pix-1 Controls Early Elongation in Parallel with mel-11 and let-502 in Caenorhabditis elegans

Cell shape changes are crucial for metazoan development. During Caenorhabditis elegans embryogenesis, epidermal cell shape changes transform ovoid embryos into vermiform larvae. This process is divided into two phases: early and late elongation. Early elongation involves the contraction of filamentous actin bundles by phosphorylated non-muscle myosin in a subset of epidermal (hypodermal) cells. The genes controlling early elongation are associated with two parallel pathways. The first one involves the rho-1/RHOA-specific effector let-502/Rho-kinase and mel-11/myosin phosphatase regulatory subunit. The second pathway involves the CDC42/RAC-specific effector pak-1. Late elongation is driven by mechanotransduction in ventral and dorsal hypodermal cells in response to body-wall muscle contractions, and involves the CDC42/RAC-specific Guanine-nucleotide Exchange Factor (GEF) pix-1, the GTPase ced-10/RAC and pak-1. In this study, pix-1 is shown to control early elongation in parallel with let-502/mel-11, as previously shown for pak-1. We show that pix-1, pak-1 and let-502 control the rate of elongation, and the antero-posterior morphology of the embryos. In particular, pix-1 and pak-1 are shown to control head, but not tail width, while let-502 controls both head and tail width. This suggests that let-502 function is required throughout the antero-posterior axis of the embryo during early elongation, while pix-1/pak-1 function may be mostly required in the anterior part of the embryo. Supporting this hypothesis we show that low pix-1 expression level in the dorsal-posterior hypodermal cells is required to ensure high elongation rate during early elongation.


Introduction
In mammals, the CDC42/RAC-specific Guanine-nucleotide exchange factor (GEF) a/b-PIX and the CDC42/RAC-specific effector kinase PAKs were shown to control cell migration, cell polarity, cytoskeleton remodeling and focal adhesion complex assembly/disassembly dynamics [1]. Their involvement in the control of epithelium morphogenesis and migration of epithelial sheets has also been recently established in mammals [2], and in model organisms such as Drosophila melanogaster and Caenorhabditis elegans [3][4][5]. Study of epithelial morphogenesis in C. elegans appears as an excellent model to better understand the function of a/b-PIX and PAKs during complex morphogenic events in living organisms.
In the nematode C. elegans, embryonic elongation involves the extension of the embryo along its longitudinal axis and a reduction of its transverse diameter, resulting in a 4-fold increase in length. This morphogenetic event involves dramatic changes in the shape of the epidermal (hypodermal) cells. Elongation is divided into an early and a late phase. The early phase, from comma to 1.75-fold stage -corresponding to embryos that are 1.75-fold in length compared to non-elongated embryos -, occurs through contraction of filamentous actin bundles (FBs) in hypodermal cells [6]. The hypodermis is composed of ventral, lateral (seam cells) and dorsal cells, which are linked by adherens junctions [7]. Contraction of FBs during early elongation is thought to be high in the seam cells and low in dorsal and ventral hypodermal cells [6,8].
The late phase of elongation involves mechanotransduction signaling from the body-wall muscles to the dorsal and ventral hypodermal cells [4]. At the 1.5-fold stage of development, muscle cells form connections, called trans-epidermal attachment structures (TEAs), with the dorsal and ventral hypodermis [9]. As embryos develop to the 1.75-fold stage, the muscles become functional and start contracting, thus inducing chemical changes in the overlying hypodermal cells through mechanical tension applied on the TEAs [4].
The signal transduction pathways that regulate early and late elongation have been extensively investigated over the last 15 years. Interestingly, many genes controlling morphological changes of the hypodermis during elongation are effectors or regulators of Rho GTPases [3,4,10,11]. Rho GTPases are molecular switches controlling a wide-range of cellular functions involving cell shape changes, cell migration, cell proliferation and differentiation [12]. They cycle between an ''ON'' GTP-bound form and an ''OFF'' GDP-bound form. When bound to GTP, they interact with specific effectors. They are regulated by three families of proteins: Guanine nucleotide-Exchange Factors (GEFs); GTPase-Activating Proteins (GAPs); and Guanine nucleotide-Dissociation Inhibitors (GDIs). To date, although three Rho GTPases (rho-1/RHOA, ced-10/RAC and mig-2/RHOG) have been implicated in pathways controlling elongation, only three of their regulators (GAPs and GEFs) have been shown to be involved in this process [4,10,11,13], suggesting that others remain to be identified.
LET-502/ROCK is an essential component of the let-502/mel-11 pathway and an essential regulator of elongation [14]. It is activated downstream of the Rho GTPase RHO-1/RHOA [15], that may itself be activated by the GEF RHGF-2 [13] and inactivated by the GAP RGA-2 [11]. Inactivation of RHO-1 by RGA-2, occurs in ventral and dorsal hypodermal cells during early elongation leading to inactivation of LET-502/ROCK and reduction of FB contractions in these cells [11]. RHO-1/RHOA and LET-502/ROCK may then be activated in the lateral hypodermal cells where most of the FB contractions may occur during early elongation. Consistent with this model, expression of MLC-4/MLC in the lateral cells can rescue mlc-4 loss-of-functionassociated elongation defects, while expression of MLC-4/MLC in ventral and dorsal cells cannot [3]. In seams cells, MEL-11/PP-1M may be inhibited by LET-502/ROCK and MRCK-1/ MRCK presumably through phosphorylation [3,8].
