Degradation of Photoreceptor Outer Segments by the Retinal Pigment Epithelium Requires Pigment Epithelium-derived Factor Receptor (PEDF-R) 1 Degradation of Photoreceptor Outer Segments by the Retinal Pigment 1 Epithelium Requires Pigment Epithelium-derived Factor Receptor (PEDF-R) 2 Jeanee Bullock1,2*, Federica Polato1*, Mones Abu-Asab3, Alexandra Bernardo-Colón1, Elma 3 Aflaki1, Martin-Paul Agbaga4, S. Patricia Becerra1a 4 1Section of Protein Structure and Function-LRCMB, National Eye Institute, National Institutes 5 of Health, Bethesda, MD; 2Department of Biochemistry and Molecular & Cellular Biology, 6 Georgetown University Medical Center, Washington D.C.; 3Section of Histopathology, National 7 Eye Institute, National Institutes of Health, Bethesda, MD, 4Departments of Cell Biology and 8 Ophthalmology, Dean McGee Eye Institute, University of Oklahoma HSC, Oklahoma City, OK 9 *These authors contributed equally to this work. 10 aCorresponding author: 11 S. Patricia Becerra 12 NIH-NEI-LRCMB 13 Section of Protein Structure and Function 14 Bg. 6, Rm. 134 15 6 Center Drive MSC 0608 16 Bethesda, MD 20892-0608 17 becerrap@nei.nih.gov 18 19 Present address: 20 JB: Fort Washington, MD, USA; FP: Washington DC, USA; EA: National Institute of Alcohol 21 Abuse and Alcoholism, NIH 22 Funding information: This work was supported by the Intramural Research Program of the 23 National Eye Institute, NIHEY000306 to SPB and by NIH/NEI R01 EY030513 to MPA. 24 25 Word count: 7951 26 27 J. Bullock, None; F. Polato, None; M. Abu-Asab, None; A. Bernardo-Colón, None; E. Aflaki, 28 None; M.P. Agbaga, None; S. P. Becerra, None. 29 30 31 Abbreviations: 32 AMD, age-related macular degeneration; BEL, bromoenol lactone; β-HB, beta hydroxybutyrate; 33 cre, cyclization recombinase; DHA, docosahexaenoic acid; loxP, locus of X-over, P1; PEDF-R, 34 pigment epithelium-derived factor receptor; PNPLA2, patatin-like phospholipase domain 35 containing 2; POS, photoreceptor outer segments; ROI, regions of interest; RPE, retinal pigment 36 epithelium; TEM, transmission electron microscopy; WT, wild type 37 mailto:becerrap@nei.nih.gov 2 Abstract 38 Purpose: To examine the contribution of PEDF-R to the phagocytosis process. Previously, we 39 identified PEDF-R, the protein encoded by the PNPLA2 gene, as a phospholipase A2 in the 40 retinal pigment epithelium (RPE). During phagocytosis, RPE cells ingest abundant phospholipids 41 and protein in the form of photoreceptor outer segment (POS) tips, which are then hydrolyzed. 42 The role of PEDF-R in RPE phagocytosis is not known. 43 Methods: Mice in which PNPLA2 was conditionally knocked out in the RPE were generated 44 (cKO). Mouse RPE/choroid explants were cultured. Human ARPE-19 cells were transfected 45 with siPNPLA2 silencing duplexes. POS were isolated from bovine retinas. The phospholipase 46 A2 inhibitor bromoenol lactone was used. Transmission electron microscopy, 47 immunofluorescence, lipid labeling, pulse-chase experiments, western blots, and free fatty acid 48 and β-hydroxybutyrate assays were performed. 49 Results: The RPE of the cKO mice accumulated lipids as well as more abundant and larger 50 rhodopsin particles compared to littermate controls. Upon POS exposure, RPE explants from 51 cKO mice released less β-hydroxybutyrate compared to controls. After POS ingestion during 52 phagocytosis, rhodopsin degradation was stalled both in cells treated with bromoenol lactone and 53 in PNPLA2-knocked-down cells relative to their corresponding controls. Phospholipase A2 54 inhibition lowered β-hydroxybutyrate release from phagocytic RPE cells. PNPLA2 knock down 55 also resulted in a decline in fatty acids and β-hydroxybutyrate release from phagocytic RPE cells. 56 Conclusions: PEDF-R downregulation delayed POS digestion during phagocytosis. The 57 findings imply that efficiency of RPE phagocytosis depends on PEDF-R, thus identifying a novel 58 contribution of this protein to POS degradation in the RPE. 59 3 A vital function of the retinal pigment epithelium (RPE) is to phagocytose the tips of the 60 photoreceptors in the neural retina. As one of the most active phagocytes in the body, RPE cells 61 ingest daily a large amount of lipids and protein in the form of photoreceptor outer segments 62 (POS) tips.1–5 On the one hand, as outer segments are constantly being renewed at the base of 63 photoreceptors, the ingestion of POS tips (~10% of an outer segment) by RPE cells serves to 64 balance outer segment renewal, which is necessary for the visual activity of photoreceptors. On 65 the other hand, the ingested POS supply an abundant source of fatty acids, which are substrates 66 for fatty acid β-oxidation and ketogenesis to support the energy demands of the RPE.6–8 The fatty 67 acids liberated from phagocytosed POS are also used as essential precursors for lipid and 68 membrane synthesis, and as bioactive mediators in cell signaling processes, e.g., the main fatty 69 acid in POS phospholipids is docosahexaenoic acid, which is involved in signaling in the retina.9 70 Rhodopsin, a pigment present in rod photoreceptors involve in visual phototransduction, is the 71 most abundant protein in POS. Approximately 85% of the total protein of isolated bovine POS is 72 rhodopsin,10 which is embedded in a phospholipid bilayer at a molar ratio between rhodopsin and 73 phospholipids of about 1:60.11 Conversely, the RPE lacks expression of the rhodopsin gene. The 74 importance of POS clearance by the RPE in the maintenance of photoreceptors was 75 demonstrated in an animal model for retinal degeneration, the Royal College Surgeons (RCS) 76 rats, in which a genetic defect in the RCS rats renders their RPE unable to effectively 77 phagocytose POS, thereby leading to rapid photoreceptor degeneration.12,13 Moreover, human 78 RPE phagocytosis declines moderately with age and the decline is significant in RPE of human 79 donors with age-related macular degeneration (AMD), underscoring its importance in this 80 disease.14 Therefore, there is increasing interest in studying regulatory hydrolyzing enzymes 81 involved in RPE phagocytosis for maintaining retina function and the visual process. 82 4 We have previously reported that the human RPE expresses the PNPLA2 gene, which encodes a 83 503 amino acid polypeptide that exhibits phospholipase A2 (PLA2) activity and termed pigment 84 epithelium-derived factor receptor (PEDF-R).15 The enzyme liberates fatty acids from 85 phospholipids, specifically those in which DHA is in the sn-2 position.16 RPE plasma 86 membranes contain the PEDF-R protein,15,17 and photoreceptor membrane phospholipids have 87 high content of DHA in their sn-2 position,9 suggesting that upon POS ingestion the substrate 88 lipid is available to interact with PEDF-R. Other laboratories used different names for the PEDF-89 R protein (e.g., iPLA2ζ, desnutrin, adipose triglyceride lipase), and showed that it exhibits 90 additional lipase activities: triglyceride lipase and acylglycerol transacylase enzymatic 91 activities.18–20 In macrophages, the triglyceride hydrolytic activity is critical for efficient 92 efferocytosis of bacteria and yeast.21 Interestingly, we and others have shown that the inhibitor of 93 calcium-independent phospholipases A2 (iPLA2s), bromoenol lactone (BEL), inhibits the 94 phospholipase and triolein lipase activities of PEDF-R/iPLA2ζ.15,18 In addition, BEL can impair 95 the phagocytosis of POS by ARPE-19 cells, associating phospholipase A2 activity with the 96 regulation of photoreceptor cell renewal.22 However, the responsible phospholipase enzyme 97 involved in RPE phagocytosis is not yet known. 98 Given that the role of PEDF-R in RPE phagocytosis has not yet been studied, here we explored 99 its contribution in this process. We hypothesized that PEDF-R is involved in the degradation of 100 phospholipid-rich POS in RPE phagocytosis. To test this hypothesis, we silenced the PNPLA2 101 gene in vivo and in vitro. Results show that with down regulation of PNPLA2 expression and 102 inhibition of the PLA2 activity of PEDF-R, RPE cells cannot break down rhodopsin, nor release 103 β-hydroxybutyrate (β-HB) and fatty acids, thus identifying a novel contribution of this protein in 104 5 POS degradation. We discuss the role that PEDF-R may play in the disposal of lipids from 105 ingested OS, and in turn in the regulation of photoreceptor cell renewal. 106 Methods 107 Animals 108 The generation of desnutrin floxed mice (hereafter referred to as Pnpla2f/f)23 and the Tg(BEST1-109 cre)Jdun transgenic line24 (which will be named BEST1-cre in this report) have been previously 110 reported. The desnutrin floxed transgenic mouse model was kindly donated to our laboratory by 111 Dr. Hei Sook Sul. The transgenic Tg(BEST1-cre)Jdun mouse model was a generous gift by Dr. 112 Joshua Dunaief. It is an RPE-specific, cre-expressing transgenic mouse line, in which the activity 113 of the human BEST1 promoter is restricted to the RPE and drives the RPE-specific expression of 114 the targeted cre in the eye of transgenic mice.24 Homozygous floxed Pnpla2 (Pnpla2f/f) mice 115 were crossed with transgenic BEST1-cre mice. The resulting mice carrying one floxed allele and 116 the cre transgene (Pnpla2f/+/cre) were crossed with Pnpla2f/f mice to generate mice with Pnpla2 117 knockout specifically in the RPE, which are homozygous floxed mice expressing the cre 118 transgene only in the RPE, Pnpla2f/f/Cre (here also termed cKO). Pnpla2f/f/cre or Pnpla2f/+/Cre were 119 also used for breeding with Pnpla2f/f to expand the colony. Pnpla2f/+ or Pnpla2f/f littermates, 120 obtained through this breeding, were used as control mice. All procedures involving mice were 121 conducted following protocols approved by the National Eye Institute Animal Care and Use 122 Committee and in accordance with the Association for Research in Vision and Ophthalmology 123 Statement for the Use of Animals in Ophthalmic and Vision Research. The mice were housed in 124 the NEI animal facility with lighting at around 280-300 lux in 12 h (6 AM-6 PM) light/12 h dark 125 (6 PM-6 AM) cycles. 126 6 DNA isolation 127 DNA was isolated from mouse eyecups using the salt-chloroform DNA extraction method25 and 128 dissolved in 200 µl of TE (Tris-EDTA composed of 10 mM Tris-HCl, pH 8, and 1 mM EDTA). 129 Aliquots (2 µl) of the DNA solution were then used for each PCR reaction using oligonucleotide 130 primers P1 and P2 (sequences kindly provided by the laboratory of Dr. Hei Sook Sul; Table 1). 