MUTATIONS WITH DOMINANT EFFECTS ON THE ... - Genetics

Copyright 0 1986 by the Genetics Society of America

MUTATIONS WITH DOMINANT EFFECTS ON THE BEHAVIOR AND MORPHOLOGY OF THE NEMATODE CAENORHABDITIS ELEGANS EUN-CHUNG PARK' A N D H. ROBERT HORVITZ Defiartment of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Manuscript received January 14, 1986 Revised copy accepted May 2, 1986 ABSTRACT We have analyzed 31 mutations that have dominant effects on the behavior or morphology of the nematode Caenorhabditis elegans. These mutations appear to define 15 genes. We have studied ten of these genes in some detail and have been led to two notable conclusions. First, loss of gene function for four of these ten genes results in a wild-type phenotype; if these genes represent a random sample from the genome, then we would estimate that null mutations in about half of the genes in C. elegans would result in a nonmutant phenotype. Second, the dominant effects of mutations in nine of these ten genes are caused by novel gene functions, and in all nine cases the novel function is antagonized by the wild-type function.

W

HILE mutations with recessive effects generally reduce or eliminate the functions of particular genes (MULLER1932), mutations with dominant effects are heterogeneous in their functional bases. Mutations with dominant effects in Drosophila melanogaster and in Caenorhabditis elegans have been determined to result in the acquisition of a novel gene activity (MULLER 1932; WATERSTON 1981), in the increase (or ectopic expression) of an essentially normal gene activity (LIFSCHYTZ and GREEN1979; GREENWALD, STERNBERG and HORVITZ1983) or in a decrease in gene activity, either by antagonizing the function of the wild-type allele (BUSSONet al. 1983; FERGUSON and HORVITZ 1985) or as a consequence of an absolute requirement for two copies of a wild-type allele for the wild-type phenotype (LINDSLEY et al. 1972; DENELL 1978). One of the aims of the present study is to examine a large number of mutations with dominant effects to determine the relative frequencies of each of these functional bases of dominance. and HORVITZ In addition, it has been suggested (SUZUKI1970; GREENWALD 1980) that members of multigene families may be most readily identified genetically by studying mutations with dominant effects, since loss-of-function mutations of these genes may not result in distinguishable mutant phenotypes. This prediction has been supported by the identification of the C. elegans gene

' Present address: Department of Molecular Biology, Massachusetts General Hospital, Boston, Massachusetts 02114.

Genetics 113: 821-852 August, 1986.

-

822

E.-C.

PARK AND H.

R.

