Portal de Periódicos da CAPES

Substrate Recognition by Nucleotide Sugar Transporters

2003; Elsevier BV; Volume: 278; Issue: 25 Linguagem: Inglês

10.1074/jbc.m302620200

1083-351X

Kazuhisa Aoki, Nobuhiro Ishida, Masao Kawakita,

Polyamine Metabolism and Applications

Human UDP-Gal transporter 1 (hUGT1) and the human CMP-Sia transporter (hCST) are similar in structure, with amino acid sequences that are 43% identical, but they have quite distinct transport substrates. To define their substrate recognition regions, we constructed various chimeras between the two transporters and demonstrated that distinct submolecular regions of the transporter molecules are involved in the specific recognition of UDP-Gal and CMP-Sia (Aoki, K., Ishida, N., and Kawakita, M. (2001) J. Biol. Chem. 276, 21555–21561). In a further attempt to define the minimum submolecular regions required for the recognition of specific substrates, we found that substitution of helix 7 of hCST into the corresponding part of hUGT1 was necessary and sufficient for a chimera to show CST activity. Additional replacement of helix 2 or 3 of hUGT1 with the corresponding hCST sequence markedly increased the efficiency of CMP-Sia transport. For UGT activity, helices 1 and 8 of hUGT1 were necessary (but not sufficient), and helices 9 and 10 or helices 2, 3, and 7 derived from hUGT1 were also required to render the chimera competent for UDP-Gal transport. The in vitro analyses of a chimera with dual specificity indicated that it transported both UMP and CMP and mediated exchange reactions between these nucleotides and nucleotide sugars that are recognized specifically by either of the parental transporters. Human UDP-Gal transporter 1 (hUGT1) and the human CMP-Sia transporter (hCST) are similar in structure, with amino acid sequences that are 43% identical, but they have quite distinct transport substrates. To define their substrate recognition regions, we constructed various chimeras between the two transporters and demonstrated that distinct submolecular regions of the transporter molecules are involved in the specific recognition of UDP-Gal and CMP-Sia (Aoki, K., Ishida, N., and Kawakita, M. (2001) J. Biol. Chem. 276, 21555–21561). In a further attempt to define the minimum submolecular regions required for the recognition of specific substrates, we found that substitution of helix 7 of hCST into the corresponding part of hUGT1 was necessary and sufficient for a chimera to show CST activity. Additional replacement of helix 2 or 3 of hUGT1 with the corresponding hCST sequence markedly increased the efficiency of CMP-Sia transport. For UGT activity, helices 1 and 8 of hUGT1 were necessary (but not sufficient), and helices 9 and 10 or helices 2, 3, and 7 derived from hUGT1 were also required to render the chimera competent for UDP-Gal transport. The in vitro analyses of a chimera with dual specificity indicated that it transported both UMP and CMP and mediated exchange reactions between these nucleotides and nucleotide sugars that are recognized specifically by either of the parental transporters. Nucleotide sugar transporters are membrane proteins localized in the endoplasmic reticulum and Golgi apparatus. They play an indispensable role in constructing the sugar chains of glycoconjugates. The transporters carry nucleotide sugars into the endoplasmic reticulum and Golgi apparatus, in which they are used by specific transferases as precursors of sugar chains (for recent reviews, see Refs. 1Kawakita M. Ishida N. Miura N. Sun-Wada G.-H. Yoshioka S. J. Biochem. (Tokyo). 1998; 123: 777-785Crossref PubMed Scopus (54) Google Scholar, 2Hirschberg C.B. Robbins P.W. Abeijon C. Annu. Rev. Biochem. 1998; 67: 49-69Crossref PubMed Scopus (309) Google Scholar, 3Berninsone P.M. Hirschberg C.B. Curr. Opin. Struct. Biol. 2000; 10: 542-547Crossref PubMed Scopus (115) Google Scholar, 4Gerardy-Schahn R. Oelmann S. Bakker H. Biochimie (Paris). 2001; 83: 775-782Crossref PubMed Scopus (85) Google Scholar, 5Hirschberg C.B. J. Clin. Invest. 2001; 108: 3-6Crossref PubMed Scopus (58) Google Scholar). More than simply functioning as a passive entrance route of nucleotide sugars into the organelles, the transporters may regulate the amounts of nucleotide sugars available in the lumen of the endoplasmic reticulum or Golgi apparatus and consequently may affect the sugar chain composition of a cell (6Kumamoto K. Goto Y. Sekikawa K. Takenoshita S. Ishida N. Kawakita M. Kannagi R. Cancer Res. 2001; 61: 4620-4627PubMed Google Scholar). To better understand the roles of transporters in regulating sugar chains and, in turn, various aspects of cellular functions, it is indispensable to define which substrate(s) each transporter recognizes and to understand how and with what stringency the transporter discriminates between its own substrate(s) and irrelevant nucleotide sugars. The functions of nucleotide sugar transporters have long been studied by analyzing the properties of transporter-deficient