Molecular weight characterization of PRG4 proteins using multi-angle laser light scattering (MALLS)

Osteoarthritis and Cartilage 21 (2013) 498e504

Molecular weight characterization of PRG4 proteins using multi-angle laser light scattering (MALLS) B.L. Steele y, M.C. Alvarez-Veronesi z, T.A. Schmidt yz * y Faculty of Kinesiology, University of Calgary, Calgary, AB, Canada z Schulich School of Engineering, Centre for Bioengineering Research & Education, University of Calgary, Calgary, AB, Canada

a r t i c l e i n f o

s u m m a r y

Article history: Received 8 June 2012 Accepted 7 December 2012

Objectives: Alternative splicing and variable post-translational modifications result in proteoglycan 4 (PRG4) proteins with historically reported apparent molecular weights (Ma) ranging from 150 to 400 kDa. The objectives of this study were to (1) identify and determine the weight averaged molecular weights (MW’s) of PRG4 proteins purified from medium with transforming growth factor-beta 1 (TGF-b1) conditioned by mature bovine articular cartilage explants and (2) to examine the effect of reduction and alkylation (RA) on PRG4. Methods: Non-reduced (NR) and RA preparations of PRG4 were separated using high performance liquid chromatography-size-exclusion chromatography with an in-line multi-angle laser light scattering (MALLS) detector, which was used for absolute determination of PRG4 MW. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), immunoblotting, and tandem mass spectrometry (MS/ MS) analysis were used to confirm the identity of separated proteins. Results: Three putative PRG4 monomers, one with previously uncharacterized MW, were identified in NR and RA PRG4 preparations of 239 (223,255), 379 (369,389), and 467 (433,501) kDa. Additionally w1 MDa putative PRG4 dimer was identified. Release of a w90 kDa PRG4 fragment was also observed on SDSPAGE after RA. Western Blotting with anti-PRG4 antibodies detected immunoreactive bands with Ma similar to MW for all species and excised bands were confirmed to be PRG4 by MS/MS. Conclusions: A variety of monomeric PRG4 proteins and a disulfide-bonded dimer/multimer are secreted by chondrocytes in bovine cartilage explants. The observed decrease in MW’s of monomeric PRG4 species upon RA may be due to the release of post-translationally cleaved fragments. Further study of these species will provide insight into the PRG4 molecular structure and function relationship. Ó 2012 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.

Keywords: PRG4 Lubricin MALLS Superficial zone protein

Introduction The proteoglycan 4 (PRG4) gene1 encodes for a family of mucinlike proteins, including lubricin and superficial zone protein (SZP). These proteins are synthesized and secreted by cells within articular joints, including synoviocytes2 and superficial zone articular chondrocytes3, and are present in synovial fluid (SF)4 and at the surface of articular cartilage5. PRG4 acts as a boundary lubricant in vitro on glass surfaces6e8 and at physiologically relevant cartilage-on-cartilage interfaces9, singularly and synergistically with hyaluronan (HA)10,11. Additionally, studies suggest that PRG4

* Address correspondence and reprint requests to: T.A. Schmidt, 2500 University Dr NW, KNB 426, University of Calgary, Calgary, AB, Canada, T2N 1N4. Tel: 403-2207028; Fax: 403-284-3553. E-mail addresses: [email protected] (B.L. Steele), cecilia.alvarezveronesi@ mail.utoronto.ca (M.C. Alvarez-Veronesi), [email protected] (T.A. Schmidt).

or its fragments create a network of “entanglements” that are responsible for increasing the relaxation time of SF, thereby allowing strain energy induced by articular joint locomotion to slowly dissipate, and for imparting chondroprotective properties to SF12. Mutations to the PRG4 gene cause the autosomal-recessive disorder camptodactyly-arthropathy-coxa vara-pericarditis (CACP), which is characterized by pain, swelling, and restricted range of motion in large joints and causes juvenile onset of noninflammatory joint failure13,14. Concurrently, PRG4 knockout mice exhibit increased cartilage wear and higher friction in diarthrodial joints than that observed in wild-type animals15. Reduction in PRG4 expression as seen after joint injury is associated with cartilage degeneration and early onset of osteoarthritis (OA)16. Thus, PRG4 is critical to the normal lubricating function of SF and to the maintenance of healthy joints. The PRG4 gene contains 12 exons17 and encodes for a highly conserved sequence of 1404 and 1445 amino acid protein in

1063-4584/$ e see front matter Ó 2012 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.joca.2012.12.002

