Wolinsky EJ (1990) Worm Breeder's Gazette "deg-1 Suppressors"

Wolinsky EJ

[Worm Breeder's Gazette, 1990]

Extragenic suppressors: I am seeking extragenic suppressors of deg-1( u38) by looking for tail touch sensitive revertants of a high-copy number, stable transformant strain. This strain (TU1191, Nature 345:410) arose after microinjection of a mixture of deg-1 dominant mutant u38 DNA and unc-22 antisense DNA into wild type oocytes. Genetically, the Deg and Twi phenotypes of this strain show X-linkage. By Southern blot analysis, at least ten intact copies of the u38 allele are present. The null phenotype of deg-1 is wild type, and after EMS mutagenesis of u38 animals, touch sensitive, intragenic revertants are readily isolated at the expected frequency for gene knock-out events. It is therefore tedious to look for rare extragenic reversion events against the background of higher frequency intragenic events. After mutagenesis of TU1191, however, touch sensitive progeny occur much less frequently. I chose to work with this strain because intragenic reversion in only one of the multiple copies of the u38 transforming DNA would be unlikely to restore tail touch sensitivity. The Deg phenotype of TU1191 is in fact partially suppressed by mec-6 mutations, thus I expect to be able to identify mutations in other genes which act as suppressors of u38. Unfortunately, EMS mutagenesis seems to stimulate excision of the transforming DNA, perhaps as a result of activation of DNA repair systems. Several touch sensitive revertants have been obtained which no longer express the Twi phenotype, and which fail to segregate the Deg phenotype when outcrossed. One candidate so far meets the criteria of tail touch sensitivity, and co-segregation of both the Twi and Deg phenotypes after outcrossing. Intragenic revertants: Most pseudo-wild type revertants of u38 behave like null alleles. I isolated one exceptional revertant, u38 u424, which confers a wild type phenotype to homozygotes, but which also acts as a semi-dominant suppressor of the original u38 mutation in trans. As shown in the graph, suppression is cold-sensitive ( animals were scored as gravid adults). Because the u38 and u38 u424 alleles both show dominant effects, I compared their DNA sequences around the region encoding the amino acid changes that Monica Driscoll and Martin Chalfie reported (E. Coast Worm Mtg.) constitute the dominant mec-4 mutations. They also reported that mec-4 and deg-1 predicted protein sequences are highly homologous. mec-4 dominant alleles and deg-1(u38) affect different neurons, but in each case affected cells undergo vacuolar degenerations which appear to be similar events. A comparison of the wild-type, u38, and u38 u424 sequences is shown below. In fact, u38 contains the same ala->val change described for mec-4(d). The double mutant u38 u424 retains this change, and an additional mutation, gly->arg is present nearby. The ala->val mutation has little effect on computer predictions of protein structure. The gly->arg change, however, reduces the hydrophobicity of the possible membrane spanning domain and removes a predicted turn in the peptide backbone. The dominance of the suppressor mutation suggests that the double mutant protein may interact with u38 molecules in a way that inactivates their cell killing activity, for example by disrupting complexes of deg-1 proteins. The gly->arg mutation is likely to change the shape of the protein at the junction between the extracellular space and the membrane, or even prevent its proper insertion into the membrane. Either effect might prevent association of deg-1 molecules. [See Figure 1]

Wolinsky EJ et al. (1988) Worm Breeder's Gazette "deg-1 is a large gene."

Chalfie M, Wolinsky EJ

[Worm Breeder's Gazette, 1988]

