Dominant negative mutations of caenorhabditis elegans daf-7 confer a novel developmental phenotype

Authors

  • Matt Crook,

    Corresponding author
    1. Department of Biochemistry and Molecular Biology, The Pennsylvania State University, University Park, Pennsylvania
    • Correspondence to: Matt Crook, Department of Biochemistry and Molecular Biology, The Pennsylvania State University, University Park, PA. E-mail: mxc83@psu.edu

    Search for more papers by this author
  • Warwick N. Grant

    1. Department of Genetics, La Trobe University, Melbourne, Victoria, Australia
    Search for more papers by this author

Abstract

Background: TGF-β signaling pathways are involved in the control of development in every member of the animal kingdom. As such, TGF-β ligands are widely divergent yet retain a set of core conserved features, specifically, a pre-protein cleavage site and several conserved ligand domain residues, the disruption of which produces a dominant negative phenotype. Results: We have extended these observations into an invertebrate system by creating a series of loss-of-function Caenorhabditis elegans daf-7 transgenes. When we tested these mutant transgenes in a daf-7/+ background, we saw a molting and excretory canal phenotype. Members of both pathways downstream of daf-4 were required for this phenotype. Conclusions: Our results show that the basic mechanisms of TGF-β function are conserved across the animal kingdom. A subset of our daf-7 mutations also produced an unexpected and novel phenotype. Epistasis experiments demonstrated that both daf-3/-5 and sma-4/-9 were downstream of our mutant daf-7 transgenes, which suggests not only a role for DAF-7 in the control of molting and the development of the excretory system but also that daf-7 and dbl-1 signaling may converge downstream of their shared Type II receptor, daf-4. Our approach may unveil new roles in development for other invertebrate TGF-β ligands. Developmental Dynamics 242:654–664, 2013. © 2013 Wiley Periodicals, Inc.

INTRODUCTION

Dauer larvae are developmentally arrested, long-lived third-stage larvae that form in response to low food availability, high con-specific population density measured as a constitutively produced pheromone, and high temperature (Riddle and Albert, 1997). DAF-7 is the ligand in the nematode TGF-β pathway that, together with the insulin-like and guanylyl cyclase pathways, controls entry into dauer development in Caenorhabditis elegans (Ren et al., 1996; Schackwitz et al., 1996; Tissenbaum and Ruvkun, 1998; Birnby et al., 2000). DAF-7, secreted by the ASI neurons in the L1 and L2 stages, suppresses entry into dauer development in the absence of dauer pheromone and at high temperature (Ren et al., 1996; Schackwitz et al., 1996; Crook et al., 2010). As such, daf-7(e1372) animals are dauer formation Constitutive (dafC) at 25°C (Swanson and Riddle, 1981).

One striking observation about DAF-7 and other TGF-β is that their broad protein primary structure is highly conserved, along with key ligand domain residues, yet sequence similarity is low, even between closely related species (Fig. 1). TGF-β ligands are secreted and form homodimer precursor polypeptides that form non-covalent bonds with the Latent TGF-β Binding Protein (LTBP) (Munger et al., 1997). Upon activation, the homodimer is cleaved at a tri-basic cleavage site by a furin-class protease (Dubois et al., 2001), releasing the active ligand (McPherron et al., 1997; Zhu et al., 2000). TGF-β dimerization is facilitated by an intermolecular disulphide bond between one of the nine conserved ligand domain cysteine residues; the remaining eight are responsible for correct protein conformation (Daopin et al., 1992). In addition, several non-cysteine ligand domain residues are conserved in all TGF-β proteins.

Naturally occurring and engineered mutations have been described in these conserved regions, which fall into two groups: single ligand domain amino acid substitutions and cleavage sequence mutations. Several types of ligand domain dominant negative substitution mutations in TGF-β have been described, including a GDF5W408R substitution (Masuya et al., 2007), a CDMP1/GDF5C400Y substitution responsible for Chondrodysplasia Grebe Type disease in humans (Thomas et al., 1997), and a phenylalanine to isoleucine substitution in the active ligand domain of Myostatin/GDF-8 in Xenopus oocytes (Wittbrodt and Rosa, 1994). In mice, Xenopus oocytes and human tissue culture, cleavage mutants produced dominant negative phenotypes by interfering with ligand secretion or through the formation of non-cleavable heterodimers (Lopez et al., 1992; Wittbrodt and Rosa, 1994; Hawley et al., 1995; Zhu et al., 2000).

