2-O-phosphorylation of xylose has been detected in the glycosaminoglycan–protein linkage region, GlcAβ1-3Galβ1-3Galβ1-4Xylβ1-O-Ser, of proteoglycans. Recent mutant analyses in zebrafish suggest that xylosyltransferase I and FAM20B, a protein of unknown function that shows weak similarity to a Golgi kinase encoded by four-jointed, operate in a linear pathway for proteoglycan production. In the present study, we identified FAM20B as a kinase that phosphorylates the xylose residue in the linkage region. Overexpression of FAM20B increased the amount of both chondroitin sulfate and heparan sulfate in HeLa cells, whereas the RNA interference of FAM20B resulted in a reduction of their amount in the cells. Gel-filtration analysis of the glycosaminoglycan chains synthesized in the overexpressing cells revealed that the glycosaminoglycan chains had a similar length to those in mock-transfected cells. These results suggest that FAM20B regulates the number of glycosaminoglycan chains by phosphorylating the xylose residue in the glycosaminoglycan–protein linkage region of proteoglycans.
- chondroitin sulfate
- glycosaminoglycan–protein linkage region
- heparan sulfate
- xylose kinase
Sulfated GAGs (glycosaminoglycans), including HS (heparan sulfate) and CS (chondroitin sulfate), are linear polysaccharides consisting of a repetition of [(-4GlcAβ1-4GlcNAcα1-)n] and [(-4GlcAβ1-3GalNAcβ1-)n] disaccharide units respectively. The assembly of GAG chains is initiated by the synthesis of the so-called common GAG–protein linkage region (GlcAβ1-3Galβ1-3Galβ1-4Xylβ1-O-Ser), which is attached to specific serine residues of different core proteins. The linkage region tetrasaccharide is formed by sequential stepwise addition of monosaccharide residues by the respective specific glycosyltransferases, XylT (xylosyltransferase), GalT-I (galactosyltransferase I), galactosyltransferase II and GlcAT-I (β1,3-glucuronyltransferase I) . The repeating disaccharide region of HS is synthesized on the linkage region by the HS co-polymerase complex of EXT1 and EXT2 . In contrast, the repeating disaccharide region of CS is synthesized on the linkage region by enzyme complexes of bifunctional chondroitin synthases-1 , -2  and -3 , and chondroitin polymerizing factor . After synthesis of the GAG sugar backbone on this tetrasaccharide, numerous modifications, including sulfation, epimerization and desulfation, are performed in a spatiotemporal manner, producing mature and functional GAG chains that exert biological functions dependent on their specific structure .
To date, structural studies have shown the presence of various modifications of the GAG–protein linkage region. One of the modifications is phosphorylation of the xylose residue at position 2 [1,6]. This modification has been detected in both HS and CS derived from cell-rich tissues [7,8], and appears to affect the transfer of galactose and D-glucuronic acid residues by GalT-I and GlcAT-I respectively [9,10]. In vitro experiments using authentic substrates showed that phosphorylation of the xylose residue prevents the transfer of a galactose residue by GalT-I . In contrast, GlcAT-I could efficiently transfer a D-glucuronic acid residue to the phosphorylated trisaccharide in vitro . These results suggested that phosphorylation of the xylose residue takes place after transfer of the first galactose residue by GalT-I and before transfer of the D-glucuronic acid residue by GlcAT-I. In fact, phosphorylation of the xylose residue is most prominent after the addition of two galactose residues [11,12], although the biological role of this modification and the enzyme responsible for the phosphorylation remain unclear.
Recently, Eames et al.  reported that FAM20B and XylT-1 drive cartilage matrix production and inhibit perichondral bone formation during endochondral ossification. In addition, double-mutant analyses indicated that FAM20B and XylT-1 operate in a linear pathway for proteoglycan production. FAM20B is a member of a family of related proteins that has been named a family with sequence similarity 20 (FAM20) with three members (FAM20A, FAM20B and FAM20C) in mammals . Although the function of FAM20 proteins is as yet unknown, they are reported to have weak similarity to a protein named Four-jointed, which was recently identified as a Golgi kinase that phosphorylates serine or threonine residues within the extracellular cadherin domain of Fat and its transmembrane ligand, Dachsous . Hence, we hypothesized that FAM20B might be a kinase that phosphorylates the xylose residue of the GAG–protein linkage region. In the present paper, we describe the identification and characterization of FAM20B as a xylose kinase.