To date, the genes controlling the pak-1 pathway during early elongation remain unknown. Moreover, the biological significance of the functional redundancy of the mel-11/let-502 and pak-1 pathways is not clear. This redundancy is intriguing since it does not appear to add robustness to the elongation system: a single perturbation in any component of the mel-11/let-502 pathway induces a high proportion of embryonic lethality [10]. This suggests that the mel-11/let-502 and pak-1 pathways have unique functions during elongation that remain to be identified.
The CDC42/RAC-specific GEF, PIX is a well-known activator of PAKs in several organisms [1]. In C. elegans, pix-1 codes for a protein homologous to the mammalian b-PIX ( Figure S1). It contains a Src-homology 3 (SH3); a GEF/dbl homology (DH); a GIT-binding (GBD), and a PDZ binding (ZB) domains ( Figure  S1). These conserved domains are shown in mammals to mediate b-PIX interaction with PAK1-3, the Rho GTPases CDC42 and RAC, the phosphatases POPX1/2, the ARFGAP GIT, the tumor suppressor Scribble and the postsynaptic density protein Shank [1]. In C. elegans, PIX-1 was shown to activate PAK-1 in a GTPase-independent manner in migrating distal tip cells (DTC) during gonad morphogenesis in larvae [17]. In this system, PAK-1 activation still appears to be at least partially dependent on CED-10/RAC [18]. PIX-1 was also shown to activate PAK-1 through the GTPase CED-10/RAC in hypodermal cells during late elongation of embryos [4].
In this study, we demonstrate that pix-1 controls early elongation in parallel with mel-11/let-502. Our data suggest that pix-1 is a novel component of the pak-1 pathway while retaining some function during early elongation independent from pak-1. We show that the pix-1/pak-1 pathway controls the anteroposterior morphology of the embryo during early elongation through regulating head width, while let-502/ROCK controls both the head and tail width. Our study proposes a novel model for early elongation where the mel-11/let-502 and pix-1/pak-1 pathways have redundant and complementary functions to shape the antero-posterior axis of the embryo.
Phenotypic characterisation of animals carrying pix-1(gk416), pix-1(ok982) and pak-1(ok448) alleles revealed low penetrance embryonic lethality (Emb) and early larval arrest (Lva) confirming previous findings (Table 1) [4]. Measurements of pix-1(gk416), pix-1(ok982), pak-1(ok448) and let-502(sb118ts) arrested larvae showed that they were significantly shorter than synchronized wt L1 larvae at 25.5uC (T-test, p,0.01) ( Figure 1A and B). To better characterize the elongation defects associated with the pix-1, pak-1 and let-502 alleles, we measured the duration of early elongation using four-dimensional light microscopy. To do so, embryos were collected through dissection of hermaphrodites grown at 25.5uC, and embryonic development was imaged at 23-24uC. The duration of early elongation was measured from the 1.2-fold stage to the beginning of late elongation -when body wall muscles started contracting ( Figure 1C). Animals carrying let-502(sb118ts) (n = 10), pix-1(gk416) (n = 25) and pak-1(ok448) (n = 20) alleles developed significantly slower than wt animals (n = 17) during early elongation ( Figure 1C). We also measured the length of the embryos at the end of early elongation ( Figure 1D), and found that mutant embryos were significantly shorter than wt embryos ( Figure 1D). These data demonstrate that the elongation rate of pix-1, pak-1 and let-502 mutant embryos is significantly reduced during early elongation, and provide the first evidence of a requirement for pix-1 at this stage. They also confirm the involvement of pak-1 during early elongation [3].
Interestingly, pix-1 and pak-1 arrested larvae had very similar morphologies and behavior. For example, they were thin and clear when compared to let-502 arrested larvae ( Figure 1A). Furthermore, they appeared to have different antero-posterior morphologies when compared to wt and let-502 larvae. When the width of the head (H3, Figure 2A and B) and the width of the tail (T3, Figure 2A and B) were measured and combined as a head/tail width ratio (H3/T3, Figure 2B), pix-1(gk416), pix-1(ok982) and pak-pix-1 Controls Early Elongation 1(ok448) had significantly higher ratios than wt and let-502 larvae (H3/T3, Figure 2B, data not shown). This indicates that pix-1 and pak-1 mutant arrested larvae present a morphological alteration characterized by an inflated anterior part of their body when compared to their posterior part. These mutants arrested larvae also had severe pharynx pumping defects and were completely paralyzed, although pumping in L3 escapers was similar to wt larvae ( Figure S2). While these later phenotypes may be physiological consequences of primary developmental phenotypes, they support the fact that a similar array of phenotypes is observed in pix-1 and pak-1 mutant animals. These data support then the hypothesis that pix-1 and pak-1 could be part of the same developmental pathway. This also suggests that pix-1 and pak-1 control the relative width of the head vs. the tail of developing embryos, a process that does not seem to require let-502.
