Our previous data demonstrated that Ras activation was necessary and sufficient for transforming growth factor-β (TGFβ)-mediated Erk1 activation, and was required for TGFβ up-regulation of the Cdk inhibitors (CKI's) p27Kip1 and p21Cip1 (KM Mulder and SL Morris, J. Biol. Chem., 267, 5029 – 5031, 1992; MT Hartsough and KM Mulder, J. Biol. Chem., 270, 7117 – 7124, 1995; MT Hartsough et al., J. Biol. Chem., 271, 22368 – 22375, 1996 and J Yue et al., Oncogene, 17, 47 – 55, 1998). Here we examined the role of Ras in TGFβ-mediated effects on a rat homolog of Smad1 (termed RSmad1). We demonstrate that both TGFβ and bone morphogenetic protein (BMP) can induce endogenous Smad1 phosphorylation in intestinal epithelial cells (IECs). The combination of transient expression of RSmad1 and TGFβ treatment had an additive effect on induction of the TGFβ-responsive reporter 3TP-lux. Either inactivation of Ras by stable, inducible expression of a dominant-negative mutant of Ras (RasN17) or addition of MAP and ERK kinase (MEK) inhibitor PD98059 to cells significantly decreased the ability of both TGFβ and BMP to induce phosphorylation of endogenous Smad1 in IECs. Moreover, either inactivation of Ras or addition of PD98059 to IEC 4-1 cells inhibited the ability of RSmad1 to regulate 3TP luciferase activity in both the presence and absence of TGFβ. Collectively, our data indicate that TGFβ can regulate RSmad1 function in epithelial cells, and that the Ras/MEK pathway is partially required for TGFβ-mediated regulation of RSmad1.
Transforming growth factor β (TGFβ) is an endogenous cytokine that regulates cell growth, differentiation, adhesion, early embryonic development, and extracellular-matrix protein synthesis (Hartsough and Mulder, 1997). The signaling receptors for TGFβ (RI and RII) and other related factors are expressed at the cell surface; they are activated when they form a heteromeric complex and are phosphorylated at Ser/Thr residues (Hartsough and Mulder, 1997; Derynck, 1994; Massagué, 1996). Recently, genetic studies in Drosophila have resulted in the identification of genes required for the function of the TGFβ superfamily member decapentaplegic (Dpp) (Raftery et al., 1995). Mammalian homologs of these signaling factors for Dpp, termed Sma and Mad homologs (Smads) have also been cloned (Hartsough and Mulder, 1997; Massagué, 1996; Derynck and Zhang, 1996; Heldin et al., 1997; Wrana and Attisano, 1996; Yue et al., in press). However, the precise mechanism for regulation of the Smads by the TGFβ superfamily members has not been fully elucidated.
Previously, we reported that TGFβ growth inhibitory activity was associated with a rapid and direct activation of Ras and of Erk1 in TGFβ-sensitive untransformed epithelial cells (Mulder and Morris, 1992; Hartsough and Mulder, 1995). Moreover, TGFβ affected a sustained activation of Erk2 and of stress-activated protein kinases (SAPKs)/c-Jun N-terminal kinases (JNKs) that was directly correlated with the growth inhibitory effects of TGFβ in human breast cancer cells (Frey and Mulder, 1997a,b). We have also reported that stable, inducible expression of a dominant-negative mutant of Ras (RasN17) resulted in a reversal of the ability of TGFβ to decrease both Cdk2 activity and cyclin A protein expression in intestinal epithelial cells (IECs) (Hartsough et al., 1996); RasN17 completely blocked TGFβ-mediated activation of Erk1 (Hartsough et al., 1996) and up-regulation of the CKIs p21Cip1 and p27Kip1 (Yue et al., 1998).
In light of our previous data and the evidence demonstrating the importance of the Smad pathway for TGFβ signaling, it was of interest to determine whether the Ras/MEK pathway was involved in TGFβ-mediated regulation of RSmad1 in IEC 4-1 cells. Our data indicate that TGFβ can phosphorylate endogenous Smad1, and that expression of RSmad1 can stimulate the TGFβ-responsive reporter 3TP-Lux in the presence of TGFβ. More importantly, however, our results demonstrate that inactivation of the Ras/MEK pathway significantly decreased both TGFβ-mediated phosphorylation of Smad1 and RSmad1-induced 3TP luciferase activity. Thus, TGFβ can signal through RSmad1, and cross-talk between the Ras/MEK and RSmad1 signaling pathways for TGFβ exists.
