Polymorphisms of cardiac presynaptic α_2C adrenergic receptors: Diverse intragenic variability with haplotype-specific functional effects

Abstract

The presynaptic α_2C adrenergic receptors (AR) act to inhibit norepinephrine release in cardiac and other presynaptic nerves. We have recently shown that a genetic variant in the α_2CAR coding region (Del322-325), which renders the receptor partially uncoupled from G_i, is a risk factor for heart failure. However, variability of heart failure phenotypes and a dominance of Del322-325 in those of African descent led us to hypothesize that other regions of this gene have functional polymorphisms. In a multiethnic population, we found 20 polymorphisms within 4,625 bp of contiguous sequence of this intronless gene encompassing the promoter, 5′ UTR, coding, and 3′ UTR. These polymorphisms occur in 24 distinct haplotypes with complex organizations, including multiple 5′-upstream polymorphisms in regions known to direct expression, a 3′ UTR substitution polymorphism within an insertion/deletion sequence, and the radical coding polymorphism that deletes four amino acids. Relatively low linkage disequilibrium between many polymorphisms, few cosmopolitan haplotypes, prevalent ethnic-specific haplotypes, and substantial genetic divergence among haplotypes was noted. The dysfunctional Del322-325 allele was partitioned into multiple haplotypes, with frequencies of 48% to 2%. The functional implications of the haplotypes were ascertained by whole-gene transfections of human neuronal cells, where haplotype was significantly related (P < 0.001) to expression levels of receptor transcript and protein. Expression varied by as much as ≈50% by haplotype, and such studies enabled haplotype clustering by phenotypic, rather than genotypic, similarities. Thus, depending on phenotype, expression-specific haplotypes may amplify, attenuate, or dominate the cardiomyopathic effect attributed to the α_2CDel322-325 marker.

The adrenergic receptors (ARs), whose endogenous ligands are epinephrine and norepinephrine, are members of the superfamily of seven transmembrane spanning G protein-coupled receptors. ARs modulate a host of functions relative to the sympathetic nervous system, including neurotransmitter release, and cardiac, vascular, pulmonary, renal, metabolic, and central nervous system function. For over a decade, it has been known that the expression and/or function of ARs varies considerably between individuals (1), even when studies are carried out under stringent conditions in normal individuals. Similarly, the physiologic or clinical response to receptor agonists and antagonists shows marked interindividual variability (2, 3). One explanation for such variability is that ARs are dynamically regulated by multiple mechanisms so as to maintain homeostasis, and that individuals may be under different environmental influences such that receptor expression or function is altered. However, considerable evidence has accumulated over the last few years suggesting that interindividual differences in receptor expression or function is based on genetic variability of the genes encoding these receptors (4).

Of recent interest has been genetic variability of the α₂AR in heart failure syndromes (5-7). The α_2A- and α_2CAR subtypes are localized to presynaptic nerve terminals and participate in a negative feedback loop regulating norepinephrine release (8). In settings of high sympathetic activity, these subtypes serve to dampen further norepinephrine release. One such setting is human heart failure of virtually all etiologies, where marked sympathetic drive develops as a response to low cardiac output and systemic perfusion (9). However, persistent stimulation by norepinephrine of postsynaptic β₁AR expressed on cardiomyocytes leads to multiple deleterious signaling events in the heart, ultimately leading to a worsening of pump function and clinical deterioration (9, 10). The critical roles of the α₂ARs in modulating norepinephrine release and its cardiac consequences have been delineated in mice lacking one or both of these subtypes. α_2A/α_2C double knockout mice develop a catecholamine-mediated dilated cardiomyopathy, indicating the necessity of this mechanism for regulation of norepinephrine release under nonstressed conditions (8). In addition, α_2CAR^-/- mice develop a lethal cardiomyopathy when subjected to pressure overload via aortic banding (7), which is consistent with uncontrolled norepinephrine release from cardiac presynaptic nerves acting to persistently stimulate the pressure-compromised heart. Thus, a genetic polymorphism of the human α_2CAR gene that leads to decreased expression or function might act to predispose individuals to heart failure, or to potentiate failure due to other causes such as myocardial infarction or hypertension.

