Introduction: With over 360 blood group antigens in systems recognized, there are antigens, such as RhD, which demonstrate a quantitative reduction in antigen expression due to nucleotide variants in the non-coding region of the gene that result in aberrant splicing or a regulatory mechanism. This study aimed to evaluate bioinformatically predicted GATA1-binding regulatory motifs in the RHD gene for samples presenting with weak or apparently negative RhD antigen expression but showing normal RHD exons. Methods: Publicly available open chromatin region data were overlayed with GATA1 motif candidates in RHD. Genomic DNA from weak D, Del or D– samples with normal RHD exons (n = 13) was used to confirm RHD zygosity by quantitative PCR. Then, RHD promoter, intron 1, and intron 2 regions were amplified for Sanger sequencing to detect potential disruptions in the GATA1 motif candidates. Electrophoretic mobility shift assay (EMSA) was performed to assess GATA1-binding. Luciferase assays were used to assess transcriptional activity. Results: Bioinformatic analysis identified five of six GATA1 motif candidates in the promoter, intron 1 and intron 2 for investigation in the samples. Luciferase assays showed an enhancement in transcription for GATA1 motifs in intron 1 and for intron 2 only when the R2 haplotype variant (rs675072G>A) was present. GATA1 motifs were intact in 12 of 13 samples. For one sample with a Del phenotype, a novel RHD c.1–110A>C variant disrupted the GATA1 motif in the promoter which was supported by a lack of a GATA1 supershift in the EMSA and 73% transcriptional activity in the luciferase assay. Two samples were D+/D– chimeras. Conclusion: The bioinformatic predictions enabled the identification of a novel DEL allele, RHD c.1–110A>C, which disrupted the GATA1 motif in the proximal promoter. Although the majority of the samples investigated here remain unexplained, we provide GATA1 targets which may benefit future RHD regulatory investigations.

The RhD antigen is clinically important and is included, alongside ABO, in routine immunohematology testing. Accurate typing of blood donors and patients as D+ or D– is required to minimize the risk of alloimmunization as anti-D is associated with mild to severe transfusion reactions [1]. Furthermore, D– pregnant women are administered prophylactic anti-D antenatally to prevent anti-D formation leading to hemolytic disease of the fetus and newborn (HDFN) where the fetus is D+. During routine testing, weak red blood cell (RBC) agglutination (2+ or less) observed in the indirect antiglobulin test can result from a weak D phenotype, where there is quantitatively reduced D antigen density on the RBC surface [2‒4]. RBCs with a Del phenotype carry very low numbers of D antigen sites, are routinely mistyped as D–, and are typically identified by adsorption-elution techniques for D antigen detection [5, 6]. Reference laboratories often perform RHD genotyping for such samples to identify the RHD allele and enable accurate D+ or D– classification.

RHD genotyping has revealed the most common weak D alleles in the Caucasian population are RHD*01W.1, RHD*01W.2, and RHD*01W.3 [3, 7, 8]. Individuals with these alleles have been reported to be safely managed as D+ to reduce unnecessary usage of D– units and prophylactic anti-D [3, 9‒11]. Otherwise, it is recommended to manage patients with other weak D alleles as D– [12]. Certain weak D alleles have been associated with different ethnic populations due to an increased prevalence in the population (e.g., RHD*09.03 (RHD*weak partial 4) in those of North African heritage [13] and RHD*15 (RHD*weak partial 15) in those of East Asian heritage [14]). Many of the weak D alleles are a result of one or more missense mutations that affect the intracellular or transmembrane region of the RhD protein [4]. For DEL alleles, the most common is the Asian type (RHD*DEL1) [15]. Patients with this allele have been reported to be safely transfused with D+ blood due to the lack of allo-anti-D production reported, although one case of anti-D production has been reported in an untransfused woman, following pregnancies [16‒18].

Currently, more than 200 weak D and DEL alleles are recognized by the International Society of Blood Transfusion (ISBT) Working Party for Red Cell Immunogenetics and Blood Group Terminology [19]. The molecular diversity giving rise to weak D and Del phenotypes has grown over time to include exon deletion [20, 21], exon duplication [22], and aberrant splicing due to nucleotide changes in intronic regions (e.g., RHD*1227+5G>C [23]) [24, 25]. A principally different background underlying weak or very weak blood group expression has been reported in several blood group systems, namely disruption of a regulatory motif in a promoter or an enhancer region.

The first example was reported in the FY system nearly 3 decades ago, where the serologically Duffy-negative phenotype on RBCs was shown to be due to a disruption in a GATA1-binding motif located in the ACKR1 (then DARC) gene promoter [26]. GATA1 is a transcription factor that binds to the DNA regulatory motif 5′-WGATAR-3′ and is key in maintaining terminal erythroid differentiation in erythropoiesis [27]. A nucleotide variant within the GATA1 motif can result in the loss of GATA1-binding as previously observed with the Fy(a–b–) phenotype arising from a c.–67T>C variant in the ACKR1 gene [26]. This was soon followed by the report of a GATA1-dependent enhancer-like motif in the ABO gene more than a decade ago to cause weakened B antigen expression, defining the Bm phenotype [28].

In contrast, it took until 2017 when a weak D sample was found to be due to a nucleotide change in the GATA1 motif located in the proximal RHD promoter [29]. Recently, Wu et al. [30] identified that a disruption in a GATA1 motif had led to an extremely reduced expression of the Knops blood group protein, Complement Receptor 1 (CR1), after confirmation with in vitro experiments. The bioinformatic pipeline used to predict regulatory GATA1 motifs in the CR1 gene had systematically identified 193 candidate regulators in 33 blood group genes. This included seven GATA1 candidate motifs in the RHD gene [30].

Given a report of a weak D sample resulting from a GATA1 motif disruption [29], it was hypothesized that there may be other disruptions in one of the proposed regulatory GATA1 motifs, resulting in reduced or negative D antigen expression. In this study, we aimed to investigate a cohort of weakened D and routinely typed D– samples with normal RHD exons (after routine reference laboratory analysis) for potential disruptions in GATA1 motif candidates.

Samples and Initial Investigation

EDTA-anticoagulated blood samples were referred to two reference laboratories for investigation of RhD/RHD status. Samples were selected for inclusion based on weak or initially undetectable D antigen despite normal RHD exon sequences according to the RHD genotyping method used in each laboratory: (1) PCR amplification and Sanger sequencing of all 10 RHD exons at the Nordic Reference Laboratory for Genomic Blood Group Typing (NRLGBT; n = 4) [31, 32]. Sanger chromatogram alignments were performed using CodonCode Aligner version 9.0.1 (CodonCode, Centerville, MA, USA) to detect RHD variants, (2) A custom, targeted exome sequencing panel that includes all 10 RHD exons at the Australian Red Cross Lifeblood (n = 9) [33, 34]. The samples tested include three weak D samples with normal RHD exons from a previous Australian blood donor study [7]. RHD variants were detected by analyzing variant call files (vcf) generated from the Binary Alignment Mapping (BAM) file using CLC Genomics Workbench version 20 (QIAGEN, Hilden, Germany) [33]. Copy number variation analysis was used to call RHD zygosity [33].

