Molecular Analysis of Rh Transcripts and Polypeptides From Individuals Expressing the D^VI Variant Phenotype: An RHD Gene Deletion Event Does Not Generate All D^VIccEe Phenotypes

THE HUMAN Rh D antigen is a collection of epitopes expressed by a complex composed of two different polypeptide chains, the 30-kD Rh D polypeptide and 50-kD Rh glycoprotein.¹ The absence of this complex from the erythrocyte membrane, which results from a deleted2 or defective³RHD gene, causes the D-negative phenotype. Defects in the structural gene that encodes the Rh glycoprotein result in the extremely rare Rh_null phenotype, indicating that its presence during biosynthesis is critical for the assembly of all Rh proteins.4 

Some D-positive individuals, initially identified by the presence of allo anti-D in their sera,5 have been classified into seven distinct categories (D^I to D^VII; with D^I being obsolete) based on the reactivity of these antisera with red blood cells of D-variant phenotypes.6 These D-variant (or partial D) individuals lack one or more D epitopes (numbered epD1 to epD9 using the 9-epitope model7,8 or epD1 to epD30 using the 30-epitope model9 ), and thus can be immunized to them on exposure to normal D-positive erythrocytes. The D^VI variant phenotype lacks most epitopes of any D-variant described (epD1, 2, 5-8 using the 9-epitope model or epD1-4, 7-22, 26-29 using the 30-epitope model), and is the most frequent of the D-variant phenotypes with gene frequencies of 0.02%, 0.02%, 0.05%, and 0.03% in English,10 United States,11 Dutch,12 and Australian13 populations, respectively. The D^VI phenotype is the only known D-variant phenotype that has made a clinically significant anti-D capable of causing severe HDN and neonatal death.14 D^VI phenotype erythrocytes type as D-positive when polyclonal anti-D typing reagents are used, which is problematic because D^VI mothers of D-positive infants should be given postpartum anti-D.

The molecular basis of the D^VI phenotype has been described previously, and is proposed to occur through two different genetic mechanisms.15 The D^VICcee phenotype appears to arise through a gene conversion or double-crossing over event resulting in a hybrid RHD-RHCE-RHD gene. The hybrid D^VI gene is composed of exons 1-3 of the RHD gene, 4-6 of RHCE, and 7-10 of the RHD gene. The D^VIccEe phenotype is proposed to result from a partial RHD gene deletion, whereby exons 4-6 are lost. Rh D transcripts that encode a truncated 266-amino acid Rh D protein have been found, and it has been assumed that this protein, when expressed at the erythrocyte surface, is responsible for the D^VIccEe phenotype.

Our findings corroborate those of Mouro et al15 concerning the D^VICe phenotype. However, we present contrasting evidence concerning the D^VIccEe phenotype. Our data suggest that this phenotype can occur as a result of a gene conversion event whereby exons 4 and 5 of the RHD gene are replaced by RhcE equivalents.

Erythrocytes and antibodies.Erythrocytes of phenotype D^VIccEe (donor Vi) were obtained from Dr Anna Pirkola (Finnish Red Cross Blood Transfusion Service, Helsinki, Finland). Erythrocytes from a further three individuals of D^VIccEe phenotype were obtained from the Institute of Haematology, Prague; and D^VICcee phenotype erythrocytes were from the Red Cross Blood Transfusion Service (Sydney, Australia) and the National Blood Service (Bristol, UK). Monoclonal antibodies (MoAbs) used were the IBGRL D typing panel (see Table 1). MoAb H41 (anti-epD3; reactive with D-positive, D^VI, and D^II phenotype erythrocytes) was obtained from Dr Hans Sonneborn (Biotest AG, Dreieich, Germany).

