|Home | About | Journals | Submit | Contact Us | Français|
Tissues that generate specialized cell-types in a production line must coordinate developmental mechanisms with physiological demand, although how this occurs is largely unknown. In the C. elegans hermaphrodite, the developmental sex-determination cascade specifies gamete sex in the distal germline, while physiological sperm signaling activates MPK-1/ERK in the proximal germline to control plasma membrane biogenesis/organization during oogenesis. We discovered repeated utilization of a self-contained negative regulatory module, consisting of NOS-3 translational repressor, FEM-CUL-2 (E3 ubiquitin ligase) and TRA-1 (Gli transcriptional repressor), which acts both in sex-determination and in physiological demand control of oogenesis, coordinating these processes. In the distal germline, where MPK-1 is not activated, TRA-1 represses the male fate as NOS-3 functions in translational repression leading to inactivation of the FEM-CUL-2 ubiquitin ligase. In the proximal germline, sperm-dependent physiological MPK-1 activation results in phosphorylation-based inactivation of NOS-3, FEM-CUL-2 mediated degradation of TRA-1 and the promotion of membrane organization during oogenesis.
A number of tissues are polarized production lines involved in the generation of highly specialized cell types. Examples include oogenesis within the gonad of many vertebrates and invertebrates and the crypt-villus axis of the mammalian gut (Simon and Gordon 1995; Ko et al. 1997). Oogenesis involves the constant production of oocytes (which are very large totipotent cells rich in cellular machinery, information molecules and nutrients) in a stepwise fashion for reproduction (Blumenfeld and Amit 1994; Matova and Cooley 2001). The generation of differentiated cells within such polarized tissue production-lines involves both developmental and physiological control mechanisms. Central to understanding polarized tissue function and homeostasis is uncovering the spatially integrated regulatory pathways that co-ordinate the developmental and physiological control of differentiated cell type production.
The germline of the adult C. elegans hermaphrodite gonad is a polarized assembly line for the production of oocytes (Figure 1; (Hirsh et al. 1976)). In the presence of sperm, major sperm protein (MSP) induces oocyte maturation/ovulation every ~23 min (McCarter et al. 1999; Miller et al. 2001) achieving continuous oocyte production, a process that requires the biogenesis and organization of plasma membranes and cytoplasmic constituents, regulating gene expression as well as progression of chromosomes through meiotic prophase. Conversely, in middle-aged adult hermaphrodites (which have exhausted their sperm) or mutant females that lack male germ cells, oocytes already produced are arrested in late meiotic prophase and oocyte production is dramatically downregulated.
The RTK-RAS-ERK pathway relays physiological and developmental extracellular signals through a conserved kinase cascade that results in phosphorylation and activation of the extracellular-signal regulated kinase (ERK) (Sundaram et al. 1996). Active ERK, in turn, controls biological processes through phosphorylation of substrate proteins (Chang and Karin 2001). The C. elegans ERK ortholog, MPK-1, controls at least seven different processes in hermaphrodite germline development, including membrane organization during oogenesis and progression of germ cell nuclei through pachytene of meiotic prophase (Lee et al. 2007). Each of the seven processes is mediated by multiple MPK-1 substrates, with additional substrates likely remaining to be identified (Arur et al. 2009). Activation of MPK-1 is (a) induced by the MSP signal and (b) spatially restricted to the medial and proximal regions of the oogenesis production line (Figure 1a, b) where MPK-1 dependent processes are executed (Miller et al. 2001; Lee et al. 2007).
An essential prerequisite for oogenesis is the developmental specification of oocytes / female fate. In the hermaphrodite germline, the male fate (sperm) is specified during larval development and female fate (oocyte) is specified throughout adulthood. Germline sexual fate in C. elegans is determined through an elaborate pathway involving more than 30 genes (Meyer 2005; Zarkower 2006; Ellis and Schedl 2007), part of which is shown in Figure 1c. Important for this study is a multi-step negative regulatory module (NOS-3/FEM-CUL-2/TRA-1 module) necessary for oocyte fate specification. NOS-3, a homolog of the Nanos RNA binding protein, binds to FBF-1 and FBF-2 (FBF), two nearly identical Pumilio RNA binding protein homologs, which together repress translation of the fem-3 mRNA (Zhang et al. 1997; Kraemer et al. 1999). FEM-3 along with the sex determination proteins FEM-1 and FEM-2, forms a subunit of the CUL-2-based E3 ubiquitin ligase complex (Starostina et al. 2007). As regulation of this CUL-2-based E3 ubiquitin ligase in the germline appears to be largely accomplished through FEM-3 levels and FEM-1, which is the substrate specificity subunit of the uniquitin ligase complex, we refer to it as the FEM-CUL-2 complex. The FEM-CUL-2 complex negatively regulates the TRA-1 Gli transcriptional repressor through ubiquitin mediated degradation (Starostina et al. 2007). In the adult hermaphrodite, NOS-3 and FBF repress the translation of fem-3 mRNA resulting in the FEM-CUL-2 E3 ubiquitin ligase complex being inactive, leading to stabilization of nuclear TRA-1 Gli. For the specification of the oocyte fate, TRA-1 represses the transcription of fog-1 and fog-3, genes that specify the sperm fate (Chen et al. 2000; Jin et al. 2001). Sexual fate is specified in the distal part of the germline, in the region of the germline stem cells (Crittenden et al. 2006; Hansen and Schedl 2006) determining the cell type generated within the gonadal production-line.