131 RNA extraction, cDNA synthesis, and quantitative RT-PCR 132 RNA was isolated from the mouse RPE following the methodology previously described.26 Total 133 RNA was purified from ARPE-19 cells using the RNeasy® Mini Kit (Qiagen, Germantown, MD) 134 following the manufacturer’s instructions. Between 100-500 ng of total RNA were used for 135 reverse transcription using the SuperScript III first-strand synthesis system (Invitrogen, Carlsbad, 136 CA). The PNPLA2 transcript levels in ARPE19 cells determined by quantitative RT-PCR were 137 normalized using the QuantiTect SYBR Green PCR Kit (Qiagen) in the QuantStudio 7 Flex 138 Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA). The primer sequences used 139 in this study are listed in Table 1. Murine PNPLA2 mRNA levels relative to HPRT transcript 140 levels were measured by the QuantStudio 7 Flex Real-Time PCR System using Taqman® gene 141 expression assays (PNPLA2, Mm00503040_m1; HPRT, Mm00446968_m1, Thermo Fisher 142 Scientific). PNPLA2 relative expression to HPRT was calculated using the comparative ΔΔCt 143 method.27 144 Eyecup flatmounts 145 Eyecup (RPE, choroid, sclera) flatmounts were prepared and processed as follows. After 146 enucleation, and removal of cornea, lens, and retina, eyecups were fixed for 1 h in 4% 147 paraformaldehyde at room temperature, and washed 3 times for 10 min each in Tris-Buffered 148 7 Saline (TBS; 25mM Tris HCl pH 7.4, 137 mM NaCl, 2.7 mM KCl). They were then blocked for 149 1 h with 10% normal goat serum (NGS) in 0.1% TBS-Ta (TBS containing 0.1% Triton-X, 150 Sigma, St. Louis, MO). Primary antibodies against cre recombinase and rhodopsin (see Table 2) 151 in 0.1% TBS-Ta containing 2% NGS were diluted and used at 4°C for 16 h. Then, the eyecups 152 were washed 3 times for 10 min each with TBS-Ta followed by incubation at room temperature 153 for 1 h with the respective secondary antibodies, using DAPI (to counterstain the nuclei) and 154 Alexa Fluor 647-phalloidin (to label the RPE cytoskeleton) diluted in 0.1% TBS-Ta containing 155 2% NGS. Eyecups were then flattened by introducing incisions and mounted with Prolong Gold 156 antifade reagent (Thermo Fisher Scientific). Images of the entire flatmounts were collected using 157 the tiling feature of the epifluorescent Axio Imager Z1 microscope (Carl Zeiss Microscopy, 158 White Plains, NY) at 20X magnification. The collected images were stitched together using the 159 corresponding feature of the Zen Blue software (Carl Zeiss Microscopy). Eyecups were also 160 imaged using confocal microscopy (Zeiss LSM 700) at 20X magnification collecting z-stacks 161 spanning 2 µm from each other and covering from the basal to the apical surface of the RPE 162 cells. The image resulting from the maximum intensity projection of the z-stacks was employed 163 for analysis. 164 Five regions of interest (ROI; 520 µm x 520 µm) were selected for each image of the flatmount 165 from cKO mice and control mice. The percentage of cre-positive cells was determined by 166 dividing the number of cells containing cre-stained nuclei by the number of RPE cells in each 167 ROI (identified by F-actin staining). 168 For phagocytosis assay, at least six ROI (320.5 µm x 320.5 µm) were analyzed per mouse. 169 Rhodopsin-stained particles were counted using Image J, after adjusting the color threshold and 170 size of the particles to eliminate the background. 171 8 Transmission electron microscopy 172 Mouse eyes were enucleated and doubly-fixed in 2.5% glutaraldehyde in PBS and 0.5% osmium 173 tetroxide in PBS and embedded in epoxy resin. Thin sections (90nm in thickness) sections were 174 generated and placed on 200-mesh copper grids, dried for 24 h, and double-stained with uranyl 175 acetate and lead citrate. Sections were viewed and photographed with a JEOL JM-1010 176 transmission electron microscope. 177 Electroretinography (ERG) 178 In dim red light, overnight dark-adapted mice were anesthetized by intraperitoneal (IP) injection 179 of Ketamine (92.5 mg/kg) and xylazine (5.5 mg/kg). Pupils were dilated with a mixture of 1% 180 tropicamide and 0.5 % phenylephrine. A topical anesthetic, Tetracaine (0.5%), was administered 181 before positioning the electrodes on the cornea for recording. ERG was recorded from both eyes 182 by the Espion E2 system with ColorDome (Diagnosys LLC, Lowell, MA, USA). Dark-adapted 183 responses were elicited with increasing light impulses with intensity from 0.0001 to 10 candela-184 seconds per meter squared (sc cd.s/m2). Light-adapted responses were recorded after 2 min 185 adaptation to a rod-saturating background (20 cd/m2) with light stimulus intensity from 0.3 to 186 100 sc cd.s/m2. During the recording, the mouse body temperature was maintained at 37°C by 187 placing them on a heating pad. Amplitudes for a-wave were measured from baseline to negative 188 peak, and b-wave amplitudes were measured from a-wave trough to b-wave peak. 189 DC ERG 190 For DC-ERG, sliver chloride electrode connected to glass capillary tubes filled with Hank’s 191 buffered salt solution (HBSS) were used for recording. The electrodes were kept in contact with 192 the cornea for 10 minutes minimum until the electrical activity reached steady-state. Responses 193 to 7-min stead light stimulation were recorded. 194 9 Cell Culture 195 Human ARPE-19 cells (ATCC, Manassas, VA, USA, Cat. # CRL-2302) were maintained in 196 Dulbecco’s modified eagle medium/Nutrient Mixture F-12 (DMEM/F-12) (Gibco; Grand Island, 197 NY) supplemented in 10% fetal bovine serum (FBS) (Gibco) and 1% penicillin/streptomycin 198 (Gibco) at 37°C with 5% CO2. For assays described below, a total of 1 x 105 cells in 0.5 ml were 199 plated per well of a 24-well plates and incubated for 3 days in DMEM/F12 with 10% FBS and 200 1% penicillin-streptomycin. ARPE-19 cells were authenticated by Bio-Synthesis (Lewisville, 201 TX) at passage 27. ARPE-19 cells in passage numbers 27-32 were used for all experiments. 202 Silencing of PNPLA2 in ARPE-19 cells using siRNA 203 Small interfering RNA (siRNA) oligo duplexes of 27 bases in length for human PNPLA2 were 204 purchased from OriGene (Rockville, MD). Their sequences, and that of a Scramble siRNA (Scr) 205 (Cat#: SR324651 and SR311349) are given in Table 3. From the six duplexes, siRNAs C, D, and 206 E consistently provided the highest silencing efficiency and therefore these three duplexes were 207 used individually for silencing experiments and referred to as siPNPLA2. ARPE-19 cells were 208 transfected by reverse transfection in 24-well tissue culture plates as follows: A total of 6 pmols 209 of siRNA was diluted in 100 µl of OptiMem (Gibco) per well, mixed with 1 µl of Lipofectamine 210 RNAiMAX (Invitrogen), and mock transfected cells received only 1 µl of Lipofectamine. Then 211 the mixture was added to each well. After incubation at room temperature for 10 min, a total of 1 212 x 105 cells in 500 µl antibiotic-free DMEM/F12 containing 10% FBS was added to each well and 213 the plate was swirled gently to mix. Assays were performed 72 h post-transfection. 214 Phagocytosis of bovine POS by ARPE-19 cells 215 POS were isolated as previously described28 from freshly obtained cow eyes (J.W. Treuth & 216 Sons, Catonsville, MD). POS pellets were stored at -80°C until use. Quantification of POS units 217 10 was performed using trypan blue and resulted in an average of 5 x 107 POS units per bovine eye. 218 The concentration of protein from purified POS was 21 pg/POS unit. Proteins in the POS 219 samples resolved by SDS-PAGE had the expected migration pattern for both reduced and non-220 reduced conditions, and the main bands stained with Coomassie Blue comigrated with 221 rhodopsin-immunoreactive proteins in western blots of POS proteins (Fig. S1). The percentage 222 of rhodopsin in the protein content of POS was estimated from the gels and revealed that 80% or 223 more of the protein content corresponded to rhodopsin. 224 Using electrospray ionization-mass spectrometry-mass spectrometry (ESI-MSMS) as previously 225 described,29 we determined the lipid composition of the POS that were fed to the ARPE-19 cells. 226 Phagocytosis assays in ARPE-19 cells were performed as follows: ARPE-19 cells (1 x 105 cells 227 per well) were attached to 24-well plates (commercial tissue culture-treated polystyrene plates, 228 TCPS,30 purchased from Corning, Corning, NY) and cultured for 3 days to form confluent and 229 polarized cell monolayers, as we reported previously.31 Ringer’s solution was prepared and 230 composed of the following: 120.6 mM NaCl, 14.3 mM NaHCO3, 4.2 mM KCl, 0.3 mM MgCl2, 231 and 1.1 mM CaCl2, with 15 mM HEPES dissolved separately and adjusted to pH 7.4 with N-232 methyl-D-glucamine. Prior to use, L-carnitine was added to the Ringer’s solution to achieve a 1 233 mM final concentration of L-carnitine. Purified POS were diluted to a concentration of 1 x 107 234 POS/ml in Ringer’s solution containing freshly prepared 5 mM glucose. A total of 500 µl of this 235 solution (medium) was added to each well and the cultures were incubated for 30 min, 60 min or 236 2.5 h, at 37°C. For pulse-chase experiments, after 2.5 h of incubation with POS (pulse), media 237 with POS were removed from the wells and replaced with DMEM/F12 containing 10% FBS and 238 continue incubation for a total of 16 h. The media were separated from the attached cells and 239 stored frozen until use, and the cells were used for preparing protein extracts and either used 240 11 immediately or stored frozen until used. For experiments using BEL (Sigma), BEL dissolved in 241 vehicle dimethyl sulfoxide (DMSO) was mixed with Ringer’s solution and the mixture added to 242 the cells and incubated for 1 h prior to starting the phagocytosis assays. The mixture was 243 removed and replaced with the POS mixture as described above containing DMSO or BEL 244 during the pulse. The assays were performed in duplicate wells per condition and each set of 245 experiments were repeated at least two times. 246 Cell viability by crystal violet staining 247 ARPE-19 cells were seeded in a 96-well plate at a density of 2 x 104 cells per well. The cells 248 were incubated at 37°C for 3 d. The medium was removed and replaced with Ringer’s solution 249 containing various concentrations of BEL and continued incubation at 37°C for 3.5 h. The 250 medium was replaced with complete medium and the cultures incubated for a total of 16 h. After 251 two washes of the cells with deionized H2O, the plate was inverted and tapped gently to remove 252 excess liquid. A total of 50 µl of a 0.1% crystal violet (Sigma) staining solution in 25% methanol 253 was added to each well and incubated at room temperature for 30 min on a bench rocker with a 254 frequency of 20 oscillations per min. The cells in the wells were briefly washed with deionized 255 H2O, and then the plates were inverted and placed on a paper towel to air dry without a lid for 10 256 min. For crystal violet extraction, 200 µl of methanol were added to each well and the plate 257 covered with a lid and incubated at room temperature for 20 min on a bench rocker set at 20 258 oscillations per min. The absorbance of the plate was measured at 570 nm. 259 Western blot 260 ARPE-19 cells plated in multiwell cell culture dishes were washed twice with ice-cold DPBS 261 (137 mM NaCl, 8 mM Na2HPO4-7H20, 1.47 mM KH2PO4, 2.6 mM KCl, 490 μM MgCl2-6H20, 262 900 μM CaCl2, pH 7.2). A total of 120 µl of cold RIPA Lysis and Extraction buffer (Thermo 263 12 Fisher Scientific) with protease inhibitors (Roche, Indianapolis, IN, added as per manufacturer’s 264 instructions) was added to each well and the plate was incubated on ice for 10 min. Cell lysates 265 were collected, sonicated for 20 s with a 50% pulse (Fischer Scientific Sonic Dismembrator 266 Model 100, Hampton, NH), and cellular debris are removed from soluble cell lysates by 267 centrifugation at 20,800 x g at 4°C for 10 min. Protein concentration in the lysates was 268 determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific) and the cell 269 lysates were stored at -20°C until use. Between 5 - 10 µg of cell lysates were used for western 270 blots. 