Two new FUT2 (fucosyltransferase 2 gene) missense polymorphisms, 739G→A and 839T→C, are partly responsible for non-secretor status in a Caucasian population from Northern Portugal

Abstract

Secretor status is defined by the expression of H type 1 antigen on gastric surface epithelium and external secretions. The H type 1 structure, and other fucosylated carbohydrates (Le^a, sialyl-Le^a, Le^b, Le^x, sialyl-Le^x and Le^y), can serve as ligands for several pathogens, including Helicobacter pylori, and are cancer-associated antigens. Secretor individuals are more susceptible to some bacterial and viral infections of the genito-urinary and digestive tracts. The aim of the present study was to examine FUT2 (fucosyltransferase 2 gene) polymorphisms in a Caucasian population of non-secretor individuals (n=36) from northern Portugal and to evaluate the activity of the mutant FUT2 enzymes. The secretor status was determined by UEAI [Ulex europaeus (gorse) lectin] histochemistry in gastric mucosa, and FUT2 polymorphisms were studied by restriction-fragment-length polymorphism and direct sequencing. The majority of non-secretors (88.9%) were homozygous for 428G→A polymorphism; 5.6% were homozygous for 571C→T and 5.6% were homozygous for two new missense polymorphisms, 739G→A (2.8%) and 839T→C (2.8%). By kinetic studies it was demonstrated that the two new FUT2 mutants (739G→A and 839T→C) are almost inactive and are responsible for some non-secretor cases.

INTRODUCTION

The secretor status is defined by the presence of H type 1 antigen in body secretions such as milk and saliva [1]. H type 1 antigen belongs to both the Lewis and the ABO(H) histo-blood-group systems and is expressed in erythrocyte membranes and in several epithelial tissues, namely the gastric mucosa, the upper respiratory tract and the lower genito-urinary tract [1,2]. Although the synthesis of H type 1 antigen is dependent on the sequential action of several glycosyltransferases, the secretor enzyme (FUT2), an α-1,2-fucosyltransferase, is responsible for the transfer of fucose in an α-1,2 linkage to form the terminal H type 1 structure [3–6]. The cell-surface fucosylated oligosacharides participate in several biological processes, such as embryogenesis, tissue differentiation, tumour metastasis, inflammation and bacterial adhesion [7].

Various pathogens, such as Helicobacter pylori, Escherichia coli and Norwalk virus, have been shown to attach themselves to either α1,2-fucosylated glycans or to precursor structures masked by fucosylation [8–10] on epithelial cells. Accordingly, several studies pointed out either the secretor or the non-secretor status as risk factors for pathogens that bind to fucosylated or precursor glycans respectively [1,11–16].

FUT2 gene polymorphisms have been described in several worldwide human populations, and some of them are responsible for non-secretor status [2,17–21]. The aim of the present study was to examine FUT2 polymorphisms in a Caucasian population of non-secretor individuals from Northern Portugal and to evaluate the implications of the mutant FUT2 enzymes.

MATERIALS AND METHODS

Population

In the present study we evaluated 36 asymptomatic or dyspeptic Caucasian individuals from northern Portugal [22]. All individuals were non-secretors, determined by UEAI [Ulex europaeus (gorse) lectin] staining in gastric epithelium. The screening of FUT2 428G→A polymorphism was made by RFLP (restriction-fragment-length polymorphism). Individuals that were not homozygous for FUT2 428G→A polymorphism were sequenced for the whole coding region of the FUT2 gene. All individuals gave their informed consert, and the study had the ethical approval of the institutions involved.

PCR-RFLP

PCR was performed using DNA extracted from blood cells as a template. The amplification of a 195 bp FUT2 fragment was performed by using the forward primer: 5′-GAGGAATACCGCCACATCCCGGGGGAGTAC-3′ and the reverse primer: 5′-AGCCGGCCGGGCACCTTTGTAGGGGTCCAT-3′. The 25 μl final reaction mixture comprised 2 μl of DNA, 2.5 μl of 10×Taq polymerase buffer, 25 pmol of each primer, 3 μl of dNTPs (10 mM) and 1 μl of Taq polymerase (Amersham). After 5 min of DNA denaturation at 95 °C, PCR was performed in 35 cycles: DNA denaturation, 94 °C, 30 s; annealing at 70 °C, 30 s, and elongation at 72 °C, 1 min; with a final extension at 72 °C for 10 min. The PCR products were verified by electrophoresis in 1.5% (w/v) agarose/TBE (90 mM Tris/85 mM boric acid/20 mM EDTA buffer, pH ??) gel and then extracted using a Gel Band DNA Extraction Kit (Amersham). The screening for 428G→A polymorphism was performed by digestion using the restriction endonuclease AvaII. The mutated sequence was not digested, and the normal sequence gave rise to two fragments (59 bp and 136 bp).