Immunoblotting with rabbit antisera against Rh synthetic peptides.Rabbit sera immunoreactive with the C-terminal (anti-Ct) and fourth external domain (anti-loop 4e) of Rh polypeptides were those described previously.16,17 Immunoblotting was performed on anti-D immunoprecipitates prepared using an MoAb strongly reactive with D^VI phenotype erythrocytes; H41, using the method of Moore et al.18 

Polymerase chain reaction (PCR) Rh D typing and analysis of Rh transcripts.Genomic DNA was prepared from lymphocytes derived from D^VIccEe and D^VICcee phenotype individuals as described,19 and analyzed using an RHD intron 4/exon 10 multiplex assay.13,20 Briefly, the assay involves the coamplification of RHD exon 10 and intron 4, and a control RHCE intron 4 PCR product. Intron 4 RHD specificity is achieved by using a reverse PCR primer that straddles the RHD-specific deletion of this intron13,21 (Fig 1). Rh transcripts were analyzed by reverse-transcription followed by PCR (RT-PCR) using total RNA templates isolated from peripheral blood using the ammonium bicarbonate/chloride lysis method.22 PCR was performed using Bio-X-Act proof-reading DNA polymerase cocktail (Bioline, London, UK) on cDNA templates prepared using standard techniques.23 The forward primer corresponded to nucleotides (nts) −81 to −56 and had the sequence 5′-TCCCTCAAGCCCTCAAGTAG-3′. The reverse primer had the sequence 5′-GTACAAATGCAGGCAACAGTG -3′, corresponding to nts 1328-1308 of the Rh30A cDNA sequence. PCR amplification was performed using the following conditions: 94°C for 1 minute, 55°C for 1 minute, and 72°C for 3 minutes for 35 cycles in a 50-μL reaction mix composed of the following:- X1 Optiperform buffer (Bioline), 2.5 mmol/L MgCl₂ 1 μmol/L each primer 1.25 mmol/L each dNTP, and 2.0 U Bio-X-Act DNA polymerase cocktail. PCR products corresponding to full-length Rh transcripts were gel-purified by agarose gel electrophoresis, and cloned into pCRII vector (Invitrogen, San Diego, CA) using the manufacturer's instructions. DNA sequencing was performed on both strands of the plasmid DNA using cycle sequencing and dye-terminator chemistry on an Applied Biosystems 373A DNA sequencer (Warrington, UK).

Serology.The D^VI phenotype was confirmed on all erythrocytes under investigation by serological testing with our panel of monoclonal anti-D to confirm the presence or absence of epitopes characteristic of this phenotype. The results obtained for D^VI phenotype erythrocytes (donors 8512 [D^VICcee] and Vi [D^VIccEe]) are illustrated in Table 1.

Immunoblotting of anti-D (H41; epD3) immunoprecipitates from D^VICcee and D^VIccEe phenotype erythrocytes. Immunoprecipitation was performed using erythrocytes obtained from four individuals expressing the D^VIccEe phenotype, an individual with the D^VICcee phenotype, and dce and DcE phenotypes. Immune complexes were analyzed by immunoblotting with anti-Ct and anti-loop 4e sera after separation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). It was found that all immunoprecipitated Rh D^VI polypeptides were of identical size (≈35 kD) and both reacted strongly with anti-Ct and anti-loop 4e (Fig 2).

Multiplex Rh D typing of D^VI genomic DNA (gDNA).Using a multiplex assay that analyzes both intron 4 and exon 10 of the RHD gene,13 we analyzed gDNA derived from 13 different individuals expressing the D^VI phenotype. All samples gave a positive RHD exon 10 band but with no RHD intron 4 band while retaining the control RHCE intron 4 band (six samples illustrated in Fig 3). This is consistent with hybrid RHD-RHCE-RHD genes being present in all individuals examined. One individual (donor number P2; of phenotype D^VIccEe) showed the presence of the RHD exon 10 band, but with a weak control RHCE intron 4 band (Fig 3). High-molecular-weight material in gel corresponds to gDNA template.

Sequence analysis of Rh transcripts derived from D^VI phenotype individuals.When Rh transcripts were isolated by subcloning and sequenced (using a pool of two different gel-purified PCR reactions) from an individual expressing the D^VICcee phenotype (donation no. 8512), nine clones were found to be derived from RHCE gene(s). Of these a cDNA encoding c and e antigens was fully sequenced and was found to encode an Rh ce polypeptide, and another six clones of this type were identified. Two transcripts proposed to express Rh Ce antigens were found in this individual. Another species of transcript was found, and was identical to that described by Mouro et al,15 ie, composed of RHD exons 1-3, RHCE exons 4-6, and RHD exons 7-10. Two clones of this type were sequenced fully and were found to be identical (Fig 4). Sequence analysis of full-length Rh transcripts isolated by RT-PCR from a D^VIccEe phenotype individual (donor Vi) (using a pool of two different gel-purified PCR reactions) showed that Rh ce and Rh cE transcripts were present. Three Rh cE transcripts and one Rh ce transcript were sequenced fully. In addition, two identical hybrid transcripts were identified and sequenced, and were found to be composed of exons 1-3 RHD,; 4-5 RHCE, and 6-10 RHD (Fig 4). The exon 5 sequence of this hybrid transcript differed from the hybrid D^VICcee transcripts inasmuch as they encoded Pro₂₂₆ , the amino acid known to be critical for Rh E expression.24 

