ER complex proteins are required for rhodopsin biosynthesis and photoreceptor survival in Drosophila and mice

Defective rhodopsin homeostasis is one of the major causes of retinal degeneration, including the disease Retinitis pigmentosa. To identify cellular factors required for the biosynthesis of rhodopsin, we performed a genome-wide genetic screen in Drosophila for mutants with reduced levels of rhodopsin. We isolated loss-of-function alleles in endoplasmic reticulum membrane protein complex 3 (emc3), emc5, and emc6, each of which exhibited defective phototransduction and photoreceptor cell degeneration. EMC3, EMC5, and EMC6 were essential for rhodopsin synthesis independent of the ER associated degradation (ERAD) pathway, which eliminates misfolded proteins. We generated null mutations for all EMC subunits, and further demonstrated that different EMC subunits play roles in different cellular functions. Conditional knockout of the Emc3 gene in mice led to mislocalization of rhodopsin protein and death of cone and rod photoreceptor cells. These data indicate conserved roles for EMC subunits in maintaining rhodopsin homeostasis and photoreceptor function, and suggest that retinal degeneration may also be caused by defects in early biosynthesis of rhodopsin.


Introduction
Retinitis pigmentosa (RP) causes the gradual degeneration of rod and cone photoreceptors and is the most common form of hereditary retinal diseases [1,2]. Mutations in the light receptor gene rhodopsin account for approximately 25% of all ADRP cases. Over 120 point mutations in the rhodopsin gene have been associated with ADRP, with most disrupting its trafficking from the endoplasmic reticulum (ER) to the plasma membrane [3]. Proper folding and targeting of rhodopsin is crucial for maintaining the structure, function, and homeostasis of photoreceptors [1].
Studying photoreceptor cells in Drosophila has provided insights into the folding and transport of rhodopsin protein [4,5]. In flies, folding of rhodopsin is tightly regulated by multiple ER chaperons, including NINAA, calnexin, and Xport [6][7][8]. Without these chaperons the rhodopsin protein fails to properly mature, leading to retinal degeneration, as seen with mutations in rhodopsin itself [7,9]. RanBP2 is a cyclophilin-related protein that acts as a chaperone for cone opsins, but many other chaperons required for rhodopsin folding are not well conserved in mammals [1,10]. In fact, no mammalian homologs have been identified for any Drosophila gene involved in rhodopsin biogenesis in the ER that cause retinal degeneration when mutated.
The ER membrane protein complex (EMC) was first identified as a multi-protein transmembrane complex in a genetic screen for the accumulation of misfolded membrane proteins in yeast. Based on a high-content proteomics strategy, this complex was then found to interact with the ER-associated degradation (ERAD) pathway, suggesting a role in the biosynthesis of transmembrane proteins [11,12]. and lung development [15][16][17][18][19][20][21][22]. Mutation of the zebrafish emc3 gene, partial optokinetic response b (pob), causes degeneration of long-wavelength photoreceptor cells, and Drosophila EMC3 is required for the stable production of rhodopsins [23,24]. Moreover, a point mutation in the human EMC1 gene has been associated with retinal dystrophy [25]. However, it has not yet been determined which steps in rhodopsin production require EMC components, and whether all EMC subunits are required for maintaining rhodopsin homeostasis and/or photoreceptor cell integrity.
In the present study, we conducted an ethyl methanesulfonate (EMS)-based genetic screen to isolate mutations that affect rhodopsin homeostasis. We identified mutations in 3 EMC subunits, emc3, emc5, and emc6 that each reduced levels of the major rhodopsin Rh1, disrupted phototransduction, and caused gradual photoreceptor degeneration. We found that EMC function was independent of ERAD, most likely regulating rhodopsin levels at an earlier step in the biosynthetic process. Furthermore, knocking out the Emc3 gene in mammalian rod photoreceptor cells led to dysregulated rhodopsin trafficking and retinal degeneration, as seen in flies, suggesting that EMC function is conserved in mammals. Our study highlights the different roles played by EMC proteins in rhodopsin biogenesis.

