Null Mutations in the lin-31 Gene Indicate Two Functions During Caenorhabditis elegans Vulval Development

Corresponding author: Leilani M. Miller, Department of Biology, 500 El Camino Real, Santa Clara University, Santa Clara, CA 95053., lmiller{at}scu.edu (E-mail)

Communicating editor: R. K. HERMAN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The lin-31 gene is required for the proper specification of vulval cell fates in the nematode Caenorhabditis elegans and encodes a member of the winged-helix family of transcription factors. Members of this important family have been identified in many organisms and are known to bind specific DNA targets involved in a variety of developmental processes. DNA sequencing of 13 lin-31 alleles revealed six nonsense mutations and two missense mutations within the DNA-binding domain, plus three deletions, one transposon insertion, and one frameshift mutation that all cause large-scale disruptions in the gene. The missense mutations are amino acid substitutions in the DNA-binding domain and probably disrupt interactions of the LIN-31 transcription factor with its DNA target. In addition, detailed phenotypic analysis of all 19 alleles showed similar penetrances for several characteristics examined. From our analysis we conclude: (1) the null phenotype of lin-31 is the phenotype displayed by almost all of the existing alleles, (2) the DNA-binding domain plays a critical role in LIN-31 function, and (3) direct screens for multivulva and vulvaless mutants will probably yield only null (or strong) alleles of lin-31.

PROTEINS belonging to the winged-helix family of transcription factors appear to play important roles in many developmental processes, including establishment of body axes, cell fate specification, differentiation of tissues, maintenance of cellular differentiation, and tumorigenesis (see KAUFMANN and KNOCHEL 1996 for review of winged-helix proteins). The LIN-31 protein, a member of this winged-helix family, is required for the proper specification of vulval cell fates during Caenorhabditis elegans vulval development (MILLER et al. 1993 ). During vulval development, six vulval precursor cells (VPCs), which are initially equivalent in developmental potential, differentiate and adopt one of three possible vulval cell fates (for review, see GREENWALD 1997 ). The different fates of the six VPCs (P3.p–P8.p) are determined in response to several extracellular signals and are adopted in the pattern 3°-3°-2°-1°-2°-3°. Cells adopting a 1° or 2° (induced) cell fate give rise to the cells that form the vulva, while cells that adopt a 3° (uninduced) cell fate give rise to cells that fuse with the hypodermal syncytium that surrounds most of the worm. Biochemical and molecular studies of the anchor cell signaling pathway have revealed a conserved receptor tyrosine kinase (RTK)/Ras/mitogen-activated protein (MAP) kinase pathway (for review, see KORNFELD 1997 ). Mutations in genes belonging to this RTK signaling pathway cause the misspecification of vulval cell fates and typically fall into two broad phenotypic classes, vulvaless (Vul) or multivulva (Muv). In Vul mutants, all six VPCs typically adopt nonvulval (3°) cell fates, due to a less active or completely inactive RTK signaling pathway. In Muv mutants, most or all of the VPCs adopt vulval fates (1° or 2°), due to inappropriate activation of the RTK pathway. In the dissecting microscope, a Muv phenotype is manifested by multiple ectopic pseudovulvae (protrusions) along the ventral side of the animal. Mutations in parallel or modulating pathways may also cause a Vul or Muv phenotype (for review, see GREENWALD 1997 ).

Animals carrying mutations in the lin-31 gene, however, display an interesting phenotype that is unique among vulval development mutants (MILLER et al. 1993 ). In contrast to the Muv and Vul mutations, which affect all six VPCs in the same way, these mutations result in a deregulated cell fate phenotype in which each of the six VPCs can adopt any one of the three possible cell fates. Furthermore, opposite cell fate transformations are often observed in the same lin-31 mutant animal, in which a cell distal to the anchor cell expresses an induced cell fate (1° or 2°), while a proximal cell simultaneously expresses an uninduced (3°) cell fate. These opposite cell fate transformations indicate that, in lin-31 mutants, the choice of cell fate is no longer strictly dependent on proximity to the anchor cell. This deregulation of vulval cell fate is manifested by variable Vul and/or Muv phenotypes. Inappropriate expression of an induced fate at P3.p, P4.p, or P8.p results in an ectopic pseudovulva (protrusion), thus causing the animal to appear Muv when observed in a dissecting microscope. Conversely, inappropriate expression of a 3° nonvulval fate by P5.p, P6.p, or P7.p can result in a Vul animal. Some lin-31 mutant animals express both Vul and Muv phenotypes simultaneously. The observation that all six VPCs require wild-type lin-31 activity suggested that lin-31 has alternative functions (MILLER et al. 1993 ). Thus, if the deregulated mutant phenotype does indeed represent the lin-31 null phenotype, the LIN-31 protein would appear to have two functions: to promote vulval induction (in P5.p, P6.p, and P7.p) and to inhibit it (in P3.p, P4.p, and P8.p). However, because incomplete penetrance is a predictable consequence of a deregulation phenotype, it has been difficult to unequivocally establish the null phenotype for lin-31 using phenotypic analysis alone.