We have recently delineated a radical human polymorphism within the third intracellular loop of the α_2CAR, which consists of an in-frame 12-nt deletion resulting in loss of Gly-Ala-Gly-Pro from this G protein-coupling domain of the receptor (11). In Chinese hamster ovary (CHO) cells recombinantly expressing wild-type and this polymorphic α_2CAR (denoted α_2CDel322-325), we have shown that the variant receptor has substantially decreased coupling to its cognate G protein G_i, with decreased inhibition of adenylyl cyclase and decreased stimulation of mitogen-activated protein kinase (11). These results have prompted several human studies to ascertain the relevance of the α_2CDel322-325 to norepinephrine release or synaptic content, and the risk for heart failure phenotypes. The polymorphism has been associated with increased cardiac [¹²⁵I]MIBG (a radiolabeled norepinephrine analog) imaging in heart failure patients (6), earlier onset of heart failure (5), and heart failure severity or progression (5, 7). In addition, we have shown (5) that the α_2CDel322-325 is an independent risk factor for the development of heart failure in African Americans (odds ratio ≈5 compared to healthy African Americans), and that such risk may be synergistic with a hyperfunctional β₁AR variant.

In the aforementioned studies, it has been recognized that, even in α_2CDel322-325 groups, there is noticeable variability in the phenotype. We have considered, then, that there are other functionally relevant polymorphisms in the α_2CAR gene, potentially including the 5′ and 3′ flanking regions that form haplotypes that regulate expression and ultimately the signaling phenotype. In the current work, this hypothesis was pursued.

Methods

Polymorphism Discovery. The reference sequence for the intronless human α_2CAR gene is that of GenBank accession number AY605898. In this report, the first nucleotide of the ORF is nucleotide 1 (corresponding to nucleotide 2638 of the reference sequence), with the 5′ UTR beginning at -1 and continuing in the negative direction. For polymorphism discovery, direct sequencing of PCR fragments spanning the α_2CAR gene amplified from genomic DNA from an index repository consisting of 30 ethnically diverse individuals was performed by using an ABI Prism 3700 Sequencer (Applied Biosystems). The analyzed sequence included an ≈1,800-bp promoter, ≈900-bp 5′ UTR, 1,386-bp coding, and the 3′ region from the stop codon to the poly(A) termination sequence (≈600 bp). Primer sets and reaction conditions used for polymorphism discovery are shown in Table 3, which is published as supporting information on the PNAS web site. For some fragments, direct sequencing yielded equivocal results, so PCR fragments were cloned into the vector pCR2.1-TOPO and multiple colonies from transformed bacteria were expanded. Isolated DNA from at least five independent clones was subsequently sequenced. Electronic data from sequencing was compared to the reference sequence by using macvector 7.2 (Accelrys, San Diego), and variants were confirmed by direct inspection of the electropherograms.

Genotype and Haplotype Determination. Unphased genotypes were determined by using an expanded index repository of 104 anonymous genomic DNA samples (Coriell Institute) representing 40 Caucasians, 40 African Americans, and 24 Asians. For the identified polymorphisms, the presence of the variant allele in a PCR product resulted in gain or loss of a restriction enzyme site, thereby providing a rapid genotyping method. See Table 4, which is published as supporting information on the PNAS web site for assay conditions. In instances of repeated homozygosity, the chromosomal organization of some polymorphisms was evident by inspection. In cases of multiply heterozygous samples, haplotypes were determined by a molecular approach. This was carried out by PCR amplification of the gene (4,625 bp) by using the Elongase Amplification System (Invitrogen) followed by subcloning of the product and subsequent genotyping of single clones, yielding phased genotypes. We also compared the composition of haplotypes delineated by this molecular approach with haplotypes imputed from unphased genotype data by an algorithmic method (12) by using the computer program phase (University of Washington, Seattle).

Constructs and Cell Transfections. For expression studies, PCR products of the 4,625-bp α_2CAR gene derived from human genomic DNA were cloned into the vector pCR2.1-TOPO. This vector lacks a eukaryotic-responsive promoter, and thus, expression of these constructs is directed by the included α_2CAR promoter sequence. Each construct was sequenced to verify the haplotype and to ensure that it differed from other constructs only at the appropriate polymorphic sites. The human neuroblastoma cell line BE(2)-C was used as the host cell for transient transfections and were grown in monolayers as described (13). Transient transfections with Lipofectamine 2000 (Invitrogen) were performed by using 12 μg of each haplotype construct as described by the manufacturer. Two days after transfection, cells were harvested for radioligand binding or real-time RT-PCR.

