P21 activated kinases (PAKs) are major downstream effectors of rac-related small GTPases that regulate various cellular processes. We have identified the new PAK gene max-2 in a screen for mutants disrupted in UNC-6/netrin-mediated commissural axon guidance. There are three Caenorhabditis elegans PAKs. We find that each C. elegansPAK represents a distinct group previously identified in other species. Here we examine their roles in the postembryonic migration of the P cell neuroblasts and the axon guidance of the ventral cord commissural motoneurons(VCCMNs). We find that the two PAKs, max-2 and pak-1, are redundantly required for P cell migration and function with UNC-73/Trio and the rac GTPases (CED-10 and MIG-2). During axon guidance of the VCCMNs, PAK-1 also acts with the rac GTPases, CED-10 and MIG-2, and is completely redundant with MAX-2. Interestingly, we find that unlike MAX-2 activity during P cell migration, for motoneuron axon guidance max-2 is also required in parallel to this PAK-1 pathway, independent of rac GTPase signaling. Finally,we provide evidence that MAX-2 functions downstream of the UNC-6/netrin receptor UNC-5 during axon repulsion and is an integral part of its signaling.

Neuronal connectivity is achieved by guidance of axonal growth cones to their targets through various environmental cues(Dickson, 2002; Tessier-Lavigne and Goodman,1996). Cellular interpretation of these cues requires the presence of proper cell surface receptors and the execution of complex signaling pathways inside the growth cones. In Caenorhabditis elegans, analysis of the factors required for axon guidance has led to the discovery of the highly conserved UNC-6/netrin signaling pathway(Hedgecock et al., 1990; Ishii et al., 1992; Serafini et al., 1994). UNC-6/netrin functions as both an attractant and a repellent. The receptor UNC-5 is required for repulsion from UNC-6/netrin, while the UNC-40 receptor is required for both repulsion and attraction(Colavita and Culotti,1998).

During development, the ventral cord commissural motoneurons (VCCMNs),which include DA, DB, DD, VD and AS neurons, are born on the ventral midline of the animal. They express both UNC-5 and UNC-40, and their commissural axons are thereby repelled by UNC-6/netrin located on the ventral side of the animal(Wadsworth, 2002). These commissures migrate circumferentially around the animal along the dorsoventral axis. The putative chemoattractant UNC-129, a TGFβ homolog, is thought to direct these axons into the dorsal cord(Yu and Bargmann, 2001). They then extend along the dorsal cord, innervating their target muscle cells. Collectively these VCCMNs function to control forward and reverse locomotion of the animals. Among them, the cholinergic (DA, DB and AS) motoneurons activate the dorsal muscles of the animals, while the GABAergic (DD and VD)motoneurons coordinate firing of the ventral and dorsal muscles(White et al., 1986). While the attractive and repulsive molecules that direct these neurons are fairly well understood, much less is known about the subcellular signaling events that interpret these cues. We have taken advantage of this system and used forward genetics to identify new components in this conserved signaling pathway (Huang et al., 2002). We have identified max-2, a C. elegans p21-activated kinase (PAK)that is necessary for the dorsal guidance of the VCCMN axons.

PAKs are a group of highly conserved signaling molecules that control cytoskeletal dynamics. Previous studies have demonstrated that PAKs are effectors of the rac/cdc42 subfamily of rho type GTPases. PAKs were initially discovered through their ability to bind activated (GTP-bound) forms of these small GTPases (Manser et al.,1994). Each PAK consists of an N-terminal GTPase-binding domain and a C-terminal serine/threonine kinase domain. The nominal activation pathway for conventional PAKs has been described, in which an activated rac/cdc42 molecule binds a PAK dimer, allowing relief from its self-inhibition and subsequent activation of its kinase domain(Bokoch, 2003). This activated PAK then goes on to target other downstream effectors, which include multiple regulators of the cytoskeleton, such as myosin light chain kinase and LIM kinase. In yeast, PAKs are required for invasive and vegetative growth(Hofmann et al., 2004). In Drosophila, Pak1 is required for photoreceptor axon guidance(Hing et al., 1999). In humans, a deficiency of PAK3 results in non-syndromic mental retardation,attributed to a loss of dendritic complexity(Allen et al., 1998).

Here, we describe and characterize the three C. elegans PAKs. We find that two C. elegans PAKs function redundantly in P cell migrations with UNC-73/Trio and the rac GTPases, but are differentially required for UNC-6/netrin-mediated commissural motor axon guidance. We provide evidence that while PAK-1 is completely redundant with MAX-2 and functions with the rac GTPases during axon guidance, the newly identified max-2is additionally required in a rac-independent pathway. Finally, we show that MAX-2 functions downstream of the receptor UNC-5 during UNC-6/netrin-mediated axon repulsion.

C. elegans and culture methods

Worm cultures were maintained as described(Brenner, 1974). For RNAi experiments, dsRNA was microinjected into the intestine or gonad of young adult animals as described (Fire et al.,1998). The following commercially available RNAi clones (Ahringer Library Clones unless otherwise specified) were used in this study: I 2J23, II 8F19, X 3E04, C09B8.7 (Open Biosystems).

The following mutant alleles were used in these studies: LG I: unc-40(e1430), dpy-5(e61), unc-73(gm33), unc-73(e936), unc-73(rh40); LG II: max-2(cy2), max-2(nv162), rol-1(e91); LG III: pag-1(ls2); LG IV: unc-5(e53), dpy-20(e1282), unc-129(ev557), eri-1(mg366); LG V: pak-2(ok332), unc-34(e566), him-5(e1490), max-1(cy2); evIs41[Pmec-7::lacZ,Pmec-7::unc-5]; LG X: oxIs12[Punc-47::GFP,lin-15(+)], pak-1(ok448), unc-6(ev400), unc-6(ju152), mig-2(mu28).

Genetic mapping

max-2(cy2) was mapped to LG II by standard SNP methods(Wicks et al., 2001). Mapping analysis placed max-2(cy2) between SNPs Y38F1A(41,752) and W09E7(4,859). dsRNA of open reading frames (ORFs) from within this genomic region were tested for their ability to elicit a phenotype(Fraser et al., 2000). Only dsRNA from the predicted ORF Y38F1A.10 was found to cause axon guidance defects. Sequencing of max-2(cy2) genomic DNA found a G-A transition at Y38F1A (69,599 bp).

Deletion library screening

The max-2(nv162) allele was isolated by PCR screening of a C. elegans UV/TMP mutagenized deletion library using the following primers: nv162.f1, CTCTCTCTCAACCCGGCAGGAAG; nv162.f1',CCGGCAGGAAGACTATATGACTC; nv162.r1, TGTCGTCGCTCTCTCACGCACCT; and nv162.r1',CACAAAGAGGGAAGAAGATCCTC. Genotyping max-2(nv162) was subsequently performed by using the inner (nv162.f1' and nv162.r1') primers along with primer nv162.r2 CCTTCTTCTGATCGGCAAGACTG, which lies inside the max-2(nv162) deletion.

Molecular biology

Cloning of DNA and generation of transgenes were accomplished by standard techniques (Hobert, 2002; Shevchuk et al., 2004). cDNAs were generated by RT-PCR from total N2 RNA samples. One max-2 cDNA was generated by ligating the 3′ end of YK651h1 (gift from Yuji Kohara)onto the 5′ end of one of our cDNAs to generate a full-length max-2 cDNA free of sequence changes. This was then cloned into a pBluescript SK(+) vector to give the clone pHJ101. To generate the max-2 rescue construct, a full-length max-2 cDNA was fused to the promoter region of max-2 with PCR fusion techniques. Briefly PCR was used to generate a partial max-2 cDNA (lacking the first two exons) out of pHJ101 with primers PMM.f1 (CAGAAGTTCAGCGGACTGCCGCAA) and PBK.r1(CAGCTATGACCATGATTACGCCA). The following primers ex.f(GTTGCGTGCATATTCCGAACAGT) and ex.r (TTGCGGCAGTCCGCTGAACTTCTG) (complimentary to PMM.f1) were used to generate a max-2 promoter region through the beginning of the third exon from genomic DNA. The resulting two PCR fragments were used as a template to generate the max-2 rescue construct by PCR with primers mx2r.f' (GCTTGTCGAGTGTGCAATTTGTC) and mx2r.r'(AATGGTTCAATTCTACAGTTTATAGATT). This PCR fusion product was cloned into a TOPO XL vector to yield pHJ102.

