REPORTS the physiological event underlying the macroscopic signs of oxidative metabolism, all of which occur within the first seconds after onset of induced activity. The spatial confinement of the NADH dip to dendrites further supports our view, since the preferred localization of cytochrome oxidase is in postsynaptic dendritic areas (28). The kinetics of the late astrocyte NADH response as an indicator of enhanced glycolysis are in accord with reports that demonstrated nonoxidative glucose consumption during induced brain activity (1, 29). Note that the temporal resolution of PET or NMR spectroscopy would not have permitted the detection of early oxidative metabolism in these studies. However, one study (30) using a microsensor with significantly higher temporal resolution described a biphasic change in extracellular lactate with an early decrease (duration 10 to 12 s), followed by a longlasting overshoot (peak after ⬃60 s). This result is in full agreement with our model, which is based on our ample resolution in both space and time. Our results confirm that early neuronal oxidative metabolism is the default response to increased neural activity. Only after a significant period (⬃10 s) with depletion of substrates for oxidative metabolism (9, 30) is astrocytic glycolysis activated. Thereby, extracellular lactate might serve as a buffer preventing activation of the astrocyte-neuron lactate shuttle during minor or short-lasting neural activity. The observation that the transient NADH production as an indicator of nonoxidative glycolysis in astrocytes exceeds neuronal NADH consumption and further increases with longer stimulation strengthens this interpretation. Our model integrates existing views of the organization of brain energy metabolism during focal neural activity (1–3, 5) and has direct implications for the design and interpretation of functional neuroimaging studies: The discovery that early oxidative metabolism is entirely neuronal strengthens the motivation for the current search for the initial dip in BOLD-fMRI, and the confinement of glycolysis to astrocytes implies that 18F-fluorodeoxyglucose–PET studies measure glucose uptake into the glial and not the neuronal compartment during focal neural activity. References and Notes

1. P. T. Fox, M. E. Raichle, M. A. Mintun, C. Dence, Science 241, 462 (1988). 2. D. Malonek, A. Grinvald, Science 272, 551 (1996). 3. F. Hyder et al., Proc. Natl. Acad. Sci. U.S.A. 93, 7612 (1996). 4. L. Pellerin, P. J. Magistretti, Proc. Natl. Acad. Sci. U.S.A. 91, 10625 (1994). 5. N. R. Sibson et al., Proc. Natl. Acad. Sci. U.S.A. 95, 316 (1998). 6. A. Schurr, J. J. Miller, R. S. Payne, B. M. Rigor, J. Neurosci. 19, 34 (1999). 7. B. Voutsinos-Porche et al., Neuron 37, 275 (2003).

8. C. P. Chih, P. Lipton, E. L. Roberts Jr., Trends Neurosci. 24, 573 (2001). 9. S. Mangia et al., Neuroscience 118, 7 (2003). 10. D. L. Rothman et al., Philos. Trans. R. Soc. London Ser. B 354, 1165 (1999). 11. W. Denk, J. H. Strickler, W. W. Webb, Science 248, 73 (1990). 12. B. D. Bennett, T. L. Jetton, G. Ying, M. A. Magnuson, D. W. Piston, J. Biol. Chem. 271, 3647 (1996). 13. G. H. Patterson, S. M. Knobel, P. Arkhammar, O. Thastrup, D. W. Piston, Proc. Natl. Acad. Sci. U.S.A. 97, 5203 (2000). 14. B. Chance, G. R. Williams, J. Biol. Chem. 217, 409 (1955). 15. X. Aubert, B. Chance, R. D. Keynes, Philos. Trans. R. Soc. London Ser. B 160, 211 (1964). 16. J. R. Williamson, B. E. Herczeg, H. S. Coles, W. Y. Cheung, J. Biol. Chem. 242, 5119 (1967). 17. B. Chance, H. Baltscheffsky, J. Biol. Chem. 233, 736 (1958). 18. L. K. Klaidman, A. C. Leung, J. D. J. Adams, Anal. Biochem. 228, 312 (1995). 19. D. K. Merill, R. W. Guynn, J. Neurochem. 27, 459 (1976) 20. D. K. Merill, R. W. Guynn, Brain Res. 221, 307 (1981) 21. P. Lipton, Biochem. J. 136, 999 (1973). 22. C. W. Shuttleworth, A. M. Brennan, J. A. Connor, J. Neurosci. 23, 3196 (2003). 23. F. F. Jobsis, M. Oconnor, A. Vitale, H. Vreman, J. Neurophys. 34, 735 (1971).

