Molecular Cloning and Characterization of a Novel 3′-Phosphoadenosine 5′-Phosphosulfate Transporter, PAPST2*

Sulfation is an important posttranslational modification associated with a variety of molecules. It requires the involvement of the high energy form of the universal sulfate donor, 3′-phosphoadenosine 5′-phosphosulfate (PAPS). Recently, we identified a PAPS transporter gene in both humans and Drosophila. Although human colonic epithelial tissues express many sulfated glycoconjugates, PAPST1 expression in the colon is trace. In the present study, we identified a novel human PAPS transporter gene that is closely related to human PAPST1. This gene, called PAPST2, is predominantly expressed in human colon tissues. The PAPST2 protein is localized on the Golgi apparatus in a manner similar to the PAPST1 protein. By using yeast expression studies, PAPST2 protein was shown to have PAPS transport activity with an apparent Km value of 2.2 μm, which is comparable with that of PAPST1 (0.8 μm). Overexpression of either the PAPST1 or PAPST2 gene increased PAPS transport activity in human colon cancer HCT116 cells. The RNA interference of the PAPST2 gene in the HCT116 cells significantly reduced the reactivity of G72 antibody directed against the sialyl 6-sulfo N-acetyllactosamine epitope and total sulfate incorporation into cellular proteins. These findings indicate that PAPST2 is a PAPS transporter gene involved in the synthesis of sulfated glycoconjugates in the colon.


Sulfation is an important posttranslational modification associated with a variety of molecules. It requires the involvement of the high energy form of the universal sulfate donor, 3-phosphoadenosine 5-phosphosulfate (PAPS). Recently, we identified a PAPS transporter gene in both humans and Drosophila.
Although human colonic epithelial tissues express many sulfated glycoconjugates, PAPST1 expression in the colon is trace. In the present study, we identified a novel human PAPS transporter gene that is closely related to human PAPST1. This gene, called PAPST2, is predominantly expressed in human colon tissues. The PAPST2 protein is localized on the Golgi apparatus in a manner similar to the PAPST1 protein. By using yeast expression studies, PAPST2 protein was shown to have PAPS transport activity with an apparent K m value of 2.2 M, which is comparable with that of PAPST1 (0.8 M). Overexpression of either the PAPST1 or PAPST2 gene increased PAPS transport activity in human colon cancer HCT116 cells. The RNA interference of the PAPST2 gene in the HCT116 cells significantly reduced the reactivity of G72 antibody directed against the sialyl 6-sulfo N-acetyllactosamine epitope and total sulfate incorporation into cellular proteins. These findings indicate that PAPST2 is a PAPS transporter gene involved in the synthesis of sulfated glycoconjugates in the colon.
Sulfation of a variety of molecules, including glycoproteins, proteoglycans, and glycolipids, is an important posttranslational modification. The process requires the involvement of the high energy form of the universal sulfate donor, namely, 3Ј-phosphoadenosine 5Ј-phosphosulfate (PAPS). 2 In higher organisms, PAPS is synthesized in the cytosol or nucleus by PAPS synthetases (1,2) and is subsequently translocated into the Golgi lumen via the PAPS transporter(s). Because most of the sulfation of glycoconjugates occurs in the Golgi apparatus, the translocation of PAPS is considered to be an essential process.
Recently, we identified and characterized a PAPS transporter in both humans and Drosophila (3). Human PAPST1 and the Drosophila ortholog SLALOM (SLL) are Golgi-localized proteins that exhibit PAPS-specific transport activity. Analysis of the sll gene by using the RNA interference (RNAi) fly demonstrated that the PAPS transporter is essential for viability in vivo (3). Furthermore, Lüders et al. (4) demonstrated that sll is involved in growth factor signaling pathways during patterning and morphogenesis. Heparan sulfate proteoglycans (HSPGs) possess glycosaminoglycan chains that contain diversely sulfated uronic acid and glucosamine residues. Cell surface HSPGs are involved in a variety of developmental signaling pathways, and the functions of HSPGs are dependent on their sulfation state (5)(6)(7)(8)(9)(10). A mutation in the sll gene caused defects in multiple signaling pathways, including Wingless and Hedgehog signaling, probably because of the lack of HSPG sulfation (4).
