Many genes that affect axon pathfinding and cell migration have been identified. Mechanisms by which these genes and the molecules they encode interact with one another in pathways and networks to control developmental events are unclear. Rac GTPases, the cytoskeletal signaling molecule Enabled, and NIK kinase have all been implicated in regulating axon pathfinding and cell migration. Here we present evidence that, in Caenorhabditis elegans, three Rac GTPases, CED-10, RAC-2, and MIG-2, define three redundant pathways that each control axon pathfinding, and that the NIK kinase MIG-15 acts in each Rac pathway. Furthermore, we show that the Enabled molecule UNC-34 defines a fourth partially redundant pathway that acts in parallel to Rac/MIG-15 signaling in axon pathfinding. Enabled and the three Racs also act redundantly to mediate AQR and PQR neuronal cell migration. The Racs and UNC-34 Ena might all control the formation of actin-based protrusive structures (lamellipodia and filopodia) that mediate growth cone outgrowth and cell migration. MIG-15 does not act with the three Racs in execution of cell migration. Rather, MIG-15 affects direction of PQR neuronal migration, similar to UNC-40 and DPY-19, which control initial Q cell polarity, and Wnt signaling, which acts later to control Q cell-directed migration. MIG-2 Rac, which acts with CED-10 Rac, RAC-2 Rac, and UNC-34 Ena in axon pathfinding and cell migration, also acts with MIG-15 in PQR directional migration.

---

IN the developing nervous system, nascent neurons must form axons, thin extensions of plasma membrane that make precise connections with targets in the nervous system. Axons are guided to their targets by the growth cone, a dynamic motile structure that detects and responds to extracellular guidance cues. These cues are detected by transmembrane receptor proteins on the surface of the growth cone and translated into changes in the dynamics and organization of the growth cone actin cytoskeleton, which underlies growth cone motility and guidance (TESSIER-LAVIGNE and GOODMAN 1996; DICKSON 2002; GALLO and LETOURNEAU 2004). In addition to axon extension, neurons must often migrate from their birthplaces to their final positions in the organism. For example, neurons born in the germinal layers of the mammalian cerebrum migrate to distal layers in the cerebral cortex (HATTEN 2002; KRIEGSTEIN and NOCTOR 2004). The process of cell migration is similar to growth cone outgrowth both morphologically and molecularly: both growth cones and leading edges of migrating cells consist of dynamic, actin-based plasma membrane protrusions called lamellipodia and filopodia; and many of the molecules that control growth cone outgrowth also control cell migration (e.g., guidance receptors and cytoskeletal signaling molecules). Many guidance receptors and their ligands have been identified (HUBER et al. 2003), and recent studies have uncovered many of the cytoplasmic signaling molecules that link guidance receptors to the actin cytoskeleton. The challenge remains to understand how pathfinding molecules work together in pathways and networks to mediate growth cone motility and cell migration. Here we demonstrate interactions between three previously identified signaling molecules that affect axon pathfinding: Rac small GTPases, the NIK kinase MIG-15, and the cytoskeletal signaling molecule UNC-34 Enabled.

Rac GTPases of the Rho subfamily regulate cell morphology and actin organization in many systems (HALL 1998). Genetic studies in Drosophila and Caenorhabditis elegans have pinpointed Rac GTPases as key regulators of growth cone outgrowth during axon pathfinding and cell migration (LUO 2000; DICKSON 2001; LUNDQUIST 2003). In C. elegans, three Rac-like GTPases CED-10 (REDDIEN and HORVITZ 2000), RAC-2, and MIG-2 (ZIPKIN et al. 1997) act redundantly in axon pathfinding (no single Rac mutation strongly affects axon pathfinding, whereas pairwise double mutants have severe defects in axon pathfinding) (LUNDQUIST et al. 2001). Here we show that two ced-10 null alleles have little effect on axon pathfinding and neuronal migration on their own, confirming the idea that Rac GTPases have overlapping roles in axon pathfinding.

The mig-15 encodes the C. elegans version of the vertebrate Nck-interacting kinase (NIK) and the Drosophila Misshapen protein (POINAT et al. 2002). MIG-15 contains an N-terminal STE20-like serine/threonine kinase domain and a regulatory C-terminal CNH domain, also found in the Citron protein in which it acts as a Rho GTPase effector domain (MADAULE et al. 2000). Drosophila Misshapen and C. elegans MIG-15 have been shown to affect axon pathfinding and cell migration (RUAN et al. 1999; SU et al. 2000; POINAT et al. 2002). Indeed, a C-terminal region of C. elegans MIG-15 interacts with the cytoplasmic domain of the alpha integrin subunit INA-1, and genetic studies place the two molecules in the same pathway in VD/DD GABAergic motor neuron axon pathfinding (POINAT et al. 2002). RNA mediated interference (RNAi) of ced-10 and mig-2 enhance a weak mig-15 mutation (POINAT et al. 2002), indicating that MIG-15 acts with the Racs in axon pathfinding of the VD/DD axons, although it is unclear if they act in the same or in parallel pathways.

The unc-34 encodes the C. elegans Enabled (Ena) protein (WITHEE et al. 2004). Ena has been shown to control actin organization and cell shape in many systems (KRAUSE et al. 2003), and unc-34 null mutations affect axon pathfinding and neuronal migration in C. elegans (GITAI et al. 2003; WITHEE et al. 2004). Indeed, UNC-34 Ena is required for the ectopic neurite formation induced by an activated UNC-40 guidance receptor in the C. elegans AVM neuron and might act in parallel to CED-10 Rac in this process (GITAI et al. 2003).

Rac GTPases, MIG-15 NIK, and UNC-34 Ena all affect axon pathfinding and cell migration, but it is unclear how these molecules interact to control these processes. Here we present evidence confirming that rac null alleles have little effect on axon pathfinding and neuronal migration and that the three rac genes have overlapping roles in these processes. We present data indicating that MIG-15 NIK and the three Rac proteins act in the same pathway to control axon pathfinding and that UNC-34 Ena defines a parallel, partially overlapping pathway separate from the Rac pathway in the control of axon pathfinding. In addition to their roles in axon pathfinding, the Racs and UNC-34 Ena also have partially overlapping function in the migration of the AQR and PQR neurons, sensory neurons that undergo long-range migration in the larva. Rac and Enabled signaling might affect a common process in axon pathfinding and cell migration, possibly the formation of lamellipodia and filopodia that underlie growth cone outgrowth and cell migration.

While MIG-15 NIK acts with the Racs in axon pathfinding, data presented here indicate that MIG-15 NIK might not affect the ability of AQR and PQR neurons to execute migrations but instead might control direction of PQR neuronal migration. Neuronal cell polarity underlies multiple aspects of neuronal development including the location of axon initiation from the cell body as well as the direction of cell body migration. MIG-2 Rac but not CED-10 or RAC-2 Rac act with MIG-15 in control of PQR direction of migration.

C. elegans genetics, transgenics, and genetic mapping:

C. elegans were cultured by standard techniques (BRENNER 1974; SULSTON and HODGKIN 1988). All experiments were performed at 20° unless otherwise noted. Germline transformation by gonadal micro-injection was performed by standard techniques (MELLO and FIRE 1995). RNAi of rac-2 was performed as described previously (LUNDQUIST et al. 2001). The following mutations and transgenic constructs were used in this work:

• LGI: dpy-5(e61), unc-40(e271)
  
• LGII: juIs73[unc-25::gfp], dpy-10(e128)
  
• LGIII: dpy-17(e124), mab-5(e1239), dpy-19(e1259)
  
• LGIV: ced-10(n1993, n3246, n3417, lq20, and tm597), nT1 IV:V, dpy-13(e184), lqIs3[osm-6::gfp]
  
• LGV: unc-34(e951, gm104, and lq17), dpy-11(e224)
  
• LGX: szT1 X:I, lqIs2[osm-6::gfp], mig-2(mu28 and lq13), unc-115(mn481 and ky275), mig-15(rh148 and rh80)
  

The lq17 mutation was assigned to a linkage group by analysis of segregation against a dpy mutation on each linkage group using the synthetic lethal phenotype of lq17 with ced-10(n1993). lq17 was found to segregate away from dpy-11 on LGV, and showed independent assortment with dpy-5 I, dpy-10 II, dpy-17 III, dpy-13 IV, and the X chromosome (data not shown). Complementation tests with unc-34(e951 and gm104) on LGV showed that lq17 failed to complement unc-34 and that lq17 was an allele of unc-34.

Where possible, double mutants were maintained as homozygous stocks. In some cases, double mutant combinations were lethal or maternal-effect lethal and could not be maintained as homozygous lines. In these cases, one or both of the mutations was maintained in a heterozygous state over a closely linked mutation or a balancer chromosome: ced-10 was balanced by the closely linked and semidominant dpy-13(e184) mutation or by the balancer nT1 qIs51; unc-34 was also balanced by nT1qIs51; and mig-15 was balanced by the balancer szT1. Double homozygotes from nT1 qIs51 strains were detected by lack of green fluorescent protein (GFP) fluorescence in the pharynx (due to the qIs51 integrated transgene on the nT1 balancer), and mig-15 homozygotes were selected from szT1 strains by the uncoordinated and egg-laying-defective phenotype exhibited by mig-15 alleles.

Scoring of PDE and VD/DD axon defects and AQR/PQR cell migration:

Axon pathfinding defects and cell-positioning defects were scored in fourth larval stage (L4) or young pregravid adult animals harboring green fluorescent protein (gfp) transgenes expressed in specific cell types. PDE neurons and axons were visualized in animals harboring an osm-6::gfp transgene (lqIs3 IV or lqIs2 X) that is expressed in all ciliated sensory neurons including PDE, AQR, and PQR (COLLET et al. 1998; STRUCKHOFF and LUNDQUIST 2003). A PDE axon was scored as misguided if it failed to directly extend to the ventral nerve cord (VNC). Those axons that grew at >45° angle to straight ventral were scored as misguided, whether the axon eventually reached the VNC or not. PDE neurons were scored as having ectopic neurites if processes in addition to the normal axon and dendrite were seen emanating from the PDE cell body or axon (axon branching).

VD/DD motor neuron morphology was scored in animals harboring an unc-25::gfp transgene (juIs73 II) (JIN et al. 1999), which is expressed in all GABAergic motor neurons including the VDs and DDs. Normally, VD/DD commissures extend directly from the VNC to the dorsal surface, where they form the dorsal nerve cord. In wild type, occasional minor misrouting of the VD/DD commissures is observed (not more than one or two per animal). The VD/DD commissural axon pathfinding of an animal was scored as defective if more than two commissural axons were misguided, terminated prematurely, or extended extra axon branches. In most cases when VD/DD pathfinding was disrupted, the defects were very severe and affected most commissural processes.

AQR and PQR neuron positioning defects were scored in animals harboring an osm-6::gfp transgene. AQR is a descendant of the QR neuroblast, which is born in the right posterior-lateral region of the animal and initiates anterior migration (SULSTON and HORVITZ 1977). Along its route, QR divides to produce AQR as well as other descendant neurons AVM and SDQR. AQR continues anterior migration into the head of the animal near the posterior pharyngeal bulb. PQR is a descendant of the QL neuroblast, which is born in left posterior-lateral region of the animal and initiates a posterior migration. Along the way, QL divides to produce PQR and other descendant neurons PVM and SDQL. PQR continues posterior migration and finally resides among the left phasmid neurons in the tail of the animal. Defects in the direction of PQR migration were scored as misplacement of PQR to the anterior of the QL birthplace (PQR was considered anteriorly misplaced if it resided anterior to the vulva). The ability of AQR and PQR to execute migration was scored as defective if AQR or PQR failed to reach their final destinations, including complete failure of migration so that the AQR and PQR cell bodies resided at the Q cells' birthplaces (between the vulva and the PDE neurons).

For each of the phenotypic parameters described above, a percentage of defective axons or animals was derived. At least 100 animals of each genotype were scored for each phenotypic class, and a standard error of the proportion was calculated and represented by the error bars in the bar graphs.

Screen for ced-10(n1993) synthetic lethal mutations:

To screen for new mutations that were synthetic lethal with ced-10(n1993), we took advantage of the behavior of extrachromosomal arrays in C. elegans, which can be lost during meiosis such that some gametes do not inherit the array (i.e., animals harboring an array give rise to both array-bearing progeny and progeny that have lost the array). We constructed an extrachromosomal array (called lqEx246) with the full-length ced-10::gfp transgene described previously (LUNDQUIST et al. 2001). This transgene contains the ced-10 promoter driving the expression of full-length CED-10 protein with GFP at the N terminus. This transgene rescued the cell corpse phagocytosis defects and distal tip cell migration defects of ced-10(n1993), ced-10(n3417), and ced-10(tm597) (data not shown), indicating that it produced a functional GFP::CED-10 molecule. Furthermore, lqEx246 rescued the axon pathfinding defects of ced-10(n1993); mig-2(mu28) double mutants and the lethality of ced-10(n3417 and tm597) and mig-2(mu28); ced-10(n1993) animals. lqEx246 did not provide maternal ced-10 activity, as ced-10(n3417), ced-10(tm597), and ced-10(n1993); mig-2(mu28) animals that did not inherit the array were zygotic embryonic lethal rather than maternal-effect lethal (i.e., maternal-effect lethality was converted to embryonic lethality due to lack of maternal gfp::ced-10 activity from lqEx246). Extrachromosomal arrays are often silenced in the germline and thus do not provide maternal gene activity (KELLY et al. 1997).