The deg-1(u38) mutation causes the PVC interneurons to degenerate and die during the late L3-early L4 stage. Thus, such mutants become insensitive to tail touch as they mature (other neurons, in the head, degenerate during the Ll stage). The phenotype conferred by u38 is dominant and temperature sensitive. We have obtained 26 pseudo-wild type revertants following EMS and transposition mutagenesis, including one deletion allele. These results suggest that the wild type function of deg-1 is dispensable, and that the u38 mutant gene product somehow 'poisons' specific neurons at specific times during development. We wish to learn more about the deg-1 gene product in order to understand the molecular basis of the u38 phenotype. As reported in the last gazette, we have cloned the gene using Tc1 tagging. Since then we have l) localized two insertions and one deletion genetically linked to deg-1 to two cosmids within a contig mapping in the center of X (cosmids and map kindly provided by A. Coulson and J. Sulston), 2) localized three different types of cDNA clones to the same two cosmids (cDNA library kindly provided by J. Ahringer and J. Kimble), and 3) begun sequencing our genomic and cDNA clones. A part of the relevant contig map is reproduced below. Our conclusions are based on Southern blots of DNA from various u38 revertants probed with the cosmids shown. We have subcloned three regions of cosmid C47C12 corresponding to three deg-1 linked DNA rearrangements. Two different Tc1 insertions map to adjacent fragments near the center of C47C12: a 3 Kb EcoRI-XbaI fragment, and a 2.3 Kb EcoRl fragment. Our deletion strain is missing DNA from an approximately 8.5 Kb EcoRl fragment hybridizing to both cosmids C47C12 and R02A8. We used our 3 deg-1 subclones to screen an approximately Ll-staged cDNA library (we have not yet detected deg-1 sequences on Northerns). Three different types of CDNA's were obtained. All three hybridize to the 2.3 Kb EcoRl fragment which is the site of one of the Tc1 insertions. Additional hybridizing sequences are found in the center of R02A8 for two cDNAs which differ in the amount of R02A8 sequences they contain (they also differ in that only one of the two contains the repetitive element of the C47C12 2.3 Kb EcoRl fragment described below. None of the cDNA's bind detectably to sequences outside the two cosmids. Thus, the R02A8 hybridizing cDNA's delineate a probable new region of deg-1 not previously identified by mapping DNA rearrangements in our revertant strains. We have begun sequencing our deg-1 cDNA and genomic clones. The largest cDNA is about 2.5 Kb, comprised of mostly R02A8 sequences. More of the C47C12 2.3 Kb EcoRl fragment is represented on another cDNA clone approximately 2 Kb in length. Our study of the structure of deg-1 has so far yielded three points of interest. l) The gene is large, extending from the center of C47C12 to the center of R02A8. It appears to be sandwiched between the vitellogenin and tubulin genes. 2) One type of deg-1 cDNA contains sequences separated on the cosmid map by more than 10 Kb, suggesting a possible large intron. 3) The 2.3 Kb EcoRl fragment from C47C12 hybridizes to 5 additional bands on high-stringency Southern blots of XbaI-EcoRI cut genomic DNA. The extra bands amy correspond to members of a gene family related to deg-1. The hypothesis of a deg- 1 gene family is a possible explanation for the non-essentiality of the gene. [See Figure 1]

Wolinsky EJ et al. (1987) Worm Breeder's Gazette "progress towards cloning the deg-1 gene."

Wolinsky EJ, Chalfie M

[Worm Breeder's Gazette, 1987]

At the Cold Spring Harbor worm meeting last May, we described the identification of a Tc1 element genetically linked to the deg-1 gene. We obtained putative insertion mutants of deg-1 by isolating revertants of the dominant allele u38 in a mut-2 background. Tc1 elements newly appearing in 3 such revertants were tested for genetic lineage with deg-1 using the flanking markers dpy-6 and dpy-7. Two putative excision strains were also obtained, by choosing phenotypically deg-1 animals from reversion strains maintained in mutator backgrounds. One particular Tc1 element is closely linked to the reversion site and is not observed in DNA from the corresponding excision strain. We have now cloned a 4.7 kb restriction fragment containing this Tc1 element. We plan to use transformation experiments to test whether the sequences we have cloned are part of the deg-1 gene. Since the deg-1 gene is defined by the dominant mutation u38, we will microinject DNA cloned from mutant u38 animals. We have therefore prepared an EMBL3 lambda library from u38 DNA, which we are currently probing with our genomic clone. Phage DNAs from the lambda library which cross-hybridize with sequences flanking the cloned Tc1 element will be microinjected into oocytes of wild-type animals to see if expression of the Deg-1 phenotype can be induced.

Wolinsky EJ et al. (1988) Worm Breeder's Gazette "the deg-1 gene may encode a membrane protein."

