Preparation of human and mouse lenses.
All experiments with mice were performed per protocols approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Alabama at Birmingham (Protocol no. 130208393). Mice were housed in a pathogen-free environment at the facility of the University of Alabama at Birmingham and dissected lenses were frozen and shipped for analysis at Oregon Health & Science University. Lenses from 3-month-old transgenic mice expressing human αA-crystallin containing a wild type (WT) sequence or human αA-crystallin containing a N101D mutation were used. Thawed lenses from single mice of 3-months of age were suspended in 0.5 ml of 50 mM ammonium bicarbonate buffer and probe sonicated for 20 sec x 2 on ice at a power setting of 2 using a Fisher Scientific Model 60 Sonic Dismembranator. Samples were centrifuged at 14,000 x g for 10 min and the supernatant removed. The pellets were resuspended in 0.2 ml of the above ammonium bicarbonate buffer, a BCA protein assay performed on both soluble supernatant and resuspended pellet fractions using BSA as a standard (Thermo Scientific), and samples stored at -80°C until further processing.
Lenses from human donors were obtained from The Lyon Eye Bank of Oregon (approved through the Institutional Review Board at Oregon Health & Science University). Lenses were dissected using a posterior approach, decapsulated, and stored at -70°C. Thawed lenses from 5-day, 23-day, 18 month, and 18-year old donors were homogenized in 1 ml of 20 mM phosphate, 150 mM NaCl, 1 mM EDTA (pH 7) buffer, centrifuged at 20,000 x g for 30 min, the supernatant removed, a BCA protein assay performed, and 100 µg portions dried by vacuum centrifugation.
Production of unlabeled and 15N-labled 6X His tagged SUMO crystallins.
Plasmids coding for the major human crystallins (αA, αB, βA3, βA2, βA4, βB1, βB2, βB3, γA, γB, γC, γD, and γS) were obtained from the DNASU plasmid repository at Arizona State University, except for γA-crystallin, which was purchased from OriGene. Coding sequences were PCR amplified and subcloned into pE-SUMO plasmid vector (LifeSensors, Inc. Malvern, PA, Cat. No. 1001A (Vector, Amp) using the methodology recommended by the manufacturer to produce chimera proteins with a N-terminal 6 X His tagged SUMO protein fused to each crystallin. PCR primers for subcloning into the pE-SUMO vector were designed so that each crystallin could be expressed either with or without N-terminal Met residues, matching the state of the proteins observed in vivo by mass spectrometry(7, 27) and as shown in Supplementary Fig. 1 for βA2, γA, and γB-crystallins. pE-SUMO crystallin plasmids were used to transform One Shot™ TOP10 chemically competent cloning E. coli cells using the method recommended by the manufacturer (Thermo Fisher Scientific, Cat. No. C404003). Sequences of each crystallin were then confirmed and plasmids used to transform One Shot™ BL21 Star™ (DE3) chemically competent E. coli cells for protein expression (Thermo Scientific, Cat. No. C601003).
For unlabeled protein purification, transformed BL21 cells expressing each pE-SUMO crystallin were grown at 37°C in LB broth (Lennox) (Sigma, Cat. No. L3022) containing 100 µg/ml ampicillin to approximately 0.5 OD at 600 nm. Cells were then induced for 3 hours following addition of 1 mM IPTG, followed by centrifugation at 3,000 x g for 15 min and pellets frozen at -80°C. For production of 15N labeled proteins, 5 ml of LB broth + ampicillin was inoculated with pE-SUMO expressing BL21 cells from frozen glycerol stocks and cells grown 5-6 hours with shaking at 37°C. Cells were then pelleted at 2500 x g for 10 min, washed with 1 ml of 15N media, pellets resuspended in 1 ml of 15N media, the suspension added to 50 ml of 15N media + ampicillin, and cells grown overnight at 25°C. The next morning, the 50 ml cell suspension was added to 1 L of 15N + ampicillin media and cells grown at 37°C with shaking for approximately 5-8 hours until they reached approximately 0.5 OD at 600 nm. They were then induced with 1 mM IPTG for 4-5 hours, harvested by centrifugation at 3,000 for 15 min and pellets frozen. 15N media was produced as previously described (20) using 15NH4Cl (> 98 atom %, Millipore Sigma, Cat. No 299251).