HORVITZ

unc-92, which was defined by mutations with dominant effects, as a member of an actin gene family (LANDELet al. 1984). However, it is not clear what fraction of genes identified by mutations with dominant effects have wild-type null phenotypes and whether all visible phenotypes resulting from mutations in genes with wild-type null phenotypes are dominant. We have addressed these questions by isolating and characterizing mutations with dominant effects in C. elegans. C. elegans is a favorable metazoan for this type of analysis. C. elegans hermaphrodites reproduce by self-fertilization or by mating with males. Self-fertilization allows the maintenance of severely defective mutants, since mating is not required for reproduction. Genetic analysis of C. elegans is straightforward and is facilitated by its small size, ease of culture and short life cycle (BRENNER 1974; HERMAN and HORVITZ1980). We have focused on mutations that affect behavior and/or morphology primarily because these phenotypes are easy to score using a dissecting microscope. This paper reports the genetic characterization of 31 mutants with dominant behavioral and/or morphological defects. MATERIALS AND METHODS Strains and genetic nomenclature: C. elegans var. Bristol strain N2 was the wild-type parent of all strains used in this study. Previously described genes and alleles that were used for mapping and in other experiments are listed below. (This list does not include those mutations generated in this study.) N 2 and most of the mutant strains have been EDCLEYand RIDDLE (1984). Other described by BRENNER(1974) and by SWANSON, mutations and strains we have used include the dominant LG I1 crossover suppressor, C l dpy-lO(e128) unc-52(e444) (HERMAN 1978); the translocations eT1 (Het') (ROSENBLUTH and BAILLIE 1981) and mnT2 (ZeX) (HERMAN, KARI and HARTMAN 1982); the LG I1 deficiencies, mnDf44, mnDf45, mnDf65, mnDf68, mnDf80, mnDf85, mnDf88, mnDf91, mnDf92, mnDf94, mnDf97, mnDfl00 (SIGURDSON, SPANIER and HERMAN 1984); unc-74(~19) Z and unc-38(~20)Z (LEWISet al. 1980); rol-6(su1006)ZZ (COXet al. 1980); unc-93(e1500 n234) ZZZ (GREENWALD and HORVITZ 1980); unc-l03@1597)ZlZ (J. HODGKIN, personal communication); unc-.58(e665 e21 12) (S. BRENNER,personal communication); let-405(~116) V, let-404(~119) V, let-402(~127)V and let-401(~193) V (T. ROGAL~KI and D. BAILLIE,unpublished results). LG I: bli-3(e767); unc-35(e259); lin-6(e1466); lin-l7(n677, n671); sup-1l(n403); dpy5(e61); dpy-l4(e188); unc-1 l(e47); unc-27(e1072); unc-l3(el091); unc-56(e403); unc54(e190). LG 11: dpy-IO(e128);tra-Z(n196); lin-5(e1348);zyg-1 l(b272);vub-9(el744);rol-6(e187); unc-4(e120);unc-52(e444);lzn-7(eI 4 13). LG 111: dpy-l7(e164);lon-l(e185);dpy-l8(e364);unc-78(e1068). LG I V dpy-l3(e184);unc-8(e49);unc-24(e138);unc-44(e362);bli-b(sc16);lin-3(e1417); dpy-ZO(e1362);unc-43(e266, e408, e755); unc-ZZ(e66);dpy-l(el166). LG V: unc-46(el77); unc-62(e644); unc-83(e1408); dpy-1I(e224); unc-70(e524); unc68(e540);unc-42(e270);sma-l(e30). LG X: unc-l(e50, e68, e94, e538, e580, e712, e719); dpy-3(e27);lon-Z(e678). This paper conforms to the standardized nomenclature for C. elegans genetics (HORVITZ et al. 1979). Apparent intragenic revertants are named as double mutants carrying the parental mutation and a new mutation generated by the reversion event, although the presence of the parental mutation in the revertant has not been proven. We adopted this nomenclature mainly because none of the revertants tested showed restoration of

