OF BIOLOGICAL Vol. 34, Issue of 5, pp. Q by The and Inc. U ... · TI11 and 89% identity with Fuc-TV...

THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 267, No. 34, Issue of December 5, pp. 24575-24584,1992 Printed in U.S.A.

Molecular Cloning of a Fourth Member of a Human a( 1,3)Fucosyltransferase Gene Family MULTIPLE HOMOLOGOUS SEQUENCES THAT DETERMINE EXPRESSION OF THE LEWIS x, SIALYL LEWIS x, AND DIFUCOSYL SIALYL LEWIS x EPITOPES*

(Received for publication, August 21, 1992)

Brent W. Weston$, Peter L. Smith§, Robert J. Kelly§, and John B. LoweNII From the $Department of Pediatrics, Division of Pediatric Hematology/Omology, the §Howard Hughes Medical Institute, and the TDepartment of Pathlogy, University of Michigan Medical School, Ann Arbor, Michigan 48109-0650

We and others have previously described the isolation of three human a(1,3)fucosyltransferase genes which form the basis of a nascent glycosyltransferase gene family. We now report the molecular cloning and expression of a fourth homologous human a(l,3)fucosyltransferase gene. When transfected into mammalian cells, this fucosyltransferase gene is capable of direct- ing expression of the Lewis x (GalBl+4[Fuca1+3] GlcNAc), sialyl Lewis x (NeuNAca2+3Ga431-4 [Fuca143]GlcNAc), and difucosyl sialyl Lewis x (NeuNAca2+3Gal@1+4[Fucal~3]GlcNAc@l+ 3Ga~1+4[Fuca1+3]GlcNAc) epitopes. The enzyme shares 86% amino acid sequence identity with Fuc- TI11 and 89% identity with Fuc-TV but differs substantially in its acceptor substrate requirements. Po- lymerase chain reaction analyses demonstrate that the gene is syntenic to Fuc-TI11 and Fuc-TV on chromosome 19. Southern blot analyses of human genomic DNA demonstrate that these four a(1,3)fucosyltransferase genes account for all DNA sequences that cross-hybridize at low stringency with the Fuc-TI11 catalytic domain. Using similar methods, a catalytic domain probe from Fuc-TIV identifies a new class of DNA fragments which do not cross-hybridize with the chromosome 19 fucosyltransferase probes. These results extend the molecular definition of a family of human a( 1,3)fucosyltransferase genes and provide tools for examining fucosyltransferase gene expression.

Many membrane-associated glycoproteins and glycolipids contain fucose linked to subterminal and/or internal N-acetylglucosamine residues. Specific sets of fucosylated glycans appear in certain mammalian tissues and cells at defined periods of development (1-7). Some of these fucosylated carbohydrate components are postulated to function in cell surface interactions during embryogenesis (8-10). Most of these

* This work was supported in part by the Howard Hughes Medical Institute and by National Institutes of Health Grants GM47455 (to J. B. L.) and GM14279 (to B. W. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) LO1 698.

)I Associate Investigator of the Howard Hughes Medical Institute. To whom correspondence should be addressed Howard Hughes Med- ical Inst., Medical Science Research Building I, Rm. 3510, 1150 West Medical Center Dr., Ann Arbor, MI 48109-0650. Tel.: 313-747-4779.

molecules are expressed in precise patterns during embryogenesis and malignant transformation (11-17), but their functions remain largely unknown.

A subset of a(l,3)-fucosylated cell surface carbohydrate molecules serve as oligosaccharide ligands for two members of the selectin family of cell adhesion receptors (E-selectin and P-selectin, reviewed in Refs. 18-22). The carbohydrate recognition domain of each of these selectins mediates cell adhesion through calcium-dependent interactions with surface-localized a(2,3)-sialylated, a(l,3)-fucosylated lactosaminoglycans represented by the sialyl Lewis x1 determinant and related monofucosylated or difucosylated moieties (23-30). a(2,3)-Sialylated a(l,4)-fucosylated glycans also function as ligands for E-selectin (31, 32). Internally fucosylated sialylated glycans may also function in cell surface adhesive interactions (10, 27). Expression of these oligosaccharide ligands is controlled in part by regulating expression of a(1,3)fucosyltransferases that participate in their biosynthesis (33, 34).

At least four human a(l,3)fucosyltransferase activities capable of generating sialyl Lewis x and related structures are discernible based on substrate specificities, cation requirements, and sensitivity to N-ethylmaleimide inhibition (35- 42). These a(l,3)fucosyltransferase “species” or “subtypes” (10) exhibit different tissue-specific expression patterns, some of which are determined by loci other than the Lewis locus (35, 37, 41, 43). With these types of analyses, however, it has not been possible to precisely define which enzyme(s) are expressed in different tissues or cells, since activity assays of homogenates cannot distinguish between single or multiple fucosyltransferase activities. Consequently, it has not been possible to define with precision the roles of specific a(l,3)fucosyltransferases in embryogenesis, oncogenesis, and in biologic processes operating in fully differentiated tissues. Cloned DNA segments derived from fucosyltransferase genes represent tools to address these questions at the molecular level.

Previously reported cloned fucosyltransferases include a cDNA encoding a human a(l,3/1,4)fucosyltransferase (Fuc- TIII, product of the human Lewis blood group locus, Ref. 44), a homologous a( 1,3)fucosyltransferase gene corresponding to

‘The abbreviations used are: Lewis x, Gal@l-+4[Fucnl+3] GlcNAc; Lewis a, Gal@1->3[Fucal->4]GlcNAc; sialyl Lewis x, NeuNAca2+3Galpl4[Fucal+3]GlcNAc; sialyl Lewis a, NeuNAca2+3Gal~l+3[Fucal4]GlcNAc; VIM-2, NeuNAcu2+ 3Gal@l4GlcNAcpl+3Galp14[Fuc~1+3]GlcNAc; difucosyl sialyl Lewis x, NeuNAca2+3Gal@l“4[Fucal+3]GlcNAc@l+3GaliJ1+ 4[Fucal+3]GlcNAc; a(l,3)fucosyltransferase, GDP-fucose:iJ-D-N- acetylglucosaminide 3-a-~-fucosyltransferase; PCR, polymerase chain reaction; kb, kilobase; bp, base pair; CHO, Chinese hamster ovary; HPLC, high performance liquid chromatography.

24575

24576 Molecular Cloning of a Fourth Human Fucosyltransferase Gene

a myeloid-expressed enzyme (Fuc-TIV) (23,26,45,46), and a third ~ ( l , 3 ) ~ c o s y l t r a n s f e r a s e gene whose cognate enzyme (Fuc-TV) exhibits properties distinct from those of Fuc-TI11 and Fuc-TIV (47). We now report the cloning and expression of a fourth distinct human a(l,3)fucosyltransferase gene, whose predicted amino acid sequence is very similar to Fuc- TI11 and Fuc-TV and s u b s ~ t i a l l y different from Fuc-TIV. Disparate acceptor substrate requirements characterize the enzymes encoded by these four genes. These homologous DNA sequences define a glycosyltransferase gene family, serve as templates for s t ~ c t u r e / ~ c t i o n anaIyses, and represent tools to study a(l,3)fucosyltransferase expression in human cells.

EXPERIMENTAL PROCEDURES

Nomenclnture-The gene described here represents the fourth distinct human ff(l,3)fucosyltransferase DNA sequence and will be called Fuc-TVI. We previously designated the human Lewis blood group a(l,3/1,4)fucosyltransferase Fuc-TI11 and a myeloid-expressed a(l,3)fucosyltransferase Fuc-TIV (47). The Fuc-TV gene encodes an a( 1,3)fucosyltransferase with properties similar, although not identical, to those ascribed to an a(l,3)fucosyltransfer~e found in human plasma (42,47).

Human Genomic Library and Screening-The preparation and screening of a human genomic DNA library has been described previously (45). The library was screened with a radiolabeled 1.7-kb XhoI-XbaI fragment isolated from the insert in pcDNAl-Fuc-TI11 (44,47).

DNA Sequence Analysis-Southern blot analyses were used to identify phage clones with strong homology to the human Fuc-TI11 probe. A 1.3-kb fragment from a strongly hybridizing phage with a disparate PstI restriction pattern was subcloned into the plasmid pTZ18R (Pharmacia LKB B i o ~ h n o l o ~ Inc.). The DNA sequence of this fragment was determined by the dideoxy chain termination method (48) using T7 DNA polymerase (Sequenase, U. S. Biochemical Corp.). Oligonucleotides were synthesized according to flanking plasmid sequences and subsequently according to insert sequence. The DNA sequence of this fragment was nearly identical to the 3' end of the Fuc-TI11 coding region but diverged immediately distal to the stop codon of the Fuc-TI11 coding segment (see "Results").