Generation of transgenic flies
The emc3, emc5, and emc6 cDNAs were amplified from the cDNA clones LD37839, RE09053, and RE35789, respectively, which were obtained from the Drosophila Genomic Resource Center. To express emc3 and emc5 under control of the ninaE (ninaE: neither inactivation nor afterpotential E) promoter, the cDNAs were subcloned into the pninaE-attB vector between the NotI and XbaI sites, and the constructs were injected into M(vas-int.Dm)ZH-2A;M (3xP3-RFP.attP)ZH-86Fb embryos. To express emc6 in photoreceptor cells, the cDNA was subcloned into the pninaE-attB vector between the NotI and XbaI sites, followed by injected into M(vas-int.Dm)ZH-2A;M(3xP3-RFP. attP) ZH-51C. The transformants were identified on the basis of eye color, and mini-white marker was eliminated by crossing to a Cre-expressing line.
To generate emc2A RNAi and emc5 RNAi flies, the following 21-nt sequences were used to generate shRNA constructs, as described.

ERG recordings
ERG recordings were performed as described [29]. Briefly, glass microelectrodes filled with Ringer's solution were placed on the compound eye and the thorax of a fly. After a brief dark adaptation period, white-eyed flies were exposed to 5 s of orange light, followed by two rounds of 5 s of blue light, and two rounds of 5 s of orange light. The time between stimulations was 5 s. ERG signals were amplified with a Warner electrometer IE-210, and recorded with a MacLab/4 s analog-to-digital converter and the clampelx 10.2 program (Warner Instruments, Hamden, USA).

Immunostaining
Retina were dissected and fixed in 4% freshly made paraformaldehyde in phosphate buffer for 2 h on ice, then dehydrated in a series of ethanol dilutions (10,25,40,55,75, 90, 100%), 30 min each. Samples were then treated in acetone solution, and embedded in LR white resin (Sigma, St Louis, MO). Sections (1 μm) were then prepared for immunostaining. Samples were blocked using PBST (0.3% Triton X-100) with 5% goat serum for 1 h, then incubated with primary antibodies at 4°C overnight. This was followed by incubation with secondary antibodies for 1 h. Samples were examined and images were recorded using a Nikon A1-R confocal microscope, and further processed using Photoshop CC2017 software.

Co-immunoprecipitation
Flies were kept in a 12h-light-12h-dark cycle at 25°C. Approximately 200 flies were collected 5 days after eclosion. The flies were frozen in liquid nitrogen to dissociate the head from the body. Heads were then sieved out and lysed using lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% NP40, 1 × proteasome inhibitor) at 4°C overnight. They were then incubated with RFP agarose beads (Chromotek) for 2 h at 4°C. After several washes, the beads were boiled in SDS loading buffer, and supernatant was loaded onto a gel. Sliver staining (Thermo Fisher Scientific) was performed and bands on the gel were cut out for further mass spectra analysis.