PCR and sequencing

All non-secretor individuals that did not show homozygosity for the 428G→A inactivating polymorphism (n=4) were screened for rarely described or for new polymorphisms by sequencing the whole coding region of the FUT2 gene.

The full-length FUT2 coding sequence was amplified using the forward primer: 5′-ATGCTGGTCGTTCAGATGCCTTTC-3′ and the reverse primer: 5′-CTGTCCCCCTTACTCAAGCACTAA-3′. PCR was performed according to the conditions described above. PCR products were purified using a Gel Band DNA Extraction Kit (Amersham) and used as templates for the sequencing reactions. Each sequencing reaction was performed in a final volume of 5 μl. The reaction mixture contained 2 μl of DNA template, 3 pmol of primer and 2 μl of TRR (terminator ready reaction) mix (ABI Prism® 3100 Sequencer; Applied Biosystems). The sequencing PCR was performed for 30 cycles: DNA denaturation, 94 °C, 30 s; annealing at 50 °C, 30 s, and elongation at 60 °C, 4 min; with a final extension at 60 °C for 20 min. The sequenced products were analysed in the ABI Prism® sequencer.

Construction of expression vectors

To evaluate enzyme activity of wt (wild-type) and polymorphic variants of FUT2 gene, expression vectors were constructed. To build the wt construct, the full-length FUT2 gene was amplified using the forward-full-length primer: 5′-ATGAAGCTTATGCTGGTCGTTCAG-3′ and the reverse-full-length primer: 5′-ATGGCGGCCGCAGTGCTTGAGTAAGG-3′ (HindIII and NotI restriction sites respectively are underlined). The PCR product (1032 bp) was restricted with the endonucleases HindIII and NotI, gel-purified and inserted into HindIII/NotI doubly digested pcDNAI. The mutant vector pcDNAI-FUT2-739G→A was made using the pcDNA-FUT2wt for site-directed mutagenesis using a nested PCR protocol. In the first PCR, two sets of primers were used to construct the mutant, namely the forward-full-length/reverse primer: 5′-GACACCTCCCACAGTGATGTGGTGT-3′ and forward primer/reverse-full-length primer: 5′-ACACCACATCACTGTGGGAGGTGTC-3′. In the second PCR the full-length mutant was amplified by using forward-full-length primer and reverse-full-length primer. The mutant vector pcDNAI-FUT2-839C→T was made by amplifying the mutant gene directly from an individual carrying the 839C→T polymorphism in homozygosity, using forward-full-length primer and reverse-full-length primer. The mutants were cloned into pcDNAI using the same protocol as that described above. The sequence of all constructs was confirmed by direct sequencing.

COS-7 cells

For the transient transfection with the expression vectors, COS-7 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) foetal-bovine serum. Transfection was performed using the cationic lipid reagent Tfx50 (Promega). A control plasmid encoding β-galactosidase was co-transfected to allow normalization of transfection efficiency. Cell extracts were prepared in 1% (v/v) Triton X-100/10% (v/v) glycerol/25 mM sodium phosphate, pH 6, 72 h after transfection. Samples were assayed for β-galactosidase (Promega) and α-1,2-fucosyltransferase. Protein concentrations from cell extracts were determined by using the BCA (bicinchoninic acid) assay reagent (Pierce Chemical Co.).

α-1,2-Fucosyltransferase activity

Fucosyltransferase assays were performed on a 20 μl reaction volume containing 3 μM GDP-[¹⁴C]fucose, 5 mM ATP, 25 mM sodium phosphate, pH 6.0, 40 μg of total protein from cell extracts and phenyl β-D-galactoside or asialofetuin as acceptor substrates. All assays were performed in duplicate, along with control reaction mixtures lacking acceptor. Reaction mixtures were incubated at 37 °C for 2 h, and termination was achieved by adding 1 ml of water to each microtube. The purification of hydrophobic fucosylated phenyl β-D-galactoside products was made by passage through a C₁₈ reverse-phase column. Radiolabelled asialofetuin products were purified by filtration through microfibre membranes (GF/C; Whatman) and radiactivity measured by liquid-scintillation counting.