In the present study, we attempted to identify the PAPS transporter that is responsible for the sulfation of glycoconjugates in the colon tissue. We found a gene that is closely related to the human PAPST1 gene by performing a BLAST search of data bases. This gene, called PAPST2, is preferentially expressed in human colon tissues. The PAPST2 protein exhibited PAPS transport activity similar to that of the PAPST1 protein. Here, we report the functional properties of this novel PAPS transporter. Ltd. (Tokyo, Japan). Fluorescein isothiocyanate (FITC)-conjugated anti-c-Myc monoclonal antibody (mAb) (9E10) and rhodamine-conjugated anti-influenza hemagglutinin epitope (HA) mAb (HA-probe, F-7) were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). All the other reagents used in the study were of the highest purity grade available commercially.

EXPERIMENTAL PROCEDURES
Isolation of Human PAPS Transporter cDNA and Construction of Expression Plasmids-The human PAPST2 gene was identified and cloned using the same procedures as described previously (3). The amino acid sequence of the open reading frame of UGTrel1 (18) was used as a query sequence for the TBLASTN search that was performed to detect novel genes. To obtain the cDNA of NM_015948, a human gene that was identified in this study, and to create recombination sites for the GATEWAY TM cloning system (Invitrogen), we used two steps of attB adaptor PCR and prepared attB-flanked PCR products. For the first gene-specific amplification, a forward template-specific primer with attB1, 5Ј-AAAAAGCAGGCTTCCATAATGGCATGGACTTG-3Ј, and a reverse template-specific primer with attB2, 5Ј-AGAAAGCT-GGGTCTACAGTCTGTGCCAGCGT-3Ј, were used. PCR was performed using Platinum Pfx DNA polymerase (Invitrogen) and a cDNA library derived from human colon tissue. The insertion of a complete attB adaptor and cloning into the pDONR TM 201 vector were performed in accordance with the manufacturer's protocol to create an entry clone for use during the subsequent subcloning steps.
The entry clone was subcloned into the appropriate expression vectors by using the GATEWAY TM cloning system in accordance with the manufacturer's protocol. A 3 ϫ HA epitope tag or a c-Myc tag was inserted into the expression vectors at the position corresponding to the C terminus of the expressing protein.
Transient Transfection and Immunofluorescence Microscopy-Transient transfection and immunofluorescence microscopy were performed by using one of two procedures. The first procedure is similar to one described previously (3,19). Briefly, HCT116 cells were subcultured onto a 4-well Lab-Tek chamber slide (Nalge Nunc International) in Dulbecco's modified Eagle's medium/Ham's nutrient mixture F-12 (1:1) containing 10% fetal bovine serum. After 24 h of subculturing, the HCT116 cells were transfected with 0.25 g/well of pCXN2 (20), pCXN2 inserted with HA-tagged PAPST1, or pCXN2 inserted with HA-tagged PAPST2 by using Lipofectamine 2000 reagent (Invitrogen) in accordance with the manufacturer's protocol. After 72 h, the cells were fixed in phosphate-buffered saline (PBS) containing 4% paraformaldehyde for 30 min at 4°C, and they were then permeabilized in a permeabilizing buffer (PBS containing 0.1% Triton X-100) for 1 h at 4°C. The cells were subsequently double immunostained with rhodamine-conjugated anti-HA mAb and anti-␤1,4-galactosyltransferase 1 mAb (21) as described previously (3,19). Finally, the cells were washed four times and mounted with PermaFluor (Thermo Shandon, Pittsburgh, PA). The fluorescence was observed using a confocal laser scanning microscope, LSM5 Pascal (Carl Zeiss, Goettingen, Germany). In the second procedure, HA-tagged PAPST2 and c-Myc-tagged PAPST1 were expressed simultaneously in the HCT116 cells. After 24 h of subculturing, the cells were transfected with 0.25 g/well of pCXN2 or pCXN2 inserted with HA-tagged PAPST2 and pCXN2 inserted with c-Myc-tagged PAPST1 by using Lipofectamine 2000 reagent. The cells were fixed, permeabilized, and immunostained with FITC-conjugated anti-c-Myc mAb and rhodamine-conjugated anti-HA mAb for 30 min at 37°C after 72 h. The remainder of the procedure was the same as that described above.