Epistasis experiments placed the lin-31 gene at or near the end of the RTK signaling pathway specifying vulval cell fate, and sequence analysis showed it to encode a winged-helix transcription factor (MILLER et al. 1993 ). Subsequently, immunolocalization experiments revealed that LIN-31 is expressed in the VPCs during the L2 and L3 larval stages and coimmunoprecipitation experiments showed that it forms a heterodimer in vivo with the LIN-1 Ets transcription factor that is disrupted upon phosphorylation of LIN-31 by MAP kinase (TAN et al. 1998 ). These results allowed expansion of the model that LIN-31 has two functions: to promote vulval induction (in its phosphorylated, undimerized form) and to inhibit induction (in its unphosphorylated, dimerized form).

The LIN-31 protein contains a winged-helix DNA-binding domain, an acidic region, a serine-rich region, and a proline-rich region (MILLER et al. 1993 ). In this study, DNA sequence analysis of 13 lin-31 mutant alleles revealed many early nonsense mutations and several large-scale disruptions of the gene, unequivocally establishing the null phenotype. DNA sequence analysis also revealed two missense mutations in the winged-helix DNA-binding domain. Although winged-helix proteins are clearly involved in many important developmental regulatory mechanisms, few missense mutations within this DNA-binding domain have been reported, making it difficult to confirm their functions in vivo. The identification of these two missense mutations suggests that the winged-helix DNA-binding domain is critical for LIN-31 function in vivo.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

General methods and strains:
Standard techniques were used for maintenance and handling of C. elegans strains (BRENNER 1974 ). Animals were grown at 20° unless otherwise indicated. The animals described as wild type were C. elegans, variety Bristol, strain N2 (BRENNER 1974 ). Unless otherwise noted, all C. elegans mutations are described in WOOD 1988 . Mutations and genetic markers were: LGII: lin-31(e1750, n301, n376, n428, n429, n435, n762, n1048, n1049, n1050, n1053) (FERGUSON and HORVITZ 1985 ), lin-31(n1282, n1290, n1291) (KIM and HORVITZ 1990 ), lin-31(ga9, ga10, ga37) (MILLER et al. 1993 ), lin-31(ga57, ga70) (gifts from D. Eisenmann); LGIII: unc-32(e189); LGV: him-5(e1490).

Male mating efficiency:
To test for male mating efficiency, double mutant strains containing a lin-31 allele and him-5(e1490) were constructed. Each lin-31 mutant allele was tested in a cross consisting of 12 lin-31; him-5 L4 males and 6 unc-32 hermaphrodites. Successful mating was indicated by the presence of non-Unc cross-progeny. Each allele was tested in at least two crosses. Male mating efficiency (ME) ratings were assigned according to WOOD 1988 , with ME0 indicating no male mating and ME3 indicating wild-type male mating levels.

L2 division phenotype:
The presence of L2 divisions in the vulval precursor cell lineage was monitored with Nomarski optics in at least nine animals from each lin-31 strain scored. Animals in the L2 larval stage were identified on the basis of size using a dissecting microscope. These animals were mounted for Nomarski observation as described in WOOD 1988 , except that 5 mM NaAzide was used to immobilize the larvae. Larval stage was confirmed by gonad size, and VPCs were spot-checked for inappropriate L2 divisions. To prevent potential confusion between inappropriate L2 VPC divisions and normal early L3 VPC divisions, only animals whose gonad arms had not yet extended past P5.p and P7.p were scored. Although this approach ensures that only animals in the L2 stage are scored, it necessarily underestimates the number of possible inappropriate L2 VPC divisions since the entire L2 stage is not scored in each animal. Mutant strains were scored as positive for the L2 division phenotype if at least one animal displayed L2 divisions.

Cloning of lin-31 alleles:
Genomic DNA for 13 lin-31 alleles was cloned by a PCR method adapted from WILLIAMS et al. 1992 . For most reactions, the PCR primers used were LM34 (5'-GGGAATTCATGAACAGGAAAATGTGTGGC-3'), and LM35 (5'-CCGAATTCGCAAAACTCTATACATATTCCC-3'). Vent DNA polymerase (0.18 units) was used in each reaction. The LM34 and LM35 primers bracket the 3901-bp genomic coding region of lin-31 (from 27 nucleotides before the start codon to 70 nucleotides past the stop codon). For some alleles where general location of mutation was known (e.g., n1049, n376, ga10, and ga37), primers internal to LM34 and LM35 were used. Gel-purified 3.9-kb fragments from two independent PCR reactions for each mutant allele were cloned using a TOPO TA cloning kit (Invitrogen, Carlsbad, CA).