271 Proteins were resolved by SDS-PAGE and transferred to nitrocellulose membranes for 272 immunodetection. The antibodies used are listed on Table 2. For PEDF-R immunodetection, 273 membranes were incubated in 1% BSA (Sigma) in TBS-Tb (50 mM Tris pH 7.5, 150 mM NaCl 274 containing 0.1% Tween-20 (Sigma) at room temperature for 1 h. Then they were incubated in a 275 solution of primary antibody against human PEDF-R at 1:1000 in 1% BSA/TBS-Tb at 4°C for 276 over 16 h. Membranes were washed vigorously with TBS-Tb for 30 min and incubated with anti-277 rabbit-HRP (Kindlebio, Greenwich, CT) diluted 1:1000 in 1% BSA/TBS-Tb at room temperature 278 for 30 min. The membranes were washed vigorously with TBS-Tb for 30 min and 279 immunoreactive proteins were visualized using the KwikQuant imaging system (Kindlebio). For 280 rhodopsin immunodetection, membranes were incubated in 5% dry milk (Nestle, Arlington, VA) 281 in PBS-T (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, pH 7.4, 0.1% Tween 282 20) at room temperature for 1 h. Then, the membranes were incubated in a solution of primary 283 antibody against human rhodopsin (Novus, Littleton, CO) at 1:5000 in a suspension of 5% dry 284 milk in PBS-T at 4°C for over 16 h. The membranes were washed vigorously with PBS-T for 30 285 min and followed with incubation in a solution of anti-mouse-HRP (Kindlebio) 1:1000 in 5% 286 13 milk in PBS-T at room temperature for 30 min. The membranes were washed vigorously with 287 PBS-T for 30 min and immunoreactive proteins were visualized using the KwikQuant imaging 288 system. For protein loading control, the antibodies in membranes as processed described above 289 were removed using Restore™ Western Blot Stripping Buffer (Thermo Fisher Scientific), 290 sequentially followed by incubation with blocking 1% BSA in TBS-T at room temperature for 1 291 h, a solution of primary antibody against GAPDH (Genetex, cat. # GTX627408, Irvine, CA) 292 1:10,000 in 1% BSA/TBS-T at 4°C for over 16 h. After washing the membranes vigorously with 293 TBS-T at room temperature for 30 min, they were incubated in a solution of anti-mouse-HRP at 294 1:1000 in 1% BSA/TBS-T at room temperature for 30 min. After washes with TBS-T as 295 described above, the immunoreactive proteins were visualized using the KwikQuant imaging 296 system. 297 β-Hydroxybutyrate quantification assay 298 In mice, the assay was performed as described before.8 Briefly, after the removal of the cornea, 299 lens and retina, optic nerve, and extra fat and muscles, the eyecup explant from one eye was 300 placed in a well of a 96-well plate containing 170 µl Ringer’s solution and the eyecup from the 301 contralateral eye in another well with the same volume of Ringer’s solution containing 5 mM 302 glucose and purified bovine POS (200 µM phospholipid content, a kind gift from Dr. Kathleen 303 Boesze-Battaglia). The eyecup explant cultures were then incubated for 2 h at 37°C with 5% CO2 304 and, the media were collected and used immediately or stored frozen until use. In ARPE-19 cells, 305 at the endpoint of the phagocytosis assay as described above, a total of 100 µl of the culturing 306 medium was collected and used immediately or stored at 80°C until use. The levels of β-307 hydroxybutyrate (β-HB) released from the RPE cells were determined in the collected samples 308 using the enzymatic activity of β-HB dehydrogenase in a colorimetric assay from the Stanbio 309 14 Beta-hydroxybutyrate LiquiColor Test (Stanbio cat. # 2440058; Boerne, TX) with β-HB 310 standards and following manufacturer’s instructions. 311 Free fatty acids quantification assay 312 A total of 50 µl of conditioned medium from ARPE-19 cell cultures were collected and used to 313 quantify free fatty acids using the Free Fatty Acid Quantification Assay Kit (Colorimetric) 314 (Abcam cat. # ab65341; Cambridge, MA) following manufacturer’s instructions. 315 Statistical analyses 316 Data were analyzed with the two-tailed unpaired Student t test or 2-way ANOVA (analysis of 317 variance), and are shown as the mean ± standard deviation (SD). P values lower than 0.05 were 318 considered statistically significant. 319 Results 320 Generation of an RPE-specific Pnpla2-KO mouse 321 To circumvent the premature lethality of PNPLA2-KO mice,32 a mouse model with RPE-specific 322 knockout of the PNPLA2 gene was designed. For this purpose, we crossed Pnpla2f/f mice23 with 323 BEST1-cre transgenic mice24 to obtain mice with conditional Pnpla2- knockout specific to the 324 RPE, hereafter referred to as cKO (or Pnpla2f/f/cre). In the cKO mice, the promoter of the RPE-325 specific gene VMD2 (human bestrophin, here referred as BEST1) drive the expression of the cre 326 (cyclization recombinase) recombinase and restrict it to the RPE. These mice carry two floxed 327 alleles in the Pnpla2 gene and a copy of the BEST1-cre transgene (Pnpla2f/f/cre). 328 We performed PCR reactions with primers P1 and P2, upstream and downstream from the loxP 329 sites flanking exon 1, respectively (Fig. 1A), with DNA extracted from cKO eyecups and found 330 that the amplimers had the expected length of 253 bp corresponding to the recombined (cKO) 331 15 allele (Fig. 1B), thus showing that the cre-loxP recombination occurred successfully and led to 332 the deletion of the floxed region (exon 1) in the RPE of cKO mice (or Pnpla2f/f/cre). Conversely, 333 we observed two PCR bands of 1749 bp and 1866 bp for littermate Pnpla2f/+ control mice 334 carrying a WT and a floxed allele, respectively (the floxed allele contains two loxP sites) (Fig 335 1B). In lanes for the cKO (or Pnpla2f/f/cre), we also observed very low intensity bands migrating 336 at positions corresponding to 1749 bp and 1866 bp, which probably resulted from a few 337 unsuccessful recombination events. 338 Reverse transcriptase PCR (RT-PCR) revealed PNPLA2 transcript levels in the RPE that were 339 lower from cKO mice than from control (with a mean that was about 32% of the control mice) 340 (Fig. 1C). We determined the percentage of RPE cells that produced the cre protein by 341 immunofluorescence of RPE whole flatmounts. Cells were visualized by co-staining with 342 fluorescein-labelled phalloidin antibody to detect the actin cytoskeleton. We observed cre-343 immunoreactivity in the RPE flatmounts isolated from cKO mice, while no cre-labeling was 344 detected in the controls (Fig. 1D). The overall distribution was patchy and mosaic, as previously 345 described for the BEST1-cre mice.24 The percentage of cre-positive cells in ROI (regions of 346 interest) of flatmounts showed nine mice with expected percentages of cre-positive cells in RPE 347 and one with low cre-positivity (Fig. 1E). The average of the mean values of cre-positive cells 348 for each cKO mouse (mouse numbers 1, 2, 4-10) was 75% (ranging between 52%-91%), which 349 was within the expected for cre positivity in the RPE of the BEST1-cre mouse.24 Cre-positive 350 cells were not detected in RPE of control animals (Fig. 1D-E). Unfortunately, further protein 351 analysis of PEDF-R in mouse retinas was not conclusive because several commercial antibodies 352 to PEDF-R gave high background by immunofluorescence and in western blots. Nevertheless, 353 the results demonstrate the successful generation of RPE-specific PNPLA2-knock-down mice. 354 16 Lipid accumulates in the RPE of Pnpla2-cKO mice 355 We examined the ultrastructure of the RPE by TEM imaging. Accumulation of large lipid 356 droplets (LDs) was observed in cKO mice as early as 3 months of age compared to the control 357 mice cohort (Fig. 2A), and LDs were still observed in the RPE of 13-month old Pnpla2-cKO 358 compared to controls (Fig. 2B). The presence of LDs was associated with either the lack 359 (normally seen in the basal side) (Fig. S2A, S2H) or the decreased thickness of the basal 360 infoldings, and with granular cytoplasm, abnormal mitochondria (Fig. S2B), and disorganized 361 localization of organelles (mitochondria and melanosomes) (Fig. S2A). In some cells, LDs 362 crowded the cytoplasm and clustered together the mitochondria and melanosomes into the apical 363 region of the cells (Figs. S2A, S2C, S2D); however, the number and expansion of LDs within 364 the cells appeared to be random (Fig. S2E). Normal apical cytoplasmic processes were lacking; 365 and degeneration in the outer segment (OS) tips of the photoreceptors was apparent (Figs. S2A, 366 S2F). Additionally, normal phagocytosis of the OS by RPE cells was not evident, implying 367 certain degree of impairment (Figs. S2A, S2E, S2G). There were apparent unhealthy nuclei with 368 pyknotic chromatin and leakage of extranuclear DNA (enDNA), indicating the beginning of a 369 necrotic process (Fig. S2B). Some RPE cells had lighter low-density cytoplasm indicating 370 degeneration of cytoplasmic components in contrast to the denser and fuller cytoplasm in the 371 RPE of the littermate controls (Fig. S2I, S2J). Thus, these observations imply that Pnpla2 down 372 regulation caused lipid accumulation in the RPE. 373 Pnpla2 deficiency increases rhodopsin levels in the RPE of mice 374 Because the RPE does not express the rhodopsin gene, the level of rhodopsin protein in the RPE 375 cells is directly proportional to their phagocytic activity.5,33 To investigate how the knock down 376 17 of Pnpla2 affects RPE phagocytic activity in mice, we compared the rhodopsin-labeled particles 377 present in the eyecup of cKO mice and those of control mice at 2-h and 5-h post-light onset in 378 vivo. The ROIs for the mutant mice were selected from areas rich in cre-positive cells. Phalloidin 379 labeled flatmounts of control mice (n=10) showed that the RPE cells had the typical cobblestone 380 morphology, while nine out of ten cKO mice had distorted cell morphology. Rhodopsin was 381 detected in all ROIs and the labeled particles were more intense and larger in size in the majority 382 of cKO flatmounts compared to those in the control mice. Representative ROIs are shown in 383 figure 3A. The observations implied that Pnpla2 knock down in the RPE prevented rhodopsin 384 degradation in vivo. 385 Ketogenesis upon RPE phagocytosis in explants from cKO mice is impaired 386 Given that RPE phagocytosis is linked to ketogenesis,8 we also measured the levels of ketone 387 body β-HB released by RPE/choroid explants of the cKO mice ex vivo and compared them with 388 those of control littermates. The experiments were performed at 5-h (11AM) and 8-h (2 PM) 389 post-light onset, a time of day in which the amount of β-HB released due to endogenous 390 phagocytosis is not expected to vary with time. A phagocytic challenge by exposure to 391 exogenous bovine OS increased the amount of β-HB released by explants from both cKO and 392 control littermates compared to the β-HB released under basal condition (without addition of 393 exogenous OS) (Fig. 3B). The OS-mediated increase in β-HB release above basal levels of the 394 cKO RPE/choroid explants (1.8 nmols at 11 AM, 0.9 nmols at 2 PM) was lower than the one of 395 the control explants (3 nmols at 11 AM and 2.5 nmols at 2 PM) (Fig. 3C). These observations 396 reveal a deficiency in β-HB production by the RPE/choroid explants of cKO mice under 397 phagocytic challenge ex vivo. 398 18 Electroretinography of the cKO mouse 399 To examine the functionality of the retina and RPE of cKO mice, we performed ERG and DC-400 ERG. Figure 4 shows histograms that revealed no differences among the animals, implying that 401 the functionality was not affected in the RPE-Pnpla2-cKO mice. 402 Phagocytic ARPE-19 cells engulf and break down POS protein and lipid 403 The complexity of the interactions that occur in the native retina makes it difficult to evaluate the 404 subcellular and biochemical changes involved in phagocytosis of POS. Cultured RPE cells 405 provide an ideal alternative to perform these studies. Accordingly, we designed and validated an 406 assay with a human RPE cell line, ARPE-19, to which we added POS isolated from bovine 407 retinas, as described in Methods. The lipid composition of the POS fed to the ARPE-19 cells 408 included phosphatidylcholine (PC) containing very long chain polyunsaturated fatty acids (VLC-409 PUFAs) that was ~27 relative mole percent of total PC species in the POS. The other major PC 410 species include PC 32:00, PC 40:06, and PC 54:10, comprising ~38 relative mole percent of the 411 total PC phospholipids. The most abundant phosphatidylethanolamine (PE) species in the POS 412 were PE 38:06, PE 40:05, and PE 40:06 that accounts for about 74 relative mole percent of the 413 total PE phospholipids. The confluent monolayer of cells was exposed to the purified POS 414 membranes for up to 2.5 h and then the ingested POS were chased for 16h for pulse-chase 415 experiments. The fate of rhodopsin, the main protein in POS, was followed by western blotting 416 of cell lysates. Rhodopsin was detected in the cell lysates as early as 30 min and its levels 417 increased at 1 h and 2.5 h during the POS pulse, and decreased with a 16 h chase (Fig. S3A). 