (A31-E5 cells) or with the pJ4 empty vector (A31-Neo cells) have been described and were kindly provided by Dr G Matlashewski. EGF, TGF-a, or PMA were added to ®nal concentrations of 25 ng/ml, and 561077 M, respectively. Immunoprecipitation and immunoblotting Activated EGF receptors were measured by immunoprecipitation and Western blotting. Extracts from 46107 E5expressing or control cells were prepared using RIPA buer containing proteinase inhibitors, and immunoprecipitation was performed with antibodies to phosphotyrosine or to the EGF receptor. After washing the immunoprecipitate, the proteins were resuspended in Laemmli buer and separated on 7.5% polyacrylamide gels. The proteins were electrotransferred onto PVDF membranes and the blots were incubated with antibodies recognizing either the EGF receptor or phosphotyrosine and detected by the ECL detection system (Amersham). Activation of MAP kinases Analysis of the phosphorylation of MAP kinases ERK1/2 was performed using the Phosphoplus kit of New England Biolabs. Total cellular extracts in Laemmli buer were prepared, proteins separated on 15% polyacrylamide gels and Western blotted with antibodies which recognize either the total ERK1/2 proteins or the phosphorylated forms only. To demonstrate the activation status of the ERK1/2 MAP kinases, in situ kinase assays were performed (Leevers and Marshall, 1992). Total cellular proteins were prepared by heating them up to 958C in Laemmli buer. Proteins were separated on 15% polyacrylamide gels containing 0.25 mg/ml of myelin basic protein. The gels were incubated with g-32P-ATP (10 mCi/ml) in a phosphorylation buer (40 mM HEPES, pH 8.0, 20 mM MgCl2, 1 mM DTT, 0.1 mM EDTA, 100 mM sodium vanadate, 30 mM ATP), washed, dried, and exposed to X-ray ®lms. The amount of incorporated radioactivity was quantitated by phosphoimaging. Measurement of PKC activity Membrane-bound PKC activity was calculated as described (Chakrabarti et al., 1991). Brie¯y, membranes were prepared after disruption of the cells by a hypotonic shock (1 mM NaHCO3, 5 mM MgCl2, 100 mM PMSF), and PKC activity was measured in the absence of phosphatidylserine in a buer containing 50 mM Tris, pH 7.5, 10 mM MgCl2, 2 mM CaCl2, 200 mM sodium vanadate, 200 mM sodium pyrophosphate, 2 mM NaF, 200 mM PMSF, 0.2 mCi g-32P-ATP and the substrate peptides. After incubation at 258C for 10 min, incorporated radioactivity was counted in a liquid scintillation counter. Two dierent peptide substrates, one corresponding to a serine-mutant of the PKC pseudosubstrate (RFARKGSLRQKNV), and a second corresponding to the PKC target of the MARCKS protein (FKKSFKL), were used.

1443


1444

Other methods RNA isolation, blotting and Northern hybridization was performed using standard methods (Sambrook et al., 1989).

Acknowledgements We greatly appreciate the excellent technical assistance of J Fischer.

