Evolutionary genetics of the human Rh blood group system

By /

  • Logo of wtpa
 Hum Genet. Author manuscript; available in PMC 2013 Jan 1.
Published in final edited form as:
PMCID: PMC3378649
EMSID: UKMS48046

Evolutionary genetics of the human Rh blood group system

George H. Perry,1,2,* Yali Xue,3,* Richard S. Smith,2,4,* Wynn K. Meyer,5 Minal Çalışkan,5 Omar Yanez-Cuna,2,6 Arthur S. Lee,2 María Gutiérrez-Arcelus,2,6 Carole Ober,5 Edward J. Hollox,7,^ Chris Tyler-Smith,3,^ and Charles Lee2,4,^
Author information ► Copyright and License information ►

Abstract

The evolutionary history of variation in the human Rh blood group system, determined by variants in the RHD and RHCE genes, has long been an unresolved puzzle in human genetics. Prior to medical treatments and interventions developed in the last century, the D-positive children of D-negative women were at risk for hemolytic disease of the newborn, if the mother produced anti-D antibodies following sensitization to the blood of a previous D-positive child. Given the deleterious fitness consequences of this disease, the appreciable frequencies in European populations of the responsible RHD gene deletion variant (for example, 0.43 in our study) seem surprising. In this study, we used new molecular and genomic data generated from four HapMap population samples to test the idea that positive selection for an as-of-yet unknown fitness benefit of the RHD deletion may have offset the otherwise negative fitness effects of hemolytic disease of the newborn. We found no evidence that positive natural selection affected the frequency of the RHD deletion. Thus, the initial rise to an intermediate frequency of the RHD deletion in European populations may simply be explained by genetic drift/ founder effect, or by an older or more complex sweep that we are insufficiently powered to detect. However, our simulations recapitulate previous findings that selection on the RHD deletion is frequency dependent, and weak or absent near 0.5. Therefore, once such a frequency was achieved, it could have been maintained by a relatively small amount of genetic drift. We unexpectedly observed evidence for positive selection on the C allele of RHCEin non-African populations (on chromosomes with intact copies of the RHD gene) in the form of an unusually high FST value and the high frequency of a single haplotype carrying the C allele. The Rhce function is not well understood, but the C/c antigenic variant is clinically relevant and can result in hemolytic disease of the newborn, albeit much less commonly and severely than that related to the D-negative blood type. Therefore, the potential fitness benefits of the RHCE C allele are currently unknown but merit further exploration.

Keywords: Blood group polymorphism, copy number variation, human evolution, balancing selection

INTRODUCTION

The human Rh blood group system is a collection of antigens expressed on erythrocyte cell membranes () that may play a role in the transport of ammonia () or carbon dioxide (; ), or not in transport but in erythrocyte membrane structure (; ). Functional (antigen) variation in the Rh blood group system is determined by insertions/ deletions, single nucleotide polymorphisms (SNPs), and gene conversion events in the RHD and RHCE genes ( Colin et al. 1991; Flegel 2011;). Homozygous deletion of the entire RHD gene results in the D-negative blood phenotype (), whereas the D-positive phenotype is conferred by the presence of either one or two intact copies of the RHD gene.

D-negative mothers may produce anti-D antibodies following exposure to red blood cells from a D-positive fetus during pregnancy or childbirth. Subsequent D-positive offspring of a D-negative mother may develop a hemolytic disease of the newborn resulting in fetal death or severe disability ( Levine et al. 1941;). Prior to treatments introduced beginning in the 1940s (culminating with Rho(D) Immune Globulin (RhoGAM) in 1968) that largely obviated these health issues (), D-negative mothers may have suffered reduced reproductive fitness; pre-treatment mortality from hemolytic disease of the newborn was reportedly 1 in every 56 births to D-negative women in a European-American population (1 in 392 births among all women, regardless of D status) (). Certain non-synonymous SNPs in the RHCE gene result in antigenic variation in the encoded RHCE protein (rs676785, Ser103Pro: C/c; rs609320, Pro226Ala: E/e) (), and differential expression of these antigens can also lead to hemolytic disease of the newborn, but with much lower frequency than that caused by anti-D antibodies, at least in European populations ().

Based on the potential consequence of hemolytic disease of the newborn, one might expect strong purifying selection to have acted against the D-negative phenotype and minor alleles of the C/c and E/e antigens. Yet in European populations the D-negative phenotype is observed at substantial frequencies, typically 0.15-0.17 and up to 0.29 in the Basque ( Touinssi et al. 2004;), and the frequency of the RHCE C allele is ~0.44 (). Similar to the effects of sickle cell and thalassemia hemoglobin heterozygosity on malarial resistance (; ; ; ), we asked whether these alleles confer an unknown fitness benefit whereby positive or balancing selection explains their otherwise surprisingly high frequencies ( Feldman et al. 1969;). Because such a history may have left detectable genomic signatures, in this study we used a population genetics framework to test evolutionary hypotheses concerning these functional genetic variants of the Rh blood group system.