Here, we report an unexpected function of the NOS-3/FEM-CUL-2/TRA-1 module in regulation of plasma membrane organization during oogenesis. We demonstrate that MPK-1 promotes membrane organization by phosphorylation and inactivation of NOS-3, which through the negative regulatory module leads to degradation of nuclear TRA-1. We show that physiological control through MPK-1 and the NOS-3/FEM-CUL-2/TRA-1 module occurs in the medial and proximal germline where oocytes are being generated. By contrast, developmental control of sexual fate specification occurs in the distal germline. Thus, developmental and physiological mechanisms are integrated within the polarized germline tissue through spatially restricting the regulation and output of the NOS-3/FEM-CUL-2/TRA-1 module.
The adult hermaphrodite germline displays a honeycomb pattern of pachytene germ cells on the surface of the gonadal tube, with an interior rachis functioning in cytoplasmic transport, followed proximally by a single file row of growing oocytes (Figure 1a). Plasma membranes of pachytene cells and oocytes are visualized by GFP::PH(PLCδ) (Figure 2; (Audhya et al. 2005), while the interior surfaces of pachytene cells and growing oocytes are visualized by anilin ANI-2 staining, where non-staining ‘windows’ represent the opening in pachytene cells for deposition of cytoplasm and the opening in growing oocytes for cytoplasm uptake (Figure S1 (Maddox et al. 2005; Wolke et al. 2007). MPK-1 signaling is required for the organization of plasma membranes and the cytoskeleton in pachytene germ cells and growing oocytes (Lee et al. 2007). In mpk-1 null (mpk-1(0)) adults, the honeycomb organization of pachytene cells and the interior rachis is completely lost; pachytene arrested nuclei, largely devoid of plasma membranes, are present in clumps and surrounded by large nuclei free regions (Figures (Figures2,2, S1). We serendipitously found that the cellular disorganization of mpk-1(0) was significantly suppressed by a null allele of nos-3, which encodes a Nanos RNA binding protein homolog. nos-3(0);mpk-1(0) germlines show restoration of plasma membranes around nuclei in a semi-honeycomb pattern on the surface of the gonadal tube (Figure 2c) and restoration of the interior rachis, with ‘windows’ observed connecting small and large pachytene cells to a common cytoplasm (Figure S1c). nos-3(0) was also found to suppress the plasma membrane/ interior rachis disorganization phenotypes of lin-45 Raf null and mek-2 Mek null mutants (data not shown), genes immediately upstream of MPK-1 in the ERK signaling cascade. The restoration of plasma membranes by loss of nos-3 suggests that inappropriate NOS-3 activity is responsible, at least in part, for the membrane disorganization phenotype of mpk-1(0). Thus, a simple hypothesis is that MPK-1 phosphorylates and inactivates NOS-3.
In nos-3(0);mpk-1(0) germlines, restoration of the plasma membranes is incomplete (e.g. some cells contain multiple nuclei) and other MPK-1 dependent processes (Lee et al. 2007; Arur et al. 2009) are not restored (e.g. nuclei remain arrested in pachytene; Figure 2c). We previously identified six distinct substrates that function to promote membrane organization (Arur et al. 2009), and it is likely that one or more of these genes function redundantly to regulate membrane organization, consistent with the partial suppression by nos-3(0) of membrane organization.
Our genetic data raised the possibility that NOS-3 is an MPK-1 substrate. Characterized MPK-1 substrates harbor one or more identified ERK docking and phospho-acceptor sites (Jacobs et al. 1999; Fantz et al. 2001; Galanis et al. 2001; Barsyte-Lovejoy et al. 2002; Arur et al. 2009). We scanned the NOS-3 sequence and found one putative ERK docking site (a D-domain) and six potential phospho-acceptors (Figure 3a). Through in vitro kinase assays we determined that full length NOS-3 is a robust substrate of active ERK2 (Figures (Figures3,3, S2). Additional in vitro kinase analysis of an N-terminal (aa 1-520) and a C-terminal fragment (aa 520-871, containing the D-domain and RNA binding domain) indicated that active ERK2 phosphorylates the C-terminal fragment. We then used site-specific mutagenesis of the C-terminal NOS-3 fragment, generating unphosphorylatable sites (Serine (S) or Threonine (T) to Alanine (A)) to identify S644 and T648 as the major phospho-acceptors utilized by ERK2 to phosphorylate NOS-3 in vitro (Figures (Figures3b,3b, S2a-e). Thus, our in vitro biochemical and genetic data together support the model that MPK-1 phosphorylates and inactivates NOS-3 in order to promote membrane organization during oogenesis.