To determine the K_m value for GDP-fucose, the reaction mixtures were supplemented with different concentrations of GDP-fucose over the range 3–400 μM. In these reaction mixtures the phenyl β-D-galactoside concentration was constant (25 mM). The K_m value for phenyl β-D-galactoside was calculated by testing different concentrations (5–50 mM) of the acceptor in reaction mixtures either containing 3 μM GDP-[¹⁴C]fucose or lacking GDP-fucose. The K_m for asialofetuin was determined by using a range of concentrations from 0.5 to 8 mg/ml.

RESULTS

Four FUT2 polymorphisms were found in our population (428G→A, 571C→T, 739G→A and 839T→C)

The 428G→A polymorphism was found in homozygosity in 32 (89%) of non-secretor individuals (Figure 1A). Individuals that were normal for 428G→A polymorphism (n=4; 11%) (Figure 1B) showed other polymorphisms after sequencing: 571C→T (n=2); 739G→A (n=1), and 839T→C (n=1). The 571C→T polymorphism inactivates enzyme activity and was previously described in Asian populations [21,23]. The polymorphisms 739G→A and 839T→C generate missense polymorphisms [739G→A (²⁴⁷Gly→Ser) and 839T→C (²⁸⁰Phe→Ser)] and are described for the first time in our population.

Figure 1. UEAI histochemistry and FUT2 genotype for 428G→A polymorphism in different individuals.

(A) Negative for UEAI and homozygous for 428G→A; (B) negative for UEAI and normal for 428G→A; (C) positive for UEAI and normal for 428G→A; and (D) positive for UEAI and heterozygous for 428G→A. All images have the gastric surface on top.

FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser proteins show impaired α-1,2-fucosyltransferase activity

After transfection, COS-7 cells showed high amounts of β-galactosidase that was used as a control for transfection: Cos-7 FUT2 wt (0.7×10⁻² unit); Cos-7 FUT2-739G→A (10⁻² unit); Cos-7 FUT2-839T→C (0.7×10⁻² unit) and Cos-7 mock (1×10⁻² unit). Cos wt cells had 0.006×10⁻² unit of β-galactosidase. The values for β-galactosidase are very similar in the cells transfected with the different constructs.

The expression of all variants of FUT2 gene (FUT2 wt, FUT2-739G→A and FUT2-839T→C) was confirmed by RT (reverse transcription) PCR (Figure 2). Mock and Cos-7 wt cells had no expression of FUT2 gene.

Figure 2. Detection of FUT2 expression by RT-PCR.

Expression of FUT2 and of GAPDH, used as an internal control, was detected by RT-PCR in Cos-7 cells either not transfected (Wt), mock transfected (mock) or transfected with variants of human FUT2 (FUT2 Wt, FUT2-739G→A and FUT2-839T→C).

Kinetic analysis showed that FUT2 wt has an apparent Michaelis–Menten constant (K_m) of 50 μM for GDP-fucose and a V_max of 21.23 pmol/h (Figure 3A). The mutants FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser had higher K_m values (178.6 and 90.9 μM respectively) and lower V_max values (8.66 and 4.37 pmol/h respectively), showing a lower α-1,2-fucosyltransferase activity under these conditions compared with FUT2 wt (Figures 3B and 3C). The K_m determined for the synthetic acceptor phenyl β-D-galactoside was 8.79 mM and the V_max was 0.89 pmol/h (Figure 3D). The α-1,2-fucosyltransferase activity assays using asialofetuin as the acceptor substrate revealed a K_m of 3.33 mg/ml and a V_max of 1.4 pmol/h for FUT2 wt (Figure 3E). It was not possible to determine the phenyl β-D-galactoside and asialofetuin K_m and V_max values for the mutants (FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser), since the reaction mixtures showed very low enzyme activity. Under all the reaction conditions tested, the cell extracts from mutants Cos-7 FUT2-²⁴⁷Gly→Ser and Cos-7 FUT2-²⁸⁰Phe→Ser exhibited levels of activity very similar to mock and lower than the cell extracts from Cos-7 FUT2 wt (Figure 4).

Figure 3. Determination of apparent Michaelis–Menten constant (K_m) and V_max for GDP-fucose, phenyl β-D-galactoside and asialofetuin in FUT2 wt, FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser.