These TGF-β mutations all conferred a dominant negative failure of developmental control. Given the striking structural conservation of TGF-β ligands among the animal kingdom, we hypothesised that these structural mutations would similarly interfere with DAF-7 signaling in nematodes, conferring a dominant negative dafC phenotype. To test this hypothesis, daf-7 transgenes containing both classes of TGF-β dominant negative mutations were made, a single point mutation resulting in the substitution of a conserved residue in the ligand domain (daf-7W264R) and a mutated cleavage site (daf-7clv mt), which should result in an unprocessed and inactive mutant:wild type heterodimer. In addition, although the literature indicates that full or partial ligand domain deletions generally result in a null recessive phenotype (Thomas et al., 1996; Grobet et al., 1997; Kambadur et al., 1997; McPherron et al., 1997; McPherron and Lee, 1997), a series of ligand domain truncations was also tested. As we expected, none of the mutations we tested resulted in a functional DAF-7 protein. However, when the daf-7T2, daf-7clv mt, and daf-7W264R transgenes were assayed in a daf-7/+ background, we observed a novel daf-7 phenotype. We will discuss the impact of our findings on the functional conservation of TGF-β ligands and a previously uncharacterized role for daf-7 in larval development.

Figure 1.

Amino acid alignment of TGF-β cleavage sites and ligand domains. Dashes represent gaps introduced to allow optimal alignment of the sequences used. Black shaded amino acid residues are conserved across all the aligned sequences. Asterisks highlight conserved cysteine residues. Arrowhead highlights the conserved W264 amino acid residue. Accession numbers are as follows: C. elegans DAF-7, AAC47389; C. briggsae DAF-7, CBG24910; C. elegans DBL-1, CCD74236; C. elegans UNC-29, CAB61007; C. elegans TIG-2, CCD65248; D. melanogaster DPP, AAC47552; M. musculus GDF5(CDMP1), AAA18778; H. sapiens Myostatin(GDF-8), ABI48514; H. sapiens TGF-β1, P01137.

Figure 2.

Effect of daf-7 transgenes on dauer formation. Percentage of gfp-positive dauers at 25°C, for (A) constructs in a daf-7(e1372) background and (B) constructs in a wild-type (WT) background. For WT and daf-7, percentage of dauers is shown. Error bars are s.e.m.

RESULTS

DAF-7 Ligand Domain Truncations and Putative Dominant Negative Mutations Are Non-Functional

We wanted to see if any of the daf-7 truncations or mutations we made were functional for two reasons: first, to expand our knowledge of which parts of DAF-7 are required for function and, second, to ensure that our manipulations were truly loss of function and thus able to act as dominant negative mutations. We found that our daf-7FLcontrol was able to completely rescue the daf-7 dafC phenotype, yet our putative dominant negative constructs were unable to rescue and are, therefore, complete loss-of-function mutations (Fig. 2A). Given the complete conservation of the ligand domain cysteine residues and their role in ligand domain folding and dimerization (Daopin et al., 1992), it was of no surprise that removing even just 31 C-terminal amino acids, containing two of the conserved cysteines, rendered DAF-7 non-functional. However, this is the first time that both correct pro-domain cleavage and a conserved non-cysteine ligand domain residue have been shown to be required for DAF-7 function.

Figure 3.

daf-7 transgene expression in lines containing daf-7W264R, daf-7clv mt, and daf-7FL arrays. MW is NEB 100-bp ladder. Lanes 1–3 and 7, daf-7W264R arrays in wild-type and daf-7 backgrounds, respectively. Lanes 4–6 and 8, daf-7clv mt arrays in wild-type and daf-7 backgrounds, respectively. Lane 9, a daf-7FL array in a daf-7 background. Lanes 10–12, N2 and CB1372 cDNA, and N2 genomic DNA, respectively. Lanes 13–15, daf-7W264R, daf-7clv mt, and daf-7FL plasmid DNA, respectively. –RT controls not shown.

DAF-7 Loss-of-Function Mutations Do Not Display a Dominant Negative Feedback in a Wild-Type Background

After we demonstrated which DAF-7 residues were required for function, and thus had the potential to produce a dominant negative dafC phenotype, we tested a subset for their effect on dauer formation in a wild-type background. We predicted that our mutant DAF-7 proteins would bind but fail to activate the DAF-7 receptor (DAF-7T1 and DAF-7T2) or sequester wild-type DAF-7 into either inactive (DAF-7W264R) or non-cleavable (DAF-7clv mt) heterodimers, which would result in reduced DAF-7 signaling and cause inappropriate dauer formation.

Despite multiple repeats, we did not detect the formation of any GFP-positive dauers (Fig. 2B). Therefore, although our daf-7 truncations and mutations are clearly loss-of-function and the literature suggests that the DAF-7W264R and DAF-7clv mt proteins should interfere with DAF-7 signaling, they do not produce a dominant negative phenotype in a wild-type background. This suggests three possible hypotheses: (1) the transgenes are not being expressed, (2) the mutated DAF-7 proteins are unable to form heterodimers with wild-type DAF-7, or (3) there is too much wild-type DAF-7 protein for the mutant DAF-7 to effectively sequester and thus reduce DAF-7 signaling sufficiently to produce a phenotype.