To assess whether NOS-3 is phosphorylated in vivo in an mpk-1 dependent manner, we generated antibodies that specifically recognize NOS-3 when phosphorylated on S644 and/or T648 (pNOS-3, Methods, Supplement). Western blot analysis of worm lysates identified a single band of an appropriate size in wild-type but not nos-3(0), demonstrating specificity (Figure 4a). In addition, no signal was present in mpk-1(0) lysates, demonstrating MPK-1 dependence (Figure 4a). In contrast, an antibody that specifically recognizes NOS-3 when neither S644 nor T648 are phosphorylated (non-pNOS-3, Methods) identifies the NOS-3 band in wild-type and mpk-1(0) whole worm lysates, but not in nos-3(0) lysates (Figure 4b). We conclude that NOS-3 is phosphorylated in vivo on S644 and/or T648 in an mpk-1 dependent manner.
To determine the spatial pattern of pNOS-3 in the germline, we dissected and co-stained wild-type hermaphrodite gonads with the anti-pNOS-3 antibody and an antibody specific for the diphosphorylated active form of MPK-1 (dpMPK-1). dpMPK-1 displays a dynamic pattern of activation in the germline (Miller et al. 2001; Lee et al. 2007) being first detected in the proximal pachytene region, followed by partial down-regulation in the loop region, and then robust activation in late stage oocytes (Figures (Figures1a,1a, ,4c).4c). The pattern of pNOS-3 largely mirrors that of dpMPK-1 activation; pNOS-3 is first detected in proximal pachytene coincident with the appearance of dpMPK-1, remains present during the rest of oogenesis, with a slight decrease in accumulation in the final oocyte (Figure 4c). pNOS-3 is not detected in nos-3(0) germlines, demonstrating specificity of the antibody (Figure S3a). Furthermore, pNOS-3 is not detected in mpk-1(0) germlines or in regions of the germline where MPK-1 is not active, such as the distal germline of wild-type hermaphrodites, or in adult females that lack MSP signal and thus MPK-1 activation (Figures (Figures4c,4c, S3b, data not shown). Together, these results demonstrate that NOS-3 phosphorylation is dependent on MPK-1 activation. In contrast, the non-pNOS-3 antibody shows a reciprocal pattern of accumulation compared to pNOS-3. Non-pNOS-3 accumulates from the distal end through mid-pachytene in wild-type germlines (Figure 4d). In addition, non-pNOS-3 accumulates throughout in the mpk-1(0) germlines (Figure S3d) due to lack of MPK-1 activity. Non-pNOS-3 is not observed in proximal pachytene and the rest of oogenesis in wild-type germlines where we detect dpMPK-1 and pNOS-3. The combined accumulation of pNOS-3 and non-pNOS-3 is consistent with total NOS-3 being present cytoplasmically throughout the adult hermaphrodite germline (Figure 4e). These results indicate, within the limits of detection by antibody staining, that phosphorylation of NOS-3 may go to completion in proximal pachytene and the remainder of oogenesis containing active MPK-1.
The presence of active MPK-1 switches NOS-3(S644/T648) from unphosphorylated to phosphorylated as germ cells transition from mid to late-pachytene, the precise region where plasma membrane defects arise in mpk-1(0) mutant germlines. Together with the genetic analysis these results suggest that MPK-1 regulates the organization and integrity of plasma membranes during oogenesis by phosphorylating and inactivating NOS-3.
As described in the Introduction, NOS-3 functions in the NOS-3/FEM-CUL-2/TRA-1 module as a part of the larger sex determination cascade in the regulation of hermaphrodite germline sexual fate, which occurs in the distal region of the tissue (Figure 1). However, phosphorylated NOS-3 along with active MPK-1 are found in more proximal germ cells whose sex is already specified and are now part of the oogenesis production line. We propose that in response to MSP/sperm signaling, MPK-1 in the proximal germline controls the activity of the NOS-3/FEM-CUL-2/TRA-1 module by phosphorylating NOS-3 to regulate a distinct essential function in germline development, plasma membrane organization during oogenesis (Figure 1). Our model predicts the following: 1) Physiological MSP signal promotes oogenesis and activates MPK-1, which in turn phosphorylates and inactivates NOS-3. Inactive NOS-3 can no longer repress fem-3 mRNA translation, which would then allow for the formation of a functional FEM-CUL-2 E3 ubiquitin ligase complex that degrades nuclear TRA-1. The absence of TRA-1 Gli family repressor from proximal pachytene nuclei would then lead to transcription of genes that promote membrane organization during oogenesis. 2) In the absence of the MSP signal oocyte production is arrested, MPK-1 is inactive and unphosphorylated NOS-3 can now repress fem-3 mRNA. The lack of fem-3 mRNA translation results in a failure to form a functional FEM-CUL-2 E3 ubiquitin ligase complex, which stabilizes nuclear TRA-1 and the repression of genes that promote membrane organization. Below, we describe our evaluation of this model through examination of NOS-3-FBF-1 binding, analysis of FEM-3 and TRA-1 accumulation in mpk-1(0) and NOS-3/FEM-CUL-2/TRA-1 module mutant backgrounds, and testing these mutants for interactions with mpk-1(0) for membrane organization phenotype.
