Fluorescence Melting Curve Analysis for Concurrent Genotyping of Three Tag SNPs in FUT3

Abstract

The synthesis of Lewis blood group antigens is governed by two fucosyltransferase genes, FUT2 and FUT3. Evidence is accumulating to suggest that functional polymorphisms of FUT2 and FUT3 are associated with a variety of clinical conditions. Fluorescence melting curve analysis (FMCA), using three different dual-labeled probes for concurrent genotyping of three single nucleotide polymorphisms (SNPs) of FUT3, c.59T>G, c.314C>T, and c.484G>A for Lewis-negative allele inference, was developed and validated using Ghanaian and Caucasian subjects. Although two other SNPs, c.55G>A, and c.61C>T, are located in the probe sequence for c.59T>G, it seems feasible to detect these two SNPs along with c.59T>G. The results obtained by probe-based FMCA were in perfect accordance with those obtained by Sanger sequencing for 106 Ghanaians and 100 Caucasians. The present method is useful and reliable for estimating Lewis-negative alleles on a relatively large scale.

1. Introduction

The Lewis blood group antigens are the ABO(H) blood group-related antigens, consisting of Lewis a (Le^a) and Lewis b (Le^b) antigens, which are present not only on red blood cells, but also in other tissues and body fluids, such as saliva [1]. Le antigens on red blood cells are thought to be acquired secondarily from plasma [2]. The synthesis of these antigens is governed by FUT2-encoded α(1,2)fucosyltransferase (Se enzyme) and FUT3-encoded α(1,3/1,4)fucosyltransferase (Le enzyme). The functional alleles, Se and Le, are dominant over the nonfunctional alleles, se and le. Thus, subjects homozygous for le, i.e., Le-negative individuals, show a Le(a-b-) phenotype regardless of secretor status. While subjects with at least one Le (Le/Le or Le/le), i.e., Le-positive individuals, show an Le(a−b+) phenotype in secretors (Se/Se or Se/se), Le(a+b−) in non-secretors (se/se), and Le(a+b+) in weak-secretors (Se^w/Se^w or Se^w/se) [3]. Conventional serological Le phenotyping is somewhat difficult because it depends on the strength and specificity of the anti-Le^a and anti-Le^b antibodies used and the skill of the observer [4]. Therefore, phenotypic inference by reliable FUT2 and FUT3 genotyping is a useful alternative method for Le phenotyping.

The FUT3 polymorphism shows a population-specific distribution [5,6], and there are several Le enzyme-inactivating SNPs within a population. As described previously, four SNPs, c.202T>C (p.Trp68Arg, rs812936), c.484G>A (p.Trp68Arg, rs28362463), c.508G>A (p.Gly170Ser, rs3745635), and c.1067T>A (p.Ile356Lys, rs3894326), are the major causal SNPs for Le enzyme inactivation. Each of them is in linkage disequilibrium with other SNPs [5,6,7,8,9,10]: c.202T>C is in near absolute linkage disequilibrium with c.314C>T (p.Thr105Met, rs778986) (the allelic distributions are almost completely dependent on each other, i.e., T-C and C-T combinations are common for 202–314, but T-T and C-C combinations rarely exist), and c.484G>A is also in near absolute linkage disequilibrium with c.13G>A (p.Gly5Ser, rs28362458) [7,8,9]. In addition, c.59T>G (p.Leu20Arg, rs28362459) is closely linked with c.508G>A or c.1067T>A. According to the Erythrogene database (https://www.erythrogene.com/) [11], FUT3 alleles with 59G and 508A account for 15.2%, and those with 59G and 1067A for 9.4%, while those with 59T and 508A account for 0.14%, and those with 59T and 1067A for 0.22%, in the global population. Furthermore, the rare 59G allele, with neither 508A nor 1067A (0.68% in global population), has been reported to exhibit a Le(a-b-) erythrocyte phenotype but is positive for Lewis antigen in saliva [12]. Therefore, 484G>A (or c.13G>A), c.59T>G (or c.508G>A plus c.1067T>A), and c.202T>C (or c.314C>T) are suggested to be useful for estimating le alleles in many populations [5,6,13,14].

