Hum. Reprod. Advance Access originally published online on February 24, 2006
Human Reproduction 2006 21(7):1743-1748; doi:10.1093/humrep/del036
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The miscarriage-associated HLA-G –725G allele influences transcription rates in JEG-3 cells
1 Department of Human Genetics and 2 Department of Obstetrics and Gynecology, The University of Chicago, Chicago, IL, USA
3 To whom correspondence should be addressed at: Department of Human Genetics, The University of Chicago, 920 E. 58th Street, CLSC 507c, Chicago, IL 60637, USA. E-mail: c-ober{at}genetics.uchicago.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussion Acknowledgements References |
|---|
BACKGROUND: HLA-G is a non-classical HLA with important immunomodulatory roles in pregnancy. A polymorphism in the promoter region, –725G, was previously associated with sporadic miscarriage in women who were unselected with respect to reproductive history. In this study, the transcription levels of different HLA-G promoter haplotypes were examined to determine whether the miscarriage-associated –725G allele influences transcription. METHODS: Five naturally occurring promoter haplotypes and three variant haplotypes created by site-directed mutagenesis were sub-cloned into luciferase expression vectors and transfected into JEG-3 cells. Expression levels of these eight haplotypes were examined in cultured cells before and after treatment with interferon-
(IFN-), cytosine-5-DNA methyltransferase (M. SssI) and 5-aza-2'-deoxycytidine. Differences in expression levels between haplotypes were determined by analysis of variance (ANOVA). RESULT: Promoter haplotypes with the miscarriage-associated –725G allele were expressed at significantly higher levels in all culture conditions compared with otherwise identical haplotypes that had a –725C or –725T allele. CONCLUSION: Variation in the HLA-G promoter region influences transcription rates. Contrary to expectations, increased expression of HLA-G may be disadvantageous in some pregnancies.
Key words: gene expression/HLA-G/methylation/miscarriage/polymorphisms
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion Acknowledgements References |
|---|
HLA-G is a non-classical class I HLA molecule with limited polymorphism in its coding region. Multiple isoforms of this unique gene are highly expressed throughout pregnancy in fetal placental cells that embed deeply into the maternal decidua, invading the spiral arteries and directly contacting maternal immune cells (reviewed in Le Bouteiller et al., 2003
; Hunt et al., 2005). Although the exact roles that HLA-G plays in pregnancy are not completely known, it is now clear that it has important immunomodulatory properties (LeMaoult et al., 2005). For example, HLA-G can protect against natural killer cell-mediated lysis (Chumbley et al., 1994; Pazmany et al., 1996), inhibit maternal CD8+ T cells (Fournel et al., 2000; Solier et al., 2002; Contini et al., 2003; Morales et al., 2003), prevent proliferation of CD4+ T cells (LeMaoult et al., 2004) and tolerize dendritic cells (Ristich et al., 2005). It is likely, therefore, that HLA-G contributes to the survival of the genetically foreign fetus.
That HLA-G is essential for successful pregnancy has not been proven. In fact, homozygosity for a naturally occurring mutation, 1597
C, that is a null for the full length transmembrane (G1) and soluble (G5) isoforms (Ober et al., 1998a) has been reported in healthy adults (Ober et al., 1998a; Castro et al., 2000), indicating that the G1 and G5 isoforms are not required for fetal survival. It is possible that, in these cases, other isoforms suffice (Ober et al., 1998a; Le Discorde et al., 2005). In fact, the 1597C allele occurs at relatively high frequency among Africans and African-Americans (frequency 5–11%), although it is rare among populations of European and Asian descent (frequency 0–3%; reviewed in Aldrich et al., 2002). The surprisingly high frequency of the null mutation among people of African ancestry led us to suggest that in areas with a high pathogen load, reduced expression of HLA-G in carriers of the null allele might be beneficial during pregnancy, perhaps allowing for a more robust maternal immune response to intrauterine pathogens (Aldrich et al., 2002). If so, the null allele may have been selected in populations with a high pathogen load and driven to higher frequencies, such as those observed in Africa. Indeed, one study of the evolutionary history of the 1597C allele supported this hypothesis and suggested that positive selection has acted on this null allele (Aldrich et al., 2002).