Proteomic analysis of the fly retina
Flies were maintained in a 12 h-light/12h-dark cycle at 25°C . Sixty female flies of each genotype were collected 5 days after eclosion. For each sample, 30 pairs of retinas were dissected in cold phosphate buffer. Two samples were generated for each genotype. Protein extraction was performed as described with several modifications [31]. Briefly, tissues were boiled and sonicated in lysis buffer (6 M guanidinium hydrochloride, 20 mM tris (2-carboxyethyl) phosphine, 50 mM chloroacetamide, 50 mM Tris, pH 8.5), followed by 20 min 18,000 g centrifugation to collect supernatant. About 50 µg protein for each sample was digested with LysC (Promega Corporation, Madison, USA) in an enzyme/protein ratio of 1:100 (w/w) for 4 h. The sample was further digested with trypsin (Promega Corporation, Madison, USA) in an enzyme/protein ratio of 1:50 (w/w) overnight at 37°C. Finally, the resulting peptides were acidified with trifluoroacetic acid and desalted by C18 column (3 M, Bracknell, UK) as described [32]. The peptide samples were labeled by TMT 10plex™ Isobaric Label Reagent Set label kit (Thermo, USA) according to the instruction. The mixed peptides from 10 samples (two replicates 50 µg each) were fractionated using a reversedphase C18 column (3 M, Bracknell, UK) as described with some modifications [33]. Briefly, 8 fractions of peptide were eluted with acetonitrile step gradients (7.5, 10, 12.5, 15, 17.5, 20, 22.5, and 50%, pH 10). Then, the 8 fractions were dried in a vacuum centrifuge and stored at −80°C until LC-MS/MS analysis.
About 2 µg of each high-pH fractionated peptides samples were separated on an in-house packed 75-μm ID × 50 cm capillary column with 2.5 μm Venusil C18 beads (Agela Technologies, China) using an EASY-nLC 1000 system (Thermo Scientific, Odense, Denmark). The column temperature was maintained at 42°C using an integrated column oven (PRSO-V1, Sonation GmbH, Biberach, Germany). The flow rate was set to 200 nL/min, and a total 240 min or 320 min gradient from 2 to 27% acetonitrile was used. Raw data was collected on Q Exactive mass spectrometer (Thermo Scientific, Bremen, Germany) using Thermo Xcalibur (2.0) control software with a datadependent MS/MS scans (TopN = 15). MS1 spectra was measured at a resolution of 70,000 at m/z 200, and full scan target was 3 × 10 6 with a maximum injection time of 50 ms. Mass range was set to 320-1500. The maximum ion injection time for the MS/MS scan was 100 ms with a resolution of 35,000 at m/z 200, and target value for fragment scans was set at 1 × 10 5 . Isolation width was set at 1.2 m/z and normalized collision energy was set at 32. Each fraction was analyzed in twice.
All raw LC−MS/MS data were submitted to Proteome Discoverer (2.2 version, Thermo Science) for TMT quantitation and database analysis using SequestHT. Data were searched against the Drosophila Melanogaster (Fruit fly) Swiss-Prot database (UP000000803, 21,922 sequences) combined with a common contaminants database (247 entries), using default processing and consensus workflow for MS2 TMT quantification method. The standard searching parameters were used: a 10 ppm MS1 error tolerance was used, with 0.02 Da error used for MS2 product ions on the Q Exactive. Trypsin was set as the enzyme, allowing for 2 missed cleavages. Methionine oxidation was set as a variable modification, and carbamidomethyl on cysteines, TMT6 plex on lysine and peptide N-terminus were set as static modifications for all searches. FDR was set to 0.01 for peptide database search.

Cell transfection
S2 cells were transfected by adding 5 µg of plasmid mixed with 2 µl VigoFect agent (Vigorous Biotechnology) into the cell media. Twenty-four hours following transfection, cells were suspended and adhered to a poly-lysine treated coverslip for 1 h and then fixed for 0.5 h in freshly made 4% paraformaldehyde in phosphate buffer. Samples were blocked using PBST (0.3% Triton X-100) with 5% goat serum for 1 h, then incubated with primary antibodies for 1 h. After washing with phosphate buffer, cells were incubated with secondary antibodies for 1 h. Samples were then imaged using a Nikon A1-R confocal microscope. Images were further processed using Photoshop CC2017 software.

Tamoxifen treatment
A total of 100 mg of tamoxifen salt (Sigma, St Louis, MO, USA) was dissolved in 10 ml of ethanol as a stock solution. On the day of injection, a 1 mg/ml working solution was prepared by mixing 10 mg/ml stock solution with corn oil (Sigma-Aldrich, St Louis, MO, USA) and mixed well. On P20, mutant and control mice were intraperitoneally injected with a daily dosage of 25 mg/kg body weight for 3 days [35].