24. R. Brandes, D. M. Bers, Biophys. J. 71, 1024 (1996). 25. L. Pellerin, P. J. Magistretti, J. Cereb. Blood Flow Metab. 23, 1282 (2003). 26. U. Lindauer et al., Neuroimage 13, 988 (2001). 27. J. K. Thompson, M. R. Peterson, R. D. Freeman, Science 299, 1070 (2003). 28. M. T. Wong-Riley, Trends Neurosci. 12, 94 (1989). 29. J. Prichard et al., Proc. Natl. Acad. Sci. U.S.A. 88, 5829 (1991). 30. Y. B. Hu, G. S. Wilson, J. Neurochem. 69, 1484 (1997). 31. We thank R. S. Balaban, L. Pellerin, and D. Piston for discussions and I. Tullis and A. Heikal for technical advice. K.A.K. thanks R. Love, F. Bergmann, S. Hess, B. T. Hyman, and A. C. Ludolph for support. Dedicated to Keiko Tsuno-Kasischke (†20 October 2003). This research was conducted in the Developmental Resource for Biophysical Imaging Opto Electronics, a NIH-NIBIB resource. Supported by the Deutsche Forschungsgemeinschaft (KA1490/2-1 and 2-2) and the National Institutes of Health (P41-EB001976-16). Supporting Online Material www.sciencemag.org/cgi/content/full/305/5680/99/DC1 Materials and Methods SOM Text Fig. S1 References 5 February 2004; accepted 21 May 2004

Inhibition of Netrin-Mediated Axon Attraction by a Receptor Protein Tyrosine Phosphatase Chieh Chang,1,2 Timothy W. Yu,2 Cornelia I. Bargmann,2* Marc Tessier-Lavigne1*† During axon guidance, the ventral guidance of the Caenorhabditis elegans anterior ventral microtubule axon is controlled by two cues, the UNC-6/netrin attractant recognized by the UNC-40/DCC receptor and the SLT-1/slit repellent recognized by the SAX-3/robo receptor. We show here that loss-of-function mutations in clr-1 enhance netrin-dependent attraction, suppressing ventral guidance defects in slt-1 mutants. clr-1 encodes a transmembrane receptor protein tyrosine phosphatase (RPTP) that functions in AVM to inhibit signaling through the DCC family receptor UNC-40 and its effector, UNC-34/enabled. The known effects of other RPTPs in axon guidance could result from modulation of guidance receptors like UNC-40/DCC. Axons in the developing nervous system respond to attractive and repulsive guidance cues of the netrin, slit, semaphorin, and ephrin families (1–3). The interpretation of a guidance signal as a repellent or an attractant is context-dependent and influenced by the activities of other signaling pathways (4, 5). Thus, the netrin receptor UNC-40 contributes to both axon attraction (acting on its own) and repulsion (in coop1 Department of Biological Sciences, Howard Hughes Medical Institute (HHMI), Stanford University, Stanford, CA 94305, USA. 2Department of Anatomy and Department of Biochemistry and Biophysics, HHMI, University of California San Francisco (UCSF), San Francisco, CA 94143, USA.

*To whom correspondence should be addressed. Email: [email protected] (C.I.B.); [email protected] (M.T.L.) †Present address: Genentech, Incorporated, 1 DNA Way, South San Francisco, CA 94080, USA.

eration with the second netrin receptor, UNC-5, or the slit receptor robo) (6–10). Receptor protein tyrosine phosphatases (RPTPs) are implicated in axon growth and guidance (11, 12). In general, the inputs that regulate RPTPs, as well as their potential targets, are unknown. Phosphatases are presumed to affect axon guidance by antagonizing kinases, and many tyrosine kinases have been implicated in axon outgrowth and guidance (13–19). The molecules that guide AVM axons toward the ventral midline in C. elegans are similar to those that direct commissural neurons toward the floor plate in vertebrate spinal cords. The AVM neuron sends its axon ventrally toward the attractant UNC-6/netrin (Fig. 1, A and B) and away from the dorsal repellent SLT-1. Mutations in either of these signaling systems result in a 30 to 40% penetrant defect in AVM ventral guidance,