Sequencing of lin-31 alleles:
For point mutations and frameshift mutation, the entire 3.9-kb insert of each mutant allele clone was sequenced on one strand for two independent clones per mutant allele. When the same mutation was identified in both clones, the other strand was also sequenced in that region to confirm the change. Two independent clones were sequenced in the region of the n1049 deletion. For each of the other deletion or insertion alleles, one clone was sequenced in the region of the mutation, and the size and position was confirmed by Southern analysis (n376 and ga37; MILLER et al. 1993 ) or PCR (ga10). Sequencing was performed on an ABI (Foster City, CA) automated DNA sequencer.

Immunofluorescence:
Rabbit anti-LIN-31 antibodies were a generous gift from P. Tan and S. Kim (TAN et al. 1998 ). Fixed animals (FINNEY and RUVKUN 1990 ) were stained with a 1:100 dilution of anti-LIN-31 antibodies as described in MILLER and SHAKES 1995 .

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Sequence of lin-31 mutant alleles:
The lin-31 alleles used in this study are independent and were obtained by several different screening methods. e1750, n301, n376, n428, n429, n435, n762, n1048, n1049, n1050, and n1053 were isolated as multivulva animals (Muvs) in a general screen for Muv or Vul animals (FERGUSON and HORVITZ 1985 ). ga10 was isolated in a screen for Muv animals (M. HERMAN, personal communication). n1282, n1290, n1291, ga9, and ga37 were isolated as transposon-induced alleles in a mutator background (KIM and HORVITZ 1990 ). ga57 and ga70 were isolated in a screen for animals with protruding vulvae (EISENMANN and KIM 2000 ).

LIN-31 is a 237-amino-acid protein containing a DNA-binding domain, an acidic region, a serine-rich region, and a proline-rich region (see Fig 1). To elucidate possible functions of these domains, 3.9 kb of genomic DNA spanning the lin-31 coding region was sequenced in 13 lin-31 mutant alleles (see MATERIALS AND METHODS). Since Taq polymerase, which was used to clone the mutant alleles, can itself induce mutations, two independent PCR clones from each noninsertion or nondeletion allele were sequenced. The indicated mutation in each case was the only change within the 3.9-kb cloned region exhibited by both clones. The sequence analysis of 13 lin-31 alleles revealed six nonsense mutations, three deletions, one transposon insertion in the 5' end of the gene, one frameshift mutation, and two missense mutations. In addition, n1291 and n1282, previously characterized as Tc1 insertions in the 5' end of the gene, and n1290, another mutator allele (MILLER et al. 1993 ), were not sequenced. The results of these experiments are shown in Fig 1 and summarized in Table 1.

View larger version (12K): In this window In a new window Download PPT slide   Figure 1. Mutant alleles within the lin-31 genomic region. Boxes indicate exons, horizontal lines between boxes indicate introns, and open boxes indicate nontranslated regions. The type of mutation is depicted by appropriate symbol. The sizes of deletions are indicated above deletion lines. DNA sequences that encode specific LIN-31 protein domains or regions are indicated by different hatching patterns. Arrows labeled LM34 and LM35 indicate primers used to clone most lin-31 mutant alleles. The total length of this region is 4593 bp.

Significantly, all six nonsense mutations map to the DNA-binding domain, with two of the alleles (n1053 and e1750) representing identical changes. One deletion (ga10) is large, eliminating the carboxyl third of the LIN-31 protein, including all of the acidic, serine-rich, and proline-rich regions. Another deletion (n376) effectively removes (or alters) the entire carboxyl third of the protein due to the deletion of the last splice acceptor site. Finally, a small deletion (n1049) eliminates the last third of the DNA-binding domain. Three additional alleles (n1050, n428, and n429) are suspected to be large deletions of the entire coding region, as PCR clones were not obtained in several independent reactions using multiple combinations of primers. The Tc1 transposon insertion is located just before the start codon, and the frameshift mutation (ga9) mapped to the end of the DNA-binding domain. Finally, the two missense mutations (ga57 and n301) mapped within the DNA-binding domain. Note that all but the two C-terminal deletions affect the DNA-binding domain (see Fig 1), although the sequenced Tc1 insertion is upstream of the DNA-binding domain and may not affect it.

