Supplemental Information Online Deimination restores inner retinal visual function in a demyelinating disease Mabel Enriquez-Algeciras, Di Ding, Fabrizio Mastronardi, Robert Marc, Vittorio Porciatti, and Sanjoy K. Bhattacharya*

*To whom correspondence should be addressed. This PDF file includes: Supplemental Materials and Methods Supplemental References Supplemental Figures 1-7 Supplemental Tables 1-2 Does not include: Supplemental Movie 1 (Submitted separately) Movie 1 legend: Movie 1 This movie shows movement of the two littermate mice, one is a heterozygous transgenic mouse for DM20, the smaller variant of the myelin proteolipid protein and the other mouse is a normal non-transgenic control. Whereas the normal non-transgenic control mouse has normal movement, its transgenic littermate has a wobbly gait with unsteady movement.

Supplemental Materials and Methods online. Tissue procurement, general fixation and immunohistochemistry. The human donor multiple sclerosis tissue (brain and eye) was procured as a research gift from Dr. Wallace W. Tourtellotte at Human Brain and Spinal Fluid Resource Center, VA West Los Angeles Healthcare Center, Los Angeles, CA. The normal and glaucomatous eyes were procured from Florida Lions Eye Bank and National Disease Research Interchange, Philadelphia, PA (Details of the donors are provided in Supplemental Table 1).

The original transgenic ND4(1, 2) mice procured from The Hospital for Sick Children, Toronto, Canada as a research gift, was subjected to rederivation using services of Charles River Laboratory (Wilmington, MA), after confirmation of the genetic status, a colony is maintained at Bascom Palmer Eye Institute. The control and Myelin Oligodendrocytes Glycoprotein (MOG) immunized [with 21 peptide residue 35-55 of MOG (3) mice eyes were procured from the colony maintained at The Hospital for Sick Children, Toronto, Canada. Control CD1 (and in addition C57BL6/J, where applicable) mice were from Bascom Palmer Eye Institute, University of Miami or from Moran Eye Center, University of Utah. All animals were used under approved protocols of Institutional Animal Care and Use Committee of respective institutions.

Human tissues were fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) within 8-24 hours after death. Mice tissues were fixed into 4% paraformaldehyde in PBS or 10% formalin or in 4% paraformaldehyde, 0.2% glutaraldehyde in cacodylate buffer immediately after euthanasia. As reported previously (4) a kit was utilized for immunohistochemical detection of protein-bound citrulline (17-347; Millipore Corporation, Billerica, MA). Confocal and immunofluorescence microscopy. Human and mice eye paraffin embedded sections (8 µm) were subjected to immunohistochemical analysis and imaged on a Leica laser scanning confocal microscope (TCS-SP5, Leica, Exton, PA). For select images stained with Thy1 (catalog number SC-9163, rabbit, Santa Cruz, CA) or MAP2 (catalog number AB5622, rabbit polyclonal, Millipore) (secondary antibody coupled with Alexa 594) and anti-citrulline (secondary antibody coupled with Alexa 488) and DAPI, the area measurement in comparable regions in retina and determination of relative intensity was performed using the Leica Application Suite, Advanced Fluorescence, 1.7.0 Build 1240 software. Counting of cells was performed in equivalent regions of retinal ganglion cell layers within an equivalent area for anti-citrulline and Thy1 (or MAP2). Antibody against Syntaxin [STX01 (HPC-1)] was from Abcam, (catalog number ab3265) and antibody for γ-synuclein, NeuN and monoclonal antibodies to all PADs were procured as research gifts from Drs. Andrei Surguchov, VA Medical Center, Kansas City, Missouri; Barbara Grimpe, Miami Project to Cure Paralysis, University of Miami, Miami, Florida and Hidenari Takahara, Department of Applied Biological Resource Sciences, School of Agriculture, Ibaraki University, Ibaraki, Japan, respectively. About 10 animals in each group were analyzed for quantification. For obtaining representative images, a series of 1 µm xy (en face) images through z plane were collected and summed for an image representing a three-dimensional projection of the entire 8 µm section. Confocal microscopic panels were composed using Adobe Photoshop version 5.5.

Gabriel Gaidosh, Department of Ophthalmology, University of Miami, Miami, Florida assisted with confocal microscopy. Imaging small molecule signatures. The small molecule signatures of cells in the RGC layer were imaged by conventional bright-field imaging as mosaic tiles of 8-bit, 1388-pixel x 1036-line frames under voltage-regulated tungsten halogen light. Images were captured with a Peltier-cooled camera (Fast 1394 Qicam; QImaging, Burnaby, BC, Canada), autotiled with a montaging system (Syncroscan; Synoptics Inc., Frederick, MD). Final tiling and multimodal registration were performed with IR-tweak, a multi-platform registration application based on a Thin Plate Spline transform. More information about IR-tweak and other related microscopy image processing applications can be found at http://www.sci.utah.edu/~koshevoy/research/. Determination of retinal ganglion cells/ Validation of retinal ganglion cell identity. Control mice were euthanized by decapitation under halothane anesthesia. Eyes were rapidly perforated with a 30 gauge needle and vitreally injected to overflow with 0.1 ml fixative, enucleated, hemisected and the posterior pole fixed for 24 hours. The fixative was 4% formaldehyde, 0.1% glutaraldehyde, 0.1 M cacodylate buffer, 1mM MgSO4, 3% sucrose, pH 7.4. Eyes were processed for epoxy embedding as described previously (5) and serially sectioned at 200 nm onto 12-spot polytetrafluoroethylene (Teflon; DuPont, Wilmington, DE)– coated slides (Cel-Line; Erie Scientific Inc., Portsmouth, NH), probed with primary IgGs targeting various molecules, and visualized with silver-intensified 1.4-nm gold granules conjugated to goat anti–rabbit IgGs (5). Deiminated cells in the RGC layer were visualized with an anti-citrulline protocol selective for protein-bound citrulline (Millipore anti-citrulline kit 17-347), and small molecule signatures of ganglion cells determined with anti-GABA YY100 and anti-glutamate E100 IgGs (Signature Immunologics, Salt Lake City, Utah). Immunoprecipitation and mass spectrometry. For Immunoprecipitation (IP) carefully dissected retina from fresh mouse or pig eyes or human donor eyes (enucleated within 6-8 hours of death) were subjected to protein extraction. About 100-200 µg of retinal extract or retinal cell extract in 10mM TrisCl, buffer pH 7.0, 50 mM NaCl and 0.1% genapol was subjected to 2, 3-butanedione and antipyrine treatment in acidic environment using citrulline kit for 10-30 minutes (Millipore Corporation, Billerica, MA). [This treatment is termed as monoxime treatment and needed to chemically modify the protein-bound citrulline residues. The chemical modification adds an adduct on protein-bound citrulline residues which is recognized by anti-citrulline-adduct antibody. Throughout this work, we have used this anti-citrulline-adduct antibody which does not recognize protein-bound citrulline unless the tissue or proteins are treated with monoxime. The antibody does not recognize free citrulline]. The acidic protein mixture immediately after incubation was neutralized using 100 mM non-adjusted Tris base solution and subjected to acetone precipitation at room temperature. Isolated retinal ganglion cells, obtained following procedure described below were also used for anti-citrulline IP following 2, 3-butanedione monoxime and antipyrine treatment to confirm identified retinal lysate proteins being RGC resident. All IP experiments used a no antibody (bead only) and non-specific antibody (anti-cochlin chicken and rabbit polyclonal) controls.