RHD Zygosity Testing, Flow Cytometry, and Extended D Phenotyping

Quantitative PCR (qPCR) analysis using RHD exon 7 was performed to confirm RHD zygosity as previously described using a target amount of 50 ng template genomic DNA (gDNA) [35]. qPCR was also used to identify D+/D– chimera samples which then underwent flow cytometry analysis. Flow cytometry was performed as described for ABO [36], except the monoclonal anti-D clone HM16 (DIAGAST, Loos, Nord, France) was used instead at a 1:5 final dilution and a polyclonal PE-labeled goat anti-human IgG secondary antibody (Jackson ImmunoResearch, West Grove, PA, USA) at a 1:20 final dilution. Flow histograms were generated using FCS Express version 6 (De Novo Software, Glendale, CA, USA). Relevant controls were included. Control gDNA homozygous, hemizygous, or negative for RHD were obtained for the qPCR from the NRLGBT or anonymized EDTA blood donor samples using the QIAamp DNA Blood Mini Kit (QIAGEN, Hilden, Germany). Extended D phenotyping was performed using the Bio-Rad ID-Partial RhD Typing Set (Bio-Rad Laboratories, Hercules, CA, USA) for sample 1 and ALBAclone® Advanced Partial RhD Typing Kit (Alba Bioscience, Edinburg, UK) for sample 2 as per manufacturer’s instructions.

Bioinformatics for Prediction of Regulatory Regions

Publicly available Assay for Transposase-Accessible Chromatin using sequencing (ATAC-seq) data indicating open chromatin regions in umbilical cord-derived CD34+ cells at different stages of erythroid development were retrieved (experiment No. GSE128266) [37]. The FASTQ files from experiments were analyzed with the Nextflow(nf)-core atacseq (version 1.2.2) pipeline [38] using human reference GRCh38.p14 (hg38). All analyses were performed on the high-performance computing platform, LUNARC, provided by Lund University together with the Swedish National Infrastructure for Computing (SNIC). The analyzed peaks from ATAC-seq data were downloaded and overlayed with the GATA1 motif predictions from Wu and colleagues [30] in Integrative Genomics Viewer (IGV) [39].

Amplification and Sequencing of RHD Regions with Predicted GATA1 Motifs

Primers to amplify RHD regions in the proximal promoter, intron 1 and intron 2, and sequence-predicted GATA1 motifs in these regions were carefully designed using a panel of resources: (1) alignment of RHD (NG_007494) and RHCE (NG_009208) reference sequences to identify mismatches and yield RHD-specific amplicons, (2) NCBI primer blast to generate or check gene-specific primers, (3) primer sequences from previous studies (online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000538469). PCR was performed with either the GoTaq Hot Start Master Mixes (Promega, Madison, WI, USA) or Phusion Hot Start II DNA polymerase (ThermoFisher Scientific, Waltham, MA, USA) according to manufacturer’s instructions on the T100 thermocycler (Bio-Rad Laboratories, Hercules, CA, USA) using the desired primer pair (Eurofins Genomics, Ebersberg, Germany) and PCR conditions (online suppl. Table 1). The PCR products were visualized on a 1–2% agarose gel and purified with QIAquick PCR Purification Kit (QIAGEN) for Sanger sequencing externally (Eurofins Genomics). Sanger sequencing chromatograms were aligned to the RHD reference sequence (NG_007494.1) to identify variants using SnapGene Software version 5.3.2 (GSL Biotech LLC, San Diego, CA, USA; available at snapgene.com). For one sample with a novel variant, FASTQ files were retrieved from the targeted exome sequencing previously performed, aligned to hg38 using Bowtie 2 on the Galaxy web platform to generate a BAM file and read alignment visualization on IGV [40, 41].

Electrophoretic Mobility Shift Assay

JASPAR analysis (10th release, 2024 version) was performed for the probe sequences using two GATA1 matrixes (MA0035.3 Gata1 for Mus musculus and MA0035.4 GATA1 for Homo sapiens) with a 70% threshold to predict the impact of the variants on the GATA1 motif [42]. Probes containing the intact or disrupted GATA1 motifs were synthesized with and without 5′ biotinylation (Eurofins Genomics; online suppl. Table 1). Complementary probes were annealed to form double-stranded DNA probes for the electrophoretic mobility shift assay (EMSA) reaction (online suppl. Table 1). Biotinylated probes (20 fmol) in Tris-EDTA (TE) buffer (Invitrogen, Waltham, MA, USA) were then added to the EMSA reaction containing reagents from the LightShift™ EMSA Optimization and Control Kit (ThermoFisher Scientific) [43]. Next, 4 μg of nuclear proteins extracted from human erythroleukemia cells (HEL), prepared as described, was added to the reaction and incubated for 20 min at room temperature [43]. For the supershift, 2 μL monoclonal mouse anti-human GATA1 (clone 4F5, Thermofisher Scientific) was added to the EMSA reaction for a further 20 min incubation at room temperature. Competition EMSA reactions and subsequent steps were performed as per Möller et al. [43] using the Chemiluminescent Nucleic Acid Detection Module Kit (ThermoFisher Scientific) and ChemiDoc Touch Image System (Bio-Rad Laboratories). Three independent experiments were performed.

Dual-Reporter Luciferase Assay Including Plasmid Construction

A total of nine DNA inserts with flanking restriction enzyme sites were prepared in forward and reverse orientations using RHD-specific PCR products as the starting template (online suppl. Table 1). For wild-type sequences, RHD gene-specific PCR products of each region were first amplified from samples with the R1R1 or R2R2 phenotype (R2R2 sample for intron 2 only) (online suppl. Table 1). Then, these PCR products were used as a template for a nested PCR using Phusion Hot Start II DNA polymerase (ThermoFisher Scientific) using a second set of primers flanked with the sequence for the restriction enzyme site required (online suppl. Table 1).

There were three 283-bp DNA inserts for the RHD promoter. This comprised of the wild-type and RHD c.1–110A>C allele PCR products, which were amplified from gDNA (online suppl. Table 1). Meanwhile, the RHD c.1–115A>C variant inserts were generated by site-directed mutagenesis using complementary 39-mer oligonucleotides as overlapping primers (online suppl. Table 1) [44]. These PCR products were then digested with BglII (ThermoFisher Scientific) and HindIII (Fermentas, Waltham, MA, USA) and ligated into the ampicillin-resistant pGL3-basic luciferase vector using T4 DNA ligase (ThermoFisher Scientific). Transformation was performed using One Shot™ TOP10 Chemically Competent Escherichia coli cells (ThermoFisher Scientific), and ampicillin-resistant cultures with the correct sequence were amplified for plasmid DNA purification using the QIAfilter Plasmid Midi Kit (QIAGEN) for the luciferase assay.

The remaining six inserts comprised two 552-bp RHD intron 1 PCR products and four 617-bp RHD intron 2 PCR products with or without the R2 haplotype variant, rs675072G>A [45]. These PCR products were digested with KpnI (New England Biolabs, Ipswich, MA, USA) and XhoI (ThermoFisher Scientific) restriction enzymes and then ligated with T4 DNA Ligase (ThermoFisher Scientific) into the wild-type RHD promoter-pGL3 construct. Ampicillin-resistant cultures with the correct sequence were then amplified and purified for the luciferase assay.