Fluorescence melting curve analysis using oligonucleotide probes (probe-based FMCA) is one of the most robust methods to detect SNPs [15]. Dual-labeled probes are commonly used for hydrolysis-probe (TaqMan) assay but can also be used for FMCA and allow multiplex assay by using different fluorochromes [16]. In the present study, we developed and validated probe-based FMCA for concurrent genotyping of c.59T>G, c.314C>T, and c.484G>A of FUT3.

2. Materials and Methods

The ethical committee of Kurume University reviewed and approved the research protocol (bioethics approval number 342).

2.1. DNA Samples

Genomic DNA previously isolated from 106 Ghanaians in Accra and genomic DNA from 100 Caucasians (HD100CAU) purchased from the Coriell Institute (Camden, NJ) were used for this study. The FUT3 coding sequences of these samples had already been determined by Sanger sequencing [13].

2.2. Probe-Based FMCA for Concurrent Genotyping of c.59T>G, c.314C>T, c.484G>A of FUT3

Since the human genome contains three paralogous genes with high sequence similarity, FUT3, FUT5, and FUT6 [17,18], primers should be designed to amplify only FUT3, not FUT5 and FUT6. Three pairs of primers for amplification of three regions containing each of c.59T>G, c.314C>T, and c.484G>A, respectively, and three self-quenching probes labeled with different fluorophores (59T-, 314C-, and 484A-probes) to determine c.59T>G, c.314C>T, and c.484G>A, are indicated in Figure 1 and

Table 1. Primers and probes for detection of SNPs in FUT3.

Primer Sequences	Position of FUT3	Differences with FUT5	Differences with FUT6	Amplicon Length
Detection of c.59T>G
forward primer: 5′-CCTCTCTCCTCTCTTCCCAGA-3′	−32 to −12 bp	1	2	179 bp
reverse primer: 5′-ACTGGGAGCCCTAGGGGAT-3′	129 to 147 bp	6	10
59T-probe: 5′-HEX-TCTGGCCGCACTGCTATTTCAGC-BHQ1-3′	48 to 70 bp
Detection of c.314C>T
forward primer: 5′-TGTGGCTCTGTCCCGCTGT-3′	225 to 243 bp	3	5	140 bp
reverse primer: 5′-GGAGGCGTGACTTAGGGTTGG-3′	344 to 364 bp	8	8
314C-probe: 5′-Cy5-ACAGGCAGACACGGTCATCGT-BHQ2-3′	303 to 323 bp
Detection of c.484G>A
forward primer: 5′-AGCCACCCCCTAACTGCCA-3′	413 to 431bp	6	9	158 bp
reverse primer: 5′-GGTCTTGGCCGAGAGGTTGA-3′	551 to 570 bp	0	0
484A-probe: 5′-FAM-CAGCAACTCCGACATCTTCA-BHQ1-3′	480 to 499 bp

The position of SNPs is shown by bold letters. “Differences with FUT5 (or FUT6)” indicate the number of base differences in each of primer with corresponding FUT5 and FUT6 sequences.

.