IP experiments for the identification of translation complexes were performed using 5µg of recombinant his-tagged deiminated and non-deiminated REFBP2. Each batch was

incubated with 1000µg of cytosolic retinal extract (Nuclear and cytoplasmic extract kit, Pierce cat# 78833) for 1 hour, incubated with about 50 µl of Ni-NTA beads and loaded onto a mini column. The column was then washed with 50 volumes of binding buffer (PBS and 5 mM imidazole) and the bound his-tagged protein was eluted using 100 mM and 250 mM imidazole. Similar IPs were also performed with cultured RGC utilizing 5µg of REFBP2 (deiminated and control) and 100 µg of nuclear extracts. Eluted proteins were either dialyzed to remove imidazole or acetone precipitated (6) and subjected to further analyses.

For eIF4B effects on REF-ribosomal complex, the product of anti-REF IP (250 µg) was subjected to subsequent IP with anti-eIF4B. The flow through of second column was retained and passed through an oligodT column. An equal amount of SNAP-25 mRNA (250 ng) was put in two equal fractions of flow through of anti-eIF4B column and incubated at room temperature for 10 minutes. Subsequently to one fraction we added about 0.1 µg of purified recombinant eIF4B or CRALBP and subjected to a second IP with anti-REF. Purified recombinant his-tagged CRALBP was prepared from a construct following published protocols (CRALBP construct was a research gift from Dr. John Crabb, Cole Eye Institute, Cleveland Clinic Foundation, Cleveland, Ohio). The PCR reaction for SNAP-25 was carried out as has been described elsewhere in this text, amplified SNAP-25 from the reaction that was included with CRALBP but not with eIF4B.

For protein identification, SDS-PAGE fractionated protein bands were excised and digested in-situ with sequencing grade trypsin (Promega Biosciences Inc, CA). Peptides were loaded onto 3 cm YMC ODS-A spherical 5-15 µm (YMC, Milford, MA) pre-columns packed in-house in 360x100 µm fused silica and washed for 5 min to desalt prior to switching inline with the analytical column (7 cm YMC-ODS AQ spherical 5 µm particles packed in-house in 360x100 µm fused silica). Peptides were eluted along a 20 min gradient of 1-80% acetonitrile in 0.1 M acetic acid into a ThermoElectron LTQ™ or LTQ-Orbitrap™ mass spectrometer (San Jose, CA) fitted with a nanospray ionization source (Proxeon, Odense, Denmark). Spectra were collected in data dependent mode with dynamic exclusion, selecting the top five most abundant ions for CID fragmentation. Peak lists were generated using Sequest v2.7 (ThermoElectron, San Jose, CA) and submitted to a clustered version of the Sequest search engine. Spectra were searched against SwissProt fasta database (EBI, Cambridge, UK) with no more than two missed tryptic cleavages. Precursor ion tolerance was set to 2 Da and fragment tolerance to 0.5 Da. Carbamidomethylation of cysteine was fixed for analysis; variable modifications of oxidized methionine and citrullinated arginine were permitted during the search. Sequest result files were loaded into Scaffold (Proteome Software, Portland, OR) for analysis through the PeptideProphet and ProteinProphet algorithms followed by manual validation of all proteins identified with at least two peptides and at a protein confidence of greater than 90%. For determination of deiminated/citrullinated peptides, the protein mixtures were subjected to chymotrypsin digestion and tandem mass spectrometry in an Orbitrap device (ThermoFinnigan, San Jose, CA). Dr. Michael Chalmers at Scripps Florida, Jupiter, Florida and Dr. Bogdan Gugiu at Beckman Research Center, City of Hope, Los Angeles, California provided consultative assistance with mass spectrometry.

Isolation and purification of REFBP2 from cells and tissues. REFBP2 was purified from isolated hippocampal neurons, retina and brain tissues. These purifications utilized wild type mouse. For purification from tissues approximately, 0.5 mg tissues were subjected to homogenization in 10mM TrisCl, pH 7.0, 50 mM NaCl and 0.1% genapol following

established procedures(7). The supernatant was centrifuged at 12000 xg for 15 minutes and decanted in a fresh tube. The solution was precipitated with 20% ammonium sulphate and supernatant was subjected to loading on a 8 cm Mono Q column (FPLC systems, GE Healthcare Inc.) and eluted with a gradient of 10-300 mM NaCl in 10 mM TrisCl buffer pH 7.0. The REFBP2 eluted in fractions that corresponded to 200mM NaCl. The fraction was dialyzed in a buffer and equilibrated to a buffer containing 10 mM NaCl, 10mM TrisCl pH 7.0 and loaded on to a 8 cm Mono S column and eluted with a gradient of 0-300 mM NaCl. The REFBP2 was isolated in a fraction corresponding to conductivity of 150 mM NaCl. About 6 million hippocampal neurons were used for isolation of REFBP2 from cells. About 12 million neurons were subjected to fractionation of nucleus and cytosolic fraction for isolation of REFBP2 from nuclear and cytosolic fraction using a kit (Pierce Biotechnology Inc., Rockford, IL) following procedures similar to that for trabecular meshwork cells(8) in our laboratory.

