Negative feedback predominates over cross-regulation to ... - Core

FEBS Letters 580 (2006) 4242–4245

Negative feedback predominates over cross-regulation to control ERK MAPK activity in response to FGF signalling in embryos Terence Gordon Smitha,*, Maria Karlssonb, J. Simon Lunna, Maxwell C. Eblaghiea, Iain D. Keenanb, Elizabeth R. Farrella, Cheryll Ticklea, Kate G. Storeya, Stephen M. Keyseb a b

Division of Cell and Developmental Biology, School of Life Sciences, University of Dundee, Dow Street, Dundee DD1 5EH, UK Cancer Research UK Molecular Pharmacology Unit, Biomedical Research Centre, Level 5, Ninewells Hospital and Medical School, Dundee DD1 9SY, Scotland, United Kingdom Received 7 June 2006; accepted 27 June 2006 Available online 5 July 2006 Edited by Richard Marais

Abstract Expression of the gene encoding the MKP-3/Pyst1 protein phosphatase, which inactivates ERK MAPK, is induced by FGF. However, which intracellular signalling pathway mediates this expression is unclear, with essential roles proposed for both ERK and PI(3)K in chick embryonic limb. Here, we report that MKP-3/Pyst1 expression is sensitive to inhibition of ERK or MAPKK, that endogenous MKP-3/Pyst1 co-localizes with activated ERK, and expression of MKP-3/Pyst1 in mice lacking PDK1, an essential mediator of PI(3)K signalling. We conclude that MKP-3/Pyst1 expression is mediated by ERK activation and that negative feedback control predominates in limiting the extent of FGF-induced ERK activity. Ó 2006 Published by Elsevier B.V. on behalf of the Federation of European Biochemical Societies. Keywords: Negative feedback; Fibroblast growth factor; Extracellular signal-regulated kinase; Mitogen activated protein kinase; MKP-3/Pyst1; Limb

1. Introduction The fibroblast growth factors (FGFs) comprise a family of ligands which initiate one of the most intensely studied and versatile signal transduction pathways [1]. The requirement of FGF signalling for cell survival and patterning during limb and neural embryonic development is well known. Removal of the apical ectodermal ridge (AER), a tissue source of FGFs in the limb, leads to truncation, which can be rescued by the application of beads soaked in FGFs [2]. Conditional mouse mutants have also demonstrated that FGFs are necessary for limb growth and patterning [3]. In neural tissue, FGFs produced by early hypoblast and Hensen’s node are required to initiate neural induction and for later patterning events in the neural plate [4].

* Corresponding author. E-mail address: [email protected] (T.G. Smith).

Abbreviations: MKP-3, MAPK phosphatase-3; ERK, extracellular signal-regulated kinase; MAPK, mitogen activated protein kinase; FGF, fibroblast growth factor; PI(3)K, phosphatidylinositol-3 kinase; FGFR, FGF receptor; AER, apical ectodermal ridge; PDK1, 3-phosphoinositide-dependent protein kinase 1

FGFs signal through receptor tyrosine kinases (FGFRs) expressed at the cell surface. Clustering of the FGFRs on binding to FGF ligand and heparan sulphate proteoglycan leads to the activation of downstream signalling pathways within the cell [5]. FGFRs can signal through both the extracellular signalregulated kinase (ERK) mitogen activated protein kinase (MAPK) pathway and the phosphatidylinositol-3 kinase (PI(3)K) signalling pathway, as well as others [6] (Fig. 1). In embryos it has been found that FGF signalling often induces the expression of genes required to transduce the FGF signal. For example, FGF-8 induces Pea-3 and Erm expression in zebrafish and chick, which are known transcription factors of the FGF pathway [7,8]. In addition, FGFs have been shown to induce the expression of negative regulators, such as sprouty and SEF, that act to restrict the output of FGF signalling [9,10]. We and others have demonstrated that the expression of the ERK-specific MAPK phosphatase MKP-3/Pyst1 is induced by FGF signalling, with reports suggesting that two possible intracellular pathways might be involved [11–17]. This data has largely been obtained in a range of tissues from different model systems, which may explain the differing results obtained. However, in the chick embryonic limb there is a direct contradiction over which intracellular pathway mediates the FGF signal. Pharmacological inhibitors of either ERK MAPK kinase or PI(3)K have been used to invoke these pathways as essential mediators of MKP-3/Pyst1 expression (Fig. 1) [11,14]. To resolve this issue, we have used several approaches in the limb bud and neural plate of embryos, and present new evidence that MKP-3/Pyst1 regulates ERK MAPK activity via negative feedback.

