The Rac GTP Exchange Factor TIAM-1 Acts with CDC-42 and the Guidance Receptor UNC-40/DCC in Neuronal Protrusion and Axon Guidance

The mechanisms linking guidance receptors to cytoskeletal dynamics in the growth cone during axon extension remain mysterious. The Rho-family GTPases Rac and CDC-42 are key regulators of growth cone lamellipodia and filopodia formation, yet little is understood about how these molecules interact in growth cone outgrowth or how the activities of these molecules are regulated in distinct contexts. UNC-73/Trio is a well-characterized Rac GTP exchange factor in Caenorhabditis elegans axon pathfinding, yet UNC-73 does not control CED-10/Rac downstream of UNC-6/Netrin in attractive axon guidance. Here we show that C. elegans TIAM-1 is a Rac-specific GEF that links CDC-42 and Rac signaling in lamellipodia and filopodia formation downstream of UNC-40/DCC. We also show that TIAM-1 acts with UNC-40/DCC in axon guidance. Our results indicate that a CDC-42/TIAM-1/Rac GTPase signaling pathway drives lamellipodia and filopodia formation downstream of the UNC-40/DCC guidance receptor, a novel set of interactions between these molecules. Furthermore, we show that TIAM-1 acts with UNC-40/DCC in axon guidance, suggesting that TIAM-1 might regulate growth cone protrusion via Rac GTPases in response to UNC-40/DCC. Our results also suggest that Rac GTPase activity is controlled by different GEFs in distinct axon guidance contexts, explaining how Rac GTPases can specifically control multiple cellular functions.


Introduction
In a developing nervous system, axonal growth happens via extension and migration of the growth cone, an actin-rich neuronal structure from which filopodial and lamellipodial structures emanate in order to sense the surrounding environment for appropriate guidance cues [1][2][3][4][5].Guidance receptors such as Deleted in Colorectal Carcinoma (DCC) and Roundabout (ROBO) are present in the leading edge of the growth cone and are activated upon ligand binding (such as Netrin and Slit, respectively) [6][7][8][9][10].Attractive and repellent guidance systems guide the growth cone toward its final target, presumably by affecting the protrusiveness of filopodial and lamellipodial structures in the growth cone [8,11].The proper control of the actin cytoskeleton is crucial for the development, maintenance and dynamics of these filopodia and lamellipodia.Failure in the proper extension of axons during development is associated with several neurological disorders, including autism spectrum disorders and dyslexia [12,13].
The intracellular signaling mechanisms linking guidance receptors to cytoskeletal alterations are an intense area of study, and many remain to be elucidated.Rac GTPases, members of the Rho subfamily of small GTPases (which also includes Rho and Cdc-42), have been well established as modulators of the actin cytoskeleton [14] and axon guidance [15].Rac GTPases and their downstream effectors modulate filopodia and lamellipodia morphology and dynamics during axon pathfinding in vitro in mammalian cell culture, as well as in vivo in the fruit fly Drosophila melanogaster and the nematode Caenorhabditis elegans [15][16][17][18].Rac GTPases act on the actin cytoskeleton via diverse downstream effectors such as the actin-filament nucleator Arp2/3 complex, the severing protein Cofilin, the Pak and LIM kinases, and the actin binding protein UNC-115/abLIM [19][20][21][22][23].In C. elegans growth cone development and axon pathfinding, the two redundant Rac GTPases CED-10/Rac1 and MIG-2/RhoG act via two distinct pathways to activate the Arp2/3 complex: CED-10/Rac1 acts with WVE-1/Wave, while MIG-2/RhoG acts with WSP-1/Wasp [24].CED-10/Rac1 is also known to act with the actin-binding protein UNC-115/abLIM via the Receptor for Activated C Kinase (RACK-1) [25].
Although much is known about the downstream effectors of the Rac GTPases in axon development, less is known about pathways upstream of Rac GTPases that control their activity in vivo.Rac GTPases switch between an inactive, GDP-bound state, and an active, GTP-bound state.Rac GTPase activation can be reached with the aid of a Guanine-nucleotide Exchange Factor (GEF), which catalyzes the exchange of GTP for GDP [26].The C. elegans genome encodes 19 predicted Dbl-homology (DH) GEF proteins (Wormbase).UNC-73/Trio is a well-characterized GEF for MIG-2/RhoG and CED-10/Rac1 in C. elegans axon pathfinding [27][28][29][30].A mutation that specifically disrupts specifically the RacGEF activity of UNC-73/Trio (rh40) disrupts axon pathfinding [30].However, pathfinding defects of mig-2; ced-10 rac double mutants are more severe than unc-73(rh40) [28], indicating that additional GEFs act with Rac GTPases to control axon pathfinding.
Here we show that TIAM-1 is a Rac GEF similar to mammalian T-cell lymphoma Invasion and Metastasis Factor 1 (Tiam1) [31] and Drosophila Still Life (Sif) [32].Previous results suggest that the UNC-73/Trio Rac GEF does not act with CED-10/Rac1 downstream of UNC-40/DCC in neuronal protrusion [9].Here we show that the TIAM-1 Rac GEF acts downstream of the UNC-40/DCC netrin receptor in neuronal protrusion, demonstrating the modular use of Rac GEFs in the control of distinct Rac-mediated events.We report that TIAM-1 functions mechanistically as a Rac-specific GEF and that TIAM-1 controls Rac GTPases in lamellipodia and filopodia formation.The roles of Rac and CDC-42 GTPases in lamellipodia and filopodia formation have been documented extensively, yet it is unclear how these molecules interact in this process.Here we show that TIAM-1 acts downstream of the CDC-42 GTPase, and might link CDC-42 and Rac signaling in a linear pathway in lamellipodia and filopodia formation downstream of UNC-40/DCC.Sif and Tiam1 have been previously implicated in axon development [33,34].We present evidence that TIAM-1 acts in the UNC-40/ DCC pathway in axon guidance.Together, these data lead us to speculate that TIAM-1 controls growth cone lamellipodia and filopodia formation during axon guidance downstream of the UNC-40/DCC guidance receptor.