DOMINANT MUTANTS OF C. ELEGANS

823

the wild-type function of its gene. However, it is possible that some of the revertants no longer carry the parental mutation and instead carry a small deletion. Genetic methods: General culture techniques have been described by BRENNER (1974).Mapping and complementation tests were done at 20" unless otherwise stated. Sources of mutants with dominant effects: All but one of the new mutants we isolated that have dominant effects were obtained from among the F1 progeny of N2 males or hermaphrodites mutagenized with ethyl methanesulfonate (EMS) (BRENNER 1974).Generally, five N 2 hermaphrodites were crossed with 10-15 mutagenized males on a Petri plate (6-cm diameter). T h e parents were transferred daily for 4 days to new Petri plates, and only one mutant of a particular phenotype was saved from each set of five Petri plates. A total of 115,000 mutagenized haploid genomes were screened (approximately 45,000cross-progeny of mutagenized males and approximately 35,000 self-progeny of mutagenized hermaphrodites). All mutants were crossed with N2 males to confirm dominance. From these heterozygous cross-progeny, animals homozygous for the mutations were then sought. All new mutations were segregated at least twice before further genetic analysis after crossing with either N2 males or males heterozygous (or hemizygous) for a marker. For those mutations that result in dominant visible and recessive lethal phenotypes, hermaphrodites heterozygous for the mutations were crossed with lon-2/0 hemizygous males. Mutant hermaphrodites from the cross were picked, and from the progeny of those segregating Lon progeny, mutant hermaphrodites were picked. These mutants were then maintained as heterozygotes. n489, n490, n491, n492, n493, n494, n495, n498, n499, n500, n501, n506, n715 (25")and n496 (20")were isolated from the progeny of mutagenized N2 males. n774, n775, n776 and n728 were isolated at 25" from the self-progeny of mutagenized N2 hermaphrodites. n1167 and n777 were isolated by M. FINNEY;nI166 was isolated by E. FERGUSON; n716 was isolated by V. AMBROS;and n773 was isolated by W. FIXSEN.All of these mutations except one (n1167) were isolated from the progeny of mutagenized N2 males, and n1167 was isolated from the progeny of mutagenized dpy-l9(e1259) males at 20".n1274 is a spontaneous mutation isolated from the N2 strain. Genetic mapping: Each mutation was mapped to a linkage group using six strains, each of which carried a recessive marker for one linkage group: dpy-5 I , dpy-10 ZI, dpy17 ZII, dpy-13 ZV, dpy-11 V, lon-2 X. New mutants were crossed with males heterozygous or hemizygous for a marker, and hermaphrodite progeny were picked. Potential crossprogeny (identified on the basis of their carrying the mapping marker) were transferred daily to a new Petri plate, and their progeny were counted. Assignment to a linkage group was based on the ratio of progeny with a dominant phenotype only to progeny with a marker phenotype only. When a mutation resulting in dominant phenotype was linked to a marker, the segregation ratio of dominant nonmarker to marker nondominant approached a 3:l ratio; in contrast, when a mutation resulting in a dominant phenotype was unlinked to a marker, the ratio was about 9:1. Once a linkage group was assigned, each mutation was mapped further by threefactor crosses, using two different protocols, which differed from the standard three1974). In protocol 1, anifactor protocols used to map recessive mutations (BRENNER mals of genotype a/bc were constructed (a represents a new mutation with a dominant effect, and 6 and c represent recessive marker mutations), and among the FI progeny those recombinants not showing the dominant phenotype [i.e., B, C or wild type ("WT")] were scored. There are three possible sets of recombinants to be obtained, depending on the relative map positions of the three genes: WT and B recombinants but no C recombinants are expected if the order is abc, B and C recombinants but no W T recombinants are expected if the order is bac and W T and C recombinants but no B recombinants are expected if the order is bca. T h e presence of two types of recombinants (e.g., B and W T in the first example) and the absence of the other type of recombinant (e.g., C in the first example) implied an unambiguous map order. The relative numbers of the two types of recombinants can be used to determine the relative distances between three markers; for example, when the order is abc, the ratio of the

824

E.

0

2 5 5c.; 7 p

2 2e

2$5

aO.O.c


845

by the following protocol: the mnT2/0 males were crossed to n494 lon-2 hermaphrodites. Among the non-Lon hermaphrodite progeny were animals like n494 homozygotes and animals like n494/+ heterozygotes. Animals displaying an n494/+ phenotype segregated both Unc and wild-type males, showing that these animals carried mnT2 chromosome (i.e., genotype of n494/mnT2). n494 homozygote-like animals segregated Unc Lon males but no wild-type males, showing that they carried the m n T l l chromosome (i.e., genotype of n494/ mnTl1). m n T l l behaved like unc-1 (e538) and unlike unc-l(ell4) when in trans to other alleles of unc-1, as shown in Table 6. These results suggest that the loss-of-function mutations are the recessive unc-Z alleles that enhance the n494 phenotype (class IV) rather than those that suppress the n494 phenotype (class 111). This conclusion is supported by the relative frequency of obtaining these two classes of mutations: there are six previously isolated unc-1 alleles that failed to complement n494, and there is one allele that suppresses dominantly; the apparent loss-of-function mutations were generated at a frequency of 2 X 10-4 (heterozygous reversion of e1598 and n774), whereas suppressor muta(as determined by isolating tions were generated at a frequency of 4.4 X dominant suppressors of n494 in the F1 generation from mutagenized n494 homozygotes). Since n494/mnTll animals display a more severe mutant phenotype than n494/+ animals, class I unc-1 alleles appear to cause the acquisition of a novel function. T h e interpretation of class I1 alleles is more complicated. n774/ mnTl1 animals display a curly-Unc phenotype, and n774/+ animals display a coiler-Unc phenotype. Since n774/n774 animals display a curly-Unc phenotype, n774 cannot simply result in an increase in normal unc-1 function. We propose two alternative interpretations of the mode of dominance of the class I1 alleles. On the one hand, a class I1 allele could antagonize a wild-type allele, so that a class ZZ/+ animal (coiler-Unc) has less wild-type function than does a null/+ animal (wild-type), and a class ZZ/null animal (curly-Unc) has less wild-type function than does a class ZZ/+ animal (coiler-Unc) and, possibly, as little wildtype function as does a nulllnull animal (curly-Unc). Alternatively, a class I1 allele may have lost its wild-type function while acquiring a novel function, so that a class ZZ/+ animal (coiler-Unc) displays a novel phenotype, but a class IZ/ null animal (curly-Unc) displays a phenotype that results from the loss of function being epistatic to the novel function. Both of these models interpret class I1 mutations as resulting in the acquisition of a novel function. In both cases, we assume that the apparent “intragenic complementation” shown by various pairs of unc-1 alleles-for example, between n494 and e114 and between n1167 and n1166-is a consequence of specific interactions between unc-l gene products. This apparent intragenic complementation suggests that the unc-1 gene product may function in a multimeric form (CRICKand ORGEL1964). unc-58(n495, e665) X: We assigned n495 as allelic to unc-58(e665) (BRENNER 1974) since n495 mutants and e665 mutants showed similar phenotypes, and an apparent intragenic revertant of n495 failed to complement an apparent intragenic revertant of unc-58(e665).

