Schauer IE et al. (1990) Worm Breeder's Gazette "Update on Early Embryonic Transcription"

Schauer IE, Wood WB

[Worm Breeder's Gazette, 1990]

In previous gazette abstracts we have reported the use of a run-on transcription reaction to quantitate and characterize transcription in pregastrulation C. elegans embryos and to identify clones of genes that are preferentially expressed during this time. We report here on further experiments addressing the validity of the in vitro assay and characterizing the early embryonic genes identified so far. The presence of 0.5% sarkosyl or 1 mg/ml heparin, inhibitors of polymerase II initiation, in the embryonic extracts and run-on transcription reactions enhances the activity by about 50-100%. The transcripts produced in the presence and absence of these initiation inhibitors have about the same size range and proportion of sequences homologous to the 3' and 5' ends of several specific genes. Thus the enhancement of activity represents a slight increase in elongation rate (as opposed to inhibition of reinitiation accompanied by a large increase in elongation rate). We conclude that little or no reinitiation occurs in the in vitro reaction. Since initiation therefore occurs in vivo, the in vitro reaction is, at that level, regulated in vivo. We have previously demonstrated correct regulation in the in vitro reaction for several temporally regulated genes (C. elegans meeting abstracts, 1989) and now also for the sex-specific gene, her-1. Radiolabeled run-on transcripts produced in extracts of him-8 embryos hybridize strongly to cloned her-1 DNA while those produced in extracts of N2 embryos do not. This result strongly suggests that the sex-specificity of her-1 transcript accumulation (C. Trent, WBG 10,3) is controlled at the level of transcription. These observations also strengthen the argument that the patterns of transcription observed in vitro do reflect in vivo regulation. We have also continued to screen for early embryonic genes using run- on transcripts to probe an early embryonic cDNA library. Several new clones, including new isolates of two of the genes found earlier, have brought the total number of identified early embryonic genes to fourteen. Hybridization to YAC grids has placed several of the new genes on the physical map. Unfortunately, and against all odds, all of them so far fall into genetically unmapped contigs. Of the fourteen total genes identified three are genetically mapped, two to LGX and one to LGIV, eight fall into genetically unmapped contigs, and three have not yet been mapped. Sequence analysis of cDNA clones of five genes has yielded open reading frames but no significant homologies. We have also extended our RNA gel blot analysis of the early embryonic genes to include RNA preparations from glp-1(q231ts) and fem-1(hc17ts) adults, both grown at the nonpermissive temperature, as well as the early embryonic, late embryonic, and L1 RNAs used previously. None of the early embryonic transcripts are present in the glp-1 (soma only) RNA, but all appear to be present at very low levels in the fem-1 (soma and germline) RNA. Levels of these transcripts increase about 20- to 50-fold in early embryonic RNA. In contrast, levels of several predominantly maternal transcripts increase only about 5- to 10-fold, a consequence of packaging of transcripts into oocytes and of a low level of continuing embryonic expression. The RNA gel blot and run-on transcription results together indicate that there is a small number of genes whose expression is limited to a period of time beginning just prior to fertilization and ending sometime early in embryogenesis. We are continuing to pursue various approaches to determining their functions in the early embryo.

Schauer IE et al. (1988) Worm Breeder's Gazette "embryonic cDNA expression libraries."

Schauer IE, Wood WB

[Worm Breeder's Gazette, 1988]