We first tested for transgene expression from daf-7FL, daf-7clv mt, and daf-7W264R arrays in wild-type and daf-7(e1372) backgrounds and detected expression from each array (Fig. 3). Therefore, the first of our hypotheses is not supported. Therefore we sought to test the function of our arrays in a daf-7/+ background.

Figure 4.

Expression of daf-7 dominant negative transgenes in a daf-7/+ background causes larval arrest. These larvae are the progeny of a cross between wild-type males carrying either daf-7FL, daf-7T1, daf-7T2, daf-7W264R, or daf-7clv mt arrays and daf-7(e1372) hermaphrodites at either 15°C or 25°C. Arrested GFP-positive larvae must therefore all be daf-7/+ as they have inherited the array from the transgenic wild-type males. n, number of GFP-positive cross progeny scored. Error bars are s.e.m. and data were analysed using Fisher's Exact test. n/s is non-significant. *P < 0.05, **P < 0.01, and ***P < 0.001, respectively.

Reducing daf-7 Gene Dosage Reveals a Novel Dominant Negative Phenotype

To test the hypothesis that the inability of the truncated DAF-7, DAF-7W264R, or DAF-7clv mt proteins to induce a dominant negative phenotype was due to wild-type DAF-7 protein being present in excess of their ability to sequester it into inactive heterodimers, we assayed arrays carrying these mutations in a daf-7/+ background. To do so, we crossed wild-type males carrying our daf-7T1, daf-7T2, daf-7W264R, or daf-7clv mt arrays into a daf-7(e1372) background. We predicted that halving the levels of wild-type DAF-7 would allow our mutant DAF-7 proteins to effectively sequester the wild-type DAF-7 and reduce DAF-7 signaling to the point where inappropriate dauer formation occurs.

When we examined our cross progeny for the presence of dauers, we did not see any GFP-positive dauer larvae for any array (data not shown). However, we occasionally observed odd-looking GFP-positive larvae on these plates at the beginning of the experiment. To pursue this further, we repeated the cross and examined early stage larvae for developmental defects. To our surprise, we found that a significant proportion of the GFP-positive cross progeny at 25°C carrying the daf-7T2, daf-7W264R, or daf-7clv mt arrays were developmentally arrested (Fig. 4). Transfer of arrested transgenic larvae to 15°C did not result in a resumption of development (data not shown). No arrested cross progeny containing the daf-7FL array were seen, which suggests that the larval arrest phenotype is the result of the action of the daf-7 dominant negative transgenes and not overexpression of daf-7. Similarly, no arrested non-transgenic cross progeny were seen when cross plates were examined at the same time point as those containing the transferred GFP-positive cross progeny.

When we examined the arrested larvae by DIC, we found a molting defect, with constrictions caused by unshed cuticle present at variable locations along the worm, and defects in their excretory canals, including frequent course reversals (Fig. 5). The percentage of arrested larvae showing molting and excretory canal defects is shown in Table 1. One concern with using extrachromosomal arrays is that mosaicism will result in the array being lost from particular cell lineages, which is of particular importance given that daf-7 is expressed in only one cell type, the ASI neuron pair (Ren et al., 1996; Schackwitz et al., 1996; Crook et al., 2010). Therefore, in arrested larvae where head neurons could be accurately identified by DIC, we also scored each larva for GFP expression in one or both ASI cell bodies. We found GFP expression in at least one ASI neuron in every arrested larvae examined (see Fig. 6 for an example), thus we can be confident that the daf-7 transgene is being expressed in these larvae.

Figure 5.

A novel molting and excretory canal defect seen in a daf-7/+ larva carrying the daf-7T2 array at 25°C. DIC micrographs of (A) a larva showing multiple lateral constrictions and unshed cuticle (×40 magnification) and (B–D) close-ups of the same larva showing molting and excretory canal defects in greater detail, in particular the excretory canal course reversions that can be seen in each micrograph (×100 magnification). Molting and excretory canal defects are highlighted by black and white arrowheads, respectively. Images were taken using a Zeiss Axioplan 2 microscope and oil immersion DIC optics. Scale bars = 20 μm.

Figure 6.

GFP expression from the rps-5::gfp co-transformation marker in an ASI neuron, the sole site of daf-7 expression, in an arrested daf-7/+; Ex daf-7T2 larvae at 25°C. A: DIC micrograph. B: GFP image with key amphidial neurons highlighted. Images were taken using a Zeiss Axioplan 2 microscope and ×100 oil immersion DIC optic. Scale bars = 20 μm.

Figure 7.

