Groucho and Tup1 are members of a conserved family of WD repeat proteins that interact with specific transcription factors to repress target genes. Here we show that mutations in WD domains of the Groucho-like protein, UNC-37, affect a motor neuron trait that also depends on UNC-4, a homeodomain protein that controls neuronal specificity in Caenorhabditis elegans. In unc-4 mutants, VA motor neurons assume the pattern of synaptic input normally reserved for their lineal sister cells, the VB motor neurons; the loss of normal input to the VAs produces a distinctive backward movement defect. Substitution of a conserved residue (H to Y) in the fifth WD repeat in unc-37(e262) phenocopies the Unc-4 movement defect. Conversely, an amino acid change (E to K) in the sixth WD repeat of UNC-37 is a strong suppressor of unc-37(e262) and of specific unc-4 missense mutations. We have previously shown that UNC-4 expression in the VA motor neurons specifies the wild-type pattern of presynaptic input. Here we demonstrate that UNC-37 is also expressed in the VAs and that unc-37 activity in these neurons is sufficient to restore normal movement to unc-37(e262) animals. We propose that UNC-37 and UNC-4 function together to prevent expression of genes that define the VB pattern of synaptic inputs and thereby generate connections specific to the VA motor neurons. In addition, we show that the WD repeat domains of UNC-37 and of the human homolog, TLE1, are functionally interchangeable in VA motor neurons which suggests that this highly conserved protein domain may also specify motor neuron identity and synaptic choice in more complex nervous systems.

In the nematode, Caenorhabditis elegans, most of the VA and VB motor neurons arise from a common precursor (Sulston and Horvitz, 1977) but adopt different morphologies and receive inputs from separate sets of interneurons in the ventral nerve cord (White et al., 1986). The unc-4 homeodomain protein (Miller et al., 1992) specifies the wild-type pattern of synaptic inputs to VA motor neurons (White et al., 1992). In unc-4 mutants, the usual inputs to VA motor neurons are replaced with a gap junction from an interneuron normally reserved for their sister VB motor neurons (Fig. 1). Neither process placement nor axonal morphology are perturbed, however, indicating that unc-4 regulates synaptic choice but not other differentiated traits of these neurons. The miswiring of the VA motor neuron circuit produces a characteristic movement defect; unc-4 mutant animals are unable to crawl backward and tend to coil dorsally when tapped on the head (Brenner, 1974; White et al., 1992).

Fig. 1.

unc-4 activity in the VA motor neurons specifies synaptic choice. In the wild type, most of the VA and VB motor neurons arise from a common precursor cell, but receive different presynaptic inputs and adopt different axonal trajectories (VA axonals project anteriorly while VB axons grow out posteriorly). In unc-4 mutants, the usual synaptic input to VA motor neurons is replaced with a gap junction from an interneuron (AVB) normally reserved for sister VB motor neurons; the morphology of both the VA and VB neurons, however, is not affected. unc-4 expression (indicated as a hatched oval) in VA distinguishes it from its VB sister with respect to presynaptic partners. The black hexagon represents interneurons AVA (gap junction, chemical synapse), AVD, AVE (chemical synapse). The grey hexagon represents interneurons AVB (gap junction) and PVC (chemical synapse). Anterior is to the left. Arrows are symbolic and do not represent axonal branches (see White et al., 1992).

Fig. 1.

unc-4 activity in the VA motor neurons specifies synaptic choice. In the wild type, most of the VA and VB motor neurons arise from a common precursor cell, but receive different presynaptic inputs and adopt different axonal trajectories (VA axonals project anteriorly while VB axons grow out posteriorly). In unc-4 mutants, the usual synaptic input to VA motor neurons is replaced with a gap junction from an interneuron (AVB) normally reserved for sister VB motor neurons; the morphology of both the VA and VB neurons, however, is not affected. unc-4 expression (indicated as a hatched oval) in VA distinguishes it from its VB sister with respect to presynaptic partners. The black hexagon represents interneurons AVA (gap junction, chemical synapse), AVD, AVE (chemical synapse). The grey hexagon represents interneurons AVB (gap junction) and PVC (chemical synapse). Anterior is to the left. Arrows are symbolic and do not represent axonal branches (see White et al., 1992).

Expression of an unc-4 transgene in VA motor neurons is sufficient to restore wild-type movement to unc-4 mutants and therefore suggests that the unc-4 homeoprotein (UNC-4) controls a specific feature of the VA motor neurons that allows interneurons to distinguish them from VB sister cells (Miller and Niemeyer, 1995). In an effort to identify other genes that might function with UNC-4 to specify this VA-specific trait, we isolated suppressors of unc-4(e2322ts), a missense mutation in the unc-4 homeobox. We identified dominant, allele-specific suppressors that restore normal backward movement to unc-4(e2322ts) animals. All four of these alleles are tightly linked to the unc-37 locus. Reversion of one of these suppressors, wd17dm, produced a recessive, loss-of-function (lf) phenotype that fails to complement unc-37(e262) (Miller et al., 1993). The uncoordinated phenotype of unc-37(e262) animals is strikingly similar to that of unc-4 mutants and is therefore consistent with a comparable wiring defect. On the basis of these genetic interactions, we have previously suggested that the suppressor mutations are unc-37 gain-of-function (gf) alleles and that unc-37 could either represent an UNC-4 target gene or encode a transcriptional cofactor protein that functions with UNC-4 in VA motor neurons (Miller et al., 1993).

In this paper, we describe the molecular characterization of the unc-37 gene. UNC-37 is highly similar to a family of transcriptional corepressor proteins that includes Drosophila Groucho (Gro) (Hartley et al., 1988) and vertebrate homologs from rat (R-esp2) (Schmidt and Sladek, 1993) mouse (ESG) (Miyasaka et al., 1993; Koop et al., 1996) and human (TLE) (Stifani et al., 1992). All of these proteins contain a conserved tandem array of the WD repeat, an evolutionarily ancient motif that mediates protein-protein interactions (Neer et al. 1994). Groucho-like proteins do not display DNA binding activity but are recruited to genes by specific transcription factors. Gro, for example, interacts with Hairy and with other similar bHLH proteins to repress target gene transcription (Paroush et al., 1994). Similarly, a related WD repeat protein in yeast, Tup1, complexes with the α-2 homeodomain (Komachi et al., 1994) and other cofactor proteins (Tzamarias and Struhl, 1995) to prevent gene expression.

Here we show that UNC-37 function depends on specific amino acid residues in the WD repeat domains. A missense mutation in WD repeat 5 produces the Unc-4-like backward movement defect of unc-37(e262) animals. Conversely, the Unc-4 suppressor activity of the unc-37(gf) mutations results from a single amino acid substitution in WD repeat 6. In addition, we show that UNC-37 expression in the VA motor neurons is sufficient to rescue the e262 movement defect. On the basis of these findings we propose that the wild-type pattern of presynaptic input to the VA motor neurons is determined by the combined activities of UNC-4 and UNC-37 and that the fifth WD repeat of UNC-37 is specifically required for this outcome. We also show that the WD domains of the human TLE1 and nematode UNC-37 proteins are interchangeable in C. elegans. We interpret this finding to indicate that this UNC-37 function is conserved in evolution and therefore may also govern motor neuron identity and neural specificity in more complex nervous systems.

Strains

Nematodes were grown as described by Brenner (1974). Unless otherwise stated, all cultures were maintained at 25°C. The wild-type strain is N2. The unc-37 alleles used in this work are: e262 (Brenner, 1974), wd14dm, wd16dm, wd17dm, wd18dm, wd17wd19, wd17wd20 (Miller et al., 1993) wd17wd21, wd17wd22 (this work). Other mutant strains were as follows. Linkage group I: dpy-5(e61), bli-4 (e937), dpy-14(e188). Linkage group II: unc-4(e2322ts). LG IV: unc-24(e927). LGX: lon-2(e768) xol-1(y70). Strain NC64[dpy-5 unc-37(wd17dm); unc-4(ts); lon-2(e768) xol-1(y70)] was used to generate revertants of unc-37(wd17dm) (Miller et al., 1993). The transgenic line, wdIs1, contains an integrated array of the unc-4-lacZ reporter plasmid, pNC4-22Lz (Miller and Niemeyer, 1995).

It should be noted that the allele designation of ‘wd’ refers to mutant alleles isolated in this laboratory and is not specific to the unc-37 gene which only coincidentally contains ‘WD’ repeat sequences.

