Glycosylation status of bone sialoprotein and its role in mineralization

Experimental Cell Research xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Experimental Cell Research journal homepage: www.elsevier.com/locate/yexcr

Glycosylation status of bone sialoprotein and its role in mineralization ⁎

⁎⁎

Lan Xua, , Zhenqing Zhangb, , Xue Sunb, Jingjing Wanga, Wei Xuc, Lv Shid, Jiaojiao Lub, Juan Tanga, Jingjing Liua, Xiong Sua a

Department of Biochemistry and Molecular Biology, Soochow University Medical College, Suzhou 215123, China College of Pharmaceutical Sciences, Soochow University, Suzhou 215123, China Department of Orthopaedic Surgery, The Second Affiliated Hospital of Soochow University, Suzhou 215004, China d Shanghai Green-Valley Pharmaceutical Co. Ltd., Shanghai 201200, China b c

A R T I C L E I N F O

A B S T R A C T

Keywords: Bone sialoprotein Mineralization Sialic acids Glycans Mass spectrometry

The highly glycosylated bone sialoprotein (BSP) is an abundant non-collagenous phosphoprotein in bone which enhances osteoblast differentiation and new bone deposition in vitro and in vivo. However, the structural details of its different glycosylation linkages have not been well studied and their functions in bone homeostasis are not clear. Previous studies suggested that the O-glycans, but not the N-glycans on BSP, are highly sialylated. Herein, we employed tandem mass spectrometry (MS/MS) to demonstrate that the N-glycanson the recombinant human integrin binding sialoprotein (rhiBSP) are also enriched in sialic acids (SAs) at their termini. We also identified multiple novel sites of N-glycan modification. Treatment of rhiBSP enhances osteoblast differentiation and mineralization of MC3T3-E1 cells and this effect could be partially reversed by efficient enzymatic removal of its N-glycans. Removal of all terminal SAs has a greater effect in reversing the effect of rhiBSP on osteogenesis, especially on mineralization, suggesting that sialylation at the termini of both N-glycans and O-glycans plays an important role in this regulation. Moreover, BSP-conjugated SAs may affect mineralization via ERK activation of VDR expression. Collectively, our results identified novel N-glycans enriched in SAs on the rhiBSP and demonstrated that SAs at both N- and O-glycans are important for BSP regulation of osteoblast differentiation and mineralization in vitro.

1. Introduction Bone sialoprotein (BSP) is one of the major extracellular matrix (ECM) proteins of the bone and belongs to the small integrin-binding ligand N-linked glycoprotein (SIBLING) family. There are five known members of the SIBLING family: BSP, osteopontin (OPN), dentin matrix protein-1 (DMP1), dentin sialophosphoprotein (DSPP), and matrix extracellular phosphoglycoprotein (MEPE) [1,2]. BSP was proposed to be physiologically important for hydroxyapatite (HA) nucleation, cell attachment and collagen binding [3,4]. Our previous study demonstrated that BSP was able to enhance osteoblastic differentiation and new bone deposition in vivo by promoting the early bone mineralization [5]. No obvious cartilage formation was observed and molecular mechanisms

for the BSP regulation of osteogenesis are not clear. Osteogenesis is regulated by activation of extracellular signal-regulated kinase (ERK), p38 mitogen-activated protein kinase (MAPK), c-Jun N-terminal kinase (JNK), and protein kinase B/Akt [2,6,7]. BSP is a highly phoshorylated and glycosylated secretory protein that is enriched in sialic acids (SAs). Although the molecular mass of the core BSP protein is 33.6 kDa, the protein is usually identified as a 75 kDa band on SDS-PAGE [1,8]. The high apparent molecular weight is due to the extensive post-translational modifications (PTMs), including glycosylation. However, structural details of BSP glycosylation have not been determined and functions of different glycosylation residues in bone homeostasis are still not clear. Early studies showed that there were about four N-glycosylation

Abbreviations: ALP, alkaline phosphatase; AKT, protein kinase B; BSP, bone sialoprotein; DSPP, dentin sialophosphoprotein; DMP, dentin matrix protein-1; ECM, extracellular matrix; ERK, extracellular signal-regulated kinase; ETD, electron transfer dissociation; JNK, c-Jun N-terminal kinase; HA, hydroxyapatite; HCD, high-energy collision dissociation; HPAEC, high performance anion-exchange chromatography; HPLC, high-performance liquid chromatography; MAPK, mitogen-activated protein kinase; MEPE, matrix extracellular phosphoglycoprotein; Neu5Ac, N-acetyl neuraminic acid; Neu5Gc, N-glycolylneuraminic acid; OC, osteocalcin; OPG, osteoprotegerin; OPN, osteopontin; PAD, pulsed amperometric detector; PTMs, post-translational modifications; qPCR, quantitative real-time PCR; Runx2, runt-related transcription factor 2; RANK, receptor activator for nuclear factor-κB; RANKL, receptor activator for nuclear factor-κB ligand; rhiBSP, recombinant human integrin binding sialoprotein; SAs, sialic acids; VDR, vitamin D receptor ⁎ Corresponding author. ⁎⁎ Correspondence to: College of Pharmaceutical Sciences, Soochow University Medical College, Suzhou 215123, China. E-mail addresses: [email protected] (L. Xu), [email protected] (Z. Zhang). http://dx.doi.org/10.1016/j.yexcr.2017.09.034 Received 9 May 2017; Received in revised form 19 September 2017; Accepted 24 September 2017 0014-4827/ © 2017 Elsevier Inc. All rights reserved.

