Technical Comments

CCR5 Promoter Alleles and Specific DNA Binding Factors

See allHide authors and affiliations

Science  09 Apr 1999:
Vol. 284, Issue 5412, pp. 223
DOI: 10.1126/science.284.5412.223a

At least 12 single nucleotide polymorphisms (SNPs) within the 5′ upstream regulatory region of the human CCR5chemokine and HIV-1 receptor gene have been described (1–4). Our recent report (1) and others (2) have shown that a common 10-SNP allele haplotype,CCR5P1, with demonstrated promoter activity (5) confers relatively rapid progression to various AIDS end points in a genetic epidemiologic analysis of 2603 patients. Although quantitative differences in expression between CCR5 coding alleles (CCR5-+ versus CCR5-Δ32) are apparent (6,7), we observed no appreciable constitutive differences among CCR5 promoter alleles (P1 andP4) in HIV-1 binding, in chemokine-mediated signal transduction, or in CCR5 quantification (1).

We have found a distinction in specific binding affinity for separate CCR5P allele sequence motifs to nuclear binding (potential transcription) factors, which suggests a possible mechanism for CCR5P1/P1 epidemiologic consequences. We used electrophoretic mobility-shift analysis (8) (EMSA) to assess DNA-protein interactions with the common CCR5P1 (frequency,f = 0.56) and CCR5P4 (f = 0.35) alleles. Synthetic allele-specific oligonucleotides (∼20 bp), representing polymorphic CCR5P sites 208 (G/T), 627 (C/T), 676 (A/G), and 303 (A/G) (2) in the context of their adjacent nucleotides, were incubated with nuclear extracts from phytohemmaglutinin (PHA)-blasted phorbolmyristate acetate (PMA) and ionomyocin-treated human T cells (8). Variants at three sites (303, 627, and 676) showed no difference in binding for alternative allele oligonucleotides; however, the T-bearing oligonucleotide at site 208 (carried in CCR5P3 andP4) (1) displayed 5- to 12-fold greater binding to a specific nuclear binding protein (or proteins) than did the G-bearing oligonucleotide (Fig. 1, complex A). Specificity for the binding was demonstrated by the fact that competitive binding of the coldCCR5P oligonucleotide (208T and 208G), but not of a nonspecific SP-1 oligonucleotide, eliminated complex A formation. Additionally, in cross-competition experiments with 208G versus 208T, as little as 10-fold excess of cold 208T eliminated complex A formation on 208G. However, 100-fold excess of cold 208G only partially competed for complex A formation on 208T.

Figure 1

Differential DNA-protein binding betweenCCR5P1/2 site 208G and CCR5P3/4 site 208T by EMSA with nuclear extracts from PHA-blasted human peripheral blood T cells. Greater DNA-protein binding affinity (complex A) was observed in the 208T site than in the 208G site. NS = non-stimulated; comp. = cold competition with double-stranded oligonucleotides; A = specific complex; B = non-specific complex; C = supershift complex induced by antisera to indicated transcription factors. Complex C specificity was demonstrated by the fact that (i) incubation of radiolabeled oligo and sera alone (no extract) or addition of an irrelevant sera did not resolve the complex, and (ii) peptide competition eliminated complex C formation. Comparable results were obtained with the use of T cell preparations from four distinct human donors. Five gel shift probes were used: 208: AGACAACAGGTTG/TTTTCCGTTTACA; 303: GAGAAAAAGGGGA/GCACAGGGTTA; 627: CGTAAATAAACC/TTCAGACCAG; 676: AGCTCAACTTAAAAA/GGAAGAACTGTTCT; and SP-1:GGGGAGGCGTGGCCTGGGCGGACTGGGGAGTGGCGA.

We examined the sequence surrounding CCR5P site 208 for sequence homology to binding sites for previously described transcription factors (9). The sequence revealed significant homology to sites capable of binding to cRel (a member of the Rel/NF-κB family). We incubated specific antibodies to cRel (10), p50 (11), and p65 (11) with the 208T oligonucleotide. We then performed EMSAs, and each antibody resolved unique complexes in addition to complex A (Fig. 1, complex C). The 208G allele also produced a weak supershift complex with the three antisera, which indicates that this sequence is also capable of binding cRel, p50, and p65 (12). The intensity of the supershift complex (complex C) relative to complex A is similar whether 208G or 208T is tested, and neither allele appreciably diminishes the major allele specific complex A, as measured by supershift experiments that use the Rel/NF-κB family antisera. Therefore, it seems the predominant interaction of CCR5P4 (that is, complex A) involves other yet to be identified binding factors in addition to the three implicated transcription factors.

The identification of differential binding of nuclear factors to oligonucleotides with CCR5P site 208T (retained byCCR5P3 and P4) as compared with CCR5Psite 208G (retained in alleles CCR5P1 and P2) raises the possibility that inclusion of the site 208G inCCR5P1 (and linkage disequilibrium of the site 303A withCCR5P1) would account for the recessive hyper-susceptibility of CCR5P1/P1 homozygotes to rapid progression to AIDS end points (1, 2). If this were so, thenCCR5P1/P1, CCR5P1/P2, andCCR5P2/P2 should each be associated with rapid AIDS progression, insofar as both CCR5P1 and P2alleles contain the 208G nucleotide residue (1). We explicitly tested this prediction by comparing cohort survival curves ofCCR5P1/P1 genotypes with CCR5P1/P2 plusCCR5P2/P2, or with the sum of CCR5P1/P1,CCR5P1/P2, and CCR5P2/P2 genotypes (Table 1). In every case, the CCR5P1/P1 genotypes alone were associated with rapid progression, while the CCR5P2-bearing genotypes did not progress more rapidly than other CCR5Pgenotypes. For this reason, we conclude that the 208G/T polymorphism differential binding to nuclear factors cannot fully explain the reported epidemiological data (1, 2).

Table 1

Influence of CCR5P-208G variant on survival to AIDS end points (outcomes) when included in genotypes bearingCCR5P1 or CCR5P2. Indicated CCR5P1 andCCR5P2 genotypes are on chromosome haplotypes that are wild type (+) for adjacent CCR2-64I and CCR5-Δ32sites. Results are adjusted for the CCR5-Δ32 and CCR2-64I protective effects as a combined variable in the Cox models.

View this table:

Nevertheless, the presence of a mixture of nuclear binding factors which discriminate among CCR5 promoter alleles remains a viable possibility to account for differential availability of CCR5 receptors in various cell populations. The nuclear factors may vary in abundance among different cell types and respond to diverse stimuli that mediate CCR5 transcription. Defining the implications of these events is an important goal of ongoing experiments.


Stay Connected to Science


Navigate This Article