Rules of Nonallelic Noncomplementation at the Synapse in Caenorhabditis elegans

Corresponding author: Erik M. Jorgensen, Assistant Professor, Department of Biology, University of Utah, 257 S. 1400 East, Salt Lake City, UT 84112-0840., jorgensen{at}biology.utah.edu (E-mail)

Communicating editor: P. ANDERSON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

Nonallelic noncomplementation occurs when recessive mutations in two different loci fail to complement one another, in other words, the double heterozygote exhibits a phenotype. We observed that mutations in the genes encoding the physically interacting synaptic proteins UNC-13 and syntaxin/UNC-64 failed to complement one another in the nematode Caenorhabditis elegans. Noncomplementation was not observed between null alleles of these genes and thus this genetic interaction does not occur with a simple decrease in dosage at the two loci. However, noncomplementation was observed if at least one gene encoded a partially functional gene product. Thus, this genetic interaction requires a poisonous gene product to sensitize the genetic background. Nonallelic noncomplementation was not limited to interacting proteins: Although the strongest effects were observed between loci encoding gene products that bind to one another, interactions were also observed between proteins that do not directly interact but are members of the same complex. We also observed noncomplementation between genes that function at distant points in the same pathway, implying that physical interactions are not required for nonallelic noncomplementation. Finally, we observed that mutations in genes that function in different processes such as neurotransmitter synthesis or synaptic development complement one another. Thus, this genetic interaction is specific for genes acting in the same pathway, that is, for genes acting in synaptic vesicle trafficking.

THE failure of two recessive alleles to complement one another for a specific phenotype indicates that the mutations are alleles of the same gene (BENZER 1955 ). However, there are instances in which alleles of two distinct genes do not complement one another even though a wild-type copy of each gene is present. This phenomenon is called nonallelic noncomplementation (also known as unlinked noncomplementation, second-site noncomplementation, extragenic noncomplementation, and transheterozygous noncomplementation). Examples of this phenomenon have been reported in a diverse number of biological processes such as cuticle development (KUSCH and EDGAR 1986 ), transcriptional regulation (RINE and HERSKOWITZ 1987 ; RANCOURT et al. 1995 ), cytoskeleton-based cell behavior (STEARNS and BOTSTEIN 1988 ; HAYS et al. 1989 ; WELCH et al. 1993 ; WANG and BRETSCHER 1997 ; HALSELL and KIEHART 1998 ), neuron outgrowth (KIDD et al. 1999 ), and oogenesis (JACKSON and BERG 1999 ). In the cases where the gene products are known, the noncomplementing loci frequently encode products that interact physically (RINE and HERSKOWITZ 1987 ; STEARNS and BOTSTEIN 1988 ; HAYS et al. 1989 ; VINH et al. 1993 ; WELCH et al. 1993 ; HILL et al. 1995 ; RANCOURT et al. 1995 ; KIDD et al. 1999 ; MOUNKES and FULLER 1999 ). Therefore nonallelic noncomplementation between mutations of two different loci is often interpreted as signifying a physical interaction between the products of the genes.

Two models have been proposed to explain nonallelic noncomplementation: the dosage model and the poison model (STEARNS and BOTSTEIN 1988 ; FULLER et al. 1989 ). In the dosage model, a decrease in dosage at a single locus does not affect the process; however, a simultaneous decrease at a second locus proves to be detrimental. For example, null mutations in slit, a ligand required for proper neuronal migration during development in Drosophila melanogaster, fail to complement null alleles of robo, a receptor for Slit protein (KIDD et al. 1999 ).

In the poison model, an altered gene product must bind and impair the protein complex with which it is normally associated. Although this first defect does not produce a visible phenotype, a second mutation in another member of the protein complex reveals a visible defect. Such interactions have been observed between - and ß-tubulin genes. For example, in Drosophila and yeast, altered -tubulins act as poisons by either sequestering ß-tubulin or disrupting the polymerization of the microtubule (STEARNS and BOTSTEIN 1988 ; FULLER et al. 1989 ; HAYS et al. 1989 ).

Proper communication between neurons relies on the regulated fusion of synaptic vesicles with the active zone, a specialized region of the plasma membrane, and the subsequent release of neurotransmitter. For this process to occur, first, synaptic vesicles and associated proteins must be transported to the synapse from the cell body. Second, the vesicles must be filled with neurotransmitter. Third, the mature synaptic vesicle must be docked and primed so that the vesicle can rapidly fuse with the plasma membrane when the neuron is depolarized. Finally, the vesicle and its associated proteins must be recovered from the plasma membrane through endocytosis to maintain a releasable pool of synaptic vesicles (Fig 1). Two proteins required for the exocytosis step are UNC-13 and syntaxin. UNC-13 is a diacylglycerol-binding protein with multiple C2 Ca²⁺-binding domains (MARUYAMA and BRENNER 1991 ; AHMED et al. 1992 ). UNC-13 proteins are required for exocytosis in Caenorhabditis elegans, mammals, and Drosophila (BROSE et al. 1995 ; BETZ et al. 1998 ; ARAVAMUDAN et al. 1999 ; AUGUSTIN et al. 1999 ; RICHMOND et al. 1999 ). Syntaxin is a member of the SNARE complex that is required for synaptic vesicle fusion with the plasma membrane (SOLLNER et al. 1993 ; HAYASHI et al. 1994 ; SCHULZE et al. 1995 ). UNC-13 and syntaxin physically interact (BETZ et al. 1997 ; AUGUSTIN et al. 1999 ; SASSA et al. 1999 ). It is thought that this interaction mediates a priming step in which the synaptic vesicle becomes fusion competent (BROSE et al. 1995 ; BETZ et al. 1998 ; ARAVAMUDAN et al. 1999 ; AUGUSTIN et al. 1999 ; RICHMOND et al. 1999 ).

View larger version (30K): In this window In a new window Download PPT slide   Figure 1. The synaptic vesicle cycle. The proteins used in this study are indicated. ACh, acetylcholine.