2. Materials and methods 2.1. Transgenic mice 3-Phosphoinositide-dependent protein kinase 1 (PDK1)/ mice were generated as described by Lawlor [18]. Whole-mount in situ hybridisation was carried out at E9.5 using standard techniques and a murine MKP-3/Pyst1 probe was used as described before [19]. 2.2. Pharmacological inhibitors AG-1X2 (BioRad Laboratories) formate-derivitized beads were incubated in one of the following inhibitors: 20 mM PD184352, 25 or 50 mM PD098059, and 20 mM LY294002 for 1 h at room temperature. DMSO beads were used as controls. Beads were grafted into HH22 limb mesenchyme in vivo, or into HH3 neural plate in EC culture, and incubated for 6 h and 4 h, respectively at 37 °C.

0014-5793/$32.00 Ó 2006 Published by Elsevier B.V. on behalf of the Federation of European Biochemical Societies. doi:10.1016/j.febslet.2006.06.081

T.G. Smith et al. / FEBS Letters 580 (2006) 4242–4245

4243 and 35 cycles. GFP (forward: 5v-ACGTGAATTCCTGACCCTGA AGTTCATCTG-3 0 and reverse: 5 0 -ACGTGCGGCCGCGTCTTGAAGTTCGCCTTGAT G-3 0 ) were used at 55 °C annealing temperature and 30 cycles.

3. Results and discussion

Fig. 1. Pathways downstream of FGF signalling. Downstream of the FGF tyrosine kinase receptor, the ERK MAP kinase pathway is activated through Grb2, and the PI(3)K pathway is activated through Gab1. We have used different reagents or mutants to selectively inhibit key points along both pathways (shown in red); pharmacological inhibitors PD184352 and PD098059 antagonise MAPK kinase and reduce ERK MAPK activity; ERK phosphorylation was specifically inhibited by expressing a human MKP-3/Pyst1:GFP fusion protein; LY294002 was used to inhibit PI(3)K signalling and this pathway was selectively blocked at the level of PDK1 in knock-out mice.