mutant cells and by in vitro characterization of the transport activity using microsomal fractions from natural sources. In recent years, the genes encoding several kinds of nucleotide sugar transporters have been cloned from several species (Refs. 7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar, 8Roy S.K. Chiba Y. Takeuchi M. Jigami Y. J. Biol. Chem. 2000; 275: 13580-13587Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar, 9Baldwin T.C. Handford M.G. Yuseff M.I. Orellana A. Dupree P. Plant Cell. 2001; 13: 2283-2295Crossref PubMed Scopus (133) Google Scholar, 10Muraoka M. Kawakita M. Ishida N. FEBS Lett. 2001; 495: 87-93Crossref PubMed Scopus (106) Google Scholar, 11Berninsone P. Hwang H.Y. Zemtseva I. Horvitz H.R. Hirschberg C.B. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 3738-3743Crossref PubMed Scopus (115) Google Scholar, 12Segawa H. Kawakita M. Ishida N. Eur. J. Biochem. 2002; 269: 128-138Crossref PubMed Scopus (76) Google Scholar and references therein) and have provided us with useful tools for studying the molecular and functional features of the transporters. For instance, functionally important subregions of the transporter molecules were defined by several investigators through analyses of randomly and site-directedly produced transporter mutants (12Segawa H. Kawakita M. Ishida N. Eur. J. Biochem. 2002; 269: 128-138Crossref PubMed Scopus (76) Google Scholar, 13Eckhardt M. Gotza B. Gerardy-Schahn R. J. Biol. Chem. 1998; 273: 20189-20195Abstract Full Text Full Text PDF PubMed Scopus (93) Google Scholar, 14Abe M. Hashimoto H. Yoda K. FEBS Lett. 1999; 458: 309-312Crossref PubMed Scopus (38) Google Scholar, 15Ishida N. Yoshioka S. Iida M. Sudo K. Miura N. Aoki K. Kawakita M. J. Biochem. (Tokyo). 1999; 126: 1107-1117Crossref PubMed Scopus (35) Google Scholar, 16Gao X.-D. Dean N. J. Biol. Chem. 2000; 275: 17718-17727Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar, 17Gao X.-D. Nishikawa A. Dean N. J. Biol. Chem. 2001; 276: 4424-4432Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar, 18Lübke T. Marquardt T. Etzioni A. Hartmann E. von Figura K. Körner C. Nat. Genet. 2001; 28: 73-76Crossref PubMed Scopus (288) Google Scholar, 19Lühn K. Wild M.K. Eckhardt M. Gerardy-Schahn R. Vestweber D. Nat. Genet. 2001; 28: 69-72Crossref PubMed Scopus (288) Google Scholar, 20Oelmann S. Stanley P. Gerardy-Schahn R. J. Biol. Chem. 2001; 276: 26291-26300Abstract Full Text Full Text PDF PubMed Scopus (81) Google Scholar). As to the substrate specificity of nucleotide sugar transporters, it had long been assumed that each of the transporter family members was specialized for one nucleotide sugar and transported the particular nucleotide sugar into the lumen of the endoplasmic reticulum or Golgi apparatus in exchange for a corresponding nucleoside monophosphate that was concomitantly exported from the organelles. This simplified view has been challenged because Muraoka et al. (10Muraoka M. Kawakita M. Ishida N. FEBS Lett. 2001; 495: 87-93Crossref PubMed Scopus (106) Google Scholar, 21Goto S. Taniguchi M. Muraoka M. Toyoda H. Sado Y. Kawakita M. Hayashi S. Nat. Cell Biol. 2001; 3: 816-822Crossref PubMed Scopus (113) Google Scholar) recently identified human and Drosophila nucleotide sugar transporters that actively transport multiple UDP-sugars. Hong et al. (22Hong K. Ma D. Beverley S.M. Turco S.J. Biochemistry. 2000; 39: 2013-2022Crossref PubMed Scopus (83) Google Scholar) also demonstrated that the Leishmania GDP-mannose transporter is active with GDP-arabinose and GDP-fucose, but the molecular mechanisms underlying such multiple substrate recognition remain obscure. In an attempt to elucidate the mechanisms of substrate recognition by nucleotide sugar transporters, we have been analyzing the properties of chimeric transporters constructed as hybrids of UDP-Gal and CMP-Sia transporters (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar, 23Aoki K. Sun-Wada G.-H. Segawa H. Yoshioka S. Ishida N. Kawakita M. J. Biochem. (Tokyo). 1999; 126: 940-950Crossref PubMed Scopus (27) Google Scholar). Human UDP-Gal transporter 1 (hUGT1) 1The abbreviations used are: hUGT1, human UDP-Gal transporter 1; hCST, human CMP-Sia transporter; PNA, peanut agglutinin. (24Miura N. Ishida N. Hoshino M. Yamauchi M. Hara T. Ayusawa D. Kawakita M. J. Biochem. (Tokyo). 1996; 120: 236-241Crossref PubMed Scopus (120) Google Scholar) and the human CMP-Sia transporter (hCST) (25Ishida N. Miura N. Yoshioka S. Kawakita M. J. Biochem. (Tokyo). 