Please cite this article as: Xu, L., Experimental Cell Research (2017), http://dx.doi.org/10.1016/j.yexcr.2017.09.034

Experimental Cell Research xxx (xxxx) xxx–xxx

L. Xu et al.

The parameters used for MS and MS/MS data acquisition under the HCD-pd-ETD mode were: positive mode; top speed mode with 3 s cycle time; FTMS: scan range (m/z) = 350–2000; resolution = 120 K; AGC target = 2 × 105; maximum injection time (ms) = 50; Filter: precursor selection range = 350–2000; include charge state = 2–8; dynamic exclude after n times = 1; Decision: precursor priority = highest charge state then most intense; FTMSn (HCD): isolation mode = quadrupole; isolation window = 1.6; collision energy (%) = 30; resolution = 30 K; AGC target = 5 × 104; maximum injection time (ms) =35; microscan = 1; Product ion trigger: at least n product ions detected = 1; top N product ions = 10; Product ion table = 138.0545, 204.0867, 366.1396; ITMSn (ETD): isolation mode = quadrupole; isolation window = 2; use calibrated charge dependent ETD parameters = true; AGC target = 5 × 104; maximum injection time (ms) = 100; microscan = 1. For direct glycopeptide identification, the HCD and ETD MS2 data were searched separately using Byonic (version 2.6.46) with the following search parameters: peptide tolerance = 20 ppm; fragment tolerance = 0.02 Da (HCD) and 0.6 Da (ETD); missed cleavages = 2; modifications: methionine oxidation (common2), N-glycan search (Nglycan 57 human plasma). Peptide-spectrum match (psm) with score ≥ 100 were accepted.

sites and eight O-glycosylation sites on BSP, and its O-glycans, rather than N-glycans are highly sialylated [8,9]. Sialic acids, also known as Nacetylneuraminic acid, is a monosaccharide found on glycoproteins or glycolipid chains. It can be attached to the ultimate ends of galactose, N-acetylgalactosamine, or to another SA. SAs play an important role in regulating interactions of glycosylated molecules in tumor metastasis [10,11]. However, the function of SAs in bone development is still not clear. BSP purified from human bone contains more SA modifications than the recombinant BSP protein in a human cell line and has higher affinity for HA, suggesting that BSP sialylation may play a role in bone homeostasis [8]. In the present study, we employed a nano-flow liquid chromatography-mass spectrometry (LC-MS) [12] method to analyze the structural details of the BSP N-glycans and to determine whether SAs are enriched at their termini. The glycopeptides were analyzed in both high-energy collision dissociation (HCD) and HCD product-dependent electron transferred dissociation (HCD-pd-ETD) modes in parallel to improve specificity and accuracy. The total sialylation of intact BSP and BSP with N-glycans removed were further quantitated by high performance anion-exchange chromatography analysis coupled with pulsed amperometric detection (HPAEC-PAD) following complete enzymatic removal of SAs, and we examined whether complete removal of all terminal SAs or removal of N-glycans are involved in the regulation of mineralization by BSP in cultured MC3T3-E1 cells. The roles of different glycosylation status of BSP in this regulation are compared and evaluated.

2.2.2. Removal of N-glycans or SAs from rhiBSP rhiBSP solutions (100 µg in 1 mL PBS) were incubated with 5 mU PNGase F (E. meningosepticum, QA-Bio, Inc) or 5 mU sialidase (recombinant from A. ureafaciens in E. Coli, E-S001, QA-Bio, Inc) at 37 °C for 24 h, respectively. The de-sialylated rhiBSP or rhiBSP with N-glycans removed was collected by centrifugal ultrafiltration (Amicon Ultra-0.5 mL, Ultracel-10 K, Millipore) at 4500g for 10 min. The protein was washed twice with 0.3 mL PBS using the same method.

2. Materials and methods 2.1. Materials The pre-osteoblast cell line MC3T3-E1 Subclone 14 cells from mouse calvaria were obtained from the Cell Bank of the Chinese Academy of Sciences. α-Minimal Essential Medium (α-MEM) and TRIzol® reagent were purchased from Life Technologies. Fetal bovine serum (FBS), Lascorbic acid, β-glycerophosphate and dexamethasone were ordered from Sigma. Alkaline phosphatase Kit (ALP kit, N1891) ordered from Sigma. The ERK inhibitor, PD0325901, was purchased from Cayman Chemical. Enzymatic DeglycoMx Kit was ordered from QA-Bio, Inc. CHO-derived recombinant human integrin binding sialoprotein (rhiBSP, AAC95490) and goat anti-BSP (AF4014, 1:1000) antibody were purchased from R & D Systems. Rabbit anti-vitamin D receptor (VDR, 3277-1, 1:1000), anti-osteopontin (OPN, 2671-1, 1:1000) antibodies were purchased from Epitomics. Goat anti-osteoprotegerin (OPG, SC-390518, 1:1000), rabbit anti-OC (OC, SC-365797, 1:1000), rabbit anti-runt-related transcription factor 2 (Runx2, YT5356, 1:1000), and mouse anti-receptor activator NFκB ligand (RANKL, SC-59982, 1:1000) antibodies were purchased from Santa Cruz Biotechnology, Inc. Mouse anti-receptor activator NFκB (RANK, 64C1385, 1:1000) was purchased from Abcam. Rabbit anti-phospho-AKT (4058, 1:1000), antitotal AKT (9272, 1:1000), anti-phospho-ERK1/2 (9101, 1:1000), and anti-total ERK1/2 (9102, 1:1000) antibodies were obtained from Cell Signaling Technology.