C11D9.1 and T21E12.2 encode a GEF similar to mammalian Tiam1 and Drosophila Still Life
The C11D9.1 gene is predicted to encode a molecule with DH and Pleckstrin homology (PH) domains most similar to mammalian T-cell lymphoma Invasion and Metastasis Factor 1 (Tiam1) and Drosophila Still Life (Sif), but lacks an apparent EVH1 domain and PDZ domain found in Tiam1 and Sif [33,35].Three cDNAs representing C11D9.1 were obtained and sequenced (courtesy of Yuji Kohara).One of the cDNAs contained sequences complementary not only to C11D9.1, but also to the predicted upstream gene T21E12.2(Figures S1 and S2), indicating that C11D9.1 and T21E12.2 are a single gene.
A search of the additional predicted sequence encoded by T21E12.2 using BLASTP and CCD [36,37] did not identify regions similar to the EVH1 domain and PDZ domain of Tiam1 and Sif.However, two regions in the C. elegans molecule were conserved in the related nematodes Ascaris Suum and Brugia malayi, and these regions from A. suum and B. Malayi were recognized by BLASTP as being similar to the EVH1 and PDZ domain regions of Tiam1 and Sif (Figure S3).While these might not represent bona fide EVH1 and PDZ domains, this evolutionary comparison revealed regions in T21E12.2/C11D9.1 similar to the EVH1 domains and PDZ domains of Tiam1 and Sif.Mammalian Tiam1 contains an N-terminal myristoylation site (Myr) [38,39], and a potential Myr site is found at the N-terminus of the predicted TIAM-1 molecule (Figure S2).We refer to the T21E12.2/C11D9.1 locus collectively as tiam-1.
The predicted molecules encoded by the tiam-1 cDNAs are shown in Figure S1.In other organisms, the DH-PH domain of Tiam1/Sif is a Rac-specific GEF [34,40].Both alleles of tiam-1 used here (ok772 and tm1556) remove at least part of the DH domain (Figures S1 and S2), and are predicted to cause frame shifts.Thus, ok772 and tm1556 are likely to be strong loss of function mutations.

Author Summary
Axons extend great distances to make precise synaptic connections in the developing nervous system.Axons are guided to their targets by the growth cone, a dynamic structure at the axon distal tip that senses extracellular cues telling the axon where to go.In response to guidance cues, growth cones alter their shape and motility resulting in outgrowth and turning.The cytoskeleton (actin and microtubules) underlies growth cone motility and guidance.The signaling mechanisms linking guidance receptors to cytoskeletal change remain mysterious.Here, we define a new signaling mechanism downstream of the guidance receptor UNC-40/DCC involving the GTPases CDC-42 and Rac, which have long been known to control growth cone protrusion.We show that CDC-42 and Rac act in a linear pathway in axon guidance; CDC-42 acts upstream of the GTPase regulatory molecule TIAM-1, which is a GTP exchange factor specific for Rac and which activates Rac signaling.We also show that TIAM-1 acts with UNC-40/DCC signaling in protrusion and axon guidance.Our results imply that Rac GTPase function in axon guidance is complex and that distinct GEFs (TIAM-1 and UNC-73/Trio) might control Rac GTPases in different aspects of axon guidance.
These studies indicate that TIAM-1 displays a genetic interaction with the Rac GTPases CED-10 and MIG-2 and the UNC-73/Trio GEF in PDE axon pathfinding.That tiam-1 enhances mig-2 and ced-10 could indicate that TIAM-1 acts in parallel to these GTPases or that TIAM-1 acts in both of the redundant MIG-2 and CED-10 pathways in pathfinding.Studies below indicate that TIAM-1 acts with both MIG-2 and CED-10.However, the mig-2; ced-10 double mutant phenotype is much stronger than tiam-1, suggesting that another molecule might regulate MIG-2 and CED-1 in parallel to TIAM-1, possibly another GEF.

A tiam-1promoter::cfp transgene is expressed in neurons
To determine where and when the tiam-1 gene might by expressed, we studied expression of a tiam-1 transcriptional reporter transgene.We generated by PCR a 4793-bp region upstream of T21E12.2 (the entire upstream region to the next predicted gene, T21E12.3)(Figure 2A), and this putative tiam-1 promoter region was placed upstream of the cyan fluorescent protein (cfp) coding region.Expression of the tiam-1promoter::cfp construct was primarily in neurons (Figure 2B-2D).Most if not all neurons expressed the construct, including neurons in the head and tail, the ventral cord commissural motorneurons, the mechanosensory neurons (ALMs, PLMs, AVM, PVM) and the CAN, PDE, and PVD neurons.tiam-1promoter::cfp expression in neurons began in embryos and lasted throughout adulthood.No obvious expression outside of the nervous system was noted.However, tiam-1 might be endogenously expressed in other tissues, as some transcriptional elements might not be present in this transgene.

The TIAM-1 and UNC-73 Rac GEF domains induce ectopic protrusion from neurons
To test the roles of TIAM-1 and UNC-73 as potential Rac GEFs in vivo, we constructed transgenes that express constitutivelyactivated forms of the TIAM-1[DHPH] and UNC-73[DHPH1] Rac GEF domains in the PDE neurons.We based these experiments on previous results with Drosophila Trio (UNC-73 in C. elegans), in which a myristoylation sequence was added to the Rac-specific DHPH domain of Trio (Figure 4A).This construct produced a DHPH domain that was localized to cellular membranes via myristoylation [42].Flies expressing this construct had retinal axon pathfinding defects due to unregulated activity of the DHPH domain.Mammalian Tiam1 has an N-terminal myristoylation site and is normally localized to membranes [38,39], and TIAM-1 in C. elegans has a potential N-terminal myristoylation site (Figure S2).Furthermore, an N-terminal truncation of mammalian Tiam1, which does not affect the DHPH domains, activates Rac more strongly than full length Tiam1 [31,43], suggesting that a deleted portion of Tiam1 might normally inhibit GEF activity of the molecule.
Based on these results, we created transgenes consisting of the TIAM-1 DHPH domain and both UNC-73 DHPH domains (Figure 4A) fused in frame to an N-terminal myristoylation sequence and to a C-terminal GFP (see Methods).Expression of these molecules in the PDE neurons was driven by the osm-6 promoter and was confirmed by DHPH::GFP fluorescence in the PDE and other osm-6-expressing neurons (data not shown).The MYR::DHPH::GFP fusion proteins displayed a punctate, peripheral accumulation pattern consistent with localization to cellular membranes as seen in other MYR::GFP fusion proteins [44].
In a wild-type background, the Rac-specific MYR::UNC-73[DHPH1] induced ectopic protrusions that emanated from the cell body and axons of PDEs (Figure 4B).MYR::TIAM-1[DHPH] also promoted ectopic lamellipodial and filopodial protrusions (Figure 4C).Axon branching and defects in pathfinding were also observed in each case (Figure 4B and 4C).We generated multiple independent extrachromosomal transgenic lines with each construct, and in every case saw robust lamellipodial protrusions (Figure S5).In Drosophila, the Rho GEF domain of Trio did not cause retinal axon pathfinding defects [42].We found that the Rho-specific MYR::UNC-73[DHPH2] of UNC-73 caused a low level of ectopic PDE protrusions in multiple transgenic extrachromosomal lines generated (Figure S5 and Figure 4D), significantly lower than the UNC-73 Rac-specific domain and MYR::TIAM-1DHPH (p#0.001),consistent with the Drosophila studies.
To test the requirement of Rac GEF activity in ectopic protrusion, we introduced point mutations into the transgenes that abolish the Rac GEF activities of UNC-73 and TIAM-1 (the unc-73(rh40) S1216F mutation [30] and the T516F mutation in TIAM-1 used in Figure 3).The ''GEF-dead'' constructs were expressed as determined by GFP fluorescence.Multiple transgenic extrachromosomal lines were produced for each, and none promoted high levels of ectopic lamellipodial protrusions (Figure S5 and Figure 4D) (p#0.001compared to non-mutant transgenes).These results suggest that Rac GEF activity was required for robust ectopic lamellipodial and filopodial induction.These results also indicate that the MYR::[DHPH] transgenes likely produced constitutively activated GEF molecules rather than dominant negative molecules.We assayed the effects of MYR::TIAM-1[DHPH] in the tiam-1(tm1556) background and observed ectopic protrusions similar to the wild-type background (data not shown), consistent with the idea that MYR::TIAM-1DHPH represents a constitutively-activated rather than a dominant negative molecule.

