Morphogenesis of neurons and glia within an epithelium

Abstract

To sense the outside world, some neurons protrude across epithelia, the cellular barriers that line every surface of our bodies. To study the morphogenesis of such neurons, we examined the C. elegans amphid, in which dendrites protrude through a glial channel at the nose. During development, amphid dendrites extend by attaching to the nose via DYF-7, a type of protein typically found in epithelial apical ECM. Here, we show that amphid neurons and glia exhibit epithelial properties, including tight junctions and apical-basal polarity, and develop in a manner resembling other epithelia. We find that DYF-7 is a fibril-forming apical ECM component that promotes formation of the tube-shaped glial channel, reminiscent of roles for apical ECM in other narrow epithelial tubes. We also identify a requirement for FRM-2, a homolog of EPBL15/moe/Yurt that promotes epithelial integrity in other systems. Finally, we show that other environmentally exposed neurons share a requirement for DYF-7. Together, our results suggest that these neurons and glia can be viewed as part of an epithelium continuous with the skin, and are shaped by mechanisms shared with other epithelia.

Keywords: Amphid; C. elegans; Dendrites; Glia; Neurodevelopment; Sensory epithelia.

Fig. 1.

Amphid neurons and glia exhibit tight junctions and distinct apical and basolateral surfaces. (A-C) Schematics showing relative positions of (A) amphid dendrite endings, (B) tight junctions (purple), and (C) apical (green) and basolateral (red) surfaces. For simplicity, only a single neuron is drawn. Neuron (Ne), pink; sheath (Sh), blue; socket (So), yellow; hypodermis (H, skin), gray. (D) Cell-specific markers were used to visualize single neurons and glia. ASER neuron (gcy-5pro, red); sheath glia (F16F9.3pro, blue); socket glia (grl-2pro, yellow). (E,F) Overlap of neuron:sheath and sheath:socket tight junctions was visualized by using these promoters to express AJM-1-YFP in the sheath and cytoplasmic mCherry plus AJM-1-CFP in (E) the neuron or (F) the socket. (D′,E′,F′) Dotted line in schematics indicates approximate outline of sheath. Images are higher magnifications of the boxed areas in D,E,F; purple arrowheads indicate the position of overlapping tight junctions. (G-I) The apical and basolateral markers ApiGreen (Ap, green) and BasoRed (Ba, red) (see Fig. S2) were expressed in (G) the neuron, (H) the sheath and (I) the socket, using the same promoters. Boxed regions in G-I are magnified in G'-I′. jxn, junction; Ap, ApiGreen; Ba, BasoRed.

Fig. 2.

DYF-7 is cleaved in vivo and its ectodomain localizes to apical surfaces. (A) Cartoon schematic of an embryo showing a single neuron (blue; cell body is marked with an asterisk) with its dendrite extending to the developing nose (dendrite ending, arrowhead). Topology of the DYF-7 reporter construct, consisting of the ZP domain (ZP), superfolder GFP tag (sfGFP), consensus furin-family cleavage site (CFCS), transmembrane segment (TM) and cytoplasmic region tagged with mCherry (mCherry). (B) A live, intact embryo expressing this construct in sensory neurons (dyf-7pro) during dendrite extension (asterisks and arrowheads indicate cell body and dendrite ending). (C,D) Embryos expressing sfGFP-DYF-7 in sensory neurons (dyf-7pro) (C) with myristyl-mCherry or (D) with immunostaining against AJM-1. Some structures appear as short linear segments (arrow) rather than rounded caps. (E-G) Embryos misexpressing sfGFP-DYF-7 in gut (pha-4pro) with myristyl-mCherry (E), ubiquitously (heat shock promoter) with myristyl-mCherry (F) or with immunostaining against AJM-1 (G). Brackets mark pharynx (p) and gut (g). Scale bars: 10 µm in main panels; 2 µm in magnified views.

Fig. 3.