Wolinsky EJ, Chalfie M

[Worm Breeder's Gazette, 1988]

The deg-1(u38) mutation results in swelling and degeneration of certain neurons (including the PVC cells, whose deaths impart a tail touch insensitive phenotype). The swelling of the affected cells suggests some type of membrane defect which eventually allows water to enter the cell. We have found that the deg-1 gene product may in fact be an integral membrane protein. This conclusion is based on cDNA sequence analysis (described below). Localization of the deg-1 gene product to the cell membrane is consistent with the idea that the deg- 1 induced degenerations result from a membrane defect. We have been sequencing three cDNA's obtained from the Ahringer/Kimble 'L1' library. All three hybridize to the same genomic fragment of the deg-1 gene under stringent conditions. The three cDNA's appear to represent transcripts which have undergone different sets of splicing events. All.three contain long stretches of identical sequence, which supports the conclusion that they represent transcripts of the same gene, rather than transcripts of other members of a hypothesized deg-1 gene family (see our abstract in last WBG). One of the cDNA's, described in June at the Boston meeting, contains a deduced open reading frame encoding at least 340 amino acids. Hydrophobicity plots of the translated protein indicate two possible membrane spanning segments, located between residues 191 to 231 and residues 275 to 316. Eleven of the fifteen cysteine residues present in the sequence are located in the first 191 residues, upstream of the first hydrophobic region. It is possible that the first 191 amino acids are part of an extracellular domain, since intracellular conditions do not favor cross-linking of cysteines. This cDNA does not contain a poly A tail. The second cDNA may be the most likely to represent a mature transcript, since it contains a poly A region of about sixty bases following a poly A addition site. This cDNA contains an open reading frame encoding 338 amino acids, which is nearly identical to that of the first cDNA, with two important exceptions. The first difference is a 72 base insertion that would result in a novel 24 amino acid sequence following amino acid 176 of the protein encoded by the first cDNA. Second, the termination codons and poly A tail of the second cDNA occur at the end of the second possible membrane spanning region encoded by the first cDNA. This would result in a protein at least twenty five amino acids shorter than that encoded by the first cDNA. The two cDNA's are probably identical at their 5' ends. However, it is possible that both are incomplete due to the usual imperfections of cDNA libraries. The third cDNA appears to represent an incompletely spliced transcript. It contains relatively short open reading frames ( some of which encode sequences found in the proteins translated from the first two cDNA's). These are flanked by A/T rich intron-like regions. Consensus splice sites can be found at several intron/exon junctions. Comparison of the sequences of the three cDNA clones suggests that deg-1 transcripts may be processed to produce different splicing patterns. In the case of the first two cDNA's, splicing differences would result in different exon usage and different protein products. It is also possible that the differences between the cDNA's arose as artifacts of library construction or subsequent cloning events. We plan to perform RNA protection experiments to see if multiple forms of deg-1 transcripts can in fact be isolated from wild type worms and from deg-1 mutants. The putative deg-1 protein sequences do not show significant homology to proteins in the data bases screened so far.

Shreffler W et al. (1995) Worm Breeder's Gazette "Embryonic Expression of unc-8 Suppressor sup-42."

Shreffler W, Wolinsky EJ

[Worm Breeder's Gazette, 1995]

