Rich regulates target specificity of photoreceptor cells and N-cadherin trafficking in the Drosophila visual system via Rab6

Abstract

Neurons establish specific synaptic connections with their targets, a process that is highly regulated. Numerous cell adhesion molecules have been implicated in target recognition, but how these proteins are precisely trafficked and targeted is poorly understood. To identify components that affect synaptic specificity, we carried out a forward genetic screen in the Drosophila eye. We identified a gene, named ric1 homologue (rich), whose loss leads to synaptic specificity defects. Loss of rich leads to reduction of N-Cadherin in the photoreceptor cell synapses but not of other proteins implicated in target recognition, including Sec15, DLAR, Jelly belly, and PTP69D. The Rich protein binds to Rab6, and Rab6 mutants display very similar phenotypes as the rich mutants. The active form of Rab6 strongly suppresses the rich synaptic specificity defect, indicating that Rab6 is regulated by Rich. We propose that Rich activates Rab6 to regulate N-Cadherin trafficking and affects synaptic specificity.

Figure 1. Isolation of 3L6 mutants in an eyFLP screen

(A) Electroretinogram traces of control flies (CTL), mutant flies (3L61, 3L62), and mutant flies carrying a genomic rescue fragment (Rescue). The ERG profiles of the mutant flies show diminished “on and off” transients but normal depolarization in response to light. (B) TEM of cartridges from a control, 3L61, 3L62, and a rescued fly. The photoreceptor terminals are marked in blue. (C) Distribution of photoreceptor terminals per-cartridge for wild-type (n=31 cartridges from 3 animals), mutant (3L61: n=52 from 3 animals; 3L62: n=32 from 3 animals) and rescued flies (n=43 cartridges from 3 animals). (D) The R7 and R8 projection patterns (24B10 staining) of a control animal, a 3L61 mutant, a 3L62 mutant, a CadN^Δ14 mutant, a Liprin α^e mutant, and a 3L61 rescued fly. The arrows indicate the mistargeted terminals. The genotypes of flies used in this figure: control flies (y w eyFLP GMR-lacZ; FRT80B iso/FRT80B cl, ubiGFP); 3L61 (y w eyFLP GMRLacZ; FRT80B 3L61/FRT80B cl, ubiGFP); 3L62 (y w eyFLP GMRLacZ; FRT80B 3L61 or 3L62/FRT80B cl, ubiGFP); rescued flies: (y w; rich-gRE::VK37/CyO; FRT80B 3L61); CadN (y w eyFLP GMRlacZ; FRT40A CadN^Δ14/FRT40A ubiGFP) and Liprin α (y w eyFLP GMRlacZ; FRT40A Liprin α^e/FRT40A cl). (See also Figure S1)

Figure 2. The R7 cells exhibit targeting defects in the 3L6 mutant animals

(A, B, A′, B′) A portion (19.7±3%, n=268) of the R7 cells in the 3L61 mutant animals fail to target to the M6 layers but instead terminate at the M3 layers. The R7 cells were labeled with pan-R7 Gal4 that drives UAS-SytGFP (green). The arrows indicate the mistargeted R7s. (C, D, C′, D′) The R8 cells have intact target patterns. The R8 cells were labeled with Rh6-GFP (green), the photoreceptor cells were labeled with 24B10 staining (red). Panels (A′ –D′) are the enlarged views of the boxed regions in panel (A –D). The genotypes of flies used in this figure: Control (A, A′) (y w eyFLP GMRlacZ; PanR7 Gal4/UAS-SytGFP; FRT80B iso/FRT80B cl); 3L61 (B, B′) (y w eyFLP GMRlacZ; PanR7 Gal4/UAS-SytGFP; FRT80B 3L61/FRT80B cl); Control (C, C′) (y w eyFLP GMRlacZ; Rh6GFP/CyO; FRT80B iso/FRT80B cl). 3L61 (D, D′) (y w eyFLP GMRlacZ; Rh6GFP/CyO; FRT80B 3L61/FRT80B cl). (See also Figure S2)