To isolate a single restriction fragment encompassing a possible novel coding region, a 7.1-kb EcoRI fragment that cross-hybridized with the Fuc-TI11 probe was isolated from this phage DNA. This fragment was cloned into the EcoRI site of pcDNAl, to yield the vector pcDNAf-Fuc-TVIE,RI. The DNA sequence of the insert in this vector was determined as described above.

To construct an expression vector consisting largely of the coding sequence of this gene, PCR (49) was used to amplify this segment from ~ D N A l - F u c - T V I ~ ~ ~ * . F i f t y - n a n o ~ a m aliquots of plasmid template were used in reactions with the GeneAmp kit (Perkin-E~mer Cetus). Twenty cycles of amplification were performed, consisting of 1.5 min at 94 "C, followed by annealing and extension for 3.5 min at 72 "C. Primers immediately flanking the open reading frame (Fig. 1) were chosen to specifically amplify the Fuc-TVI coding sequence (upper strand primer: g c g ~ T C C T C T C T C C C C A C ~ C C C A G - AGACTCTGA, nucleotides -32 through -3, extraneous base pairs in lower case, synthetic HindIII site underlined; lower strand primer: GGTGAAGCTTCAGGCAAACGAGTCCTTAGG, corresponding to reverse complement of nucleotides 1154-1183; endogenous HindIII site underlined). The PCR product was digested with HindIII, and the 1.2-kb fragment was cloned into the HindIII site of pcDNAl. A representative plasmid containing the insert in the sense orientation with respect to the plasmid's cytomegalovirus promoter was designated pcDNAl-Fuc-TVI. The insert in plasmid pcDNAl-Fuc-TVI was sequenced in its entirety to exclude possible PCR errors. Se- quence analyses were performed using the sequence analysis software package of the University of Wisconsin Genetics Computer Group (50) and the MacVector version of the IBI Pustell Sequence Analysis Software package (Kodak/IBI).

Transfection of COS-1 and CHO-T Cells-COS-1 cells, cultured as described (23, 51), were transfected with pfasmid DNAs (pcDNAl- Fuc-TVI, this work; pcDNAl-Fuc-TIII, Refs. 44 and 47; pcDNAl- Fuc-TIV, Ref. 45; pcDNAl-Fuc-TV, Ref. 47; or pcDNAl, Invitrogen) using a DEAE-dextran procedure (44, 52). A Chinese hamster ovary cell line (53) that stably expresses the polyoma large T antigen (CHO-

T cells)' and that replicates polyoma replication origin-containing plasmids (54, 55) was transfected with plasmid DNAs using a lipo- some-based reagent ~iV-[l-(2,3-diol~yloxy)propyl]-N,iV,iV-tri- methylammoniummethylsulfate, DOTAP, Boehringer Mannheim) and a modification of the manufacturer's protocol.

Antibodies-Anti-Lewisx antibody (anti-SSEA-1, IgM, Ref. 3) was provided by Dr. Davor Solter (Wistar Institute, Philadelphia). Anti- H and anti-lewis a monoclonal antibodies (IgM) were purchased from Chembiomed Ltd. (Edmonton, Alberta, Canada). A monoclonal anti-VIM-2 antibody (IgM) was purchased from An-der-Grub (Kaum- berg, Austria). Anti-sialyl Lewis x monoclonal antibody CSLEXl (IgM) and anti-sialyl Lewis a monoclonal antibody CSLEAl (IgG3) were purchased from the UCLA Tissue Typing Laboratory (Los Angeles, CA). The IgM monoclonal antibody FH6 was provided by Dr. Reiji Kannagi (Aichi Cancer Research Institute, Nagoya, Japan). FH6 recognizes difucosyl sialyl Lewis x (16, 17, 24, 56). Fluorescein- conjugated goat anti-mouse IgM and IgG antibodies were purchased from Sigma.

Flow C y ~ o ~ ~ ~ Analysis-COS-1 or CHO-T cells transfected with fucosyltransferase expression vectors, or with a control vector, were harvested (57) 72 h after transfection and stained with monoclonal antibodies diluted in staining media (23,44,45,47). Anti-Lewis a and anti-H antibodies were used at 10 pg/ml. Anti-SSEA-1 was used at a dilution of k1000, Anti-sialyl Lewis x was used at 10 pg/ml. Anti- sialyl Lewis a was used at a dilution of 1:500 (10.8 pgfml). Anti-VIM- 2 was used at 1:50. Anti-difucosyl sialyl Lewis x (FH6) was used at 1:5. Cells were then stained with fluorescein isothiocyanate-conjugated goat anti-mouse IgM or anti-mouse I& and analyzed on a Becton-Dickinson FACScan (23,44,45, 47). Approximately 27-32% of the population of COS-1 and CHO-T cells transfected with each fucosyltransferase vector express the Lewis x determinant. Over 95% of control transfectants (pcDNAl-transfected COS-1 and CHO-T cells) stained with the anti-Lewis x antibody exhibit a fluorescent intensity below the gated settings.

F ~ o s y l t r a ~ f e r ~ e Assays-Cell extracts containing 1% Triton X- 100 were prepared from transfected COS-1 cells (44, 47). Fucosyl- transferase assays with neutral acceptor substrates were performed

6.2), 5 mM ATP, 10 mM L-fucose, 20 mM MnC12, 3 p~ GDP-["C] in a volume of 20 pL1 and contained 25 mM sodium cacodylate (pH

fucose, and 10 pg of cell extract protein. Acceptor substrates were added to a final concentration of 20 mM (iV-acetyllactosamine, Gal@14GlcNAc; lactose, Gal@G+4Glc; and facto-N-biose I, Gal@l+ 3GlcNAc) or 5 mM (2'-fucosyllactose, Fucal-+2Gal@l4Glc). Con- trol assays with no added acceptor were performed using the same conditions. Reactions were incubated at 37 "C for periods of time to yield linear rates (<20% of GDP-fucose consumed during the course of the reaction; typically 1 hf. Assays were terminated by addition of 20 pl of ethanol, followed by addition of 560 p1 of distilled water. The terminated assays were centrifuged at 12,000 X g for 5 min, and the supernatants were collected. To determine total radioactivity in the reaction, an aliquot of each terminated reaction supernatant was subjected to scintillation counting. Another aliquot was applied to a column containing Dowex 1x2-400, formate form (44,471. To quan- titate incorporation of radioactive fucose into product, the flow- through fraction, and 2 ml of a subsequent water elution, were collected, pooled, and counted.

Amine adsorption HPLC (44,581 was used to confirm the structure of the product formed with N-acetyllactosamine. The neutral product in the Dowex eluate was lyophilized, resuspended in 130 p1 of 70% acetonitrile, and subjected to fractionation on a Dynamax 60A (primary amine) column equilibrated in acetonitrile/water (7030). Com- pounds were efuted isocratically with acetonitrile/water (7030) a t a flow rate of 1 mllmin. The trisaccharide product of these reactions exhibited a retention time identical to a previously described Gal@l+ 4[Fucal-+3]GlcNAc standard (47). The a anomeric configuration of fucosidic linkages was confirmed by a-L-fucosidase digestion, which was performed as described previously (47) using the neutral HPLC- purified product formed with iV-acetyllactosam~e. The product of digestion was identified by comparison to a radiolabeled fucose standard (47), using the amine adsorption HPLC method described above.

(NeuNAca2+3Gal/314GlciVAc, Oxford GlycoSystems), or an ac- Fucosyltransferase assays using a(2,3)sialyl-N-acetyllactosamine

companying N-acetyllactosamine control (above), were performed in a volume of 20 pl and contained 25 mM sodium cacodylate (pH 6.21, 5 mM ATP, 10 mM L-fucose, 3 @M GDP-['*C]fucose, 20 mM MnCL, 20 mM acceptor substrate, and 10 pg of cell extract protein. Reactions

P. L. Smith and J. B. Lowe, manuscript in preparation.

Molecular Cloning of a Fourth Human Fucosyltransferase Gene 24577

were incubated at 37 “C for 1 h and then terminated by dilution with 580 pl of 5 mM sodium phosphate (pH 6.8). The terminated assays were centrifuged at 12,000 X g for 5 min, and the supernatants were collected. A 200-pl aliquot of this material was subjected to scintillation counting to determine the total amount of radioactivity in the terminated diluted assay. A second 2 0 0 4 aliquot was applied to a Dowex 1-X8 (PO:-) column (500 pl packed volume) equilibrated as previously described (59). The flow-through fraction and three sepa- rate l-ml water elutions were collected, pooled, and an aliquot of this material was counted as a measure of product formation. A portion of this material was also lyophilized to dryness and then resuspended in 130 pl of 75% acetonitrile, 25% 2 M triethylamine acetate (pH 5.5) for subsequent product analysis by HPLC as described below. To verify that the Dowex chromatography procedure effectively fractionated radioactive product from radioactive GDP-fucose substrate, a third aliquot of the terminated diluted assay was lyophilized to dryness, resuspended in 130 p1 of 75% acetonitrile, 25% 2 M triethylamine acetate (pH 5.5), and was also subjected to analysis by HPLC, as described below.