MATERIALS AND METHODS

Genotyping

DNA samples and cell lines from the HapMap individuals were obtained from the Coriell Institute for Medical Research. To genotype the RHD deletion we used a previously-developed TaqMan quantitative PCR (qPCR) assay (), in which the forward primer for the RHD amplicon (all 5′-3′; CCTCTCACTGTTGCCTGCATT) maps to the very 3′ end of the gene and the reverse primer (AGTGCCTGCGCGAACATT) maps to the 5′ end of the segmental duplication flanking and unique to RHD (Figure 1a). Therefore, the amplification is specific to RHD (i.e., there is no complicating signal from RHCE). Primers and the internal probe ([FAM]TACGTGAGAAACGCTCATGACAGCAAAGTCT[TAMRA]) were purchased from Integrated DNA Technologies (Coralville, IA). To calibrate for minor input DNA quantity differences across samples we multiplexed the RHD assay with a second TaqMan assay from the single-copy control gene RNasePwith a VIC-labeled probe (Applied Biosystems, Foster City, CA). Quadruplicate reactions of 8 ng DNA in 10 μL volumes with TaqMan Genotyping Master Mix (Applied Biosystems) were run on Applied Biosystem’s 7900HT Real-Time PCR System in a 384-well plate. One individual with 2 RHD copies, NA10857, was run on each plate and used to calculate the estimated diploid copy numbers for all unknowns based on ΔΔCT (Supplemental Table 1). To genotype the RHCE SNPs rs676785 (C/c) and rs609320 (E/e) we developed an assay based on co-amplification of these regions (for both RHCE and the corresponding regions of RHD) using dye-labeled primers followed by digestion of the internal products with the restriction enzyme MnlI (New England Biolabs). Products were digested variably depending on rs676785 and rs609320, and the different-sized products were compared quantitatively to estimate Cc and Ee genotypes, taking RHD gene copy number into account. Details of the assay are provided in  Supplemental Figure 1.

Figure 1

RHD deletion genotyping and fiber-FISH validation

We validated the TaqMan-estimated RHD genotypes by checking consistency with Mendelian inheritance in the CEPH and Yoruba (these samples are comprised of parent-offspring trios), comparison to genotype estimates from analysis of data from the Affymetrix 6.0 SNP genotyping array (), and high-resolution fluorescence in situ hybridization on stretched DNA fibers (fiber FISH). For fiber FISH analyses we used a fosmid probe G248P84657G6 (labeled with digoxigenin-11-dUTP; green in images) that maps to the 5′ end of RHD but also hybridizes strongly to RHCE due to the high homology between these genes. We also created a PCR probe (labeled with biotin-16-dUTP; red in images) for the pair of segmental duplications that flank RHD only and contain the deletion breakpoint (Figure 1a). We first used long-range PCR to amplify nearly the entire segmental duplications (primers general to both copies; TAAATGCTCTTCTGAAGGCTGATACG and TTTACAAAGGGGAGAACGGTAAGAAG) from NA10857 in a 25 μL reaction with 100 ng DNA and TripleMaster Taq Polymerase (Eppendorf) with an initial step of 93° C for 3 min followed by 40 cycles of 93° C for 30 sec, 64° C for 30 sec, and 68° C for 10 min. Nested PCR was then used to amplify two internal, overlapping fragments ~3.5 kb each in size (fragment 1: TCTTCTGAAGGCTGATACGACA and ATGATAGGGTTTGGTTGTGTCC; fragment 2: GGACACAACCAAACCCTATCAT and AATCACCGTCAAGGAGTCAGAT) using 0.1 μL of the long-range PCR product in 25 μL reactions with HotMaster Taq Polymerase (Eppendorf) for 3 min at 93° C followed by 40 cycles of 93° C for 30 sec, 60° C for 30 sec, and 70° C for 5 min. Probe labeling, preparation of the fibers and slides, hybridization, washes, detection, and imaging were performed as described previously () except that the fosmid probe was labeled for 12 rather than 5 hours.

Nucleotide sequence analyses

We used long-range PCR to amplify two unique (i.e., not duplicated) regions of the RHD/RHCE gene locus, each approximately 6 kb in size. One region is directly upstream of the segmental duplication at the 5′ end of the RHD gene, while the second region is directly downstream of the segmental duplication at the 3′ end of RHD, and between the RHD and RHCE genes (Figure 1a). The ~6kb amplified products were then used in nested PCR reactions to amplify products ~600 bp in size with overlap ~280 bp, which were then sequenced in two directions as described previously(). PCR primers are provided in Supplemental Table 2. Putative SNPs were flagged using Mutation Surveyor (SoftGenetics, State College, PA, USA) and then manually checked; each SNP call was supported by at least four reads. Summary statistic tests of neutrality were carried out and their P values estimated based on the percentile of the test value in the null distribution from 1000 simulations conditioned on the number of segregating sites using the program ms () with the best-fit demographic model () via a custom Perl script as described previously ().