B.L. Steele et al. / Osteoarthritis and Cartilage 21 (2013) 498e504

human and bovine, respectively, with molecular weights (MW’s) of w150 kDa for the unsubstituted core proteins5,18,19. PRG4 contains a central mucin-like domain, which is extensively glycosylated with O-linked oligosaccharide substitutions that account for greater than 50% of the mass of the expressed protein, and contains globular, cysteine-rich domains at the N- and C-termini. The core protein also contains attachment sites for chondroitin and keratan sulfate chains2; however, it has been shown that not all PRG4 molecules are substituted with these glycosaminoglycans20. PRG4 has also been shown to form disulfide-bonded multimers in normal SFs21, with intermolecular bond formation believed to occur via the N-terminus of the protein22,23. Reduction and alkylation (RA) of PRG4 reduces its lubricating ability in vitro8,24 and also reduces its ability to bind to cartilage surfaces8,22, indicating that multimerization of PRG4 is a key structural component for biological function. Alternative splicing of exons 2e5 of the PRG4 gene1,19,23,25e27 and variable post-translational modifications have lead to the separate identification of several homologous PRG4 species3,19,25. Transforming growth factor-beta 1 (TGF-b1) has been shown to regulate alternative splicing of PRG4 by chondrocytes in cartilage explants cultures23, as well as increase synthesis and secretion of PRG419,28. Initially isolated in 1977 from the lubricating protein fraction of bovine SF, the protein was first termed lubricating glycoprotein (LGP-I)29 and then renamed “lubricin”30. The MW of lubricin was determined to be 227.5 kDa by sedimentation-equilibrium measurements31 and 206 kDa by Rayleigh light scattering30, but exhibits an apparent migration on sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) of 280 kDa6. In 1994, Schumacher et al. described a “novel” proteoglycan, which was synthesized and secreted exclusively by chondrocytes in the superficial zone of articular cartilage and was thus named SZP, with an apparent molecular weight (Ma) of 345 kDa by SDS-PAGE2. Due to differences in cellular origin2,3,5, in migration on SDSPAGE2,5,6 and in buoyant densities from sedimentation2,8, lubricin and SZP were initially identified as distinct SF constituents. However, cDNA sequencing and homology analysis revealed that SZP and lubricin are both products of the megakaryocyte stimulating factor (MSF) gene3,19,25, now referred to as the PRG4 gene1. Initial cDNA cloning studies on the MSF gene were conducted by Merberg et al. using the w25 kDa urinary MSF metabolite, which was shown to be a partial peptide derived from a larger, glycosylated precursor protein with a theoretical MW of 400 kDa17. However, the exact biosynthetic pathways and molecular processing that result in the production and secretion of diverse protein products of the PRG4 gene, herein collectively referred to as PRG4, are not well understood. Multi-angle laser light scattering (MALLS) is a powerful technique that accurately determines molecular size and mass of biomolecules in a non-destructive manner. Unlike traditional methods for determining MW including size-exclusion chromatography (SEC), gel electrophoresis, analytical ultracentrifugation and time-of-flight mass spectrometry, MALLS does not require the use of relative standards, is relatively fast, and can be used to characterize very high MW proteins and aggregates32,33. As such, MALLS can confirm the previously reported molecular weights Ma of PRG4 species, which may have inherent inaccuracies when based on SDS-PAGE due to the high extent of PRG4 glycosylation34. Therefore, the objectives of this study were to (1) use high performance liquid chromatography (HPLC)-SEC with in-line MALLS to identify and determine the weight averaged MW’s of PRG4 proteins purified from medium with TGF-b1 conditioned by mature bovine cartilage explants and (2) to examine the effect of RA on PRG4.

499

Methods Materials Ethylenediaminetetraacetic acid (EDTA) and Tris-base were purchased from Fisher Scientific (Fairlawn, NJ, USA), and bicinchoninic acid assay (BCA) kit was purchased from Thermo Fisher Scientific (Rockford, IL, USA). Sodium chloride, guanidine hydrochloride (GdHCl), and dithiothreitol (DTT) were purchased from SigmaeAldrich (St. Louis, MO, USA). Sodium acetate was purchased from MP Biomedicals (Santa Ana, CA, USA), and iodoacetamide was obtained from Acros Organics (Geel, Belgium). Diethylaminoethyl (DEAE) Sepharose, ECL Prime Substrate, and polyvinylidene fluoride (PVDF) were purchased from GE Healthcare (Piscataway, NJ, USA). Blotting grade non-fat milk and Tween-20 were obtained from Bio-Rad Laboratories (Hercules, CA, USA). Unless otherwise specified, all other gel and Western Blotting reagents and apparatus were purchased from Invitrogen (Carlsbad, CA, USA), as well as cartilage culture media and reagents. Human, recombinant TGF-b1 was obtained from Peprotech (Rocky Hill, NJ, USA). Anti-PRG4 polyclonal antibody LPN, which is specific to the C-terminal sequence aa1356e1373 was purchased from Thermo Fisher Scientific. Anti-sera Jl08N, raised against N-terminal sequence aa75-95, was a gift from Dr. M. L. Warman (Harvard Medical School, Boston, MA, USA). Goat anti-rabbit horseradish peroxidase (HRP) coupled IgG was purchased from Chemicon (Temecula, CA, USA). PRG4 preparation & purification As described previously9,28, PRG4 was prepared from culture media, supplemented with 10 ng/mL of TGF-b1, conditioned by cartilage disks harvested from the patellofemoral groove of mature bovine stifle joints. PRG4 was purified from pooled media from the 4-week culture using anion exchange chromatography. Briefly, media was applied to a DEAE Sepharose column equilibrated with 0.15 M NaCl, 0.005 M EDTA, and 0.05 M sodium acetate, pH ¼ 6. The column was washed with three volumes of 0.15 M NaCl buffer, then eluted with a step gradient of 0.3 M and 0.615 M NaCl buffer. The PRG4 enriched 0.615 M eluate and PRG4 containing 0.3 M eluate were collected, concentrated, buffer exchanged into water with a 30 kDa MW cutoff filter (Millipore; Billerica, MA, USA), and quantified by BCA. Reduction and alkylation of PRG4 Dried PRG4 was reconstituted in 4.0 M GdHCl, 50 mM Tris-base, 1 mM EDTA, pH ¼ 8.5 at a concentration of 2 mg/mL and incubated with (RA) or without (non-reduced: NR) 10 mM DTT at 60 C for 2 h, followed by 40 mM iodoacetamide for 2 h in the dark at room temperature. The samples were then stored at 20 C or were ethanol precipitated for SDS-PAGE analysis21. To assess if complete reduction was occurring, samples were incubated with 10 mM DTT for up to 5 h. Time points were taken every 30 mine1 h, and either addition of alkylation reagent or ethanol precipitation was used to stop the reaction. Extent of reduction was determined by examining the relative intensity of reduced product bands on SDS-PAGE. HPLC-SEC-MALLS NR and RA PRG4 species from the 0.615 M eluate were separated using HPLC (Perkin Elmer-Series 200) with a size exclusion column (WTC-050S5, 500Å pore size; Wyatt Technology Corp, Santa Barbara, CA, USA) equilibrated in 4.0 M GdHCl and with in-line UV (Perkin Elmer-Series 200), MALLS, and refractive index detectors

