Human α1,3/4-Fucosyltransferases

Amino acid sequence alignment of human α1,3/4-fucosyltransferases (FucTs) demonstrates that three highly conserved Lys residues are present in the catalytic domain of FucTs III, IV, V, and VI. Two of these sites are conserved in FucT VII, with the third located within the α1,3-FucT motif as a conservative change to Arg at position 223. Site-directed mutagenesis experiments were conducted to change Lys255 of FucT V (equivalent to Arg223 of FucT VII) to either Arg255 or Ala255. Enzyme assays demonstrate that the FucT V K255R mutant has a 34-fold lower specific activity than native FucT V and that the K255A mutant is inactive. Site-directed mutagenesis of FucT VII was also conducted to change Arg223 to Lys223 for analysis of the effect on enzyme kinetic parameters. No differences in acceptor specificities orK m values for either substrate were observed between native FucT VII and the R223K mutant; however, the purified R223K mutant enzyme had a 2-fold increased specific activity compared with purified native FucT VII. No change in GDP-fucose-protectable pyridoxal-P/NaBH4 inactivation was observed for native or mutant FucT V or VII, further supporting the absence of involvement of this residue in sugar nucleotide binding. The results indicate that a basic residue in this position is required for enzyme activity, with a Lys residue providing higher intrinsic activity. The lack of influence of this site on substrate binding parameters and its location within the α1,3-FucT motif suggest that at least some of the residues within this motif are involved in catalysis rather than substrate binding.

12, and 12 Lys residues, respectively. In contrast, FucTs VII and IV contain only two and five Lys residues, respectively. Sequence alignment analysis of these forms demonstrate that nine of the Lys residues present in FucT III, V, and VI are in equivalent positions. Of these, three Lys residues can be aligned with Lys residues found in the FucT III, IV, V, and VI sequences. The FucT VII sequence contains two of these Lys residues, and a conservative substitution (Arg 223 ) occurs at the third site. The sites corresponding to Arg 223 and Lys 240 of FucT VII are located in the so-called ␣1,3-FucT motif of amino acids that are highly conserved among FucT enzymes (15,16). This distribution of Lys/Arg residues in human FucTs is shown in Fig. 1. The highly conserved nature of these three amino acid sites strongly suggests that one or more are critical for enzymatic activity.
We have previously shown that at least one catalytically essential Lys residue is present in human FucTs (17) and is protected from PLP/NaBH 4 inactivation by GDP or GDP-fucose. We have initiated a series of site-directed mutagenesis studies to investigate the importance of the Lys (or Arg) residues that are highly conserved among the FucTs. Initial mutagenesis studies were done to evaluate Lys 255 in FucT V, which corresponds to Arg 223 of FucT VII, changing it to an Arg or Ala residue. Further site-directed mutagenesis experiments were conducted to convert Arg 223 to Lys 223 in FucT VII. The results of these investigations demonstrate that this Lys/Arg residue is critical for FucT activity and that even conservative amino acid substitutions significantly affect enzyme activity.
In the first step, a specific mutation was introduced by using a pair of mutagenic primers and a pair of opposite end flanking primers. For the K255R mutant, for example, one PCR mixture included the B14 and V3K255R primers to generate the N-terminal half of the "mutated" FucT V; another PCR mixture included V5K255R and C6 primers to generate the C-terminal half of the "mutated" FucT V. The PCR products were gel-purified and used together with both flanking primers for the second-step PCR mixture. The resulting PCR product was a 916base pair DNA fragment encoding the truncated FucT V (Leu 76 -Thr 374 ). The 916-base pair DNA fragment was then purified, cut with EcoRI, and subcloned into pPROTA. Bacterial clones containing plasmids in the correct orientation were screened with HindIII and StuI double digestions. The mutation was confirmed by DNA sequencing using an Applied Biosystems model 373A DNA sequencing system.