Simulations

We simulated selection at the RHD locus in a population of infinite size, considering the following: i) probability of sensitization if the mother is D-negative and the child is D-positive (ps), ii) probability that an individual child of a sensitized mother would die from hemolytic disease of the newborn (pa), iii) mean family size in families not potentially affected by hemolytic disease of the newborn (fs), and iv) maximum number of attempts to conceive (na). In turn, we fixed three of the parameters at reasonable values (including the assumption of some negative effect on individual fitness due to hemolytic disease of the newborn: ps = 0.2 and pa = 0.9) and varied the fourth, calculating the change in RHD deletion frequency over one generation given different initial RHD deletion population frequencies

Family size in families not potentially affected by hemolytic disease of the newborn was modeled as a binomial, with n = na and p = fs/na. For families potentially affected by hemolytic disease of the newborn, we took a random draw from this binomial as the initial number of offspring and then applied selection due to hemolytic disease of the newborn. Heterozygous offspring of D-negative mothers paired with D-positive fathers were split into “pre-” and “post-sensitization” sets, with sensitization occurring at the x-th pregnancy, where x is a random draw from a geometric distribution with p = ps. Selection occurred on the heterozygous post-sensitization offspring (if any), with the number of surviving heterozygous post-sensitization offspring determined by a random draw from a binomial with n = the number of possible heterozygous offspring remaining, and p = 1 – pa. In families in which the father was heterozygous for the RHD deletion, each offspring was assigned to be homozygous or heterozygous with equal probability. We performed 10,000 simulations as described above and calculated the mean number of heterozygous offspring from families with D-negative mothers paired with m1) D-positive fathers homozygous for the non-deletion allele and m2) D-positive fathers heterozygous for the deletion. We defined selection against heterozygous offspring of m1 as s1 = 1 – (mean number of offspring of m1 couples/fs), and selection against heterozygous offspring of m2 as s2 = 1 – (mean number of heterozygous offspring of m2 couples / (fs/2)). We then assumed an initial frequency of the RHD deletion and calculated the change in frequency over one generation, assuming a population of infinite size, as follows. We started with a population at Hardy-Weinberg equilibrium and calculated the proportion of families of each type potentially affected by hemolytic disease of the newborn (m1 and m2). We then reduced the frequency of RHD deletion heterozygotes in the subsequent generation by the expected proportion of heterozygous offspring from each potentially affected family type, multiplied by 1 – selection against heterozygotes of that family type (s1 or s2). We calculated the new frequency of the RHD deletion in the population after selection had occurred and subtracted the initial frequency to obtain the change in frequency over one generation.

RESULTS

RHD/ RHCE genotype frequencies

We genotyped the RHD gene deletion and the RHCE gene SNPs marking the C/c and E/e antigen variants in individuals from four HapMap human populations: European-Americans from Utah (CEU; n = 90), Yoruba from Ibadan, Nigeria (YRI; n = 90), Chinese Han from Beijing (CHB; n = 45), and Japanese from Tokyo (JPT; n = 45). In a subset of these individuals (n = 68; CEU, n=23; YRI, n=23; and CHB, n=23), we also resequenced two genomic regions, each ~6 kb in size: one region flanked the 5′ end of the RHD gene, and the second region was located between the 3′ ends of RHD and RHCE (Figure 1a). We used the HapMap population samples to integrate our RHD and RHCE genotype data with the HapMap Phase II SNP data () for genome-wide comparisons of population differentiation and analyses of extended SNP haplotypes.

We estimated RHD deletion frequencies of 0.43, 0.19, and 0.07 in the CEU, YRI, and CHB+JPT population samples, respectively (Table 1; Figure 1b,c), consistent with previous studies (; ; ). Unexpectedly, we estimated three RHD copies per diploid cell for one YRI (NA19204) and one JPT (NA18952) individual. Fluorescent in situ hybridization analysis on stretched DNA fibers (fiber FISH) confirmed the presence of an RHD duplication allele (Figure 1d), the formation of which was likely mediated by non-allelic homologous recombination of the RHD-flanking segmental duplications (i.e., the reciprocal product of the RHD deletion). The genotypes of an additional YRI parent-offspring trio initially seemed inconsistent with a Mendelian pattern of inheritance, but this result was explained by the presence of both the RHD duplication and RHD deletion alleles in the father (Figure 1e). A similar duplication has been reported previously in a different Japanese population sample (). Such a duplication may also explain qPCR results of greater RHDthan RHCE copy number in eight individuals from Germany with relatively large densities of the D antigen in red blood cells despite Rhesus box (the RHD-flanking segmental duplications; see Figure 1a)-based genotyping assay suggestions of heterozygosity for the RHD deletion in those individuals (). If so, then this result would indirectly suggest that the RHD duplication (which may cause Rhesus box-based RHD genotyping methods to fail) has a functional effect on antigen density.

Table 1

RHD and RHCE derived allele frequencies in HapMap populations

For RHCE, the estimated frequency of the derived C allele was considerably higher in the CHB+JPT (0.78) and CEU (0.45) population samples than in YRI (0.08). Frequencies of the derived E allele were less variable among populations: 0.25, 0.09, and 0.08 in CHB+JPT, CEU, and YRI, respectively (Table 1).

Evolutionary analyses

We first examined the relative degree of population differentiation (FST) for the RHD and RHCE functional variants compared to genome-wide SNPs that have been genotyped in the same samples (). Extreme FST values may reflect past population-specific positive selection ( Barreiro et al. 2008; Sabeti et al. 2006;) (but see ref. ()). While the RHD deletion and the E/e variant of RHCE are not unusual in this respect, the (CHB+JPT)-YRI FSTvalue for the C/c variant of RHCE is exceptional (FST = 0.64; genome-wide percentile = 98.8; Figure 2a).

Figure 2

Evidence for positive selection on the C allele of RHCE

It is possible that FST values of variants within a duplicated locus, which may be subject to recurrent gene conversion (), might not be comparable to those of unique variants. Indeed, the RHCE C allele itself is likely the result of a gene conversion event that transferred sequence from RHD to RHCE(). To examine this issue, we resequenced ~12 kb of this genomic region from 68 HapMap individuals. When we integrated the identified SNP genotype data from this resequencing effort with the RHD deletion and RHCE C/c and E/e genotype data and examined the estimated haplotypes, it was clear that the high frequency of the C allele in non-African populations is explained almost exclusively by a single mutation event (Figure 2b; see discussion below), alleviating this concern.