Scoring of animals

DD and VD commissural defects were scored as previously described(Huang et al., 2002). Briefly,the number of wild-type or defective commissures from a sample population was counted and percent defects were determined by summing all the defective commissures and dividing this by the number of commissures expected in the same number of wild-type animals (17 commissures/animal). A commissure was deemed to fail to reach the dorsal cord only if no part of the commissure could be observed to connect to the dorsal cord. For constitutively active rac experiments, the number of animals scored (n) was the combination of at least two independently generated lines. All lines assayed showed the same trends. Constitutively activated rac [rac(GF)] lines were crossed into the mutant [pak-1(ok448) or max-2(cy2)] background and rac(GF);wild type and rac(GF);mutant animals were isolated. The rac(GF);wild type and rac(GF);mutant animals were scored as first cousins. For rac(GF) RNAi experiments transgenic lines were generated in an eri-1(mg366)IV;oxIs12X background. Transgenic animals were picked to a single plate and approximately half were injected with dsRNA. The progeny of the injected and non-injected animals were then scored. The rac(GF) constructs were generously donated by Erik Lundquist,and each contains a rac gene with the equivalent to the canonical G12V mutation under the control of the unc-115 promoter(Struckhoff and Lundquist,2003).

P cell migration defects were scored by counting the number of laterally displaced VD neurons. VD neurons were identified by the expression of GFP in an oxIs12[Punc-47::GFP] background and by their gross morphology. VD neurons were counted as being laterally displaced if the cells were more than two cell widths from the ventral cord or had no visible connection to the ventral cord.

For the suppression of ectopic UNC-5 experiments, animals were fixed and stained as described (Colavita and Culotti,1998). Touch receptor cells were scored as reaching the dorsal cord if any part of the neuron reached the dorsal cord. Only the anteriormost touch receptor cells (the ALMs and the AVM) were scored. Only animals in which all of the anterior touch receptor cells could be seen to extend long axons(or were directed into the DC) were scored.

max-2 is required for commissural axon guidance of the VCCMNs and represents a gene with homology to p21-activated kinases

In a forward genetic screen for C. elegans mutants in motor axon guidance we have identified max-2(cy2)(Huang et al., 2002). max-2(cy2) animals have defects in the guidance of the axons of the VCCMNs (Fig. 1). All max-2 animals show some percent of misguided commissures, yet have only subtle defects in locomotion and appear superficially wild type. Newly hatched max-2 animals exhibit defects in the embryonically connected DA, DB and DD motoneurons (Fig. 1D,H), as well as in the postembryonic VD motoneurons(Fig. 1F). In max-2mutant animals, many motoneuron commissures turn either anteriorly or posteriorly before reaching the dorsal cord(Fig. 1D,F,H). Those commissures that fail to reach the dorsal cord often possess the length necessary to reach the dorsal cord had they been guided in the proper direction. This is consistent with max-2 being required for the dorsal guidance of these motoneurons and not solely required for axonal outgrowth.

We mapped the max-2(cy2) mutation to an approximately 200 kb region on chromosome II (Fig. 2A). We next used a reverse genetics candidate screen to identify the max-2 gene. dsRNA from only one of the genes in this area elicited an axon guidance phenotype. This predicted ORF Y38F1A.10 has homology to PAKs (Hofmann et al.,2004). A partial cDNA clone (a gift from Yuji Kohara) was found to span this predicted ORF and an adjacent one (F18A11.4) indicating that they are, in fact, a single gene (Fig. 2A boxed region). Comparison with a syntenic region of C. briggsae and knowledge of published PAK sequences allowed us to predict a probable ORF consisting of parts of these two annotated ORFs. Isolation and analysis of cDNAs confirmed this and demonstrated the genomic organization of the max-2 gene (Fig. 2B). A minigene consisting of a 5 kb upstream element fused to a max-2 cDNA (Fig. 2B)was found to rescue the max-2(cy2) defect(Fig. 2C).

max-2(cy2) and max-2(nv162) are likely null alleles

The max-2(cy2) allele has a missense mutation in the ATP-binding region of the kinase domain. The resulting amino acid sequence change is expected to convert a highly conserved glycine to a glutamate(Fig. 2B, Fig. 4A). This glycine is the third in a group of glycines (gly-X-gly-X-X-gly) that are in the conserved region of kinase subdomain I (Hanks et al., 1988). This change probably impairs the kinase activity by altering the structure of the ATP-binding pocket. To further examine the function of max-2, we analyzed the phenotype of max-2(RNAi) animals and isolated an additional mutant allele max-2(nv162). We found that RNAi of max-2 caused a weaker but similar phenotype than either of our two mutants(Fig. 2D).

Fig. 1.

max-2 is required for ventral cord commissural motoneuron guidance. (A,B) A schematic of the DD and VD cell bodies(solid dots) and commissural axons in wild-type (A) and max-2(cy2) (B) animals. In each panel, a cross-section is on the left and a lateral view of the entire animal is on the right. (C-H)Confocal images of representative animals: anterior is to the left and dorsal is up. (C,D) First larval stage (L1) animals before the formation of the VD neurons. In wild-type animals (C), the dorsal commissures of the DD neuron all reach and enter the dorsal cord, but the DD commissures in max-2(cy2) animals (D) often fail to reach the dorsal cord.(E,F) L4 animals, after the migration of the VD commissures. In wild-type animals, all commissures reach the dorsal cord (E). In max-2(cy2) animals, many commissures fail to reach the dorsal cord (F). All animals in C-F are in the oxIs12[Punc-47::GFP]X background to visualize the DD and VD motoneurons. (G,H) max-2 animals have defects in the DA and DB motoneurons. Shown here are examples of commissural axons in wild-type (G)and max-2(cy2) animals (H) at early L4 stage. All animals in G,H are in the evIs82[Punc-129::GFP] background to visualize the DA and DB neurons. Scale bar: 10 μm.

Fig. 1.

max-2 is required for ventral cord commissural motoneuron guidance. (A,B) A schematic of the DD and VD cell bodies(solid dots) and commissural axons in wild-type (A) and max-2(cy2) (B) animals. In each panel, a cross-section is on the left and a lateral view of the entire animal is on the right. (C-H)Confocal images of representative animals: anterior is to the left and dorsal is up. (C,D) First larval stage (L1) animals before the formation of the VD neurons. In wild-type animals (C), the dorsal commissures of the DD neuron all reach and enter the dorsal cord, but the DD commissures in max-2(cy2) animals (D) often fail to reach the dorsal cord.(E,F) L4 animals, after the migration of the VD commissures. In wild-type animals, all commissures reach the dorsal cord (E). In max-2(cy2) animals, many commissures fail to reach the dorsal cord (F). All animals in C-F are in the oxIs12[Punc-47::GFP]X background to visualize the DD and VD motoneurons. (G,H) max-2 animals have defects in the DA and DB motoneurons. Shown here are examples of commissural axons in wild-type (G)and max-2(cy2) animals (H) at early L4 stage. All animals in G,H are in the evIs82[Punc-129::GFP] background to visualize the DA and DB neurons. Scale bar: 10 μm.

The allele max-2(nv162) was found by screening a deletion library for N-terminal deletions in the max-2 gene. The max-2(nv162) allele contains a deletion that removes all of exons 1-4, including the start codon (Fig. 2B). The first in frame methionine would lead to an interrupting stop codon before any wild-type protein could be made. The next in frame methionine does not occur until the middle of the kinase domain. max-2(nv162) is therefore almost certainly a null.