www.sciencemag.org SCIENCE VOL 305 2 JULY 2004

103

REPORTS whereas mutations in both systems together result in a near-complete failure of ventral guidance (20). Because there are two sources of guidance information, mutants that potentiate signaling through one guidance pathway might suppress the effects caused by loss of the other pathway. A slt-1(null) strain is therefore a useful genetic background to search for mutations allowing enhanced signaling through the unc-6 – unc-40 pathway. We undertook two parallel approaches to the suppressor screen. First, we generated double mutants between slt-1 and 30 genes implicated in axon guidance or cell migration and examined the resulting AVM phenotypes. Second, we clonally screened a mutagenized slt-1 null mutant (⬃1000 genomes), selecting mutants that decreased the penetrance of its AVM guidance defect. Both approaches identified a single strong suppressor, in each case a mutation in the clr-1 gene. The temperature-sensitive mutation clr1(e1745), tested in the candidate screen, resulted in an almost-complete suppression of the AVM ventral guidance defect of slt-1 null mutants (Fig. 1, C to E). A single mutation from the genetic screen, cy14, yielded a similar suppression of the ventral guidance defect and mapped to chromosome II near the

clr-1 locus. Complementation tests showed that cy14 fails to complement the clr1(e1745) mutant phenotype. Sequencing the clr-1 open reading frame in the cy14 mutant revealed a molecular lesion for cy14 (21), confirming that cy14 represents a previously unknown clr-1 allele. cy14 is a G-to-A transition in the splice acceptor of intron 5 of clr-1 that leads to the use of a cryptic splice acceptor and consequently to an 18-base-pair deletion in exon 6 (fig. S1, A and B). The mutant allele lacks part of the single immunoglobulin (Ig) domain in the extracellular region of CLR-1 (fig. S1C). CLR-1 encodes a transmembrane protein tyrosine phosphatase that antagonizes fibroblast growth factor receptor (FGFR) signaling (22). An alignment of vertebrate RPTP and CLR-1 catalytic domains assigns CLR-1 to the R5 subdivision of RPTP family members (23) (fig. S1D). However, the extracellular region of CLR-1 is more like the leukocyte common antigen-related protein subfamily of vertebrate RPTPs, because it contains an Ig domain and FN III (fibronectin type III) repeats (fig. S1D). These sequence features indicate that CLR-1 can be classified as an RPTP but not as the clear ortholog of a particular vertebrate RPTP.

Fig. 1. clr-1 regulates unc-6/unc-40 –dependent guidance. (A) Schematic diagram of wild-type and mutant AVM axons. D indicates dorsal; V, ventral; A, anterior; P, posterior. (B) SAX-3/robo and UNC-40/DCC guidance receptors in AVM (green) guide axons toward ventral UNC-6/netrin (blue) and away from dorsal SLT-1/slit (red). (C) clr-1 mutation suppresses AVM guidance defects of slt-1 and sax-3 mutations but not unc-6 and unc-40 mutations. Because clr-1 is essential for viability, we used a

104

Several models could explain the suppression of the slt-1 mutant phenotype by the clr-1 mutation. To distinguish between various possibilities, we examined the effects of the clr-1 mutation in mutant backgrounds that disable elements of different guidance pathways. The clr-1 null phenotype is lethal, and clr-1(cy14) is subviable, so these genetic studies were performed with the use of the temperature-sensitive clr-1 allele e1745ts. The clr-1 mutation suppressed the AVM defects of a sax-3 mutant (Fig. 1C), arguing that clr-1 function is sax-3–independent. By contrast, the clr-1 mutation failed to significantly suppress AVM defects of unc-6, unc40, unc-6 slt-1, or unc-40 slt-1 mutants (Fig. 1, C, F, and G). Because AVM guidance defects in unc-6, unc-40, slt-1, or sax-3 mutants are comparable in severity, suppression of slt-1 and sax-3 appears to be pathwayspecific. These results suggest that CLR-1 requires UNC-6 and UNC-40 to affect ventral guidance, consistent with the model that CLR-1 acts on UNC-6/UNC-40 signaling. One way in which CLR-1 could antagonize netrin signaling would be to limit the production or distribution of UNC-6/netrin. To address this possibility, we examined an alternative response to UNC-6. The DA/DB

temperature-sensitive partial loss-of-function clr-1 allele, e1745ts. All other mutations were strong loss-of-function alleles. Strains were grown at 20°C except the sax-3 and the clr-1; sax-3 mutants. Asterisks indicate data significantly different from clr-1(⫹) controls (P ⬍ 0.001). AVM axon trajectories (arrow) labeled by zdIs5[mec-4::gfp] in (D) slt-1, (E) clr-1; slt-1, (F) unc-6, and (G) clr-1; unc-6. Arrowhead, ALM cell body. Anterior is to the left; dorsal, up. Scale bar, 20 ␮m.

