Identification and Characterization of the Human Xylosyltransferase I Gene Promoter Region*

Human xylosyltransferase I catalyzes the initial and rate-limiting step in the biosynthesis of glycosaminoglycans and proteoglycans. Furthermore, this enzyme has been shown to play a major role in the physiological development of bone and cartilage as well as in pathophysiological processes such as systemic sclerosis, dilated cardiomyopathy, or fibrosis. Here, we report for the first time the identification and characterization of the XYLT1 gene promoter region and important transcription factors involved in its regulation. Members of the activator protein 1 (AP-1) and specificity protein 1 (Sp1) family of transcription factors are necessary for the transcriptional regulation of the XYLT1 gene, which was proven by curcumin, tanshinone IIA, mithramycin A, and short interference RNA treatment. A stepwise 5′ and 3′ deletion of the predicted GC-rich promoter region, which lacks a TATA and/or CAAT box, revealed that a 531-bp core promoter element is able to drive the transcription on a basal level. A binding site for transcription factors of the AP-1 family, which is essential for full promoter activity, was identified by site-directed mutagenesis located 730 bp 5′ of the translation initiation site. The ability of this site to bind members of the AP-1 family was further verified by electrophoretic mobility shift assays. A promoter element containing this binding site was able to drive the transcription to about 79-fold above control in SW1353 chondrosarcoma cells. Our findings provide a first insight into the regulation of the XYLT1 gene and may contribute to understanding the processes taking place during extracellular matrix formation and remodeling in health and disease.

Proteoglycans represent an important group of glycosylated macromolecules, which are expressed in the extracellular matrix as well as on virtually every animal cell surface. They are essentially involved in cell-cell interactions, cell adhesion, extracellular matrix deposition, tumor cell growth, biomechanical lubrication, viral infections, cell proliferation and differentiation, as well as in neurite outgrowth (1). Proteoglycans are composed of a core protein to which varying numbers of glycosaminoglycans (GAGs) 2 are covalently attached. The bio-logical roles of proteoglycans are closely related to the presence and biosynthesis of GAG chains because they interact with various proteins such as growth factors, cytokines, and extracellular matrix components through specific saccharide sequences embedded in their chains (2). GAGs such as chondroitin sulfate, dermatan sulfate, heparan sulfate, and heparin are anionic linear polysaccharides. They consist of alternating disaccharide units that are attached to specific serine residues within the core protein by a uniform tetrasaccharide linker, D-GlcA-␤1,3-Gal-␤1,3-Gal-␤1,4-Xyl-Ser. The formation of GAG chains occurs successively by a complex interaction of different enzymes. However, the initial and rate-limiting step is catalyzed by the UDP-D-xylose:proteoglycan core protein ␤-D-xylosyltransferase I (XT-I, EC 2.4.2.26), which transfers ␤-D-xylose from UDP-xylose to selected serine residues of the core protein (3)(4)(5).
The gene coding for human XT-I is located on the short arm of chromosome 16 at the region 16p12.3. Although the XYLT1 cDNA was already cloned in 2000 (6) and the involvement of this enzyme in many physiological and pathophysiological processes is well established, little is known about the transcriptional activation and regulation of XT-I. Therefore, the aim of this work was to identify and characterize the promoter region of the XYLT1 gene. Hereby, we determined that a 531-bp promoter element is able to drive the transcription on a basal level. Additionally, we present evidence that members of both the Sp1 and the AP-1 family of transcription factors are necessary to drive the transcription of the XYLT1 gene at full strength. This work extends our knowledge of the transcriptional regulation of XT-I and may therefore contribute to a better understanding of extracellular matrix formation and remodeling in tissue development and during pathophysiological processes.