References Ahn NG, Seger R, Bratlien RL, Dieltz CD, Tonks NK and Krebs EG. (1991). J. Biol. Chem., 266, 4220 ± 4227. Alessi DR, Saito Y, Campbell DG, Cohen P, Sithanandam G, Rapp U, Ashworth A, Marshall CJ and Cowley S. (1994). EMBO J., 7, 1610 ± 1619. Bouvard V, Matlashewski G, Gu Z, Storey A and Banks L. (1994). Virology, 203, 73 ± 80. Burkhardt J, Willingham M, Gay C, Jeang K and Schlegel R. (1989). Virology, 170, 334 ± 339. Cao X, Koski RA, Gashler A, McKiernan M, Morris CF, Ganey R, Hay RV and Sukhatme VP. (1990). Mol. Cell Biol., 10, 1931 ± 1939. Carpenter G and Wahl MI. (1990). The epidermal growth factor family. In pp 69 ± 171, Springer-Verlag. Berlin. Chakrabarti BR, Bussey A, Whit®elf JF, Sikorska M, Williams RE and Durkin JP. (1991). Anal. Biochem., 196, 144 ± 150. Cobb MH and Goldsmith EJ. (1995). J. Biol. Chem., 270, 14843 ± 14846. Conrad M, Bubb VJ and Schlegel R. (1993). J. Virol., 67, 6170 ± 6178. Conrad M, Goldstein D, Andresson T and Schlegel R. (1994). Virology, 200, 796 ± 800. Crews CM, Alessandrini A and Erikson RL. (1992). Science, 258, 478 ± 480. Davis R. (1993). J. Biol. Chem., 268, 14553 ± 14556. Danesch U, Weber PC and Sellmayer A. (1994). J. Biol. Chem., 269, 27258 ± 27263. DiMaio D. (1991). Adv. Cancer Res., 56, 133 ± 159. DuÈrst M, Gissmann L, Ikenberg H and zur Hausen H. (1983). Proc. Natl. Acad. Sci. USA, 80, 3812 ± 3815. Faulkner Valle G and Banks L. (1995). J. Gen. Virol., 76, 1239 ± 1245. Gardner AM, Vaillancourt RR and Johnson GL. (1993). J. Biol. Chem., 268, 17896 ± 17901. Gillmann MZ. (1988). Genes & Develop., 2, 394 ± 402. Gilmore T and Martin GS. (1983). Nature, 306, 487 ± 492. Goldstein DT, Kulke R, DiMaio D and Schlegel R. (1992). J. Virol., 66, 405 ± 413. Gu Z and Matlashewski G. (1995). J. Virol., 69, 8051 ± 8056. Harvat BL and Wharton W. (1995). Cell Growth & Di., 6, 955 ± 964. Hill CS and Treisman R. (1995). EMBO J., 14, 5037 ± 5047. Howsley MD. (1991). Eur. J. Biochem., 195, 9 ± 27. Hwang ES, Nottoli T and DiMaio D. (1995). Virology, 211, 227 ± 233. Kolch W, Heidecker G, Kochs G, Hummel R, Vahidi H, Mischak H, Finkenzeller G, MarmeÂ D and Rapp UR. (1993). Nature, 364, 249 ± 252.

Leechanachai P, Banks L, Moreau F and Matlashewski G. (1992). Oncogene, 7, 19 ± 25. Leptak C, RamoÂn y Cajal S, Kulke R, Horwitz BH, Riese II DJ, Dotto GP and DiMaio D. (1991). J. Virol., 65, 7078 ± 7083. Leevers SJ and Marshall CJ. (1992). EMBO J., 11, 569 ± 574. Malarkey K, Belham CM, Paul A, Graham A, McLees A, Scott PH and Plevin R. (1995). Biochem. J., 309, 361 ± 375. Marais R, Wynne J and Treisman R. (1993). Cell, 73, 381 ± 393. Marshall CJ. (1995). Cell, 80, 179 ± 185. Misra-Press A, Rim CS, Yao H, Roberson MS and Stork PJS. (1995). J. Biol. Chem., 270, 14587 ± 14596. Mulcahy LS, Smith MR and Stacey DW. (1985). Nature, 313, 241 ± 243. Newton AC. (1995). J. Biol. Chem., 270, 28495 ± 28498. Oelze I, Kartenbeck J, Crusius K and Alonso A. (1995). J. Virol., 69, 4489 ± 4494. Pim D, Collins M and Banks L. (1992). Oncogene, 7, 27 ± 32. Robbins DJ, Cheng M, Zhen E, Vanderbildt C, Fieg LA and Cobb MH. (1992). Proc. Natl. Acad. Sci. USA, 89, 6924 ± 6928. Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular Cloning. Cold Spring Harbor Laboratory Press, CSH. Satoh T, Endo M, Nakafuku M, Akiyama T, Yamamoto T and Kaziro Y. (1990). Proc. Natl. Acad. Sci. USA, 87, 7926 ± 7929. Seger R, Biener Y, Feinstein R, Hanoch T, Gazit A and Zick Y. (1995). J. Biol. Chem., 270, 28325 ± 28330. Schiller JT, Vass WC, Vousden KH and Lowy DR. (1986). J. Virol., 57, 1 ± 6. Straight SW, Hinkle PM, Jewers RJ and McCance DJ. (1993). J. Virol., 67, 4521 ± 4532. Straight SW, Herman B and McCance DJ. (1995). J. Virol., 69, 3185 ± 3192. Stone RL and Dixon JE. (1994). J. Biol. Chem., 269, 31323 ± 31326. Tomic S, Greiser U, Lammers R, Kharitonenkov A, Imyanitov E, Ullrich A and BoÈhmer FD. (1995). J. Biol. Chem., 270, 21277 ± 21284. zur Hausen H. (1991). Virology, 184, 9 ± 13. zur Hausen H and Schneider A. (1987). The role of papillomaviruses in human anogenital cancers. In The Papovaviridae. Salzman NP and Howley P, eds. pp 245 ± 263. Plenum Press, New York.