Embryonic Expression of unc-8 suppressor sup-42 Wayne Shreffler and Eve Wolinsky Dept. Biochemistry, NYU Medical School, *0 First Ave., NY 10012 Strong unc-8 (IV) dominant mutations produce ventral motorneuron swelling, coiling, and inability to back. These phenotypes are suppressed by mec-6 mutations (WBG13(3):77, Genetics, in press), suggesting afunctional similarity to members of the deg-1 ion channel family. The strong unc-8 phenotype allows ready isolation of suppressor mutations that restore the ability to move backwards. One such mutation, sup-42(lb88) X, obtained after EMS mutagenesis of unc-8(n491), recessively restores backing ability and largely suppresses the motorneuron swelling phenotypes of unc-8 strong dominant mutations, although it does not restore fully wild-type locomotion. Cultures of unc-8(n491); sup-42(lb88) double mutants also grow surprisingly slowly despite this amelioration of the Unc-8 phenotype. sup-42(lb88) in the absence of other mutations confers an extremely mild sluggish Unc phenotype. The slow growth of the double mutant, may, however, be explained by embryonic lethality of sup-42(lb88). 25 percent of sup-42(lb88) eggs and 44percent of unc-8(n491);sup 42(lb88) eggs fail to hatch, as compared to 1percent of wild-type strain N2 and 8percent of unc-8(n491) single mutant eggs. At least 10percentof both early and late sup 42(lb88) and unc-8(n491);sup-42(lb88) embryos contain large swollen cells. These embryonic swellings could be due to an osmotic defect or reflect secondary leakiness of cells dying from other causes. Larval lethality is comparable to wild-type; with the exception of unusual sickly newly-hatched animals, cell swel!ing is not observed in larvae or adults. Thus, sup-42 appears to function during embryogenesis, as well as interacting with unc-8 in the nervous system after hatching. sup-42(lb88) suppression of unc-8(n491) in a mec-6(+) background argues against functional redundancy with mec-6. sup-42 might be a functional homolog of the deg-1 family, in which case suppression of sup-42(lb88) embryonic lethality by mec-6 is predicted. Interactions of sup-42 with other deg-1 family members and unc-8 suppressor mutations are being tested. Perhaps sup-42(lb88) will prove to identify a novel channel regulator. In order to clone unc-8 and its hypodermally-active suppressor gene sup 40 (I), we have been injecting nearby cosmids. Both genes are identified by dominant mutations with deleterious effects, therefore a multi-copy array of wild type DNA may be harmful. It has proven, in fact, extremely difficult to obtain any transformants (as marked by coinjected antisense Roller plasmid), and no stably transmitting arrays have yet been obtained from either the LGI or LGIV cosmids injected as a group. We are currently trying to identify an individual "killer" cosmid from each group (which may, however, encode a harmful sequence unrelated to sup-40 or unc-8).

Shreffler W et al. (1995) Worm Breeder's Gazette "Genes Controlling Ion Permeability in C. elegans"

Wolinsky EJ, Shreffler W

[Worm Breeder's Gazette, 1995]

Genetic data suggest that unc-8 IV encodes a subunit of a Na+ channel homologous to the passive leak Na+ channels subserving transepithelial salt and water transport in mammals. Physiological, biochemical, and genetic experiments indicate that the ion channel is a multi-subunit structure, of which two mammalian components have been molecularly identified: ENaC's and CFTR (Stutts et. al., Science 269:847). To identify other genes encoding proteins with important roles in ion channel function, extragenic unc-8 suppressor and enhancer mutations were sought. Two unc-8 suppressor loci, sup-40 I and sup-41 IV were described previously (Shreffler et. al. Genetics 139:1261). We describe here gene interactions of two new unc-8 supressor loci, sup-42(lb88) X and sup-43(lb141) II, and an enhancer locus, enu-2(lb140). Sup-43(lb141) is an unusual allele specific suppressor of unc-8(e49). Single mutants of either sup-43(lb141) or unc-8(e49) express a coiler Unc phenotype, while the double mutant sup-43(lb141);unc-8(e49) behaves as wild-type. These three mutations were initially isolated by the sole criterion of amelioration or exacerbation of unc-8 locomotion defects, however, each displays unselected cell swelling phenotypes which further suggest an effect on membrane permeability: sup-42(lb88) causes swelling of embryonic cells, while sup-43(lb141) and enu-2(lb140) cause vacuoles within body wall muscle, similar in appearance to those caused by unc-105(n490). Waterston et. al. have shown that unc-105 is a member of the ENaC gene family. The phenotypic similarities between sup-43(lb141), enu-2(lb140) and unc-105(n490) are underscored by the observation that each exhibits striking synthetic lethality in combination with a fourth mutation, ndg-4(lb108). Ndg-4(lb108) confers strong resistance to nordihydroguairetic acid (NDG), a non-specific lipoxygenase inhibitor; dominant unc-8 mutations are moderately NDG resistant. Strikingly, many unc-105(n490);ndg-4(lb108) mutant embryos are severely vacuolated, while embryonic vacuoles are not observed in either single mutant. While the biochemical basis of ndg-4(lb108) drug resistance is not yet understood, interactions between ndg-4(lb108) and mutations in three different genes causing vacuoles within body wall muscle strongly suggest some commonality of action. Interactions with ndg-4 are particularly intriguing in the light of reports of effects of lipo- and epoxygenase inhibitors on CFTR (e.g., Kersting et. al. PNAS 90:4047). We speculate that unc-8 suppressor mutations may act by up-regulating an endogenous eicosanoid inhibitor of ENaC activity. We are currently analysing a novel unc-8 suppressor strain which expresses a pale egg phenotype identical to that of ndg-4. The gene expression patterns inferred from single and double mutant phenotypes in this and previous work suggest that overlapping sets of gene products are available for construction of Na+ leak channels in motorneurons, and body wall muscle, and early development. Phenotypic effects are exerted in motorneurons by unc-8(dom.), sup-40-43, and mec-6(l-o-f), in body wall muscle by unc-8(dom. or null), sup-43(lb141), enu-2(lb140), unc-105(n490), and ndg-4(lb108), and in developing eggs or embryos by sup-40(lb130), sup-42(lb88), sup-43(lb141), enu-2(lb140), unc-105(n490), and ndg-4(lb108).