Frozen pellets from both 15N labeled and unlabeled cells were then subjected to 3 freeze/thaw cycles and suspended by probe sonication in 10 ml lysis buffer per g pellet containing 50 mM Tris (pH =8), 100 mM NaCl, 4 µg/ml DNAse I (Millipore Sigma, Cat No. DN25), 0.2 ug/ml RNAse A (Millipore Sigma, Cat. No. R5125), and 1 Pierce protease inhibitor mini tablet, EDTA Free (Thermo Scientific, Cat. No. A32955). Following rotation at room temperature for 30 min, the lysed cells were centrifuged at 20,000 x g for 30 min at 4°C and the supernatant and pellet recovered. SUMO-crystallins were then purified using HisPur™ cobalt spin columns (Thermo Scientific, Cat. No. 89969) by gravity flow using the method recommended by the manufacturer. Initial experiments indicated that all SUMO crystallins expressed in LB media remained largely in the soluble fraction following cell lysis, except βA4-crystallin, which was instead recovered in the pellet following centrifugation. This pellet was resuspended by probe sonication in the recommended HisPur column equilibration buffer containing 8 M deionized urea to solubilize the βA4-crystallin. Following this solubilization, the extract was subjected to an additional 20,000 x g centrifugation for 30 min at 4°C before cobalt column purification in buffers containing 8 M urea.
Following expression of crystallins in normal media, all were expressed again, but in 15N media. This again resulted in insoluble SUMO tagged βA4, as well as rendering largely insoluble Sumo tagged αA, αB, and βB3, all three of which were extracted and purified in 8M urea containing buffers. Following loading and washing of cobalt columns, SUMO tagged crystallins were eluted in 3-4 column volumes and pooled based on either SDS-PAGE or whole mass spectrometric analysis of fractions. SDS-PAGE was performed using Novex 10% tricine, 1 mm, 12 well gels (Cat No. EC66752BOX), MES running buffer (Cat. No. NP0002), and SeeBlue Plus2 prestained molecular weight markers (Cat. No. LC5925) (Invitrogen).
Whole masses of expressed crystallins were measured by injecting approximately 1 µg of protein onto a micro trap cartridge (Optimize Technologies, Cat. No. 10-04815-TM) at a flow rate of 50 µl/min in 98% mobile phase A containing 0.1% formic acid, and 2% mobile phase B containing acetonitrile (ACN) and 0.1% formic acid. After 2 min, the flow from the trap was diverted to a 1 x 75 mm PoroShell 300SB-C18 column (Agilent, Cat No. 661750-902) and protein eluted by increasing mobile phase B from 2-50% over 8 min, 50-90% over 2 min, holding at 90% for 2 min, and re-equilibration at 2% B for 5 min. Masses of eluted protein were determined by on-line electrospray ionization using an LTQ Velos Pro linear ion trap (Thermo Scientific) fitted with a HESI-II probe, 5.8 kV source voltage, 325°C ion transfer tube temperature, and sheath gas setting of 15. The instrument acquired full MS scans in profile mode over a range of m/z = 350-2000 in enhanced resolution mode while averaging 10 microscans. Spectra acquired during elution of protein peaks were averaged and deconvoluted using the Manual Respect module for isotopically unresolved data in Protein Deconvolution 4.0 software (Thermo Scientific).
Following mass analysis, crystallins were concentrated by ultrafiltration using Amicon Ultra-15 centrifugal filter units with a 10,000 MW cut off (Sigma Millipore, Cat No. UFC901024) and a BCA protein assay performed (Thermo Scientific). For cleavage of the SUMO tag from crystallins, proteins were incubated with 6X His tagged ulp-1 SUMO protease at a 1:50 ratio of protease:substrate following the addition of 2 mM DTT. The incubation was performed for 1 hour at 30°C, then overnight at 4°C. SUMO tagged crystallins purified in 8M urea were diluted to 2M urea concentration before cleavage. 6X His tagged ulp-1 was purified to over 90% purity by expression from B21(DE3) cells as described above for SUMO tagged crystallins, followed by storage at -80°C in 50 mM Tris, pH 7.5, 150 mM NaCl, 5 mM DTT,10% glycerol. Plasmid for 6X His tagged ulp-1 was obtained from Addgene #31122. 15N-labled 6 X His tagged SUMO crystallins were used as standards without removing SUMO tags. Tag removal was performed on unlabeled crystallins to test the methodology to produce untagged crystallins for future structure/function studies, while tag removal from 15N-labeled crystallins was performed to only investigate their molecular weights as a quality control measure before their use as standards.