α_2CAR Transcript and Protein Expression. Reverse transcription of total RNA from BE(2)-C cells was performed after DNase I digestion using murine leukemia virus (MuLV) reverse transcriptase and random hexamers. Identical incubations were also performed in the absence of reverse transcriptase. Forward and reverse primers (5′-TTTGCACCTCGTCGATCGT and 5′-CCTGCGTCACCGACCAGTA) for quantitative real-time PCR amplified a 66-bp fragment from an invariant portion of the α_2CAR coding region. PCRs consisted of a 50-fold dilution of the reverse transcriptase reaction (≈10 ng), 300 nM each forward and reverse primers, and 25 μl of Sybr Green Master Mix (Applied Biosystems). Cycle conditions were 50°C for 2 min, 95°C for 10 min, and 40 cycles of 95°C for 15 sec and 60°C for 1 min. Each sample was assayed in triplicate. For each amplification, a standard curve was generated by using 0.1-100 ng cDNA prepared from stock BE(2)-C total RNA to determine the relative levels of α_2CAR present in each sample. Reactions containing primers corresponding to endogenous GAPDH (5′-ATGGAAATCCCATCACCATCTT and 5′-CGCCCCACTTGATTTTGG) were also performed to control for small variations in the amount of template in each reaction. α_2CAR protein expression was determined by [³H]yohimbine radioligand binding with cell membranes as described (11).

Miscellaneous. Agreement between genotypes and those predicted by Hardy-Weinberg equilibrium were assessed by χ² tests with 1 df. Linkage disequilibrium was calculated as Δ (14). Haplotype-tagged polymorphisms were determined by using the computer program snp-tagger (15). Comparisons of results from radioligand binding and real-time RT-PCR were analyzed by repeated measures ANOVA and post hoc t tests. Significance was considered when P < 0.05 after correction for multiple comparisons by the Newman-Keuls method. Data are reported as mean ± SE.

Results and Discussion

Multiple Polymorphic Sites Are Present Throughout the α_2CAR Gene. Polymorphisms of the α_2CAR gene were ascertained from a total of 104 individuals composed of three ethnic groups. The interrogated sequence of this intronless gene consisted of ≈4,625 bp, representing contiguous sequence of promoter, 5′ UTR, coding, and the 3′ UTR. In doing so, multiple sequence variants were identified with allele frequencies of ≥1% compared to the reference sequence in at least one of the ethnic groups. The allele frequencies, locations, and identifier codes for the polymorphisms are shown in Table 1. Three single nucleotide polymorphisms (SNPs) that do not alter the encoded amino acid (synonymous SNPs) were noted in the coding region at nucleotide positions 933, 996, and 1167, but are not included in Table 1 and are not further discussed. Of the other 20 variants, 18 were SNPs, and two were deletion polymorphisms. Within the α_2CAR coding region, we identified only one nonsynonymous variant (position q, Table 1), which is the previously described 12-bp in-frame deletion polymorphism of nucleotides 964-975 that results in loss of amino acids 322-325 (11). In the noncoding regions of the α_2CAR gene, 13 sequence variants were found within the α_2CAR promoter region, three variants were found within the 5′ UTR, and three variants were found within the 3′ UTR. In the 3′ UTR, a 21-bp deletion of nucleotides 1483-1503 was identified 94 bp downstream of the α_2CAR translational stop site (location r). Interestingly, another 3′ UTR SNP at position 1486 (location s) occurs within this deletion region; thus, in the presence of the r polymorphism, the s SNP cannot be present on the same chromosome. The distribution of homozygous and heterozygous alleles in each ethnic group was not different from that predicted by the Hardy-Weinberg equilibrium (P > 0.10).

Table 1. Locations, base changes, and allele frequencies of variable loci within the α_2CAR gene.

				Minor allele frequency
Region	Location code	Nucleotide	Alleles	Ca	AA	As
Promoter	a	−2579	T/C	8.8	6.3	35.4
	b	−2416	C/G	0	10.0	0
	c	−2357	C/T	0	10.0	0
	d	−2280	G/T	0	5.0	0
	e	−2069	C/T	7.5	2.5	35.4
	f	−2064	C/T	0	0	2.1
	g	−1926	G/A	1.3	1.3	2.1
	h	−1692	T/G	6.3	38.8	8.3
	i	−1513	T/G	2.5	0	0
	j	−1262	C/A	0	7.5	0
	k	−965	G/C	0	11.3	0
	l	−940	G/A	7.5	2.5	35.4
	m	−933	C/A	1.3	1.3	2.1
5′ UTR	n	−696	C/G	11.3	0	0
	o	−241	C/G	0	7.5	0
	p	−230	T/C	5.0	2.5	6.3
Coding	q	+964 to +975	Ins/Del	6.3	42.5	12.5
3′ UTR	r	+1483 to +1503	Ins/Del	8.8	5.0	42.5
	s	+1486	T/C*	34.2	71.1	30.4
	t	+1757	G/C	31.3	67.5	20.8

Ca, Caucasian; AA, African-American; As, Asian; Ins, insertion; Del, deletion.