Shreffler W et al. (1994) Worm Breeder's Gazette "The unc-8 Suppressor Locus sup-40 Regulates Ion Channel Function in Motorneurons, Hypodermis and Developing Oocytes"

Wolinsky EJ, Hsu M, Shreffler W

[Worm Breeder's Gazette, 1994]

The unc-8 suppressor locus sup40 regulates ion channel function in motorneurons, hypodermis and developing oocytes. Wayne Shreffler, Mei Hsu, and Eve Wolinsky Dept. Biochemistry, N.Y.U. Medical School, 550 First Avenue, New York NY 10016 Maintenance of osmotic balance is a vital cell function. Cell volume and membrane potential both depend on proper ion distribution across the plasma membrane. Specialized functions of some cell types, such as neuronal signalling and trans-epithelial transport are also critically dependent on electrochemical gradients. Mutational analysis can be used to identify new ion channel genes and to examine the relationship between ion channel structure and function. We devised a drug-resistance screen which identifies such genes in C elegans. Both unc-8 and its newly-identified dominant suppressor locus sup40 (LGI) can mutate to confer increased survival after exposure to a toxic level of nordihydroguairetic acid (NDG), a non-specific lipoxygenase inhibitor shown to block opening of Aplysia S-K+ channelsl. A separate class of NDG-resistant mutants, identifying at least two other genes, was found, which are behaviorally normal but which exhibit excess liquid in the pseudocoelom of adults, abnormally swollen coelomocytes, and leaky sterility. Dominant mutations of the unc-8 gene cause defective locomotion2. We observed that unc-8(elS) and n491 mutations also cause motorneuron swelling, suggesting an osmotic defect. Dominant missense mutations in three previously isolated genes, degl, mec4 and mec-105, also cause neuronal swelling. degl, mec-4, and mec-10 are members of a gene family with significant sequence6,7 and functional8 homologies to the amiloride-sensitive epithelial sodium channel of mammals. We also found that both the uncoordinated locomotion and swollen motorneuron phenotypes of unc-8 mutants are fully suppressed by mutations of the mec-6 gene. mec-6 mutations also suppress the dominant cell-swelling phenotype caused by degl, mec-4, and mec-10 mutations3,5. We found as well that, like the degl locus, unc-8 is a dispensable gene9 and exhibits interallelic interactions suggesting that its product functions as a multimer . Furthermore, we isolated a dominant suppressor mutation of unc-8(n491) which cross suppresses degl(u38). Thus, by genetic criteria, unc-8 appears to encode a new member of the degl family of ion channel components. We also identified a different unc-8 suppressor locus, sup40 (LGI), that can mutate to dominantly suppress the unc-8 motorneuron defect. Thus, sup-40 appears to regulate motorneuron ion channel function. sup-40(1bl30) also confers three recessive phenotypes: NDG resistance, swelling of hypodermal nuclei and almost complete disruption of oogenesis. Those oocytes which do form are abnormally large, misshapen, and frequently burst apart, as if under internal pressure. These recessive phenotypes suggest that sup40 is involved in control of membrane permeability in hypodermis and oocytes as wdl as in motorneurons. mec-6 mutations do not suppress the recessive sup40 phenotypes, nor does the sup-40(lbl30) cross-suppress dcg-l(u38). The drug-resistance of unc-8 dominant mutations and of the sup40(1bl30) mutation suggest that NDG may deleteriously alter the functioning of the wild-type channel. The biochemical mechanism of NDG toxicity in C clegans is not yet known, however, NDG can act as a lipoxygenase inhibitor in other systems. It is thus possible that lipoxygenase products of arachidonic acid, such as HPEI'E or its metabolites may affect channel function, as they have been reported to do at S-type K+ channels of Aplysia1. NDG does affect the C elegans nervous system: brief exposure of wild-type worms to NDG induces egg-laying, and unc-8 dominant mutants display significantly elevated basal egg-laying as compared to wild-type. This phenotype is also sup40-suppressible. Thus, the unc-8 ion channel and its sup40 partner are components of egg-laying as well as locomotory circuitry. Although many questions remain about the actions of NDG in C elegans, we have found NDG-resistance to be a useful criterion for expanding the collection of mutations affecting neuronal signalling and cell membrane permeability. 1. Neuron 101079. 2. Genetics 77:71. 3. Nature 34S:410. 4. Nature 349588. 5. Nature 367:467. 6. Nature 367:463. 7. Nature 361:504 8. Nature 367:467. 9. Genetics 113:821.