Genetics

The alleles wd26dm, wd27dm and wd28dm are intragenic revertants of unc-37(e262) and were isolated as suppressors of the Unc-37 backward movement defect. The method of selection depends on the restoration of backward locomotion to animals that are mutant for unc-37(e262) (no backing) and unc-24(e927) (limited forward movement) (see Miller et al., 1993). F2 progeny of EMS-treated unc-37; unc-24 animals were applied to chemotaxis selection plates. Animals with restored backward movement (Unc-37 suppressor) or forward movement (Unc-24 suppressor) can traverse a 100 mm Petri plate to reach a patch of bacteria in an overnight selection period whereas the parental unc-37; unc-24 animals are ‘Stuck’ and cannot crawl to food. Unc-37-suppressed animals were mated with + +/dpy-5(e61) unc-37(e262) males and hermaphrodite F1 progeny allowed to reproduce by selfing. Intragenic revertants failed to segregate 3/4 Dpy-nonUnc progeny predicted by random segregation (wd26dm, wd27dm, and wd28dm are dominant). Extragenic unc-37(e262) suppressors were also isolated and will be described elsewhere (J. Y.-J. Meir and D. M. Miller, unpublished data).

Putative unc-37 null alleles wd19, wd20, wd21 and wd22 were generated by reversion of wd17dm as described by Miller et al. (1993) and are symbolized as unc-37(0). The unc-37(e262)-rescuing transgene, pAP5.1, (see below) also complements unc-37(0) alleles in the following test: dpy 14(e188) males (grown at 16°C) were mated with bli-4(e937) unc-37(e262) animals carrying wdEx25, a transgenic array of pAP5.1 and the dominant Roller plasmid, pRF4 (Mello et al., 1991). non-Rol male cross progeny were mated with + dpy14/unc-37(0) + hermaphrodites (Miller et al., 1993). Some of these males are non-Rol due to the instability of the transgene in somatic cells but can be expected to transmit the transgene in germ line cells. non-Dpy her-maphrodite progeny showing the Rol phenotype were picked for selfing. The desired transgenic line, unc-37(0) wdEx25, was obtained from successive rounds of picking individual Rol non-Unc progeny that did not segregate Dpy-14 or Bli-4. wdEx25 fully rescued wd19, wd21 and wd22 homozygous animals but only weakly rescued wd20 (only a few subviable progeny were produced).

Isolation and sequencing of unc-37 genomic clones

For germline transformation experiments, the syncytial gonad of unc-37(e262) hermaphrodites was injected with pools of cosmids (1 μg/ml each) (Barnes et al., 1996) and pRF4 (50 μg/ml) (Mello et al., 1991). Transgenic animals carry the injected DNA as an extrachromosomal array and display the roller (Rol) phenotype. Rol-non-Unc F2 animals were scored as rescued for unc-37(e262). Cosmid pool 37/1 (F57B10, F48A9, C03F5, K01B2, T23A10, K10E3, C17F3) does not rescue unc-37(e262) whereas pool 37/2 (C17F3, F29E2, T23B3, W02D3, K01B11, QQQB5) contains unc-37 rescuing activity. Subsequent injections of individual cosmids from pool 37/2 showed that only W02D3 can rescue unc-37(e262). The unc-37(e262)-rescuing plasmid, pAP5.1, contains a 5.5 kb KpnI to BamHI fragment subcloned from W02D3.

4.7 kb of the pAP5.1 insert was sequenced on both strands by primer walking from the 3′ end of the dihydroorotase reductase (DhRd; K. Anders and P. Anderson, personal communication) and from a T7 site in the flanking vector sequence (Sambrook et al., 1989). All genomic and cDNA sequences were generated by PCR-cycle sequencing in an Applied Biosystems automated sequencer in the Vanderbilt Cancer Center.

Isolation and sequencing of unc-37 cDNA clones

Ten positive clones were detected from about 150,000 plaques of a mixed stage cDNA library (Barstead and Waterston, 1989) probed with the adjacent 1.6 kb EcoRV fragments shown in Fig. 2C. Two cDNA clones with inserts approximately equal in size to the unc-37 mRNA detected on northerns (2.1 kb) (data not shown) were sequenced by primer walking from T3 and T7 sites in flanking vector sequence (Sambrook et al., 1989). Identical sequences were obtained from both cDNA clones.

Fig. 2.

Cloning of unc-37. (A) unc-37 maps to the genetic interval between the cloned genes smg-5 on the left (P. Mains, personal communication) (W02D3) and unc-87 (K01B11) and dpy-14 (T14D10) on the right. W02D3 is the only cosmid that rescues the backward movement defect of unc-37(e262) in germline transformation experiments. (B) A BamHI digest of W02D3 rescues unc-37(e262) indicating that BamHI does not cut in the unc-37 transcription unit. Plasmid TR258, a subclone of the 18 kb BamHI fragment, rescues unc-37(e262). TR259 complements smg-5 mutations (K. Anders and P. Anderson, personal communication) but does not rescue unc-37(e262) whereas the pAP5.1 subclone does rescue unc-37(e262). Restriction enzyme sites are B, BamHI; H, HindIII; K, KpnI; P, PstI; S, SalI. (C) The 2.1 kb unc-37 cDNA is shown, boxes depict exons and lines, introns. pAP-5.1 includes the 3′ terminal exon of the adjacent gene (grey box), dihydroorotase reductase (DhRd). Arrows indicate the direction of transcription. Shown below the unc-37 gene are the 1.6 kb EcoRV (RV) genomic fragments used to probe a cDNA library.

Fig. 2.

Cloning of unc-37. (A) unc-37 maps to the genetic interval between the cloned genes smg-5 on the left (P. Mains, personal communication) (W02D3) and unc-87 (K01B11) and dpy-14 (T14D10) on the right. W02D3 is the only cosmid that rescues the backward movement defect of unc-37(e262) in germline transformation experiments. (B) A BamHI digest of W02D3 rescues unc-37(e262) indicating that BamHI does not cut in the unc-37 transcription unit. Plasmid TR258, a subclone of the 18 kb BamHI fragment, rescues unc-37(e262). TR259 complements smg-5 mutations (K. Anders and P. Anderson, personal communication) but does not rescue unc-37(e262) whereas the pAP5.1 subclone does rescue unc-37(e262). Restriction enzyme sites are B, BamHI; H, HindIII; K, KpnI; P, PstI; S, SalI. (C) The 2.1 kb unc-37 cDNA is shown, boxes depict exons and lines, introns. pAP-5.1 includes the 3′ terminal exon of the adjacent gene (grey box), dihydroorotase reductase (DhRd). Arrows indicate the direction of transcription. Shown below the unc-37 gene are the 1.6 kb EcoRV (RV) genomic fragments used to probe a cDNA library.

To identify unc-37 5′ termini, total RNA was purified from mixed stage wild-type animals (Miller et al., 1992) and submitted to RACE-PCR according to the protocol described by the manufacturer (Life Technology). The RACE-PCR generated cDNA fragment was gel purified and sequenced. The 5′ end cDNA fragment included overlapping sequences from both the SL1 and SL2 leaders. The existence of two populations of unc-37 transcripts transpliced to either SL1 or SL2 leaders was confirmed by separate reverse-transcription PCR reactions containing an unc-37 coding sequence primer and either an SL1-specific or SL2-specific primer (Blumenthal, 1995).

For viable unc-37 mutants, cDNA templates generated by reverse transcription-PCR were gel purified and directly sequenced. The complete unc-37 coding region was sequenced for wd14dm, wd18dm, and e262 but only the 3′ terminal region spanning the WD repeats was sequenced for the mutations wd16dm, wd17dm, wd26dm, wd27dm and wd28dm. For the severe loss-of-function alleles, wd19, wd20, wd21 and wd22, single worm PCR (Williams et al., 1992) was carried out to amplify genomic DNA fragments spanning each of the exons and exon/intron boundaries. The PCR-amplified genomic fragments were gel purified for sequencing.

Sequence analysis

Sequence similarity of UNC-37 to Gro family members and to other WD repeat proteins was initially detected from a BLAST search of the NCBI GenBank database (Altschul et al., 1990). Alignments between UNC-37 domains and Groucho-like proteins sequences were generated using Clustal.