The requirement of Ras for phosphorylation of Smad1 by TGFβ
Previous studies involving expression of exogenous, epitope-tagged Smad1 suggested that Smad was generally involved in BMP, but not TGFβ, signaling (Liu et al., 1996; Kretzschmar et al., 1997a; Hoodless et al., 1996). However, it is not clear whether endogenous Smad1 is regulated by TGFβ in TGFβ-sensitive epithelial cells. We have cloned a rat homolog of Smad1 (RSmad1) by screening a cDNA library constructed from the highly TGFβ-sensitive, untransformed rat IEC 4-1 cell line (Yue et al., 1999; Mulder et al., 1993). We were interested in whether TGFβ could regulate endogenous Smad1 phosphorylation in the IEC 4-1 cells from which the cDNA library was constructed. For these experiments, we used BMP2 as a positive control.
The specificity of the anti-Smad1 antibody used in our studies has been described previously (Kretzschmar et al., 1997a). We have also shown that this Smad1 antibody does not recognize RSmad1 mutated at the four Erk consensus phosphorylation sites in the linker region of Smad1, although it recognizes wild-type RSmad1 (data not shown). Our result may indicate that this Smad1 antibody recognizes a region near these sites. Similar sites are present in Smads 5 and 8. Thus, these data suggest that this Smad1 antibody does not cross-react with Smads other than Smads 5 or 8, and these two Smads migrate at different positions than does Smad1 (Chen et al., 1997; Nishimura et al., 1998).
As shown in Figure 1a, both TGFβ3 and BMP2 treatment resulted in an increase in phosphorylation of endogenous Smad1 after 15 – 30 min of ligand addition to IEC 4-1 cells. No basal level of Smad1 phosphorylation was observed in these cells. Hence, the results in Figure 1a demonstrate that in addition to BMP2, TGFβ3 can result in a rapid phosphorylation of endogenous Smad1. Similar results were obtained with the TGFβ1 isoform. Moreover, incubation of IEC 4-1 cells with [125I]TGFβ1, followed by chemical cross-linking with disuccinimidyl suberate, demonstrated the presence of RI and RII type receptors in anti-Smad1 immunocomplexes (data not shown). Thus, our data suggest that TGFβ phosphorylation of endogenous Smad1 is mediated by TGFβ-specific binding proteins. Although two groups previously reported that TGFβ could phosphorylate endogenous Smad1 (Lechleider et al., 1996; Yingling et al., 1996), the antibodies employed in their systems were not specific for Smad1.
As mentioned previously, we have demonstrated that Ras was partially required for the effects of TGFβ on Cdk2 activity and cyclin A expression, whereas Ras was essential for TGFβ-mediated activation of Erk1 and up-regulation of the CKIs p21Cip1 and p27Kip1 (Hartsough et al., 1996; Yue et al., 1998). Hence, there appears to be a tight link between the cell cycle effects of TGFβ and activation of the Ras/MAPK pathway. Since activation of the Smad pathway has been implicated in TGFβ-mediated growth inhibition (Liu et al., 1997), it was of interest to examine whether the Ras/MEK pathway was required for TGFβ-induced phosphorylation of endogenous Smad1 in IEC 4-1 cells.
Previously, we stably transfected the IEC 4-1 cells with a dominant-negative mutant of Ras (RasN17) under the control of an inducible metallothionein promoter, and selected several positive clones (E3, C5 and C6) (Hartsough et al., 1996). Here, we cultured the E3 cells in the presence or absence of ZnCl2 for 36 h to induce RasN17 expression and then treated the cells with TGFβ3 for 30 min. As shown in Figure 1b, both TGFβ3 and BMP2 resulted in a significant increase in endogenous Smad1 phosphorylation in the absence of ZnCl2. Moreover, in similarity to the results observed in the parental IEC 4-1 cells, no basal Smad1 phosphorylation was observed in the E3 cells. In contrast, in the presence of ZnCl2, phosphorylation of Smad1 was significantly decreased by 85% or 58% after treatment with either TGFβ or BMP, respectively. However, despite the loss of endogenous Ras function, some Smad1 phosphorylation was still observed after ligand addition. Our results suggest that Ras/MEK-independent mechanisms also contribute to the phosphorylation of Smad1 by TGFβ or BMP. Thus, both ligands can phosphorylate endogenous Smad1, and Ras/MEK pathway is partially required for these ligand-dependent events.