^*SNP at s is located within the insertion sequence site r only, so in the deletion form of r, the s genotype is not applicable.

Examination of the frequencies of each α_2CAR polymorphism revealed that some were cosmopolitan (present in all ethnic groups), whereas others displayed a high degree of population specificity (Table 1). Eleven polymorphisms (≈55%) were present in all ethnic groups and represented variation in all regions of the α_2CAR gene. For many cosmopolitan polymorphisms, there were nevertheless substantial differences in the allele frequencies between ethnic groups. In African Americans, polymorphisms at positions h, q, s, and t were present at significantly higher allele frequencies compared to Caucasians and Asians (P < 0.001). As discussed below, these four variants also show a high degree of linkage disequilibrium within this population. In Asians, polymorphisms at positions a, e, l, and r are significantly more common in this group as compared to Caucasians and African Americans (P < 0.001). No single cosmopolitan variant was more common in Caucasians compared to either African Americans or Asians. When examining population-specific variants, African Americans had the highest number (six), with two present in Caucasians, and a single population-specific variant observed for Asians. Overall, the frequencies of polymorphisms (both cosmopolitan and population-specific) in each group ranged from 0 to as high as 70%. Indeed, SNP locations s and t in African Americans are very common, in which case the minor alleles in the total population are actually the major alleles in African Americans. The diversity of the α_2CAR gene is greater than that reported from resequencing DNA from a multiethnic population at 313 genes randomly selected from the genome (16). Given that the α_2CAR has only one nonsynonymous coding variation, it is clear that the majority of the genetic plasticity of this gene has evolved because of polymorphisms in the noncoding regions.

Fig. 1 A-C shows a graphical display of α_2CAR genotypes for each DNA sample grouped by population. Blue and yellow indicate homozygotes for the reference and variant alleles, respectively, and red indicates heterozygotes. Individual DNA samples are clustered to show sites with similar genotype patterns. Genotype data were used to calculate the degree of linkage disequilibrium (Δ) between all pairs of polymorphic loci (Fig. 1 D-F). Of the 20 variants, relatively few pairs exhibit high levels of linkage disequilibrium (defined as Δ value ≥0.8). As shown in Fig. 1, among all ethnic groups linkage disequilibrium was observed for a/e/l/r, g/m, h/p/q, and s/t. At sites a/e/l/r and h/p/q, the degree of linkage disequilibrium observed in Caucasians and Asians was high between all possible pairs. However, in African Americans, the Δ values for these pairs were significantly lower. For example, Δ was 0.89 for h/p in Caucasians, but was 0.20 for African Americans. Unique combinations of polymorphisms showing strong linkage disequilibrium only in African Americans were also noted and included the paired combinations of b, c, j, and k. When considering the Del322-325 polymorphism (q), the SNP at location h is in strong linkage disequilibrium with this polymorphism in all ethnic groups and thus may represent an alternate “causative” SNP for the cardiac phenotype ascribed to Del322-325. Interestingly, in Caucasians and Asians only, the SNP at location p also showed high linkage disequilibrium with Del322-325.

Fig. 1.

Visual depiction of the organization of α_2CAR polymorphisms into haplotypes stratified by ethnic group. In A-C, each column represents a polymorphism location as labeled, and each horizontal row represents one of the 104 individuals in the cohort (not labeled). Homozygosity for the reference allele is colored blue, heterozygosity is red, and homozygosity for the polymorphism is yellow. Gray indicates that, in the presence of the homozygous deletion at position r, the s genotype is not applicable. In D-F, the degree of linkage disequilibrium (Δ) between any two polymorphisms is shown (note scale). White indicates that the genotype data are noninformative in the indicated cohort, and thus linkage disequilibrium is indeterminate. Ca, Caucasian; AA, African-American; As, Asian.