500

B.L. Steele et al. / Osteoarthritis and Cartilage 21 (2013) 498e504

(DAWN Heleos II and Optilab reX, respectively, Wyatt Technology Corp) for MW characterization. For each run, 250 mg of NR or RA PRG4 were injected (n ¼ 3; with three technical repeats each) and eluted with a flow rate of 0.5 mL/min. UV, MALLS, and RI data were collected and analyzed using ASTRAÔ software (Wyatt Technology), and a dn/dc value of 0.160 mL/g was used for PRG430. Calculated MW of the peaks are presented as mean with a 95% confidence interval (lower limit, upper limit), which was calculated by multiplying the standard error of the mean by 1.96. SDS-PAGE and Western Blotting Two to 4 mg of PRG4 were resuspended in NuPAGE LDS sample buffer and were heated at 70 C for 10 min. The samples were loaded onto a Novex 3e8% Tris-Acetate gel with 10 mL of HIMarkÒ Pre-Stained High Molecular Weight Protein Standard. Electrophoresis was conducted in NuPAGE Tris-Acetate SDS running buffer at 4 C for 75 min at 150 V in a vertical gel apparatus. Gels were stained using Coomassie G-250 (SimplyBlue SafeStain) per manufacturer’s instructions, or gels were electroblotted to 0.2 mm PVDF at a constant current of 200 mA for 2hr in NuPAGE Transfer Buffer (0% methanol). Membranes were blocked with 5% non-fat dry milk in Tris-buffered saline with Tween (TBST) (20 mM Tris-base, 137 mM NaCl, pH ¼ 7.6 with 0.1%(v/v) Tween-20) for 60 min at room temperature and then incubated overnight at 4 C with rocking with 2 mg IgG/mL LPN primary antibody or 1:1000 J108N serum in 3% milk/TBST. Membranes were washed in TBST (3  10 min), incubated in 0.5 mg/mL goat anti-rabbit HRP antibody in 3% milk/ TBST with rocking for 1hr at room temperature, and washed again in TBST (3  10 min). To test for non-specific absorption of the secondary antibody, blots were treated as described above, except primary antibody was omitted for LPN or 1:1000 dilution of rabbit serum was used in place of J108N. Immunoreactive bands were detected using ECL Prime Chemiluminescent Substrate and CHEMI Genius 2 Bio Imaging System (SYNGENE; Frederick, MD, USA). MS/MS analysis Stained protein bands of interest with Ma corresponding to those determined for PRG4 species by MALLS were excised from the 0.615 M lanes and submitted for identification by tandem mass spectrometry (MS/MS) to the Southern Alberta Mass Spectrometry Centre (University of Calgary), as described previously21. For data analysis, Mascot searches using MudPIT scoring with P < 0.05 were conducted against bovine PRG4 sequences. Results HPLC-SEC-MALLS HPLC-SEC was able to distinguish different MW species within the 0.615 M PRG4 preparations, as indicated by representative, normalized UV, MALLS, and RI chromatograms for NR PRG4 [Fig. 1(A)]. Three major peaks were observed in the UV trace for NR and RA PRG4 [Fig. 1(B)]. After RA, the relative intensity of peak 1 decreased, while the relative intensities of peak 2 and peak 3 increased in the normalized UV trace [Fig. 1(B)]. The increase in relative intensity for peak 2 is not readily observed in the figure, due to normalization of the absorbance spectra to this peak. Calculated MW of the peaks (n ¼ 3) are shown (Table I). For NR and RA peak 1, the calculated Mw are not statistically different when analyzed by an unpaired t-test, which may indicate incomplete reduction of the PRG4 samples. Concurrently, a reduction time course showed that a weak protein band with Ma corresponding to w1 MDa on SDS-PAGE was present at an incubation time of 2 h and

Fig. 1. HPLC-SEC-MALLS on PRG4. UV (e), RI (e e) and MALLS () traces from HPLCSEC of NR PRG4 (A). UV chromatograms for NR (e) and RA (e) PRG4 with peaks analyzed for MW determination by MALLS numbered (B).