pix-1 and pak-1 control the head to tail width of the embryos during early elongation Since pix-1 and pak-1are also involved in the control of late elongation, we investigated whether their function during early elongation is required to control the head to tail width of the animal. To do so, we measured the H/T ratio of pix-1(gk416), pak-1(ok448) and let-502(sb118ts) embryos at the beginning, 1.2-fold stage (H1/T1, Figure 2A and B), and at the end of the early elongation (H2/T2, Figure 2A and B). These experiments were done using the same populations of embryos characterized previously ( Figure 1C and D). While no change, or a reduced H/T ratio was observed in pix-1, pak-1 and let-502 mutants when compared to wt animals at the beginning of early elongation (H1/ T1, Figure 2A), all three mutants showed a significantly higher H/ T ratio than wt embryos at the end of early elongation (T-test pvalues,0.006; H2/T2, Figure 2A). This shows that the inflated head vs tail morphology observed in pix-1 and pak-1 arrested larvae is also observed at the end of early elongation in pix-1, pak-1  and let-502 mutant embryos, suggesting that pix-1, pak-1 and let-502 control the head to tail width of the embryos at that stage. We then measured the reduction of head (H) and tail (T) width during early elongation. To do so, we compared the width of the head and the width of the tail of the embryos at the beginning and the end of early elongation (H1/H2 and T1/T2 respectively; Figure 2C). In wt embryos the head and the tail appeared to be 1.4-and 1.16-fold wider at the beginning than at the end of early elongation ( Figure 2C). This shows that the head width reduces more than the tail width during early elongation in wt animals. We found that the head width of the pix-1 and pak-1 mutant embryos reduced significantly less during early elongation than wt embryos ( Figure 2C left panel). We also found that the tail width reduced similarly in pix-1 and pak-1 mutant embryos than in wt embryos. Interestingly, the degrees of head and tail width reduction in let-502(sb118ts) embryos were significantly lower than wt embryos and not significantly different than 1, suggesting that in let-502 mutant animals neither the head, nor the tail width of the animal reduce during early elongation ( Figure 2C). Together, these data suggest that the higher reduction of the head versus the tail width of the embryo during early elongation requires pix-1, pak-1 and let-502 function. These three genes appear to be required for reduction of the head width of the embryo, while only let-502 may be required for reduction of the tail width.
As previously shown, 100% of mel-11(it26); let-502(sb118ts); pak-1(ok448) embryos failed to hatch and displayed early elongation arrest ( Table 2, line 12) [3]. These results suggest that pix-1 functions in parallel with the let-502/mel-11 pathway during early elongation, as shown for pak-1. However, the increased phenotypic severity observed in pak-1 vs. pix-1 mutants suggests that pix-1 functions redundantly with a gene or a group of genes in parallel with mel-11/let-502 during early elongation. The different relationship observed between mel-11 and pix-1 or pak-1 mutants also suggests that pix-1 and pak-1 have independent functions during early elongation. Whether this is the case during late elongation, could not be ascertained from our data. PIX-1::GFP is homogeneously distributed in the cytoplasm and at cell periphery of hypodermal cells pix-1 was previously shown to be involved in mechanotransduction signaling in ventral and dorsal hypodermal cells upon muscle Early elongation arrest include embryos arresting between comma and 1.75-fold with or without rupturing. * arrested larvae present a pix-1(gk416) specific morphology ( Figure 1A). ** arrested larvae present a let-502 (sb118ts) specific morphology ( Figure 1A). *** late elongation arrest without hatching.
$ all arrested embryos rupture. contraction during late elongation events [4]. The localisation of PIX-1 at the TEA in dorsal and ventral hypodermal cells supports this function [4]. Our data suggest that pix-1 also controls early elongation events. These events are thought to be directed by the contraction of filamentous actin bundles (FBs) at the cell periphery and at the most apical part of the lateral hypodermal cells [3]. To better understand the function of PIX-1 during early elongation, we characterized its subcellular localization in hypodermal cells during early elongation. To do so, we examined the localisation of PIX-1::GFP expressed under the control of the pix-1 endogenous promoter in pix-1(gk416) animals immuno-stained with MH27 antibodies (staining of hypodermal adherens junctions; Figure 4) or expressing the filamentous actin-binding probe VAB-10 ABD ::m-Cherry in hypodermal cells ( Figure S4). Confocal microscopy analysis revealed that during early elongation, PIX-1::GFP is expressed in the dorsal ( Figure 4A-C, Figure S4A-F), lateral ( Figure 4D-F) and ventral hypodermal cells ( Figure 4G-I, Figure  S4D-F). Throughout early elongation, PIX-1::GFP is located at the TEA in dorsal and ventral hypodermal cells, as previously reported (Figure 4 E, arrow-head) [4] and is also homogeneously distributed in the cytoplasm and at the cell periphery of all expressing cells (Figure 4, Figure S4). Interestingly, at the comma stage, PIX-1::GFP expression appears to be reduced in several posterior dorsal hypodermal cells (Figure 4 B, arrowhead). Concomitant with the fusion of these cells, around the 1.2-fold stage, the expression of PIX-1::GFP appears to be reduced in the fused cells ( Figure 5A). Considering that pix-1 controls reduction of the head but not the tail width of the embryos during early elongation, we were intrigued by the reduction of the expression of PIX-1::GFP in the dorsal-posterior hypodermal cells in the transgenic animals. We then quantified PIX-1::GFP expression in dorsal-anterior (DA), dorsal-posterior (DP), ventral-anterior (VA), ventral-posterior (VP), lateral-anterior (LA) and lateral-posterior (LP) hypodermal cells (Figure 5 A). This study revealed a significant lower ratio of DP/ DA and DP/V expression when compared to LP/LA, VP/VA, DA/V and DA/L ratios, which were not significantly different than 1 (Figure 5 B). Similar data were obtained using two independent transgenic lines expressing translational fusions of PIX-1::GFP under the control of the pix-1 endogenous promoter, one transgenic line carrying an extrachromosomal array (pix-1(gk416);sajEx1) and one stable transgenic line carrying an integrated array (unc-119(ed3);pix-1(gk416);sajIs1 see Methods). We also found that all measured ratios were constant throughout early elongation ( Figure S5).
