INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

1,3/4-Fucosyltransferases (FucTs)¹ can bind a wide variety of acceptor substrates (see Ref. 1 and references therein) and catalyze the synthesis of glycoconjugates with different immunological and functional properties including blood group antigens and selectin ligands (2-9). Interestingly, the acceptor substrate specificity of different forms of the human FucTs with highly homologous amino acid sequences can differ significantly (1, 10-13). For example, FucT III has a very high level of activity with a type 1 disaccharide acceptor (Gal1,3GlcNAc), whereas FucT VI appears to be a true 1,3-FucT and utilizes exclusively type 2 acceptors (1, 10-15). It is important to point out that different investigators have obtained somewhat different results in terms of the acceptor substrate specificity of FT III and that this enzyme appears to utilize type 2 acceptors when assayed in vitro with high concentrations of simple acceptors and that antibody staining of cells transfected with full-length FT III bind antibodies that recognize type 2 substrates (e.g. anti-Le^x) (e.g. see Ref. 10 and "Discussion" of Ref. 1). Nevertheless, FucTs do have significantly different acceptor substrate specificities when assayed in vitro with disaccharide acceptors despite the fact that they share >90% amino acid sequence homology (10-13). In addition, some of these enzymes can bind the same acceptor substrates but produce different products. For example, mammalian cells transfected with FucT V can produce cell surface antigens recognized by anti-VIM 2 and anti-difucosyl-sLe^x, whereas cells transfected with FucT VI express antigens recognized by anti-difucosyl-sLe^x but not anti-VIM 2 (11, 13). In a previous study, we demonstrated that truncated forms of FucT III and FucT V, containing ~300 amino acids and differing at less than 25 positions, have dramatically different acceptor substrate specificities when assayed with simple disaccharide substrates (16). Furthermore, we found that the swapping of an NH₂-terminal segment of FucT III, containing eight amino acids unique to FucT III, for the corresponding area of the FucT V amino acid sequence produced a protein (see Fig. 1) that could efficiently transfer fucose to both type 1 and type 2 disaccharides (16). In the current study, we have produced a series of FucT V mutants that contain one or more of the eight FucT III amino acid substitutions and characterized them for acceptor substrate specificity. The results demonstrate that a FucT V mutant containing two amino acids characteristic of the FucT III amino acid sequence has significantly higher levels of 1,4-FucT activity compared with FucT V. All of these studies were completed with truncated, soluble forms of the enzymes that are fused to the IgG-binding domain of protein A. This allowed us to utilize highly purified forms of the enzymes for our kinetic analyses. In a previous study (16), we demonstrated that the attachment of the fusion partner does not significantly alter the acceptor substrate specificity of the enzymes.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Materials-- Rabbit IgG-agarose, anti-goat IgG-alkaline phosphatase conjugate, goat IgG, 5-bromo-4-chloro-3-indolyl phosphate, and nitro blue tetrazolium were obtained from Sigma. Fetal calf serum was obtained from Hyclone Laboratories Inc. (Logan, Utah). Trans-blot transfer medium (0.45 µm), kaleidoscope prestained standards, 12% Tris/glycine gels, and Dowex AG 1-X2 resin (200-400-mesh Cl form) were obtained from Bio-Rad. GDP-fucose was kindly provided by Dr. Ole Hindsgaul (Department of Chemistry, University of Alberta) or purchased from Calbiochem. GDP-[³H]fucose (7.57 Ci/mmol) was purchased from NEN Life Science Products. All other reagents were of the highest purity commercially available.