Amino acid analyses and linear mode solid state ionization mass spectrometry. We performed amino acid analyses of purified recombinant and recombinant in vitro deiminated REFBP2 as well as REFBP2 purified from nuclear and cytosolic fractions (from wild type and ND4 mice; Supplemental Table 2). About 25 µg of proteins were subjected to overnight acid hydrolysis followed by analyses on a Hitachi L-8900 amino acid analyzer following established procedures. The isolated REFBP2 protein from wild type and ND4 mouse brain was also subjected to mass spectrometry on a MALDI-TOF device (Voyager DE Pro, ABI Inc.) analyzed in linear mode. This yielded m/z ratio of 23730.29 corresponding to unmodified REFBP2 (m/z) 23730.47) and a large peak for 23737.32 corresponding to modification of 7 arginine residues. Relatively lower intensity ions were also recorded between 23737.29-23737.32 indicating presence of intermediate modifications between 0 and 7 arginine residues as well. Molecular modeling of REF protein. REF was modeled using PyMol Program and NMR structure accession number 2F3J corresponding to SwissProt accession number Q9JJW6 (Supplemental Figure 2 D, E). Dr. Arun Malhotra, Department of Biochemistry and Molecular Biology, University of Miami, Miami, Florida assisted with molecular modeling. The N-terminus of REF has an arginine rich region, spanning residues 17- 48, which may present as an RNA affinity site and may potentially interfere with mRNA binding by the bona fide RNA binding motif of REFBP2, which spans between residues 74-154 (Figure 4C) of the protein. The deimination of N-terminal arginines into citrulline, a neutral charge, will result in an increase of the overall negative charge in this region, thus preventing potential interference of RNA binding to bona fide RNA binding motif by the N-terminal arginines. Another prominent change is arginine residue 24. In solution, structure residue 24 faces arginine 37, which is a highly repulsive destabilizing interaction (Supplemental Figure 2D, E). The change of arginine residue 24 into citrulline is expected to render this interaction into an energetically relative favorable one. The residue 21 modification into citrulline will encourage RNA binding. With neighboring proline (P28 and P33) residues, this is expected to result in a more rigid structure that is encouraging a situation where two arginine residues, 24 and 37, face each other. V8 protease digestion of native REFBP2 suggests the linker region to be a flexible structure. Deimination of REFBP2 arginines is expected to have consequences for structure (Supplemental Figure 2) and for function.

Gel electrophoresis and Western analyses. Gel electrophoresis (1D and 2D) were performed as described previously (7). Western transfers from 2D gels on PVDF blots were carried out following established procedures and blots were probed with antibodies to REFBP2 (catalog number ab6141, Abcam Inc., Cambridge, MA) and anti-citrulline from the kit as described above. The 2D analysis was performed to confirm lack of REFBP2 deimination in ND4 retinal lysates. Western blot analyses were performed for anti-citrulline IPs and also for anti-his (recombinant REFBP2) IPs to identify translation product complexes. Anti-citrulline IP was probed with anti-citrulline and corresponding bands were subjected to mass spectrometry. Western analyses of anti-his (for recombinant REFBP2) (catalog number AB.3517 chicken anti polyhistidine tag antibody; Chemicon Inc., Temecula, CA) were performed after transfer of IP products on PVDF membranes, which were probed with rabbit polyclonal antibodies to eIF4B, eIF4E, PABP1, Phospho-eIF4B, Phospho eIF4E (catalog numbers #3592, #9742, #4992, #3591 and # 9741 respectively from Cell Signaling Technologies, Beverly Hills, CA) All membranes bearing IP products were also probed with mouse monoclonal to Aly (catalog number ab6141, Abcam Inc., Cambridge, MA) and rabbit polyclonal anti-citrulline (Millipore #07-390) respectively. Cloning, purification, RNA isolation and binding experiments. REFBP2 clone (EMM1002-96824126) in plasmid pExpress1 procured from OpenBiosystems, Huntsville, AL, was subcloned in pQE1 vector (Qiagen, Valencia, CA) using restriction enzymes, EcoRI, XhoI and a primer set: (5’-CGTGGAGCCGAATTCATGGCCGACAAGATGGACATG-3’, 5’-CACCACCCGCTCGAGGCTGGTGTCCCTCCTTGCATTG-3’). The clone was sequenced and transformed in E. coli M15 cells, induced using 0.1 mM IPTG at between absorbance values of 0.55-0.8 at 600 nm. Recombinant his-tagged REFBP2 was purified using two rounds of Ni-NTA column (Qiagen, Valencia, CA). Briefly, the cell lysate prepared by sonication was centrifuged at 12000 xg for 15 minutes at 4ºC followed by binding of the clear soluble proteins onto a Ni-NTA column in PBS. The column was washed with 50 volumes of PBS and recombinant REFBP2 was eluted with 100 and 250 mM imidazole in the first round. In the second round the purified product of first round was extensively dialyzed using a 3500 MWCO membrane (Sigma Chemical Co., St. Louis) in PBS and subjected to bind to Ni-NTA column and re-purified eluting with only 100 mM imidazole. The final purified product to render free of imidazole was subjected to dialysis.

