The Groucho ortholog UNC-37 interacts with the short Groucho-like protein LSY-22 to control developmental decisions in C. elegans

Abstract

Transcriptional co-repressors of the Groucho/TLE family are important regulators of development in many species. A subset of Groucho/TLE family members that lack the C-terminal WD40 domains have been proposed to act as dominant-negative regulators of Groucho/TLE proteins, yet such a role has not been conclusively proven. Through a mutant screen for genes controlling a left/right asymmetric cell fate decision in the nervous system of the nematode C. elegans, we have retrieved loss-of-function alleles in two distinct loci that display identical phenotypes in neuronal fate specification and in other developmental contexts. Using the novel technology of whole-genome sequencing, we find that these loci encode the C. elegans ortholog of Groucho, UNC-37, and, surprisingly, a short Groucho-like protein, LSY-22, that is similar to truncated Groucho proteins in other species. Besides their phenotypic similarities, unc-37 and lsy-22 show genetic interactions and UNC-37 and LSY-22 proteins also physically bind to each other in vivo. Our findings suggest that rather than acting as negative regulators of Groucho, small Groucho-like proteins may promote Groucho function. We propose that Groucho-mediated gene regulatory events involve heteromeric complexes of distinct Groucho-like proteins.

Fig. 1.

Left/right asymmetry of the ASE neurons is controlled by a complex network of regulatory factors, including lsy-5 and lsy-22. (A) Summary of the previously known factors controlling the laterality of the ASEL/ASER neurons in C. elegans. Red indicates ASEL fate and blue indicates ASER fate. A bistable feedback loop controls expression of the downstream terminal differentiation genes, of which only a selected subset is shown. The negative regulation of die-1 expression by cog-1 involves 3′UTR regulation of die-1 (Chang et al., 2004) as well as uncharacterized regulatory interactions (Didiano et al., 2010). (B) Loss of lsy-22 and lsy-5/unc-37 show a similar effect on downstream gfp reporters, in which expression of the ASER fate marker gcy-5::mCherry is lost and the ASEL fate marker lim-6::gfp is ectopically expressed in ASER (the `two ASEL' phenotype). See C for quantification of data. Red and blue circles indicate ASEL and ASER neurons, respectively. (C) Quantification of laterality defects in lsy-22 and lsy-5/unc-37 mutants. We refer to lsy-5 as unc-37 (see text). Shown is the percentage of animals within the population with a given phenotype. Circles represent ASEL and ASER and green shading indicates whether and where the respective fate marker is expressed. lim-6 marker gene expression for alleles ot244, ot114, ot37, ot240 and e262 was rescored and found to be similar to that previously reported (Chang et al., 2003; Sarin et al., 2007). ¹Scored with lim-6::gfp (otIs114) (^*) and/or gcy-7::gfp (otIs3 or otIs4) (^**). If scored with both, the lim-6::gfp data are shown. ²Scored with gcy-5::gfp (ntIs1) (^#)or gcy-5::mCherry (otIs220) (^##). ³Homozygous unc-37 or lsy-22 worms were scored from the heterozygous balanced strain lsy-22 or unc-37/hT2[bli-4(e937)let-?(q782)qIs48]. ⁴cog-1(OE) is an extrachromosomal array, otEx4124 Ex[ceh-36::cog-1, elt-2::gfp], that ectopically expresses cog-1 in ASEL and ASER. ⁵Data from Sarin et al. (Sarin et al., 2007) and shown for comparison purposes only.

Fig. 2.

Molecular identity of lsy-22 and lsy-5. (A) Whole-genome sequencing (WGS) data for the lsy-22(ot114) and lsy-5(ot240) alleles. ^*, Variants were considered as background if they were also found in other WGS datasets from the laboratory (see Materials and methods). ^**, 1/7 was found to be a heterozygous variant. (B) LSY-22 and UNC-37, showing the location of the various mutant alleles. Some previously isolated unc-37 alleles (Pflugrad et al., 1997) that we used here are indicated in gray text. Detailed sequence comparisons, including coiled-coil domains, are shown in Fig. S1 in the supplementary material. (C) lsy-22::venus and unc-37::yfp reporters generated by fosmid recombineering are broadly expressed during embryogenesis and larval development. Representative ∼1.5-fold stage C. elegans embryos are shown.