Recombinant Enzymes-- Mutant forms of FucT V were prepared as described previously (17), using a plasmid containing the coding region corresponding to amino acids 76-374 of FucT V as the template. A series of synthetic oligonucleotides were designed to create six FucT V mutant constructs. Mutagenic oligonucleotides were used in PCR reactions to replace specific amino acid(s) of FucT V with the corresponding amino acid(s) in the FucT III amino acid sequence. The primers used were as follows: for the mutant containing Asn⁸⁶  His, Thr⁸⁷  Ile, and Pro⁹²  Ser substitutions, 5'-GGAATTCGCTACTGATCCTGCTGTGGACGTGGCCTTTTCACATACCCGTGGCTCTGTCTAGATGCTCAGA-3' and 5'-CATGATATCCCAGTGGTGCACGAT-3'; for the mutant containing Asn⁸⁶  His and Thr⁸⁷  Ile substitutions, 5'-GGAATTCGCTACTGATCCTGCTGTGGACGTGGCCTTTTCACA TACCCGTGGCTCTGTCTAGA-3' and 5'-CATGATATCCCAGTGGTGCACGAT-3'; for the mutant containing Asn⁸⁶  His substitutions, 5'-GGAATTCGCTATTAATCCTGCTGTGGACGTGGCCTTTTCACA CACCC-3' and 5'-CATGATATCCCAGTGGTGCACGAT/3'; for the mutant containing Thr⁸⁷  Ile substitutions, 5'-GGAATTCGCTATTAATCCTGCTGTGGACGTGGCCTTTTAACATACCC-3'and 5'-CATGATATCCCAGTGGTGCACGAT-3'; for the mutant containing Ala¹⁰¹  Thr, Asn¹⁰⁵  His, Ser¹¹⁰  Arg, Ser¹¹¹  Lys, and Ala¹¹⁸  Thr substitutions (mutant A101T,N105H,S110R,S111K,A118T), 5'-TCCCCGCGGTGCTCAGAGATGGTGCCCGGTACCGCCGAGTGCCACATCACTGCCGACCGCAAGGTGTACC-3' and 5'-CATGATATCCCAGTGGTGCACGATGACCGT-3'; and for the mutant containing Ser¹¹⁰  Arg, Ser¹¹¹  Lys, and Ala¹¹⁸  Thr substitutions, 5'-GGAATTCGCTACTGATCCTGCTGTGGACGT-3' and 5'-CATGATATCCCAGTGGTGCACGATGACCGTGTCTGCCTGTGGGTACACCTTGCGGTCCGCGGTGATGTGG-3'. The PCR product was gel-purified and ligated into the vector, pPROTA (1), and sequences of the inserts were confirmed by sequencing (San Francisco State University Sequencing Facility) both strands of the constructs (results not shown). COS-7 cells were used for enzyme expression. The recombinant enzymes were expressed as soluble proteins with an NH₂-terminal protein A, IgG binding domain. The fusion proteins were purified from the cell culture medium by IgG-agarose column chromatography. The recombinant FucTs were eluted from the IgG-Agarose resin as follows. IgG-agarose beads were briefly suspended in 0.1 M citrate buffer, pH 4.3, and the resulting supernatant was immediately removed and placed in a tube containing neutralizing buffer (1 M bis-Tris-propane, pH 8.0, and 1 mg/ml bovine serum albumin). The enzyme was either used immediately or stored at 4 °C and used within 7-10 days. Enzymes retained full activity during storage under these conditions. Recombinant FucTs were detected and quantified via Western blot analysis.

Western Blot Analysis of Fucosyltransferases-- Recombinant FucTs were detected and quantified via Western blot analysis (16). Proteins were separated on 12% Tris-glycine polyacrylamide gels, transferred to nitrocellulose membranes, and incubated with an IgG-alkaline phosphatase conjugate, and the blot was developed with 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium. Quantitative analysis of the resulting bands was done by scanning the blot with a Hewlett Packard ScanJet II C scanner and analyzing the resulting image with the NIH Image 1.41 program (National Institutes of Health public domain program). Quantification of the amount of FucT was accomplished by comparing the band intensities of samples to those obtained for known quantities of IgG.

Fucosyltransferase Assays-- The standard reaction mixture contained 50 mM MOPS/NaOH buffer, pH 6.5, 6.25 mM MnCl₂, 0.05% bovine serum albumin, 3.0 nmol of GDP-fucose, 0.01 µCi of GDP-[³H]fucose, varying amounts of acceptor, and 5 µl of soluble enzyme. Reactions were terminated by adding 400 µl of water, and the radioactive substrate and product were separated by Dowex anion exchange chromatography and quantified as described previously (16). The kinetic constants given here were determined with the computer program KinetAssyst. The data were fit to the Michaelis-Menten equation by an iterative fitting routine. Assays with glycolipid acceptors were done as previously reported (18).

