Polymorphisms in the FcεRIβ Promoter Region Affecting Transcription Activity: A Possible Promoter-Dependent Mechanism for Association between FcεRIβ and Atopy

Sequencing analysis

Genomic DNA was prepared from peripheral blood using a DNA quick kit (Dainippon Pharmaceutical, Osaka, Japan). Nucleotide sequencing of FcεRIβ was conducted as previously reported for the α-chain promoter (12). The following oligonucleotides were used as primers to amplify the β-chain gene: 5′-CCACCAATTCCTGAAGAC-3′ (−1098/−1081) and 5′-CACCGTGACTATGACTTC-3′ (+225/+208) to amplify the promoter region; and 5′-AGCGTGGTGGCAGGTACCTGAGGTT-3′ (+6203/+6227) and 5′-TTTCCAGTAGCCCCTTAACAAAACC-3′ (+7062/+7038) to amplify exon 7. Oligonucleotides 5′-CCCATTCTTGCCACTGT-3′ (−667/-651), 5′-ACAGTGGCAAGAATGGG-3′ (−651/-667), 5′-CCAGAAGAAGGGCACATCTC-3′ (−279/-260), 5′-GAGATGTGCCCTTCTTCTGG-3′ (−260/-279), and 5′-ATGACAGAGAGCGTGAGACCCAGA-3′ (+6770/+6793) were used as the sequencing primers.

Subjects

All subjects were unrelated Japanese volunteers from the University of Tokyo (Departments of Biotechnology and Applied Biological Chemistry, Biotechnology Research Center) and from Juntendo University School of Medicine (Departments of Dermatology and Immunology, Atopy Research Center). We selected healthy individuals who had no symptoms or history of atopic dermatitis, atopic asthma, or allergic rhinitis for the determination of the cell surface expression of FcεRI. This study was approved by the ethics committee of Juntendo University School of Medicine.

Measurement of FcεRI level on blood basophils by flow cytometry

Basophils were enriched from 5 ml of peripheral blood using Polymorphoprep (AXIS-SHIED PoC, Oslo, Norway). As previously reported (13), cells were analyzed by flow cytometry on a FACSCalibur (BD Biosciences, Mountain View, CA) after staining with PE-conjugated anti-human CD3, CD7, CD9, CD14, and CD19 Abs (BD Pharmingen, San Diego, CA) and with FITC-conjugated anti-human FcεRI α-chain Ab CRA1, which does not compete with IgE (Cosmo Bio, Tokyo, Japan) or FITC-conjugated mouse IgG2b isotype control (BD Pharmingen). Mean fluorescence intensity (MFI) of the anti-FcεRIα staining on CD3⁻CD7⁻CD9⁻CD14⁻CD19⁻ basophils was determined using CellQuest software (BD Biosciences).

Statistical analysis.

The p value was determined using a paired t test and <0.05 was considered to indicate statistical significance. The χ² test was performed to analyze linkage between each polymorphism using SPSS software (version 11.0; SPSS, Chicago, IL). A p value of exact significance was used in the χ² test. Multiple linear regression analysis was performed to test for predictors of MFI by entering genotypes of SNPs into the model (SPSS).

Cell lines

Human mast cell lines, HMC-1 and KU812, a megakaryocyte cell line HML/SE, and a cervical carcinoma cell line, HeLa, were maintained and characterized as described previously (14, 15).

Reporter assay for promoter activity

Reporter plasmids carrying the whole β-chain promoter of a major allele (−752T/−654C/−426T/−109T) and a minor allele (−752C/−654T/−426C/−109C) were generated from the reporter plasmid pGLβ(−1079/+102) (14) by site-directed mutagenesis using a QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Three tandem repeats of each polymorphic sequence were introduced upstream of the human β-chain minimal promoter (14) in a luciferase reporter plasmid (see Fig. 3⇓C). In brief, 3 x (−121/−97) sequence, 3 x (−438/−414), 3 x (−666/−642), and 3 x (−764/−740) were used to examine the effect of −109C/T, −426T/C, −654C/T, and −752C/T SNPs, respectively, on promoter activity. The expression plasmid pCR-YY1 (16) was used for coexpression analysis. Cells were transfected with 5 μg of reporter plasmid to compare the effect of each SNP, or they were infected with 3 μg of reporter plasmid and 3 μg of expression plasmid for coexpression analysis. Transfection was performed by electroporation using a Bio-Rad Gene Pulsar II (Bio-Rad, Hercules, CA) as described previously (17). The luminescence of each cell lysate was measured using MicroLumat Plus (Berthold, Postfach, Germany) as described previously (14).