The role of downstream daf-7 and dbl-1 pathway effectors on larval arrest in a daf-7/+ background. DAF-3 and DAF-5 are downstream of and antagonized by DAF-7 signaling; daf-3 and daf-5 mutants are dauer defective. SMA-4 and SMA-9 are their equivalents in the dbl-1 pathway which controls body size. Percentage larval arrest in (A) daf-7/+ cross progeny homozygous for daf-3 or daf-5 and (B) daf-7/+ cross progeny fed sma-4 or sma-9 RNAi. One daf-7W264R array (from WG496) and two daf-7clv mt arrays (from WG508 and WG509) were tested in each experiment. n, number of GFP-positive cross progeny scored. Error bars are s.e.m. and data were analysed using Fisher's Exact test. n/s is non-significant. *P < 0.05, **P < 0.01, and ***P < 0.001, respectively.

Table 1. Percentage of Arrested Larvae Showing Molting And/ Or Excretory Duct Defectsa
ArrayMolting defectExcretory duct defect
  1. a

    Number of larvae presenting each defect is shown in parentheses.

daf-7T1040% (3)
daf-7T270% (10)16% (5)
daf-7W264R25% (13)28% (14)
daf-7clv mt24% (11)21% (10)

The Larval Arrest Phenotype Occurs Through Both daf-7 and dbl-1 Signaling Pathways

We have shown that expression of daf-7 dominant negative transgenes in a daf-7/+ background results in two novel phenotypes: a molting defect where the cuticle fails to be shed and an excretory canal phenotype. However, neither of these phenotypes has ever been associated with the daf-7 signaling pathway so the mechanism by which they occur is unknown. Two hypotheses may explain this phenomenon; first, these phenotypes are the result of the disruption of daf-7 signaling and, second, they occur due to the disruption of another signaling pathway. We decided to focus on the daf-7W264R and daf-7clv mt transgenes for these experiments.

To test our first hypothesis, we investigated the role of daf-3, a Co-Smad (Patterson et al., 1997), and daf-5, a Sno/Ski homolog (da Graca et al., 2004; Tewari et al., 2004), which together promote dauer development and are antagonized by daf-7 signaling. daf-3 and daf-5 mutants are dauer defective and block the dafC phenotype of daf-7 mutants (Patterson et al., 1997; Tewari et al., 2004). Therefore, if our daf-7 dominant negative transgenes act through the daf-7 signaling pathway, we would predict that mutations in either daf-3 or daf-5 would block the early larval arrest phenotype we see in a daf-7/+ background. However, we found only a partial suppression of early larval arrest, with the greatest reduction seen in the daf-5 mutant (Fig. 7A), which suggests that our dominant negative transgenes do not act solely through the daf-7 pathway.

The daf-7 pathway shares its Type II receptor, DAF-4, with the dbl-1 TGB-β signaling pathway that controls body size and male tail development (Estevez et al., 1993; Savage et al., 1996). Thus, our dominant negative daf-7 transgenes could be disrupting early larval molting and excretory canal formation via the dbl-1 signaling pathway. To test this hypothesis, we knocked down sma-4 or sma-9 in our transgenic cross progeny and looked for early larval arrest. We found that sma-4 or sma-9 RNAi almost completely suppressed the early larval arrest phenotype for both transgenes (Fig. 7B). To further explore the role of the dbl-1 signaling pathway, we repeated the daf-7W264R and daf-7clv mt crosses into a dbl-1 background to generate dbl-1/+ transgenic cross progeny. We found significant levels of larval arrest in dbl-1/+; Ex daf-7W264R cross progeny and detectable but non-significant larval arrest in dbl-1/+; Ex daf-7clv mt cross progeny (Fig. 8).

Figure 8.

Expression of daf-7W264R and daf-7clv mt arrays in a dbl-1/+ background also causes early larval arrest but at a much lower level. Two daf-7W264R arrays (from WG496 and WG497) and two daf-7clv mt arrays (from WG508 and WG509) were tested in each background. n, number of GFP-positive cross progeny scored. Error bars are s.e.m. and data were analysed using Fisher's Exact test. n/s is non-significant. *P < <0.05, **P < 0.01, and ***P < 0.001, respectively.

Figure 9.

Schematic of the transgenes used in this study. C represents conserved cysteine residues. Peaked lines represent ligand domain C-terminal truncations. The amino acid substitutions in the daf-7W264R or daf-7clv mt transgenes are shown in their respective schematics.

DISCUSSION

A striking and often commented upon observation about TGF-β ligands is the contrast between their highly conserved primary protein structure and their low levels of sequence similarity, even between the same ligand in sister species. Given the wide range of developmental processes different TGF-β signaling pathways are important in, even in the invertebrate C. elegans, this structural conservation suggests that the central processes of TGF-β ligand protein folding, activation, and receptor binding remain the same across the animal kingdom. Support for this hypothesis comes from dominant negative mutations in a wide variety of TGF-β ligands from several different vertebrate species, all of which target either conserved ligand domain residues (Wittbrodt and Rosa, 1994; Thomas et al., 1997; Masuya et al., 2007) or the cleavage site (Lopez et al., 1992; Wittbrodt and Rosa, 1994; Hawley et al., 1995; Zhu et al., 2000).