Figure 3. 3L6 encodes an evolutionarily conserved gene CG9063 (rich)

(A) Schematic showing the cytological interval and the position of the P-elements and the overlapping deficiencies used to identify the genomic region of the 3L6. The recombination rate is 5/876 for BG02493 and 0/820 for BG01780. “−” indicates “fail to complement” and “+” indicates “complements” 3L61 and 2. The lower panel shows genes in the region identified by deficiency mapping and the genomic rescue fragment of rich (Red box). (B) Schematic of the Rich protein showing the predicted domains and the locations of the stop codons in 3L61 and 3L62. The bottom panel shows the homology (identity %; similarity %) of WD40 repeats domain, RIC1 domain, and overall protein sequence between Rich and its homologs in humans, frogs, worms, and yeast.

Figure 4. Rich is expressed in the developing eye

(A) Hrp (green) and HA (red) staining of an optic-lobe of a control animal (y w) at the 3^rd instar larval stage. The HA antibody does not detect non-specific staining in the brain. (B) The optic lobe of a rich mutant animal carrying a HA-Rich-genomic rescue construct (rescue animal: HA-Rich-gRE::VK37; FRT80B rich¹) at the 3^rd instar larval stage. At this stage, Rich is enriched in the developing lamina (LA) and medulla (ME). (C, D) The optic lobe of rescued animals at 15 hrs and 50 hrs after pupa formation (APF). Rich is enriched in the lamina and medulla neuropil. (E) The eye disc of the rescued animal at 50 hrs APF. Rich is expressed in the PR cells. (See also Figure S3).

Figure 5. rich is required for synaptic specificity of R7

Single optic section taken through a medulla of either (A, A′) control (ey3.5FLP; Act>Gal4 UAS-SytGFP/UAS-SytGFP; FRT80B M (3), tub-Gal80/FRT80B iso) or (B, B′) rich mutant (ey3.5FLP; Act>Gal4 UAS-SytGFP/UAS-SytGFP; FRT80B M (3), tub-Gal80/FRT80B rich¹) animals. (A′, B′) are the enlarged images of the boxed region in (A, B). Many of the mutant R7 cells fail to target to the M6 layers (arrows). (C) Quantification of the R7 targeting defects in the ey3.5 animals (0 in control n=159;74±16% in rich n=175) illustrated in A and B. (D) 62.5 ±22% (n=37) R7 cells fail to target to the correct layer even when they are surrounded by the wild-type R7 and R8 neighbors. The small inset illustrates the single R7 mutant clone (see also Figure S4).

Figure 6. Rich is enriched in Golgi apparatus and co-localizes with Rab6

(A–C; A′–C′) S2 cells expressing V5-tagged Rich were stained with anti-V5 (red) and different Golgi markers (green): GM130 (A, A′), SYX16 (B, B′), and 120KD protein (C, C′). (A′–C′) are the enlarged images of the boxed regions in image (A–C). Noticed that Rich is expressed in punctae that are co-localized or juxtaposed to the Golgi markers, indicating that Rich is enriched in Golgi apparatus. (D, D′) S2 cells expressing V5-tagged Rich and YFP tagged Rab6 were stained with anti-V5 (red) and anti-GFP (green) antibody. Rich and Rab6 are substantially co-localized in S2 cells. (D′) is the enlarged image of the boxed region in (D). (E, E′) Staining of salivary gland cells from a rescued animal (HA-Rich-gRE::VK37; FRT80B rich¹) with anti-HA (red: shows the pattern of Rich) and anti-120KD protein (green). Rich partially colocalizes with the 120KD protein, indicating Rich is enriched in the Golgi when it is expressed at the endogenous level. (E′) is the enlarged image of the boxed region in (E). (F, F′) Rich is enriched in the Golgi of photoreceptor cells. Eye disc of a rescued animal (HA-Rich-gRE::VK37; FRT80B rich¹) at 24 hrs after pupal formation was stained with anti-HA (red), anti-120KD (green). (F′) The colocalized signals of HA and 120KD protein shown in (F) were analyzed with the ImageJ and indicated with a pseudo-color (white).