In this context then, it was surprising that two studies in primarily Caucasian couples reported an increased frequency of the 1597C allele in women with recurrent miscarriage (Aldrich et al., 2001; Pfeiffer et al., 2001). Although rare in Caucasian controls (<1%; Pfeiffer et al., 2001; Aldrich et al., 2002), 3.5% of North American and 6.4% of European women with recurrent miscarriage carried this allele. Thus, although the G1 and G5 isoforms are not essential for fetal survival, decreased levels of these isoforms in fetuses carrying the 1597C allele may be a risk factor for miscarriage. More recently, the presence of an insertion of 14 bp in the untranslated exon 8 has been associated with IVF failures (Hviid et al., 2004a). This variant had previously been associated with lower HLA-G mRNA levels in placental cells (Hviid et al., 2003), further suggesting that HLA-G genotype and reduced expression of HLA-G are risk factors for pregnancy failure.
To further explore the possibility that other variation in HLA-G that could influence expression levels was associated with miscarriage, we and others re-sequenced the region upstream of exon 1, including all of the known promoter elements, in individuals with nearly all of the known HLA-G alleles (Hviid et al., 1999; Ober et al., 2003; Tan et al., 2005). We identified 27 single-nucleotide polymorphisms (SNPs) and showed that one, –725G, was associated with sporadic miscarriage in the Hutterites, a founder population of European descent who have participated in a prospective study of pregnancy outcome since 1986 (Ober et al., 1992, 1998b, 2003). The odds ratio for miscarriage among couples in whom both partners carried the –725G allele was 2.72 (95% CI, 1.08–6.87), slightly higher than the risk associated with maternal age in the same population (Ober et al., 2003).
The presence of a G at this site resulted in the creation of a CpG dinucleotide, which was methylated in DNA derived from blood mononuclear cells. We speculated that the –725G allele was associated with reduced expression of HLA-G, perhaps because the methylated dinucleotide interfered with binding of the interferon response factor-1 (IRF-1) to a nearby interferon-specific regulatory element (ISRE)-binding site (Figure 1). In this study, we report the results of expression studies of five naturally occurring HLA-G promoter haplotypes and three in vitro mutated haplotypes and show that, contrary to expectations, the miscarriage-associated –725G allele is associated with increased transcription in JEG-3 cells.
|
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion Acknowledgements References |
|---|
We used luciferase reporter assays to determine whether variation in the promoter region of HLA-G influences transcription. A 1500-bp fragment from each of five unique promoter haplotypes was sub-cloned into a pCR4-TOPO vector (TOPOP TA cloning kit; Invitrogen, Carlsbad, CA, USA). Clones with each of the five haplotypes were identified by sequencing, which also confirmed that no mutations were introduced during the PCR. To determine the effects of different nucleotides at position –725 and –1121 on different haplotype backgrounds, we used site-directed mutagenesis (QuickChange XL; Stratagene, La Jolla, CA, USA) to create haplotypes that do not exist in nature and included them in the luciferase experiments.
A 1379 bp fragment from –1412 to –33 from the translational start site (ATG in exon 1) was PCR-amplified from each of these clones using the primers: forward CAC-GGTACC-ACTGGAGTGTTTTAG GTGGAGA and reverse CAC-CTCGAG-GTGAGCGAGGACTTTA GAACCA. The forward primer included a XhoI site and the reverse primer an MluIF. PCR products were cloned into the promoterless pGL3-basic vector, upstream of the firefly luciferase gene (Promega, Madison, WI, USA). One microgram of the reporter construct (or pGL3 empty vector) was transfected into sub-confluent JEG-3 cells. Cells were harvested and lysed at 40 h post-transfection. Transfection efficiency was normalized by co-transfecting 12 ng of pRL.SV40 Renilla luciferase vector (Promega). The dual luciferase assay was performed by measuring the firefly and Renilla luciferase activities of the same sample, and the results were expressed as the ratio of firefly to Renilla luciferase activity.