Quantification of 15N-labled crystallin standards.
SDS was added to each 15N-labeled crystallin standard at a final 5% concentration and each dialyzed overnight against a solution containing 5% SDS using 3.5K MWCO Slide-A-Lyzer® G3 Dialysis Cassettes (Thermo Scientific, Cat No. A52966) and the approximate protein concentrations of the dialyzed proteins determined by BCA assay. Since not all 15N labeled crystallins were purified to homogeneity, they were instead quantified by addition of a BCA assayed, homogeneously purified unlabeled SUMO tagged βB2 crystallin. Two µg of unlabeled SUMO βB2 was added as a light standard to approximately 2 µg of each heavy (15N labeled) SUMO tagged crystallin, the mixtures trypsin digested using S-trap™ micro columns (Protifi, Cat. No. C02-micro) using the method recommended by the manufacturer, and 50 ng of each digest analyzed by LC/MS using an Orbitrap Eclipse mass spectrometer (Thermo Scientific) as described below in the mass spectrometric analysis of peptide digests section, except using an overall 60 min method, 5-35% acetonitrile gradient, and parallel reaction monitoring method to target light and heavy 15N labeled forms of doubly charged SUMO tryptic peptides 37-47 (VSDGSSEIFFK) and 74-80 (FLYDGIR). Precursor MS scans were performed at 120,000 resolution, m/z 400-900 range, maximum injection time (MIT) 246 ms, automatic gain control (AGC) target = 4 x 10E5, run start EASY-IC activated, and data collected in profile mode. Targeted MS2 scans used a quadrupole isolation window of 1.6, high-energy collision dissociation (HCD) at 30%, MIT = 22 ms, AGC = 5 x 10E4, 15,000 Orbitrap resolution, and centroid data collection. Data analysis and export of heavy to light ratios of integrated peptide peaks was performed using Skyline (Version 23.1.0.455). Both precursor and 3 intense fragment ions were used for ratio calculations. Three independent tryptic digests were performed for each crystallin to estimate the precision of the 15N crystallin standard quantifications. To create the Skyline method a spectral library for SUMO tryptic peptides was produced from a S-trap tryptic digest of purified SUMO and a data-dependent acquisition (DDA) LC/MS analysis using the Orbitrap Eclipse instrumental setup described below by injecting 25 ng of digest on the system. The DDA analysis used a 90 min method with a 7.5-30% acetonitrile LC gradient and identical precursor MS settings as above, except using a m/z 375-1500 scan range. Data-dependent acquisition MS2 scans used the MIPS filter set to peptide, charge state inclusion 2-7, exclusion duration = 60 sec, high and low mass tolerances of 10 ppm, isotope exclusion on, minimum intensity threshold of 5 x10E3, quadrupole isolation window = 1.6, HCD fragmentation at 35% collision energy, rapid scan rate in the ion trap, MIT = 35 ms, AGC target 2 x 10E3, no scan averaging, and centroid data collection. The resulting MS2 scans were searched using Sequest software within the Protein Discoverer 1.4 software package (Thermo Scientific). Searches used a UniProt E. coli canonical reference proteome downloaded in March, 2024 containing 4362 entries, supplemented with the SUMO sequence. A precursor mass tolerance of 1.25 Da, fragment ion tolerance of 1.005 Da, a static modification of 57.021 applied to cysteine, and a dynamic modification of 15.995 applied to methionines. A reverse sequence database search strategy was used to estimate peptide false discovery rates using Percolator, and peptides with q-scores <0.05 were used to create a biblioSpec library in Skyline.
Preparation of digests from human lenses.