Shreffler W et al. (1994) Worm Breeder's Gazette "Enhancers of unc-8 Mutations."

Shreffler W, Wolinsky EJ, Shekdar K

[Worm Breeder's Gazette, 1994]

Enhancers of unc-8 mutations.

Tavernarakis NN et al. (1996) Worm Breeder's Gazette "unc-8 revealed"

Wang SL, Driscoll MA, Tavernarakis NN, Shreffler W

[Worm Breeder's Gazette, 1996]

Degenerins are proteins postulated to encode Na+ ion channel subunits that can mutate to induce necrotic-like cell death. Touch cell degenerins mec-4 and mec-10 are hypothesized to be subunits of a mechanically-gated ion channel expressed in the touch receptor neurons. Several !new! degenerin family members have been identified by the C. elegans sequencing project. We are interested in deciphering the functions of these degenerins. Cosmid R13A1 on chromosome IV harbors a predicted gene (R13A1_4) with a high degree of similarity to degenerins (62%). We analyzed the spatial and temporal pattern of R13A1_4 expression by generating stable transgenic strains harboring fusions of candidate 5! regulatory regions of R13A1_4 to lacZ and GFP. R13A1_4 is expressed in about 35 neurons that have cell bodies situated primarily in the ventral cord and nerve ring. Staining is most intense in L1 and L2 animals, although it is evident at late embryonic stages and persists into adulthood. The neuronal processes of some stained cells are clearly detectable, which has facilitated identification of some of the cells that express this degenerin family member. The PDA, PDB, DA, DB, DD, VD, VC, VB, HSN, ASH, BDU, SABD, PVM, PVC, AVB, AVA, AVD, AVE, AVJ, AVG, AIB, AIN and AIZ neurons express R13A1_4placZ and R13A1_4pGFP. Additional neurons in the nerve ring also strongly express this gene. The posterior touch receptor neurons (PLMs) clearly do not express these fusions. To get hints about the potential mutant phenotype of R13A1_4, we fused R13A1_4 regulatory regions to the toxic mec-4(d), so that the cells in which this new degenerin family member is expressed undergo degeneration. These ablation studies help both in identification of the cells in which a degenerin is expressed and in determination of possible mutant phenotypes (provided that the degenerin is needed for cell function). Animals harboring the R13A1-4 ablation fusion are extremely uncoordinated. The most severely affected animals are nearly paralyzed, cannot back up and coil when prodded with a pick. Transgenic animals also appear to be touch abnormal and somewhat hypercontracted at the posterior.. When we compared the genetic and physical maps we realized that R13A1 mapped in the vicinity of unc-8. unc-8 had been proposed to encode a member of the degenerin gene family (Shreffler et al., 1995) because dominant gain-of-function mutations cause swelling of ventral cord neurons, a phenotype suppressed by mutations in mec-6 (which also suppress swelling and death induced by toxic alleles of degenerin family members mec-4, mec-10 and deg-1). The phenotype of the R13A1_4pmec-4(d) ablation fusion line very closely resembles the phenotype of the semi-dominant unc-8(n491) allele. This ablation phenotype is fully suppressed by mec-6 mutations. Several lines of evidence support that R13A1_4 is unc-8. First, injection of cosmid R13A1 into the semi-dominant unc-8(e15) background strongly suppressed the Unc phenotype in transgenic animals (E. Wolinsky and W.S., unpublished results). Second, we PCR amplified fragments including R13A1_4 from both wild type and from semi-dominant unc-8(n491) and found that when the PCR products were injected into N2 animals, only transgenic lines containing amplified unc-8(n491) DNA were uncoordinated. Third, we identified Tc1 transposon insertions within R13A1_4 in transposon-tagged unc-8 strains (provided by E. Wolinsky and W.S.). Coupled with the expression pattern of R13A1_4, these data strongly argue that R13A1_4 is indeed unc-8. We are currently sequencing the mutant unc-8 alleles e15, e49 and n491 to identify the semi-dominant mutations. Since the motorneuron swelling phenotype of unc-8 mutant strains harboring these alleles is distinct from degeneration induced by mec-4(d), the mutations might map to sites other than those where dominant mutations are located in toxic mec-4(d)alleles. Identifying these mutations should provide insight into the function of degenerins.