This report describes the characterization of Rh transcripts and Rh D^VI polypeptides in individuals expressing the Rh D^VI phenotype. Although our findings confirm those of Mouro et al,15 concerning the D^VICcee phenotype, we were unable to show the existence of truncated Rh D polypeptide species in individuals expressing the D^VIccEe phenotype, which was proposed to occur by these workers. In recent years it has been suggested that Rh antigen expression is possible on shortened Rh polypeptide isoforms that arise by alternate splicing of Rh heterogeneous nuclear RNA (hnRNA). A model for the differential expression of Rh C or c by an alternately spliced transcript of the RHCE gene has been proposed while Rh E or e antigen expression is proposed to occur on full-length Rh polypeptides.24-26 This hypothesis is limited inasmuch as no evidence for the existence of such protein structures at the erythrocyte surface has ever been found. Two recent studies have refuted this hypothesis. Firstly, expression of Rh antigens using an ex vivo system indicates that alternate splicing is not necessary for Rh c or E expression.27 Secondly, analysis of immunopurified Rh C and c proteins indicate that they exist as full-length Rh polypeptides in the erythrocyte membrane.17 These findings indicate that Rh C or c and E or e antigen expression both occur on the same polypeptide.

It is now clear that Rh CcEe and probably D^VI phenotype expression does not occur on shortened Rh isoforms. Our studies described here provide unequivocal evidence that Rh D^VI proteins purified from four D^VIccEe individuals studied here are not expressed on truncated Rh protein isoforms. We have shown that immunopurified Rh D proteins from all D^VI phenotype erythrocytes studied express the fourth external domain of Rh proteins, which would be absent on the hypothetical shortened Rh D^VIcE protein isoform (Fig 2). These proteins are of identical molecular size to other Rh polypeptides (confirming earlier findings15 ), and suggest that all Rh D^VI proteins are full-length (416-amino acid) structures. We have previously shown that our SDS-PAGE analysis of Rh proteins and proteolytic fragments is sufficient to resolve amino acid differences of 50 amino acids or less,17 which refutes the hypothesis that proposed truncated Rh protein isoforms cannot be resolved by conventional SDS-PAGE. Data presented here cast some doubt as to whether the previously proposed15 truncated (266-amino acid) D^VIcE protein is expressed at the erythrocyte membrane surface. The truncated protein could not be shown in any individual expressing the D^VIcE haplotype in this study. The high degrees of interaction between the Rh proteins and Rh glycoprotein4,17 would indicate that disruptions to either Rh glycoprotein or Rh protein structure (in-frame stop codons, missplicing resulting in truncated polypeptide products, and/or altered reading frames) will prevent expression. It is possible that all shortened Rh transcripts described previously represent misspliced versions of Rh hnRNA primary transcripts.

Illustration of full-length hybrid RHD-RHCE-RHD transcripts in an individual expressing the D^VIccEe phenotype (donor Vi) confirms that a deleted RHD gene is not present in this individual. The Rh D^VI transcripts appear to be derived from a hybrid RHD-RHCE-RHD gene comprising exons 1-3 RHD, 4-5 RHCE, and exons 6-10 RHD (Fig 5). This hybrid gene differs from that found in individuals expressing the D^VICcee phenotype (exons 1-3 RHD, 4-6 RHCE, and 7-10 RHD; Fig 3). This indicates that residues encoded by RHD exons 4 and 5 are involved in epD1-2, 5-8 expression (using the 9-epitope model7,8 ) or epD1-4, 7-22, and 26-28 (using the recent 30-epitope model9 ). Exon 6 of the RHD gene encodes only two amino acid differences to the RHCE gene (Val₃₀₆ and Tyr₃₁₁ ), and these are likely to be located toward the internal face of the erythrocyte membrane (Fig 5). This finding indicates that RHD-specific exon 6 encoded residues are not directly involved in expression of epitopes absent from D^VI phenotype erythrocytes. It is possible that other genetic mechanisms account for the D^VI phenotype, but we assume that all involve rearrangement of at least exons 4 and 5 of the RHD gene, because all individuals studied were RHD intron 4-negative when typed by the multiplex assay. All D^VI polypeptides are likely to be full-length protein structures similar to those depicted in Fig 5.