2.2.3. Quantitation of terminal SAs of rhiBSP The solution (100 µg/mL) of intact rhiBSP, the N-glycan removed rhiBSP and de-sialyated rhiBSP were incubated with 5 mU sialidase at 37 °C for 24 h, respectively. The protein was collected by centrifugal ultrafiltration (Amicon Ultra-0.5 mL, Ultracel-3 K, Millipore, Billerica, MA) at 4500g for 10 min. The filter was washed twice with 0.3 mL PBS. The filtrates were combined with wash solution and were further diluted in 2 mL water prior to HPAEC-PAD analysis [13,14]. Two types of SAs, N-Acetyl neuraminic acid (Neu5Ac) and N-glycolylneuraminic acid (Neu5Gc) were used as standards in this work. Neu5Ac and Neu5Gc were dissolved in ultrapure water to make a stock solution of 2 mg/mL. The standard curve of Neu5Ac was prepared at five different concentrations (0.05, 0.25, 0.50, 1, 5 µg/mL). An equally mixed solution (Neu5Ac/Neu5Gc) was prepared and diluted to 5 µg/mL for each ingredient. The analysis was performed on a Metrohm 850 Professional System with a 919 IC auto-sampler plus, dual pumps and coupled with a pulsed amperometric detector (PAD, Herisau, Switzerland). SAs are separated in a Dionex Amino Trap trap column (4 × 50 mm) in line with a Dionex Carbopac PA10 analytical column (4 × 250 mm); using an isocratic mobile phase consisted of 50 mM NaOH and 150 mM NaOAc as described [14]. Data was acquired and analyzed using MagIC Net 2.4 (Herisau, Switzerland) software. All analysis was performed in duplicate.

2.2. Methods 2.2.1. Nano LC-MS/MS-Orbitrap analysis and data processing The solution of rhiBSP (25 μg in 0.1 mL PBS) was digested by protease E (~ 0.2 U, Sigma) at 37 °C overnight and analyzed by nanospray LC-MS/MS on an Orbitrap Fusion Tribrid (Thermo Scientific) coupled to an EASY-nanoLC System (Thermo Scientific) without trap column. Glyco-peptide mixtures were loaded onto a C18 column (15 cm × 50 μmi.d.) and separated at a flow rate of 300 nL/min using a gradient of 5–22% solvent B (100% acetonitrile with 0.1% formic acid) in 40 min, followed by an increase to 90% B in 10 min and held for another 5 min. Solvent A was 0.1% formic acid in water and solvent B was 0.1% formic acid in acetonitrile.

2.2.4. Cell culture MC3T3-E1 Subclone 14 cells were cultured in α-MEM containing 10% FBS, 100 units/mL penicillin G sodium, and 100 μg/mL streptomycin sulfate at 37 °C with 5% CO2. Media was replaced every 3 days. Cells at around 90% confluency were replated at a concentration of 1.5 × 105 cells/well in 6 well cell culture plates. Differentiation was stimulated by incubation in conditioned media containing L-ascorbic acid (50 μg/mL), β-glycerophosphate (10 mM), and dexamethasone (10−8 M) for up to 14 days as described [15]. 2

Experimental Cell Research xxx (xxxx) xxx–xxx

L. Xu et al.

Fig. 1. Analysis of glycopeptides of rhiBSP by MS/ MS in conjunction with HCD-pd-ETD. (A) HCD MS/ MS spectrum of a representative glycopeptide of rhiBSP. (B) ETD MS/MS spectrum of the representative glycopeptides of rhiBSP.

2.2.5. ALP enzyme activity assay ALP enzyme activity assay was performed spectraphotometrically and was adapted to a 96 well plate format. Briefly, 5 mL pNPP substrate buffer were prepared by dissolving one pNPP tablet of and one Tris buffer tablet in distilled water. Cell lysates were added in each well, followed by injection of 200 μL pNPP substrate solution. The plate was incubated in the dark at room temperature for 30 min. The plate was read at 405 nm in the end of the incubation using a Thermo fisher multi-well plate reader.

9.5 μL ddH2O, 1 μL primer F, 1 μL primer R, 1 μLcDNA. Then the samples in 96 well plates were mixed and proceed to qPCR by the ABI PRISM 7500 sequence detection system. The PCR amplification program was set up as follows: denaturation (95 °C，60 s), annealing (95 °C, 15 s), extension (60 °C, 60 s). Relative quantification of mRNA expression was analyzed using the 2-ΔΔCt method. The GAPDH was served as an internal control. All qPCR reactions were performed in triplicate. Primers used in these experiments were listed in Supplementary meterial Table S1.

2.2.6. ALP staining ALP staining was performed using phosphatase substrate (Sigma) according to the manufacturer's instructions. Briefly, cells were washed twice with PBS and were fixed by incubating with 2 mL of fix solution for 2 min at room temperature. The fix solution was then aspirated. Cells were washed twice again with PBS and then incubated with 1 mL of fresh AP staining solution in the dark for 20 min. The staining was stopped by aspirating the staining solution and washing the cells with PBS. The cells were then covered with mounting medium and stored at 4 °C before imaging. Differential expression of ALP will result in a red or purple stain. Microscopy Image was obtained using on an Eclipse TS100Inverted Routine Microscope (Nikon Instruments Inc.) [16].