We next performed tests of neutrality on the patterns of nucleotide diversity and the haplotypes that were estimated from the RHD/RHCE genotypes combined with the ~12 kb resequencing data. These tests evaluate whether or not the observed allele and haplotype frequency distributions are consistent with neutrality, given the demographic history of the populations (). We observed significant deviations from neutral expectations in non-African populations (Table 2), including an excess of variants in external branches of the phylogeny in the CEU (Fu and Li’s D test (); P = 0.038) and a single haplotype that has risen to unexpectedly high frequency in both CEU and CHB (Figure 2b; 47 out of 90 chromosomes; P = 0.045 by the common haplotype frequency test()). This result is not explained by haplotypes containing the RHD deletion. Rather, the high-frequency haplotype is comprised of RHD-positive chromosomes that carry the C allele of RHCE (Figure 2b). In total, we sampled 90 CHB and CEU chromosomes, 56 of which contain the RHCE C allele. Of these 56 chromosomes, 47 are identical across the ~12 kb region, and an additional four chromosomes differ from this common haplotype by only one nucleotide each.

Table 2

Summary statistics for RHD/RHCE region resequencing data

Finally, we integrated the RHD deletion and RHCE C/c and E/e genotypes with HapMap Phase II SNP haplotype data to examine the patterns of linkage disequilibrium extending from these functional genetic variants. Positive selection may drive beneficial genetic variants to high frequencies at faster rates than neutral variants (which increase in frequency only by genetic drift). If the selection was recent and strong, then such variants might be associated with relatively long, low-diversity haplotypes, because recombination has not had sufficient opportunity to break down the extended haplotype on which the mutation occurred (). To quantify the unusualness of extended haplotypes, we computed iHS scores () for the RHD and RHCE functional variants. None of the variants in any population had |iHS| > 2.5, a cutoff corresponding to the highest 1% of genome-wide iHS scores for SNPs in that population. The iHS score for the RHD deletion variant in CEU was 2.12, a moderate outlier, with the positive iHS score indicating that a relatively long, low-diversity haplotype is associated with the ancestral (non-deletion) allele. No other RHD or RHCE |iHS| score exceeded 1.05. The |iHS| scores associated with the C allele in CEU and CHB+JPT were 0.069 and 0.670, respectively.

DISCUSSION

Potential evidence for positive selection on the C allele of RHCE

Several observations from our evolutionary analyses – an unusually high (CHB+JPT)-YRI FST value for the RHCE C/c variant, the unusual Fu and Li’s D value in the CEU and the high frequency of a single haplotype carrying the derived C allele in non-Africans – suggest a departure from neutral evolution at the RHD/RHCE locus. These findings all point towards the same conclusion: that the high frequency of the C allele in European and especially East Asian populations may reflect a history of positive natural selection. Enthusiasm for this hypothesis is somewhat tempered by the lack of a positive selection signal from the analysis of associated extended haplotypes. This discrepancy, however, might be explained by the moderate frequency of the putatively-selected haplotype or the different temporal resolutions of the population genetic tests we used. The FST statistic and the allele and haplotype frequency analyses from resequencing data are generally more sensitive to older selection events that have reached intermediate to high frequency, while the IHS test is most powerful for detecting evidence of recent, strong, positive selection when there is good SNP density across the region of haplotype breakdown ().

The derived C allele resulted from an RHD to RHCE gene conversion event () that effectively reduced the difference between the two encoded proteins, raising the possibility that the C allele reached high frequency in non-African populations by conferring protection against RHD-related hemolytic disease of the newborn to D-negative mothers (i.e., if an antigenic response to exposure to blood from a D-positive fetus is less likely or less severe if the mother’s RhCE antigen is more similar to the child’s RhD). However, this idea can be dismissed based on two lines of evidence. One, the RHCE C allele is found almost exclusively on chromosomes with an intact RHD gene, consistent with previous observations (). Among the 90 CHB+CEU chromosomes for which we collected the ~12 kb resequencing data, 56 were found to carry the RHCE C allele. Of these 56 chromosomes, 55 (98%) were associated with intact RHD genes while only one (2%) was associated with the RHD deletion (Figure 2b). Two, the estimated RHD deletion frequency is < 0.10 in CHB+JPT (Table 1). The majority of D-negative blood type cases in an East Asian population resulted from RHD deletion (). Together, these results suggest a very low occurrence of D-negative blood type and thus a naturally low rate of RHD-related hemolytic disease of the newborn in East Asian populations, consistent with previous observations ().

We hope that our report of possible positive selection on the RHCE C allele will motivate more detailed studies of the functional effects of this variant, both for those associated with the antigenicity phenotype and for any not associated with antigenicity. While Ser103Pro (which was the focus of our genotyping via the underlying SNP rs676785) explains C/c antigenicity, there are three other amino acid variants of RHCE in linkage disequilibrium with Ser103Pro (Cys16Trp, Ile60Leu, and Ser68Asn) () that should also be considered candidates for fitness-relevant functional effects. In the absence of a current functional and testable hypothesis, we consider the suggestion of positive selection on the RHCE C allele to be preliminary but intriguing, especially given the previously-demonstrated importance of functional variation of other red blood cell membrane proteins in human evolution, for example of the Duffy antigen (FY gene) and susceptibility to Plasmodium vivax malaria ().