To amplify the region containing c.484G>A, the same PCR primers as the Eprobe-based FMCA for detection of c.508G>A were used [19]. The asymmetric PCR reaction mixture has a final volume of 10 µL and contains 5 µL of Premix Ex Taq (Probe qPCR) (Takara, Tokyo, Japan), 50 nM of 59F-primer, 500 nM of 59R-primer, 50 nM of 314F-primer, 500 nM of 314R-primer, 50 nM of 484F-primer, 500 nM of 484R-primer, 100 nM of 59T-probe, 200 nM of 314C-probe, 200 nM of 484A-probe, and 2–20 ng of genomic DNA. The PCR was conducted on a LightCycler 480 instrument II (Roche Diagnostics, Tokyo, Japan) with the following thermal conditions: pre-denaturation at 95 °C for 30 s, then 50 cycles of denaturation at 95 °C for 5 s, and annealing/extension at 60 °C for 15 s. The PCR products were then heated to 95 °C for 1 min, cooled to 40 °C for 1 min, and fluorescence data were obtained during heating from 40 to 80 °C at a 0.1 °C/s ramp rate with 3 acquisitions/s using the FAM filter (Excitation-Emission: 465 nm–510 nm) for the 484A-probe, VIC/HEX/Yellow 555 filter (533 nm–580 nm) for the 59T-probe and Cy5/Cy5.5 filter (618 nm–660 nm) for the 314C-probe. Melting curve genotyping and melting temperature (Tm) calling analyses were carried out using LightCycler 480 gene scanning software.

3. Results

3.1. Probe-Based FMCA for Genotyping of c.59T>G, c.314C>T, and c.484G>A of FUT3

In this study, we first considered five SNPs, c.13G>A, c.59T>G, c.202T>C, c.314C>T, and c.484G>A, of FUT3, as candidates for probe-based FMCA analysis to infer le alleles in various populations. Among them, c.484G>A and c.314C>T showed more favorable analysis results than c.13G>A and c.202T>C, respectively, which could be the respective surrogate SNPs. Accordingly, the three SNPs—c.59T>G, c.314C>T, and c.484G>A—were finally selected for le allele inference (data not shown). Therefore, in this study, a triplex probe-based FMCA analysis for c.59T>G using a 179-bp amplicon, HEX-labeled 59T-probe, for c.314C>T using a 140-bp amplicon and Cy5-labeled 314C-probe, and for c.484G>A using a 158-bp amplicon and FAM-labeled 484A-probe, was carried out (Figure 1a–c).

According to our previous results [13], as well as the results of Erythrogene [11], two other SNPs, c.55G>A (Ala19Thr, rs146199130) and c.61C>T (synonymous SNP, rs28362460), were located in the 59T-probe sequence (Figure 1a). c.55G>A was present at a much lower frequency in Europeans and Africans, and c.61C>T appeared to be specific to Africans and Americans. Because 55A or 61T was located at the 59G allele, ten haplotype combinations were expected, but only seven samples could be subjected to probe-based FMCA analysis because DNA samples with the other three haplotype combinations were not available (

Table 2. Haplotype combinations of 55-59-61 and observed Tm values of probe-based FMCA analysis for 59T>G.

	55-59-61/55-59-61	Tm (°C)
Subject 1	G-T-C/G-T-C	68.0
Subject 2	G-T-C/G-G-C	68.0/61.5
Subject 3	G-G-C/G-G-C	61.5
Subject 4	G-T-C/G-G-T	68.5/53.5
Subject 5	G-G-C/G-G-T	61.5/53.5
Subject 6	G-G-T/G-G-T	53.5
Subject 7	G-T-C/A-G-C	68.0/56.5
DNA not available	G-G-C/A-G-C	-
DNA not available	A-G-C/A-G-C	-
DNA not available	G-G-T/A-G-C	-

). Probe-based FMCA analysis could clearly separate the seven subjects using a VIC/HEX/Yellow 555 filter (Figure 2a). Tm values for each peak are shown in Table 2.

Next, for c.314C>T, subjects with genotypes of C/C (wild type), with Tm about 70.5 °C, T/C, with two Tm peaks of about 61 °C and 70.5 °C, and T/T with Tm about 61 °C for c.314C>T, were distinguished from each other using a Cy5 filter (data not shown). Finally, for c.484G>A, subjects with genotypes of G/G (wild type) with a Tm of about 57.5 °C, G/A with two Tm peaks of about 57.5 °C and 64.5 °C, and A/A with a Tm of about 64.5 °C for c.484G>A, were also clearly distinguished from each other using the FAM filter (data not shown).