To test this and to compare the defects of these two mutant alleles, we scored the severity of each mutant defect in the dorsal guidance of the DD and VD neurons. We found that the phenotypes of the two mutants were of similar nature and severity (Fig. 2D). Next we scored the severity of max-2(cy2)/max-2(nv162) animals and found that they were not significantly different from animals homozygous for either allele (data not shown). Taken together, these results indicate that both max-2(cy2) and max-2(nv162) are likely genetic nulls.

max-2 expression is required in neurons for axon guidance

To examine the expression of max-2, we generated promoter-GFP fusions (Chalfie et al., 1994; Fire et al., 1990). GFP expression is nearly ubiquitous in the early embryo. At early comma stage expression becomes intensely focused in the anterior of the embryo(Fig. 3A). Strong expression is observed from the pharynx and some unidentified head neurons beginning around the 1.5-fold stage. After hatching, GFP expression is present in the ALM and PLM neurons (Fig. 3D). Beginning in late L1 stages expression comes on in the PVD neurons and some time later in the AVM (Fig. 3D,E). Despite using the same upstream DNA elements utilized for our rescue constructs with these promoter-GFP constructs, we did not observe significant expression of GFP in VCCMNs.

To test whether max-2 functions cell autonomously to guide motoneurons we expressed max-2 cDNAs under the control of tissue-specific promoters. There are 26 GABAergic neurons, 19 of these (six DD and 13 VD) send commissures dorsally and have defects in max-2mutants. We used promoters from the characterized unc-25 and unc-47 genes (Jin et al.,1999; McIntire et al.,1997) to drive expression of a max-2 cDNA specifically in the GABAergic neurons of max-2(cy2) animals. Both these promoter-cDNA fusions could rescue the defects of max-2(cy2)mutants (Fig. 3F). We conclude that MAX-2 functions cell autonomously in these neurons.

The C. elegans genome contains three p21-activated kinases

A database search of the C. elegans genome demonstrates the existence of three C. elegans PAKs, including max-2. Caenorhabditis elegans pak-1 has previously been cloned by degenerate PCR and has been shown to co-localize with CED-10 and CDC-42 at late embryonic stages (Chen et al., 1996). Few data exist on the other PAK (C45B11.1), which we find is SL1 transspliced and will be hereafter referred to as C. elegans pak-2. Presented in Fig. 4A is a kinase domain alignment of the three C.elegans PAKs and some closely related PAKs from other species. The overall structural differences of the C. elegans PAKs are also compared in Fig. 4B. These differences are striking with respect to the site and number of the putative SH3-binding motifs. An analysis of the PAK sequences from worms, flies and humans demonstrates that C. elegans PAK-1 is most closely related to group I/A PAKs and C. elegans PAK-2 is most closely related to group II/B PAKs (Fig. 4C). The kinase domain of MAX-2 is most closely related to group I/A PAKs(Fig. 4A), while the N-terminal region is more divergent than classic group I/A PAKs(Zhao and Manser, 2005). In this respect, MAX-2 is most similar to Drosophila DmPAK3.

Fig. 2.

max-2 is a single gene that includes the two predicted ORFs F18A11.4 and Y38F1A.10. (A) A diagram of a genomic region on the distal right arm of chromosome II (LGII) that contains the max-2gene. Cosmid and YAC coverage in this region are diagrammed with black bars. The boxed region of the schematic is enlarged and shows the region that contains the two predicted ORFs F18A11.4 and Y38F1A.10 (boxed region is not to scale). (B) A diagram of the max-2 genomic organization(upper) and the minigene construct (lower) used for the rescue experiments in(C). The boxed area indicates the region deleted in the max-2(nv162) allele, and the asterisk indicates the site of the point mutation identified in the max-2(cy2) allele. (C) max-2 mutants are rescued by injecting the rescue construct depicted in B. The graph shows the percentage of the DD and VD motor commissures that fail to reach the dorsal cord (DC). Two independent transgenic lines expressing the minigene construct znex130-131[Pmax-2::max-2(cDNA)] are shown here;non-transgenic siblings that lack the transgene are also shown for znex131. (D) max-2(cy2) and max-2(nv162) show similar severity in the defects of dorsal guidance of commissural axons. Injecting double-stranded RNA (dsRNA) of the max-2 gene causes less severe axon guidance defects that are otherwise similar to the max-2 mutants. The numbers (n) of animals used for each experiment are shown; bars represent the standard error(C,D).

Fig. 2.

max-2 is a single gene that includes the two predicted ORFs F18A11.4 and Y38F1A.10. (A) A diagram of a genomic region on the distal right arm of chromosome II (LGII) that contains the max-2gene. Cosmid and YAC coverage in this region are diagrammed with black bars. The boxed region of the schematic is enlarged and shows the region that contains the two predicted ORFs F18A11.4 and Y38F1A.10 (boxed region is not to scale). (B) A diagram of the max-2 genomic organization(upper) and the minigene construct (lower) used for the rescue experiments in(C). The boxed area indicates the region deleted in the max-2(nv162) allele, and the asterisk indicates the site of the point mutation identified in the max-2(cy2) allele. (C) max-2 mutants are rescued by injecting the rescue construct depicted in B. The graph shows the percentage of the DD and VD motor commissures that fail to reach the dorsal cord (DC). Two independent transgenic lines expressing the minigene construct znex130-131[Pmax-2::max-2(cDNA)] are shown here;non-transgenic siblings that lack the transgene are also shown for znex131. (D) max-2(cy2) and max-2(nv162) show similar severity in the defects of dorsal guidance of commissural axons. Injecting double-stranded RNA (dsRNA) of the max-2 gene causes less severe axon guidance defects that are otherwise similar to the max-2 mutants. The numbers (n) of animals used for each experiment are shown; bars represent the standard error(C,D).

MAX-2 and PAK-1 function with partial redundancy to guide the VCCMNs

The motor axon defect of max-2(cy2) is the first mutant phenotype reported for any of the C. elegans PAKs. To begin to understand the in vivo functions of the other C. elegans PAKs, we created promoter-GFP fusions to examine their expression (see Fig. S1 in the supplementary material). We confirmed previous studies that pak-1 is expressed in the VCCMNs (Iino and Yamamoto, 1998). pak-1 was also expressed in the migrating distal tip cell (DTC), in the developing uterus and later in the vulval muscle cells (see Fig. S1 in the supplementary material, and data not shown). pak-2 did not appear to be neuronally expressed (see Fig. S1 in the supplementary material). To further address whether all three C. elegans PAKs are involved in UNC-6/netrin axon guidance, we analyzed pak-1 and pak-2 mutants for neuronal defects similar to those found in max-2 mutants. pak-1(ok448) and pak-2(ok332) mutant alleles were obtained from the C. elegans Gene Knock Out Consortium. Both these mutants are expected nulls. The pak-1(ok448) allele has a deletion that removes the majority of the kinase coding region and results in a frame shift that is expected to cause a premature stop codon. The pak-2(ok332)allele has a deletion that removes the start codon. Both these PAK mutants appear superficially wild type and we have found them to be wild type for DD and VD axon guidance (Fig. 5E).

The double PAK mutants max-2(cy2);pak-1(ok448) were severely defective for DD and VD axon guidance (Fig. 5D,E). These animals were also uncoordinated, defective in egg laying and in DTC migrations, and exhibited ventral enclosure defects. RNAi was used to confirm these results independently (data not shown). max-2(cy2);pak-2(ok332) animals were superficially wild type and did not have DD and VD defects significantly worse than max-2(cy2) animals(Fig. 5E). pak-1(ok448);pak-2(ok332) animals had defects in embryogenesis and exhibited L1 lethality, but escapers were wild type for VCCMN guidance and appeared relatively coordinated. We conclude that PAK-2 does not play a role in the axon guidance of the VCCMNs, while max-2 and pak-1 function with some redundancy to control axon guidance.

PAK mutants are phenotypically similar to unc-73/Trio and rac mutants

PAKs are activated by racs. Racs are rho family GTPases, implicated in controlling cell migrations and axon guidance by acting as molecular switches that can relay and amplify cellular signals(Hall, 1998; Lundquist, 2003; Luo et al., 1997). Racs are activated by guanine exchange factors (GEFs). Of the rac activators, which coordinate rac activity during axon guidance, UNC-73/Trio is the best known(Steven et al., 1998). In C. elegans, unc-73/Trio mutants have major defects in cell migrations, axon outgrowth and axon guidance. The three C. elegansracs (MIG-2, CED-10 and RAC-2) are activated during multiple developmental processes by the GEF UNC-73/Trio(Lundquist et al., 2001; Wu et al., 2002; Zipkin et al., 1997). Consistent with C. elegans PAKs being used in this UNC-73/Trio-rac pathway, we found that max-2(cy2);pak-1(ok448) double mutants were phenotypically very similar to unc-73 mutants and also ced-10(n1993);mig-2(mu28) double rac mutant animals with respect to their mutant phenotypes in DD and VD axon guidance and also in P cell migrations(Fig. 5 and Table 1). We chose to examine in detail the roles that the C. elegans PAKs play during these two different cellular processes. We first present data on PAK function in P cell migrations and then examine PAK function in the UNC-6/netrin VCCMN axon repulsion pathway.