418 Quantification revealed that rhodopsin levels were 21% of those detected after 2.5 h of POS 419 supplementation (Fig. S3B). 420 19 Free fatty acid and β-HB levels were also determined in the culture media during the pulse. The 421 levels of free fatty acids in the medium of POS-challenged ARPE-19 cells were 7-, 5-, and 3-fold 422 higher at 30 min, 60 min and 2.5 h of incubation, respectively, relative to those in the medium of 423 cells not exposed to POS (Fig. S3C). The β-HB levels released into the medium after POS 424 addition also increased by 10-, 2.5- and 4-fold after 30 min, 60 min and 2.5 h incubations, 425 respectively, relative to those observed in the medium of cells not exposed to POS (Fig. S3D). 426 Altogether, these results show that under the specified conditions in this study, the batch of 427 ARPE-19 cells phagocytosed, i.e., engulfed and digested bovine POS protein and lipid 428 components. 429 Bromoenol lactone blocks the degradation of POS components in phagocytic 430 ARPE-19 cells 431 We investigated the role of PEDF-R PLA2 activity in RPE phagocytosis. As we have previously 432 described, a calcium-independent phospholipase A2 inhibitor, bromoenol lactone (BEL), inhibits 433 PEDF-R PLA2 enzymatic activity.15 First, we determined the concentrations of BEL that would 434 maintain viability of ARPE-19 cells. Figure 5A shows the concentration response curve of BEL 435 on ARPE-19 cell viability. The BEL concentration range tested was between 3.125 and 200 μM 436 and the Hill plot estimated an IC50 (concentration that would lower cell viability by 50%) of 437 30.3 μM BEL. Therefore, to determine the effects of BEL on the ARPE-19 phagocytic activity, 438 cultured cells were preincubated with the inhibitor at concentrations below the IC50 for cell 439 viability prior to pulse-chase assays designed as described above. Pretreatment with DMSO 440 alone without BEL was assayed as a control. Interestingly, the inhibitor at 10 μM and 25 μM 441 blocked more than 90% of the degradation of rhodopsin during POS chase for 16 h in ARPE-19 442 cells (Figs. 5B-5C). Similar blocking effects of BEL (25 µM) were observed with time up to 24 443 20 h during the chase (Figs. 5D-5E). The inhibitor did not appear to affect rhodopsin ingestion. The 444 rhodopsin levels in pulse-chase assays with cells pretreated with DMSO alone were like those 445 without pretreatment (compare Figs. 5B and S3A). The cells observed under the microscope 446 after the chase point and prior to the preparation of cell lysates had similar morphology and 447 density among cultures with and without POS, and cultures before and after pulse. Moreover, 448 BEL blocked 40% of the β-HB releasing activity of ARPE-19 cells, whereas DMSO alone did 449 not affect the activity (Fig. 5F). These observations demonstrate that while binding and 450 engulfment were not affected by BEL under the conditions tested, phospholipase A2 activity was 451 required for rhodopsin degradation and β-HB release by ARPE-19 cells during phagocytosis. 452 PNPLA2 down regulation in ARPE-19 cells impairs POS degradation 453 We also silenced PNPLA2 expression in ARPE-19 cells to investigate the possible requirement 454 of PEDF-R for phagocytosis. First, we tested the silencing efficiency of six different siRNAs 455 designed to target PNPLA2, along with a Scrambled siRNA sequence (Scr) as negative control 456 (see sequences in Table 3). The siRNA-mediated knockdown of PNPLA2 resulted in significant 457 decreases in the levels of PNPLA2 transcripts (siRNA A, C, D and E, Figs. 6A and S5) with a 458 concomitant decline in PEDF-R protein levels (siRNA C, D and E, Fig. 6D) in ARPE-19 cell 459 extracts. The siRNAs with the highest efficiency of silencing PNPLA2 mRNA (namely C, D, and 460 E) were individually used for subsequent experiments, and denoted as siPNPLA2 (Fig. 6A). A 461 time course of siPNPLA2 transfection revealed that the gene was silenced as early as 24 h and 462 throughout 72 h post-transfection and parallel to pulse-chase (98.5 h, Figs. 6B, S5). There was 463 no significant difference between mock transfected cells and cells transfected with Scr (Fig. 6C). 464 Examining the cell morphology under the microscope, we did not notice differences between the 465 scrambled and siPNPLA2-transfected cells. Western blots showed that protein levels of PEDF-R 466 21 in ARPE-19 membrane extracts declined 72 h post- transfection (Fig. 6D). Thus, subsequent 467 experiments with cells in which PNPLA2 was silenced were performed 72 h after transfection. 468 Second, we tested the effects of PNPLA2 silencing on ARPE-19 cell phagocytosis. Here we 469 monitored the outcome of rhodopsin in pulse-chase experiments. Interestingly, while PNPLA2 470 knock down did not affect ingestion, the siPNPLA2-transfected cells failed to degrade the 471 ingested POS rhodopsin (88% and 24% remaining at 16 h and at 24h, respectively), while Scr-472 transfected cells were more efficient in degrading them (21% and 12% remaining at 16 and 24 h 473 respectively) (Figs. 7A-7B). 474 Third, we also determined the levels of secreted free fatty acids and β-HB production in PNPLA2 475 silenced cells at 0.5 h, 1 h, and 2.5 h following POS addition. Free fatty acid levels in the culture 476 medium were lower in siPNPLA2-transfected cells than in cells transfected with Scr at 30 min 477 post-addition of POS, and no difference was observed between siPNPLA2 and Scr at 1 h and 2.5 478 h post-addition (Fig. 7C). Secreted β-HB levels in the culture medium were lower in siPNPLA2 479 cells than in Scr-transfected cells at all time points (Fig. 7D). To determine the effect of PNPLA2 480 knockdown on lipid and fatty acid levels in the ARPE-19 cells fed POS membranes, we used 481 electron spray ionization-mass spectrometry (ESI/MS/MS) and gas chromatography-flame ion 482 detection to identify and quantify total lipids and fatty acid composition of the ARPE-19 cells at 483 2.5 and 16 h post POS feeding. Our results did not show any significant differences in the 484 intracellular lipid and fatty acid levels in the siPNPLA2 knockdown in Scr and WT control cells 485 at both 2.5 and 16 h after POS addition (data not shown). Taken together, these results 486 demonstrate that digestion of POS protein and lipid components was impaired in PNPLA2 487 silenced ARPE-19 cells undergoing phagocytosis. 488 22 Discussion 489 Here, we report that PEDF-R is required for efficient degradation of POS by RPE cells after 490 engulfment during phagocytosis. This conclusion is supported by the observed decrease in 491 rhodopsin degradation, in fatty acid release and in β-HB production upon POS challenge when 492 the PNPLA2 gene is downregulated or the PEDF-R lipase is inhibited. These observations occur 493 in RPE cells in vivo, ex vivo and in vitro. The findings imply that RPE phagocytosis depends on 494 PEDF-R for the release of fatty acids from POS phospholipids to facilitate POS protein 495 hydrolysis, thus identifying a novel contribution of this enzyme in POS degradation and, in turn, 496 in the regulation of photoreceptor cell renewal. 497 This is the first time that the PNPLA2 gene has been studied in the context of RPE phagocytosis 498 of POS. Previously, we investigated its gene product, termed PEDF-R, as a phospholipase-linked 499 cell membrane receptor for pigment epithelium-derived factor (PEDF), a retinoprotective factor 500 encoded by the SERPINF1 gene and produced by RPE cells.15,17,34,35 Like RPE cells, non-501 inflammatory macrophages are phagocytic cells, but unlike RPE cells, they are found in all 502 tissues, where they engulf and digest cellular debris, foreign substances, bacteria, other microbes, 503 etc.36,37 The Kratky laboratory reported data on the effects of PNPLA2 silencing in efferocytosis 504 obtained using PNPLA2-deficient mice (termed atgl-/- mouse), and demonstrated that their 505 macrophages have lower triglyceride hydrolase activity, higher triglyceride content, lipid droplet 506 accumulation, and impaired phagocytosis of bacterial and yeast particles,21 and that in these 507 cells, intracellular lipid accumulation triggers apoptotic responses and mitochondrial 508 dysfunction.38 We have shown that PNPLA2 gene knockdown causes RPE cells to be more 509 responsive to oxidative stress-induced death.39 PNPLA2 gene silencing, PEDF-R peptides 510 blocking ligand binding, and enzyme inhibitors abolish the activation of mitochondrial survival 511 23 pathways by PEDF in photoreceptors and other retinal cells.17,34,40 Consistently, overexpression 512 of the PNPLA2 gene or exogenous additions of a PEDF-R peptide decreases both the death of 513 RPE cells undergoing oxidative stress and the accumulation of biologically detrimental 514 leukotriene LTB4 levels.31 The fact that PEDF is a ligand that enhances PEDF-R enzymatic 515 activity, suggests that exposure of RPE to this factor is likely to enhance phagocytosis. These 516 implications are unknown and need further study. Exogenous additions of recombinant PEDF 517 protein to ARPE-19 cells undergoing phagocytosis did not provide evidence for such 518 enhancement (JB personal observations). This suggests that heterologous SERPINF1 519 overexpression in cells and/or an animal model of inducible knock-in of Serpinf1 may be useful 520 to focus on the role of PEDF/PEDF-R in RPE phagocytosis unbiased by the endogenous 521 presence of PEDF. 522 To investigate the consequences of PNPLA2 silencing in POS phagocytosis, we generated a 523 mouse model with a targeted deletion of Pnpla2 in RPE cells in combination with the BEST-cre 524 system for its exclusive conditional silencing in RPE cells (cKO mouse). These mice are viable 525 with no apparent changes in other organs and in weight compared with control littermates and 526 wild type mice. The cKO mice live to an advanced age, in contrast to the constitutively silenced 527 PNPLA2-KO mice in which the lack of the gene causes premature lethality (12-16 weeks) due to 528 heart failure associated with massive accumulation of lipids in cardiomyocytes.32 The RPE cells 529 of the cKO mouse have large lipid droplets at early and late age (Figs. 2A, S2) consistent with a 530 buildup of substrates for the lipase activities of the missing enzyme. In cKO mice, lipid 531 accumulation associates with lack of or the decreased thickness of the basal infoldings, granular 532 cytoplasm, abnormal mitochondria and disorganized localization of organelles (mitochondria and 533 melanosomes) in some RPE cells (Fig. S2). Taken together, the TEM observations in 534 24 combination with the greater rhodopsin accumulation and decline in β-HB release in cKO mice 535 support that PEDF-R is required for lipid metabolism and phagocytosis in the RPE. However, 536 interestingly, the observed features do not seem to affect photoreceptor functionality (Fig. S3) 537 and appear to be inconsequential to age-related retinopathies in the Pnpla2-cKO mouse. This 538 unanticipated observation suggests that the remaining RPE cells expressing Pnpla2 gene 539 probably complement activities of those lacking the gene, thereby lessening photoreceptor 540 degeneration and dysfunction in the cKO mouse. We note that the cKO mouse has a mosaic 541 expression pattern with non-cre-expressing RPE cells, as shown before for the BEST1-cre 542 transgenic line.24 At the same time, the ERG measurements performed correspond to global 543 responses of the photoreceptors and RPE cells, thereby missing individual cell evaluation. The 544 lack of photoreceptor dysfunction with RPE lipid accumulation due to PNPLA2 down regulation 545 also suggests that during development a compensatory mechanism independent of 546 Pnpla2/PEDF-R is likely to be activated, thereby minimizing retinal degeneration in the cKO 547 mouse. Further study will be required to understand the implications of these unexpected 548 findings. Animal models of constitutive heterozygous knockout or inducible knockdown of 549 PNPLA2 may be instrumental to address the role of PNPLA2/PEDF-R in mature photoreceptors 550 unbiased by compensatory mechanisms due to low silencing efficiency or during development. 