During our studies of neurotransmission in the nematode C. elegans, we observed nonallelic noncomplementation between mutations in UNC-13 and syntaxin. These proteins are encoded by the genes unc-13 and unc-64, respectively (MARUYAMA and BRENNER 1991 ; OGAWA et al. 1998 ; SAIFEE et al. 1998 ). Homozygous unc-13(n2813) and unc-64(e246) mutations cause animals to be uncoordinated. These mutations are recessive and heterozygotes are wild type in behavior. Surprisingly, worms doubly heterozygous for these alleles are uncoordinated (genotype unc-13(n2813)/+; unc-64(e246)/+) and thus exhibit nonallelic noncomplementation. We used this interaction as a starting point to determine the rules of nonallelic noncomplementation at the synapse. In these experiments, we used a drug assay to quantify the effects of mutations on synaptic function. Sensitivity to Aldicarb, an inhibitor of acetylcholinesterase, is a measure of the amount of acetylcholine released into the synaptic cleft. First, we used null and hypomorphic alleles of both unc-13 and unc-64 to distinguish between the dosage and poison models for this interaction. Second, we determined the boundaries of this interaction. Specifically, we determined whether nonallelic noncomplementation could be observed between proteins that physically interact, between proteins that do not interact but are found together in a larger complex, between proteins that simply function in the same pathway and are not in the same complex, and between proteins that are not in the same pathway.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

Strains:
N2 var. Bristol was used as the wild-type strain. Worms were cultured and maintained as previously described (BRENNER 1974 ). The following mutations were used in this study:

Linkage group (LG) I: unc-13(e51); unc-13(n2813); unc-29 (e1072)

LG II: unc-104(e1265)

LG III: unc-64(js115), unc-64(e246), bli-5(e518)

LG IV: cha-1(p1152), unc-17(e245); unc-17(ox51)

LG V: snb-1(md247), snb-1(js124); dpy-11(e224)

LG X: unc-18(e81), dpy-23(e840); syd-2(ju37)

Generation of single and double heterozygotes:
To generate heterozygous worms for a single mutation, L4 hermaphrodites homozygous for the relevant genotype were mated to wild-type males, except in the cases of unc-64(js115) and snb-1 (js124). Since these mutations are homozygous lethal, they are maintained as heterozygotes balanced by bli-5(e518) and dpy-11(e224) mutations, respectively.

To generate worms heterozygous for two mutations, heterozygous males of one genotype were mated with homozygous L4 hermaphrodites of the second genotype except in the case of unc-64(js115). In this case, unc-64(js115) +/+ bli-5(e518) males were crossed to homozygous strains to generate the double heterozygote.

Drug resistance assay:
Aldicarb (2-methyl-2-[methylthio]propionaldehyde O-[methylcarbamoyl]oxime; Chem Services, West Chester, PA) was solubilized in acetone and then diluted to a working stock solution of 17.5 mM in M9. Standard NGM worm plates were treated with Aldicarb to a final concentration of 0.7 mM and allowed to dry at room temperature. For each experiment, animals that have been adults for <2 days were picked to pretreated plates and scored for resistance after 12 hr of exposure to the drug. Worms were considered resistant if they could move on their own or respond when prodded by a platinum wire. In some experiments, the worms were scored for movement every 4 hr (±30 min) over a 16-hr period. In other experiments the worms were scored in the 12th hour of exposure.

Statistical analysis:
Since all of the double heterozygotes were generated by mating heterozygous males with homozygous hermaphrodites, the progeny from any given cross are of two types and thus the relevant genotype is not observable directly. The resistance exhibited by the worms that result from each cross is the average resistance of the two genotypes produced by segregation. To analyze the levels of drug resistance of the relevant genotype in any cross we adopted a maximum-likelihood framework (see the Appendix) in which we were able to estimate the resistance of the double heterozygote when given direct measurements of the background genotype. For most cases, including the crosses involving unc-64(js115), the genotype of resistant animals was confirmed by picking resistant animals and observing the genotypes of the progeny. Furthermore, our framework allowed us to compare hypotheses about the relationships between the resistances of different genotypes. In addition, the framework allowed us to test whether or not the level of resistance exhibited by the double heterozygote could be an additive effect of each of the single heterozygotes. The outcomes of these hypothesis tests were used as a basis for distinguishing significance in the level of drug resistance among the double heterozygote genotype and each single heterozygote genotype. We report the maximum-likelihood estimates with the corresponding 2-unit support boundaries in parentheses [i.e., 4% (2–12%)]. The 2-unit support boundaries are often used in likelihood analysis in lieu of confidence limits (EDWARDS 1992 ) and are approximately the same as the upper and lower 95% confidence intervals for population proportions (SOKAL and ROHLF 1987 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

A quantitative assay:
In the process of making double mutants we observed that worms doubly heterozygous for unc-13(n2813) and unc-64(e246) (genotype unc-13/+; unc-64/+) were uncoordinated. Thus, these mutations exhibited nonallelic noncomplementation. Further characterization of this phenomenon required a quantitative assay. Since the severity of an uncoordinated phenotype is a subjective measure, we chose to use a pharmacological assay to quantify allelic interactions. The inhibitor of cholinesterase, Aldicarb, blocks the breakdown of acetylcholine and causes acetylcholine to accumulate in the synaptic cleft. In wild-type worms, this accumulation results in constitutive depolarization and hypercontraction of the muscle, which leads to paralysis and eventually to death. Mutations that disrupt the release of acetylcholine are resistant to Aldicarb (BRENNER 1974 ; Rand AND RUSSELL 1985 ; HOSONO et al. 1989 ; MILLER et al. 1996 ). Therefore we used sensitivity to Aldicarb as a measure of synaptic activity.

To maximize the sensitivity of our assay, we identified conditions under which wild-type worms were fully sensitive to Aldicarb, but unc-13(n2813) worms were fully resistant. The lowest concentration of drug at which all wild-type animals were paralyzed was 0.7 mM Aldicarb (Fig 2A). unc-13(n2813) animals were completely resistant to all concentrations of the drug tested. Thus, this concentration provided the greatest sensitivity to changes in response to Aldicarb and was consistent with results obtained in other studies (JORGENSEN et al. 1995 ).

View larger version (20K): In this window In a new window Download PPT slide   Figure 2. Resistance to Aldicarb reveals a poisonous effect of hypomorphic unc-13 and unc-64 synaptic function alleles. (A) Resistance to Aldicarb at different concentrations. Worms were treated with different concentrations of the drug (0.3, 0.5, 0.7, 0.9, and 1.1 mM) and scored for resistance after 8 hr of exposure. (•) Wild type, n = 3(19); () unc-13(n2813), n = 3(15); and () unc-13(n2813)/+, n = 3(18). (B) Adult worms were treated with 0.7 mM Aldicarb and assayed for resistance after 4, 8, and 12 hr of exposure. (•) Wild type, n = 3(20); () unc-13(n2813), n = 3(16); () unc-13(n2813)/+, n = 3(18); () unc-13(e51), n = 6(52); and () unc-13(e51)/+, n = 5(30). (C) Adult worms were treated with 0.7 mM Aldicarb and assayed for resistance after 4, 8, and 12 hr of exposure. (•) Wild type, n = 3(20); () unc-64(e246), n = 9(32); () unc-64(e246)/+, n = 6(39); and () unc-64(js115)/+, n = 9(20). n = 3(19) indicates that three plates, each representing one experiment, were counted with an average number of 19 animals per plate. The error bars represent the maximum-likelihood analog of 95% confidence intervals.

To determine the time of exposure that provided the highest sensitivity to changes in response to Aldicarb, we performed a time course of resistance. When exposed to 0.7 mM Aldicarb, wild-type worms quickly succumbed to the effects of the drug and by 12 hr of exposure almost all the animals were paralyzed (Fig 2B). By contrast, homozygous unc-13(n2813) worms remained resistant at 12 hr (Fig 2B) and exhibited only a slight decrease in resistance at 16 hr (data not shown). Since a 12-hr exposure to 0.7 mM Aldicarb provided a satisfactory level of distinction between wild-type and mutant worms we used these parameters to assess synaptic efficacy among all double heterozygotes in this study.