Stable Transfection and Subcellular Fractionation-HCT116 cells were subcultured onto 6-cm dishes in Dulbecco's modified Eagle's medium/Ham's nutrient mixture F-12 (1:1) containing 10% fetal bovine serum. After 24 h, the cells were transfected with 8 g of pCXN2 vector, pCXN2 inserted with HA-tagged PAPST1, or pCXN2 inserted with HA-tagged PAPST2 by using Lipofectamine 2000 reagent in accordance with the manufacturer's protocol. The transfectants were selected by the addition of 600 g/ml of geneticin (Invitrogen) to the medium and cultured for 1 month after 48 h.
Subcellular fractionation was performed as described previously (3,19). The cells were harvested and suspended in 10 mM HEPES-Tris (pH 7.4) containing 0.25 M sucrose, 1 mM phenylmethylsulfonyl fluoride, and 1 g/ml each of leupeptin, aprotinin, and pepstatin A. The cells were then homogenized using a Dounce homogenizer. The lysate was centrifuged at 1,000 ϫ g for 10 min to remove the unlysed cells and cell wall debris. The supernatant was then centrifuged at 7,700 ϫ g for 10 min at 4°C, and the supernatant was further centrifuged at 100,000 ϫ g to yield a pellet of P100 membrane fraction.
Subcellular Fractionation of Yeast and Transport Assay-Yeast (Saccharomyces cerevisiae) strain W303-1a (MATa, ade2-1, ura3-1, his3-11,15, trp1-1, leu2-3,112, and can1-100) was transformed by the lithium acetate procedure (22) using a yeast expression vector, YEp352GAP-II (23). These transformed yeast cells were grown at 30°C in a synthetic defined medium, which did not contain uracil, for selecting transformants. Subcellular fractionation and nucleotide sugar transport assays were performed as described previously (3,19). The cells were harvested, washed with ice-cold 10 mM NaN 3 , and converted into spheroplasts by incubation at 37°C for 30 min in spheroplast buffer (1.4 M sorbitol, 50 mM potassium phosphate (pH 7.5), 10 mM NaN 3 , 40 mM 2-mercaptoethanol, and 1 mg of Zymolyase 100T/g of cells). The spheroplasts were pelleted using a refrigerated centrifuge and washed twice with 1.0 M ice-cold sorbitol to remove traces of zymolyase. The cells were suspended in ice-cold lysis buffer (0.8 M sorbitol in 10 mM triethanolamine (pH 7.2), 5 g/ml of pepstatin A, and 1 mM phenylmethylsulfonyl fluoride) and subsequently homogenized using a Dounce homogenizer. The lysate was centrifuged at 1,000 ϫ g for 10 min to remove the unlysed cells and cell wall debris. The supernatant was then centrifuged at 10,000 ϫ g for 15 min at 4°C, which yielded a pellet of P10 membrane fraction. The supernatant was further centrifuged at 100,000 ϫ g to yield a pellet of P100 membrane fraction. Each fraction (200 g of protein) was then incubated in 100 l of reaction buffer (20 mM Tris-HCl (pH 7.5), 0.25 M sucrose, 5.0 mM MgCl 2 , 1.0 mM MnCl 2 , and 10 mM 2-mercaptoethanol) that contained 1 M radiolabeled substrate at 30°C for 5 min. After incubation, the radioactivity incorporated in the microsomes was trapped using a 0.45-m nitrocellulose filter and measured using liquid scintillation. The amount of incorporated radioactivity was calculated as the difference from the background value obtained from the same assay at 30°C for 0 min for each sample.