All experiments were also performed after culturing of JEG-3 cells with 1000 U/ml interferon (IFN-; Chemicon, Temecula, CA, USA) to determine whether allelic variants at –725 influenced binding of IRF-1 to the ISRE-binding site and using plasmids treated with 12 U/µg CpG methylase (M. SssI; New England Biolabs, Ipswich, MA, USA) and cells cultured with 5 µM 5-aza-2'-deoxycytidine (Sigma, St. Louis, MO, USA) to determine whether haplotype-specific methylation patterns influence expression levels. All experiments were performed in triplicate. All experiments in cells cultured in IFN- and/or treated with SssI were performed in parallel with an equal number of experiments in untreated cultured cells (baseline) and were compared against these parallel experiments in the statistical analysis. Altogether, 14 independent experiments (each in triplicate) were performed at baseline. Two experiments each were performed after culturing with IFN- alone and in combination with M. SssI, and six experiments were performed after treatment with M. SssI alone (all in triplicate).
Promoter activity was compared using two-way analysis of variance (ANOVA) with repeated measures, considering each replicate individually and including ‘experiment’ as a covariate.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion Acknowledgements References |
|---|
We previously reported 27 SNPs in the approximately 1300 bp upstream of exon 1 that defined 13 unique haplotypes (Tan et al., 2005
). Seven of these haplotypes, defined by 22 SNPs, are present in the Hutterites (Table I). Seven of the 22 SNPs, including –725C/G/T, either create or destroy a CpG dinucleotide (denoted by superscript letter ‘a’ in Table I), potentially resulting in differential methylation patterns among promoter haplotypes. To study the transcriptional properties of the different promoters, we sub-cloned five of these promoters upstream of a luciferase reporter gene.
|
Two promoters carried a –725G allele (G*010101b and G*010101c), one carried a –725T allele (G*010301a) and two carried a –725C allele (G*010101a and G*010102). The results of these experiments are shown in Figure 2A. Overall, there were significant differences in expression levels between promoters (ANOVA; P = 0.0012). However, contrary to our expectations, the promoters with the miscarriage-associated –725G allele had significantly higher expression levels compared with promoters carrying a –725C (G*010101b, G*010101c versus G*010101a, G*010102, P < 0.001) or a –725T (versus G*010301, P < 0.001; Figure 2A). There were no differences in expression between promoters with a –725T compared with those carrying a –725C (G*010101a, G*010102 versus G*010103, P = 0.94).
|
The expression of all promoter haplotypes increased in cells cultured in IFN-
(Figure 2B; baseline vs. treated for each haplotype, P < 0.001), similar to a previous study (Lefebvre et al., 2001). Expression of all haplotypes significantly decreased after methylation with M. SssI (Figure 2C) but increased with the combined treatment (Figure 2D). Expression of all promoters increased after demethylation with 5-aza-2'-deoxycytidine (not shown), as previously reported (Moreau et al., 2003). The number of CpG dinucleotides (i.e. methylation targets) on the different promoters was not associated with expression levels after treatment with M. SssI (Sign test, P = 0.812). In fact, despite having the most CpG sites, the G*010301 decreased on average by only 36% compared with average decreases between 58% and 76% for the other haplotypes. Moreover, regardless of treatment, promoters with a –725G allele remained the most highly expressed. Because the G*010101a promoter is identical to the G*010101b and G*010101c promoters except at nucleotides –725 and –1121, these results suggested that the –725G allele, and perhaps the –1121T allele, accounted for the observed increase in expression. To explore this possibility, we next used site-directed mutagenesis to examine the specific effects of the –725C/G/T and –1121C/T alleles on different promoter backgrounds (Figure 3). Regardless of the promoter background, the –725G allele was associated with increased expression levels. The expression of the relatively low-expressing promoter, G*010102, increased on average by 37% when the C at –725 was mutated to a G (P = 0.012), and the expression of the high-expressing promoter, G*010101c, decreased on average by 44% and 58% when the G at –725 was mutated to a C (P = 0.0008) or to a T (P = 0.0037), respectively.