2 JULY 2004 VOL 305 SCIENCE www.sciencemag.org

REPORTS motor axons use UNC-6 as a guidance cue, but they are repelled rather than attracted by ventral UNC-6 (Fig. 2A). Repulsion of DA/DB axons uses both UNC-40 and UNC-5 guidance receptors (24). unc-5 mutants have a severe defect in DA/DB dorsal guidance, whereas unc-40 defects in DA/ DB guidance are milder (Fig. 2B). If CLR-1 normally acts to decrease netrin availability, a clr-1 mutation might sup-

press the mild dorsal guidance defects of an unc-40 mutant by making more UNC-6/ netrin available for detection by UNC-5. Contrary to this prediction, the clr-1 mutation resulted in a significant enhancement of unc-40 mutant phenotypes in DA and DB axons (Fig. 2, B to D). This finding suggests a positive role for CLR-1 in netrin repulsion, in contrast with its inhibitory role in netrin attraction. CLR-1 also pro-

Fig. 2. CLR-1 potentiates UNC-5– dependent repulsion from UNC-6/netrin. (A) Schematic diagram of wild-type (wt) and mutant DA/DB motor axon guidance. (B) clr-1 potentiates DA/DB dorsal guidance. unc-5 data are from (34). Asterisks, data significantly different from clr-1(⫹) control (P ⬍ 0.001). DA and DB neurons in L4 stage animals visualized with evIs82[unc-129::gfp] in (C) unc-40 and (D) clr-1; unc-40. DA/DB axons in the dorsal nerve cord (red arrowheads) are misrouted to dorsolateral positions (white arrowheads). White arrows indicate lateral seam cells. Open arrowheads, ventral DA/DB cell bodies. Dorsal is up and anterior at left. Scale bar, 20 ␮m.

motes UNC-6 – dependent dorsal mesodermal cell migrations (25). These genetic results argue against a common effect of CLR-1 on the netrin ligand. By implication, they suggest that, in AVM ventral guidance, CLR-1 normally limits the activity of the UNC-40 receptor. CLR-1 could inhibit UNC-40 by acting either nonautonomously, for example as a transmembrane ligand for UNC-40, or autonomously in the netrin-responsive cell. A clr-1::gfp fusion gene is expressed in many cells whose normal guidance requires unc40, including AVM, HSN, DD, VD, and mesodermal cells, suggesting an autonomous role (fig. S2). To ask where CLR-1 acts to regulate AVM attraction to netrin, we expressed a clr-1 cDNA under the mec-7 promoter, which is expressed only in AVM, ALM, PVM, and PLM neurons. mec-7::clr-1 rescued the effect of the clr-1 mutation in a clr-1; slt-1 double mutant, indicating that CLR-1 acts cell-autonomously in AVM to inhibit UNC-40 activity (Fig. 3A). mec-7::clr-1 did not cause AVM defects by itself (Fig. 3A) or in a slt1(eh15)/⫹ background, indicating that overexpression of clr-1 did not disregulate its activity. DCC signaling in AVM is mediated by two downstream signaling pathways, one involving UNC-34, an enabled (Ena) homolog, and the other CED-10, a Rac guanosine triphosphatase, and UNC-115, an actin-binding protein similar to human abLIM/limatin (26, 27). To address whether either of these pathways might mediate the CLR-1 effect, we disrupted each branch in a slt-1 mutant background, where only unc-40 guidance signaling is active in AVM, and asked whether the clr-1 mutation could still modify the AVM phenotype. The clr-1 mutation significantly suppressed the AVM defects of slt-1; ced-10 and slt-1; unc-115 but not slt-1; unc-34 double mutants (Fig. 3B). Thus the slt-1;

Fig. 3. CLR-1 acts through UNC-34 in AVM. (A) clr-1 acts cell-autonomously in AVM. Animals expressing cyEx[mec-7::clr-1, odr-1::dsred]; zdIs5[mec-4::gfp] were significantly different from sibling animals that spontaneously lost the mec-7::clr-1 transgene (asterisk, P ⬍ 0.001). (B) CLR-1 requires UNC-34 to inhibit UNC-40 signaling. AVM was visualized with zdIs5[mec-4::gfp]. Asterisks indicate data significantly different from clr-1(⫹) controls (P ⬍ 0.001).