We wanted to extend this hypothesis to invertebrates to see if these same TGF-β mutations produce a dominant negative phenotype. To do so, we chose C. elegans and the DAF-7 signaling pathway for two reasons: first, for the ease of creating and assaying transgenic animals and, second, because daf-7 animals have a completely penetrant dafC phenotype at 25°C, thus providing a clear readout of any putative dominant negative mutation. We tested the effect on dauer formation of a range of mutant daf-7 constructs, with successive ligand domain deletions, a W264R substitution equivalent to that in GDF5 (Masuya et al., 2007), and a mutated cleavage site.

Although originally intended purely as loss-of-function controls, the fact that none of the mutant daf-7 constructs were able to rescue the daf-7 dafC phenotype proves that the regions deleted or mutated are essential for DAF-7 function. All of the previously described TGF-β C-terminal truncations, such as those in Double Muscled Belgian Blue cattle (Grobet et al., 1997; Kambadur et al., 1997; McPherron and Lee, 1997), have removed almost the entire ligand domain and are unsurprisingly null. This is the first demonstration that even a small 31–amino acid C-terminal deletion, containing two of the conserved cysteine residues, renders DAF-7 null. However, further work, such as creating individual substitutions of these cysteine residues, similar to that observed in Double Muscled Piedmontese cattle and Chondrodysplasia Grebe type disease (McPherron and Lee, 1997; Thomas et al., 1997), is needed to confirm their requirement for DAF-7 function. In addition, this work is also the first demonstration of the requirement of a functioning cleavage site and the conserved tryptophan264 residue in an invertebrate TGF-β ligand. The requirement of a functioning cleavage site in DAF-7 indicates that the basic activation mechanism of TGF-β ligands, i.e., release of the active domain by the cleavage of the inactive pre-protein dimer, is conserved across the animal kingdom, even in TGF-β ligands with atypically short pro-domains (Freitas et al., 2007). The conserved tryptophan264 residue has previously been suggested to be part of a conserved pocket responsible for ligand:receptor interactions (Wittbrodt and Rosa, 1994). Although this work clearly shows that it is required for function, further work would be need to elucidate its precise role in DAF-7.

The role of different conserved DAF-7 features was only half of the question we sought to answer. We were especially interested in the ability of these mutations to produce a dominant negative daf-7 dafC phenotype in a wild-type background. We were initially disappointed as none of the mutations tested caused inappropriate dauer formation. However, one possible explanation of the inability of these mutations to affect DAF-7 signaling was that endogenous DAF-7 was present at levels that exceeded the ability of the mutant DAF-7 to sequester it. Therefore, we repeated our dauer formation experiments in a daf-7/+ background and were disappointed to find that, again, our mutant proteins were unable to cause inappropriate dauer formation. However, we noticed odd-looking early stage larvae at the beginning of one experiment and realized that our test for dauer formation in this background would automatically preclude us from discovering earlier developmental phenotypes. When we looked at larval development before the L3 stage we noticed that a large percentage of daf-7/+ larvae carrying the daf-7T2, daf-7W264R, or daf-7clv mt transgenes at 25°C were developmentally arrested, a phenotype we had never seen before in either wild-type or daf-7 backgrounds. We then examined these larvae more closely and found that the most likely reason for their developmental arrest was a molting defect, with unshed cuticle either constricting the worm or preventing its feeding and defecation, along with an excretory canal defect. Although the percentages of arrested larvae with these defects varied greatly, especially between the daf-7T2 and the daf-7W264R, or daf-7clv mt arrays, only alive arrested larvae could be scored for either defect. Thus, the numbers scored are small and are most likely an underestimation of the penetrance of each defect. It is not possible to know if these two defects are directly related or occur independently as the result of aberrant daf-7 signaling. Although this molting defect isn't the phenotype we predicted, it does confirm that these mutations, which produce dominant negative TGF-β phenotypes in vertebrates, are also capable of producing a dominant negative phenotype in C. elegans.

However, the question remained: how do these dominant negative daf-7 transgenes cause these molting and excretory canal defects? One hypothesis was that these dominant negative daf-7 transgenes act through the daf-7 signaling pathway. A second hypothesis was that they act through a separate signaling pathway, of which the dbl-1 signaling pathway was a prime candidate, due to their common Type II receptor, daf-4. To test these hypotheses, we carried out a series of genetic epistasis experiments, using mutants or RNAi of genes downstream of daf-4 in each pathway. We found that although daf-3 and daf-5 knockouts partially suppressed the early larval arrest phenotype, sma-4 or sma-9 RNAi was able to almost completely suppress it. When we looked at these arrays in a dbl-1/+ background, we saw much lower levels of larval arrest than in a daf-7/+ background, which suggests that the DBL-1 ligand is not the main target of our dominant negative transgenes but that reduced wild-type signaling through the DAF-4 receptor may play a role in this phenotype.