We then observed the elongation rate of embryos at the single animal level and attempted to see if there was any correlation between a given expression pattern of the transgene and the ability of the transgene to efficiently rescue the mutant elongation rate defect. We found that transgenic embryos with similar elongation rates to wt embryos (elongation$288 nm/min, Figure 6D) had DP/DA and DP/V ratios of PIX-1::GFP intensity that were significantly lower than embryos that elongated slower (elongation,288 nm/min, p-value,0.01; Figure 6D). No significant difference was observed for the LP/LA, VP/VA DA/V and DA/L ratios between the two populations of embryos ( Figure 6D). These data show that PIX-1::GFP expression was significantly reduced in dorsal-posterior cells when compared to dorsal-anterior and ventral cells in rescuing animals. Such reduction of PIX-1::GFP expression in dorsal-posterior cells was not observed in non-rescuing animals. Importantly, animals presenting a DP/DA ratio higher than 0.5, elongated almost 3 times slower than wt animals ( Figure 6A). Overall, the speed of early elongation appeared to be negatively correlated to the DP/DA ratio of PIX-1::GFP expression (Spearman correlation coefficient R 2 .0.427; pvalue,0.002; Figure 6C), but not to any other PIX-1::GFP intensity ratio measured in hypodermal cells ( Figure S6). Similar results were obtained using two independent pixp::pix-1::GFP expressing transgenic lines, one carrying an extrachromosomal array (pix-1(gk416);sajEx1), and a stable transgenic line carrying an integrated array (unc-119(ed3);pix-1(gk416);sajIs1). These data suggest that transgenic animals expressing PIX-1::GFP homogeneously in dorsal-anterior, lateral and ventral hypodermal cells, but two times less in dorsal-posterior cells, elongate at a wt-rate during early elongation. However, animals expressing the transgene in dorsal-posterior cells at a similar/higher level when compared to other cells elongated significantly slower.
PIX-1::GFP is expressed in cells other than the hypodermis, and we cannot exclude the possibility that differential expression of PIX-1::GFP in other cells may affect the rescuing ability of the transgene. To test this possibility, we generated transgenic animals expressing PIX-1::GFP under the control of the hypodermal specific promoter, lin-26p. This promoter is thought to drive the expression of coding sequences only in hypodermal cells and in a homogenous manner [3]. We confirmed this later assumption through measurements of the DP/DA fluorescence intensity ratios in lin26p::vab-10(ABD)::GFP and lin-26p::vab-10(ABD)::mCherry transgenic animals carrying integrated arrays expressing actinbinding fluorescent probes in all hypodermal cells under the control of lin-26p [3]. We showed that these ratio were not significantly different than 1 ( Figure 5 C and D).
As observed with pix-1p::pix-1::GFP expressing animals, transgenic animals carrying an integrated array expressing PIX-1::GFP under the control of lin-26p segregated into two populations. The embryos of the first population displayed an elongation rate similar to wt (elongation rate$288 nm/min, Figure S7) and had DP/DA and DP/L intensity ratios that were significantly lower than the second population of embryos that elongated slower (elongation rate,288 nm/min, Figure S7). As shown for transgenic animals expressing pix-1p::pix-1::GFP, all animals expressing PIX-1::GFP under the control of lin-26p express similar levels of the transgene in DA, L and V cells ( Figure S7). These data support our hypothesis that a DP/DA PIX-1::GFP expression ratio higher than 0.5 may hinder early elongation.
To determine if the lower DP/DA, DP/L and DP/V intensity ratios observed in embryos with a wt-like elongation rate were due to a reduced expression of PIX-1::GFP in DP cells or to an increased expression of PIX-1::GFP in DA, L and V cells, we measured the fluorescence intensity of PIX-1::GFP in DA and DP cells in lin-26p::pix-1::GFP expressing animals. We found that the PIX-1::GFP fluorescence intensity was not significantly different in the DA cells of embryos developing at a wt-rate when compared to those developing at slower rates (Figure 6 E). However, PIX-1::GFP fluorescence intensity appeared to be significantly lower in the DP hypodermal cells in the subpopulation of embryos that elongated faster. This phenomenon was observed in two independent stable transgenic lines. These data suggest that the DP/DA ratios in embryos that elongated at a wt-rate was not due to an increased expression of PIX-1::GFP in the DA cells, but to decreased PIX-1::GFP expression in the DP cells.