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Preparation of FucT V Mutants-- Swapping an NH₂-terminal segment of the FucT III catalytic domain, containing eight amino acids unique to FucT III, for the corresponding area of the FucT V amino acid sequence produces a protein (see Fig. 1) that can efficiently transfer fucose to both type 1 and type 2 disaccharides (16) (Fig. 2, Domain Swap). To determine which of the amino acids in this segment contributed the new acceptor substrate specificity, a series of FucT V mutants were prepared (see "Experimental Procedures"). Initially, four FucT V mutants were prepared containing amino acids at the amino (HIS substituted for NTP and HI for NT) or the carboxyl (THRKT substituted for ANSSA and RKT for SSA) terminus of the sequence of the original domain swap. After an initial characterization of these mutants, two single amino acid mutants (His substituted for Asn and Ile for Thr) were prepared and analyzed.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Diagrammatic representation of FucTs III and V and a recombinant FucT produced by domain swapping. The portions of the diagram that are underlined with a single line represent FucT III sequences, whereas those underlined with a double line represent FucT V sequences.

View larger version (36K): [in this window] [in a new window]   Fig. 2.   FucT activity measurements of various FucTs with disaccharide substrates. Assays were done as described under "Experimental Procedures" using a type 1 (white bars) or type 2 (gray bars) disaccharide substrate. The y axis is expressed in percentage of activity with each disaccharide acceptor substrate (type 1 versus type 2). The top of the figure shows the amino acid differences between FucT III and V at the NH2 terminus of their catalytic domain. Domain Swap refers to the recombinant FucT (III8/V see Fig. 1). H, I, HI, HIS, RKT, and THRKT refer to the FucT V mutants described under "Experimental Procedures."

Acceptor Substrate Specificity Analyses-- Time course analyses with type 1 and 2 disaccharide acceptors were conducted with each mutant. All of the mutants catalyzed transfer of fucose to both acceptor substrates but at vastly different rates (see Fig. 2 and Table I). Under the assay conditions used, FucT III has no detectable activity with a type 2 disaccharide, whereas FucT V has very low activity with a type 1 compared with a type 2 acceptor substrate (Fig. 2A). In contrast, the domain swap construct has substantial activity with both type 1 and 2 disaccharides (Fig. 2A). The results indicate that FucT V mutants containing amino acids from the NH₂-terminal region (HIS and HI) of the original domain swap had significant activity with both type 1 and 2 disaccharides (Fig. 2B). In contrast, mutants with amino acids from the COOH-terminal region (THRKT and RKT) of the original domain swap had significant activity with the type 2 disaccharide but very little activity with the type 1 disaccharide (Fig. 2C). Based on these results, two other mutants (H and I) were created. As shown in Fig. 2D and Table I, neither single amino acid FucT V mutant had the acceptor substrate properties of the HI mutant. Thus, the two adjacent amino acids (equivalent to positions His⁷³ and Ile⁷⁴ of FucT III) must be present to give rise to a FucT that can catalyze fucose transfer to a type 1 acceptor substrate at a significant rate.

View larger version (69K): [in this window] [in a new window]   Fig. 3.   Thin layer chromatographic separation of products produced by various FucTs from Lc4 and nLc4. Samples were visualized by autoradiography of the radioactive products formed. Odd numbered lanes are products formed with Lc4, and even numbered lanes are products formed with nLc4. Products formed with FucT III (lanes 1 and 2), FucT V (lanes 3 and 4), FucT V HIS mutant (lanes 5 and 6), FucT V HI mutant (lanes 7 and 8), FucT V I mutant (lanes 9 and 10) are shown. Results shown in lanes 5 and 6 were from a different autoradiography from those shown in other lanes, accounting for differences in band intensity. The amount of product formed in the reactions shown in lanes 5 and 6 was nearly identical to that shown in lanes 7 and 8.

Kinetic Analyses-- Comparison of the kinetic properties of the FucT V mutants and the parent enzyme resulted in GDP-fucose saturation curves for each of the enzymes that were essentially identical, resulting in K_m values for each of 35 ± 5 µM (Fig. 4).