|
Thus, a G at nucleotide position –725 in the promoter region of HLA-G results in increased transcription rates on haplotype backgrounds differing at 15 of 27 polymorphic sites and representing two evolutionarily divergent promoter sequences (Tan et al., 2005
). Moreover, the –1121T allele was not sufficient to account for the increased expression of the G*010101c promoter. However, although the change of –725C to –725G increased expression of the G*010102 promoter, expression levels remained lower than the naturally occurring haplotypes with a –725G, indicating that other sequence differences between the G*010101 and G*010102 promoters also influence expression. Nonetheless, these studies demonstrate that the miscarriage-associated –725G allele results in increased expression level on divergent haplotype backgrounds and that the mechanism for increased expression is unrelated to either methylation status or response to IFN-, as we previously hypothesized (Ober et al., 2003). |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion Acknowledgements References |
|---|
Assuring survival of the fetus in mammalian pregnancy required the evolution of adaptive mechanisms that allowed for immunologic tolerance of the fetal allograft while at the same time maintaining immune competence against invading pathogens. These adaptations likely involve many pathways, but the expression of HLA-G in human pregnancy may represent one such mechanism for protecting the fetus from cytotoxic maternal immune responses. Because fetuses carrying a null allele for the HLA-G1 and HLA-G5 isoforms (Ober et al., 1998a
) are less likely to survive to term (Aldrich et al., 2001; Pfeiffer et al., 2001), we hypothesized that increased expression of HLA-G is protective and facilitates successful pregnancy. The results of the studies presented here indicate otherwise and suggest that extreme levels of expression, whether they are high or low, may be deleterious under different conditions of pregnancy.
Although we do not know the causes of miscarriages in the Hutterites, all were sporadic losses interspersed between many successful pregnancies and nearly all occurred in the first trimester (Ober et al., 2003). In fact, the Hutterites have one of the highest natural fertility rates ever recorded (Sheps, 1965; Ober et al., 1999), despite having a relatively high frequency of the –725G allele (approximately 16%). Thus, similar to the 1597C allele, many fetuses carrying the –725G allele survive to term. Although the association of a high-expressing HLA-G allele with miscarriage seems paradoxical, this finding is consistent with our previous study demonstrating positive selection on the 1597C allele in Africans (Aldrich et al., 2002), as discussed above, and further suggests that high levels of expression of HLA-G may be a risk factor for miscarriage in some pregnancies. Selection against both high- and low-expressing promoter haplotypes in different environments could contribute to the unusual genetic features of the HLA-G promoter region and help maintain two evolutionarily divergent haplotype groups (Tan et al., 2005).
A potential limitation of these studies is that they were performed in a choriocarcinoma cell line, JEG-3. These cells are known to have altered properties and do not always mimic the responses seen in primary cells. Therefore, we cannot rule out the possibility that, in primary trophoblast cells, the –725G allele will be associated with decreased transcription. We think this is unlikely, however, because we have observed similar patterns of expression, that is increased expression of promoters with –725G, in other, non-trophoblast cell lines (unpublished data). Nonetheless, until these studies are conducted in primary trophoblast cells, it remains a possibility.
The –725G variant is present on the most common HLA-G allele, G*010101, and accounts for 30% of G*010101 alleles in the Hutterites, 20% in outbred European Americans, 36% in African-Americans and 9% in Han Chinese (Ober et al., 2003; Tan et al., 2005). The G*010101 allele also carries the 14 bp deletion allele in the untranslated exon 8, which has been previously associated with increased mRNA expression in placental cells compared with the insertion allele that is present on the second most common, G*010102, allele (Hviid et al., 2003). Although the G*010101a promoter that is present on most G*010101 alleles and the G*010102 promoter did not show differences in expression levels in our study, it is possible that the higher expression levels of the subset of G*010101 alleles carrying –725G contribute to the observed differences between the 14 bp deletion and insertion alleles, due to the extensive linkage disequilibrium among variants in this gene (Hviid et al., 1999; Ober et al., 2003; Hviid et al., 2004b). However, it is also possible that different polymorphisms in the HLA-G gene influence transcription, splicing efficiencies and mRNA stability.