551 Results obtained from experiments using RPE cell cultures further establish that PEDF-R 552 deficiency affects phagocytosis. It is worth mentioning that the data obtained under our 553 experimental conditions were essentially identical to those typically obtained in assays 554 performed with cells attached to porous permeable membranes, and this provides an additional 555 advantage to the field by requiring shorter time to complete (see Fig. S4). On one hand, the 556 decrease in the levels of β-HB and in the release of fatty acids (the breakdown products of 557 25 phospholipids and triglycerides) upon POS ingestion by cells pretreated with BEL as well 558 as transfected with siPNPLA2 relative to the control cells indicates that PNPLA2 participates in 559 RPE lipid metabolism. On the other hand, the fact that PEDF-R inhibition and PNPLA2 down 560 regulation impair rhodopsin break down from ingested POS in RPE cells implies a likely 561 dependence of PEDF-R-mediated phospholipid hydrolysis for POS protein proteolysis. In this 562 regard, we envision that proteins in POS are mainly resistant to proteolytic hydrolysis, because 563 the surrounded phospholipids block their access to proteases for cleavage. Phospholipase A2 564 activity would hydrolyze these phospholipids to likely liberating the proteins from the 565 phospholipid membranes and become available to proteases, such as cathepsin D, an aspartic 566 protease responsible for 80% of rhodopsin degradation.41 It is important to note that the findings 567 cannot discern whether PEDF-R is directly associated to the molecular pathway of rhodopsin 568 degradation, or indirectly involved in downregulating cathepsin D or other proteases. It is also 569 possible that PNPLA2 deficiency results in the alteration of critical genes regulating the 570 phagocytosis pathway, such as LC3 and genes of the mTOR pathway. Animal models deficient 571 in such genes display retinal phenotypes such as impaired phagocytosis and lipid accumulation, 572 similar to those observed in PEDF-R deficient cells.42–44 These implications need further 573 exploration. 574 Given that BEL is an irreversible inhibitor of iPLA2 it has been used to discern the involvement 575 of iPLA2 in biological processes. Previously, we demonstrated that BEL at 1 to 25 µM blocks 20 576 – 40% of the PLA activity of human recombinant PEDF-R.15 Jenkins et al showed that 2 µM 577 BEL inhibits >90% of the triolein lipase activity of human recombinant PEDF-R (termed by this 578 group as iPLA2ζ).18 In cell-based assays, Wagner et al showed BEL at 20 µM inhibits 40% of 579 this enzyme’s triglyceride lipase activity in hepatic cells.45 In the present study, to minimize 580 26 cytotoxicity and ensure inhibition of the iPLA2 activity of PEDF-R in ARPE-19 cells, we 581 selected 10 µM and 25 µM BEL concentrations that are below the IC50 determined for ARPE-582 19 cell viability (30.2 µM BEL; Fig. 5A). We note that these BEL concentrations are within the 583 range used in an earlier study on ARPE-19 cell phagocytosis.22 We compared our results to those 584 by Kolko et al 22 regarding BEL effects on phagocytosis of ARPE-19 cells. Using Alexa-red 585 labeled-POS, they reported the percent of phagocytosis inhibition caused by 5 – 20 µM BEL as 586 24% in ARPE-19 cells. However, the authors did not specify the time of incubation for this 587 experiment and, based on the other experiments in the report, the time period may have lasted at 588 least 12 h of pulse, implying inhibition of ingestion of POS, and lacking description of the effects 589 of BEL on POS degradation. With unmodified POS in pulse-chase assays, our findings show a 590 percent of inhibition after chase of >90% for 10 µM and 25 µM BEL, indicating more effective 591 inhibition of POS digestion. The effect of BEL on POS ingestion under 2.5 h was insignificant 592 and over 2.5 h remains unknown (pulse). In addition, we show that pretreatment with BEL 593 results in a decrease in the release of β-HB, which is produced from the oxidation of fatty acids 594 liberated from POS. Thus, our assay provides new information -e.g., pulse-chase, use of 595 unmodified POS, β-HB release- to those reported by Kolko et al. It is concluded that BEL can 596 impair phagocytic processes in ARPE-19 cells. While BEL is recognized as a potent inhibitor of 597 iPLA2, it can also inhibit non-PLA2 enzymes, such as magnesium-dependent phosphatidate 598 phosphohydrolase and chymotrypsin.46,47 Consequently , a complementary genetic approach 599 targeting PEDF-R is deemed reasonable and appropriate to investigate its role in RPE 600 phagocytosis. The complex and highly regulated phagocytic function of the RPE also serves to 601 protect the retina against lipotoxicity. By engulfing lipid-rich POS and using ingested fatty acids 602 for energy, the RPE prevents the accumulation of lipids in the retina, particularly phospholipids, 603 27 which could trigger cytotoxicity when peroxidized.48,49 In this regard, the lack of observed 604 differences in intracellular phospholipid and fatty acids between PEDF-R-deficient RPE and 605 control cells lead us to speculate that in ARPE-19 cells exposed to POS the undigested lipids 606 remain within the cells and contribute to the total lipid and fatty acid pool, some of which may 607 be converted to other lipid byproducts to protect against lipotoxicity. Also, the duration of the in 608 vitro chase is shorter than what pertains in vivo, where undigested POS accumulate and overtime 609 coalesce to form the large lipid droplets observed in the RPE in vivo. Thus, future experiments 610 aimed at detailed time-dependent characterization of specific lipid species and free fatty acid 611 levels in the RPE in vivo, and in media and cells in vitro will allow us to have a better 612 understanding of classes of lipids and fatty acids that contribute to the lipid droplet accumulation 613 in the RPE in vivo due to PNPLA2 deletion. Nonetheless, a role of PEDF-R in POS degradation 614 agrees with the previously reported involvement of a phospholipase A2 activity in the RPE 615 phagocytosis of POS22, and with the role of providing protection of photoreceptors against 616 lipotoxicity. 617 In conclusion, this is the first study to identify a role for PEDF-R in RPE phagocytosis. The 618 findings imply that efficient RPE phagocytosis of POS requires PEDF-R, thus highlighting a 619 novel contribution of this protein in POS degradation and its consequences in the regulation of 620 photoreceptor cell renewal. 621 Acknowledgements 622 This work was supported by the Intramural Research Program of the National Eye Institute, NIH 623 (Project #EY000306) to SPB and by NIH/NEI R01 EY030513 to MPA. We thank the NEI 624 animal house, Histopathology, Visual Function, Genetic Engineering and Biological Imaging 625 28 Core facilities for technical support. We thank Dr. Hei Sook Sul, University of California, 626 Berkeley, for kindly providing sequences for primers of Pnpla2 and the Desnutrin flox mouse; 627 Dr. Joshua Dunaief, University of Pennsylvania for kindly providing the transgenic Tg(BEST1-628 cre)Jdun mouse model; Dr. Kathleen Boesze-Battaglia’s laboratory for kindly providing POS; 629 Drs. Eugenia Poliakov and Sheetal Uppal for help in isolating POS; Dr. Kiyoharu J Miyagishima 630 for performing the dcERG experiments; Dr. Preeti Subramanian for technical assistance with cell 631 culture and microscopy; and Dr. Ivan Rebustini for proofreading the manuscript and providing 632 feedback and reagents for RT-PCR. 633 29 Table 1. Primers used for qRT-PCR 634 Gene (Human) Forward Primer Reverse Primer PNPLA2 5’-AGCTCATCCAGGCCAATGTCT-3’ 5’-TGTCTGAAATGCCACCATCCA-3’ 18S 5’-GGTTGATCCTGCCAGTAG-3’ 5’-GCGACCAAAGGAACCATAAC-3’ P1 and P2 5’-GCTTCAAACAGCTTCCTCATG-3’ 5’-GGACTTTCGGTCATAGTTCCG-3’ 635 30 Table 2. Antibodies used in the study 636 Antibody Type & host Application Dilution Company Catalog number GAPDH Monoclonal mouse WB 1:10,000 GeneTex GTX627408 PEDF-R Polyclonal rabbit WB IF 1:1000 1:250 Protein tech 55190-1-AP Rhodopsin (A531) Monoclonal mouse WB IF 1:5000 1:800 Novus Biologicals NBP2-25159 Rhodopsin (B630) Monoclonal mouse IF 1:1000 Novus Biologicals NBP2-25160 cre Recombinase Monoclonal rabbit IF 1:800 Cell Signaling Technology 15036 Alexa Fluor 488 Goat anti-Mouse IgG (H+L) IF 1:500 ThermoFisher Scientific A-11001 Alexa Fluor 555 Goat anti-Rabbit IgG (H+L) IF 1:500 ThermoFisher Scientific A-21428 Alexa Fluor 647 - phalloidin IF 1:100 Cell Signaling Technology 8940 637 31 Table 3. siRNA duplex sequences 638 siRNA Duplex Identifier Duplex sequences SR311349A A rCrGrCrCrArArArGrCrArCrArUrGrUrArArUrArArArUrGCT SR311349B B rGrGrCrArCrArUrArUrArGrArArCrGrUrArCrUrGrCrArUrUCC SR311349C C rGrCrCrUrGrArGrArCrGrCrCrUrCrCrArUrUrArCrCrArCTG SR324651A D rCrCrArArGrUrUrCrArUrUrGrArGrGrUrArUrCrUrArArAGA SR324651B E rCrUrGrCrCrArCrUrCrUrArUrGrArGrCrUrUrArArGrArACA SR324651C F rCrUrUrGrGrUrArArArUrArArArArArCrGrArArArArUrGTT 639 32 References 640 1. Goldman AI, Teirstein PS, O’Brien PJ. The role of ambient lighting in circadian disc 641 shedding in the rod outer segment of the rat retina. Investigative Ophthalmology & Visual 642 Science. 1980;19(11):1257-1267. 643 2. LaVail MM. Circadian nature of rod outer segment disc shedding in the rat. Investigative 644 Ophthalmology & Visual Science. 1980;19(4):407-411. 645 3. Strauss O. The Retinal Pigment Epithelium. Physio Rev. 2005;85(3):845-881. 646 4. Kevany BM, Palczewski K. Phagocytosis of Retinal Rod and Cone Photoreceptors. 647 Physiology. 2010;25(1):8-15. doi:10.1152/physiol.00038.2009 648 5. Mazzoni F, Safa H, Finnemann SC. Understanding photoreceptor outer segment 649 phagocytosis: use and utility of RPE cells in culture. Exp Eye Res. 2014;126:51-60. 650 doi:10.1016/j.exer.2014.01.010 651 6. Fliesler AJ, Anderson RE. Chemistry and metabolism of lipids in the vertebrate retina. 652 Progress in Lipid Research. 1983;22(2):79-131. doi:10.1016/0163-7827(83)90004-8 653 7. Chen H, Anderson RE. Differential incorporation of docosahexaenoic and arachidonic acids 654 in frog retinal pigment epithelium. Journal of Lipid Research. 1993;34(11):1943-1955. 655 8. Reyes-Reveles J, Dhingra A, Alexander D, Bragin A, Philp NJ, Boesze-Battaglia K. 656 Phagocytosis-dependent ketogenesis in retinal pigment epithelium. J Biol Chem. 657 2017;292(19):8038-8047. doi:10.1074/jbc.M116.770784 658 9. SanGiovanni JP, Chew EY. The role of omega-3 long-chain polyunsaturated fatty acids in 659 health and disease of the retina. Progress in Retinal and Eye Research. 2005;24(1):87-138. 660 doi:10.1016/j.preteyeres.2004.06.002 661 10. Obin MS, Jahngen-Hodge J, Nowell T, Taylor A. Ubiquitinylation and Ubiquitin-dependent 662 Proteolysis in Vertebrate Photoreceptors (Rod Outer Segments): EVIDENCE FOR 663 UBIQUITINYLATION OF Gt AND RHODOPSIN. Journal of Biological Chemistry. 664 1996;271(24):14473-14484. doi:10.1074/jbc.271.24.14473 665 11. Palczewski K. G protein-coupled receptor rhodopsin. Annu Rev Biochem. 2006;75:743-767. 666 doi:10.1146/annurev.biochem.75.103004.142743 667 12. Strauss O, Stumpff F, Mergler S, Wienrich M, Wiederholt M. The Royal College of 668 Surgeons Rat: An Animal Model for Inherited Retinal Degeneration with a Still Unknown 669 Genetic Defect. Cells Tissues Organs. 