846

E.-C. PARK AND H. R. HORVITZ


847

We isolated one spontaneous revertant of n495 (n495 n1273) from n495 homozygotes. This animal showed an Unc phenotype, but was larger and much more mobile than n495 animals. T h e new mutation was tightly linked to n495 and suppressed n495 in cis: phenotypically wild-type hermaphrodites were found when 72495 n1273 hermaphrodites were crossed with N 2 males, and n495 nl274/++ animals failed to segregate n495-like progeny among approximately 9000 self-progeny, indicating that the new mutation was tightly linked to n495 (the recombination frequency was less than 0.02% ( p I1/9000). S. BRENNER(personal communication) and GREENWALD and HORVITZ (1 980) isolated apparent intragenic revertants of e665 by EMS, and we showed that one of these revertants (e665 e2112) failed to complement the Unc phenotype of n495 n1274: when n495 n1274/0 males were crossed to e665 e2112 hermaphrodites, all the progeny, both males and hermaphrodites, were like 72495 n1274 (or e665 e2112) homozygotes. Based on this complementation result, we assigned n495 to unc-58. Summary Table 7 summarizes the phenotypes caused by various alleles of the ten genes we have characterized. DISCUSSION

We have isolated 24 mutations that have dominant effects on the behavior and/or morphology of C. elegans and have studied several additional mutations with dominant effects isolated by others. These mutations appear to define 15 genes. Five of these genes (unc-108, unc-109, unc-105, sma-8 and egl-36) are newly identified; two (unc-8 and unc-43) previously had been defined only by recessive mutations; and eight (egl-30, unc-54, rol-6, unc-103, bli-6, unc-70, unc1 and unc-58) previously had been assigned mutations with dominant effects. We have identified two or more mutations with dominant effects in four of these 15 genes. In most of the cases we have examined, mutations with dominant effects do not appear to be null mutations. Furthermore, in most cases the null phenotypes appear to be different from the dominant phenotypes: four genes (unc105, rol-6, unc-103 and unc-8) appear to have wild-type null phenotypes, and three genes (unc-43, unc-70 and unc-1) appear to have null phenotypes that are recessive and unlike the dominant phenotypes. We have determined the modes of dominance of 18 mutations in ten genes by analyzing the behavior of mutations with dominant effects when present in trans to apparent null mutations and to the wild-type allele of the gene. Mutations with dominant effects in nine of these genes (egl-30, unc-54, unc-105, rol-6, unc-8, unc-70, unc103, unc-43 and unc-1) appear to result in the expression of novel functions, since when in trans to apparent null mutations of these genes the dominant phenotypes are enhanced as compared to when in trans to wild-type alleles. In addition, the heterozygous mutant phenotypes of two of these nine genes (unc54 and egl-30) resemble the phenotypes caused by loss-of-function mutations

848

E.-C.