3.2. Validation of Probe-Based FMCA for Genotyping of Three SNPs

We then analyzed 106 Ghanaians and 100 Caucasians, and the results of c.59T>G (also c.55G>A and c.61C>T) and c.314C>T were in perfect accordance with previous Sanger sequencing results [13]. Representative melting curve genotyping results for c.59T>C and c.484G>A in Ghanaians are shown in Figure 2b,d, and for 314C>T in Caucasians in Figure 2c.

We confirmed the repeatability of the present triplex probe-based FMCA method by two independent assays of 106 Ghanaians and 100 Caucasians.

3.3. Inferred Le Phenotypes from Combinations of Three SNPs Genotypes

Although we did not perform Le phenotyping in the present subjects, we inferred Le phenotypes from combinations of three SNPs genotypes. Thirty-five of 106 Ghanaians (33%) and 10 of 100 Caucasians (10%) were likely to have Le-negative phenotype (

Table 3. Inferred Lewis genotypes and phenotypes from three SNPs.

c.59T>G	c.314C>T	c.484G>A	Lewis Genotype	Lewis Phenotype	Number of Ghanaians	Number of Caucasians
T/T	C/C	G/G	Le/Le	positive	23	51
T/T	C/C	G/A	Le/le	positive	23	0
T/G	C/C	G/A	le/le	negative	9	0
T/T	C/T	G/A	le/le	negative	7	0
T/T	C/C	A/A	le/le	negative	2	0
T/G	C/C	G/G	Le/le	positive	17	8
T/G	C/T	G/G	le/le	negative	3	7
G/G	C/C	G/G	le/le	negative	14	1
T/T	C/T	G/G	Le/le	positive	8	31
T/T	T/T	G/G	le/le	negative	0	2

59G contained 55A-59G-61C and 55G-59G-61T.

). However, as shown in

Table 4. FUT3 alleles whose Lewis phenotype may be misjudged by the present FMCA method.

SNPs	Inferred	Actual	Frequency (%)
			Africa	America	E. Asia	Europe	S. Asia	Global
59T-508A	Le	le	0.23	0	0.20	0	0.20	0.14
59T-1067A	Le	le	0.08	0	0.60	0.10	0.30	0.22
202C-314C	Le	le	0.38	0	0	0	0.40	0.18
59G-484A-1067A	le	le	0	0	0.10	0	0	0.02
Uncharacterized	Le	?	1.60	0.99	0.60	1.10	1.32	1.16

59T-508A: 59T alleles with 508A; 59T-1067A: 59T alleles with 1067A (including 59T-202C-314C-1067A with frequency of 0.10% in S. Asia); 202C-314C: 314C alleles with 202C (including 59T-202C-314C-1067A with frequency of 0.10% in S. Asia); 314T allele with 202T was not observed in the database; 59G-484A-1067A: This allele itself is le, but when subjected to test for Le/le^59,484,1067 was erroneously determined to be le^59,1067/le⁴⁸⁴ plus other SNPs; Uncharacterized: alleles with only uncharacterized SNPs; Any allele may contain alleles with or without other additional SNPs. E. Asia: East Asia, S. Asia: South Asia. Allele frequencies were obtained from Erythrogene [11]; "?" means undetermined (may contain both both Le and le).

, a limitation of the present assay is that there is a possibility of misclassification of about 1–2% subjects in the estimation of the Le phenotype, including alleles with SNPs that have not yet been characterized.

4. Discussion

Evidence is accumulating to suggest that FUT3 polymorphisms or Le phenotypes are associated with a variety of pathologic conditions, including Helicobacter pylori infection, ischemic heart diseases, intestinal infections, inflammatory bowel diseases, ankylosing spondylitis, COVID-19 susceptibility, and autoimmune neutropenia [20,21,22,23,24,25]. Large-scale replication studies are needed to confirm these associations, and reliable FUT3 genotyping methods are desirable for this purpose.