TIAM-1 and Rac GTPases co-localize in C. elegans neurons
To determine the subcellular localization of TIAM-1 in neurons, we fused the longest isoform A tiam-1 cDNA (yk730h9 in Figure S1 and S2) with mCherry at the C-terminus, and placed this construct under the control of the unc-25 promoter (for VD/ DD motor neurons) and the osm-6 promoter (for PDE expression).The yk730h9 cDNA contained two errors relative to the reference gene models that led to frame shifts.These were repaired before proceeding (see Methods and Figure S2).This construct likely produced a functional TIAM-1::mCherry mole- cule, because it was required for ectopic protrusion induced by activated CDC-42(G12V) in the PDE neurons (see below and Figure 6), and it induced ectopic protrusions on its own (Figure 5A  and 5B).
In the PDE and VD/DD neurons, TIAM-1::mCherry localized to puncta primarily at the periphery of the cell bodies, consistent with plasma membrane localization, as well as in the axons (Figure S6).Consistent with this observation, mammalian Tiam1 associates with membranes [38].We noted that full-length TIAM-1::mCherry produced a low percentage (5%) of PDE and VD/DD neurons with ectopic neurites and protrusions (Figure 5A and 5B), similar to the TIAM-1 MYR::DHPH construct, suggesting that this molecule is active and drives ectopic protrusions.This is also consistent with the low level of PDE axon pathfinding defects observed upon tiam-1(+) expression in Figure 1A.TIAM-1::mCherry accumulated strongly in these ectopic protrusions (Figure 5A and 5B).TIAM-1::mCherry also accumulated in the growth cones of developing VD axons (Figure 5C-5E), consistent with a role of TIAM-1 in the growth cone during outgrowth.
We constructed mig-2/RhoG and ced-10/Rac1 transgenes tagged with gfp driven by the unc-25 promoter.MIG-2 contains an Nterminal myristoylation site and a C-terminal CAAX prenylation site.We placed gfp immediately after the Myr sequence in frame with mig-2, such that both myristoylated and prenylated forms of MIG-2 could be tagged with GFP.CED-10 was N-terminally tagged with GFP.GFP::CED-10 and GFP::MIG-2 accumulated at the cell periphery as previously described [28,45] (data not shown).GFP::MIG-2 and TIAM-1::mCherry co-localized to the periphery of the ectopic protrusions of VD/DD neurons induced by expression of these molecules (Figure 5F-5H).GFP::CED-10 also co-localized with TIAM-1::mCherry (data not shown).Similar to TIAM-1::mCherry, GFP::CED-10 and GFP::MIG-2 were also found in the VD growth cones (data not shown).While TIAM-1::mCherry and GFP::MIG-2 and GFP::CED-10 showed colocalization, it was not completely overlapping co-localization (compare GFP in Figure 5F to mCherry in Figure 5G).This suggests that TIAM-1 and the GTPases MIG-2 and CED-10 might interact in specific regions in the cell (e.g.ectopic protrusions at the cell periphery).As expected, expression of GFP::CED-10 in fibroblasts caused marginal lamellipodial ruffling, a phenotype canonically associated with expression of vertebrate Rac1 in these cells (Figure 5I and Figure S7).GFP::CED-10 localized to intracellular regions around the nucleus that might represent membranous organelles, as well as to the plasma membrane, particularly in lamellipodial regions (Figure 5I and Figure S7A).Expression of mCherry::TIAM-1 did not affect cell shape as did GFP::CED-10, but did accumulate at the cell margins and intracellular membranous organelles (Figure S7B).These patterns are likely to represent CED-10 and TIAM-1 specific localization, as mCherry and GFP expression alone did not show these patterns (Figure S7C and S7D).

TIAM-1 might act in the UNC-6/Netrin pathway in ventral axon guidance
We have shown evidence that the Rac GEF TIAM-1 regulates the Rac GTPases CED-10 and MIG-2 in response to CDC-42 signaling.We next sought to determine which dorsal-ventral axon guidance ligand/receptor pathway might be upstream of TIAM-1.Distinct parallel ventral guidance pathways control dorsal ventral guidance in C. elegans.UNC-6/Netrin is a secreted ligand expressed in the ventral side of the animal that attracts axons that express its receptor UNC-40/DCC [49].SLT-1/Slit is expressed primarily in the dorsal side and is thought to act as a repellent for ventrally-guided axons that express its receptor SAX-3/Robo [50].Ephrins have also been involved in directing proper axon pathfinding [51].VAB-1/EphR is the only C. elegans Ephrin receptor [52] and has been shown to work in parallel with UNC-6/Netrin and SLT-1/Slit in axon ventral guidance [53].