Requirement of Ras for Smad1-mediated 3TP-Lux regulation
The current model for activation of the Smad1 pathway suggests that ligand-dependent phosphorylation may lead to regulation of gene transcription in the nucleus (Hartsough and Mulder, 1997; Heldin et al., 1997). Accordingly, it was on interest to determine whether Smad1 could regulate transcriptional events and if so, whether the Ras/MEK pathway could participate in such events. For these studies, cells were transiently transfected with the RSmad1 that we cloned from an IEC cDNA library (Yue et al., 1999), together with the TGFβ-responsive luciferase reporter, p3TP-Lux (Wrana et al., 1992). 3TP-Lux contains three AP-1 sites and 400 bp of the PAI-1 promoter (Wrana et al., 1992). Although 3TP-lux does not represent a minimal Smad regulated site, and is regulated by growth factors in addition to TGFB through the Smad components (Yingling et al., 1997; De Caestecker et al., 1998), it is still a useful indicator of general TGFβ responsiveness.
As shown in the left panel of Figure 2, TGFβ increased 3TP luciferase activity, and ZnCl2 treatment, alone, did not cause any significant change in 3TP luciferase activity either in the presence or absence of TGFβ. As shown in the right panel of Figure 2, expression of RSmad1 by transient transfection potentiated the ability of TGFβ to induce 3TP luciferase activity in RasN17 E3 cells. Moreover, the induction of RasN17 by ZnCl2 treatment decreased the ability of RSmad1 to induce 3TP luciferase activity in the presence of TGFβ by 40% (Figure 2, right panel). These results indicate that Ras is partially required for the RSmad1-mediated 3TP luciferase induction in intestinal epithelial cells.
Requirement of MEK for Smad1 regulation by TGFβ
It was also of interest to determine whether MEK1, a downstream component of Ras, was involved in Smad1 phosphorylation and RSmad1-mediated transcriptional regulation by TGFβ. For these studies, we employed the MEK1 inhibitor PD98059 to block Erk1 activation by TGFβ in IEC 4-1 cells. As shown in Figure 3a, Erk1 was activated by twofold within 30 min of TGFβ3 addition to IEC 4-1 cells. Addition of the MEK1 inhibitor PD98059, at concentrations of either 10 μM or 15 μM, completely inhibited the ability of TGFβ3 to activate Erk1. Thus, PD98059, at a concentration of 10 μM could be used to determine whether MEK1 was involved in Smad1 regulation by TGFβ. Similar results were obtained for BMP (data not shown).
We have examined the effect of PD98059 on Smad1 phosphorylation by TGFβ and BMP. As shown in Figure 3b, addition of the MEK1 inhibitor PD98059 (10 μM) resulted in a 57% or 65% inhibition of Smad1 phosphorylation by TGFβ or BMP treatment, respectively. Hence, our data indicate that MEK is partially required for Smad1 phosphorylation by both ligands.
We have also examined whether MEK was involved in RSmad1-mediated transcriptional regulation. As depicted in the left panel of Figure 3c, TGFβ3 increased 3TP luciferase activity, and PD98059, alone, did not have a significant effect on 3TP luciferase activity either in the absence or presence of TGFβ. As shown in the right panel of Figure 3c, expression of RSmad1 in the parental IEC 4-1 cells induced 3TP luciferase activity, and potentiated the ability of TGFβ to induce 3TP activity. Addition of the MEK1 inhibitor PD98059 decreased the ability of RSmad1 to induce 3TP activity both in the absence and presence of TGFβ, by 30% or 40%, respectively (Figure 3c, right panel). These data demonstrate that MEK1 is also partially required for RSmad1-induced 3TP luciferase activity in IEC 4-1 cells. The ability of both TGFβ and RSmad1 to regulate 3TP luciferase activity in the IEC 4-1 cells (Figure 3c) was greater than that in the N17E clone (Figure 2). This difference may be due to leaky expression of the RasN17 in the E3 clone.