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Vo versus [GDP-fucose] plot for various FucTs. Assays were done with a type 2 disaccharide as the acceptor substrate (60 mM). , FucT V; , FucT V HIS; , FucT V HI.

N-Linked Glycosylation Sites in FucTs-- The predicted amino acid sequences for FucT III and V contain two and four potential N-linked glycosylation sites, respectively (10, 11). Two of these sites are conserved (Asn¹⁶⁷ and Asn¹⁹⁸ in FucT V; Asn¹⁵⁴ and Asn¹⁸⁵ in FucT III). The other two potential N-linked glycosylation sites exist at Asn⁶⁰ and Asn¹⁰⁵ of FucT V. Asn⁶⁰ occurs outside of the catalytic domain of FucT V (as would the corresponding Ser in FucT III), and Asn¹⁰⁵ of FucT V is substituted by a His residue in FucT III (His⁹²). Lowe and co-workers (10) have demonstrated that when FucT III is expressed in an in vitro translation system it is glycosylated with endoglycosidase H-sensitive carbohydrate residues. Their results suggest that both potential N-linked glycosylation sites are utilized. Similar information for FucT V is unavailable. One of the FucT V mutants (THRKT) does not contain the potential N-linked site at residue 105 of FucT V due to a His substitution at this site. This provided an opportunity to determine if this substitution affected the overall molecular weight of the FucT V mutant compared with FucT V. Plasmids containing the FucT V and the THRKT mutant (both as protein A chimeras) were transfected into COS cells, and the resulting proteins were purified by IgG-agarose affinity chromatography and analyzed by Western blot. No significant difference in the relative mobility of the wild-type and mutant proteins occurred, suggesting that Asn¹⁰⁵ of FucT V is not glycosylated in the COS cell expression system (data not shown).

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

The initial cloning of FucTs III and V demonstrated that these two enzymes shared substantial amino acid sequence homology yet differed in their acceptor substrate specificities (10, 11). Subsequent cloning of a third, highly homologous FucT (i.e. FucT VI) demonstrated a similar relationship (12, 13). Thus, each of these enzymes has distinct acceptor substrate specificities although they have remarkably similar predicted amino acid sequences. Our previous work has focused on defining the minimal catalytic domains of FucT enzymes (16). Our characterization of the catalytic domains of FucTs III and V demonstrated that these distinct acceptor substrate specificities are due to differences occurring at 23 out of ~300 amino acids (16). Most notably, FucT III was found to be essentially an 1,4-FucT, whereas FucT V is much more active with type 2 than type 1 acceptor substrates. Using a domain swapping approach, we (16) were able to demonstrate that the incorporation of as few as eight amino acids unique to the NH₂ terminus of the catalytic domain of FucT III into FucT V produced a FucT with both type 1 and type 2 acceptor specificity (Fig. 1). In the current study, we have shown that two of these eight amino acids can account for the enhanced type 1 acceptor substrate specificity. It is interesting to note that these two amino acids lie next to each other (i.e. His⁷³ and Ile⁷⁴) in the FucT III sequence. Substitution of only one of these two amino acids (His or Ile) into the FucT V sequence is insufficient to incorporate the new acceptor substrate specificity into the resulting mutant. Thus, in the FucT V sequence Asn (the normal amino acid in FucT V) cannot substitute for His even when the other site is an Ile residue, and Thr (the normal amino acid in FucT V) cannot substitute for Ile even when the other site is a His residue. Based on these results, we can conclude that a polar, uncharged amino acid substitution at either position is insufficient to produce the enhanced acceptor substrate specificity obtained with the FucT V HI mutant. Additional mutational studies will be required to determine if a specific or general (e.g. negatively charged or neutral) amino acid side chain substitution at these sites is required for the new acceptor substrate specificity.