We used this transgene to screen for new EMS-induced mutations that were synthetic lethal with ced-10. Using standard techniques of ethyl methansulfonate (EMS) mutagenesis (ANDERSON 1995), we mutagenized ced-10(n1993) animals harboring the gfp::ced-10 transgene in an extrachromosomal array (ced-10(n1993); lqEx246 animals), allowed P0 hermaphrodites to self-fertilize, and placed individual F1 ced-10(n1993); lqEx246 animals on single plates. Two F2 ced-10; lqEx246 animals from each F1 were placed on single plates, and the F3 broods were inspected for GFP fluorescence. Normally, we observed both array-bearing (GFP+) and non-array-bearing (GFP-) adult F3 progeny from F2 animals. If, however, the F2 harbored a homozygous mutation that was synthetic lethal with ced-10(n1993), we observed only array-bearing adult F3 and few non–array-bearing adults [i.e., animals that lost lqEx246 were now homozygous for both ced-10(n1993) and the new mutation and were lethal]. We screened the F2 progeny of 3500 F1 animals and found three mutations, lq13, lq17, and lq20, that displayed synthetic defects with ced-10. Through a combination of genetic mapping experiments using the synthetic lethality with ced-10 and complementation tests with known genes, we determined that lq13 was a new allele of mig-2, an expected result of the screen. We found that lq20 was a new mutation in ced-10 such that the ced-10 gene in these animals carried both the n1993 lesion and the lq20 lesion. Finally, we determined that lq17 was a novel allele of the unc-34 gene, which encodes the C. elegans ortholog of Enabled.

Mapping of mig-2(lq13) and ced-10(n1993lq20):

We determined that lq13 was linked to the X linkage group by crossing lq13; ced-10(n1993) hermaphrodites balanced with an extrachromosomal array bearing ced-10(+) DNA to ced-10(n1993) males and observing that all resulting adult male cross-progeny harbored the array and none had lost it (i.e., males that lost the array were hemizygous for lq13 and homozygous for ced-10 and thus depended upon the ced-10 activity provided by the array for their survival). Because mig-2 is on the X linkage group and mig-2 was previously known to be synthetic lethal with ced-10(n1993), we conducted a complementation test with mig-2(mu28) and lq13. Upon outcrossing the lq13; ced-10(n1993); array strain to wild type, we noted that lq13 homozygotes in a ced-10(n1993) heterozygous background [lq13; ced-10(n1993)/+] were strongly uncoordinated (Unc), indicating that loss of one copy of the ced-10 gene enhanced lq13 to cause uncoordination whereas loss of both copies of ced-10 enhanced lq13 to lethality. We crossed hemizygous mig-2(mu28) males to lq13; ced-10(n1993); array animals and found that all non–array-bearing cross-progeny hermaphrodites of the genotype lq13/mig-2(mu28); ced-10(n1993)/+ were strongly Unc, indicating that mig-2(mu28) failed to complement lq13.

Outcrossing and genetic mapping experiments indicated that the lq20 mutation was tightly linked to ced-10(n1993) (e.g., we could not recover the n1993 allele alone from the n1993lq20 strain). This prompted us to directly determine the nucleotide sequence of the ced-10 gene in n1993lq20 animals. lq20 was a new allele in the ced-10 gene itself (see RESULTS), and mutants carried both the n1993 lesion and the lq20 lesion in the ced-10 gene.

Molecular biology:

Recombinant DNA, polymerase chain reaction (PCR), and other molecular biology techniques were performed using standard procedures (SAMBROOK et al. 1989). All primer sequences used in PCR are available upon request. The molecular lesions associated with mig-2(lq13), ced-10(n1993lq20), unc-34(lq17), mig-15(rh148), mig-15(rh80), and mig-15(rh326) were identified as follows: the coding regions of each gene were amplified by PCR from genomic DNA isolated from the mutant strain. The PCR products were directly sequenced using gene-specific primers and standard techniques. The nucleotide sequence of each exon was determined, as well as the flanking intronic sequences (30–50 nucleotides) to identify possible splice site mutations. The mutant sequence was compared to wild-type sequence and to the other mutant sequences to identify potential lesions. The sequencing procedure was repeated on each mutant to independently confirm the nature of the lesion. In each case, the lesion associated with a particular mutation was found in neither wild-type nor distinct mutant strains.

Construction and analysis of mig-2 and ced-10 mutant transgenes:

The point mutations corresponding to ced-10(lq20 and n3246) and mig-2(lq13) were introduced into the previously described ced-10 and mig-2 transgenes, respectively (STRUCKHOFF and LUNDQUIST 2003). The transgene consisted of the wild-type ced-10 and mig-2 coding region controlled by the osm-6 promoter, which drives expression in ciliated sensory neurons including PDE. Point mutations were introduced into the coding regions of these transgenes by site-directed mutagenesis (Quikchange kit, Stratagene, La Jolla, CA). Animals were made transgenic with wild-type and mutant transgenes by gonadal micro-injection of a 1-ng/µl transgene solution with an osm-6::gfp transgene (25 ng/µl) as a transformation marker. Transformation with wild-type ced-10 and mig-2 transgenes at this concentration caused few PDE axon defects, whereas the mutant transgenes at this concentration caused significant PDE defects as well as slow growth and some lethality. In fact, these transgenes were difficult to maintain as stable lines and were gradually lost as they were transmitted, possibly due to selection against lethality and slow growth. Defects were noted in the first three generations after the initial transformation experiment.

Generation of unc-34 cDNAs from unc-34(lq17) mutants:

We used reverse transcription PCR (RT–PCR) to generate unc-34 cDNAs from unc-34(lq17) mutants to assess the effect of lq17 on unc-34 splicing. We isolated total RNA from wild-type and unc-34(lq17) animals by standard techniques (KRAUSE 1995) and subjected 1 µg of total RNA to reverse transcription using AMV reverse transcriptase and a primer specific to the unc-34 3' UTR (see Figure 8B). Using a nested 3' UTR primer and a 5' primer specific to the start of the unc-34 coding region (Figure 8B), we amplified unc-34 RT products by PCR (35 cycles) and separated the PCR products by agarose gel electrophoresis. The 1.4-kb RT–PCR products from wild type and unc-34(lq17) were sequenced directly using gene-specific primers and standard techniques. The RT–PCR was repeated twice to confirm these results.

View larger version (56K): In this window In a new window Download PPT slide   FIGURE 8.— The unc-34(lq17) mutation affects splicing of the unc-34 transcript. (A) A schematic of the 454-residue UNC-34 polypeptide (GenBank accession no. NP_503360), which contains an N-terminal EVH1 domain, a proline-rich domain (PRD), and a C-terminal EVH2 domain. The locations of the gm104 (WITHEE et al. 2004) and lq17 lesions are indicated. (B) The structure of the unc-34 locus. Shaded boxes represent exons. 5' is to the left. The arrows represent the positions of the primers used to generate unc-34 cDNAs by RT–PCR. The sequence of the 5' of the last unc-34 intron is shown above the gene structure. The lq17 mutation is indicated (a G-to-A transition of the first nucleotide of the intron), and the cryptic splice site used in unc-34(lq17) mutants is underlined. (C) The sequence of the unc-34(lq17) cDNA generated by RT–PCR. The nucleotide sequence and conceptual translation are shown. Numbering represents nucleotide/amino acid residue. Block C of the EVH2 domain is underlined. The 13-nucleotide insertion due to use of a cryptic splice site in unc-34(lq17) is shown in lower-case letters. The insertion results in a premature in frame stop codon, which truncates block C of the EVH2 domain.

 

Microscopy and imaging:

Animals were mounted for microscopy in a drop of M9 buffer (SULSTON and HODGKIN 1988) on a 2% agarose pad, both containing 5 mM sodium azide as an anesthetic. A coverslip was placed over the sample, and the slides were analyzed by epifluorescence microscopy for GFP or by differential interference contrast (DIC) microscopy (Leica DMR compound microscope). Images were captured by a Hamamatsu Orca digital camera and analyzed using Openlab software.

Null alleles of ced-10 Rac synergize with mig-2 Rac mutations and rac-2(RNAi) in axon pathfinding:

Previous studies indicated that the null mig-2(mu28) allele and the hypomorphic ced-10(n1993) allele caused few axon defects on their own (LUNDQUIST et al. 2001). However, ced-10(n1993); mig-2(mu28) double mutants displayed severe axon pathfinding defects, indicating that ced-10 and mig-2 have overlapping roles in axon pathfinding (LUNDQUIST et al. 2001; WU et al. 2002).

We characterized a new ced-10 null allele. The ced-10(tm597) allele (S. Mitani, National Bioresource Project for the Experimental Animal "Nematode C. elegans") was a 612-bp deletion of the ced-10 locus that removed all of exon 2 and part of exon 3 (Figure 1A). The previously characterized ced-10(n3417) deletion removed all of exon 2 and 3 (Figure 1A) (LUNDQUIST et al. 2001). Both ced-10(n3417) and ced-10(tm597) are likely to be ced-10 null alleles. ced-10(tm597) animals displayed a phenotype similar to ced-10(n3417) animals: tm597 animals were maternal-effect embryonic lethal with severe defects in morphogenesis, including early defects in gastrulation [a Gex phenotype (SOTO et al. 2002)] and embryonic elongation (data not shown). tm597 homozygotes with wild-type ced-10 contribution were viable and had defects in gonad morphogenesis due to distal tip cell migration errors and defects in phagocytosis of cells undergoing programmed cell death (data not shown).

View larger version (38K): In this window In a new window Download PPT slide   FIGURE 1.— Mutations in the ced-10 and mig-2 loci. (A) A diagram of the ced-10 gene. Left to right is 5'–3'. Shaded boxes represent exons. The extents of two ced-10 deletions are shown. n3417 removes bases 34554 to 35530 and tm597 removes bases 34880–35491 of the ced-10 region relative to cosmid C09G12 (GenBank accession no. AF038608). (B and C) Nucleotide sequences of the mig-2 and ced-10 coding regions (GenBank accession nos. CAB01691 and AAC25821, respectively). The amino acid sequence is shown below the nucleotide sequence. Numbering represents nucleotide sequence/amino acid sequence. The switch 1 and switch 2 regions of the polypeptide sequence are underlined. Nucleotide lesions associated with each mutant are represented above the nucleotide sequence, as well as the allele name and the resulting amino acid substitution. mig-2(mu28) is a G179-to-A transition resulting in a premature stop codon; mig-2(lq13) is a C224-to-T transition resulting in an S75-to-F substitution; ced-10(n1993) is a T569-to-G transition resulting in a V190-to-G substitution; ced-10(lq20) is a C86-to-T transition resulting in a P29-to-L substitution; and ced-10(n3246) is a G178-to-A transition resulting in a G60-to-R substitution. (D) A diagram of general structure of MIG-2, CED-10, and RAC-2 GTPases. Indicated are the P-loop, the switch 1 and switch 2 regions, two guanine nucleotide binding domains (GNB) and the C-terminal CAAX sequence that mediates prenylation. MIG-2 has an N-terminal myristoylation sequence (not shown) that is not found on CED-10 or RAC-2. The relative locations of the ced-10(lq20 and n3246) and mig-2(lq13) mutations are indicated.

 
We analyzed the effects of ced-10(tm597) on pathfinding of the axons of two populations of neurons, the PDEs and the VD and DD motor neurons. The cell bodies of the bilateral PDE neurons reside in the postdeirid ganglia in the posterior-lateral region of the animal and can be visualized and unambiguously identified with an osm-6::gfp transgene (Figure 2A) (COLLET et al. 1998; STRUCKHOFF and LUNDQUIST 2003). The PDEs extend a single dendrite dorsally and a single unbranched axon straight ventrally to the VNC, where the axon bifurcates and extends anteriorly and posteriorly in the VNC (WHITE et al. 1986). The VD and DD motor neuron cell bodies reside along the VNC and normally extend axons anteriorly in the VNC, which then turn dorsally and extend as single-axon commissures to the dorsal nerve cord (Figure 2C) (WHITE et al. 1986). We found that ced-10(tm597M+) animals alone displayed few defects in PDE and VD/DD axon pathfinding (Figure 3A). Furthermore, trans-heterozygous combinations of ced-10(tm597)with ced-10(n1993) lacking maternal wild-type ced-10, which were viable and fertile, also showed few axon defects, similar to ced-10(n3417)/ced-10(n1993) (Figure 3A). Single mutant genotypes displayed some weak defects in VD/DD axon pathfinding [e.g., mig-2(mu28) displayed 20%; Figure 3A]. However, these defects were not as severe as those seen in the double mutants described below. Thus, neither ced-10(n3417) nor ced-10(tm597), both likely to be null mutations, had strong effects on PDE and VD/DD axon pathfinding.

