Endoderm development in Caenorhabditis elegans: the synergistic action of ELT-2 and -7 mediates the specification→differentiation transition

Abstract

The transition from specification of cell identity to the differentiation of cells into an appropriate and enduring state is critical to the development of embryos. Transcriptional profiling in Caenorhabditis elegans has revealed a large number of genes that are expressed in the fully differentiated intestine; however, no regulatory factor has been found to be essential to initiate their expression once the endoderm has been specified. These gut-expressed genes possess a preponderance of GATA factor binding sites and one GATA factor, ELT-2, fulfills the expected characteristics of a key regulator of these genes based on its persistent expression exclusively in the developing and differentiated intestine and its ability to bind these regulatory sites. However, a striking characteristic of elt-2(0) knockout mutants is that while they die shortly after hatching owing to an obstructed gut passage, they nevertheless contain a gut that has undergone complete morphological differentiation. We have discovered a second gut-specific GATA factor, ELT-7, that profoundly synergizes with ELT-2 to create a transcriptional switch essential for gut cell differentiation. ELT-7 is first expressed in the early endoderm lineage and, when expressed ectopically, is sufficient to activate gut differentiation in nonendodermal progenitors. elt-7 is transcriptionally activated by the redundant endoderm-specifying factors END-1 and -3, and its product in turn activates both its own expression and that of elt-2, constituting an apparent positive feedback system. While elt-7 loss-of-function mutants lack a discernible phenotype, simultaneous loss of both elt-7 and elt-2 results in a striking all-or-none block to morphological differentiation of groups of gut cells with a region-specific bias, as well as reduced or abolished gut-specific expression of a number of terminal differentiation genes. ELT-2 and -7 synergize not only in activation of gene expression but also in repression of a gene that is normally expressed in the valve cells, which immediately flank the termini of the gut tube. Our results point to a developmental strategy whereby positive feedback and cross-regulatory interactions between two synergistically acting regulatory factors promote a decisive and persistent transition of specified endoderm progenitors into the program of intestinal differentiation.

FIGURE 1. elt-7::GFP is expressed strongly and exclusively in the endoderm

(A) Arrows point to the two gut progenitor cells that have just entered the interior of the embryo during the onset of gastrulation in this DIC image of a wild-type embryo. (B) Fluorescent image of the embryo in (A) shows expression of an elt-7::GFP transcriptional fusion reporter that is first detectable at the late 2E cell stage. (C) DIC image of a wild-type embryo at the 16E cell stage, just prior to elongation. The gut cells are outlined. (D) Fluorescence image of the embryo in (C) shows that expression of the reporter has greatly increased by this stage. (E–F) Reporter expression persists throughout the organ as seen in this L2 stage larva.

FIGURE 2. ELT-7 is sufficient to activate ectopic gut differentiation

Early embryos carrying hs-elt-7 were heat-shocked and allowed to develop overnight. (A) Wild-type embryos that are not heat-shocked continue development and hatch. (B) Virtually all cells in arrested transgenic embryos following a heat-shock contain nuclei with the “fried egg” morphology characteristic of differentiated intestinal cells. (C) MH33 staining of wild type IFB-2 expression is gut-specific in an elongated embryo. (D) MH33 staining is seen throughout an arrested hs-elt-7 embryo as a series of rings of varying size. (E) 1CB4 stains the intestine of wild-type embryos beginning at the onset of elongation, similar to MH33; pharyngeal gland cells also stain with this antibody, seen just anterior to the gut. (F) 1CB4 staining in an arrested hs-elt-7 embryo extends throughout the embryo. Strong staining, similar in intensity to that seen in the gut lumen, is observed around individual cells. (G) pep-2::GFP in a wild-type L1 larva. (H) pep-2::GFP fills an arrested hsp::ELT-7 embryo. (I) pho-1::lacZ::GFP expression begins shortly before hatching and remains high throughout the rest of development, as shown in this L3 larva. (J) Virtually all nuclei of an arrested hs-elt-7 embryo express pho-1::lacZ::GFP.

FIGURE 3. Activation of elt-7::GFP and elt-2::GFP by gut-specific GATA factors

(A–D) Expression of an elt-7::lacZ::GFP transcriptional fusion reporter is driven by integrated hs-end-1 (A), hs-end-3 (B), hs-elt-2 (C) and hs-elt-7 (D) constructs following heat shock of early embryos and overnight development. The reporter is expressed throughout the embryo in each case. (E) Expression of an elt-2::lacZ::GFP reporter (as described in (Fukushige et al., 1999)) after being treated similarly in an embryo carrying hs-elt-7 on a separate extrachromosomal element. (F) Expression of elt-2::lacZ::GFP throughout the entire body of an L1 larva observed several hours after it was subjected to heat shock.