1996; 120: 1074-1078Crossref PubMed Scopus (92) Google Scholar) are suitable for this study because they are similar membrane proteins with 43% amino acid identity and with 10 transmembrane helices (26Eckhardt M. Gotza B. Gerardy-Schahn R. J. Biol. Chem. 1999; 274: 8779-8787Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar) located in corresponding regions of the molecules, but have clearly distinct substrate specificity. In one of our previous studies (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar), we constructed some chimeric transporters that exhibit both UGT and CST activities and proposed that different regions of a transporter protein critically contribute to the recognition of UDP-Gal and CMP-Sia. Here, we continued our attempt to define more precisely the submolecular regions that are necessary for substrate recognition, and we were able to identify regions that are indispensable for specific recognition of UDP-Gal and CMP-Sia as well as groups of submolecular regions whose subsets in particular combinations, rather than all of these regions, are required for establishing or improving the recognition of a particular substrate. We also compared the kinetic as well as substrate recognition properties of chimeric and parental transporters and found that both UDP-Gal and CMP-Sia transporters are intrinsically competent to recognize both UMP and CMP as the antiport substrate. Materials—UDP-[6-3H]Gal (60 Ci/mmol), CMP-[6-3H]Sia (20 Ci/mmol), and [5-3H]CMP (20 Ci/mmol) were purchased from American Radiolabeled Chemicals (St. Louis, MO), and [5-3H]UMP (19 Ci/mmol) was purchased from Moravek Biochemicals, Inc. (Brea, CA). An anti-hUGT1 antibody and an anti-hCST antibody were prepared by immunizing rabbits with synthetic peptides representing the C-terminal amino acid sequences of each protein (377RGDLITEPFLPKSVLVK393 and 321TSIQQGETASKERVIGV337 for hUGT1 and hCST, respectively) and affinity-purified as described previously (27Yoshioka S. Sun-Wada G.-H. Ishida N. Kawakita M. J. Biochem. (Tokyo). 1997; 122: 691-695Crossref PubMed Scopus (42) Google Scholar, 28Ishida N. Ito M. Yoshioka S. Sun-Wada G.-H. Kawakita M. J. Biochem. (Tokyo). 1998; 124: 171-178Crossref PubMed Scopus (43) Google Scholar). An Alexa 546-conjugated goat anti-rabbit IgG antibody was purchased from Molecular Probes, Inc. (Eugene, OR). Construction of Chimeric cDNAs—The chimeras analyzed and discussed in this study as well as the ones used as intermediates in the construction of these chimeras are summarized in Table I.Table IChimeras used in this workChimeraChimeric fragmentshUGT1UGT-(1-393)hCSTCST-(1-337)B1UGT-(1-68)/CST-(46-337)E2UGT-(1-68)/CST-(46-237)/UGT-(263-393)E3UGT-(1-68)/CST-(46-194)/UGT-(220-393)G1UGT-(1-68)/CST-(46-120)/UGT-(145-219)/CST-(195-237)/UGT-(263-393)C(7)/UUGT-(1-219)/CST-(195-237)/UGT-(263-393)H1UGT-(1-113)/CST-(90-120)/UGT-(145-219)/CST-(195-237)/UGT-(263-393)H2UGT-(1-68)/CST-(46-64)/UGT-(88-219)/CST-(195-237)/UGT-(263-393)H1-aUGT-(1-113)/CST-(90-337)H2-aUGT-(1-68)/CST-(46-64)/UGT-(88-393)I1UGT-(1-68)/CST-(46-237)/UGT-(263-289)/CST-(265-337)I2UGT-(1-68)/CST-(46-237)/UGT-(263-320)/CST-(296-337)I3UGT-(1-68)/CST-(46-237)/UGT-(263-341)/CST-(317-337)I4UGT-(1-68)/CST-(46-194)/UGT-(220-289)/CST-(265-337)I5UGT-(1-144)/CST-(121-194)/UGT-(220-289)/CST-(265-337)I5-aUGT-(1-144)/CST-(121-194)/UGT-(220-393)I5-bCST-(1-194)/UGT-(220-393) Open table in a new tab Chimeras B1, E2, E3, G1, and C(7)/U (previously designated as G1-b) were constructed as described previously (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). Chimera H1 was constructed by ligation of the large fragment of chimera H1-a (see below and Table I) with the small fragment of chimera G1 obtained after digestion with BstEII and NotI. Chimera H2 was constructed by ligating the large fragment of chimera C(7)/U with the small fragment of chimera H2-a after digestion with EcoRI and PstI. Chimeras I1–I5 and other intermediate chimeras were constructed by the two-step PCR method described previously (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). Briefly, two fragments to be connected were amplified separately by the first PCR. The inside primers for each first PCR were 36 nucleotides long, and each consisted of an 18-nucleotide sequence located directly upstream and downstream, respectively, from the junctions of the intended chimera. Thus, the two fragments amplified in the first PCR had overlapping sequences around the junction. The N-terminal outside primers contained an EcoRI site as well as a hemagglutinin tag sequence, and the C-terminal outside primers possessed a NotI site. The second PCR was performed using the two fragments amplified by the first PCR as templates and the two outside primers as primers. The templates used in the first PCR to construct the N- and C-terminal fragments, respectively, of each chimera were as follows: for I1–I3, E2 and hCST; for I4, E3 and hCST; for I5, I5-a and hCST; for I5-a, I5-b and hUGT1; for I5-b, hUGT1 and hCST; for H1-a, hUGT1 and hCST; and for H2-a, B1 and hUGT1. All chimeric cDNAs constructed by this method were sequenced to rule out PCR-induced mutations. Cell Cultures and Transfection—Chinese hamster ovary cell strain K1 and its derivatives Lec2 (a CST-deficient mutant) (29Deutscher S.L. Nuwayhid N. Stanley P. Briles E.I. Hirschberg C.B. Cell. 1984; 39: 295-299Abstract Full Text PDF PubMed Scopus (198) Google Scholar) and Lec8 (a UGT-deficient mutant) (30Deutscher S.L. Hirschberg C.B. J. Biol. Chem. 1986; 261: 96-100Abstract Full Text PDF PubMed Google Scholar) were cultured in α-minimal essential medium supplemented with 10% fetal calf serum (31Stanley P. Siminovitch L. Somatic Cell Genet. 