Generation of anti-UNC-37 antibodies and immunofluorescence microscopy

A PCR-generated unc-37 cDNA fragment extending 667 bp from a BamHI site in the 5′-terminal primer to an internal XbaI site was cloned into pGEX-KG to create a GST C-terminal fusion protein in the vector, GST2. The chimeric GST-UNC-37 fusion protein (approx. 52×103Mr) spans the UNC-37 N-domain and a portion of V-region but does not include the WD repeat sequences (see Fig. 3A). The GST2 fusion protein was purified (Smith and Johnson, 1988) from a 2L culture (2XTY, 200 μg/ml ampicillin) after a 3 hour induction with IPTG (0.2 mg/ml). Rabbit antisera were prepared by Charles River PharmServices according to standard methods. The crude GST-UNC-37 antiserum was immunoadsorbed with an acetone powder made from IPTG-induced bacteria carrying vector alone (pGEX-KG) (Miller and Shakes, 1995) and then affinity-purified against GST-UNC-37 coupled to cyanogen-bromide activated sepharose; the adsorbed antibody was eluted at low pH (0.1 M glycine, pH 2.5) (Harlow and Lane, 1988). Immunoblots of 10% SDS-PAGE gels of bacterial and C. elegans homogenates (Miller et al., 1986) were incubated with anti-UNC-37 then goat anti-rabbit peroxidase (Sigma) and developed by ECL according to methods described by the manufacturer (Amersham). Animals were fixed and stained (Finney and Ruvkun, 1990) with affinity-purified rabbit anti-GST2 antibody (0.88 μg/ml) and FITC-coupled goat anti-rabbit secondary antibody (1/160) (Sigma) to detect UNC-37-expressing cells. To detect lacZ-positive neurons (e.g. DAs and VAs) in the unc-4-lacZ integrated line, wdIs1, fixed animals were also stained with mouse monoclonal anti-β-galactosidase (1/50 dilution) (Promega) and rhodamine-coupled goat antimouse secondary antibody (1/190 dilution) (Sigma). The secondary antibodies were preadsorbed with an acetone powder of wild-type C. elegans to reduce non-specific staining. DAPI (1 μg/ml) was included in the mounting media to counterstain cell nuclei (Miller and Shakes, 1995). Gonads were dissected and stained for immunofluorescence microscopy as described in Francis et al. (1995).

Fig. 3.

unc-37 cDNA sequence and predicted protein. cDNA numbering begins at the splice site (arrow) of SL1 and SL2 transpliced leader sequences (see text). Arrowheads mark exon boundaries. Asterisks mark a presumptive polyadenylation signal. Locations of potential CKII (dashed line) and cdc2 phosphorylation sites (double underline) and nuclear localization signal (box), are shown. Grey boxes mark individual WD repeats in the conserved WD region (bold letters). P-282 is highlighted to indicate the fusion junction between N-terminal UNC-37 and C-terminal TLE1 in pJM4. Schematic of UNC-37 domains. N-terminal domain (N) (a.a. 1-102), Variable domain (V) (a.a. 103-278), WD repeat domain (a.a. 279-612). The percentage identity of UNC-37 to Groucho (Drosophila), ESG (mouse) and TLE1 (human) is shown for the amino acid sequence in each domain. Differences in length of proteins are principally due to differences in size of V region. Approximate locations of unc-37 mutations are indicated.

Fig. 3.

unc-37 cDNA sequence and predicted protein. cDNA numbering begins at the splice site (arrow) of SL1 and SL2 transpliced leader sequences (see text). Arrowheads mark exon boundaries. Asterisks mark a presumptive polyadenylation signal. Locations of potential CKII (dashed line) and cdc2 phosphorylation sites (double underline) and nuclear localization signal (box), are shown. Grey boxes mark individual WD repeats in the conserved WD region (bold letters). P-282 is highlighted to indicate the fusion junction between N-terminal UNC-37 and C-terminal TLE1 in pJM4. Schematic of UNC-37 domains. N-terminal domain (N) (a.a. 1-102), Variable domain (V) (a.a. 103-278), WD repeat domain (a.a. 279-612). The percentage identity of UNC-37 to Groucho (Drosophila), ESG (mouse) and TLE1 (human) is shown for the amino acid sequence in each domain. Differences in length of proteins are principally due to differences in size of V region. Approximate locations of unc-37 mutations are indicated.

We performed three experiments to establish that the anti-UNC-37 antibodies are specific for UNC-37. First, similar patterns of staining were observed for affinity-purified antisera derived from three different rabbits. Second, preadsorption of affinity-purified anti-UNC-37 with the GST2 fusion protein eliminated immunoblot and immunofluorescence staining. Third, gonads dissected from animals homozygous for the severe unc-37(lf) allele, wd21, did not stain with affinity-purified anti-UNC-37 (data not shown).

Construction of plasmids and transgenic lines

pNE-1 was constructed by subcloning the unc-4 promoter, a 3.2 kb PstI-SmaI fragment (Miller and Niemeyer, 1995), into the gfp-expression plasmid pPD95.67 (A. Fire, S. Xu, J. Ahnn and G. Seydoux, personal communication). The resultant unc-4-gfp chimeric gene encodes a fusion protein containing the first 53 amino acids of the UNC-4 N-terminal sequence but does not include the UNC-4 homeodomain. pJM1 and pJM1.1 contain a translational fusion of the unc-4 promoter fragment with the intact unc-37 cDNA and 3′ UTR with the exception that the amino terminal methionine of UNC-37 is changed to a valine. In pJM1.1, a gfp-encoding fragment from pPD95.67 is inserted at the unc-37 N terminus. pJM-TLE1.2 includes the complete coding sequence of TLE1 (Stifani et al., 1992) with an N-terminal fusion to the 3.2 kb PstI-SmaI unc-4 promoter fragment and a C-terminal fusion to the unc-37 3′ UTR. To create pJM4, the 2.8 kb HindIII-SmaI fragment from the unc-4 promoter region (Miller and Niemeyer, 1995) was fused to the N-terminal sequence of UNC-37 and the WD repeat domain of unc-37 was replaced with the homologous C-terminal sequence of TLE1 immediately after UNC-37 amino acid residue 281 (Fig. 3A). A cDNA fragment spanning the e262 mutation was inserted into pJM1.1 to produce pJM1.1-e262 and a cDNA fragment spanning the wd17dm mutation was inserted into pJM1.1 to produce pJM1.1-wd17dm. In all of these plasmids, the fidelity of the ligation junctions and PCR-generated fragments were confirmed by sequencing. Details of the construction of these plasmids will be provided on request. To create transgenic lines, unc-37(e262) hermaphrodites were injected with pJM1 and the dominant rol-6 plasmid, pRF4, as described above. non-Dpy progeny of unc-37(e262); dpy-20(e1282) animals carrying the pJM1.1, pNE-1, pJM-TLE1.2, pJM4, or pJM1.1-e262 transgenes and non-Dpy progeny of unc-4(e2322ts); dpy-20(e1282) hermaphrodites carrying pJM1.1-wd17dm were generated by coinjection with the dpy-20 (+) plasmid, pMH86 (Clark et al., 1995). non-Dpy progeny of the e262 line carrying pJM1.1 that showed wild-type backward movement were anesthetized with 25% ethanol in M9 buffer (Brenner, 1974) for photography in a Zeiss Axioplan microscope equipped with an FITC filter set.

Cloning unc-37

We cloned the unc-37 gene by relying on the known physical map locations of the nearby genetic loci (Fig. 2). Cosmids from the unc-37 region (Coulson et al., 1995) were tested for their ability to restore backward movement to unc-37(e262) animals in microinjection/transformation experiments. Ultimately, unc-37-rescuing activity was localized to a 5.6 kb KpnI-BamHI fragment from cosmid W02D3. This genomic clone also rescued the maternal effect lethal phenotype of presumptive unc-37 null alleles (see Materials and Methods) which is indicative of transgene expression in germ line tissues (Kelly et al., 1997).

A single transcript of 2.1 kb was detected on northern blots (data not shown) probed with 1.6 kb EcoRV subfragments (Fig. 2C) of the 5.6 kb rescuing clone. We screened a cDNA library (Barstead and Waterston, 1989) with the 1.6 kb EcoRV probes and sequenced two independently identified clones. The cDNA insert in each clone is 2.1 kb and encodes a single open reading frame of 612 amino acids. The DNA sequence of the overlapping genomic region confirmed the cDNA sequence and identified five introns with strong consensus splice acceptor/donor sequences.