Table 1.

MAX-2 and PAK-1 function with redundancy in the UNC-73-rac pathway to control P cell migrations

GenotypePercentage of animals with laterally displaced VDs*Average number of laterally displaced VDs/animaln
wt 80 
max-2(cy2)II 64 
pak-1(ok448)X 84 
max-2(cy2)II; pak-1(ok448)X 50 0.71 133 
unc-73(rh40)I 27 0.38 52 
max-2(cy2)II; unc-73(rh40)I 35 0.42 122 
pak-1(ok448)X; unc-73(rh40)I 31 0.45 96 
unc-73(gm33)I 100 4.8 46 
max-2(cy2)II; unc-73(gm33)I 100 4.8 58 
ced-10(n1993)IV 100 
mig-2(mu28)X 80 
ced-10(n1993)IV; mig-2(mu28)X 100 3.4 59 
max-2(cy2)II; ced-10(n1993)IV 100 
max-2(cy2)II; mig-2(mu28)X 6.3 0.11 95 
pak-1(ok448)X; ced-10(n1993)IV 5.5 0.05 55 
pak-1(ok448); mig-2(mu28)X 1.5 0.01 68 
GenotypePercentage of animals with laterally displaced VDs*Average number of laterally displaced VDs/animaln
wt 80 
max-2(cy2)II 64 
pak-1(ok448)X 84 
max-2(cy2)II; pak-1(ok448)X 50 0.71 133 
unc-73(rh40)I 27 0.38 52 
max-2(cy2)II; unc-73(rh40)I 35 0.42 122 
pak-1(ok448)X; unc-73(rh40)I 31 0.45 96 
unc-73(gm33)I 100 4.8 46 
max-2(cy2)II; unc-73(gm33)I 100 4.8 58 
ced-10(n1993)IV 100 
mig-2(mu28)X 80 
ced-10(n1993)IV; mig-2(mu28)X 100 3.4 59 
max-2(cy2)II; ced-10(n1993)IV 100 
max-2(cy2)II; mig-2(mu28)X 6.3 0.11 95 
pak-1(ok448)X; ced-10(n1993)IV 5.5 0.05 55 
pak-1(ok448); mig-2(mu28)X 1.5 0.01 68 
*

The number of animals with a laterally displaced VD neuron divided by the total number of animals scored.

Animals were scored for the number of laterally displaced VD neurons and the sum of these numbers was divided by the total number of animals.

Fig. 3.

max-2 expression is dynamic and is required cell autonomously for dorsal commissure guidance. (A-E) The expression pattern of max-2 from znex135[Pmax-2::GFP] transgenic animals. Promoter element (5′-3′) contains the 5 kb region from 5557 to 262 on cosmid F18A11. (A-C) At the comma and later embryonic stages, GFP expression is mainly in the anterior region of the embryo and along the ventral cord. (D) After hatching, GFP is expressed in the pharynx and several head and tail neurons. Expression of GFP in the ALM neurons (arrow) and the postembryonic PVD neuron can also be seen. (E) A young adult animal, showing GFP expression in the postembryonic AVM neuron (arrow and left inset) and the elaborate dendritic connections of a mature PVD neuron (right inset).(F) A max-2 cDNA expressed in the DD and VD motoneurons cell-autonomously rescues the max-2(cy2) defects in these neurons. For the Punc-25::max-2 rescue, the data presented are the combined data from four independently generated transgenic lines. A Punc-47::max-2 rescue line (znex133) and its non transgenic siblings (znex133 sibs) are also shown here. The unc-25promoter used for this experiment contained 2 kb of the 5′ region through the start codon. The unc-47 promoter used for this experiment contained 1.7 kb of the region 5′ to the start codon through the 50th codon. The numbers (n) of animals used for each experiment are shown;bars represent the standard error in F. Asterisks indicate a P-value of less than 0.001 (Student's t-test). In all pictures, anterior is to the left and dorsal is up. Scale bar: 10 μm.

Fig. 3.

max-2 expression is dynamic and is required cell autonomously for dorsal commissure guidance. (A-E) The expression pattern of max-2 from znex135[Pmax-2::GFP] transgenic animals. Promoter element (5′-3′) contains the 5 kb region from 5557 to 262 on cosmid F18A11. (A-C) At the comma and later embryonic stages, GFP expression is mainly in the anterior region of the embryo and along the ventral cord. (D) After hatching, GFP is expressed in the pharynx and several head and tail neurons. Expression of GFP in the ALM neurons (arrow) and the postembryonic PVD neuron can also be seen. (E) A young adult animal, showing GFP expression in the postembryonic AVM neuron (arrow and left inset) and the elaborate dendritic connections of a mature PVD neuron (right inset).(F) A max-2 cDNA expressed in the DD and VD motoneurons cell-autonomously rescues the max-2(cy2) defects in these neurons. For the Punc-25::max-2 rescue, the data presented are the combined data from four independently generated transgenic lines. A Punc-47::max-2 rescue line (znex133) and its non transgenic siblings (znex133 sibs) are also shown here. The unc-25promoter used for this experiment contained 2 kb of the 5′ region through the start codon. The unc-47 promoter used for this experiment contained 1.7 kb of the region 5′ to the start codon through the 50th codon. The numbers (n) of animals used for each experiment are shown;bars represent the standard error in F. Asterisks indicate a P-value of less than 0.001 (Student's t-test). In all pictures, anterior is to the left and dorsal is up. Scale bar: 10 μm.

The C. elegans PAKs function redundantly in the UNC-73/Trio-RAC pathway to guide P cell migrations

Immediately after hatching, the 12 P cells migrate from their lateral positions down into the ventral cord and subsequently divide, generating the P cell lineage. A failure in this migration can lead to the ectopic placement of cells in the P cell lineage. Among the P cell descendents are the VD neurons(Sulston and Horvitz, 1977). We used ectopically placed VD neurons as an indicator of defects in P cell migrations. Two racs (CED-10 and MIG-2) are implicated in acting with UNC-73 during the migration of the P cells(Spencer et al., 2001; Wu et al., 2002). Both unc-73 and double rac mutants have defects in P cell migrations. We found that, while no single PAK mutant exhibited defects, the double PAK mutant max-2(cy2);pak-1(ok448) had defects in P cell migrations similar to those of the unc-73 mutants and the ced-10(n1993);mig-2(mu28) double rac mutants (Fig. 5B-D,F and Table 1). As with the racs, we found that the PAKs, max-2 and pak-1, were completely redundant with each other for this process.

We found that either max-2(cy2) or pak-1(ok448) only slightly increased the P cell migration defect of a weak unc-73 allele [unc-73(rh40)]. Interestingly, a double mutant of max-2(cy2) and a strong allele of unc-73 [unc-73(gm33)] did not increase the average number of ectopic P cells per animal found in the unc-73(gm33) allele alone(Table 1). It is important to note that null alleles of unc-73 are lethal, which precludes a definite conclusion that the PAKs act entirely with UNC-73 during P cell migrations. However, our observation that a loss of max-2 did not enhance the strong unc-73 allele, along with the weak enhancement observed in PAK;unc-73(rh40) double mutants, indicates that the two C. elegans PAKs (MAX-2 and PAK-1) act with UNC-73 in guiding P cell migrations. Finally, we found that any combination of rac;PAKdouble mutants either did not enhance, or only weakly enhanced, the defects of the single mutants (Table 1). This indicates that the PAKs function linearly with the racs to guide this process.