For binding experiments, total RNA from 3 month old ND4 and control mouse brain was isolated using Triazol method using standard protocol and quality verified by electrophoretic separation on a commercial gel (catalog number 57238;1.25% Seakem Gold Agarose in MOPS buffer, Latitude RNA precast gels, Cambrex Bioscience Rockland Inc., Rockland, ME). RNA binding experiments were performed using 10 µg of purified his-tagged recombinant non-deiminated or deiminated REFBP2 incubated with about 400 µg total RNA, subsequently the mixture was incubated with 100 µl of Ni-NTA beads and loaded onto a mini column. The column was washed with 50 volumes of binding buffer and bound protein was released using 100 mM immidazole in the binding buffer. The eluted RNA product was precipitated using carrier BSA and subsequently chloroform extracted and subjected to microarray analysis or separated on a commercial gel or in-house prepared composite 0.5% agarose-1% polyacrylamide gel. Quantitative (dissociation constant measurements)

electrophoretic mobility shift experiments utilized 10 µg recombinant or deiminated recombinant his-tagged REFBP2 and varying amounts of up to 0.5 mg of isolated purified total RNA. The confirmatory qualitative experiments for bound RNA species (SNAP-25, VAMP2) were performed using total RNA derived from retina and RGCs as well in a similar manner except using 2 µg of REFBP2 protein (recombinant and control) and 50 µg of total RNA. In vitro deimination. For in vitro deimination of recombinant REFBP2, a recombinant PAD2 was prepared. Briefly, the PAD2 clone (MHS1010-7507607) was procured from openbiosystems and cloned as a (Glutathione S transferase) GST fusion (using ligation of GST generated by PCR amplification) gene in pQE1 vector backbone. The GST column purified PAD2 was incubated under deiminating conditions(9) with dialyzed REFBP2, the product of the first Ni-NTA column described above. The recombinant REFBP2 (100 µg) was deiminated in vitro using 10 µg of recombinant PAD2 and re-purified using Ni-NTA column, dialyzed and quantified using Bradford’s method. Virus construct and PAD2 expression in animals eyes. The Thy1 promoter region is defined by 2000bp upstream of Thy1 start codon. The sequence was identified using UCSC Genome Bioinformatics database. The BAC clone with the insert of genomic DNA covering this region was obtained from Children's Hospital Oakland Research Institute (CHORI). Primers (Forward 5’ AAAAAAACGCG TAATCCAGTCCAGAAATGGGGGTG3'; Reverse 5’GTGGGGGCTAGCGGACAAAGAAAACTGCACAATA3’) including restriction sites for Mlu1 and Nhe1 were prepared to amplify the region 0-2000bp upstream of Thy1 start codon. PCR products were subjected to verification by agarose gel electrophoresis. The PCR bands were excised and extracted from the agarose gel using a QIAGEN Gel Extraction Kit (Qiagen, Valencia, CA). The concentration of gel purified DNA was determined using a UV/visible spectrophotometer. The insert and vector plionII were simultaneously digested using restriction enzymes which would excise the CMV region on the vector. The digested inserts, plionII were again subjected to gel electrophoresis, and were extracted using the Gel Extraction Kit. The vector was dephosphorylated with calf intestinal alkaline phosphatase (CIP). The digested vector and insert DNA were ligated in a ratio of 1:3 by incubating with T4 ligase at 37˚C for 2 hours. The ligated constructs were used for transformation of TOP10 competent E. coli cells. The colonies of transformants were picked and isolated DNA was subjected to confirmation for presence of insert by double digestion of the plasmid. The construct was further confirmed by DNA sequencing. The Thy1 containing pLionII was further subjected to digestion with Pme1 and Not1 restriction enzymes. PAD2 was amplified using the clone vector (open biosystem) with primers (Forward 5’ATATAAGTTTAAAC ATGCTGCGCGAGCGGACCG Reverse 5’TTTTGCGGCCGCTTACAGAGGAAAGCTGCT C3’) containing the Pme1 and Not1 restriction sites. PAD2 was ligated with the Thy1-pLionII vector using the same procedure as above. Virus production and transfection. Thy1-PAD2 pLionII and CMV-YFP pLionII constructs were transfected separately to 293 Human Embryonic Kidney (HEK) cells using FuGENE 6 Transfection Reagent (Roche - Cat. No. 11 815 091 001). Briefly, structural vector (pCI-VSVG), envelope vector ( pCPRΛEnv) and transfer vector (Thy1-PAD2 pLionII, empty

pLionII vector with thy1 promoter only or CMV-YFP pLionII) were mixed with 10:10:1 ratio. 21 µg of mixed DNA was incubated with 500µl Opti-MEM (GIBCO) containing 5 µl FuGENE and mixed with 293 HEK cells. Media was changed every 3 days. Old media was collected and filtrated with 0.4µm Super Membrane (PALL). The filtrated media was mixed with 100% PEG (Polyethylene Glycol 6000, USB Corporation, Solon, OH) according to 6:4, V:V ratio, and centrifuged for 20min at 3500rpm at 4°C. The pellet was precipitated and was re-suspended in 300µl DMEM. 10 µl of Thy1-PAD2 pLionII lentivirus was mixed with 1 µl of CMV- YFP pLionII lentivirus and the whole mixture was added to 293 HEK cells. The neuronal cells were transfected using the lipofectamineTM 2000 (Invitrogen-Cat. No. 11668-019). Ocular Injections. The mice were anesthetized with intraperitoneal ketamine (50 mg/kg) and xylazine (5mg/kg). Subretinal injections were performed under anaesthesia following established methods either using a 5 μl Hamilton syringe connecting to Ultra Micro Pump II (UMPII; World Precision Inc, Sarasota, FL) to deliver 0.5-1 μl viral construct. Reproducible injections are achieved by the UMPII device. An ointment containing antibiotics and gentamycin was applied to the injection site to prevent infection after injections.