View larger version (105K): In this window In a new window Download PPT slide   FIGURE 2.— Axon pathfinding defects of ced-10(tm597); mig-2(mu28) double mutants. (A and B) Fluorescent micrographs of fourth larval stage (L4) animals with osm-6::gfp expression in the PDE neuron. (C and D) Micrographs of L4 animals with unc-25::gfp expression in the VD and DD motor neurons. Anterior is left and dorsal is up. (A) A wild-type PDE neuron. The PDE axon extended directly to the ventral nerve cord (dashed line) without wandering or forming ectopic branches. (B) The PDE axon of a mig-2(mu28); ced-10(tm597M+) animal failed to extend directly to the ventral nerve cord, wandered laterally, and exhibited ectopic branches. (C) A wild-type animal. The VD and DD cell bodies resided along the ventral nerve cord (arrowheads) and the commissural axons extended directly to the dorsal nerve cord without wandering or exhibiting extra branches (arrows). (D) A mig-2(mu28); ced-10(tm597M+) animal. The commissural VD and DD axons wandered laterally, terminated prematurely without reaching the dorsal cord, and exhibited ectopic branches (arrows). VD or DD cell bodies were displaced laterally from their normal position in the ventral cord (arrowheads). Bars: A and B, 10 µm; C and D, 20 µm.

 

View larger version (27K): In this window In a new window Download PPT slide   FIGURE 3.— Null ced-10 alleles synergize with mig-2 and rac-2 in axon pathfinding and neuronal migration. (A) The graph displays the percentage of axon pathfinding defects (y-axis) of various genotypes (x-axis). Shaded bars represent PDE axon guidance defects and solid bars represent VD/DD axon pathfinding defects (see MATERIALS AND METHODS). M+ indicates that the mutants had wild-type maternal contribution for that locus. (B) The graph displays the percentage of AQR and PQR neuronal migration defects along their normal routes (y-axis) in various genotypes (x-axis). See MATERIALS AND METHODS for scoring of AQR and PQR migration defects along their normal routes. M+ indicates that the mutants had wild-type maternal contribution for that locus. In all cases except mig-2(rh17), AQR and PQR were equally affected. In mig-2(rh17) mutants, PQR was predominantly affected. Error bars represent the standard error of the proportion.

 
We found that both ced-10(n3417) and ced-10(tm597) displayed severe axon defects in double mutant combinations with mig-2(mu28) (Figures 2 and 3A). Defects in PDE axon pathfinding in ced-10(n3417M+); mig-2(mu28) and ced-10(tm597M+); mig-2(mu28) double mutants included axon guidance errors, indicated by failure of the axon to extend directly to the VNC and instead wander laterally (69% and 72% of PDEs from ced-10(n3417M+); mig-2(mu28) and ced-10(tm597M+); mig-2(mu28), respectively); and formation of ectopic neurites emanating from the cell body or from the axon (Figure 2B). Defects in VD/DD axon pathfinding in ced-10(n3417M+); mig-2(mu28) and mig-2(mu28); ced-10(tm597M+) double mutants included failed guidance, indicated by commissural axons that wandered laterally and sometimes failed to reach the dorsal cord; axon termination, when axon extension terminated prematurely before reaching the dorsal cord; and formation of ectopic neurites emanating from the commissural axon (Figures 2D and 3A). Every ced-10(n3417M+); mig-2(mu28) and ced-10(tm597M+); mig-2(mu28) mutant animal analyzed showed VD/DD axon pathfinding defects. A third Rac GTPase, RAC-2, acts redundantly with ced-10(n1993) and mig-2(mu28) (LUNDQUIST et al. 2001). We found that ced-10(n3417) and ced-10(tm597) animals treated with rac-2 RNAi displayed similar but weaker axon pathfinding defects as described for the mig-2 doubles (Figure 3A). We used RNAi to silence rac-2 because neither of the two extant rac-2 alleles caused a phenotype in double mutant combinations with mig-2 and ced-10 (E.A.L., unpublished results). We have strong evidence that rac-2 RNAi specifically targets rac-2 depletion (LUNDQUIST et al. 2001) and that the two rac-2 alleles are not null (E.A.L., unpublished results). However, these results with rac-2 RNAi should be interpreted with caution until the null phenotype of rac-2 is clarified. With previous results, these data indicate that null alleles of neither ced-10 nor mig-2 strongly affect axon pathfinding and that ced-10, mig-2, and rac-2 have overlapping roles in axon pathfinding.

mig-2, ced-10, and rac-2 have overlapping roles in AQR and PQR neuronal migration:

Previous results indicate that ced-10, mig-2, and rac-2 have overlapping roles in migration of the CAN neurons (LUNDQUIST et al. 2001). We found that the three rac genes also redundantly control the migrations of the AQR and PQR neurons. The AQR neuron is born from the Q neuroblast on the right side of the animal (QR) in the posterior midbody region and migrates anteriorly to a position near the posterior pharynx (Figure 4A and MATERIALS AND METHODS). The PQR neuron is born from the left Q neuroblast (QL) and migrates posteriorly to a position near the phasmid neurons in the left lumbar ganglion (Figure 4A and MATERIALS AND METHODS). The Q cells respond to left-right asymmetry cues by extending processes anteriorly (QR) or posteriorly (QL), which polarizes the cells to migrate anteriorly or posteriorly, respectively (HONIGBERG and KENYON 2000). After an initial Q cell migration, Wnt signaling induces the expression of the mab-5 gene in QL, which causes QL and its descendants to continue their posterior migrations (SALSER and KENYON 1992; MALOOF et al. 1999). QR and its descendants, which do not express mab-5, continue their anterior migrations. Mutations that disrupt the orientation of initial QL polarity and initial migration as well as those that disrupt later mab-5 expression cause QL and/or its daughter cells to migrate anteriorly instead of posteriorly (HONIGBERG and KENYON 2000). We monitored PQR direction of migration by scoring misplacement of PQR anterior to the vulva, indicating that PQR or its precursors migrated anteriorly instead of posteriorly. We also monitored the ability of AQR and PQR to migrate along their normal routes by assaying misplacement of AQR and PQR along their correct migration routes (e.g., a PQR neuron that migrated posteriorly but stopped before reaching its final destination near the phasmid neurons).

View larger version (81K): In this window In a new window Download PPT slide   FIGURE 4.— AQR and PQR neuronal migration defects in rac double mutants. Anterior is to the left and dorsal is up. All are fluorescence micrographs of fourth larval stage (L4) animals with osm-6::gfp expression in the AQR, PQR, and other neurons. The position of the vulva is indicated by a "V." (A) A wild-type animal. The position of the birthplace of QL and QR is indicated, as are the migration routes of QL and descendants (PQR) (posterior arrow) and QR and descendants (AQR) (anterior arrow). The final positions of PQR and AQR are indicated. (B) A ced-10(tm597M+); mig-2(mu28) animal with mispositioned AQR and PQR neurons. The neurons failed to complete their migrations along their normal paths (AQR migrated anteriorly and PQR migrated posteriorly, but both failed to reach their normal positions). (C) The posterior region of ced-10(tm597M+); mig-2(mu28) animal with a mispositioned PQR neuron on the left side of the animal. Bars: A and B, 20 µm; C, 10 µm.

 
No single loss-of-function rac mutation strongly affected the direction of PQR migration or the ability of AQR and PQR to migrate (Figure 3B), although mig-2(mu28) showed weak but significant defective AQR and PQR migration along their normal routes (5%; Figure 3B), consistent with previous results demonstrating that mig-2(mu28) affects the migrations of other Q neuroblast daughters.

In pairwise double rac mutant combinations, including those with rac-2(RNAi) and the putative null ced-10 deletion alleles, strong defects in AQR and PQR migration along their normal routes were observed (Figure 3B and Figure 4, B and C). For example, ced-10(tm597M+) and mig-2(mu28) alone showed 3% and 5% AQR/PQR migration defects respectively, whereas the ced-10(tm597M+); mig-2(mu28) double mutant displayed 67% AQR/PQR migration defects. The effects observed with rac-2(RNAi) were generally weaker than those observed with ced-10 and mig-2 mutations. Thus, ced-10, mig-2, and rac-2 have overlapping roles in AQR and PQR neuronal migration. We detected no significant defect in the direction of PQR migration in these double mutants, suggesting that the three rac genes might have overlapping roles in the ability of AQR and PQR to migrate but not in the direction of PQR migration. We also observed lateral misplacement of VD/DD motor neuron cell bodies in double mutants (Figure 2D) consistent with previous results showing that the rac genes redundantly control the ventral migrations of the P cells, the progenitors of the VD and DD neurons (LUNDQUIST et al. 2001). The previously described constitutively active mig-2(rh17) allele (ZIPKIN et al. 1997) caused defects in PDE axon pathfinding and AQR and PQR migration (Figure 3B). While PQR was more often affected than AQR in mig-2(rh17) animals, defects in AQR migration were also observed.

Previous studies implicated the Dbl-homology GTP exchange factor (DH-GEF) UNC-73 Trio as acting in all three Rac pathways in PDE axon pathfinding and CAN cell migration (LUNDQUIST et al. 2001; WU et al. 2002). Furthermore, unc-73 mutations affect the ability of Q cells to polarize and the ability of Q cells and their descendants to execute migrations. However, unc-73 does not affect orientation of Q-cell polarity as do unc-40, dpy-19, and mab-5 (i.e., unc-73 mutants do not display QL direction of migration defects) (HONIGBERG and KENYON 2000). Consistent with these results, we found that unc-73 mutants displayed defects in AQR and PQR migration along their normal routes (Figure 3B) but did not affect PQR direction of migration. These data indicate that the Racs and UNC-73 control the ability of Q cells or descendants to migrate along their normal routes. No defects in direction of PQR migration were observed, suggesting that UNC-73 and the Racs are not involved in controlling the direction of PQR migration. The unc-73(rh40) allele specifically eliminates the Rac GEF activity of UNC-73. The severity of the defects of unc-73(rh40) were less than those of mig-2; ced-10 double mutants, indicating that UNC-73 might not be the only Rac GEF involved in the migrations of Q cells or their descendants.

MIG-15 NIK controls axon pathfinding:

Vertebrate NIK (SU et al. 1997) and Drosophila Misshapen (TREISMAN et al. 1997) are serine/threonine kinases of the GCK family and contain an N-terminal STE20-like kinase domain and a C-terminal CNH (citron/NIK homology) domain, which is thought to interact with Rho-family GTPases. In C. elegans, the mig-15 gene encodes a NIK homolog (Figure 5A) (POINAT et al. 2002), and RNAi of ced-10, rac-2, and mig-2 increase the frequency of axon pathfinding defects in the VD/DD commissural axons of a hypomorphic mig-15 mutation (POINAT et al. 2002), suggesting that mig-15 and the rac genes interact in axon pathfinding, although it is unclear if they act in the same pathway or in parallel pathways.

View larger version (71K): In this window In a new window Download PPT slide   FIGURE 5.— Effects of mig-15 and rac mutants on PDE axon pathfinding and AQR/PQR migration. (A) A diagram of the predicted 1087-residue MIG-15 polypeptide (N terminus to the left) (GenBank accession no. AF087131; Wormbase Gene Model ZC504.4a). The N terminus of the polypeptide contains a region similar to a serine-threonine kinase of the STE20 class (STE20 S/T kinase), and the C-terminus contains a region similar to a regulatory domain of the protein Citron (CNH). The central region of the polypeptide contains a proline-rich domain (PRD). The locations of the mig-15 alleles are shown above the structure. Below the structure is the polypeptide sequence in the region of the rh148 lesion. Residues in this region that contribute to the ATP binding pocket of the kinase domain are boxed (residues outside of this region also contribute to the ATP binding pocket). rh148 was a T-to-A transversion at position 503 in the mig-15 open reading frame resulting in a V169E missense mutation; rh80 was a G-to-A transition at position 2694 in the mig-15 open reading frame resulting in an opal stop (TGG to TGA); and rh326 was a G-to-A transition at position 1315 in the mig-15 open reading frame resulting in an ochre stop (CAA to TAA). (B) The graph depicts the percentage of PDE axon guidance defects (y-axis) in various genotypes (x-axis) (see MATERIALS AND METHODS for PDE scoring). (C) The graph represents percentages of defects (y-axis) in various genotypes (x-axis). Gray bars represent the percentage of anteriorly displaced PQR neurons (anterior to the vulva), indicating a defect in the direction of PQR migration. Black bars represent the percentage of AQR neurons that migrated anteriorly but failed to reach their normal final position. M+ indicates that the mutants had wild-type maternal contribution for that locus. The asterisks (*) denote that in mab-5(e1751) and mig-15(rh148); mab-5(e1751) animals, AQR is misdirected posteriorly and resides among the phasmid neurons 86% and 73% of mutants, respectively (see RESULTS). Error bars represent the standard error of the proportion.