Two putative DNA-binding mutants:
DNA sequencing of lin-31 showed that the N-terminal half of the LIN-31 protein consists of a DNA-binding domain similar to that of the rat transcription factor HNF-3 (MILLER et al. 1993 ). X-ray crystallography studies of the HNF-3/DNA cocrystal structure by CLARK et al. 1993 subsequently showed that the HNF-3 DNA-binding domain has a winged-helix motif, which resembles the structure of histone H5. Its three -helices in the amino-terminal half adopt a compact structure that presents the third (recognition) helix to the major groove, which interacts with its DNA target through both direct and water-mediated major and minor groove base contacts. The carboxyl-terminal half of the motif includes antiparallel ß-structure and random coils ("wings") that interact with the minor groove. The DNA-binding domain of the LIN-31 protein displays a striking similarity to the DNA-binding domains of the fork head/HNF-3-related winged-helix family of proteins (Fig 2; MILLER et al. 1993 ; KAUFMANN and KNOCHEL 1996 ). Thus, the overall three-dimensional structures of these DNA-binding domains should also be very similar.

View larger version (79K): In this window In a new window Download PPT slide   Figure 2. Multiple sequence alignment of selected winged-helix DNA-binding domains. LIN-31 (MILLER et al. 1993 ) was compared with the founding members of the winged-helix family: fork head (FKH), cloned from Drosophila (WEIGEL et al. 1989 ), and HNF-3, HNF-3ß, and HNF-3, cloned from rats (LAI et al. 1990 , LAI et al. 1991 ). Also included in this alignment are several proteins from the same winged-helix subclass as LIN-31 (defined in KAUFMANN and KNOCHEL 1996 ): FD4 from Drosophila (HACKER et al. 1992 ), FKH-4 from mouse (KAESTNER et al. 1993 ), and XFD-5 from Xenopus (LEF et al. 1996 ), and several winged-helix DNA-binding domains from C. elegans: PHA-4 (AZZARIA et al. 1996 ), PES-1(HOPE 1994 ), and DAF16 (LINET AL. 1997 ; OGG et al. 1997 ). Lines indicate positions of -helices (H1–H3), boldface lines indicate positions of ß-pleated sheets (S1–S3), and wing-like loop regions are indicated by W1 and W2 (CLARK et al. 1993 ). H3 is the DNA recognition helix. Asterisks indicate positions of the two lin-31 missense alleles, n301 and ga57, respectively. Shading indicates sequence identity to LIN-31. DNA-binding domain percentage identity to LIN-31 is indicated at the beginning of each alignment.

From our sequencing studies, we identified two missense alleles of the lin-31 gene, which are located in the conserved winged-helix DNA-binding domain (Fig 2 and Table 1). One mutant allele, ga57, results in the substitution of the amino acid isoleucine for the highly conserved asparagine at the end of the recognition helix (helix 3) within the DNA-binding domain. The HNF-3/DNA cocrystal structure revealed that this particular asparagine makes contact with its DNA target via hydrogen bonds with water molecules (CLARK et al. 1993 ). Changing it to a hydrophobic isoleucine presumably disrupts this interaction. The importance of this particular asparagine side chain is underscored by its conservation in all but a few members of known winged-helix proteins (KAUFMANN and KNOCHEL 1996 ). The other allele, n301, also lies within the DNA-binding domain (in the middle of helix 3) and results in the substitution of a glutamine for an extremely highly conserved arginine. Although the X-ray crystallography studies of HNF-3 did not indicate that this arginine is specifically involved in binding DNA, this substitution may still affect how the mutant protein interacts with the DNA. This particular arginine is very likely to play an especially critical role in the function of these proteins, as it is completely conserved in all known winged-helix proteins (KAUFMANN and KNOCHEL 1996 ).

It is important to determine whether these two mutant proteins are defective because they are no longer able to bind their DNA target or because the amino acid substitutions have destabilized the protein enough to cause unfolding and subsequent degradation. Using immunolocalization studies, we were able to rule out the second possibility by showing that the LIN-31 protein is present in the proper cells at the proper stage of development. Fig 3 shows wild-type and lin-31(ga57) animals stained with antibodies to LIN-31 during the late L2 or early L3 stage, when LIN-31 protein is known to be present in the nucleus of each vulval precursor cell (TAN et al. 1998 ). In these animals, the LIN-31 protein is observed in the nuclei of the VPCs. The additional staining cells in the ga57 animal represent either extra VPCs that are often generated during the L2 larval stage in lin-31 mutant animals or normal VPC divisions in an early L3 animal. lin-31(n301) animals displayed similar results (data not shown). Thus, both ga57 and n301 mutant strains express LIN-31 protein at the proper time and place, suggesting that their defects lie in their inability to bind the target DNA, either by disruption of specific protein:DNA interactions or through a folding defect that prevents binding but does not result in significant degradation of the protein.

View larger version (67K): In this window In a new window Download PPT slide   Figure 3. LIN-31 expression patterns in wild-type (N2) and ga57 animals. (A) Wild-type (N2) and (B) lin-31(ga57) mutant animals stained with anti-LIN-31 antibodies during the late L2 or early L3 larval stage. Figure shows midbody region. Dorsal is to the left, ventral to the right, anterior toward the bottom, and posterior toward the top. Figures are negative images of FITC immunofluorescence. Both animals show staining of LIN-31 protein within the nuclei of the VPCs.

