Title: Inhibition of ERK-MAP kinase signaling by RSK during Drosophila development
Abstract: Article8 June 2006free access Inhibition of ERK-MAP kinase signaling by RSK during Drosophila development Myungjin Kim Myungjin Kim National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Jun Hee Lee Jun Hee Lee National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Hyongjong Koh Hyongjong Koh National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Soo Young Lee Soo Young Lee National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Cholsoon Jang Cholsoon Jang National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Cecilia J Chung Cecilia J Chung National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Jung Hwan Sung Jung Hwan Sung National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author John Blenis John Blenis Department of Cell Biology, Harvard Medical School, Boston, MA, USA Search for more papers by this author Jongkyeong Chung Corresponding Author Jongkyeong Chung National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Myungjin Kim Myungjin Kim National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Jun Hee Lee Jun Hee Lee National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Hyongjong Koh Hyongjong Koh National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Soo Young Lee Soo Young Lee National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Cholsoon Jang Cholsoon Jang National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Cecilia J Chung Cecilia J Chung National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Jung Hwan Sung Jung Hwan Sung National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author John Blenis John Blenis Department of Cell Biology, Harvard Medical School, Boston, MA, USA Search for more papers by this author Jongkyeong Chung Corresponding Author Jongkyeong Chung National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea Search for more papers by this author Author Information Myungjin Kim1, Jun Hee Lee1, Hyongjong Koh1, Soo Young Lee1, Cholsoon Jang1, Cecilia J Chung1, Jung Hwan Sung1, John Blenis2 and Jongkyeong Chung 1 1National Creative Research Initiatives Center for Cell Growth Regulation and Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Yusong, Taejon, Korea 2Department of Cell Biology, Harvard Medical School, Boston, MA, USA *Corresponding author. Department of Biological Sciences, Korea Advanced Institute of Science and Technology, 373-1 Kusong-Dong, Yusong, Taejon 305-701, Korea. Tel.: +82 42 869 2620; Fax: +82 42 869 8260; E-mail: [email protected] The EMBO Journal (2006)25:3056-3067https://doi.org/10.1038/sj.emboj.7601180 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Although p90 ribosomal S6 kinase (RSK) is known as an important downstream effector of the ribosomal protein S6 kinase/extracellular signal-regulated kinase (Ras/ERK) pathway, its endogenous role, and precise molecular function remain unclear. Using gain-of-function and null mutants of RSK, its physiological role was successfully characterized in Drosophila. Surprisingly, RSK-null mutants were viable, but exhibited developmental abnormalities related to an enhanced ERK-dependent cellular differentiation such as ectopic photoreceptor- and vein-cell formation. Conversely, overexpression of RSK dramatically suppressed the ERK-dependent differentiation, which was further augmented by mutations in the Ras/ERK pathway. Consistent with these physiological phenotypes, RSK negatively regulated ERK-mediated developmental processes and gene expressions by blocking the nuclear localization of ERK in a kinase activity-independent manner. In addition, we further demonstrated that the RSK-dependent inhibition of ERK nuclear migration is mediated by the physical association between ERK and RSK. Collectively, our study reveals a novel regulatory mechanism of the Ras/ERK pathway by RSK, which negatively regulates ERK activity by acting as a cytoplasmic anchor in Drosophila. Introduction The p90 ribosomal protein S6 kinase (RSK) is well conserved among the metazoan systems (reviewed in Blenis, 1993; Nebreda and Gavin, 1999). RSK was originally identified as a direct target of the ribosomal protein S6 kinase/extracellular signal-regulated kinase (Ras/ERK)-MAP kinase signaling pathway, which regulates the most critical cellular responses including cell proliferation, differentiation, and metabolism (reviewed in Blenis, 1993; Pearson et al, 2001). RSK specifically