Immunostaining of mouse retinas
For immunohistochemistry, enucleated eyes were removed, marked at the nasal side for orientation, and fixed for 3 h in 4% paraformaldehyde in 100 mM phosphate buffer (PB) (pH 7.4) and then cryoprotected in 30% sucrose. Tissues were embedded in optimal cutting temperature solution (OCT) and frozen on dry ice for sectioning. Sections (10 micron) were blocked and permeabilized with 10% normal goat serum and 0.1% Triton X-100 in phosphate buffer for 60 min. Labelling with various antibodies was performed as previously described [36]. Primary antibodies were diluted in phosphate buffer containing 5% normal goat serum and 0.1% Triton X-100. Sections were incubated with primary antibodies overnight. Then, the sections were washed with PB three times and labelled for 1 h with secondary antibody and counterstained with DAPI.
Quantification of mislocalized M-Opsin in the cell bodies of cone cells was performed as follows. P30 and P50 cross-sections of retinas were stained using M-Opsin opsin and DAPI, and higher-magnification images were captured on a Leica SP8 confocal microscope. The number of cones in which M-Opsin mislocalized to the inner segment and to the cell body was counted. Slides were photographed on a Zeiss LSM 800 confocal scanning microscope.
Histology and measurement of outer nuclear layer of the mouse retinas For H&E staining, eyes from WT and knockout (KO) mice were removed, marked on the nasal side for orientation, and fixed overnight in 1.22% glutaraldehyde and 0.8% paraformaldehyde in 0.08 M phosphate buffer. They were embedded in paraffin, and cut into 5-µm sections. H&E stained sections were used to count the rows of photoreceptors in the outer nuclear layer [37]. Three measurements of the outer nuclear layer were taken every 200 μm from the optic nerve and averaged. The optic nerve was designated as 0 µm.

EMC3, EMC5, and EMC6 are required for rhodopsin expression
To identify genes involved in rhodopsin homeostasis, we performed EMS mutagenesis and screened chromosome 2 and 3 using the Rh1::GFP ey-flp/hid system, which utilizes the FLP-FRT recombination system to generate eyes homozygous for mutations, while using GFP-tagged Rh1 to track photoreceptor cell integrity [26,38]. Among the~20 mutants which displayed low GFP fluorescence ( Figure S1), four were mutations in genes encoding different EMC components (Fig. 1a). The emc3 G7 and emc3 G10 mutations are two new emc3 alleles, containing nonsense and missense mutation, respectively. One new (mutant emc5 G2 ) introduces a premature stop codon in the CG15168 gene, which encodes an EMC5 homologue. The emc6 N10 allele introduces a nonsense mutation in the CG11781 gene, which encodes an EMC6 homologue ( Fig. 1b and Figure S2).
We confirmed via western blotting that endogenous Rh1 protein levels were greatly reduced in emc3 G7 , emc3 G10 , emc5 G2 , and emc6 N10 mutant animals. Moreover, restoration of EMC3, EMC5, and EMC6 protein in emc3 G7 , emc3 G10 , emc5 G2 , and emc6 N10 photoreceptor cells via the ninaE promoter, completely restored Rh1 levels ( Fig. 1c and Figure S3). These results indicate that EMC is essential for rhodopsin biosynthesis. To analyse the specificity of EMC3, EMC5, and EMC6, we measured levels of TRP and INAD, exclusively function in the phototransduction cascade.
INAD, a cytosolic scaffold protein has little changes in all 3 mutants, indicating the integrity of photoreceptor cells. However, TRP protein levels were moderately reduced in emc5 G2 and emc3 G7 mutants, but dramatically downregulated in emc6 N10 mutants ( Fig. 1c and Figure S3). This suggests that EMC3, EMC5, and EMC6 are involved in regulating the levels of other transmembrane proteins in photoreceptor cells, albeit to different extents.