Altogether, these data suggest that homogenous expression of PIX-1::GFP in dorsal-anterior, lateral and ventral hypodermal cells, and decreased expression in dorsal-posterior cells significantly rescued the elongation defects observed in the pix-1(gk416) animals. They also suggest that expression of the transgene in dorsal-posterior cells above a threshold corresponding to half the expression in the other hypodermal cells decreases the efficiency of early elongation. This suggests that pix-1 may be submitted to a tight control of its expression level in hypodermal cells, particularly to decrease its level in the dorsal-posterior cells.

Discussion
Embryonic elongation transforms the ovoid embryo into the long, thin vermiform nematode. This morphogenetic process occurs in two phases. The early elongation is driven by the contraction of Filamentous actin Bundles (FBs) in hypodermal cells. The late elongation is driven by a mechanotransduction pathway in the dorsal and ventral hypodermis resulting from contraction of the underlying muscle cells.
Two parallel pathways control early elongation. One pathway involves two kinases: the RHO-1/RHOA effector LET-502/ ROCK and the ortholog of the CDC-42-effector human myotonic dystrophy kinase MRCK-1/MRCK [8,10]. Both kinases are thought to inhibit the function of MEL-11/PP-1M thus permitting phorphorylation of myosin-light chain (MLC-4/MLC) and contraction of FBs. The second pathway involves the CDC42/ RAC-effector PAK-1, the PP2C phosphatase FEM-2/POPX2 and a RHO/RAC GTP nucleotide-exchange factor (GEF) UNC-73/ TRIO. The function of these two later genes in the regulation of MLC-4/MLC phosphorylation and/or PAK-1 function remains unknown [3,8,10,16]. Downstream of the two parallel pathways LET-502/ROCK and PAK-1 are thought to phosphorylate MLC-4/MLC in the hypodermal cells and to control contraction of FBs. This contraction is assumed to be high in the lateral and  [11].
In this study, we show that the CDC42/RAC-GEF pix-1 controls early elongation in parallel with the mel-11/let-502 pathway. Our data suggest that pix-1 controls this process as part of the pak-1 pathway in parallel with a gene (or a group of genes) that remains to be identified. pix-1 may also function independently from pak-1. Our data suggest that PIX-1/PAK-1 and LET-502/ROCK have different function along the antero-posterior axis of the embryo during early elongation: PIX-1, PAK-1 and LET-502/ROCK appear to control the constriction of the head while only LET-502/ROCK controls the constriction of the tail of the elongating embryo. Our study also revealed that PIX-1::GFP fusion protein is homogeneously distributed in the cytoplasm and at the cell periphery of hypodermis during early elongation. We showed that this fusion protein rescues the reduced elongation rate observed pix-1(gk416) in a subset of transgenic animals expressing the transgene homogenously in dorsal-anterior, lateral and ventral cells and at a lower level in dorsal-posterior cells. Our results also suggest that PIX-1 expression above a certain threshold in dorsalposterior hypodermal cells is detrimental for early elongation.

pix-1 functions with unidentified genes in parallel with mel-11/let-502
Both pix-1 and pak-1 mutants have similar elongation phenotypes and both aggravate the elongation defects observed in mel-11; let-502 double mutant suggesting that these two genes control together a subset of developmental mechanisms in parallel with mel-11; let-502 during early elongation. However, pix-1 mutant aggravates elongation defects associated with mel-11; let-502 to a lesser extent than the pak-1 mutant. This suggests that pix-1 functions with other genes that act in parallel with mel-11/let-502 pathway (Figure 7).
In both mammals and C. elegans bPIX/PIX-1 was shown to activate PAK-1 kinase activity in a GTPase-dependent (canonical) or GTPase-independent (non-canonical) manner [4,17,18]. M. Labouesse's laboratory (IGBMC, Illkirch, France) showed using GTPase pull-down assays that the level of activation of the Rho GTPase CED-10/RAC was significantly lower in pix-1(gk416) than in wt embryos, suggesting that PIX-1 regulates CED-10/ RAC activity during embryonic development [4]. We can then hypothesize, that PIX-1 may activate PAK-1 in a canonical manner through CED-10/RAC during early elongation as shown during late elongation [4]. Supporting this hypothesis, CED-10/ RAC was located at both cell junctions and at the TEA within hypodermal cells in elongating embryos [4,20]. However, we cannot exclude the possibility that PIX-1 may also activate PAK-1 during early elongation in a GTPase-independant/non-canonical manner as shown during gonad morphogenesis [17,18]. Interestingly, the RHO/RAC specific GEF unc-73/TRIO was also shown to control early elongation in parallel to mel-11/let-502 [10]. The allele used in that study, rh40, consists of a missense mutation that eliminates its exchange activity towards CED-10/RAC, RAC-2/ RAC and MIG-2/RHOG without affecting its activity towards RHO-1/RHOA [15,21]. A UNC-73/TRIO -CED-10/RAC -PAK-1 pathway would then be an excellent candidate pathway controlling early elongation in parallel with a non-canonical PIX-1 -PAK-1 or a canonical PIX-1 -CED-10/RAC -PAK-1 pathway (Figure 7). Careful study will be required to test these hypotheses and to better understand the molecular mechanisms involving PIX-1 and controlling the activity of PAK-1 during early elongation.