PARK AND H.

R.

HORVITZ

of these genes, indicating that, in these cases, dominance is a result of an antagonistic effect of the mutant gene products on the activities of the wildtype products. We failed to identify revertants of u n c - l 0 8 ( n 5 0 1 ) after an extensive search, suggesting either that the null phenotype of this locus is dominant lethality or that n501 is itself a null allele. In either case, two copies of the unc-108(+) allele must be necessary for the wild-type phenotype. Dominant null mutations (e.g., in Minute genes and in the Notch locus) have previously been identified in Drosophila (LINDSLEYand GRELL 1968; LINDSLEYet al. 1972). In each of the genes defined by multiple mutations with dominant effects, all mutations show the same mode of dominance. Thus, a single most common mode of dominance is characteristic of each of these genes. This observation suggests that the wild-type function of a gene may determine the modes of dominance of mutations that may be obtained. For instance, a gene with a product that functions within a multimeric complex could have mutations that cause a dominant inactivation of the complex. However, more complex possibilities exist. For example, the E. coli l a d gene displays two different modes of dominance, lacZ" and l a d d (JACOB and MONOD1961; WILLSONet al. 1964; DAVIESand JACOB 1968). A priori one might expect that any gene might be mutated to result in a dominant overproduction of a normal gene activity. However, in our studies we did not identify any such mutations. Mutations of this class have been identified in other C. elegans genes: lin-12 (GREENWALD, STERNBERG and HORVITZ1983), lin-14 (AMBROSand HORVITZ1984), tra-l (HODGKIN1983) and her-l (TRENT,TSUNG and HORVITZ1983). T h e apparent absence of such mutations in our study may suggest that the functions of genes involved in determining the phenotypes we have examined are relatively insensitive to increased levels of gene activity; alternatively, these genes we have studied may be relatively difficult to mutate to cause a substantial increase in gene activity. T h e frequent occurrence of mutations with dominant effects in certain genes may reflect certain structural and functional characteristics of their gene products. In the case of unc-1, there are a number of types of interactions between alleles of the gene (see Table 6). These interactions suggest that the unc-1 product functions in multimeric forms. Similarly, several mutations with dominant effects in C. elegans have been identified in genes encoding structural et al. proteins of muscle, such as a myosin heavy chain (unc-54) (MACLEOD 1977), paramyosin (unc-15) (WATERSTON, FISHPOOLand BRENNER1977) and actin (unc-92) (WATERSTON, HIRSHand LANE 1984). These data suggest that a susceptibility to mutations with dominant effects may be a characteristic of genes encoding proteins of multimeric assemblies. Seven of the genes we have studied are defined only by mutations with dominant effects. This characteristic of these genes probably reflects their null phenotypes. Of the four of these seven genes for which apparent null phenotypes have been determined, three (unc-105, unc-8, unc-103) appear to have wild-type null phenotypes; in these cases, mutations causing a reduction or loss of gene function would not be expected to result in a mutant phenotype. One