The C. elegans PAKs function differentially to control UNC-6/netrin-mediated VCCMN axon guidance

Of all the C. elegans rac and PAK single mutants, only max-2 animals had significant defects in UNC-6/netrin-mediated VCCMN axon guidance (Fig. 6A). To determine if these molecules act together or in parallel, we systematically analyzed double mutant combinations. Double mutants of either of the two rac genes (ced-10 or mig-2) with pak-1 did not cause defects in VCCMN guidance greater than any single mutant(Fig. 6A). This is consistent with PAK-1 and the racs acting linearly in this VCCMN guidance pathway. Surprisingly, we found that max-2 double mutants with either of these two rac genes greatly enhanced the VCCMN defects of max-2(Fig. 6A). This indicates that MAX-2 functions in parallel to the racs and PAK-1 during VCCMN guidance.

Fig. 4.

Sequence analysis of the three C. elegans p21-activated kinases. (A) An amino acid sequence comparison of the kinase domains from the three C. elegans PAKs and representative PAKs from other species. The asterisk indicates the site of the max-2(cy2) mutation. (B) A diagram of the structural organization of the three C. elegans PAKs. The red boxes indicate the P21-binding domain, the blue box indicates the site of a clearly conserved autoinhibitory domain, the green ovals indicate the kinase domain and the vertical lines indicate the sites of putative SH3-interacting (PXXP) domains.(C) An unrooted phylogenetic tree of the entire amino acid sequence of all the known PAKs from C. elegans (Ce), D. melanogaster(Dm) and H. sapiens (Hs). The group I/A and group II/B PAKs are boxed with blue and green, respectively. This tree was generated with the ClustalW alignment tool from the European Bioinformatics Institute(www.ebi.ac.uk/clustalw/). Sequences for max-2 and pak-2 were deposited in GenBank as DQ523832 and DQ523831, respectively.

Fig. 4.

Sequence analysis of the three C. elegans p21-activated kinases. (A) An amino acid sequence comparison of the kinase domains from the three C. elegans PAKs and representative PAKs from other species. The asterisk indicates the site of the max-2(cy2) mutation. (B) A diagram of the structural organization of the three C. elegans PAKs. The red boxes indicate the P21-binding domain, the blue box indicates the site of a clearly conserved autoinhibitory domain, the green ovals indicate the kinase domain and the vertical lines indicate the sites of putative SH3-interacting (PXXP) domains.(C) An unrooted phylogenetic tree of the entire amino acid sequence of all the known PAKs from C. elegans (Ce), D. melanogaster(Dm) and H. sapiens (Hs). The group I/A and group II/B PAKs are boxed with blue and green, respectively. This tree was generated with the ClustalW alignment tool from the European Bioinformatics Institute(www.ebi.ac.uk/clustalw/). Sequences for max-2 and pak-2 were deposited in GenBank as DQ523832 and DQ523831, respectively.

It has previously been reported that the racs function with (and probably downstream of) UNC-73 to guide the DD and VD motoneuron axons(Wu et al., 2002). We found that double pak mutants were phenotypically similar to unc-73mutants, yet single PAK mutants enhanced non-lethal unc-73 alleles for DD and VD axon guidance defects (Fig. 5B,D, Fig. 6A and data not shown). This has previously been reported for the racs as well(Wu et al., 2002).

To further address whether this parallel rac-independent pathway exists, we tested whether a loss of function in PAK gene activity could suppress the defects resulting from constitutively active racs. We hypothesized that the defects caused by a constitutively active molecule would be decreased when one of its downstream effectors was lost, or conversely the loss of parallel activity would cause an increase in these defects. We generated transgenic animals that express constitutively active C. elegans racs under a pan-neuronal promoter (Struckhoff and Lundquist, 2003). We found that loss of function in pak-1(ok448) significantly suppressed the axon guidance defects caused by the constitutively active rac, MIG-2(Fig. 6B). We also found consistent but nonsignificant suppression of the defects caused by constitutively active CED-10 when we used RNAi to knock out pak-1function. Generally, RNAi yields a weaker phenotype than null mutants,particularly in neurons. For this reason we cannot conclude that PAK-1 does not function downstream of CED-10. However, our results are consistent with PAK-1 acting downstream of at least some of the rac GTPases.

Unlike pak-1, a loss of function in max-2 considerably enhanced the defects caused by constitutively active racs in all cases(Fig. 6C). This indicates that max-2 functions in a rac-independent pathway during commissural axon guidance. It is important to note that this does not exclude MAX-2 from functioning downstream of the racs as well. In our assay, the loss of MAX-2 parallel activity may have masked any suppression [of the rac(GF) induced defects] that was occurring. Collectively,our results indicate that PAK-1 functions with the racs, downstream of MIG-2(at least) and that MAX-2 functions in parallel (at least partly) to this UNC-73-rac-PAK-1 pathway to guide the VCCMN axons.

Fig. 5.

max-2;pak-1 double mutants are phenotypically similar to unc-73 and double rac mutants. (A-D) Representative morphology of the DD and VD neurons in unc-73, rac and PAK mutants. Arrows point to displaced VD neurons and open arrowheads point to abnormal commissures. (A) A wild-type animal. (B) An unc-73(e936)I animal. (C) A ced-10(n1993)IV;mig-2(mu28)Xanimal. (D) A max-2(cy2)II;pak-1(ok448)Xanimal. (E) DD and VD commissural guidance defects in PAK mutants.(F) P cell migration defects in PAK, rac and unc-73 mutants. The numbers (n) of animals used for each experiment are shown; bars represent the standard error in E,F. All animals are in the oxIs12[Punc-47::GFP]X. Scale bar: 10 μm.

Fig. 5.

max-2;pak-1 double mutants are phenotypically similar to unc-73 and double rac mutants. (A-D) Representative morphology of the DD and VD neurons in unc-73, rac and PAK mutants. Arrows point to displaced VD neurons and open arrowheads point to abnormal commissures. (A) A wild-type animal. (B) An unc-73(e936)I animal. (C) A ced-10(n1993)IV;mig-2(mu28)Xanimal. (D) A max-2(cy2)II;pak-1(ok448)Xanimal. (E) DD and VD commissural guidance defects in PAK mutants.(F) P cell migration defects in PAK, rac and unc-73 mutants. The numbers (n) of animals used for each experiment are shown; bars represent the standard error in E,F. All animals are in the oxIs12[Punc-47::GFP]X. Scale bar: 10 μm.

MAX-2 likely acts with UNC-6/netrin during DD and VD commissural axon guidance

As max-2 mutants exhibited defects specifically in an axon guidance process controlled by UNC-6/netrin, we analyzed max-2genetic interactions with genes that are implicated in this pathway. Mutations in the dorsally expressed orphan ligand unc-129(Colavita et al., 1998)exhibited an additive enhancement in a max-2 background. However, max-1 and unc-34 mutants were dramatically enhanced in a max-2 background (Fig. 6D). Double mutants of unc-40/DCC (the UNC-6/netrin receptor) and max-2 exhibited synergistic genetic interactions(Fig. 6D). This indicates that the two (unc-40 and max-2) are required in parallel but interconnected signaling pathways. Null mutants of unc-6/netrin and unc-5 have VCCMN axon guidance defects that are absolutely severe. Because of this, unc-6 and unc-5 double mutant combinations with max-2 were less than informative. As an alternative we tested max-2 dosage interactions with these genes. max-2 mutants synergized with a hypomorphic allele of unc-6 to nearly complete severity [from 20% failure in unc-6(ju152) to 75% in max-2(nv162);unc-6(ju152)]. Additionally,we found that max-2 mutants were dramatically enhanced when they had only a single copy of unc-6 or unc-5. A much weaker enhancement was observed when only a single copy of unc-40 was present (Fig. 6D). Collectively, our data indicate that MAX-2 functions with UNC-6/netrin and its repulsion receptor UNC-5 in parallel to UNC-40 to mediate commissural motor axon guidance.