The PCR product from the second step PCR mixture was briefly treated with mung bean nuclease and cloned into the EcoRV site of pZErO-1 plasmid, and the mutation was confirmed by DNA sequencing using the Sequenase version 2.0 sequencing system. The insert containing the desired mutation was then excised with EcoRI and cloned into pPROTA as described above for FucT V mutants. The correct orientation was established by HindIII/MscI double digestion. Plasmids were transiently transfected into COS-7 cells grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum using the DEAE-dextran method (20), and the expressed protein was isolated on IgG-agarose beads as described previously (21).
Full-length FucT VII-The pCDM8 FucT VII plasmid contains the full coding sequence of FucT VII plus approximately 0.78 kilobase pair of 3Ј-untranslated region. A NotI cassette of approximately 1.5 kilobase pairs, encompassing NotI (nucleotide 329) within the FucT VII gene through the end of the 3Ј-untranslated region was excised from parental pCDM8-FucT VII and replaced with a 0.718-kilobase pair NotI cassette from pZErO-1-FucT VII-R223K mutant. This cassette represents the same coding portion of the FucT VII gene (including the R223K mutation), minus the 3Ј-untranslated sequence. Orientation was established via EcoRI/MscI digestion. The resultant plasmid, pCDM8-FucT VII-R223K, along with native pCDM8-FucT VII, was subsequently used in transfections using the DEAE-dextran method (20) and harvested after 5 days.
Enzyme Assays-FucT activity using glycolipid acceptors was determined in reaction mixtures containing 2.5 mol of HEPES buffer, pH 7.2, 1 mol of MnCl 2 , 100 g of taurodeoxycholic acid, 20 g IV 3 NeuAcnLc 4 or VI 3 NeuAcnLc 6 , 15 nmol of GDP-[ 14 C]fucose (15,000 cpm/nmol), and 300 -400 g of protein in a total volume of 0.1 ml. Cells expressing full-length enzymes were homogenized by two strokes of a Potter-Elvehjem homogenizer in a buffer composed of 50 mM HEPES, pH 7.2, containing 25% glycerol. The reaction mixture was incubated at room temperature for 2 h, terminated by the addition of 0.1 ml of CHCl 3 /CH 3 OH, 2:1, and quantitated as described previously (22).
FucT enzyme activities utilizing oligosaccharide acceptors were determined in reaction mixtures composed of 1 mol of HEPES buffer, pH 7.2, 6 nmol of GDP-[ 14 C]fucose (15,000 cpm/nmol), 0.4 mol of LacNAc or 0.072 mol of 8-methoxycarbonyloctyl glycoside derivatives, 2 mol of NaCl, 0.125 mol of MnCl 2 , 10 g of bovine serum albumin, 0.01 mol of ATP, and chimeric enzyme bound to IgG-agarose beads in a total volume of 0.02 ml. The reaction mixture was incubated for 1 h at room temperature and stopped and quantitated as described previously utilizing Dowex-1 for oligosaccharide acceptors (26) or C 18 columns for 8-methoxycarbonyl glycoside derivatives as acceptors (26).
Western Blot Analysis of pPROTA-expressed Enzymes-The pPROTAexpressed recombinant FucT enzymes were separated on 12% Tris/ glycine polyacrylamide gels, transferred to nitrocellulose membranes, and probed as described in an accompanying paper (34).
PLP Inactivation of Native and Mutant Enzymes-0.5 g of PLP, was initially purified by Dowex-1 chromatography with elution by an acetic   PLP treatment of enzymes expressed as pPROTA fusion proteins bound to IgG-agarose beads were conducted in reaction mixtures containing 5 l of packed enzyme beads in a phosphate-buffered saline (8.1 mM Na 2 HPO 4 , 1.5 mM KH 2 PO 4 , 137 mM NaCl, 2.7 mM KCl, pH 7.4) slurry and PLP varying from 0.5 to 2 mM final concentration in a total volume of 50 l. In some reaction mixtures, GDP-fucose was added at a final concentration of 300 M. After incubation for 15 min at room temperature, a 15-fold excess of NaBH 4 dissolved in water was added, and the reaction mixtures were incubated for an additional 15 min. The enzyme activity remaining was then determined using assay conditions described above.