TIAM-1 might act downstream of UNC-40/DCC
Previous studies indicated that the Rac GTPase CED-10 acts downstream of the UNC-6/Netrin receptor UNC-40/DCC in the AVM neuron [9].A myristoylated form of the cytoplasmic tail of the Netrin receptor UNC-40 missing the extracellular and transmembrane domains induced axon defects and ectopic lamellipodia and filopodia in the AVM neurons due to constitutive activation of UNC-40 signaling (see diagram in Figure 8A) [9].
We assayed AVM axon guidance in single and double loss-offunction mutants involving unc-6, unc-40, and tiam-1 (Figure 8E).unc-6(ev400) and unc-40(n324) both displayed defective AVM ventral guidance as expected, but unc-6(ev400) was significantly stronger than unc-40(n324) (41% versus 14%; p,0.001).Both are predicted null mutations, suggesting that UNC-40 might not be the only receptor for UNC-6 in AVM ventral guidance.AVM defects in unc-6(ev400) mutants were scored previously by another group [54], and fewer defects were noted in this study (,25% compared to 41% in our hands).We scored AVM guidance as defective if the axon failed to reach the ventral cord or if the axon wandered more than 45 degrees away from straight ventrally.When we consider only axons that failed to reach the ventral nerve cord, we see 27% defects in unc-6(ev400) (data not shown), similar to those previously described [54].
We used unc-40(n324) and unc-40(RNAi) and the previous study used unc-40(e1430) [54], raising the possibility that the stronger enhancement of unc-6(ev400) that we report is due to different unc-40 alleles used.
In sum, these genetic interactions suggest that UNC-6 and UNC-40 act partially in parallel and also suggest that UNC-6 might not be the only ligand for UNC-40/DCC in AVM ventral guidance.They are also consistent with TIAM-1 acting in the UNC-40 pathway in AVM ventral axon guidance.That unc-40 has a more severe phenotype than tiam-1 suggests that additional molecules might act in parallel to TIAM-1 downstream of UNC-40.

Discussion
In recent years, attention has focused on how guidance receptors engage cytoplasmic signaling mechanisms to control growth cone outgrowth and turning during axon guidance.In this report we present evidence of a signaling pathway (Figure 9A) involving the previously uncharacterized C. elegans molecule TIAM-1, which acts specifically as a GEF for Rac GTPases and interacts with them in axon guidance and neuronal protrusion.We show that TIAM-1 regulates neuronal protrusion downstream of the CDC-42 GTPase and upstream of the two redundant Rac GTPases MIG-2/RhoG and CED-10/Rac1.Genetic interactions of these molecules suggest that that they act in a common pathway in axon pathfinding.Furthermore, we provide evidence that TIAM-1 acts with the UNC-6/Netrin receptor UNC-40/DCC in axon guidance and downstream of UNC-40/DCC in neuronal protrusion.Previous studies have implicated Rac GTPases and the Rac GEF Trio as players in Netrin signaling and interaction with the Netrin receptor DCC [55][56][57].Here we present evidence that the Rac GEF TIAM-1, along with CDC-42, CED-10/Rac1, and MIG-2/RhoG, acts downstream of UNC-40/DCC in the formation of neuronal protrusions, which might reflect a role in protrusive activity in the growth cone during outgrowth.

TIAM-1 links CDC-42 and Rac GTPases in neuronal protrusion downstream of UNC-40/DCC
GTP exchange assays indicated that TIAM-1 is a Rac-specific GEF (Figure 3).The activated Rac-specific MYR::DHPH domains of TIAM-1 and UNC-73 induced ectopic protrusions from PDE neurons that resembled lamellipodia and filopodia.Protrusions were dependent upon a functional DH GEF domain in these molecules as well as on the presence of MIG-2/RhoG and CED-10/Rac1.TIAM-1 was required for ectopic protrusions induced by CDC-42 and activated MYR::UNC-40; and CDC-42 was required for MYR::UNC-40-induced protrusions.Together, these data suggest a pathway of neuronal lamellipodial and filopodial protrusion wherein TIAM-1 links CDC-42 and Rac GTPase signaling downstream of UNC-40/DCC.
A caveat with the studies using DHPH Rac GEF domains of TIAM-1 and UNC-73 is that the GEF domains were not in the context of the full-length molecules, and thus these MYR::DHPH constructs might not reflect the normal cellular roles of these molecules in neurons.These studies do indicate, however, that TIAM-1 and UNC-73 have the ability to act as Rac GEFs in vivo, whether or not their normal role is in protrusion.However, we found that transgenic expression of full-length TIAM-1::mCherry also induced ectopic protrusions in the PDE neurons similar to but at a much lower level than the MYR::DHPH construct, suggesting that regulation of protrusion is a normal cellular role of TIAM-1.This is consistent with tiam-1 suppression of protrusion caused by CDC-42(G12V) and MYR::UNC-40.In contrast, unc-73 suppressed neither of these, suggesting that UNC-73/Trio might not normally regulate protrusive events downstream of UNC-40.We speculate that UNC-73 might control Rac activity upstream of guidance receptors to control their cell surface trafficking (Figure 9B and see below).
TIAM-1 displayed genetic interactions with the Rac GTPases in PDE axon guidance.tiam-1 mutations enhanced mutations in mig-2 and ced-10, which have redundant roles in axon guidance [15,24,28,41].Thus, TIAM-1 might act with both molecules.This idea is supported by gain-of-function studies showing that both MIG-2 and CED-10 are required for the ectopic protrusions induced by the activated MYR::DHPH domain of TIAM-1.TIAM-1 might act in parallel to another molecule, possibly another GEF, that regulates Rac GTPases in axon guidance, as the mig-2; ced-10 double mutant phenotype is more severe than tiam-1.