EMSA

The following oligonucleotides labeled with FITC at the 5′ site were used as the probes for EMSA: 5′-GAGACTAACACAC/TACTCACTCACAT-3′ corresponding to −438/-414 for −426C/T and 5′-CCATTCTTGCCAC/TTGTAAAGATCTA-3′ corresponding to −666/-642 for −654C/T. Nuclear extracts and double-stranded oligonucleotide probes were prepared as described previously (18, 19). Abs against YY1 and USF1 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). In vitro transcription and translation were performed with a TnT T7 Quick coupled transcription/translation system (Promega) using pCR-YY1 (16) as the template. The FITC-labeled probe was mixed with nuclear extract or the in vitro transcription/translation mixture and subjected to 4% PAGE as described in our previous report (14, 20). Fluorescent images were analyzed on a FluorImager 595 (Molecular Dynamics, Sunnyvale, CA).

Quantification of β-chain mRNA

Basophils were purified by the MACS separation technique using a Basophil Isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany) and a Vario MACS (Miltenyi Biotec) according to the manufacturer’s instructions. In brief, PBMC from 60 ml of peripheral blood, which was isolated by density gradient centrifugation, were treated with isotonic ammonium chloride buffer to lyse erythrocytes and then washed with PBS supplemented with 0.5% BSA and 2 mM EDTA. The PBMC suspension treated with reagents from a Basophil Isolation kit (FcR Blocking Reagent, Hapten-Antibody Cocktail, and MACS Anti-Hapten BicroBeads) was applied onto an LS⁺ column in the magnetic field of a Vario MACS. The purity of the isolated basophils, which were collected as negative fraction using the above magnetic separation system, was confirmed to be ∼95% by flow cytometry. Total RNA was prepared from the basophils using TRIzol reagent (Invitrogen Life Technologies, Leek, The Netherlands). Reverse transcription was performed using a SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen Life Technologies). The Advantage 2 polymerase mix (BD Clontech, Palo Alto, CA) was used for semiquantitative PCR with the same primer sets for human FcεRI β-chain and β-actin as used in our previous study (14). The thermal cycling condition was as follows: denaturalization at 95°C for 30 s, annealing at 55°C for 30 s, and extension at 68°C for 1 min. The amount of FcεRI β-chain mRNA was also quantified using a 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA) with an Assays-on-Demand gene expression product (no. Hs00175091_m1# for human FcεRI β-chain) and TaqMan Universal PCR Master Mix (Applied Biosystems). The expression level of the FcεRI β-chain was shown as a ratio to that of GAPDH by calculation of cycle threshold (Ct) values in amplification plots with 7500 SDS software. The β-chain mRNA expression level in Fig. 5⇓B is the ratio to that of no. 1 of the −654C/C individual.

The β-chain mRNA expression level (%) = 2^{((Ct value of FcεRI β-chain) − (Ct value of GAPDH)) − ((Ct value of FcεRI β-chain of −654C/C no. 1) −(Ct value of GAPDH of −654C/C no. 1))} × 100.

Correlation between the FcεRIβ Glu237Gly polymorphism and cell surface expression level of FcεRI on blood basophils