www.sciencemag.org SCIENCE VOL 305 2 JULY 2004

105

REPORTS Fig. 4. Model for CLR-1 regulation of netrin attraction. Arrows indicate a positive effect and bars a negative effect. UNC40, UNC-34, or an unknown effector is phosphorylated by an unknown tyrosine kinase during netrin signaling and dephosphorylated by CLR-1 to inhibit signaling. The cytoplasmic domain of CLR-1 produced in vitro can associate with GST:UNC-40 cytoplasmic domain, consistent with a direct interaction of the proteins (33). EGF, epidermal growth factor; PTK, protein tyrosine kinase; PXXP, proline motif. Question marks indicate undefined signaling components.

ced-10 double mutants retain a pathway that is inhibited by CLR-1 and can respond to the mutation, whereas the slt-1; unc-34 double mutants have lost the CLR-1-sensitive signaling pathway. These results therefore suggest that clr-1 exerts its negative effect in netrin attraction through the unc34– dependent pathway. In the case of DA and DB axon dorsal guidance, the clr-1 mutation enhanced unc-40 mutant phenotypes but not those of unc-34 mutants (Fig. 2B), indicating that clr-1 also exerts its positive effect in netrin repulsion through an unc-34– dependent pathway. In contrast to its role in attractive and repulsive axon guidance, clr-1 did not affect an outgrowthpromoting activity of unc-40 (fig. S3). Thus, the RPTP CLR-1 functions as a negative regulator of the netrin attractive guidance pathway, acting with or downstream of the netrin receptor UNC-40/DCC in the UNC-34/enabled pathway. The suppression of ventral guidance defects in slt-1mutants by clr-1 is remarkably complete: Nearly all AVM axons are guided normally in clr-1; slt-1 double mutants. This result indicates that UNC-6/netrin is sufficient for accurate guidance of AVM even without the dorsal repellent SLT-1, provided the CLR-1 negative regulatory influence is removed. Reducing clr-1 activity in a wild-

106

type background did not cause guidance defects, but the null phenotype of clr-1 is lethal, so only nonnull alleles were examined; stronger axon defects might be observed if clr-1 function was eliminated. The rate of axon extension is enhanced in neurons from RPTP␴⫺/⫺ mice (28) and in Xenopus retinal ganglion cell neurons expressing a dominant negative CRYP (RPTP␴) (29), suggesting a conserved role for RPTPs in inhibiting axon growth. Our results suggest a model in which a protein tyrosine kinase is a positive regulator of netrin-mediated attraction and CLR-1 is a negative regulator. Such antagonistic effects might reflect opposite effects on the tyrosine phosphorylation state of UNC-40 or of an UNC-40 effector, perhaps UNC-34/enabled (Fig. 4) (30). The enhancement of UNC-40 –mediated guidance by removal of CLR-1 identifies a mechanism for negatively regulating netrin attraction and suggests that effects of RPTPs in axon guidance might result from regulation of key axon guidance receptors. References and Notes

1. M. Tessier-Lavigne, C. S. Goodman, Science 274, 1123 (1996). 2. T. W. Yu, C. I. Bargmann, Nat. Neurosci. 4, 1169 (2001). 3. B. J. Dickson, Science 298, 1959 (2002).