(A) K_m and V_max for GDP-fucose in FUT2 wt; (B) K_m and V_max for GDP-fucose in FUT2-²⁴⁷Gly→Ser; (C) K_m and V_max for GDP-fucose in FUT2-²⁸⁰Phe→Ser; (D) K_m and V_max for phenyl β-D-galactoside in FUT2 wt; and (E) K_m and V_max for asialofetuin in FUT2 wt. The polymorphic enzymes FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser showed higher K_m values and lower V_max values for GDP-fucose compared with the FUT2 wt enzyme. It was not possible to calculate the K_m and V_max for phenyl β-D-galactoside and asialofetuin in FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser, since the reaction mixtures showed very low enzyme activity.

Figure 4. Transfer of GDP-fucose in different concentrations of acceptor and donor substrates for FUT2 wt, mock, FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser.

(A) Incorporation of GDP-Fucose in different concentrations of cold GDP-fucose onto phenyl-β-D-galactoside; (B) incorporation of GDP-fucose in different concentrations of phenyl-β-D-galactoside, and (C) incorporation of GDP-fucose in different concentrations of asialofetuin. FUT2 wt enzyme showed high α-1,2-fucosyltransferase activity whereas FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser enzymes were very similar to mock.

DISCUSSION

In the present study we have shown that all the non-secretors cases, determined by histology of gastric epithelium using UEAI, were homozygous for one nonsense (428G→A or 571C→T) or missense (739G→A or 839T→C) polymorphism located in the catalytic domain, generating nearly inactive FUT2 enzymes.

The 428G→A polymorphism was found in an allelic frequency of 0.89 in the population of UEAI-negative individuals in the present study. The 428G→A polymorphism was also shown to be the main one responsible for non-secretors in Caucasian populations from the United States [24], Sweden [17,19], Denmark [25] and South Africa [18]. In the case of 571C→T, 739G→A and 839T→C polymorphisms, it was not possible to calculate the allelic frequency, since only four individuals were screened. As far as we are aware, the 571C→T polymorphism was found in one Caucasian individual from South Africa, in an allelic frequency of 0.005 [18]. Furthermore, it is associated in a relatively low allelic frequency with non-secretors from Indonesia (0.03 [21]), Polynesia (0.13 [17]), Asian [aboriginal Taiwanese (from 0.02 to 0.2 [21]), Chinese (0.005, 0.007 and 0.008 [20,21,26]), Thai (0.006 [21]), Filipino (0.014 and 0.13 [21,27]) and Japanese (0.005 [2]) populations. The two new polymorphisms (FUT2-739G→A and FUT2-839T→C), identified for the first time in the present study, were found to be almost inactive by kinetic studies. The 739G→A mutation, at least, is not characteristic of the Portuguese population, since we found it in the heterozygous state in two Chinese individuals out of 50 (results not shown). Both FUT2-739G→A and FUT2-839T→C polymorphisms lead to an exchange between two neutral amino acids, ²⁴⁷Gly→Ser and ²⁸⁰Phe→Ser respectively. These changes can lead to conformational deviations in the enzyme catalytic domain that are incompatible with α-1,2-fucosyltransferase activity. Unfortunately, it is not possible to determine the approximate conformation of the polymorphic enzymes, since the FUT2 crystallized structure is not known.

The GDP-fucose K_m determined for FUT2 wt (50 μM) is lower than that described by other authors (197 μM) for α-1,2-fucosyltransferase supposedly encoded by the Secretor locus [24,28]. The higher affinity of our enzyme for the substrate GDP-fucose may be attributable to the fact that we cloned the full-length FUT2 and not only the secretor enzyme. The polymorphic enzymes (FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser) had higher K_m values, namely 178.6 and 90.9 μM respectively, showing a lower affinity for GDP-fucose. The K_m value determined for phenyl β-D-galactoside (8.79 mM) is very close to the values (15.1 and 11.4 mM) determined for α-1,2-fucosyltransferase in other studies [24,28]. Although asialofetuin is a heterogeneous substrate, the α-1,2-fucosyltransferase activity showed a very good correlation, and the K_m was 3.33 mg/ml. For both polymorphic enzymes (FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser) it was not possible to calculate the K_m for phenyl β-D-galactoside or for asialofetuin, since the reaction mixtures showed a very low enzyme activity. These results may have been obtained because the α-1,2-fucosyltransferase activity of these mutant proteins, if it exists, is very low, despite occurring in a linear way.

In conclusion, we have described two new polymorphisms in the FUT2 gene (FUT2-739G→A and FUT2-839T→C) that lead to the expression of two inactive FUT2 variants (FUT2-²⁴⁷Gly→Ser and FUT2-²⁸⁰Phe→Ser), and are responsible for some non-secretor cases in a Portuguese population.