RESULTS
Site-directed Mutagenesis-Site-directed mutagenesis of FucT V was performed by replacing the sequence for codon 255 (AAG) with AGG for Arg or GCG for Ala and with FucT VII to change the sequence for codon 223 (CGC) to AAA for Arg. Sequencing on both strands of all mutants confirmed that there were no other nucleotide modifications compared with the respective native enzyme (results not shown).
Effect of the K255R and K255A Mutations on FucT V Activity- Table I shows the effect of the K255R and K255A mutations of FucT V on enzyme specific activity. For this comparison, the enzymes were expressed as soluble fusion proteins of the catalytic domain of the enzymes with the protein A Ig binding domain using the pPROTA vector. These enzyme forms have the important advantage of allowing convenient protein isolation and protein determinations needed for accurate specific activity measurements of the purified enzyme. Replacement of the Lys residue by Arg reduced the enzyme's specific activity 34-fold compared with the native enzyme. Replacement by an Ala residue inactivated the enzyme. Western blot analysis demonstrated that each mutant was expressed at levels comparable with that of the wild type enzyme and that each produced a single band with a molecular weight corresponding to that of the wild type enzyme (results not shown). The dramatic decrease in specific activity for the K255R mutant precluded a kinetic analysis of this enzyme to accurately determine the nature of the changes in kinetic parameters caused by this mutation. To test kinetic parameters and to determine if a Lys residue in this site was generally optimal for activity, an Arg to Lys mutant at the corresponding site of FucT VII (Arg 223 ) was prepared for a kinetic evaluation.
Effect of the R223K Mutation on FucT VII Substrate Specificity and Activity-An analysis of acceptor specificity and enzyme-specific activity of native FucT VII and the R223K mutant was conducted by expression of the catalytic domains for both enzymes in the pPROTA vector. The results shown in Table II confirm the requirement of a terminal sialyl substitution on the acceptor for optimal activity for both the native FucT VII and the R223K mutant. Product formation occurred with 3Ј-sulfo-LacNAc at a rate 20 -25% of that found with 3Ј-sialyl-LacNAc. This demonstrates the importance of a negative charge at the nonreducing end of the acceptor for activity. The FucT VII R223K enzyme form had 2-fold higher specific activity than native FucT VII when assayed with sialyl-Lac-NAc-R 1 acceptor. Overall, the results demonstrate that the change of Arg 223 to Lys 223 in FucT VII had no effect on acceptor specificity but had a substantial effect on the intrinsic activity of the enzyme in a manner predicted by the results with native FucT V and its K255R mutant.
Transfer of fucose into multiple glycolipid acceptors was also tested with full-length forms of the native FucT VII and R223K mutant enzymes expressed using the pCDM8 vector in COS-7 cells. Fig. 2 shows results from transfer of labeled fucose into glycolipid acceptors using crude homogenates of COS-7 cells expressing the full-length enzymes. Fucose transfer was only observed into sialylated acceptors IV 3 NeuAcnLcOse 4 Cer and VI 3 NeuAcnLcOse 6 Cer. No detectable product was observed with nLcOse 4 Cer as an acceptor or with mock-transfected COS-7 cell extracts (results not shown). Qualitatively identical results were obtained with both the native and R223K enzymes. However, transfer was always much greater with homogenates of COS-7 cells transfected with cDNA encoding the full-length R223K enzyme compared with native FucT VII, consistent with the observation from the pPROTA-expressed enzyme forms that the R223K mutant was an inherently more active enzyme. Only monofucosylated products were obtained with each enzyme, with transfer to VI 3 NeuAcnLcOse 6 Cer (lanes 3 and 4) yielding exclusively the product VI 3 NeuAc-V 3 FucnLcOse 6 Cer (results not shown; see Ref. 22).