Isolation of polysomes and ribosomes. Polyosomes were prepared from a total of 2 ×106 isolated RGC preparations that were washed with ice-cold isolation buffer R (20 mM Tris, pH 7.4, 100 mM KCl, 10 mM MgCl2, 2 mM DTT containing protease mix (0.5 mg/ml Pefabloc SC, 2 mg/ml leupeptin, 2 mg/ml pepstatin, 0.2% aprotinin and 0.5 mg/ml heparin). Isolation buffer R containing 0.5% Igepal-600 (Sigma) was used to lyse RGCs and cytoplasmic extracts were obtained after centrifugation at 25,000 × g for 15 min at 4°C. Cytoplasmic extracts so prepared were centrifuged at 100,000 × g for 3 h at 4°C for generation of a polysome-enriched pellet (P100) and a post-polysomal supernatant (S100). For preparation of ribosomes, all steps were identical except RGCs were suspended in isolation buffer R with protease mix and 100 μg/ml cycloheximide (Sigma Chemical Co., St. Louis,MO) to block at the elongation step. Cytosolic extract were prepared using 0.5% Igepal-600 and 100 μg/ml cycloheximide and cytoplasmic extracts were obtained after centrifugation at 25,000 × g for 15 min at 4°C. The clear extract was loaded onto a linear 10–50% (w/v) sucrose gradient in isolation buffer R, and centrifuged at 100,000 × g for 3 h at 4°C. Gradients were fractionated by upward displacement with 70% sucrose, and absorbance at 260 nm was monitored continuously by using a UV monitor. Ribosomal subunits (60S, 40S) were fractionated by re-suspending the polysomes in 20 mM Tris, pH 7.5, 500 mM KCl, 3 mM MgCl2, 2 mM DTT and subjecting them to centrifugation for 24 h at 100,000 × g at 4°C. The mouse or rat retinal polysome and ribosomes were also prepared. The retina from freshly euthanized animals were quickly removed and placed in ice-cold isolation buffer R containing 100 μg/ml cycloheximide. All subsequent operations were performed at 4°C. Tissues were lysed and homogenized in this buffer by using a hand held Kontes homogenizer. A supernatant (postmitochondrial fraction) was prepared by two consecutive centrifugations of the homogenate at 30,000 × g for 15 min. Triton X-100 was added to a concentration of 1% to this supernatant and incubated for 15 min. Subsequently the supernatant was layered onto a step gradient containing 7.5 ml of 0.7 M (w/v) sucrose on top of 8 ml of 1.6 M (w/v) sucrose (both in isolation buffer R), and centrifuged at 100,000 × g for 18 h. This contained polysome-enriched pellet which was resuspended in isolation buffer R

containing 100 μg/ml cycloheximide and layered onto 10–50% sucrose gradients prepared in the same buffer. The gradients were centrifuged and fractionated as described above. To dissociate ribosomal subunits (60S and 40S) from rat tissues, polysome-enriched pellets were resuspended in 20 mM Tris, pH 7.4, 500 mM KCl, 3 mM MgCl2, 2 mM DTT and 1mM puromycin (the latter was added to a final concentration of 1 mM). Samples were incubated at 37°C for 15 min and centrifuged twice for 15 min at 30,000 × g, and the supernatant was loaded on a linear sucrose gradient (10–30%) in the isolation buffer R. The gradients were centrifuged at 100,000 × g for 6 h at 4°C, and fractions were collected by upward displacement with 70% sucrose as mentioned above. Fractions corresponding to particular subunits were pooled, diluted with isolation buffer R, and subjected to centrifugation for 24 h at 100,000 × g for precipitation. RNA was isolated from different fractions using established methods and subjected to analysis. UV cross- linking experiments. UV cross-linking was performed using 0.5µg of TUNP, a RNA binding protein, (a research gift from Dr. Ralf Landgraf, Department of Biochemistry and Molecular Biology, University of Miami, Miami, Florida) and his-tagged purified REFBP2 using 250ng of SNAP-25 mRNA purified using an antisense oligonucleotide (5 ′-ATGTCTGCGTCCTCGGCCAT-3′) column coupled to beads or total mRNA. Briefly, the RNA protein mixture was incubated for 15 minutes at room temperature followed by exposure to 365nm UV light for 45 seconds. The proteins were resolved on a 4-20% gel and probed with radio-labeled oligonucleotides against SNAP-25 mRNA. The gel bands were also excised subjected to RNA extraction with Qiagen kit # 75142 and PCR amplification for a region of SNAP-25 mRNA. Statistics. All values are presented as the mean± standard deviation (SD). Comparisons among groups were made using analysis of variance followed by paired or unpaired, nonparametric Student's t test as indicated in each individual figure. Differences were considered statistically significant at p < 0.05. Other cell cultures and ELISA, siRNA experiments. The mixed cell culture was prepared from whole retina/brain cells suspension (p7) as described above. The deimination level was detected by IHC in different cell types as indicated by different cell markers such Thy 1 and GFAP. The quantification was carried out by ELISA measurements using cells purified using immunomagnetic methods following published procedures(10, 11). Briefly, protein extract (10 µg) was prepared from 104-105 cells. After treatment with acid as described above, the extract was detected using citrulline antibody. Primary cultures of astrocytes were prepared from mouse pups with suitable modification of previously established protocols for astrocyte culture (12, 13) and further purification were carried out employing published immunomagnetic methods (11). Purified deiminated MBP and CRALBP (10 µg of protein each) was used as positive and negative anti-citrulline immunoreactivity controls respectively. The 2 day old rat RGC neurons were used for siRNA experiments. The siRNA experiments were performed using siRNA against PAD2 and PAD4 sequences [Stealth Select RNAi for rat (Rattus norvegicus) PAD2 (catalog numbers, Oligo id#, RSS309706, RSS309707, RSS309705) and for PAD4 (catalog numbers, Oligo id#, RSS309711, RSS309713, RSS309712) from Invitrogen Corporation, Carlsbad, CA]. Control siRNA (catalog number