[γ-32P]ATP (3000 Ci/mmol) was purchased from PerkinElmer. Unlabelled ATP and bovine liver β-glucuronidase (EC 220.127.116.11) were obtained from Sigma. Shrimp alkaline phosphatase was purchased from Roche Molecular Biochemicals. α-TM (α-thrombomodulin) with a truncated linkage region tetrasaccharide, GlcAβ1-3Galβ1-3Galβ1-4Xyl, was purified and structurally characterized as described previously [16,17]. Galβ1-3Galβ1-4Xyl(2-O-phosphate)β1-O-Ser was chemically synthesized . Galβ1-3Galβ1-4Xylβ1-O-Ser was prepared by digestion of Galβ1-3Galβ1-4Xyl(2-O-phosphate)β1-O-Ser with alkaline phosphatase.
Construction of a soluble form of FAM20B
The cDNA fragment of a truncated form of FAM20B, lacking the first 30 N-terminal amino acids, was amplified by PCR with KIAA 0475 cDNA obtained from the Kazusa DNA Research Institute (Chiba, Japan) as a template using a forward primer (5′-CGGGATCCTCAGCTGCCAACCGGGAGGAC-3′) containing an in-frame BamHI site and a reverse primer (5′-CGGGATCCACTTCCAATCCATCTCATACC-3′) containing a BamHI site located 67 bp downstream of the stop codon. PCR was carried out with KOD-Plus DNA polymerase (Toyobo) for 30 cycles at 94 °C for 30 s, 58 °C for 30 s and 68 °C for 150 s in 5% (v/v) DMSO. The PCR fragment was subcloned into the BamHI site of pGIR201protA , resulting in the fusion of the insulin signal sequence and the Protein A sequence present in the vector, as described in .
Expression of a soluble form of FAM20B and enzyme assays
The expression plasmid (6.0 μg) was transfected into COS-1 cells on 100-mm-diameter plates using FuGENE™ 6 (Roche Molecular Biochemicals) according to the manufacturer's instructions. At 2 days after transfection, 1 ml of the culture medium was collected and incubated with 10 μl of IgG–Sepharose (GE Healthcare) for 1 h at 4 °C. The beads recovered by centrifugation at 150 g for 2 min were washed with and then resuspended in the assay buffer, and tested for kinase activity, as described below. Assays for kinase activity were carried out using α-TM, Galβ1-3Galβ1-4Xyl(2-O-phosphate)β1-O-Ser or Galβ1-3Galβ1-4Xyl-O-Ser as an acceptor and ATP as a phosphate donor. Kinase reactions were incubated in reaction mixtures containing the following constituents in a total volume of 20 μl: 1 nmol of α-TM, Galβ1-3Galβ1-4Xyl(2-O-phosphate)β1-O-Ser or Galβ1-3Galβ1-4Xyl-O-Ser, 10 μM [γ-32P]ATP (1.11×105 d.p.m.), 50 mM Tris/HCl (pH 7.0), 10 mM MnCl2, 10 mM CaCl2, 0.1% BSA and 10 μl of the resuspended beads. The mixtures were incubated for 4 h at 37 °C.
Characterization of the reaction products
Products of kinase reactions on α-TM were isolated by gel filtration on a Superdex peptide column with 0.2 M ammonium bicarbonate as the eluent. The 32P-labelled tetrasaccharide chains were released from α-TM by mild alkaline treatment using 0.5 M LiOH, derivatized with 2-aminobenzamide and then isolated by gel-filtration HPLC as described in . The isolated tetrasaccharides were digested with β-glucuronidase in 50 mM sodium acetate (pH 5.2) or alkaline phosphatase in 50 mM Tris/HCl (pH 8.0) and 1 mM MgCl2 at 37 °C. Periodate oxidation was performed in 0.02 M NaIO4 and 0.05 M sodium formate (pH 3.0) as described in . An aliquot of the compound was subjected to gel-filtration chromatography on a column (7.6 mm×500 mm) of Asahipak GS-320 (Asahi Chemical Industry) using 50 mM ammonium acetate as the eluent at a flow rate of 1 ml/min as described in .
