Introduction

The human beta-adrenergic receptors (β-AR) are Gs-protein-coupled receptors that bind catecholamine neurotransmitters and signal transduce by raising intracellular levels of cyclic AMP.1 β-AR are implicated in a variety of catecholamine-mediated physiological functions and in the pathophysiology of obesity,2 asthma3 and cardiovascular disorders.4 β-AR have been classified into β1, β2 and β3 subgroups. β1 receptors are expressed in the heart, kidney, blood vessels and regulate heart rate and vascular tone. β2 receptors are widely distributed in the respiratory tract, and relax smooth muscle in small airways. β3 receptors are found mainly in adipose tissue, where they stimulate lipolysis and thermogenesis.5 Genes encoding the three β-AR subtypes (ADRB1, hCG39839; ADRB2, hCG36934 and ADRB3, hCG21141) are located on chromosomes 10, 5 and 8, respectively. ADRB1 and ADRB2 are nonintronic, and are 3.2 and 3.4 kb, respectively. ADRB3 has one intron and is 3.7 kb in length.

Functional loci have been identified at each of the β-AR genes. Two abundant ADRB1 missense variants Ser49Gly and Gly389Arg6 alter in vitro receptor coupling7 but have no clear in vivo significance. Some positive results were reported using these markers in linkage studies of cardiomyopathy,8 heart and renal failure,7, 9 and hypertension.10 However, no relationship was detected to the response of healthy subjects to drugs acting through the β1-AR11 nor in the hemodynamic response of hypertensive subjects to chronic β1-AR blockade.12 Two obesity studies with the Gly389Arg marker yielded conflicting results.13, 14

Within the coding region of ADRB2, nine SNPs were identified,15 five of which are synonymous. Missense substitutions were Arg16Gly, Gln27Glu, Val34Met, and Thr164Ile.16, 17 Among them, two common alleles have been shown to be functional in vitro: Gly16 leads to enhanced agonist-mediated downregulation, and Glu27 reduces such regulation.18 These polymorphisms have been associated with a variety of β-AR-related phenotypes, but association results have been inconsistent across different studies. Arg16Gly was associated with obesity,19, 20 diabetes21 and cystic fibrosis,22 but not with plasma norepinephrine concentration23 or agonist-induced beta 2AR desensitization.24 Evidence regarding the Arg16Gly polymorphism's relationship with asthma is conflicting.25, 26 Gln27Glu was associated with hypertriglyceridemia27 and obesity in Spanish men28 but not in the Tongan population.29

ADRB3 Trp64Arg is located in the first intracellular loop of the receptor. Arg64 has higher allele frequencies in Pima Indians [0.31] as compared to Mexican Americans [0.13], African Americans [0.12], and Caucasians [0.08]30 supporting the idea that this variant could impair activation of thermogenesis in adipose cells, contributing to the high frequency of obesity and adult onset diabetes in the Pima Indians.31 However, the linkage studies are contradictory.32

Taken together, this evidence illustrates that a consistent picture of β-AR genotype–phenotype relationships has yet to emerge. Other functional loci may be present, including polymorphisms which are known but which have not yet been recognized to be functional. A haplotype approach combining known functional polymorphisms with a series of loci chosen for haplotype informativeness could comprehensively capture the potential information content on β-AR functional variants of moderate abundance.33 In this study, we report a haplotype map for each of the β-AR genes for two populations, American Caucasians and African Americans, by genotyping a panel of SNP markers and the known functional polymorphisms in these populations. For each gene, we also describe marker panels that maximize haplotype information content.

Materials and methods

Participants

A total of 192 unrelated subjects were genotyped, including 96 individuals from each of two populations: US Caucasians and African Americans. Informed consent was obtained according to human research protocols approved by the human research committees of the recruiting institutes, including the National Institute on Alcohol Abuse and Alcoholism, National Institute of Mental Health and Rutgers University. All participants had been psychiatrically interviewed and none had been diagnosed with a psychiatric disorder.

SNP markers

The physical position and frequency of minor alleles (>0.05) from a commercial database (Celera Discovery System, CDS, September, 2003) were used to select SNPs (including known nonsynonymous substitutions). 5′ nuclease assays (vide infra) were designed for seven ADRB1, 11 ADRB2 and nine ADRB3 SNPs and optimized. These markers were nearly equally spaced and covered the entire genes plus 2.5–6 kb upstream and 2.5–6 kb downstream from each gene.

Genomic DNA

Genomic DNA was extracted from lymphoblastoid cell lines, diluted to a concentration of 10 ng/μl. Aliquots of 1 μl were dried in 384-well plates.