We have constructed two C. elegans embryonic cDNA libraries in lambda gt11, one from predominantly early embryonic RNA and one from total embryonic RNA of him-8 embryos. Characterization of these cDNA libraries to date is as follows. The early embryonic cDNA library was made with poly A+ RNA from an N2 embryo population that was about 85% <30 cells. The library was amplified from 2x10+E6 independent recombinants in two batches. The final library is 82-85% recombinants. Before amplification, 16 random recombinants were picked and DNA was prepared. All but 2 of these contained detectable inserts ranging from 0.4 -4 kb, with an average of 1.1 kb. Eight of the inserts were used as probes on C. elegans genomic Southern blots; 6 hybridized to single bands and 2 to pairs of bands. The library was screened with four previously characterized clones of unidentified genes, three that are expressed in the early embryo and one expressed maternally, and positives were found for all of them. As might be expected, the maternally expressed gene is especially abundant in this library. One of these cDNAs has so far been characterized further and found not to be full length. The him-8 cDNA library was made with poly A+ RNA from a him-8 embryo population that ranged from very early to well into morphogenesis. This library was also amplified in two batches from a total of 1x10+E6 independent recombinants. The final libraries are 37% and 55% recombinants. Random recombinants were not checked from this library, but positive recombinants have been identified with a her-1 probe and a generic actin probe. For two her-1 positives, identified by Carol Trent, the inserts are about 65% and 75% full length. The generic actin probe yielded positives at a frequency of . 05-.07%, as might be expected for an abundant mRNA. Five positives were plaque-purified, and DNA was prepared. Insert sizes ranged from 0.8 to 1.7 kb (full-length actin mRNA's are 1.6 to 1.7 kb, depending on the gene). These five inserts were also used as probes on genomic Southern blots. Four yielded only the expected actin bands, while the fifth had a single additional band which may be unrelated to actin. Thus the majority of recombinants in both libraries appear to have inserts derived from single cDNA species. There is, however, one known problem with the him-8 library and possibly, to a lesser degree, with the early embryonic library. The two her-1 inserts and two or three of the five actin inserts cannot be cut out with EcoRI. Together with the low frequency of full length clones, this suggests that there may have been some EcoRI* activity during linker removal. In summary, both libraries contain a reasonable frequency of recombinants containing simple inserts of C. elegans cDNA. The frequency of full length clones is low, and analysis of inserts may be complicated by loss of EcoRI sites. These libraries are available upon request.

Schauer IE et al. (1986) Worm Breeder's Gazette "P-granule gene screen."

Wood WB, Schauer IE, Hall B

[Worm Breeder's Gazette, 1986]

We have used four of S. Strome's anti-P-granule monoclonal antibodies, L416, L271, K76, and N123, to screen a modified gtll genomic library from wild-type C. by R. Barstead and R. Waterston. A screen of about 8 genome equivalents yielded 37 positive clones that rescreened as positives with one or more anti-P-granule antibodies, but were negative with a control IgM. Although the antibodies used for the screen had been preadsorbed against gt11-infected E. coli and showed little or no reactivity to E. coli proteins on 'Western' blots or on immunoblots of negative plaques, most of the positive clones were found to contain inserts of bacterial DNA. This result suggests that these antibodies react with common epitopes that are also found at low abundance in bacteria. The result also indicates that the library contains some bacterial DNA in addition to nematode DNA. While bacterial DNA has been found in some other screens with this library (T. Blumenthal, personal communication) , it has apparently not been a problem when polyclonal antibodies were used (R. Barstead, personal communication). Of 25 positive clones tested by Southern blotting, four were found that hybridize to C. DNA. Though these could also code for non-P- granule proteins that share an epitope with a P-granule protein, they are being further characterized. Two classes were found. One is represented by a single clone, KL3a. It reacts only with the K76 antibody and hybridizes to a single band on C. Southern blots at high stringency. On 'Northern' blots it hybridizes to a single 2kb mRNA of moderate abundance, which appears to be most abundant in adults. Lysogens of KL3a make a very unstable fusion protein of about 120 kd (of which only 80 kd is -galactosidase in this modified gtll vector) that is recognized by both K76 and anti- -galactosidase antibody. The insert is being subcloned into a better expression vector to allow purification of the fusion protein and generation of a polyclonal antibody. This antibody will be used initially to determine whether the corresponding protein is a P-granule component or simply a cross- reacting protein. The other class of positive clones containing C. epresented by three independent clones, LN4c, LN7a, and LN7b, whose inserts cross-hybridize. Proteins produced by these clones react with all four antibodies, and their DNA hybridizes to about 30 bands on high stringency genomic Southern blots. We have not yet identified transcripts for these clones by 'Northern' blot analysis. Lysogens of these positive clones, however, produce fusion proteins of 120-170 kd, indicating open reading frames of 1-2 kb. One of these clones produces a very abundant and stable fusion protein which reacts on immunoblots with all four monoclonal antibodies and anti- -galactosidase antibody. This fusion protein has been purified and used to immunize rabbits. Preliminary results indicate that the immune serum contains antibodies against the fusion protein, but fails to recognize P-granules. Therefore, the inserts in positive clones of this class probably are not derived from genes for P-granule proteins.