2.2.9. Western blotting analysis Total proteins were extracted. Then, aliquots of 15 μg protein samples were mixed with protein loading buffer (Beyotime, China) (5:1, v/v) and boiled for 10 min. Proteins were separated by 12% SDS-PAGE for 60 min and then transferred to nitrocellulose membrane (NC membrane) (Millipore, USA) in 90 min. Next, the NC membrane was blocked with 5% (v/v) skimmed-milk for 2 h at room temperature. After three times of washing with TBST (Tris-HCL Buffered Saline Tween 20), the membrane was incubated with primary antibody at 4 °C overnight. Then, the membrane was washed three times again, incubated with secondary anti-mouse or anti-rabbit IgG conjugated to horseradish peroxidase (Sigma, St Louis, MO) at room temperature for 2 h. Protein brands were scanned and quantified by Image J software.

2.2.7. Von Kossa staining The media were aspirated and the cells were washed twice with PBS, fixed in 95% ethanol for 10 min, and washed three times with ddH2O. Cells were then incubated with 1% silver nitrate solution for 45 min under ultra-violet light, and then washed three times with ddH2O and then placed under running ddH2O for 10 min. The sodium thiosulfate reaction was allowed to proceed for 5 min. Cells were counterstained with van Gieson solution for 5 min. The cells were then fixed in 95% ethanol again for 10 min prior to air dry and storage. Mineralized nodules resulting from the silver deposit are black. The number of mineralized nodule were counted manually [17].

2.2.10. Statistical analysis All statistical analysis were performed using the Graph-Pad Prism 5.0 program. The data are presented as means ± SEM and statistically significant differences between mean values were determined using unpaired Student's t-tests, or ANOVA tests. A P value < 0.05 was considered statistically significant. 3. Results 3.1. Characterization of N-glycans in rhiBSP The enriched glycopeptides digested from rhiBSP were analyzed by MS/MS in HCD-pd-ETD mode. If two or more specific diagnostic ions, such as m/z 138.0545 (HexNAc fragment), 204.0867 (HexNAc), and 366.1396 (HexNAc-Hex), were precisely detected by HCD mode, ETD MS/MS mode would be triggered and minimize mistargeting of

2.2.8. RNA extraction and quantitative real-time PCR (qPCR) Total RNA was isolated using TRIzol reagent. RNA concentration was detected by Nanodrop2000. cDNA was synthesized from 500 ng of total RNA using RNA reverse transcription kit (TaKaR, Japan). Realtime PCR use a 25 μL system: 12.5 μL SYBR Green (Roche, Japan), 3

Experimental Cell Research xxx (xxxx) xxx–xxx

L. Xu et al.

glycopeptide due to interfering isobaric or isomeric ions. Fig. 1 showed the MS/MS spectra of a representative glycopeptide, acquired at HCD and ETD mode in parallel. Fig. 1A is the HCD MS/MS spectrum of the representative glycopeptide. The specific fragment ions observed at m/z 138.1, 204.1, and 336.1 were assigned as diagnosed ions to trigger the ETD MS/MS. They are fragment ions of cross-ring cleaved HexNAc, HexNAc and HexNAc-Hex, respectively. In addition, the fragment ions observed at m/z 274.1 and 292.1 were assigned as anhydro-SA and SA residue, confirming the presence of SA residue in this glycopeptide. As the energy used in HCD was too high to obtain the fragment ions with intact peptide, glycan and linkage domains, ETD was applied in parallel. Fig. 1B is the ETD MS/MS spectrum of the representative glycopeptide. The fragment ions observed in HCD MS/MS spectrum also presented in ETD MS/MS spectrum. Furthermore, the triply charged molecular ion was observed at 896.4. It is a doubly sialylated and singly fucosylated bi-antennary glycopeptide. The fragment ions observed at m/z 454.2, 657.3 and 819.5 were assigned as fragment ions with sialylation. In addition, the fragment ions with intact linkage domain were assigned at m/z 538.3, 684.4, 1049.6, 1065.5, 1211.6, 1262.5, 1373.8, 1577.8 and 1738.4 to confirm the glycosylation site. All assignments were labeled in the Fig. 1B. Byonic software offers glycopeptide analysis following the roles described above, and allows one N-glycan modification on the N – X (not P) – S/T consensus motif per peptide. Twenty eight different glycoforms of the glyco-peptides at four N-glycosylation sites (N104, N177, N182, N190) were identified with mass and composition chosen from its N-glycan 57 human plasma database by searching of Byonic software. These glycoforms are listed in Supplementary material Table S2. Four MS/MS spectra of glycopeptides represented each glycosylation site were shown in Fig. 2. It is the first time to confirm the presence of high sialylation in N-glycans of rhiBSP experimentally.