It is noteworthy that both RasN17 and PD98059 were able to inhibit 3TP-Lux activity induced by TGFβ addition in the presence of exogenous RSmad1 expression, yet blockade of this induction was not observed in the absence of exogenous RSmad1 ex-pression. These results suggest two additional points of potential interest. First, our current results are similar to our previous data indicating that a dominant-neg-ative mutant of ERK2 did not inhibit the ability of TGFβ to stimulate 3TP-Lux activity (Frey and Mulder, 1997b). Thus, our data suggest that the activation of the Ras/MEK pathway by TGFβ may not directly regulate AP-1 activity, at least in the context of the 3TP-Lux reporter. Second, endogenous Smad1 should be affected by Ras/MEK blockade. However, in the absence of transfection of exogenous RSmad1, endogenous Smad 2 (which does not contain Erk consensus phosphorylation sites but is also activated by TGFβ) may still regulate 3TP-Lux activity.
Collectively, our data herein demonstrate that both RasN17 and PD98059 partially inhibited the ability of RSmad1 to induce 3TP-Lux activity in both the absence and presence of TGFβ. Although 3TP-Lux does not represent a minimal Smad regulated site, our results may still suggest that two parallel pathways emanate from the TGFβ receptors to indirectly regulate 3TP-Lux activity. Activation of the Ras/MEK pathway potentiates 3TP-Lux stimulation by the Smad pathway. Each pathway may also regulate additional signaling components that ultimately affect 3TP-Lux activity.
Our results prodvide the first clear evidence of cross-talk between the Smad1 and Ras/MEK signaling pathways for TGFβ in epithelial cells. Although Atfi and coworkers (Atfi et al., 1997) reported that activation of the SAPKs/JNKs was required for induction of 3TP luciferase activity after coexpression with Smad3 plus Smad4 in MDCK cells, TGFβ did not activate the SAPKs until 8 h after TGFβ addition to their system. Since the activation of the SAPKs was delayed in their system, the effects of these kinases on the ability of Smad3 plus Smad4 to regulate 3TP luciferase activity would be expected to be indirect. This is in contrast to the rapid (within 3 – 30 min) activation of Ras, Erks, and SAPKS that we have shown here and reported previously (Mulder and Morris, 1992; Hartsough and Mulder, 1995, 1997; Frey and Mulder, 1997a,b).
Our results also differ from a report which indicated that epidermal growth factor and hepatocyte growth factor (HGF) could phosphorylate Smad1 though an Erk-mediated pathway to inhibit BMP-mediated nuclear accumulation and transcriptional activation of Smad1 (Kretzschmar et al., 1997b), However, this report did not examine the direct effects of BMP on Erk activation and subsequent Smad1 regulation. Here, we have demonstrated the direct activation of ERK both by TGFβ (Figure 3a) and BMP (data not shown). We have also demonstrated that Ras and MEK are partially required for Smad1 phosphorylation by both TGFβ and BMP. Our results appear to be similar to the situation in Xenopus, for which BMP4 can regulate the Ras/Raf/AP-1 pathway (Xu et al., 1996). In addition, we have demonstrated that Ras and MEK are partially required for the ability of TGFβ to regulate Smad1 transcriptional effects on the 3TP-Lux reporter. Accordingly, consideration of the direct effects of TGFβ superfamily members on the Ras/MEK pathway is necessary to fully interpret the results obtained from Smad signaling studies.
In summary, our results are consistent with a model for TGFβ signaling whereby TGFβ activates Ras and MAPKs. Ras and MAPKs are partially required for both TGFβ-induced Smad1 phosphorylation and Smad1-mediated transcriptional regulation. Although Ras is necessary and sufficient for TGFβ regulation of the CKIs (Yue et al., 1998), Ras/MEK-independent pathways are also required for Smad1 phosphorylation and transcriptional activation by TGFβ. In addition, it is expected that other intersecting pathways are activated downstream of the TGFβ receptors, each of which contributes to the diverse cellular outcomes of TGFβ.
transforming growth factor-beta
bone morphogenetic protein
dominant-negative Ras mutant
intestinal epithelial cells
mitogen-activated protein kinase
Sma and Mad homologs
MAP and ERK kinase
- SDS – PAGE:
SDS polyacrylamide gel electrophoresis
cyclin-dependent kinsase inhibitors
extracellular signal-regulated kinase
stress-activated protein kinase/c-Jun N-terminal kinase
Supplemental McCoys 5A medium
Atfi A, Buisine M, Mazars A and Gespach C. . 1997 J. Biol. Chem. 272: 24731–24734.
Chen Y, Bhushan A and Vale W. . 1997 Proc. Natl. Acad. Sci. USA 94: 12939–12943.
De Caestecker M, Parks WT, Frank CJ, Castagnino P, Bottaro DP, Roberts AB and Lechleider RJ. . 1998 Genes & Dev. 12: 1587–1592.