1998;162(2-3):101-111. doi:10.1159/000046474 670 13. D’Cruz PM, Yasumura D, Weir J, et al. Mutation of the receptor tyrosine kinase gene Mertk 671 in the retinal dystrophic RCS rat. Human Molecular Genetics. 2000;9(4):645-651. 672 doi:10.1093/hmg/9.4.645 673 33 14. Inana G, Murat C, An W, Yao X, Harris IR, Cao J. RPE phagocytic function declines in age-674 related macular degeneration and is rescued by human umbilical tissue derived cells. J 675 Transl Med. 2018;16(1):63-63. doi:10.1186/s12967-018-1434-6 676 15. Notari L, Baladron V, Aroca-Aguilar JD, et al. Identification of a Lipase-linked Cell 677 Membrane Receptor for Pigment Epithelium-derived Factor. Journal of Biological 678 Chemistry. 2006;281(49):38022-38037. doi:10.1074/jbc.M600353200 679 16. Pham TL, He J, Kakazu AH, Jun B, Bazan NG, Bazan HEP. Defining a mechanistic link 680 between pigment epithelium–derived factor, docosahexaenoic acid, and corneal nerve 681 regeneration. Journal of Biological Chemistry. 2017;292(45):18486-18499. 682 doi:10.1074/jbc.M117.801472 683 17. Subramanian P, Locatelli-Hoops S, Kenealey J, DesJardin J, Notari L, Becerra SP. Pigment 684 epithelium-derived factor (PEDF) prevents retinal cell death via PEDF Receptor (PEDF-R): 685 identification of a functional ligand binding site. J Biol Chem. 2013;288(33):23928-23942. 686 doi:10.1074/jbc.M113.487884 687 18. Jenkins CM, Mancuso DJ, Yan W, Sims HF, Gibson B, Gross RW. Identification, Cloning, 688 Expression, and Purification of Three Novel Human Calcium-independent Phospholipase A2 689 Family Members Possessing Triacylglycerol Lipase and Acylglycerol Transacylase 690 Activities. Journal of Biological Chemistry. 2004;279(47):48968-48975. 691 doi:10.1074/jbc.M407841200 692 19. Villena JA, Roy S, Sarkadi-Nagy E, Kim K-H, Sul HS. Desnutrin, an Adipocyte Gene 693 Encoding a Novel Patatin Domain-containing Protein, Is Induced by Fasting and 694 Glucocorticoids: ECTOPIC EXPRESSION OF DESNUTRIN INCREASES 695 TRIGLYCERIDE HYDROLYSIS. Journal of Biological Chemistry. 2004;279(45):47066-696 47075. doi:10.1074/jbc.M403855200 697 20. Zimmermann R, Strauss JG, Haemmerle G, et al. Fat Mobilization in Adipose Tissue Is 698 Promoted by Adipose Triglyceride Lipase. Science. 2004;306(5700):1383. 699 doi:10.1126/science.1100747 700 21. Chandak PG, Radovic B, Aflaki E, et al. Efficient phagocytosis requires triacylglycerol 701 hydrolysis by adipose triglyceride lipase. J Biol Chem. 2010;285(26):20192-20201. 702 doi:10.1074/jbc.M110.107854 703 22. Kolko M, Wang J, Zhan C, et al. Identification of Intracellular Phospholipases A2 in the 704 Human Eye: Involvement in Phagocytosis of Photoreceptor Outer Segments. Investigative 705 Ophthalmology & Visual Science. 2007;48(3):1401-1409. doi:10.1167/iovs.06-0865 706 23. Ahmadian M, Abbott MJ, Tang T, et al. Desnutrin/ATGL is regulated by AMPK and is 707 required for a brown adipose phenotype. Cell Metab. 2011;13(6):739-748. 708 doi:10.1016/j.cmet.2011.05.002 709 34 24. Iacovelli J, Zhao C, Wolkow N, et al. Generation of Cre transgenic mice with postnatal RPE-710 specific ocular expression. Invest Ophthalmol Vis Sci. 2011;52(3):1378-1383. 711 doi:10.1167/iovs.10-6347 712 25. Müllenbach R, Lagoda P, Welter C. An efficient salt-chloroform extraction of DNA from 713 blood and tissues. Trends in genetics : TIG. 1989;5(12):391. 714 26. Xin-Zhao Wang C, Zhang K, Aredo B, Lu H, Ufret-Vincenty RL. Novel method for the 715 rapid isolation of RPE cells specifically for RNA extraction and analysis. Exp Eye Res. 716 2012;102:1-9. doi:10.1016/j.exer.2012.06.003 717 27. Livak KJ, Schmittgen TD. Analysis of Relative Gene Expression Data Using Real-Time 718 Quantitative PCR and the 2−ΔΔCT Method. Methods. 2001;25(4):402-408. 719 doi:10.1006/meth.2001.1262 720 28. Schertler GFX, Hargrave PA. [7] Preparation and analysis of two-dimensional crystals of 721 rhodopsin. In: Methods in Enzymology. Vol 315. Academic Press; 2000:91-107. 722 doi:10.1016/S0076-6879(00)15837-9 723 29. Agbaga M-P, Stiles MA, Brush RS, et al. The Elovl4 Spinocerebellar Ataxia-34 Mutation 724 736T>G (p.W246G) Impairs Retinal Function in the Absence of Photoreceptor 725 Degeneration. Molecular Neurobiology. Published online August 11, 2020. 726 doi:10.1007/s12035-020-02052-8 727 30. Lerman MJ, Lembong J, Muramoto S, Gillen G, Fisher JP. The Evolution of Polystyrene as a 728 Cell Culture Material. Tissue Engineering Part B: Reviews. 2018;24(5):359-372. 729 doi:10.1089/ten.teb.2018.0056 730 31. Subramanian P, Mendez EF, Becerra SP. A Novel Inhibitor of 5-Lipoxygenase (5-LOX) 731 Prevents Oxidative Stress-Induced Cell Death of Retinal Pigment Epithelium (RPE) Cells. 732 Invest Ophthalmol Vis Sci. 2016;57(11):4581-4588. doi:10.1167/iovs.15-19039 733 32. Haemmerle G, Lass A, Zimmermann R, et al. Defective Lipolysis and Altered Energy 734 Metabolism in Mice Lacking Adipose Triglyceride Lipase. Science. 2006;312(5774):734. 735 doi:10.1126/science.1123965 736 33. LaVail MM. Rod outer segment disc shedding in relation to cyclic lighting. Experimental 737 Eye Research. 1976;23(2, Part 2):277-280. doi:10.1016/0014-4835(76)90209-8 738 34. Comitato A, Subramanian P, Turchiano G, Montanari M, Becerra SP, Marigo V. Pigment 739 epithelium-derived factor hinders photoreceptor cell death by reducing intracellular calcium 740 in the degenerating retina. Cell Death Dis. 2018;9(5):560-560. doi:10.1038/s41419-018-741 0613-y 742 35. Hernández-Pinto A, Polato F, Subramanian P, et al. PEDF peptides promote photoreceptor 743 survival in rd10 retina models. Experimental Eye Research. 2019;184:24-29. 744 doi:10.1016/j.exer.2019.04.008 745 35 36. Mayerson PL, Hall MO. Rat retinal pigment epithelial cells show specificity of phagocytosis 746 in vitro. J Cell Biol. 1986;103(1):299-308. doi:10.1083/jcb.103.1.299 747 37. Finnemann SC, Rodriguez-Boulan E. Macrophage and retinal pigment epithelium 748 phagocytosis: apoptotic cells and photoreceptors compete for alphavbeta3 and alphavbeta5 749 integrins, and protein kinase C regulates alphavbeta5 binding and cytoskeletal linkage. J Exp 750 Med. 1999;190(6):861-874. doi:10.1084/jem.190.6.861 751 38. Aflaki E, Radovic B, Chandak PG, et al. Triacylglycerol accumulation activates the 752 mitochondrial apoptosis pathway in macrophages. J Biol Chem. 2011;286(9):7418-7428. 753 doi:10.1074/jbc.M110.175703 754 39. Subramanian P, Becerra SP. Role of the PNPLA2 Gene in the Regulation of Oxidative Stress 755 Damage of RPE. In: Bowes Rickman C, Grimm C, Anderson RE, Ash JD, LaVail MM, 756 Hollyfield JG, eds. Retinal Degenerative Diseases. Springer International Publishing; 757 2019:377-382. 758 40. Kenealey J, Subramanian P, Comitato A, et al. Small Retinoprotective Peptides Reveal a 759 Receptor-binding Region on Pigment Epithelium-derived Factor. J Biol Chem. 760 2015;290(42):25241-25253. doi:10.1074/jbc.M115.645846 761 41. Rakoczy PE, Baines M, Kennedy CJ, Constable IJ. Correlation Between Autofluorescent 762 Debris Accumulation and the Presence of Partially Processed Forms of Cathepsin D in 763 Cultured Retinal Pigment Epithelial Cells Challenged with Rod Outer Segments. 764 Experimental Eye Research. 1996;63(2):159-167. doi:10.1006/exer.1996.0104 765 42. Dhingra A, Bell BA, Peachey NS, et al. Microtubule-Associated Protein 1 Light Chain 3B, 766 (LC3B) Is Necessary to Maintain Lipid-Mediated Homeostasis in the Retinal Pigment 767 Epithelium. Front Cell Neurosci. 2018;12:351-351. doi:10.3389/fncel.2018.00351 768 43. Cheng S-Y, Cipi J, Ma S, et al. Altered photoreceptor metabolism in mouse causes late stage 769 age-related macular degeneration-like pathologies. Proc Natl Acad Sci U S A. 770 2020;117(23):13094-13104. doi:10.1073/pnas.2000339117 771 44. Go Y-M, Zhang J, Fernandes J, et al. MTOR-initiated metabolic switch and degeneration in 772 the retinal pigment epithelium. The FASEB Journal. 2020;34(9):12502-12520. 773 doi:10.1096/fj.202000612R 774 45. Wagner C, Hois V, Pajed L, et al. Lysosomal acid lipase is the major acid retinyl ester 775 hydrolase in cultured human hepatic stellate cells but not essential for retinyl ester 776 degradation. Biochim Biophys Acta Mol Cell Biol Lipids. 2020;1865(8):158730-158730. 777 doi:10.1016/j.bbalip.2020.158730 778 46. Balsinde J, Dennis EA. Bromoenol Lactone Inhibits Magnesium-dependent Phosphatidate 779 Phosphohydrolase and Blocks Triacylglycerol Biosynthesis in Mouse P388D1 Macrophages. 780 Journal of Biological Chemistry. 1996;271(50):31937-31941. doi:10.1074/jbc.271.50.31937 781 36 47. Jenkins CM, Han X, Mancuso DJ, Gross RW. Identification of Calcium-independent 782 Phospholipase A2 (iPLA2) β, and Not iPLA2γ, as the Mediator of Arginine Vasopressin-783 induced Arachidonic Acid Release in A-10 Smooth Muscle Cells: ENANTIOSELECTIVE 784 MECHANISM-BASED DISCRIMINATION OF MAMMALIAN iPLA2s. Journal of 785 Biological Chemistry. 2002;277(36):32807-32814. doi:10.1074/jbc.M202568200 786 48. Ueta T, Inoue T, Furukawa T, et al. Glutathione peroxidase 4 is required for maturation of 787 photoreceptor cells. J Biol Chem. 2012;287(10):7675-7682. doi:10.1074/jbc.M111.335174 788 49. Imai H, Matsuoka M, Kumagai T, Sakamoto T, Koumura T. Lipid Peroxidation-Dependent 789 Cell Death Regulated by GPx4 and Ferroptosis. In: Nagata S, Nakano H, eds. Apoptotic and 790 Non-Apoptotic Cell Death. Springer International Publishing; 2017:143-170. 791 doi:10.1007/82_2016_508 792 793 37 Figure legends 794 Figure 1. 795 Generation of RPE-specific PNPLA2-cKO mice. (A) Scheme of Pnpla2 floxed and cre-796 mediated recombined allele. The loxP sites flank Exon 1. P1 and P2 are the primers homologous 797 to sequences outside the floxed (flanked by the loxP sites) region used to detect cre-mediated 798 recombination (generating recombined alleles) on genomic DNA. The sizes of the amplicons 799 obtained by PCR using P1 and P2 are indicated. (B) Gel electrophoresis of PCR reaction 800 products obtained using primers P1 and P2 and genomic DNA isolated from mouse eyecups 801 from either cKO or control (Ctr) mice (Pnpla2f/+); lane 1 (MW) corresponds to molecular weight 802 markers (GeneRuler DNA Ladder Mix). One eyecup per lane from a 4-month old mouse, n=2 803 cKO, n=2 Ctr. (C) Pnpla2 expression (vs. HPRT) in RPE from month-old cKO (Pnpla2f/f/cre) 804 relative to control littermates (Pnpla2f/f). Each data point corresponds to the average of six PCR 805 reactions per eyecup, six eyes from three cKO mice and six eyes from three control mice at 5 – 7 806 months old. (D) cre (red) and phalloidin (yellow) labeling of RPE/choroid flatmounts from 807 control (Pnpla2f/f) (left) and littermate cKO (Pnpla2f/f/cre) (right). The scale corresponds to 20 808 µm. (n=2 images from individual mouse eyecup at 11-14 months old). (E) Plot of percentage of 809 cre-positive RPE cells in cKO animals (Pnpla2f/f/cre, n=10, age was 10.5-18.5 months old) as 810 indicated in x-axis. Each data point corresponds to percentage of cre-positive RPE cells from an 811 ROI, each bar corresponds to a flatmount of an individual cKO mouse, and the bar for control 812 (Pnpla2f/f) has data from 10 mice. 813 Figure 2. 814 Lipid accumulation in the RPE of Pnpla2-cKO mice. Electron microscopy micrographs 815 showing the RPE structure of 3- (A) and 13 (B) month-old cKO mice and control animals. LD: 816 lipid droplets; BI: basal infoldings. Scale bar corresponds to 2 µm. The representative images 817 were selected among examinations of micrographs from 8 eyes of cKO (PNPLA2f/f cre+) mice, 818 from 7 eyes of control (PNPLA2f/f) mice at 1.75 - 3.75-month-old; and from 3 eyes of cKO mice 819 and 3 eyes of control mice at 12.5 - 13-month-old. 820 Figure 3. 