1977; 3: 391-405Crossref PubMed Scopus (124) Google Scholar, 32Briles E.B. Li E. Kornfeld S. J. Biol. Chem. 1977; 252: 1107-1116Abstract Full Text PDF PubMed Google Scholar). Transfections for both transient expression and stable expression were performed with LipofectAMINE reagent (Invitrogen) following the manufacturer's instructions. Stable transformants were selected in 0.5 mg/ml G418 for 10 days and cloned by limiting dilution. Assessment of the UGT and CST Activities—UGT and CST activities of chimeric molecules were assessed as described previously (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). To examine the UGT activity, UGT-deficient Lec8 cells were transfected with a chimeric cDNA, cultured for 1 day, transferred onto a chamber slide (Nalge Nunc International Corp., Rochester, NY), incubated overnight, and then fixed with methanol at –20 °C for 6 min. The fixed cells were stained with 20 μg/ml fluorescein isothiocyanate-conjugated Griffonia simplicifolia lectin II (EY Laboratories, Inc., San Mateo, CA) for 30 min and washed twice with phosphate-buffered saline for 10 min. Subsequently, they were incubated with an appropriate primary antibody in 3% bovine serum albumin and phosphate-buffered saline for 1 h, washed twice with phosphate-buffered saline for 10 min, and then incubated with a fluorescent secondary antibody for 1 h. The cells were washed twice with phosphate-buffered saline and once with water and mounted with Permafluor (Immunotech, Marseilles, France). Fluorescence labeling was visualized under a Carl Zeiss LSM510 laser-scanning confocal microscope. The CST activity of chimeras was assessed by a similar method in which CST-deficient Lec2 cells and fluorescein isothiocyanate-conjugated peanut agglutinin (PNA) (EY Laboratories, Inc.) were used instead of Lec8 cells and G. simplicifolia lectin II, respectively. Subcellular Fractionation of Yeast—Saccharomyces cerevisiae strain YPH500 (MATα ura3-52 lys2-801 ade2-101 trp1-Δ63 his3-Δ200 leu2-Δ1) was transformed by a lithium acetate procedure (33Ito H. Fukuda Y. Murata K. Kimura A. J. Bacteriol. 1983; 153: 163-168Crossref PubMed Google Scholar). Subcellular fractionation was performed as described previously (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar) with slight modifications. The cells were cultured to a density of A600 = 0.4, and cupric sulfate was added to the medium at a final concentration of 2 mm at 3 h before harvesting. The cells were harvested, washed twice with ice-cold 10 mm NaN3, resuspended in spheroplast solution (1.4 m sorbitol, 50 mm potassium phosphate (pH 7.5), 10 mm NaN3, and 40 mm 2-mercaptoethanol) containing 2 mg of zymolyase 100T (Seikagaku Corp., Tokyo, Japan) per g of packed cells, and incubated at 37 °C for 40 min. The spheroplasts were collected by centrifugation at 1000 × g for 5 min, resuspended in 4 volumes of lysis buffer (0.8 m sorbitol, 10 mm HEPES/Tris (pH 7.4), and 1 mm EDTA) containing a protease inhibitor mixture (Roche Applied Science, Mannheim, Germany), and disrupted by passing the suspension three times through a thin pipette tip. The homogenate was centrifuged at 1500 × g for 10 min to remove unlysed cells and debris. The membrane fraction was collected by centrifugation at 10,000 × g for 10 min and resuspended in lysis buffer. The protein concentrations were determined with BCA reagent (Pierce). In Vitro Assessment of Transport Activity—The in vitro transport assay was performed essentially as described previously (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar) with slight modifications. The reaction was started by addition of the membrane preparation (50 μg of protein) to the reaction mixture (0.8 m sorbitol, 10 mm Tris-HCl (pH 7.0), 1 mm MgCl2, 0.5 mm dimercaptopropanol, and 3H-labeled nucleotide sugar or nucleotide (0.64 μCi/reaction)). The reaction mixture (100 μl) was incubated at 30 °C, diluted with 1 ml of ice-cold stop buffer (0.8 m sorbitol, 10 mm Tris-HCl (pH 7.0), and 1 mm MgCl2) to stop the reaction, and poured onto a nitrocellulose filter (0.45 μm; Millipore Corp., Bedford, MA). The filter was washed three times with stop buffer and dried, and the radioactivity trapped on the filter was measured in a toluene-based scintillator. Minimum Structural Requirements for CST Activity—In one of our previous studies, we demonstrated that chimeric transporters E2 and G1 are able to transport both UDP-Gal and CMP-Sia (Fig. 1) (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). This indicates that helices 2, 3, and 7 of hCST in the hUGT1 context are sufficient for a chimera to exhibit CST activity. To investigate whether any of these helices are nonessential for the CST activity, we constructed a number of diagnostic chimeras