The 5′ end of each of the cDNA clones includes the last seven nucleotides of the SL1 transpliced leader sequence (Blu-menthal, 1995). We used RACE-PCR to show that unc-37 transcripts are transpliced to either the SL1 or SL2 leaders at a position five nucleotides upstream of the ATG start. This finding is consistent with the location of a consensus 3′ splice site (TTTCAG) in the genomic sequence abutting the short 5′ UTR (data not shown). The presence of the SL2 leader indicates that unc-37 is likely to be embedded in an operon in which a tandem array of individual genes are co-transcribed from a common promoter element (Blumenthal, 1995). The unc-37 transcript is polyadenylated and the 3′ UTR includes a consensus polyadenylation signal AATAAA (Fig. 3A).

unc-37 encodes a Groucho-like protein

The unc-37 coding sequence is closely related (~40% overall identity) to the Groucho (Gro) family of transcriptional cofactor proteins (Fig. 3B). All of these proteins share a similar amino-terminal domain (N) and a highly conserved carboxyterminal region of six tandem WD repeats (Neer et al. 1994). The greatest sequence differences occur within a region (V) of variable length between the N and WD domains. A potential nuclear localization signal and presumptive casein kinase II (CKII) and cdc2 phosphorylation sites are located in the V region. Comparable sites have been identified in other Gro family members (Fig. 3A) (Stifani et al., 1992; Koop et al., 1996). The functional significance of the conserved N-terminal sequence is not known but it may include a dimerization domain (Pinto and Lobe, 1996).

The UNC-37 WD domain is the most highly conserved region (approx. 60% identity) (Figs 3B, 4). The WD motif was originally identified in G-protein β and is named for the ‘Trp Asp’ dipeptide that usually terminates each repeat. Different versions of this repeating unit are now known to occur in over 50 different proteins and are believed to mediate proteinprotein interactions (Neer et al., 1994). In the recently determined crystal structure of G-protein β (Wall et al., 1995; Sondek et al., 1996), a tandem array of WD repeats forms a propeller-like structure. Each of the WD repeats folds into a series of four β-pleated sheets that are arranged in an antiparallel fashion to create the individual ‘blades’ of the propeller domain (Fig. 4). The primary sequences of WD-like domains in other proteins, including those of the Groucho family, display a repeating pattern of conserved amino acid side chains that is also compatible with this basic structure (Fig. 4). Our analysis of unc-37 mutations is consistent with this prediction (see below).

Fig. 4.

The WD repeats of UNC-37 are highly conserved and mutations in this region affect function. (A) Amino acid sequences of WD repeats 2, 5 and 6. Dots indicate amino acid identity between UNC-37 and Groucho family members. Residues highlighted in red are predicted to form a hydrogen bonding network (H, S, D) or provide hydrophobic interactions (W) that stabilize the fifth blade of the WD-repeat propeller structure. For the WD repeat consensus sequence x = any amino acid; h = hydrophobic residue; s = S, G, A, T, or C with s’ including Y (Neer et al., 1994). The gap in consensus sequence and corresponding location in WD2 and WD5 accommodates an extra amino acid in this position in WD6. The gap in the WD2 repeat of UNC-37 maximizes alignment with homologs in this region. Predicted locations of β-pleated sheets in each WD repeat are indicated by colored arrows labeled d, a, b, c and by colored lines above each WD repeat. (B) UNC-37 WD domain is likely to form a propeller-like structure (see text). Each propeller blade is composed of four antiparallel β-pleated sheets. WD repeats are staggered with respect to each blade such that β-pleated sheet d in each WD repeat is grouped with β-pleated sheets a, b, and c of the previous WD repeat. The propeller is held together in its toroidal form by β-sheet d from WD repeat 1 interacting with sheets a, b and c of the sixth and C-terminal WD repeat. See text for descriptions of mutations.

Fig. 4.

The WD repeats of UNC-37 are highly conserved and mutations in this region affect function. (A) Amino acid sequences of WD repeats 2, 5 and 6. Dots indicate amino acid identity between UNC-37 and Groucho family members. Residues highlighted in red are predicted to form a hydrogen bonding network (H, S, D) or provide hydrophobic interactions (W) that stabilize the fifth blade of the WD-repeat propeller structure. For the WD repeat consensus sequence x = any amino acid; h = hydrophobic residue; s = S, G, A, T, or C with s’ including Y (Neer et al., 1994). The gap in consensus sequence and corresponding location in WD2 and WD5 accommodates an extra amino acid in this position in WD6. The gap in the WD2 repeat of UNC-37 maximizes alignment with homologs in this region. Predicted locations of β-pleated sheets in each WD repeat are indicated by colored arrows labeled d, a, b, c and by colored lines above each WD repeat. (B) UNC-37 WD domain is likely to form a propeller-like structure (see text). Each propeller blade is composed of four antiparallel β-pleated sheets. WD repeats are staggered with respect to each blade such that β-pleated sheet d in each WD repeat is grouped with β-pleated sheets a, b, and c of the previous WD repeat. The propeller is held together in its toroidal form by β-sheet d from WD repeat 1 interacting with sheets a, b and c of the sixth and C-terminal WD repeat. See text for descriptions of mutations.

unc-37 mutations alter the structure of WD domains

We determined the sequence defects in twelve independently derived unc-37 mutations (Table 1). As homozygotes, the severe lf alleles, wd19, wd20, wd21, and wd22 display a highly pleiotrophic phenotype (Unc-4-like movement defect, protruding vulva, distorted gonad, defective oogenesis, etc; McKim et al. 1992; Miller et al., 1993). A few embryos are produced but are never viable which indicates that these mutations are maternal effect lethals. wd22 is a frameshift mutation in exon five. wd19 and wd21 alter splice acceptor sequences in introns five and two, respectively. All three of these point mutations are predicted to lead to premature termination of unc-37 translation (wd22) or to delete portions of the unc-37 coding sequence (wd19, wd21) and are therefore likely to represent the null phenotype.

Table 1.

unc-37 alleles

unc-37 alleles
unc-37 alleles

The hypomorphic mutation e262 is a histidine-539 to tyrosine substitution in WD repeat 5 (Fig. 4). In the crystal structure of G-protein β (Wall et al., 1995; Sondek et al., 1996), histidine residues in this position contribute to a hydrogen-bonding network with serine and aspartate residues that are also conserved in WD repeat 5 (S557 and D561 in Fig. 4). The severe lf allele wd20, alters a glutamate residue (E394K) in the turn region between β-pleated sheet b and c. In this location, E394 may be needed to stabilize the conformation of WD repeat 2 via intra-domain hydrogen bonds. The strong and highly pleiotropic phenotype of the wd20 mutation is similar to that of the presumptive unc-37 null mutations (Table 1) and may therefore indicate that WD repeat 2 provides an essential UNC-37 function. In contrast, homozygous e262 animals show only the Unc-4 like backward movement defect that all unc-37(lf) mutations display, but are otherwise fertile and morphologically normal (Miller et al., 1993). We interpret this finding to mean that WD repeat 5 is required for unc-37 activity in VA motor neurons (see below) but is apparently dispensable for UNC-37 functions in many other cells. In addition, the Unc-4-like backward movement defect of e262 animals is consistent with the idea that WD repeat 5 mediates a process that also depends on UNC-4 function.

Suppressor mutations alter a conserved residue in the sixth WD repeat of UNC-37

We have previously described the isolation of dominant suppressors of unc-4(e2322ts), which map to unc-37 (Miller et al., 1993). All four of these mutations carry a glutamate-580 to lysine (E to K) substitution in the C-terminal (sixth) WD repeat. We have also shown that an unc-37 transgene (pJM1-wd17dm) bearing the E580K mutation restores normal movement to unc-4(e2322ts) animals whereas a wild-type unc-37 transgene (pJM1) does not (data not shown). This experiment confirms that the E580K mutation accounts for the suppressor activity of unc-37(gf) mutations. Although the E-580 residue is conserved in all known groucho family members (Stifani et al. 1992; Koop et al., 1996) it does not provide an essential function since these suppressor mutants do not display a visible phenotype in a wild-type genetic background (Miller et al., 1993). In the hypothetical structure of the the UNC-37 WD domain, the E-580 side chain is located on the ‘top’ surface of the sixth propeller blade and does not have an obvious structural role, but could be involved in interactions with other proteins (Fig. 4B).