For the luciferase assay, 2 × 106 HEL cells were electroporated with 10 μg plasmid DNA and 1 μg control Renilla plasmid DNA as previously described [30]. Cells were then incubated in RPMI 1640 media supplemented with 10% fetal bovine serum (Gibco, ThermoFisher Scientific, Carlsbad, CA, USA) for 40–44 h in 37°C, 5% CO2. Cell lysates were then prepared and analyzed in triplicate using the Dual-Luciferase Reporter Assay System (Promega) and GloMax Discover Luminometer (Promega) according to manufacturer’s instructions. Three independent assays were performed. The luminescence ratio between the Firefly and Renilla control was calculated and then normalized to the RHD wild-type promoter construct.

We report here the investigation into 13 samples sent to two reference laboratories for resolution of apparent phenotype/genotype discrepancies involving RhD/RHD status. As a starting point for the study, the reported samples were selected on the basis that they displayed normal RHD sequences in combination with weak D, Del or D– (by routine phenotyping methods) phenotypes (Table 1, where samples are identified as numbers 1–13, and used in the following).

Table 1.

Sample background and findings from RHD sequencing of GATA1 motifs in the proximal promoter, intron 1 and intron 2 predicted by Wu et al. [30]

Sample No. used in studySample information obtained for the studySample typeRhCE phenotypeFinding
Weak D Pregnant C+E–c–e+ No GATA1 disruptions 
Weak D Not known C+E–c–e+ No GATA1 disruptions 
Weak D, D– by immediate spin. Weakly positive by indirect antiglobulin test Not known C+E–c+e+ D chimera 
Weak D, 2+ on RhD vision (LDM3/ESD1 blend) and 2+ on DVI gel card Pregnant C+E–c+e+ D chimera 
D variant query. Possible D on BeadChip Patient C+E–c–e+ No GATA1 disruptions 
(NGS prediction) 
Routinely phenotyped as weak D. Possible D on RHD BeadChip™ Patient C+E–c+e+ No GATA1 disruptions 
(NGS prediction) 
Routinely phenotyped as r”r, Possible D on RHD BeadChip™. Del phenotype found with polyclonal anti-D plasma Donor C–E+c+e+ Novel GATA1 mutation RHDc.1–110A>C 
Routinely phenotyped as r”r. Possible D on RHD BeadChip™ Donor C–E+c+e+ RHD*DEL5, RHD c.148+1G>A 
Routinely phenotyped as weak D. Possible D on RHD BeadChip™ Donor C+E–c+e+ No GATA1 disruptions 
10 Routinely phenotyped as r”r. Possible D on RHD BeadChip™ Donor C–E+c+e+ No GATA1 disruptions 
11 Weak D sample previously published McGowan et al. [7] 2017. Weak reaction with P3X21223B10, P3X241 and LHM57/17 Donor C+E–c–e+ No GATA1 disruptions 
12 Weak D sample, previously published McGowan et al. [7] 2017 Donor C+E–c+e+ No GATA1 disruptions 
13 Weak D sample, previously published McGowan et al. [7] 2017. Weak reaction with P3X21223B10, P3X241, and LHM57/17 Donor C+E–c–e+ No GATA1 disruptions 
Sample No. used in studySample information obtained for the studySample typeRhCE phenotypeFinding
Weak D Pregnant C+E–c–e+ No GATA1 disruptions 
Weak D Not known C+E–c–e+ No GATA1 disruptions 
Weak D, D– by immediate spin. Weakly positive by indirect antiglobulin test Not known C+E–c+e+ D chimera 
Weak D, 2+ on RhD vision (LDM3/ESD1 blend) and 2+ on DVI gel card Pregnant C+E–c+e+ D chimera 
D variant query. Possible D on BeadChip Patient C+E–c–e+ No GATA1 disruptions 
(NGS prediction) 
Routinely phenotyped as weak D. Possible D on RHD BeadChip™ Patient C+E–c+e+ No GATA1 disruptions 
(NGS prediction) 
Routinely phenotyped as r”r, Possible D on RHD BeadChip™. Del phenotype found with polyclonal anti-D plasma Donor C–E+c+e+ Novel GATA1 mutation RHDc.1–110A>C 
Routinely phenotyped as r”r. Possible D on RHD BeadChip™ Donor C–E+c+e+ RHD*DEL5, RHD c.148+1G>A 
Routinely phenotyped as weak D. Possible D on RHD BeadChip™ Donor C+E–c+e+ No GATA1 disruptions 
10 Routinely phenotyped as r”r. Possible D on RHD BeadChip™ Donor C–E+c+e+ No GATA1 disruptions 
11 Weak D sample previously published McGowan et al. [7] 2017. Weak reaction with P3X21223B10, P3X241 and LHM57/17 Donor C+E–c–e+ No GATA1 disruptions 
12 Weak D sample, previously published McGowan et al. [7] 2017 Donor C+E–c+e+ No GATA1 disruptions 
13 Weak D sample, previously published McGowan et al. [7] 2017. Weak reaction with P3X21223B10, P3X241, and LHM57/17 Donor C+E–c–e+ No GATA1 disruptions 

RHD Zygosity and Identification of D+/D– Chimera

RHD zygosity results were not available for all samples but were performed here for the whole cohort with the same method to confirm zygosity. Ratios of RHD exon 7 and albumin gene (ALB) mean quantity identified nine samples to be hemizygous as expected. Thus, the RHD zygosity calls previously obtained by targeted exome sequencing and copy number variation analysis were confirmed as RHD hemizygous for samples 5–13 (shown in Fig. 1a). Two samples gave values indicating homozygosity (samples 1 and 2) despite the very weak/partial D reactivity profile for sample 1 and somewhat weak D profile for sample 2 (Fig. 1b). Meanwhile, two samples (3 and 4) resulted in zygosity values that were neither compatible with hemi- nor homozygosity but too high for a D– sample. This suggested the presence of a D+/D– chimera (shown in Fig. 1a). Flow cytometry analysis for sample 4 showed two populations with 78.4% being D– RBCs and 20.7% being D+ RBCs (shown in Fig. 1c). RBCs from sample 3 were not available for flow cytometry analysis.

Fig. 1.

a Ratio of mean RHD exon 7/Albumin gene quantity obtained by a previously described quantitative PCR method [35]. Range for hemizygosity and homozygosity are indicated by the dashed light orange and dark orange lines, respectively. Controls with known RHD zygosity are shown as light gray columns and the samples in dark gray. Samples 1–2 were classified as homozygous, 3–4 as possible D+/D– chimeras and samples 5–13 as hemizygous. b Reactivity profiles for samples 1 and 2 using two types of monoclonal anti-D kits on a 0–4 scale. c Flow cytometry histogram analysis for the D+/D– chimera detected in sample 4. Marker gates for “D– RBCs” were set based on the D– RBC control (black) and the “D + RBCs” gate was based on the D+ control (blue). In sample 4, the D– population was present at 78.4% and D+ at 20.7% (red).

