Three frequent genetic polymorphisms in the human high-affinity IgE receptor α-subunit (FcεRIα) were shown to be associated with allergic disorders and/or total serum IgE levels in allergic patients. Two of these were previously demonstrated to affect FcεRIα expression while the third –18483A>C (rs2494262) has not yet been subjected to functional studies. We hypothesized that the –18483A>C variant affects transcriptional activity of the FcεRIα distal promoter in monocytes in which FcεRIα transcription is driven through that regulatory region. Indeed, we confirmed preferential binding of the YY1 transcription factor to the –18483C allele, resulting in lower transcriptional activity when compared with the –18483A allele.
Considering the crucial role of the human high-affinity IgE receptor (FcεRI; Zhang et al. 2007), it is not surprising that FcεRI α-subunit (FcεRIα) gene variability was recently studied in the context of allergies and allergic disorders. FcεRIα –344(–315)1 C>T (rs2427827) polymorphism was found to be associated with aspirin-induced urticaria (Bae et al. 2007) and with total serum IgE levels in different groups of allergic subjects (Shikanai et al. 2002; Potaczek et al. 2006, 2007a, 2007b; Bae et al. 2007). Its close genomic neighbor, FcεRIα proximal promoter –95(–66)1 T>C (rs2251746) polymorphism, was shown to be associated with atopic dermatitis (Hasegawa et al. 2003b). Both of these frequent polymorphisms were demonstrated to strongly affect FcεRIα expression in mast cells and/or basophils in an additive manner (Hasegawa et al. 2003b; Bae et al. 2007; Kanada et al. 2008), thus providing a partial mechanistic background for the genetic associations described above. In addition, the frequent –18483A>C (rs2494262) polymorphism was also found to be associated with total serum IgE levels in allergic subjects (Potaczek et al. 2007a, 2007b); however, no functional studies on that genetic variant have been conducted to date. Here, we report the results of functional analysis of FcεRIα –18483A>C polymorphism.
The FcεRIα gene is composed of five coding exons, spanning about 6 kbp, directly preceded by the proximal promoter while two additional untranslated exons (1A and 2A) are localized further upstream (about 12 and 18 kbp, respectively; Nishiyama et al. 2001) and are under transcriptional regulation of the distal promoter (Hasegawa et al. 2003a; Fig. 1a). In contrast to basophils and/or mast cells, in which the proximal promoter strongly works with the GATA-1/PU.1 transcription factor-dependent activating mechanism, in monocytes, transcription of FcεRIα is also driven by the distal promoter (Hasegawa et al. 2003a). Hence, in monocytes, any potential effects of –18483A>C single-nucleotide polymorphism (SNP) on FcεRIα expression through the distal promoter would not be easily masked by strong activity of the proximal promoter and the potent effects of –95T>C and/or –344C>T proximal promoter polymorphisms (Hasegawa et al. 2003b; Bae et al. 2007; Kanada et al. 2008).
Coverage of the FcεRIα gene by the –18483A>C, –344C>T, and –95T>C polymorphisms is shown in Fig. 1a. Linkage disequilibrium (LD) data between FcεRIα –18483A>C, –344C>T, and –95T>C SNPs, estimated in representative population groups of Polish Caucasians (n = 104) and Japanese East Asians (n = 102), are presented Fig. 1b while the allelic frequencies and the distribution of haplotypes of the three FcεRIα polymorphisms are given in Fig. 1c. In all cases, three pairwise haplotypes of –18483A>C, –344C>T, and –95T>C polymorphisms account for 97.7–100.0% of all haplotypes (Fig. 1c). In addition, all of the pairwise D' values are high, ranging from 0.85–1.0 (Fig. 1b). Therefore, it can be assumed that there were no past pairwise recombination events (–344C>T|–95T>C in Poles and Japanese and –18483A>C|–95T>C in Japanese) or that their rates were low (–18483A>C|–344C>T in Poles and Japanese and –18483A>C|–95T>C in Poles). As a result, only four three-loci haplotypes account for 98.4% and 97.7% of all three-locus haplotypes in Poles and Japanese, respectively (Fig. 1b, c).