Lyophilized resuspended samples derived from unfractionated assays or from Dowex column-fractionated assays were subjected to ion-suppressed amine-adsorption HPLC using a Varian AX-5 column equilibrated in 75% acetonitrile, 25% 2 M triethylamine acetate (pH 5.5) (60). Products were eluted isocratically with 75% acetonitrile, 25% 2 M triethylamine acetate (pH 5.5) for 12 min at a flow rate of 1 ml/min and then with a gradient of 50% acetonitrile, 25% 2 M triethylamine acetate (pH 5.5), 25% water (at 12.1 min) to 5% acetonitrile, 25% 2 M triethylamine acetate (pH 5.5), 70% water (at 60 min), for 47.9 min at 1 ml/min. The column eluant was monitored with an on-line radioisotope detector (Beckman Instruments). Prod- ucts were identified by comparison to elution times exhibited by radiolabeled standards (~-[“C]fucose, 9 min; Gal~l4( [ ’*C]Fucal+ 3)GlcNAc, 27 min; NeuNAc2+3Gal~14([14C]Fuca1+3)GlcNAc, 32 min; GDP-[“C]fucose, 51 min). Parallel HPLC separations made with the Dowex column-fractionated assays, and the unfractionated assays indicated that the Dowex column procedure removes more than 95% of the radiolabeled GDP-fucose in the assay, yet does not retain detectable amounts of the sialylated product.

The structure predicted for the fucosylated product generated with cu(2,3)sialyl-N-acetyllactosamine (NeuNAca2+3Gal~14[Fucal~ 31GlcNAc) was confirmed by neuraminidase digestion followed by HPLC analysis. The HPLC-purified fucosylated tetrasaccharide was digested with Clostridiumperfringens neuraminidase (Sigma Type X; 4 milliunits/10,000 cpm of tetrasaccharide product) for 8 h at 37 “C in a buffer consisting of 20 mM sodium cacodylate (pH 6.2) and 10 mM L-fucose. The digest was then subjected to fractionation by amine adsorption HPLC. The radiolabeled product of neuraminidase digestion co-eluted with a 3-fucosyl-N-acetyllactosamine standard (47).

PCR Analysis of Human Chromosomal Somatic Cell Hybrids- Genomic DNAs prepared from two panels of hamster-human somatic cell hybrids (47) were obtained from BIOS Corp. (New Haven, CT). Murine microcell hybrids containing human chromosome 11 (61,62) and 7 (63) have been described previously (47). Hybrid cell genomic DNAs were analyzed using the PCR primers detailed above for the cloning of pcDNAl-Fuc-TVI (Fig. 1). Amplification using these two primers yields a 1.2-kb fragment from the Fuc-TVI gene. Thirty cycles of amplification were performed, consisting of 1.5 min at 94 “C, followed by annealing and extension for 3.5 min at 72 “C. Multiple negative controls with no added genomic DNA template were run in parallel with test samples to exclude the presence of contaminating DNA sequences. Other negative controls included hamster and mouse genomic DNAs, which were used to rule out the possibility of non- specific amplification of homologous sequences across species. Posi- tive controls for these experiments consisted of genomic DNA from the Lewis-positive individual who served as the source of the genomic DNA library and other human genomic DNAs. Cloned segments of Fuc-TIII, Fuc-TIV, Fuc-TV, and Fuc-TVI were used to confirm primer specificity for Fuc-TVI. A portion of each PCR reaction was subjected to restriction endonuclease digestion with NcoI (which does not cleave Fuc-TI11 or Fuc-TV at the site shown in Fig. 1) to verify the identity of amplified fragments. PCR products and restriction digests were analyzed on 1.2% agarose gels.

Southern Blot Analyses of Human Genomic DNA-Genomic DNA was prepared using a previously described procedure (64). To control for DNA sequence polymorphisms previously found in human a(1,3)fucosyltransferase genes: genomic DNA was prepared from the

B. W. Weston, R. Mollicone, and J. B. Lowe, unpublished data.

same individual whose DNA was used to construct the original genomic library (45,47). Genomic DNA was digested with restriction endonucleases, fractionated through 0.8% agarose gels, and subjected to Southern transfer using previously described methods (45,64). For comparability in subsequent experiments with several probes, multiple blots were prepared from identical sets of restriction digests electrophoresed on a single gel. Hybridization and wash conditions are as described previously (45, 64) and as detailed in the legend to Fig. 4. The hybridized and washed blots were subjected to autoradiography for 48-120 h, and the resulting autoradiograms were then scanned using an OmniMedia 6cx scanner (XRS, Torrence, CA). The images were grouped and overall contrast was adjusted using Adobe Photoshop 2.01, followed by labeling with Adobe Illustrator 3.2 (Mountain View, CA).

Southern blots were probed with “gene-specific” probes or with probes that recognize conserved portions of the genes that correspond to their putative catalytic domains (“catalytic domain” probes). Gene- specific probes were derived from regions 3’ to the highly homologous coding portions of each gene (Fig. 2, A and B ) . Each gene-specific probe was chosen to have less than 70% homology to the other fucosyltransferase sequences or to other sequences in the GenBank data base, and human repetitive sequences were avoided (44, 65). Catalytic domain probes, derived from either Fuc-TI11 or from Fuc- TIV, were chosen from regions of highest DNA sequence conservation among the four human fucosyltransferase genes (Figs. 1 and 2; see also Refs. 45 and 47). The Fuc-TV gene-specific probe was isolated by XbaI digestion of a clone encompassing 1.3 kb of Fuc-TV 3’ untranslated region (immediately downstream of the insert in pc- DNA1-Fuc-TV, Ref. 47).4 The remaining probes were generated using PCR amplification of cloned templates as described above and the primers listed in Table I. The positions and sizes of each probe are illustrated in Fig. 2B.

RESULTS

Molecular Cloning of a Human Genomic DNA Segment Homologous to, but Distinct from, the Human Fw-TIII, -IV, and - V Genes-Four or more distinct a( 1,3)fucosyltransferase activities have been described previously in biochemical analyses of human tissues and cells (35-42). Candidate genes encoding fucosyltransferases similar to three of these activities have been reported (26, 44-47). Two of these enzymes, Fuc-TI11 and Fuc-TV, maintain extraordinarily high levels of protein and nucleic acid similarity, as depicted schematically in Fig. 2A but display marked differences in acceptor substrate specificities (47). Another gene termed Fuc-TIV (45, 46) or ELFT (26) is less similar in its sequence, and its corresponding enzyme also exhibits disparate substrate requirements (45). The Fuc-TIV and Fuc-TV genes were isolated from a human genomic DNA phage library screened at low stringency with the cDNA encoding Fuc-TI11 (45, 47). Eighteen phages isolated from this screening were characterized. Nine of these exhibited relatively weak hybridization to the Fuc-TI11 probe and proved to contain the Fuc-TIV gene. Nine other phages exhibited strong hybridization signals. Seven of these phages proved to contain the Fuc-TV gene, whereas one other was found to contain sequences co-linear with the Fuc-TI11 cDNA. The remaining phage exhibited a unique PstI restriction pattern, suggesting that it represents a sequence distinct from the other a(1,3)fucosyltransferase genes. A 7.1-kb hybridization-positive EcoRI fragment from this phage was subcloned and sequenced.

DNA sequence analysis of this genomic fragment identifies a single long open reading frame (Fig. 1, bp 1-1080) predicted to yield a protein with striking amino acid sequence similarity to Fuc-TI11 (Fig. 1). This reading frame begins at a methio- nine codon found within a sequence context corresponding to Kozak’s consensus rules for translation initiation sites (66). A single strongly hydrophobic peptide segment corresponding

‘ B. W. Weston, P. L. Smith, R. J. Kelly, and J. B. Lowe, unpublished data.

24578 Molecular Cloning of a Fourth Human Fucosyltransferase Gene TABLE I

PCR primers used to generate Southern blot probes Probe Primers"

Fuc-TI11 catalytic domain U: CCCAGCAGAAGCAACTACGAGAGGTTCCTGCC L : GGCAGATGAGGTTCCCGGCAGCCCAGGCAC

Fuc-TI11 gene-specific

Fuc-TIV catalytic domain

Fuc-TIV gene specific

U: CCACCCGGGAGTGATGGTTCTGGCTGATTT L : CGATCCCACCTGTACCCTATAAGTGGTGGT U: GGTGCCCGAAATTGGGCTCCTGCACAC L: CCAGAAGGAGGTGATGTGGACAGCGTA U: CAGCTGGTTCGAGCGGTGAAGCCGCGCT L: CAGAAAAACGTGAATCGGGAACAGTTGTGT

Fuc-TVI gene-specific U: AAAGGGGGAGTGTCCCTTTCTGAGTGTGCAA L: CCTGTCTCCCACCCAGATCCAGCGTCCCAA

a U, upper; L, lower.