FIGURE 4. Synergistic requirement of ELT-7 and ELT-2 in morphological gut differentiation

(A, B) A typical elt-7(tm840) L1 larva has a smooth, well-defined gut lumen and brush border, as observed by DIC microscopy (A) and autofluorescent gut granules evident throughout the intestine (B), appearing essentially wild-type. (C) The blockage and swelling of the brush border surrounding the gut lumen (arrow) is apparent in an elt-2(ca15) L1 larva, but morphological differentiation of the entire gut occurs normally. Note the lumen progressing continuously from the pharynx (arrowhead) throughout the length of the gut. (D) Autofluorescent gut granules are present throughout the intestine of the same elt-2(ca15) larva. (E–G) A representative elt-7(tm840); elt-2(ca15) L1 larva lacks an evident brush border, lumen, and rhabditin granules in sporadic patches in the region between the pharynx (arrowhead in (E)) and rectum (arrow in (E)). These patches show no apparent signs of differentiation. (G) Magnified view of a portion of the larva in (E) shows that the pharynx lumen is continuous with a small anterior portion of the gut lumen and brush border, which end abruptly (arrowhead). Only a single small patch of brush border is present more posteriorly in this animal (arrow). Birefringent (E, G) and autofluorescent (F) gut granules are also observed only sporadically. (H) The average frequencies of visible lumen (left panel) and gut granules (right panel) are dramatically reduced in elt-7(tm840); elt-2(ca15) L1 larvae compared to those in elt-2(ca15) or elt-7(tm840) single mutant larvae.

FIGURE 5. ELT-2 and ELT-7 function synergistically to activate markers of intestinal fate

(A–C) Expression of itx-1::GFP. (A) Wild-type worms and (B) elt-2(ca15) L1 larvae show similar levels of itx-1::GFP expression. (C) itx-1::GFP is expressed sporadically and at reduced levels in elt-2(ca15);elt-7(RNAi) L1 larvae. (D–H) MH33 staining of IFB-2. (D) A wild-type L1 shows uniform staining with antibody MH33. (E) MH33 staining appears wild-type in elt-2(ca15) L1 larvae. (F) Staining with MH33 reveals only sporadic patches of IFB-2 in elt-7(tm840); elt-2(ca15) double mutants. Images (D) and (E) were taken with a 100 ms exposure time; image (F) was taken with a 300 ms exposure. (G) elt-2(ca15) worms stain for IFB-2 along the entire length of the lumen, while elt-7(tm840); elt-2(ca15) worms show significant reduction in frequency of staining across the gut. (H) An even greater difference between single and double mutants is observed when comparing the frequency of MH33 staining that shows typical lumen-like morphology. Anterior is to the left in all images.

FIGURE 6. ELT-2 and ELT-7 are required to form normal apical junctions within intestinal cells

MH27 staining of intestinal AJM-1 is not significantly different in wild-type (A) and elt-2(ca15) L1 larvae (B). (C) MH27 reveals that only sporadic patches of AJM-1 are present in elt-7(tm840); elt-2(ca15) larvae and shows an intense valve-cell-like staining pattern at the anterior (bracket). (D) Quantification of MH27 staining frequency across the length of the intestine reveals and anterior bias in the double mutants. (E) AJM-1::GFP expression is observed in elt-2(ca15) mutants (top worm) and L1 worms rescued for the elt-2 mutation (bottom worm). Rescued worms show wild type expression, while elt-2(ca15) mutants show strong reporter expression extending caudally into the region of the anterior intestine (bracket). (F) Caudal extension of intense reporter expression is also virtually always seen in elt-7(tm840); elt-2(ca15) L1s (bracket), though the reporter expression is sporadic along the length of the gut. (G) Analysis of reporter signal reveals a strong anterior bias along the length of the intestine. Arrowheads mark the position of the pharyngeal-intestinal valves. (A–C) are confocal images.

FIGURE 7. ELT-7 synergizes with ELT-2 to repress markers of valve cell fate in the terminal gut regions

(A–D) Expression of the cdf-1::GFP reporter. (E–F) In situ hybridization of cdf-1 transcripts. (A) Wild-type animals show strong cdf-1::GFP expression in the pharyngeal-intestinal (arrowhead) and rectal-intestinal (arrow) valve cells. The expression pattern is not substantially altered in elt-7(RNAi) (B) or elt-2(ca15) (C) animals. (D) Expression is greatly enhanced (brackets) at both termini of the intestine in elt-2(ca15);elt-7(RNAi) animals, while it remains very low in the middle of the organ. (E) cdf-1 transcripts detected by in situ hybridization are restricted to the pharyngeal and rectal valve cells in all elt-2(ca15) worms (both with and without the elt-2(+) rescuing array, see Materials and Methods). (F) Typical expansion of cdf-1 mRNA hybridization seen in elt-7(tm840); elt-2(ca15) animals. (G) A fraction of elt-7(tm840); elt-2(ca15) worms, corresponding to the percentage of non-rescued worms in this strain, showed strong expansion of cdf-1 mRNA (brackets) into the terminal gut regions. * Chi square test p-value = 0.0007.

FIGURE 8. Proposed regulatory network of endoderm-specific GATA factors

Solid grey arrows indicate a relationship implied by overexpression data. Solid black arrows denote a relationship demonstrated through overexpression data and either gel-shift analysis (END-1 and elt-2; ELT-2 and ges-1), nuclear-spot assay (ELT-2 and elt-2), or loss-of-function (END-1 and END-3 → elt-2). All other black arrows are based on loss-of-function studies. Other factor(s) proposed to be involved in intestinal differentiation (see Discussion) are denoted X, and proposed relationships indicated with dashed grey lines.