Western Blot Analysis-Fifty micrograms of protein from each sample was added to 3ϫ sodium dodecyl sulfate (SDS) sample buffer (New England Biolabs Inc., Beverly, MA) and subsequently incubated at room temperature for 2 h. The samples were fractionated on a 2-15% SDSpolyacrylamide gel gradient (Daiichi Pure Chemicals Co., Ltd., Tokyo, Japan). The separated proteins were electrotransferred onto a polyvinylidene difluoride membrane. The HA-tagged proteins were immunostained with anti-HA mouse mAb and horseradish peroxidase-conjugated anti-mouse IgG mAb. Bound horseradish peroxidase was detected using ECL plus (Amersham Biosciences) in accordance with the manufacturer's instructions.
Quantitative Analysis of the PAPST2 Transcript in Human Tissues by Real-time PCR-The amount of PAPST1 and PAPST2 transcripts in human tissues was determined by real-time PCR. Total RNA was extracted from human tissues by the method of Chomczynski and Sacchi (24). First strand cDNA was synthesized using a Superscript II First Strand Synthesis kit (Invitrogen) in accordance with the manufacturer's instructions. Real-time PCR was performed using a qPCR Mastermix (QuickGoldStar; Eurogentec, Seraing, Belgium) and ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Foster, CA). The PCR primer pair sequences and TaqMan probes used for each gene were as follows. For the detection of PAPST1, the forward primer 5Ј-GGCAG-GCCCTGAAGCT-3Ј, reverse primer 5Ј-TGCGGGTCATCACTCTT-TC-3Ј, and probe 5Ј-CCACAGGGCTCCAGGTGTCTTATCTG-3Ј were used. For the detection of PAPST2, the forward primer 5Ј-GAT-TAGGCCCTGCAGTAACATT-3Ј, reverse primer 5Ј-ATCCAGTGA-GGGAAAAAAGGA-3Ј, and probe 5Ј-TGTGCAAAGAATCCAGTT-CGGACCTA-3Ј were used. The probes were labeled at the 5Ј-end with the reporter dye FAM and at the 3Ј-end with the quencher dye TAMRA. The relative amounts of PAPST1 and PAPST2 transcripts were normalized with respect to the human glyceraldehyde 3-phosphate dehydrogenase (GAPDH) transcripts present in the same cDNA.
Northern Blot Analysis-PAPST2 cDNA probe (full-length open reading frame) was prepared by random priming using [␣-32 P]dCTP and BcaBest TM labeling kit (Takara Bio Inc., Shiga, Japan) in accordance with the manufacturer's instructions. Poly(A) ϩ RNA was isolated using Oligotex-dT30 (super) mRNA purification kit (Takara Bio Inc.) in accordance with the manufacturer's instructions. Poly(A) ϩ RNA derived from each sample was separated by 1.2% agarose gel containing 2.2 M formaldehyde and then transferred onto Hybond-XL nylon membrane (Amersham Biosciences). The membranes were prehybridized in the hybridization solution (5ϫ SSPE (standard saline phosphate with EDTA; 150 mM NaCl, 10 mM NaH 2 PO 4 , 1 mM EDTA), 5ϫ Denhardt's solution, and 0.2 mg/ml of sermon sperm DNA) for 2 h at 42°C. The membranes were hybridized overnight in hybridization solution containing 2 ϫ 10 6 cpm/ml of each radiolabeled probe at 42°C. Following hybridization, the membrane was washed in 2ϫ SSPE containing 0.1% SDS at room temperature and in 0.2ϫ SSPE containing 0.1% SDS at 50°C. The radiolabeled materials were detected by using a BAS-2000 imaging analyzer (Fuji Photo Film). The membrane was reprobed with a human GAPDH cDNA probe.