Genetic interactions suggest that TIAM-1 acts in the UNC-40/DCC axon guidance pathway
The UNC-6/Netrin ligand is expressed in ventral cells, and growth cones are attracted to or repelled from the ventral source In the absence of UNC-6/Netrin ligand, the extracellular domains of UNC-40 prevent dimerization of the cytoplasmic regions and activation of receptor signal transduction.A myristoylated version of the cytoplasmic domain lacking the transmembrane domain and extracellular domains is believed to constitutively dimerize and is constitutively active, resulting in ectopic protrusion from the AVM neuron.Ig = immunoglobulin domain; FNIII = fibronectin type II domain; TM = transmembrane domain; P1-P3 = conserved proline-rich domains; MYR = covalent myristoyl group.(B) A wild-type AVM neuron visualized with soluble GFP from a mec-4::gfp transgene, with an arrow pointing to the unbranched ventrally-directed axon.The dashed line indicates the ventral nerve cord.(C) An AVM neuron with expression of unc-86p::myr::unc-40, a transgene that drives a myristoylated version of the cytoplasmic domain of UNC-40 in AVM.Fluorescence is soluble GFP expressed from mec-4::gfp to highlight cell shape.A large arrow points to an ectopic lamellipodium, and a smaller arrow points to a misguided and branched AVM axon.In all micrographs, dorsal is up and anterior is to the left.The scale bar represents 5 mm for B and C. (D) A graph charting the percentage of AVM axons with ectopic lamellipodial protrusions (X axis) in different genotypes (Y axis).Unless otherwise noted, all backgrounds are wild type.[myr::unc-40] represents animals harboring a transgene that expresses the myristoylated UNC-40 cytoplasmic domain under the mec-7 promoter (mec-7p::myr::unc-40).M+ indicates wild-type maternal gene function.Alone, none of the single mutants tiam-1(ok772), cdc-42(gk388M+), or unc-73(rh40) displayed ectopic protrusions.(E) A graph plotting the percentage of animals with AVM ventral axon guidance defects in different genotypes.Reported here are defects in AVM guidance that include failure of the axon tom reach the ventral nerve cord and axon wandering.If we just consider failure to reach the ventral nerve cord, similar significant results are found (unc-40 and tiam-1 enhance unc-6(ev400), and tiam-1 does not enhance unc-40).At least 100 AVM neurons were scored for each genotype, and p value significance was determined using Fisher's Exact analysis.Error bars represent 26 standard error of the proportion in both directions.Dashed lines indicate comparisons between genotypes not marked with a p value to those marked with p values.NS = not significantly different.doi:10.1371/journal.pgen.1002665.g008[58].A series of loss-of-function genetic interactions suggest that CDC-42, TIAM-1, MIG-2/RhoG, and CED-10/Rac1 act in the UNC-40/DCC axon guidance pathway.In PDE axon guidance (Figure 8D), tiam-1 enhanced slt-1/Slit and vab-1/EphR but not unc-6/Netrin.In AVM axon guidance (Figure 8E), tiam-1 enhanced unc-6, suggesting cell specific differences in the roles of these molecules in axon guidance.However, tiam-1 did not enhance unc-40 in AVM axon guidance, suggesting that TIAM-1 might act with UNC-40.Consistent with this idea, unc-40 also enhanced unc-6 in AVM guidance, similar to tiam-1.However, tiam-1 mutations were less severe than unc-40, suggesting that other molecules act in parallel to TIAM-1 in UNC-40 signaling.Indeed, the TRIM molecule MADD-2 acts with UNC-6 and UNC-40 signaling in AVM ventral guidance, but is not required for the protrusive effects of MYR::UNC-40 [54].We find that TIAM-1 is required for MYR::UNC-40 protrusive effects but not axon guidance defects caused by MYR::UNC-40, suggesting that distinct pathways might act downstream of UNC-40 to mediate distinct roles of UNC-40 in axon guidance (i.e.protrusive versus nonprotrusive events).
These results have implications about UNC-6/Netrin signaling in axon guidance.First, unc-6 AVM axon guidance defects were more severe than unc-40, suggesting that UNC-6 signals through other receptors in addition to UNC-40/DCC.Second, that unc-40 enhances unc-6 suggests that these molecules act at least in part in parallel, and suggest that UNC-6 might not be the only ligand for UNC-40/DCC in AVM ventral guidance.Possible candidates for such interactions might be the SLT-1/Slit ligand and its receptor SAX-3/Robo, which have been shown to functionally interact with Netrin and DCC [59].In any case, genetic interactions suggest that TIAM-1 acts in the UNC-40 pathway in AVM axon guidance.

The relationship of lamellipodial and filopodial protrusion to axon guidance
Our results demonstrate that TIAM-1 is required for axon guidance and that TIAM-1 regulates lamellipodial and filopodial protrusions in neurons.Growth cones, including those in C. elegans, display dynamic lamellipodial and filopodial protrusions [19,60].Our recent findings indicate that the Netrin receptors UNC-40/ DCC and UNC-5 modulate growth cone lamellipodial and filopodial protrusion that correlates with attractive versus repulsive responses to UNC-6/Netrin (i.e.UNC-6/Netrin inhibits protrusion in repelled growth cones via UNC-5, and stimulates protrusion in attracted growth cones via UNC-40/DCC) [61].We speculate that TIAM-1 normally stimulates growth cone lamellipodial and filopodial protrusion in response to UNC-40/ DCC in UNC-6/Netrin attractive axon guidance (Figure 9).While it is possible that TIAM-1 controls axon guidance by a mechanism distinct from lamellipodia and filopodia formation, this model is consistent with our data presented here.

Distinct roles of TIAM-1 and UNC-73/Trio in axon guidance
Our results and those of others [9] suggest that UNC-73/Trio acts in parallel to UNC-6/Netrin signaling in PDE axon guidance, and might not act downstream of UNC-40/DCC in AVM  [55,62].In Drosophila, Trio and the Abelson kinase Abl interact with the Netrin receptor Frazzled (the fly equivalent of UNC-40 and DCC) to control guidance events involving Netrin [56].Trio was required for the ectopic midline crossing of axons caused by a hybrid chimeric Frazzled-Robo receptor composed of the extracellular and transmembrane domains of Robo and the cytoplasmic domain of Frazzled.In vertebrates, Trio is required for Netrin mediated activation of Rac1 in neurons, and is required for neurite extension in response to Netrin [57].
How might our results fit with these known roles of Trio in Netrin signaling?Cell-specific and context-specific differences in Trio function downstream of guidance receptors are possible explanations.In other words, Trio might act with UNC-40/DCC in some neurons and not in others, or might display distinct, cellspecific interactions with guidance receptors.
An alternative but not mutually exclusive explanation, consistent with previous results, is that Trio and TIAM-1 affect distinct aspects of Rac function in axon guidance.It is becoming apparent that the role of Rac GTPases in axon pathfinding is complex.Rac GTPases clearly regulate cytoskeletal events in the growth cone in response to guidance cues.For example, in C. elegans, Rac GTPases engage the actin-nucleating Arp2/3 complex and the actin-binding protein UNC-115/abLIM to promote filopodial extension [19,24,41].It is also clear that Rac GTPases (e.g.MIG-2/RhoG) control the accumulation of guidance receptors (UNC-40/DCC) at the cell surface via trafficking [55,62].Thus, Rac GTPases might act both upstream and downstream of guidance receptors (Figure 9B).This might also explain in part why the axon guidance defects of mig-2; ced-10 doubles are more severe than tiam-1 alone.Rac GTPases might affect multiple aspects of axon guidance, whereas TIAM-1 might specifically regulate them in downstream cytoskeletal effects.
Trio might be a GEF that controls Rac activity in receptor trafficking upstream of UNC-40/DCC, possibly in a feedback mechanism (Figure 9B