Rhodopsin is properly folded and is not degraded through the ERAD pathway in emc mutants
To determine which step of the rhodopsin biosynthetic process requires EMC, we examined the localization of endogenous Rh1 in resin-embedded retina sections from emc mutants. As with wild-type flies, the majority of Rh1 The emc3, emc5, and emc6 loci and mutations associated with the emc3 G7 , emc3 G10 , emc5 G2 , and emc6 N10 alleles. c Rhodopsin levels were reduced in emc3 G7 , emc3 G10 , emc5 G2 , and emc6 N10 flies, and were restored by ninaE-emc3, ninaE-emc5, and ninaE-emc6 transgenes, respectively. Protein extracts from 1/4 head of each genotype (1 d after eclosion) were loaded and probed with antibodies against Rh1, ß-actin, TRP, and INAD localized to the rhabdomere (the tightly-packed microvilli in which phototransduction occurs) in emc3 G7 and emc5 G2 flies, although the Rh1 signal was weak compared to rescued flies. This indicates that Rh1 trafficking was normal in both emc3 and emc5 mutants (Fig. 3a). However, Rh1 levels were much lower in emc6 N10 mutants, and Rh1 was detected in the cell body (Fig. 3a). Restoring EMC3 and EMC5 protein levels rescued Rh1 signal in emc3 G7 and emc5 G2 mutants, respectively, and EMC6 restoration rescued both Rh1 signal and localization in emc6 N10 flies, albeit not fully mimic wild-type morphology, which may due to different sectioning angle and depth. During biosynthesis, Rh1 is transiently glycosylated in the ER lumen, and this modification is progressively removed as it is transported to the rhabdomere. If Rh1 quality control or trafficking is blocked, such as in ninaA [6] or dmppe e02905 [44] mutant flies, immature Rh1, which has a higher molecular weight, can be detected [38,45]. The major Rh1 protein isolated from emc3 G7 and emc5 G2 flies was about the same size as in wild-type flies. In emc6 N10 flies, the major Rh1 band was the mature form, but immature Rh1 was also detected (Fig. 3b). This is consistent with the detection of Rh1 in the cell body, and suggests that EMC6 may have distinct functions.
It has been shown that EMCs functionally interacts with the ERAD pathway, which degrades unassembled polypeptides on the ER [11]. Thus, it is possible that the massive degradation of newly synthesized rhodopsin is responsible for the reduced levels of Rh1 in emc mutants. To test this hypothesis, we generated a deletion mutation for the hrd1 gene ( Figure S4), which encodes the major E3 ligase in the ERAD pathway [46]. Eliminating hrd1 did not affect Rh1 levels, and more importantly, loss of hrd1 failed to rescue Rh1 levels in emc3 G7 , emc5 G2 , and emc6 N10 mutants (Figs. 3c-e). These data suggest that EMCs regulate rhodopsin biosynthesis independent of the ERAD pathway.