We first examined the correlation between FcεRIβ polymorphism at Glu237Gly and the expression level of FcεRI on peripheral blood basophils from nonatopic healthy individuals whose genotypes were determined by direct sequencing. Most of the Japanese population carried either GAA/GAA (Glu/Glu) (48 of 70 individuals) or GAA/GGA (Glu/Gly) (21 of 70 individuals) at +6960 (237). No basophils of the GGA/GGA (Gly/Gly) genotype were subjected to this analysis due to its extremely low allelic frequency (1 of 70 individuals). As shown in Fig. 1⇓, basophils of the Glu/Gly (A/G) genotype expressed significantly higher levels of FcεRI (MFI = 759 ± 78) than those of the Glu/Glu (A/A) genotype (MFI = 480 ± 46; p = 0.0037). However, we and others have previously demonstrated that the amino acid substitution Glu237Gly itself has no direct effect on the expression or function of FcεRI (10, 11). In this context, we hypothesized that the Glu237Gly (+6960A/G) polymorphism might be linked with another polymorphism in the promoter region that was able to directly affect β-chain gene expression.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Correlation between the FcεRIβ polymorphism at Glu237Gly and cell surface expression levels of FcεRI on peripheral blood basophils. Peripheral blood basophils from 237 Glu/Glu (+6960A/A) and 237 Glu/Gly (+6960A/G) healthy volunteers were stained with FITC-conjugated anti-human FcεRI α-chain mAb, which does not compete with IgE (see Materials and Methods). Data represent the MFI determined by flow cytometry. A p value was determined by a paired t test.

Linkage between polymorphisms in the FcεRIβ promoter and Glu237Gly polymorphism

We sequenced over 1 kb of FcεRIβ promoter region from 70 individuals to identify polymorphic sites. As shown in Fig. 2⇓, we found four SNPs at −109, −426, −654, and −752. Subsequently, we performed statistical analysis to examine their linkage to the +6960A/G polymorphism (Tables I–IV⇓⇓⇓⇓). When the polymorphisms were scored for association with each other, we found a high degree of linkage disequilibrium: −426T was closely associated with +6960A, while −426C was closely associated with +6960G (Table II⇓, χ² = 117.81 with 4 df; p = 2.39 × 10⁻¹³). In addition, −654C and T were also closely associated with +6960A and +6960G, respectively (Table III⇓, χ² = 110.01 with 4 df; p = 1.96 × 10⁻¹¹). A SNP at −109 was also significantly associated with that at +6960 (Table I⇓, χ² = 20.17 with 4 df; p = 0.00018). In contrast, no statistically significant association was found between −752T/C and +6960A/G (Table IV⇓, p = 0.083). By the exact test from 2 × 2 contingency tables excluding heterozygous genotypes (Table V⇓), significant association of the Glu237Gly polymorphism was observed with SNPs at −426 (p = 0.021) and −654 (p = 0.021) but not with those at −109 (p = 0.069) and −752 (p = 0.31). Association between the polymorphism at Glu237Gly (+6960A/G) and those in the promoter is summarized in Fig. 3⇓A. We also found that the −752T/−654C/−426T/−109T/+690A genotype constitutes a typical major allele while the −752C/−654T/−426C/−109C/+6960G genotype constitutes a typical minor allele in the Japanese population.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Polymorphisms in the human FcεRIβ promoter region. Polymorphic sites in the 5′ flanking region of human FcεRIβ are shown on the nucleotide sequence (accession number AB080913) (14 ). The transcription start site is indicated as +1. The TATA box and Oct-1 binding sites are shown in boxes.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Effect of SNPs in the FcεRIβ promoter on its function. A, Schematic presentation of the association between the Glu237Gly polymorphism and promoter region polymorphisms. Significant associations were observed between +6960 and −426 and between +6960 and −654. Associations between +6960 and −109 and that between +6960 and −752 are not significant. Details are shown in Tables I–V⇓⇓⇓⇓⇓. B, Effect of polymorphisms in the promoter region on its transcriptional activity. A reporter assay was performed in KU812 cells transfected with luciferase reporter plasmid carrying the whole major (TCTT) or minor (CTCC) allele promoter. The relative luciferase activity is represented as the ratio to the activity of pGL3-Basic. The results are expressed as the mean + SE of triplicate samples. A p value was determined by a paired t test. C, The indicated reporter plasmids with three tandem repeats of each polymorphic site in the FcεRIβ promoter region were transiently transfected into HML/SE cells. The relative luciferase activity is represented as the ratio to the activity of pGL3-Basic. The results are expressed as the mean + SE of triplicate samples. A p value was determined by a paired t test.