In this work, we isolated three dominant, intragenic suppressors of unc-37(e262). We find that all of these e262 suppressors also correspond to the unc-37 E580K substitution that suppresses unc-4(e2322ts) and other specific unc-4 missense mutations (A. R. Winnier and D. M. Miller, unpublished data). Thus, the alteration of a single amino acid in UNC-37 is sufficient to suppress an Unc-4-like movement defect that arises from specific missense mutations in either UNC-4 or UNC-37. This observation supports the notion that e262 perturbs a structure necessary for UNC-4 activity. The E580K mutation is not sufficient, however, to rescue the severe Unc-37 phenotype of wd20 animals. wd20 was obtained by reverting wd17dm; DNA sequencing confirmed that these animals still carry both mutations (data not shown). Therefore, the E580K substitution can suppress a specific UNC-37 hypomorphic allele but cannot compensate for mutations that produce the presumptive unc-37 null phenotype.

UNC-37 is ubiquitously expressed and nuclearlocalized

To identify the cells in which UNC-37 is expressed, we raised rabbit polyclonal antibodies to a GST-UNC-37 fusion protein (Fig. 5). Immunofluorescence staining with the affinity-purified antibody detected UNC-37 in the nuclei of apparently all somatic cells throughout development (Fig. 5 and data not shown). A transgene expressing an UNC-37-gfp fusion protein is also localized to the nucleus and displays a similar pattern of widespread expression (Kelly et al., 1997). Ubiquitous, nuclearlocalized expression has been previously reported for Drosophila Gro (Delidakis et al., 1991) and human TLE proteins (Stifani et al., 1992; Dehni et al., 1995).

Fig. 5.

UNC-37 is nuclear-localized and ubiquitously expressed. (A) Immunoblot showing that anti-UNC-37 stains the GST2 fusion protein in bacterial extracts and a single approx. 70×103Mr band in C. elegans wild-type (N2) homogenate. (B) Anti-UNC-37 stained and (C) DIC image of wild-type hermaphrodite gonad. Enlarged, brightly stained oocyte nuclei occupy the proximal arm of the gonad. Arrows denote sheath cell nuclei. (D) Distal arm of wild-type hermaphrodite gonad stained with anti-UNC-37. Asterisk marks distal tip cell. (E) Wild-type embryo (seven cell nuclei) and (F) L1 larva stained with anti-UNC-37. (G) Posterior region of late L1/early L2 larva stained with DAPI (blue) to mark all cell nuclei and with anti-β-galactosidase (red) to identify lacZ-positive motor neurons in the unc4-lacZ transgenic strain, wdIs1. unc-4-lacZ is expressed in all DA and VA motor neurons in the ventral nerve cord; however, VA8, VA11, and VA12 do not express lacZ in this animal. lacZ-stained cells in the preanal ganglion (PAG) are DA8 and DA9. (H) Same animal as in G, stained with anti-UNC-37 (green). Cell nuclei of DA and VA motor neurons are marked with white lines. Horizontal scale bars, 20 μm.

Fig. 5.

UNC-37 is nuclear-localized and ubiquitously expressed. (A) Immunoblot showing that anti-UNC-37 stains the GST2 fusion protein in bacterial extracts and a single approx. 70×103Mr band in C. elegans wild-type (N2) homogenate. (B) Anti-UNC-37 stained and (C) DIC image of wild-type hermaphrodite gonad. Enlarged, brightly stained oocyte nuclei occupy the proximal arm of the gonad. Arrows denote sheath cell nuclei. (D) Distal arm of wild-type hermaphrodite gonad stained with anti-UNC-37. Asterisk marks distal tip cell. (E) Wild-type embryo (seven cell nuclei) and (F) L1 larva stained with anti-UNC-37. (G) Posterior region of late L1/early L2 larva stained with DAPI (blue) to mark all cell nuclei and with anti-β-galactosidase (red) to identify lacZ-positive motor neurons in the unc4-lacZ transgenic strain, wdIs1. unc-4-lacZ is expressed in all DA and VA motor neurons in the ventral nerve cord; however, VA8, VA11, and VA12 do not express lacZ in this animal. lacZ-stained cells in the preanal ganglion (PAG) are DA8 and DA9. (H) Same animal as in G, stained with anti-UNC-37 (green). Cell nuclei of DA and VA motor neurons are marked with white lines. Horizontal scale bars, 20 μm.

We also observed strong UNC-37 expression in germ line nuclei. As shown in Fig. 5B, UNC-37 is expressed in an apparent gradient in the proximal arm of the gonad with the highest levels of UNC-37 in the enlarged nuclei of oocytes adjacent to the spermatheca. A lower level of UNC-37 staining was observed in germ line nuclei in the distal arm of the gonad (Fig. 5B,D). We note that Gro is abundantly expressed in the Drosophila oocytes and nurse cells (Tata and Hartley, 1993).

The widespread expression of UNC-37 thoughout somatic and germ line tissues is consistent with the highly pleiotropic phenotype of the putative unc-37 null alleles. For example, the elimination of UNC-37 expression in the distal tip cell (Kimble, 1981) (Fig. 5D) could account for the highly distorted gonad morphology of the severe unc-37 alleles (Miller et al., 1993; and data not shown). Embryonic development is likely to depend on maternally provided UNC-37, since homozygous unc-37 animals arising from heterozygous mothers reach adulthood but produce dead embryos. These homozygous unc-37 adults, however, also exhibit multiple defects in postem-bryonically derived cells (e.g. motor neurons, germ line and somatic gonad, vulva) which may be indicative of an essential role for zygotic unc-37 expression during larval development. It is noteworthy that in Drosophila, embryogenesis has been shown to require maternally derived Gro (Paroush et al., 1994) whereas postembryonic development relies on zygotic Gro expression (Delidakis et al., 1991).

Backward movement depends on UNC-37 expression in VA motor neurons

A transgenic line, wdIs1, carrying chromosomally integrated copies of an unc-4-lacZ reporter plasmid (Miller and Niemeyer, 1995), was double-stained with anti-β-galactosidase (red) and with anti-UNC-37 (green) to show that UNC-4 and UNC-37 are coexpressed in the DA and VA motor neurons in the ventral nerve cord (Fig. 5 G,H). UNC-37 expression in the DA and VA motor neurons was also observed in wild-type (N2) animals (data not shown). The backward movement defect of unc-4 mutants has been attributed to the absence of unc-4 function in the VA motor neurons (Miller and Niemeyer, 1995). The DA motor neurons are not miswired in unc-4 mutants, however, and appear to function normally (White et al., 1992).

Despite the coincident expression of UNC-4 and UNC-37 in VA motor neurons, the widespread appearance of UNC-37 throughout the nervous system raised the possibility that the uncoordinated movements displayed by unc-37(e262) could be due to defects in other neuronal cells. To address this question, we expressed the UNC-37 protein under the control of the unc-4 promoter region (pJM1) in unc-37(e262) animals (Fig. 6A). Backward locomotion is restored to unc-37(e262) animals carrying pJM1. In these transgenic animals, expression of wildtype UNC-37 protein should be restricted to cells that also express UNC-4 (Miller and Niemeyer, 1995). To confirm that this pattern of expression does occur in rescued animals, we modified pJM1 to create pJM1.1 in which the green fluorescent protein (gfp; Chalfie et al., 1994) was fused to the N terminus of UNC-37. pJM1.1 also rescued the e262 backward movement defect and is expressed in DA and VA motor neurons in the ventral nerve cord and in certain other neurons in ganglia at either end of the cord (Fig. 6) (Miller and Niemeyer, 1995). We interpret these experiments to mean that the unc-37(e262) movement defect is a direct result of UNC-37 dysfunction in the VA motor neurons and that rescue depends on UNC-37 expression in these cells.

Fig. 6.