This poses a quandary. How can a mutated DAF-7 protein disrupt early larval molting and excretory canal formation, predominantly in a background with reduced wild-type DAF-7 levels, yet do so largely via SMA-4 and SMA-9 of the dbl-1 signaling pathway? One explanation is that the correct functioning of both signaling pathways downstream of DAF-4 is required for successful early larval molts and excretory canal formation and that both suppression of DAF-3 and DAF-5 activity and activation of SMA-9 activity are essential. The presence of a cuticle collagen organizing gene, adt-2, downstream of dbl-1 supports a role for dbl-1 signaling in molting (Fernando et al., 2011). When the levels of endogenous DAF-4 ligands are reduced and DAF-7 binding is disrupted, DAF-3, DAF-5, and SMA-9 activity are upregulated, which leads to the incorrect transcription of genes involved in molting and excretory canal formation. If this is the case, a daf-8; daf-14 double mutant containing either a sma-9 overexpression array or constitutively active sma-2 and sma-3 mutants should recapitulate this phenotype. Even so, this would not explain why our mutant DAF-7 proteins cause the DAF-4 receptor to selectively phosphorylate SMA-2 and SMA-3, yet not DAF-8 and DAF-14.

However, given the requirement of sma-4 and sma-9 for our mutant phenotype, it is not entirely clear that our mutant daf-7 transgenes are truly dominant negative (antimorph) but may instead have acquired novel characteristics (neomorph). Thus, an alternative explanation to the convergent signaling outlined above is that our mutant DAF-7 proteins bind to and inappropriately activate the DAF-4 receptor independently of their forming heterodimers with wild-type DAF-7. The ability of the daf-7W264R and daf-7clv mt transgenes to produce low levels of early larval arrest in a dbl-1/+ background and the requirement of sma-4 and sma-9 support our mutations being neomorphs. In addition, given the proposed role of the W264 residue in ligand:receptor interactions (Wittbrodt and Rosa, 1994), one could envisage a mechanism by which a substitution at this residue affects receptor binding or specificity. In contrast, the appearance of a phenotype in a m/+ background and an absence of a phenotype in a wild-type background are key characteristics of an antimorph mutation. Our mutant proteins also act at least partially through the same pathway as the wild-type protein. In addition, it is hard to think how an unprocessed pro-protein, such as that formed by DAF-7clv mt, would gain a novel phenotype when it is structurally identical to the wild-type pro-protein. As such, we propose that our daf-7 mutations are antimorphs and thus act in a dominant negative fashion, yet their action reveals a hitherto unknown relationship between the daf-7 and dbl-1 signaling pathways.

Thus, not only have we been able to induce a daf-7 dominant negative phenotype in C. elegans using mutations that induce TGF-β dominant negative phenotypes in higher organisms, demonstrating animal kingdom wide conservation of function, but doing so has revealed a novel phenotype for daf-7 and, potentially, a novel interaction between two related TGF-β signaling pathways. There are three other TGF-β ligands in C. elegans (Patterson and Padgett, 2000), involved in other developmental processes such as body length determination (Suzuki et al., 1999) and axon guidance (Colavita et al., 1998). It would be of great interest to see if these or other dominant negative mutations can be replicated in dbl-1, unc-129, or tig-2, further supporting this conservation of function, and also, possibly, discovering novel roles for these genes in the process.

EXPERIMENTAL PROCEDURES

Worm Maintenance

C. elegans N2 (Bristol wild type), CB1372 daf-7(e1372), CB1376 daf-3(e1376), CB1386 daf-5(e1386), LT121 dbl-1(wk70) (Caenorhabditis Genetics Centre), GR1269 daf-7(e1372); daf-3(e1376), and GR1278 daf-5(e1386); daf-7(e1372) (Garth Patterson, Rutgers University, New Bruswick, NJ) were maintained on NGM plates seeded with E. coli OP50 and kept at 15°C, as described previously (Sulston and Hodgkin, 1988).

Transgene Construction

See Tables 2 and 3 for a full list of primers and constructs, respectively, and Figure 9 for a graphic representation of the constructs used. Using overlap extension PCR, we first carried out a series of truncations, to produce deletions of 31, 44, 72, 100, and 115 amino acids from the C-terminal of DAF-7. To reproduce the GDF5W408R substitution (Masuya et al., 2007), primers were designed to produce a single tryptophan to arginine amino acid change at the highly conserved W264 residue (Fig. 1). Finally, to create a non-cleavable dominant negative form of DAF-7 that would heterodimerise with wild type DAF-7, as demonstrated for myostatin (Zhu et al., 2000), primers were designed to change the DAF-7 cleavage site from RKRR to GLDG. As a positive control, we cloned the 3.1-kbp genomic region containing daf-7 and its regulatory sequences. All constructs used the same 1.2-kbp daf-7 upstream region and 521-bp wild type daf-7 3′UTR.