Schauer IE et al. (1988) Worm Breeder's Gazette "early embryonic transcription in C elegans."

Schauer IE, Wood WB

[Worm Breeder's Gazette, 1988]

We have been analyzing transcriptional activity in embryos by preparing extracts and assaying run-on transcription in the presence of [32P]-UTP (WBG Nov. 1987). We reported that early embryos (<30 cells) are transcriptionally active, having about 50% of the activity of later embryos. We have now further characterized these reactions with regard to linearity, developmental timing of specific gene transcripts, amanitin sensitivity, and comparative rates of incorporation in extracts from different embryonic stages. Incorporation of [32P]-UTP is linear to about 15-20 minutes and plateaus by about 60 minutes. The time course of incorporation does not differ significantly between extracts made from embryos at different stages. Amanitin-sensitive incorporation is consistently 80- 90% of the total, but this has not yet been checked for extracts of the earliest embryos analyzed (99% <30 cells). Early and late run-on transcripts (from extracts of 95% <30-cell embryos and >95%o >30-cell embryos, respectively) have been used to probe Southern blots of a collection of genes with known or predicted patterns of developmental regulation during embryogenesis. Histone, tubulin, and actin genes are already being transcribed in the early extracts, and transcription increases significantly in the late extracts. In contrast, unc-54 and col-1 are expressed in late but not in early extracts. Finally, vit-5 transcripts are not detected in either extract. Thus run-on transcription in vitro appears to follow the expected patterns of developmental regulation, at least for these genes. Further quantitation of incorporation by extracts of embryos at different stages suggests that very early embryos are as active transcriptionally as later embryos on a per nucleus basis. Extracts of early embryos (99% <30 cells) incorporate at least as much [32P]- UTP per nucleus as extracts of post-gastrulation embryos. These incorporation rates at 22 C correspond to 300-900 kb of RNA (or about 100-300 mRNA molecules) per nucleus per min. The variation in this number probably reflects the difficulty of accurately determining concentrations of nuclei rather than reflecting stage-specific variations in incorporation rates. Incorporation in these extracts was also normalized to DNA concentration, determined by fluorimetry, and again no significant stage-specific variation of incorporation rates was found. However, DNA readings were 2-3 fold higher than expected based on nuclear estimates. This is presumably due at least in part to high levels of mitochondrial DNA in embryos. Because of this discrepancy the two methods of normalization give different absolute values for [32P]-UTP incorporation/nucleus. However, by both methods early embryos are as active transcriptionally as later embryos. This conclusion appears to be at odds with those of Hecht et al. ( Dev. Biol. 83:374, 1981), who used in situ hybridization of a [3H]- poly(U) probe to squashes of embryos at different stages to estimate levels of nuclear poly(A)+ mRNA per embryo compared to total cellular poly(A)+ mRNA. These authors reported that nuclear poly(A)+ is first detectable around the 90-cell stage, suggesting that transcription of the embryonic genome begins at this time. However, their data (Fig. 3) also show that the average number of grains per nucleus stays approximately constant from the 100-cell to about the 500-cell stage, consistent with our results. Furthermore the level of about 100 poly( A)+ mRNA molecules per nucleus that can be estimated from their data during this time is roughly consistent with our incorporation levels, assuming a short time for transit of nuclear poly(A)+ molecules to the cytoplasm. Therefore, the principal discrepancy between our results and theirs is simply that we find evidence for equivalent levels of transcription in early and late embryos, whereas they do not. One possible explanation is that in our experiments, despite the evidence presented above for apparently normal control of selected transcripts, the earliest incorporation we observe in extracts is an artifact not representative of in vivo transcription. If our results do reflect the in vivo situation, then the discrepancy most probably results from the substantial differences between the two studies in what is being measured: nucleotide incorporation rates vs levels of accumulated nuclear poly(A) present, respectively. Other explanations could be, for example, that: 1) early transcription does not produce mature mRNA's, or produces only poly(A)- mRNA's; or 2) early transcription produces poly(A)+ mRNA's but their transit time to the cytoplasm is shorter than in later embryos, so that the steady state level of nuclear poly(A)+ is below the level of detection by the methods of Hecht et al. We are attempting to distinguish between the various alternatives by analyzing RNA synthesis in permeablized whole early embryos.