The cDNA fragment encoding FAM20B was amplified using a forward primer (5′-CGGAATTCCAGGGGAGAAGGAAAAGAG-3′) containing an EcoRI site and a reverse primer (5′-CGGGATCCAAGTGTGAGAGAGGCATCCT-3′) containing a BamHI site. PCR was carried out with KOD-Plus DNA polymerase for 30 cycles at 94 °C for 30 s, 53 °C for 42 s and 68 °C for 180 s in 5% (v/v) DMSO. The PCR fragment was subcloned into the pEGFP-N1 expression vector (Clontech). The Golgi marker vector (pDsRed-Golgi) was constructed using the pECFP-Golgi vector (Clontech) that harbours a sequence encoding the N-terminal 81 amino acids of human β1-4-galactosyltransferase . The region from pECFP-Golgi was digested with NheI and BamHI, and subcloned into pDsRed-N1. Combinations of GFP (green fluorescent protein)-tagged and DsRed-Monomer-tagged expression vectors (3.0 μg each) were transfected into HeLa cells on glass-bottom dishes (Matsunami Glass) using FuGENE™ 6 according to the manufacturer's instructions. Fluorescent images were obtained using a laser-scanning confocal microscope, FluoView (Olympus).
Creation of stably transfected cell lines
The cDNA fragment encoding FAM20B was amplified from KIAA 0475 cDNA as a template using a forward primer (5′-CGGAATTCGTGTCGAGAATTACAAGTGTG-3′) containing an EcoRI site and a reverse primer (5′-CGGGATCCACTTCCAATCCATCTCATACC-3′) containing a BamHI site. PCR was carried out with KOD-Plus DNA polymerase for 30 cycles at 94 °C for 30 s, 53 °C for 42 s and 68 °C for 180 s in 5% (v/v) DMSO. The PCR fragments were subcloned into the EcoRI/BamHI site of the pCMV expression vector (Invitrogen). The expression plasmid was transfected into HeLa cells, and colonies were selected as described in .
FAM20B silencing in HeLa cells was performed using MISSION shRNA (short hairpin RNA) (Sigma). A hairpin construct identified by The RNAi Consortium clone number TRCN0000138872 was used. The shRNA plasmid (6.7 μg) was transfected into HeLa cells on 100-mm-diameter plates using FuGENE™ 6 according to the manufacturer's instructions. Transfectants were cultured in the presence of 0.4 μg/ml puromycin. Resultant colonies were then picked up and propagated for experiments.
Quantitative real-time RT (reverse transcription)–PCR
The cDNA was synthesized from total RNA extracted from HeLa cells as described in . Primer sequences were as follows: FAM20B, forward primer 5′-AGAGATCAAACCTGTCGCC-3′ and reverse primer 5′-CCAAAGTGTGACAGATCCCT-3′; and glyceraldehyde-3-phosphate dehydrogenase, forward primer 5′-ATGGGTGTGAACCATGAGAAGTA-3′ and reverse primer 5′-GGCAGTGATGGCATGGAC-3′. Quantitative real-time RT–PCR was performed as described in .
GAGs from HeLa cells were prepared as described in . The purified GAG fraction was digested with chondroitinase ABC or a mixture of heparinase and heparitinase, and then the digests were derivatized with 2-aminobenzamide and analysed by HPLC as described in .
Gel-filtration chromatography of GAGs
To determine the chain length of GAGs, the purified GAG fraction was subjected to reductive β-elimination using NaBH4/NaOH, and then analysed by gel-filtration chromatography on a column (10 mm× 300 mm) of Superdex 200 eluted with 0.2 M ammonium bicarbonate at a flow rate of 0.4 ml/min. Fractions were collected at 3.0 min intervals, freeze-dried and digested with chondroitinase ABC or a mixture of heparinase and heparitinase. The digests were derivatized with 2-aminobenzamide, and then analysed by HPLC on an amine-bound PA-03 column .
RESULTS AND DISCUSSION
FAM20B consists of 409 amino acids with type II transmembrane protein topology and shows weak similarity (21.4%) to a protein named Four-jointed, which was recently identified as a Golgi kinase that phosphorylates a subset of cadherin domains (see Supplementary Figure S1 at http://www.BiochemJ.org/bj/421/bj4210157add.htm) . To examine whether FAM20B could phosphorylate xylose on the GAG–protein linkage region, a soluble form of FAM20B was generated by replacing the first 30 amino acids of FAM20B with a cleavable insulin signal sequence and a Protein A IgG-binding domain, as described in the Experimental section, and then soluble FAM20B was expressed in COS-1 cells as a recombinant protein fused with the Protein A IgG-binding domain. When the expression plasmid containing the putative kinase–Protein A fusion was expressed in COS-1 cells, an approx. 76 kDa protein was secreted, as shown by Western blotting (see Supplementary Figure S2 at http://www.BiochemJ.org/bj/421/bj4210157add.htm). The fused putative kinase expressed in the medium was adsorbed on IgG–Sepharose beads to eliminate endogenous kinases, and then the protein-bound beads were used as an enzyme source. The bound fusion protein was assayed for xylose kinase activity using various linkage region compounds as acceptor substrates. As shown in Table 1, marked kinase activity was detected with α-TM containing a linkage region tetrasaccharide, GlcAβ1-3Galβ1-3Galβ1-4Xylβ1, on native core protein  and non-phosphorylated trisaccharide–serine Galβ1-3Galβ1-4Xylβ1-O-Ser, but not with the phosphorylated counterpart Galβ1-3Galβ1-4Xyl(2-O-phosphate)β1-O-Ser as the acceptor substrate. These findings indicate that the expressed protein is a kinase that phosphorylated the GAG–protein linkage region.