binds to ERK (Scimeca et al, 1992) and is activated through phosphorylation by activated ERK (Sturgill et al, 1988; Chung et al, 1991). As the activity of RSK tightly correlates with that of ERK, RSK has been thoroughly studied as one of the critical downstream effectors of ERK. Indeed, various physiologically important molecules such as lamin-C, glycogen synthase kinase 3, cAMP-responsive binding-element protein (CREB), histone 3B, anaphase-promoting complex (APC), C/EBP beta, Bub1, c-Fos, filamin A, and tuberous sclerosis complex (TSC) were suggested as putative targets mediating the molecular function of RSK (Schwab et al, 2001; Roux et al, 2004; Woo et al, 2004; reviewed in Frodin and Gammeltoft, 1999). Although these targets seem to appropriately explain the roles of RSK as a downstream effector of the Ras/ERK signaling pathway, there has been insufficient evidence to consider them as the actual downstream targets of RSK in vivo. Furthermore, some recent studies suggest that RSK has an inhibitory role in the Ras/ERK pathway while neither the physiological relevance nor the molecular mechanism has yet been sufficiently addressed (Myers et al, 2004). The existence of many RSK isoforms (RSK 1–4) in the mammalian genome has hampered extensive genetic researches on the physiological functions of RSK (Alcorta et al, 1989; Moller et al, 1994; Yntema et al, 1999). In this study, we took advantage of the Drosophila system, which has only a single RSK gene in the genome (Wassarman et al, 1994). We generated null flies for Drosophila RSK and succeeded in characterizing its in vivo function. Surprisingly, we found that RSK is devoted in negatively regulating nuclear ERK function, restraining ERK in the cytoplasm during Drosophila eye and wing development, by physical association with ERK. Results RSK null flies survive to the adult stage with some developmental abnormalities To understand the role of RSK at the organism level, we generated a null mutant for Drosophila RSK, a single orthologue of the mammalian RSK gene family (Wassarman et al, 1994). Surprisingly, the RSK null flies (RSKD1) that completely lack RSK transcripts and proteins (Figure 1B–D) survived to the adult stage. However, RSKD1 mutants displayed several developmental abnormalities including some developmental delay, reduced fertility, and shortened longevity, as well as previously reported learning defects (Supplementary Figure 1 and data not shown; Putz et al, 2004). Since RSK is ubiquitously expressed throughout all developmental stages (Figure 1E) and tissues (Figure 1F), we inferred that RSK is involved in general developmental processes that are not essential for the survival of the organism. Figure 1.Ectopic differentiation of retinal and wing vein cells in RSK null mutant. (A) Genomic organization of the Drosophila RSK locus and the molecular status of the RSK mutants. The EP element insertion in EP(X)7363 (triangle) and 3.1 kb genomic deletion in RSKD1 are indicated. (B) RSK expression in RSKD1 mutant. The amount of RSK transcripts was visualized through RT–PCR. rp49 was used as a control. (C) Northern blot analyses of RSKD1 mutant. 18S rRNA was used as a control. (D) Lysates from flies with indicated genotypes were immunoprecipitated using guinea-pig anti-RSK antibody. The immune complexes were analyzed by immunoblot using rabbit anti-RSK antibody. Tubulin (tub) was used as a loading control. (E) RT–PCR analysis of RSK expression in wild-type flies at various developmental stages (rp49, loading control). (F) In situ hybridization of wild-type embryo with RSK antisense and sense probes. (G–K) SEM images (left panels) and tangential cross sections (right panels) of Drosophila adult eye. Left bottom panels are magnified versions ( × 600) of SEM images. Arrowheads point to ectopic photoreceptor-containing ommatidia. (L) Quantified data of (I) to (K). The mean number of photoreceptor cells per ommatidium was calculated as described in Materials and methods. Error bar indicates the standard deviation (s.d.) between ommatidia. (M, N) Pupal eyes were stained with anti-Arm antibody and rhodamine-labeled phalloidin (green and red, respectively), or with anti-Dlg antibody (green). Boxed area contains a cluster of ommatidial precursor cells, which is magnified on the right panels. Genotypes are: WT (w1118, G, I, M), RSKD1 (RSKD1/Y, H, J, N), and RSKD1; da>RSKWT (RSKD1/Y; UAS-RSKWT/+; da-Gal4/+, K). Flies of (I–K) and (M, N) were cultured at 29°C. Download figure Download PowerPoint