4. H. J. Song, G. L. Ming, M. M. Poo, Nature 388, 275 (1997). 5. K. Hong, M. Nishiyama, J. Henley, M. Tessier-Lavigne, M. M. Poo, Nature 403, 93 (2000). 6. M. Hamelin, Y. Zhou, M. W. Su, I. M. Scot, J. G. Culotti, Nature 364, 327 (1993). 7. K. Hong et al., Cell 97, 927 (1999). 8. K. Keleman, B. J. Dickson, Neuron 32, 605 (2001). 9. E. Stein, M. Tessier-Lavigne, Science 291, 1928 (2001). 10. T. W. Yu, J. C. Hao, W. Lim, M. Tessier-Lavigne, C. I. Bargmann, Nat. Neurosci. 5, 1147 (2002). 11. C. J. Desai, J. G. Gindhart Jr., L. S. Goldstein, K. Zinn, Cell 84, 599 (1996). 12. D. Van Vactor, Curr. Opin. Cell Biol. 10, 174 (1998). 13. U. Drescher et al., Cell 82, 359 (1995). 14. M. Nakamoto et al., Cell 86, 755 (1996). 15. S. McFarlane, E. Cornel, E. Amaya, C. E. Holt, Neuron 17, 245 (1996). 16. F. S. Walsh, P. Doherty, Annu. Rev. Cell Dev. Biol. 13, 425 (1997). 17. F. B. Gertler, R. L. Bennett, M. J. Clark, F. M. Hoffmann, Cell 58, 103 (1989). 18. H. E. Beggs, P. Soriano, P. F. Maness, J. Cell Biol. 127, 825 (1994). 19. W. R. Morse, J. G. Whitesides 3rd, A. S. LaMantia, P. F. Maness, J. Neurobiol. 36, 53 (1998). 20. J. C. Hao et al., Neuron 32, 25 (2001). 21. Information on materials and methods is available on Science Online. 22. M. Kokel, C. Z. Borland, L. DeLong, H. R. Horvitz, M. J. Stern, Genes Dev. 12, 1425 (1998). 23. J. N. Andersen et al., Mol. Cell. Biol. 21, 7117 (2001). 24. E. M. Hedgecock, J. G. Culotti, D. H. Hall, Neuron 4, 61 (1990). 25. D. C. Merz, G. Alves, T. Kawano, H. Zheng, J. G. Culotti, Dev. Biol. 256, 173 (2003). 26. Z. Gitai, T. W. Yu, E. A. Lundquist, M. Tessier-Lavigne, C. I. Bargmann, Neuron 37, 53 (2003). 27. E. A. Lundquist, R. K. Herman, J. E. Shaw, C. I. Bargmann, Neuron 21, 385 (1998). 28. K. M. Thompson et al., Mol. Cell. Neurosci. 23, 681 (2003). 29. K. G. Johnson, I. W. McKinnell, A. W. Stoker, C. E. Holt, J. Neurobiol. 49, 99 (2001). 30. The egl-17 FGF and the egl-15 FGFR antagonize clr-1 function in sex myoblast migration (31). egl-17 null mutations do not affect AVM guidance, whereas egl-15 mutations partially antagonize clr-1 (fig. S4). However, this effect could be indirect, because egl-15 affects axon outgrowth by acting in the epidermis (32) and egl-15 expression in AVM did not rescue the defect (33). 31. C. Z. Borland, J. L. Schutzman, M. J. Stern, Bioessays 23,1120 (2001). 32. H. E. Bulow, T. Boulin, O. Hobert, Neuron 42, 367 (2004). 33. C. Chang, C. I. Bargmann, M. Tessier-Lavigne, data not shown. 34. A. Colavita, J. G. Culotti, Dev. Biol. 194, 72 (1998). 35. We thank G. Garriga, S. Clark, and the Caenorhabditis Genetics Center for strains; A. Fire for vectors; M. Stern for the clr-1::GFP reporter; H. Nguyen for technical support; A. Colavita for critical reading of the manuscript; and members of the Bargmann and Tessier-Lavigne labs for helpful discussions. C.C. is an American Cancer Society Postdoctoral Fellow; T.W.Y. was a MINDS Predoctoral Fellow and a UCSF Medical Scientist Training Program student; and C.I.B. is and M.T.-L. was an Investigator with HHMI. Supporting Online Material www.sciencemag.org/cgi/content/full/305/5680/103/ DC1 Materials and Methods Figs. S1 to S4 References 19 February 2004; accepted 1 June 2004

2 JULY 2004 VOL 305 SCIENCE www.sciencemag.org

Supporting Online Material Materials and Methods Strains Nematodes were cultivated according to standard protocols and maintained at 20°C unless stated otherwise (1).

The following alleles were used: LGI, unc-40(e1430),

zdIs5[mec-4::gfp, lin-15(+)]; LGII, clr-1(e1745), clr-1(cy14); LGIV, ced-10(n1993), evIs82[unc-129::gfp, pMH86], zdIs4[mec-4::gfp, lin-15(+)]; LGV, unc-34(gm104); LGX, unc-6(ev400), sax-3(ky123), slt-1(eh15), slt-1(ok255), unc-115(ky275), egl15(n484), egl-15(n1477), egl-17(e1313), egl-17(n1377). Transgenes maintained as extrachromosomal

arrays

included:

kyEx637[mec-7::myr::unc-40(∆P2), 40(∆P1), odr-1::dsred].

kyEx456[mec-7::myr::unc-40,

odr-1::dsred],

and

str-1::gfp],

kyEx639[mec-7::myr::unc-

Strains that were not derived in the Tessier-Lavigne and

Bargmann laboratories were kindly provided by Scott Clark (zdIs4 and zdIs5), Gian Garriga (unc-34(gm104)), or Theresa Stiernagle of the Caenorhabditis Genetics Center. Double and triple mutants were constructed in the absence of marker mutations using standard genetic methods and confirmed by either complementation tests or genotyping. Detailed information is available upon request.