did not completely disappear until after 4 h of incubation (data not shown). The polydispersity (the ratio of the MW to the number averaged molecular weight (MN)) was 1.03e1.05 for NR and RA peaks 2e3 and was 1.2 for NR and RA peak 1. SDS-PAGE and Western Blotting Protein staining of the 0.3 M and 0.615 M NaCl eluates showed bands with Ma consistent with the calculated MW from MALLS for HPLC-SEC peaks 1e3 for NR [Fig. 2(A)] and peaks 2e3 for RA PRG4 [Fig. 2(C)]. At the given load (w2 mg of protein per lane) in Fig. 2, a band corresponding to RA1 was not observed, however; when the concentration of RA protein was increased four-fold, a band corresponding to RA1 was observed on SDS-PAGE with protein staining (data not shown). In addition, while only the 0.615 M NaCl wash was used for HPLC-SEC-MALLS, visualization of NR and RA peak 3 required analysis of the 0.3 M NaCl wash due to low apparent abundance of these species on SDS-PAGE in the higher salt wash. Western Blotting of the 0.3 and 0.615 M NaCl eluates showed LPN immunoreactive reactive bands that co-migrated with NR1e3 [Fig. 2(B)] and J108N immunoreactive bands corresponding to RA 2,3 [Fig. 2(D)]. A band corresponding to NR3 was difficult to detect using LPN, therefore two 0.615 M blots [Fig. 2(B)] are shown to illustrate immunoreactivity of this species. After RA, an Maz90 kDa band appeared on SDS-PAGE and was J108N immunoreactive [Fig. 2 (C, D); F90].

B.L. Steele et al. / Osteoarthritis and Cartilage 21 (2013) 498e504

501

Table I Summary of SDS-PAGE, immunoreactivity, and MS/MS on PRG4 proteins. The calculated weight averaged MW’s from MALLS (Mw, n ¼ 3; with three technical repeats each) for HPLC-SEC peaks 1e3 for NR and RA PRG4 are given with 95% confidence intervals (lower limit, upper limit), along with a summary of SDS-PAGE, J108N and LPN bands observed with apparent molecular weight, Ma, corresponding to a given Mw. Peptides identified from MS/MS of excised SDS-PAGE bands are also listed with their corresponding starting amino acid sequence (bovine PRG4) Western Blot

NR 1

Molecular weight (Mw) (kDa)

LPN

1180 (867, 1493)

X

J108N

SDS-PAGE Protein Stain (Ma)

MS/MS

Peptide sequences bovine PRG4

X

X

79

RGRECDCDSDCKK KVIESEEITEEHSVSENQESSSSSSSSSSTIRK RITEVWGIPSPIDTVFTRC 1315 KQEPTQKC 1385 RGLPNVVTSAISLPNIRK 1413 KDQYYNIDVPSRT 91 KYGKCCSDYGAFCEAVHNPTTPPPPKT 370 KKPAPTTPKE 1225 RITEVWGIPSPIDTVFTRC 1259 RFTNDIKD 1385 RGLPNVVTSAISLPNIRK 1413 KDQYYNIDVPSRT 1385 RGLPNVVTSAISLPNIRK 1225 RITEVWGIPSPIDTVFTRC 1248 KTFFKG 1253 KGSQYWRF 1325 RRPAINYSVYGETAQVRR 685 KEPAPTSPKEPAPTSPKE 146

1225

NR 2

467 (433, 501)

X

NR3 RA 1

239 (223, 255) 1060 (890, 1230)

X

RA 2 RA 3

379 (369, 389) 150 (108, 192)

X

X X

MS/MS analysis Bands with migrations corresponding to NR1e3 and RA 1e3 were excised from SDS-PAGE, trypsin digested, and submitted for MS/MS analysis. Identified peptides were searched against the bovine PRG4 sequences (Accession: DAA21008.1), and all bands were confirmed to be PRG4 except for RA3. The peptide matches for each band with corresponding starting amino acid position are listed in Table I. Discussion The results of this study demonstrate that chondrocytes in articular cartilage explants predominately produce and secrete three distinct PRG4 species, using SEC-MALLS. Two previously uncharacterized PRG4 species with calculated weight averaged MW’s of w1 MDa and 467 (403,501) kDa were identified, in addition to a 239 (222,256) kDa species. The 467 kDa species is likely the full

X

X

X X

X X

X X

X

length, glycosylated form of PRG4 previously identified as MSF precursor17, but with a MW w70 kDa larger than predicted. Additionally, the calculated MW’s for RA2 and RA3 were determined to be 379 (369,389) and 150 (108,192) kDa, respectively. After RA, the relative abundance of the w1 MDa peak was significantly diminished, with no statistical difference arising between the calculated MW’s of 1.18 (0.906,1.493) and 1.06 (0.876,1.230) MDa for NR1 and RA1, respectively, indicating that this species is a disulfide-bonded PRG4 dimer or multimer. These results combined with a preliminary report from our laboratory35 are the first to demonstrate that bovine articular cartilage chondrocytes synthesize and secrete high MW, disulfide-bonded multimers of PRG4. A limitation of this study was the resolution of the SEC column, which was designed for the separation of very large biomolecules (w15 kDae5 MDa), as the MW’s of co-eluting species in each peak are averaged together to calculate the MW during MALLS analysis. Improved resolution of the multimer and monomer peaks could potentially be achieved using a column with a larger pore size, but

Fig. 2. SDS-PAGE and Western Blotting on PRG4. SDS-PAGE with Coomassie blue staining (A, C) and Western Blot analysis with antibodies LPN (B) and J108N (D) of NR (A, B) and RA (C, D) PRG4 species on 0.3 M and 0.615 M NaCl eluates. Bands with Ma corresponding to Mw for a given peak are marked. F90 corresponds to the release of fragment, confirmed to be PRG4 by MS/MS, of w90 kDa after RA.