2.3. Limb dissections and Western blotting Distal, medial and proximal pieces were pooled from between 13 and 20 limb buds at stages HH21-23 and snap frozen at 80 °C prior to analysis. Pooled fractions were thawed on ice for 5 min, followed by centrifugation (Eppendorf centrifuge 13 K for 2 min at 4 °C). Excess media was then removed and samples resuspended in 60 ll of SDS sample buffer followed by heating to 95 °C for 3 min and sonication for 4 s. Samples were then reheated (3 min at 95 °C) and cleared by centrifugation (Eppendorf centrifuge 13 K for 2 min at room temperature) before analysis using the NuPage electophoresis system (4–12%, Invitrogen). Following electrophoresis, proteins were transferred to a 0.45 lm polyvinylidene difluoride membrane (Millipore). Immunoblotting was then performed using the following antibodies: MKP-3 C20 and tubulin H235 (Santa Cruz), ERK, phospho-ERK and phospho-AKT (S473) (Cell Signalling Technologies). This analysis was performed at least three times with reproducible results and a representative blot is illustrated. 2.4. Electroporation and RT-PCR HH21 limbs were microinjected with a mixture of 2 lg/ll hMKP3GFP and Fast Green. One platinum electrode was placed to the side of the limb bud (positive) and another on the neural tube (negative) before five 60 V, 50 ms pulses were delivered using a CUY-21 electroporator. Electroporated limbs were identified by GFP fluorescence and dissected using a sharpened tungsten needle under UV light. Only tissue that was GFP positive was included for RNA extraction. Approximately 7–13 embryos were collected in four independent experiments and processed using Qiagen RNA Easy according to the manufacturer’s instructions. Reverse transcription reaction was performed as previously described [15]. Briefly, 15 ll of the extracted RNA was heated to 65 °C for 3 min, then cooled on ice. One microliter of reverse transcriptase was used in a final volume of 30 ll and incubated at 42 °C for 1 h in a PCR machine. The cDNA was stored at 20 °C. PCR was performed using Taq polymerase from BIOLINE in a total reaction volume of 30 ll. 5 ll from 30 ll reaction volumes were used in analytical agarose gels. cDNA produced from RNA extraction was normalised using GAPDH primers (forward: 5 0 -AGTCATCCCTGAGCTGAATG-3 0 and reverse: 5 0 AGGATCAAGTCCACAACAC G-3 0 ). A linear amplification range was provided by using 20, 25 and 30 cycles and an annealing temperature of 55 °C. Chick Mkp-3 specific primers (forward: 5 0 -ACGTGCGGCCGCATGCTAGATACGTTCAGACCCGTC-3 0 and reverse: 5 0 -ACGTGAATTCTCACGTGGACTGCAGGGAGTCCACC-3 0 ) were used at 59.8 °C annealing temperature

Two MAPK kinase inhibitors, PD184352 and PD098059 (Fig. 1), have both been used previously in the chick limb to address how MKP-3/Pyst1 transcription is regulated, but have not given consistent results [11,14]. We therefore began by comparing the effects of these two different inhibitors on MKP-3/Pyst1 transcription. In the chick neural plate as we have shown previously, beads soaked in 20 mM PD184352 had a clear repressive effect on MKP-3/Pyst1 (Fig. 2A, n = 5/ 6) [11]. In contrast, beads soaked in 25 mM (Fig. 2C, n = 6/ 6) and 50 mM (data not shown, n = 3/3) of PD098059 had no detectable effect. Similar results were obtained in the limb bud. Beads soaked in 20 mM PD184352 inhibited MKP-3/ Pyst1 transcription [11], whereas beads soaked in 50 mM PD098059 had no effect, even after 20 h (data not shown, n = 6/6). The failure of PD098059 to inhibit MKP-3/Pyst1 transcription confirms previous results in chick and mouse embryos [12,14]. Since PD184352 is a more soluble, potent and specific inhibitor of the MAPK pathway than PD098059, this may explain the different effects of the two drugs [20,21]. Therefore PD183252 appears to be the most reliable tool to study the MAPK pathway in vivo. To circumvent problems associated with the use of inhibitors, we manipulated ERK MAPK activity by expressing MKP-3/Pyst1 itself in the chick limb bud (Fig. 1) [22]. In contrast to some other MAPK phosphatases, MKP-3/Pyst1 is highly specific for ERK MAPK [22]. In order to distinguish between ectopic and endogenous MKP-3/Pyst1 transcripts, we expressed a human MKP-3/Pyst1:GFP fusion protein (hMKP3-GFP), which inactivates ERK in vivo and in vitro [11,15], and then used RT-PCR to discriminate between the human and chick (cMkp-3) mRNAs [15]. This allowed us to ask whether direct inactivation of ERK MAPK affects cMkp-3 transcription. After electroporation into HH21 limb buds, mesenchyme expressing hMKP3-GFP was dissected from non-transfected tissue under UV fluorescence (Fig. 3A),

Fig. 2. ERK MAPK activity is effectively repressed with PD184352 and not PD098059. Effects on MKP-3/Pyst1 expression in the neural plate by two MAPK kinase inhibitors after 4 h. (A) Beads soaked in 20 mM PD184352 decreased MKP-3/Pyst1. (B) DMSO control beads did not affect expression of MKP-3/Pyst1. (C) There was no detectable change in MKP-3/Pyst1 expression with beads soaked in 25 mM PD098059. The beads are marked with asterisks.