Phenotype of lin-31 alleles:
Genetic criteria for establishing a null phenotype include ease of mutant isolation, similarity among strong alleles, phenotype in trans to a deficiency, and the ability to be suppressed by a nonsense suppressor. Previous studies showed allelic similarity for the vulval cell fate deregulation phenotype and also showed that a strong allele acts like a deficiency in trans to another allele (MILLER et al. 1993 ). To determine if lin-31 alleles display allelic similarity for all known lin-31 mutant phenotypes, we analyzed three phenotypic characteristics: the deregulation of VPC fates, male mating efficiency, and inappropriate cell division of VPCs during the L2 larval stage.

In lin-31 deregulated mutants, VPCs that normally choose induced cell fates sometimes choose uninduced fates and VPCs that normally choose uninduced cell fates sometimes choose induced cell fates (MILLER et al. 1993 ). These mistakes in cell fate determination are manifested at the dissecting microscope level as Vul or Muv phenotypes, respectively. Because each vulval precursor cell in a lin-31 mutant animal appears to make an independent decision, these animals can also sometimes display the Vul and Muv phenotypes simultaneously. All 19 lin-31 mutant strains produce a high percentage of non-wild-type animals (Muv, Vul, or Muv/Vul), except for n1291, which expressed a somewhat weaker deregulation phenotype (see Table 2). In addition, each mutant strain also exibited both Muv and Vul phenotypes (data not shown), indicating that each allele is defective for appropriate adoption of both induced and uninduced cell fates.

In addition to the vulval precursor cell deregulation phenotype in hermaphrodites, lin-31 mutant males are also mating incompetent (FERGUSON and HORVITZ 1985 ; MILLER et al. 1993 ), probably due to formation of defective mating spicules (S. BAIRD, personal communication). The 19 lin-31 mutant strains were tested for the efficiency of male mating and all 19 were found to be completely mating defective, with a mating efficiency rating of "0" (ME0; Table 2).

A third phenotype that is observed in lin-31 mutant animals is the occasional inappropriate cell division of a VPC during the L2 larval stage to give rise to either an extra VPC or an extra Pn.a-like neuroblast cell (MILLER et al. 1993 ). VPCs in wild-type animals are quiescent during the L2 larval stage. When cell division of a VPC during the L2 results in the creation of an extra VPC in lin-31 mutant animals, its fate is also deregulated. At least one representative allele from each class of mutation (missense, nonsense, deletion, insertion, and frameshift) was analyzed for this phenotype. L2 divisions of VPCs were observed in all lin-31 mutant strains tested (Table 2).

DNA sequence analysis supports the conclusion that at least some of the lin-31 mutant alleles represent the null phenotype. Since all 13 of these sequenced mutant alleles, as well as two previously mapped Tc1 insertions and the three deletions that likely delete the entire coding region, show similar mutant phenotypes, this suggests that LIN-31 protein function is similarly defective in 18 of the 19 mutant strains. The n1291 strain was the only one to show a weaker phenotype in any of the phenotypic analyses. Its 54% vulval cell fate deregulation phenotype fell significantly below the next highest allele of 80% (see Table 2), indicating that it may represent a somewhat weaker allele. It was, however, indistinguishable from other alleles in the remaining two phenotypic analyses (male mating and inappropriate VPC division during the L2 larval stage). Earlier experiments characterized n1291 as a Tc1 insertion in the 5' end of the gene, resulting in a smaller mRNA (1.4 vs. 1.5 kb observed in wild-type animals; MILLER et al. 1993 ). Thus, with the possible exception of n1291, all existing lin-31 alleles display the null phenotype and are likely to be null mutations.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this article, we report the molecular and phenotypic characterization of 19 mutant alleles of the gene that encodes the winged-helix transcription factor LIN-31. All 19 alleles displayed similar phenotypes, and DNA sequencing results indicated that most of these alleles must be null alleles. In addition, we show that two missense alleles in the DNA-binding domain are likely to disrupt the interaction of the transcription factor with its DNA target, revealing the critical importance of the DNA-binding domain to the function of the LIN-31 protein.

The null phenotype of lin-31:
Due to the nature of the lin-31 mutant phenotype (presence of both Muv and Vul animals and less than 100% penetrance for deregulation of VPCs), it has been difficult to unequivocally establish the null phenotype by phenotypic characteristics alone. In this article, allelic similarity was demonstrated for three phenotypic characteristics, whereas previous analyses had relied on only one phenotype. In addition, six early nonsense mutations, four insertions in the 5' end of the gene, one early frameshift mutation, and at least three deletions were identified. Furthermore, P. TAN (personal communication) has shown, using antibody studies, that there is no detectable LIN-31 protein for at least two alleles, n1053 and ga37.