#12935-200; Invitrogen Corporation) unrelated to any known mammalian sequence was also used. Neurite lengths of RGC neurons (and also generic neurons) were determined following published procedures (14) without and with control siRNA treatment. Efficacy of PAD2 siRNA down-regulation was determined as performed previously on astrocytes. All siRNA used here were carefully evaluated and were found to down regulate greater than 85% of the PAD2 protein(15). The PAD2 siRNA affected translation of a handful of mRNA at dendrites/neurites (SNAP-25, Complexin1 and VAMP2) and at mitochondrial surface (for example, ATP5b) but no appreciable change in global protein translation was detected for 26 other mRNAs that were also sampled. The mixed population of cells that were derived from carefully dissected Retinal ganglion cell layer under a light microscope were used to stain cells with HPC1/syntaxin, and for RGC markers Thy1, γ-synuclein, or NeuN and anti-citrulline. Although we did not purify amacrine cells, however, in the mixed population of cells only cells that were positive for RGC markers (Thy1, NeuN) were also positive for anti-citrulline. Hippocampal neurons were prepared using established protocols. Low-density primary embryonic E18 rat hippocampal neuronal cultures were prepared following published protocols (Brewer et al., 1993; Banker and Goslin, 1998). Briefly, after complete removal of the meninges and microdissection to remove the hippocampus, intact hippocampi were incubated in calcium ion free Hibernate medium (BrainBits, LLC, Springfield, IL) containing 0.25% Trypsin (Cat. #15090, Invitrogen Corporation) and 600 µg/mL DNase for 15 min at 37oC. The tissue was then rinsed extensively with calcium ion free Hibernate medium supplemented with 1X B-27 (Invitrogen Corporation), followed by trituration with a fire-polished Pasteur pipette to obtain hippocampal neurons. Retinal ganglion and Hippocampal cells are generic neurons which were used for siRNA experiments described here. Detection of SNAP-25 and β-tubulin mRNA in laser captured neurites of mouse RGCs was performed using a non-radioactive quantifiable detection system using probes directly labeled with horseradish perooxidase (HRP) or biotinylated probes (16). The SNAP-25 antisense oligonucleotide 5'-ATGTCTGCGTCCTCGGCCAT-3' (against Rattus norvegicus NM_011428) and mis-sense control oligonucleotide 5'-ATCTCAGCGTGCTTCGCCTT-3' and an additional control oligonucleotide consisted of the same base composition as the antisense oligonucleotide but arranged in a different sequence (5'-TAGCTTCGGCTCGCTCGCTA-3') was used as probe. For β-tubulin a biotinylated or direct HRP labeled probe sequence 5’-TCTCGGCCTCGGTGAACTC-3’ and an antisence control sequence 5’-AAGGCCTTCCTGCACTGGTA-3’ (against Mus muculus NM_023279) was utilized for quantitative estimations. Supplemental references. 1. Johnson, R.S., Roder, J.C., and Riordan, J.R. 1995. Over-expression of the DM-20

myelin proteolipid causes central nervous system demyelination in transgenic mice. J Neurochem 64:967-976.

2. Mastronardi, F.G., Ackerley, C.A., Arsenault, L., Roots, B.I., and Moscarello, M.A. 1993. Demyelination in a transgenic mouse: a model for multiple sclerosis. J Neurosci Res 36:315-324.

3. Sun, D., Whitaker, J.N., Huang, Z., Liu, D., Coleclough, C., Wekerle, H., and Raine, C.S. 2001. Myelin antigen-specific CD8+ T cells are encephalitogenic and produce severe disease in C57BL/6 mice. J Immunol 166:7579-7587.

4. Bhattacharya, S.K., Sinicrope, B., Rayborn, M.E., Hollyfield, J.G., and Bonilha, V.L. 2008. Age-related reduction in retinal deimination levels in the F344BN rat. Aging Cell 7:441-444.

5. Jones, B.W., Watt, C.B., Frederick, J.M., Baehr, W., Chen, C.K., Levine, E.M., Milam, A.H., Lavail, M.M., and Marc, R.E. 2003. Retinal remodeling triggered by photoreceptor degenerations. J Comp Neurol 464:1-16.

6. Patel, N., Solanki, E., Picciani, R., Cavett, V., Caldwell-Busby, J.A., and Bhattacharya, S.K. 2008. Strategies to recover proteins from ocular tissues for proteomics. Proteomics 8:1055-1070.

7. Bhattacharya, S.K., Crabb, J.S., West, K.A., Gu, X., Sun, J., Bonilha, V.L., Smejkal, G., Shadrach, K., Hollyfield , J.G., and Crabb, J.W. 2006. Optic nerve fractionation for proteomics. In Separation Methods in Proteomics. G.B. Smejkal, and A. Lazarev, editors. Boca Raton: CRC Press. 135-155.

8. Picciani, R.G., Diaz, A., Lee, R.K., and Bhattacharya, S.K. 2009. Potential for transcriptional upregulation of cochlin in glaucomatous trabecular meshwork: a combinatorial bioinformatic and biochemical analytical approach. Invest Ophthalmol Vis Sci 50:3106-3111.

9. Senshu, T., Sato, T., Inoue, T., Akiyama, K., and Asaga, H. 1992. Detection of citrulline residues in deiminated proteins on polyvinylidene difluoride membrane. Anal Biochem 203:94-100.

10. Tezel, G., and Wax, M.B. 2000. Increased production of tumor necrosis factor-alpha by glial cells exposed to simulated ischemia or elevated hydrostatic pressure induces apoptosis in cocultured retinal ganglion cells. J Neurosci 20:8693-8700.

11. Wright, A.P., Fitzgerald, J.J., and Colello, R.J. 1997. Rapid purification of glial cells using immunomagnetic separation. J Neurosci Methods 74:37-44.

12. Zhuang, Z., Yang, B., Theus, M.H., Sick, J.T., Bethea, J.R., Sick, T.J., and Liebl, D.J. 2010. EphrinBs regulate D-serine synthesis and release in astrocytes. J Neurosci 30:16015-16024.

13. Gregorios, J.B., Mozes, L.W., Norenberg, L.O., and Norenberg, M.D. 1985. Morphologic effects of ammonia on primary astrocyte cultures. I. Light microscopic studies. J Neuropathol Exp Neurol 44:397-403.

14. Shelly, M., Cancedda, L., Heilshorn, S., Sumbre, G., and Poo, M.M. 2007. LKB1/STRAD promotes axon initiation during neuronal polarization. Cell 129:565-577.

15. Bhattacharya, S.K., Crabb, J.S., Bonilha, V.L., Gu, X., Takahara, H., and Crabb, J.W. 2006. Proteomics implicates peptidyl arginine deiminase 2 and optic nerve citrullination in glaucoma pathogenesis. Invest. Ophthalmol. Vis. Sci. 47:2508-2514.

16. Pollard-Knight, D., Read, C.A., Downes, M.J., Howard, L.A., Leadbetter, M.R., Pheby, S.A., McNaughton, E., Syms, A., and Brady, M.A. 1990. Nonradioactive nucleic acid detection by enhanced chemiluminescence using probes directly labeled with horseradish peroxidase. Anal Biochem 185:84-89.