849

gene (unc-108) appears to have a dominant null phenotype, and thus, mutations causing a reduction or loss of gene function would be expected to be dominant. A significant portion of the genes we have identified as susceptible to mutations with dominant effects have wild-type null phenotypes. If these genes represent a random sample from the genome, we would estimate that null mutations in about half of the genes in C. elegans would result in a nonmutant phenotype. T h e functional bases of wild-type null phenotypes are not well understood. Possible explanations for apparent wild-type null phenotypes include (1) the gene in question is a member of functionally redundant gene family or is involved in a functionally redundant pathway (SUZUKI1970; GREENWALD and HORVITZ1980; LANDELet al. 1984); (2) the loss of function of the wild-type allele leads to an abnormal phenotype not observed by our methods; and (3) the locus is normally silent, and mutations having dominant effects act by causing its expression (the mutationally induced expression of the silent gene suc 2 " of yeast is an example of this class; CARLSON, OSMOND and BOTSTEIN 1981). In this third case, the wild-type allele should behave indistinguishably from apparent null alleles. Such behavior is not consistent with our data, as a heterozygote between the wild-type allele and the mutant allele of each gene we have studied shows a phenotype between that of the wild type and the mutant homozygote. In yeast, genes with wild-type null phenotypes that reflect functional redunet al. 1979; POWERSet al. dancy within gene families are known (HEREFORD 1984; DEFEO~ONES et al. 1983); for example, rasl and ras2 (TATCHELL et al. 1984) and the genes encoding histones H2A and H2B (RYKOWSKI et al. 1981). In C. elegans, a mutation in either the ace-I gene or the ace-2 gene does not cause an obvious phenotype, whereas ace-1 ; ace-2 double mutants have a strong Unc phenotype (CULOTTIet al. 1981; JOHNSON et al. 1981). Mutations with dominant effects have been described in three other C. elegans genes that appear to have wild-type null phenotypes, unc-93 (GREENWALD and HORVITZ 1980), unc-92 (WATERSTON, HIRSHand LANE 1984) and sup-7 (WATERSTON 1981). unc-92 is one of four actin genes found in C. elegans (LANDELet al. 1984), and sup-7 is a tRNA gene (BOLTONet al. 1984). Thus, the four genes we have found to have wild-type null phenotypes may also represent members of multigene families. However, mutations in these genes, unlike mutations in the unc-92 gene, are semidominant, and apparent null mutations enhance their phenotypes. Although mutations in members of a simple multigene family could have dosage-dependent dominant effects, null alleles in this case would not be expected to enhance the dominant phenotype. Indeed, in the case of unc-92 (WATERSTON, HIRSH and LANE 1984), apparent null alleles behave indistinguishably from the wild-type allele when tested in heterozygotes bearing dominant alleles. However, genetic behavior resembling that of unc-92 is not necessarily expected of multigene families having distinctly regulated or nonidentical members. Our finding of a wild-type null phenotype for rol-6 is unusual in that this gene is represented both by alleles with dominant effects and by a comparable number of alleles with recessive effects (COX et al. 1980). Thus, not all mu-

850

E.-C. PARK AND H. R . HORVITZ

tations in a gene with a wild-type null phenotype necessarily have dominant effects. The rol-6 mutations with no apparent dominant effects appear to define a distinct class of mutations with recessive effects. These mutations appear unlikely to result in a reduction or loss of gene function. We thank L. AVERY,D. BOTSTEIN,E. FERCUSON, G. FINK,M. FINNEY,W. FIXSEN,E. LANDER, B. MEYER,D. STINCHCOMB, J. THOMAS and especially D. FINLEYand R. HERMAN for comments concerning the manuscript. This work was supported by United States Public Health Service research grants GM24663 and GM24943 and Research Career Development Award HD00369 to H.R.H. E.-C.P. was supported in part by a Biogen Fellowship granted to the Department of Biology, Massachusetts Institute of Technology. LITERATURE CITED

AMBROS,V. and H. R. HORVITZ,1984 Heterochronic mutants of the nematode Caenorhabditis elegans. Science 226: 409-416. D. A. FISCHHOFF and R. H. WATERSTON, 1984 The sup-7(st5) BOLTON,S. L., P. POWELL-ABEL, X gene of Caenorhabditis elegans encodes a tRNATrP(UAA)amber suppressor. Proc. Natl. Acad. Sci. USA 81: 6784-6788. S., 1974 The genetics of Caenorhabditis elegans. Genetics 77: 71-94. BRENNER,

BUSSON,D., M. GANS,K. KOMITOPOULOUS and M. MASSON,1983 Genetic analysis of three dominant female-sterile mutations located on the X-chromosome of Drosophila melanoguster. Genetics 105: 309-325. CARLSON,M., B. C. OSMOND and D. BOTSTEIN,1981 Genetic evidence for a silent SUC gene in yeast. Genetics 98: 41-54.