Table 2. Oligonucleotide Primers (5'-3') Used in This Studya
Poly(dT)18VTTTTTTTTTTTTTTTTTTV
  1. a

    For linker primers denoted /CRx, the reverse primer is the reverse complement of the primer shown. Italics represent c-myc tag. Bases in bold represent introduced base changes from wild type sequence.

External primers 
Ce daf-7 CF1GTAAACACTTTTCAGATTTTTCAAATTTCAG
Ce daf-7 CR5AAAACTCTAAAGAAAATGTGTAACAGCACTCA
Linker primers 
Ce daf-7 CF3-1/ CR6GATACTGCTGCCATCCAACCTAAATTCCACTTCCGATTGTCACAAA
Ce daf-7 CF3-2/ CR7CAAGATCATGCGCGCAGCTCTAAATTCCACTTCCGATTGTCACAAA
Ce daf-7 CF3-3/ CR8CACCACGTTACAATGCCTACATGTAAATTCCACTTCCGATTGTCACAAA
Ce daf-7 CF3–4/ CR9CGAAGCACAGTCCAAGGGGTAAATTCCACTTCCGATTGTCACAAA
Ce daf-7 CF3–5/ CR10GGTCCCGCAAACGCCGTTAAATTCCACTTCCGATTGTCACAAA
Ce daf-7 CF3–7GAACAAAAACTTATTTCTGAAGAAGATCTGTAAATTCCACTTCCGATTGTCACAAA
Ce daf-7 CF7/ CR14AAAGATCGGAAGGGACTGGATCGT
Ce daf-7 CF8GGCCTAGACGGTTCTCATGCCA
Ce daf-7 CR15CCGTCTAGGCCGGACCCCTT
Ce daf-7 CR16CAGATCTTCTTCAGAAATAAGTTTTTGTTCTGAGCAACCGCATTTCTTGGC
RT-PCR primers 
Ce-act-3 QF1CCACGAGACTTCTTACAACTCCATCATG
Ce-act-3 QR1CGATGATCTTGATCTTCATGGTTGATG
Ce-daf-7 QF3CAGCTCTCCACTCGTCCAAAGG
c-myc QR1CAGATCTTCTTCAGAAATAAGTTTTTGTTC
Table 3. List of Constructs Used in This Study, the Inserts and Primers Used in Their Construction, and the Lines Containing Them
Plasmid nameInsertExternal primersLinker primersLine (background)
pWG1daf-7FL (wild type)Ce daf7 CF1Ce daf7 CF3–7WG534 (e1372)
  Ce daf7 CR5Ce daf7 CR16WG450 (N2)
pWG6.1daf-7T1 (Δ31aa)Ce daf7 CF1Ce daf-7 CF3-1WG459 (N2)
  Ce daf7 CR5Ce daf-7 CR6WG493 (N2)
    WG460 (e1372)
pWG6.2daf-7T2 (Δ44aa)Ce daf7 CF1Ce daf-7 CF3-2WG452 (N2)
  Ce daf7 CR5Ce daf-7 CR7WG461 (e1372)
    WG462 (e1372)
pWG6.3daf-7T3 (Δ72aa)Ce daf7 CF1Ce daf-7 CF3-3WG494 (e1372)
  Ce daf7 CR5Ce daf-7 CR8WG495 (e1372)
pWG6.4daf-7T4 (Δ100aa)Ce daf7 CF1Ce daf-7 CF3–4WG463 (e1372)
  Ce daf7 CR5Ce daf-7 CR9 
pWG6.5daf-7T5 (Δ115aa)Ce daf7 CF1Ce daf-7 CF3–5WG471 (e1372)
  Ce daf7 CR5Ce daf7 CR10WG472 (e1372)
pWG7daf-7W264RCe daf7 CF1Ce daf7 CF7WG496 (N2)
  Ce daf7 CR5Ce daf7 CR14WG497 (N2)
   Ce daf7 CF3–7WG498 (N2)
   Ce daf7 CR16WG520 (e1372)
pWG8daf-7clv mtCe daf7 CF1Ce daf7 CF8WG508 (N2)
  Ce daf7 CR5Ce daf7 CR15WG509 (N2)
   Ce daf7 CF3–7WG510 (N2)
   Ce daf7 CR16WG527 (e1372)

Plasmid DNA was prepared for microinjection by LiCl precipitation and injected into daf-7(e1372) hermaphrodites at 50 ng.μl−1 with 25 ng.μl−1 of pPD129.51 (Fire et al., 1990; Addgene Plasmid 1655, www.addgene.org), a nuclear localized gfp reporter under control of the C. elegans rps-5 promoter, as a co-transformation marker.