No evidence for positive selection on the RHD deletion allele

Given that D-negative mothers – prior to now-common medical interventions – would have faced the potential fitness-reducing consequences of hemolytic disease of the newborn, we were somewhat surprised that in Europeans, with RHD deletion allele frequency = 0.43, our evolutionary analyses failed to uncover any strong evidence of positive selection on this allele that may have resulted from some unknown, offsetting fitness benefit. Our tests may have had insufficient power to reject neutrality for this variant if its high frequency initially resulted from a relatively ancient selective event. A second selective sweep in a genomic region (e.g., on the RHCE C allele) might also make it more difficult to detect an earlier selective event (e.g., on the RHD deletion). Extending this notion further, we note that there is considerable functional diversity at the RHD/ RHCE gene locus even beyond the common, well-known variants studied here. For example, a pseudogene-causing (and D-negative phenotype-producing) 37 bp frameshift duplication in RHD exon 4 has a reported frequency of 7% in at least one African population (), and there are numerous, rare antigen-affecting SNPs, gene conversion events, and partial deletions of RHD and RHCE ( Avent and Reid 2000;). Such a pattern of diversity could reflect low functional constraint at this locus, but if these mutations also resulted in or contributed to fitness-beneficial phenotypes, then their collective frequency increases may have further clouded any genomic signatures of positive selection associated with the RHD deletion itself (). Therefore, based only on the results of our analyses, it is not possible to determine that selection for the RHD deletion allele did not occur.

However, as an alternative to the above scenarios, the negative fitness effect of the D-negative blood type in the past might have been less strong than previously thought; the combined probabilities of (i) sensitization to a D-positive fetus and (ii) penetrance/ morbidity of hemolytic disease of the newborn may have been too low to prevent the RHD deletion from reaching high frequency in Europe by genetic drift/ founder effect alone, given the effective size of this population. There are few appropriate, pre-medical intervention data with which to address this issue: Knox and Walker () used antenatal antibody test data collected from 1947 to 1956 in the United Kingdom to estimate that the risk of sensitization for D-negative women was 1 in 17 pregnancies with a D-positive fetus. Between 1940 and 1946 at the Chicago Lying-In Hospital in the United States, incidence of hemolytic disease among the newborns of D-negative women was 1 in 37 births (1:252 when considering all women regardless of D status), with approximately two-thirds of these cases resulting in newborn death or stillbirth ().

While the above data suggest a higher mortality rate for the children of D-negative mothers, it is less clear whether this affected differences in the total number of surviving children between D-positive and D-negative mothers.  analyzed and reviewed multiple birth statistic datasets from the 1940s to 1960s in the United Kingdom, Canada, and United States and observed a significant association between D status and fertility in only one population sample, in which there were fewer live births to couples with a D-negative mother and a D-positive father than to couples with a D-positive mother or a D-negative father or both.

To address this issue further, we genotyped the RHD deletion in 942 Hutterites from South Dakota, U.S.A. The Hutterites are a founder population of European descent with high reproductive rates and large family sizes ( Hostetler 1974;). These data were combined with Rh blood group types for 504 Hutterites, which were previously determined by serology and birth and pedigree information for 1623 Hutterites (including all individuals with either Rh genotypes or serotypes) (). We considered all families with at least one child born prior to 1968 (the first year of RhoGAM availability) and compared potentially affected families (defined as families with a D-negative mother and D-positive father) to unaffected families (defined as families with a D-positive mother and a D-positive father, a D-positive mother and a D-negative father, or a D-negative mother and a D-negative father, an unknown D status mother and a D-negative father, or a D-positive mother and an unknown D status father). There was no significant difference between the two groups in either total number of children (t-test; P = 0.73; Figure 3a) or mean interbirth interval (t-test; P = 0.75; Figure 3b).

Figure 3

Fertility rate and inter-birth interval analysis by D status in the Hutterites

The lack of a significant difference in family size or mean interbirth interval between these two groups of families may reflect the small sample size of potentially affected families (there were only 24 families with a D-negative mother and a D-positive father). We also caution that some (non-first) children included in this analysis were born after 1968 (Figure 3c), and therefore RhoGAM treatment may have helped to prevent sensitization in some women if sensitization had not occurred in earlier pregnancies. In addition, recent general medical advances may have lessened the fitness disadvantage for D-negative mothers, even without RhoGAM treatment. Still, our results are consistent with most previous analyses that also failed to observe significant differences in fertility rates between D-negative and D-positive mothers ().