Expression of UNC-37 or UNC-37-TLE1 in VA motor neurons rescues the backward movement defect of unc-37(e262). The expression plasmids pJM1, pJM1.1, pNE-1 and pJM-TLE1.2, contain a 3.2 kb genomic PstI-SmaI fragment and pJM4 contains a 2.8 kb HindIII-SmaI fragment both of which span the unc-4 upstream region and exon1 and part of exon2 (black vertical rectangles) but do not include the unc-4 homeodomain (Miller and Niemeyer, 1995). pNE-1 does not rescue e262 which indicates that the N-terminal UNC-4 fragment does not account for e262 rescue by pJM1, pJM1.1, or pJM4. Plasmid pJM4 encodes a fusion protein containing the N and V domains of UNC-37 (N-unc-37) and the WD domain of TLE1 (TLE1-C).A rescued L2 larva carrying the pJM1.1 transgene showing gfppositive motor neurons in the ventral nerve cord. The gonad is marked with an asterisk. Magnified image of posterior region of animal shown in B. Horizontal scale bars, 20 μm.

Fig. 6.

Expression of UNC-37 or UNC-37-TLE1 in VA motor neurons rescues the backward movement defect of unc-37(e262). The expression plasmids pJM1, pJM1.1, pNE-1 and pJM-TLE1.2, contain a 3.2 kb genomic PstI-SmaI fragment and pJM4 contains a 2.8 kb HindIII-SmaI fragment both of which span the unc-4 upstream region and exon1 and part of exon2 (black vertical rectangles) but do not include the unc-4 homeodomain (Miller and Niemeyer, 1995). pNE-1 does not rescue e262 which indicates that the N-terminal UNC-4 fragment does not account for e262 rescue by pJM1, pJM1.1, or pJM4. Plasmid pJM4 encodes a fusion protein containing the N and V domains of UNC-37 (N-unc-37) and the WD domain of TLE1 (TLE1-C).A rescued L2 larva carrying the pJM1.1 transgene showing gfppositive motor neurons in the ventral nerve cord. The gonad is marked with an asterisk. Magnified image of posterior region of animal shown in B. Horizontal scale bars, 20 μm.

Immunostaining experiments with anti-UNC-37 indicate that unc-4 mutations have no detectable effect on UNC-37 expression in the VA motor neurons (data not shown). Similarly, unc-4 expression is apparently normal in unc-37(e262;Miller et al., 1993). These findings rule out the possibility that either unc-4 or unc-37 regulate each other’s expression and favor a model in which UNC-4 and UNC-37 are independently expressed.

The WD domains of human TLE1 and nematode UNC-37 are functionally interchangeable in VA motor neurons

The human genome contains four separate but highly similar groucho-like genes, TLE1, TLE2, TLE3, and TLE4 (Stifani et al., 1992). In order to determine if a human TLE protein and UNC-37 are functionally homologous, full length TLE1 was expressed in C. elegans. Although expression of the intact TLE1 protein in VA motor neurons does not rescue e262, a chimeric protein containing N and V domains from UNC-37 fused to the WD domain of TLE1 is sufficient to restore normal movement to e262 animals (Fig. 6A). This finding parallels the degree of sequence similarity between the TLE and UNC-37 proteins which is highest in the WD domain and indicates that the function of the WD repeat domain of UNC-37 is conserved in its vertebrate homologs.

VA motor neuron differentiation depends on UNC-4 and UNC-37

unc-4 encodes a homeodomain protein (Miller et al., 1992) that specifies synaptic input to one class of motor neurons in the C. elegans ventral nerve cord (White et al., 1992). In unc-4 mutants, VA motor neurons adopt synaptic inputs normally reserved for their lineal sisters, the VB motor neurons (Fig. 1). As a result of this wiring defect, unc-4 animals are unable to crawl backward (White et al., 1992). Expression of an unc-4 transgene in the VA motor neurons restores backward locomotion to unc-4 mutants. We have therefore proposed that UNC-4 normally specifies a VA trait that allows presynaptic partners to distinguish VA motor neurons from their VB sisters (Miller and Niemeyer, 1995).

Loss-of-function alleles of the unc-37 gene produce a backward movement defect that is identical to that of unc-4 mutants (Miller et al., 1993). In this work, we show that expression of the unc-37 gene in VA motor neurons restores wild-type movement to unc-37 mutant animals. These findings indicate that UNC-37 functions in the VA motor neurons to specify the same phenotypic trait as UNC-4. We have shown, however, that unc-37 does not regulate unc-4 expression (Miller et al., 1993) and, conversely, that unc-4 mutations do not affect UNC-37 immunostaining in VA motor neurons (this work). We conclude that neither UNC-4 nor UNC-37 are upstream of each other in a linear genetic pathway and favor the alternative idea that UNC-4 and UNC-37 are independently regulated and simulaneously active in VA motor neurons.

What is the mechanism of UNC-4-UNC-37 interaction?

We consider two possible models of UNC-4 and UNC-37 action:

  1. UNC-4 and UNC-37 act in parallel pathways to regulate separate sets of target genes, both of which are required for the specification of synaptic input to VA motor neurons. This model is consistent with the observation that lf mutations in either unc-37 or unc-4 produce an identical backward movement defect (Miller et al., 1993). The allele specificity of the unc-37(gf) mutation could be explained by the assumption that the suppressible unc-4 alleles retain partial function which can be enhanced by elevated unc-37 activity. In this scenario, unc-4 missense and null alleles that do not retain residual unc-4 function are not suppressed.

  2. In the alternative model, UNC-4 and UNC-37 regulate a common set of target genes. In this case, the unc-37(gf) suppressor allele restores functional interactions between UNC-4 and UNC-37 that are otherwise perturbed by either the unc-37(e262) mutation or by specific unc-4 missense mutations (Miller et al., 1993; A. R. Winnier and D. M. Miller, unpublished data). unc-4 null alleles are thus not suppressible by the unc-37 E580K protein because no mutant UNC-4 is present to interact with modified UNC-37.

The UNC-37-like protein, Gro, functions as a corepressor protein with Hairy and closely related bHLH proteins encoded by the Drosophila enhancer of split [E(spl)] locus (Paroush et al., 1994; Fisher et al., 1996). Similarly, in yeast, the WD repeat protein, Tup1, interacts with the α-2 homeodomain to block transcription of target genes (Komachi et al., 1994). In both of these cases, Gro and Tup1 are in physical contact with specific transcription factors. However, we have not detected direct association between UNC-4 and UNC-37 under conditions in which Gro and Hairy bind to each other with high affinity (A. Pflugrad, J. Y.-J. Meir and D. M. Miller, unpublished findings). Thus, if UNC-4 and UNC-37 are members of a common transcription complex, as we have hypothesized, it must follow that the molecular mechanism of UNC-4-UNC-37 interaction differs from that of Gro and Hairy. Perhaps accessory proteins are required to stablize UNC-4 interactions with UNC-37. The WD repeat protein, dTAF1180, for example, is tightly associated with the multisubunit RNA II polymerase complex but fails to bind individual protein components (Dynlacht et al., 1993).

Parallels to groucho function in Drosophila

The ubiquitous expression of UNC-37 and the highly pleiotropic unc-37 null phenotype parallel similar findings for Gro in Drosophila and suggest that Groucho-like proteins provide essential functions in a variety of different cells. Genetic evidence implicates groucho in multiple pathways including sex determination, segmentation and neurogenesis (Delidakis et al., 1991; Paroush et al., 1994). Groucho functions downstream of Notch, for example (Heitzler et al., 1996), and is proposed to act as a corepressor protein with E(spl) bHLH proteins to block expression of achaete scute genes. Hairy-dependent repression of the pair rule genes fushi tarazu (ftz) and engrailed (en) requires maternal gro function in the embryo (Paroush et al., 1994). The observation that Groucho-like proteins are found in tissues that do not express Hairy or E(spl) bHLH proteins (Delidakis et al., 1991; Koop et al., 1996) is indicative of functional interactions with still other classes of specific transcription factors. The homeoprotein, eve, for example, requires gro function to repress ftz transcription in the Drosophila embryo (Paroush et al., 1994).

We have extended the known repertoire of gro functions to include an explicit role in the specification of motor neuron fate in C. elegans. The finding that a rare hypomorphic gro allele selectively alters photoreceptor fate in Drosophila (FischerVize et al., 1992) is analogous to our observation that the unc-37(e262) mutation disables UNC-37 function in the VA motor neurons but does not produce obvious defects in other cells. The e262 missense mutation is predicted to destabilize the structural domain encompassed by WD repeat 5. Apparently, the WD5 domain is required for an UNC-4 dependent trait but its structural integrity is less important for other UNC-37 functions. These findings are consistent with the idea that each of the WD domains could mediate UNC-37 interactions with different transcription factors or components of the transcription machinery (Komachi et al., 1994).