To evaluate the effect of each polymorphism in the promoter on cell surface FcεRI expression, we reanalyzed the data in Fig. 1⇑ using a multiple linear regression model. A significant multiple linear regression equation was obtained (p = 0.0065), and the −109T (p = 0.0024), −426C (p = 0.0058), and −654T (p = 0.031) SNPs were significantly associated with higher FcεRI expression. These results suggested that the higher expression of FcεRI on basophils carrying the +6960A/G genotype was due to the effect of polymorphisms in the promoter that were tightly linked to the +6960A/G SNP.

Effect of SNPs on promoter activity

To examine the effect of the four SNPs on the promoter activity, we performed the luciferase reporter assay in FcεRIβ-positive KU812 cells (14), which were transiently transfected with the reporter plasmids containing the whole promoter (−1079/+102) of the major allele carrying −752T/−654C/−426T/−109T (TCTT), or the minor allele, carrying −752C/−654T/−426C/−109C (CTCC), just upstream of the luciferase gene. As shown in Fig. 3⇑B, the minor allele (CTCC) promoter exhibited significantly higher activity than the major allele (TCTT) promoter in KU812 cells (p = 0.014).

To identify the SNP affecting the promoter activity, we constructed four pairs of reporter plasmids, in which three tandem repeats of 25-bp sequence around each SNP were placed just upstream of the minimal β-chain promoter (−95/+102) containing a TATA box and Oct-1 binding sites (see Fig. 2⇑) (14) as shown in Fig. 3⇑C. FcεRIβ-positive HML/SE cells (14) were transfected with these reporter plasmids, and luciferase activity in the cell lysates was measured. The sequence with −654T exhibited markedly higher cis-enhancing activity than that with −654C. The sequence with −426C also showed moderately higher activity than that with −426T. By contrast, no significant difference was observed between −109C and −109T or between −752C and −752T. These results indicate that both −426C and −654T in the minor allele, which were tightly linked with the +6960G genotype (Fig. 3⇑A), had higher cis-enhancing activity than −426T and −654C in the major allele.

Specific transactivation of the −654T allele by YY1

The significant effects of SNPs at −426 and −654 on the promoter activity suggested that some unidentified transcription factor may be differentially recognizing these SNPs. We then performed EMSA with nuclear extracts from FcεRIβ-positive cell lines (HMC-1, KU812, and HML/SE) or a negative cell line (HeLa) using double-stranded 25-bp oligonucleotides corresponding to −426T/C (−438/−414) or −654C/T (−666/−642) as probes. No apparent difference was observed with the −426C and −426T probes (data not shown). In contrast, a specific band was observed with the −654T probe but not with the −654C probe (Fig. 4⇓A, indicated by arrow). This band specifically disappeared by the addition of excess unlabeled −654T probe but not −654C probe as a competitor (Fig. 4⇓B). These results indicate the presence of a nuclear protein, which specifically bound to the −654T allele. This nuclear protein was found not only in FcεRIβ-positive cells but also in FcεRIβ-negative HeLa cells (Fig. 4⇓A), suggesting that it is ubiquitously expressed. We noticed that the nucleotide substitution of T for C at −654 generated a potential YY1-binding motif, CCAT, at this position. To test the involvement of YY1, we added anti-YY1 Ab to the EMSA mixture and found that the target band disappeared in the presence of anti-YY1 Ab but was still present after addition of irrelevant anti-USF1 Ab to the assay mixture (Fig. 4⇓B). Furthermore, when YY1 produced by in vitro transcription/translation was used instead of nuclear extract, a band showing the same mobility was observed (Fig. 4⇓C, indicated by arrow). This band disappeared in the presence of anti-YY1 Ab.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Specific binding and transactivation of the −654T minor allele by YY1. A, Differential binding of a nuclear protein to −654C/T alleles. Nuclear proteins prepared from the indicated cell lines were subjected to EMSA with a probe containing the −654C or −654T allele. A specific band for the −654T allele is shown by the arrow. B, Discrimination of −654C/T polymorphism by YY1. A competitive binding assay was performed with a 100- or 200-fold excess of unlabeled competitor containing −654T or −654C in EMSA with the KU812 nuclear extract and −654T probe. In some lanes, anti-YY1 or anti-USF1 Ab was added to confirm the binding of YY1 to the −654T probe. The arrow indicates the YY1 band. C, Binding of in vitro-translated YY1 to the −654T allele. In vitro-translational products from YY1 or mock expression vector and KU812 nuclear extract were subjected to EMSA with the −654T probe. Anti-YY1 Ab was added to some lanes. The arrow indicates the YY1 band. D, Specific transactivation of the −654T allele by YY1. A reporter assay was performed in KU812 cells transfected with the luciferase reporter plasmid carrying the minimal promoter with (−654C)x3 or (−654T)x3 along with YY1 or mock expression vector. The relative luciferase activity is represented as the ratio to the activity of pGL3-Basic plus mock. The results are expressed as the mean + SE of three independent experiments with triplicate samples. The p values were determined by a paired t test.