Rho GTPases and Rho GEFs in axon guidance
Many studies have shown that Rac and CDC-42 GTPases control lamellipodia and filopodia formation, yet it has been unclear how these molecules interact to regulate protrusion.Our results indicate that they act in a linear pathway, with CDC-42 controlling Rac activity mediated by TIAM-1.However, the ced-10; mig-2 double mutant is much more severe than tiam-1, so it is likely that another molecule regulates MIG-2 and CED-10 in parallel to TIAM-1 in axon pathfinding, possibly another GEF.
Rho-family GTPases control many cellular functions, and it has been assumed that upstream regulators confer specificity to Rho GTPase activity.Indeed, genomes encode many GTP exchange factors for Rho GTPases (e.g. the C. elegans genome encodes 19 DH GEF proteins).Our results provide experimental evidence for this idea, as we have shown that the TIAM-1 GEF but not the UNC-73/Trio GEF controls Rac function downstream of UNC-6/Netrin signaling in protrusion.UNC-73/Trio might act upstream of UNC-40/DCC and regulate its trafficking [55,62] (Figure 9B).Future studies will be aimed at understanding how UNC-40/DCC links to CDC-42 to regulate TIAM-1, as well as understanding the roles of Rho GTPases and GEFs in repulsive axon guidance in response to UNC-6/Netrin.
All micrographs were obtained on a Leica DMRE microscope with a Qimaging Rolera MGi EMCCD camera or a Qimaging Retiga CCD camera.Openlab and IPlab software were used to obtain images.PDE axon pathfinding.The cell bodies of the PDE neurons (PDEL and PDER) are situated in the posterior lateral post-deirid ganglion.PDEs extend an axon ventrally to the ventral nerve cord, which then bifurcates and extends anteriorly and posteriorly in the ventral nerve cord.If the axon failed to reach the ventral nerve cord or wandered beyond a 45u angle from a straight line ventrally from the cell body, it was considered mutant.

Scoring of PDE, VD/DD, and AVM axon defects
VD/DD axon pathfinding.A VD/DD commissural axon that failed to reach the dorsal nerve cord or that wandered laterally before reaching the dorsal nerve cord was considered mutant.The percent of animals with pathfinding defects was noted, and the percentage of defective axons was noted.
AVM axon pathfinding.The AVM cell body is situated laterally on the right side, in the anterior half of the body of the animal.The axon extends ventrally towards the ventral nerve cord, which then extends anteriorly until the first pharyngeal bulb.Similarly to PDE axons, if the axon failed to reach the ventral nerve cord or wandered beyond a 45u angle from a straight line ventrally, it was considered mutant.p-value significance was determined using Fisher Exact Analysis.
Ectopic PDE and AVM protrusions.Ectopic lamellipodial and filopodial protrusions were identified as any thin ruffle-like or finger-like protrusion from the cell body, which is normally smooth and round, or axon of the neurons.

Transgenes for expression in C. elegans
All C. elegans expression plasmids utilized the standard Fire expression vectors with the unc-54 39 UTR (A.Fire, Stanford University).Primer sequences used to amplify C. elegans genomic and cDNA sequences to construct these plasmids are shown in Figure S9.For each transgene, at least three independent transgenic lines were generated, each with similar results.A tiam-1 transcriptional promoter fusion plasmid was generated by amplifying by PCR the tiam-1 upstream promoter region, a 4793bp region upstream of T21E12.2 (the entire upstream region to the next predicted gene, T21E12.3)(Figure 2), and placing it upstream of the cyan fluorescent protein (cfp) coding region.For cell specific expression of full length TIAM-1, the open reading frame of the yk370h9 tiam-1 cDNA was amplified by PCR and fused to mCherry at the 39 end to create a C-terminally tagged TIAM-1::mCherry molecule (Figure 5).tiam-1::mCherry expression plasmids contained the osm-6 promoter (to drive expression in the PDEs) [66] and the unc-25 promoter (VD/DDs) [67] in separate constructs.The yk730h9 cDNA contained errors relative to the reference gene models that resulted in frame shifts.These were repaired in the yk730h9 cDNA by site-directed mutagenesis.A G residue after base 1039 relative to the start codon in Figure S2 was removed, and the A residue at position 2478 was added (see Figure S2).
A gfp::mig-2 fusion plasmid was constructed by placing gfp in frame after the N-terminal myristoylation sequence of mig-2 and before the remaining genomic coding region, allowing for GFP tagging of both myristoylated and prenylated forms of MIG-2 (Figure 5).gfp::ced-10 was constructed by placing gfp at the Nterminus as previously described [48] (Figure 5).gfp::mig-2 and gfp::ced-10 expression was driven by the osm-6 or unc-25 promoter.
A plasmid containing an activated cdc-42(G12V) genomic coding region under the control of the osm-6 promoter was constructed using site directed mutagenesis as described previously for ced-10(G12V) [41].A myr::unc-40(cyto) plasmid under the control of the mec-7 promoter was constructed as described in [9] (Figure 8A).The myr::GEF(DHPH) plasmids with the osm-6 promoter were used as means of overactivating the GEFs as previously described [42] (Figure 4A).Three distinct DHPH GEF domains were fused in frame to an N-terminal myristoylation sequence as described in [44] and to a C-terminal GFP: the DHPH region of TIAM-1 (residues 488-856 relative to the A isoform); the Rac-specific UNC-73 DHPH domain (DHPH1) (residues 1193-1527 relative to the F55C7.7Aisoform); and the Rho-specific UNC-73 DHPH domain (DHPH2) (residues 1795-2129).The coding regions of transgenes were sequenced after PCR to ensure that no mutations had been introduced.

Mant-GTP based exchange assay
We amplified the DHPH domains of TIAM-1 (residues 488-856 relative to isoform A), placed it into the pET-28b vector (Novagen), which has a T7 IPTG-inducible promoter and a Cterminal 6HIS-tag, and transformed this construct into BL21(DE3) Gold competent cells (Agilent Technologies).For the TIAM-1[DH(T516F)PH] construct, we used site directed mutagenesis to make a point mutation in the wild-type construct using the Quikchange kit (Stratagene).We purified TIAM-1[DHPH]::6HIS using Ni-column based chromatography (Qiagen).For the mant-GTP assay, we used the RhoGEF Exchange Assay Biochem Kit (Cytoskeleton, Inc) and followed instructions in the kit.We used 12.5 mM of TIAM-1[DHPH] in these experiments.
N-methylanthraniloyl-GTP (mant-GTP) is a fluorescently labeled nucleotide analog that can be incorporated by small GTPases.Using a Varian Cary Eclipse fluorescent spectrophotometer, we measured the amount of mant-GTP incorporated by the GTPase due to the spectroscopic difference between bound and free mant-GTP (when in the GTPase binding pocket, mant-GTP increases its fluorescence).The GTPase (human Rac1, Cdc-42 or RhoA) was incubated in buffer and mant-GTP for five minutes, after which time solvent (control) or the GEF (TIAM-1[DHPH]) was added.Experiments were recorded for one half hour.