 
To understand the relationship of MIG-15 NIK and Rac signaling in PDE axon pathfinding, we first analyzed the effects of mig-15 mutants on PDE axon development. Alone, mig-15(rh148) and mig-15(rh80) displayed moderate defects in PDE axon pathfinding, including ventral PDE axon guidance errors (25% and 64%, respectively; Figures 5B and 6A) and formation of ectopic neurites (Figure 6A). We were unable to maintain the mig-15(rh326) allele as a homozygous strain due to sterility and larval lethality caused by the mutation. However, mig-15(rh326) homozygotes from a mig-15(rh326)/+ heterozygous mother (wild-type mig-15 maternal contribution) displayed 62% PDE ventral axon pathfinding errors, compared to 23% for mig-15(rh80) animals with wild-type maternal contribution (Figure 5B).

View larger version (66K): In this window In a new window Download PPT slide   FIGURE 6.— PDE axon pathfinding and AQR/PQR cell migration defects in mig-15 mutants. Anterior is to the left and dorsal is up. All are fluorescence micrographs of fourth larval stage (L4) animals with osm-6::gfp expression in the PDEs, AQR, PQR, and other neurons. The position of the vulva is indicated by a "V." (A) A PDE axon of a mig-2(rh80) animal failed to extend directly to the ventral surface (dashed line) and instead wandered laterally. The axon also exhibited ectopic neurites. (B) A left-lateral view of a mig-15(rh148) animal. The PQR neuron migrated anteriorly instead of posteriorly. (C) A left-lateral view of a mig-15(rh80) mutant. The PQR neuron migrated anteriorly, and the AQR neuron did not complete its anterior migration. Bars: A, 10 µm; B and C, 20 µm.

 
On the basis of PDE axon defects (Figure 5B) and on viability and growth, the three mig-15 alleles formed an allelic series from least severe to most severe: rh148 < rh80 < rh326. We determined the molecular nature of the three mig-15 alleles. The weakest allele, rh148, resulted in a missense change of valine 169 to glutamic acid (V169E) (Figure 5A). On the basis of comparison with the crystal structures of other GCK kinase family members, V169 of MIG-15 might contribute to the hydrophobic ATP binding pocket of the putative kinase domain (Figure 5A) (LOWE et al. 1997; NIEFIND et al. 1998; MARCHLER-BAUER et al. 2005). In rh148, a charged glutamic acid residue at this position might interfere with ATP binding and affect the activity of the kinase. rh80 resulted in a nonsense change of tryptophan 898 to an opal stop codon (W898STOP) (Figure 5A) in the middle of the conserved CNH domain coding region, which is thought to mediate interaction with Rho and Rac GTPases (MADAULE et al. 1995; MARCHLER-BAUER et al. 2005). rh326 resulted in a nonsense change of glutamine 439 to an ochre stop codon (Q439STOP) (Figure 5A). rh326 is a good candidate for a mig-15 null mutation, as rh326 resulted in an early predicted stop and had the strongest PDE axon pathfinding defects. Both rh148 and rh80 likely retain some mig-15 activity.

MIG-15 acts with the Racs in axon pathfinding:

The PDE axon pathfinding defects of hypomorphic mig-15(rh148) were enhanced by ced-10(n1993). For example, mig-15(rh148) displayed 25% PDE ventral guidance errors, whereas mig-15(rh148); ced-10(n1993) doubles displayed 69% (Figure 5B). Double mutants of ced-10 with mig-15(rh148) were viable and fertile, whereas ced-10(n1993) double mutants with mig-15(rh80) and mig-15(rh326) arrested as larvae. Despite wild-type maternal mig-15 (indicated by M+ in the genotype), double mutants of ced-10(n1993); mig-15(rh80M+) displayed 63% PDE ventral axon guidance errors compared to 23% for mig-15(rh80M+) alone (Figure 5B). ced-10(n1993); mig-15(rh326M+) displayed 74% PDE axon pathfinding defects compared to 62% for mig-15(rh326M+) alone (Figure 5B), suggesting that ced-10(n1993) might slightly enhance mig-15(rh326). However, ced-10(n1993); mig-15(rh326M+) doubles were often larval lethal and had severe body morphology defects that might have contributed to apparent PDE axon defects. mig-2(mu28) also enhanced the PDE axon defects of hypomorphic mig-15(rh148) to a level similar to mig-15(rh80) and mig-15(rh326M+), and rac-2(RNAi) slightly enhanced mig-15(rh148) but had no effect on mig-15(rh80).

Both ced-10 and mig-2 mutations enhanced the PDE axon defects of the weak mig-15(rh148) allele to levels similar to the stronger mig-15(rh80 and rh326) alleles, and ced-10 enhanced the defects of mig-15(rh80) with wild-type maternal mig-15 contribution, suggesting that both rh148 and rh80 are hypomorphic alleles. No defects in rac; mig-15 doubles were strikingly stronger than the putative null allele mig-15(rh326M+) alone. These data are consistent with a model in which mig-15 and the three rac genes might participate in the same pathway (i.e., the three Racs have overlapping roles but mig-15 might act in each Rac pathway). However, ced-10(n1993) might have slightly enhanced the mig-15(rh326) null allele, raising the possibility that mig-15 and ced-10 act in parallel pathways. In any case, these data indicate that mig-15 and three rac genes ced-10, mig-2, and rac-2 act together to control PDE axon pathfinding.

MIG-15 NIK and MIG-2 Rac control direction of PQR migration:

We found that mig-15 mutants alone affected the direction of PQR migration. In 46% of mig-15(rh148), 81% of mig-15(rh80), and 47% of mig-15(rh326M+) animals, PQR was positioned anterior to the vulva, indicating that PQR, its precursors from QL, or QL itself migrated anteriorly rather than posteriorly (Figure 5C and Figure 6, B and C). mig-15(rh326M+) was weaker than mig-15(rh80), possibly because of wild-type maternal mig-15 activity. Indeed, rh80 with maternal mig-15(+) was weaker than mig-15(rh326M+) (Figure 5C). mig-15 mutants also displayed defects in AQR migration along its normal route (11%, 71%, and 60% in rh148, rh80, and rh326M+ respectively; Figure 5C and Figure 6C).

AQR/PQR defects of mig-15 mutants were not affected by ced-10(n1993) or by rac-2(RNAi) (Figure 5C). In contrast, the mig-2(mu28) mutation significantly enhanced the PQR directional migration defects and the AQR migration defects of mig-15(rh148) to a level similar to that of mig-15(rh80) alone (Figure 5C). Alone, mig-2(mu28) had no effect on directional PQR migration, but did display weak defects in AQR and PQR migration along their normal routes (5%; Figure 3B). These data indicate that MIG-2 might act in the MIG-15 pathway in controlling direction of PQR migration, and that CED-10 and RAC-2 are not involved in this process. The AQR migration defects of mig-15 were also enhanced by mig-2(mu28) but not by ced-10(n1993) or rac-2(RNAi). Constitutively active mig-2(rh17), which affected the ability of AQR and PQR to migrate along their normal routes (Figure 3B), did not affect AQR or PQR direction of migration (data not shown).

MIG-15 might act upstream of MAB-5 in controlling PQR direction of migration:

Our data indicate that mig-15 controls the direction of PQR migration, which could be due to defects in the polarity of QL cell or its daughters. Q-cell polarity occurs in at least two phases: an initial morphological polarity and short migration, where QL polarizes to the posterior and QR polarizes to the anterior; and a later Wnt signal that instructs posterior migration of QL and daughters but not QR and daughters (HONIGBERG and KENYON 2000). Genes known to control initial QL polarity include unc-40 and dpy-19 (HONIGBERG and KENYON 2000). We found that the PQR neuron was misdirected anteriorly in 25% of unc-40(e271) mutants (n = 100) and 33% of dpy-19(e1259) mutants (n = 100). The misdirected PQR neurons in these mutants migrated anteriorly to a position near the normal final anterior position of AQR (a complete anterior migration). In contrast, the migration of the misdirected PQR neurons in mig-15 often failed along their anterior routes (Figure 6, B and C). Mutations in Wnt signaling genes also cause incomplete migrations of misdirected QL daughters (WHANGBO and KENYON 1999; HONIGBERG and KENYON 2000; HERMAN 2001).

After the initial polar posterior migration of QL that is dependent upon UNC-40 and DPY-19, Wnt signaling induces the expression of the homeodomain transcription factor MAB-5 in QL but not in QR (SALSER and KENYON 1992; WHANGBO and KENYON 1999; HONIGBERG and KENYON 2000). MAB-5 expression in QL and daughters controls these cells' further posterior migrations. As expected, the QL daughter PQR was misplaced anteriorly in 75% (n = 100) of mab-5(e1239) loss-of-function mutants. The misdirected PQR neurons reached the normal anterior position of AQR, similar to dpy-19 and unc-40 mutants. The gain-of-function mab-5(e1751) allele causes constitutive mab-5 activity in both QL and QR and daughter cells independent of Wnt signaling (SALSER and KENYON 1992). As expected, PQR placement was normal (in the posterior) in all mab-5(e1751gf) animals examined (n = 100) (Figure 5C) and AQR was mispositioned in the posterior among the phasmid neurons in 86% (n = 100) of mab-5(e1751gf) animals.

These data show that PQR was mispositioned anteriorly in mig-15 mutants and was normally positioned in mab-5(e1751gf) mutants. In mig-15(rh148); mab-5(e1751gf) double mutants, AQR and PQR placement resembled that of mab-5(e1751) alone: PQR was positioned normally in the posterior in all mig-15(rh148); mab-5(e1751gf) animals examined (n = 100) (Figure 5C) and AQR was found in the posterior among the phasmid neurons in 73% (n = 100) of double mutants. This result indicates that mab-5 activation in QL compensated for loss of mig-15 and that mig-15 might act upstream of mab-5 in PQR migration.

mig-2(lq13), ced-10(n1993lq20), and ced-10(n3246) are gain-of-function rac mutations:

To identify other genes that act redundantly with ced-10 in axon pathfinding, we undertook a genetic screen for mutations that were synthetic lethal with ced-10 with the idea that the genes identified by these mutations might also act in parallel to ced-10 in axon pathfinding (see MATERIALS AND METHODS for details of the screen).

The lq13 mutation was a new allele of mig-2 and lq20 was a new allele of ced-10 (see MATERIALS AND METHODS). ced-10(n1993); mig-2(lq13) double mutants arrested as embryos and larvae with defects in gastrulation, similar to ced-10 null animals and ced-10(n1993); mig-2(mu28) animals (data not shown). ced-10(n1993lq20) mutants displayed variable lethality characteristic of other ced-10-related lethality (defects in gastrulation; data not shown). However, n1993lq20 also gave rise to nonlethal "escaper" animals that were severely Unc. We found that lq13 was a C-to-T change that results in a missense mutation of serine 75 of MIG-2 to phenylalanine (S75F) (Figure 1B), and that lq20 was a C-to-T missense mutation resulting in a proline 29-to-leucine change (P29L) in CED-10 (Figure 1C). We also confirmed the presence of the n1993 lesion in this strain (a G-to-T transition resulting in a change of valine 190 to glycine). ced-10(n1993) is a hypomorphic allele that retains some ced-10 activity. Because n1993 is not a null allele, the potential gain-of-function effect of lq20 could be expressed in the n1993 background.

In contrast to mig-2 and ced-10 null mutants, mig-2(lq13) animals were slightly uncoordinated and ced-10(n1993lq20) were severely uncoordinated (data not shown). Furthermore, mig-2(lq13) and ced-10(n1993lq20) mutants displayed PDE axon pathfinding defects not observed in mig-2 or ced-10 loss-of-function mutants (Figure 7A): the mig-2(mu28) null mutant displayed 1% PDE axon guidance errors whereas mig-2(lq13) displayed 24%; and ced-10(n3417 and tm597) null mutants displayed 3–5% PDE axon defects compared 38% in ced-10(n1993lq20). The previously described mutation ced-10(n3246) is a G-to-A transition resulting in a change of glycine 60 of CED-10 to arginine (G60R) (Figure 1C) (REDDIEN and HORVITZ 2000). ced-10(n3246) animals also displayed slight uncoordination (data not shown) as well as defects in PDE and VD/DD axon pathfinding (Figures 3A and 7A) not observed in ced-10 null mutants.