Microscopic Examination of Axon Trajectories Axonal processes of the touch cell neurons and DA/DB motorneurons were visualized with the integrated mec-4::gfp transgene zdIs5 and unc-129::gfp transgene evIs82, respectively. Animals were placed on 5% Noble Agar pads in M9 buffer containing 20 mM sodium azide and examined with a Plan-NEOFLUAR 40x objective, using

1

fluorescence optics in a Zeiss Axioplan 2 imaging system. Images were captured using a SPOT camera (RT Slider Diagnostic Instruments, Inc.). The statistics was performed by comparing two proportions from different mutants.

clr-1 Allele Sequencing To identify the molecular lesion in clr-1(cy14), PCR was performed on genomic DNA from mutants, and the products purified and sequenced. The mutation was identified by aligning the sequence with the reported genomic sequence from the C. elegans Genome Sequencing Consortium. The consequence of the splice site mutation was determined by PCR amplification of reverse-transcribed RNA from mutants and wild-type animals. PCR amplification for 40 cycles of wild-type and clr-1(cy14) reverse-transcribed RNA was performed with two pairs of primers surrounding intron 5. RT-PCR products were sequenced or separated on a 3% agarose gel.

Molecular Biology Standard molecular biology techniques were used.

mec-7::clr-1 was generated by

cloning the clr-1 cDNA into NheI and KpnI sites of pPD96.41, which contains mec-7 promoter. pPD96.41 was a gift of Andrew Fire (Carnegie Institute of Washington). The clr-1 cDNA was provided by Michael Stern (Yale University).

Transgenic Animals Germline transformation of C. elegans was performed using standard techniques (2). The mec-7::clr-1 promoter fusion was injected at 100 ng/µl along with the coinjection marker

2

odr-1::dsred at 50 ng/µl. Transgenic lines were maintained by following odr-1::dsred fluorescence. For the clr-1 cell autonomy experiment, mec-7::clr-1 was injected with odr-1::dsred into slt-1(eh15) background. The resulting transgenic lines were crossed to clr-1 mutants to generate clr-1; slt-1; cyEx(mec-7::clr-1) or crossed into wild type to generate cyEx(mec-7::clr-1). Two independently isolated transgenic lines were analyzed. The data shown are from one representative line.

Binding Assays The full-length clr-1 cDNA was cloned into XbaI/KpnI sites of pCDNA3 and the partial clr-1 cDNA, which encodes the cytoplasmic domain 767-1409 amino acid residues, was cloned into NcoI/KpnI sites of pSPUTK in vitro translation vector (Stratagene, La Jolla, California). Both cDNAs are under the control of the SP6 promoter to generate a 35[S]methionine-labeled probe by in vitro transcription/translation (TNT SP6 Quick Coupled Transcription/Translation Kit, Promega, Madison, WI). Expression vectors for UNC-40GST (1106-1415), SAX-3-GST (978-1224) or GST proteins were constructed in pGEX4T-1 (Amersham, Piscataway, New Jersey). GST proteins were expressed in E. coli BL21-CodonPlus RIL (Stratagene, La Jolla, California), purified using glutathione agarose beads (Sigma), and quantified by SDS-PAGE and Coomassie staining. For binding assays, 2 µg of GST protein was mixed with 10 µl of 35[S]-methionine-labeled CLR-1 probes in 10mM Tris 7.5, 150 mM NaCl, 0.1% Triton and 0.05% BSA. Samples were incubated for two hours at 4°C and then washed three times in the same buffer. BSA was omitted from the last wash. Bound proteins were separated by SDS-PAGE, soaked in Amplify (Amersham), dried down under vacuum and exposed to film.