fem-3 mRNA has been shown to be regulated by NOS-3 and FBF, where binding of NOS-3 to FBF-1 appears critical for the function of NOS-3 in translational repression (Zhang et al. 1997; Kraemer et al. 1999). The minimal region of NOS-3 required for binding to FBF-1 is aa 429-681 (Kraemer et al. 1999), which includes the residues phosphorylated by MPK-1 (S644 & T648). Since genetic analysis indicates that MPK-1 phosphorylation inactivates NOS-3, one possibility is that phosphorylated NOS-3 may be impaired in FBF-1 binding. To examine this possibility, we conducted in vitro pull-down assays with S35 labeled wild-type FBF-1 (produced by in vitro transcription/translation) combined with wild-type and mutant versions of bacterially generated recombinant FLAG::NOS-3. Wild-type and unphosphorylatable FLAG::NOS-3 (S644A,T684A) bind strongly to FBF-1 (Figure 5a). In contrast, the phospho-mimetic FLAG::NOS-3 mutant (S644, T648 to Glutamic Acid (E)) binds poorly to FBF-1 (Figure 5a), showing that mimicking the MPK-1 phosphorylation state (by a charged residue) reduces the ability of NOS-3 to bind FBF-1 in vitro and, by extension, supporting the model that NOS-3 phosphorylation reduces its binding to FBF-1 and thus its function in translational repression of fem-3 mRNA in vivo.
Genetic and molecular data indicate that NOS-3-FBF-1/-2 repress fem-3 mRNA translation (Barton et al. 1987; Zhang et al. 1997; (Kraemer et al. 1999) however, there is no information on FEM-3 accumulation in vivo. The germline sex determination pathway predicts that FEM-3 should be low in the distal germline of adult hermaphrodites where the oocyte fate is specified, while in the model for physiological control of membrane organization during oogenesis FEM-3 should be high in the presence of active MPK-1 and pNOS-3 in the proximal end (Figure 1). Using anti-FEM-3 antibodies we find that FEM-3 is absent from the distal wild-type adult germline, a pattern consistent with its function in the sex determination cascade. In physiological control, FEM-3 accumulates in proximal pachytene and diplotene/diakinesis oocytes, mirroring the accumulation of dpMPK-1 and pNOS-3 (Figures (Figures5b,5b, S4). Furthermore, FEM-3 accumulation depends on nos-3 and mpk-1 activity. In nos-3(0), FEM-3 accumulates throughout the distal germline, supporting a role for NOS-3 mediated repression of fem-3 translation in this region (Figure 5c). In mpk-1(0) germlines, FEM-3 fails to accumulate, consistent with active NOS-3 mediated translational repression throughout. Finally, in the nos-3(0);mpk-1(0) double mutant germlines FEM-3 accumulates through the germline, consistent with NOS-3 functioning downstream to MPK-1 to regulate FEM-3 levels (Figure S4d). Importantly, no staining is observed in fem-3(0) germlines (Figure S4c). These results confirm the genetic interpretation that MPK-1 phosphorylation inactivates NOS-3, allowing fem-3 mRNA to be translated during oogenesis.
FEM-3 functions as part of a CUL-2 E3 ubiquitin ligase complex to regulate TRA-1 accumulation during sex determination (Starostina et al. 2007). In the wild-type adult hermaphrodite high levels of nuclear TRA-1 are observed in the distal germline, consistent with TRA-1 promoting the oocyte fate (Starostina et al. 2007). However, for physiological control of membrane organization during oogenesis, our model predicts that when MPK-1 is active, TRA-1 levels will be low due to high levels of the FEM-CUL-2 E3 ubiquitin ligase complex, owing to the absence of NOS-3 mediated translational repression (Figure 1). This is exactly what we observe; in wild-type hermaphrodite germlines high distal nuclear TRA-1 levels fall dramatically as MPK-1 is activated (Figure 6). We next assayed fem-3(0), fem-1(0) and cul-2(RNAi) germlines, from mated animals to activate MPK-1, and find that nuclear TRA-1 persists much further proximally until the end of pachytene, indicating that nuclear TRA-1 level is regulated by FEM-CUL-2 complex in this region (Figure 6, data not shown, also see below). Consistent with the activity of the FEM-CUL-2 E3 ubiquitin ligase complex being regulated by FEM-3 levels, we do not detect nuclear TRA-1 in the germline of nos-3(0) (which has high FEM-3 levels), while nuclear TRA-1 accumulates throughout in mpk-1(0) germlines (Figure 6).