4244

Fig. 3. Chick MKP-3/Pyst1 transcription is regulated by ERK MAPK in the limb bud. Semi-quantitative RT-PCR analyses after electroporation of either GFP control or hMKP3-GFP in the limb bud after 5 h. (A) hMKP3-GFP fluorescence. (B) RT-PCR was performed on normalised cDNA to detect cMkp3, GFP and GAPDH. Lane 1: GFP control. Lane 2: hMKP3-GFP. The transcription of cMkp3 was inhibited in presence of hMKP3-GFP.

and RNA was extracted and analysed using RT-PCR. Our results clearly demonstrate that hMKP3-GFP expression substantially reduced endogenous cMkp-3 transcription when compared to GFP controls (lanes 1 and 2, Fig. 3B). This is consistent with ERK MAPK governing MKP-3/Pyst1 expression in chick limb mesenchyme. If MKP-3/Pyst1 is regulated in a negative feedback loop, ERK MAPK must be activated in cells transcribing MKP-3/ Pyst1. To address this, limb bud mesenchyme was separated from the AER, cut into proximal, middle and distal pieces (Fig. 4A), and analysed by Western blotting. The active form of ERK MAPK is phosphorylated on critical threonine and tyrosine residues, and we used an antibody specific for this active state in our experiments as well as one that detected the unphosphorylated form of ERK MAPK. In contrast to a previous report [14], we reproducibly detected phospho-ERK across the entire limb bud mesenchyme (Fig. 4B). PhosphoAKT/PKB, a downstream target of the PI(3)K pathway was also present in this tissue indicating that FGF signalling may be transduced via either or both of these pathways. Furthermore, MKP-3/Pyst1 protein was present in a distal-to-proximal gradient, similar to its mRNA expression pattern (Fig. 4A). In support of our data concerning the presence of activated ERK in chick limb bud, Corson et al. have also observed a distal-to-proximal gradient of phospho-ERK MAPK

Fig. 4. Phospho-ERK is detected in limb bud mesenchyme, overlapping with MKP-3/Pyst1. Analysis of phospho-ERK (p-ERK) and p-AKT levels in three fractions of the limb bud mesenchyme. (A) Diagram showing limb dissection in relation to MKP-3/Pyst1 transcripts. Mesenchyme cut away from the AER and separated into D, distal, M, middle, and P, proximal mesenchyme. (B) Western analysis of extracts of the three regions. p-ERK and p-AKT levels are present throughout the limb mesenchyme. MKP-3/Pyst1 protein is graded from distal to proximal, similar to its mRNA expression pattern. Tubulin served as a loading control.


in mouse limb bud mesenchyme using whole-mount immunohistochemistry [23]. These results are consistent with the regulation of MKP-3/Pyst1 by phospho-ERK MAPK. If, as previously suggested in the chick hindlimb and mouse isthmus, [12,14], MKP-3/Pyst1 expression is regulated by PI(3)K, this would represent crosstalk between two intracellular pathways downstream of the FGFR rather than negative feedback. The PI(3)K antagonist LY294002 (Fig. 1) inhibits MKP-3/Pyst1 expression in some embryonic tissues [12,14], but not others [11,15]. We have shown that beads soaked in LY294002 can attenuate ectopic induction of MKP-3/Pyst1 expression by FGFs in neural plate and limb bud, but a single bead has no effect on endogenous MKP-3/Pyst1 expression after short or long incubations in either tissue ([11], data not shown). We reinvestigated the effects of this drug and found that endogenous MKP-3/Pyst1 expression could only be reduced when two beads soaked in 20 mM LY294002 were coimplanted into the chick wing bud (Fig. 5A and B, n = 3/3). However, these high levels of LY294002 possibly inhibited MKP-3/Pyst1 expression through non-specific effects. To probe further the role of PI(3)K, we explored the expression of MKP-3/Pyst1 in mutant embryos (Fig. 1). Signalling downstream of PI(3)K is blocked in mice lacking the PDK1 kinase [18]. These mice do not survive after day 9.5, and do not form limb buds or forebrains [18]. However MKP-3/Pyst1 transcripts were still present in tailbud, caudal neural plate