Driscoll MA et al. (1990) Worm Breeder's Gazette "mec-4 and deg-1, Genes That Can Mutate to Induce Neuronal Degeneration, Are Highly Homologous"

Driscoll MA, Chalfie M

[Worm Breeder's Gazette, 1990]

Neuronal degeneration in C. elegans can be caused by mutations in two genes, mec-4 and deg-1. Dominant alleles e1611, u214, and u231 of the mec-4 gene induce the degeneration of the touch receptor neurons. The dominant allele u38 of the deg-1 gene causes degeneration of a different group of neurons (including the PVC interneurons). Although different cells are affected by these mutations, the deaths induced appear physically similar--characterized by a vacuolar appearance of the dying cells before their disappearance. In addition, since mutations in the mec-6 gene suppress both mec-4 and deg-1 degenerations, it appears that the mec-4 and deg-1 gene products act similarly to induce degeneration in different cells. We are interested in deciphering the normal function of these gene products and in determining the nature of the changes that can culminate in abnormal cell death. We report here that mec-4 and deg-1 are members of a gene family and encode highly homologous proteins. The mec-4 gene was mapped to cosmid T20B9 by standard transposon tagging techniques. A 4 kb EcoRI restriction fragment within this cosmid was shown to be the site of three transposon insertions that are absent in revertants. In addition, a gamma ray-induced allele and an EMS-induced allele show polymorphisms in this fragment. 5x10+E6 plaques from Chris Martin's cDNA library (2-3 kb size fraction) were screened and a single 1.6 kb cDNA that included the sequences covered by the polymorphisms was isolated. We have sequenced this cDNA and the corresponding genomic DNA. The cDNA contains an open reading frame that encodes a protein of 494 amino acids. It contains a poly A tail, so we are confident that the 3' end of the gene is included in the cDNA clone. The 494 amino acids are encoded in 12 exons in the genomic DNA. We suspect that the mec-4 gene product may be larger than 494 amino acids since there are substantial open reading frames in the genomic sequence 5' to the position where the cDNA begins. We are currently attempting to identify the real 5' end of the gene through RACE amplification experiments. The deduced amino acid sequence of the mec-4 protein can be roughly divided into three domains. The first 430 amino acids include many cys and pro residues. In addition, this 'domain' has five sites for possible N-linked glycosylation. The next 30 amino acids are largely hydrophobic and are likely to constitute a trans-membrane domain. The carboxy terminal domain contains many charged residues. Although the deduced mec-4 protein does not have extensive homology with any entries in the current GenBank database, the mec-4 protein exhibits striking homology to the deg-1 protein sequence. These proteins are 50% identical over most of their lengths. We have seen evidence of other cross-homologous C. elegans genes on Southern blots. About eight cross-homologous bands are seen using a mec-4 probe. Similar cross-homology has been reported for the deg-1 gene (Wolinsky and Chalfie, WBG 10(2), Chalfie and Wolinsky, Nature, in press). Thus, it appears that the C. elegans genes that can mutate to cause degeneration are members of a gene family. We have constructed libraries of genomic DNA from the three dominant mec-4 mutants and have isolated clones of the e1611, u214, and u231 alleles. We are currently sequencing these alleles to determine the amino acid changes that are associated with degeneration. In addition, we have isolated other cross-homologous genes from C. elegans to continue our study of this gene family.