α_2CAR Polymorphisms Are Organized into 24 Haplotypes. These 20 variable loci of the α_2CAR gene were found to be organized into a total of 24 haplotypes within the entire cohort as determined by the definitive molecular method (Table 2). The raw unphased genotypes from individuals with polymorphism with frequencies ≥5% in any one population were also used to infer haplotypes by the algorithmic method, and we found complete concordance between the two approaches. The haplotypic variation of this gene is substantial, particularly in African Americans and Asians, where the most common haplotype represents only 24% and 40% of individuals, respectively. As shown in Table 2, there is substantial ethnic variation in α_2CAR haplotype frequencies. Only three haplotypes (1, 2, and 8) are present in all three ethnic groups, and the frequencies at which these haplotypes occur in each group varies considerably. Haplotypes 1 and 2 are the most common haplotypes in all groups and differ in frequency between groups by up to ≈2-fold. For haplotype 8, there is greater disparity, with allele frequencies of 27% in Asians compared to 2.5% and 7.5% in African Americans and Caucasians, respectively. In terms of population-specific haplotypes, there are 10 in African Americans and four each in Caucasians and Asians. The extent of haplotype divergence within the α_2CAR gene is shown in Fig. 2. Here, for clarity, the deletions are considered one “site.” The number of sites that differ between any two haplotype pairs varied from a low of one site (e.g., haplotypes 9 and 3) to a high of 11 sites (e.g., haplotypes 22 and 5). Haplotype divergence, calculated based on actual differences in the number of nucleotides, ranged from 0.022% to 0.86%.

Table 2. Hyplotypes of the α_2CAR gene in 104 individuals.

Haplotype no.	Location code
	a	b	c	d	e	f	g	h	i	j	k	l	m	n	o	p	q	r	s	t	AA	Ca	As
1	T	C	C	G	C	C	G	T	T	C	G	G	C	C	C	T	Ins	Ins	T	G	23.8	58.8	39.6
2	T	C	C	G	C	C	G	T	T	C	G	G	C	C	C	T	Ins	Ins	C	C	13.8	13.8	12.5
3	T	C	C	G	C	C	G	G	T	C	G	G	C	C	C	T	Del	Ins	C	C	21.3	0	8.3
4	T	C	C	G	C	C	G	G	T	C	G	G	C	C	G	T	Del	Ins	C	C	7.5	0	0
5	T	G	T	G	C	C	G	T	T	A	C	G	C	C	C	T	Ins	Ins	C	C	7.5	0	0
6	T	C	C	G	C	C	G	T	T	C	G	G	C	C	C	T	Del	Ins	C	C	5	0	0
7	T	C	C	T	C	C	G	G	T	C	G	G	C	C	C	T	Del	Ins	C	C	5	0	0
8	C	C	C	G	T	C	G	T	T	C	G	A	C	C	C	T	Ins	Del	NA	G	2.5	7.5	27.1
9	T	C	C	G	C	C	G	G	T	C	G	G	C	C	C	C	Del	Ins	C	C	2.5	3.75	0
10	T	C	C	G	C	C	G	T	T	C	G	G	C	C	C	T	Ins	Del	NA	G	1.25	0	2.1
11	T	C	C	G	C	C	G	G	T	C	G	G	C	C	C	T	Ins	Ins	T	G	2.5	0	0
12	T	G	T	G	C	C	G	T	T	C	C	G	C	C	C	T	Ins	Ins	C	C	2.5	0	0
13	T	C	C	G	C	C	G	T	G	C	G	G	C	C	C	T	Ins	Ins	T	G	0	2.5	0
14	T	C	C	G	C	C	G	T	T	C	G	G	C	G	C	T	Ins	Ins	C	C	0	11.3	0
15	C	C	C	G	C	C	G	G	T	C	G	G	C	C	C	T	Del	Ins	C	C	1.25	0	0
16	C	C	C	G	C	C	G	T	T	C	G	G	C	C	C	T	Ins	Ins	C	C	1.25	0	0
17	C	C	C	G	C	C	G	T	T	C	G	G	C	C	C	T	Ins	Del	NA	G	1.25	0	0
18	T	C	C	G	C	C	A	T	T	C	G	G	A	C	C	T	Del	Ins	T	G	1.25	0	0
19	C	C	C	G	C	C	G	G	T	C	G	G	C	C	C	C	Del	Del	NA	C	0	1.25	0
20	T	C	C	G	C	C	A	G	T	C	G	G	A	C	C	T	Del	Ins	C	C	0	1.25	0
21	C	C	C	G	T	C	G	T	T	C	G	A	C	C	C	C	Ins	Del	NA	G	0	0	4.2
22	C	C	C	G	T	C	A	T	T	C	G	A	A	C	C	T	Ins	Del	NA	G	0	0	2.1
23	C	C	C	G	T	T	G	T	T	C	G	A	C	C	C	T	Ins	Del	NA	G	0	0	2.1
24	T	C	C	G	C	C	G	T	T	C	G	G	C	C	C	C	Ins	Ins	T	G	0	0	2.1