Fig. 1.

a Ratio of mean RHD exon 7/Albumin gene quantity obtained by a previously described quantitative PCR method [35]. Range for hemizygosity and homozygosity are indicated by the dashed light orange and dark orange lines, respectively. Controls with known RHD zygosity are shown as light gray columns and the samples in dark gray. Samples 1–2 were classified as homozygous, 3–4 as possible D+/D– chimeras and samples 5–13 as hemizygous. b Reactivity profiles for samples 1 and 2 using two types of monoclonal anti-D kits on a 0–4 scale. c Flow cytometry histogram analysis for the D+/D– chimera detected in sample 4. Marker gates for “D– RBCs” were set based on the D– RBC control (black) and the “D + RBCs” gate was based on the D+ control (blue). In sample 4, the D– population was present at 78.4% and D+ at 20.7% (red).

Close modal

Bioinformatically Predicted GATA1 Motifs in the RHD Gene

Five of the six predicted GATA1 motifs were found to overlay with open chromatin regions in the RHD promoter, intron 1 or intron 2 regions (shown in Fig. 2a). One GATA1 motif prediction (peak 2810), despite their high FIMO and JASPAR motif scores, did not have a high peak score and did not overlay with the open chromatin regions (shown in Fig. 2a). It was therefore considered a less likely target for transcription factor binding. This resulted in the investigation of five predicted GATA1 motifs (one motif for peak 2806 in the promoter, two motifs for peak 2808 in intron 1 and two motifs for peak 2811 in intron 2).

Fig. 2.

a Diagrammatic representation of hg38 regions overlayed with predicted FIMO and JASPAR GATA1 motifs (light and dark purple boxes) [30] and open chromatin regions obtained from public databases (blue boxes with increasing intensity based on peak score). The peak and motif scores for the open chromatin regions and GATA1 motifs are provided next to the boxes. GATA1 motifs within open chromatin regions were selected for further investigation in weak D samples and luciferase assays (red box). The rs675072G>A nucleotide change previously associated with the R2 haplotype was also selected for further analysis as it was within the open chromatin region and in proximity to GATA peak 2811. RHD regions used for the luciferase assay are shown in red boxes. b Relative luciferase activity for predicted GATA1 motifs in open chromatin regions. The promoter showed activating transcriptional activity. An enhancement in transcriptional activity was observed when open chromatin sequences from intron 1 (pale pink) or intron 2 with rs675072A (dark pink) were present in the forward (F) or reverse (R) orientation. The transcriptional activity for the RHD intron 2 region with rs675072G had minor differences compared to the RHD promoter.

Fig. 2.

a Diagrammatic representation of hg38 regions overlayed with predicted FIMO and JASPAR GATA1 motifs (light and dark purple boxes) [30] and open chromatin regions obtained from public databases (blue boxes with increasing intensity based on peak score). The peak and motif scores for the open chromatin regions and GATA1 motifs are provided next to the boxes. GATA1 motifs within open chromatin regions were selected for further investigation in weak D samples and luciferase assays (red box). The rs675072G>A nucleotide change previously associated with the R2 haplotype was also selected for further analysis as it was within the open chromatin region and in proximity to GATA peak 2811. RHD regions used for the luciferase assay are shown in red boxes. b Relative luciferase activity for predicted GATA1 motifs in open chromatin regions. The promoter showed activating transcriptional activity. An enhancement in transcriptional activity was observed when open chromatin sequences from intron 1 (pale pink) or intron 2 with rs675072A (dark pink) were present in the forward (F) or reverse (R) orientation. The transcriptional activity for the RHD intron 2 region with rs675072G had minor differences compared to the RHD promoter.

Close modal

Transcriptional Activity in Open Chromatin Regions

Luciferase assays showed the RHD promoter to activate transcription when compared to the vector only control, as expected (shown in Fig. 2b; 100%). Upon addition of the open chromatin candidate regions in intron 1 or intron 2 to the RHD promoter vector, an enhancement in transcriptional activity was observed in candidate regions from RHD intron 1 (forward orientation 180%, reverse 246%) and RHD intron 2 with rs675072A (forward 242%, reverse 146%). In contrast, introduction of rs675072G in the open chromatin region of RHD intron 2 resulted only in minor transcriptional differences compared to the RHD promoter alone (forward 118%, reverse 77%).

Analysis of Predicted GATA1 Regions

Twelve of the 13 weak D samples tested did not have nucleotide variants in or around 40 to 120-bp of the predicted GATA1 motifs. One of the samples had a novel A>C nucleotide change at RHD c.1–110A>C (chr1:25,272,438) which disrupted the 3′-end of the GATA1 motif located in the proximal promoter (sample 7) (shown in Fig. 3a; GenBank accession No. OR296440). A previously reported weak D sample carrying RHD c.1–115A>C allele disrupted the same GATA1 motif, but at the 5′-end [29]. Reanalysis of previous targeted exome sequencing data showed there were nine out of twelve reads with the A>C variant at RHD c.1–110 (Fig. 3b). Of the remaining three reads, two of the reads with the C>A nucleotide change at chr1:25,272,426 were suspected to be misaligned RHCE reads, while one read did not include this C>A site because it started 4 bp downstream of the chr1:25,272,426 coordinate (Fig. 3b). This variant reflected the G>T nucleotide change at the corresponding RHCE position in hg38 at the reverse orientation (chr1:25,420,908) (Fig. 3b).

Fig. 3.

a Sanger chromatogram alignment to the RHD reference sequence for the RHDc.1–110A>C allele detected in sample 7. The nucleotide A>C change at RHD c.1–110 is indicated with the red box. b Read alignment of targeted exome sequencing data from sample 7. The A>C variant was present in nine out of 12 reads (red box). Two reads with a C>A nucleotide variant may be misaligned reads from RHCE (top blue box). This variant was reflected in the corresponding RHCE nucleotide position (chr1:25,420,908) where 18 out of 18 reads aligned showed there was a G>T nucleotide variant (bottom blue box; note that RHCE is in the reverse orientation in hg38).

Fig. 3.

a Sanger chromatogram alignment to the RHD reference sequence for the RHDc.1–110A>C allele detected in sample 7. The nucleotide A>C change at RHD c.1–110 is indicated with the red box. b Read alignment of targeted exome sequencing data from sample 7. The A>C variant was present in nine out of 12 reads (red box). Two reads with a C>A nucleotide variant may be misaligned reads from RHCE (top blue box). This variant was reflected in the corresponding RHCE nucleotide position (chr1:25,420,908) where 18 out of 18 reads aligned showed there was a G>T nucleotide variant (bottom blue box; note that RHCE is in the reverse orientation in hg38).

Close modal

Although not located in a predicted GATA1 motif, a RHD c.148+1G>A (RHD*DEL5) nucleotide variant was detected upon reanalysis of sample 8. The allele was indeed detected by targeted exome sequencing at Lifeblood after a modification to include splice sites in the data analysis was implemented.