Resistance to Aldicarb provided a more sensitive means of measuring synaptic efficacy than a behavioral analysis. Strong unc-13 mutants, such as the nonsense mutant unc-13(e51), are severely paralyzed as homozygotes. The unc-13(e51) mutation is an early stop in the long transcript encoding the synaptically localized isoform of UNC-13 (the short transcript encodes the axonally localized isoform; NURRISH et al. 1999 ; KOHN et al. 2000 ). By contrast, worms homozygous for the weak unc-13(n2813) mutation move forward quite well and exhibit only a jerky phenotype when prodded to move backward. The unc-13(n2813) mutation is a missense mutation in the C-terminal region of the UNC-13 protein that is common to all splice forms (KOHN et al. 2000 ). However, unc-13(n2813) mutants exhibit almost the same level of drug resistance as unc-13(e51) mutants, 92% (88–96%) and 93% (90–96%), respectively (Fig 2B). So although unc-13(n2813) homozygotes are often difficult to distinguish from wild-type worms on the basis of their behavioral phenotype, they can easily be identified on the basis of drug resistance. Therefore a small decrease in neurotransmitter release can translate into a significant level of resistance to Aldicarb.

Poison alleles:
To test the dosage and poison models, we needed null and poison alleles of the unc-13 and unc-64 loci. The unc-13(e51) and unc-13(n2813) alleles are recessive; specifically, heterozygotes exhibit a wild-type behavioral phenotype. Surprisingly, we detected a poisonous effect of the weak allele, n2813, but not the nonsense allele, e51, in our Aldicarb sensitivity experiments. unc-13(e51)/+ heterozygotes were as sensitive to Aldicarb as wild-type worms at 8 and 12 hr of exposure (Fig 2B). By contrast, unc-13(n2813) heterozygotes were weakly resistant to Aldicarb, particularly at intermediate concentrations (Fig 2A) or at short times of drug exposure (Fig 2B). These data indicated that the unc-13(n2813) allele has a weakly poisonous effect on synaptic transmission.

Similarly, we detected a poisonous effect of the weak allele of unc-64/syntaxin, e246, but not the null allele, js115, in our drug sensitivity assay. The unc-64(e246) mutation is a missense change in the SNARE domain of the UNC-64 protein. The unc-64(js115) null lesion is an early stop. As with the unc-13 null heterozygotes, unc-64(js115)/+ heterozygotes were as sensitive to Aldicarb as wild-type worms at all times of exposure to the drug (Fig 2C). By contrast, worms heterozygous for the weak allele [unc-64(e246)/+] were weakly resistant to Aldicarb at 8–12 hr of exposure (Fig 2C). Thus, like the unc-13(n2813) allele, the unc-64(e246) allele exhibited a weakly poisonous effect on synaptic transmission in a heterozygote. We used these null and poison alleles of unc-13 and unc-64 to test the dosage and poison models for nonallelic noncomplementation.

Dosage vs. poison:
We did not observe resistance to Aldicarb in heterozygous strains for null alleles of unc-13 or unc-64 (Fig 2B and Fig C). Assuming that there is no compensation for the lowered levels of gene product, then a 50% reduction in the concentrations of either of these proteins does not impair neurotransmission. However, in the dosage model of nonallelic noncomplementation, the detrimental effect is observed only when both loci are decreased at the same time. Therefore, to test the dosage model, we constructed worms that had a reduction in gene dosage at both loci simultaneously; that is, we constructed double heterozygotes [unc-13(e51)/+; unc-64(js115)/+] and measured the level of drug resistance exhibited by these worms. The double heterozygotes were not resistant to Aldicarb [1% (0–6%); Fig 3A].

View larger version (29K): In this window In a new window Download PPT slide   Figure 3. Noncomplementation between UNC-13 and syntaxin requires the presence of a poison allele. Worms singly or doubly heterozygous for unc-13 and unc-64 were constructed and tested for resistance to 0.7 mM Aldicarb after 12 hr of exposure and compared to the wild type [first column of each graph, n = 20(30)]. (A) Synaptic function is not dosage sensitive. unc-13(e51)/+, n = 7(34); unc-64(js115)/+, n = 12(18); and unc-13(e51)/+; unc-64(js115)/+, n = 5(27). (B) The two hypomorphic alleles exhibit nonallelic noncomplementation. unc-13(n2813)/+, n = 13(32); unc-64(e246), n = 9(33); and unc-13(n2813)/+; unc-64(e246)/+, n = 9(33). (C) The hypomorphic unc-13 allele reveals dosage sensitivity to unc-64 gene product levels. unc-13(n2813)/+, n = 12(30); unc-64(js115)/+, n = 12(18); and unc-13(n2813)/+; unc-64(js115)/+, n = 8(29). (D) The unc-64 hypomorphic allele reveals dosage sensitivity to unc-13 gene product levels. unc-13(e51)/+, n = 7(34); unc-64(e246), n = 9(33); and unc-13(e51)/+; unc-64(e246)/+, n = 7(29).

Alternatively, in the poison model, nonallelic noncomplementation is observed only if altered gene products impair the protein complexes with which they are normally associated. To test the poison model we constructed worms that were doubly heterozygous for the poison alleles of unc-13 and unc-64 [genotype unc-13(n2813)/+; unc-64(e246)/+]. The double heterozygote exhibited a significant increase in the level of Aldicarb resistance compared to either heterozygote alone [42% (35–50%) for unc-13(n2813)/+; unc-64(e246)/+, 4% (2–7%) for unc-13(n2813)/+, and 8% (5–11%) for unc-64(e246)/+; Fig 3B]. This result suggests that nonallelic noncomplementation between these loci obeys the poison model. In addition, the level of resistance exhibited by the double heterozygote is not due simply to an additive effect of each single heterozygote (Table 1, Appendix).

To test whether a single poison gene product can sensitize the strain to further mutations, we constructed worms heterozygous for a poison allele at one locus and a null allele at the other locus. In these cases, we still observed nonallelic noncomplementation (Fig 3C and Fig D). In addition, noncomplementation occurred regardless of which locus encoded the poison allele or the null allele. The level of Aldicarb resistance exhibited by worms heterozygous for the unc-13(n2813) poison allele and the unc-64(js115) null allele [genotype unc-13(n2813)/+; unc-64(js115)/+] was 56% (43–69%; Fig 3C). Conversely, the level of Aldicarb resistance exhibited by worms heterozygous for the unc-13(e51) nonsense allele and the unc-64(e246) poison allele was 17% [10–24%; genotype unc-13(e51)/+; unc-64(e246)/+]. This is significantly different than the level of resistance exhibited by the single heterozygotes alone and is not simply an additive effect of the single heterozygotes (Table 1, Appendix). These data imply that the presence of one altered synaptic component is sufficient to sensitize synaptic transmission to the levels of interacting gene products.