RSK is a negative regulator of the retinal cell fate determination To find out more about endogenous RSK function, we mainly focused on the eye differentiation phenotypes of RSKD1 mutants since these phenotypes are highly correlated with the Ras/ERK pathway (Dickson et al, 1992; Brunner et al, 1994). As a result, we unexpectedly found RSKD1 adult flies to have disarrayed eye structure and some ommatidia with increased number of photoreceptor rhabdomeres (R cells) (Figure 1H). When RSKD1 mutants were grown at a higher temperature, the number of R cells was further increased and other eye defects were also exacerbated (compare Figure 1J with I). To further examine these phenotypes in earlier developmental stages, we stained the pupal retinal cells with anti-Armadillo (Arm, Drosophila β-catenin) antibodies to mark adherens junctions. A characteristic staining pattern of seven ‘dots’ forming a circle at the center of each ommatidium was observed in the wild-type eye (Figure 1M; Hong et al, 2003). However, in RSKD1 mutants, we found some pupal ommatidia aberrantly containing extra anti-Arm antibody-stained dots (Figure 1N), consistent with the extra R-cell phenotype of the adult ommatidia in the mutant (Figure 1H and J). These results strongly suggested that RSK plays a negative role in photoreceptor cell differentiation during eye development. Additionally, the RSK-null mutants displayed severe irregularities in ommatidial spacing (Figure 1J), presumably reflecting defects in the differentiation of non-neuronal retinal cells such as cone and pigment cells. To analyze the non-neuronal retinal cells, we utilized anti-Discs large (Dlg) antibodies to stain the membrane structure of cone and pigment cells (Woods et al, 1997). In the wild-type pupal retina, the regular arrangement of four cone cells and 11 pigment cells was discernable in each ommatidium (Figure 1M). However, RSK-null flies displayed an increased number of cone and pigment cells with highly disrupted structures (Figure 1N) under the same developmental conditions (34 h after puparium formation). These results strongly supported that RSK regulates not only the neuronal photoreceptor cell differentiation but also the non-neuronal retinal cell development during Drosophila eye morphogenesis. To provide further evidence for our hypothesis, the transallelic mutants of RSKD1 with Df(1)R8A, an RSK deficiency allele, also displayed identical phenotypes to RSKD1 mutants (data not shown). Furthermore, weak ubiquitous expression of the wild-type RSK (RSKWT) transgene by the da-Gal4 driver (Supplementary Figure 2) fully rescued the eye phenotypes of RSKD1 mutants (Figure 1K and L), demonstrating that the defects in Drosophila retinal cell differentiation indeed resulted from the absence of RSK. As null mutation of RSK ectopically induced differentiation of both neuronal and non-neuronal retinal cells, we hypothesized RSK to be a general negative regulator of retinal cell fate determination. To further confirm this possibility, we expressed RSKWT by using eye-specific Gal4 drivers. As expected, expression of RSKWT in Drosophila eye by the gmr-Gal4 (Supplementary Figure 2) induced a roughened eye phenotype with dramatically reduced number of photoreceptor cells (Figure 2B, compared to Figure 2A). Moreover, RSKWT-overexpressing pupal eyes further revealed inhibited differentiation of both neuronal and non-neuronal retinal cells (Figure 2D and F). Figure 2.Inhibition of retinal cell differentiation by RSK. (A, B) SEM image (left panels) and tangential cross section (right panels) of Drosophila adult eye. The numbers indicate the mean number of photoreceptor cells per ommatidium±s.d. (C–F) Pupal eyes stained with anti-Arm antibody (C, D) and with anti-Dlg antibody (E, F). Boxed area contains a cluster of ommatidial precursor cells. Right panels are magnified views of the boxed areas, respectively. Genotypes are: gmr-Gal4 (gmr-Gal4/+, A, C, E) and gmr>RSKWT (gmr-Gal4 UAS-RSKWT/+, B, D, F). Download figure Download PowerPoint Collectively, these RSK null and overexpression experiments consistently demonstrated that RSK is a negative regulator of retinal cell differentiation in Drosophila. RSK is a negative regulator of the wing vein differentiation To address the possibility of the involvement of RSK in differentiation processes other than retinal cell development, we carefully observed other adult structures such as the thorax, abdomen, legs, and wings. Notably, some RSKD1 mutants (∼16%) had