Kinetics of Native FucT VII and the R223K Mutant-The increased specific activity of the FucT VII R223K mutant compared with the native form made it possible to conduct a kinetic analysis of these enzymes to determine if the increased specific activity of the R223K enzyme was due to a K m or V max effect on the enzyme. Saturation kinetics with both donor and acceptor substrates was studied for both pPROTA-expressed enzyme forms (Fig. 3). No differences in the K m for GDP-fucose was found for the mutant versus wild type enzyme, with observed K m values of 9.2 Ϯ 0.3 and 10.8 Ϯ 1.4 M for the native FucT VII and R223K enzyme, respectively. Similarly, the K m for sialyl-LacNAc-R 1 was the same for both enzymes at 0.70 Ϯ 0.02 and 0.70 Ϯ 0.05 mM for the native FucT VII and R223K enzymes, respectively. Thus, the increased specific activity of the R223K mutant is a V max effect and indicates that a Lys residue in this position creates an inherently more active enzyme.
PLP Inactivation of Native and Mutant Enzymes-The presence of a PLP/NaBH 4 -sensitive Lys residue that lies in or near the GDP-fucose binding site has previously been demonstrated (17). To determine if Lys 255 of FucT V is this residue, a series of PLP inactivation studies were conducted with pPROTAexpressed mutant FucT V and VII enzymes. These enzymes were treated with increasing concentrations of PLP followed by NaBH 4 reduction to determine if the Lys group in this position was associated with GDP-fucose-protectable enzyme inactivation. The results are shown in Fig. 4. Panels A-C show the GDP-fucose-protectable inactivation of FucTs III, IV, and VI as positive controls confirming the general property of PLP/ NaBH 4 sensitivity among human FucTs. Panels D and E demonstrate that both the native and K255R FucT V and native and R223K FucT VII enzymes, respectively, are inactivated by PLP/NaBH 4 to a similar extent. The extent of the inactivation of the FucT VII enzymes was less than that seen for the expressed pPROTA constructs of the other FucT enzymes at the same PLP concentrations (panels A-D), suggesting that the native and R223K FucT VII enzymes have a lower affinity for PLP compared with other FucTs. Analysis of the K i for PLP for both the native and R223K FucT VII enzymes demonstrate that each enzyme had a K i of approximately 2 mM (results not shown). This compares to a K i for PLP of 105 M previously reported for the enzyme derived from NCI-H69 cells (17). Thus, the observed inactivation results demonstrate that PLP inactivation of native and R223K FucT VII enzymes occurs over a wider PLP concentration range. Fig. 4 also demonstrates that PLP/NaBH 4 inactivation of all enzymes tested is protectable by GDP-fucose present at 0.3 mM final concentration. DISCUSSION Genes encoding ␣1,3-FucT enzymes have been cloned from a variety of species (8 -16), including humans (1-7). Alignments of the deduced amino acid sequence of these enzymes demonstrates a substantial sequence homology, primarily within the C-terminal catalytic domain region. The multiplicity of enzymes that exists, for example in humans, probably accounts for subtle differences in enzyme function in vivo and gives rise to differing acceptor properties and transfer specificities. An increasing amount of information is becoming available to show how specific amino acid residues within the more variable regions of the catalytic domain of these proteins influence specific properties of a given enzyme form (Refs. 17, 21, and 23-32; see also accompanying papers (34,36)). The results presented in this paper address a region of the catalytic domain that is instead rather highly conserved among enzyme forms.
Although variable numbers of Lys residues are present among human FucTs, highly conserved sites exist. In particular, three Lys residues are in identical positions in FucTs III, IV, V, and VI. Two of these positions are occupied by the only Lys residues present in the FucT VII enzyme. The third site is a conservative change to an Arg residue (Arg 223 ). The results presented in this paper focus on enzymatic properties inherent to this Lys/Arg site among human FucT enzymes.