Only one other study has been published in which the acceptor specificity of FucTs has been evaluated with respect to the enzymes' amino acid sequence. Legault et al. (19) used a domain swap approach with full-length forms of FucTs III and VI to identify a discrete sequence of amino acids that distinguishes acceptor specificity (i.e. type 1 versus type 2). Based on an extensive set of experiments in which various coding sequences of FucTs III and VI were swapped, Legault et al. (19) concluded that segments encoding amino acid differences between residues at 105-151 of FucT III and residues 104-150 of FucT VI (referred to as subdomains 4 and 5) influence type 1 acceptor specificity. Thus, the substitution of subdomains 4 and 5 of FucT VI for the corresponding regions in FucT III eliminated type 1 acceptor specificity, whereas the complementary construct (i.e. subdomains 4 and 5 of FucT III substituted into the FucT VI coding region) produced an enzyme with both type 1 and 2 acceptor specificity. However, it should be noted that the latter chimera had very poor activity compared with the wild-type enzymes. Since these studies utilized FucT VI in combination with FucT III sequences, it is difficult to directly compare them to our results. However, Legault et al. (19) did report on the fucosylated products formed by a single FucT III/FucT V chimera (designated Fuc-TC21) they produced in which subdomains 4 and 5 of FucT III were substituted with subdomains 4 and 5 of FucT V. This construct is very similar to the FucT domain swap we produced (III⁸/V Fig. 1), which has both type 1 and 2 acceptor substrate specificity (Table I and Fig. 2A). The only difference is a single amino acid substitution in the catalytic domains of Fuc-TC21 and III⁸/V (i.e. Fuc-TC21 contains Ala¹¹⁸ of FucT V, whereas III⁸/V contains a Thr residue). When the acceptor substrate specificity (measured as the cell surface expression of fucosylated antigens) of Fuc-TC21 was characterized, it was found that it could utilize both type 1 and 2 acceptors (Fig. 4a of Ref. 19). In contrast, FucT V was shown to have only type 2 specificity when transfected into COS cells (Fig. 2 of Ref. 11). Thus, the incorporation of the equivalent of subdomains 1-3 of FucT III into FucT V produced an enzyme with both type 1 and 2 acceptor specificity. Subdomain 2 of FucT III contains the amino acids (i.e. HI) that we have demonstrated in this study to be important for type 1 acceptor specificity.

Although the substitution of FucT III amino acids (i.e. HI or HIS) in place of the corresponding amino acids of FucT V produces enzymes capable of catalyzing the transfer of fucose to a type 1 acceptor, the K_m value(s) of the resulting proteins for a type 1 acceptor are not equivalent to that obtained for FucT III. The lower affinity for a type 1 substrate indicates that other amino acids in the FucT III sequence contribute to the overall binding properties of FucT III. Interestingly, the K_cat is similar for FucT III and the FucT V mutants indicating that, although the binding affinity of the mutants for a type 1 substrate is lower for the mutants, the limiting reaction(s) occurs at a similar rate.

It is interesting that the ratio (i.e. type 1 versus type 2) of product formation from disaccharides versus glycolipids is different for the FucT V mutants (Table I). This may be a reflection of differences in the in vitro assay conditions or the substrate structure. For example, the glycolipid reaction mixture requires the addition of a detergent to solubilize the acceptor substrates, whereas a solubilizing agent is not required for disaccharide substrates. It is clear from previous studies that detergents can alter product formation by FucTs (20). Further studies with different detergents will be necessary to determine the affect they have on product formation. It is also possible that the additional structure (i.e. LacCer) at the nonreducing end of the glycolipids could influence product formation. Regardless of the reason for the difference in product ratio, the important observation is that the FucT V HI and HIS mutants have a substantially higher activity with a type 1 acceptor (disaccharide or glycolipid) than FucT V. This demonstrates that the acceptor substrate specificity of FucTs can be altered by changing as few as two amino acids at the NH₂ terminus of the catalytic domain.

The results presented in this study are the first to pinpoint the location of amino acids associated with the binding of acceptor substrates by FucTs. It is interesting that they lie close (10 and 11 amino acids away) to the NH₂ terminus of the catalytic domain, which we have shown begins at amino acid residue 76 of FucT V. In an accompanying article (21), we have shown that an amino acid that lies near the COOH terminus of FucT V alters the enzyme's binding properties for an H-type 1 acceptor substrate. Thus, amino acids occurring at the two extremes of the catalytic domain of FucT V affect acceptor substrate binding. This finding correlates well with the location of amino acids (residues 68 and 356 in FucT III) that are known to be essential for FucT catalytic activity (22-26). Two other naturally occurring, inactivating mutants have been located at a more central position, residues 146 and 170 of FucT III, equivalent to residues 159 and 183 of FucT V, respectively. Interestingly, these two amino acids lie in close proximity to Cys¹⁵⁶ of FucT V, which we have found to affect GDP-fucose binding (27). Thus, there is a close correlation between these amino acids that have been found to be catalytically important and those that affect substrate binding.