Nuclear Extracts and Gel Retardation Assays-Nuclear extracts were obtained from SW1353 chondrosarcoma cells using NE-PER nuclear and cytoplasmic extraction reagents (Pierce) containing a proteinase inhibitor mixture (Sigma) according to the manufacturer's protocol. Protein concentration in the nuclear fraction was determined by a BCA assay (Sigma). A 32-bp 5Ј-biotinylated oligonucleotide containing the potentially active AP-1 binding motif of the XT-I promoter was used (5Ј-GGG ACC AGA GAA GTG ACT CAG TGA ACA CTT AG-3Ј and complement). Detection of the AP-1 oligonucleotide complex was performed using the LightShift chemiluminescent EMSA kit (Pierce). Briefly, nuclear protein extract (4 g) or recombinant c-Jun protein (Promega, Mannheim, Germany) was incubated in a buffer containing 10 mM Tris, 50 mM KCl, 1 mM dithiothreitol (pH 7.5), 7.5% glycerol, and 75 ng/l poly(dI⅐dC) for 20 min on ice. Afterward, 20 fmol of the biotinylated oligonucleotide was added, and the reaction mixture was allowed to incubate at room temperature for another 20 min. Finally, the DNA⅐protein complexes were resolved on a 6% polyacrylamide gel (Invitrogen) in 0.5ϫ Tris borate-EDTA buffer for 1.5 h. For competition and supershift experiments, unlabeled probes or antibodies were added to the reaction mixture 20 min before the addition of the labeled probe. The antic-Jun antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
Synthesis of Promoter Construct Inserts-The first 600 bp 5Ј of the translation start site, including an XhoI restriction site at the 5Ј and an EcoRI restriction site at the 3Ј end of the putative XYLT1 promoter, were generated in a gene synthesis and inserted into a pUC57 vector (Biocat, Heidelberg, Germany). Afterward, this fragment was restricted by the respective enzymes and ligated into a pGl4.10 promoterless luciferase reporter vector (Promega). All other promoter fragments were obtained by PCR amplification from genomic DNA by standard means. The PCR products were subcloned into a pCR2.1 TOPO vector (Invitrogen) and afterward digested with Acc65I (New England Biolabs). The promoter fragments were analyzed by 1.5% agarose gel electrophoresis, gel-purified using the NucleoSpin Extract II kit, and subsequently ligated into the pGl4.10 luciferase reporter vector containing the synthesized 600-bp fragment. To ensure the fidelity of the cloned promoter fragments, all final constructs were sequenced using the vector specific primers pGl4F (5Ј-CTA GCA AAA TAG GCT GTC CC-3Ј) and pGl4R (5Ј-CTT AAT GTT TTT GGC ATC TTC CA-3Ј).
Transient Plasmid Transfections-Plasmids were purified from bacterial cultures using the GeneElute (HP) Plasmid Midiprep kit (Sigma). SW1353 cells were seeded on 4.0-cm 2 12-well plates. After 24 h, the cells were transiently transfected with a mixture of 1 g of the particular promoter fragment containing pGl4.10 plasmid and 10 ng of the pGl4.74 vector (Promega) using Lipofectamine 2000 transfection reagent (Invitrogen) in accordance with the manufacturers' instructions. The pGl4.74 vector, which contains the thymidine kinase promoter from herpes simplex virus and the Renilla luciferase reporter gene, was used as an internal control to normalize for the efficiency of the transient transfection. After 48 h, the luciferase activity was assayed with the Dual Luciferase Reporter assay system (Promega) on a Lumat LB9705 luminometer (EG&G, Berthold, Bad Wilbad, Germany). As a control for vector backbone-based luciferase expression, the promoterless pGl4.10 vector was used. Where appropriate, transfected cells were treated with medium supplemented with 5 ng/ml TGF␤ 1 24 h before promoter activity was assayed.
Determination of Enzymatic Xylosyltransferase Activity-The enzymatic xylosyltransferase activity was detected in the cell culture supernatant by a radiochemical xylosyltransferase assay as described previously (7).
Statistical Analysis of Promoter Activity Assay Data and Real Time Quantitative RT-PCR data-Promoter activity data were calculated as the ratio of the luminescence values for each XYLT1 promoter firefly luciferase reporter construct and the corresponding value for the cotransfected pGl4.74 vector for normalization of transfection efficiency compared with the respective promotorless vector. All assays were performed in triplicate, and each experiment was repeated a minimum of three times. Relative expression values and fold changes in expression of real time quantitative RT-PCR data were calculated using the equation published by Pfaffl et al. (8), which considers the PCR efficiency. Activity and expression levels were compared by the Mann-Whitney U test. p values were considered significant below 0.05. All tests were calculated with GraphPad Prism 4.0 (GraphPad Prism software, San Diego, CA).

