Fig 1.
Drosophila retinal structure and expression of CC-restricted prosPSG-GAL4.
(A) Diagram of an individual eye unit (ommatidium) from the Drosophila compound eye, highlighting CC (green)-PR (blue) interfaces throughout the retina. CCs cap the rhabdomeres distally, send interretinular fibers that contact the PR cell bodies, and form a pore at the base of the retina, around which the PR axons exit the retina. Pigment cells (PC, orange) form a fenestrated membrane which separate the retina from the brain, and through which PR axons exit the eye. Epithelial glia in the optic lobe (purple, brain) that support photoreceptor neurotransmitter recycling are also indicated. (B) CC expression of a prosPSG-GAL4 driven reporter (UAS-GFP, green) during larval, pupal, and adult stages. Elav (blue) is included in the larval stage to highlight the formation of CCs after neurogenesis in the eye imaginal disc (oriented left/anterior-to-right/posterior [youngest to oldest ommatidia]). The diagram represents the strategy of the current study, in which the prosPSG-GAL4 driver was used to express RNAi constructs to genetically test the role of CCs in regulating PR morphology or function and to isolate CCs for molecular (transcriptomic) analysis.
Fig 2.
Histological and electrophysiological evidence for CCs providing structural and functional support for Drosophila photoreceptors.
(A-F) Toluidine blue-stained coronal (A,C,E) and sagittal (B,D,F) thin sections from adult control (prosPSG-GAL4>UAS-nGFP, A, B), prosRNAi (prosPSG-GAL4>UAS-prosRNAi, C, D), and Pax2RNAi (prosPSG-GAL4>UAS-dPax2RNAi, E,F) eyes. In B,D,and F, the blue dashed lines highlight rhabdomeres and filled arrow indicates the fenestrated membrane. F’ magnifies a subretinal region in Pax2RNAi heads, highlighting relatively intact PR clusters (G-I) CCs (marked with Fas3, green, arrows) remain associated with PRs (marked with phalloidin, magenta) in both control and dPax2RNAi flies, even though many PRs have fallen into the brain in dPax2RNAi flies (I, I’). The interommatidial bristle (*) also expresses Fas3 (H) in adult eyes. (J-M) Averaged ERGs from dark-adapted control (prosPSG-GAL4; UAS-nGFP), prosRNAi (prosPSG-GAL4; UAS-prosRNAi), and Pax2RNAi (prosPSG-GAL4; UAS-dPax2RNAi) excited with 5x5sec pulses of 490 nm (blue-green) light (n = 5 animals). Standard deviation is shown by the shaded area, and day-matched control recordings are included in grey for each experimental condition. (N) Averaged PR responses from J-M, normalized to day-matched controls. Ccontrol, sev14, and dPax2RNAi flies are indistinguishable, while prosRNAi flies have significantly reduced PR responses (*p<0.001). (O) Control or prosRNAi flies raised for 7 days in dark or light and PR structure were analyzed by toluidine blue stained semi-thin plastic sections, showing a light-dependent degeneration only in prosRNAi flies raised under continuous light. (P) DPP presence was quantified daily for 7 days from control or prosRNAi flies raised in continuous light (* p<0.001 by day 3, n = 3 replicates of 25 flies each). Error bars = S.D.
Fig 3.
Gliogenic genes are expressed and required in Drosophila cone cells for photoreceptor activity.
(A) TMM-normalized log2 expression levels of transcription factors frequently associated with fly and vertebrate glial specification, plotted from larval, pupal and adult CCs (green) and adult PRs (blue). (B) Relative PR activity (normalized to day-matched controls, see Methods and S4A and S4B Fig) from ERG recordings of prosPSG-GAL4 > nuclear GFP (con), and knockdowns using lines for prosPSG-GAL4 driven gcmRNAi-1, repoRNAi-1, and OliRNAi-1 lines indicate that all three factors are required in CCs for full PR activity. (C) Relative PR activity (solid bars) and relative “on” transient measurements (striped bars) from heterozygous or homozygous repo1 animals, or prosPSG- (CC-), repo- (glia-), and otd1.6- (PR-) GAL4 lines driving gcmRNAi-1, repoRNAi-1, OliRNAi-1, and the Rh1/ninaERNAi-1 flies (n = 5 flies, 5x5sec light flashes each). “on” transients were calculated as the ratio of the “on” transient strength to the maximal PR response (see Methods and S4C Fig). N.B. We did not detect the early reversed polarity phenotype initially reported in very young repo1 flies (<1 d old [78]), but instead observed the phenotypes observed with older repo1 flies [97]. (D) Relative “on” transient size from each of 5 light pulses from control and pros>OliRNAi-1 flies tested for significance by 1-way ANOVA. Error bars = SEM, * p<0.001.
Fig 4.
Transcriptome comparisons with Drosophila cone cells/photoreceptors and glial/neuronal cell populations.
(A) Heatmap of TMM-normalized log2 expression levels from larval, pupal and adult CCs for 109 genetically-identified factors involved in fly glial differentiation (http://www.sdbonline.org/sites/fly/aimorph/glia.htm) [98]. Examples of genes that cluster based on different expression dynamics or levels are indicated by brackets. (B) Relative overlap of genes present in different cell and tissue types [1000 genes with highest relative expression] in larval, pupal, and adult CCs (green), PRs (blue), mature nervous system (purple), and mature digestive system (orange)] with 109 “glial” genes [98] or 2309 gcm downstream factors [those deregulated in both loss- and gain-of-function gcm experiments [84]]. Separate analysis (right panels) using the 1000 most CC- vs. PR-enriched sets (CC>PR, PR>CC) shows specific enrichment for glial genes in CCs but not PRs. Enrichment = actual number of overlapping genes/expected number of overlapping genes; *p<0.02, ** p<0.002. (C) Relative enrichment values (as in B) for Drosophila CC vs PR-enriched genes analyzed for overlap with murine retinal and forebrain neuronal and glial cell types. (D) Venn diagram representing commonly associated glial genes between cone cells, Müller glia, astrocytes, and oligodendrocytes.
Fig 5.
Functional dependence of cone cell-expressed glial effectors on photoreceptor physiology.
(A) Table of 12 conserved glial genes tested for CC-dependent PR activity support functions. (B) TMM-normalized RNA expression levels of candidate genes from (C) in adult cone cells. (C) Relative PR activity (normalized to day-matched controls, +/- SEM) for CC knockdowns with prosPSG-GAL4 (*p<0.001). Eight show significant reductions in PR activity, including those that are typically associated ion maintenance (AtpαRNAI-1, nrv2RNAI-1, nrv3RNAi-1 and Irk2RNAi-1), metabolism (dLDHRNAi-1 and Glut1RNAi-1) and histamine and glutamate recycling (pyd3RNAi-1 and EAAT1RNAi-1). Four (Irk1RNAi-1, Irk3RNAi-1, EAAT2RNAi-1 and GS2RNAi-1) were statistically similar to controls. (D) Schematic of the glial factors expressed in and required by CCs for supporting retinal integrity and function. The transient expression of gcm, repo, and Oli are reflected by lighter font. Parentheses indicate glial factors with abundant mRNA expression in CCs, but whose function was not defined here. The role of epithelial glia in neurotransmitter recycling, a synaptic event that occurs in the brain and which we found to be genetically separable from CCs, is also noted.