502

B.L. Steele et al. / Osteoarthritis and Cartilage 21 (2013) 498e504

at the expense of resolution of smaller MW species. Despite not achieving baseline resolution in the SEC separation, the PRG4 peaks analyzed in this study had a polydispersity of w1 indicating that the masses of the eluting species were fairly monodisperse. In addition to the three main PRG4 peaks analyzed in both NR and RA chromatograms, several small, poorly defined peaks with MW’s less than 100 kDa were observed, but eluted with inconsistent retention times thus precluding further analysis. These small peaks may be contaminants remaining after DEAE purification of conditioned media (future studies could use a 100 kDa filter to reduce such non-PRG4 contaminants). Additionally, reduction time course experiments verified that RA of the peak 1, PRG4 multimer was incomplete during the 2 h incubation period used in this study (data not shown) and did not disappear until after 4 h, suggesting that longer incubation times or increased concentration of reducing agent should be used in future experiments. As PRG4 has a putative HA binding domain17, samples were also treated with hyaluronidase prior to RA and HPLC-SEC-MALLS, which did not significantly change the calculated Mw values (data not shown), indicating that an interaction with any remaining HA in the preparation was not occurring. Some PRG4 species with high relative abundances in the SEC-UV spectra were difficult to detect using SDS-PAGE with protein staining and immunoblotting. The negatively charged carbohydrates of glycoproteins have been shown to hinder the binding of both Coomassie blue36 and silver ions37 during SDS-PAGE staining, resulting in unreliable protein detection. Accordingly, densitometric analysis of the NR and RA 0.615 M, enriched PRG4 lanes did not show significant peaks corresponding to the 239 kDa and 150 kDa peaks observed in the UV traces (data not shown). As such, the 0.15e0.3 M NaCl wash from the DEAE purification, which has a higher abundance of proteins with lower charge densities, was probed in addition to the 0.3e0.615 M NaCl, PRG4 enriched wash that was used in SEC-MALLS experiments. Due to poor transfer efficiency and low concentration, a band with Ma corresponding to RA1 was not observed with J108N immunoblotting and required a four-fold increase in the amount of protein typically loaded for visualization on SDS-PAGE with Coomassie blue staining (data not shown). An immunoreactive band corresponding to NR3 was not always detected, perhaps due to loss of proper conformation of the binding epitope for LPN, which might be remedied by the use of another anti-PRG4 antibody such as 6A1, a monoclonal antibody (mAb) specific to the C-terminal of PRG4 that binds under NR and RA conditions19. In a previous study, Western Blotting with mAb 6A1 under non-reducing conditions showed immunoreactive bands at 140e150 and 60e70 kDa19, but these bands were not typically observed in our NR preparations. Using J108N, PRG4 immunoreactive bands at w150 kDa were intermittently observed after RA in the 0.3 M and 0.615 M washes. Excised SDS-PAGE bands with Ma corresponding to each calculated MW from MALLS were submitted for MS/MS analysis, which confirmed that NR1e3 and RA1e2 were indeed PRG4. However, a band with Ma corresponding to RA3 could not be positively identified as PRG4 by MS/MS; thus, it remains to be determined whether the 150 kDa, RA3 peak with a MW similar to that of the PRG4 core protein is in fact a partial, glycosylated or unglycosylated PRG4 product or an impurity present in this preparation. Additionally, MS/MS analysis on the 60 kDa band [0.3 M, Fig. 2 (C)] did not display any peptide matches to PRG4, and analysis of the Ma z 460 kDa, reduced band [0.615 M, Fig. 2 (C)] matched to PRG4 sequences 146KVIESEEITEE HSVSENQESSSSSSSSSSTIRK, 370KKPAPTTPKE, 1225RITEVWGIPSPID TVFTRC, and 1259RFTNDIKD. The results of this study agree with and extend previous studies on the molecular organization of PRG4. The reported MW’s in this study correspond with historically reported masses for lubricin31