GATA1-Binding and Transcriptional Activity in Disrupted Motifs

GATA1-binding and transcription activity for the novel RHD c.1–110A>C allele was then compared to the RHD wild-type and the previously reported neighbor variant, RHD c.1–115A>C allele, which disrupted the same GATA1 motif at the 5′-end. The relative scores from the JASPAR analysis using MA0035.3 Gata1 and MA0035.4 GATA1, respectively, were: 0.937 and 0.905 for the RHD wild-type motifs, 0.815 and 0.772 for RHD c.1–115A>C and 0.813 and 0.764 for RHD c.1–110A>C. Overall, the relative scores for the mutants were lower than the wild-type RHD probes and lower than or close to the default threshold (0.80). EMSA analysis showed there was GATA1-binding with the intact GATA1 motif in the wild-type RHD sequence in all three independent experiments (shown in Fig. 4). In contrast, the RHD c.1–115A>C variant probes disrupting the GATA1 motif at the 5′-end showed HEL nuclear extract binding, but a weak anti-GATA1 supershift was visualized in only one of the three experiments. This suggested GATA1 as a borderline binder in the presence of the RHD c.1–115A>C variant. The RHD c.1–110A>C probes also showed HEL nuclear extract binding but no GATA1 supershifts were observed in all three independent experiments. Luciferase assays showed transcriptional activity for the RHD promoter (100%) and for RHD c.1–115A>C construct (96%) (shown in Fig. 5). The activity for the RHD c.1–110A>C construct was lower at 73%, indicating that this could be compatible with weakened D expression.

Fig. 4.

Comparison of GATA1-binding between the intact GATA1 motif in the wild-type RHD sequence and the variant RHD alleles. a Probe sequences used for analysis of GATA1-binding to the RHD proximal promoter sequence with a known ABO GATA1 sequence used as positive control. b Blot from electrophoretic mobility shift assay (EMSA) showing a shift for reactions with the nuclear extract only (lanes 2, 6, and 10; orange arrow) and a supershift in lanes with anti-GATA1 antibodies added to the reaction (lanes 3, 7, and 11; dark orange arrow). For the RHD c.1–110A>C probes, there was a band to indicate nuclear extract binding (lane 14; dark orange arrow) but a supershift from the addition of anti-GATA1 antibodies was not observed (lane 15). c Two replicate EMSA blots showing nuclear extract binding to the wild-type RHD probes, RHD c.1–115A>C, and RHD c.1–110A>C probes (lanes 2, 6, and 10; orange arrow). The supershift from the anti-GATA1 antibody was only observed the intact GATA1 motif found in the RHD wild-type probes (lane 3; dark orange arrow).

Fig. 4.

Comparison of GATA1-binding between the intact GATA1 motif in the wild-type RHD sequence and the variant RHD alleles. a Probe sequences used for analysis of GATA1-binding to the RHD proximal promoter sequence with a known ABO GATA1 sequence used as positive control. b Blot from electrophoretic mobility shift assay (EMSA) showing a shift for reactions with the nuclear extract only (lanes 2, 6, and 10; orange arrow) and a supershift in lanes with anti-GATA1 antibodies added to the reaction (lanes 3, 7, and 11; dark orange arrow). For the RHD c.1–110A>C probes, there was a band to indicate nuclear extract binding (lane 14; dark orange arrow) but a supershift from the addition of anti-GATA1 antibodies was not observed (lane 15). c Two replicate EMSA blots showing nuclear extract binding to the wild-type RHD probes, RHD c.1–115A>C, and RHD c.1–110A>C probes (lanes 2, 6, and 10; orange arrow). The supershift from the anti-GATA1 antibody was only observed the intact GATA1 motif found in the RHD wild-type probes (lane 3; dark orange arrow).

Close modal
Fig. 5.

Comparison of the relative luciferase activity in wild-type RHD promoter to the two RHD alleles disrupting the GATA1 motif, RHD c.1–115A>C, and RHD c.1–110A>C constructs. A decrease in transcriptional activity was evident in the RHD c.1–110A>C allele with a Del phenotype compared to the previously reported RHD c.1–115A>C with a weak D phenotype and the wild-type sequence.

Fig. 5.

Comparison of the relative luciferase activity in wild-type RHD promoter to the two RHD alleles disrupting the GATA1 motif, RHD c.1–115A>C, and RHD c.1–110A>C constructs. A decrease in transcriptional activity was evident in the RHD c.1–110A>C allele with a Del phenotype compared to the previously reported RHD c.1–115A>C with a weak D phenotype and the wild-type sequence.

Close modal

This study has applied bioinformatically predicted GATA1 regions to samples with weak D, Del and routinely typed D– phenotypes and normal RHD exons. The bioinformatic prediction of the GATA1 motif in the proximal promoter was consolidated with previous literature and further supported by a case report of a weak D sample arising from a RHD c.1–115A>C variant disrupting the GATA1 motif at the 5′-end [29, 46]. The remaining GATA1 bioinformatic predictions were located in introns 1 and 2. When an increase in transcriptional activity was observed in vitro, in comparison to the RHD promoter sequence, it suggested that these GATA1 motifs would function as enhancers when intact.

Although most of the samples were hemizygous, there were two homozygous RHD samples with one showing a weak and partial D reactivity profile and the other showing a weak D profile. This raises the possibility that these two samples may be homozygous or compound heterozygous for an RHD variant allele in a non-coding region which has yet to be investigated. Additionally, the qPCR showed two other samples had an RHD dosage out of range for the three expected outcomes (homozygous, hemizygous, or negative), consistent with a D+/D– chimera. Zygosity testing by qPCR has been previously discussed to be capable in detecting D+/D– chimera [47, 48]. Phenotypically, D+/D– chimeras can be identified using serologically mixed-field reactions but may be mistaken for weak D samples. However, flow cytometry can show the percentage of RBCs which are D+ or D– in the same cell population. This example of a D+/D– chimera, initially referred as a weak D sample in a pregnant patient had D+ RBCs present at 21%, whereas the D+/D– chimeras previously detected in a cohort of D– donors had approximately 6% of D+ RBCs present in the RBC population [25]. Our study highlights two points in this regard: (1) One should suspect a possible chimera if zygosity tests show unexpected values, (2) flow cytometry easily differentiates a D+/D– chimera with few D+ RBCs from a sample with weak D expression.

Applying the bioinformatic GATA1 motif predictions provided a targeted approach to navigate non-coding regions for regulatory motifs [30]. The RHD*DEL5 allele in one sample highlights the importance of including and checking non-coding regions when investigating the RHD gene in weak D, Del, and D– samples, as this allele was not identified until a modification to include splice sites in the data analysis was made. In another donor sample, a novel RHDc.1–110A>C allele that disrupted the 3′-end of the GATA1 motif at the proximal promoter was revealed. The low coverage at this nucleotide position may have caused the quality parameters to exclude it from the variant call file of targeted exome sequencing. In addition, the high homology between RHD and RHCE added to the complexity of the analysis with RHCE reads misaligning to RHD. This emphasizes the importance of manually analyzing the read alignment for weakened D or D– samples with normal RHD exons, at least until more sophisticated automated algorithms are in place.