Apart from any uncertainty about the underlying fitness effect of the D-negative blood type, an additional evolutionary aspect that should be considered is the effect of frequency-dependent selection on the RHD deletion. Specifically, once the deletion reaches intermediate frequency, D-negative children of D-negative mothers and heterozygous D-positive fathers would have an advantage over D-positive children (). Previous models of under-dominance at the RHD locus ( Feldman et al. 1969; Levin 1967; Li 1953;) were developed primarily to derive equilibrium frequencies of the allele, and few besides  have provided illustrations of the effects of variation in relevant parameters at all frequencies of the deletion. Therefore, to evaluate how such variables might influence the strength of frequency-dependent selection, we simulated selection at the RHD locus in a population of infinite size, considering the following (see Methods): i) probability of sensitization, ii) probability of death from hemolytic disease of the newborn, iii) mean family size, and iv) maximum number of attempts to conceive. We fixed three of the parameters at reasonable values and varied the fourth, calculating the change in RHD deletion frequency over one generation has given different initial RHD deletion population frequencies. We found that the strength of selection was affected by the probability of sensitization (Figure 4a), the probability of child death due to hemolytic disease of the newborn (Figure 4b), and family size (Figure 4c); due to modeling assumptions and the region of parameter space explored, the strength of selection was unaffected by the maximum number of attempts to conceive (Figure 4d).

Figure 4

Influence of parameter variation on frequency-dependent selection at the RHD locus

These simulations confirm what  and others ( Feldman et al. 1969; Levin 1967;) demonstrated through mathematical evaluation: that selection for the RHD deletion, under a model in which there a potential fitness effect on D-negative mothers and in populations of infinite size, is frequency-dependent. If deletion frequencies are initially < 0.5, then the population frequency becomes lower still in subsequent generations. If deletion frequencies are initially > 0.5, then the population frequency becomes higher in subsequent generations. However, at intermediate frequencies – near RHD deletion frequency of ~0.5, or similar to that observed in populations of European ancestry – the effect of selection is relatively small (at frequency = 0.5, there is no change; Figure 4). In populations of finite size, weak selective pressures may have been overwhelmed by the stochastic effects of genetic drift. In addition, the simulations demonstrate that family size has a dramatic effect on the strength of selection (also see ); if family sizes were small during the initial rise in frequency of the RHD deletion allele, then this also may have reduced the magnitude of selection.

Overall, evolutionary interpretations of the RHD gene deletion and the D-negative phenotype are mixed. Pre-treatment reported hemolytic disease mortality for births to D-negative mothers was 1 in 56 (). Once sensitization occurred, all subsequent pregnancies with a D-positive fetus would have been at risk, yet the observed fertility rates in the Hutterites generally did not differ between D-positive and D-negative mothers. None of our population genetic analyses revealed convincing evidence of past positive selection on the RHD deletion. While this negative result could be explained by a number of evolutionary scenarios, our current working explanation for the initial rise to a relatively high frequency of the RHD deletion and the D-negative phenotype in European populations must be the null hypothesis: genetic drift/ founder effect. The level of purifying selection acting against this phenotype may have been less strong than thought, or simply not strong enough to overcome the effects of genetic drift in founding European populations. However, once an intermediate frequency had been achieved, selective pressures would have been relatively weak and potentially overcome by small amounts of genetic drift, even with considerable negative fitness consequences for individual D-negative mothers.

  

Supplementary Material

Supplementary Figure1: RHCE genotyping assays.

Supplementary Table 1: RHD/CE genotype data for HapMap populations.

Supplementary Table 2: Amplification and resequencing of RHD flanking regions – primers.

Supplementary Table 3: Genotype data from resequencing of RHD flanking regions.

ACKNOWLEDGMENTS

We thank Rachael Cartlidge for assistance with initial development of the RHCE genotyping assays, David Hopkinson for DNA samples with known Rh serotype from the former MRC Blood Group Unit, Luis Barreiro for the HapMap Phase II FST value database, Joe Pickrell for assistance with the IHS test, Richard Hudson for assistance with simulations, the Sanger Faculty Small Sequencing Projects Group for generating the sequence data, Steve McCarroll, Pardis Sabeti, and Molly Przeworski for helpful discussions, and two reviewers for insightful comments and suggestions on the manuscript. We acknowledge the participants who contributed samples for this study. This work was funded by National Institutes of Health Grant P41-HG004221 (to C.L.), The Wellcome Trust (WT098051, C.T.-S., and Y.X.), Medical Research Council New Investigator Award GO801123 (to E.J.H.), and National Institutes of Health Grant R01-HD21244 (to C.O.).