(NR3) and SZP2,5 (RA2). Additionally, the identification of w1 MDa PRG4 protein supports the findings of a previous study in which disulfide-bonded dimers were isolated from normal bovine SF21. Furthermore, RA of PRG4 was hypothesized to result in the collapse of disulfide-bonded multimers into one or two easily identifiable PRG4 monomer(s) and/or fragment(s). However, after RA rather than only observing a shift in the relative intensities for NR and RA peaks 1e3, a decrease in MW of w90 kDa occurred between NR2 and RA2 and between NR3 and RA3. The release of low Ma fragments upon reduction of PRG4 was observed using SDS-PAGE and Western Blotting (F90, Fig. 2), including a Ma w90 kDa J108N immunoreactive band, which may or may not correspond to the observed shift in MW. While MS/MS analysis of F90 only matched to the human PRG4 sequence (1029KKPSTKKPKThuman similar to 1071KKPSTKKPRT-bovine), the fragment release is consistent with a previous report where an w70 kDa fragment, confirmed to be PRG4 by MS/MS, was released after RA and was speculated to be a truncated and/or under glycosylated PRG4 species21. In another study, the formation of a w90 kDa band on SDS-PAGE in reduced samples of PRG4 was observed, along with w50 and w28 kDa fragments that were believed to be breakdown products due to sample heating6. As such, future studies in which RA of PRG4 preparations of isolated NR1 and NR2 species are examined may provide more information on fragment release and help clarify the mechanisms of PRG4 assembly and processing. Other studies using mouse and human recombinant constructs of PRG4 have demonstrated release upon RA of a 12e14 kDa PRG4 fragment that is a C-terminal peptide generated from posttranslational subtilisin-like proprotein convertase (SPC) cleavage that remains disulfide-bonded to the main protein chain of the C-terminal globular domain22,38. Mutations to either the conserved arginine cleavage site (RFERA) or to one of two cysteines responsible for the intra-chain disulfide bond prevents SPC mediated cleavage. Cleavage is predicted to induce a normal structural conformation that is critical to the function of PRG4 in SF, which is supported by the detection of immunoreactive SPC fragments of PRG4 in healthy porcine and bovine SF and by the identification of a mutation to PRG4 in a subset of patients with CACP that prevents SPC cleavage38. As such, the shift in PRG4 MW after reduction observed in this study may also represent the loss of several disulfide-bonded fragments that are formed from SPC mediated cleavage in combination with additional, uncharacterized processing events on the termini of PRG4 or in combination with partial degradation of PRG4 by other proteases. PRG4 has been shown to be susceptible to degradation by matrix metallo-, serine, and cysteine proteases39,40. The release of multiple fragments is supported by the appearance of several low Ma, J108N immunoreactive fragments on SDS-PAGE after reduction of PRG4. However, an immunoreactive band corresponding to the release of the SPC cleavage fragment was not detected, as it would migrate with the front in the gel system employed in this study, which is optimized for the detection of high MW proteins (w31e460 kDa). TGF-b1 was used here to increase synthesis and secretion of PRG4 by chondrocytes in explant culture19,28. While TGF-b1 has been shown to increase synthesis of an isoform lacking exons 4 and 5 and to decrease expression of the full length isoform23, the change in MW based on the primary amino acid sequences between the smallest and largest predicted PRG4 isoforms would be only w20 kDa. Thus, alternative splicing alone is not enough to explain the differences in the observed MW’s of secreted PRG4 products. In all known PRG4 isoforms, the mucin-like domain that is encoded by exon 6 is conserved; therefore, differences in PRG4 MW are most likely due to differences in O-glycosylation and in glycosaminoglycan substitution of the core protein. In a previous study, treatment of human PRG4 with Ma w350 kDa with sialidase caused

B.L. Steele et al. / Osteoarthritis and Cartilage 21 (2013) 498e504

a decrease of w70 kDa in mass, and further treatment with O-glycanase resulted in two PRG4 proteins with Ma of w225 and w250 kDa on SDS-PAGE. Yet, a plausible explanation for the presence of two isoforms, both of which have Ma greater than predicted for the PRG4 core, is that incomplete deglycosylation of PRG4 may have occurred41. Another study showed that deglycosylation of human PRG4 with Ma of w280 kDa resulted in a 120 kDa product, but the extent of deglycosylation was not known3. Since deglycosylation results in diminished lubricating ability of PRG47 and since changes in the extent of PRG4 glycosylation have been detected in different disease states41, glycosylation is a critical structural component to the biological function of PRG4. As the level of TGFb1 used in the culture media is within the range observed to occur in arthritic human SF42, the distribution of secreted MW species observed in this study may represent a pathophysiological state. Therefore, future studies that evaluate the effect of TGF-b1 stimulation of cultured articular cartilage on the MW of secreted PRG4 and of its reduction products to those observed in the absence of growth factor may clarify differences in PRG4 synthesis and proteolysis between healthy and diseased states. In conclusion, the results presented in this study illustrate that the biosynthetic pathways leading to the secretion of PRG4 species with varying MW’s are complex and involve alternative splicing, glycosylation, post-translational cleavage, and potentially proteolytic activation/degradation. The methods developed here allow for the separation and accurate characterization of PRG4 species with differing molecular masses. Further study of these species will provide insight into the PRG4 molecular structure and function relationship through examination of their, potentially differential, ability to bind to physiologically relevant surfaces and provide boundary lubrication, and/or to contribute to the formation of “entanglement” networks in SF that dissipate strain energy. In addition, studies examining chondroitinase and keratanase treatment and/or deglycosylation may also provide new insights into the functional roles of these modifications on the different PRG4 species. Collectively, these results could contribute to the understanding of the roles of these different PRG4 species in normal and pathological SF and, ultimately, in possible OA biotherapeutic strategies involving PRG443,44. Author contributions All authors contributed to the conception and design of the original study and approved the final submitted manuscript. BLS and CA were responsible for data acquisition and analysis. The article was first drafted by BLS, and critically reviewed by CA and TS. TS obtained funding for the study, and takes full responsibility for the integrity of the work as a whole. Role of funding source Study sponsors had no role in collection, interpretation, or summary of results reported here. Conflict of interest The authors do not have any conflicts of interest to disclose. Acknowledgments We would like to thank Morgan Khan and the Southern Alberta Mass Spectrometry Centre for their assistance with MS/MS analysis. We also would like to thank Dr M. L. Warman of Harvard Medical School, Boston, MA for generously donating anti-sera J108N. This work was supported by funding from the National Science and Engineering Research Council of Canada Discovery Grant Programme and the Collaborative Research and Training Experience