ectopic veins in various regions of the wing (Figure 3B), suggesting that differentiation of wing vein cells is also promoted by RSK null mutation. The number of ectopic veins was significantly increased when the flies were grown at 29°C (Figure 3D and F). Moreover, RSKD1 allele over Df(1)R8A also led to identical wing vein phenotypes to those of RSK-null flies (data not shown), and weak ubiquitous expression of RSKWT transgene by the da-Gal4 driver completely rescued the ectopic wing vein phenotypes of RSKD1 flies (Figure 3E and F), verifying that the ectopic differentiation of wing vein cells resulted from the absence of RSK. Figure 3.Inhibition of wing vein cell differentiation by RSK. (A–E, G–J) Microscopic views of Drosophila adult wings. Longitudinal veins (L1–L5) as well as anterior (AC) and posterior (PC) cross veins were indicated in wild-type wings. Right panels of (B) and (D) are magnified views of the boxed areas, respectively. (F) Quantified data of (C) to (E). More than 100 wings for each genotype were used for quantification. Flies of (C–E) were cultured at 29°C. Details of all indicated genotypes are described in Supplementary data. Download figure Download PowerPoint Consistently, overexpression of RSKWT by the e16E-Gal4 (induces gene expression in the posterior part of wing) or the MS1096-Gal4 (induces gene expression in the whole wing) dramatically induced compartment-specific vein-loss phenotypes (Figure 3H and J). In sum, our data demonstrated that RSK negatively regulates vein cell differentiation during Drosophila wing development. Genetic interaction between RSK and the Ras/ERK pathway As the Ras/ERK pathway positively regulates the developmental processes of retinal and wing vein cell formation (Dickson et al, 1992; Brunner et al, 1994), and because all of our genetic studies consistently suggested RSK as a negative regulator of retinal and wing vein cell differentiation, we doubted the reliability of the established hypothesis on the downstream role of RSK in the Ras/ERK pathway in Drosophila. Moreover, RSK-null embryos frequently displayed partial deletion of the abdominal denticle belts (Supplementary Figure 3), which is highly similar to those caused by gain-of-function mutation of ERK (Brunner et al, 1994) and Torso (Klingler et al, 1988). Therefore, we suspected that RSK may have an antagonizing function against the Ras/ERK pathway in Drosophila. To substantiate whether RSK indeed suppresses the Ras/ERK signaling activity in Drosophila, we performed various genetic interaction assays between RSK and the components of the Ras/ERK pathway. First, we tested whether downregulation of Ras/ERK signaling enhances the RSK-overexpression phenotypes. Although the eyes with heterozygotic mutation of Ras, Raf, MEK, or ERK itself showed no change in the number of R cells (Kim et al, 2004 and data not shown), overexpression of RSKWT with heterozygotic mutation of Ras (Figure 4A), Raf (Figure 4B), MEK (Figure 4C), or ERK (Figure 4D) resulted in more severe roughened-eye phenotypes with further decreased number of R cells, when compared to the eyes overexpressing RSKWT in a wild-type genetic background (Figure 2B). Next, we examined whether RSK-null mutation enhances gain-of-function phenotypes of the Ras/ERK pathway components. As previously reported, ectopic expression of constitutively active Ras (RasV12) or Raf (RafF179) caused roughened eye phenotypes with extra R cells in some ommatidia (Supplementary Figure 4D, F, H, and J). Interestingly, these phenotypes were further enhanced in heterozygotic or hemyzygotic backgrounds of RSKD1 (Supplementary Figure 4E, G, I, and K). These two series of experiments consistently suggested that RSK negatively regulates the Ras/ERK pathway. Figure 4.Genetic interaction between RSK and the components of the Ras/ERK pathway. (A–N) SEM images (upper panels) and tangential cross sections (lower panels) of Drosophila adult eyes. Detailed genotypes of (A–N) are described in Supplementary data. The numbers indicate the mean number of photoreceptor cells per ommatidium±s.d. (O) RSK null mutation rescued the lethality of hemizygote RafHM7 males at 25°C. Adult viability was scored as previously described (Kim et al, 2004). Download figure Download PowerPoint We next questioned whether downregulation of Ras/ERK signaling suppresses the RSK-null phenotypes. Interestingly, RSKD1 phenotypes were significantly suppressed by heterozygotic mutation of Ras or ERK (Supplementary Figure 4B and