CD1 ND4

A

MAP2

NeuN

Brn3b

0

20

40

60

80

100

120

RG

C n

umbe

r (in

0.2

squ

are

mm

)

Thy1 MAP2 NeuN Brn3b

B

CD1

ND4

PSD 95

PSD 95

Syntaxin-1

Syntaxin-1

C D

E F

0

0.2

0.4

0.6

0.8

1.0

PSD95 Syntaxin-1

Rel

ativ

e ab

sorb

ance

(405

nm)

*

*

G

Rel

ativ

e ab

sorb

ance

(405

nm

)

H

I

CD1 ND4 0

0.2

0.4

0.6

0.8

1.0

*

J

Supplemental Figure 1 Representative staining of neuronal markers for retina ganglion cells in the GCL. (A) Representative CD1 and ND4 mouse retinal GCL layer stained for the markers as indicated. Bar= 75 µm (B) RGC numbers for the markers as indicated in 0.2 square mm area of retinal sections. Mean SD were calculated from three sections per individual donor for a total of 5 donors in each group. (C, E, D, F) Representative detection of synaptic markers (PSD95 and Syntaxin-1) in CD1 and ND4 mouse retina as indicated. Bar= 20 µm. (G) ELISA analyses of 10µg of retinal protein extract, the immunoreactivity as absorbance at 405nm for PSD95 and Syntaxin-1 from control (hollow bars) and ND4 mice (solid bars) as indicated. (H, I) Representative staining of central retina from a CD1 (H) and ND4 (I) mouse using anti-SNAP-25 antibody. Bar= 100 µm. (J) ELISA analysis for relative amounts of SNAP-25 in retinal extract (10 µg). Absorbance at 405nm from control (hollow bars) and ND4 mice (solid bars) as indicated. Mean ± SD from at least three independent measurements (G, J) have been presented, data were subjected to two-tailed paired t-test with respect to controls (B, G, J), *p<0.05.

Rel

ativ

e in

tens

ity

A

m/z

B

R/R/N/R/P/Ab

y 279.6349.97447.58603.43717.44873.591029.89

604.37874.59

271.3 328.25 424.82 581.52 737.72 851.83 1008.15 1105.48

R/R/N/R/P/Ab

y 279.6349.97447.58603.43717.44873.591029.89

604.37874.59

271.3 328.25 424.82 581.52 737.72 851.83 1008.15 1105.48C

e D

24 37 24 37

Supplemental Figure 2 Identification and modeling of REFBP2 deimination sites. (A) spectrum of RRVNRGGGPRRNRPA from the pepsin digested REFBP2. The deiminated and wild type peptide co-eluted and were often found in equal ratios. The magnified view of Y7 ion is shown in the inset. (B) The Y5 and Y7 ions showing M+1Da ions. (C) The experimental b and y ions of the six amino acid sequence containing the deimination sites. The identified M+1Da ion corresponding to Y-ions have been shown as underlined italicized numbers. (D) The model of the entire REFBP2 with modified residues identified using ball and stick. (E) The charge distribution in native (unmodified) REF protein. Arginine residues 24 and 37 are face to face (arrow), which is a highly repulsive destabilizing interaction. The N- and C-terminal of the protein have been indicated. The probably reduction in repulsive destabilization due to arginine modification in residues 24 and 37 has been illustrated.

E

N-

C-

N-

C-

Supplemental Figure 3 Select REFBP2 bound RNAs from microarray analysis and reduction in levels of SNAP25 and Vamp2 expression after REFBP2 shRNA transfection. (A) Identification of select RNAs that bound to REFBP2 using Affymatrix Mouse Genome 430 2.0 Array chip. Microarray data has been submitted to GEO database, accession number GSE11843. Bound mRNA to deiminated REF (hollow bar) and non-deiminated REF (solid bar) are compared by signal intensity relative to starting materials. The (*) indicates mRNA that belongs to SNARE complex or neuronal membrane vesicles. (B) Reduction in levels of SNAP-25 and Vamp2 protein expression after Refbp2 shRNA transfection. SNAP-25 and Vamp2 expression at dendritic sites in primary neurons was compared between control (hollowed bars) and transfected (solid bars) using ELISA analyses. TGF-beta2 was used as control. (C) Determination of levels of SNAP-25 relative to β-tubulin mRNA in laser captured control and ND4 retinal ganglion cell neurites. The mRNA was extracted from pools of 1000 laser captured neurite for each experiment and probed for SNAP-25 and β-tubulin using non-radioactive labeling. Neurites were captured from isolated cells. The bars shows the ratio of SNAP-25/β-tubulin signals. In panels B, C, all comparisons were made from at least three independent measurements, mean SD has been presented. All data were subjected to two-tailed paired t-test with respect to controls, *P≤0.05.

Relative levels -500 0 500 1000 1500 2000 2500 3000 3500 4000 4500

Arg2

Ldhb

Doc2a

Bet1l*

Gosr2*

Napa*

Napb*

Cplx3*

Ykt6*

Snap25*

Vti1a*

Stx16*

Vamp2*

Vti1b*

Cplx1* 0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

SNAP-25 Vamp2 Refbp2 TGF-beta2

Rel

ativ

e ab

sorb

ance

*

*

*

A B

SNAP

25/B

eta

Tubu

lin ra

tio

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

ND4

*

C

Control

cit

B

Merged

C

Rel

ativ

e ab

sorb

ance

(405

nm

)

0

0.2

0.4

0.6

0.8

Control +shRNA

D

*

Rel

ativ

e ab

sorb

ance

(405

nm

)

Untreated +PAD2

E

0

0.2

0.4

0.6

0.8

*

*

REF

A

Supplemental Figure 4 Localization of deiminated REF in cultured wild type mice primary RGCs (A) REF in elongated neurites; anti-REFBP2, Alexa 488 (green) (B) detection for deimination (cit) anti-citrulline-monoxime adduct, Alexa594 (red) and (C) merged picture of DAPI (blue) with anti-REFBP2 and anti-citrulline-monoxime adduct immunoreactivity. Arrows indicate the co-localization of REF with deimination in neurites. (D) ELISA analyses on equal amounts (10 µg) of neuronal cell extracts using anti-REFBP2 antibody from control and cells treated with shRNA against REFBP2. Average absorbance at 405nm corrected for background (no primary antibody) from three independent experiments standard deviation has been shown. (E) ELISA analyses on 10 µg of retina (hollow bars) and optic nerve (hashed bars) extracts using anti-SNAP-25 antibody from untreated ND4 mice and PAD2 injected ND4 animals (n= 3 in each group). Average absorbance at 405nm corrected for background (no primary antibody) from three independent experiments± standard deviation has been shown. All data were subjected to two-tailed paired t-test with respect to controls, *P≤0.05.