Noncomplementation among interacting loci:
Biochemical studies have demonstrated that syntaxin binds to the integral synaptic vesicle membrane protein synaptobrevin and plasma membrane protein SNAP-25 to form a SNARE complex (SOLLNER et al. 1993 ). The crystal structure of the SNARE complex suggests that these three proteins are intertwined when the synaptic vesicle is positioned at the active zone (SUTTON et al. 1998 ). To determine whether other proteins that interact with syntaxin exhibit nonallelic noncomplementation, we tested interactions between the unc-64/syntaxin and the C. elegans homolog of synaptobrevin, snb-1. We tested noncomplementation of unc-64 using the snb-1 hypomorphic allele, md247 (NONET et al. 1998 ). We observed that the double heterozygote exhibited a marked increase in Aldicarb resistance, [48% (29–71%) for unc-64(e246)/+; snb-1(md247)/+; Fig 4A] compared to the resistance of either heterozygote alone [8% (5–11%) for unc-64(e246)/+ and 4% (2–7%) for snb-1 (md247)/+]. This result suggests that nonallelic noncomplementation is a more general phenomenon among synaptic genes and is not specific to the unc-13(n2813) mutation.

View larger version (31K): In this window In a new window Download PPT slide   Figure 4. Nonallelic noncomplementation occurs between genes that encode components of the same complex. (A) Syntaxin and synaptobrevin mutations exhibit nonallelic noncomplementation. unc-64(e246)/+, n = 9(46); snb-1 (md247)/+, n = 7(27); and unc-64(e246)/+; snb-1(md247)/+, n = 4(17). (B) UNC-13 and synaptobrevin mutations exhibit nonallelic noncomplementation. unc-13(n2813)/+, n = 12(30); snb-1(md247)/+, n = 7(27); and unc-13(n2813)/+; snb-1 (md247)/+, n = 5(27). (C) UNC-13 and UNC-18 mutations exhibit nonallelic noncomplementation. unc-13(n2813)/+, n = 12(30); unc-18(e81)/+, n = 9(28); and unc-13(n2813)/+; unc-18(e81)/+, n = 9(30). Insets depict gene products at the synapse. In all graphs, wild-type n = 20(30) and resistance was measured after 12 hr of exposure to 0.7 mM Aldicarb.

Noncomplementation among proteins of the same complex:
Nonallelic noncomplementation is often interpreted to signify a direct physical interaction between the products of the loci involved. To test whether nonallelic noncomplementation requires a direct physical association between the noncomplementing gene products, we constructed double heterozygotes between unc-13(n2813) and mutations in proteins that are not known to bind to UNC-13. Since SNB-1/synaptobrevin associates only indirectly with UNC-13 as part of a complex with UNC-64/syntaxin (BETZ et al. 1997 ), we tested whether unc-13(n2813) would complement snb-1. Double heterozygotes between unc-13(n2813) and the snb-1(md247) weak allele exhibited a level of drug resistance that is significantly different from either heterozygote alone [19% (11–28%) for unc-13(n2813)/+; snb-1(md247)/+, 4% (2–7%) for unc-13(n2813)/+, and 4% (2–7%) for snb-1(md247)/+; Fig 4B]. Therefore nonallelic noncomplementation at the synapse does not require a direct physical interaction between the gene products. However, as observed between unc-13 and unc-64, there was no interaction between null alleles of these loci. Specifically, unc-13(e51) fully complements the null allele snb-1(js124) (data not shown).

Furthermore, we tested whether unc-13 exhibited nonallelic noncomplementation with unc-18. UNC-18 also binds to UNC-64/syntaxin (HATA et al. 1993 ), but does not bind when UNC-13 or synaptobrevin is bound (PEVSNER et al. 1994 ; SASSA et al. 1999 ). We observed that the double heterozygote exhibited more resistance than either heterozygote alone [19% (12–27%) for unc-13(n2813)/+; unc-18(e81)/+, 4% (2–7%) for unc-13(n2813)/+, and 1% (0–2%) for unc-18(e81)/+; Fig 4C]. Since unc-18 heterozygotes (unc-18(e81)/+) are wild type for Aldicarb sensitivity (Fig 4C), the e81 mutation does not appear to have a poisonous effect. Furthermore, unc-18(e81) is likely to be a null allele since it is a nonsense mutation and since there is no immunoreactivity in unc-18(e81) homozygous worms (GENGYO-ANDO et al. 1993 ). Thus, unc-13(n2813) can sensitize synaptic function to decreases in the dosage of proteins that are not components of the SNARE complex.

Boundaries of nonallelic noncomplementation:
So far our results have demonstrated that a poison allele of unc-13 can sensitize synaptic transmission to the dosage of components required for synaptic vesicle exocytosis. The availability of gene products at the synaptic terminal requires transport from two routes: via kinesin-mediated transport from the cell body and via endocytosis from the terminal itself. Therefore, to explore the ability of a poison allele to reveal dosage sensitivity in the transport machinery, we tested genetic interactions between UNC-13 and synaptic vesicle kinesin and between UNC-13 and the clathrin adaptor complex, proteins that control vesicle trafficking at the synapse.

unc-104 encodes a kinesin motor protein that transports synaptic vesicles and proteins to the terminal (HALL and HEDGECOCK 1991 ; OTSUKA et al. 1991 ). There is no evidence to support a direct interaction between UNC-13 and UNC-104 proteins, nor is it suspected that these products participate in the same complex. Worms homozygous for the weak allele, unc-104(e1265), are nearly paralyzed, whereas heterozygotes behave like the wild type. In the presence of Aldicarb, unc-104/+ single heterozygotes also behave like wild-type animals. However, double heterozygotes exhibited a significantly increased level of Aldicarb resistance compared to either heterozygote [19% (11–28%) for unc-13(n2813)/+; unc-104(e1265)/+, 2% (0–5%) for unc-104(e1265)/+; Fig 5A]. This result suggests that a poison allele in the fusion machinery can reveal perturbations in trafficking components.

View larger version (25K): In this window In a new window Download PPT slide   Figure 5. Nonallelic noncomplementation occurs between unc-13(n2813) and genes involved in synaptic vesicle trafficking. (A) UNC-13 and synaptic vesicle kinesin mutations exhibit nonallelic noncomplementation. unc-13 (n2813)/+, n = 6(39); unc-104(e1265)/+, n = 6(38); and unc-13(n2813)/+; unc-104(e1265)/+, n = 6(38). (B) UNC-13 and clathrin adaptor complex mutations exhibit nonallelic noncomplementation. unc-13 (n2813)/+, n = 6(39); dpy-23(e840)/+, n = 2(31); and unc-13(n2813)/+; dpy-23(e840), n = 6(22). In all graphs, wild-type, n = 10(32) and resistance was measured after 12 hr of exposure to 0.7 mM Aldicarb.