A combination of JASPAR scores and previous literature strongly suggested that the GATA1 disruption from RHD c.1–110A>C would lead to a reduction in RHD gene transcription. For the RHD c.1–115A>C allele, borderline GATA1-binding observed in the EMSA suggested that there may be scenarios where GATA1 binds or does not bind. For the RHD c.1–115A>C allele studies, it may be possible that the variation in luciferase activity between our study and Fennell et al. [29] further demonstrate the lack of GATA1-binding stability at the motif. Meanwhile, the combination of disrupted GATA1-binding and reduced transcription levels of the new RHD c.1–110A>C variant allele appeared to correspond with a Del phenotype. In correspondence with the JASPAR relative score predictions, these findings showed that the GATA1 motif disruption at the 3′-end had greater impact on RhD antigen expression compared to the disruption at the 5′-end of the GATA1 motif.

One of the limitations of this study was that it was performed retrospectively with finite amounts of gDNA. Additional gDNA could have helped identify alternative reasons underlying the weak D expression. There are many theoretical explanations possible, including the fact that we only looked for GATA1 motifs and not binding sites for other erythroid transcription factors. For the two samples that appeared to be homozygous for RHD, but also for the others, it could have been of interest to test whether an RHAG variant within the coding or non-coding region (or possibly variants of other protein complex partners) is involved, as an RHAG variant has previously been shown to have a weak D phenotype and normal RhCE protein expression [49]. A collection of fresh blood samples to perform quantitative RHD mRNA analysis would provide an important lead to show whether the molecular basis was regulatory related. The availability of RNA from a fresh blood sample also provides an opportunity to sequence RHD cDNA, which has previously led to the identification of a variant transcript sequence for a sample with a very weak D phenotype [50]. The 80-bp insert identified in the RHD cDNA transcript sequence was tracked back to an RHD c.149–2632T>A nucleotide change deep in intron 1 of the RHD gene [50]. Normal RHD transcript sequences at reduced levels in a sample with a weak or very weak D phenotype would strongly suggest a regulatory related molecular mechanism.

Another approach to resolving the weakened D or D– samples with intact GATA1 motifs would be to sequence the RHD gene using long-range gene-specific amplicons and/or long-read sequencing technologies [51]. This approach would capture the whole RHD gene, allowing a non-biased approach to variant detection. More importantly, the whole RHD gene sequence would provide the intronic sequences required to identify nucleotide variants with a potential to disrupt regulatory motifs or cause dysfunctional transcripts as mentioned above. The incorporation of previously established RHD intronic SNVs linked with R2 and R1, R0, or RZ haplotypes [45] enabled us to investigate the transcriptional activity of open chromatin and GATA1 motif candidate regions in RHD depending on the Rh haplotype. For example, the inclusion of rs675072G>A variant associated with the R2 haplotype in the luciferase assay constructs for the GATA1 motif in RHD intron 2 showed an approximate 2-fold increase in transcription activity for rs675072G>A construct compared to the wild-type rs675072G construct. These haplotype differences are reminiscent of the suppressive effect on RhD antigen expression from the RHCE*C allele as the transcriptional activity was similar to the promoter when the R2 haplotype variant was absent (rs675072G) [52, 53].

The remaining weakened D and D– samples with no GATA1 motif disruptions at the promoter, intron 1 or intron 2 showed that detecting variants in non-coding regions is difficult and, if regulatory related, could likely extend beyond erythroid transcription factors other than GATA1 [54, 55] or involve remote regulatory sites such as that for GATA1 with the Xg(a–) phenotype [43, 56]. The full epigenetic landscape, including histone modification and 3D genome interactions, adds to this complexity [57, 58]. Further development of this pipeline to broaden the range of known blood group gene transcription factors and inclusion of histone modification markers may benefit the other weakened D or D– samples with normal RHD exons from the study.

In conclusion, this study has evaluated GATA1 candidate regulatory regions predicted by a bioinformatics approach in samples with a weak D, Del and D– phenotype despite normal RHD exons. It detected and functionally validated a novel RHD c.1–110A>C allele leading to a Del phenotype from the disruption of a GATA1 motif at the proximal promoter. Haplotype SNVs in the introns of the RHD gene are important to consider when investigating disruptions in regulatory motifs as they can impact transcriptional activity. The candidate GATA1 motifs from this study may be useful in future investigations of weak D samples with normal RHD exons. However, despite resolving four of 13 samples in this small cohort of unexplained weakened or D– phenotypes, the majority remains unresolved, further highlighting the complexity and diversity of molecular mechanisms or potential protein interactions that have yet to be discovered for the RhD antigen.

The authors would like to thank Dr. Annika Hult for performing the flow cytometry and Prof. Jill Storry for performing the ALBAclone phenotyping. We are also grateful to Dr. Yew-Wah Liew and Ms. Glenda Millard from Lifeblood’s Red Cell Reference Laboratory for retrieving donor and patient information required for this study. We acknowledge that Australian Governments fund Australian Red Cross Lifeblood for the provision of blood, blood products, and services to the Australian community.

We report here the anonymized results of reference laboratory work-ups of patient and donor samples displaying weakened D expression on their RBCs but no deviations in their RHD exons compared to the consensus RHD sequence. As control and validation material, anonymized, EDTA-anticoagulated blood samples collected as part of the routine blood donation procedure were obtained from random healthy blood donors at the Department of Clinical Immunology and Transfusion Medicine, Office for Medical Services, Region Skåne, Sweden. According to the Swedish research law, using fully anonymized and/or pooled biological waste material obtained for other purposes does not require ethics approval. Thus, no approval was required for this study, apart from the permission from the Department of Clinical Immunology and Transfusion Medicine to obtain the anonymized blood from waste material (ref. No. 2020:16). Samples from the Australian cohort were provided with the approval of the Australian Red Cross Lifeblood Ethics Committee, under the Project with reference number 2017#25 and with an amendment approved by Lifeblood Ethics on August 12, 2022.

M.L.O. declares no actual conflict of interest in relation to this study but in accordance with the journal’s conflict of interest policy, the following is reported: M.L.O. is an inventor of patents on Vel blood group genotyping and own 50% of the shares in BLUsang AB, an incorporated consulting firm that receives royalties for said patents. All other authors state no conflict of interest.

This study was supported by the Knut and Alice Wallenberg Foundation (2020.0234 to MLO), the Swedish Research Council (2019-01683 to MLO), governmental ALF grants to the university healthcare in Region Skåne, Sweden (ALFSKANE-446521 to MLO), and the Royal Physiographic Society of Lund, Sweden (2023-152651 to PCW).

E.C.M. wrote the manuscript, designed primers, performed PCR assays, EMSA and luciferase assay, and data analysis, including the sequencing data from PCRs performed by P.C.W. P.C.W. helped write the manuscript, performed the analysis that provided the GATA1 candidates, designed primers, performed PCR assays, and constructed vectors for the luciferase assay. Å.H. oversaw the NRLGBT samples and analysis of the D+/D– chimeras. G.H.L. and C.A.H. provided and managed the data from the Lifeblood samples. E.C.M., P.C.W., and M.L.O. interpreted and reviewed the data. M.L.O. supervised the study. All authors contributed to the preparation and review of this manuscript and approved the manuscript for publication.

All data generated or analyzed during this study are included in this article and its supplementary material files. Further inquiries can be directed to the corresponding author.