and examined their transport activity. The chimeras discussed in this work are summarized in Fig. 1. To evaluate the UGT and CST activities, a chimeric cDNA was transiently expressed in Lec8 or Lec2 cells, and the cells were examined to determine whether their characteristic mutant phenotype, viz. binding of G. simplicifolia lectin II or PNA lectin on their surface, was abolished as a result of the restoration of nucleotide sugar transport (Fig. 2; also summarized in Fig. 1). Fig. 2A shows that chimeras H1 and H2 exhibited CST activity. This indicates that either one of two hCST-derived helices, helix 2 or 3, may be dispensable and substitutable with its hUGT1 counterpart provided that the other is available. Upon further analysis, it was finally shown that chimera C(7)/U, in which only helix 7 is derived from hCST, was competent for CMP-Sia transport as judged by the complementation of the CST-deficient phenotype of Lec2 cells. This is shown convincingly in Fig. 2B in an experiment using stable transformants rather than transient transformants. Fig. 2B shows the PNA binding property and the expression of the chimeric protein in a cell line stably expressing C(7)/U protein. The cells bound virtually no PNA on their surface. Therefore, the hCST sequence in helix 7 is necessary and qualitatively sufficient for eliciting CST activity. However, it should be noted that the CST activity of C(7)/U was quantitatively lower than that of chimeras H1 and H2 as judged by their effectiveness in complementing the Lec2 phenotype. The presence of hCST-derived helix 2 or 3 together with helix 7 of hCST rendered the transporters able to exhibit potent CST activity. We reported previously that a chimera with hCST-derived helices 4–7in the hUGT1 context is unable to transport CMP-Sia (chimera F1 in Ref. 7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar); but subsequently, stable transformant clones expressing the chimera were isolated, and the clones were shown to be able to correct the genetic defect of Lec2 cells (data not shown). It should be noted that helices 4–6 do not increase the ability of a chimera to transport CMP-Sia as do helices 2 and 3.Fig. 2Analyses of UGT and CST activities of chimeric transporters. A, Lec8 and Lec2 cells were transfected with the indicated chimeric cDNAs to assess their UGT and CST activities, respectively. Two days after transfection, the cells were double-stained with fluorescein isothiocyanate-conjugated lectins (G. simplicifolia lectin II (GS-II) for Lec8 cells and PNA for Lec2 cells; green) and antibodies (an anti-hUGT1 antibody for chimeras H1, H2, and C(7)/U and an anti-hCST antibody for chimeras I1–I5). The binding of the antibodies was detected with an Alexa 546-conjugated anti-rabbit IgG antibody (red). The two fluorescence images and a Nomarski image of the same field are merged. Bars = 20 μm. The expected terminal oligosaccharide structures on cells with a given phenotype and their reactivity for lectins are schematically shown below the images. B, Lec2 cells were transfected with chimeric C(7)/U cDNA, and stable transformants were cloned. The cells of a representative clone were stained with PNA and the anti-hUGT1 antibody as described for A to assess the expression of the chimeric transporter and its CST activity. Bar = 20 μm.View Large Image Figure ViewerDownload (PPT) Minimum Structural Requirements for UGT Activity—It was shown previously that chimeras containing either hUGT1-derived helices 1, 8, 9, and 10 or hUGT1-derived helices 1–8 exhibit UGT activity, but those with hCST-derived sequences in either helix 1 or 8 do not (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). This indicates the importance of these two helices, but it remains to be examined whether the presence of hUGT1-derived helices 1 and 8 is sufficient for recognizing UDP-Gal as a substrate. Chimeras I1–I5 were thus constructed to define the submolecular regions contributing to the recognition of UDP-Gal. Chimera I1 had only helices 1 and 8 with hUGT1 sequences in the hCST background, whereas chimeras I2–I5 had additional hUGT1-derived helices together with these two helices. As shown in Figs. 1 and 2, chimera I1 could not transport UDP-Gal. Helices 1 and 8 of hUGT1 are therefore necessary (but not sufficient) for chimeras to elicit UGT activity. Chimera I2 did not express UGT activity, but chimera I3 did, indicating that the C-terminal loop region is not required for the UGT activity, but that the region from helices 8 to 10 is likely to be. However, the necessity for helix 9 has not yet been conclusively proven. It should be noted that the requirement for helix 10 (and probably helix 9) is not absolute because chimera I5 manifested UGT activity. This indicates that helices 2, 3, and 7 of hUGT1 could substitute for helices 9 and 10 to make a chimera competent for UDP-Gal transport. hUGT1-derived helices 2 and 3 were necessary for the UGT activity in this context because chimera I4 was