503

Programme (NSERC-CREATE), Canadian Arthritis Network, The Arthritis Society, Alberta Innovates Health Solutions OA Team Grant, Faculty of Kinesiology and Schulich School of Engineering’s Center for Bioengineering Research and Education at the University of Calgary. References 1. Ikegawa S, Sano M, Koshizuka Y, Nakamura Y. Isolation, characterization and mapping of the mouse and human PRG4 (proteoglycan 4) genes. Cytogenet Genome Res 2000;90(3e4): 291e7. 2. Schumacher BL, Block JA, Schmid TM, Aydelotte MB, Kuettner KE. A novel proteoglycan synthesized and secreted by chondrocytes of the superficial zone of articular cartilage. Arch Biochem Biophys 1994;311(1):144e52. 3. Jay GD, Britt DE, Cha CJ. Lubricin is a product of megakaryocyte stimulating factor gene expression by human synovial fibroblasts. J Rheumatol 2000;27(3):594e600. 4. Swann DA, Silver FH, Slayter HS, Stafford W, Shore E. The molecular structure and lubricating activity of lubricin isolated from bovine and human synovial fluids. Biochem J 1985;225: 195e201. 5. Schumacher BL, Hughes CE, Kuettner KE, Caterson B, Aydelotte MB. Immunodetection and partial cDNA sequence of the proteoglycan, superficial zone protein, synthesized by cells lining synovial joints. J Orthop Res 1999;17(1):110e20. 6. Jay GD. Characterization of a bovine synovial fluid lubricating factor I. Chemical, Surface activity and lubricating properties. Connect Tissue Res 1992;28(1e2):71e88. 7. Jay GD, Harris DA, Cha C-J. Boundary lubrication by lubricin is mediated by O-linked b(1-3)Gal-GalNAc oligosaccharides. Glycoconj J 2001;18(10):807e15. 8. Swann DA, Hendren RB, Radin EL, Sotman SL. The lubricating activity of synovial fluid glycoproteins. Arthritis Rheum 1981;24(1):22e30. 9. Schmidt TA, Gastelum NS, Nguyen QT, Schumacher BL, Sah RL. Boundary lubrication of articular cartilage: role of synovial fluid constituents. Arthritis Rheum 2007;56(3):882e91. 10. Jay GD, Lane BP, Sokoloff L. Characterization of a bovine synovial fluid lubricating factor III. The interaction with hyaluronic acid. Connect Tissue Res 1992;28(4):245e55. 11. Kwiecinski JJ, Dorosz SG, Ludwig TE, Abubacker S, Cowman MK, Schmidt TA. The effect of molecular weight on hyaluronan’s cartilage boundary lubricating ability e alone and in combination with proteoglycan 4. Osteoarthritis Cartilage 2011;19(11):1356e62. 12. Jay GD, Torres JR, Warman ML, Laderer MC, Breuer KS. The role of lubricin in the mechanical behavior of synovial fluid. Proc Natl Acad Sci 2007;104(15):6194e9. 13. Marcelino J, Carpten JD, Suwairi WM, Gutierrez OM, Schwartz S, Robbins C, et al. CACP, encoding a secreted proteoglycan, is mutated in camptodactyly-arthropathy-coxa vara-pericarditis syndrome. Nat Genet 1999;23(3):319e22. 14. Basit S, Iqbal Z, Umicevic-Mirkov M, Kamran Ul-Hassan Naqvi S, Coenen M, Ansar M, et al. A novel deletion mutation in proteoglycan-4 underlies camptodactyly-arthropathy-coxavara-pericarditis syndrome in a consanguineous Pakistani family. Arch Med Res 2011;42(2):110e4. 15. Jay GD, Torres JR, Rhee DK, Helminen HJ, Hytinnen MM, Cha C-J, et al. Association between friction and wear in diarthrodial joints lacking lubricin. Arthritis Rheum 2007;56(11):3662e9. 16. Elsaid KA, Fleming BC, Oksendahl HL, Machan JT, Fadale PD, Hulstyn MJ, et al. Decreased lubricin concentrations and markers of joint inflammation in the synovial fluid of patients