Schauer IE et al. (1987) Worm Breeder's Gazette "transcription in early embryos of C elegans."

Schauer IE, Wood WB

[Worm Breeder's Gazette, 1987]

We are characterizing early transcription in embryos of C. elegans. While general embryonic transcription appears to begin at about the 100-cell stage, midway through gastrulation (Hecht et al.), there is evidence for some transcription as early as the four-cell stage (L Edgar, personal communication). We have used run-on transcription in extracts of staged N2 embryos and gel blot analysis of separated RNAs from adults, oocytes, L1's and staged embryos to analyze the pattern of embryonic transcription and to look for genes that are preferentially expressed during early embryogenesis. To obtain embryos, L1 hermaphrodites are synchronized by hatching overnight in M9 without food. The starved L1s are then plated at 10-20,000/plate on NGM plates and kept well fed for 34 days at 16-20 C. The worms are observed as they reach adulthood. 'Early'' embryos are harvested by hypochlorite treatment when the most mature worms contain 4-6 eggs. The result is a low yield of embryos of which 90-95% are pregastrular, with <30 cells. 'late' embryos are obtained by hypochlorite treatment of heavily gravid adults followed by 2 hr of aeration of the embryos in M9. Of these embryos, 95-99% have >30 cells; usually about half the embryos are premorphogenesis and half are in the process of morphogenesis. Extracts prepared from early and late embryos are able to incorporate [32P]-UTP into run-on transcripts. In vitro transcription is 100% rifampicin-resistant and ~80% amanitin-sensitive, indicating that most of the incorporation results from eukaryotic pol II activity. The transcripts include sequences homologous to the actin gene act-1, but not to the vitellogenin gene vit-5, consistent with normal developmental regulation. Labeled run-on transcripts from early and late embryonic extracts were used to probe duplicate filter lifts from an N2 genomic library ( prepared by C. Link). With the reservations that there may be artifacts in in vitro transcription and that some later embryos are present in the early embryo preparations, the following observations suggest that substantial transcription occurs prior to the 30-cell stage. 1) Early extracts incorporate labeled nucleotide at about 50% of the rate seen in late extracts, on a per nucleus basis. 2) When library filters are probed, early run-on transcripts hybridize as strongly as late run-on transcripts to 25-50% of the plaques. 3) RNA gel blot analysis of actin gene expression indicates that act-1 and act-3, but interestingly not act-4, are already being expressed in early embryos. When duplicate library filters are probed with early and late running transcripts, about 0.02% of the plaques hybridize more strongly to early transcripts. Several such phage have been isolated and further characterized by gel blot analysis of RNAs from the following developmental stages: adults with oocytes (temperature sensitive fer-1 adults raised at the non-permissive temperature), early embryos, late embryos, and L1's. Nine independently isolated phage clones have been found to recognize transcripts that are present primarily in early embryos. Three of these transcripts were recognized by more than one of the clones, suggesting that the number of primarily early transcripts may not be large. Six different transcripts were recognized by the nine clones. These six transcripts fall-roughly into the following three classes: 1) Low abundance in adults with oocytes, high in early embryos, low in late embryos (one transcript). 2) Absent from adults with oocytes, high in early embryos, barely detectable in late embryos (one transcript). 3) Absent in adults with oocytes, high in early embryos, 2- to 5-fold lower in late embryos (four transcripts). None of the transcripts persist as far as early L1. All six are small, ranging from 0.7 to 2 kb. We are currently looking for cDNA clones that correspond to these transcripts and planning to use the cDNA's for in situ hybridization and sequencing. Five of the six sequences have been placed in contigs (J. Sulston and A. Coulson, personal communication), but none of the relevant contigs are mapped or marked. Plans to begin genetic analysis must, therefore, await further developments.

