The function and regulation of the GATA factor ELT-2 in the C. elegans endoderm

Abstract

ELT-2 is the major regulator of genes involved in differentiation, maintenance and function of C. elegans intestine from the early embryo to mature adult. elt-2 responds to overexpression of the GATA transcription factors END-1 and END-3, which specify the intestine, as well as to overexpression of the two GATA factors that are normally involved in intestinal differentiation, ELT-7 and ELT-2 itself. Little is known about the molecular mechanisms underlying these interactions, how ELT-2 levels are maintained throughout development or how such systems respond to developmental perturbations. Here, we analyse elt-2 gene regulation through transgenic reporter assays, ELT-2 ChIP and characterisation of in vitro DNA-protein interactions. Our results indicate that elt-2 is controlled by three discrete regulatory regions conserved between C. elegans and C. briggsae that span >4 kb of 5' flanking sequence. These regions are superficially interchangeable but have quantitatively different enhancer properties, and their combined activities indicate inter-region synergies. Their regulatory activity is mediated by a small number of conserved TGATAA sites that are largely interchangeable and interact with different endodermal GATA factors with only modest differences in affinity. The redundant molecular mechanism that forms the elt-2 regulatory network is robust and flexible, as loss of end-3 halves ELT-2 levels in the early embryo but levels fully recover by the time of hatching. When ELT-2 is expressed under the control of end-1 regulatory elements, in addition to its own endogenous promoter, it can replace the complete set of endoderm-specific GATA factors: END-1, END-3, ELT-7 and (the probably non-functional) ELT-4. Thus, in addition to controlling gene expression during differentiation, ELT-2 is capable of specifying the entire C. elegans endoderm.

Keywords: Caenorhabditis elegans; ChIP-Seq; ELT-2; Endoderm development; GATA factor; Transcription.

Fig. 1.

Regulatory network consisting of the four zygotically expressed endoderm-specific GATA-type transcription factors that specify and differentiate the C. elegans early endoderm (E lineage). Time scale (minutes after first cell division at 20°C) is shown on the left. In the centre are images of three early stages of embryogenesis: the 1E, 2E and 4E cells are indicated by white dots. The current model for the roles and regulatory relations between the various transcription factors is shown on the right.

Fig. 2.

Transcriptional regulation of the elt-2 gene. (A) The expression of transgenic reporters accurately reflects the in vivo expression of ELT-2. The top row shows the normal endogenous expression patterns of ELT-2 in early C. elegans embryos, as detected by immunofluorescence using the anti-ELT-2 monoclonal antibody 455-2A4. The middle row shows expression patterns in early to mid-stage embryos of a transgenic nuclear-localised GFP reporter construct driven by the 5048 bp 5′ flanking region of the elt-2 gene. Egg shells are outlined (dashed line). Beneath is a differential interference contrast (DIC) image of an L1 larva, with the fluorescence from the transgenic elt-2::GFP reporter superimposed. Scale bar: 20 μm. In each image, fluorescence signal is adjusted to high contrast to emphasize expression patterns and the lack of non-intestinal expression. (B) Identifying conserved regions in the elt-2 5′ control region by sequence alignments. The dot matrix plot (EMBOSS/dotmatcher, www.ebi.ac.uk/tools/emboss) compares 6 kb upstream of the ATG initiation codon for C. elegans (horizontal axis) and C. briggsae (vertical axis), revealing three blocks of conserved sequences (CR I, CR II and CR III). These conserved blocks are aligned with the genomic locus of the C. elegans elt-2 gene, showing (to scale and from left to right) the upstream elt-4 gene, the apparent ORF C39B10.7 and the elt-2 coding region with the transcriptional start site (TSS) indicated. Also shown are two genomic deletions (ca16 and gk153), TGATAA sites (filled triangles) and WGATAR sites that are not TGATAA (open circles).

Fig. 3.

Enhancer activities of the three conserved regions identified in the 5′ flanking region of elt-2. (A) The enhancer activity of CR I, CR II and CR III was tested individually and in combinations. Reporter expression patterns from multiple independent transgenic lines are summarised as: +++, similar pattern and intensity as the intact 5 kb promoter; ++ and +, decreasing (∼by half) steps in intensity; −, no detectable reporter expression. (B) Importance of conserved TGATAA sites in CR III and CR I for elt-2 enhancer activity. Mutated TGATAA sites are marked by ‘X’.

Fig. 4.

DNA-protein interactions in the elt-2 control region. Electrophoretic mobility shift assays to show that END-1, ELT-2 and ELT-7 proteins can all bind directly to each of the four conserved TGATAA sites in CR III of the elt-2 promoter. The same set of labelled probes was used for all three proteins, with the coordinates of the individual conserved TGATAA sites shown at the top.

Fig. 5.

ELT-2 ChIP-Seq on the elt-2, ges-1 and cpr-6 loci. (A) ELT-2 ChIP-Seq tracks (dark blue) from L3 larval worms are shown on and around the elt-2 gene, with significant MACS2 peaks highlighted above (dark blue bars). ChIP-Seq reads were normalised with respect to read depth and IgG-only controls. The average of three replicates is shown. Individual replicates are shown in Fig. S5. The corresponding RNA-Seq (light blue) results obtained from the same chromatin preparation (whole L3 worms) are shown below. Regions of poor mappability owing to genomic repeats are depicted in the Repeatmasker (www.repeatmasker.org) trace (grey). The location of CR I, CR II and CR III and the occurrences of TGATAA motifs are also shown. (B,C) ELT-2 ChIP-Seq and RNA-Seq tracks at the (B) ges-1 and (C) cpr-6 loci.

Fig. 6.

ELT-2 can replace END-1, END-3, ELT-7 and ELT-4. (A) Embryos from the rescued quadruple mutant strain JM229 (black circles) hatch later than embryos from the control strain JM230 (white circles). Other symbols represent hatching curves measured for four different local versions of N2 wild-type worms (including a recent thaw). Time on the x-axis is minutes at 20°C from the 1- to 4-cell stage of embryogenesis. (B) DIC image of an L1 larva from the rescued quadruple mutant strain JM229. Average length of JM229 L1 larvae is 242±20 μm. (C) Fluorescent image of the same larva as in B, showing expression of an elt-2prom::GFP reporter incorporated into the integrated rescuing array caIs85. (D) DIC image of an L1 larva from N2 wild-type control. Average length of N2 L1 larvae is 278±18 μm.