View larger version (129K): In this window In a new window Download PPT slide   FIGURE 7.— PDE axon pathfinding defects and AQR/PQR migration defects in rac gain-of-function mutants. (A) The graph represents the percentage of defects (y-axis) in various genotypes (x-axis). Gray bars represent the percentage of PDE axon guidance defects, and the black bars represent percentages of AQR and PQR migration defects along their normal routes (see MATERIALS AND METHODS for scoring). In each genotype, PQR was affected more often than AQR. Error bars represent the standard error of the proportion. (B–E) Fluorescence micrographs of osm-6::gfp expression in PDE neurons in animals harboring gain-of-function rac transgenes (see MATERIALS AND METHODS). Anterior is to the left and dorsal is up. The dashed line indicates the ventral midline. (B) The wild-type PDE axon extends directly to the ventral midline. (C–E) The axons of animals transgenic for mig-2(lq13), ced-10(lq20), and ced-10(n3246) were misguided and displayed ectopic neurites. Bar for B–E, 10 µm.

 
mig-2(lq13), ced-10(n1993lq20), and ced-10(n3246) affected migration of AQR and PQR neurons along their normal routes (Figure 7A) [e.g., 5% for mig-2(mu28) compared to 27% for mig-2(lq13)]. While execution of AQR and PQR migration was affected, these mutations had no effect on direction of AQR or PQR migration (data not shown). Together, these data indicate that mig-2(lq13), ced-10(n1993lq20), and ced-10(n3246) are not simple loss-of-function mutations and are likely to be gain-of-function mutations. However, each of these mutations was recessive to wild-type for Unc and for PDE axon defects and AQR/PQR migration (data not shown).

To further test the idea that lq13, n1993lq20, and n3246 were gain-of-function mutations, we used a transgenic assay to study the consequences of expression of MIG-2 with the lq13 mutation and CED-10 with the lq20 and n3246 mutations specifically in the PDE neuron. We constructed transgenes consisting of the mig-2 or ced-10 genomic coding regions (exons and introns; STRUCKHOFF and LUNDQUIST 2003) downstream of the osm-6 promoter, which is expressed in all ciliated sensory neurons including PDE (see MATERIALS AND METHODS). Consistent with previous results (STRUCKHOFF and LUNDQUIST 2003), we found that transgenic expression of wild-type MIG-2 or CED-10 in the PDE caused few defects in axon pathfinding or PDE morphogenesis (data not shown). In contrast, transgenic expression of MIG-2(lq13), CED-10(lq20), and CED-10(n3246) caused severe defects in PDE axon pathfinding: the PDE neurons displayed ectopic neurites and misguided axons (Figure 7C) similar to the phenotype of mig-2(mu28); ced-10(n1993) loss-of-function mutants. However, MIG-2(lq13), CED-10(lq20), and CED-10(n3246) expression did not result in robust formation of ectopic lamellipodia-like and filopodia-like structures caused by transgenic expression of constitutively active MIG-2(G16V) or CED-10(G12V) (STRUCKHOFF and LUNDQUIST 2003).

In summary, these data suggest that mig-2(lq13), ced-10(n1993lq20), and ced-10(n3246) are gain-of-function alleles: in contrast to the null alleles of mig-2 and ced-10, each caused axon pathfinding and cell migration defects; and transgenic expression of these mutant molecules caused dominant PDE axon defects whereas transgenic expression of wild-type CED-10 and MIG-2 did not.

unc-34(lq17) affects splicing of the unc-34 transcript:

The ced-10 synthetic lethal mutation lq17 was found to be a hypomorphic allele of unc-34, which encodes the C. elegans homolog of the cytoskeletal signaling protein Enabled (Figure 8A) (WITHEE et al. 2004). ced-10(n1993); unc-34(lq17) animals that lost the ced-10(+) array were embryonic lethal with a Gex phenotype similar to ced-10 null mutants (data not shown). Thus, UNC-34 Enabled might act redundantly with Rac signaling in gastrulation.

On the basis of the synthetic lethality with ced-10(n1993), we determined that the genetic position of lq17 was on the left arm of linkage group V (see MATERIALS AND METHODS), the location of the unc-34 gene that has previously been shown to affect axon pathfinding and neuronal migration. We found that lq17 failed to complement the Unc phenotype of three unc-34 alleles, e315, gm104, and, e951, suggesting that lq17 was an allele of unc-34 (data not shown). We determined the nucleotide sequence of unc-34 from lq17 mutants and found a G-to-A transition that affected the first nucleotide residue of the last unc-34 intron and that was predicted to abolish use of this site as a splice donor (Figure 8B). To determine the effect of the lq17 mutation on the unc-34 transcript, we used RT–PCR to generate cDNAs representing the unc-34 transcript from lq17 mutants. In wild type, we observed the expected cDNA product of approximately 1.4 kb, and sequencing confirmed that this product represented the unc-34/Y50D4C.1a mRNA. Using the same RT–PCR protocol, we obtained a cDNA of approximately 1.4 kb from unc-34(lq17) animals. Sequencing of the 1.4-kb lq17 transcript revealed that a cryptic donor splice site 14 nt to the 3' of the normal intron 5' splice donor site was used in this transcript, resulting in a 13-bp insertion into the unc-34 transcript (Figure 8, B and C). The 13-bp addition introduced an in-frame stop codon that would result in truncation of the C-terminal 43 residues of the UNC-34 polypeptide.

The C-terminus of UNC-34 contains a predicted EVH2 domain (Figure 8A), which consists of A, B, and C functional subdomains (KRAUSE et al. 2003; WITHEE et al. 2004). The lq17 mutation leads to a 43-residue truncation of the UNC-34 polypeptide that is predicted to remove most of the EVH2 C domain while leaving other domains unaffected (Figure 8C). The EVH2 C domain is thought to mediate tetramerization of the Enabled molecule, which facilitates its actin-binding ability (HAFFNER et al. 1995; BACHMANN et al. 1999; ZIMMERMANN et al. 2002). Thus, an UNC-34 polypeptide produced in lq17 animals might lack the ability to tetramerize but might retain other functions of the molecule.

UNC-34 Ena acts in parallel to CED-10 Rac, RAC-2 Rac, and MIG-2 Rac in axon pathfinding and neuronal migration:

To assess the effect of the new lq17 mutation on these processes, we isolated lq17 away from ced-10(n1993) and the ced-10(+) array and scored PDE axon pathfinding and AQR/PQR migration in these animals. Alone, lq17 animals were slightly Unc but less severely so than extant null unc-34 alleles (data not shown). Furthermore, lq17 animals showed weak defects in PDE axon pathfinding and few AQR/PQR migration defects (Figure 9A). In contrast, other unc-34 mutants e951 and gm104, both predicted to be null alleles (WITHEE et al. 2004), showed stronger defects in PDE axon guidance (9%) and AQR/PQR migration along their normal routes (4–5%) (Figure 9A), but did not affect direction of PQR migration. Thus, lq17 is likely to be a hypomorphic allele of unc-34.

View larger version (61K): In this window In a new window Download PPT slide   FIGURE 9.— PDE axon pathfinding and AQR/PQR migration defects in unc-34 mutants. (A) The graph represents percentages of defects (y-axis) in various genotypes (x-axis). Shaded bars represent PDE ventral axon guidance errors, and solid bars represent AQR/PQR migration defects along their normal routes. M+ indicates that the animals had wild-type maternal contribution for that locus. Error bars represent the standard error of the proportion. (B and C) Fluorescent micrographs of animals with osm-6::gfp expression. (B) A unc-34(lq17M+); ced-10(n1993M+) animal had a misguided PDE axon that failed to extend directly to the ventral midline (dashed line). Anterior is left; dorsal is up. (C) A unc-34(lq17); mig-2(mu28) animal, viewed from a ventral perspective, displayed a PQR neuron that failed to migrate to its final position. Anterior is to the left and the left side of the animal is up. Bars: B, 10 µm; C, 20 µm.

 
To determine the interactions between unc-34 and the rac genes ced-10, mig-2, and rac-2, we constructed double mutants of ced-10 and mig-2 with three unc-34 alleles, including the hypomorphic lq17 allele and the null alleles e951 and gm104. Double mutants of unc-34 and ced-10 were maintained as double heterozygotes over the balancer nT1, which balances both unc-34 and ced-10. Double homozygotes from this balanced strain were viable, but in each case maternal-effect embryonic lethal with the Gex phenotype (data not shown). We scored PDE development and AQR/PQR migration in double mutants with wild-type unc-34 and ced-10 maternal contribution. ced-10(M+); unc-34(e951M+) and ced-10(M+); unc-34(gm104M+) displayed synergistic defects in PDE axon pathfinding and in AQR/PQR migration along their normal routes (Figure 9A). For example, unc-34(e951) displayed 9% PDE axon guidance defects whereas unc-34(e951M+); ced-10(n1993M+) displayed 40%. ced-10(n1993M+); unc-34(lq17M+) displayed a lesser degree of synergy (Figure 9, A–C), supporting the idea that lq17 is a hypomorphic allele. No defects in the direction of PQR migration were observed.

We analyzed mig-2(mu28); unc-34 double mutants, which were viable and fertile and did not show strong maternal-effect embryonic lethality. unc-34; mig-2 animals were severely uncoordinated, and each mig-2; unc-34 double mutant combination displayed synergistic defects in PDE axon pathfinding and AQR/PQR migration along their normal routes (Figure 9A). Direction of PQR migration was unaffected. These data indicate that while unc-34 and mig-2 have overlapping roles in axon pathfinding and neuronal migration, they do not have overlapping roles in embryonic gastrulation.

Finally, to determine if the third C. elegans rac gene rac-2 also acted with unc-34, we knocked down rac-2 function using RNAi in unc-34(lq17 and e951). We found synergistic defects in PDE axon pathfinding and AQR/PQR migration in each case (Figure 9A). Direction of PQR migration was not affected in rac-2(RNAi); unc-34 doubles, and defects with unc-34(lq17) were less severe than with unc-34(e951).

These data indicate that mutations in ced-10, mig-2, and rac-2(RNAi) enhance axon pathfinding and AQR/PQR migration defects of null unc-34 mutants, suggesting that UNC-34 Enabled acts in parallel to ced-10, rac-2, and mig-2 in axon pathfinding and neuronal migration. Interestingly, ced-10(n1993) and rac-2(RNAi) synergized with unc-34 in gastrulation and embryonic elongation but mig-2 did not, suggesting that mig-2 and unc-34 might act in a common pathway in parallel to ced-10 and rac-2 in this event.

MIG-15 NIK and the actin binding protein UNC-115 abLIM act in parallel to UNC-34 Ena in axon pathfinding:

The above data are consistent with the idea that three redundant Rac pathways, each using MIG-15 NIK, act in parallel to UNC-34 Ena in axon pathfinding and neuronal migration. If this were the case, we would expect that unc-34 and mig-15 mutations would display a synergistic phenotype. In fact, doubles mutants of unc-34(e951) with both mig-15(rh148) and mig-15(rh80) were larval lethal, even with wild-type maternal mig-15 activity. We were unable to score axon defects in these animals. However, the double hypomorphic mutant mig-15(rh148); unc-34(lq17) was viable and severely uncoordinated, and gave rise to some larval lethal animals. The severity of the PDE axon guidance defects in these animals (91%; Figure 5B) was higher than in mig-15(rh80) alone and higher than in any rac; mig-15 double mutant combination. These data indicate that MIG-15 and UNC-34 act in parallel, partially redundant pathways to control axon pathfinding [i.e., unc-34 enhanced mig-15 mutations to a level higher than mig-15(rh80) alone, whereas rac mutations did not]. unc-34(lq17) did not enhance defects in the direction of PQR migration or AQR migration of mig-15(rh148) (Figure 5C).

The actin-binding protein UNC-115 abLIM (Figure 10A) was previously shown to act downstream of RAC-2 in C. elegans axon pathfinding (STRUCKHOFF and LUNDQUIST 2003). If UNC-34 Ena acts in a parallel pathway with Rac signaling, we expect that mutations in unc-115 and unc-34 would display synergistic defects in axon pathfinding. We found that double mutants of unc-115 and unc-34 were viable and displayed no larval lethality. unc-115; unc-34 mutants displayed synergistic defects in PDE axon guidance (Figure 10, B and C). For example, unc-115(ky275), a null allele, displayed few PDE axon guidance defects, unc-34(e951) displayed 9% defects, and unc-115(ky275); unc-34(e951) displayed 33% defects. Double mutants of unc-115 with unc-34(lq17) showed weaker defects, consistent with the idea that lq17 is a hypomorphic unc-34 allele.

