The WAVE complex drives the morphogenesis of the photoreceptor outer segment cilium

Abstract

The photoreceptor outer segment is a modified cilium filled with hundreds of flattened "disc" membranes responsible for efficient light capture. To maintain photoreceptor health and functionality, outer segments are continuously renewed through the addition of new discs at their base. This process is driven by branched actin polymerization nucleated by the Arp2/3 complex. To induce actin polymerization, Arp2/3 requires a nucleation promoting factor. Here, we show that the nucleation promoting factor driving disc morphogenesis is the pentameric WAVE complex and identify all protein subunits of this complex. We further demonstrate that the knockout of one of them, WASF3, abolishes actin polymerization at the site of disc morphogenesis leading to formation of disorganized membrane lamellae emanating from the photoreceptor cilium instead of an outer segment. These data establish that, despite the intrinsic ability of photoreceptor ciliary membranes to form lamellar structures, WAVE-dependent actin polymerization is essential for organizing these membranes into a proper outer segment.

Keywords: Arp2/3; WAVE complex; actin cytoskeleton; outer segment; photoreceptor.

Fig. 1.

The pentameric WAVE complex is the nucleation promoting factor in photoreceptor outer segments. (A) Specific protein isoforms of the WAVE complex identified by mass spectrometry in immunoprecipitates from lysed purified mouse rod outer segments using either anti-WASF3 or anti-ABI1 antibodies. Proteins included in the table had at least a twofold enrichment over control immunoprecipitates with anti-IgG antibodies. See Dataset S2 for the full list of identified proteins in each experiment. (B) Western blot showing co-precipitation of BRK1 with WASF3 from lysed rod outer segments using the anti-WASF3 antibody. Immunoblotting for rhodopsin kinase, an outer segment protein not associated with the WAVE complex, was used as a negative control. Whole retinal lysate was used as a reference. See also SI Appendix, Fig. S2. (C) A ribbon diagram of the WAVE complex labeled with the protein isoforms identified in the photoreceptor outer segment. The structure of the WAVE complex was determined by X-ray crystallography (23) and the ribbon diagram downloaded from World Wide Protein Data Bank (DOI: 10.2210/pdb3P8C/pdb).

Fig. 2.

Subunits of the WAVE complex, including WASF3, are highly expressed in photoreceptor cells. (A) mRNA levels of all known nucleation promoting factors expressed in mouse photoreceptor cells. The values are quantified as fragments per kilobase million (FPKM). The data were obtained from a published RNAseq dataset of flow-sorted rod cells isolated at various time points between P2 and P28 (24). Raw data were acquired from NCBI Gene Expression Omnibus series GSE74660. (B) A comparison of the mRNA expression profiles for Wasf3, three outer segment proteins (Cnga1, Cngb1 and Rom1), actin (Actb) and a housekeeping gene (Hsc70). For each gene, the data were normalized to their maximal expression level at any time point. The data in both panels are shown as mean ± SD.

Fig. 3.

F-actin is specifically lost from the site of disc formation in Wasf3^−/− mice. (A) A retinal cross section, centered at the site of disc formation, from a P21 WT mouse immunostained with anti-WASF3 antibody (green). The sections were co-stained with phalloidin to label F-actin (red) and WGA to label outer segments (magenta). Arrowheads denote F-actin puncta present at the site of disc formation that overlap with distinct puncta of WASF3. (Scale bar, 5 µm.) (B) WASF3 staining (green) overlapping with F-actin (red) at the outer segment base of a cone labeled with PNA (magenta). (Scale bar, 5 µm.) (C) Retinal cross sections, centered at the site of disc formation, of P21 WT and Wasf3^−/− mice stained with phalloidin to label F-actin (red). (Scale bars, 10 µm.) (D) Whole retinal cross sections of P21 WT and Wasf3^−/− mice stained with phalloidin. The outer nuclear layer (ONL), outer plexiform layer (OPL), inner nuclear layer (INL), inner plexiform layer (IPL) and ganglion cell layer (GCL) are labeled. (Scale bars, 20 µm.) See also Movies S1 and S2.

Fig. 4.

WASF3 knockout results in normal protein trafficking but disrupted outer segment morphology. (A) Retinal cross sections of P21 WT and Wasf3^−/− mice stained with antibodies against indicated outer segment proteins (green). Nuclei were stained with Hoechst (blue). (B) Expanded view of the outer segment layer in retinal cross sections of P21 WT and Wasf3^−/− mice stained with anti-ROM1 antibodies. (C) Expanded view of the outer segment layer in retinal cross sections of P21 WT and Wasf3^−/− mice stained with fluorophore-conjugated PNA. (D) Western blot for three representative outer segment proteins from whole retinal lysates of WT and Wasf3^−/− mice at P21. Three mice of each genotype were analyzed. Each lane was loaded with 20 µg total protein. HSC70 serves as a loading control. Scale bars in all panels are 10 µm.

Fig. 5.

Wasf3^−/− mice form disorganized ciliary membrane layers rather than outer segments. Electron micrographs of photoreceptor outer segments from WT mice or ciliary membrane layers from Wasf3^−/− mice at indicated ages. Retinal sections were contrasted with tannic acid/uranyl acetate to discern disc membranes exposed to the extracellular space (darkly stained membranes) from those enclosed within the cell (lightly stained membranes). (Scale bars, 1 µm.)

Fig. 6.

Photoreceptor cells of Wasf3^−/− mice progressively degenerate. Thin plastic-embedded cross sections of WT and Wasf3^−/− mouse retinas stained by Toluidine blue and imaged by light microscopy. Mice were analyzed at P21 (A), P45 (B) and P200 (D). (Scale bars, 20 µm.) The quantification of photoreceptor degeneration is represented by spider diagram graphs showing the number of photoreceptor nuclei counted at P45 (C) and P200 (E) from eight 100 µm-wide segments of the section taken at 500 µm intervals from the optic nerve.

Fig. 7.

The early knockout of Arp2/3 disrupts retinal development while preserving the ability of photoreceptors to produce ciliary membrane layers. (A) Thin plastic-embedded cross sections of P17 WT and Six3Cre/ArpC3^f/f mouse retinas stained by Toluidine blue and imaged by light microscopy. (Scale bars, 20 µm.) (B) Electron micrographs of photoreceptor ciliary membrane layers from P17 Six3Cre/ArpC3^f/f mice. (Scale bars, 1 µm.)

Fig. 8.

Cartoon illustrating the role of the WAVE complex in initiating photoreceptor disc formation and functions of proteins implicated in membrane bending and trans-membrane adhesion. The WAVE complex cooperates with Arp2/3 to drive actin polymerization evaginating each nascent disc. A more detailed illustration of WAVE-Arp2/3-driven actin polymerization is presented in SI Appendix, Fig. S1. Extended oligomeric complexes of peripherin-2/ROM1 support the highly curved edges along the entire circumference of mature discs. The leading edges of expanding nascent discs are enriched with prominin-1, which is thought to promote membrane bending at this location. Homotypic rhodopsin interactions have been hypothesized to provide membrane-to-membrane adhesion across the extracellular space between nascent discs and the intradiscal space of mature discs. PDE6 and GARP2 have been hypothesized to connect discs across the cytoplasm throughout the disc surface and near the disc rim, respectively. Proteins whose role in membrane bending or adhesion remains hypothetical are denoted with a question mark. See Discussion for details.