The above results support the model that NOS-3 is inactivated by MPK-1 phosphorylation, leading to fem-3 mRNA translation, which in turn targets nuclear TRA-1 degradation (Figure 1). As a direct test of the effect of NOS-3 phosphorylation status on regulation of FEM-3 and TRA-1, we generated transgenic worms carrying wild-type, phospho-mimetic and unphosphorylatable NOS-3 mutant proteins (Methods and Supplement). These transgenes were assayed for function in nos-3(0), where FEM-3 accumulates throughout the germline while nuclear TRA-1 is not detected (Figures (Figures5c,5c, ,6).6). The wild-type NOS-3 transgene rescues nos-3(0), restoring fem-3 translational repression and nuclear TRA-1 accumulation in the distal germline, with high FEM-3 accumulation and the absence of nuclear TRA-1 in proximal pachytene (Figures (Figures7,7, S5). The phospho-mimetic NOS-3(S644/T648>E) mutant is predicted to be inactive throughout the germline, mimicking continuous phosphorylation by MPK-1, while the unphosphorylatable NOS-3(S644/T648>A) mutant is predicted to be constitutively active for fem-3 translation repression, being refractory to downregulation by MPK-1. Consistent with these predictions, the phospho-mimetic NOS-3 mutant is unable to rescue nos-3(0), failing to restore fem-3 translational repression and nuclear TRA-1 accumulation, while the unphosphorylatable NOS-3 mutant displays fem-3 translational repression and nuclear TRA-1 accumulation throughout the germline (Figures (Figures7,7, S5). These data show that mimicking MPK-1 mediated phosphorylation by a charged residue disrupts NOS-3 function in translational repression of fem-3 mRNA in vivo.
nos-3(0) adult hermaphrodites display only a very low level of germline masculinization (Kraemer et al. 1999; data not shown). This is surprising as we observe high cytoplasmic FEM-3 (Figures (Figures5c,5c, ,7b,7b, S4) and low nuclear TRA-1 levels (Figures (Figures6,6, S5) in the distal germlines of nos-3(0) and nos-3(0) containing the phospho-mimetic nos-3 mutant transgene. The reasons that nos-3(0) adult hermaphrodite germlines are not masculinized despite the high FEM-3 and low TRA-1 levels is unclear. One speculative possibility is that NOS-3 also acts as a translational repressor of a gene that promotes the oocyte fate, downstream or in parallel to TRA-1, which in nos-3(0) is overexpressed and counteracts the high FEM-3 / low TRA-1 levels for germline sex determination.
While the protein accumulation patterns of FEM-3 and TRA-1 suggest a role for these proteins in oogenesis downstream to mpk-1, we directly investigated their function in this process. So far, we have shown that in mpk-1(0), nuclear TRA-1 levels are high in pachytene cells because FEM-3 levels are low, while in nos-3(0) and nos-3(0);mpk-1(0) nuclear TRA-1 levels are low because of high FEM-3 (Figures (Figures5b,5b, ,6,6, S7). These findings lead to two predictions. First, that suppression of the disorganized plasma membrane phenotype of mpk-1(0) by nos-3(0) is due to high FEM-3 accumulation (Figure S4) and thus high FEM/CUL-2 E3 ubiquitin ligase activity. Second, that the mpk-1(0) disorganized plasma membrane phenotype is caused, in part, by high nuclear TRA-1 in pachytene cells. If the first prediction is correct, then loss of FEM-CUL-2 E3 ubiquitin ligase activity should reverse the suppression of mpk-1(0) by nos-3(0). We find that reduction or elimination of fem-1, fem-3 or cul-2 activity reverses the suppression of mpk-1(0) by nos-3(0) (Table 1, lines 12-14). We note that fem-1(0), fem-3(0) and cul-2(RNAi) alone do not show a significant disorganized membrane phenotype; this is likely because there are MPK-1 substrates, in addition to NOS-3, that function in plasma membrane organization (Arur et al. 2009). If the second prediction is correct, that high nuclear TRA-1 contributes to the disorganized plasma membrane phenotype, then loss of tra-1 activity should suppress the mpk-1(0) membrane disorganization phenotype. However, tra-1 loss-of-function results in masculinization of the soma, precluding analysis of oogenesis phenotypes within the hermaphrodite gonad. We therefore performed tra-1 RNAi in mpk-1(0) that is also mutant for rrf-1, which causes RNAi to only function in the germline. rrf-1 encodes an RNA-dependent RNA polymerase necessary for RNAi in the soma but not the germline (Sijen et al. 2001). In such somatically female animals, the plasma membrane disorganization phenotype of mpk-1(0) was significantly suppressed following tra-1 RNAi (Table 1, line 9), similar to the nos-3(0);mpk-1(0) mutant (Figure 2). Analysis of additional mutant combinations (Table 1) indicates (a) that NOS-3 functions solely through TRA-1 (the extent of suppression of mpk-1(0) membrane disorganization is equivalent in doubles with nos-3(0) or tra-1(RNAi) or in the triple with nos-3(0) and tra-1(RNAi), lines 8, 9 and 15) and (b) that FEM-3 acts solely on TRA-1 (restoration of suppression in nos-3(0);mpk-1(0);fem-3(0), following tra-1(RNAi), lines 12 and 17) for plasma membrane organization during oogenesis. The genetic and molecular results together demonstrate that MPK-1, through phosphorylation of NOS-3, regulates TRA-1 accumulation in the control of membrane organization during oogenesis.