View larger version (121K): In this window In a new window Download PPT slide   FIGURE 10.— unc-115 acts with unc-34 in PDE axon pathfinding but not AQR/PQR migration. (A) A schematic of the 639-residue UNC-115 polypeptide (GenBank accession no. CAA9005), which contains three N-terminal LIM domains and a C-terminal actin-binding villin headpiece domain (VHD). The locations of the unc-115 mutations are indicated [mn481 is a Tc1 transposable element insertion and ky275 causes a premature stop (LUNDQUIST et al. 1998)]. (B) The graph represents percentages of defects (y-axis) in various genotypes (x-axis). Gray bars represent PDE ventral axon guidance defects and black bars represent AQR/PQR migration defects along their normal routes. Error bars represent the standard error of the proportion. (C) A fluorescent micrograph of an L4 animal with osm-6::gfp expression in the PDE neuron. The PDE axon of the unc-115(ky275); unc-34(e951) animal failed to extend directly to the ventral midline (dashed line). Bar in C, 10 µm.

 
unc-115; unc-34 double mutants did not display enhanced AQR/PQR migration defects (Figure 10B), nor did they display defects in PQR direction of migration. This result is consistent with previous studies indicating that unc-115 does not affect cell migration (e.g., unc-115 and rac mutations showed no enhanced CAN cell migration defects) (LUNDQUIST et al. 2001). Thus, UNC-115 might act in parallel to UNC-34 Ena signaling in axon pathfinding but not cell migration.

While many signaling molecules that affect axon pathfinding and cell migration have been identified, it is unclear how many of these molecules relate to one another in pathways and networks to control these events. Three Rac GTPases, the MIG-15 NIK kinase, and UNC-34 Enabled have all been shown to affect axon pathfinding and cell migration on their own. Here we present evidence that parallel, partially-redundant pathways control axon pathfinding and neuronal migration in C. elegans: three pathways defined by three Rac GTPases and the NIK kinase MIG-15; and another defined by the cytoskeletal signaling molecule UNC-34 Enabled (Figure 11, A and B). The partial redundancy of these pathways in axon navigation indicates that they affect the same cellular process in these events, possibly the organization and dynamics of the actin cytoskeleton of the growth cone that mediates the formation of lamellipodia and filopodia that control growth cone outgrowth and steering. While the three Racs and UNC-34 also redundantly control AQR and PQR migration along their normal routes, MIG-15 was not involved in this event. Rather, MIG-15 acted with MIG-2 to control the direction of PQR migration (Figure 11C).

View larger version (48K): In this window In a new window Download PPT slide   FIGURE 11.— Models of Rac signaling in axon pathfinding and neuronal migration. (A) PDE ventral axon guidance. Three Rac pathways redundantly control PDE axon guidance. MIG-15 is in parentheses, indicating that MIG-15 might act upstream or downstream of the Racs. UNC-115 acts downstream of RAC-2 Rac. UNC-34 Ena acts in a redundant pathway parallel to the Racs. (B) AQR/PQR migration along their normal routes. The involvement of UNC-34 Ena and the Racs is similar to PDE axon guidance; however, MIG-15 NIK and UNC-115 abLIM are not involved in this process. (C) PQR direction of migration. Initial QL polarity and migration is governed by UNC-40 and DPY-19, and later PQR is polarized by a Wnt signal, which induces MAB-5 expression in PQR and subsequent posterior migration. MIG-15 acts upstream of MAB-5 and might be required for either or both of these polarity events. MIG-2 acts with MIG-15 in controlling PQR direction of migration (upstream or downstream), and another unidentified molecule (?) likely acts in parallel to MIG-2, as mig-2 mutations alone have no effect on PQR direction of migration. CED-10 Rac, RAC-2 Rac, UNC-73 Trio, UNC-34 Ena, and UNC-115 abLIM are not involved in PQR direction of migration.

 

Three Rac GTPases have overlapping roles in axon pathfinding and neuronal migration:

Previous results using a null allele of mig-2 and a hypomorphic allele of ced-10 suggested that neither ced-10 nor mig-2 had a strong effect on axon pathfinding alone but that the two genes (and rac-2) had overlapping roles in axon pathfinding and neuronal migration (LUNDQUIST et al. 2001). Here, we analyzed the effects of two deletion alleles of ced-10 on axon pathfinding, both of which are likely to be null for ced-10 activity. We found that neither deletion allele resulted in strong defects in axon pathfinding and neuronal migration. While PDE and VD/DD axon outgrowth and AQR/PQR migration occur relatively late in development (in the larva), these events are still subject to maternal contribution [e.g., mig-15(rh80) with wild-type maternal mig-15(+) activity showed fewer axon and cell migration defects than mig-15(rh80) maintained as a homozygous line] (Figure 5B). Therefore, it was possible that ced-10(+) maternal activity was masking axon and cell migration defects in ced-10 null homozygous adults. This is unlikely, as trans heterozygotes of ced-10 deletion alleles and ced-10(n1993), which were viable and fertile and which would have reduced maternal ced-10 activity, also showed few axon defects (Figure 3A). The ced-10 deletion alleles strongly synergized with both mig-2 and rac-2 to cause severe axon defects, indicating that ced-10, mig-2, and rac-2 do indeed have overlapping, redundant roles in axon pathfinding and cell migration (Figure 3A).

While the rac genes controlled the migration of AQR and PQR along their normal routes, they did not affect the direction of PQR migration. unc-73 mutations had a similar effect. Thus, our results are consistent with previous results indicating that UNC-73 does not control orientation of Q-cell polarization as do UNC-40 and DPY-19, but rather that UNC-73 is necessary for Q cells to polarize in either orientation (HONIGBERG and KENYON 2000). The Racs along with UNC-73 might affect the ability of the Q cells to undergo polarization or the ability of the Q cells or their descendants to migrate along their normal routes.

mig-2(lq13), ced-10(n1993lq20), and ced-10(n3246) are gain-of-function rac alleles:

We describe the isolation and characterization of gain-of-function alleles of ced-10 and mig-2. Unlike null alleles of each locus, ced-10(n1993lq20), ced-10(n3246), and mig-2(lq13) alone affect axon pathfinding and neuronal migration (Figure 7A). Furthermore, transgenic expression of ced-10 and mig-2 harboring these mutants led to PDE axon pathfinding defects (Figure 7, B–E). ced-10(n1993lq20) was isolated in a hypomorphic ced-10(n1993) loss-of-function background. Possibly, the lq20 mutation could be expressed because of the residual ced-10 activity in ced-10(n1993), and in a wild-type background the lq20 mutation might have a much stronger effect. Indeed, transgenic expression of ced-10(lq20) without the n1993 lesion caused strong defects in PDE axon pathfinding. These gain-of-function rac mutations also affected AQR and PQR migrations along their normal routes but had no effect on direction of PQR migration.

Constitutive activation of each of the three Racs CED-10, MIG-2, and RAC-2 causes ectopic formation of neurites and ectopic lamellipodia and filopodia formation (STRUCKHOFF and LUNDQUIST 2003), whereas loss of rac function (in double mutants) caused axon guidance defects and ectopic neurite formation. The effects of ced-10(n3246 and n1993lq20) and mig-2(lq13) mutations in mutants and in transgenic animals resembled rac double loss-of-function mutants: axon guidance errors and ectopic neurites with few if any ectopic lamellipodia and filopodia formation (Figure 7). ced-10(lq20), ced-10(n3246), and mig-2(lq13) might be dominant-negative rather than a constitutively active alleles, although further experiments will be required to assess the specific effect of each allele. Interestingly, the mutations affect residues in or near the switch regions of the Rac GTPases (Figure 1D). The switch 1 and switch 2 regions mediate interaction of Rac GTPases with their upstream regulators, GTP exchange factors of the DH-GEFs (ERICKSON and CERIONE 2004). In fact, mig-2(lq13) affected a residue conserved in human Rac1 that makes a van der Waal's contact with the DH domain of the Rac GEF Tiam1 (WORTHYLAKE et al. 2000). We have shown previously that the DH-GEF UNC-73 Trio acts in each of the Rac pathways in axon pathfinding and cell migration, and mutations that affect the Rac GEF activity of UNC-73 Trio cause axon and cell migration defects similar to rac double mutants. The n3246, lq20, and lq13 mutations might cause aberrant interactions of the Rac molecules with UNC-73 Trio that perturb UNC-73 Trio activity in all three Rac pathways. Indeed, dominant-negative T17N Rho GTPase mutants have been shown to interfere with the activity of DH-GEFs, which might be required for multiple Rho GTPase signaling pathways (DEBRECENI et al. 2004).

MIG-15 might act in each of the three Rac pathways in axon pathfinding:

mig-15 encodes a molecule similar to NIK kinase and has previously been shown to affect axon pathfinding (POINAT et al. 2002). RNAi of rac genes enhanced mig-15 mutations, suggesting that mig-15 and the racs act together to control axon pathfinding. On the basis of phenotypic and molecular analyses of the mig-15 mutations presented here, rh80 and rh148 likely retain some mig-15 activity. rh148 was a missense mutation in the ATP binding pocket of the kinase domain. The residual activity of mig-15(rh148) could be due to other domains of the protein (e.g., the CNH domain) or could be due to partial impairment of the kinase domain. rh80 caused a premature stop in the C-terminal CNH domain. Possibly, the residual activity of mig-15(rh80) is due to the activity of the kinase alone (i.e., rh80 might specifically eliminate the function of the CNH domain but not the kinase domain). rh326 caused a premature stop in the middle of the reading frame and might be a null mig-15 allele.

Here we provide genetic evidence that is consistent with the notion that MIG-15 acts in each of the three Rac pathways in axon pathfinding: mutations in ced-10, mig-2, and rac-2 enhance a weak mig-15 allele to resemble the stronger mig-15(rh80 and rh326) alleles; and no rac; mig-15 double mutant is strikingly more affected than the strong mig-15(rh326M+) mutant, although ced-10(n1993) might slightly enhance mig-15(rh326M+), raising the possibility that ced-10 and mig-15 act in parallel pathways (Figure 5B). MIG-15 physically associates with the INA-1 -integrin cytoplasmic tail and genetically interacts with INA-1 in axon pathfinding (POINAT et al. 2002). Possibly, MIG-15 acts downstream of Racs and affects INA-1 function in response to Rac signaling. Alternatively, INA-1 and MIG-15 might act upstream of Racs and control Rac activity in response to integrin signaling during axon pathfinding.

ced-10 null mutants and ced-10(n1993); mig-2(mu28) doubles display synthetic lethality with a Gex phenotype characteristic of defects in cell movements during gastrulation. In contrast, mig-15 mutants alone and in combination with mig-2(mu28) and ced-10(n1993) did not display the Gex phenotype, although ced-10(n1993); mig-15(rh80) animals were larval lethal. While MIG-15 acts with the Racs in axon pathfinding, it might not act with the Racs in gastrulation cell movements. However, MIG-15 might have a role in this process that is obscured by redundant gene function.

The Drosophila NIK kinase Misshapen is also involved in cytoskeletal signaling during axon pathfinding (ERICKSON and CERIONE 1993; RUAN et al. 1999; SU et al. 2000). Furthermore, vertebrate NIK interacts with the SH2-SH3 adapter protein Nck (SU et al. 1997), which is a component of a protein complex regulated by Rac activity (EDEN et al. 2002; INNOCENTI et al. 2004). This complex contains the Arp2/3 activator WAVE/Scar, which induces Arp2/3-dependent actin nucleation in response to Rac signaling (SUETSUGU et al. 2001). Furthermore, Drosophila Misshapen interacts with the Nck-like molecule Dreadlocks in Drosophila axon pathfinding (RUAN et al. 1999). In C. elegans, MIG-15 and the Racs might act via a similar signaling mechanism in axon pathfinding.

MIG-15 and MIG-2 control direction of PQR neuronal migration:

mig-15 mutations caused strong anterior misplacement of the QL descendant PQR (Figure 5C), indicating that PQR and/or its precursor QL migrated anteriorly rather than posteriorly. The antennapedia-like homeodomain transcription factor MAB-5 is a key determinant in direction of Q-cell migration: loss of MAB-5 in QL and daughters causes anterior migration rather than the normal posterior; and ectopic expression of MAB-5 in QR and daughters causes posterior migration rather than normal anterior (SALSER and KENYON 1992). At least two events in Q-cell development are required for MAB-5 expression in QL: an initial orientation of QL polarity and migration to the posterior controlled by UNC-40 and DPY-19; and a later posteriorly derived Wnt signal that induces MAB-5 expression in QL and not QR (HONIGBERG and KENYON 2000).