Mutating EMC components caused different phenotypes
Since the emc5, emc3, and emc6 mutants exhibited similar phenotypes, we assumed that they function through forming the EMC complex. We expressed mCherry-tagged EMC5 in photoreceptor cells, and performed an immunoprecipitation assay with the RFP-Trap followed by mass spectra identification ( Figure S5A). We found that most EMC subunits, including EMC3, EMC4, EMC7, EMC8/9, and Fig. 2 Loss of PDA in emc3, emc5, and emc6 mutants. a ERG recordings in wild-type flies (eyflp rh1-GFP;FRT40A/GMR-hid CL FRT40A) showed that PDA was induced by blue light and terminated by orange light (arrows). b PDA was eliminated in ninaE P332 flies and in c emc3 G7 , e emc3 G10 , g emc5 G2 , and i emc6 N10 flies. PDA was restored in (d) emc3 G7 ;ninaE-emc3, f emc3 G10 ;ninaE-emc3, h emc5 G2 ;ninaE-emc5, and j ninaE-emc6;emc6 N10 flies. Flies 1day after eclosion were darkadapted for 2 min and subsequently exposed to 5-s pulses of orange (O) or blue (B) light. At least 10 flies of each genotype were tested EMC10, were pulled down ( Figure S5B and Supplementary data). This agreed with other mass spectra data indicating that different EMC subunits interact with one another [11,19].
To further demonstrate that different EMC subunits have distinct molecular functions, we performed proteomic analysis of emc mutants from different phenotypic categories. Because null emc2A and emc4 mutations completely lacked retinal cells, we generated weak alleles by expressing RNAis of emc2A, emc4, or emc5 in compound eyes using the GMR-Gal4 driver [28]. The efficiency of emc2A, emc4, and emc5 knockdown was >60% ( Figure S8A), and eye morphologies were normal. Importantly, Rh1 levels were significantly reduced in each of these mutants, confirming that these RNAis reduced the function of EMC2A, EMC4, and EMC5, respectively ( Figure S8B). These data also demonstrated that EMC2A and EMC4 are required for Rh1 biogenesis. For EMC7, we used null mutants for proteomic analyses. We dissected retinas from control, emc2A RNAi , emc4 RNAi , emc5 RNAi and emc7 1 flies for the proteomic assay (Fig. 4b). A total of 6,390 proteins were identified, and compared with controls, 50, 44, 39, and 61 proteins were reduced > 50% in emc2A, emc4, emc5, and emc7 mutants, respectively (Supplementary Tables 1-4 and Supplementary data). Importantly, a majority of these proteins contained transmembrane domain/signal peptide, and Rh1 levels were reduced in emc2A, emc4, and emc5, but not emc7 samples, validating the quantitative proteomic experiments.
Among the reduced proteins, 13 were downregulated in all 4 groups, suggesting that the EMC subunits function cooperatively in the biogenesis of some membrane proteins.
Mutations in emc3, emc5, and emc6 lead to retinal degeneration Because mutations in emc3, emc5, and emc6 largely reduced Rh1 protein levels and disrupted phototransduction, we examined if photoreceptor integrity depends on these genes. Wild-type ommatidia consistently exhibit seven photoreceptor cells with intact rhabdomeres (Fig. 5a). Young (~1 day old) emc5 G2 , emc3 G7 , and emc6 N10 mutants exhibited normal retinal morphology with all 7 rhabdomeres, although the rhabdomeres were smaller in emc6 N10 flies, reflecting an early stage of retinal degeneration. Both emc5 G2 and emc3 G7 mutants contained intact rhabdomeres and photoreceptor cells at 5 days of age, whereas 35-dayold flies exhibited retinal degeneration with prominent vacuoles and obvious loss of rhabdomeres (Figs. 5b, c). Thirty-five-day-old emc5 G2 and emc3 G7 flies exhibited a decrease in ERG amplitude ( Figure S9), also reflecting mild retinal degeneration associated with emc5 G2 and emc3 G7 flies. Consistent with previous results, more severe photoreceptor cell death was associated with emc6 N10 mutant flies, which exhibited disruption of rhabdomeres and the accumulation of abnormal vesicles shortly after eclosion (Fig. 5d). These results demonstrated that EMC3, EMC5, and EMC6 are required for photoreceptor cell survival.

Loss of Emc3 in mammalian cone cells leads to conephotoreceptor defects
We next assessed the functions of EMC in mammalian photoreceptor cells. A conditional knockout allele of Emc3 was generated using the CRISPR/Cas9 system ( Figure S10).
In mouse retina, EMC3 is mainly localized to the outer segment (OS), similar to Rho ( Figure S11). Weak staining was also observed in the inner segment of photoreceptor cells ( Figure S11). One possible explanation of EMC3 localization to OS might be binding of EMC3 to OS targeted membrane proteins, similar to interaction of EMC3 with ABCA3 in the lung [22]. To investigate the role of Emc3 in photoreceptor cells, we generated a conephotoreceptor knockout line using HRGP-Cre mice (Emc3 loxp/loxp HRGP-Cre: Emc3 cone-KO) in which Cre is driven by the human red/green pigment gene promoter [34]. HRGP-Cre mice were used as controls (WT). In Emc3 cone-KO retinas, M-opsin levels were not affected on postnatal day 30 (P30) (Figs. 6a, b). However, by P50 very little M-opsin protein could be detected in Emc3 cone-KO retina sections, suggesting that EMC3 is important for Mopsin stability (Figs. 6c, d).