-120

1

121

235

355

175

595

715

135

955

1072

FIG. 1. Comparison of DNA and derived protein sequences of Fuc-TVI and Fuc-TIII. Residue 1 of the nucleotide sequence of the Fuc-TVI genomic DNA segment is assigned to the A residue of the putative initiation codon. The derived protein sequence of Fuc-TVI, in single letter code, is displayed above the Fuc-TVI DNA sequence. Dashed lines below the numbered Fuc-TVI DNA sequence denote DNA sequence identity between Fuc-TI11 and Fuc-TVI. Nucleotide sequence differences are indicated within the dashed line representing the Fuc- TI11 sequence. Blank areas correspond to regions of Fuc-TVI that have no counterpart in the Fuc-TI11 sequence or vice versa. The derived amino acid sequences for Fuc-TI11 and Fuc-TVI are identical except where indicated by the inclusion of differing amino acids (in single letter code) underneath the corresponding codons in Fuc-TIII. The doubly underlined amino acids in Fuc-TVI correspond to a putative transmembrane domain. The positions of the two PCR primers used to construct pcDNAl-Fuc-TVI and to analyze chromosomal hybrid DNAs are underlined. The location of the NcoI site in Fuc-TVI used to confirm the identity of the PCR product is italicized and labeled. Circled asparagine residues in the Fuc-TVI protein sequence represent potential asparagine-linked glycosylation sites. The DNA sequence in the 5'- untranslated region highlighted by stippled overlining corresponds to a sequence motif consistent with a splice acceptor consensus sequence. Similar analyses for Fuc-TIV and Fuc-TV have been published previously (45,47).

to amino acid residues 15-34 is flanked by basic amino acids amino acids (data not shown). Because of this strong sequence and corresponds to the signal-anchor sequence characteristic similarity to previously described human a(l,3)fucosyltrans- of mammalian glycosyltransferases (33, 34) and other type I1 ferases, we term the derived protein sequence Fuc-TVI. transmembrane proteins (67). Marked sequence discontinuities occur between these derived

The open reading frame predicts a 358-amino acid protein protein sequences in a region corresponding to base pairs 121- that is identical to Fuc-TI11 at 306 of 361 corresponding amino 449 of Fuc-TVI (Fig. 1). This region of Fuc-TVI does not acid positions (Fig. 1) and identical to Fuc-TV at 334 of 374 contain the 33-bp nucleotide sequence insertion found in FUC-

~oZecular Cloning of a Fourth Human ~ucosy l t r~ns fe r~e Gene 24579

T V relative to Fuc-TI11 (47). A total of 97 other amino acid sequence differences between Fuc-TIII, Fuc-TV, and FUC- TVI are observed in this region; by contrast, only 14 amino acid sequence differences are noted between the COOH-terminal halves of these three proteins (Fig. 2A) . Four asparagine residues that represent potential N-glycosylation sites are found within the Fuc-TVI predicted protein (Fig. 1). Two of these, located at amino acid residues 153 and 184, correspond to N-linked glycosylation sites present a t precisely corresponding positions in Fuc-TI11 and Fuc-TV (Fig. 2 A ) . The other two, located at positions 46 and 91, are not found in corresponding locations in Fuc-TIII, but are present in Fuc- T V (Fig. 2A).

The Fuc-TVI DNA sequence diverges from the Fuc-TI11 sequence at a position approximately 18 base pairs 5‘ to the initiator codons of the two sequences (data not shown). A sequence corresponding to the consensus for splice acceptor sites (68,691 is found in this region of the Fuc-TVI sequence,

A.

F~c-Tt l l v v f 361 88

FuC-TV 1 I 374 a0

v v v , FUC-TVI B I O mas

Fuc-TtV 405 a8

B. m7u

m E ~ R V W I

111 4 4 trazsw

BnH

mo z m o

EcoRl t.( ”” M w ~ m V I - I P I I - 1 1 ” -

tnn

VI

2(Dap - EA” m I m EooRvElwl

I 9 I _ I -, 4 4

i ”

am o mt+

” I

7 - FIG. 2. Schematic depictions of sequences from the four

cloned human a(l.3)fucosyltransferases. A, derived amino acid sequence comparisons of human c~(1,3)fucosyltransferases using the Fuc-TI11 polypeptide as prototype. The putative transmembrane domains for each predicted protein are shuded, and vertical bars denote amino acid residues differing from the Fuc-TI11 template. The primary structure for Fuc-TIV, which is less homologous and non- syntenic to the other sequences, is represented in the bottom schematic (405 ami40 acids). The sizes of the predicted polypeptides are shown approximately to scale. The positions of potential asparagine- linked glycosylation sites in the four sequences are indicated by stick figures. B, restriction maps of human a(l,3)fucosyltransferase genomic clones 111, V, VI, and IV. The positions of restriction endonuclease cleavage sites relevant for the analyses in Fig. 4 are indicated above each schematic DNA sequence, which are shown approximately to scale except where indicated (//). The open rectangular bars represent the coding regions for each gene. Primers used to generate catalytic domain and gene-specific probes by PCR (see Table I and “Experimental Procedures”) are indicated by arrows above and below each sequence. Catalytic domain probes are represented by dotted r e c ~ ~ g ~ s below the Fuc-TI11 and Fuc-TIV schematics. The ur(l,3)fucosyltransferase gene-specific probes are depicted as solid rectangles below each sequence.

beginning at a position 96 base pairs 5’ to the initiator codon (Fig. 1). The Fuc-TI11 and Fuc-TVI DNA sequences also diverge at a position 57 base pairs 3’ to their respective termination codons.

In summaryt the cross-hybridizing open reading frame has the potential to encode a type I1 transmembrane protein with marked amino acid sequence similarity to two other cloned human fucosyltransferases. This relationship is most salient in a region corresponding to the catalytic domain of these latter two enzymes and is analogous to the structural similar- ities previously described for Fuc-TI11 and Fuc-TV (Fig. 24 and Ref. 47). Given the sequence discontinuities identified in the DNA sequence flanking the open reading frame, these comparisons suggest that this cross-hybridizing sequence represents the coding region of an a{l,3)fucosyltransferase gene.

The Genomic DNA Restriction Fragment Determines Expression of a Related but Distinct a(l,3)Fucosyltransferme- To determine if this open reading frame encodes a functional a( 1,3)fucosyltransferase, the putative coding region segment was cloned into a mammalian expression vector and tested in transfection experiments. The expression plasmid pcDNA1- Fuc-TVI (see “Experimental Procedures”) was first transfected into the COS-1 cell line, a mammalian host that does not normally express a( 1,3)fucosyltransferase activity (44). Extracts prepared from transfected cells were tested for a(1,3)- and a(l,4)fucosyltransferase activity in assays containing low molecular weight oligosaccharides previously shown to function in vitro as acceptor substrates (44-47). Extracts prepared from pcDNAl-Fuc-TVI-transfected COS- 1 cells were found to contain an a(l,3)fucosyltransferase activity that efficiently used N-acetyllactosam~ne to generate 3-fucosyl-N-acetyllactosamine (specific activity of 880 pmol/ mg/h).Underthesameassayconditions,thea(1,3)fucosyltransferase activity in this same extract preparation did not use two other neutral type I1 acceptor molecules, 2’-fucosyIlactose and lactose (specific activity less than 0.5 pmol/mg/h). In addition, no detectable a(1,4)fucosyltransferase activity was found in multiple assays with the type I substrate lacto-N- biose I (specific activity less than 0.5 pmol/mg/h). As noted previously (44,45,47), extracts prepared from pcDNAl-transfected COS cells contain no detectable fucosyltransferase activity with any of the four acceptors tested.

A second set of experiments were completed to determine the relative efficiency with which the enzyme could use the sialylated acceptor a(2,3)sialyl-N-acetyllactosamine. Extracts from pcDNAl-Fuc-TVI-transfected COS-1 cells used a(2,3)sialyl-N-acetyllactosamine with a specific activity of 1048 pmol/mg/h. The fucosyltransferase activity in these same extracts consistently used the nonsialylated acceptor N-acetyllactosamine with slightly less efficiency (932 pmol/ mg/h). Control extracts prepared from pcDNA1-transfected COS-1 cells transferred no detectable fucose to a(2,3)sialyl- N-acetyllactosamine. These results indicate that the genomic DNA fragment in pcDNA1-Fuc-TVI encodes an a(l,3)fucosyltransferase.