Burkholder A et al. (1990) Worm Breeder's Gazette "An Embryonic Histone-Binding Protein?"

Wood WB, Schauer IE, Burkholder A

[Worm Breeder's Gazette, 1990]

We have identified an embryonic nuclear antigen that has several similarities to the Xenopus histone-binding protein N1 (see Kleinschmidt and Seiter, EMBO 7:1605,1988). The antigen was identified in a screen of C. elegans expression libraries with sera from rabbits immunized with homogenates of early embryos. Plaques of positive clones were used to affinity purify cognate antibodies from the sera, which were then assayed for staining of fixed embryos by indirect immunofluorescence microscopy (Burkholder and Wood, C. elegans meeting abstr. 1989). We obtained an antibody preparation that recognizes a diffuse nuclear antigen, present in oocytes beginning at pachytene and in all nuclei of early embryos, but becoming more restricted later in development. In larvae, the staining is less intense and only in subsets of nuclei that correlate with the positions of proliferating cells, and in adults only maturing oocytes contain the antigen. In mitotic cells the antigen is diffuse throughout the cytoplasm, and then appears to reassemble into the nucleus neither associated with chromatin or the nuclear envelope. We have studied the cloned gene encoding this nuclear antigen, which is named D1. The gene encodes a 2kb messenger RNA which is relatively abundant in early embryos, apparently resulting from the accumulation of maternally supplied transcripts. Late embryos and L1's have very low levels of message. The message encodes a 583 a.a. protein with a predicted MW of 63kd, although the antibody detects a protein of 1 00kd on an immunoblot of embryonic proteins. This discrepancy is likely due to the highly charged nature of the protein, which has 25% acidic residues including one region where 13/18 residues are negatively charged. A highly basic region near the C-terminus ( including the sequence TKKRKS) may function as the nuclear localization signal. In a search of the Genbank protein database the most homologous protein was the Xenopus N1 protein, although sequence identity between the two proteins is only 20% over 572 residues. This level of homology alone is not very convincing, but combined with the similar size, arrangement of highly charged regions, cellular location and developmental expression of the two proteins, we are exploring the possibility that the C. elegans D1 protein has a histone-binding function similar to the Xenopus N1 protein. Maternal stores of histones H3 and H4 are found in a complex with N1 protein, which is thought to promote chromatin formation during the rapid cleavage cycles of Xenopus embryos. The D1 gene maps to YAC Y11G11 on V (A. Coulson and J. Sulston, pers. comm.), which places it on LGV between myo-3 and rrs-1 on the genetic map. In that interval the only obvious candidate for a mutation in the D1 gene is the maternal-effect lethal mutation par-1; however, par-1 probably maps to the right of D1 according to physical mapping data (A. Telfer, pers. comm.).