Cox, G. N., J. S. LAUFER,M. KUSCH and R. S. EDGAR,1980 Genetic and phenotypic characterization of roller mutants in C . elegans. Genetics 95: 317-339. CRICK,F. H. and L. E. ORCEL,1964 The theory of inter-allelic complementation. J. Mol. Biol. 8: 161-165. CULOTTI,J., G. VON EHRENSTEIN, M. R. CULOTTIand R. L. RUSSELL, 1981 A second class of acetylcholinesterase-deficient mutants of the nematode Caenorhabditis elegans. Genetics 97: 281-305. DAVIES, J. and F. JACOB,1968 Genetic mapping of the regulator and operator genes of the lac operon. J. Mol. Biol. 36: 413-417. DEFEO-JONES, D., E. SCOLNICK, R. KOLLERand R. DHAR,1983 ras-related gene sequences identified and isolated from Saccharomyces cermisiae. Nature 3 0 6 707-709. DENNELL, R. E., 1978 Homeosis in Drosophila. 11. A genetic analysis of polycomb. Genetics 90: 277-289. and S. BRENNER, 1974 A mutant affecting the heavy chain of myosin EPSTEIN,H., R. WATERSTON in Caenorhabditis elegans. J. Mol. Biol. 90: 291-300. E. and H. R. HORVITZ,1985 Identification and characterization of 22 genes that FERGUSON, affect the vulval cell lineages of the nematode Caenorhabditis elegans. Genetics 110: 17-72.

GREENWALD, I. and R. HORVITZ,1980 unc-93(eI500) 111: a behavioral mutant of Caenorhabditis elegans that defines a gene with a wild-type null phenotype. Genetics 96: 147-164. GREENWALD, I., P. STERNBERC and R. HORVITZ,1983 The lin-12 locus specifies cell fates in C. elegans. Cell 34: 435-444. L., K. FAHRNER, J. WOOLFORD, JR., M. ROSBASH and D. B. KABACK, 1979 Isolation HEREFORD, of yeast histones H2A and H2B. Cell 18: 1261-1271.


851

HERMAN, R., 1978 Crossover suppressors and balanced recessive lethals in Caenorhabditis elegans. Genetics 88: 49-65. HERMAN, R. and R. HORVITZ,1980 Genetic analysis of Caenorhabditis elegans. pp. 227-261. In: Nematodes as Biological Models, Edited by B. ZUCKERMAN.Academic Press, New York. HERMAN, R. K., C. K. KARIand P. S. HARTMAN, 1982 Dominant X-chromosome nondisjunction mutants of Caenorhabditis elegans. Genetics 102: 379-400. HODGKIN, J., 1983 T w o types of sex determination in a nematode. Nature 3 0 4 267-268. HORVITZ,R., S. BRENNER, J. HODGKIN and R. HERMAN, 1979 A uniform genetic nomenclature for the nematode Caenorhabditis elegans. Mol. Gen. Genet. 175: 129-133. HORVITZ,R. and J. SULSTON, 1980 Isolation and genetic characterization of cell lineage mutants of the nematode Caenorhabditis elegans. Genetics 96: 435-454. F. and J. MONOD,1961 Genetic regulatory mechanisms in the synthesis of proteins. J. Mol. Biol. 3: 318-356.

JACOB,

C. D., J. G. DUCKETT, J. G. CULDTTI,R. K. HERMAN, P. M. MENEELY and R. L. RUSSELL, 198 1 An acetylcholinesterase-deficient mutant of the nematode Caenorhabditis elegans. Genetics 97: 261-279.

JOHNSON,

LANDEL,C. P., M. KRAUS, R. WATERSTON and D. HIRSH, 1984 DNA rearrangements of the actin gene cluster in C . elegans accompanying reversion of three muscle mutants. J. Mol. Biol. 180: 497-513. LEWIS,J., C.-H. Wu, H. BERGand J. LEVINE,1980 Genetics of levamisole resistance in the nematode Caenorhabditis elegans. Genetics 95: 905-928. LIFSCHYTZ, E. and M. M. GREEN,1979 Genetic identification of dominant overproducing mutations: the Beadex gene. Mol. Gen. Genet. 171: 153-159. LINDSLEY, D. L. and E. H. GRELL,1968 Genetic variation of Drosophila melanogaster. Carnegie Inst. Wash. Publ. 627. LINDSLEY, D. L., L. SANDLER, B. S. BAKER,A. T. C. CARPENTER, R. E. DENELL, J. C. HALL,P. A. JACOBS,G. L. G. MIKLOS,B. K. DAVIS,R. C. GETHMANN, R. W. HARDY,A. HESSLER, S. M. MILLER,H. NOZAWA, D. M. PERRYand M. GOULDSOMERO, 1972 Segmental aneuploidy and the genetic gross structure of the Drosophila genome. Genetics 71: 157-184.