Site-directed mutagenesis experiments changing the equivalent Lys 255 of FucT V to an Arg residue (K255R mutant) resulted in an enzyme with dramatically reduced specific activity compared with the native enzyme. Changing this site to a nonconservative Ala residue (K255A mutant) abolished activity. Thus, a basic group in this position is required for activity. The poor activity of the resulting K255R mutant required an alternate approach to accurately determine the kinetic effect of this type of substitution on the enzyme. Since native FucT VII has an Arg residue in this position, changing it to a Lys residue and analyzing its affect was deemed to be an ideal approach. Presumably, a more active enzyme would result, allowing kinetic comparisons of both the native and mutant enzymes.
Analysis of native FucT VII and the R223K mutant confirmed that this substitution had no effect on acceptor specificity properties of the enzyme, including the enzyme's behavior with sialylated acceptors containing multiple GlcNAc residues. Further, this amino acid change had no effect on the K m for either the donor GDP-fucose or the acceptor oligosaccharide. The only effect caused by this amino acid change was to create an enzyme with inherently greater activity. Taken together, the results strongly suggest that this site does not directly function in substrate binding; however, a basic group in this site is required for activity. Enzymes containing a Lys residue in this position are inherently more active compared with those with an Arg residue, suggesting the ionization state of the side chain is significant in maintaining optimal enzyme conformation or may be involved in catalysis. This is consistent with a previous report that suggested that a basic group on human FucT enzymes (i.e. FucT III and VI) is involved in a critical hydrogen bond donor interaction with reactive acceptor hydroxyl groups during the glycosyl transfer reaction (33).
Recently, cloning of FucT enzymes from H. pylori and a comparison of aligned sequences from a wide variety of species have led to the identification of an "␣1,3-FucT motif" (15,16). This motif is composed of certain fully conserved amino acid residues within a portion of the catalytic domain between residues Tyr 254 and Lys 272 of human FucT V and between residues Tyr 222 and Lys 240 of FucT VII. Interestingly, the Lys/Arg site described in this paper occurs within this segment of the catalytic domain. The behavior of the native and mutant enzymes described in this paper suggests that this motif may have a more complex function in FucT enzymes by directly participating in catalysis, in addition to a possible involvement in binding GDP-fucose or Mn 2ϩ (15). A more complete evaluation of the role of this ␣1,3-FucT motif is required before firm conclusions can be drawn relating to the functions of the differing amino acids that are present.
Previous results using enzyme derived from NCI-H69 cells demonstrated that PLP is a competitive inhibitor with respect to GDP-fucose and, further, that a PLP/NaBH 4 -sensitive, GDPfucose-protectable Lys residue is present (17). The present results confirm that this is a common property of FucT enzymes. Thus, it is reasonable to conclude that this is a property of at least one of the highly conserved Lys residues present in FucTs. To determine if FucT VII was also PLP/NaBH 4 -sensitive and if the site corresponding to Lys 255 of FucT V is the reactive site, PLP/NaBH 4 inactivation studies were conducted on the native and mutant enzymes. The results confirmed that FucT VII was also inactivated by PLP/NaBH 4 , although with a lower efficiency than other enzyme forms. The K i for PLP inhibition with respect to GDP-fucose was determined to be 2 mM, compared with 105 M reported previously (17). Comparisons of the results from both FucT V and VII native and mutant enzymes demonstrate that there was no difference in the GDP-fucose-protectable PLP/NaBH 4 inactivation regardless of whether a Lys or Arg residue occupied this amino acid position. Thus, this position does not correspond to the PLP/ NaBH 4 -sensitive site. This is consistent with the results from a kinetic analysis of the native and mutant FucT VII enzymes, which demonstrate that the change from an Arg to Lys had no affect on the K m for GDP-fucose.