Otsuka AJ (1990) Worm Breeder's Gazette "ISOLATION AND CHARACTERIZATION OF UNC-44 CDNA CLONES"

Otsuka AJ

[Worm Breeder's Gazette, 1990]

Seven unc-44 cDNA clones were isolated from 2.6 million phage (1.4 million phage screened from the Barstead-Waterston bank, 0.6 million clones each from the Schauer-Wood early embryonic and him-8 embryonic libraries). A 5'-probe (the 9 kb BamHI fragment from cosmid B0350 corresponding to the site of transposon insertion in unc-44(q331) and a 3'-probe (the 3.7-kb SalI fragment from phage DD#LRF1 corresponding to the site of Tc1 insertion in unc-44(rf1042)) were used in combination to screen the libraries. The results are shown below: [See Figure 1] Because of the ease of releasing the plasmid portion of the ZAP II phagemid vector, the characterization of the clones from the Barstead library has progressed further than those from the Schauer library. Although isolated in different screens, clones DD#LA013 and DD#LA045 have identical SstI restriction patterns. DD#LA049, which hybridizes only to the 5' probe, overlaps with DD#LA013 and, assuming the absence of cloning artifacts, predicts a minimum size of 4 kilobases for the unc-44 message (diagrammed below). DNA sequencing of the cDNA clones is currently in progress. [See Figure 1]

Otsuka AJ (1990) Worm Breeder's Gazette "ISOLATION AND CHARACTERIZATION OF UNC-104 CDNA CLONES"

Otsuka AJ

[Worm Breeder's Gazette, 1990]

Although the unc-104 fragments isolated from an Arhinger-Kimble library were instrumental in obtaining approximately 4 kb of the 6 kb mRNA sequence, it was desirable to obtain full length cDNA clones to both complete the cDNA sequence and to produce the recombinant protein product in bacteria. The discrepancy between the abundance of unc-104 cDNA clones in the Arhinger-Kimble library (approximately 1 per 10,000) and in the Kim-Horvitz (none in 200,000) and Thacker-Capecci libraries (none in 100,000), led us to screen the Barstead-Waterston and Schauer-Wood libraries. The abundance of cDNA clones in these additional libraries suggests that an insufficient number of clones was screened in the cases of the Kim and Thacker libraries. cDNA clones were isolated from 2.6 million phage (1.4 million phage screened from the Barstead-Waterston bank, 0.6 million the Schauer- Wood early embryonic and him-8 embryonic libraries). The probe was the 5'-most cDNA fragment obtained from the library. The results are shown below: [See Figure 1] The two cDNA clones from the Barstead library overlap and cover a region of 4.7 kb. DNA sequencing of the cDNA clones is currently in progress.

Pettitt J et al. (1993) Worm Breeder's Gazette "Investigating the vets"

Pettitt J, Wood WB

[Worm Breeder's Gazette, 1993]