Synaptic gene products are recycled from the plasma membrane via clathrin-mediated endocytosis. Clathrin assembly on the plasma membrane requires the AP-2 adaptor complex. The µ-subunit of the C. elegans AP-2 complex is encoded by dpy-23 (P. BAUM and G. GARRIGA, personal communication). Worms homozygous for the null allele dpy-23(e840) are severely uncoordinated and exhibited an intermediate level of resistance to Aldicarb [55% (46–63%), data not shown]. Although dpy-23 heterozygotes were not drug resistant, double heterozygotes were markedly resistant [42% (29–58%) for unc-13(n2813)/+; dpy-23(e840)/+, 4% (2–7%) for unc-13(n2813)/+, and 0% (0–3%) for dpy-23(e840)/+; Fig 5B]. This level of resistance is similar to the level of drug resistance exhibited by dpy-23(e840) homozygotes. Thus, the poison mutation of unc-13 can sensitize synaptic function to decreased levels of components required for endocytosis.

Mutations in distant pathways:
To determine if unc-13(n2813) could sensitize neurotransmission to mutations that do not affect the exocytotic machinery, we analyzed interactions with mutations affecting the levels of neurotransmitter in synaptic vesicles. cha-1 encodes choline acetyltransferase, the biosynthetic enzyme for acetylcholine (Rand AND RUSSELL 1984 ). Null mutations in cha-1 are lethal whereas a hypomorphic mutation [cha-1(p1152)] is viable and paralyzed as homozygotes. Even though the cha-1 heterozygote has decreased levels of choline acetyltransferase activity (Rand AND RUSSELL 1984 ), heterozygous worms exhibited wild-type levels of Aldicarb resistance [1% (0–4%) for cha-1(p1152)/+; Fig 6A]. Furthermore, the double heterozygote was barely distinguishable from the unc-13 heterozygote alone and any increase in the level of resistance of the double heterozygote could be attributed to an additive effect of both mutations [5% (0–14%) for unc-13(n2813)/+; cha-1(p1152)/+ and 4% (2–7%) for unc-13(n2813)/+; Fig 6A, Appendix]. Therefore, a poison UNC-13 protein does not sensitize synaptic function to deficits in neurotransmitter supply.

View larger version (20K): In this window In a new window Download PPT slide   Figure 6. Nonallelic noncomplementation does not occur between distant members of the synaptic vesicle pathway. (A) unc-13 complements a mutation of the biosynthetic enzyme for acetylcholine. unc-13(n28- 13)/+, n = 6(39); cha-1(pr1152)/+, n = 4(29); unc-13(n2813)/+; cha-1(pr- 1152)/+, n = 5(25). (B) unc-13 complements a mutation of the acetylcholine vesicular transporter. unc-13(n2813)/+, n = 6(39); unc-17(e245)/+, n = 4(22); and unc-13(n2813)/+; unc-17(e245)/+, n = 8(31). For both graphs, wild-type, n = 10(32); and resistance was measured after 12 hr of exposure to 0.7 mM Aldicarb.

Furthermore, we analyzed genetic interactions between unc-13(n2813) and unc-17. unc-17 encodes the vesicular acetylcholine transporter (VAChT) that is required for loading synaptic vesicles with acetylcholine (ALFONSO et al. 1993 ). Worms homozygous for the null allele, ox51, are lethal, whereas worms homozygous for the weak allele, e245, are severely uncoordinated and coiled. unc-17(e245) heterozygotes exhibited a small amount of resistance. However, the double heterozygote did not exhibit a level of Aldicarb resistance that was significantly different from the single heterozygotes [3% (0–6%) for unc-13(n2813)/+; unc-17(e245)/+, 4% (2–7%) for unc-13(n2813)/+, and 1% (0–5%) for unc-17(e245)/+; Fig 6B]. Similar results were obtained for the lethal allele, ox51, of unc-17 (data not shown). Thus, a poison UNC-13 protein does not significantly enhance defects in the loading of acetylcholine into synaptic vesicles.

Neurotransmitter from the synaptic vesicle is released into the synaptic cleft and diffuses to receptors on the postsynaptic cell. We observed that unc-13(n2813)/+ does not complement a mutation in the acetycholine receptor subunit encoded by the unc-29 gene [26% (19–34%) for unc-13(n2813)/+; unc-29(e1072)/+; 1% (0–5%) for unc-29(e1072)/+, and 4% (2–7%) for unc-13(n2813)/+; Fig 7]. Therefore, perturbations in unc-13 enhance reductions in the postsynaptic receptors that bind acetylcholine but not reductions in the machinery that loads the vesicle with acetylcholine. These data suggest that vesicle loading is not a rate-limiting step in synaptic transmission.

View larger version (31K): In this window In a new window Download PPT slide   Figure 7. Nonallelic noncomplementation occurs between genes that function across the synapse. unc-13 does not complement a mutation in unc-29, which encodes an acetylcholine receptor subunit. Wild type, n = 10(32); unc-13(n2813)/+, n = 6(39); unc-29(e1072)/+, n = 2(48); and unc-13(n2813)/+; unc-29(e1072)/+, n = 6(22). Resistance was measured after 12 hr of exposure to 0.7 mM Aldicarb.

Finally, a mutation that affects synaptic development does not display nonallelic noncomplementation with unc-13(n2813). syd-2 encodes a liprin protein, which is a family of proteins that interact with receptor tyrosine phosphatases. syd-2 is required for the development of synaptic termini (ZHEN and JIN 1999 ). Specifically, homozygous syd-2 mutants exhibit a mislocalization of presynaptic proteins and have deformed active zones. syd-2 mutant animals are moderately resistant to Aldicarb. However, unc-13(n2813) fully complements syd-2(ju37) [7% (0–16%) for unc-13(n2813)/+; syd-2(ju37)/+, 4% (1–10%) for syd-2(ju37)/+, and 4% (2–7%) for unc-13(n2813)/+; Fig 8]. These results indicate that the compromises in synaptic machinery caused by unc-13(n2813) do not synergistically interact with mutations in the developmental gene syd-2.

View larger version (20K): In this window In a new window Download PPT slide   Figure 8. Nonallelic noncomplementation does not occur between unc-13 and genes required for development. unc-13(n2813) complements a mutation in syd-2, which is required for active zone formation. Wild type, n = 10(32); unc-13(n2813)/+, n = 6(39); syd-2(ju37)/+, n = 3(26); and unc-13(n2813)/+; syd-2(ju37)/+, n = 6(23). Resistance was measured after 12 hr of exposure to 0.7 mM Aldicarb.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

Nonallelic noncomplementation occurs between synaptic function genes in C. elegans. Our analysis of this genetic interaction demonstrated that hypomorphic mutations, such as unc-13(n2813) and unc-64(e246), which are recessive in behavioral assays, can act as weak poisons as heterozygotes in quantitative drug sensitivity assays. These poisons sensitize the process of neurotransmission to perturbations at other synaptic loci, resulting in nonallelic noncomplementation. In addition, it is the presence of these poisons rather than a simple decrease in the dosage of the gene product that is essential for nonallelic noncomplementation interactions at the synapse. Furthermore, in the cases where we observed nonallelic noncomplementation we have demonstrated that the increase in drug resistance of the double heterozygote is a synergistic effect and not simply an additive effect of the noncomplementing mutations.