Evolutionary conservation of function in the Groucho family

The fundamental domain structure and amino acid sequence of Groucho-like proteins are remarkably well conserved from nematodes to humans. The sequence identity of the WD repeat domain approaches 90% for Drosophila Gro and 60% for C. elegans UNC-37 in comparison to human TLE proteins (Stifani et al., 1992). Furthermore, we have shown that the WD repeat domains of UNC-37 and of a human Groucho-like proteins are functionally interchangeable in nematodes. The failure of the intact TLE1 protein to complement unc-37(e262) reflects the lower level of conservation of the N-terminal UNC-37 sequence and indicates that some primordial Gro functions have not been conserved in evolution.

Our finding that the WD domain of a human Gro protein can substitute for the UNC-37 WD domain in nematode motor neurons suggests that the capacity to interact with relevant components of the C. elegans transcriptional machinery has been conserved in vertebrate Gro proteins. At least one of the TLE proteins has been detected as a nuclear-localized antigen in differentiating motor neurons of the vertebrate spinal cord (Dehni et al., 1995). In addition, apparent sequence homologs (approx. 90% identity) of the unc-4 homeodomain have been reported in mammals (Saito et al., 1996; Rovescalli et al., 1996). Thus, it is reasonable to suppose that orthologs of UNC-37 and UNC-4 may function together to define motor neuron identity and synaptic specificity in the vertebrate CNS.

UNC-4 and UNC-37 inhibit VB fate?

By analogy to the known function of Gro in Drosophila (Paroush et al., 1994), we propose that unc-37 and unc-4 repress VB specifying genes in the VA motor neurons (Fig. 7). In this model, gene products that promote VA differentiation are controlled by other regulatory factors. If either UNC-4 or UNC-37 activity is eliminated, the ‘default’ VB pathway is derepressed and is dominant to VA traits. The net result is that loss of either unc-37 or unc-4 function is sufficient to respecify a VA motor neuron to accept synaptic inputs normally reserved for their VB sister neurons.

Fig. 7.

Model of UNC-4 and UNC-37 action in VA motor neurons. We propose that UNC-4 and UNC-37 function together as negative regulators of genes that specify the VB pattern of synaptic input.

Fig. 7.

Model of UNC-4 and UNC-37 action in VA motor neurons. We propose that UNC-4 and UNC-37 function together as negative regulators of genes that specify the VB pattern of synaptic input.

Although we have not directly demonstrated that VA motor neurons are miswired with VB-specific inputs in unc-37(lf) mutants, we have ruled out other possible defects that could also produce Unc-4 like movement in these animals. VA morphology and axonal polarity, for example, are not detectably altered by the e262 mutation (Miller et al., 1993). A definitive resolution of this issue must await EM reconstruction of the VA motor neuron circuit in unc-37(e262) animals.

Our data are compatible with a model in which UNC-4 and UNC-37 act in parallel pathways to regulate different groups of target genes both of which are required to specify VA inputs or a case in which UNC-4 and UNC-37 are members of a common transcription complex that regulates a single set of downstream genes. It should be possible to distinguish between these models by performing co-immunoprecipitation experiments with specific UNC-4 and UNC-37 antibodies.

We thank D. Greenstein for comments on the manuscript; Lucy Liaw for Fig. 5B; K. Anders and P. Anderson for plasmids TR258 and TR259 and information prior to publication; A. Coulson and R. Shownkeen for cosmids; R. Barstead for a C. elegans cDNA library; N. Emambokus for constructing plasmid pNE-1; A. Fire for plasmid pPD95.67; S. Stifani for TLE1 cDNA, and F. Schachat, M. Sheetz and G. Ruvkun for their support. Some of the strains used in this work were provided by the Caenorhabditis Genetics Center (University of Minnesota, St. Paul), which is funded by the NIH NCRR. This work was supported by a grant from NIH/NINDS (R01NS26115) to D. M. M. and by the Lucille P. Markey Charitable Trust.