In most cases, high D' values are not, however, accompanied by r2 values of a similar magnitude, which reflects the substantial pairwise differences in allelic frequencies between FcεRIα SNPs particularly for pairs including the –95T>C SNP in Japanese (Fig. 1b, c). One exception is the r2 value for the –18483A>C|–344C>T pair in Poles, although this is diminished by the presence of recombinant haplotype –18483A|–344 T (1.0%), despite the pairwise difference in allelic frequencies between –18483A>C and –344C>T SNPs in that ethnic group being relatively small. Nevertheless, in both Caucasians and East Asians, the genomic relationships between the three FcεRIα SNPs are close as reflected by the distribution of haplotypes and high D' values.
We speculated that the FcεRIα –18483A>C polymorphism affects the transcriptional activity of the distal promoter localized immediately upstream (Nishiyama et al. 2001; Hasegawa et al. 2003a). To verify our hypothesis, we examined the potential of the –18483A>C polymorphism to affect the binding of transcription factors using electrophoretic mobility shift assay (EMSA) as described previously (Kanada et al. 2008; Fig. 2). Oligonucleotides, antibodies, and basic vectors used in the present study are shown in Supplementary Materials.
EMSA conducted using nuclear extracts from human monocytic U937 cells with competitive oligonucleotides showed preferential binding of nuclear protein(s) to the –18483C allele (Fig. 2a). Thus, we analyzed the sequence surrounding the –18483C allele and noted that it almost perfectly matched the YY1 binding consensus repressive motif while its homology to the consensus YY1 activating site was comparatively low (Shrivastava and Calame 1994; Fig. 3). Indeed, subsequent EMSA with specific antibodies confirmed that the transcription factor binding to –18483C allele was YY1 (Fig. 2b), which was further confirmed using in vitro translated YY1 protein (Fig. 2b) obtained as described previously (Hasegawa et al. 2003a; Kanada et al. 2008). Similar results were obtained using nuclear extracts from human monocytic THP-1 cells (Fig. 2c).
In order to confirm YY1 binding in vivo, we performed chromatin immunoprecipitation (ChIP) assay as described previously (Wang et al. 2008). ChIP assay was conducted using U937 cells confirmed by direct sequencing (Potaczek et al. 2008) to possess –18483CC (Fig. 4). Significantly larger amounts of chromosomal DNA around –18483 were immunoprecipitated by YY1 antibody (Ab) as compared with control Ab, which demonstrated the occurrence of YY1 binding around the –18483C allele in vivo (Fig. 4, center). Although two cis-control regions at ∼350 bp upstream (Fig. 4, left) and downstream (Fig. 4, right) exhibited significant binding with YY1, possibly due to the presence of YY1-bindable sequences (marked with stars; Nishiyama et al. 2001; Hasegawa et al. 2003a), the SNP site showed the highest amount, thus suggesting that detection of the SNP site was not dependent on YY1 binding to the YY1 sequences in the distal promoter and in intron 1A. One cannot, however, exclude the possibility that some of the SNP site detection was derived from the fragments containing binding motifs in the distal promoter or intron 1A.
Next, we examined the potential effects of –18483A>C substitution on transcriptional activity by performing luciferase reporter assay as described previously (Kanada et al. 2008; Wang et al. 2008). We used tandem repeated constructs based on the pGL4.23 [luc2/miniP] plasmid to evaluate the effect of this SNP on transcriptional activity. Briefly, triple motif CCT(A/C)CATGCTACTAAG (–18486/–18471), containing the –18483A>C polymorphism within its genomic neighborhood and covering in length for repressive or activator YY1 consensus binding sites (Shrivastava and Calame 1994), was XhoI/HindIII subcloned upstream of the minimal promoter of pGL4.23 [luc2/miniP] plasmid. In human monocytic THP-1 cells, the luciferase activity of the –18483C allele-specific construct was significantly lower when compared with that of both the –18483A allele-specific construct and the basic (containing no genomic insert) plasmid (Fig. 5a).