The acceptor substrate preferences exhibited by pcDNA1- Fuc-TVI-transfected COS-1 cells contrast with those described previously for the other cloned human a(l,3)fucosyltransferases. Fuc-TI11 can use all of the five acceptors tested with high efficiency, although highest relative activity is found with the type I acceptor lacto-I?-biose I (44, 47). Fuc-TIV utilizes N-acetyllactosamine, but not the low molecular weight acceptor substrates a(2,3)sialyl-N-acetyllactosamine, ~‘-fucosyl~actose, or lacto-~-biose I (45, 46). Both Fuc-TV and Fuc-TVI fucosylate N-acetyllactosamine and ~(2,3)sialyl-N-acetyllactosamine with similarly high efficien-


cies. Fuc-TV, however, uses 2‘-fucosyllactose at approximately 42% of the efficiency with which it uses N-acetyllactosamine (Ref. 47). Fuc-TV also shows low but reproducible activity with lactose and lacto-N-biose I (approximately 10% of the activity demonstrated with N-acet~llactosamine, Ref. 47). By contrast, Fuc-TVI shows no detectable activity with any of these other neutral type I1 or type I acceptors. Thus, these results indicate that Fuc-TVI maintains catalytic properties that are similar to, but nonetheless distinct from, those exhibited by the other cloned a(1,3)fucosyltransferases.

To examine acceptor substrate specificities in uiuo, COS-1 cells transfected with human ~cosyltransferase vectors were analyzed by flow cytometry. COS-1 cells maintain substrate levels of GDP-fucose and glycosylated acceptors and can construct surface-localized a(1,3)- and a(l,4)-fucosylated oligosaccharides when transfected with fucosyltransferase expression vectors (23, 44, 45, 47). As shown in Fig. 3A, COS-1 cells transfected with pcDNAl-Fuc-TVI exhibit surface expression of the Lewis x and sialyl Lewis x determinants but do not express the Lewis a or sialyl Lewis a antigens. By contrast, COS-1 cells transfected with pcDNAl-Fuc-TI11

A. 8W

a w n

600

400

200

” pcONA-1 FuoTIII Fuc-TIV FUC-TV FUC-TVI

6.

anli-H

anli-sLex anti.Lex

anti-VIM.2 anli-difuc-sbx

pcDNA-1 Fuc-Tiit Fw-Tlv Fuc-TV Fuc-TVf

FIG. 3. Flow cytometry histograms of COS-1 and CHO-T cells transfected with human ~~(1,3)fucosyltransferase expression vectors. COS-1 cells ( A ) and CHO-T cells ( B ) were transfected with ~ D ~ A l - F u c - T I I I , -Fuc-TIV, -Fuc-TV, -Fuc-TVI, or with the negative control plasmid pcDNAl. After a 72-h expression period, the cells were stained with the monoclonal antibodies listed in the insets of A and B and subjected to flow cytometry analysis as described under ‘‘Experimen~l Procedures.” The data presented here are the mean fluorescence intensities of the antigen-positive population of transfected cells (gated at 1 X 10’ units for COS-1 cells and 1.5 X 10’ units for CHO-T cells). To allow direct intra-vector comparison of mean fluorescence intensities, the same population of transfected cells for each vector and host cell line was analyzed in parallel with all of the antibodies listed in either A or B.

stain with antibodies specific for the Lewis x, sialyl Lewis x, Lewis a, and sialyl Lewis a determinants (Fig. 3A and Ref. 44). The results obtained with pcDNAl-Fuc-TVI also differ from those obtained when COS-1 cells are transfected with pcDNA1-Fuc-TIV, which constructs surface-localized Lewis x determinants when expressed in COS-1 cells, but does not determine expression of the sialyl Lewis x antigen (Fig. 3A and Ref. 45). Comparison of the flow cytometric profiles for pcDNA1-Fuc-TVandpc~NAl-Fuc-TVI in COS-1 cells shows that these two cloned fucosyltransferases are functionally more similar to each other than to either Fuc-TI11 or Fuc- TIV; both direct expression of Lewis x and sialyl Lewis x molecules, but not Lewis a or sialyl Lewis a determinants (Fig. 3A ) .

To further examine the cloned a(l,3)fucosyltransferases for their abiIities to fucosy~ate terminal and internal ~-acetyIglu- cosamine residues, CHO cells containing the polyoma large T-antigen (CHO-T cells, see “Experimental Procedures’’ and Refs. 54 and 55) were transiently transfected with each expression plasmid. Unlike COS-1 cells, CHO cells are known to contain glycosylated type I1 precursors for the VIM-2 determinant (45, 53) but do not construct type I precursors (23, 30, 45). CHO cells also maintain type I1 precursors for the Lewis x and sialyl Lewis x epitopes. As shown in Fig. 3B, all four cloned fucosyltransferase sequences are capable of determining surface-localized Lewis x in CHO cells. Fuc-TIII, -IV, and -V, but not Fuc-TVI, are able to construct the VIM- 2 structure. In addition, plasmids pcDNAl-Fuc-TIII, -Fuc- TV, and -Fuc-TVI each direct expression of sialyl Lewis x and difucosylated sialyl Lewis x determinants, whereas pc- ~NAl-Fuc-TIV does not.

Thus, the pattern of sialylated lactosaminoglycan fucosylation shared by Fuc-TIII, -V, and -VI differs from that demonstrated by the less homologous Fuc-TIV, which is un- able to fucosylate the terminal N-acetylglucosamine on sialylated precursors regardless of the presence or absence of fucose on the internal N-acetylglucosamine residue. These results suggest a correlation between shared structural determinants in Fuc-TIII, -V, and -VI and their abilities to express sialyl Lewis x and difucosyl sialyl Lewis x moieties. Likewise, these results suggest a correlation between enzyme-specific peptide motifs and the enzymes’ disparate abilities to determine expression of the VIM-2 epitope. Such differences in the ability to construct the VIM-2 determinant likely depend, a t least in part, upon variance in the rates with which each enzyme can fucosylate internal and terminal N-acetylglucosamine residues found in a(2,3)-sialylated lactosaminoglycans and upon modification of such rates by prior fucosylation at alternate positions (differences in order of fucosylation, see “Discussion”).

The Fuc-TVI Gene Is Syntenic to the Homologous Fuc-TIII and Fuc-TV Genes on Chromosome 19-Biochemical and genetic analyses indicate that human a(l,3)fucosyltransferase activities can be assigned to loci on chromosomes 19 (43) and 11 (35, 41). We reported previously the molecular localization of Fuc-TI11 and Fuc-TV to chromosome 19 and Fuc-TIV to chromosome 11 (47). To determine the chromosomal location of Fuc-TVI, PCR was used to analyze a series of somatic cell hybrid DNAs informative for each human chromosome (see “Experimental Procedures” and Ref. 47). Diagnostic fragment sizes for Fuc-TVI (1.2 kb) and confirming NcoI restriction digests (450- and 750-bp fragments, data not shown) were obtained only when using genomic DNA template from hybrids containing human chromosome 19 (cell lines 683, 750, 756, 860, 867, 1006, and 1099; Ref. 47). All other human chromosomes were excluded, since these fragments were not


detected with hybrid DNAs containing, in aggregate, every other human chromosome. These data assign the Fuc-TVI locus to chromosome 19.

Two Distinct Classes of Human Genomic DNA Fragments Are Defined by Homology with the Catalytic Domains of Fuc- TIZZ and Fuc-TIV-Several human a(l,3)fucosyltransferase activities have been described (10, 35-43), suggesting the existence of an equivalent number of a( 1,3)fucosyltransferase genes. We used Southern blot analyses and probes derived from each of the four cloned human a(l,3)fucosyltransferase genes to determine if these sequences fully represent all structurally similar human a(l,3)fucosyltransferase genes or if others remain to be isolated.

One pair of probes (catalytic domain probes, see “Experi- mental Procedures”) was used to sample the genome for additional DNA sequences that cross-hybridize to the most highly conserved segments of the four a(l,3)fucosyltransferase genes. Another set of probes (gene-specific probes, see “Experimental Procedures”) was used to assign specific genes to bands detected with the catalytic domain probes. Restric- tion enzymes used to digest the genomic DNA were chosen so that, for each known gene, a single restriction fragment would be identified by its gene-specific probe, allowing comparison with the catalytic domain probe analyses. Any band(s) detected with a catalytic domain probe, but not detected with any of the gene-specific probes, would represent new cross- hybridizing sequence(s) and would thus be candidateb) for novel fucosyltransferase gene segment@.).

The DNA sequence of the Fuc-TI11 catalytic domain probe is 96 and 95% identical, respectively, to corresponding positions in the Fuc-TV and Fuc-TVI genes and is 62% identical to the corresponding segment in the Fuc-TIV gene. This probe identifies several bands when used under conditions of decreased hybridization stringency (Fig. 4A). Regardless of the restriction enzyme used, however, each of the bands detected with the Fuc-TI11 catalytic domain probe used at low stringency can be assigned to one of the four cloned a(1,3)fucosyltransferase genes, by comparison to the blots probed at high stringency with the four gene-specific probes (Fig. 48). This result indicates that the Fuc-TI11 catalytic domain probe is not detecting any novel cross-hybridizing DNA sequences and suggests that the Fuc-TIV, -V, and -VI genes fully represent all existing human genes with substan- tial DNA sequence similarity to the Fuc-TI11 gene.