In conclusion, we demonstrated that a single nucleotide in the promoter region of HLA-G significantly influences expression levels in a trophoblast cell line. Because this same variant was previously associated with sporadic miscarriage in a healthy, unselected population, we suggest that in some pregnancies increased expression levels of HLA-G may be deleterious. We speculate that these may be pregnancies with sub-clinical infection and that the increased levels of HLA-G in these pregnancies may compromise maternal cellular immunity to invading pathogens. Moreover, we did not observe significant differences in transcription levels between HLA-G promoters with a –725C or –725T, even though they differed from each other at >10 polymorphic sites (G*010101a versus G*010102 versus G*010301a), and could not, in general, demonstrate haplotype-specific responses to IFN- or after chemical methylation and demethylation, although the G*010301 promoter was associated with a relatively smaller decrease in expression following treatment with M. SssI. It is possible, however, that additional differences between haplotypes would become apparent in primary trophoblast cells or after treatment with other cytokines or modulators of expression. Studies addressing all of these possibilities are currently underway.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion Acknowledgements References |
|---|
We thank Lin Pan, Paul Kogut, Geoffrey Wool, Katinka Vigh, Sunita Shuklas and Lauren Weiss for technical assistance. This work was supported by NIH grants HD21244 and HL72414.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion Acknowledgements References |
|---|
Aldrich CL, Stephenson MD, Karrison T, Odem RR, Branch DW, Scott JR, Schreiber JR, Ober C. (2001) HLA-G genotypes and pregnancy outcome in couples with unexplained recurrent miscarriage. Mol Hum Reprod 7:1162–1172.
Aldrich C, Wambebe C, Odama L, Di Rienzo A, Ober C. (2002) Linkage disequilibrium and age estimates of a deletion polymorphism (1597C) in HLA-G suggest non-neutral evolution. Hum Immunol 63:405–412.[CrossRef][Medline]
Castro MJ, Morales P, Rojo-Amigo R, Martinez-Laso J, Allende L, Varela P, Garcia-Berciano M, Guillen-Perales J, Arnaiz-Villena A. (2000) Homozygous HLA-G*0105N healthy individuals indicate that membrane-anchored HLA-G1 molecule is not necessary for survival. Tissue Antigens 56:232–329.[CrossRef][ISI][Medline]
Chumbley G, King A, Robertson K, Holmes N, Loke YW. (1994) Resistance of HLA-G and HLA-A2 transfectants to lysis by decidual NK cells. Cell Immunol 155:312–322.[CrossRef][ISI][Medline]
Contini P, Ghio M, Poggi A, Filaci G, Indiveri F, Ferrone S, Puppo F. (2003) Soluble HLA-A,-B,-C and -G molecules induce apoptosis in T and NK CD8+ cells and inhibit cytotoxic T cell activity through CD8 ligation. Eur J Immunol 33:125–134.[CrossRef][ISI][Medline]
Fournel S, Aguerre-Girr M, Huc X, Lenfant F, Alam A, Toubert A, Bensussan A, Le Bouteiller P. (2000) Cutting edge: soluble HLA-G1 triggers CD95/CD95 ligand-mediated apoptosis in activated CD8+ cells by interacting with CD8. J Immunol 164:6100–6104.
Hunt JS, Petroff MG, McIntire RH, Ober C. (2005) HLA-G and immune tolerance in pregnancy. FASEB J 19:681–693.
Hviid TV, Sorensen S, Morling N. (1999) Polymorphism in the regulatory region located more than 1.1 kilobases 5' to the start site of transcription, the promoter region, and exon 1 of the HLA-G gene. Hum Immunol 60:1237–1244.[CrossRef][ISI][Medline]
Hviid TV, Hylenius S, Rorbye C, Nielsen LG. (2003) HLA-G allelic variants are associated with differences in the HLA-G mRNA isoform profile and HLA-G mRNA levels. Immunogenetics 55:63–79.[ISI][Medline]
Hviid TV, Hylenius S, Lindhard A, Christiansen OB. (2004a) Association between human leukocyte antigen-G genotype and success of in vitro fertilization and pregnancy outcome. Tissue Antigens 64:66–69.[CrossRef][ISI][Medline]
Hviid TV, Larsen LG, Hoegh AM, Bzorek M. (2004b) HLA-G expression in placenta in relation to HLA-G genotype and polymorphisms. Am J Reprod Immunol 52:212–217.
Le Bouteiller P, Legrand-Abravanel F, Solier C. (2003) Soluble HLA-G1 at the materno-foetal interface – a review. Placenta 24:Suppl. A, S10–S15.
Le Discorde M, Le Danff C, Moreau P, Rouas-Freiss N, ED Carosella. (2005) HLA-G*0105N null allele encodes functional HLA-G isoforms. Biol Reprod 73:280–288.