Confirmatory analyses were conducted in human basophilic KU812 cells (Fig. 5b), rat basophilic RBL-2H3 cells (Fig. 5c), and mouse mastocytic PT18 cells (Fig. 5d). In all cases, reporter activity of the –18483C allele-specific vector was significantly lower when compared with that of both the –18483A allele-specific construct and the basic plasmid (Fig. 5b–d). In KU812 and PT18 cells, the luciferase activity of the –18483A allele-specific vector was also lower when compared with that of the basic plasmid (Fig. 5b, d). If the C allele is present in the –18483 locus, the nucleotide sequence surrounding the –18483A>C polymorphism possesses strong homology with the YY1 recognition repressor sequence (Shrivastava and Calame 1994) while replacement of C by A in the core YY1 binding motif, corresponding to replacement of the –18483C allele by the –18483A variant, results in a 3–4-fold reduction in YY1 binding (Javahery et al. 1994; Fig. 3). Therefore, YY1 binding to the –18483C allele would be expected to result in a lower transcription rate when compared with the –18483A allele. Indeed, in all four cell lines, the luciferase activity of the –18483C allele-specific vector was lower when compared with both the –18483A allele-specific and basic vectors (Fig. 5a–d).
In the case of KU812 and PT18 cells, –18483A allele-specific constructs also demonstrated lower luciferase activity when compared with the basic vector (Fig. 5b, d). This may be explained by low-affinity binding of the YY1 transcription factor to the –18483A allele, which could be also observed in EMSA (Fig. 2a, c), resulting in a repressive effect on transcriptional activity in some cells. Nevertheless, in all the cases, luciferase activity of –18483C allele-specific constructs was lower than that of –18483A allele-specific vectors (Fig. 5a–d). Although the –18483A>C polymorphism is located downstream of the distal promoter transcription start site, it would not be surprising that binding of YY1 downstream of the transcription initiation site can affect gene expression (Griffioen et al. 2000).
Finally, in order to confirm the effects of YY1 on –18483C allele-mediated transcriptional suppression, reporter assay was performed under YY1 knockdown conditions. Briefly, YY1 siRNA or control siRNA was introduced into THP-1 cells with a reporter plasmid using Nucleofector II (Amaxa, Cologne, Germany) as described previously (Maeda et al. 2006). As shown in Fig. 6, luciferase activity driven by a plasmid carrying the C allele was significantly upregulated by the introduction of YY1 siRNA, whereas the other two promoters, basic and carrying the A allele, were not affected by YY1 siRNA, thus suggesting that transcriptional suppression depending on the C allele at the –18483 locus is regulated by YY1.
YY1 was previously reported to be involved in regulation of several important allergy-related genes. Briefly, Silverman et al. (2004) demonstrated that, being associated with variant YY1 binding and thus altering transcriptional activity, the –509C>T polymorphism of transforming growth factor-β1 was associated with asthma. Guo et al. (2001) and Mordvinov et al. (1999) showed that YY1 transcription factor regulates T cell cytokine gene expression and allergic immune responses. Finally, some members of our group demonstrated the involvement of YY1 in transcriptional regulation of FcεRI subunit expression. Briefly, variant YY1 factor binding to FcεRI β-subunit (FcεRIβ) –654C>T polymorphism was shown to affect its expression (Nishiyama et al. 2004). Moreover, YY1 was demonstrated to contribute to the regulation of FcεRIα transcription through proximal and distal promoters (Nishiyama et al. 2001, 2002; Hasegawa et al. 2003a).
In summary, we demonstrated preferential binding of YY1 to the FcεRIα gene –18483C allele resulting in lower transcriptional activity. The differences in transcriptional activity between –18483C and –18483A alleles are not apparently striking, but they may potentially be of some biological importance in monocytes in which the potential effect of –18483A>C variant on the distal promoter would not be easily masked by strong activity of the proximal promoter (Hasegawa et al. 2003a) and/or potent influences of –344C>T/–95T>C polymorphisms (Hasegawa et al. 2003b; Bae et al. 2007; Kanada et al. 2008). The associations between FcεRIα polymorphism and serum IgE levels and/or allergic disorders (Shikanai et al. 2002; Hasegawa et al. 2003b; Potaczek et al. 2006, 2007a, 2007b; Bae et al. 2007) may result from haplotypic interplay between functional (Hasegawa et al. 2003b; Bae et al. 2007; Kanada et al. 2008) –344C>T, –95T>C, and –18483A>C polymorphisms. A hypothetical mechanism by which alterations in FcεRI(α) expression could affect IgE synthesis/levels remains unknown. Therefore, further functional studies focusing on this issue are necessary.
Polymorphisms in the present study are numbered according to the translation start site; –95T>C and –344C>T polymorphisms were also numbered according to the FcεRIα gene proximal promoter transcription start site as –66T>C and –315C>T, respectively (Hasegawa et al. 2003b; Kanada et al. 2008).