Different Mi-2 complexes for various developmental functions in Caenorhabditis elegans

Abstract

Biochemical purifications from mammalian cells and Xenopus oocytes revealed that vertebrate Mi-2 proteins reside in multisubunit NuRD (Nucleosome Remodeling and Deacetylase) complexes. Since all NuRD subunits are highly conserved in the genomes of C. elegans and Drosophila, it was suggested that NuRD complexes also exist in invertebrates. Recently, a novel dMec complex, composed of dMi-2 and dMEP-1 was identified in Drosophila. The genome of C. elegans encodes two highly homologous Mi-2 orthologues, LET-418 and CHD-3. Here we demonstrate that these proteins define at least three different protein complexes, two distinct NuRD complexes and one MEC complex. The two canonical NuRD complexes share the same core subunits HDA-1/HDAC, LIN-53/RbAp and LIN-40/MTA, but differ in their Mi-2 orthologues LET-418 or CHD-3. LET-418 but not CHD-3, interacts with the Krüppel-like protein MEP-1 in a distinct complex, the MEC complex. Based on microarrays analyses, we propose that MEC constitutes an important LET-418 containing regulatory complex during C. elegans embryonic and early larval development. It is required for the repression of germline potential in somatic cells and acts when blastomeres are still dividing and differentiating. The two NuRD complexes may not be important for the early development, but may act later during postembryonic development. Altogether, our data suggest a considerable complexity in the composition, the developmental function and the tissue-specificity of the different C. elegans Mi-2 complexes.

Figure 1. Two NuRD complexes and a MEC complex are present in C. elegans.

(A) LET-418 interacts with HDA-1 and LIN-53, but not with CHD-3 nor with HDA-2 and HDA-3. Extracts from wild-type mixed-stage worms were precipitated with either α-LET-418 or no antibodies (negative control). Inputs and immunoprecipitates were subjected to Western analysis and immunoblotted with antibodies directed against proteins indicated next to each panel. (B) HDA-1 binds to CHD-3. Extracts from wild-type or chd-3(eh4) mixed-stage worms were precipitated with either α-HDA-1 or no antibodies (negative control). Inputs and immunoprecipitates were subjected to Western analysis and immunoblotted with antibodies directed against CHD-3. (C) LIN-53::GFP interacts with LET-418, HDA-1 and CHD-3. Extracts from lin-53::gfp or lin-53(-) mixed-stage worms were precipitated with either α-GFP or no antibodies (negative control). Inputs and immunoprecipitates were subjected to Western analysis and immunoblotted with antibodies directed against proteins indicated next to each panel. (D) LIN-40::GFP interacts with LET-418, HDA-1 and CHD-3, but not with MEP-1. Extracts from lin-40::gfp or lin-40(+) mixed-stage worms were precipitated with either α-GFP or no antibodies (negative control). Inputs and immunoprecipitates were subjected to Western analysis and immunoblotted with antibodies directed against proteins indicated next to each panel. (E) EGL-27::GFP does not interact with LET-418 nor with HDA-1. Extracts from egl-27::gfp or egl-27(-) mixed-stage worms were precipitated with either α-GFP or no antibodies (negative control). Inputs and immunoprecipitates were subjected to Western analysis and immunoblotted with antibodies directed against proteins indicated next to each panel. The membrane was reprobed with anti-GFP as a positive control. (F) MEP-1 interacts with LET-418 and HDA-1, but not CHD-3 nor with LIN-53. Extracts from wild-type mixed-stage worms were precipitated with either α-MEP-1 or no antibodies (negative control). Inputs and immunoprecipitates were subjected to Western analysis and immunoblotted with antibodies directed against proteins indicated next to each panel. Input: 5% of the immunoprecipitate; IP: 100% of the immunoprecipitate (1 mg of proteins). All co-immunoprecipitation experiments were reproducibly performed at least twice.

Figure 2. The MEC complex negatively regulates lag-2::gfp expression in the gut.

The lag-2::gfp transgene is expressed in the gut of let-418(RNAi), mep-1(RNAi) and hda-1(RNAi) L3 larvae. (A) L3 worms carrying lag-2::gfp transgene and fed bacteria containing empty vector (RNAi control) (A) show expression in the Distal Tip Cells (DTC) and in the ventral nerve cord. No ectopic expression is observed. (B, D–E) lag-2::gfp is ectopically expressed in the gut of let-418(RNAi) (B), mep-1(RNAi) (D) and hda-1(RNAi) (E) L3 larvae. (C) No ectopic expression is observed in chd-3(RNAi) L3 larvae. (F) lag-2::gfp is mainly expressed in the epidermis in lin-53(RNAi) L3 larvae. (G) lag-2::gfp is ectopically expressed only in two cells in the most anterior part of the intestine in lin-40(RNAi) L3 larvae. The asterisk marks the DTC.

Figure 3. LET-418 and MEP-1 regulate common target genes.

A strong correlation is observed between the genes deregulated in let-418(RNAi) and mep-1(RNAi) L1 larvae. The fold change of each of the common genes was plotted on the graph (X-axis: fold change of let-418(RNAi) genes, Y-axis: fold change of the mep-1(RNAi)). Each circle represents a single gene. A standard correlation factor of R = 0.98 was calculated according to the linear regression.

Figure 4. LET-418 and MEP-1 regulate germline and early embryonic genes.

Pie charts show repartition and functional annotation of the common deregulated genes (A) and the early genes (B) according to gene ontology (GO) annotations. The GO terms are indicated on the right of each pie chart. (A) The pie chart shows the repartition of the 914 common deregulated genes compared to the list of 4699 germline genes. The common genes are divided into four groups: upregulated germline (rose slice) and non germline (gray slice) genes; and downregulated germline (pink slice) and non germline (blue slice) genes. (B) The pie chart shows the repartition of the 228 early genes compared to the common upregulated genes and to the list of 4699 germline genes. 49 common upregulated genes are also early genes. They are subdivided in germline (pink slice) and non germline (rose slice) genes. The remaining early genes, that are not targets of MEC, are subdivided in germline (blue slice) and non germline (gray slice) genes.

Figure 5. CHD-3 acts as a component of the NuRD complex to regulate fbxa-103.

Fold change of fbxa-103 mRNA was analyzed by qRT-PCR in diverse RNAi-treated L1 larvae, as indicated on the right of the panel. fbxa-103 is upregulated in chd-3(RNAi) and lin-53(RNAi) but not in let-418(RNAi) nor in mep-1(RNAi) L1 larvae.

Figure 6. let-418 and chd-3 transcriptional reporters have different expression patterns.

(A-B) Young adults animals carrying both a transcriptional let-418p::Venus (first lane) and a transcriptional chd-3p::DsRed2 (second lane) reporter genes were analyzed by confocal microscopy. The third lane is the merged picture from both let-418p::Venus and chd-3p::DsRed2 reporters. Expression is shown (A) in the head and (B) in the vulva and somatic gonad. v: vulva (ventral view); g: somatic gonad; vn: ventral nerve cord.

Figure 7. Summary figure of the complexes and their proposed roles.

The two C. elegans Mi-2 proteins are members of at least three different complexes, two distinct NuRD complexes and a MEC complex.

Figure 8. Phylogenetic comparison of the C. elegans, Drosophila and human Mi-2 orthologues.

The regions of the different Mi-2 proteins corresponding to residues 320 to 1076 of the CeLET-418 sequence were used to construct the phylogenetic tree. Numbers refer to bootstrap values supporting particular groupings.