1.
Reid
ME
,
Lomas-Francis
C
,
Olsson
ML
.
The blood group antigen factsbook
. 3rd ed.
Academic Press
;
2012
.
2.
Daniels
G
.
Human blood groups
. 3rd ed.
John Wiley & Sons
;
2013
.
3.
Wagner
FF
,
Frohmajer
A
,
Ladewig
B
,
Eicher
NI
,
Lonicer
CB
,
Müller
TH
, et al
.
Weak D alleles express distinct phenotypes
.
Blood
.
2000
;
95
(
8
):
2699
708
.
4.
Wagner
FF
,
Gassner
C
,
Müller
TH
,
Schönitzer
D
,
Schunter
F
,
Flegel
WA
.
Molecular basis of weak D phenotypes
.
Blood
.
1999
;
93
(
1
):
385
93
.
5.
Körmöczi
GF
,
Gassner
C
,
Shao
CP
,
Uchikawa
M
,
Legler
TJ
.
A comprehensive analysis of DEL types: partial DEL individuals are prone to anti-D alloimmunization
.
Transfusion
.
2005
;
45
(
10
):
1561
7
.
6.
Nuchnoi
P
,
Thongbus
J
,
Srisarin
A
,
Kerdpin
U
,
Prachayasittikul
V
.
Clinical and laboratory update on the DEL variant
.
Lab Med
.
2014
;
45
(
4
):
285
90
.
7.
McGowan
EC
,
Lopez
GH
,
Knauth
CM
,
Liew
YW
,
Condon
JA
,
Ramadi
L
, et al
.
Diverse and novel RHD variants in Australian blood donors with a weak D phenotype: implication for transfusion management
.
Vox Sang
.
2017
;
112
(
3
):
279
87
.
8.
Denomme
GA
,
Wagner
FF
,
Fernandes
BJ
,
Li
W
,
Flegel
WA
.
Partial D, weak D types, and novel RHD alleles among 33,864 multiethnic patients: implications for anti-D alloimmunization and prevention
.
Transfusion
.
2005
;
45
(
10
):
1554
60
.
9.
Flegel
WA
.
How I manage donors and patients with a weak D phenotype
.
Curr Opin Hematol
.
2006
;
13
(
6
):
476
83
.
10.
Daniels
G
.
Variants of RhD – current testing and clinical consequences
.
Br J Haematol
.
2013
;
161
(
4
):
461
70
.
11.
Flegel
WA
,
Denomme
GA
,
Queenan
JT
,
Johnson
ST
,
Keller
MA
,
Westhoff
CM
, et al
.
It’s time to phase out “serologic weak D phenotype” and resolve D types with RHD genotyping including weak D type 4
.
Transfusion
.
2020
;
60
(
4
):
855
9
.
12.
Sandler
SG
,
Flegel
WA
,
Westhoff
CM
,
Denomme
GA
,
Delaney
M
,
Keller
MA
, et al
.
It’s time to phase in RHD genotyping for patients with a serologic weak D phenotype. College of American Pathologists Transfusion Medicine Resource Committee Work Group
.
Transfusion
.
2015
;
55
(
3
):
680
9
.
13.
Ouchari
M
,
Srivastava
K
,
Romdhane
H
,
Jemni Yacoub
S
,
Flegel
WA
.
Transfusion strategy for weak D Type 4.0 based on RHD alleles and RH haplotypes in Tunisia
.
Transfusion
.
2018
;
58
(
2
):
306
12
.
14.
Flegel
WA
.
Proceed with care: the “uncommon” serologic weak D phenotypes
.
Blood Transfus
.
2021
;
19
(
4
):
272
6
.
15.
Yin
Q
,
Flegel
WA
.
DEL in China: the D antigen among serologic RhD-negative individuals
.
J Transl Med
.
2021
;
19
(
1
):
439
.
16.
Ji
Y
,
Luo
Y
,
Wen
J
,
Sun
Y
,
Jia
S
,
Ou
C
, et al
.
Patients with Asian-type DEL can safely be transfused with RhD-positive blood
.
Blood
.
2023
;
141
(
17
):
2141
50
.
17.
Ohto
H
,
Ito
S
,
Srivastava
K
,
Ogiyama
Y
,
Uchikawa
M
,
Nollet
KE
, et al
.
Asian-type DEL (RHD*DEL1) with an allo-anti-D: a paradoxical observation in a healthy multiparous woman
.
Transfusion
.
2023
;
63
(
8
):
1601
11
.
18.
Shao
CP
.
Transfusion of RhD-positive blood in “Asia type” DEL recipients
.
N Engl J Med
.
2010
;
362
(
5
):
472
3
.
19.
International Society of Blood Transfusion Red Cell Immunogenetics and Blood Group Terminology
.
(ISBT 004) RHD blood group alleles v6.4 31-JUL-2023
[cited 26/10/2023]. Available from: https://www.isbtweb.org/resource/004rhd.html.
20.
Fichou
Y
,
Chen
J-M
,
Le Maréchal
C
,
Jamet
D
,
Dupont
I
,
Chuteau
C
, et al
.
Weak D caused by a founder deletion in the RHD gene
.
Transfusion
.
2012
;
52
(
11
):
2348
55
.
21.
Lopez
GH
,
Turner
RM
,
McGowan
EC
,
Schoeman
EM
,
Scott
SA
,
O’Brien
H
, et al
.
A DEL phenotype attributed to RHD Exon 9 sequence deletion: slipped-strand mispairing and blood group polymorphisms
.
Transfusion
.
2018
;
58
(
3
):
685
91
.
22.
Fichou
Y
,
Parchure
D
,
Gogri
H
,
Gopalkrishnan
V
,
Le Maréchal
C
,
Chen
JM
, et al
.
Molecular basis of weak D expression in the Indian population and report of a novel, predominant variant RHD allele
.
Transfusion
.
2018
;
58
(
6
):
1540
9
.
23.
Wen
J
,
Jia
S
,
Wang
Z
,
Chen
J
,
Liang
Q
,
Wei
L
, et al
.
Molecular and serological analysis of the D variant in the Chinese population and identification of seven novel RHD alleles
.
Transfusion
.
2023
;
63
(
2
):
402
14
.
24.
Scott
SA
,
Nagl
L
,
Tilley
L
,
Liew
YW
,
Condon
J
,
Flower
R
, et al
.
The RHD(1227G>A) DEL-associated allele is the most prevalent DEL allele in Australian D– blood donors with C+ and/or E+ phenotypes
.
Transfusion
.
2014
;
54
(
11
):
2931
40
.
25.
Wagner
FF
,
Frohmajer
A
,
Flegel
WA
.
RHD positive haplotypes in D negative Europeans
.
BMC Genet
.
2001
;
2
(
1
):
10
.
26.
Tournamille
C
,
Colin
Y
,
Cartron
JP
,
Le Van Kim
C
.
Disruption of a GATA motif in the Duffy gene promoter abolishes erythroid gene expression in Duffy-negative individuals
.
Nat Genet
.
1995
;
10
(
2
):
224
8
.
27.
Gutiérrez
L
,
Caballero
N
,
Fernández-Calleja
L
,
Karkoulia
E
,
Strouboulis
J
.
Regulation of GATA1 levels in erythropoiesis
.
IUBMB Life
.
2020
;
72
(
1
):
89
105