With the establishment of the null phenotype, the transformation of both induced and uninduced vulval cell fates in lin-31 null mutants indicates that LIN-31 plays an important role in the specification of both types of cell fates. This is consistent with models proposing that LIN-31 functions both to promote and to inhibit vulval induction (MILLER et al. 1993 ; TAN et al. 1998 ). Thus, a cell lacking all lin-31 activity would be incapable of correctly specifying either an induced or uninduced cell fate. This would require the cell to adopt a VPC fate in a stochastic manner, resulting in the deregulated pattern of cell fate specification observed in lin-31 mutants. The adoption of a specific induced or uninduced cell fate in the absence of LIN-31 rather than a default or intermediate cell fate can be explained by a model in which target genes of the LIN-31 transcription factor are autoregulated, and that stochastic perturbances can cause one of the target proteins to rise above a threshold level, triggering autoactivation and subsequent deactivation of other target genes, thus "locking in" a particular cell fate (MILLER et al. 1993 ; TAN et al. 1998 ).

A winged-helix DNA-binding domain is essential for function:
While the winged-helix family of DNA-binding proteins is clearly an important one in developmental biology, most of its eukaryotic members have been identified simply by sequence homology, and very few missense mutations within the DNA-binding domain have been identified through genetic screens. Thus, definitive proof of this domain's in vivo importance in eukaryotes is limited. For example, site-directed mutation studies in HNF3 have identified specific residues in the DNA-binding domain important for its stability (STEVENS et al. 2000 ), but those studies were limited to transcription activation assays in tissue culture and in vitro binding experiments. Other studies have shown that targeted null mutations in HNF3 and HNF3ß are lethal in vivo, but those experiments deleted almost the entire protein-coding region (ANG and ROSSANT 1994 ; WEINSTEIN et al. 1994 ) or at least the entire DNA-binding domain (KAESTNER et al. 1999 ). Furthermore, some transcription factors, such as fushi tarazu, are capable of regulating transcription without a DNA-binding domain (FITZPATRICK et al. 1992 ). In this study, we show that the winged-helix domain is indispensable to the function of the LIN-31 protein.

One indication that the LIN-31 DNA-binding domain is critical for function is the number of mutations that lie within it (Fig 1). Eleven of the 16 alleles with known positions are located within the DNA-binding domain, including two missense mutations and a small deletion. The most compelling argument for the importance of the DNA-binding domain is the existence of the two missense mutations within this domain that leave the rest of the protein intact. Since these mutations lie in or near the recognition helix and since LIN-31 protein is not degraded due to unfolding caused by the amino acid substitutions in these two mutant strains, it is likely that the only defect these mutant proteins display is an inability to bind the correct target DNA. Furthermore, at least one of these missense alleles represents a change in an amino acid that, in a similar winged-helix homolog, is known to make contact with the target DNA. The null phenotype of this allele strongly suggests that the ability to bind target DNA is a critical function of the LIN-31 protein. Finally, since the phenotypes of all lin-31 alleles affecting the DNA-binding domain are equivalent to the lin-31 null phenotype, it is clear that this winged-helix domain is essential for protein function. The importance of this domain in LIN-31 is consistent with a study showing that a missense mutation in helix 1 of the DNA-binding domain of a human winged-helix protein (TTF-2) results in impaired DNA binding and loss of transcriptional function (CLIFTON-BLIGH et al. 1998 ).

Standard genetic screens yield only null alleles:
All six existing lin-31 nonsense mutations map to the DNA-binding domain, located in the N-terminal half of the protein. Interestingly, no late nonsense mutations were observed, indicating that a LIN-31 protein truncated late in the protein may not result in a phenotype detectable by standard Muv and Vul screens. While there are more codons in the N-terminal half of the LIN-31 protein that are capable of being mutated to a stop codon in a single step (37 vs. 27 in the C-terminal half), this difference does not adequately explain the complete absence of late nonsense mutations. The most likely explanation for the lack of nonsense mutations in the latter part of the protein is that a mutation leaving more than half of LIN-31 intact results in a protein with some residual function. These weak alleles would not have been detected in a standard screen for Muvs or Vuls, since the penetrance for either phenotype would probably be well below the 31–76% Muv penetrance or the 15–41% Vul penetrance (L. MILLER, unpublished results) exhibited by null alleles. Thus, the isolation of weak or other nonnull lin-31 alleles awaits the development of nonstandard mutant screening procedures or in vitro mutagenesis experiments. Such studies (in progress) may reveal the functional domains responsible for the two different activities of LIN-31 postulated by the model discussed above.