Fig. 3. Quantitative analysis of sialic acids in rhiBSP by HPAEC. (A) The chromatogram of mixed standards of Neu5Ac and Neu5Gc. (B) The chromatograms of sialic acid released from intact, de-N-glycosylated and de-sialylated rhiBSP from top to bottom.

abundance. We employed HPAEC-PAD approach to identify and quantitate the SAs on the rhiBSP. Two SA standards, Neu5Ac and Neu5Gc, were used (Fig. 3A). Compared to the SA standards, only Neu5Ac was detected in filtrates containing sialic acids released from intact rhiBSP and the rhiBSP with N-glycan removed (Fig. 3B). 0.26 ± 0.05 μg Neu5Ac was released from per μg rhiBSP protein and 0.17 ± 0.03 μg SA was released from per μg protein, respectively. Little Neu5Ac was detected after additional sialidase treatment of the desialylated rhiBSP, suggesting completion of the enzymatic digestion (Fig. 3B). Thus, treatment with sialidase for 24 h was used to obtain desialylated rhiBSP in the following activity assays. 3.3. Effects of N-glycosylation and terminal sialylation of BSP on osteogenesis

3.2. Quantitation of terminal Neu5Ac in rhiBSP

We examined the dose dependent effects of rhiBSP on differentiation of MC3T3-E1 Subclone 14 cells to osteoblasts. Osteogenic stimulation of the cells was initiated in the presence or absence of rhiBSP at 0, 1.0, 2.5 and 5.0 µg/mL. rhiBSP at 2.5 µg/mL significantly increased

Neu5Ac and Neu5Gc are the two most common sialic acids on recombinant glycoproteins and O-Acetylated SA is usually at very low

Fig. 2. MS/MS spectra of representative glycopeptides at each glycosylation site with HCD mode. (A) MS/MS spectrum of a representative glycopeptide at N104; (B) MS/MS spectrum of a representative glycopeptide at N177; (C) MS/MS spectrum of a representative glycopeptide at N182; (D) MS/MS spectrum of a representative glycopeptide at N190.

4

Experimental Cell Research xxx (xxxx) xxx–xxx

L. Xu et al.

Fig. 4. Effects of rhiBSP and rhiBSP with N-glycans or sialic acids removed on osteogenesis in MC3T3-E1 Subclone 14 cells. (A) ALP enzyme activity assay for cells cultured in conditioned medium for 10 days without (CTL) or with rhiBSP (2.5 µg/mL) (BSP) or 2.5 µg/mL of rhiBSP pre-digested by N-glycosidase (BSP-N) and Sialidase (BSP-Sia), respectivelly. (B) ALP staining for cells cultured in conditioned medium for 10 days without (CTL) or with rhiBSP (2.5 µg/mL) (BSP) or 2.5 µg/mL of rhiBSP pre-digested by N-glycosidase (BSP-N) and Sialidase (BSPSia), respectivelly. (C) von Kossa staining for cells cultured in conditioned medium for 14 days without (CTL) or with rhiBSP (2.5 µg/mL) (BSP) or 2.5 µg/mL of rhiBSP pre-digested by Nglycosidase (BSP-N) and Sialidase (BSP-Sia), respectivelly. *P < 0.05, vs CTL group; ***P < 0.001, vs CTL group; ##P < 0.01, vs BSP group; ###P < 0.001, vs BSP group.

3.5. Regulation of osteogenesis by terminal sialylation of BSP via ERK activation

the differentiation of osteoblasts as indicated by ALP staining and ALP enzymatic activity assay on day 10 after rhiBSP treatment (Fig. 4A and B). The stimulatory effect of rhiBSP was reduced after removal of all SA residues as evaluated by ALP enzyme activity assay (Fig. 4A). In consistent with its effect on osteogenesis, the von Kossa staining showed that mineralized nodules were also enhanced when treated with 2.5 µg/ mL rhiBSP. But its stimulatory effect on mineralization was abolished by removal of N-glycans residues especially removal of SA residues after 14 days incubation (Fig. 4C).

ERK signaling plays important role in integrin regulation of osteogenesis [6,18], we examined whether rhiBSP activates ERK and whether this activation requires rhiBSP N-glycosylation and terminal sialylation. Treatment of cells with 2.5 µg/mL rhiBSP activated ERK in MC3T3-E1 Subclone 14 cells 3 days post differentiation. However, rhiBSP with N-glycans removed was not effective, especially rhiBSP with SA residues removed was not able to activate ERK (Fig. 6A). Akt activation was not observed upon rhiBSP treatment (data not shown). To evaluate the potential involvement of ERK activation in the regulation of osteogenesis by BSP, we examined the effect of PD0325901, a specific inhibitor of ERK [19,20] on osteogenesis. Treatment of PD0325901 for 3 or 6 days in differentiating MC3T3-E1 Subclone 14 cells significantly decreased levels of VDR protein, and also VDR decrease was partially reversed in the presence of rhiBSP (Fig. 6B and C). Moreover, treatment of PD0325901 for 3 or 6 days did not affect the protein levels of OPG and RANKL (data not shown).

3.4. Effects of N-glycosylation and terminal sialylation of BSP on osteogenesis-related gene expression We next examined the effect of rhiBSP and rhiBSP with N-glycans or terminal SAs removed on the expression of some key osteogenic genes. rhiBSP significantly increased the mRNA levels of OC, OPG and VDR, and the effects were reversed by removal of N-glycansor terminal SAs of rhiBSP (Fig. 5A). Similar to the mRNA levels, rhiBSP addition also resulted in increase in the protein levels of OC, OPG and VDR, and the effects were reversed by removal of terminal SAs of rhiBSP, removal of N-glycans of rhiBSP decreased protein levels of OC and VDR on day 15 (Fig. 5B). mRNA as well as the protein levels of OPN, RANK, RANKL were not affected. Protein, but not mRNA, level of the early transcription factor Runx2 was increased by rhiBSP treatment and the effects were reversed by removal of N-glycans or terminal SAs on day 10. Interestingly, removal of terminal SAs of rhiBSP decreased the protein levels of OC and OPG as comparing with the untreated control group (Fig. 5B).