Figure 1 .
Figure 1.TIAM-1 acts with the Rac GTPases in axon pathfinding.A) A graph representing PDE axon pathfinding defects (X axis) in different genotypes (Y axis).At least 100 PDE neurons were scored, and p value significance was determined by Fisher's exact analysis.[tiam-1(+)] indicates a transgene that harbors a wild-type genomic copy of the tiam-1 locus on fosmid clone WRM0633ch01.The error bars represent 26 standard error of the proportion in each direction.M+ indicates that the genotype had wild-type maternal gene function.The inset is a model of the predicted TIAM-1 isoform A molecule (856 amino acid residues) showing the regions of the molecule missing in two deletion alleles.Unless otherwise noted, all backgrounds are wild-type.p value significance as determined by Fisher Exact Analysis is shown.Dashed lines indicate comparisons between genotypes not marked with a p value to those marked with p values.(B-D) Fluorescent micrographs of osm-6::gfp in the PDE neurons of young adults.In all micrographs, dorsal is up and anterior is to the left.The arrows point to axons, which are misguided in mutants in (B) and (D).The ventral nerve cord is represented by a dashed line.The scale bar in (B) represents 5 mm for (B-D).doi:10.1371/journal.pgen.1002665.g001

Figure 2 .
Figure 2. The tiam-1 promoter drives expression in the nervous system.A) A diagram of the tiam-1 promoter::cfp transgene.The tiam-1 promoter included 4793 bp upstream of the tiam-1 A cDNA initiator codon.(B-D) Fluorescent micrographs of young adults harboring tiam-1p::cfp.In all micrographs, dorsal is up and anterior is to the left.B) tiam-1p::cfp expression in ventral cord motor neurons (ventral aspect).C) Expression in the post-deirid ganglion neurons PVD and PVM (lateral aspect) The diagram below depicts the positions of the neurons in the micrograph.D) Expression in the PDE neuron (lateral aspect), with a depiction to the right.The scale bar in (B) represents 10 mm, and the scale bar in (C) represents 2 mm for (C-D).doi:10.1371/journal.pgen.1002665.g002

Figure 3 .
Figure 3.The TIAM-1 DHPH domain is a Rac1-specific GTP exchange factor.Graphs plotting arbitrary fluorescence units (A.U.) over time in a mant-GTP exchange assay (Cytoskeleton, Inc.) (see Methods).Dark, cross-hatched points represent exchange activity of the GTPase after the addition of vehicle without the DHPH domain of TIAM-1.Light, diamond-shaped points represent exchange after the addition of the DHPH domain of TIAM-1 in vehicle.The dashed vertical line represents the time of addition of vehicle or vehicle plus TIAM-1 DHPH.A) Exchange activity on human Rac1.B) Exchange activity on human RhoA.C) Exchange activity of human Cdc-42.D) Exchange activity on human Rac1 of a TIAM-1 DHPH domain with the T516F point mutation predicted to eliminate Rac GEF activity (TIAM-1(DEAD)).Error bars represent standard deviations of three independent experiments which each had consistent results.doi:10.1371/journal.pgen.1002665.g003

Figure 4 .
Figure 4. Ectopic protrusions induced by overactive TIAM-1 and UNC-73 were suppressed by loss-of-function of Rac GTPases.(A) Diagram of membrane-tethered, Rac-specific MYR::DHPH domains of UNC-73 and TIAM-1 inducing ectopic protrusions.(B-C) Fluorescent micrographs of PDE neurons with ectopic lamellipodial and filopodial protrusions (arrows) induced by MYR::DHPH constructs.Fluorescence is from soluble GFP expressed from an osm-6::gfp co-injection marker for cell shape.Compare to the wild-type PDE neurons in Figure 1.In all micrographs, dorsal is up and anterior is to the left.The scale bar in (B) represents 5 mm for (B-C).D) A graph plotting percentage of ectopic lamellipodial protrusions from the PDE neurons of young adult animals harboring transgenes driving the expression of MYR::DHPH domains of UNC-73 and TIAM-1.MYR::TIAM-1[DHPH] contains the TIAM-1 DHPH region (lqIs165).MYR::UNC-73[DHPH1] (lqIs123) contains the Rac specific GEF domain of UNC-73.A second DH GEF domain of UNC-73 is Rho specific (MYR::UNC-73[DHPH2]) (lqIs125).UNC-73[DH(DEAD)PH1] (lqIs134) and TIAM-1[DH(DEAD)PH] (lqIs193) are transgenes that harbor the point mutations that eliminate Rac GEF activity of the DH domains of UNC-73 DH1 and TIAM-1 (S1216F and T516F, respectively).Unless otherwise noted, all backgrounds are wild-type.At least 100 PDE neurons for each genotype were scored, and p value significance was determined using Fisher's exact analysis.Error bars represent 26standard error of the proportion in both directions.The asterisk ''*'' indicates transgenes that have been crossed away from the ced-10 or mig-2 mutations and re-scored in a wild-type background.Dashed lines indicate comparisons between genotypes not marked with a p value to those marked with p values.doi:10.1371/journal.pgen.1002665.g004

Figure 6 .
Figure 6.TIAM-1 acts cell autonomously downstream of CDC-42.(A) Fluorescent micrograph of a PDE neuron from an animal with an integrated cdc-42(G12V) transgene expressed with the osm-6 promoter.An osm-6::gfp marker transgene was included to label the PDE cell body and processes.Dorsal is up and anterior is to the left.An arrow points to an ectopic lamellipodial protrusion.The scale bar in represents 5 mm.(B) A graph charting the percentage of PDE axons with ectopic lamellipodial and filopodial protrusions (X axis) in different genotypes (Y axis).Unless otherwise noted, all backgrounds are wild type.M+ indicates that the genotype had wild-type maternal gene function.[cdc-42(G12V)] indicates a transgene that harbors activated cdc-42(G12V) driven by the osm-6 promoter in the PDEs.[osm-6p::tiam-1(+)::mCherry] indicates a transgene that harbors the corrected tiam-1 cDNA yk730h9 fused in frame to mCherry and driven by the osm-6 promoter in the PDE neurons (see Methods and Figure S2).At least 100 PDE neurons were scored for each genotype, and p value significance was determined using Fisher's Exact analysis.Error bars represent 26 standard error of the proportion in both directions.Dashed lines indicate comparisons between genotypes not marked with a p value to those marked with p values.doi:10.1371/journal.pgen.1002665.g006