The axon guidance defects caused by ectopic expression of the UNC-6/netrin receptor UNC-5 are partially suppressed by a loss of max-2

To test directly whether max-2 functions downstream of UNC-5 and UNC-6/netrin, we determined whether a loss of max-2 function could suppress the axon guidance effects of ectopically expressed UNC-5. Previously,Hamelin and colleagues (Hamelin et al.,1993) demonstrated that if UNC-5 is ectopically expressed in the touch receptor cells their axonal processes are re-routed away from UNC-6/netrin. This effect absolutely requires UNC-6/netrin and demonstrates that the machinery necessary for UNC-5 signaling is present in these cells. This is the basis for a classic screen that has identified many genes required for UNC-5 repulsion (Colavita and Culotti,1998). Our max-2 expression studies indicate that MAX-2 is present in several of these touch receptor cells. In particular, high expression was observed in the anteriormost touch receptor cells (the ALMs and the AVM) (Fig. 3D,E). We reasoned that MAX-2 might be part of the signaling machinery usurped by the ectopically expressed UNC-5, to misguide these axonal projections.

max-2(cy2) was crossed into the ectopic UNC-5 expressing strain NW798 (a gift from Joseph Culotti). max-2(cy2)mutants suppressed the guidance effects caused by the ectopic UNC-5(Fig. 7A-C). We also found that in the max-2 mutant background the touch cell receptor axons that reach the dorsal cord take a less direct route, often stretching anteriorly for approximately half their length before turning toward the dorsal cord. The lack of absolute suppression observed from a loss of max-2 was quite similar to the defects observed in max-2 mutants. In both cases an apparently relatively small portion of UNC-5 signaling was lost. Our ectopic UNC-5 suppression results are therefore highly correlative to the situation found with the guidance of the VCCMNs, where a loss of max-2 is far less severe than a loss of unc-5. This confirms previous observations that there are multiple signaling pathways downstream of UNC-5, in UNC-6/netrin-induced axonal repulsion, and implicates MAX-2 as being integral to at least one of these cascades.

We have identified MAX-2 as a p21-activated kinase. There are three PAK genes in the C. elegans genome. Sequence analysis indicates that each C. elegans PAK is representative of a distinct group of PAKs previously identified in other species. Importantly, none of the C. elegans PAK mutants is embryonic lethal. With these mutants, we can therefore perform in vivo genetic analysis on phenotypes to study specific PAK functions. Our genetic and expression pattern studies demonstrate that pak-1 and max-2 are key players in regulating cell migration and axon guidance: they function redundantly as rac effectors in P neuroblast migration and differentially for VCCMN axon guidance. We conclude that these C. elegans PAKs function both redundantly and independently to regulate multiple cellular processes.

Each of the three C. elegans PAKs represents a distinct group of PAKs

In mammals, the PAK family is divided into two subgroups (I/A and II/B)based on structural differences. Members from both PAK groups are implicated in neuronal morphogenesis. The human PAK3, a group I/A PAK, is mutated in an X-linked mental retardation (Allen et al.,1998). Knockout mice deficient for PAK4, a group II/B PAK, show clear neuronal migration defects (Qu et al., 2003).

Drosophila melanogaster has three PAKs. The Drosophilagroup I/A PAK (Pak1; previously know as Dpak1) has been shown to be required along with NCK/Dock and UNC-73/Trio, for the guidance of photoreceptor axons(Hing et al., 1999; Newsome et al., 2000) and has also been shown to function downstream of the chemorepellent slit in midline axon guidance (Fan et al.,2003). The Drosophila group II/B PAK (Mbt), is believed to regulate neuronal morphogenesis, rather than axon guidance(Melzig et al., 1998). A third Drosophila PAK (Pak3) has sequence similarity to both group I/A and group II/B PAKs (Mentzel and Raabe,2005). In vivo functional data on Drosophila Pak3 have yet to be reported.

We found that there are three PAKs in the C. elegans genome. Humans and mice contain six PAKs. Interestingly, we have found that the C. elegans PAKs consist of single representative members from the two mammalian PAK groups (I/A and II/B) and also a third more divergent member of the PAK family. This phenomenon has previously been reported of D. melanogaster (Mentzel and Raabe,2005). The differences in PAK representatives between mammals and invertebrates may imply that an expanded repertoire of PAKs in both groups can replace the functions of more divergent PAKs. The expansion probably suggests a specialization of PAK activity for members within each group. Studying and comparing PAK functions will lead to a better understanding of the signaling specificity of different PAK member activities in vivo.

A model for PAK activity in P cell migration and VCCMN axon guidance

The C. elegans PAKs function redundantly in the UNC-73/Trio-rac pathway to control the migration of P cells. Mutations in the individual PAK genes did not lead to defects. However, double mutants of pak-1 and max-2 did have defects in P cell migration. PAK mutants only weakly enhanced unc-73(rh40). Mutations in max-2 did not enhance the defects of a strong unc-73/Trio allele. These data indicate that these PAKs function with UNC-73/Trio (see Fig. S2 in the supplementary material). Loss of either of two of the rac genes(mig-2 or ced-10) in either a max-2 or pak-1 mutant background either weakly enhanced or had no defect greater than any of the single mutants. These data indicate that the PAKs function with the racs during this process. By examining the average number of ectopic cells per animal, we noted that in the double PAK mutant the severity of the P cell migration defects were less than that of the double racmutant, which in turn was less severe than the strong unc-73(gm33) allele (Fig. 5F and Table 1). This is consistent with a model in which a fraction of the P cell migration activity of UNC-73 is controlled by the racs, and a fraction of this rac activity is controlled by the PAKs (see Fig. S2 in the supplementary material). This supports previous findings that UNC-73/Trio also stimulates rho activity (in addition to racs), which in turn activates the rho kinase(LET-502), facilitating the migration of P cells(Spencer et al., 2001).

Fig. 6.

pak-1 and max-2 are differentially required for commissural axon guidance. Genetic interaction studies on the DD and VD commissural guidance defects in PAK mutants and other mutants in genes that are known to play roles in the unc-6/netrin commissural guidance pathway. (A) Axon guidance defects of rac mutants are greatly enhanced by a loss of max-2 but not by a loss of pak-1. (B)The defects caused by constitutively active rac(GF)s can be suppressed by the loss of pak-1. The asterisk denotes a P-value less than 0.001(Student's t test). (C) The defects caused by constitutively active rac(GF)s are enhanced by a loss of function in the max-2 gene. The max-2 data was also independently confirmed with dsRNA as in the ced-10(GF) result. (D) max-2(cy2) enhances the defects of mutants from genes that are known to play roles in the unc-6/netrin commissural guidance pathway. The numbers (n) of animals used for each experiment are shown; bars represent the standard error.

Fig. 6.

pak-1 and max-2 are differentially required for commissural axon guidance. Genetic interaction studies on the DD and VD commissural guidance defects in PAK mutants and other mutants in genes that are known to play roles in the unc-6/netrin commissural guidance pathway. (A) Axon guidance defects of rac mutants are greatly enhanced by a loss of max-2 but not by a loss of pak-1. (B)The defects caused by constitutively active rac(GF)s can be suppressed by the loss of pak-1. The asterisk denotes a P-value less than 0.001(Student's t test). (C) The defects caused by constitutively active rac(GF)s are enhanced by a loss of function in the max-2 gene. The max-2 data was also independently confirmed with dsRNA as in the ced-10(GF) result. (D) max-2(cy2) enhances the defects of mutants from genes that are known to play roles in the unc-6/netrin commissural guidance pathway. The numbers (n) of animals used for each experiment are shown; bars represent the standard error.

The C. elegans PAKs act both rac-dependently and rac-independently in the guidance of VCCMN axons. In VCCMN axon guidance, as with the migrations of the P cells, the rac genes are reported to function downstream of the GEF UNC-73/Trio (Wu et al.,2002). Indeed, the phenotypes of double rac mutants and unc-73/Trio are remarkably similar(Fig. 5B,C). We found that, as has been reported for the racs, PAK mutants enhanced unc-73 mutant defects for DD and VD axon guidance. This may indicate that there is another rac-GEF acting in parallel to UNC-73/Trio during this process (see Fig. S2 in the supplementary material). Although the C. elegans PAKs(pak-1 and max-2) appeared to function in a typical manner with this GEF-rac pathway to guide the P cell migrations, the two did not function solely in this GEF-rac pathway for the dorsal guidance of commissural axons. While PAK-1 acts with the racs in guiding these axons, we found that MAX-2 has a unique role outside this pathway. Because max-2 can act redundantly with pak-1, we conclude that max-2 is probably also acting in part downstream of the racs in VCCMN guidance (see Fig. S2 in the supplementary material). Interestingly, max-2 mutants enhanced unc-73/Trio axon guidance defects much more than pak-1mutants did (Fig. 6A). This is consistent with the rac-independent activity of MAX-2 acting in parallel to the GEF, UNC-73/Trio.