Supplemental Figure 5 Enlarged panels of Figure 1 C, D and I, J

I C

J

GCL

INL

ONL

Normal

GCL

ONL

INL

Multiple sclerosis

D

I

ND4 control

H G GCL

IPL

INL

OPL

ONL

PR Normal control

ND4 Deimination Normal Deimination

F E

Supplemental Figure 6 Enlarged panels of Figure 1 E, G and F, H

Supplemental Figure 7 Enlarged and separated panels of Figure 4 I and J. The staining with anti-Citrulline and anti-REF antibody are as indicated.

I

L

J

K

Cit

REF

Merged

DAPI

HSB #* Age Gender Structure Autolysis Grams Neuropathological Diagnoses

2778 61 F Plaque, Parietal 38.0 0.8 Multiple Sclerosis2946 59 M Plaque, Parietal 15.0 0.7 Multiple Sclerosis3185 50 M Plaque,

Temporal13.8 0.8 Multiple Sclerosis, Chronic

MS plaque formation3816 47 F Plaque, Frontal 20.7 1.0 Multiple Sclerosis, Chronic-

active MS plaque formation3840 61 F Plaque, Frontal 22.8 1.0 Multiple Sclerosis, Active

MS plaque formation2096 76 M Plaque, Occipital 25.0 0.4 Multiple Sclerosis

2743 59 F Plaque, Occipital 23.0 1.2 Multiple Sclerosis

3010 50 F Plaque, Frontal 15.0 1.4 Multiple Sclerosis3025 76 M Plaque, Parietal 23.0 1.3 Multiple Sclerosis3161 51 F Plaque, Occipital 19.5 1.4 Multiple Sclerosis, Chronic

active MS plaque formation, Hypoxia, acute

HSB # Age Gender Structure Autolysis Grams Neuropathological Diagnoses

2765 51 F Eye, Left 9.0 9.6 Multiple Sclerosis

2771 64 F Eye, Left 17.3 8.5 Multiple Sclerosis

3022 69 F Eye, Left 52.0 18.7 Multiple Sclerosis3025 76 M Eye, Left 23.0 13.4 Multiple Sclerosis

3056 61 M Eye, Left 10.0 14.2 Multiple Sclerosis

HSB #* Age Gender Structure Autolysis Grams Neuropathological Diagnoses

3221 90 M NAWM, Frontal 17.8 1.0 Normal Aging3236 76 M NAWM, Frontal 16.3 0.7 Normal 3276 54 M NAWM, Frontal 19.0 1.1 Normal, Atherosclerosis,

basilar artery

3280 81 F NAWM, Frontal 11.3 1.0 Normal, Atherosclerosis

Normal donors

Supplemental Table 1: Donor information

Multiple Sclerosis (MS) donorsBrain

Eyes

Brain

3322 90 F NAWM, Frontal 11.8 0.7 Normal3346 91 F NAWM, Frontal 10.0 1.2 Normal, cerebroVascular

Disease, basilar artery, atherosclerosis

3371 52 M NAWM, Frontal 16.0 0.9 Normal3397 72 F NAWM, Frontal 19.5 1.4 Normal3401 82 M NAWM, Frontal 14.0 1.3 Normal Aging3406 72 F NAWM, Frontal 20.1 1.2 Normal

FLEB#* Age Gender Race Autolysis Grams Neuropathological Diagnoses

0804 014 82 M W N/A N/A No history of Neurological or ocular disease

0706-047 49 M W N/A N/A No history of Neurological or ocular disease

0706-046 76 M W N/A N/A No history of Neurological or ocular disease

0704-046 81 F W N/A N/A No history of Neurological or ocular disease

0703-017 87 M W N/A N/A No history of Neurological or ocular disease

NI#* Age Gender Race Time (h) C/D** Clinical Diagnosis/Scaling***

1 59 M W 9 0.8 Glaucoma (POAG); severity= 2

2 55 F W 4.5 0.7 Glaucoma (POAG); severity= 2

3 60 F W 4 0.9 Glaucoma (POAG); severity= 3

4 79 F W 10 0.7 Glaucoma (POAG); severity= 2

* Our internal sample identification number, W indicates Caucasian. **C/D = Cup to disc ratio. POAG= primary open angle glaucoma. Clinical diagnosis and Glaucoma scaling is based on a static perimetry threshold test (24-2), glaucomatous hemifield test. No significant decrease in retinal ganglion cell (RGC) numbers were detected using Thy1, Brn3b, MAP2 and NeuN staining in individual retinas of comparable age between normal and MS groups (see Supplemental Figure 1). Averages were calculated from three sections each of an individual donor (5 donors for each group).

Eyes

Glaucoma Donor Eyes

Arginine residues

Approx. %

Citrulline residues

Approx. %

Recombinant* His-tagged REF 26 11.9 0 0.0Recombinant His-tagged REF; 1h 20 9.2 6 2.8Recombinant His-tagged REF;10h 11 5 15 6.4

Unmodified REFBP2** 26 11.9 0 0.0Nuclear REFBP2 23 10.5 1 0.4Cytosolic REFBP2 16 7.2 7 3.2

CD1 derived REFBP2 16 7.2 7 3.2ND4 derived REFBP2 22 11 3 1.3

*The recombinant REF (REFBP2) had a tag: MGHHHHHHHHHHSSGHIDDDDKH. ** Theoretical estimates. The amino acid analyses were performed on a Hitachi L-8900 amino acid analyzer. The analysis was performed using 25µg of protein in a 50µl volume. Approximate molar percentage has been shown. The purified REFBP2 is derived from neurons from animals as indicated. Experimental result is derived from at least three animals. The concurrent values were used to calculate arginine and citrulline residues.

Supplemental Table 2: Amino acid analyses of Recombinant and purified REF proteins

Analyses of recombinant and in vitro modified protein

Analyses of in vivo purified protein from CD1 mice

Analyses of cytosolic purified protein from control CD1 and ND4 mice