Altschul
,
S. F.
,
Gish
,
W.
,
Miller
,
W.
,
Myers
,
E. W.
and
Lipman
,
D. J.
(
1990
).
Basic local alignment search tool
.
J. Mol. Biol.
215
,
403
410
.
Barnes
,
T. M.
,
Jin
,
Y.
,
Horvitz
,
H. R.
,
Ruvkun
,
G.
and
Hekimi
,
S.
(
1996
).
The Caenorhabditis elegans behavioral gene unc-24 encodes a novel bipartite protein similar to both erythrocyte band 7.2 (Stomatin) and nonspecific lipid transfer protein
.
J. Neurochem.
67
,
46
57
.
Barstead
,
R. J.
and
Waterston
,
R. H.
(
1989
).
The basal component of the nematode dense-body is vinculin
.
J. Biol. Chem.
264
,
10177
10188
.
Blumenthal
,
T.
(
1995
).
Trans-splicing and polycistronic transcription in Caenorhabditis elegans
.
Trends Genet.
11
,
132
136
.
Brenner
,
S.
(
1974
).
The genetics of Caenorhabditis elegans
.
Genetics
77
,
71
94
.
Chalfie
,
M.
,
Tu
,
Y.
,
Euskirchen
,
G.
,
Ward
,
W. W.
and
Prasher
,
D. C.
(
1994
).
Green fluorescent protein as a marker for gene expression
.
Science
263
,
802
805
.
Clark
,
D.V
,
Suleman
,
D. S.
,
Beckenbach
,
K. A.
,
Gilchrist
,
E. J.
and
Baillie
,
D. L.
(
1995
).
Molecular cloning and characterization of the dpy-20 gene of Caenorhabditis elegans
.
Mol. Gen. Genet.
247
,
367
378
.
Coulson
A.
,
Huynh
C.
,
Kozono
Y.
, and
Shownkeen
R.
(
1995
).
The physical map of the Caenorhabditis elegans genome
. In
C. elegans: Modern Biological Analysis of an Organism.
(ed.
H.F.
Epstein
and
D.C.
Shakes
) Methods in Cell Biology
48
,
533
550
.
Dehni
,
G.
,
Liu
,
Y.
,
Husain
,
J.
and
Stifani
,
S.
(
1995
).
TLE expression correlates with mouse embryonic segmentation, neurogenesis, and epithelial determination
.
Mechanisms of Development
53
,
369
381
.
Delidakis
,
C.
,
Preiss
,
A.
,
Hartley
,
D. A.
and
Artavanis-Tsakonas
,
S.
(
1991
).
Two genetically and molecularly distinct functions involved in early neurogenesis reside within the Enhancer of split locus of Drosophila melanogaster
.
Genetics
129
,
803
823
.
Dynlacht
,
B. D.
,
Weinzierl
,
R. O.
,
Admon
,
A.
and
Tjian
R.
(
1993
).
The dTAFII80 subunit of Drosophila TFIID contains beta-transducin repeats. Nature
363
,
176
-
179
, 1993
Finney
,
M.
and
Ruvkun
,
G.
(
1990
).
The unc-86 gene product couples cell lineage and cell identity in C. elegans
.
Cell
63
,
895
905
.
Fischer-Vize
,
J. A.
,
Vize
,
P. D.
and
Rubin
,
G. M.
(
1992
).
A unique mutation in the Enhancer of split gene complex affects the fates of the mystery cells in the developing Drosophila eye
.
Development
115
,
89
101
.
Fisher
,
A. L.
,
Ohsako.
S.
and
Caudy
,
M.
(
1996
).
The WRPW motif of the hairy-related basic helix-loop-helix repressor proteins acts as a 4-amino-acid transcription repression and protein-protein interaction domain
.
Mol. Cell. Biol.
16
,
2670
2677
.
Francis
,
R.
,
Barton
,
M. K.
,
Kimble
,
J.
and
Schedl
,
T.
(
1995
).
gld-1, a tumor suppressor gene required for oocyte development in Caenorhabditis elegans
.
Genetics
139
,
579
606
.
Harlow
,
E.
and
Lane
,
D.
(
1988
).
Antibodies: A Laboratory Manual
Cold Spring Harbor, New York
:
Cold Spring Harbor Laboratory Press
.
Hartley
,
D. A.
,
Preiss
,
A.
and
Artavanis-Tsakonas
,
S.
(
1988
).
A deduced gene product from the Drosophila neurogenic locus, enhancer of split, shows homology to mammalian G-protein beta subunit
.
Cell
55
,
785
795
.
Heitzler
,
P.
,
Bourouis
,
M.
,
Ruel
,
L.
,
Carteret
,
C.
and
Simpson
,
P.
(
1996
).
Genes of the Enhancer of split and achaete-scute complexes are required for a regulatory loop between Notch and Delta during lateral signalling in Drosophila
.
Development
122
,
161
71
.
Kelly
,
W. G.
,
Xu
,
S. Q.
,
Montgomery
,
M. K.
and
Fire
,
A.
(
1997
).
Distinct requirements for somatic and germline expression of a generally expressed
C. elegans gene. Genetics in press.
Kimble
,
J.
(
1981
).
Alterations in cell lineage following laser ablation of cells in the somatic gonad of Caenorhabditis elegans
.
Dev. Biol.
87
,
286
300
.
Komachi
,
K.
,
Redd
,
M. J.
and
Johnson
,
A. D.
(
1994
).
The WD repeats of Tup1 interact with the homeo domain protein alpha2
.
Genes Dev.
8
,
2857
2867
.
Koop
,
K. E.
,
MacDonald
,
L. M.
and
Lobe
,
C. G.
(
1996
).
Transcripts of Grg4, a murine groucho-related gene, are detected in adjacent tissues to other murine neurogenic gene homologues during embryonic development
.
Mech. Dev.
59
,
73
87
McKim
,
K.
,
Starr
,
S. T.
and
Rose
,
A. M.
(
1992
).
Genetic and molecular analysis of the dpy-14 region in Caenorhabditis elegans
.
Mol. Gen. Genet.
233
,
241
251
.
Mello
,
C. C.
,
Kramer
,
J. M.
,
Stinchcomb
,
D.
and
Ambros
,
V.
(
1991
).
Efficient gene transfer in C. elegans: extrachromosomal maintenance and integration of transforming sequences
.
EMBO J.
10
,
3959
3970
.
Miller
,
D. M.
,
Shen
,
M. M.
,
Shamu
,
C. E.
,
Bürglin
,
T. R.
,
Ruvkun
,
G.
,
Dubois
,
M. L.
,
Ghee
,
M.
and
Wilson
,
L.
(
1992
).
C. elegans unc-4 gene encodes a homeodomain protein that determines the pattern of synaptic input to specific motor neurons
.
Nature
355
,
841
845
.
Miller
,
D. M.
, III
,
Niemeyer
C.J.
, and
Chitkara
,
P.
(
1993
).
Dominant unc-37 mutations suppress the movement defect of a homeodomain mutation in unc-4, a neural specificity gene in Caenorhabditis elegans
.
Genetics
135
,
741
753
.
Miller
,
D.M.
, III
and
Niemeyer
,
C.J.
(
1995
).
Expression of the unc-4 homeoprotein in C. elegans motor neurons specifies presynaptic input
.
Development
121
,
2877
2886
.
Miller
,
D. M.
, III and
Shakes
,
D. C.
(
1995
). Immunofluorescence Microscopy. In
C. elegans: Modern Biological Analysis of an Organism
(ed.
H. F.
Epstein
and
D.C.
Shakes
) Methods in Cell Biology.
48
,
365
392
.
Miller
,
D. M.
,
Stockdale
,
F.
and
Karn
,
J.
(
1986
).
Immunological identification of the genes encoding the four myosin heavy chains of Caenorhabditis elegans
.
Proc. Nat. Acad. Sci. USA
83
,
2305
230
.
Miyasaka
,
H.
,
Choudry
,
B. K.
,
Hou
,
E. W.
, and
Li
,
S. S.-L.
(
1993
).
Molecular cloning and expression of mouse and human cDNA encoding AES and ESG proteins with strong similarity to Drosophila enhancer of split groucho protein
.
Eur. J. Biochem.
216
,
343
352
.
Neer
,
E. J.
,
Schmidt
,
C. J.
,
Nambudripad
,
R.
and
Smith
,
T. F.
(
1994
).
The ancient regulatory-protein family of WD-repeat proteins
.
Nature
371
,
297
300
Paroush
,
Z.
,
Finley
,
R. L.
,
Kidd
,
T.
,
Wainwright
,
S. M.
,
Ingham
,
P. W.
,
Brent
,
R.
and
Ish-Horowicz
,
D.
(
1994
).
Groucho is required for Drosophila neurogenesis, segmentation, and sex determination and interacts directly with hairy-related bHLH proteins
.
Cell
79
,
805
815
.
Pinto
,
M.
and
Lobe
,
C. G.
(
1996
).
Products of the grg (Groucho-related gene) family can dimerize throught the amino-terminal Q domain
.
J. Biol. Chem.
271
,
33026
33031
.
Rovescalli
,
A. C.
,
Asoh
,
S.
and
Nirenberg
,
M.
(
1996
).
Cloning and characterization of four murine homeobox genes
.
Proc. Natl. Acad. Sci. USA
93
,
10691
10696
.
Saito
,
T.
,
Lo
,
L.
,
Anderson
,
D. J.
and
Mikoshiba
,
K.
(
1996
).
Identification of a novel paired homeodomain protein related to C. elegans UNC-4 as a potential downstream target of MASH1
.
Dev. Biol.
180
,
143
155
Sambrook
,
J.
,
Fritsch
,
E. F.
and
Maniatis
,
T.
(
1989
).
Molecular Cloning: A Laboratory Manual
.
New York
:
Cold Spring Harbor Laboratory
.
Schmidt
,
C. J.
and
Sladek
,
T. E.
(
1993
).
A rat homolog of the Drosophila enhancer of split (groucho) locus lacking WD-40 repeats
.
J. Bio. Chem.
268
,
25681
25686
.
Smith
,
D. B.
and
Johnson
,
K. S.
(
1988
).
Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene
67
,
31
40
.
Sondek
,
J.
,
Bohm
,
A.
,
Lambright
,
D. G.
,
Hamm
,
H. E.
and
Sigler
,
P. B.
(
1996
).
Crystal structure of a G-protein beta gamma dimer at 2.1A resolution
.
Nature
379
,
369
374
.
Stifani
,
S.
,
Blaumueller
,
C. M.
,
Redhead
,
N. J.
,
Hill
,
R. E.
and
Artavanis-Tsakonas
,
S.
(
1992
).
Human homologs of a Drosophila enhancer of split gene product define a novel family of nuclear proteins
.
Nature Genetics
2
,
119
127
.
Sulston
,
J. E.
and
Horvitz
,
H. R.
(
1977
).
Post-embryonic cell lineages of the nematode, Caenorhabditis elegans
.
Dev. Biol.
56
,
110
156
.
Tata
,
F.
and
Hartley
,
D. A.
(
1993
).
The role of the Enhancer of split complex during cell fate determination in Drosophila
.
Development
Supplement,
139
148
.
Tzamarias
,
D.
and
Struhl
,
K.
(
1995
).
Distinct TPR motifs of Cyc8 are involved in recruiting the cyc8-Tup1 corepressor complex to differentially regulated promoters
.
Genes Dev.
9
,
821
831
.
Wall
,
M. A.
,
Coleman
,
D. E.
,
Lee
,
E.
,
Iniguez-Lluhi
,
J. A.
,
Posner
,
B. A.
,
Gilman
,
A. G.
, and
Sprang
,
S. R.
(
1995
).
The structure of the G protein heterotrimer Gi alpha 1 beta 1 gamma 2
.
Cell
83
,
1047
1058
(
1995
).
White
,
J. G.
,
Southgate
,
E.
,
Thomson
,
J. N.
and
Brenner
,
S.
(
1986
).
The structure of the nervous system of the nematode Caenorhabditis elegans
.
Phil. Trans. R. Soc. Lond. B
314
,
1
340
.
White
,
J. G.
,
Southgate
,
E.
and
Thomson
,
J. N.
(
1992
).
Mutations in the Caenorhabditis elegans unc-4 gene alter the synaptic input to ventral cord motor neurons. Nature
355
,
838
841
.
Williams
,
B. D.
,
Schrank
,
B.
,
Huynh
,
C.
,
Shownkeen
,
R.
and
Waterston
,
R. H.
(
1992
).
A genetic-mapping system in Caenorhabditis elegans based on polymorphic sequence-tagged sites
.
Genetics
131
,
609
624
.