LITERATURE CITED

  • Abney M, McPeek MS, Ober C. Estimation of variance components of quantitative traits in inbred populations. American journal of human genetics. 2000;66:629–50. [PMC free article] [PubMed]
  • Allen SJ, O’Donnell A, Alexander ND, Alpers MP, Peto TE, Clegg JB, Weatherall DJ. alpha+-Thalassemia protects children against disease caused by other infections as well as malaria. Proc Natl Acad Sci U S A. 1997;94:14736–41. [PMC free article] [PubMed]
  • Allison AC. The distribution of the sickle-cell trait in East Africa and elsewhere, and its apparent relationship to the incidence of subtertian malaria. Trans R Soc Trop Med Hyg. 1954;48:312–8.[PubMed]
  • Avent ND, Reid ME. The Rh blood group system: a review. Blood. 2000;95:375–87. [PubMed]
  • Barreiro LB, Laval G, Quach H, Patin E, Quintana-Murci L. Natural selection has driven population differentiation in modern humans. Nat Genet. 2008;40:340–5. [PubMed]
  • Carritt B, Kemp TJ, Poulter M. Evolution of the human RH (rhesus) blood group genes: a 50-year-old prediction (partially) fulfilled. Hum Mol Genet. 1997;6:843–50. [PubMed]
  • Chen JM, Cooper DN, Chuzhanova N, Ferec C, Patrinos GP. Gene conversion: mechanisms, evolution and human disease. Nat Rev Genet. 2007;8:762–75. [PubMed]
  • Colin Y, Cherif-Zahar B, Le Van Kim C, Raynal V, Van Huffel V, Cartron JP. Genetic basis of the RhD-positive and RhD-negative blood group polymorphism as determined by Southern analysis. Blood. 1991;78:2747–52. [PubMed]
  • Coop G, Pickrell JK, Novembre J, Kudaravalli S, Li J, Absher D, Myers RM, Cavalli-Sforza LL, Feldman MW, Pritchard JK. The role of geography in human adaptation. PLoS Genet. 2009;5:e1000500. [PMC free article] [PubMed]
  • Edward V, Cartron JP, Ripoche P, Gros G. RhAG protein of the Rhesus complex is a CO2 channel in the human red cell membrane. FASEB J. 2008;22:64–73. [PubMed]
  • Feldman MW, Nabholz M, Bodmer WF. Evolution of the Rh polymorphism: a model for the interaction of incompatibility, reproductive compensation, and heterozygote advantage. Am J Hum Genet. 1969;21:171–93. [PMC free article] [PubMed]
  • Fisher RA, Race RR. Rh gene frequencies in Britain. Nature. 1946;157:48–9. [PubMed]
  • Flegel WA. Molecular genetics and clinical applications for RH. Transfus Apher Sci. 2011;44:81–91.[PMC free article] [PubMed]
  • Flint J, Hill AV, Bowden DK, Oppenheimer SJ, Sill PR, Serjeantson SW, Bana-Koiri J, Bhatia K, Alpers MP, Boyce AJ, et al. High frequencies of alpha-thalassemia are the result of natural selection by malaria. Nature. 1986;321:744–50. [PubMed]
  • Fu YX, Li WH. Statistical tests of neutrality of mutations. Genetics. 1993;133:693–709.[PMC free article] [PubMed]
  • Haldane JBS. Selection against heterozygosis in man. Annals Eugenics. 1942;11:333–40.
  • Hostetler J. Hutterite Society Johns Hopkins University Press; Baltimore, MD: 1974.
  • Hudson RR. Generating samples under a Wright-Fisher neutral model of genetic variation. Bioinformatics. 2002;18:337–8. [PubMed]
  • International HapMap Project Consortium A second generation human haplotype map of over 3.1 million SNPs. Nature. 2007;449:851–61. [PMC free article] [PubMed]
  • Knox G, Walker W. Nature of the determinants of rhesus isoimmunization. Br J Prev Soc Med. 1957;11:126–30. [PMC free article] [PubMed]
  • Kustu S, Inwood W. Biological gas channels for NH3 and CO2: evidence that Rh (Rhesus) proteins are CO2 channels. Transfus Clin Biol. 2006;13:103–10. [PubMed]
  • Kwiatkowski DP. How malaria has affected the human genome and what human genetics can teach us about malaria. Am J Hum Genet. 2005;77:171–92. [PMC free article] [PubMed]
  • Levin BR. The effect of reproductive compensation on the long-term maintenance of the Rh polymorphism: the Rh crossroad revisited. Am J Hum Genet. 1967;19:288–302. [PMC free article][PubMed]
  • Levine P, Vogel P, Katzin EM, Burnham L. Pathogenesis of Erythroblastosis Fetalis: Statistical Evidence. Science. 1941;94:371–372. [PubMed]
  • Li CC. Is Rh facing a crossroad? A critique of the compensation effect. Am Nat. 1953;87:257–61.
  • Lo YM, Hjelm NM, Fidler C, Sargent IL, Murphy MF, Chamberlain PF, Poon PM, Redman CW, Wainscoat JS. Prenatal diagnosis of fetal RhD status by molecular analysis of maternal plasma. N Engl J Med. 1998;339:1734–8. [PubMed]
  • Luettringhaus TA, Cho D, Ryang DW, Flegel WA. An easy RHD genotyping strategy for D-East Asian persons applied to Korean blood donors. Transfusion. 2006;46:2128–37. [PubMed]
  • Marini AM, Matassi G, Raynal V, Andre B, Cartron JP, Cherif-Zahar B. The human Rhesus-associated RhAG protein and a kidney homologue promote ammonium transport in yeast. Nat Genet. 2000;26:341–4. [PubMed]
  • McCarroll SA, Kuruvilla FG, Korn JM, Cawley S, Nemesh J, Wysoker A, Shapero MH, de Bakker PI, Maller JB, Kirby A, Elliott AL, Parkin M, Hubbell E, Webster T, Mei R, Veitch J, Collins PJ, Handsaker R, Lincoln S, Nizzari M, Blume J, Jones KW, Rava R, Daly MJ, Gabriel SB, Altshuler D. Integrated detection and population-genetic analysis of SNPs and copy number variation. Nat Genet. 2008;40:1166–74. [PubMed]