RNAi of PAPST2 Gene-A sequence of small interfering RNA (siRNA) for each gene was designed as described previously (25). Twenty-five base pairs of stealth RNAs were designed and purchased from Invitrogen. The PAPST1-813 siRNA (initiated at the 813 nucleotide position from the start codon) has the sequence 5Ј-CCUCAUCU-UACUGGCAGGUUAUAUU-3Ј. The sequences of PAPST2-342 siRNA (initiated at the 342 nucleotide position from the start codon) and PAPST2-513 siRNA (initiated at the 513 nucleotide position from the start codon) are 5Ј-CCUUACCUUAGUGCAGUUUGCCUUU-3Ј and 5Ј-CCAAGUCAUCUUCAAGUGCUGCAAA-3Ј, respectively. As control siRNA, stealth RNAi negative control duplex (Invitrogen) was used. The HCT116 cells were subcultured onto 6-cm dishes at a con-centration of 1 ϫ 10 6 cells/dish in Dulbecco's modified Eagle's medium/ Ham's nutrient mixture F-12 (1:1) containing 10% fetal bovine serum 24 h prior to the transfection. The cells were transfected with 10 or 100 nM siRNA by using Lipofectamine 2000 reagent. RNA was extracted using TRIzol reagent (Invitrogen), and the first strand cDNA was synthesized using a Superscript II First Strand Synthesis kit (Invitrogen).
Metabolic Labeling of Colon Cancer Cell Line-Radiolabeling of sulfated residues in cell macromolecules was performed using procedures similar to those described previously (26). HCT116 cells were subcultured onto 10-cm dishes at a concentration of 1.5 or 2 ϫ 10 6 cells/dish in the inorganic sulfate-free Dulbecco's modified Eagle's medium/Ham's nutrient mixture F-12 (1:1) containing 10% fetal bovine serum and 100 Ci/ml of carrier-free [ 35

RESULTS
Molecular Cloning of Human PAPST2 cDNA-We identified a cDNA sequence (GenBank TM Accession number NM_015948) homologous to the putative nucleotide sugar transporter gene UGTrel1 by using the same procedure that was employed for the PAPST1 gene (3). Initially, it was described as CGI-19 protein in the GenBank TM data base. We named it PAPST2 and cloned the open reading frame as described under "Experimental Procedures." The phylogenetic tree of PAPS and the nucleotide sugar transporter genes in humans and Drosophila indicated that PAPST1, PAPST2, and UGTrel1 are classified in the same group (Fig. 1A). An alignment of the amino acid sequences of these genes is shown in Fig. 1B. PAPST2 comprised 401 amino acids with a calculated mass of 44.6 kDa. Hydrophobicity analyses of the amino acid sequences indicate that the PAPST2 protein is a type III transmembrane protein with nine transmembrane domains similar to that of PAPST1, although the transmembrane topology has yet to be verified experimentally.
PAPST2 showed 22.4 and 21.7% identity to UGTrel1 and PAPST1, respectively. The structural similarity suggested that PAPST2 is a PAPS transporter gene similar to PAPST1. There are eight potential N-glyco- PAPST2 Is a Golgi-localized Protein-PAPST1 is a membrane protein that is localized on the Golgi apparatus (3). Because PAPST2 shares similarity to PAPST1, we expected that PAPST2 would also be a Golgilocalized PAPS transporter.
First, we investigated the subcellular localization of the PAPST2 protein. Human colon cancer HCT116 cells were transiently transfected with a mammalian expression vector, pCXN2, that contained HA-tagged PAPST2 or HA-tagged PAPST1 gene and were double immunostained with anti-HA mAb and anti-␤1,4 galactosyltransferase 1 mAb. The results of immunofluorescence microscopy of the cells are shown in Fig. 2A. The HA-tagged PAPST1 protein was observed to be colocalized with ␤1,4 galactosyltransferase 1, which is a protein that is typically localized in the trans-Golgi (21), and this observation is consistent with a previous report (3). In this study, HA-tagged PAPST2 was also observed to be colocalized with ␤1,4 galactosyltransferase 1, thereby indicating trans-Golgi localization.
To confirm whether PAPST1 and PAPST2 proteins possess identical subcellular localization, an experiment was performed wherein HAtagged PAPST2 and c-Myc-tagged PAPST1 were expressed simultaneously in HCT116 cells. The cells were double immunostained with rhodamine-conjugated anti-HA mAb and FITC-conjugated anti-c-Myc mAb. As shown in Fig. 2B, PAPST2-HA was completely colocalized with PAPST1-c-Myc. These results indicate that the PAPST2 protein localizes on the Golgi apparatus like the PAPST1 protein.