	FOOTNOTES

¹ Present address: Yale University School of Medicine, 333 Cedar St., SHM CE30, P.O. Box 208024, New Haven, CT 06520-8024.

	ACKNOWLEDGMENTS

We thank L. Corbett and D. Schatz for help with DNA sequencing; P. Tan and S. Kim for the generous gift of LIN-31 antibodies; S. Kim, D. Eisenmann, and M. Herman for lin-31 alleles; K. Warren for identification of the ga70 deletion; J. Holzman, K. Warren, R. Childers, L. Treat, and K. Donnelly for help with male mating experiments; members of the Kim lab for antibody advice; and P. Schatz, D. Preuss, D. Eisenmann, P. Tan, J. Grainger, C. Stephens, and B. Meyer for comments on the manuscript. This work was supported by National Institutes of Health (NIH) research grant 1 R15 GM52636-01, National Science Foundation CAREER grant IBN-9600864, and the Clare Boothe Luce Foundation. Some nematode strains were supplied by the Caenorhabditis Genetics Center, which is supported by a contract between the NIH National Center for Research Resources and the University of Minnesota.

Manuscript received November 16, 1999; Accepted for publication August 3, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ANG, S.-L. and J. ROSSANT, 1994  HNF-3ß is essential for node and notochord formation in mouse development. Cell 78:561-574[Medline].

AZZARIA, M., B. GOSZCZYNSKI, M. A. CHUNG, J. M. KALB, and J. D. MCGHEE, 1996  A fork head/HNF-3 homolog expressed in the pharynx and intestine of the Caenorhabditis elegans embryo. Dev. Biol. 178:289-303[Medline].

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

CLARK, K. L., E. D. HALAY, E. LAI, and S. K. BURLEY, 1993  Co-crystal structure of the HNF-3/fork head DNA-recognition motif resembles histone H5. Nature 364:412-420[Medline].

CLIFTON-BLIGH, R. J., J. M. WENTWORTH, P. HEINZ, M. S. CRISP, and R. JOHN et al., 1998  Mutation of the gene encoding human TTF-2 associated with thyroid agenesis, cleft palate and choanal atresia. Nat. Genet. 4:399-401.

EISENMANN, D. M. and S. K. KIM, 2000  Protruding vulva mutants identify novel loci and Wnt signaling factors that function during C. elegans vulva development. Genetics 156:1097-1116[Abstract/Free Full Text].

FERGUSON, E. L. and H. R. HORVITZ, 1985  Identification and characterization of 22 genes that affect the vulval cell lineages of the nematode Caenorhabditis elegans.. Genetics 110:17-72[Abstract/Free Full Text].

FINNEY, M. and G. RUVKUN, 1990  The unc-86 gene product couples cell lineage and cell identity in C. elegans.. Cell 63:895-905[Medline].

FITZPATRICK, V. D., A. PERCIVAL-SMITH, C. J. INGLES, and H. M. KRAUSE, 1992  Homeodomain-independent activity of the fushi tarazu polypeptide in Drosophila embryos. Nature 356:610-612[Medline].

GREENWALD, I., 1997 Development of the vulva, pp. 519–541 in C. elegans II, edited by D. L. RIDDLE, T. BLUMENTHAL, B. J. MEYER and J. R. PRIESS. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

HÄCKER, U., U. GROSSNIKLAUS, W. J. GEHRING, and H. JÄCKLE, 1992  Developmentally regulated Drosophila gene family encoding the fork head domain. Proc. Natl. Acad. Sci. USA 89:8754-8758[Abstract/Free Full Text].

HOPE, I. A., 1994  PES-1 is expressed during early embryogenesis in Caenorhabditis elegans and has homology to the fork head family of transcription factors. Development 120:505-514[Abstract].

KAESTNER, K. H., K. H. LEE, J. SCHLÖNDORFF, H. HIEMISCH, and A. P. MONAGHAN et al., 1993  Six members of the mouse forkhead gene family are developmentally regulated. Proc. Natl. Acad. Sci. USA 90:7628-7631[Abstract/Free Full Text].

KAESTNER, K. H., J. KATZ, Y. LIU, D. J. DRUCKER, and G. SCHÜTZ, 1999  Inactivation of the winged helix transcription factor HNF3 affects glucose homeostasis and islet glucagon gene expression in vivo. Genes Dev. 13:495-504[Abstract/Free Full Text].

KAUFMANN, E. and W. KNÖCHEL, 1996  Five years on the wings of fork head. Mech. Dev. 57:3-20[Medline].

KIM, S. K. and H. R. HORVITZ, 1990  The Caenorhabditis elegans gene lin-10 is broadly expressed while required specifically for the determination of vulval cell fates. Genes Dev. 4:357-371[Abstract/Free Full Text].