DYF-7 forms extracellular fibrils. (A) Image of a wild-type embryo after high-pressure freezing, fixation and flat embedding. Flat embedding allowed selection of embryos at the desired stage and orientation for sectioning. Serial sectioning led to identification of the (i-iii) left- and (iv-vi) right-hand amphids. Three serial sections of each amphid are shown. Neurons, red; sheath, blue; socket, yellow. Basal bodies/centrioles are annotated in iii with arrowheads. Extracellular fibrils are marked with brackets. (B) Three serial sections (vii-ix) of an amphid in a dyf-7 embryo at the same developmental stage. The sheath forms a pocket around the neurons and no fibrils are visible. Basal bodies/centrioles appear normal (arrowheads). (C) Drosophila S2 cells were transfected with sfGFP-DYF-7. A representative field of cells is shown. GFP-labeled structures appeared as (i) a plume at the cell surface, (ii) complex tangles, (iii) puncta and fibrils that appear to encircle the cells, and (iv,v) small webs.

Fig. 4.

Disruption of the amphid glial channel in dyf-7 mutants. (A) Schematic showing the relative positions of the socket (So), sheath (Sh) and neurons (Ne) in a dyf-7 mutant. A posterior process from socket to sheath developed in 16/48 amphids. Regions of interest are boxed: (i) socket:skin junction, shown in B-C; (ii) sheath:socket junction, shown in D-E; (iii) neuron:sheath junction, shown in F-H. (B-F″) Three-color merged images showing tight junctions: AJM-1-YFP expressed in sheath (B″,D″,F″) plus AJM-1-CFP (B′,D′,F′) and cytoplasmic mCherry expressed in (B,D) socket or (F) neuron. Purple arrowheads indicate positions of relevant tight junctions; asterisks indicate cell bodies. (C,E,G,H) ApiGreen and BasoRed expressed in (C,E) socket, (G) neuron or (H) sheath. All markers and promoters are the same as in Fig. 1.

Fig. 5.

Loss of FRM-2 enhances weak dyf-7 defects. (A) Schematic of genetic strategy used to isolate enhancers of the hypomorphic missense mutant dyf-7(ns117). Parental (P0) strain exhibits ∼35% short amphids (red) and ∼65% full-length amphids (green). Following mutagenesis, clonal third-generation (F3) progeny bearing additional mutations (mut) were visually screened for enhanced penetrance. (B) Schematic of the FRM-2 protein showing location of hmn162 nonsense mutation. FRM-2 is roughly 50-60% identical to human EPB41L5, zebrafish moe and Drosophila Yurt throughout the FERM B-lobe (FB), FERM C-lobe (FC) and FERM-adjacent (FA) domains, with a more divergent C-terminal sequence. (C) frm-2(hmn162) enhances the amphid dendrite extension defects of dyf-7(ns117). frm-2(+) transgene indicates animals bearing a transgene with a wild-type frm-2 genomic fragment. n≥100 amphids per genotype. Data are mean±s.e.m. (D) frm-2pro drives expression in sensory neurons (dyf-7pro, arrowheads) at the time of dendrite extension, as well as bright expression in pharynx and gut. (E) A rescuing FRM-2-GFP construct expressed by its endogenous promoter localizes to developing dendrite endings (arrow).

Fig. 6.

Sensory neurons within epithelia exhibit a shared dependence on DYF-7. (A,B) Cell-specific markers were used to visualize sensory neurons within epithelia – amphid (AWC, odr-1pro), CEP (dat-1pro), OL (OLQ, ocr-4pro) and IL (IL2, klp-6pro) – as well as URY (tol-1pro), which is not within an epithelium, in (A) wild type and (B) dyf-7(ns119) null mutants. Arrowheads indicate position of dendrite endings. (C) Quantification of dendrite lengths as a fraction of nose length for each neuron in wild-type (left column) and dyf-7 (right column) animals. Colored bars represent individual dendrites (n≥48 per column); black bars represent means. Shaded areas represent the wild-type mean±5 standard deviations (s.d.) for each neuron type; the percentage of dendrites in this range (‘full-length dendrites’) is given.