It is interesting to note that several lines of evidence indicate a subtle preference by the racs for the PAKs. We found that for P cell migrations double mutants of ced-10 and pak-1 were weakly enhanced,while double mutants of mig-2 and pak-1 were not. We found that max-2 had a reversed relationship(Table 1). Additionally, we found that a loss of pak-1 significantly suppressed the defects caused by a constitutively active MIG-2 but did so to a lesser extent for CED-10. Conversely, a loss of max-2 greatly enhanced the defects of constitutively active MIG-2 but did so to a lesser extent for CED-10(Fig. 6). Finally, double mutants of mig-2 and max-2 were enhanced much more than double mutants of ced-10 and max-2(Fig. 6A). A subtle opposite relationship could also be seen for pak-1(Fig. 6A). These data collectively argue that the preferred PAK for CED-10 is MAX-2, while the preferred PAK for MIG-2 is PAK-1. However, the fact that double mutants of max-2;mig-2 or pak-1;ced-10 do not equal a double PAK mutant indicate that either rac can use either PAK(Table 1).

Fig. 7.

MAX-2 is required for the axon guidance effects induced by ectopic expression of UNC-5 in the anterior touch cell receptor neurons.(A-C) DIC images of adult animals after fixation and staining for the presence of lacZ. In all images, dorsal is up and anterior is to the left. Black arrowheads mark the dorsal cord, white arrowheads mark touch receptor cell bodies and black arrows mark touch receptor cell axons. For clarity not all cell bodies and axons present are marked in each image. Animals[evIs41[Pmec7::lacZ,Pmec7::unc-5];pag-1(ls2)]that ectopically express UNC-5 and lacZ under the control of the mec-7 promoter have the anterior touch receptor cell axons (ALMs and AVM)guided to the dorsal cord (A). (B,C) In a max-2 background, the guidance effects from ectopic expression of UNC-5 are partially suppressed. The image in B is two superimposed images from different focal planes of the same animal. Occasionally we observed ventral migration of the AVM axon in a max-2 background, as shown in C. We never observed this in the ectopic UNC-5 strain alone. (D) A graphical representation of the quantification of ALM and AVM guidance in the two populations. The numbers(n) of animals used for each experiment are shown; bars represent the standard error. The asterisk denotes a P-value less than 0.001(Student's t-test). Scale bar: 10 μm.

Fig. 7.

MAX-2 is required for the axon guidance effects induced by ectopic expression of UNC-5 in the anterior touch cell receptor neurons.(A-C) DIC images of adult animals after fixation and staining for the presence of lacZ. In all images, dorsal is up and anterior is to the left. Black arrowheads mark the dorsal cord, white arrowheads mark touch receptor cell bodies and black arrows mark touch receptor cell axons. For clarity not all cell bodies and axons present are marked in each image. Animals[evIs41[Pmec7::lacZ,Pmec7::unc-5];pag-1(ls2)]that ectopically express UNC-5 and lacZ under the control of the mec-7 promoter have the anterior touch receptor cell axons (ALMs and AVM)guided to the dorsal cord (A). (B,C) In a max-2 background, the guidance effects from ectopic expression of UNC-5 are partially suppressed. The image in B is two superimposed images from different focal planes of the same animal. Occasionally we observed ventral migration of the AVM axon in a max-2 background, as shown in C. We never observed this in the ectopic UNC-5 strain alone. (D) A graphical representation of the quantification of ALM and AVM guidance in the two populations. The numbers(n) of animals used for each experiment are shown; bars represent the standard error. The asterisk denotes a P-value less than 0.001(Student's t-test). Scale bar: 10 μm.

MAX-2 is involved in UNC-6/netrin-mediated commissural axon repulsion

The dorsal guidance of the VCCMN axon is controlled by the chemorepellent UNC-6/netrin and its receptor UNC-5. Null mutants of the two genes showed a complete absence of the commissurals. Although the absolute severity of null mutants precludes double mutant analysis, several lines of evidence indicate that MAX-2 acts downstream of UNC-6/netrin and UNC-5. First, the neurons that showed defects in max-2 mutants appeared to be disrupted specifically in their dorsal commissural migrations. In max-2 mutants the neuronal processes in the ventral cord of the VCCMNs were normal. Second, max-2 mutants synergized with other mutants involved in this pathway. MAX-1 has been shown to mediate UNC-5 signaling in the VCCMN guidance. UNC-40 is another UNC-6/netrin receptor that is partially required in this guidance. Double mutants of max-2 with max-1 or unc-40 showed profound commissural guidance defects, suggesting that max-2 acts in parallel to both max-1 and unc-40. Interestingly, max-2 was also found to enhance a weak unc-6 allele,suggesting that MAX-2 does not function downstream of UNC-6 with a simple linear relationship. Finally, max-2 was required for axon guidance effects of ectopically expressed UNC-5. The axons of touch neurons were guided dorsally when UNC-5 was ectopically expressed in these cells. This phenotype absolutely requires UNC-6/netrin and is suppressed in max-2mutants. Taken together, our studies demonstrate that MAX-2 acts downstream of UNC-6/netrin and UNC-5 to regulate the dorsal guidance of VCCMN axons.

We thank Marc Tessier-Lavigne, Joseph Culotti, Eric Lundquist, Eric Struckhoff, Yishi Jin, Andy Fire, Yuji Kohara, Alan Coulson, Antonio Colavita,Scott Kennedy, Gary Ruvkun, Robert Steven, the C. elegans Gene Knockout Consortium and the Caenorhabditis Genetics Center for materials. We thank Judy Nguyen for technical help. We thank Dan Starr,Lesilee Rose, Jin Lee, Damien O'Halloran and members of the UC Davis C. elegans Group for comments and advice. This work was partially supported by an NIH Training Grant to M.S.L. and by grants from the Klingenstein Fund and the Sloan Foundation to H.-J.C.