We have given the name vet (very early transcript) to a set of genes originally identified by Irene Schauer(1) in molecular screens for genes that were preferentially transcribed in pre-gastrulation embryos. Irene isolated sixteen vet genes, eight of which have been placed on the physical map. Sequence analysis of the six vet genes for which cDNA clones exist (vet-1 to -6) revealed that only one gene (vet-1 ) had any similarity to other sequences in the Genbank databases. We are extending the sequence analysis of these genes and have determined the entire coding sequences for vet-1 and vet-6 ,both of which are trans-spliced with SL1 .This work has established that vet-1 encodes a product which is likely to be almost entirely coiled-coil in structure, whereas vet-6 appears to encode a novel protein. We have begun to characterize the spatial expression patterns of selected vet genes, in order to establish which ones, if any, show lineage specific expression patterns. We are beginning by making lacZ constructs of the six vet genes which we know most about. The results with our first lacZ fusion construct showed that at least one vet gene has a spatially restricted expression pattern. This fusion construct (termed pVET6 B)consist of a 6 kb genomic DNA fragment, derived from the vet-6 gene, fused to the lacZ gene in Andy Fire's pPD22 .11vector. We have used four lines, each carrying an integrated array of pVET6 B,to study its expression pattern. The original screen in which vet-6 was isolated was biased for highly transcribed genes, and pVET6 Bappears to be highly expressed as well, producing patterns that are apparent after only 15 minutes of staining with X-Gal. The fusion is expressed only relatively briefly and staining first appears at the 28 cell stage and disappears by the end of gastrulation. We have not yet established which cells express the fusion; however, the expression begins in a few cells and spreads to include nearly all the cells of 60-100 cell embryos. The staining pattern is reduced to a subset of these cells in subsequent divisions prior to its disappearance. The fusion protein appears to be cytoplasmic in all the cells that express it. We are currently using Geraldine Seydoux's in situ hybridization protocol to try to confirm that the expression of pVET6 Bcorrelates with the presence of vet-6 transcripts. We have also made a lacZ fusion using a 3.5 kb genomic fragment derived from vet-1 ,but we have not been able to obtain any expression from this construct. Further lacZ fusion constructs using larger fragments from vet-1 ,as well as constructs for other vet genes are underway. (1) Schauer and Wood (1990) Development 110,1303.

Kiff JE et al. (1985) Worm Breeder's Gazette "further studies on unc-54."

Kiff JE, Moerman DG, Waterston RH

[Worm Breeder's Gazette, 1985]

We have expanded our collection of missense alleles of unc-54 that suppress unc-22 induced twitching to a total of 17 unc-22 suppressors. These mutations can be split into two classes, an s74-like class with t4 members (see Cell 29, 773-781) and a new class with 3 members, st130, st132 and st148. This new class of mutations also act as dominant suppressors of unc-22(s12) and unc-22(st137::tc1), and a worm heterozygous for any of these mutations moves like N2 and has normal muscle structure. However, an animal homozygous for st130, st132 or st148 is slow, does not lay eggs, and the muscle cells exhibit significant A-band disorganization. It is this last trait that distinguishes this new class of mutation from mutations of the s74- like class which affect movement but have little, or no affect on A- band organization. We have carried out an extensive series of reversion experiments using s74, s95, st130 and st132 to see if there are either inter- or intragenic sites that interact with these mutations. The results are summarized below. [See Figure 1] Only st150, an s95 revertant, proved to be an extragenic suppressor and it is probably a sup-3 allele. The other revertants are the result of intragenic events. The frequency at which intragenic revertants of st132 occur as compared to s95 (a member of the s74-like class) further suggests that st132 defines a new class of mutation in unc-54. The st130 revertant and 2 of the st132 revertants are partial revertants while the other st132 revertants are full revertants to wild type. This latter set has been divided into 2 groups - of 13 tested mutations, 9 are dosage dependent, ie st132/st132, st132 m/st132 +, and st132 m/st132 m have 3 discernable movement phenotypes, and 5 are dosage independent, ie, restoration of wild type movement is fully dominant. The dosage response and the high reversion frequency of st132 indicates that compensatory second-site mutations in unc-54 can occur. The s95 intragenic mutations are full revertants and 5 of these were analyzed at the DNA level. The s95 mutation is at position 2340 in the nucleotide sequence and changes CCA (gly) to ACA (arg) (J.M.B. 183, 543-551). Fortuitously s95 disrupts a HpaII site (CCCC to CCAC). Of the 5 intragenic revertants examined by Southern analysis, only one restores the HpaII site. After cloning and sequencing the other 4 revertants we were surprised to find that all 4 have the same nucleotide change at position 2342 thus altering ACA (arg) to ACT (ser) . Our inability to find compensatory mutations in another part of unc- 54 suggests that s95 may disrupt interactions with a molecule other than myosin. The proximity of s95 to the putative ATP binding site ( EMB0 1, 945-951; P.N.A.S. 80, 4253-4257) makes ATP a likely candidate.