Dauer Phenotype Assay

daf-7(e1372) animals are dauer constitutive (dafC) at 25°C. To test for a DafC phenotype, 15–20 gfp-positive gravid hermaphrodites were picked for three NGM-OP50 plates, which had been pre-warmed to 25°C, and were left to lay eggs at 25°C. After 3 hr, the adults were removed and the plates incubated at 25°C for 3 days. Worms were scored for developmental fate (adults or dauers) and for gfp expression. Percentage rescue was calculated as number of gfp-positive dauers/total number of gfp-positive worms in a daf-7(e1372) genetic background. Each dauer phenotype assay was repeated at least twice per line. Once the lack of function of each construct was verified by their failure to rescue the daf-7(e1372) dafC phenotype, the daf-7T1, daf-7T2, daf-7clv mt, and daf-7W264R arrays were crossed into the N2 wild type background and assayed for a dominant negative dafC phenotype.

RT-PCR

To test for daf-7 transgene expression in a subset of lines, containing the daf-7FL, daf-7clv mt, and daf-7W264R arrays, RNA was purified by trizol extraction from synchronized L1 populations hatched at 25°C in M9. cDNA was prepared using a poly dT(V) primer and Superscript reverse transcriptase (Invitrogen, Carlsbad, CA), and amplified by PCR (94°C for 3 min, followed by 35 cycles of 94°C for 30 sec, 60°C for 30 sec, 72°C for 1 min, with a final extension of 72°C for 7 min) with Ce-act-3 QF1 and QR1 primers to verify cDNA integrity. The PCR was then repeated with transgene-specific primers Ce-daf-7 QF3 and c-myc QR1, to produce 416- and 467-bp bands from cDNA and array DNA, respectively.

Assays in a daf-7/+ Background

To test for possible dominant negative phenotypes in a daf-7/+ background, GFP-positive males containing the daf-7FL (from WG450), daf-7T1 (from WG459), daf-7T2 (from WG452), daf-7W264R array (from WG496 and 497) or the daf-7clv mt (from lines WG508 and 509) arrays were crossed into a daf-7(e1372) background at 15°C or 25°C. GFP-positive progeny from this cross both contain the array and are daf-7 heterozygotes. Dauer formation in these cross progeny was assayed by allowing the cross to proceed for 2 days at 25°C or 3 days at 15°C (approximately equivalent to L4 stage for the first progeny produced) after which the cross plates were flooded with 1% sodium dodecyl sulphate (SDS) solution and incubated at 20°C for 30 min. Plates were scored for the number of GFP-positive dauers and L4 larvae.

A second test, looking at the development of pre-L3 stage larvae, was carried out by removing the adults from the 25°C and 15°C cross plates after 24 and 48 hr, respectively. GFP-positive cross progeny were then transferred to new plates 24 hr after the adults were removed and scored for arrested larvae 48 hr after that, by which point all normally developing larvae should be L3/dauer stage or older. Any arrested larvae were counted and examined by DIC microscopy for defects and presence of GFP expression in one or both ASI neurons. Each cross was repeated at least in triplicate.

Requirement of daf-7 and dbl-1 Downstream Pathway Members for the daf-7/+ Dominant Negative Phenotype

To determine in which TGF-β pathway our dominant negative daf-7 transgenes are acting, we took two approaches. For the daf-7 pathway, we generated daf-3(e1376) and daf-5(e1386) males carrying the daf-7W264R array (from WG496) or the daf-7clv mt array (from lines WG508 and 509) and crossed these males with daf-7(e1372); daf-3(e1376) and daf-5(e1386); daf-7(e1372) hermaphrodites, respectively, at 15°C and 25°C as above. For the dbl-1 pathway, we repeated the original crosses with wild-type males carrying daf-7W264R or daf-7clv mtarrays into a daf-7(e1372) background on sma-4 and sma-9 RNAi plates (Source BioSciences, Cambridge, UK) at 15°C and 25°C, as above. We also crossed wild-type males carrying daf-7W264R (from WG496 and 497) or daf-7clv mt (from WG508 and 509) arrays into a dbl-1(wk70) background at 15°C and 25°C.

ACKNOWLEDGMENTS

This work was funded by an AR&C fellowship from AgResearch Ltd., New Zealand. M.C. is currently supported by NIH grant number R01GM086786. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources (NCRR). We thank Kirsten Grant for carrying out all of the microinjections, Matthew Beuchner for his help analyzing the DIC images, Garth Patterson and Mary Anne Royal for the daf-7; daf-3 and daf-5; daf-7 strains, and Wendy Hanna-Rose for comments on the manuscript.

Ancillary