KORNFELD, K., 1997  Vulval development in Caenorhabditis elegans.. Trends Genet. 13:55-61[Medline].

LAI, E., V. R. PREZIOSO, E. SMITH, O. LITVIN, and R. H. COSTA et al., 1990  HNF-3A, a hepatocyte-enriched transcription factor of novel structure is regulated transcriptionally. Genes Dev. 4:1427-1436[Abstract/Free Full Text].

LAI, E., V. R. PREZIOSO, W. F. TAO, W. S. CHEN, and J. E. DARNELL, JR., 1991  Hepatocyte nuclear factor 3 alpha belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head. Genes Dev. 5:416-427[Abstract/Free Full Text].

LEF, J., P. DEGE, M. SCHEUCHER, V. FORSBACH-BIRK, and J. H. CLEMENT et al., 1996  A fork head related multigene family is transcribed in Xenopus laevis embryos. Int. J. Dev. Biol. 40:245-253[Medline].

LIN, K., J. B. DORMAN, A. RODAN, and C. KENYON, 1997  daf-16: an HNF-3/forkhead family member that can function to double the life-span of Caenorhabditis elegans. Science 278:1319-1322. [see comments][Abstract/Free Full Text].

MILLER, D. M., and D. C. SHAKES, 1995 Immunofluorescence microscopy, pp. 365–394 in Methods in Cell Biology, Volume 48: Modern Biological Analysis of an Organism, edited by H. F. EPSTEIN and D. C. SHAKES. Academic Press, San Diego.

MILLER, L. M., M. E. GALLEGOS, B. A. MORISSEAU, and S. K. KIM, 1993  lin-31, a Caenorhabditis elegans HNF-3/fork head transcription factor homolog, specifies three alternative cell fates in vulval development. Genes Dev. 7:933-947[Abstract/Free Full Text].

MILLER, L. M., D. A. WARING, and S. K. KIM, 1996  Mosaic analysis using a ncl-1(+) extrachromosomal array reveals that lin-31 acts in the Pn.p cells during Caenorhabditis elegans vulval development. Genetics 143:1181-1191[Abstract].

OGG, S., S. PARADIS, S. GOTTLIEB, G. I. PATTERSON, and L. LEE et al., 1997  The Fork head transcription factor DAF-16 transduces insulin-like metabolic and longevity signals in C. elegans.. Nature 389:994-999[Medline].

STEVENS, K., L. CIRILLO, and K. S. ZARET, 2000  Creating temperature-sensitive winged helix transcription factors: amino acids that stabilize the DNA binding domain of HNF3. J. Biol. Chem. 275:30471-30477[Abstract/Free Full Text].

TAN, P. B., M. R. LACKNER, and S. K. KIM, 1998  MAP kinase signaling specificity mediated by the LIN-1 Ets/LIN-31 WH transcription factor complex during C. elegans vulval induction. Cell 93:569-580[Medline].

WEIGEL, D., G. JÜRGENS, F. KUTTNER, E. SEIFERT, and H. JÄCKLE, 1989  The homeotic gene fork head encodes a nuclear protein and is expressed in the terminal regions of the Drosophila embryo. Cell 57:645-658[Medline].

WEINSTEIN, D. C., A. R. ALTABA, W. S. CHEN, P. HOODLESS, and V. R. PREZIOSO et al., 1994  The winged-helix transcription factor HNF-3ß is required for notochord development in the mouse embryo. Cell 78:575-588[Medline].

WILLIAMS, B. D., B. SCHRANK, C. HUYNH, R. SHOWNKEEN, and R. H. WATERSTON, 1992  A genetic mapping system in Caenorhabditis elegans based on polymorphic sequence-tagged sites. Genetics 131:609-624[Abstract].

WOOD, W. B., 1988 The Nematode Caenorhabditis elegans. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

This article has been cited by other articles:

				E. C. Andersen, X. Lu, and H. R. Horvitz C. elegans ISWI and NURF301 antagonize an Rb-like pathway in the determination of multiple cell fates Development, July 15, 2006; 133(14): 2695 - 2704. [Abstract] [Full Text] [PDF]

				J. L. Goldstein, D. Glossip, S. Nayak, and K. Kornfeld The CRAL/TRIO and GOLD Domain Protein CGR-1 Promotes Induction of Vulval Cell Fates in Caenorhabditis elegans and Interacts Genetically With the Ras Signaling Pathway Genetics, February 1, 2006; 172(2): 929 - 942. [Abstract] [Full Text] [PDF]

				V. Hsu, C. L. Zobel, E. J. Lambie, T. Schedl, and K. Kornfeld Caenorhabditis elegans lin-45 raf Is Essential for Larval Viability, Fertility and the Induction of Vulval Cell Fates Genetics, February 1, 2002; 160(2): 481 - 492. [Abstract] [Full Text] [PDF]