The DNA sequence of the Fuc-TIV catalytic domain probe is 73, 72, and 70% identical, respectively, to corresponding positions in the Fuc-TIII, -V, and -VI genes. Like the Fuc- TI11 catalytic domain probe, this probe also identifies several bands when used under conditions of decreased hybridization stringency (compare Fig. 4, C and A ) . Most, but not all, of these fragments may be assigned to one of the four cloned a( 1,3)fucosyltransferase genes. Specifically, this probe identifies several cross-hybridizing restriction fragments (Fig. 4C, asterisks) that do not correspond to bands detected with the gene specific probes (Fig. 4B). These fragments exhibit cross- hybridization intensities similar to those observed when the Fuc-TIV gene is detected at low stringency with the Fuc-TI11 probe (Fig. 4A) but are not detected with the Fuc-TI11 catalytic domain probe using the same hybridization conditions (Fig. 4C). These results indicate that the Fuc-TIV catalytic domain probe is detecting novel cross-hybridizing DNA sequences and suggest the possibility that these sequences may correspond to one or more additional, as yet uncloned, fucosyltransferase genes, whose sequences may be more similar to Fuc-TIV than to the chromosome 19-localized genes Fuc- TIII, -V, and -VI.

DISCUSSION

Subsets of a(1,3)- and a(l,4)-fucosylated cell surface oligosaccharides figure importantly in the inflammatory process in their roles as ligands for the selectin family of cell adhesion proteins (23-32). These molecules, and related ones, may also maintain other important functions, including ones critical to morphogenic processes during early mammalian embryogenesis (1-7). Such other postulated functions remain to be definitively demonstrated, however. Expression of these molecules is controlled in large measure by tissue-specific and developmentally regulated expression of a( 1,3)fucosyltransferases (10,33,34). The number of these human enzymes and corresponding genes has not been defined. Similarly, the mechanisms that control their expression, and the functions of the oligosaccharides whose biosynthesis they determine are not completely understood. We have sought to study these issues by isolating and characterizing human a(l,3)fucosyltransferase genes. Previous reports have described three members of a structurally related a(l,3)fucosyltransferase gene family (26, 44-47). In this report, we describe the isolation and sequence of a fourth member of this family, its expression in mammalian cells, and an analysis of its functional and structural relationship to known and unknown sequences in the human genome.

This gene, termed Fuc-TVI, is most similar in its structure to the Fuc-TV gene. Both genes maintain intronless coding segments, whose nucleic acid sequences share 91% identity. Their respective protein products are also very similar, shar- ing 89% primary sequence identity. These enzymes are nearly identical in their COOH-terminal regions, and most sequence differences lie in a region between the enzymes’ transmembrane segments and the NH2-terminal ends of their catalytic domains. This is a relationship we previously noted when comparing the sequences of Fuc-TI11 and Fuc-TV (47), whose sequences also differ most in an analogously-positioned “hypervariable” region. The repetition of this relationship in a group of three distinct a( 1,3)fucosyltransferases reinforces the concept (47) that residues within this hypervariable region play important roles in determining the efficiency with which particular a( 1,3)fucosyltransferases use different acceptor substrates and, by inference, participate in binding interactions between the enzymes and acceptor molecules during fucosylation reactions.

The genes encoding Fuc-TV and Fuc-TVI are localized to human chromosome 19, along with the Fuc-TI11 gene. This latter gene also maintains an intronless coding region with a high degree of sequence identity to the coding portions of Fuc-TV and Fuc-TVI (although these latter two sequences are more similar to each other than they are to Fuc-TIII). By contrast, the other known human a(1,3)fucosyltransferase gene (Fuc-TIV) is located on human chromosome 11 and is substantially less similar to the three chromosome 19-localized a( 1,3)fucosyltransferase genes. These observations are consistent with a hypothesis that a relatively distant gene duplication event generated two distinct a( 1,3)fucosyltransferase genes on two different chromosomes and that their sequences have diverged substantially in the interim. It seems probable that more recent duplicative events have yielded the family of structurally similar chromosome 19-localized a(1,3)fucosyltransferase genes, whose sequences have not yet diverged substantially. This notion is also consistent with analyses of pedigrees with individuals deficient in a(l,3)fucosyltransferase activities, wherein Lewis blood group-negative phenotype is genetically linked to plasma a(l,3)fucosyltransferase deficiency (38). It is also consistent with preliminary Southern blot analyses of the chromosome


9.5-

6.7-

4.3-

2.3- 2.0-

1.35-

0- 1 1 1

a@- VI m-

- - - IV

- V *-

'Ill

. V ' IL

111 VI IV

V

I FUC-TIII I I FUC-TIII FUC-TIV FUC-TV FUC-TVI I FUC-TIV I Catalytic Domain Probe Gene-specific Probes Catalytic Domain Probe

A. B. C. FIG. 4. Southern blot analyses of human genomic DNA using a(l,3)fucosyltransferase catalytic domain and gene-specific

probes. Human genomic DNA was digested with the restriction enzymes BamHIIEcoRI, EcoRVIEcoRI, or EarI, electrophoresed on agarose gels, and subjected to Southern blot analysis. The results shown are derived from a master gel and blot containing reiterated sets of the three digests (10 pg of digested DNA/lane). The master blot was subdivided into strips after transfer, and each component strip containing three digests was separately hybridized with gene-specific or catalytic domain probes. Hybridizations were performed in a described previously solution (64) for a t least 18 h. For hybridizations with catalytic domain probes, the temperature was maintained at 35 "C. Blots hybridized with gene-specific probes were maintained a t 42 "C. Following hybridization, the blots were rinsed three times with room temperature 2 X SSC, washed using the final conditions noted below, and subjected to autoradiography (see "Experimental Procedures"). Roman numeral labels denote genomic fragments corresponding to specific cloned a(l,3)fucosyltransferases. A , Fuc-TI11 catalytic domain probe. The final wash was with 2 X SSC, 1% sodium dodecyl sulfate, a t 60 "C for 30 min. B, gene-specific probes. Fuc-TIII, final wash was 0.1 X SSC, 0.1% sodium dodecyl sulfate, 65 "C for 45 min. Fuc-TIV, final wash was 0.5 X SSC, 0.2% sodium dodecyl sulfate, 65 "C for 30 min. Fuc-TV, final wash was 0.5 X SSC, 0.2% sodium dodecyl sulfate, 65 "C for 30 min. Fuc-TVI, final wash was 0.1 X SSC, 0.1% sodium dodecyl sulfate, 65 "C for 45 min. C, Fuc-TIV catalytic domain probe. The first three separated lanes derive from a blot which was washed with 2 X SSC, 1% sodium dodecyl sulfate, at 63 "C for 30 min. The block of juxtaposed lanes at the far right in C was washed identically except that the temperature was maintained a t 60 "C. Asterisks mark fragments that do not correspond to previously cloned fucosyltransferase genes.

19-localized a( 1,3)fucosyltransferase genes? suggesting that these genes maintain close physical linkage. In this context, it will be interesting to define the genomic organization of these latter three genes to more precisely determine their physical relationships and to aid in the definition of the molecular basis for null alleles a t these loci.

Important issues that remain to be addressed include defin- ing the total number of human a( 1,3)fucosyltransferase genes and assigning existing (and potential) a(l,3)fucosyltransferase genes to the different a( 1,3)fucosyltransferase activities that have been described. Biochemical (37,39,40,42, 70-74) and genetic (38) analyses suggest that at least five different a( 1,3)fucosyltransferase activities may be found in human tissues or cells. I t is difficult to directly compare many of these results, however, because of differences in assay conditions used or substrates tested, and because the activities assayed have variously been derived from tissue or cell homogenates or enzymes variably purified from tissues and fluids. Nonetheless, tentative molecular correlates for some of these enzyme activities may be proposed, based on acceptor substrate specificities of a( 1,3)fucosyltransferase activities generated with cloned a(1,3)fucosyltransferase genes.

' B. W. Weston, unpublished data.

The Fuc-TI11 gene has been proposed to cqrrespond to the human Lewis blood group locus, based on its chromosomal assignment, and on the ability of its corresponding enzyme to exhibit both a( 1,3)fucosyltransferase and a( 1\4)fucosyltransferase activities (44, 71). Definitive,confirmation of this assignment will require a demonstration of tight linkage between a molecular defect in this gene and null alleles at the Lewis blood group locus. This is especially true in light of recent evidence indicating that a( 1,4)fucosyltransfer ase activity may be detected in some rare Lewis blood group- negative individuals (72).