Fig. 5. PI(3)K is not responsible for MKP-3/Pyst1 expression in mouse neural tissue. Expression of MKP-3/Pyst1 in the presence of beads soaked in PI(3)K inhibitor LY294002 after 6 h. (A) View of the contralateral control limb showing normal MKP-3/Pyst1 expression in the limb bud. (B) Loss of MKP-3/Pyst1 transcripts when two beads of 20 mM LY294002 were implanted. (C) MKP-3/Pyst1 expression in wild-type and PDK1/ embryos at day E9.5. Arrows indicate persistent MKP-3/Pyst1 expression in anterior neural structures and in the tail bud.


and anterior neural tissues of PDK1/ embryos (Fig. 5C, n = 4/4). This suggests that PI(3)K is not absolutely required for MKP-3/Pyst1 expression in these tissues. We conclude that in the developing limb bud and neural plate, cells receiving FGF signals co-express MKP-3/Pyst1 to control ERK MAPK signal gain through a negative feedback mechanism. This illustrates a fundamental way to shape signalling in biological systems, where the regulators of intracellular signalling pathways are transcriptionally controlled by their target pathways [24].

4245

[12]

[13]

[14] Acknowledgements: We thank Margaret Lawlor and Dario Alessi for the PDK1/ mice. C.T., T.G.S., M.C.E. and E.R.F. were supported by The Royal Society, The MRC and The Anatomical Society. S.K. is supported by Cancer Research UK. K.G.S. is an MRC SNCRF (G9900177) and J.S.L. was supported by a BBSRC studentship.

[15]

References [16] [1] Ornitz, D.M. and Itoh, N. (2001) Fibroblast growth factors. Genome Biol. 2, REVIEWS3005. [2] Niswander, L., Jeffrey, S., Martin, G.R. and Tickle, C. (1994) A positive feedback loop coordinates growth and patterning in the vertebrate limb. Nature 371, 609–612. [3] Sun, X., Mariani, F.V. and Martin, G.R. (2002) Functions of FGF signalling from the apical ectodermal ridge in limb development. Nature 418, 501–508. [4] Streit, A., Berliner, A.J., Papanayotou, C., Sirulnik, A. and Stern, C.D. (2000) Initiation of neural induction by FGF signalling before gastrulation. Nature 406, 74–78. [5] Ornitz, D.M. (2000) FGFs, heparan sulfate and FGFRs: complex interactions essential for development. Bioessays 22, 108–112. [6] Schlessinger, J. (2004) Common and distinct elements in cellular signaling via EGF and FGF receptors. Science 306, 1506–1507. [7] Brent, A.E. and Tabin, C.J. (2004) FGF acts directly on the somitic tendon progenitors through the Ets transcription factors Pea3 and Erm to regulate scleraxis expression. Development 131, 3885–3896. [8] Roehl, H. and Nusslein-Volhard, C. (2001) Zebrafish pea3 and erm are general targets of FGF8 signaling. Curr. Biol. 11, 503– 507. [9] Furthauer, M., Lin, W., Ang, S.L., Thisse, B. and Thisse, C. (2002) Sef is a feedback-induced antagonist of Ras/MAPKmediated FGF signalling. Nat. Cell Biol. 4, 170–174. [10] Hanafusa, H., Torii, S., Yasunaga, T. and Nishida, E. (2002) Sprouty1 and Sprouty2 provide a control mechanism for the Ras/ MAPK signalling pathway. Nat. Cell Biol. 4, 850–858. [11] Eblaghie, M.C., Lunn, J.S., Dickinson, R.J., Munsterberg, A.E., Sanz-Ezquerro, J.J., Farrell, E.R., Mathers, J., Keyse, S.M.,