A low-frequency antigen of the Rh system, BARC (Rh52), is associated with the majority of D^VICcee (a small number of D^VICcee phenotype individuals are BARC-negative) but not D^VIccEe phenotypes.28,29 Anti-BARC does not react with D^VIccEe phenotype erythrocytes, thus sequence differences in the Rh D proteins of these two phenotypes must account for the expression of this antigen. The only exofacial difference between the D^VICe and D^VIcE proteins is that of residue 226, localized on the fourth external domain (Fig 5). It is possible that Ala₂₂₆ (expressed by the D^VICe Rh D protein, and known to be involved in Rh e antigen expression24 ), in combination with the unique external domain structure of D^VI proteins, results in BARC antigen expression. Lack of BARC expression on D^VIcE, Rh cE, and Rh D proteins indicate that the antigen must be discontinuous, and that the presence of Pro₂₂₆ on the D^VIcE protein prevents expression. Alternatively, the D^VIcE Rh D protein lacks the third palmitoylation site (Cys₃₁₁ -Leu-Pro) in common with normal Rh D proteins. The D^VICe Rh D protein has an RHCE-specific palmitoylation site at this position (Fig 5). It is possible that conformational differences induced by differing degrees of palmitoylation results in expression of the BARC antigen.

In the United Kingdom, D-typing reagents (ie, monoclonal anti-D) are adopted to deliberately type D^VI patients as D-negative. Thus, after pregnancy D^VI phenotype mothers are given prophylactic anti-D. However, in instances where D^VI phenotype individuals are identified as D-positive by use of inappropriate D typing reagents (eg, polyclonal anti-D), D^VI mothers may become alloimmunized to their normal D-positive infants where prophylactic anti-D is not administered. In cases where alloimmunization has occurred in such cases, prenatal detection of a fetus carrying a normal RHD gene can be achieved by analysis of fetal DNA with the multiplex assay. A normal RHD intron 4–derived band would indicate that the fetus carries a paternally inherited RHD gene. The absence of the RHD intron 4–derived band and the presence of the RHD exon 10–derived band would indicate that the fetus carries only the maternally derived RhD^VI gene, and would hence not be at risk of HDN. Using the multiplex assay described here, we have analyzed 357 genomic DNAs derived from D-positive, D-negative, and D-variant phenotypes.3 This study illustrates that for routine antenatal diagnosis of Rh D status, analysis of at least two regions of the RHD gene is essential: there are numerous instances where false-positive and false-negative results would be obtained when using a single RHD typing assay.

The D^VIcE polypeptide does not express C antigen, but it is interesting to note that the protein is almost identical to Rh polypeptide species purported to express Rh C antigen (Fig 5; ie, Rh Ce and CE proteins). This protein encodes Ser₁₀₃ , which is known to be crucial for C expression,24 but clearly not C antigen. The only exofacial differences between the D^VIcE protein and Rh Ce/CE proteins are those located in the sixth external domain. It is possible that the RHCE-specific residues of this loop are essential for correct expression of C antigen. Alternatively, the D^VIcE protein encodes Trp₁₆ , whereas the Rh Ce/CE proteins all encode Cys₁₆ , which has also been proposed to be crucial for C expression.24 The lack of Cys₁₆ expression on the D^VIcE protein may also prevent C expression. With the recent success of an in vitro system to study the expression of Rh antigens27 it is now possible to directly study the involvement of predicted critical amino acid residues in Rh antigen expression.

We thank Dr Geoff Daniels for useful discussion, Dr Anna Pirkola for D^VIccEe phenotype erythrocytes, and Dr Hans Sonneborn for monoclonal anti-D.