It was shown that the fem-2 null allele induces weak larval arrest with non-elongated larvae similar to that observed in pix-1 and pak-1 mutants [10]. The fem-2 mutant can also aggravate mel-11; let-502 double mutants and suppress the mel-11 rupturing phenotype. [10,16]. This suggests that pix-1, and fem-2/POPX2 function together to control early elongation in parallel with mel-11/let-502. Interestingly, the PP2C-like serine/threonine phosphatases, POPX2/Protein Phosphatase 1F, the closest homolog of FEM-2 in mammals, was shown to interact with b-Pix (ortholog of PIX-1) and to dephosphorylate and inactivate PAK1 (ortholog of PAK-1) kinase activity in mammalian cells [22]. It was suggested that the b-PIX -POPX2 complex controls the activation/inactivation turnover of PAK1 in mammalian cells [22,23]. Considering that these proteins are highly conserved between C. elegans and mammals, we hypothesize that PIX-1, FEM-2/POPX2 and PAK-1 may have functional relationships similar to their homologs in mammals. Following this hypothesis, PIX-1 and FEM-2/POPX2 may control the activation/inactivation turnover dynamics of PAK-1 during early elongation. Considering that pix-1 and fem-2 mutant have similar interactions with mel-11 and let-502 during early elongation, this hypothesis suggests that a reduced or a sustained activation of PAK-1 alters similarly early elongation process. While this hypothesis would explain the genetic data obtained with pix-1, fem-2 and pak-1 mutants during elongation, it still remains to be confirmed through a careful analysis of a possible relationship existing between PAK-1 activation/inactivation dynamics and the regulation of myosin contraction in C. elegans hypodermal cells during early elongation.

pix-1 may control early elongation in a pak-1independent manner
Considering the molecular analysis of mel-11(it26) allele [19], we cannot exclude the possibility that this allele may not be completely null even at 25.5uC. We can then hypothesize that pix-1 may function together with pak-1 in parallel with mel-11 and may also function upstream of mel-11 independently of pak-1. This would constitute an alternative explanation for the suppression of mel-11(it26)-inducing rupturing by pix-1 but not by pak-1 allele. Interestingly, the kinase MRCK-1/MRCK, whose mammalian ortholog, is an effector of CDC-42 [24] was shown to control early elongation upstream of mel-11 [3]. CDC-42 being expressed in hypodermal cells during early elongation [20], and being also a potential target of PIX-1 GEF activity, we can hypothesize that PIX-1 may activate MRCK-1 through CDC-42 upstream of MEL-11 and may also activate PAK-1 in parallel with mel-11/let-502 (Figure 7).
The pix-1/pak-1 pathway mainly controls the constriction of the head of the embryos during early elongation Observation of the head to tail morphology of the elongating embryo showed that the anterior part of the embryo at 1.2-fold stage is much wider than the posterior part. This suggests that contraction forces applied on the anterior part of the embryo is higher than in the posterior part at that stage. We showed that the mechanism leading to a faster reduction of the head width during early elongation is dependent on pix-1, pak-1 and let-502/ROCK and that reduction of the tail width is dependent on let-502/ ROCK but neither on pix-1 nor on pak-1. This suggests that contraction forces controlled by let-502/ROCK are required for the embryo morphogenesis along the antero-posterior axis, while those controlled by the pix-1/pak-1 pathway may mostly be required in the anterior part of the embryo. This also suggests that either the pix-1/pak-1 pathway induces contractions mostly in the anterior part of the embryo, or this pathway is redundant with another pathway (in addition to let-502/mel-11) specifically involved in the control of contractions in the posterior part of the embryo.
Interestingly, we found that animal expressing PIX-1::GFP at a reduced level in the dorsal posterior cells present an elongation rate similar to wt. Higher expression of PIX-1::GFP in the dorsalposterior cells is also shown to be detrimental for the elongation rate of embryos during early elongation. This suggests that the function of pix-1 is required at a low level in the dorsal-posterior hypodermal cells to ensure optimal elongation rate. This supports the hypothesis that the pix-1/pak-1 pathway may induce more contractions in the anterior part of the embryo than in the posterior part. However, this hypothesis will have to be tested through careful measurement of contraction forces induced by the let-502/mel-11 and pix-1/pak-1 pathways in individual sets of hypodermal cells during early elongation.