The Fuc-TIV gene is known to correspond to an a(1,3)fucosyltransferase activity found in myeloid-lineage cells, based on acceptor substrate specificity and Northern blot analyses (26, 39, 45, 46). This gene is also transcribed in a wide variety of other tissues, however, as is the Fuc-TI11 gene, and transcripts from both genes are detectable in some tissues? These latter observations imply that caution must be exercised when interpreting the results of a(l,3)fucosyltransferase assays on tissue extracts because of the very real possibility that the "activity" detected in a tissue may be a composite of multiple distinct a( 1,3)fucosyltransferases.

B. W. Weston, P. L. Smith, and J. B. Lowe, unpublished data.


Assignment of the Fuc-TV and Fuc-TVI gene products to specific previously described a( 1,3)fucosyltransferase activities remains problematic, however. One prominent candidate is an a(1,3)fucosyltransferase activity found in the plasma of all except rare individuals (39, 40). Preliminary analyses of the acceptor substrate specificity of this activity indicate that i t does not correspond either to the Lewis blood group enzyme (Fuc-TIII) nor to the myeloid type enzyme (Fuc-TIV) (38- 40). While a minor fraction of a(l,3)fucosyltransferase activity in human plasma appears to be derived from myeloid lineage cells (approximately IO%, Ref. 39), the cellular origin(s) of the remainder of the circulating a(l,3)fucosyltransferase activity are not yet well defined (39). Nonetheless, analysis of pedigrees with individuals deficient in this activity are consistent with the hypothesis that it corresponds to a single gene product (38, 40, 41). This activity has recently been purified to homogeneity (42), and its acceptor substrate specificity has been examined in detail. The plasma enzyme efficiently uses N-acetyllactosamine and a(2,3)sialyllactosamine (apparent K,,, values of less than 1 mM, Ref. 42) but does not use lactose or 2’-fucosyllactose with significant efficiency (apparent K,,, values in excess of 60 and 26 mM, respectively, Ref. 42). This enzyme exhibits no detectable a( 1,4)fucosyltransferase activity when assayed with 1acto-N- biose I (42). As we have noted previously, these properties are similar, but not identical, to those maintained by Fuc-TV (47); this enzyme operates efficiently with both N-acetyllactosamine and a(2,3)sialyllactosamine, but it can also utilize 2’-fucosyllactose to some extent. Perhaps most importantly, Fuc-TV can also use lacto-N-biose I at a low but significant rate (47).

By contrast, the properties of the purified plasma enzyme are in may ways more similar to those we describe here for Fuc-TVI. Specifically, Fuc-TVI uses both N-acetyllactosamine and a(2,3)sialyllactosamine with high efficiency but does not operate on 2’-fucosyllactose nor on lacto-N-biose I. Since nonidentical methods (44, 45, 47) and acceptors (39, 40, 42, 47) have been used to analyze the purified and nonpurified plasma enzyme and the recombinant Fuc-TVI, direct comparison is difficult, and we therefore cannot yet be sure if Fuc-TVI in fact corresponds to the plasma enzyme.

It should also be noted that the activity of Fuc-TVI also resembles that of an a(l,3)fucosyltransferase in lung carci- noma cells (74, reviewed in Ref. 10). This latter enzyme maintains an acceptor substrate specificity that closely resembles that of the plasma enzyme, yet exhibits physical and affinity chromatographic properties distinct from the plasma a(l,3)fucosyltransferase (10). It is interesting to note that the lung enzyme does not form the VIM-2 determinant when tested in vitro with glycolipid acceptor substrates (74), pre- sumably because this enzyme prefers to perform its initial fucosylation reaction using the terminal N-acetylglucosamine moiety on a(2,3)-sialylated polylactosaminoglycans. Thus, terminal fucosylation to form the sialyl Lewis x determinant occurs first and is followed by fucosylation of the internal N- acetylglucosamine to form difucosyl sialyl Lewis x. The reverse reaction, with the potential to form a singly internally fucosylated product corresponding to the VIM-2 moiety, does not occur to any significant degree. This result is analogous to the results with Fuc-TVI reported here, which also does not construct the VIM-2 molecule when expressed in transfected cells that maintain glycoprotein-based (53) VIM-2 precursors. In particular, Fuc-TVI yields high amounts of surface-localized mono- and di-fucosylated sialyl Lewis x determinants but little if any VIM-2 epitopes. These data suggest that Fuc-TVI maintains an ordered fucosylation preference

analogous to that proposed for the lung a(l,3)fucosyltransferase. In any event, further detailed biochemical and molecular genetic studies will be required to confirm this prediction and to determine if the Fuc-TVI gene corresponds to the plasma and/or lung a( 1,3)fucosyltransferaseactiv- ities or to an enzyme expressed elsewhere.

It also remains possible that the plasma and lung a( 1,3)fucosyltransferase activities, and those in other tissues, are encoded by other as yet uncharacterized human a(1,3)fucosyltransferase genes. Such genes may include the sequences identified by the Southern blot analyses reported here that were detected on blots probed at low hybridization stringency with the Fuc-TIV probe. Alternatively, a( 1,3)fucosyltransferase genes may exist whose primary structures are distinct from the a(l,3)fucosyltransferases defined to date and which will require expression cloning approaches, or protein purification methods, for their isolation. Once isolated, this group of molecular reagents should prove useful as tools to define the tissue-specific expression patterns of a(l,3)fucosyltransferase genes, the mechanisms that regulate them, and ultimately, the functions of their oligosaccharide products. They should also prove useful in experiments de- signed to address the amino acid sequence determinants that allow these enzymes to quantitatively distinguish among a variety of oligosaccharide acceptor substrates.

Acknowledgments-We thank Reiji Kannagi and Davor Solter for providing antibody reagents, Mathew Thayer and Mitchell Drumm for providing human chromosome microcell hybrids, and Marlene DenHouter for assistance with Fig. 4.

REFERENCES 1. Bird, J. M., and Kimber, S. J. (1984) Deu. Biol. 104,449-460 2. Fenderson. B. A.. Holmes. E. H.. Fukushi. Y.. and Hakomori. S. (1985)

Deu. Bid. 114,’ 12-21 ’

5565-5569

Chem. 267,14865-14874

Exp. Morphol. 90,337-361

M. J. (1981) Nature 292 , 156-158

, , , , ,

3. Solter, D., and Knowles, B. B. (1978) Proc. Natl. Acud. Sei. U. S. A. 76 ,

4. Kannagi, R., Nudelman, E., Levery, S. B., and Hakomori, S. (1982) J. Biol.

5. Pennington, J. E., Rastan, S., Roelcke, D., and Feizi, T. (1985) J. Embryol.

6. Gooi, H. C., Feizi, T., Kapadia, A., Knowles, B. B., Solter, D., and Evans,

7. Feizi, T., and Childs, R. A. (1987) Biochem. J. 246 , l -11 8. Saitoh, O., Wang, W. C., Lotan, R., and Fukuda, M. (1992) J. Biol. Chem.

267.5700-5711 9. Moore, K. L., Varki, A., and McEver, R. P. (1991) J. Cell Biol. 112 , 491-

~,~ -~ ~ ~~

499 10.

11.

12. 13.

14.

15.

16.

17.

18.

20. 19.

Macher, B. A,, Holmes, E. H., Swiedler, S. J., Stults, C. L. M., and Srnka,

Fox, N., Damjanov, I., Knowles, B. B., and Solter, D. (1983) Cancer Res.

Feizi, T. (1985) Nature 314,53-57 Fukushima, K., Hirota, M., Terasaki, P. I., Wakisaka, A,, Togashi, H.,

Chia, D., Suyama, N., Fukushi, Y., Nudelman, E., and Hakomori, S. (1984) Cancer Res. 44,5279-5285

Sakamoto, J., Furukawa, K., Cordon-Cardo, C., Yin, B. W. T., Rettig, W. J., Oettgen, H. F., Old, L. J., and Lloyd, K. 0. (1986) Cancer Res. 4 6 ,

Kim, Y. S., Itzkowitz, S. H., Yuan, M., Chung, Y., Satake, K., Umeyama, 1553-1561

Muroi, K., Suda, T., Nojiri, H., Ema, H., Amemiya, Y., Miura, Y., Nakauchi, K., and Hakomori, S. (1988) Cancer Res. 48,475-482

Hoff, S. D., Irimura, T., Matsushita, Y., Ota, D. M., Cleary, K. R., and H., Singhal, A., and Hakomori, S. (1992) Blood 79,713-719

Feizi, T. (1991) Trends Biochem. Sci. 16,84-86 Hakomori, S. (1990) Arch. Surg. 126,206-209

Sprin er, T. A., and Lasky, L. A. (1991) Nature 349 , 196-197 Branfley, B. K., Swiedler, S. J., and Robbins, P. W. (1990) Cell 6 3 , 861-

C. A. (1991) Glycobiology 1,577-584

43,669-678

863 21. Springer, T. A. (1990) Nature 346,425-434 22. Lowe, J. B. (1993) Handbook of Immunophurmucolngy: Adhesion Molecules