Our results suggest that MIG-15 acts upstream of MAB-5 in QL. The gain-of-function mab-5(e1751) mutation causes expression of MAB-5 in both QL and QR independent of Wnt signaling, and mig-15(rh148); mab-5(e1751gf) double mutants resembled mab-5(e1751gf) alone (posterior PQR and posterior AQR). MIG-15 might act in QL downstream of the Wnt signal, or MIG-15 might act in initial Q cell polarity orientation with UNC-40 and DPY-19. Alternatively, MIG-15 might act nonautonomously (e.g., MIG-15 might be required for an inducing cell to secrete a Wnt signal). In Drosophila, the MIG-15-like protein Misshapen acts downstream of the Wnt receptor Frizzled and the Wnt pathway molecule Disheveled in planar cell polarity (PARICIO et al. 1999).

MIG-15 likely has additional roles in Q-cell development, as both anteriorly directed AQR and PQR often fail to migrate to their normal locations, a defect rarely seen in mab-5, unc-40, and dpy-19 mutants. The AQR and PQR migration defects seen in mig-15 mutants might be due to a defect in cell polarity rather than to a defect in the ability of the cells to migrate. Consistent with this idea, the migration defects of mig-15 mutants were not enhanced by ced-10, rac-2, or unc-34. Possibly, CED-10, RAC-2, and UNC-34 Ena affect the ability of migrating cells to generate protrusive structures necessary for cell migration (e.g., lamellipodia and filopodia) whereas MIG-15 might be involved in the generation and/or maintenance of cell polarity during migration.

Our data suggest that MIG-2 Rac is involved in both MIG-15–dependent cell polarity and CED-10/RAC-2/UNC-34–dependent cell migration, as mig-2(mu28) enhanced the PQR direction of migration defects of mig-15(rh148) and synergized with ced-10(n1993), rac-2(RNAi), and unc-34 for AQR and PQR migration along their normal routes (Figures 3 and 5). MIG-2 has been previously implicated in execution of Q-cell migration and axon pathfinding (ZIPKIN et al. 1997; LUNDQUIST et al. 2001) as well as regulating direction of migration of the ray 1 cell in the C. elegans male tail in response to semaphorin/plexin signaling (DALPE et al. 2004). Differing levels of Rac activity (including MIG-2) apparently determine whether semaphorin/plexin signaling in ray 1 is attractive (high Rac activity) or repulsive (low Rac activity). Rac activity in ray 1 might involve guided cell and growth cone migration rather than cell polarity per se (DALPE et al. 2004). Interestingly, the constitutively-active mig-2(rh17) mutation, which presumably causes higher-than-normal levels of MIG-2 activity, affected the ability of Q cells and daughters to execute migrations, similar to mig-2 loss of function, but did not affect direction of Q-cell migration. Possibly, MIG-2 activity in Q-cell direction of migration is independent of the ability of MIG-2 to cycle between GTP- and GDP-bound forms.

Our data indicate that MIG-15 participates in multiple morphogenetic processes (axon pathfinding and cell polarity), similar to Drosophila Misshapen. In fact, Misshapen acts with distinct signaling networks in the control of axon pathfinding and dorsal closure: axon pathfinding requires Misshapen interaction with the Nck-like protein Dreadlocks, whereas dorsal closure does not (SU et al. 2000). Our data suggest that MIG-15 also acts with distinct signaling complexes in different events: MIG-15 acts with the three Rac proteins CED-10, MIG-2, and RAC-2 in axon pathfinding but not in cell migration, and MIG-15 acts with the Rac GTPase MIG-2 in PQR direction of migration, which does not apparently involve CED-10 or RAC-2 (Figure 11).

UNC-34 Ena acts in a parallel, partially overlapping pathway with Rac in axon pathfinding and neuronal migration:

In an unbiased screen for new mutations that were synthetic lethal with ced-10(n1993), we identified a new allele of the unc-34 gene, which encodes the C. elegans Enabled molecule (WITHEE et al. 2004). Enabled family members have been broadly implicated in actin organization and cellular morphogenesis, including axon pathfinding (KRAUSE et al. 2003). Furthermore, unc-34 Ena has been previously implicated in axon pathfinding (GITAI et al. 2003; WITHEE et al. 2004). We found that null unc-34 alleles, which were viable and fertile, displayed synthetic lethality with ced-10(n1993), and that unc-34; ced-10 embryos arrested with the Gex phenotype characteristic of defects in cell movements during gastrulation (SOTO et al. 2002). Thus, UNC-34 and CED-10 might have partially overlapping roles in gastrulation. In contrast, unc-34; mig-2 doubles were viable and fertile, indicating that UNC-34 and MIG-2 might act in the same pathway in gastrulation or that they both have minor effects on gastrulation compared to CED-10.

We found that unc-34(lq17) was a weak hypomorphic allele that had few axon defects on its own but that synergized with ced-10, mig-2, and rac-2 for axon pathfinding and AQR/PQR migration along their normal routes (Figure 9A). Previous studies showed that unc-34 and ced-10 act redundantly in AVM axon pathfinding (GITAI et al. 2003). The lq17 mutation resulted in the use of a cryptic splice donor site in the last unc-34 intron (Figure 8, B and C). The polypeptide produced from this transcript is predicted to be missing block C of the EVH2 domain, which is required for Ena tetramerization (HAFFNER et al. 1995; BACHMANN et al. 1999; ZIMMERMANN et al. 2002). Thus, lq17 might affect the ability of UNC-34 Ena to tetramerize but might leave other functions of the molecule intact. These data suggest that Ena tetramerization is important for the function of the molecule in vivo.

unc-34 did not affect direction of PQR migration. Rather, UNC-34 Ena might act in parallel to the Racs and UNC-73 in execution of AQR and PQR migrations along their normal routes. Null alleles of unc-34, e951, and gm104 also synergized with ced-10, mig-2, and rac-2 for this phenotype. Our data suggest that UNC-34 Ena defines a parallel pathway with partially overlapping function with Rac signaling in axon pathfinding and neuronal migration. In accordance with this model, we found that unc-34(lq17) enhanced the axon pathfinding defects of the weak mig-15(rh148) allele to a severity higher than the putative null mig-15(rh326M+) or any mig-15; rac double mutant combination (Figure 5B).

That the Racs and UNC-34 Ena act in a partially overlapping manner in axon pathfinding and neuronal migration suggest that they affect the same process during axon pathfinding. Possibly, both the Racs and UNC-34 Ena drive the formation of protrusive structures (lamellipodia and/or filopodia) in the axonal growth cone and leading edge of a migrating cell in response to guidance and outgrowth signals. Indeed, Enabled controls filopodia formation in neurons in response to axon guidance signals (LEBRAND et al. 2004), and Rac activity induces both lamellipodia and filopodia in C. elegans neurons (STRUCKHOFF and LUNDQUIST 2003). The actin-binding protein UNC-115 abLIM controls lamellipodia and filopodia formation in response to Rac signaling (STRUCKHOFF and LUNDQUIST 2003; YANG and LUNDQUIST 2005), and our results indicate that UNC-34 Ena and UNC-115 act in a partially overlapping manner in axon pathfinding.

The data presented here are consistent with the models presented in Figure 11. In axon pathfinding (Figure 11A), three Rac molecules act redundantly, the NIK kinase MIG-15 might act in each Rac pathway, and the UNC-34 Enabled molecule defines a parallel pathway that acts in a partially overlapping manner with Rac/MIG-15. In AQR/PQR neuronal migration (Figure 11B), the Racs and UNC-34 Ena display similar relationships as in axon pathfinding, although MIG-15 and UNC-115 are not involved in this event. Finally, MIG-2 Rac might act with MIG-15, upstream of MAB-5, in PQR direction of migration in parallel to an unidentified factor. CED-10 Rac, RAC-2 Rac, UNC-73 Trio, UNC-34 Ena, and UNC-115 abLIM are not involved in PQR polarity. In most genotypes described here, many PDE axons are still correctly guided to the VNC, and AQR and PQR are often in their normal locations. In the most severely affected genotype, mig-15(rh148); unc-34(lq17), nearly 10% of PDE axons are still correctly guided to the VNC. These observations suggest that there could be additional parallel pathways that control axon pathfinding and neuronal migration in a partially overlapping manner with Rac/MIG-15 and UNC-34 Enabled.

The authors thank E. Struckhoff and S. Mariano for technical assistance, M. Prochaska for assistance with mig-15 allele sequencing, M. Herman for critical reading of the manuscript, members of the Lundquist lab for helpful discussions, G. Garriga for unc-34 strains, S. Mitani and the National Bioresource Project for the Experimental Animal "Nematode C. elegans" for ced-10(tm597), and the Caenorhabditis Genetics Center, sponsored by the National Center for Research Resources, for C. elegans strains. This work was supported by grants from the National Institutes of Health (NIH) (NS40945) and National Science Foundation (IOB93192) to E.A.L., a grant from the National Science Foundation (EPS236913) in support of M.A.S., and NIH grant P30 HD02528.

ANDERSON, P., 1995 Mutagenesis. Methods Cell Biol. 48: 31–58.[Medline]

BACHMANN, C., L. FISCHER, U. WALTER and M. REINHARD, 1999 The EVH2 domain of the vasodilator-stimulated phosphoprotein mediates tetramerization, F-actin binding, and actin bundle formation. J. Biol. Chem. 274: 23549–23557.[Abstract/Free Full Text]

BRENNER, S., 1974 The genetics of Caenorhabditis elegans. Genetics 77: 71–94.[Abstract/Free Full Text]

COLLET, J., C. A. SPIKE, E. A. LUNDQUIST, J. E. SHAW and R. K. HERMAN, 1998 Analysis of osm-6, a gene that affects sensory cilium structure and sensory neuron function in Caenorhabditis elegans. Genetics 148: 187–200.[Abstract/Free Full Text]

DALPE, G., L. W. ZHANG, H. ZHENG and J. G. CULOTTI, 2004 Conversion of cell movement responses to Semaphorin-1 and Plexin-1 from attraction to repulsion by lowered levels of specific RAC GTPases in C. elegans. Development 131: 2073–2088.[Abstract/Free Full Text]

DEBRECENI, B., Y. GAO, F. GUO, K. ZHU, B. JIA et al., 2004 Mechanisms of guanine nucleotide exchange and Rac-mediated signaling revealed by a dominant negative trio mutant. J. Biol. Chem. 279: 3777–3786.[Abstract/Free Full Text]

DICKSON, B. J., 2001 Rho GTPases in growth cone guidance. Curr. Opin. Neurobiol. 11: 103–110.[CrossRef][Medline]

DICKSON, B. J., 2002 Molecular mechanisms of axon guidance. Science 298: 1959–1964.[Abstract/Free Full Text]

EDEN, S., R. ROHATGI, A. V. PODTELEJNIKOV, M. MANN and M. W. KIRSCHNER, 2002 Mechanism of regulation of WAVE1-induced actin nucleation by Rac1 and Nck. Nature 418: 790–793.[CrossRef][Medline]

ERICKSON, J. W., and R. A. CERIONE, 1993 Regulation of the cGMP phosphodiesterase in bovine rod outer segments. Use of resonance energy transfer to distinguish between associative and dissociative activation mechanisms. J. Biol. Chem. 268: 3328–3333.[Abstract/Free Full Text]

ERICKSON, J. W., and R. A. CERIONE, 2004 Structural elements, mechanism, and evolutionary convergence of Rho protein-guanine nucleotide exchange factor complexes. Biochemistry 43: 837–842.[Medline]

GALLO, G., and P. C. LETOURNEAU, 2004 Regulation of growth cone actin filaments by guidance cues. J. Neurobiol. 58: 92–102.[CrossRef][Medline]

GITAI, Z., T. W. YU, E. A. LUNDQUIST, M. TESSIER-LAVIGNE and C. I. BARGMANN, 2003 The netrin receptor UNC-40/DCC stimulates axon attraction and outgrowth through Enabled and, in parallel, Rac and UNC-115/AbLIM. Neuron 37: 53–65.[CrossRef][Medline]

HAFFNER, C., T. JARCHAU, M. REINHARD, J. HOPPE, S. M. LOHMANN et al., 1995 Molecular cloning, structural analysis and functional expression of the proline-rich focal adhesion and microfilament-associated protein VASP. EMBO J. 14: 19–27.[Medline]

HALL, A., 1998 Rho GTPases and the actin cytoskeleton. Science 279: 509–514.[Abstract/Free Full Text]

HATTEN, M. E., 2002 New directions in neuronal migration. Science 297: 1660–1663.[Abstract/Free Full Text]

HERMAN, M., 2001 C. elegans POP-1/TCF functions in a canonical Wnt pathway that controls cell migration and in a noncanonical Wnt pathway that controls cell polarity. Development 128: 581–590.[Abstract]