RESULTS
Identification of XYLT1 Promoter Region and Activity Analysis of Promoter Luciferase Constructs-Using Genomatix ModelInspector software (9) the putative XYLT1 gene promoter region was identified. A 1638-bp genomic fragment 5Ј of the translation initiation start site was cloned upstream of a firefly luciferase reporter gene in a pGL4.10 vector, as described under "Experimental Procedures" (Fig. 1). Different 5Ј and 3Ј truncations of this putative promoter fragment were generated and transfected into SW1353 chondrosarcoma cells. The results of the dual luciferase assays are shown in Fig. 2A. The highest activities were observed for fragments spanning 1031 and 797 nucleotides of the 5Ј-flanking region, respectively, with an induction around 50-fold above control. A further trunca-tion of 88 nucleotides dramatically decreases the promoter activity to a basal induction of about 5-fold above control. This residual activity is completely lost when the promoter is truncated to 218 nucleotides. Interestingly, a 3Ј truncation of 218 nucleotides of the 797-bp and the 531-bp fragment nearly increases the promoter activity by 180% compared with the full-length construct (Fig. 2B) to 78.5 (Ϯ3.5)-fold and 8.3 (Ϯ0.2)-fold (mean Ϯ S.E.) above control, respectively. On the basis of these findings, the sequence between Ϫ797 and Ϫ708 was investigated in detail. MatInspector analysis (9) highlighted one AP-1 binding site located at Ϫ730 bp. Therefore, this binding site was mutated and deleted in both the Ϫ797-bp fulllength and in the 218-nucleotide 3Ј truncated Ϫ797-bp fragment. In both cases the activity decreased dramatically to the level measured for the constructs between Ϫ218 and Ϫ709 nucleotides, which are all lacking the AP-1 binding site, to about 5-fold above control (Fig. 2C).
EMSA and Supershift Analysis-The results of the EMSA and supershift assay using a 32-bp biotin-labeled probe containing the AP-1 binding site of the XYLT1 promoter (bold letters in Fig. 1) are shown in Fig. 3, A and B. The probe was either incubated with SW1353 nuclear extracts or with recom-  NOVEMBER 6, 2009 • VOLUME 284 • NUMBER 45 JOURNAL OF BIOLOGICAL CHEMISTRY 30777 binant c-Jun protein, which is a component of the AP-1 transcription factor complex. In both cases, the same specific EMSA profile with a considerable single-shifted band of the 5Ј-biotinylated probe was observed (lanes 2, Fig. 3, A and B,  respectively). The specificity of the transcription factor binding was proven by competition experiments, in which a molar excess ranging from 50-to 200-fold of unlabeled AP-1 probe was added to the binding reaction (Fig. 3A, lanes  3-6). The lowest concentration of unlabeled probe was sufficient to compete for the binding of the transcription factor to the biotin-labeled probe, demonstrating the specificity of protein binding. In experiments using a specific antibody against c-Jun protein, supershifted bands were observed accompanied by a considerable reduction of the shifted band (Fig. 3B, lanes 3 and 5).

Identification of XT-I Promoter
AP-1 Inhibitor Treatment-To determine the effects of an AP-1 inhibition on the XT-I mRNA expression, SW1353 cells were treated with curcumin (20 M) and tanshinone IIA (100 M), respectively, for 6 h. In both cases, the XT-I mRNA expression was reduced significantly to 41.8% (Ϯ8.5%) and 13.8% (Ϯ4.1%), respectively (Fig. 4) as detected by real time quantitative RT-PCR.
Mithramycin A Treatment-Because the region 5Ј of the XYLT1 translation initiation site is very GCrich and contains several binding sites for transcription factors of the Sp1 family, as highlighted by Mat-Inspector (Fig. 1), we determined whether members of this transcription factor family are involved in the regulation of the XYLT1 gene. Therefore, SW1353 cells were treated with different concentrations (10 nM to 1 M) of mithramycin A for 24 h, which prevents the binding of transcription factors to GC-rich sequences (10). All concentrations used led to a highly significant decrease of XT-I mRNA levels compared with nontreated cells (Fig. 5). In fact, the lowest final concentration of 10 nM mithramy-cin A in the cell culture supernatant was sufficient to reduce the XT-I mRNA expression to 14% (Ϯ0.45%).
TGF␤ 1 Treatment-To elucidate a putative TGF␤ 1 -mediated induction of the XT-I promoter activity, cells transfected with the 1638-bp, 797-bp, and the 797-bp AP-1 site mutated XT-I promoter construct were treated with medium supplemented with TGF␤ 1 (5 ng/ml) for 24 h. None of the constructs investigated exhibited an increased promoter activity after TGF␤ 1 treatment compared with untreated controls as measured by the Dual Luciferase assay (data not shown). To ensure a TGF␤ 1 inducibility of the native XYLT1 promoter in SW1353 cells, the XT-I mRNA expression of untransfected cells was measured by real time RT-PCR with and without TGF␤ 1 treatment for the same period of time. Here, an increase of 2.9 (Ϯ0.4) (mean Ϯ S.D.) compared with controls was detected.