unable to transport UDP-Gal. Substrate Specificity of hUGT1, hCST, and UDP-Gal/CMP-Sia-transporting Chimera G1—To examine the biochemical properties of hUGT1, hCST and their hybrids, the transporters were expressed heterologously in the budding yeast S. cerevisiae, and the transport of various substrates was studied directly using the microsomal vesicles obtained from the yeast cells expressing these transporters. The dependence of the transport activity of hUGT1, hCST, and chimera G1 on substrate concentration is illustrated in Fig. 3. The Km value of hUGT1 for UDP-Gal and that of hCST for CMP-Sia were 5.5 and 60 μm, respectively. Chimera G1 showed almost the same affinity as hCST for CMP-Sia, with a Km value of 45 μm, whereas its affinity for UDP-Gal was much lower than that of hUGT1, which was 180 μm. The kcat values of the transport reactions cannot be readily compared because the amounts of transporters expressed in the microsomes are not the same and do not allow direct comparison. It is believed that UGT exports UMP from the Golgi apparatus in exchange for UDP-Gal, whereas CST exchanges CMP for CMP-Sia. To clarify which nucleoside monophosphate the chimeric transporter G1 can transport in exchange for UDP-Gal and CMP-Sia, we examined the transport of UMP and CMP by chimera G1 using microsomal vesicles obtained from yeast cells. Fig. 4 shows that chimera G1 transported UMP in an amount comparable to hUGT1. This implied that the chimera was able to mediate the exchange between UMP and either of its nucleotide sugar substrates. We also tried to determine whether the yeast microsomal vesicles showed CMP uptake dependent on heterologously expressed nucleotide sugar transporters and their chimeras, but this attempt was unsuccessful because negative control vesicles prepared from the yeast cells transformed with the vector alone incorporated almost the same amount of CMP as those from yeast cells expressing hUGT1, hCST, or chimera G1 (data not shown). Because of this endogenous CMP transport activity in yeast microsomes, we were unable to directly assess the ability of chimera G1 and its parental transporters to transport CMP. To gain further information concerning the substrate specificity of the nucleotide sugar-nucleoside monophosphate antiport systems, we carried out "exchange assays." In this type of assay, a radiolabeled substrate was first incorporated into the microsomal vesicles at 30 °C for 1 min, and then another nonradioactive substance at 100 times higher concentration was added to the reaction mixture. If exchanges between the first and second substrates took place, radioactive substrate that had been incorporated in the vesicles during the first step of incubation would be released during the second step of incubation. Fig. 5 (A–C) shows that the transporters carried out exchanges not only between nucleotide sugars and nucleoside monophosphates, but also between nucleotide sugars. The vesicles from the yeast cells expressing hUGT1 readily released radioactive UDP-Gal when UMP, CMP, UDP-Gal, or UDP-GalNAc was added to the external medium as the second substrate (Fig. 5A). A moderate release of UDP-Gal was also observed upon addition of AMP, UDP-GlcNAc, or UDP-Glc. The exchange properties of chimera G1 could not be determined probably because the specific incorporation of UDP-Gal at 10 μm was very low due to its low affinity for UDP-Gal (data not shown). Fig. 5 (B and C) shows the exchanges between CMP-Sia and various nucleotides or nucleotide sugars by hCST and chimera G1. CMP-Sia was released from vesicles expressing hCST in exchange for AMP, UMP, CMP, or CMP-Sia, but not for other nucleotide sugars, whereas chimera G1 mediated the exchange between CMP-Sia and nucleotide sugars recognized by hUGT1, including UDP-Gal, UDP-GalNAc, as well as CMP-Sia, the nucleotide sugar recognized by hCST. It is worth pointing out that these constitute the sum of the substances that serve as substrates for hUGT1 and hCST. Both hUGT1 and hCST showed a qualitatively similar specificity for nucleoside monophosphates. It was rather unexpected that CMP promoted UDP-Gal release from hUGT1-expressing vesicles and that UMP caused CMP-Sia release from hCST-expressing vesicles. Even a purine nucleotide, AMP, significantly promoted the release of nucleotide sugars, although GMP was completely inert in this respect. We reported above that hCST-expressing vesicles did not accumulate UMP significantly when incubated with 10 μm [3H]UMP. The affinity of hCST for UMP may be lower than that of hUGT1, so the movement of UMP across the microsomal membranes was observable only at higher concentrations of UMP. As shown in Fig. 5 (D and E), the exchange of UMP with various nucleoside monophosphates and nucleotide sugars was assayed using vesicles prepared from yeast cells transformed with hUGT1 and chimera G1. Fig. 5D shows that AMP, UMP, CMP, UDP-Gal, and UDP-GalNAc induced the release of UMP from hUGT1-expressing vesicles. In addition to these substances, CMP-Sia was also active in releasing UMP from the chimera G1-expressing vesicles. These results indicate that chimera G1 acquired the combined repertoires of hUGT1 and hCST and transported UMP and CMP as well as UDP-Gal, UDP-GalNAc, and CMP-Sia. They also indicate that these substrates could be exchanged in any combination between the two compartments across the membrane. In our previous studies, we constructed chimeric transporters between hUGT1 and hCST to elucidate the mechanisms underlying the substrate recognition by nucleotide sugar transporters and found that some chimeras exhibit both UGT and CST activities (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar, 23Aoki K. Sun-Wada G.