Deletion of Emc3 in adult animals leads to rod cell degeneration
To further assess the role of EMC3 in mature photoreceptor cells, we used tamoxifen-activated CAG-CreER to remove Emc3 in adult mice. We generated Emc3 loxp/loxp ; CAG-CreER (named Emc3 iKO) mice and treated the mutant and control (CAG-CreER or Emc3 loxp/loxp ) animals with tamoxifen at P20 and assessed the retina 7 and 11 days later (P27 and P31). Tamoxifen induction efficiently deleted Emc3 in all retinal cells, and EMC3 levels were reduced by 70% at P27 compared with control retinas ( Figure S11 and S12). We examined H&E stained retinal sections to assess the morphology of mutant retinas at P27 and P31. At P27, no obvious differences were observed between the mutant and control mice (Figures S13). At P31, however, a thinner outer nuclear layer was seen in the mutant retina (Fig. 7a). The mutant outer nuclear layer (ONL) was reduced by approximately 50% compared with controls, indicating degeneration of Rod photoreceptor cells (Fig. 7a). To investigate if the severe reduction in OS length in Emc3 KO retinas was caused by defective RHO biosynthesis, we examined retina sections via immunohistochemistry using antibodies against RHO. In control retinas, RHO localized exclusively to the OS. In iKO retinas, RHO levels were reduced, and RHO could be detected in the inner segment and in cell bodies at P27 ( Figure S13) and P31 (Fig. 7b). These data suggest that RHO trafficking to the OS was compromised in Emc3 iKO retinas. Fig. 7 Retinal degeneration and RHO mislocalization in Emc3 inducible KO mice. a Paraffin sections of mice retinas 11 days after induction (P31). Sections were stained with H&E, and quantification of outer nuclear layer is shown on the left. Scale bar is 50 μm. H&E staining of paraffin sections and quantification of outer nuclear layer (ONL) nuclei revealed that the mutant ONL was reduced bỹ 50% compared to controls 11 days after induction. On top is an illustration of the whole mouse retina. b Immunofluorescence labeling of retina cryosections for control (WT) and mutant (KO) littermates at P31. Sections were labeled using antibodies against RHO (green) and DAPI (blue). NaK ATPase antibodies were used to label the inner segment. RHO was mislocalized to the inner segment and cell bodies in KO retinas (arrows). Scale bar is 25 μm