C), demonstrating that RSK-null phenotypes result from upregulation of the Ras/ERK pathway. Conversely, we hypothesized that if RSK-null mutation indeed upregulates the Ras/ERK pathway, it should relieve the phenotypes caused by a low activity in the Ras/ERK signaling pathway. As expected, RSKD1 mutation strongly suppressed the eye roughness and R-cell number decrease phenotypes in a hypomorphic Raf mutant (RafHM7; Figure 4N, compared to Figure 4M), and even rescued the temperature-sensitive lethality (although RafHM7 hemizygote males are viable at 18°C, they are lethal at 25°C; Melnick et al, 1998) of RafHM7 mutants at 25°C (Figure 4O). Collectively, these results coherently showed that RSK is a negative regulator of the Ras/ERK pathway in vivo. Finally, we assessed whether RSK overexpression suppresses the phenotypes caused by the upregulation of the Ras/ERK pathway. Strikingly, the gain-of-function phenotypes induced by a receptor tyrosine kinase (sevs11; Figure 4E), Ras (sev>RasV12; Figure 4F), and Raf (sev>RafF179; Figure 4G) were almost completely suppressed by expression of RSKWT (Figure 4I–K, respectively), supporting that RSK inhibits the Ras/ERK pathway at the downstream of these molecules. However, interestingly, RSKWT (Figure 4L) entirely failed to suppress the gain-of-function phenotypes of ERK (sev>ERKSem; Figure 4H), showing the epistatic relationship between RSK and ERK. RSK inhibits in vivo activities of ERK by preventing its nuclear localization Next, we questioned how RSK inhibits the Ras/ERK pathway. From our genetic results, we found that RSK is epistatic to hyperactive RTK, Ras, and Raf, but not ERK. Hence, we presumed the direct regulation of ERK activity by RSK, without involving upstream components. Interestingly, it has been previously demonstrated that mammalian ERK is constitutively associated with RSK in quiescent cells and that the key residues of the docking domains of the two proteins play a crucial role in the interaction (mutation in the conserved Asp319 to Asn in the common docking (CD) domain of ERK (Dimitri et al, 2005; corresponding to ERKSem mutation in Drosophila) or a mutation in the conserved Arg742 (Arg902 in Drosophila RSK; Figure 6B) of RSK to Ala in the ERK-docking (D) domain (Roux et al, 2003) nullified their interaction in mammals). We predicted that the physical association between ERK and RSK also occurs in Drosophila since their interaction domains are highly conserved and both ERK and RSK are colocalized in the cytoplasm of larval eye discs (Figure 5A). Indeed, our hypothesis was confirmed by the fact that RSK physically associated with ERK under overexpressed (Figure 5B) or endogenous (Figure 5C) conditions. Moreover, ERKSem (ERKD334N) and RSKR902A mutants failed to bind RSK and ERK, respectively (Figures 5B and 6C, respectively). These results demonstrated that RSK–ERK interaction in Drosophila occurs via key residues in the D domain of RSK and the CD domain of ERK, as in mammals. Figure 5.Inhibition of ERK activity and its nuclear localization by RSK. (A) The eye discs of gmr>ERK+RSKWT (gmr-Gal4 UAS-RSKWT/UAS-ERK) flies co-expressing Myc-tagged RSK and HA-tagged ERK were stained with anti-Myc (red) or anti-HA (green) antibody. (B) Physical interaction between RSK and ERK. Lysates from ∼600 adult fly heads of the gmr-Gal4 (gmr-Gal4/+), gmr>RSKWT (gmr-Gal4 UAS-RSKWT/+), gmr>RSKWT+ERKSem (gmr-Gal4 UAS-RSKWT/UAS-ERKSem), and gmr>RSKWT+ERK (gmr-Gal4 UAS-RSKWT/UAS-ERK) flies were subjected to co-immunoprecipitation assay using anti-HA antibody. The anti-HA immune complexes were analyzed by immunoblot using anti-Myc antibody. Anti-Myc and -HA immunoblots were also performed using whole-cell lysates (WCL). Myc-tagged RSKWT, HA-tagged ERK and HA-tagged ERKSem were used. (C) Lysates of RSKD1 (RSKD1), WT (w1118), and hs>RSKWT (hs-Gal4/UAS-RSKWT) flies were subjected to co-immunoprecipitation assay using guinea-pig anti-RSK antibody. The anti-RSK immune complexes were analyzed by immunoblot using rabbit anti-ERK antibody or rabbit anti-RSK antibody. Anti-ERK immunoblots were also performed using WCL. Tubulin (tub) was used as a loading control. (D, E) Subcellular localization of RSK and ERK was determined in Drosophila eye discs. Eye discs with indicated genotypes were stained with anti-HA (D) or anti-dpERK (E) antibody (green), anti-Myc antibody (red), and Hoechst 33258 (blue). Detailed genotypes are described in Supplementary data. Each magnified image ( × 60 000) of the eye discs was shown