Allen, K. M., Gleeson, J. G., Bagrodia, S., Partington, M. W.,MacMillan, J. C., Cerione, R. A., Mulley, J. C. and Walsh, C. A.(
1998
). PAK3 mutation in nonsyndromic X-linked mental retardation.
Nat. Genet.
20
,
25
-30.
Bokoch, G. M. (
2003
). Biology of the p21-activated kinases.
Annu. Rev. Biochem.
72
,
743
-781.
Brenner, S. (
1974
). The genetics of Caenorhabditis elegans.
Genetics
77
,
71
-94.
Chalfie, M., Tu, Y., Euskirchen, G., Ward, W. W. and Prasher, D. C. (
1994
). Green fluorescent protein as a marker for gene expression.
Science
263
,
802
-805.
Chen, W., Chen, S., Yap, S. F. and Lim, L.(
1996
). The Caenorhabditis elegans p21-activated kinase (CePAK)colocalizes with CeRac1 and CDC42Ce at hypodermal cell boundaries during embryo elongation.
J. Biol. Chem.
271
,
26362
-26368.
Colavita, A. and Culotti, J. G. (
1998
). Suppressors of ectopic UNC-5 growth cone steering identify eight genes involved in axon guidance in Caenorhabditis elegans.
Dev. Biol.
194
,
72
-85.
Colavita, A., Krishna, S., Zheng, H., Padgett, R. W. and Culotti, J. G. (
1998
). Pioneer axon guidance by UNC-129, a C. elegans TGF-beta.
Science
281
,
706
-709.
Dickson, B. J. (
2002
). Molecular mechanisms of axon guidance.
Science
298
,
1959
-1964.
Fan, X., Labrador, J. P., Hing, H. and Bashaw, G. J.(
2003
). Slit stimulation recruits Dock and Pak to the roundabout receptor and increases Rac activity to regulate axon repulsion at the CNS midline.
Neuron
40
,
113
-127.
Fire, A., Harrison, S. W. and Dixon, D. (
1990
). A modular set of lacZ fusion vectors for studying gene expression in Caenorhabditis elegans.
Gene
93
,
189
-198.
Fire, A., Xu, S., Montgomery, M. K., Kostas, S. A., Driver, S. E. and Mello, C. C. (
1998
). Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans.
Nature
391
,
806
-811.
Fraser, A. G., Kamath, R. S., Zipperlen, P., Martinez-Campos,M., Sohrmann, M. and Ahringer, J. (
2000
). Functional genomic analysis of C. elegans chromosome I by systematic RNA interference.
Nature
408
,
325
-330.
Hall, A. (
1998
). Rho GTPases and the actin cytoskeleton.
Science
279
,
509
-514.
Hamelin, M., Zhou, Y., Su, M. W., Scott, I. M. and Culotti, J. G. (
1993
). Expression of the UNC-5 guidance receptor in the touch neurons of C. elegans steers their axons dorsally.
Nature
364
,
327
-330.
Hanks, S. K., Quinn, A. M. and Hunter, T.(
1988
). The protein kinase family: conserved features and deduced phylogeny of the catalytic domains.
Science
241
,
42
-52.
Hedgecock, E. M., Culotti, J. G. and Hall, D. H.(
1990
). The unc-5, unc-6, and unc-40 genes guide circumferential migrations of pioneer axons and mesodermal cells on the epidermis in C. elegans.
Neuron
4
,
61
-85.
Hing, H., Xiao, J., Harden, N., Lim, L. and Zipursky, S. L.(
1999
). Pak functions downstream of Dock to regulate photoreceptor axon guidance in Drosophila.
Cell
97
,
853
-863.
Hobert, O. (
2002
). PCR fusion-based approach to create reporter gene constructs for expression analysis in transgenic C. elegans.
Biotechniques
32
,
728
-730.
Hofmann, C., Shepelev, M. and Chernoff, J.(
2004
). The genetics of Pak.
J. Cell Sci.
117
,
4343
-4354.
Huang, X., Cheng, H.-J., Tessier-Lavigne, M. and Jin, Y.(
2002
). MAX-1, a novel PH/MyTH4/FERM domain cytoplasmic protein implicated in netrin-mediated axon repulsion.
Neuron
34
,
563
-576.
Iino, Y. and Yamamoto, M. (
1998
). Expression pattern of the C. elegans P21-activated protein kinase, CePAK.
Biochem. Biophys. Res. Commun
.
245
,
177
-184.
Ishii, N., Wadsworth, W. G., Stern, B. D., Culotti, J. G. and Hedgecock, E. M. (
1992
). UNC-6, a laminin-related protein,guides cell and pioneer axon migrations in C. elegans.
Neuron
9
,
873
-881.
Jin, Y., Jorgensen, E., Hartwieg, E. and Horvitz, H. R.(
1999
). The Caenorhabditis elegans gene unc-25 encodes glutamic acid decarboxylase and is required for synaptic transmission but not synaptic development.
J. Neurosci.
19
,
539
-548.
Lundquist, E. A. (
2003
). Rac proteins and the control of axon development.
Curr. Opin. Neurobiol
.
13
,
384
-390.
Lundquist, E. A., Reddien, P. W., Hartwieg, E., Horvitz, H. R. and Bargmann, C. I. (
2001
). Three C. elegans Rac proteins and several alternative Rac regulators control axon guidance, cell migration and apoptotic cell phagocytosis.
Development
128
,
4475
-4488.
Luo, L., Jan, L. Y. and Jan, Y. N. (
1997
). Rho family GTP-binding proteins in growth cone signalling.
Curr. Opin. Neurobiol
.
7
,
81
-86.
Manser, E., Leung, T., Salihuddin, H., Zhao, Z. S. and Lim,L. (
1994
). A brain serine/threonine protein kinase activated by Cdc42 and Rac1.
Nature
367
,
40
-46.
McIntire, S. L., Reimer, R. J., Schuske, K., Edwards, R. H. and Jorgensen, E. M. (
1997
). Identification and characterization of the vesicular GABA transporter.
Nature
389
,
870
-876.
Melzig, J., Rein, K. H., Schafer, U., Pfister, H., Jackle, H.,Heisenberg, M. and Raabe, T. (
1998
). A protein related to p21-activated kinase (PAK) that is involved in neurogenesis in the Drosophila adult central nervous system.
Curr. Biol.
8
,
1223
-1226.
Mentzel, B. and Raabe, T. (
2005
). Phylogenetic and structural analysis of the Drosophila melanogaster p21-activated kinase DmPAK3.
Gene
349
,
25
-33.
Newsome, T. P., Schmidt, S., Dietzl, G., Keleman, K., Asling,B., Debant, A. and Dickson, B. J. (
2000
). Trio combines with dock to regulate Pak activity during photoreceptor axon pathfinding in Drosophila.
Cell
101
,
283
-294.
Qu, J., Li, X., Novitch, B. G., Zheng, Y., Kohn, M., Xie, J. M.,Kozinn, S., Bronson, R., Beg, A. A. and Minden, A. (
2003
). PAK4 kinase is essential for embryonic viability and for proper neuronal development.
Mol. Cell. Biol.
23
,
7122
-7133.
Serafini, T., Kennedy, T. E., Galko, M. J., Mirzayan, C.,Jessell, T. M. and Tessier-Lavigne, M. (
1994
). The netrins define a family of axon outgrowthpromoting proteins homologous to C. elegans UNC-6.
Cell
78
,
409
-424.
Shevchuk, N. A., Bryksin, A. V., Nusinovich, Y. A., Cabello, F. C., Sutherland, M. and Ladisch, S. (
2004
). Construction of long DNA molecules using long PCR-based fusion of several fragments simultaneously.
Nucleic Acids Res.
32
,
e19
.
Spencer, A. G., Orita, S., Malone, C. J. and Han, M.(
2001
). A RHO GTPase-mediated pathway is required during P cell migration in Caenorhabditis elegans.
Proc. Natl. Acad. Sci. USA
98
,
13132
-13137.
Steven, R., Kubiseski, T. J., Zheng, H., Kulkarni, S.,Mancillas, J., Ruiz Morales, A., Hogue, C. W., Pawson, T. and Culotti, J.(
1998
). UNC-73 activates the Rac GTPase and is required for cell and growth cone migrations in C. elegans.
Cell
92
,
785
-795.
Struckhoff, E. C. and Lundquist, E. A. (
2003
). The actin-binding protein UNC-115 is an effector of Rac signaling during axon pathfinding in C. elegans.
Development
130
,
693
-704.
Sulston, J. E. and Horvitz, H. R. (
1977
). Post-embryonic cell lineages of the nematode, Caenorhabditis elegans.
Dev. Biol.
56
,
110
-156.
Tessier-Lavigne, M. and Goodman, C. S. (
1996
). The molecular biology of axon guidance.
Science
274
,
1123
-1133.
Wadsworth, W. G. (
2002
). Moving around in a worm: netrin UNC-6 and circumferential axon guidance in C. elegans.
Trends Neurosci
.
25
,
423
-429.
White, J., Southgate, E., Thomson, J. and Brenner, S.(
1986
). The structure of the nervous system of the nematode Caenorhabditis elegans.
Philos. Trans. R. Soc. Lond. Biol. Sci
.
314
,
1
-340.
Wicks, S. R., Yeh, R. T., Gish, W. R., Waterston, R. H. and Plasterk, R. H. (
2001
). Rapid gene mapping in Caenorhabditis elegans using a high density polymorphism map.
Nat. Genet.
28
,
160
-164.
Wu, Y. C., Cheng, T. W., Lee, M. C. and Weng, N. Y.(
2002
). Distinct rac activation pathways control Caenorhabditis elegans cell migration and axon outgrowth.
Dev. Biol.
250
,
145
-155.
Yu, T. W. and Bargmann, C. I. (
2001
). Dynamic regulation of axon guidance.
Nat. Neurosci. Suppl.
4
,
1169
-1176.
Zhao, Z. S. and Manser, E. (
2005
). PAK and other Rho-associated kinases - effectors with surprisingly diverse mechanisms of regulation.
Biochem. J.
386
,
201
-214.
Zipkin, I. D., Kindt, R. M. and Kenyon, C. J.(
1997
). Role of a new Rho family member in cell migration and axon guidance in C. elegans.
Cell
90
,
883
-894.

Supplementary information