Effect of siRNA Treatment on XT-I mRNA Expression and Xylosyltransferase Activity-Based on the results obtained from the mithramycin A treatment, SW1353 cells were transfected with siRNA specifically targeting two members of the Sp1 family, Sp1 and Sp3, respectively. As shown in Fig. 6A, the mRNA level of both Sp1 and Sp3 was significantly reduced to less than 30% (p Ͻ 0.0001) compared with controls treated with a scram-bled siRNA for a total of 96 h after transfection. The siRNA-mediated knockdown of Sp1 had no detectable effect on the XT-I mRNA levels, whereas the knockdown of Sp3 led to a highly significant reduction to Ͻ40% compared with controls (p Ͻ 0.0001) (Fig. 6B). Similar results were obtained with two other siRNA oligonucleotides targeting Sp3. We also observed a considerable decrease of 51% in the enzyme xylosyltransferase activity in the cell culture supernatant of Sp3 siRNA-transfected cells (57 Ϯ 9 dpm) compared with controls (116 Ϯ 7 dpm) 96 h after transfection.

DISCUSSION
Human XT-I is nearly ubiquitously expressed in all tissues except for liver and a few cell lines (4,6), albeit at comparatively low levels. The enzyme can be used as a diagnostic marker for systemic sclerosis (11,12), participates in bone development (13), and was recently proven to exhibit a key role in cartilage repair (14). Additionally, Hurtado et al. (15) demonstrated that a knockdown of XT-I significantly improves axon growth through scar tissue in the adult rat spinal chord. However, to date, little is known about the transcriptional regulation and possibilities of modulating the expression of the XYLT1 gene. To gain an insight into the transcriptional activation, we identified and characterized the promoter region of the XYLT1 gene and determined important transcription factors involved in its regulation.
As a first step, the in silico analysis of the sequence immediately upstream of the XYLT1 translation start site revealed that the investigated sequence does not contain a TATA box. In contrast, especially the first 300 nucleotides 5Ј of the translation start exhibit a very high GC content of about 78% containing several GC boxes. The absence of a TATA box and the presence of CpG islands are often encountered within housekeeping gene promoters (16 -18). This might be an explanation for the wide distribution of XT-I in different tissues, although the strength of the gene expression varies considerably in these tissues, also indicating tissue-specific regulation mechanisms (4,6). As shown in Fig. 1, among the assumed transcription FIGURE 2. Functional analysis of human XYLT1 promoter constructs. A, functional analysis of human XYLT1 promoter activity using 5Ј deletion constructs in SW1353 cells. Bars show fold increase in luciferase activity for the XYLT1 promoter constructs cloned into pGL4.10 vector compared with promoterless pGL4.10 vector (negative control). The numbers on the left indicate the 5Ј ends of the constructs relative to the translation initiation start site. The values are the mean Ϯ S.E. of triplicates from at least three independent experiments. B, effect of 218-bp 3Ј deletions on XYLT1 promoter activity for selected promoter constructs. C, comparison of the activities of the wild-type (wt) XYLT1 promoter fragment ranging from Ϫ797 to the translation start site (Ϫ797 ϩ 1 wt) and the equivalent 218-bp 3Ј truncated form (Ϫ797 Ϫ218 wt) to constructs in which the AP-1 binding site is either mutated (mut) or deleted (del). The mutated bases are underlined and printed in bold letters; the deletion is indicated by a gap. factor binding sites are several belonging to the Sp1 and AP-1 families. These binding motifs were also found to be active in many other ECM protein-coding genes, i.e. collagens such as Col11a2 (19) and Col24a1 (20), proteoglycan core proteins such as biglycan (21,22), decorin (23), glypican-3 (24) or murine glypican-4 (25). Binding sites for Sp1 and AP-1 are also involved in the transcriptional regulation of other enzymes participating in glycosaminoglycan biosynthesis, like UDP-glucose dehydrogenase (26,27). Because XT-I catalyzes the initial step in GAG biosynthesis (3)(4)(5) and a similar regulation of genes involved in the same biosynthetic pathway is very probable, an important participation of members of the Sp1 and AP-1 protein family in the regulation of the XYLT1 expression is very likely.
The analysis of the activity of 5Ј and 3Ј truncated XYLT1 promoter constructs revealed that the highest transcriptional activity was detected for the AP-1 consensus sequence TGACT containing fragments Ϫ1031 ϩ 1 and Ϫ797 ϩ 1 ( Fig. 2A). All larger promoter constructs exhibited only about half the activity. This might indicate the presence of negatively regulating cis-acting elements in this region. Further truncated promoter constructs, completely lacking the AP-1 binding site, even revealed a more dramatic decrease in promoter activity. By sitedirected mutagenesis, EMSA, and supershift experiments, we additionally confirmed the participation of the AP-1 binding motif in driving the XYLT1 transcription. Using an antibody against c-Jun as a part of the AP-1 transcription factor complex, we clearly identified this protein as the binding species.