The dosage model and the poison model have been proposed to explain how nonallelic mutations could exhibit noncomplementation. In the dosage model a decrease in expression at two loci cripples the process. In the poison model an altered gene product poisons the protein complex with which it is associated; thus there are not enough functional complexes. In the poison model the number of functional protein complexes is the limiting factor, not the amount of a single protein. Examples of both dosage-based (RANCOURT et al. 1995 ; JACKSON and BERG 1999 ; KIDD et al. 1999 ) and poison-based (STEARNS and BOTSTEIN 1988 ; HAYS et al. 1989 ; WELCH et al. 1993 ) nonallelic noncomplementation have been reported. In the instances of dosage-based nonallelic noncomplementation the gene products are required to coordinate developmental programs. For example, neuronal outgrowth across the CNS midline in Drosophila is compromised by a decrease in dosage of both slit and robo, the ligand and receptor, respectively, of this guidance pathway (BROSE et al. 1999 ; KIDD et al. 1999 ). Thus a limiting factor in neuronal outgrowth is the concentration of its signals and receptors. This sensitivity to protein concentrations is a common feature of developmental pathways that must respond to protein gradients.

By contrast, we have demonstrated that synaptic function is not sensitive to gene dosage; however, synaptic function is sensitive to altered gene products. These observations suggest that neurotransmission is not as sensitive a process as developmental pathways to changes in protein concentrations. However, neurotransmission depends on the formation of a limited number of protein complexes. In processes requiring the correct assembly of protein complexes, a single faulty subunit can render a large number of gene products inactive by participating in and poisoning protein complexes.

How could poison alleles such as unc-13(n2813) or unc-64(e246) affect synaptic transmission? Synaptic transmission requires the action of a docking complex and then a fusion complex. We imagine that if a component of a complex is a poison protein, then the complex as a whole would not function or would function inefficiently. Therefore, poisons of the docking or fusion complexes could hinder vesicles from associating with the plasma membrane.

Not surprisingly, nonallelic noncomplementation was not observed between genes involved in processes removed from the exocytotic machinery. For example, unc-13(n2813) fully complemented mutations in genes responsible for acetylcholine synthesis or loading of neurotransmitter into vesicles. In addition unc-13(n2813) fully complemented a mutation in syd-2, which is required for development of neuromuscular junctions. Thus, nonallelic noncomplementation with unc-13 is observed with other proteins in the UNC-13 pathway but no interactions were observed with genes involved in distant pathways such as neurotransmitter synthesis or synaptic development. Nonetheless, it is surprising that unc-13(n2813) can sensitize synaptic transmission to UNC-104 and DPY-23, proteins that are not involved in the formation of the SNARE complex. Heterozygotes of these genes must affect the complex indirectly by stressing the vesicular transport machinery. Thus, nonallelic noncomplementation with unc-13 extends to processes involved in synaptic vesicle trafficking. Taken together, our results suggest that nonallelic noncomplementation would be a useful screening tool for uncovering new genes required for both SNARE complex formation as well as for genes acting at more distant steps in synaptic transmission.

	FOOTNOTES

¹ Present address: Department of Biochemistry, University of Oxford, Oxford OX1 3QU, England.
² Present address: Department of Zoology, University of Toronto, Toronto, Ontario M5S 3G5, Canada.

	ACKNOWLEDGMENTS

We are very grateful to Becky Eustance Kohn, Jim Rand, Paul Baum, and Gian Garriga for kindly sharing unpublished data. We thank the members of the Jorgensen Lab and the participants of Theory Lunch for providing invaluable discussions and support. In particular, we thank Mark Hammarlund, Todd W. Harris, and Wayne Davis for very helpful comments on the manuscript. We thank the Caenorhabditis Genetics Center and Mike Nonet for providing strains. This work was supported by a National Institutes of Health (NIH) genetics training grant to K.J.Y. and NIH grant RO1 NS34307 to E.M.J.

Manuscript received October 3, 2000; Accepted for publication January 29, 2001.

	APPENDIX

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

MAXIMUM-LIKELIHOOD MODELS OF GENETIC EFFECTS
The analysis of data in this article required statistical calculations to estimate the effects of particular mutations to test genetic hypotheses. We adopted a maximum-likelihood approach, which allows both parameter estimation and hypothesis testing using the same technique. Parameter estimations can determine the contributions to resistance by both the experimental genotypes and the balancer genotypes, which will appear in our experiments but are not informative. There are two hypotheses for the resistance of the double heterozygote: It is either due to the independent resistance from each locus—that is it is additive—or they interact synergistically. Note that nonallelic noncomplementation could also be caused by the independent additivity of resistance of the genotypes; that is, the double heterozygote will have greater resistance than either of the single heterozygotes. However, this is not true for the interactions observed.

Estimating gene effects:
The unc-64 null allele is homozygous lethal; thus it can be maintained only as a heterozygous strain balanced by another mutation [genotype unc-64(js115)+/+ bli-5(e518)]. Therefore, the effect of the unc-64(js115) null allele could not be measured directly; populations of unc-64(js115)/+ heterozygotes are always accompanied by bli-5(e518)/+ worms. This problem is compounded in the generation of double heterozygotes; populations of double heterozygotes containing the lethal unc-64 allele are accompanied by double heterozygotes containing the balancing allele. Since the lethal and balancing mutations segregate in a Mendelian fashion, the total probability that a worm carries the lethal mutation is easily calculated. Since the worms are isogenic, each worm is treated as an independent trial for a genotype and thus we adopted a binomial model of drug resistance. This model has no error terms, even though experimental error is introduced by variations in drug batches, humidity, agar, etc. These factors are unlikely to affect our statistical interpretations because wild-type worms were treated and measured in parallel with each of the drug trials. Data were included in the analysis only on days when wild-type worms, used as our control, behaved consistently with previously measured data.

Maximum-likelihood statistics involve the probability of observing the data given a model and model parameters. With the model described above, the probability of observing a given data set is

(A1)

where p₁ is the probability that an individual with the balancing allele is resistant; p₂ is the probability that an individual with the lethal allele is resistant; n_1,j and i_1,j are the starting number, that is, the initial number of worms on each plate, and ending number, that is, the number of resistant worms on each plate, from the jth trial for the balancing allele; and n_2,j and i_2,j are the starting and ending numbers of worms for the jth trial for the lethal allele. The likelihood function is the negative of the natural logarithm of the probability. The maximum-likelihood estimator (MLE) is the parameter value that maximizes the likelihood function. This can be found by solving for the derivative of the likelihood function, with respect to the parameters, equal to 0. For the two-parameter model described by Equation A1, that is the simultaneous solution of

and

which yields

(A2)

(A3)

where i_k = _ji_k,j, n_k = _jn_k,j, and p^*_i is the maximum-likelihood estimator for genotype i.