Tissue Distribution of PAPST1 and PAPST2 Transcripts-Next, we analyzed the tissue distribution of PAPST2 transcripts using real-time PCR. All transcript levels are shown relative to that of GAPDH. As shown in Fig. 3A, PAPST2 is highly expressed in the colon, whereas the expression level of PAPST1 is low. PAPST2 transcripts are also widely expressed in other tissues when compared with that of PAPST1 (Fig.  3A). To confirm the expression of PAPST2 transcripts in the colon, we performed Northern blot analysis of RNAs that were derived from colon cancer cell lines HCT116 and DLD-1 and normal human colon tissue. The PAPST2 mRNA was detected as a single band corresponding to 2 kbp (Fig. 3B), which is consistent with the length of the identified sequence, namely, NM_015948 (2068 bp). We expected that PAPST2 would be a PAPS transporter gene that functions by compensating for the insufficient expression of PAPST1 during the synthesis of sulfated glycoconjugates in the colon.
PAPST2 Is a PAPS Transporter Gene-The substrate specificity of the PAPST2 protein was examined by yeast expression in a manner similar to that used for PAPST1 (3). The yeast expression vector YEp352GAP-II was inserted with HA-tagged PAPST2 and introduced into W303-1a yeast for the preparation of the Golgi-enriched P100 membrane fraction that expressed the PAPST2 protein. The HA-tagged PAPST2 protein was detected in the yeast P100 membrane fraction by Western blotting using an antibody against the HA epitope tag (Fig. 4A).  The substrate specificity of the PAPST2 protein was examined using the P100 membrane fraction and radiolabeled substrates. The transport activity of PAPS into the P100 membrane fraction is shown in Fig. 4B. The P100 membrane fraction prepared from yeast cells that expressed PAPST2 showed PAPS transport activity that was significantly higher than that observed in the mock cells (1.7 Ϯ 0.2 versus 2.6 Ϯ 0.3 pmol/mg of protein, respectively, mean Ϯ S.E. from six independent experiments; p Ͻ 0.05, Student's t-test), although the yeast Golgi-enriched fraction had relatively high endogenous PAPS transport activity. The substrate concentration dependence of PAPS transport by PAPST2 is shown in Fig. 4C. PAPST2 showed a saturated PAPS transport activity with an apparent K m value that was estimated to be 2.2 M.
Both PAPST1 and PAPST2 Transport PAPS in the Human Colon Cancer Cell Line-We also tested the PAPS transport activity of PAPST2 protein by using a mammalian cell line. For measuring the PAPS transport activity, we analyzed the Golgi-enriched fractions of colon cancer HCT116 cells that stably expressed HA-tagged PAPST1, HA-tagged PAPST2, or the mock vector (pCXN2 vector alone).
As shown in Fig. 5A, the levels of PAPST1 and PAPST2 transcripts in the PAPST1 and PAPST2 transfectant cells, respectively, showed a marked increase over those obtained from the mock transfectants. No difference was observed between the mock and PAPST1 transfectants with respect to PAPST2 transcript levels or between the mock and PAPST2 transfectants with respect to PAPST1 transcript levels (Fig.  5A). In each transfectant, HA-tagged PAPST1 or HA-tagged PAPST2 protein was detected in the Golgi-enriched fraction by Western blotting using an antibody against the HA epitope tag (Fig. 5B). The PAPS transport activity that was obtained from each Golgi-enriched fraction is shown in Fig. 5C. These Golgi-enriched fractions also showed relatively high endogenous PAPS transport activity; however, the value observed in the PAPST2 transfectants was significantly higher than that observed in the mock transfectants (1.9 Ϯ 0.2 versus 4.2 Ϯ 0.3 pmol/mg of protein, respectively, mean Ϯ S.E. from three independent experiments; p Ͻ 0.05, Student's t-test). These results indicate that both PAPST1 and PAPST2 act as PAPS transporters in the human colon cancer cell line.