Polymerase chain reaction (PCR) amplification

Genotyping was performed by the 5′ nuclease method34 using fluorogenic allele-specific probes. Oligonucleotide primer and probe sets were designed based on gene sequence from the CDS, September 2003. Primers and detection probes for each locus in each gene are listed in Table 1a–c.

Table 1 Primer and probe sequences for 5′ nuclease genotyping
Full size table

Reactions were in a 5 μl volume containing 2.375 μl TE, 2.5 μl Master Mix (ABI, Foster City, CA, USA) with AmpliTaq Gold® DNA Polymerase, dNTPs, Gold Buffer and MgCl2, 10 ng genomic DNA, 900 nM of each forward and reverse primer and 100 nM of each reporter and quencher probe. DNA was incubated at 50°C for 2 min and at 95°C for 10 min, and amplified on an ABI 9700 device for 40 cycles at 95°C for 30 s and 60°C for 75 s. Allele-specific signals were distinguished by measuring end point 6-FAM or VIC fluorescence intensities at 508 and 560 nm, respectively, and genotypes were generated using Sequence Detection V.1.7 (ABI).

Genotyping error rate was directly determined by regenotyping 25% of the samples, randomly chosen, for each locus. The overall error rate was <0.005. Genotype completion rate was 0.99.

Haplotype analysis

Haplotype frequencies were estimated using a Bayesian approach implemented with PHASE.35 These frequencies closely agreed with results from a maximum likelihood method implemented via an expectation-maximization (EM) algorithm.36 Haploview version 2.0.2 (Whitehead Institute for Biomedical Research, USA) was used to produce LD matrices.

Results and discussion

Of a total of 27 markers in three β-AR genes, 23 were polymorphic both in US Caucasians and African Americans. ADRB3 marker #2 (rs4999) was monomorphic in Caucasians, and ADRB3 markers 7–9 (rs4993, rs802162 and rs13258937) were monomorphic in both populations. Dramatic interpopulation differences in allele frequencies were observed for many of the markers. Allele frequencies of all markers and their locations in the genes are shown in Table 2a–c. For ADRB1, two functional nonsynonymous polymorphisms (Ser49Gly and Ala389Gly) are located in the exon, one marker is located in the gene 3′ UTR region, and the rest of the markers are in the intergenic region upstream and downstream of ADRB1 (Figure 1a). For ADRB2, two functional nonsynonymous polymorphisms (Arg16Gly and Gln27Glu) and two synonymous polymorphisms are located in the exon, one marker is located in the gene 5′ UTR region, and the rest of the markers are in the intergenic region upstream and downstream of ADRB2 (Figure 1b). For ADRB3, one functional nonsynonymous polymorphism (Arg64Trp) is located in exon 1, one marker is located in the 5′ UTR region (exon 1), two markers are located in the gene 3′ UTR region (exon 2), and the rest of the markers are in the intronic sequence and intergenic region upstream and downstream of ADRB3 (Figure 1c).

Table 2 Locations and allelic frequencies in 96 individuals from each of two populations
Full size table
Figure 1
figure 1

Locations of single-nucleotide polymorphisms genotyped in ADRB1, ADRB2 and ADRB3. Coding exons are shown as solid blocks. Physical locations are from the Celera Discovery System [CDS] database, September 2003. *ADRB3 is transcribed in the reverse direction.

Full size image

Within the ADRB1, ADRB2 and ADRB3 regions, a single conserved haplotype block spanned each gene in both Caucasians and African Americans (Figure 2a–c) and the block boundaries extend beyond the region we have evaluated. In African Americans, the ADRB2 block may be smaller; the first and last SNPs were in lower linkage disequilibrium (LD) [D′<0.8] with all other markers. Definition of haplotype blocks and block boundaries is inexact. Isolated nucleotide substitutions can occur within nonrecombined blocks. On the other hand, some disruptions of LD occurring within blocks are attributable to low allele frequencies that lead to increased variance in estimation of LD. We discounted low D′ values which might have originated from this cause. In the ADRB1, ADRB2 and ADRB3 haplotype block regions, D′ was generally >0.80 from one end of the region to the other. Average D′ values within haplotype blocks in Caucasians and African Americans were, respectively, ADRB1: 0.98 and 0.84, ADRB2: 0.98 and 0.87, ADRB3: 1.00 and 0.74. Median D′ values within the haplotypes blocks from both Caucasians and African Americans were high: ADRB1: 1.00 and 1.00, ADRB2: 1.00 and 1.00, and ADRB3: 1.00 and 0.93, indicating that most pairs of loci within these regions are in very high LD.

Figure 2
figure 2

(a–c) Haplotype block organization of ADRB1, ADRB2 and ADRB3. Each box represents % LD [D′] between pairs of markers, as generated by Haploview (Whitehead Institute for Biomedical Research, USA). D′ is color coded, red box indicating complete [1.00] D′ between locus pairs. *ADRB3 marker #2, monomorphic in Caucasians, is excluded from the Caucasian haplotype map, but included, for comparability to African Americans, in all haplotypes (see Table 3c).