4. Discussion It has been well documented that BSP enhances osteogenesis [21–23] and promotes early bone mineralization in vivo [5], but the underlying molecular mechanisms have not been studied. Mineralization of the extracellular matrix in bone requires a well-defined matrix structure and HA nucleation. The process is complex and poorly understood. The highly anionic BSP protein undergoes extensive PTMs, 5

Experimental Cell Research xxx (xxxx) xxx–xxx

L. Xu et al.

Fig. 5. Regulation of osteogenesis-related gene expression and protein levels by rhiBSP and rhiBSP with N-glycans or sialic acids removed. (A) The mRNA expression was measured using the 2-ΔΔCt method.The GAPDH was served as an internal control. Cells were cultured for 10 days in conditioned medium (CTL), conditioned medium with rhiBSP (2.5 µg/mL) (BSP) and conditioned medium with rhiBSP pre-digested by N-glycosidase (BSP-N) and Sialidase (BSP-Sia), respectivelly. (B, C) Cells were cultured for 10 days or 14 days in conditioned medium (CTL), conditioned medium with rhiBSP (2.5 µg/mL) (BSP) and conditioned medium with rhiBSP pre-digested by N-glycosidase (BSP-N) and Sialidase (BSP-Sia), respectivelly. Whole cell lysates were subjected to immunoblot analysis utilizing the indicated antibodies, and the bands were quantified by densitometry. *P < 0.05, vs CTL group; **P < 0.01, vs CTL group; ***P < 0.001, vs CTL group; #P < 0.05, vs BSP group; ##P < 0.01, vs BSP group; ###P < 0.001, vs BSP group.

rhiBSP supplementation, and desialylation of rhiBSP abolished its stimulatory effects on mineralization by von Kossa staining. Moreover, rhiBSP supplementation increases mRNA and the steady state protein levels of OC, VDR and OPG, while these effects were abolished by removal of SA modifications. Importantly, the protein level of Runx2, a master transcription factor for osteogenesis [29], was also increased by rhiBSP, which could be reversed by removal of N-glycans or terminal SAs. Collectively, our results suggested that SAs on both N-glycans and O-glycans contribute to BSP function in osteogenesis. Interestingly, rhiBSP but not the desialylated form activates ERK phosphorylation. Pharmaceutical inhibition of ERK significantly decreased the protein level of VDR, but not OPG or RANKL, suggesting that rhiBSP-stimulated mineralization is likely due to ERK activation of VDR expression. VDR is a key nuclear receptor that binds nutritionally derived ligands and is important to maintain bone mineral homeostasis [30]. Vitamin D not only promotes mineral deposition in osteoblasts, but also enhances the synthesis of matrix collagen fibers in the absence of osteoblasts. Understanding the specific functions of O-glycans in osteogenesis requires precise removal of O-glycans without affecting N-glycans and its terminal sialylation, but the current method to remove O-glycans require the combined action of both O-glycosidase and sialidase, which will unavoidably lead to non-specific cleavage of the functional SA residues of the N-glycans. In our study, we first time identified the presence of terminal SAs in N-glycans and sialidase pre-treatment leads to the removal of terminal SA in both O-glycans and N-glycans, and we tried to use O-glycosidase to remove O-glycans in the rhiBSP without sialidase pre-treatment and estimated its efficiency using HCD and HCD-pd-ETD modes and identify O-glycosylation sites (Supplementary

which dramatically shift its molecular weight on SDS-PAGE [24]. The phosphorylation of BSP is not critical for its effect on osteogenesis in vitro and PTMs such as sulfation were rarely identified in the protein [25,26]. In addition, BSP protein is heavily glycosylated and sialylated, Miwa et al. reported a reduction in the apparent molecular masses of BSP in the Galnt1-null mice, indicating that ppGalNAc-T1 is indispensable for O-glycosylation at specific sites of BSP in the bone [8,27]. Moreover, human BSP proteins purified from different human bone or cells have different SA modifications and affinity for HA [8]. However, their structural details and functions are still not clear. In this work, we employed the MS/MS technique in combination with HCD-pd-ETD to identify terminal SA residues on all N-glycans. The sialylation on both O- and N-glycans were further confirmed by HPAECPAD analysis. The results are consistent with previous computational prediction [3], but inconsistent with previous MS studies [8]. One possibility is that recombinant BSP produced in CHO cells used in this study and that from human cell lines as used in previous study may have distinct distribution of differentially conjugated SAs, leading to different functions. Another possibility is that the SAs were released from glycopeptides at relative high temperature used for hydrazinolysis in those studies [8,28]. In our study, we identified four new N-glycosylation sites (N104, N177, N182, N190) on the rhiBSP, which provide valuable information to help understand specific functions of different BSP glycosylation sites and their terminal SA residues in osteogenesis. Our current study also demonstrated that rhiBSP and the terminal SA residues of rhiBSP are important regulated of osteogenesis. rhiBSP significantly affects differentiation of osteoblasts and this effect was reversed by removal of SA residue assessed by ALP enzyme activity. However, sialylation is critical for the enhanced bone mineralization by 6