Discussion
Although EMCs have been implicated in a variety of physiological functions, their molecular function has remained unexplored. EMC components were first identified based on their ability to interact with ERAD components, suggesting that EMC is required for protein folding and/or ERassociated degradation [11,12]. Later, mutations to EMC subunits were shown to reduce expression of acetylcholine receptors in worms, and rhodopsin in flies, supporting a role for EMC in quality control for a number of multipass transmembrane proteins [24,47]. Although rhodopsin levels are much lower in emc3, emc5, and emc6 mutant animals, rhodopsin is still transported to rhabdomeres, and the mature, deglycosylated form of rhodopsin is primarily detected, indicating that rhodopsin is folded properly in the ER and transported to the plasma membrane. Our data suggest that EMC functions during the de novo biosynthesis of its client protein, rhodopsin, before the protein-folding procedure.
The ERAD pathway is a specialized protein degradation process that removes proteins that fail to fold appropriately in the ER [48]. If loss of EMC destabilizes unfolded rhodopsin in the ER, and/or EMC is connected to the ERAD pathway, disruption of ERAD would be expected to increase rhodopsin levels. However, loss of hrd1 failed to increase rhodopsin levels in emc3, emc5, and emc6 mutant animals, suggesting that EMC has limited function in the quality control of immature rhodopsin, as reported [24,49]. Our finding that EMC functions during rhodopsin biosynthesis, rather than during later post-translational modifications, is supported by recent reports that EMC is involved in the initial biogenesis of membrane proteins [13,14,50].
Since its first identification, EMC components have been shown to interact with each other in several genetic screens [11,12,14,17], and have been suggested to function together as a unit [19,21]. However, a genome-wide genetic screen in C. elegans found that EMC6 is required for the expression of acetylcholine receptors, whereas other EMC subunits are not [47]. Consistent with this notion that different EMC components have distinct functions, several other studies have indicated that individual EMC subunits functions in autophagy, viral assembly, and tissue repair [16,51,52]. In our genome-wide screen for genes that regulate rhodopsin levels, the only EMC components to emerge were emc3, emc5, and emc6. Moreover, by generating an entire EMC KO library, we were able to measure the function and importance of each EMC subunit in different contexts, ranging from animal viability, cellular viability, rhodopsin levels, and phototransduction. Our data revealed that emc1, emc2A, emc4 and emc8/9 are required to maintain basic cellular functions. This may also explain why emc1, emc2, emc4 and emc8/9 mutants were not identified in our EMS screen. In contrast, no phenotype was observed in emc7 and emc10 null mutants, suggesting a less important role of these two EMC components in cellular functions. The weak effects of emc7 and emc10 mutations may be explained by functional redundancy. However, this is unlikely as no other EMC7 or EMC10 homologues can be found in flies based on protein sequence.
In weak emc2 and emc4 alleles (expressing RNAis), Rh1 levels were significantly decreased as well as applying emc5 RNAi, indicating that rhodopsin is a common substrate of EMC complexes. Whereas EMC7 and EMC10 are not required for the EMC complex to participate in rhodopsin biogenesis. Quantitative proteomic analysis to compare changes in protein levels in emc2, emc4, emc5, and emc7 mutants revealed that 13 proteins were downregulated in all 4 mutants, but most (including rhodopsin) were downregulated in ≤3 emc mutants. These data provide additional support that EMC subunits are differentially required for the biogenesis of distinct membrane proteins. Different EMC subunits have been shown to function together in a number of cellular processes. Compared with emc5 G2 and emc3 G7 null mutant animals, emc6 N10 null mutants exhibited lower levels of rhodopsin and TRP proteins, more disrupted rhabdomere morphology, and more severe retinal degeneration. Moreover, EMC5 and EMC3 localized to the ER, whereas EMC6 was also concentrated in intracellular vesicles. However, EMC6 localized to the ER when co-expressed with EMC5, indicating that EMC6 is a bona fide component of the EMC complex, while also performing independent functions outside of the ER. The independent functions of EMC6 may explain its role in the biosynthesis of acetylcholine receptors and in autophagy [16,47]. In S2 cells, EMC6, but not EMC5 and EMC3, partially co-localized with the late endosome maker Rab7, supporting a role for EMC6 in regulating autophagy.
Both rhodopsin and TRP rely on EMC for their biogenesis. However, the emc3, emc5, and emc6 mutant flies only lack the PDA, as seen for ninaE mutants. These are consistent with previous reports that low levels of TRP is enough for maintaining phototransduction, suggesting that loss of rhodopsin is the major cause of the defective visual responses [53]. Rhodopsin also plays structural roles in maintaining the morphology of photoreceptor cells, therefore loss of rhodopsin in most ninaE alleles results in retinal degeneration [54,55]. Photoreceptor cell degeneration associated with the emc mutants could be explained by a disruption in rhodopsin biosynthesis.
The importance of EMC in various human diseases has emerged in recent years. Mutations in EMC1 are associated with global developmental delay, hypotonia, scoliosis, visual impairment and cerebellar atrophy [56,57]. In addition, EMC is indispensable for West-Nile-Virus-induced cell death [21], and EMC1-dependent stabilization plays an essential role in membrane penetration of a nonenveloped virus [51]. We showed that removal of Emc3 from photoreceptor cells led to their degeneration. Removal of Emc3 from the adult retina resulted in the accumulation of RHO in the inner segment or in cell bodies, and the death of rods. We did not observe mislocalized cone opsin in Emc3 mutants. One potential explanation is that Emc3 is not essential for cone opsin trafficking. To the best of our knowledge, our study provides the first evidence that an ER membrane protein is essential for mammalian photoreceptor structure. 2018YSZH0020 to LZ). The funders had no role in the study design, data collection or analysis, decision to publish, or preparation of the manuscript. The authors alone are responsible for the content and writing of the paper.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.