as an inset. (F) The eye discs of sev>ERK (sev-Gal4 UAS-ERK/+) and RSKD1; sev>ERK (RSKD1; sev-Gal4 UAS-ERK/+) were stained with anti-HA antibody (green) and Hoechst 33258 (blue). (G) The eye discs of WT (w1118) and RSKD1 (RSKD1) flies were stained with anti-dpERK antibody (green) and Hoechst 33258 (blue). (H) Nuclear localization rates of HA-ERK (left graph, quantified from (F)) and dpERK (right graph, quantified from (G)) were presented as bar graphs (*P=6.83 × 10−5; **P=6.71 × 10−5). Error bars indicate the s.d. between eye discs. (I) Anti-β-galactosidase antibody staining of gmr-Gal4 rho-lacZ (gmr-Gal4/+; rho-lacZ/+) and gmr>RSKWT rho-lacZ (gmr-Gal4 UAS-RSKWT/+; rho-lacZ/+) eye discs. (J) Northern blot showing the transcription level of the pnt-P1 gene in WT (w1118), RSKD1 (RSKD1) and hs>RasV12 (hs-Gal4/UAS-RasV12) flies. 18S rRNA was used as a loading control. Download figure Download PowerPoint Figure 6.Essential roles of the physical association between RSK and ERK. (A) Various RSK mutants were generated as described in Materials and methods. Each black box indicates the domain containing a mutation, which changes the key amino-acid residue in the N-terminal kinase domain (NTKD) or the D domain. (B) Alignments of the D domain of Drosophila (d) RSK, human (h) RSK1-4, and avian (av) RSK1. The black boxes indicate the highly conserved residues among species. The asterisk indicates the key ERK-docking residue, Arg902 in Drosophila. This arginine was mutated to alanine in order to obtain RSKRA. (C) Co-immunoprecipitation assay was performed using lysates from ∼600 adult fly heads with indicated genotypes. The anti-HA immune complexes were analyzed by immunoblot using anti-Myc antibody. Anti-Myc and -HA immunoblots were also performed using whole-cell lysates (WCL) All RSK proteins were Myc-tagged, and ERK was HA-tagged. (D) RSK from flies with indicated genotypes was subjected to kinase assays using S6 protein as a substrate (Kim et al, 2000). The autophosphorylated RSK protein (**) and transphosphorylated S6 protein (*) were visualized by autoradiography. Anti-Myc and -tubulin (tub) immunoblots were performed using WCL. (E) Tangential cross sections of adult eyes (upper panels) and anti-Myc immunostaining of larval eye discs (lower panels). The numbers indicate the mean number of photoreceptor cells per ommatidium±s.d. (F) Microscopic views of adult wings (upper panels) and larval wing discs stained with anti-Myc antibody (lower panels). Detailed genotypes of (C–F) are described in Supplementary data. Download figure Download PowerPoint As various ERK-binding molecules modulate ERK activity by altering its subcellular localization (Kwon et al, 2002; Burack and Shaw, 2005; reviewed in Tanoue and Nishida, 2003) and because retinal cell fate determination is fully dependent on the nuclear localization of activated ERK in Drosophila (Kumar et al, 2003), we questioned whether RSK affects the intracellular localization pattern of ERK during Drosophila eye development. Thus, we expressed a hemagglutinin (HA)-tagged form of Drosophila ERK in the eye disc to detect ERK localization using anti-HA antibody. HA-ERK was predominantly distributed in the cytoplasm (Figure 5F), but hyperactive Ras strongly induced migration of HA-ERK into the nucleus (Figure 5D). Surprisingly, co-expression of RSKWT dramatically suppressed the Ras-induced nuclear localization of HA-ERK (Figures 5D and 7K), while co-expression of RSKWT/RA completely failed to generate the same result as RSKWT (Figure 7H and K). Consistently, even though RSKWT and RSKWT/RA were expressed at similar levels in the eye and wing, expression of RSKWT/RA could not inhibit retinal and wing vein cell differentiation (Figure 6E and F), showing that RSK–ERK association is essential for the physiological functions of RSK. Figure 7.Kinase-independent inhibition of ERK by RSK. (A–G) RSKKR functionally interacts with components of the Ras/ERK pathway. Upper panels show SEM images. Lower panels show tangential cross sections of adult eyes. The numbers indicate the mean number of photoreceptor cells per ommatidium±s.d. (H) Eye discs stained with anti-HA antibody (green), anti-Myc antibody (red), and Hoechst 33258 (blue). Each magnified image ( × 60 000) of the eye discs was shown as an inset. SEM images of Drosophila adult eyes were shown in bottom panels. (I) Anti-β-galactosidase antibody staining of eye discs. (J) The eye discs stained with anti-dpERK antibody (green), anti-Myc antibody (red),