Fig. 3.

lsy-22 and unc-37 control similar developmental processes. (A) lsy-22 synergizes with unc-37 for sterility/lethality. unc-37(hypo) is the hypomorphic mutant allele e262, unc-37(null) is the null allele wd17wd22 and lsy-22(null) is the null allele ot244. Null mutant animals that were maternally rescued for viability were picked based on their Lsy phenotype and their progeny scored for viability. For more information on RNAi, see Table S1 in the supplementary material. Control RNAi refers to the L4440 vector clone. (B) Unc phenotypes of lsy-22 and lsy-5/unc-37 (observed in 50/50 animals of each genotype; 0/50 wild-type animals show comparable locomotory tracks). Single homozygous adults were placed in the middle of an E. coli lawn and observed after 5 minutes. The red circle is where the animal was originally placed and tracks were drawn over a photo of the animal's tracks as an estimation of its path. One representative example is shown. (C) lsy-22 and unc-37 mutants show ectopic expression of the wdIs3 (del-1::gfp) reporter in the presumptive VA motoneurons [21/21 animals for lsy-22(ot244) and 37/40 animals for unc-37(e262)]. The reporter is expressed only in the VB motoneurons in wild-type animals at the L2 stage. (D) unc-37 and lsy-22 animals display protruding vulva (Pvl) phenotypes (observed in 50/50 animals of each mutant genotype). The ventral surface is indicated by a dotted line and the position of the vulva is indicated with an asterisk. (E) Removal of maternal lsy-22 and unc-37 by RNAi results in ectopic expression of the qIs19(lag-2::gfp) reporter in embryos. An asterisk indicates ectopic expression in certain AB descendants; dotted lines encircle endogenous expression in AB and MS descendants. Images are lateral views, anterior to the left. Percentages of ectopic expression observed were: control (L4440), 4% (n=28); unc-37(RNAi), 97% (n=28); lsy-22(RNAi), 81% (n=26); unc-37(RNAi) lsy-22(RNAi), 95% (n=21). (F) pkd-2::gfp (bxIs14) reveals missing rays in both unc-37 and lsy-22 maternally rescued null mutant males [20/20 and 9/9 for lsy-22(ot244) and unc-37(wd17wd22), respectively].

Fig. 4.

LSY-22 physically interacts with UNC-37. (A) Yeast two-hybrid assay with UNC-37 and LSY-22 fused to the GAL4-activating domain (AD) or DNA-binding domain (BD) shows interaction of these two proteins. Additionally, the assay indicates that LSY-22 is able to interact with itself, and COG-1 is able to interact with UNC-37. (B) Co-immunoprecipitation of UNC-37 and LSY-22 from transgenic C. elegans expressing 2×FLAG-VENUS epitope-tagged LSY-22 (otEx4125). The lysate column indicates whether lysates were prepared from wild-type animals (WT) or from otEx4125 animals. Control refers to beads with non-specific antibody. Endogenous UNC-37 was detected using an anti-UNC-37 antibody (Winnier et al., 1999). (C) Model for UNC-37–LSY-22 interaction. Given the reported Q-rich domain-mediated oligomerization of UNC-37/Groucho in other species, we suggest that the interaction occurs via their respective Q-rich domains, but interaction via other domains, such as the GP domain, is conceivable as well. This might occur in a tetrameric configuration, as previously reported for Groucho (Chen et al., 1998). Our data indicating a lack of interaction between UNC-37 and itself suggest that LSY-22 may be required for UNC-37 co-repressor complex formation. The interaction of UNC-37 and COG-1 might occur via a WD40/EH1 domain interaction, as suggested by crystallographic studies of such domains (Pickles et al., 2002). We propose that a broadly expressed LSY-22–UNC-37 complex interacts in a tissue-specific manner with specific DNA-binding transcription factors to control cell fate decisions; in ASE, this transcription factor is COG-1, in the VA motoneurons it might be the EH1 domain-containing UNC-4 protein (Winnier et al., 1999), and in other cellular contexts it might be any one of the dozens of EH1 domain-containing proteins (marked X) present in C. elegans (Copley, 2005) (or any other transcription factor interacting with UNC-37 in a non-EH1 domain-dependent manner).