An interesting result was obtained for a 3Ј truncation of 218 nucleotides of the Ϫ797 ϩ 1 and Ϫ531 ϩ 1 promoter constructs, which both nearly doubled the transcriptional activity of the respective fragment (Fig. 2B). This is possibly caused by the binding of negative regulatory elements within the first 218   nucleotides 5Ј of the translation initiation site. Another possible reason might be the formation of distinct secondary structures caused by the high GC content, leading to a reduced transcriptional level in the full-length constructs. These assumptions are further supported by the fact that the Ϫ218 bp construct itself has no detectable promoter activity (Fig. 2B).
To investigate the participation of AP-1 on the native XYLT1 promoter, SW1353 cells were treated with the two different AP-1 inhibitors, curcumin and tanshinone IIA (28 -31). As shown in Fig. 4, a significant reduction of XT-I mRNA compared with untreated controls could be observed for both inhibitors, again confirming an essential role of this transcription factor in driving the transcription of the XYLT1 gene.
Because transcription factors of the Sp1 family are frequently involved in the basal expression of extracellular matrix genes, as well as the transcription of many TATA-less promoters depending on Sp1 family proteins (18,32), we investigated a possible participation of these proteins in the transcriptional regulation of the XYLT1 gene. Especially the first 650 nucleotides 5Ј of the translation initiation start site contain several predicted binding sites for Sp1 family members. In the majority of promoters containing binding sites for Sp1 family members, these transcription factors provide a basal level of gene expression (32,33). This is in concordance with the observation that a basal transcriptional level was detected for promoter constructs spanning the first 531-709 nucleotides 5Ј of the translation initiation site. Therefore, SW1353 cells were treated with mithramycin A, a well known GC-specific DNA-binding drug, which prevents the binding of Sp1 family transcription factors (10). Even very low concentrations at the nanomolar level of mithramycin A were sufficient to reduce the XT-I mRNA amount to less than 15%. To gain a more detailed insight into which member of the Sp1 family might participate in the transcriptional regulation of the XYLT1 gene, we used specific siRNA, targeting the ubiquitous Sp1 family members, Sp1 and Sp3. Interestingly, only the selective inhibition of Sp3 caused a significant reduction of the XT-I mRNA level to less than 40%, whereas the specific Sp1 inhibition showed no detectable effect. This indicates that the transcriptional regulation of XYLT1 may depend primarily on Sp3. Nevertheless, the transcriptional activity of Sp1 family proteins, and in particular that of Sp3, is thought to be dependent on the cellular context, the structure and arrangement of the recognition sites, as well as on the Sp1:Sp3 ratio within the cell (34,35). Changes in the cellular environment and response to external stimuli have the potential to alter the respective levels of these transcription factors and may modulate their function (34 -37). In addition, it cannot be excluded that a knockdown of Sp3 leads to overall changes in the transcription of the cell and its normal function, causing a down-regulation of XT-I gene expression in a manner unrelated to Sp3 directly regulating XT-I promoter activity. Therefore, an identification of the exact binding site has to be elucidated in future studies.
Although the mRNA data obtained in this study for SW1353 cells and a previously published study for human cardiac fibroblasts (12) provide evidence that the native XYLT1 promoter is inducible by TGF␤ 1 , as detected by real time quantitative RT-PCR, we did not detect a TGF␤ 1 -mediated induction of the transfected XYLT1 promoter constructs. This indicates that the TGF␤ 1 responsible element of the XYLT1 promoter is not included within the constructs analyzed and may be located upstream of the investigated fragments. The occurrence of such responsive elements in far distance to the proximal promoter is not uncommon and was reported among others for the COL1A2 or JUNB promoter regions, respectively (38,39).
In conclusion, we have for the first time identified and characterized the proximal promoter region of XYLT1. We showed that a 531-nucleotide promoter fragment is sufficient to drive the transcription on a basal level. Furthermore, we demonstrated that the c-Jun/AP-1 transcription factor is essential for full XYLT1 promoter activity in SW1353 cells. We also provided the first hints that transcription factors of the Sp1 family, especially Sp3, are very likely involved in the regulation of the XT-I mRNA expression. Taken together, our findings present new insights into the regulation of XT-I and may contribute to understanding of the regulatory mechanisms of this enzyme and for extracellular matrix formation in health and disease. Moreover, these results might help to find novel ways of modulating the XYLT1 gene expression, especially in the development of therapeutic strategies for the treatment of fibrotic remodeling processes or cartilage repair.