HONIGBERG, L., and C. KENYON, 2000 Establishment of left/right asymmetry in neuroblast migration by UNC-40/DCC, UNC-73/Trio and DPY-19 proteins in C. elegans. Development 127: 4655–4668.[Abstract]

HUBER, A. B., A. L. KOLODKIN, D. D. GINTY and J. F. CLOUTIER, 2003 Signaling at the growth cone: ligand-receptor complexes and the control of axon growth and guidance. Annu. Rev. Neurosci. 26: 509–563.[CrossRef][Medline]

INNOCENTI, M., A. ZUCCONI, A. DISANZA, E. FRITTOLI, L. B. ARECES et al., 2004 Abi1 is essential for the formation and activation of a WAVE2 signalling complex. Nat. Cell Biol. 6: 319–327.[CrossRef][Medline]

JIN, Y., E. JORGENSEN, E. HARTWIEG and H. R. HORVITZ, 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.[Abstract/Free Full Text]

KELLY, W. G., S. XU, M. K. MONTGOMERY and A. FIRE, 1997 Distinct requirements for somatic and germline expression of a generally expressed Caenorhabditis elegans gene. Genetics 146: 227–238.[Abstract]

KRAUSE, M., 1995 Techniques for analyzing transcription and translation. Methods Cell Biol. 48: 513–529.[Medline]

KRAUSE, M., E. W. DENT, J. E. BEAR, J. J. LOUREIRO and F. B. GERTLER, 2003 Ena/VASP proteins: regulators of the actin cytoskeleton and cell migration. Annu. Rev. Cell Dev. Biol. 19: 541–564.[CrossRef][Medline]

KRIEGSTEIN, A. R., and S. C. NOCTOR, 2004 Patterns of neuronal migration in the embryonic cortex. Trends Neurosci. 27: 392–399.[CrossRef][Medline]

LEBRAND, C., E. W. DENT, G. A. STRASSER, L. M. LANIER, M. KRAUSE et al., 2004 Critical role of Ena/VASP proteins for filopodia formation in neurons and in function downstream of netrin-1. Neuron 42: 37–49.[CrossRef][Medline]

LOWE, E. D., M. E. NOBLE, V. T. SKAMNAKI, N. G. OIKONOMAKOS, D. J. OWEN et al., 1997 The crystal structure of a phosphorylase kinase peptide substrate complex: kinase substrate recognition. EMBO J. 16: 6646–6658.[CrossRef][Medline]

LUNDQUIST, E. A., 2003 Rac proteins and the control of axon development. Curr. Opin. Neurobiol. 13: 384–390.[CrossRef][Medline]

LUNDQUIST, E. A., R. K. HERMAN, J. E. SHAW and C. I. BARGMANN, 1998 UNC-115, a conserved protein with predicted LIM and actin-binding domains, mediates axon guidance in C. elegans. Neuron 21: 385–392.[CrossRef][Medline]

LUNDQUIST, E. A., P. W. REDDIEN, E. HARTWIEG, H. R. HORVITZ and C. I. BARGMANN, 2001 Three C. elegans Rac proteins and several alternative Rac regulators control axon guidance, cell migration and apoptotic cell phagocytosis. Development 128: 4475–4488.[Medline]

LUO, L., 2000 Rho GTPases in neuronal morphogenesis. Nat. Rev. Neurosci. 1: 173–180.[Medline]

MADAULE, P., T. FURUYASHIKI, M. EDA, H. BITO, T. ISHIZAKI et al., 2000 Citron, a Rho target that affects contractility during cytokinesis. Microsc. Res. Tech. 49: 123–126.[CrossRef][Medline]

MADAULE, P., T. FURUYASHIKI, T. REID, T. ISHIZAKI, G. WATANABE et al., 1995 A novel partner for the GTP-bound forms of rho and rac. FEBS Lett. 377: 243–248.[CrossRef][Medline]

MALOOF, J. N., J. WHANGBO, J. M. HARRIS, G. D. JONGEWARD and C. KENYON, 1999 A Wnt signaling pathway controls hox gene expression and neuroblast migration in C. elegans. Development 126: 37–49.[Abstract]

MARCHLER-BAUER, A., J. B. ANDERSON, P. F. CHERUKURI, C. DEWEESE-SCOTT, L. Y. GEER et al., 2005 CDD: a Conserved Domain Database for protein classification. Nucleic Acids Res. 33: D192–D196.[Abstract/Free Full Text]

MELLO, C., and A. FIRE, 1995 DNA transformation. Methods Cell Biol. 48: 451–482.[Medline]

NIEFIND, K., B. GUERRA, L. A. PINNA, O. G. ISSINGER and D. SCHOMBURG, 1998 Crystal structure of the catalytic subunit of protein kinase CK2 from Zea mays at 2.1 A resolution. EMBO J. 17: 2451–2462.[CrossRef][Medline]

PARICIO, N., F. FEIGUIN, M. BOUTROS, S. EATON and M. MLODZIK, 1999 The Drosophila STE20-like kinase misshapen is required downstream of the Frizzled receptor in planar polarity signaling. EMBO J. 18: 4669–4678.[CrossRef][Medline]

POINAT, P., A. DE ARCANGELIS, S. SOOKHAREEA, X. ZHU, E. M. HEDGECOCK et al., 2002 A conserved interaction between beta1 integrin/PAT-3 and Nck-interacting kinase/MIG-15 that mediates commissural axon navigation in C. elegans. Curr. Biol. 12: 622–631.[CrossRef][Medline]

REDDIEN, P. W., and H. R. HORVITZ, 2000 CED-2/CrkII and CED-10/Rac control phagocytosis and cell migration in Caenorhabditis elegans. Nat. Cell Biol. 2: 131–136.[CrossRef][Medline]

RUAN, W., P. PANG and Y. RAO, 1999 The SH2/SH3 adaptor protein dock interacts with the Ste20-like kinase misshapen in controlling growth cone motility. Neuron 24: 595–605.[CrossRef][Medline]

SALSER, S. J., and C. KENYON, 1992 Activation of a C. elegans Antennapedia homologue in migrating cells controls their direction of migration. Nature 355: 255–258.[CrossRef][Medline]

SAMBROOK, J., E.F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SOTO, M. C., H. QADOTA, K. KASUYA, M. INOUE, D. TSUBOI et al., 2002 The GEX-2 and GEX-3 proteins are required for tissue morphogenesis and cell migrations in C.elegans. Genes Dev. 16: 620–632.[Abstract/Free Full Text]

STRUCKHOFF, E. C., and E. A. LUNDQUIST, 2003 The actin-binding protein UNC-115 is an effector of Rac signaling during axon pathfinding in C. elegans. Development 130: 693–704.[Abstract/Free Full Text]

SU, Y. C., J. HAN, S. XU, M. COBB and E. Y. SKOLNIK, 1997 NIK is a new Ste20-related kinase that binds NCK and MEKK1 and activates the SAPK/JNK cascade via a conserved regulatory domain. EMBO J. 16: 1279–1290.[CrossRef][Medline]

SU, Y. C., C. MAUREL-ZAFFRAN, J. E. TREISMAN and E. Y. SKOLNIK, 2000 The Ste20 kinase misshapen regulates both photoreceptor axon targeting and dorsal closure, acting downstream of distinct signals. Mol. Cell. Biol. 20: 4736–4744.[Abstract/Free Full Text]

SUETSUGU, S., H. MIKI, H. YAMAGUCHI, T. OBINATA and T. TAKENAWA, 2001 Enhancement of branching efficiency by the actin filament-binding activity of N-WASP/WAVE2. J. Cell Sci. 114: 4533–4542.[Medline]

SULSTON, J., and J. HODGKIN, 1988 Methods, pp. 587–606 in The Nematode Caenorhabditis elegans, edited by W. B. WOOD. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SULSTON, J. E., and H. R. HORVITZ, 1977 Post-embryonic cell lineages of the nematode, Caenorhabditis elegans. Dev. Biol. 56: 110–156.[CrossRef][Medline]

TESSIER-LAVIGNE, M., and C. S. GOODMAN, 1996 The molecular biology of axon guidance. Science 274: 1123–1133.[Abstract/Free Full Text]

TREISMAN, J. E., N. ITO and G. M. RUBIN, 1997 Misshapen encodes a protein kinase involved in cell shape control in Drosophila. Gene 186: 119–125.[CrossRef][Medline]

WHANGBO, J., and C. KENYON, 1999 A Wnt signaling system that specifies two patterns of cell migration in C. elegans. Mol. Cell 4: 851–858.[CrossRef][Medline]

WHITE, J. G., E. SOUTHGATE, J. N. THOMSON and S. BRENNER, 1986 The structure of the nervous system of the nematode Caenorhabditis elegans. Philos. Trans. R. Soc. Lond. 314: 1–340.[Abstract/Free Full Text]

WITHEE, J., B. GALLIGAN, N. HAWKINS and G. GARRIGA, 2004 Caenorhabditis elegans WASP and Ena/VASP proteins play compensatory roles in morphogenesis and neuronal cell migration. Genetics 167: 1165–1176.[Abstract/Free Full Text]

WORTHYLAKE, D. K., K. L. ROSSMAN and J. SONDEK, 2000 Crystal structure of Rac1 in complex with the guanine nucleotide exchange region of Tiam1. Nature 408: 682–688.[CrossRef][Medline]

WU, Y. C., T. W. CHENG, M. C. LEE and N. Y. WENG, 2002 Distinct rac activation pathways control Caenorhabditis elegans cell migration and axon outgrowth. Dev. Biol. 250: 145–155.[CrossRef][Medline]

YANG, Y., and E. A. LUNDQUIST, 2005 The actin-binding protein UNC-115/abLIM controls formation of lamellipodia and filopodia and neuronal morphogenesis in Caenorhabditis elegans. Mol. Cell. Biol. 25: 5158–5170.[Abstract/Free Full Text]

ZIMMERMANN, J., D. LABUDDE, T. JARCHAU, U. WALTER, H. OSCHKINAT et al., 2002 Relaxation, equilibrium oligomerization, and molecular symmetry of the VASP (336–380) EVH2 tetramer. Biochemistry 41: 11143–11151.[CrossRef][Medline]

ZIPKIN, I. D., R. M. KINDT and C. J. KENYON, 1997 Role of a new Rho family member in cell migration and axon guidance in C. elegans. Cell 90: 883–894.[CrossRef][Medline]

Communicating editor: K. KEMPHUES

This article has been cited by other articles:

				C. J. Locke, B. B. Kautu, K. P. Berry, S. K. Lee, K. A. Caldwell, and G. A. Caldwell Pharmacogenetic Analysis Reveals a Post-Developmental Role for Rac GTPases in Caenorhabditis elegans GABAergic Neurotransmission Genetics, December 1, 2009; 183(4): 1357 - 1372. [Abstract] [Full Text] [PDF]

				P. J. Vanderzalm, A. Pandey, M. E. Hurwitz, L. Bloom, H. R. Horvitz, and G. Garriga C. elegans CARMIL negatively regulates UNC-73/Trio function during neuronal development Development, April 1, 2009; 136(7): 1201 - 1210. [Abstract] [Full Text] [PDF]

				K. L. Schmidt, N. Marcus-Gueret, A. Adeleye, J. Webber, D. Baillie, and E. G. Stringham The cell migration molecule UNC-53/NAV2 is linked to the ARP2/3 complex by ABI-1 Development, February 15, 2009; 136(4): 563 - 574. [Abstract] [Full Text] [PDF]

				M. A. Shakir, K. Jiang, E. C. Struckhoff, R. S. Demarco, F. B. Patel, M. C. Soto, and E. A. Lundquist The Arp2/3 Activators WAVE and WASP Have Distinct Genetic Interactions With Rac GTPases in Caenorhabditis elegans Axon Guidance Genetics, August 1, 2008; 179(4): 1957 - 1971. [Abstract] [Full Text] [PDF]

				J. Gates, J. P. Mahaffey, S. L. Rogers, M. Emerson, E. M. Rogers, S. L. Sottile, D. Van Vactor, F. B. Gertler, and M. Peifer Enabled plays key roles in embryonic epithelial morphogenesis in Drosophila Development, June 1, 2007; 134(11): 2027 - 2039. [Abstract] [Full Text] [PDF]

				Y. Yang, J. Lu, J. Rovnak, S. L. Quackenbush, and E. A. Lundquist SWAN-1, a Caenorhabditis elegans WD Repeat Protein of the AN11 Family, Is a Negative Regulator of Rac GTPase Function Genetics, December 1, 2006; 174(4): 1917 - 1932. [Abstract] [Full Text] [PDF]

International Access Link

Online ISSN: 1943-2361  Print ISSN: 0016-6731
Copyright 2009 by the Genetics Society of America
phone: 412-268-1812   fax: 412-268-1813   email: genetics-gsa{at}andrew.cmu.edu