In germline sex determination, TRA-1 promotes the female fate through transcriptional repression of fog-1 and fog-3, genes that repress the sperm fate (Figure 1). In principle, membrane organization could have been mediated by fog-1 and fog-3, where the absence of TRA-1 in proximal pachytene cells would lead to their expression. However, fog-1 and fog-3 are not expressed in the oogenic germline (Chen and Ellis 2000; Thompson et al. 2005). Thus, TRA-1 likely regulates currently uncharacterized gene(s) for membrane organization during oogenesis (see Discussion).
The adult hermaphrodite germline has two physiological states: in the presence of sperm and the MSP signal, MPK-1 is activated and the production line generates oocytes while in the absence of sperm and the MSP signal, MPK-1 is not activated, oocytes are arrested in late prophase, and the production line is dramatically slowed. To assess the role of the absence of MSP signaling on the control of FEM-3 and TRA-1 levels, we examined fog-1(0) females and middle-aged adult hermaphrodites that had exhausted their self sperm. As predicted by the model when oogenesis is arrested (Figure 1), we observed nuclear TRA-1 throughout pachytene while FEM-3, pNOS-3 and dpMPK-1 levels are low (Figure S6, data not shown). Mating restores the pattern to that observed in wild-type adult hermaphrodites; low nuclear TRA-1 and high FEM-3, pNOS-3 and dpMPK-1 in proximal pachytene. (In contrast, in fem-1(0) or fem-3(0) mutants the observed pattern of TRA-1 accumulation, high nuclear throughout, is not altered in the presence or absence of sperm and activated MPK-1 because the FEM-CUL-2 complex is defective.) These results indicate that under conditions where the generation of oocytes is limited (no sperm), nuclear TRA-1 in proximal pachytene represses the transcription of genes that promote membrane organization. By contrast, in adult hermaphrodites with sperm and where oocytes are actively produced, TRA-1 is degraded, permitting genes that promote membrane organization to be transcribed.
In the sibling species C. briggsae, fem-1, fem-3 and tra-1 orthologs have conserved functions in somatic sex determination. However, for germline sex determination Cb-fem-1 and Cb-fem-3 are not required. For example, Cb-fem-3(0) animals are self-fertile hermaphrodites (Haag et al. 2002; Hill et al. 2006). To begin to examine if membrane organization during oogenesis in C. briggsae is controlled by MPK-1 and the NOS-3/FEM-CUL-2/TRA-1 module we examined Cb-mpk-1 function in germline development and examined dpMPK-1, pNOS-3 and TRA-1 accumulation in the germline using cross-species reacting antibodies.
RNAi of Cb-mpk-1 results in a spectrum of adult germline phenotypes that are very similar to those observed in C. elegans mpk-1 loss-of-function and null mutants (Table S1, Lee et al. 2007). Specifically, the honeycomb organization of pachytene cells on the surface and the interior rachis are completely lost with clumped nuclei arrested in pachytene (Figures S8, S9). Thus MPK-1 orthologs regulate membrane organization, as well as the progression through pachytene, during oogenesis in both C. briggsae and C. elegans.
Staining for dpMPK-1, pNOS-3 and TRA-1 in adult C. briggsae germlines gives essentially identical patterns as observed for the C. elegans orthologs (Figures S8, S9). Cb-pNOS-3 and Cb-dpMPK-1 are coincident, rising to a high level in proximal pachytene and extending through diakinesis oocytes, while nuclear Cb-TRA-1 accumulates from the distal end to mid-pachytene and then is not detected in proximal pachytene nuclei. To determine if the absence of TRA-1 from proximal pachytene was due to FEM-CUL-2 E3 ubiquitin ligase activity, we examined Cb-TRA-1 in the Cb-fem-3 null mutant nm63 (Hill et al. 2006). We find that Cb-TRA-1 is observed in proximal pachytene nuclei in Cb-fem-3(0) hermaphrodites (Figure S8), analogous to what is observed in C. elegans fem-1(0) or fem-3(0) mutants. These results support the idea that MPK-1 regulates the NOS-3/FEM-CUL-2/TRA-1 module to control membrane organization during oogenesis in C. briggsae as well as in C. elegans. It is currently unclear if the ancestor to C. elegans acquired fem function for germline sex determination, in which case the role the NOS-3/FEM-CUL-2/TRA-1 module in membrane organization during oogenesis is ancestral, or if the ancestor to C. briggsae lost fem function for germline sex determination. However, since these results indicate that Cb-fem-3 regulates Cb-TRA-1 stability in proximal pachytene, it is not obvious why similar Cb-fem-3 regulation of Cb-TRA-1 does not also have a function in C. briggsae sex determination.