MACLEOD,A. R., R. H. WATERSTON and S. BRENNER,1977 An internal deletion mutant of a myosin heavy chain in C. elegans. Proc. Natl. Acad. Sci. USA 74: 5336-5340. MACLEOD,A. R., R. H. WATERSTON, R. M. FISHPOOLand S. BRENNER, 1977 Identification of the structural gene for a myosin heavy chain in C . elegans. J. Mol. Biol. 114: 133-140. MULLER,H. J., 1932 Further studies on the nature and causes of gene mutations. pp. 213-255. In: Proceedings of the Sixth International Congress of Genetics, Edited by DONALDF. JONES. Brooklyn Botanic Gardens, Menasha, Wisconsin.

PARK,E.-C. and H. R. HORVITZ,1986 C . elegans unc-105 mutations affect muscle and are suppressed by other mutations that affect muscle. Genetics 113: 853-867. POWERS, S., T. KATAOKA, 0. FASANO,M. GOLDFARB, J. STRATHERN, J. BROACHand M. WIGLER, 1984 Genes in S. cerevisiae encoding proteins with domains homologous to the mammalian ras proteins. Cell 3 6 607-612. ROSENBLUTH, R. E. and D. L. BAILLIE,1981 The genetic analysis of a reciprocal translocation, eTl(III;V), in Caenorhabditis elegans. Genetics 9 9 4 15-428. RYKOWSKI, M., J. WALLIS,J. CHOEand M. GRUNSTEIN, 1981 Histone H2B subtypes are dispensible during the yeast life cycle. Cell 25: 477-487. SIGURDSON, D. C., G. J. SPANIER and R. HERMAN, 1984 C . elegans deficiency mapping. Genetics 108: 331-345.

852

E.-C. PARK AND H. R. HORVITZ

SULSTON, J. E. and H. R. HORVITZ,1977 Post-embryonic cell lineages of the nematode Caanorhabditis elegans. Dev. Biol. 56: 110-156. SUZUKI, D., 1970 Temperature-sensitive mutations in Drosophila melanogaster. Science 170: 695706. SWANSON, M. M., M. L. EDGLEY and D. L. RIDDLE,1984 Caenorhabditis elegans. pp. 286-299. In: Genetic Maps, edited by S. J. O’BRIEN.Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. TATCHELL, K., D. T. CHALEFF, D. DEFEO-JONES and E. M. SCOLNICK, 1984 Requirement of either of a pair of ras-related genes of Saccharomyces cerevisiae for spore viability. Nature 309 523527.

TRENT,C., N. TSUNG and H. R. HORVITZ,1983 Egg-laying defective mutants of the nematode Caenorhabditis elegans. Genetics 104: 61 9-647. WATERSTON, R. H., 1981 A second informational suppressor, sup7 X, in C. elegans. Genetics 97: 307-325. WATERSTON, R. H. and S. BRENNER, 1978 A suppressor mutation in the nematode acting on specific alleles of many genes. Nature 275: 715-719. WATERSTON, R., R. FISHPOOL and S. BRENNER, 1977 Mutants affecting paramyosin in Caenorhabditis elegans. J. Mol. Biol. 117: 679-697. WATERSTON, R. H., D. HIRSH and T. R. LANE, 1984 Dominant mutations affecting muscle structure in Caenorhabditis elegans that map near the actin gene cluster. J. Mol. Biol. 180: 473-496. WILLSON,C., D. PERRIN,M. COHEN,F. JACOB and J. MONOD, 1964 Non-inducible mutants of the regulator gene in the “lactose” system of Escherichia coli. J. Mol. Biol. 8: 582-592. Communicating editor: R. K. HERMAN