[17]

[18]

[19]

[20] [21]

[22] [23] [24]

Storey, C. and Tickle, C. (2003) Negative feedback regulation of FGF signaling levels by Pyst1/MKP3 in chick embryos. Curr. Biol. 13, 1009–1018. Echevarria, D., Martinez, S., Marques, S., Lucas-Teixeira, V. and Belo, J.A. (2005) Mkp3 is a negative feedback modulator of Fgf8 signaling in the mammalian isthmic organizer. Dev. Biol. 277, 114–128. Gomez, A.R., Lopez-Varea, A., Molnar, C., de la Calle-Mustienes, E., Ruiz-Gomez, M., Gomez-Skarmeta, J.L. and de Celis, J.F. (2005) Conserved cross-interactions in Drosophila and Xenopus between Ras/MAPK signaling and the dual-specificity phosphatase MKP3. Dev. Dyn. 232, 695–708. Kawakami, Y., Rodriguez-Leon, J., Koth, C.M., Buscher, D., Itoh, T., Raya, A., Ng, J.K., Esteban, C.R., Takahashi, S., Henrique, D., Schwarz, M.F., Asahara, H. and Izpisua Belmonte, J.C. (2003) MKP3 mediates the cellular response to FGF8 signalling in the vertebrate limb. Nat. Cell Biol. 5, 513–519. Smith, T.G., Sweetman, D., Patterson, M., Keyse, S.M. and Munsterberg, A. (2005) Feedback interactions between MKP3 and ERK MAP kinase control scleraxis expression and the specification of rib progenitors in the developing chick somite. Development 132, 1305–1314. Tsang, M., Maegawa, S., Kiang, A., Habas, R., Weinberg, E. and Dawid, I.B. (2004) A role for MKP3 in axial patterning of the zebrafish embryo. Development 131, 2769–2779. Vieira, C. and Martinez, S. (2005) Experimental study of MAP kinase phosphatase-3 (Mkp3) expression in the chick neural tube in relation to Fgf8 activity. Brain Res. Brain Res. Rev. 49, 158– 166. Lawlor, M.A., Mora, A., Ashby, P.R., Williams, M.R., MurrayTait, V., Malone, L., Prescott, A.R., Lucocq, J.M. and Alessi, D.R. (2002) Essential role of PDK1 in regulating cell size and development in mice. Embo J. 21, 3728–3738. Dickinson, R.J., Eblaghie, M.C., Keyse, S.M. and Morriss-Kay, G.M. (2002) Expression of the ERK-specific MAP kinase phosphatase PYST1/MKP3 in mouse embryos during morphogenesis and early organogenesis. Mech. Dev. 113, 193–196. Davies, S.P., Reddy, H., Caivano, M. and Cohen, P. (2000) Specificity and mechanism of action of some commonly used protein kinase inhibitors. Biochem. J. 351, 95–105. Sebolt-Leopold, J.S., Dudley, D.T., Herrera, R., Van Becelaere, K., Wiland, A., Gowan, R.C., Tecle, H., Barrett, S.D., Bridges, A., Przybranowski, S., Leopold, W.R. and Saltiel, A.R. (1999) Blockade of the MAP kinase pathway suppresses growth of colon tumors in vivo. Nat. Med. 5, 810–816. Keyse, S.M. (2000) Protein phosphatases and the regulation of mitogen-activated protein kinase signalling. Curr. Opin. Cell Biol. 12, 186–192. Corson, L.B., Yamanaka, Y., Lai, K.M. and Rossant, J. (2003) Spatial and temporal patterns of ERK signaling during mouse embryogenesis. Development 130, 4527–4537. Niehrs, C. and Pollet, N. (1999) Synexpression groups in eukaryotes. Nature 402, 483–487.