Each quantified 15N-labled crystallin standard was then diluted and mixed in proportions approximately matching relative abundances of crystallins found in the lens of the 23-day old donor based on iterative experiments to achieve 1:1 light to heavy ratios for each crystallin in initial digests. One hundred µg dried portions of soluble human lens proteins were then dissolved in 100 µl of 5% SDS by vigorous shaking for 5 min and the protein content of each determined again using a BCA assay performed in quadruplicate. Based on this assay, the soluble protein from lenses of 5-day, 23-day, 18-month, and 18-year-old donors were diluted to 0.25 µg/µl in 5% SDS and 24 µl of each (6 µg) mixed with 6.7 µg of 15N labeled standard in a 16 µl volume containing all 13 crystallin species and each diluted to a final 75 µl volume by addition of 5% SDS, 50 mM triethyl ammonium bicarbonate (TEAB). Samples were reduced by adding 3.4 ul of 0.5M DTT and heating at 95°C for 5 min, and then alkylated by adding 6.8 µl of iodoacetamide and incubation for 30 min at room temperature in the dark. Each sample was then acidified by adding 8.5 µl of 12% phosphoric acid, followed by addition of 562 µl of S-trap binding/wash buffer containing 100 mM TEAB, 90% methanol. The precipitated protein suspension was applied to S-trap™ micro columns and washed using centrifugation, as recommended by the manufacturer. Each sample was digested by applying 2 µg of sequencing grade modified trypsin (Promega, Cat. No. V5117), and incubation for 2 hours at 47°C. Peptides were recovered by successive washing of S-traps using 40 µl volumes of 1) 50 mM TEAB, 2) 0.2% formic acid, and 3) 50% acetonitrile. The combined eluents were then dried by vacuum centrifugation.
Preparation of digests from mouse lenses.
Based on a BCA assay of soluble and insoluble protein isolated from 3-month-old mice expressing either WT human αA, or αA N101D mutant protein, 50 µg of protein was removed, the volume adjusted to 50 µl by addition of 50 mM ammonium bicarbonate, and 50 ul of 10% SDS added. Samples were then vortexed, and an additional BCA assay performed. Based on this assay, 12 µg of protein was removed,1.36 µg of 15N labeled human SUMO- αA standard added, and the final volume set to 75 µl by addition of 5% SDS, 100 mM TEAB buffer. The samples were then reduced and alkylated, trypsin digested using S-traps as described above, and recovered peptides dried by vacuum centrifugation.
Mass spectrometric analysis of peptide digests.
Peptide digests containing 15N labeled crystallin standards were analyzed by liquid chromatography-mass spectrometry (LC-MS) using a Dionex NCS-3500RS UltiMate RSLCnano UPLC system for peptide separation and an Orbitrap Eclipse Tribrid instrument with an EasySpray nano source for mass analysis (Thermo Scientific). Peptides were dissolved by shaking in 5% formic acid and 500 ng of human or mouse lens digests were injected onto an Acclaim PepMap 100 μm x 2 cm NanoViper C18, 5 μm trap column at a 5 µl/min flow rate using a loading pump and 2% acetonitrile (ACN), 0.1% formic acid mobile phase. After 5 min of loading and washing, the trap column was connected to a PepMap RSLC C18, 2 μm, 75 μm x 25 cm EasySpray column (Thermo Scientific) using a switching valve. Peptides were then separated using a 2–30% ACN gradient over 60 min, 30-98% ACN gradient over 1 min, 98% ACN for 5 min, 98-2% ACN over 1 min, and 2% ACN for 18 min for equilibration (90 min total) in a mobile phase containing 0.1% formic acid at a 300 nl/min flow rate. Survey MS scans from m/z = 380-985 were collected using the instrument’s Orbitrap mass analyzer at a resolution of 60,000, with maximum injection time (MIT) = 100 ms, normalized automatic gain control (AGC) target at 4x10E5, the run start easy internal calibration (IC) feature activated, and data collected in profile mode. MS2 peptide fragment spectra were collected by data-independent acquisition (DIA) using a quadrupole 10 m/z isolation window with 1 m/z overlap, HCD fragmentation with 30% collision energy, 15,000 Orbitrap resolution, 145-1450 m/z scan range, 200% normalized AGC target, MIT = 40 ms, centroid data collection, 60 scan events, and loop count = 30. Initial data-dependent acquisition runs of human and mouse lens digests were performed to create spectral