Experimental Cell Research xxx (xxxx) xxx–xxx

L. Xu et al.

Fig. 6. Regulation of ERK activation by rhiBSP and rhiBSP with N-glycans or sialic acids removed. (A) Cells were cultured in conditioned medium (CTL), conditioned medium with rhiBSP (2.5 µg/mL) (BSP) and conditioned medium with rhiBSP pre-digested by N-glycosidase (BSP-N) and Sialidase (BSP-Sia), respectively, for 30 min after MC3T3-E1 Subclone 14 was cultured for 3 days. ** P < 0.01, vs CTL group; ###P < 0.001, vs BSP group. (B, C) Cells were cultured in conditioned medium without PD0325901 or without BSP (CTL), Cells were cultured in conditioned medium with various concentration of PD0325901 (PD) (0, 10, 100 ng/mL). Cells were cultured in conditioned medium with PD0325901 (0, 10, 100 ng/mL) and rhiBSP (0, 2.5 µg/mL) (PD+BSP) for 3 or 6 days. Whole cell lysates were subjected to immunoblot analysis using the indicated antibodies, and the bands were quantified by densitometry.* P < 0.05, vs 0 ng/mL PD; ** P < 0.01, vs 0 ng/mL PD.

assistance.

material Table S3). We also found that treatment with O-glycosidase alone partially removed O-glycans from rhiBSP and BSP O-glycosylation may play a role in osteoblasts differentiation and mineralized nodule formation (data not shown). However, complete understanding of BSP O-glycosylation in osteogenesis will require identification of specific Oglycosylation sites and expression and purification of recombinant BSP proteins with O-glycosylation sites mutated. This work still need to be further investigated.

Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.yexcr.2017.09.034. References

5. Conclusion

[1] L.W. Fisher, D.A. Torchia, B. Fohr, M.F. Young, N.S. Fedarko, Flexible structures of SIBLING proteins, bone sialoprotein, and osteopontin, Biochem. Biophys. Res. Commun. 280 (2001) 460–465. [2] C. Qin, O. Baba, W.T. Butler, Post-translational modifications of sibling proteins and their roles in osteogenesis and dentinogenesis, Crit. Rev. Oral Biol. Med. 15 (2004) 126–136. [3] C.E. Tye, G.K. Hunter, H.A. Goldberg, Identification of the type I collagen-binding domain of bone sialoprotein and characterization of the mechanism of interaction, J. Biol. Chem. 280 (2005) 13487–13492. [4] G.S. Baht, G.K. Hunter, H.A. Goldberg, Bone sialoprotein-collagen interaction promotes hydroxyapatite nucleation, Matrix Biol. 27 (2008) 600–608. [5] L. Xu, A.L. Anderson, Q. Lu, J. Wang, Role of fibrillar structure of collagenous carrier in bone sialoprotein-mediated matrix mineralization and osteoblast differentiation, Biomaterials 28 (2007) 750–761. [6] R.K. Jaiswal, N. Jaiswal, S.P. Bruder, G. Mbalaviele, D.R. Marshak, M.F. Pittenger, Adult human mesenchymal stem cell differentiation to the osteogenic or adipogenic lineage is regulated by mitogen-activated protein kinase, J. Biol. Chem. 275 (2000) 9645–9652. [7] N. Bokui, T. Otani, K. Igarashi, J. Kaku, M. Oda, T. Nagaoka, M. Seno, K. Tatematsu, T. Okajima, T. Matsuzaki, K. Ting, K. Tanizawa, S. Kuroda, Involvement of MAPK signaling molecules and Runx2 in the NELL1-induced osteoblastic differentiation, FEBS Lett. 582 (2008) 365–371. [8] M. Wuttke, S. Muller, D.P. Nitsche, M. Paulsson, F.G. Hanisch, P. Maurer, Structural

In this study, we have identified multiple novel N-glycosylation sites on the rhiBSP protein and demonstrated that both O- and N-glycans of rhiBSP are highly sialylated. The terminal SA residues of the rhiBSP are important in regulating osteogenesis. Conflict of interest The authors have no conflict of interest. Acknowledgements This work was supported by the National Natural Science Foundation of China (31470796, 81472105, 31401173), Suzhou Basic and Applied Medical Research Plan (SYS201673), and a project funded by Priority Academic Programme Development of Jiangsu Higher Education Institutions. We thank Ms. Xiaorui Shi for technical 7