Here we report an unexpected role for a module within the C. elegans sex determination cascade, NOS-3/FEM-CUL-2/TRA-1, to regulate plasma membrane organization during oogenesis. MPK-1 ERK regulates this multi-step negative regulatory module by phosphorylating and inactivating NOS-3 function which allows translation of fem-3 mRNA. FEM-3 then participates in the CUL-2 E3 ubiquitin ligase complex to help degrade nuclear TRA-1 Gli, resulting in the transcription of genes that promote membrane organization during oogenesis (Figure 1). The function of the NOS-3/FEM-CUL-2/TRA-1 module in membrane organization is spatially distinct from its role in germline sex determination and has different transcriptional targets for output. Furthermore, it is likely that the MPK-1 regulation of this module for membrane organization is conserved in the related nematode C. briggsae, even though the function of the module in germline sex determination is not (Haag et al. 2002; Hill et al. 2006). Below we discuss our findings.
Reversible phosphorylation is an important mechanism for regulating protein function. Examples of ERK phosphorylation affecting substrate function include changing protein activity (e.g. activating TIF-1A, (Zhao et al. 2003)), or protein function (e.g. converting LIN-1 Ets from a repressor to a transcriptional activator, (Leight et al. 2005)) and changing subcellular localization (EPLIN translocating from stress fibers to peripheral ruffles, (Han et al. 2007)). We show that MPK-1 phosphorylation of NOS-3 inactivates its function in fem-3 translational repression based on three lines of evidence: 1) genetic suppression of mpk-1(0) by nos-3(0), 2) failure of a phospho-mimetic NOS-3 version to translationally repress fem-3 mRNA in vivo, and 3) substantial reduction of in vitro binding of a phospho-mimetic version of NOS-3 to co-factor FBF-1, which binds the fem-3 mRNA.
Phosphorylation and inactivation of NOS-3 is restricted to the region of the oogenesis production line where mpk-1 dependent membrane organization is occurring based on staining with phospho-specific antibodies. Phosphorylation did not noticeably alter the cytoplasmic subcellular localization of NOS-3. We found, within the limits of sensitivity of antibody staining, that all detectable NOS-3 above background is converted to the phosphorylated form in regions of the germline containing active MPK-1, dividing the germline into zones of unphosphorylated and phosphorylated NOS-3. This conversion of NOS-3 to the inactive pNOS-3 form is likely important to relieve fem-3 translational repression and thus producing sufficient FEM-3 to efficiently degrade nuclear TRA-1.
During the physiological control of oocyte production, we propose that TRA-1 represses the transcription of genes that promote membrane organization under conditions where the production line is slowed down (absence of sperm and loss of MPK-1 activation), but is degraded to allow transcription of these genes when the production line is active (presence of sperm and MPK-1 activation). fog-1 and fog-3, two genes whose germline expression is repressed by TRA-1, for sex determination, are unlikely to function as TRA-1 transcriptional targets for membrane organization as they are not expressed in the oogeneic germline (Chen and Ellis 2000; Lamont et al. 2004). As a preliminary approach to assess genes that may be repressed by TRA-1 to modulate membrane organization during oogenesis we scanned two sets of potential target genes for the presence of the TRA-1 DNA binding sites (Methods). First, we examined a set of genes whose germline expression is MPK-1 responsive (Leacock and Reinke 2006). This analysis identified meg-1 and meg-2 as containing the TRA-1 site in their 5′ intergenic regions. meg-1 and -2 encode P-granule components required maternally for fertility (Leacock and Reinke 2008) but are not known to function in membrane organization; however, these genes have not been examined in the appropriate sensitized backgrounds to overcome redundancy among genes that function downstream of MPK-1 in membrane organization (Arur et al. 2009). Importantly, additional MPK-1 responsive genes, which may contain the TRA-1 binding sequence, likely remain to be identified as the temperature sensitive mpk-1 mutant used in the original expression study (Leacock and Reinke 2006) is not fully mutant at the restrictive temperature and not fully wild-type at the permissive temperature (Lee et al. 2007) thus reducing the sensitivity of the microarray analysis. This finding, although preliminary, supports the proposal that the MPK-1 transcriptional output is, in part, through indirect regulation of TRA-1 via the action of NOS-3/FEM-CUL-2. We next asked whether genes that encode MPK-1 substrates known to function in membrane organization during oogenesis (Arur et al. 2009) contain TRA-1 binding sites. These genes did not contain TRA-1 binding sites suggesting that are not direct transcriptional targets of TRA-1, although they could be indirect TRA-1 transcriptional targets.
We previously found that the seven MPK-1 controlled biological processes are mediated by multiple substrates and that different processes are executed by distinct sets of substrates (Arur et al. 2009). The membrane disorganization phenotype of mpk-1(0) is only partially suppressed by nos-3(0). We identified six distinct substrates of MPK-1 that function to promote membrane organization of pachytene cells (Arur et al. 2009), suggesting that there may be redundancy among a number of genes governing membrane organization during oogenesis downstream to MPK-1. Furthermore, since TRA-1 regulation does not appear to be acting directly through known membrane organization MPK-1 substrates as transcriptional targets (see above), the NOS-3/FEM-CUL-2/TRA-1 module is likely acting through a distinct gene(s), in parallel to the known substrates, to control membrane organization. Additionally, since nos-3(0) (and tra-1 RNAi) does not suppress other phenotypes in mpk-1(0) (e.g. progression of nuclei through pachytene), function of NOS-3, and thus the NOS-3/FEM-CUL-2/TRA-1 module, may be limited to membrane organization.