To identify kinase reaction products, the representative acceptor substrate, α-TM, was labelled by the kinase reaction using [γ-32P]ATP, and the linkage tetrasaccharide was liberated from α-TM by mild alkaline treatment, as described in the Experimental section. The isolated linkage tetrasaccharide fraction was derivatized with a fluorophore, 2-aminobenzamide, and fractionated by gel-filtration HPLC, resulting in two fluorescent components: 32P-labelled (peak I) and unlabelled tetrasaccharides (peak II) (Figure 1A). Peak I was shifted to a position corresponding to 2-aminobenzamide-derivatized GlcAβ1-3Galβ1-3Galβ1-4Xyl by alkaline phosphatase digestion. Peak II was co-eluted with authentic 2-aminobenzamide-derivatized GlcAβ1-3Galβ1-3Galβ1-4Xyl when they were co-injected, indicating that it was derived from the unused acceptor substrate. Thus peak I was analysed below. The 32P-labelled tetrasaccharide (peak I) was converted into 32P-labelled trisaccharide after treatment with β-glucuronidase (Figure 1B). This 32P-labelled trisaccharide was resistant to periodate oxidation (Figure 1C). In contrast, after phosphate removal by alkaline phosphatase (Figure 1D), the trisaccharide became sensitive to periodate oxidation (Figure 1E). As expected, the 2-aminobenzamide-derivatized peaks in Figures 1(D) and 1(E) were eluted at the positions of the 2-aminobenzamide-derivatized Galβ1-3Galβ1-4Xyl derived from peak II (Figure 1A) treated with β-glucuronidase (Figure 1F) and the products obtained after periodate oxidation of the 2-aminobenzamide-derivatized Galβ1-3Galβ1-4Xyl (Figure 1G) respectively. As the 2-aminobenzamide was attached to C-1 of xylose, the phosphate must have been attached to C-2, making the C-1/C-2 and the C-2/C-3 glycols resistant to periodate oxidation. These findings clearly showed that FAM20B is a xylose kinase that phosphorylates C-2 of the xylose residue in the GAG–protein linkage region of the tetrasaccharide sequence of α-TM.
To examine the intracellular localization of xylose kinase, a full-length form of FAM20B fused with EGFP (enhanced GFP) at the C-terminus (FAM20B–EGFP) was generated as described in the Experimental section. FAM20B–EGFP was then co-expressed with a Golgi marker (Golgi–DsRed) or an ER (endoplasmic reticulum) marker (ER–DsRed) in HeLa cells and analysed by confocal microscopy. FAM20B–EGFP (Figure 2A) was co-localized with the Golgi–DsRed marker (Figure 2C), whereas FAM20B–EGFP (Figure 2D) was not completely co-localized with the ER–DsRed marker (Figure 2F). These results suggested that FAM20B acts as a xylose kinase in the Golgi apparatus.
To examine further the physiological relevance of the xylose kinase, we investigated whether overexpression or knockdown of FAM20B changes the amount of GAG in HeLa cells. As shown in Table 2, the disaccharide composition and the amount of GAG isolated from each of the representative stable clones were analysed by HPLC, as described in the Experimental section. The results showed a correlation between the expression levels of FAM20B and the total amounts of both CS and HS in these stable clones. In addition, while no significant change was detected in the disaccharide composition of HS among these stable clones, differences in that of CS were observed corresponding to the expression level of FAM20B. Notably, the increase in the expression level of FAM20B was concomitant with the increase in the proportion of CS 6-sulfate to 4-sulfate. Together, results these indicated that the xylose kinase regulates the sulfation profile of CS chains as well as the total amount of GAG synthesized in cells.