While the solutions in Equation A2 and Equation A3 represent the only critical point for the likelihood function, they sometimes yield solutions outside of the range of possible susceptibilities [0, 1]. In that case there is no critical point and the MLE is on the boundary. If Equation A3 yields a p^*₂ < 0, then the solution is

(A4)

(A5)

Likewise, if Equation A3 yields a p^*₂ > 1, then the solution is

(A6)

(A7)

The confidence intervals were calculated by finding the range of parameters for which the likelihood function was within 2 of the maximum value. The range of parameters was determined by numerically finding the value of p₁ that produced a likelihood score 2 units less than the maximum, while the likelihood function was maximized in p₂. The bounds for p₂ were calculated in a symmetric way.

Testing genetic hypotheses:
Since single heterozygotes can exhibit a low level of drug resistance on their own, it was important to determine if the resistance of the double heterozygote could be explained simply by the two loci acting independently or if the resistance of the double heterozygote is caused by a synergistic interaction of the two loci. To this end we compared two hypotheses for the resistance of the double heterozygote. We call the first hypothesis the additive hypothesis, that is, that the resistance of the double heterozygote is given by p₃ = 1 - (1 - p₁)*(1 - p₂). The other hypothesis, which we call the synergistic hypothesis, is that all three genotypes have independent resistances. The additive hypothesis could be rejected in favor of the synergistic hypothesis for two potential reasons: The resistance of the double heterozygote was either lower or higher than expected under additivity. If it is higher than expected under additivity then we can say that there is a synergistic effect. This is similar to a two-tailed test. There were no cases of synergism where the resistance was less than expected under the additive hypothesis.

The likelihood of a given model is analogous to Equation A1 for each genotype, and the total likelihood of a model is the sum of the likelihoods for each genotype (EDWARDS 1992 ). We used likelihood-ratio tests [analogous to the G-test (EDWARDS 1992 )] to determine which hypothesis best represents the data. Since the different models have different numbers of parameters, we apply the Akaike information criterion (AIC; SAKAMOTO et al. 1986 ) and deduct one log-likelihood point per free parameter. The results from the hypotheses testing for the genotypes compared in this article are displayed in Table 1 (underlined values represent the accepted model).

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX LITERATURE CITED

AHMED, S., I. N. MARUYAMA, R. KOZMA, J. LEE, and S. BRENNER et al., 1992  The Caenorhabditis elegans unc-13 gene product is a phospholipid-dependent high-affinity phorbol ester receptor. Biochem. J. 287:995-999.

ALFONSO, A., K. GRUNDAHL, J. S. DUERR, H. P. HAN, and J. B. RAND, 1993  The Caenorhabditis elegans unc-17 gene: a putative vesicular acetylcholine transporter. Science 261:617-619[Abstract/Free Full Text].

ARAVAMUDAN, B., T. FERGESTAD, W. S. DAVIS, C. K. RODESCH, and K. BROADIE, 1999  Drosophila UNC-13 is essential for synaptic transmission. Nat. Neurosci. 2:965-971[Medline].

AUGUSTIN, I., A. BETZ, C. HERRMANN, T. JO, and N. BROSE, 1999  Differential expression of two novel Munc13 proteins in rat brain. Biochem. J. 337:363-371.

BENZER, S., 1955  Fine structure of a genetic region in bacteriophage. Proc. Natl. Acad. Sci. USA 41:344-354[Free Full Text].

BETZ, A., M. OKAMOTO, F. BENSELER, and N. BROSE, 1997  Direct interaction of the rat unc-13 homologue Munc13-1 with the N terminus of syntaxin. J. Biol. Chem. 272:2520-2526[Abstract/Free Full Text].

BETZ, A., U. ASHERY, M. RICKMANN, I. AUGUSTIN, and E. NEHER et al., 1998  Munc13-1 is a presynaptic phorbol ester receptor that enhances neurotransmitter release. Neuron 21:123-136[Medline].

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

BROSE, N., K. HOFMANN, Y. HATA, and T. C. SUDHOF, 1995  Mammalian homologues of Caenorhabditis elegans unc-13 gene define novel family of C2-domain proteins. J. Biol. Chem. 270:25273-25280[Abstract/Free Full Text].

BROSE, K., K. S. BLAND, K. H. WANG, D. ARNOTT, and W. HENZEL et al., 1999  Slit proteins bind Robo receptors and have an evolutionarily conserved role in repulsive axon guidance. Cell 96:795-806[Medline].

EDWARDS, A. W. F., 1992 Likelihood. Johns Hopkins University Press, Baltimore.

FULLER, M. T., C. L. REGAN, L. L. GREEN, B. ROBERTSON, and R. DEURING et al., 1989  Interacting genes identify interacting proteins involved in microtubule function in Drosophila. Cell. Motil. Cytoskeleton 14:128-135[Medline].

GENGYO-ANDO, K., Y. KAMIYA, A. YAMAKAWA, K. KODAIRA, and K. NISHIWAKI et al., 1993  The C. elegans unc-18 gene encodes a protein expressed in motor neurons. Neuron 11:703-711[Medline].

HALL, D. H. and E. M. HEDGECOCK, 1991  Kinesin-related gene unc-104 is required for axonal transport of synaptic vesicles in C. elegans.. Cell 65:837-847[Medline].

HALSELL, S. R. and D. P. KIEHART, 1998  Second-site noncomplementation identifies genomic regions required for Drosophila nonmuscle myosin function during morphogenesis. Genetics 148:1845-1863[Abstract/Free Full Text].

HATA, Y., C. A. SLAUGHTER, and T. C. SUDHOF, 1993  Synaptic vesicle fusion complex contains unc-18 homologue bound to syntaxin. Nature 366:347-351[Medline].

HAYASHI, T., H. MCMAHON, S. YAMASAKI, T. BINZ, and Y. HATA et al., 1994  Synaptic vesicle membrane fusion complex: action of clostridial neurotoxins on assembly. EMBO J. 13:5051-5061[Medline].

HAYS, T. S., R. DEURING, B. ROBERTSON, M. PROUT, and M. T. FULLER, 1989  Interacting proteins identified by genetic interactions: a missense mutation in alpha-tubulin fails to complement alleles of the testis-specific beta-tubulin gene of Drosophila melanogaster.. Mol. Cell. Biol. 9:875-884[Abstract/Free Full Text].

HILL, K. K., V. BEDIAN, J. L. JUANG, and F. M. HOFFMANN, 1995  Genetic interactions between the Drosophila Abelson (Abl) tyrosine kinase and failed axon connections (fax), a novel protein in axon bundles. Genetics 141:595-606[Abstract].