Figure 7 .
Figure 7. tiam-1 and cdc-42 interactions with axon guidance mutations.A graph representing PDE axon pathfinding defects (X axis) in different genotypes (Y axis).At least 100 PDE neurons were scored, and p value significance was determined by Fisher's exact analysis.The error bars represent 26 standard error of the proportion in each direction.M+ indicates that the genotype had wild-type maternal gene function.Dashed lines indicate comparisons between genotypes not marked with a p value to those marked with p values.doi:10.1371/journal.pgen.1002665.g007

Figure 8 .
Figure 8. TIAM-1 acts with UNC-6/netrin and UNC-40/DCC in axon guidance and protrusion.(A) A diagram illustrating attractive axon guidance via the UNC-40/DCC receptor.In the absence of UNC-6/Netrin ligand, the extracellular domains of UNC-40 prevent dimerization of the cytoplasmic regions and activation of receptor signal transduction.A myristoylated version of the cytoplasmic domain lacking the transmembrane domain and extracellular domains is believed to constitutively dimerize and is constitutively active, resulting in ectopic protrusion from the AVM neuron.Ig = immunoglobulin domain; FNIII = fibronectin type II domain; TM = transmembrane domain; P1-P3 = conserved proline-rich domains; MYR = covalent myristoyl group.(B) A wild-type AVM neuron visualized with soluble GFP from a mec-4::gfp transgene, with an arrow pointing to the unbranched ventrally-directed axon.The dashed line indicates the ventral nerve cord.(C) An AVM neuron with expression of unc-86p::myr::unc-40, a transgene that drives a myristoylated version of the cytoplasmic domain of UNC-40 in AVM.Fluorescence is soluble GFP expressed from mec-4::gfp to highlight cell shape.A large arrow points to an ectopic lamellipodium, and a smaller arrow points to a misguided and branched AVM axon.In all micrographs, dorsal is up and anterior is to the left.The scale bar represents 5 mm for B and C. (D) A graph charting the percentage of AVM axons with ectopic lamellipodial protrusions (X axis) in different genotypes (Y axis).Unless otherwise noted, all backgrounds are wild type.[myr::unc-40] represents animals harboring a transgene that expresses the myristoylated UNC-40 cytoplasmic domain under the mec-7 promoter (mec-7p::myr::unc-40).M+ indicates wild-type maternal gene function.Alone, none of the single mutants tiam-1(ok772), cdc-42(gk388M+), or unc-73(rh40) displayed ectopic protrusions.(E) A graph plotting the percentage of animals with AVM ventral axon guidance defects in different genotypes.Reported here are defects in AVM guidance that include failure of the axon tom reach the ventral nerve cord and axon wandering.If we just consider failure to reach the ventral nerve cord, similar significant results are found (unc-40 and tiam-1 enhance unc-6(ev400), and tiam-1 does not enhance unc-40).At least 100 AVM neurons were scored for each genotype, and p value significance was determined using Fisher's Exact analysis.Error bars represent 26 standard error of the proportion in both directions.Dashed lines indicate comparisons between genotypes not marked with a p value to those marked with p values.NS = not significantly different.doi:10.1371/journal.pgen.1002665.g008

Figure 9 .
Figure 9.A model of the distinct roles of the Rac GEFs TIAM-1 and UNC-73/Trio in axon guidance.Arrows represent functional or genetic interactions between molecules and do not imply physical interactions.A) TIAM-1 controls protrusive events downstream of UNC-40/DCC.Results presented here suggest that TIAM-1 acts downstream of UNC-40/DCC and CDC-42, and upstream of the Rac GTPases CED-10 and MIG-2, in ectopic lamellipodial and filopodia formation and attractive axon guidance.B) TIAM-1 and UNC-73/Trio might regulate distinct aspects of Rac function in UNC-40/DCC-mediated axon guidance.TIAM-1 might regulate Rac GTPases downstream of UNC-40/DCC involving cytoskeletal protrusive events.UNC-73/Trio might control Rac GTPases upstream of UNC-40/DCC in trafficking of UNC-40/DCC to the membrane [55,62].The dashed arrow indicates a potential feedback loop where by UNC-40/DCC might control its own membrane trafficking via UNC-73.Ig = immunoglobulin I domain; FNIII = fibronectin type III domain; TM = transmembrane domain; P1-P3 = conserved proline-rich domains; EVH1 = EVH1 domain; PDZ = PDZ domain; DH = Dbl-homology domain; PH = Pleckstrin homology domain.doi:10.1371/journal.pgen.1002665.g009 ).In other words, UNC-6/Netrin activation of UNC-40/DCC might alter UNC-40/DCC trafficking to the cell surface, mediated by UNC-73/Trio and Rac.While the studies of Trio and Rac function in flies and vertebrates discussed above [56,57] did not specifically address UNC-40/DCC trafficking, each can be explained by Trio and Rac regulating cell surface accumulation of Frazzled/DCC and the chimeric Frazzled-Robo receptor, similar to UNC-73/Trio and VAB-8L in C. elegans [55,62].In our studies, the myristoylated UNC-40 cytoplasmic domain is likely trafficked to the cell surface by a mechanism that is distinct from the endogenous UNC-40 molecule and might be independent of UNC-73/Trio and Rac-regulated cell surface accumulation, explaining why unc-73 did not suppress MYR::UNC-40.Thus, suppression of MYR::UNC-40 might target those molecules that act downstream of UNC-40/DCC but not those involved in it's trafficking (Figure 9B).Our studies showing that tiam-1 suppressed MYR::UNC-40 indicate that TIAM-1 might regulate Rac GTPases specifically downstream of UNC-40/DCC to mediate cytoskeletal responses.

Fluorescence
microscopy was used to score axon pathfinding defects of fourth larval stage (L4) or young adult hermaphrodite animals expressing a green fluorescent protein transgene for specific cells.To score and visualize PDE axons, the integrated transgenes lqIs2 X and lqIs3 IV [osm-6 promoter::gfp] were used[41,66].To score and visualize VDs and DDs axons, the integrated transgene juIs76 II [unc-25 promoter::gfp] was used[67].To score and visualize AVM axons, the integrated transgene zdIs4 IV [mec-4 promoter::gfp] was used.

Figure S9
Figure S9 Primer sequences used to amplify C. elegans genomic or cDNA fragments to construct expression plasmids.The fragment of genomic DNA or cDNA amplified is underlined.(+) indicates the sense strand primer relative to the gene, and (2) indicates the anti-sense strand primer relative to the gene.Spacers and restriction enzyme sites added to the ends of the primers to aid in cloning are in italics.(PDF)