Figure 7. Rich interacts with Rab6 and positively regulates Rab6 activity in fly eyes

(A) The ERG profiles of control animals (CTL) (y w eyFLP GMRlacZ; FRT40A iso/FRT40A cl) and Rab6 mutant animals (y w eyFLP GMRlacZ; FRT40A Rab6^D23D/FRT40A cl). Rab6 mutants have no “on” and “off” transients but exhibit a normal depolarization. (B, B′, C, C′) A portion (18.8±3%, n=265) of the R7 cells in the Rab6 mutant animals (y w eyFLP GMRlacZ; FRT40A Rab6^D23D/FRT40A cl; PanR7 Gal4/UAS-SytGFP) fail to target to the M6 layers as the controls (y w eyFLP GMRlacZ; FRT40A iso/FRT40A cl; PanR7 Gal4/UAS-SytGFP), but instead terminate at the M3 layers. The R7 cells were labeled with pan-R7 GFP (Green). The arrows indicate mistargeted R7s. (B′, C′) are the enlarged images of the boxed regions in image (B, C). (D) Quantification of the mistargeting R7s in Rab6 mutants. (E–H) Loss of one copy of Rab6 enhanced the phenotype of the hypomorphic allele rich² but not the phenotype of the null allele rich¹, indicating that Rich and Rab6 function in the same pathway. Single optic sections of medulla in rich¹ (E) (y w eyFLP GMRlacZ; PanR7 Gal4/UAS-SytGFP; FRT80B rich^1/FRT80B cl); rich2 (F) (y w eyFLP GMRlacZ; PanR7 Gal4/UAS-SytGFP; FRT80B rich^2/FRT80B cl); rich² with one copy of Rab6 (G) (y w eyFLP GMRlacZ/UAS-SytGFP; PanR7 Gal4/FRT40A Rab6^D23D; FRT80B rich^2/FRT80B cl), and rich¹ with one copy of Rab6 (H) (y w eyFLP GMRlacZ/UAS-SytGFP; PanR7 Gal4/FRT40A Rab6^D23D; FRT80B rich^1/FRT80B cl). The R7 cells were labeled with pan-R7 GFP (green) and 24B10 staining indicates the projection patterns of the photoreceptor cells. (I–J) Expression of Rab6CA but not Rab6DN suppresses the targeting phenotypes of rich¹ mutant, indicating Rich positively regulate Rab6. Single optic sections of medulla in control animals (I) (y w eyFLP; Act>Gal4 UAS-SytGFP/UAS-SytGFP; FRT80B M (3), tub-Gal80/FRT 80B iso), rich¹ mutants (J) (y w eyFLP; Act>Gal4 UAS-SytGFP/UAS-SytGFP; FRT80b M (3), tub-Gal80/FRT 80b rich¹), rich¹ mutants in which the constitutively active form of Rab6 (Rab6CA) is expressed in the mutant cells (y w eyFLP/UAS-YFPRab6 Q71L; Act>Gal4 UAS-SytGFP/CyO; FRT80B M (3), tub-Gal80/FRT 80B rich¹), and rich¹ in which the dominant negative form of Rab6 (Rab6DN) is expressed in the mutant cells (eyFLP; Act>Gal4 UAS-SytGFP/UAS-YFPRab6 T26N; FRT80B M (3), tub-Gal80/FRT 80B rich¹). Projection patterns are shown with 24B10 staining. The arrows indicate the mistargetted R cells. (M) Quantification of the R7 targeting phenotypes in the mutants illustrated in (E–H). 19.7±3% (n=268) R7 cells in rich¹ mutant (E); 8.5±2% (n=329) R7 cells in rich² mutants (F) fail to target to the correct layer. Upon removing one copy of Rab6, 19.0±3% (n=308) R7 cells in rich¹ (H) and 17.0±3% (n=483) R 7 cells in rich² (G) fail to target to the correct layers. (N) Quantification of the R7 targeting phenotypes in the flies illustrated in (I–L). In rich¹ mutants (J), 20.24±2% (n=252) R7 cells fail to target to the corresponding layer. In the rich¹ mutant with Rab6CA expression animals (K), 7.69±3% (n=296) R7 cells fail to target to the correct layer. In the rich¹ mutant with Rab6DN expression animals (L), 20.8±4% (n=233) R7 cells fail to target to the correct layer. (O) GST pull-down experiments show that both the RIC1 domain and the WD40 domain interact with Rab6 physically. The RIC1 and WD40 domain have high affinity for the wild-type Rab6 and Rab6DN but have low affinity for the Rab6CA. 1/10 input of the RIC1 domain and WD40 domain were loaded as control (see also Figure S5).