Prompted by these observations, we next examined whether the increase in the amount of GAG in FAM20B-overexpressing cells is caused by an increased number of GAG chains or by greater elongation of individual GAG chains. For this analysis, the length of CS and HS chains obtained from FAM20B-overexpressing and mock-transfected cells was compared. Gel-filtration analysis using a Superdex 200 column revealed that the length of CS and HS chains in FAM20B-overexpressing cells was similar to that in mock-transfected cells (Figure 3), although the short CS chains in the overexpressing cells were particularly augmented. These findings indicated that the increase in the amount of CS and HS in FAM20B-overexpressing cells was mainly caused by the increased number of CS and HS chains. Although we cannot rule out the possibility that FAM20B may act on other substrates that could contribute to the enhanced effect on GAG biosynthesis, since a direct effect at the cellular level on xylose phosphorylation remains to be formally established, these results suggest that FAM20B regulates the amount of GAG chains by controlling the number of GAG chains and plays an important role in the biosynthesis of GAG.
We demonstrated recently that GlcAT-I could efficiently transfer a D-glucuronic acid residue to the phosphorylated trisaccharide–serine Galβ1-3Galβ1-4Xyl(2-O-phosphate)β1-O-Ser rather than to the non-phosphorylated counterpart Galβ1-3Galβ1-4Xylβ1-O-Ser . In addition, in rat articular cartilage explants, the introduction of GlcAT-I enhanced GAG synthesis was attributable to an increase in the abundance rather than the length of GAG chains, whereas antisense inhibition of GlcAT-I expression impaired GAG synthesis . Moreover, Bai et al.  showed that the transfection of Chinese-hamster ovary cell mutants defective in GlcAT-I with GlcAT-I cDNA augmented GAG synthesis to levels approximately double that in wild-type cells, suggesting that GlcAT-I regulates the expression of GAGs. Hence, phosphorylation of the xylose residue may be required for biosynthetic maturation of the linkage region tetrasaccharide sequence, which may be a prerequisite for the initiation and efficient elongation of the repeating disaccharide region of GAG chains. Xyl-2-O-phosphate has been found in both HS and CS from Drosophila to mammals [1,6,23]. In fact, a homologue of human FAM20B is present in Drosophila, suggesting that the possible involvement of phosphorylation of the xylose residue by FAM20B in the processing and maturation of the growing linkage region might be conserved during evolution.
To date, in addition to FAM20B, two FAM members (FAM20A and FAM20C) have been reported in mammals . Database searches suggested that the amino acid sequence of FAM20B displays 35.7 and 39.1% identity with FAM20A and FAM20C respectively. It was also reported that the FAM20A gene displays the most restricted expression pattern, whereas FAM20B and FAM20C are expressed in a wider variety of tissues and their expression patterns are very similar . It is therefore possible that FAM20A and FAM20C are also involved in phosphorylation of the xylose residue in the linkage region, but exhibit distinct or overlapping acceptor substrate specificities. Characterization of FAM20A and FAM20C is now in progress.
Toshiyasu Koike and Tomomi Izumikawa performed the research, Hiroshi Kitagawa designed the research, Toshiyasu Koike, Tomomi Izumikawa and Hiroshi Kitagawa analysed the data, Jun-ichi Tamura contributed new reagents, and Tomomi Izumikawa and Hiroshi Kitagawa wrote the paper.
This work was supported in part by a Scientific Research Promotion Fund from the Japan Private School Promotion Foundation, Core Research for Evolutional Science and Technology (CREST) of the Japan Science and Technology (JST) Corporation (to H. K.), and Grants-in-aid for Scientific Research-B [grant numbers 19390025 and 21390025] (to H. K.).
The nucleotide sequence data reported in this paper for FAM20B have been submitted to the DDBJ, EMBL, GenBank® and GSDB Nucleotide Sequence Databases under accession number AB480690.
Abbreviations: CS, chondroitin sulfate; EGFP, enhanced green fluorescent protein; ER, endoplasmic reticulum; FAM20, family with sequence similarity 20; GAG, glycosaminoglycan; GalT-I, galactosyltransferase I; GFP, green fluorescent protein; GlcAT-I, β1,3-glucuronyltransferase I; HS, heparan sulfate; RT, reverse transcription; shRNA, small hairpin RNA; α-TM, α-thrombomodulin; XylT, xylosyltransferase
- © The Authors Journal compilation © 2009 Biochemical Society