We also showed that loss of let-502 function suppresses the increased head/tail width ratio observed in pix-1 arrested larvae when compared to wt and that this suppression occurs only in embryos expressing an active mel-11 ( Figure 3D). These data support our observations that let-502; pix-1 arrested larvae display let-502 morphology, while let-502; mel-11; pix-1 larvae display pix-1 morphology (Table 2). To explain these results, we may hypothesize that loss of let-502 may induce a severe reduction of contraction forces in hypodermal cells due to the activation of mel-11. This activated mel-11 may then strongly suppress the contraction forces induced by the pix-1/pak-1 pathway and consequently the difference of contraction forces applied on the head and tail during early elongation. According to this hypothesis, in absence of both let-502 and mel-11, deletion of pix-1 function is associated with the arrest of larvae displaying a similar increase of H/T width ratio than observed in single pix-1(gk416) mutants. These data confirm that MEL-11 function antagonizes both contraction forces induced by LET-502 and those induced by the PIX-1/PAK-1 pathway (Figure 7) [3,14].
In summary, our study demonstrates a function for pix-1 during early elongation within the pix-1/pak-1 pathway in parallel with mel-11/let-502. At that stage, let-502 may drive the contraction forces leading the reduction of the embryo circumference along the antero-posterior axis, while the pix-1/pak-1 pathway may mainly control the contraction forces applied on the anterior part of the embryo.

Phenotyping mutant animals and 4-dimensional microscopy
To score Emb and Lva phenotypes, worms were synchronized by hypochlorite treatment. After synchronisation 10-20 worms were deposited on NGM agar with OP50 as a source of food Worms were allowed to lay eggs at 18uC or 25.5uC for 4 to 5 hours and were washed off the plate with M9 medium. After 24 and 48 hours dead eggs and arrested L1 larvae were counted and observed at high magnification. The stage of embryonic arrest was confirmed in mutant animals using four-dimension microscopy. Embryos dissected from adult hermaphrodites were mounted on 3% agarose pads in M9 buffer and coverslips were sealed with drawing gum (pébéo). Elongation was recorded using 4-dimensional microscopy (3D and time), which recorded a Z-stack every 2 minutes during 10 hours at 23-24uC using a Leica DM5500 microscope equipped with a 63X oil immersion objective upon differential interference contrast illumination (DIC). Images were captured using the Leica LAS AF imaging software. These recording were used to measure the duration of early elongation for at least 20 eggs from 1.2-to the end of early elongationidentified as the moment when body-wall muscles start contracting, the length of the embryos at the end of early elongation, the width of the head (measured at equidistance from the tip of the nose to the pharynx-intestinal valve), the width of the tail (measured at equidistance from the pharynx-intestinal valve to the distal extremity of hyp10) at 1.2-fold and at the end of early elongation in the different mutant animals. These measurements were done using Leica LAS AF6000 imaging analysis tools. The reproducibility of these measurements was tested as detailed in the supplementary Figure 3. Length and head/tail width measurements were also done on mutant arrested larvae and wild-type (wt) L1 arrested by starvation after hypochlorite treatment. The head/ tail ratio was calculated as the ratio of the head width with the tail width in mm per animal. The head (and tail) width reduction ratio was calculated as the ratio of the head width (or tail width) at 1.2fold stage over the head width (or tail width) at the end of early elongation. Statistical significance was calculated using unpaired Student's T-test.

Immunostaining of embryos
For indirect immunofluorescence, embryos were fixed using 3% paraformaldehyde at room temperature for 10 minutes. Following washes with PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 , 1.47 mM KH 2 PO 4 Adjust to a final pH of 7.4.), embryos were incubated 10 minutes in cold methanol, and extensively washed with PBS. Fixed embryos were incubated overnight at 4uC with appropriate dilutions of primary antibodies in culture media (4X eggs salt, 0.5% Hepes 1M, 5% goat serum). Mouse anti-MH-27 antibodies (The Developmental Studies Hybridoma Bank, University of Iowa) were used at 1:10 dilution, rabbit anti-GFP antibodies were used at 1:500 (Invitrogen). After three washes with culture media (4X egg salt, 0.5% HEPES 1M, goat serum 5%), embryos were incubated at room temperature for 1 h with 1:200 dilution of either anti-mouse IgG conjugated to TRITC or anti-rabbit IgG conjugated to FITC (Jackson ImmunoResearch) in 1X-PBS. Embryos were washed three times with culture media and incubated with 100 ng/ml of DAPI for 1 minute at room temperature, washed with water, resuspended in 40 ml of Mowiol (sigma Aldrich) and mounted on slides.