Figure 8. Rich, Rab6, and CadN function in a common pathway

(A–F′) N-Cadherin is reduced in the lamina of rich and Rab6 mutants. Single optic slides of lamina of the fly eye at 24h APF stained with anti-CadN (red), 24B10 (blue), and GFP (green, mark the mutant cells). (A–C, A′–C′) show a side view of the lamina. (D–F, D′–F′) show a frontal view of the lamina. (A, A′, D, D′) control (y w eyFLP; Act>Gal4 UAS-SytGFP/UAS-SytGFP; FRT80B M (3), tub-Gal80/FRT 80B iso). (B, B′, E, E′) rich¹ mutant (y w eyFLP; Act>Gal4 UAS-SytGFP/UAS-SytGFP; FRT80b M (3), tub-Gal80/FRT 80B rich¹). (C, C′, F, F′) Rab6 mutant (y w eyFLP; FRT40A tub-Gal80/FRT40A Rab6^D23D; tubGal4/UAS-SytGFP). (A′–F′) are red channels (CadN) of (A–F). (G–H′) CadN distribution is not affected in rich1 mutant cone cells. The pupal eye discs at 36APF were stained with anti-GFP (green) and anti-CadN (red) antibody. The mutant cells were marked with CD8GFP. The animal genotypes are: CTL (y w eyFLP; Act>UAS-SytGFP/UAS-CD8GFP; FRT80B M (3), tub-Gal80/FRT80B iso). rich¹ (y w eyFLP; Act>UAS-SytGFP/UAS-CD8GFP; FRT80B M (3), tub-Gal80/FRT 80B rich¹). (I) The R7 cell targeting patterns of rich¹ and rich² upon removal of one copy of CadN (y w UAS-SytGFP; FRT40A CadN^405/CyO; FRT80B rich¹/TM6b; and y w UAS-SytGFP; FRT40A CadN^405/CyO; FRT80B rich^2/TM6B). Compared with Figure 7(E) and (F), removal of one copy of CadN enhances the R7 targeting phenotypes of the hypomorphic allele rich² but not the phenotypes of the null allele rich¹, indicating Rich and CadN function in a same pathway to regulate R7 targeting. (J) The quantification of R7 targeting phenotypes of rich¹ mutants (Figure 7E), rich² mutants (Figure 7F) and mutants of rich¹ and rich² with one copy of CadN. Upon removal of one copy of CadN, 19.6 ± 2% (n=384) of R7 cells in rich¹ and 19.0 ± 2 % (n=493) of R7 cells in rich² mutants fail to target to the correct layer. (K) The CadN, rich¹ double mutants have similar eye targeting phenotypes as CadN single mutants. The genotypes are: (CTL) y w eyFLP;; FRT80B cl ubi-GFP/FRT80B iso. (rich¹): y w eyFLP;; FRT80B cl ubi-GFP/FRT80B rich¹. (CadN): y w eyFLP; FRT40A CadN^Δ14/FRT40A ubi-GFP. (CadN; rich¹): y w eyFLP; FRT40A rich-gRE::VK37/FRT40A CadN^Δ14; FRT80B rich¹/3L6 deletion. (L) Quantification of mistargeted R7 of control flies (CTL): 0%, n=458; rich¹ mutants, 18.9± 3.9%, n=450; CadN mutants, 24.6±2.2%, n= 244; CadN; rich¹ mutants: 24.4±1.6%, n=237. (See also Figure S6–S8)