PAPST2 Is Involved in the Synthesis of Sulfated Glycoconjugates in the Human Colon-We revealed that PAPST2 is a PAPS transporter that is expressed in the colon. To elucidate the role of PAPST2 in the synthesis of sulfated glycoconjugates in the colon, we performed a flow cytometric analysis of the colon cancer cell line by using an antibody, G72, recognizing sialyl 6-sulfo N-acetyllactosamine (sialyl 6-sulfo galactose ␤1,4 N-acetyl D-glucosamine: Sia ␣2,3 Gal ␤1,4 GlcNAc that is sulfated at the C-6 position of GlcNAc) (27). We attempted to reduce PAPST2 expression in the colon cancer cell line HCT116, which is moderately reactive to G72 antibody, by using RNAi. Based on their ability to suppress gene expression, two siRNA sequences that would target the PAPST1 and PAPST2 genes, i.e. PAPST1-813 and PAPST2-513, respectively, were selected. These siRNAs were synthesized as doublestranded stealth RNAs (Invitrogen) possessing 25 nucleotides, and the stealth RNAi negative control duplex (Invitrogen) was used as a control siRNA. Lipofection was used to transfect the HCT116 cells with 10 nM PAPST1-813 and PAPST2-513 siRNAs.  1 and 2) or HA-tagged PAPST2 (lanes 3 and 4), was subjected to SDS-polyacrylamide electrophoresis. Western blot analysis was performed using a monoclonal antibody against the HA epitope. Arrow indicates HA-tagged PAPST2. B, substrate specificity of PAPST2. Each P100 fraction (200 g of protein) was incubated in 100 l of reaction buffer containing 1 M substrate at 30°C for 5 min, and the incorporated radioactivity was measured. The indicated values are the mean Ϯ S.E. obtained from six independent experiments. Each experiment was performed with two transformant clones. Open bars, Mock; solid bars, PAPST2. C, substrate concentration dependence. Each P100 fraction (200 g of protein) was incubated in 100 l of reaction buffer containing different concentrations of [ 35 S]PAPS at 30°C for 5 min, and the incorporated radioactivity was measured. Specific incorporation was calculated by subtracting the value of the mock transfection from each of the corresponding values. Right panel shows the double-reciprocal plot used to determine the K m value.

DISCUSSION
We identified a novel human PAPS transporter that is predominantly expressed in the colon. Human colonic epithelial tissues express many sulfated glycoconjugates. Silencing of the PAPST2 gene by RNAi caused a reduction in the amount of sulfated structures in the colon cancer cell line. Expression of small amounts of PAPST1 mRNA in the colon suggests that PAPST2 plays a major role in the sulfation of glycoconjugates in this organ.
The PAPST2 gene is closely related to the PAPST1 gene, and the genomic locus of the PAPST2 gene is located close to that of PAPST1 (6p24.3 and 6p11.2-12.1, respectively). Despite their functional conservation, the genomic structures of PAPST2 and PAPST1 are not similar (eleven exons versus four exons, respectively). Furthermore, both PAPST1 and PAPST2 have a single ortholog in Drosophila, namely, sll and NM_140697, respectively (Fig. 1A). 3 These findings indicate that during the evolutionary process, the PAPST1 and PAPST2 genes diverged relatively early and thereafter evolved independently.
Recently, the nucleotide sugar transporter genes have been classified into the solute carrier family 35 (SLC35) (28). The PAPST1 and PAPST2 genes have also been classified into the SLC35 subfamily B. In Gen-Bank TM , UGTrel1, PAPST1, and PAPST2 have been defined as members of B1 (SLC35B1), B2 (SLC35B2), and B3 (SLC35B3), respectively. More recently, Ashikov et al. (29) reported that another member of subfamily B, namely, SLC35B4, is a UDP-xylose/UDP-GlcNAc transporter gene. Data bases mention that SLC35B4 is a human ortholog of yeast UDP-GlcNAc transporter genes, namely, Kluyveromyces lactis mnn2-2 (30) and S. cerevisiae yea4 (31). This implies that the subfamily B comprises members with entirely different substrate specificities despite their structural conservation.