-H. Segawa H. Yoshioka S. Ishida N. Kawakita M. J. Biochem. (Tokyo). 1999; 126: 940-950Crossref PubMed Scopus (27) Google Scholar). This indicate that UDP-Gal and CMP-Sia are critically recognized at distinct submolecular regions in a transporter protein. In the present work, we further examined the properties of these and other chimeras to define the requirements for dual substrate recognition more precisely and to determine their substrate spectra in more detail. We have shown that a helix 7-containing segment of hCST alone was sufficient in an otherwise hUGT1 context to confer the ability to recognize CMP-Sia on the chimera C(7)/U. So far, every chimera with an hCST-derived helix 7-containing region was found to be competent in transporting CMP-Sia. Although the presence of the hCST stretch containing helix 7 was qualitatively sufficient for eliciting the CST activity, it should be noted that additional substitution of either helix 2 or 3 of the C(7)/U chimera by the hCST counterpart markedly enhanced the efficiency of CMP-Sia transport, as inferred from the fact that chimeras H1 and H2 were much more effective in correcting the genetic defect of Lec2 cells. This indicates that multiple submolecular regions contribute to recognizing appropriate substrates properly and efficiently, some of which are indispensable or particularly important for the occurrence of specific recognition, whereas others are involved in increasing the strength of the specific interaction. The fact that chimera C(7)/U recognized CMP-Sia implies that the helix 7 region of the hCST sequence contains the CMP-Sia-recognizing motif. We are currently trying to define in detail the CMP-Sia-recognizing motif by site-directed mutagenesis. The requirements for specific recognition of UDP-Gal are more complicated because the recognition process appears to involve multiple submolecular regions. We showed previously that helices 1 and 8 of hUGT1 are indispensable for UGT activity (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). The present study demonstrated that these helices alone in the hCST context are not sufficient for a chimera to recognize UDP-Gal. Either a region containing helices 2, 3, and 7 or one containing helices 9 and 10 together with helices 1 and 8 is required to confer the UDP-Gal transport activity on a chimeric transporter. The motifs involved in the recognition of UDP-Gal are likely dispersed spatially in the transporter molecule, and the nucleotide sugar is recognized as a result of interactions at multiple points in the substrate-binding site. Some of these contact points, including those on helices 1 and 8, are critically important, whereas either the subset on helices 2, 3, and 7 or the one on helices 9 and 10 may be dispensable provided the other subset is available. The affinities of chimera G1, as a representative of the dual substrate-specific chimeras, for its substrates were compared with those of the parental transporters, hUGT1 and hCST, using microsomal membrane vesicles prepared from S. cerevisiae cells expressing these heterologous transporters. The Km value of hUGT1 was in the range reported before, but that of hCST determined here was considerably higher than those reported previously for reasons that are unclear at the moment (12Segawa H. Kawakita M. Ishida N. Eur. J. Biochem. 2002; 269: 128-138Crossref PubMed Scopus (76) Google Scholar, 29Deutscher S.L. Nuwayhid N. Stanley P. Briles E.I. Hirschberg C.B. Cell. 1984; 39: 295-299Abstract Full Text PDF PubMed Scopus (198) Google Scholar, 30Deutscher S.L. Hirschberg C.B. J. Biol. Chem. 1986; 261: 96-100Abstract Full Text PDF PubMed Google Scholar, 34Carey D.J. Sommers L.W. Hirschberg C.B. Cell. 1980; 19: 597-605Abstract Full Text PDF PubMed Scopus (59) Google Scholar, 35Milla M.E. Hirschberg C.B. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 1786-1790Crossref PubMed Scopus (39) Google Scholar, 36Milla M.E. Clairmont C.A. Hirschberg C.B. J. Biol. Chem. 1992; 267: 103-107Abstract Full Text PDF PubMed Google Scholar, 37Berninsone P. Eckhardt M. Gerardy-Schahn R. Hirschberg C.B. J. Biol. Chem. 1997; 272: 12616-12619Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar, 38Sun-Wada G.-H. Yoshioka S. Ishida N. Kawakita M. J. Biochem. (Tokyo). 