Several methods for detection of FUT3 SNPs, such as PCR-RFLP, PCR using sequence-specific primers (PCR-SSP), an allele-specific oligonucleotide hybridization method, TaqMan assay, Sanger sequencing of PCR amplicons, and a multiplex SNaPshot assay, were reported [13,22,26,27,28]. In addition, we have recently developed three independent high resolution melting (HRM) analyses that detect SNPs, c.13G>A, c.59T>G, and c.202T>C, in FUT3 and probe-based FMCA genotyping at positions of c.508G>A and c.1067T>A using a unique fluorescence probe (Eprobe) [14,19,29]. Compared to other methods, HRM, probe-based FMCA, and TaqMan assay do not require post-PCR processing. The TaqMan assay is an established and frequently used method for SNP detection. However, it requires two dual-labeled probes to detect one SNP, whereas the probe-based FMCA requires only one dual-labeled probe to detect one SNP [15]. Accordingly, the general TaqMan assay requires multiple assays to detect SNPs at multiple sites, whereas probe-based FMCA allows multiplex assay that can detect SNPs at multiple sites in a single assay by employing one site-specific fluorescence probe. Therefore, probe-based FMCA has the potential to reduce assay cost and time compared to TaqMan assay.

HRM is an effective method for detecting heterozygotes of rare variants [30]. However, it does not seem to be suitable for detecting individual homozygotes for SNPs that are present at relatively high frequencies since the difference in their respective Tm values is often very small (usually less than 1 °C). On the other hand, probe-based FMCA is suitable for detection of the homozygotes for each of the relatively high-frequency SNP because of the large difference in Tm values (usually 4–10 °C) between homozygotes for each of them [16]. Eprobe-based FMCA is an excellent method to detect SNPs [29], but its disadvantage compared to the dual-labeled probe-based FMCA is that the multiplex assay is limited because, at present, only three fluorochromes (oxazole yellow, thiazole orange, and thiazole pink) can be used, and the cost of probe synthesis is high.

The presence of SNPs other than the target SNP in the probe sequences would affect the melting curve profile obtained by probe-based FMCA analysis. Therefore, we searched the Erythrogene database to see if the three probes contained other SNPs and found that only the probe for c.59T>G detection contained two SNPs (c.55G>A and c.61C>T). The c.55G>A was found in European and African populations with quite low frequency, and c.61C>T was found in Africans and Americans (4.24% in Africans, 1.30% in Americans, and 1.30% in global populations); 61T was likely to be in complete linkage disequilibrium with both 59G and 508A. Although the frequency of 55A was quite low, this SNP appeared to be linked with the 59G allele and with neither 508A nor 1067T in Europeans [13]. Probe-based FMCA is likely to be able to detect other SNPs even if they are nearby, while HRM seems to be difficult when other SNPs are nearby. Thus, for large-scale association studies of FUT3, examination probe-based FMCA is more suitable than HRM analysis. In fact, accurate genotyping of c.59T>G was difficult with HRM analysis, especially for African samples containing c.61C>T, in addition to c.59T>G [14]. However, in this probe-based FMCA, the allele containing c.61C>T (55G-59G-61T) was clearly distinguished from the other alleles (55G-59T-61C, 55G-59G-61C, and 55A-59G-61C).

Recently, we have developed a probe-based FMCA to identify the three major FUT2 SNPs involved in Se enzyme inactivation [31]. Simultaneous implementation of two probe-based FMCA methods for detection of FUT2 and FUT3 SNPs would allow for more accurate estimation of Le phenotypes.

In conclusion, the present probe-based FMCA for c.59T>G, c.314C>T, and c.484G>A is valid and feasible for inference of Le phenotypes and for association studies of FUT3 in populations around the world.