Experimental Cell Research xxx (xxxx) xxx–xxx

L. Xu et al.

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19] S.D. Barrett, A.J. Bridges, D.T. Dudley, A.R. Saltiel, J.H. Fergus, C.M. Flamme, A.M. Delaney, M. Kaufman, S. LePage, W.R. Leopold, S.A. Przybranowski, J. SeboltLeopold, K. Van Becelaere, A.M. Doherty, R.M. Kennedy, D. Marston, W.A. Howard Jr, Y. Smith, J.S. Warmus, H. Tecle, The discovery of the benzhydroxamate MEK inhibitors CI-1040 and PD 0325901, Bioorg. Med. Chem. Lett. 18 (2008) 6501–6504. [20] Q.L. Ying, J. Wray, J. Nichols, L. Batlle-Morera, B. Doble, J. Woodgett, P. Cohen, A. Smith, The ground state of embryonic stem cell self-renewal, Nature 453 (2008) 519–523. [21] P. Bianco, L.W. Fisher, M.F. Young, J.D. Termine, P.G. Robey, Expression of bone sialoprotein (BSP) in developing human tissues, Calcif. Tissue Int. 49 (1991) 421–426. [22] H.Y. Zhou, H. Takita, R. Fujisawa, M. Mizuno, Y. Kuboki, Stimulation by bone sialoprotein of calcification in osteoblast-like MC3T3-E1 cells, Calcif. Tissue Int. 56 (1995) 403–407. [23] L.F. Cooper, P.K. Yliheikkila, D.A. Felton, S.W. Whitson, Spatiotemporal assessment of fetal bovine osteoblast culture differentiation indicates a role for BSP in promoting differentiation, J. Bone Miner. Res. 13 (1998) 620–632. [24] S.L. Cheng, C.F. Lai, S.D. Blystone, L.V. Avioli, Bone mineralization and osteoblast differentiation are negatively modulated by integrin alpha(v)beta3, J. Bone Miner. Res. 16 (2001) 277–288. [25] J.A. Gordon, C.E. Tye, A.V. Sampaio, T.M. Underhill, G.K. Hunter, H.A. Goldberg, Bone sialoprotein expression enhances osteoblast differentiation and matrix mineralization in vitro, Bone 41 (2007) 462–473. [26] K. Vincent, M.C. Durrant, A structural and functional model for human bone sialoprotein, J. Mol. Graph. Model. 39 (2013) 108–117. [27] H.E. Miwa, T.A. Gerken, O. Jamison, L.A. Tabak, Isoform-specific O-glycosylation of osteopontin and bone sialoprotein by polypeptide N-acetylgalactosaminyltransferase-1, J. Biol. Chem. 285 (2010) 1208–1219. [28] S. Takasaki, T. Mizuochi, A. Kobata, Hydrazinolysis of asparagine-linked sugar chains to produce free oligosaccharides, Methods Enzymol. 83 (1982) 263–268. [29] S. Vimalraj, B. Arumugam, P.J. Miranda, N. Selvamurugan, Runx2: structure, function, and phosphorylation in osteoblast differentiation, Int. J. Biol. Macromol. 78 (2015) 202–208. [30] M.R. Haussler, C.A. Haussler, L. Bartik, G.K. Whitfield, J.C. Hsieh, S. Slater, P.W. Jurutka, Vitamin D receptor: molecular signaling and actions of nutritional ligands in disease prevention, Nutr. Rev. 66 (2008) S98–S112.

characterization of human recombinant and bone-derived bone sialoprotein. Functional implications for cell attachment and hydroxyapatite binding, J. Biol. Chem. 276 (2001) 36839–36848. J. Zaia, R. Boynton, D. Heinegard, F. Barry, Posttranslational modifications to human bone sialoprotein determined by mass spectrometry, Biochemistry 40 (2001) 12983–12991. R. Kannagi, Molecular mechanism for cancer-associated induction of sialyl Lewis X and sialyl Lewis A expression-the Warburg effect revisited, Glycoconj. J. 20 (2004) 353–364. S. Mathieu, M. Prorok, A.M. Benoliel, R. Uch, C. Langlet, P. Bongrand, R. Gerolami, A. El-Battari, Transgene expression of alpha(1,2)-fucosyltransferase-I (FUT1) in tumor cells selectively inhibits sialyl-Lewis x expression and binding to E-selectin without affecting synthesis of sialyl-Lewis a or binding to P-selectin, Am. J. Pathol. 164 (2004) 371–383. S.W. Wu, T.H. Pu, R. Viner, K.H. Khoo, Novel LC-MS(2) product dependent parallel data acquisition function and data analysis workflow for sequencing and identification of intact glycopeptides, Anal. Chem. 86 (2014) 5478–5486. P.G. Stanton, Z. Shen, E.A. Kecorius, P.G. Burgon, D.M. Robertson, M.T. Hearn, Application of a sensitive HPLC-based fluorometric assay to determine the sialic acid content of human gonadotropin isoforms, J. Biochem. Biophys. Methods 30 (1995) 37–48. Z. Zhang, N.M. Khan, K.M. Nunez, E.K. Chess, C.M. Szabo, Complete monosaccharide analysis by high-performance anion-exchange chromatography with pulsed amperometric detection, Anal. Chem. 84 (2012) 4104–4110. D. Wang, K. Christensen, K. Chawla, G. Xiao, P.H. Krebsbach, R.T. Franceschi, Isolation and characterization of MC3T3-E1 preosteoblast subclones with distinct in vitro and in vivo differentiation/mineralization potential, J. Bone Miner. Res. 14 (1999) 893–903. J. Wang, M.J. Glimcher, J. Mah, H.Y. Zhou, E. Salih, Expression of bone microsomal casein kinase II, bone sialoprotein, and osteopontin during the repair of calvarial defects, Bone 22 (1998) 621–628. J. Rungby, M. Kassem, E.F. Eriksen, G. Danscher, The von Kossa reaction for calcium deposits: silver lactate staining increases sensitivity and reduces background, Histochem. J. 25 (1993) 446–451. M.K. Shin, M.K. Kim, Y.S. Bae, I. Jo, S.J. Lee, C.P. Chung, Y.J. Park, S. Min do, A novel collagen-binding peptide promotes osteogenic differentiation via Ca2+/calmodulin-dependent protein kinase II/ERK/AP-1 signaling pathway in human bone marrow-derived mesenchymal stem cells, Cell Signal. 20 (2008) 613–624.

8