Silencing of PAPST2 gene expression in the colon cancer cell line revealed the involvement of PAPST2 in the synthesis of a sulfated epitope, sialyl 6-sulfo galactose ␤1,4 N-acetyl D-glucosamine (Fig. 6). The fucosylated derivative, namely, sialyl 6-sulfo Le x , is a major L-selectin ligand present on the high endothelial venules of the human peripheral lymph nodes for the initial adhesion of leukocytes to endothelial cells (27,(32)(33)(34). It has been proposed that sialyl 6-sulfo Le x serves as a ligand for selectins in routine trafficking of leukocytes, whereas nonsulfated sialyl Le x plays a major role in the inflammatory response (35,36). PAPST2 may be involved in the function of these physiological routine processes, including the recruitment of gut-homing helper memory T lymphocytes, by regulating the sulfation of carbohydrate epitopes.
It is common knowledge that the risk of malignancy and recurrence of colorectal cancer is correlated with an increase in sialylation (13,15,16) and a decrease in the sulfation of carbohydrate epitopes (11)(12)(13)(14). The sialyl 6-sulfo Le x (14) and 3Ј-sulfo-Le a (13,17) are predominantly expressed in nonmalignant tissues but are not detected in the malignant tissues in colorectal cancers. Kumamoto et al. (37) reported that in colorectal cancers the expression of the UDP-Gal transporter increases significantly and results in the supply of the donor substrate for the synthesis of sialyl Le a and sialyl Le x . On the other hand, the colon cancer cell line HCT116 showed relatively low PAPST2 expression when compared with that of normal colon tissue. We could not find any colon cancer cell line that highly expressed PAPST2. The possibility that PAPST2 decreases the expression associated with the progression of colorectal cancer, as reported in the case of certain sulfotransferases (38 -40) and a sulfate transporter (41), may be considered.
The RNAi of PAPST2 gene in the HCT116 cells reduced total sulfate incorporation into cellular proteins. The sulfotransferases have long been believed to be the rate-limiting components of the sulfation process; however, sulfation is also controlled by the components involved in the earlier steps, such as the sulfate transporters and PAPS synthetases (42). For instance, PAPS synthetases modulate PAPS levels in high endothelial venules and thus play a key role in the control of the sulfation state of L-selectin ligands and its functional activities (43)(44)(45). Furthermore, mutations in some genes involved in PAPS synthesis, such as diastrophic dysplasia sulfate transporter (DTDST) (46 -49) and PAPS synthetase 2 (PAPSS2) (2,50), are responsible for a form of humaninherited osteochondrodysplasias that results because of down-regulated sulfation of chondroitin sulfate in the cartilage. This evidence implies that PAPS production is a rate-limiting step for sulfation in certain tissues. Our investigation also demonstrated that the PAPS transporter is a rate-limiting factor in the sulfation of glycoconjugates in the colon cells. The PAPST2 protein showed a relatively low apparent K m value for PAPS (2.2 M), although the precise kinetics of PAPST2 protein, excluding other factors, remains uncertain. The effect of the PAPS transporter would be significant if the K m value for PAPS would be lower than the K m value of the sulfotransferases in the Golgi lumen.
In the previous study, we reported that sll, the PAPST1 ortholog in Drosophila, is essential for viability of the fly (3). Lüders et al. (4) demonstrated that sll gene is involved in the signaling pathways and that the mutation of sll gene exhibits disrupted Wingless gradient and reduced Hedgehog signaling activity. Here, both PAPST1 and PAPST2 are shown to be Golgi-localized PAPS transporters. Most sulfation reactions of glycoconjugates, including proteoglycans (51), glycoproteins (52), and glycolipids (53), occur in the Golgi apparatus. Furthermore, tyrosine O-sulfation of the proteins occurs in the trans-Golgi network (54,55). In most human tissues, the expression of PAPST2 is comparable with that of PAPST1 (Fig. 3). The PAPST1 and PAPST2 proteins have comparable K m values for PAPS (0.8 versus 2.2 M, respectively). PAPST1 and PAPST2 may function in a manner that is complementary to each other. Further investigations are required to clarify the involvement and significance of PAPST1 and PAPST2 in these pathways.