  • Moncharmont P, Dupraz F Juron, Vignal M, Rigal D, Meyer F, Debeaux P. Haemolytic disease of the newborn infant. Long-term efficiency of the screening and the prevention of alloimmunization in the mother: thirty years of experience. Arch Gynecol Obstet. 1991;248:175–80. [PubMed]
  • Mouro I, Colin Y, Cherif-Zahar B, Cartron JP, Le Van Kim C. Molecular genetic basis of the human Rhesus blood group system. Nat Genet. 1993;5:62–5. [PubMed]
  • Ober C, Hyslop T, Hauck WW. Inbreeding effects on fertility in humans: evidence for reproductive compensation. Am J Hum Genet. 1999;64:225–31. [PMC free article] [PubMed]
  • Pennings PS, Hermisson J. Soft sweeps III: the signature of positive selection from recurrent mutation. PLoS Genet. 2006;2:e186. [PMC free article] [PubMed]
  • Perry GH, Dominy NJ, Claw KG, Lee AS, Fiegler H, Redon R, Werner J, Villanea FA, Mountain JL, Misra R, Carter NP, Lee C, Stone AC. Diet and the evolution of human amylase gene copy number variation. Nat Genet. 2007;39:1256–60. [PMC free article] [PubMed]
  • Potter EL. Rh… Its relation to congenital hemolytic disease and to intragroup transfusion reactions.Year Book Publishers; Chicago: 1947.
  • Reed TE. Does reproductive compensation exist? An analysis of Rh data. Am J Hum Genet. 1971;23:215–24. [PMC free article] [PubMed]
  • Sabeti PC, Reich DE, Higgins JM, Levine HZ, Richter DJ, Schaffner SF, Gabriel SB, Platko JV, Patterson NJ, McDonald GJ, Ackerman HC, Campbell SJ, Altshuler D, Cooper R, Kwiatkowski D, Ward R, Lander ES. Detecting recent positive selection in the human genome from haplotype structure. Nature. 2002;419:832–7. [PubMed]
  • Sabeti PC, Schaffner SF, Fry B, Lohmueller J, Varilly P, Shamovsky O, Palma A, Mikkelsen TS, Altshuler D, Lander ES. Positive natural selection in the human lineage. Science. 2006;312:1614–20.[PubMed]
  • Schaffner SF, Foo C, Gabriel S, Reich D, Daly MJ, Altshuler D. Calibrating a coalescent simulation of human genome sequence variation. Genome Res. 2005;15:1576–83. [PMC free article] [PubMed]
  • Singleton BK, Green CA, Avent ND, Martin PG, Smart E, Daka A, Narter-Olaga EG, Hawthorne LM, Daniels G. The presence of an RHD pseudogene containing a 37 base pair duplication and a nonsense mutation in Africans with the Rh D-negative blood group phenotype. Blood. 2000;95:12–8.[PubMed]
  • Suto Y, Ishikawa Y, Hyodo H, Uchikawa M, Juji T. Gene organization and rearrangements at the human Rhesus blood group locus revealed by fiber-FISH analysis. Hum Genet. 2000;106:164–71.[PubMed]
  • Touinssi M, Chiaroni J, Degioanni A, De Micco P, Dutour O, Bauduer F. Distribution of rhesus blood group system in the French Basques: a reappraisal using the allele-specific primers PCR method. Hum Hered. 2004;58:69–72. [PubMed]
  • Urbaniak SJ, Greiss MA. RhD hemolytic disease of the fetus and the newborn. Blood Rev. 2000;14:44–61. [PubMed]
  • Voight BF, Kudaravalli S, Wen X, Pritchard JK. A map of recent positive selection in the human genome. PLoS Biol. 2006;4:e72. [PMC free article] [PubMed]
  • Wagner FF, Flegel WA. RHD gene deletion occurred in the Rhesus box. Blood. 2000;95:3662–8.[PubMed]
  • Wagner FF, Moulds JM, Tounkara A, Kouriba B, Flegel WA. RHD allele distribution in Africans of Mali. BMC Genet. 2003;4:14. [PMC free article] [PubMed]
  • Westhoff CM. The Rh blood group system in review: a new face for the next decade. Transfusion. 2004;44:1663–73. [PubMed]
  • Westhoff CM, Wylie DE. Transport characteristics of mammalian Rh and Rh glycoproteins expressed in heterologous systems. Transfus Clin Biol. 2006;13:132–8. [PubMed]
  • Xue Y, Daly A, Yngvadottir B, Liu M, Coop G, Kim Y, Sabeti P, Chen Y, Stalker J, Huckle E, Burton J, Leonard S, Rogers J, Tyler-Smith C. Spread of an inactive form of caspase-12 in humans is due to recent positive selection. Am J Hum Genet. 2006;78:659–70. [PMC free article] [PubMed]
  • Xue Y, Sun D, Daly A, Yang F, Zhou X, Zhao M, Huang N, Zerjal T, Lee C, Carter NP, Hurles ME, Tyler-Smith C. Adaptive evolution of UGT2B17 copy-number variation. Am J Hum Genet. 2008;83:337–46. [PMC free article] [PubMed]
  • Xue Y, Zhang X, Huang N, Daly A, Gillson CJ, Macarthur DG, Yngvadottir B, Nica AC, Woodwark C, Chen Y, Conrad DF, Ayub Q, Mehdi SQ, Li P, Tyler-Smith C. Population differentiation as an indicator of recent positive selection in humans: an empirical evaluation. Genetics. 2009;183:1065–77. [PMC free article] [PubMed]
  • Yokoyama S. Family size and evolution of Rh polymorphism. J Theor Biol. 1981;92:119–25.[PubMed]
  • Yu X, Wagner FF, Witter B, Flegel WA. Outliers in RhD membrane integration are explained by variant RH haplotypes. Transfusion. 2006;46:1343–51. [PubMed]

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top