Lefebvre S, Berrih-Aknin S, Adrian F, Moreau P, Poea S, Gourand L, Dausset J, ED Carosella, Paul P. (2001) A specific interferon (IFN)-stimulated response element of the distal HLA-G promoter binds IFN-regulatory factor 1 and mediates enhancement of this nonclassical class I gene by IFN-beta. J Biol Chem 276:6133–6139.
LeMaoult J, Krawice-Radanne I, Dausset J, ED Carosella. (2004) HLA-G1-expressing antigen-presenting cells induce immunosuppressive CD4+ T cells. Proc Natl Acad Sci USA 101:7064–7069.
LeMaoult J, Rouas-Freiss N, ED Carosella. (2005) Immuno-tolerogenic functions of HLA-G: relevance in transplantation and oncology. Autoimmun Rev 4:503–509.[Medline]
Morales PJ, Pace JL, Platt JS, Phillips TA, Morgan K, Fazleabas AT, Hunt JS. (2003) Placental cell expression of HLA-G2 isoforms is limited to the invasive trophoblast phenotype. J Immunol 171:6215–6224.
Moreau P, Mouillot G, Rousseau P, Marcou C, Dausset J, ED Carosella. (2003) HLA-G gene repression is reversed by demethylation. Proc Natl Acad Sci USA 100:1191–1196.
Ober C, Hyslop T, Hauck WW. (1999) Inbreeding effects on fertility in humans: evidence for reproductive compensation. Am J Hum Genet 64:225–231.[CrossRef][ISI][Medline]
Ober C, Elias S, Kostyu DD, Hauck WW. (1992) Decreased fecundability in Hutterite couples sharing HLA-DR. Am J Hum Genet 50:6–14.[ISI][Medline]
Ober C, Aldrich C, Rosinsky B, Robertson A, Walker MA, Willadsen S, Verp MS, Geraghty DE, Hunt JS. (1998a) HLA-G1 protein expression is not essential for fetal survival. Placenta 19:127–132.[CrossRef][ISI][Medline]
Ober C, Hyslop T, Elias S, Weitkamp LR, Hauck WW. (1998b) Human leukocyte antigen matching and fetal loss: results of a 10-year prospective study. Hum Reprod 13:33–38.
Ober C, Aldrich CL, Chervoneva I, Billstrand C, Rahimov F, Gray HL, Hyslop T. (2003) Variation in the HLA-G promoter region influences miscarriage rates. Am J Hum Genet 72:1425–1435.[CrossRef][ISI][Medline]
Pazmany L, Mandelboim O, Vales-Gomez M, Davis DM, Reyburn HT, Strominger JL. (1996) Protection from natural killer cell-mediated lysis by HLA-G expression on target cells. Science 274:792–795.
Pfeiffer KA, Fimmers R, Engels G, van Der Ven H, van Der Ven K. (2001) The HLA-G genotype is potentially associated with idiopathic recurrent spontaneous abortion. Mol Hum Reprod 7:373–378.
Ristich V, Liang S, Zhang W, Wu J, Horuzsko A. (2005) Tolerization of dendritic cells by HLA-G. Eur J Immunol 35:1133–1142.[CrossRef][ISI][Medline]
Sheps MC. (1965) An analysis of reproductive patterns in an American isolate. Popul Stud 19:65–80.
Solier C, Mallet V, Lenfant F, Bertrand A, Huchenq A, Le Bouteiller P. (2001) HLA-G unique promoter region: functional implications. Immunogenetics 53:617–625.[CrossRef][ISI][Medline]
Solier C, Aguerre-Girr M, Lenfant F, Campan A, Berrebi A, Rebmann V, Grosse-Wilde H, Le Bouteiller P. (2002) Secretion of pro-apoptotic intron 4-retaining soluble HLA-G1 by human villous trophoblast. Eur J Immunol 32:3576–3586.[CrossRef][ISI][Medline]
Tan Z, Shon AM, Ober C. (2005) Evidence of balancing selection at the HLA-G promoter region. Hum Mol Genet 14:3619–3628.
Submitted on November 29, 2005; resubmitted on January 6, 2006; accepted on January 19, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||