libraries of crystallin derived tryptic peptides. These runs used a similar chromatographic method and identical instrument as used in the DIA method above, with digests of soluble proteins from 23-day old human, and 3-month-old mouse lenses, except with the following instrument parameters. Survey MS scans were from m/z 375-1500 in the Orbitrap mass analyzer at 60,000 resolution, MIT = 118 ms, normalized AGC target as above, run start EASY-IC activated, and profile mode data. MS2 peptide fragment ion spectra were collected using data-dependent acquisition (DDA), with MIPS filter set to peptide, charge state inclusion 2-7, exclusion duration = 60 sec, high and low mass tolerance of 10 ppm, isotope exclusion on, minimum intensity threshold of 5 x10E4, quadrupole isolation window = 1.6, HCD fragmentation with 30% collision energy, 15,000 Orbitrap resolution, MIT = 22 ms, and AGC target = 5 x10E4. MS2 scans from these DDA runs were searched using Sequest software within the Protein Discoverer 1.4 software package (Thermo Scientific). Searches of the human lens digest used a UniProt human canonical reference proteome downloaded in Oct, 2022 containing 20,535 entries, and the mouse lens digest used a UniProt mouse canonical reference proteome downloaded in June, 2023 containing 21,937 entries supplemented with the human αA-crystallin sequence. A precursor mass tolerance of 10 ppm was used for both searches, fragment ion tolerances of 0.2 and 0.05 Da used for human and mouse searches, respectively, and a static modification of 57.021 applied to cysteine residues. A reverse sequence database search strategy was used to estimate peptide false discovery rates using Percolator, and peptides with q-scores <0.05 used to create biblioSpec libraries. Spectral libraries for crystallin peptides were then imported into Skyline software. (version 23.1.0.455) to define the tryptic peptides used to measure ratios of light (endogenous)/heavy (15N labeled standard) peptides (L/H). Quantification used precursors (M, M+1, and M+2 peaks) and the 3 most intense fragment ions. These peptides were manually inspected in Skyline to remove low intensity outliers and select only single charge states with the highest intensity for each peptide.
The peptides were further screened to omit ones that did not provide a linear response at extreme L/H ratios. This was done by varying the amount of both soluble protein from the 23-day-old human lens and standards that were mixed before digestion. Approximately equal amounts of protein from 23-day-old soluble protein and 15N standards were mixed to achieve ratios of approximately 1:1 for each crystallin. These volumes were then varied by either keeping the amount of lens soluble protein constant and adding 2-fold less increments of 15N-labeled standards or keeping the amount of 15N-labled standard constant and adding successively 2-fold less lens soluble protein. This achieved mixed L/H ratios from 1:16, 1:8, 1:4, 1:2, 1:1, 2:1, 4:1, 8:1 and 16:1. Since the measured 1:1 ratios were sometimes slightly lower than 1 for some crystallins, the data was normalized so that the average peptide ratio for the 1:1 mixture equaled 1 and other ratios correspondingly adjusted. This dataset was generated using single LC/MS runs for each of the 9 ratios using the high resolution survey/DIA method described above.
Measurement of crystallin abundances in the lenses of the 4 human donors was then performed using 4 independent digests and LC/MS analyses. The absolute weight of each endogenous crystallin in the digests was calculated using the following equation: (µg heavy standard added to digest) x (MW untagged crystallin)/(MW SUMO tagged crystallin) x (averaged L/H ratios).
In mouse lenses, the amount of both endogenous mouse αA-crystallin and transgene expressed human αA-crystallin was measured by preparing digests spiked with heavy human SUMO-αA standard from 4 different mice expressing either WT human αA, or N101D human αA-crystallin mutant, for both lens water-soluble and water-insoluble fractions. The amount of mouse αA-crystallin was calculated using measured L/H ratios for αA-crystallin tryptic peptides unique to human and shared between mouse and human separately and subtracting the amount of human αA-crystallin from the amount total αA-crystallin.