Full size image

Haplotype frequencies for ADRB1, ADRB2 and ADRB3 in both populations are shown in Table 3a–c. For each population and haplotype block, 2–5 common (frequency ⩾0.05) haplotypes accounted for most of the total: 88–100% of Caucasian and 88–96% of African-American haplotypes. For US Caucasians and African Americans, the numbers of common (frequency ⩾0.05) haplotypes were: in ADRB1, 3 and 5; in ADRB2, 4; in ADRB3, 2 and 4, respectively. Population differences in haplotype frequencies are clearly illustrated in Figure 3a–c.

Table 3 Frequencies of haplotypes
Full size table
Figure 3
figure 3

(a–c) Frequencies of common β-AR haplotypes in US Caucasians and African Americans.

Full size image

The marker panels we genotyped were sufficient to capture diversity in all blocks in the two populations we studied. We evaluated haplotype diversity within each block by successively subtracting SNPs from the haplotypes to evaluate the increment/decrement in diversity contributed by each SNP. SNPs were serially subtracted in that order that minimized the decrement in diversity at each step, and until only a single SNP (ie the SNP with the highest heterozygosity) remained. The chosen measure of diversity (haplotype frequencies and diplotype heterozygosity) was recalculated for each size SNP panel [n, n−1,…,1]. At some point for each haplotype block and for each population, adding or subtracting a SNP does not appreciably alter diversity, as shown in Figure 4, panels A-C. For ADRB1and ADRB2, haplotype diversity was highest in African Americans. A similar number of markers (2–4) was sufficient to capture maximum diversity in either population. This number represents an optimal panel, itself derived from the larger panel of SNP markers we genotyped. The minimum SNP set necessary to maximize haplotype diversity was also determined using SNPTagger.37 The SNPs that constitute this minimal set are indicated in Table 2a–c.

Figure 4
figure 4

(a–c) Effect of successive subtraction/addition of SNPs on β-AR haplotype diversity in two populations. SNPs were successively subtracted from haplotypes in such a way as to minimize loss of diversity (diplotype heterozygosity, Y-axis). Panel (a) ADRB1, panel (b) ADRB2 and panel (c) ADRB3. For each block, marker panels are sufficient to maximize diversity, and diversity can in fact be maximized with 2–4 optimal markers. For each haplotype panel, addition of the functional β-AR locus (or loci) yields no further increment in diversity.

Full size image

For each β-AR gene, extensive amounts of resequencing have been performed and missense polymorphisms are known within each gene.38, 39, 40, 41, 42, 43 However, resequencing has been largely confined to the coding regions and to only a few populations. Although a complete inventory of common missense variants may be available in Caucasians and African Americans, unknown loci affecting function may be present, and some loci that are known may have unrecognized functional significance. Individual SNP loci provided some ability to capture information on the missense polymorphisms known at each gene (r2 values ranged in Caucasians and African Americans, respectively: ADRB1 Gly49Ser: 0.04–1.00 and 0.13–0.71, ADRB1 Gly389Ala: 0.03–0.84 and 0.03–0.59, ADRB2 Gly16Arg: 0.01–0.83 and 0.06–0.5, ADRB2 Gln27Glu: 0.14–0.94 and 0.02–0.23, ADRB3 Trp64Arg: 1.00 and 0.01–0.95, and as shown in Table 4a–c). Haplotypes enabled high sensitivity of detection of the missense substitutions (when a missense allele was present a particular haplotype(s) was present) and specificity of detection (when a haplotype(s) was present the missense allele was present). For each of the three β-AR genes, the haplotype was capable of capturing all or almost all the information provided by directly genotyping the missense loci, in either population (Table 5a, b). It is therefore likely that the SNP panels covering β-AR gene regions would capture information on unknown functional alleles. Certainly, genotyping of polymorphisms that affect gene expression and/or function is highly important in association/linkage studies. However, there is a possibility that an unrecognized functional locus contributes to a phenotype. The focus of the haplotype-based approach to analyzing case–control populations has been to detect the effects of every functional locus, known or unknown.

Table 4 R2 values for functional markers versus noncoding SNPs
Full size table
Table 5 Effect of functional markers on haplotype diversity
Full size table

For the β-AR genes, we have created multilocus SNP panels to define haplotype structure across each gene region. Each panel is sufficient to capture the signal of the moderately abundant missense alleles and unknown functional loci. The β-AR gene haplotype maps and marker panels provide a basis for future studies to investigate the role of genetic variation in physiology and pathophysiology related to β-AR function.