3

Supporting Figures

Fig. S1. Identification of the molecular lesion of clr-1(cy14). (A) Sequence of the clr-1(cy14) mutation. The upper panel shows the genomic sequence around the exon 5–intron 5 and intron 5–exon 6 boundaries. 4

Exons are shown in

uppercase lettering, introns in lowercase. Upper panel, wild type; lower panel, clr1(cy14). The large downward-pointing arrow indicates the 5' location of the wild-type splice. The upward-pointing arrow indicates the 3' splice junction. The mutation is indicated by carets. The 3' splice acceptor has shifted due to the mutation. (B) RT-PCR analysis of wild-type and clr-1(cy14) RNA. Lanes: 1, 1 kb ladder; 2 and 4, amplification of wild-type cDNA; 3 and 5, amplification of clr-1(cy14) cDNA. Correctly spliced mRNA yields a product of 299 or 270 bp (correct joining of intron 5 and exon 6), depending on which primer set was used. In clr-1(cy14) animals, a 281 or a 252 bp product results from usage of a downstream cryptic splice acceptor at nucleotide 1,534 of cosmid F56D1.4a and removal of 18 bp from exon 6. (C) Sequence alignment of Ig domains from CLR-1, Human MHC class I antigen, and Mouse MHC class II antigen. Dark gray boxes indicate amino acid identities, and light gray boxes indicate conservative changes. Seven amino acids deleted in clr-1(cy14) are underlined. (D) Phylogram of the evolutionary relationships of C. elegans CLR-1 and five representative subtypes of vertebrate RPTPs.

Shown is an unrooted tree derived from the alignment of five

subtypes of vertebrate RPTPs with CeCLR-1 and CePTP-3, indicating the similar origin of CeCLR-1 and R5 subtype. The tree was drawn by the neighbor-joining method (3) and the proteins were selected to reflect the clustering in the neighbor-joining tree. The horizontal distance represents the degree of sequence divergence, and the scale bar at the bottom corner corresponds to 10% substitution events.

5

Fig. S2. clr-1 expression analysis. Expression pattern of clr-1 promoter was obtained from an extrachromosomal array containing GFP with a nuclear localization sequence (NLS) (NH#950) (a gift from M. Stern). 4.5kb of clr-1 promoter drives expression in a restricted subset of motile neurons and mesodermal cells. (A-C) clr-1 is expressed in the AVM neuron from mid-L1, a stage

6

when AVM sends out its axon. clr-1::gfp expression in the AVM neuron is enriched in the nucleus because of the NLS (A). mec-4::dsred was used to label the AVM neuron mainly in its cytosol (B). The superimposed image identifies the GFP expressing cell as the AVM neuron (C). (D-F) 4.5kb of clr-1 upstream sequence exhibits promoter activity in many other cells, including distal tip cells (DTC) (D), DD/VD neurons (E), and HSN neuron (F).

7

Fig. S3. CLR-1 regulates the guidance rather than the outgrowth function of UNC40. (A) Schematic of UNC-40 and its signaling pathways. The conserved cytoplasmic P1 and P2 motifs mediate distinct downstream pathways for UNC-40 function in the axon outgrowth:

UNC-34 in one pathway, and CED-10 and UNC-115 in the other (4).

MYR::UNC-40 consists of the UNC-40 cytoplasmic domain fused to a membranetargeting myristoylation signal (black circle). The P1 and P2 domains of MYR::UNC-40 are deleted in ∆P1 and ∆P2, respectively. (B) Mutations in clr-1 do not enhance the excessive outgrowth phenotypes associated with MYR::UNC-40 expression.

The

percentage of excess AVM outgrowth, detected with a mec-4::gfp transgene, was determined for animals carrying the MYR::UNC-40, MYR::UNC-40 ∆P1, or MYR::UNC-40 ∆P2 transgene alone or in combination with the clr-1(e1745) mutation.

8

Fig. S4. egl-15/FGFR promotes AVM ventral guidance. egl-15(n484) is a moderate hypomorphic allele and egl-15(n1477ts) is a temperature sensitive allele affecting the FGF receptor EGL-15; egl-17(e1313) and egl-17(n1377) are candidate null alleles of one of its two FGF ligands. egl-15 mutations could significantly enhance the defect of either slt-1 or clr-1; slt-1 strains (asterisks).

9

Supporting References and Notes 1. S. Brenner, Genetics 77, 71 (1974). 2. C. Mello, A. Fire, Methods Cell Biol. 48, 451 (1995). 3. N. Saitou, M. Nei, Mol. Biol. Evol. 4, 406 (1987). 4. Z. Gitai, T. W. Yu, E. A. Lundquist, M. Tessier-Lavigne, C. I. Bargmann, Neuron 37, 53 (2003).

10