HOSONO, R., T. SASSA, and S. KUNO, 1989  Spontaneous mutations of trichlofon resistance in the nematode Caenorhabditis elegans.. Zool. Sci. 6:697-708.

JACKSON, S. M. and C. A. BERG, 1999  Soma-to-germline interactions during Drosophila oogenesis are influenced by dose-sensitive interactions between cut and the genes cappuccino, ovarian tumor and agnostic.. Genetics 153:289-303[Abstract/Free Full Text].

JORGENSEN, E. M., E. HARTWIEG, K. SCHUSKE, M. L. NONET, and Y. JIN et al., 1995  Defective recycling of synaptic vesicles in synaptotagmin mutants of Caenorhabditis elegans.. Nature 378:196-199[Medline].

KIDD, T., K. S. BLAND, and C. S. GOODMAN, 1999  Slit is the midline repellent for the robo receptor in Drosophila.. Cell 96:785-794[Medline].

KOHN, R. E., J. S. DUERR, J. R. MCMANUS, A. DUKE, and T. RAKOW et al., 2000  Expression of multiple UNC-13 proteins in the C. elegans nervous system. Mol. Biol. Cell 11:3441-3452[Abstract/Free Full Text].

KUSCH, M. and R. S. EDGAR, 1986  Genetic studies of unusual loci that affect body shape of the nematode Caenorhabditis elegans and may code for cuticle structural proteins. Genetics 113:621-639[Abstract/Free Full Text].

MARUYAMA, I. N. and S. BRENNER, 1991  A phorbol ester/diacylglycerol-binding protein encoded by the unc-13 gene of Caenorhabditis elegans.. Proc. Natl. Acad. Sci. USA 88:5729-5733[Abstract/Free Full Text].

MILLER, K. G., A. ALFONSO, M. NGUYEN, J. A. CROWELL, and C. D. JOHNSON et al., 1996  A genetic selection for Caenorhabditis elegans synaptic transmission mutants. Proc. Natl. Acad. Sci. USA 93:12593-12598[Abstract/Free Full Text].

MOUNKES, L. C. and M. T. FULLER, 1999  Molecular characterization of mutant alleles of the DNA repair/basal transcription factor haywire/ERCC3 in Drosophila. Genetics 152:291-297[Abstract/Free Full Text].

NONET, M. L., O. SAIFEE, H. ZHAO, J. B. RAND, and L. WEI, 1998  Synaptic transmission deficits in Caenorhabditis elegans synaptobrevin mutants. J. Neurosci. 18:70-80[Abstract/Free Full Text].

NURRISH, S., L. SEGALAT, and J. M. KAPLAN, 1999  Serotonin inhibition of synaptic transmission: Galpha(0) decreases the abundance of UNC-13 at release sites. Neuron 24:231-242[Medline].

OGAWA, H., S. HARADA, T. SASSA, H. YAMAMOTO, and R. HOSONO, 1998  Functional properties of the unc-64 gene encoding a Caenorhabditis elegans syntaxin. J. Biol. Chem. 273:2192-2198[Abstract/Free Full Text].

OTSUKA, A. J., A. JEYAPRAKASH, J. GARCIA-ANOVEROS, L. Z. TANG, and G. FISK et al., 1991  The C. elegans unc-104 gene encodes a putative kinesin heavy chain-like protein. Neuron 6:113-122[Medline].

PEVSNER, J., S. C. HSU, J. E. BRAUN, N. CALAKOS, and A. E. TING et al., 1994  Specificity and regulation of a synaptic vesicle docking complex. Neuron 13:353-361[Medline].

RANCOURT, D. E., T. TSUZUKI, and M. R. CAPECCHI, 1995  Genetic interaction between hoxb-5 and hoxb-6 is revealed by nonallelic noncomplementation. Genes Dev. 9:108-122[Abstract/Free Full Text].

RAND, J. B. and R. L. RUSSELL, 1984  Choline acetyltransferase-deficient mutants of the nematode Caenorhabditis elegans.. Genetics 106:227-248[Abstract/Free Full Text].

RAND, J. B. and R. L. RUSSELL, 1985  Pharmacogenetics: a preclinical and clinical perspective. Psychopharmacol. Bull. 21:623-630[Medline].

RICHMOND, J. E., W. S. DAVIS, and E. M. JORGENSEN, 1999  UNC-13 is required for synaptic vesicle fusion in C. elegans.. Nat. Neurosci. 2:959-964[Medline].

RINE, J. and I. HERSKOWITZ, 1987  Four genes responsible for a position effect on expression from HML and HMR in Saccharomyces cerevisiae.. Genetics 116:9-22[Abstract/Free Full Text].

SAIFEE, O., L. WEI, and M. L. NONET, 1998  The Caenorhabditis elegans unc-64 locus encodes a syntaxin that interacts genetically with synaptobrevin. Mol. Biol. Cell 9:1235-1252[Abstract/Free Full Text].

SAKAMOTO, Y., M. ISHIGURO and G. KITAGAWA, 1986 Akaike Information Criterion Statistics. Kluwer Academic, Boston.

SASSA, T., S. HARADA, H. OGAWA, J. B. RAND, and I. N. MARUYAMA et al., 1999  Regulation of the UNC-18-Caenorhabditis elegans syntaxin complex by UNC-13. J. Neurosci. 19:4772-4777[Abstract/Free Full Text].

SCHULZE, K. L., K. BROADIE, M. S. PERIN, and H. J. BELLEN, 1995  Genetic and electrophysiological studies of Drosophila syntaxin-1A demonstrate its role in nonneuronal secretion and neurotransmission. Cell 80:311-320[Medline].

SOKAL, R. F., and F. J. ROHLF, 1987 Introduction to Biostatistics. W. H. Freeman, New York.

SOLLNER, T., M. K. BENNETT, S. W. WHITEHEART, R. H. SCHELLER, and J. E. ROTHMAN, 1993  A protein assembly-disassembly pathway in vitro that may correspond to sequential steps of synaptic vesicle docking, activation, and fusion. Cell 75:409-418[Medline].

STEARNS, T. and D. BOTSTEIN, 1988  Unlinked noncomplementation: isolation of new conditional-lethal mutations in each of the tubulin genes of Saccharomyces cerevisiae.. Genetics 119:249-260[Abstract/Free Full Text].

SUTTON, R. B., D. FASSHAUER, R. JAHN, and A. T. BRUNGER, 1998  Crystal structure of a SNARE complex involved in synaptic exocytosis at 2.4 A resolution. Nature 395:347-353[Medline].

VINH, D. B., M. D. WELCH, A. K. CORSI, K. F. WERTMAN, and D. G. DRUBIN, 1993  Genetic evidence for functional interactions between actin noncomplementing (Anc) gene products and actin cytoskeletal proteins in Saccharomyces cerevisiae.. Genetics 135:275-286[Abstract].

WANG, T. and A. BRETSCHER, 1997