The NOS-3/FEM-CUL-2/TRA-1 module is a multi-step negative regulatory cassette that is utilized in the C. elegans hermaphrodite germline to integrate developmental and physiological mechanisms for function during oocyte production. At the distal end, the module acts in developmental control as part of the sex determination pathway to specify the oocyte fate, when NOS-3 and TRA-1 are active. The generation of oocytes, in the medial and proximal parts of the germline tissue, is controlled through a physiological mechanism: when the oogenesis production line is active in the presence of the MSP sperm signal, active MPK-1 ERK phosphorylates and inactivates NOS-3, leading to inactivation/degradation of TRA-1 and expression of genes that promote membrane organization necessary for generating oocytes. When the oogenesis production line is inactive in the absence of the MSP sperm signal, MPK-1 is inactive, NOS-3 and TRA-1 are active leading to repression of oogenesis membrane organization genes as part of downregulation of oocyte production. Spatial control of the NOS-3/FEM-CUL-2/TRA-1 module within the germline occurs at two levels. First, the physiological control mechanism, MPK-1 activation, is restricted to the medial/proximal proximal region and does not occur in the distal germline where the oocyte sexual fate is specified. Second, the transcriptional output from TRA-1 is distinct, for sex determination it is repression of fog-1 and fog-3 while for oogenesis it is repression of as yet to be identified membrane organization genes.
Wild-type and each of the corresponding mutant germlines were dissected from animals that were synchronized at L4/ adult molt and allowed to develop for 24 hours, except in some cases where deviation from this were dictated by the experimental design. Dissected germlines were then processed as described before (Lee and Schedl 2001; Lee et al. 2007; Arur et al. 2009) and treated with the respective primary antibody at 4°C overnight and secondary for 2 hours at room temperature.
All images were taken on a Zeiss compound microscope using AxioPlan 2.0 imaging software and Hamamatsu camera. Each dissected and stained gonad was captured as a montage, with overlapping cell boundaries and processed as described earlier (Lee and Schedl 2001; Lee et al. 2007; Arur et al. 2009).
Purified proteins were dialyzed into the kinase assay buffer for 2-4 hours (50 mM Tris-HCl, 10 mM MgCl2, 2 mM DTT, 1 mM EGTA, 0.01 % Brij 35 at pH 7.5). Purified ERK2 kinase was purchased from NEB (cat # P6080L) and reactions carried out as described previously (Jacobs et al. 1998; Jacobs et al. 1999; Fantz et al. 2001; Arur et al. 2009). Kinetic analysis was performed using seven concentrations of protein ranging from 0.2μṂ to 2μM, with results an average of 5 separate experiments. Km and Vmax were calculated from the Lineweaver-Burke plot and in each case the data closely approximated a straight line (Figure S2). Vmax was determined by assaying the amount of total phosphate incorporation per unit time/ enzyme amount (usually calculated to around 0.03pm). Total phosphate incorporation was calculated by using the measured CPM, the specific activity of P32, and the reaction volume. Relative acceptor ratio is Vmax/ Km and was normalized by setting value for myelin basic protein to 1.0.
Wild-type and mutant Flag-tagged NOS-3 were generated as described in the supplement. Purified FLAG::NOS-3 (wt / mutant) was then bound to Flag M2 beads (Sigma) for 1 hour at room temperature in binding buffer (20mM Tris, pH 7.5, 500 mM NaCl, 0.5% NP 40 ). The beads were then washed extensively in wash buffer (10mM Tris 7.5, 150mM NaCl). FBF-1 was generated in parallel using Ambion in vitro transcription translation coupled system with S35 labeled methionine. The S35 labeled FBF-1 extract was then combined with FLAG::NOS-3 beads for 2 hours at room temperature and the extensively washed in the wash buffer. Each FBF-1, FLAG::NOS-3 tube also contained 100 μM fem-3 3′ UTR (supplement). The bound protein, were then eluted with 3x Flag peptide as described earlier (Lee and Schedl, 2001) and run out on a 10% SDS-PAGE gel, dried and exposed to the autoradiograph. Quantitation of binding was determined by densitometry of the autoradiograph.
We thank Siqun Xu for assistance in generating transgenic lines. We thank Jim Skeath for helpful comments and Edward Kipreos and lab members for helpful discussions. We thank Alessandro Puoti for the anti-FEM-3 antibody, Eric Haag for the Cb-fem-3(nm63) mutant and Karen Oegema for the anti-ANI-1 antibody. We thank WormBase and the Caenorhabditis Genetics Center for resources. This work was possible through funding from the National Institutes of Health GM085150 and GM63310 to TS, GM053099 to DZ and T32HD007480 to MB.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.