504

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

B.L. Steele et al. / Osteoarthritis and Cartilage 21 (2013) 498e504

with anterior cruciate ligament injury. Arthritis Rheum 2008;58(6):1707e15. Merberg DM, Fitz LJ, Temple P, Giannotti J, Murtha P, Fitzgerald M, et al. A comparison of vitronectin and megakaryocyte stimulating factor. In: Preissner KT, Rosenblatt S, Kost C, Wegerhoff J, Mosher DF, Eds. Biology of Vitronectins and Their Receptors. BV: Elsevier Science; 1993:45e52. Zimin AV, Delcher AL, Florea L, Kelley DR, Schatz MC, Puiu D, et al. A whole-genome assembly of the domestic cow, Bos taurus. Genome Biol 2009;10(4):R42. Flannery CR, Hughes CE, Schumacher BL, Tudor D, Aydelotte MB, Kuettner KE, et al. Articular cartilage superficial zone protein (SZP) is homologous to megakaryocyte stimulating factor precursor and is a multifunctional proteoglycan with potential growth-promoting, cytoprotective, and lubricating properties in cartilage metabolism. Biochem Biophys Res Comm 1999;254(3):535e41. Lord MS, Estrella RP, Chuang CY, Youssef P, Karlsson NG, Flannery CR, et al. Not all lubricin isoforms are substituted with a glycosaminoglycan chain. Connect Tissue Res 2012;53(2):132e41. Schmidt TA, Plaas AHK, Sandy JD. Disulfide-bonded multimers of proteoglycan 4 (PRG4) are present in normal synovial fluids. Biochim Biophys Acta Gen Subj 2009;1790(5):375e84. Jones ARC, Gleghorn JP, Hughes CE, Fitz LJ, Zollner R, Wainwright SD, et al. Binding and localization of recombinant lubricin to articular cartilage surfaces. J Orthop Res 2007;25(3):283e92. DuRaine GD, Chan SMT, Reddi AH. Effects of TGF-b1 on alternative splicing of superficial zone protein in articular cartilage cultures. Osteoarthritis Cartilage 2011;19(1):103e10. Zappone B, Greene G, Oroudjev E, Jay GD, Israelachvili J. Molecular aspects of boundary lubrication by human lubricin: effect of disulfide bonds and enzymatic digestion. Langmuir 2008;24(4):1495e508. Jay GD, Tantravahi U, Britt DE, Barrach HJ, Cha C-J. Homology of lubricin and superficial zone protein (SZP): products of megakaryocyte stimulating factor (MSF) gene expression by human synovial fibroblasts and articular chondrocytes localized to chromosome 1q25. J Orthop Res 2001;19(4):677e87. DuRaine G, Neu CP, Chan SMT, Komvopoulos K, June RK, Reddi AH. Regulation of the friction coefficient of articular cartilage by TGF-b1 and IL-1b. J Orthop Res 2009;27(2): 249e56. Grad S, Lee CR, Wimmer MA, Alini M. Chondrocyte gene expression under applied surface motion. Biorheology 2006;43(3):259e69. Schmidt TA, Gastelum NS, Han EH, Nugent-Derfus GE, Schumacher BL, Sah RL. Differential regulation of proteoglycan 4 metabolism in cartilage by IL-1a, IGF-I, and TGF-b1. Osteoarthritis Cartilage 2008;16(1):90e7. Swann DA, Radin EL. The molecular basis of articular lubrication. J Biol Chem 1972;247(24):8069e73.

30. Swann DA, Slayter HS, Silver FH. The molecular structure of lubricating glycoprotein-I, the boundary lubricant for articular cartilage. J Biol Chem 1981;256(11):5921e5. 31. Swann DA, Sotman S, Dixon M, Brooks C. The isolation and partial characterization of the major glycoprotein (LGP-I) from the articular lubricating fraction from bovine synovial fluid. Biochem J 1977;161:473e85. 32. Wyatt PJ. Light scattering and the absolute characterization of macromolecules. Anal Chim Acta 1993;272(1):1e40. 33. Oliva A, Llabrés M, Fariña J. Applications of multi-angle laser light-scattering detection in the analysis of peptides and proteins. Curr Drug Discov Tech 2004;1(3):229e42. 34. Leach BS, Collawn JF, Fish WW. Behavior of glycopolypeptides with empirical molecular weight estimation methods 1. In sodium dodecyl sulfate. Biochemistry 1980;19(25):5734e41. 35. Alvarez M, Kooyman J, Schmidt TA. Synthesis of proteoglycan 4 (PRG4) disulfide-bonded multimers by chondrocytes in cartilage explants (abstract). Trans Orthop Res Soc 2010;35: 850. 36. Osset M, Piñol M, Fallon MJM, De Llorens R, Cuchillo CM. Interference of the carbohydrate moiety in Coomassie Brilliant Blue R-250 protein staining. Electrophoresis 1989;10(4):271e3. 37. Jay GD, Culp DJ, Jahnke MR. Silver staining of extensively glycosylated proteins on sodium dodecyl sulfate-polyacrylamide gels: enhancement by carbohydrate-binding dyes. Anal Biochem 1990;185(2):324e30. 38. Rhee DK, Marcelino J, Al-Mayouf S, Schelling DK, Bartels CF, Cui Y, et al. Consequences of disease-causing mutations on lubricin protein synthesis, secretion, and post-translational processing. J Biol Chem 2005;280(35):31325e32. 39. Elsaid KA, Jay GD, Warman ML, Rhee DK, Chichester CO. Association of articular cartilage degradation and loss of boundary-lubricating ability of synovial fluid following injury and inflammatory arthritis. Arthritis Rheum 2005;52(6): 1746e55. 40. Jones ARC, Hughes CE, Wainwright SD, Flannery CR, Little C, Caterson B. Degradation of PRG4/SZP by matrix proteases. Trans Orthop Res Soc 2003;28:133. 41. Estrella RP, Whitelock JM, Packer NH, Karlsson NG. The glycosylation of human synovial lubricin: implications for its role in inflammation. Biochem J 2010;429:359e67. 42. Fava R, Olsen N, Keski-Oja J, Moses H, Pincus T. Active and latent forms of transforming growth factor beta activity in synovial effusions. J Exp Med 1989;169(1):291e6. 43. Jay GD, Fleming BC, Watkins BA, McHugh KA, Anderson SC, Zhang LX, et al. Prevention of cartilage degeneration and restoration of chondroprotection by lubricin tribosupplementation in the rat following anterior cruciate ligament transection. Arthritis Rheum 2010;62(8):2382e91. 44. Flannery CR, Zollner R, Corcoran C, Jones AR, Root A, RiveraBermúdez MA, et al. Prevention of cartilage degeneration in a rat model of osteoarthritis by intraarticular treatment with recombinant lubricin. Arthritis Rheum 2009;60(3):840e7.