821 Phagocytosis and β-hydroxybutyrate production in the RPE of Pnpla2-cKO mice. (A) 822 Representative ROI of the eyecup from one control and one cKO animal isolated at 2 h (8 AM) 823 and 5 h (11 AM) post light onset (6 AM) after immunolabeling for rhodopsin (in green) 824 38 phalloidin (in yellow) and cre (in red). The column to the right shows magnification of an area. 825 The mean of rhodopsin immunolabel intensity in micrographs (n ≥ 6 ROIs) from flatmounts (as 826 indicated in x-axis) relative to control at 2h was determined among three mice per condition and 827 shown in the plot. Age of mice was 10.5 – 18.5 months. (B) Ex-vivo β-HB release by the RPE of 828 Pnpla2-cKO eyecups upon ingestion of outer segments (OS) in comparison to that of controls. 829 Eyecups were isolated at 5 h (11 AM) and 8 h (2 PM) after light onset (6 AM). Statistical 830 significance was calculated using 2-way ANOVA for the 2 groups (controls and cKO mice) with 831 and without treatment (second variance) for each time after light onset (* p=0.02; ** p=0.006; 832 *** p=0.0001); ns, not significant. (n =6 eyecups from 3 control (f/+) mice at 3.5 months; n=4 833 eyecups from 2 control (f/f cre-) mice at 3.5 months; n=10 eyecups from 5 mice (f/f cre+) at 2.75 834 – 3.5 months) (C) The OS-mediated increase in β-HB release above basal levels of the cKO 835 RPE/choroid explants was calculated from the data in Panel (C) and plotted. 836 Figure 4. 837 RPE and Retinal functionality in RPE-Pnpla2-cKO mice. (A) Histogram showing the 838 amplitude (mean, standard deviation) of the c-wave, fast oscillation (FO), light peak (LP) and 839 off-response (OFF) measured by DC-ERG of 11-week-old cKO (n=4, empty histograms) and 840 control mice ((Pnpla2f/f and Pnpla2f/+, n=5, filled histograms). (B) Electroretinograms showing 841 amplitude (y-axis) of scotopic a- and b-wave, and photopic b-wave, as a function of light 842 intensity (x-axis) of 3 and 12-month-old cKO mice (empty circle) and littermate controls 843 (Pnpla2f/f, filled circles) (n=3/genotype). 844 Figure 5. 845 Phagocytosis in ARPE-19 cells pretreated with BEL. (A) ARPE-19 cells were incubated with 846 BEL at the indicated concentrations for 3.5 h. Then the mixture was removed, washed gently 847 with PBS, and incubated with complete medium for a total of 16 h. Cell viability was assessed 848 by crystal violet staining and with three replicates per condition. (B) Representative immunoblot 849 of total lysates of cells, which were pretreated with DMSO alone, 10 or 25 µM BEL/DMSO for 1 850 h prior to pulse-chase of POS, as described in methods. Extracts of cells harvested at the 851 indicated times (top of blot) were resolved by SDS-PAGE followed by immunoblotting with 852 anti-rhodopsin. Migration position of rhodopsin is indicated to the right of the blot. (C) 853 Quantification of rhodopsin from total lysates of cells of the pulse-chase experiments as in panel 854 (B). Samples from each biological replicate were resolved in duplicate by SDS-PAGE from two 855 39 experiments and single for the third experiment for quantification. Intensities of the 856 immunoreactive bands were determined and the percentage of the remaining rhodopsin after 16-857 h chase relative to rhodopsin at 2.5 h-pulse was plotted. (D) Representative immunoblot of total 858 lysates of cells, as in panel B to determine the effects of BEL at 16 h and 24 h of chase (as 859 indicated). (E) Quantification of rhodopsin from two independent experiments of the pulse-chase 860 experiments as in panel D. Samples from each biological replicate were resolved in duplicate by 861 SDS-PAGE for quantification. Intensities of the immunoreactive bands were determined and the 862 percentage of the remaining rhodopsin after 16-h chase relative to rhodopsin at 2.5 h-pulse was 863 plotted. (F) Cells were preincubated with DMSO alone, 10 or 25 µM BEL/DMSO in Ringer’s 864 solution at 37°C for 1 h. Then, the mixture was removed, and cells were incubated with Ringer’s 865 solution containing 5 mM glucose and POS (1x107 units/ml) with DMSO alone, 10 or 25 µM 866 BEL/DMSO for the indicated times (x-axis). Media were removed to determine the levels of β-867 HB secretion, which were plotted (y-axis). (n=3) Data are presented as means ± S.D. **p<0.01, 868 ***p<0.001. 869 Figure 6. 870 Knockdown of PNPLA2 in ARPE-19 cells. ARPE-19 cells were transfected with Scr 871 (Scrambled siRNA control) or siRNAs targeting PNPLA2, and mRNA levels and protein were 872 tested. (A) RT-qPCR to measure PNPLA2 mRNA levels in ARPE-19 cells 72 h post-transfection 873 with Scr and six different siRNAs (as indicated in the x-axis) was performed and a plot is shown. 874 PNPLA2 mRNA levels were normalized to 18S. All siRNA are represented as the percentage of 875 the scrambled siRNA control. n = 3 (B) A plot is shown for a time course of PNPLA2 mRNA 876 levels following transfection with Scr and siPNPLA2-C. n = 3 (C) RT-qPCR of mock-transfected 877 cells, cells transfected with Scr, and siPNPLA2-C (x-axis) at 72 h after transfection. mRNA 878 levels were normalized to the 18S RNA (y-axis). n = 3 (D) Total protein was obtained from cells 879 harvested 72 h after transfection and resolved by SDS-PAGE followed by western blotting with 880 anti-PNPLA2 and anti-GAPDH (loading control). The siRNAs used in transfections are 881 indicated at the top, and migration positions for PEDF-R and GAPDH are to the right of the blot. 882 Data are presented as means ± S.D. **p<0.01, ***p<0.001***p<0.001 883 Figure 7. 884 Phagocytosis and fatty acid metabolism in siPNPLA2 cells. ARPE-19 cells were transfected 885 with Scr or siRNAs targeting PNPLA2. At 72 h post-transfection, ARPE-19 cells were incubated 886 40 with POS (1 x 107 units/ml) in 24-well tissue culture plates for pulse-chase experiments. (A) 887 Representative immunoblot of total lysates of ARPE-19 cells at 0.5 h, 1 h, and 2.5 h of POS 888 pulse and at a 16-h and 24-h chase period, as indicated at the top of the blot. Proteins in cell 889 lysates were subjected to immunoblotting with anti-rhodopsin followed by reprobing with anti-890 GAPDH as the loading control. (B) Quantification of rhodopsin from duplicate samples and 3 891 blots of cell lysates from pulse-chase experiments and time periods (indicated in the x-axis) as 892 from panel. Data are presented as means ± S.D. ns, not significant, **p<0.01. (A). Intensities of 893 the immunoreactive bands were determined and the percentage of the remaining rhodopsin after 894 16-h and 24-h chase relative to rhodopsin at 2.5 h-pulse was plotted (y-axis). (C-D) Levels of 895 secreted free fatty acids (C) and β-HB (D) were measured in culture media of cells transfected 896 with Scr or siPNPLA2 following incubation with POS for the indicated periods of times (x-axis). 897 (n =3) Data are presented as means ± S.D. * p < 0.05, **p<0.01. Duplex siPNPLA2 C was used 898 to generate the data (see Table 3 for sequences of duplexes). 899 41 Supplementary Information 900 Figure S1. Proteins in the POS samples were determined and resolved by SDS-PAGE in the 901 same gel in two sets: one with 5 µg and another with 0.1 µg protein per lane. For each set, one 902 sample was non-reduced and the other was reduced with DTT. After electrophoresis, the gels 903 were cut in half lengthwise. The gel portion with 5 µg of protein was stained with Coomassie 904 Blue and the other portion with 0.1 µg protein was transferred to a nitrocellulose membrane for 905 immunostaining using anti-rhodopsin antibodies (as described in Methods). Photos of the stained 906 gel and western blot are shown. 907 The proteins of POS isolated from bovine retina had the expected migration pattern for both 908 reduced and non-reduced conditions, and the main bands stained with Coomassie Blue 909 comigrated with rhodopsin-immunoreactive proteins in western blots of POS proteins. 910 Figure S2. Electron microscopy micrographs. Panels A-J show electron microscopy 911 micrographs of RPE structures of 3-month-old RPE cKO prepared as described in the main text 912 and Figure 2. Magnification is indicated for each image. 913 The presence of LDs was associated with lack (Fig. S2A) of or the decreased thickness of the 914 basal infoldings, and with granular cytoplasm, abnormal mitochondria (Fig. S2B), and 915 disorganized localization of organelles (mitochondria and melanosomes) (Fig. S2A). In some 916 cells, the large LDs crowded the cytoplasm and clustered together the mitochondria and 917 melanosomes into the apical region of the cells (Figs. S2A, S2C, S2D); however, LDs number 918 and expansion within the cells appeared to be random and their expansion could go into any 919 direction (Fig. S2E). Normal apical cytoplasmic processes were lacking; however, degeneration 920 in the outer segment (OS) tips of the photoreceptors was visible (Figs. S2A, S2F). Additionally, 921 normal phagocytosis of the OS was lacking indicating an impaired RPE phagocytosis (Figs. 922 S2A, S2E, S2G). There were apparent unhealthy nuclei with pyknotic chromatin and leakage of 923 extranuclear DNA (enDNA), indicating that the beginning of the necrotic process had started 924 (Fig. S2B). Some RPE cells lacked basal infoldings, normally seen at the basal side (Fig. S2H). 925 Occasionally some RPE cells had lighter low-density cytoplasm indicating degeneration of 926 cytoplasmic components in contrast to the denser and fuller cytoplasm in the RPE of the 927 littermate control (Fig. S2I, S2J). 928 42 Figure S3. 929 Phagocytosis in ARPE-19 cells. ARPE-19 cells were cultured in 24-well plates for 3 days, and 930 then exposed to POS at 1x107 units/ml for up to a 2.5-h pulse followed by a 16-h chase period as 931 described in Methods. (A) Representative immunoblots of total cell lysates during pulse-chase 932 (times indicated at the top of the blot) with anti-rhodopsin followed by reprobing with anti-933 GAPDH as the loading control are shown. Migration positions of rhodopsin and GAPDH are 934 indicated to the right of the blot. Duplicate biological replicates were performed. (B) 935 Quantification of rhodopsin from duplicate samples per condition from pulse-chase experiments 936 at time periods indicated in the x-axis as from panel (A). Intensities of the immunoreactive bands 937 from duplicate samples of cell lysates were determined. The percentage of the remaining 938 rhodopsin after 16-h chase relative to rhodopsin at 2.5 h-pulse was plotted. (C-D) Levels of free 939 fatty acids (C) and β-HB (D) measured in culture media of cells incubated with and without POS 940 for the indicated periods of time (x-axis) were plotted and shown. n = 3 Data are presented as 941 means ± S.D. * p < 0.05, ***p<0.001. 942 Figure S4. Phagocytosis in ARPE-19 cells in porous membranes. ARPE-19 cells were treated 943 with 1x107 POS/ml. (A) Representative immunoblot showing rhodopsin internalization from 944 total cell lysates of ARPE-19 cells following 30, 60, and 150 min of POS incubation following 945 plating in 12-well transwell inserts for 3 weeks. Cell extracts were resolved by SDS-PAGE 946 followed by immunoblotting with anti-rhodopsin. The blot was stripped and reprobed with anti-947 GAPDH as a loading control. (B) Levels of B-HB secreted towards the apical membrane of 948 ARPE-19 cells following POS incubation for 30, 60, and 150 min. (n = 3) Data are presented as 949 means ± S.D. 950 Methods: 951 To demonstrate a functional assay to study phagocytosis in ARPE-19 cells we perform the assay 952 with confluent cells attached on porous membranes 953 ARPE-19 cells seeded on porous membranes were incubated for 3 weeks in culturing media. 954 Then the media was removed and replaced with Ringer’s solution alone or Ringer’s solution 955 containing 1 x 107 POS/ml and 5 mM glucose for the indicated time points. Rhodopsin was 956 detected by western blotting. 957 43 Rhodopsin levels in the lysates of cells incubated with POS were detected in as little as 30 min 958 and up to 2.5 h following POS incubation, while rhodopsin was undetectable in cells without 959 POS (Fig. S4A). β-HB levels released into the media of the apical chamber of transwells 960 following POS incubation increased four-fold and three-fold after 1 h and 2.5 h, respectively, 961 while released β-HB levels from cells incubated with Ringer’s solution alone did not increase 962 (Fig. S4B). 