Shown are the allele frequencies (in %) of the haplotypes for each ethnic group. See Table 1 for location code and alleles. AA, African-American; Ca, Caucasian; As, Asian; NA, not applicable because of deletion at r; Ins, insertion; Del, deletion.

Fig. 2.

Divergence between haplotype pairs of the α_2CAR. Each cell represents a haplotype pair in the matrix and shows the number of different polymorphic sites that differ between the two haplotypes.

We found a complex distribution of the cardiomyopathic α_2CDel322-325 polymorphism within the context of these haplotypes. Del322-325 is present primarily in haplotypes 3, 4, 6, and 7 (haplotype frequencies ≥5%) and at lower frequencies in haplotypes 9, 15, and 18, in the African-American group. In Caucasians, Del322-325 is present only in haplotypes 9, 19, and 20, whereas in Asians it is present exclusively in haplotype 3. Overall, the Del322-325 polymorphism is partitioned into nine different haplotypes. Fig. 3 shows the number of individuals with a specific Del322-325 containing haplotype as a percentage of all those with the Del322-325 genotype. Approximately 48% of those with the Del322-325 are haplotype 3, and the remainder are distributed amongst several other haplotypes with frequencies varying from ≈14% to ≈2% of the Del322-325 population. Thus, heterogeneity within the context of the cardiomyopathic Del322-325 polymorphism is substantial. The minimum number of polymorphic sites that need to be interrogated to define all of the haplotypes (haplotype-tagged polymorphisms) is 15. Excluding haplotypes with an allele frequency <5% in all three ethnic groups results in nine haplotypes (1, 2, 3, 4, 5, 6, 7, 8, and 14), which requires eight genotypes to so define (positions a, b, d, h, n, o, q, and t).

Fig. 3.

The α_2CDel322-325 polymorphism is partitioned into nine haplotypes. Shown is the distribution of Del322-325-containing haplotypes relative to all subjects in the index population with the Del322-325 polymorphism. Haplotype 3 encompasses 43% of all Del322-325 individuals, and the rest are represented by other haplotypes with frequencies between ≈14 and ≈2%.

Haplotypes Direct Cellular Expression of α_2CAR Transcript and Protein. The α_2CAR promoter and 5′ UTR polymorphisms identified in this report lie within an ≈2-kb region that has been shown to direct transcription of a luciferase reporter construct in human cell lines (17). For the rat α_2CAR, translational processing has been shown to be regulated by a region in the 3′ UTR sequence (18), analogous to where we found the human polymorphisms. We were thus prompted to ascertain the effects of polymorphic variation within the α_2CAR gene on receptor expression. For such studies, we felt that it was critical to consider polymorphisms within a haplotype context, because it is the net effect from multiple variable sites that potentially direct a phenotype. Thus, we used a strategy whereby the constructs for transfections are devoid of viral or other promoters, but rather consist of the entire 4,625-bp α_2CAR sequence encompassing the various haplotypes. Expression of the α_2CAR is therefore directed/maintained by its own promoter, 5′ UTR, and 3′ UTR, and the effects of polymorphisms as they occur in haplotypes within these regions on expression of α_2CAR transcripts and protein can thus be ascertained. A neuronal cell line, BE(2)-C, which is of human origin and expresses low levels of α₂AR, was used as the host cell for transfections. All haplotypes that had allele frequencies of ≥5% in any one of the three ethnic groups were studied.