1998; 123: 912-917Crossref PubMed Scopus (39) Google Scholar, 39Tiralongo J. Abo S. Danylec B. Gerardy-Schahn R. von Itzstein M. Anal. Biochem. 2000; 285: 21-32Crossref PubMed Scopus (11) Google Scholar, 40Chiaramonte M. Koviach J.L. Moore C. Iyer V.V. Wagner C.R. Halcomb R.L. Miller W. Melançon P. Kuchta R.D. Biochemistry. 2001; 40: 14260-14267Crossref PubMed Scopus (17) Google Scholar). The Km value of chimera G1 for UDP-Gal (180 μm) was much higher than that of hUGT1, whereas its affinity for CMP-Sia was almost the same as that of hCST. Substitution of helices 2, 3, and 7 of hUGT1 thus resulted in a significant reduction of affinity for UDP-Gal. This is quite reasonable because these helices constitute a subset of hUGT1 helices that participate together in recognizing UDP-Gal. We showed previously that chimera G1 transports both UDP-Gal and CMP-Sia, but the activity for UDP-Gal under the standard conditions (at 1 μm substrate) is much lower than that of the parental transporter hUGT1 (7Aoki K. Ishida N. Kawakita M. J. Biol. Chem. 2001; 276: 21555-21561Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). The present results suggest that this difference in UDP-Gal transport activity may be largely explained in terms of the difference in affinity for UDP-Gal, although the kcat values of the transport reaction cannot be readily compared because the amounts of transporters expressed in the microsomes are not the same and because we have no reliable measure of the amount of the transporter protein. Gao et al. (17Gao X.-D. Nishikawa A. Dean N. J. Biol. Chem. 2001; 276: 4424-4432Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar) recently reported mutations that affect the affinity of the yeast GDP-Man transporter for its substrate. Although chimera G1 was not as active as the parental transporters, in particular hUGT1, it was able to work in cells efficiently enough to complement the defects of both Lec2 and Lec8 cells. It should be noted here that intracellular nucleotide sugar concentrations are estimated to be 60–200 μm for UDP-Gal and 70–500 μm for CMP-Sia (40Chiaramonte M. Koviach J.L. Moore C. Iyer V.V. Wagner C.R. Halcomb R.L. Miller W. Melançon P. Kuchta R.D. Biochemistry. 2001; 40: 14260-14267Crossref PubMed Scopus (17) Google Scholar, 41Tomiya N. Ailor E. Lawrence S.M. Betenbaugh M.J. Lee Y.C. Anal. Biochem. 2001; 293: 129-137Crossref PubMed Scopus (173) Google Scholar). In view of these values, the Km value of chimera G1 for UDP-Gal may not be too high for the transporter to work with reasonable efficiency in cells. Decreased affinity of a chimeric transporter for CMP-Sia would similarly be tolerated by Lec2 cells in complementation for the defective glycosylation system of the cells. It was unexpected that UDP-Gal and CMP-Sia were exported via hUGT1 and hCST, respectively, in exchange for both UMP and CMP. hUGT1 and hCST may not strictly discriminate between UMP and CMP, although the affinity of hCST for UMP may not be very high because direct measurement of transport did not detect accumulation of UMP in CST-expressing membrane vesicles. Chiaramonte et al. (40Chiaramonte M. Koviach J.L. Moore C. Iyer V.V. Wagner C.R. Halcomb R.L. Miller W. Melançon P. Kuchta R.D. Biochemistry. 2001; 40: 14260-14267Crossref PubMed Scopus (17) Google Scholar) recently examined the substrate specificity of CST in detail and reported that CST can transport UMP, although less efficiently than CMP. We reported recently that UGT is not as specific for UDP-Gal as had been widely believed and is also able to transport UDP-GalNAc across the membranes (12Segawa H. Kawakita M. Ishida N. Eur. J. Biochem. 2002; 269: 128-138Crossref PubMed Scopus (76) Google Scholar). The present "substrate exchange" assay further revealed that UDP-GlcNAc and perhaps UDP-Glc added to the external aqueous phase at 1 mm were exchanged with UDP-Gal as well as UMP trapped inside the membrane vesicles. Because these nucleotide sugars are not transported by hUGT1 in a direct uptake assay carried out at lower substrate concentrations (1 μm), their affinity for hUGT1 must be significantly lower than that of UDP-Gal. The physiological significance of these low affinity substrates remains to be elucidated, but it may not be easily dismissed in view of the fact that the cytoplasmic concentration of nucleotide sugars is of the order of 10–4m. In contrast to hUGT1, the specificity of hCST for nucleotide sugar substrates was quite strict, and hCST did not use nucleotide sugars other than CMP-Sia even at the higher concentrations at which the substrate exchange assays were carried out. The molecular basis for this strict specificity as opposed to the rather broad specificity of hUGT1 is intriguing, but quite obscure at present. Further analysis of the significance of the CMP-Sia specificity-conferring stretch (40 amino acid residues containing helix 7), which was identified in this study, would provide valuable information on this issue.