963 Figure S5: ARPE-19 cells were transfected with siScramble siRNA control or siRNAs targeting 964 PNPLA2 (siPNPLA2 A). RT-qPCR to measure PNPLA2 mRNA levels in ARPE-19 cells at (A) 965 72 h post-transfection and (B) 98.5h post transfection equivalent to pulse (2.5h) and chase (24h) 966 was performed with siRNA duplexes (as indicated in the x-axis). Treatment of cells in panel B 967 was as for pulse-chase (see diagram in Fig S3). PNPLA2 mRNA levels were normalized to 18S. 968 n =3 biological replicates, each data point corresponds to the average of triplicate PCR reactions. 969 The RT-PCR was repeated twice per biological replicate. Values that fell out of the standard 970 curve were not included in the plot. 971 The data shows that siPNPLA2 duplex silenced PNPLA2 in ARPE-19 at 72 h post-transfection 972 and that silencing was maintained throughout a 2.5 h and pulse-chase of 24 h. 973 Floxed allele Cre-recombined allele 1866 bp 253 bp MW cKO Ctr cKO Ctr Figure 1. B. A. C. D. E. co ntr ol 1 2 3 4 5 6 7 8 9 10 0 40 80 120 cKO mouse # C re p os iti ve c el ls (% ) control cKO cre/phalloidin Co ntr ol cK O 0.0 0.5 1.0 1.5 P np la 2/ H P R T Figure 2. BI LD A. B. BI LD BI BI Co nt ro l cK O Figure 3. A. 2 ho ur s (8 A M ) cK O Co nt ro l rhodopsin/cre/phalloidin 5 ho ur s (1 1 A M ) cK O Co nt ro l 5 hours (11 AM) 8 hours (2 PM) 0 2 4 6 β -h yd ro xy bu ty ra te (n m ol ) Control Control +OS cKO cKO+OS ****** ** ** 5 h (11 AM) 8 h (2 pm) 0 1 2 3 Time after light onset ∆ β -H B (n m ol ) Control cKO * * B. C. co ntr ol cK O co ntr ol cK O 0 1 2 3 Mouse R ho do ps in (r el at iv e to c on tro l 2 h) 2 h 5 h Scotopic a-wave Scotopic b-wave Photopic b-wave -5 -4 -3 -2 -1 0 1 2 0 200 400 600 800 1000 -5 -4 -3 -2 -1 0 1 2 0 100 200 300 -2 -1 0 1 2 0 100 200 300 400 Control cKO -2 -1 0 1 2 0 100 200 300 400 -5 -4 -3 -2 -1 0 1 2 0 200 400 600 800 1000 -5 -4 -3 -2 -1 0 1 2 0 100 200 300 3 m on th 12 m on th Light intensity log (cd/s.m2) A m pl itu de (µ V) c-wave FO LP OFF 0.0 0.5 1.0 1.5 2.0 A m pl itu de (µ V) Figure 4. A. B. Figure 5. A. B. C. D. 1 10 100 1000 0.0 0.2 0.4 0.6 BEL (µM) C el l v ia bi lit y (A bs 57 0n m ) 0.5 1 2.5 0.5 1 2.5 0.5 1 2.5 0 1 2 3 Time (h) β -H B (n m ol es ) 0 10 25 ** *** ** ** *** *** BEL (µM) -Rhodopsin BEL (µM) 0 10 25 E. -Rhodopsin BEL (µM) 0 25 0.5 1 2.5 16 24 0.5 1 2.5 16 24 (h) F. 0.5 1 2.5 16 0.5 1 2.5 16 0.5 1 2.5 16 (h) 0 10 25 0 50 100 150 BEL (µM) R ho do ps in r em ai ni ng (% re la tiv e to 2 .5 h ) 16 h ✱✱✱ ✱✱ 2.5 16 24 2.5 16 24 0 50 100 150 Time (h) R ho do ps in r em ai ni ng (% re la tiv e to 2 .5 h ) BEL (µM) 0 25 ns **** *** Figure 6. B. A. C. D. 24 48 72 0.00 0.04 0.08 0.12 0.16 Time (h) P N P LA 2/ 18 S Scr siPNPLA2 *** *** *** No ne Sc r PN PL A2 0.00 0.15 0.30 0.45 siRNA PN PL A 2/ 18 S n.s. *** Scr A B C D E F 0 50 100 150 siRNA P N P LA 2/ 18 S (% ) *** *** *** *** *** ** None Scr C D E siRNA -GAPDH -PEDF-R Figure 7. A. C. D. Scr siPNPLA2 -GAPDH -Rhodopsin 0.5 1 2.5 16 24 0.5 1 2.5 16 24 (h) B. 0.5 1.0 2.5 0.0 0.4 0.8 1.2 Time (h) β -h yd ro xy bu ty ra te (n m ol ) Scr siPNPLA2 * * ** 0.5 1.0 2.5 0.0 0.5 1.0 1.5 Time (h) Fr ee fa tt y ac id s (p m ol ) Scr siPNPLA2 * 16 24 0 30 60 90 120 Time (h) R ho do ps in r em ai ni ng (% re la tiv e to 2 .5 h ) Scr siPNPLA2 ✱✱ ✱✱✱ Degradation of Photoreceptor Outer Segments by the Retinal Pigment Epithelium Requires Pigment Epithelium-derived Factor Receptor (PEDF-R) Jeanee Bullock, Federica Polato, Mones Abu-Asab, Alexandra Bernardo-Colón, Ivan Rebustini, Elma Aflaki, Martin-Paul Agbaga, S. Patricia Becerra Supplementary Figures POS (µg) 5 5 0.1 0.1 DTT - + - + ~260 ~140 ~100 ~70 ~50 ~40 ~35 ~25 ~15 MW x 10-3 Coomassie Blue Ab-Rhodopsin Proteins in the POS samples were determined and resolved by SDS-PAGE in the same gel in two sets: one with 5 µg and another with 0.1 µg protein per lane. For each set, one sample was non-reduced and the other was reduced with DTT. After electrophoresis, the gels were cut in half lengthwise. The gel portion with 5 µg of protein was stained with Coomassie Blue and the other portion with 0.1 µg protein was transferred to a nitrocellulose membrane for immunostaining using anti-rhodopsin antibodies (as described in Methods). Photos of the stained gel and western blot are shown. The proteins of POS isolated from bovine retina had the expected migration pattern for both reduced and non-reduced conditions, and the main bands stained with Coomassie Blue comigrated with rhodopsin-immunoreactive proteins in western blots of POS proteins. Figure S1. SDS-PAGE and western blot of bovine POS A. B. C. D. E. F. G. H. I. J. Figure S2. TEM of RPE in RPE-Pnpla2-cKO mice The presence of LDs was associated with lack (Fig. S2A) of or the decreased thickness of the basal infoldings, and with granular cytoplasm, abnormal mitochondria (Fig. S2B), and disorganized localization of organelles (mitochondria and melanosomes) (Fig. S1A). In some cells, the large LDs crowded the cytoplasm and clustered together the mitochondria and melanosomes into the apical region of the cells (Figs. S2A, S2C, S2D); however, LDs number and expansion within the cells appeared to be random and their expansion could go into any direction (Fig. S2E). Normal apical cytoplasmic processes were lacking; however, degeneration in the outer segment (OS) tips of the photoreceptors was visible (Figs. S2A, S2F); . Additionally, normal phagocytosis of the OS was lacking indicating an impaired RPE phagocytosis (Figs. S2A, S2E, S2G). There were apparent unhealthy nuclei with pyknotic chromatin and leakage of extranuclear DNA (enDNA), indicating that the beginning of the necrotic process had started (Fig. S2B). Some RPE cells lacked basal infoldings, normally seen at the basal side (Fig. S2H). Occasionally some RPE cells had lighter low-density cytoplasm indicating degeneration of cytoplasmic components in contrast to the denser and fuller cytoplasm in the RPE of the littermate control (Fig. S2I, S2J). Figure S3. A. C. D. 2.5 16 24 0 25 50 75 100 Time (h) Rh od op si n re m ai ni ng (% re la tiv e to 2 .5 h ) 0.5 1 2.5 0 10 20 30 Time (h) Fr ee fa tt y ac id s (p m ol ) - POS + POS *** *** * 0.5 1 2.5 0 2 4 6 Time (h) β- hy dr ox yb ut yr at e (n m ol ) - POS + POS *** *** *** B. 0.5 1 2.5 16 24 (h) -GAPDH -Rhodopsin - 3 days 0 0.5h 1h 2.5h 16h 24h Plate cells +107 POS/ml Remove POS Add complete media Media → FFA, β-HB Cells → WB Pulse Chase Figure S3. Phagocytosis in ARPE-19 cells. ARPE-19 cells were cultured in 24-well plates for 3 days, and then exposed to POS at 1x107 units/ml for up to a 2.5-h pulse followed by an upto 24-h chase period as described in Methods. (A) Representative immunoblots of total cell lysates during pulse-chase (times indicated at the top of the blot) with anti-rhodopsin followed by reprobing with anti-GAPDH as the loading control are shown. Migration positions of rhodopsin and GAPDH are indicated to the right of the blot. Duplicate biological replicates were performed. (B) Quantification of rhodopsin from duplicate samples per condition from pulse-chase experiments at time periods indicated in the x-axis as from panel (A). Intensities of the immunoreactive bands from duplicate samples of cell lysates were determined. The percentage of the remaining rhodopsin after 16-h chase relative to rhodopsin at 2.5 h-pulse was plotted. (C-D) Levels of free fatty acids (C) and -HB (D) measured in culture media of cells incubated with and without POS for the indicated periods of time (x-axis) were plotted and shown. n = 3 Data are presented as means ± S.D. * p < 0.05, ***p<0.001. 30 60 150 30 60 150 min - POS +POS -GAPDH -Rhodopsin A. Cells on porous membranes B. Cells on porous membranes 30 60 150 0 2 4 6 8 Time (min) S ec re te d β- H B (n m ol es ) - POS + POS 30 60 150 30 60 150 min - POS +POS -Rhodopsin -GAPDH C. Cells on plastic Figure S5. Phagocytosis in ARPE-19 cells in porous membranes. ARPE-19 cells were treated with 1x107 POS/ml. (A) Representative immunoblot showing rhodopsin internalization from total cell lysates of ARPE-19 cells following 30, 60, and 150 min of POS incubation following plating in 12-well transwell inserts for 3 weeks. Cell extracts were resolved by SDS-PAGE followed by immunoblotting with anti-rhodopsin. The blot was stripped and reprobed with anti-GAPDH as a loading control. (B) Levels of B-HB secreted towards the apical membrane of ARPE-19 cells following POS incubation for 30, 60, and 150 min. Data are presented as means ± S.D. ARPE-19 cells plated on porous membranes engulf bovine outer segments To demonstrate a functional assay to study phagocytosis in ARPE-19 cells we perform the assay with confluent cells attached on porous membranes Methods: ARPE-19 cells seeded on porous membranes were incubated for 3 weeks in culturing media. Then the media was replaced with Ringer’s solution alone or Ringer’s solution containing 1 x 107 POS/ml and 5 mM glucose for the indicated time points. Rhodopsin was detected by western blotting. Rhodopsin levels in the lysates of cells incubated with POS were detected in as little as 30 min and up to 2.5 h following POS incubation, while rhodopsin was undetectable in cells without POS (Fig. S4A). B-HB levels released into the media of the apical chamber of transwells following POS incubation increased four-fold and three-fold after 60 and 150 min, respectively, while released B-HB levels from cells incubated with Ringer’s solution alone did not increase (Fig. S4B). Figure S4. siScramble siRNA A 0.0 0.2 0.4 0.6 0.8 1.0 P np la 2/ 18 S **** siScramble siRNA A 0.0 0.5 1.0 1.5 2.0 P np la 2/ 18 S **** ARPE-19 cells were transfected with siScramble siRNA control or siRNAs targeting PNPLA2 (siPNPLA2 A). RT-qPCR to measure PNPLA2 mRNA levels in ARPE-19 cells at (A) 72 h post-transfection and (B) 98.5h post transfection equivalent to pulse (2.5h) and chase (24h) was performed with siRNA duplexes (as indicated in the x-axis). Treatment of cells in panel B was as for pulse-chase (see diagram in Fig S3). PNPLA2 mRNA levels were normalized to 18S. n =3 biological replicates, each data point corresponds to the average of triplicate PCR reactions. The RT-PCR was repeated twice per biological replicate. Values that fell out of the standard curve were not included in the plot. The data shows that siPNPLA2 duplex silenced PNPLA2 in ARPE-19 at 72 h post-transfection and that silencing was maintained throughout a 2.5 h and pulse-chase of 24 h. Figure S5. A. 72h post transfection B. 98.5 h post transfection, parallel to pulse-chase aCorresponding author: S. Patricia Becerra NIH-NEI-LRCMB Section of Protein Structure and Function Bg. 6, Rm. 134 6 Center Drive MSC 0608 Bethesda, MD 20892-0608 becerrap@nei.nih.gov PEDF-R in phagocytosis 9-22-20 MS REVISED 12-29-20.pdf aCorresponding author: S. Patricia Becerra NIH-NEI-LRCMB Section of Protein Structure and Function Bg. 6, Rm. 134 6 Center Drive MSC 0608 Bethesda, MD 20892-0608 becerrap@nei.nih.gov Phagocytosis and PEDF-R Figures 9-22-20 revised 12-22-2020.pdf Slide Number 1 Slide Number 2 Slide Number 3 Slide Number 4 Slide Number 5 Slide Number 6 Slide Number 7 Supplemmentary information 9-15-20 REVISED 12-26-20.pdf Degradation of Photoreceptor Outer Segments by the Retinal Pigment Epithelium Requires Pigment Epithelium-derived Factor Receptor (PEDF-R)�Jeanee Bullock, Federica Polato, Mones Abu-Asab, Alexandra Bernardo-Colón, Ivan Rebustini, Elma Aflaki, Martin-Paul Agbaga, S. Patricia Becerra Slide Number 2 Slide Number 3 Slide Number 4 Slide Number 5 Slide Number 6 Rhodopsin in RPE of cKO and control mice Slide Number 8 Slide Number 9 Slide Number 10 Slide Number 11