In addition to these amino acid residues, another catalytically essential amino acid has been identified in FucT VI. Mollicone et al. (28) have shown that Glu²⁴⁷ of FucT VI, when mutated to a Lys residue, leads to an inactive enzyme. Sequence alignments indicate that this residue occurs at a position in FucT VI that is within six amino acid residues of Lys²⁵⁵ of FucT V, a residue identified in an accompanying article (29) to be catalytically essential in FucT V. As pointed out under "Discussion" in that article, these amino acids occur within the highly homologous amino acid sequence designated the "1,3-FucT motif" (30, 31). Thus, several segments of the FucT amino acid sequence have been shown to be important for substrate binding and catalytic activity.

In the domain swap study reported by Legault et al. (19), those authors created recombinant FucTs (designated Fuc-TC7 and Fuc-TC12) that were only altered by a single amino acid change, which resulted in the substitution of a His (found in FucT III at residue 92) for an Asn residue (found in FucT VI at residue 91). These mutants were identified by Legault et al. (19) to have a swap of subdomain 3, and the mutation introduced would alter one of the four potential N-linked glycosylation sites of FucT VI (12, 13). They reported that swapping subdomain 3 did not affect acceptor substrate specificity. FucT V also contains this potential N-linked site (11), and our FucT V THRKT mutant would contain the His residue from FucT III at this site, eliminating the potential N-linked glycosylation site. As shown in Fig. 2 and Table I, this mutant's acceptor substrate specificity was not altered compared with FucT V. Interestingly, there was no indication that this site was glycosylated by COS cells. Oulmouden et al. (32) have recently cloned a bovine FucT with significant sequence homology to human FucTs, which may be an ancestor of the human FucT III-VI enzymes. It is interesting to note that this enzyme does not contain the potential N-linked glycosylation site found at amino acid residues 105 and 91 of FucT V and VI, respectively. Therefore, this site may not serve as a glycosylation site in the FucT V and VI enzymes. Further analyses are required to determine which of the potential N-linked glycosylation sites are utilized. Recently, Nagai et al. (33), Martina et al. (34), and Toki et al. (35) have demonstrated that the N-linked chains of N-acetylglucosaminyltransferase III, GD₃ synthase, and core 2 N-acetylglucosaminyltransferase are required for enzyme activity and subcellular localization of these enzymes. Thus, it will be important to identify the N-linked glycosylation sites of FucTs and determine their function in enzyme activity and subcellular localization of this class of glycosyltransferases.

Legault et al. (19) also pointed out that FucT VI contains 11 unique amino acids, occurring in the region of the sequence they designated subdomains 4 and 5, which may contribute to a type 2 acceptor substrate specificity. Since FucT V can also catalyze the fucosylation of type 2 acceptors efficiently, amino acids in its sequence must also contribute to its type 2 acceptor substrate binding site. Based on the results of our previous domain swap experiments (16), it is unlikely that amino acids affecting type 2 substrate binding would be found near the NH₂ terminus of the FucT catalytic domain. For example, we have found that the domain swap V⁸/III (FucT V-(76-123)/FucT III-(111-361)), which has an NH₂ terminus containing eight amino acids found in FucT V attached to the remaining sequence of FucT III, does not catalyze the fucosylation of a type 2 disaccharide (16). In addition, the domain swap III⁸/V (FucT III-(62-110)/FucT V-(124-374)) is still capable of catalyzing the fucosylation of a type 2 disaccharide (16). Therefore, it seems likely that amino acids that affect type 2 substrate binding will be found among the 15 residues that differ between the FucT III and V sequence that occur after amino acid 124 of the FucT V sequence (similar to subdomains 4 and 5 of Legault et al. (19)). Demonstration of this possibility awaits further mutational studies.