23 Lowe, J. B., Stoolman, L. M., Nair, R. P., Larsen, R. D., Berbend, T., and (Wegner, C. D., ed) Academic Press, New York, in press

24. Phillips, M. L., Nudelman, E., Gaeta, F. C. A., Perez, M., Singhal, A. K., Marks, R. M. (1990) Cell 63,475-484

25. Walz, G., Aruffo, A., Kolanus, W., Bevilacqua, M., and Seed, B. (1990) Hakomori, S., and Paulson, J. C. (1990) Science 260,1130-1132

26. Goelz, S. E., Hession, C., Goff, D., Griffiths, B., Tizard, R., Newman, B., Science 2 6 0 , 1132-1137

27. Tiemeyer, M., Swiedler, S. J., Ishohara, M., Moreland, M., Schweingruber, Chi-Rosso, G., and Lobb, R. (1990) Cell 6 3 , 1349-1356

H., Hirtzer, P., and Brandley, B. K. (1991) Proc. Natl. Acud. Sci. U. S. A. 88, 1138-1142

28. Larsen. E., Palabrica, T., Sajer, S., Gilbert, G. E., Wagner, D. D., Furie, B. C., and Furie, B. (1990) Cell 63,467-474

24584 Molecular Cloning of a Fourth Human Fucosyltransferase Gene 29. Polley, M. J., Phillips, M. L., Wayner, E., Nudelman, E., Singhal, A. K.,

Hakomori, S., and Paulson, J. C. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,6224-6228

30. Zhou, Q.? MOOR, K. L., Smith, D. F., Varki, A., McEver, R. P., and

31. Berg, E. L., Robinson, M. K., Mansson, O., Butcher, E. C., and Magnani, Cummmgs, R. D. (1991) J. Cell Bzol. 116,557-564

32. Takada, A,, Ohmori, K., Takahashi, N., Tsuyuoka, K., Yago K., Zenita J. L. (1991) J. Biol. Chem. 266, 14869-14872

K., Hasegawa, A., and Kannagi, R. (1991) Biochem. Bioph;. Res. Corn: mcm. 179,713-719

33. Paulson, J. C., and Colley, K. J. (1989) J. Biol. Chem. 264, 17615-17618

35. Tetteroo, P. A. T., de Heij, H. T., Van den El~nden, D. H., Visser, F. J., 34. Lowe, J. B. (1991) Semin. Cell Biol. 2, 289-307..

Schoenmaker, E., and Geurts van Kessel, A. H. M. (1987) J. Diol. Chem.

36. Potvin, B., Kumar, R., Howard, D. R., and Stanley, P. (1990) J. Biol. Chem.

37. Foster, C. S., Gillies, D. R. B., and Glick, M. C. (1991) J. Biol. Chem. 266,

38. Caillard, T., LePendu, J., Ventura, M., Mada, M., Rault, G., Mannoni, P.,

39. Mollicone, R., Gibaud, A., Francois, A., Ratcliffe, M., and Oriol, R. (1990) and Oriol, R. (1988) Exp. Clin. Immunogenet. 6,15-23

40. Mollicone, R., Candelier, J. J., Mennesson, B., Couillin, P., Venot, A. P., Eur. J. Biochem. 191,169-176

41. Couillin, P., Mollicone, R., Grisard, M. C., Gibaud, A,, Ravise, N., Feingold, and Oriol, R. (1992) Carbohydr. Res. 228,265-276

42. Sarnesto, A., Kohlin, T., Hindsgaul, O., Vogele, K., Blaszczyk-Thurin, M., J., and Oriol, R. (1991) Cytogenet. Cell Genet. 66,108-111

43. Ball, S. P., Tongue, N., Gibaud, A,, LePendu, J., Mollicone, R., Gerard, G., and Thurin, J. (1992) J. Biol. Chem. 267,2745-2752

44. Kukowska-Latallo, J. F., Larsen, R. D., Nair, R. P., and Lowe, J. B. (1990) and Oriol, R. (1991) Ann. Hum. Genet. 66,225-233

45. Lowe, J. B., Kukowska-Latallo, J. F., Nair, R. P., Larsen, R. D., Marks, R. Genes & Deu. 4, 1288-1303

M., Macher, B. A., Kelly, R., and Ernst, L. K. (1991) J. Biol. Chem. 266, 17467-17477

46. Kumar, R., Potvin, B., Muller, W. A., and Stanley, P. (1991) J. Diol. Chem. 266. 21777-21783

262,15984-15989

266,1615-1622

3526-3531

47. Weston, B. W., Nair, R. P., Larsen, R. D., and Lowe, J. B. (1992) J.

48. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sei.

" . I ~ ~ . .

Chem. 267,4152-4160

U. S. A. 74.5463-5467 49. Saiki, R. K.,-Scharf, ST,-Faloona, F., Mullis, K. B., Horn, G. T., Erlich, H.

A,, and Arnheim, N. (1985) Science 230,1370-1374

50. Devereux, J., Haeberli, P., and Smithies, 0. (1984) Nucleic Acids Res. 12, 387-395

51. Rajan, V. P., Larsen, R. D., Ajmera, S., Ernst, L. K., and Lowe, J. B. (1989) J. Biol. Chem. 264, 11158-11167

52. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1989) Molecular Cloning: A Moratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Hnrhnr NY

53.

54.

55. 56.

57.

Smith D. F. Larsen R. D. Mattox S. Lowe, J. B., and Cummings, R. D.

Muller, W. J., Naujokas, M. A., and Hassell, J. A. (1984) Mol. Cell. Bid . 4,

Hefferman, M., and Dennis JW. (1991) Nucleic Acids Res. 19,85-92 Fukushi, Y., Kannagi, R., Hakomori, S., Shepard, T., Kulander, B., and

Singer, J. W. (1985) Cancer Res. 46,3711-3717 Larsen, R. D., Ra~an, V. P., Ruff, M. M. Kukowska-Latallo, J. Cummings

8231 R. D., and Lowe, J. B. (1989) Proc. k t l . Acad. Sei. U. S. A. 86, 82271

-------, - . - (1940) J. biol. d m . 266,6225-6234

2406-2412

58. Mellisi S. J., and Baenziger, J. U. (1981) Anal. Biochem. 114,276-280 59. Paulson, J. C., Rearick, J. I., and Hill, R. L. (1977) J. Biol. Chem. 262,

61. Lu 0, T G Handelin B. Killary, A. M., Housman, D. E., and Fournier, 60. Mellis, S. J., and Baenziger, J. U. (1983) Anal. Biochem. 134,442-449

62. Thayer, M. J. and Weintraub, H. (1990) Cell 63, 23-32 63. Collins, F. S.,'Drumm, M. L Cole, J. L. Luckwood W. K. VandeWonde,

64. Ernst L. K., Rajan, V. P Larsen, R. D., Ruff, M. M., and Lowe, J. B. G. F., and Iannuzzi, M. C.'\1987) S c i e k 236,1646-104b

65. Jelinek, W. R., Toomey, T . P., Leinwand, L., Duncan, C. H., Biro, P. V., (19k9) J. Biol. Chem. 264,3436-3447

Choudary, P. V., Weiasman, S. M., Rubin, C. M., Houck, C. M., Dein- lnger, P. L., and Schmld, C. W. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 1398-1402

2363-2371

d. E. i ~ . 71987) Mol.'cejl. ~ i o l . 7,2814-2820

67. Wickner W. T., and Lodish, H. F. (1985) Scierrce 230,400-407 66. Kozak, M. (1989) J. CellBiol. 108,229-241

68. Mount, 6. M. (1982) Nucleic Acids Res. 10,459-472 69. Green, M. R. (1986) Annu. Reu. Genet. 20,671-708 70. Johnson, P. H., Yates, A. D., and Watkins, W. M. (1981) Biochern. Biophys.

71. Prieels, J. P., Monnom, D., Dolmans, M., Beyer, T. A,, and Hill, R. H.

72. Orntoft, T. F., Holmes, E. H., Johnson, P., Hakomori, S., and Clausen, H.

73. Mollicone, R., Caillard, T., LePendu, J., Francois, A., Sansonetti, N.,

74. Holmes, E!? H., Ostrander, G. K., and Hakomori, S. (1986) J. Biol. Chem.

Res. Commun. 100,1611-1618

(1981) J. Biol. Chem. 266,10456-10463

(1991) Blood 77, 1389-1396

Villarro a, H , and Oriol, R. (1988) Blood 71, 1113-1119

261,3737-3743

OF BIOLOGICAL Vol. 34, Issue of 5, pp. Q by The and Inc. U ... · TI11 and 89% identity with Fuc-TV...

Documents

Transcript of OF BIOLOGICAL Vol. 34, Issue of 5, pp. Q by The and Inc. U ... · TI11 and 89% identity with Fuc-TV...