Confocal fluorescence microscopy
The expression pattern of PIX-1::GFP in living animals was observed using a Nikon A1R confocal microscope with 100X oil CFI NA 1.45 Plan Apochromat l objective. All images were captured with a pinhole size of 59.1 mm, with a calibration of 0.12 mm/pixel (radial resolution of 0.20 mm) and a Z-step of 0.15 mm. Images were captures using NIS-element software (Nikon). Deconvolution was done using Autoquant 3X, 3D deconvolution software. Orthogonal views, and fluorescence quantifications were generated using ImageJ software. For fluorescence quantification, individual Z-steps were extracted from Z-stacks recording of elongating embryos. DP, DA, LP, LA, VP, VA hypodermal cells were selected using ImageJ 'polygon selections' tool. Nuclei of cells in these areas were also selected using the same tool. The Raw intensities of selected areas were measured using area and gray value function. Raw Intensities of nuclei were subtracted to raw intensities of selected hypodermal sections. The pixel mean value was then calculated dividing this adjusted raw intensity by the area size selected excluding also size of nuclei. Intensity ratios and raw intensities were calculated and used using the resulting value. Comparison of raw intensities between different animals was done on image captured within a week, to limit the impact of laser fluctuation. This fluctuation was controlled through comparison of intensities obtained for the same sample at different days. Fluorescence intensity was compared on images captured using the same laser power and adjusted photomultiplier tube (PMT). Statistical significance of quantification was assessed using the unpaired student T-test. Spearman correlation coefficients and statistical tests for significance of correlation between two quantitative variables were calculated using spearman cor.test function in R (Bioconductor).  Figure S2 pix-1(ok982) and pak-1(ok448) arrested larvae present severe pharynx pumping defects. 48 hours after egg-laying, pharynx pumping rates were counted on arrested L1 animals and escaper L3 animals moving freely on a bacterial lawn. At least 10 animals per genotype were examined during 15-sec periods. N = 3. ** T-test p (mutant/N2),0.001 (TIF) Figure S3 Establishment of embryo width measurement as a robust metrics to characterize embryonic elongation. A) We tested the robustness and reproducibility of head width measurement of embryos. To do so, head width was measured five times on a given population of wt embryos at 1.2fold stage (n = 12 embryos). Means and standard deviation were calculated and Brown-Forsythe test (using R statistical package) was used to test for homogeneity of variances among the five different groups of measurement. This test revealed no significant variance difference amongst the measurements (F-test p-value.0.5). B) The repeatability and batch effect of our measurements were assessed through measurement of the head width of wt embryos at 1.2-fold stage from 4D-recording done at three different days (n = 12 embryos). Means and standard deviation were calculated and Brown-Forsythe test was used to test for homogeneity of variances among the three different groups of measurements. This test revealed no significant variance difference amongst the measurements (F-test p-value.0.5). Similar results were obtained for tail width measurements and for measurement done at different stages of early elongation (data not shown). These data indicate that the significant differences observed between genotypes using head-width, tail-width and head/tail width ratio measurements are not due to measurement variability and batch effect. (TIF) Figure S4 PIX-1 is homogeneously distributed in the cytoplasm and at the cell periphery of hypodermal cells during early elongation. Confocal microscopy analysis of pix-1(gk416) embryos carrying sajEx1[pix-1p::pix-1::gfp; rol-6]; mcI-s40[lin-26p::ABDvab-10::mCherry + myo-2p::gfp]. PIX-1::GFP is observed in B and E (green in C, F) and VAB-10 ABD ::mCherry in A and D (red in C, F). Embryos are oriented anterior to the left and dorsal up. Enlarged views (lower panels) show areas indicated by white rectangles in upper panels. Apical and basolateral membrane are indicated by arrow and arrowhead, respectively (L, lower panel). Scale bars upper panels: 10 mm; lower panels 5 mm. (TIF) Figure S5 PIX-1::GFP intensity ratio are constant throughout early elongation. Dorsal-posterior/dorsal-anterior (DP/DA), lateral-posterior/lateral-anterior (LP/LA), ventralposterior/ventral-anterior (VP/VA), dorsal-posterior/ventral (DP/V), dorsal-anterior/ventral (DA/V) and dorsal-anterior/ lateral (DA/L) fluorescence intensity ratio were measured as detailed in methods and in Figure 5 in pix-1(gk416); unc-119; sajIs2[lin-26p::pix-1::GFP,unc-119 R ] embryos during early elongation. Bar correspond to the mean and error bars to the standard deviation. * T-test,0.05 (TIF) Figure S6 Only DP/DA ratio inversely correlates with the elongation rate of the embryos during early elongation. Scatter plot representing the relationship between the lateral-posterior/lateral-anterior (LP/LA), ventral-posterior/ ventral-anterior (VP/VA), dorsal-posterior/dorsal-anterior (DP/ DA) and dorsal-anterior/ventral (DA/V) intensity ratio of PIX-1::GFP and the elongation rate in mm/min during early elongation of pix-1(gk416); sajEx1[pix-1p::pix-1::GFP, rol-6] embryos (n = 20). The spearman correlation (R2) between the elongation rate and the PIX-1::GFP intensity ratio are indicated, as well as the p-values rejecting the null hypothesis being that the two values are not significantly correlated. Similar results were obtained from two independent transgenic lines. (TIF) Figure S7 High expression of PIX-1::GFP in dorsalposterior hypodermis is detrimental for elongation rate of embryos. Dorsal-posterior/dorsal-anterior (DP/DA), lateral/ ventral (L/V), dorsal-anterior/lateral (DA/L) and dorsal-posterior/lateral (DP/L) fluorescence intensity ratio were measured as detailed in methods and in Figure 5 in pix-1(gk416); unc-119; sajIs2[lin-26p::pix-1::GFP;unc-119 R ] embryos elongating at a wt-rate (elongation$288 nm/min) or elongating slower (elongation,288 nm/min) during early elongation. Bar correspond to the mean and error bars to the standard deviation. ** T-test pvalue,0.01. (TIF)