To evaluate the effect of YY1 on −654C/T SNP, we performed cotransfection analysis using reporter plasmids carrying three tandem −654 C or T sequences just upstream of the FcεRIβ minimal promoter (see Fig. 3⇑C). The luciferase activity driven by the promoter with (−654T)x3 sequence was significantly higher than that driven by the (−654C)x3 sequence in KU812 cells (p = 0.029; Fig. 4⇑D) as well as in HML/SE cells (Fig. 3⇑C). Overexpression of YY1 moderately up-regulated the promoter with the (−654C)x3 sequence (p = 0.058; Fig. 4⇑D). This was probably due to a nonspecific effect of exogenous YY1, because a similar effect of exogenous YY1 was observed when a promoter-less plasmid, pGL3-Basic, was used as the reporter plasmid (data not shown). In contrast, the promoter with (−654T)x3 sequence was markedly up-regulated by overexpression of YY1 (p = 0.00053; Fig. 4⇑D). When the fold induction was compared between (−654C)x3 and (−654T)x3, the effect of coexpressed YY1 on (−654T)x3 (2.27-fold) was statistically higher than that on (−654C)x3 (1.77-fold; p = 0.021). These results suggested that the higher transcriptional activity of the minor allele promoter was partly due to specific transactivation by YY1 via the site containing −654T.

Expression level of the FcεRI β-chain in blood basophils from −654C/C and −654C/T individuals

To analyze the effect of the SNPs on the expression level of the FcεRI β-chain, we compared the amount of FcεRI β-chain mRNA in basophils between −654C/C individuals (n = 4) and −654C/T individuals (n = 5). The amount of FcεRI β-chain mRNA in basophils isolated from peripheral blood was analyzed by semiquantitative PCR (Fig. 5⇓A) and real-time PCR (Fig. 5⇓B). The expression of FcεRI β-chain mRNA in basophils from individuals who have the minor heterozygous allele was significantly higher than that of those with the major homozygous genotype, as determined by quantitative PCR (Fig. 5⇓B). A similar conclusion was obtained using semiquantitative PCR, as interpreted using the intensity of the bands in the agarose gel profile (Fig. 5⇓A). These results indicate that the minor allele at −654 of FcεRIβ increases FcεRI β-chain transcription in basophils.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Expression level of FcεRI β-chain mRNA in blood basophils. A, Semiquantitative PCR to compare the expression level of FcεRI β-chain mRNA in basophils from −654C/C (n = 4) and −654C/T (n = 5). PCR products obtained after 23, 26, 29, 32, and 35 cycles (for β-chain) and 20, 23, 26, 29, and 32 cycles (for β-actin), respectively, were applied onto agarose gel and detected by ethidium bromide staining. For detection of the β-chain transcript, 10-fold cDNA was used as template compared with β-actin. Numbers at the top of the agarose gel profiles indicate the individuals. B, Quantification of FcεRI β-chain mRNA by Real Time PCR. The numbers at each circle indicate the individuals corresponding to those in A. The expression level of the FcεRI β-chain (FcεRI β-chain/GAPDH) of each individual is represented as the ratio to that of no. 1 of −654C/C by setting individual no. 1 of −654C/C at 100%. A p value was determined by a paired t test.