Results from transfection studies where mRNA was quantitated by real-time RT-PCR are shown in Fig. 4. α₂AR mRNA expression was significantly related to haplotype (P < 0.001 by ANOVA, n = 6 experiments). Haplotype 14 had statistically greater expression than all other haplotypes, which was most apparent compared to low expressing haplotypes 2 and 4 and the intermediate expressing haplotypes 6 and 7 (P < 0.001). Haplotypes 2 and 4 had lower expression than all other haplotypes (P values between 0.01 and 0.001). Thus, this grouping leaves haplotypes 3 and 5 with essentially wild-type (haplotype 1) expression phenotypes. Haplotype 8 also had relatively higher expression as compared to haplotypes 2, 4 (P < 0.001), 7 (P < 0.01), and 6 (P < 0.05). Given the potential limitations of this transfection approach in representing a haplotype-dependent effect that might be manifested in various cells in an intact organism, expression of α_2CAR protein was also considered as a phenotypic endpoint. α_2CAR receptor density was determined in the transfected cells by quantitative radioligand binding using the α₂AR ligand [³H]yohimbine. There was a significant relationship between α_2CAR protein expression and haplotype (P < 0.001 by ANOVA, n = 9 experiments, Fig. 5). There was reasonable overall agreement between the haplotype effect on mRNA and protein expression. One notable difference was with haplotype 3, which had a low level of protein expression that was similar to haplotypes 2 and 4 in the radioligand binding studies. One might consider, then, that haplotypes 2, 3, and 4 represent a low expression “cluster.”

Fig. 4.

α_2CAR haplotypes display differential mRNA expression profiles. Neuronal BE(2)-C cells were transfected with the indicated haplotype constructs, and α_2CAR transcripts were quantitated by real-time RT-PCR as described in Methods. Data represent six independent experiments. α_2CAR mRNA was significantly related to haplotype (P < 0.001 by ANOVA). See text for individual statistical comparisons.

Fig. 5.

α_2CAR haplotypes display differential receptor expression profiles. Neuronal BE(2)-C cells were transfected with the indicated haplotype constructs, and α_2CAR protein expression was quantitated by [³H]yohimbine radioligand binding as described in Methods. Data represent nine independent experiments. α_2CAR protein expression was significantly related to haplotype (P < 0.001 by ANOVA). See text for individual statistical comparisons.

It is interesting to note that the coding polymorphism Del322-325, which renders the receptor hypofunctional, is present in two of the three low expressing haplotypes (Table 2). Thus, ≈60% of the Del322-325-containing haplotypes might be expected to have an exaggerated phenotype, because receptor expression and function are both decreased. On the other hand, if the human “pathologic threshold” for the Del322-325 phenotype requires low expression as well, haplotypes 6 and 7 (representing 18% of all Del322-325-containing haplotypes) may act at the physiologic level as “wild-type.” A corollary to this concept is that the low expression imparted by haplotypes 2, 3, and 4 may be sufficient to evoke a pathologic phenotype regardless of the coding genotype. As such, if one knows only the genotype at the Del322-325 position, a number of the pathologic haplotypes would be misclassified. These scenarios emphasize that the multiple α_2CAR haplotypes have the potential for increased predictive value in association studies, as compared to the single site coding polymorphism. The implications of the high-expression haplotype 14 may be most evident in studies designed to ascertain a hyperfunctional α_2CAR phenotype. In heart failure, this might be seen as a protective effect on development of the syndrome or lower synaptic or systemic norepinephrine levels.

Conclusions. These data reveal a highly variable α_2CAR gene with complex intragenic organization of 20 polymorphic sites representing 24 haplotypes. This includes a common 3′ UTR substitution SNP within an insertion/deletion sequence, a radical coding polymorphism that deletes four amino acids, relatively low linkage disequilibrium between many polymorphisms, few cosmopolitan haplotypes, and prevalent ethnic-specific haplotypes. There was substantial divergence between certain haplotypes and a partitioning of the cardiomyopathic Del322-325 polymorphism into multiple haplotypes. Furthermore, we show that the haplotypic variation has biological relevance as assessed in studies of receptor transcript and protein expression. This degree of genetic and functional diversity in a gene with critical functions throughout the body may have implications for diseases other than heart failure, such as cardiac arrhythmias, hypertension, behavioral and learning disorders, obesity, and diabetes mellitus. Current studies have shown that the Del322-325 polymorphism is associated with physiologic or pathologic cardiac phenotypes (5-7). However, given the diversity of the Del322-325-containing haplotypes, a further refinement of its role, or that of other polymorphisms or haplotypes of the α_2CAR gene in heart failure and its response to pharmacologic intervention, is in order. Such studies will require larger populations (which include ethnic diversity), as compared to those studied to date. Our results also suggest that haplotype clustering may be appropriate in association studies based on the expression profiles.