Derynck R and Zhang Y. . 1996 Curr. Biol. 6: 1226–1229.
Derynck R. . 1994 TIBS 19: 548–553.
Frey RS and Mulder KM. . 1997a Cancer Res. 57: 628–633.
Frey RS and Mulder KM. . 1997b Cancer Lett. 117: 41–50.
Gille H, Sharrocks AD and Shaw PE. . 1992 Nature 358,: 414–417.
Hartsough MT, Frey R, Zipfel P, Buard A, Cook S, McCormick F and Mulder KM. . 1996 J Biol. Chem. 271: 22368–22375.
Hartsough MT and Mulder KM. . 1995 J. Biol. Chem. 270: 7117–7124.
Hartsough MT and Mulder KM. . 1997 Pharmacol. Ther. 75: 21–42.
Heldin C-H, Miyazono K and ten Dijke P. . 1997 Nature 390: 465–471.
Hoodless P, Haerry T, Abdollah S, Stapleton M, O'Connor MB, Attisano L and Wrana JL. . 1996 Cell 85: 489–500.
Kretzschmar M, Liu F, Hata A, Doody J and Massagué J. . 1997a Genes & Dev. 11: 984–995.
Kretzschmar M, Doody J and Massagué J. . 1997b Nature 389: 618–622.
Lechleider RJ, de Caestecker MP, Dehejia A, Polymeropoulos MH and Roberts AB. . 1996 J. Biol. Chem. 271: 17617–17620.
Liu F, Hata A, Baker JC, Doody J, Cárcamo J, Harland RM and Massagué J. . 1996 Nature 381: 620–623.
Liu X, Sun Y, Constantinescu SN, Karam E, Weinberg RA and Lodish HF. . 1997 Proc. Natl. Acad. Sci. USA 94: 10669–10674.
Massagué J. . 1996 Cell 85: 947–950.
Mulder KM, Segarini PR, Morris SL, Ziman JM and Choi HG. . 1993 J. Cell. Physiol. 154: 162–174.
Mulder KM and Morris SL. . 1992 J. Biol. Chem. 267: 5029–5031.
Nishimura R, Kato Y, Chen D, Harris SE, Mundy GR and Yoneda T. . 1998 J. Biol. Chem. 273: 1872–1879.
Raftery LA, Twombly V, Wharton K and Gelbart WM. . 1995 Genetics 139: 241–254.
Wrana JL, Attisano L, Cárcamo J, Zentella A, Doody J, Laiho M, Wang X-F and Massagué J. . 1992 Cell 71: 1003–1014.
Wrana JL and Attisano L. . 1996 Trends Genet. 12: 493–496.
Xu R-H, Dong Z, Maeno M, Kim J, Suzuki A, Ueno N, Sredni D, Colburn NH and Kung H-F. . 1996 Proc. Natl. Acad. Sci. USA 93: 834–838.
Yingling JM, Das P, Savage C, Zhang M, Padgett RW and Wang X-F. . 1996 Proc. Natl. Acad. Sci. USA 93: 8940–8944.
Yue J, Buard A and Mulder KM. . 1998 Oncogene 17: 47–55.
Yue J, Hartsough MT, Frey R, Frielle T and Mulder KM. . 1999 J. Cell. Physiol. in press.
We wish to thank M Morin (Pfizer Pharmaceuticals, Groton, CT, USA) for generously supplying the TGFβ3, J Massagué (Memorial Sloan-Kettering Cancer Center, New York, NY, USA) for the anti-Smad1 antibody and p3TP-Lux and V Rosen (Genetics Institute, Cambridge, MA, USA) for BMP2. We also thank Andrew Stevenson for assisting with the preparation of the manuscript. This work was supported by National Institutes of Health Grants CA51425, CA54816, and CA68444 to KMM. KMM is a recipient of National Institutes of Health Research Career Development Award K04 CA59552.
About this article
Nature Communications (2019)
Methylglyoxal, a glycolysis metabolite, triggers metastasis through MEK/ERK/SMAD1 pathway activation in breast cancer
Breast Cancer Research (2019)
Acta Biochimica et Biophysica Sinica (2018)
Testosterone promotes tube formation of endothelial cells isolated from veins via activation of Smad1 protein
Molecular and Cellular Endocrinology (2017)
Journal of Pathology and Translational Medicine (2015)