.
28.
Sano
R
,
Nakajima
T
,
Takahashi
K
,
Kubo
R
,
Kominato
Y
,
Tsukada
J
, et al
.
Expression of ABO blood-group genes is dependent upon an erythroid cell–specific regulatory element that is deleted in persons with the Bm phenotype
.
Blood J Am Soc Hematol
.
2012
;
119
(
22
):
5301
10
.
29.
Fennell
K
,
Hoffman
R
,
Yoshida
K
,
Iwamoto
S
,
Govender
L
,
Vather
K
, et al
.
Effect on gene expression of three allelic variants in GATA motifs of ABO, RHD, and RHCE regulatory elements
.
Transfusion
.
2017
;
57
(
11
):
2804
8
.
30.
Wu
PC
,
Lee
YQ
,
Möller
M
,
Storry
JR
,
Olsson
ML
.
Elucidation of the low-expressing erythroid CR1 phenotype by bioinformatic mining of the GATA1-driven blood-group regulome
.
Nat Commun
.
2023
;
14
(
1
):
5001
.
31.
Gassner
C
,
Schmarda
A
,
Kilga-Nogler
S
,
Jenny-Feldkircher
B
,
Rainer
E
,
Müller
TH
, et al
.
RHD/CE typing by polymerase chain reaction using sequence-specific primers
.
Transfusion
.
1997
;
37
(
10
):
1020
6
.
32.
Pardi
C
,
Hellberg
Å
,
Kapadzha
M
,
Olsson
ML
,
Isakson
P
.
Novel RHD variant causing RhD negative phenotype identified in a pregnant woman
.
Transfusion
.
2022
;
62
(
8
):
E37
9
.
33.
Schoeman
EM
,
Lopez
GH
,
McGowan
EC
,
Millard
GM
,
O’Brien
H
,
Roulis
EV
, et al
.
Evaluation of targeted exome sequencing for 28 protein-based blood group systems, including the homologous gene systems, for blood group genotyping
.
Transfusion
.
2017
;
57
(
4
):
1078
88
.
34.
Roulis
E
,
Schoeman
E
,
Hobbs
M
,
Jones
G
,
Burton
M
,
Pahn
G
, et al
.
Targeted exome sequencing designed for blood group, platelet, and neutrophil antigen investigations: proof-of-principle study for a customized single-test system
.
Transfusion
.
2020
;
60
(
9
):
2108
20
.
35.
Grootkerk-Tax
MG
,
Maaskant-van Wijk
PA
,
van Drunen
J
,
van der Schoot
CE
.
The highly variable RH locus in nonwhite persons hampers RHD zygosity determination but yields more insight into RH-related evolutionary events
.
Transfusion
.
2005
;
45
(
3
):
327
37
.
36.
Hult
AK
,
Olsson
ML
.
Many genetically defined ABO subgroups exhibit characteristic flow cytometric patterns
.
Transfusion
.
2010
;
50
(
2
):
308
23
.
37.
Schulz
VP
,
Yan
H
,
Lezon-Geyda
K
,
An
X
,
Hale
J
,
Hillyer
CD
, et al
.
A unique epigenomic landscape defines human erythropoiesis
.
Cell Rep
.
2019
;
28
(
11
):
2996
3009.e7
.
38.
Patel
H
,
Ewels
P
,
Espinosa-Carrasco
J
,
Peltzer
A
,
Behrens
D
,
Gabernet
G
, et al
.
Nf-core/atacseq: NfCore/atacseq v1.2.2 - iron ossifrage
[cited 03/11/2023]. Available from: https://zenodo.org/records/6544493.
39.
Robinson
JT
,
Thorvaldsdóttir
H
,
Winckler
W
,
Guttman
M
,
Lander
ES
,
Getz
G
, et al
.
Integrative genomics viewer
.
Nat Biotechnol
.
2011
;
29
(
1
):
24
6
.
40.
Galaxy Community
.
The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2022 update
.
Nucleic Acids Res
.
2022
;
50
(
W1
):
W345
51
.
41.
Langmead
B
,
Salzberg
SL
.
Fast gapped-read alignment with Bowtie 2
.
Nat Methods
.
2012
;
9
(
4
):
357
9
.
42.
Rauluseviciute
I
,
Riudavets-Puig
R
,
Blanc-Mathieu
R
,
Castro-Mondragon Jaime
A
,
Ferenc
K
,
Kumar
V
, et al
.
JASPAR 2024: 20th anniversary of the open-access database of transcription factor binding profiles
.
Nucleic Acids Res
.
2023
;
52
(
D1
):
D174
82
.
43.
Möller
M
,
Lee
YQ
,
Vidovic
K
,
Kjellström
S
,
Björkman
L
,
Storry
JR
, et al
.
Disruption of a GATA1-binding motif upstream of XG/PBDX abolishes Xg(a) expression and resolves the Xg blood group system
.
Blood
.
2018
;
132
(
3
):
334
8
.
44.
Heckman
KL
,
Pease
LR
.
Gene splicing and mutagenesis by PCR-driven overlap extension
.
Nat Protoc
.
2007
;
2
(
4
):
924
32
.
45.
Tounsi
WA
,
Madgett
TE
,
Avent
ND
.
Complete RHD next-generation sequencing: establishment of reference RHD alleles
.
Blood Adv
.
2018
;
2
(
20
):
2713
23
.
46.
Chérif-Zahar
B
,
Le Van Kim
C
,
Rouillac
C
,
Raynal
V
,
Cartron
JP
,
Colin
Y
.
Organization of the gene (RHCE) encoding the human blood group RhCcEe antigens and characterization of the promoter region
.
Genomics
.
1994
;
19
(
1
):
68
74
.
47.
Engelfriet
C
,
Reesink
H
,
Fontao-Wendel
R
,
Kormoczi
G
,
Mayr
W
,
Panzer
S
, et al
.
Testing for weak D
.
Vox Sanguinis
.
2006
;
90
(
2
):
140
53
.
48.
Krog
GR
,
Clausen
FB
,
Dziegiel
MH
.
Quantitation of RHD by real-time polymerase chain reaction for determination of RHD zygosity and RHD mosaicism/chimerism: an evaluation of four quantitative methods
.
Transfusion
.
2007
;
47
(
4
):
715
22
.
49.
Wen
J
,
Verhagen
OJHM
,
Jia
S
,
Liang
Q
,
Wang
Z
,
Wei
L
, et al
.
A variant RhAG protein encoded by the RHAG*572A allele causes serological weak D expression while maintaining normal RhCE phenotypes
.
Transfusion
.
2019
;
59
(
1
):
405
11
.
50.
Floch
A
,
Vege
S
,
Berardi
P
,
Hannon
J
,
Ochoa-Garay
G
,
Lomas-Francis
C
, et al
.
An intron c.149-2632T>A change in RHD is associated with aberrant transcription and very weak D phenotype
.
Transfusion
.
2022
;
62
(
2
):
E14
-
6
.
51.
Thun
GA
,
Gueuning
M
,
Mattle-Greminger
M
.
Long-read sequencing in blood group genetics
.
Transfus Med Hemother
.
2023
;
50
(
3
):
184
97
.
52.
Levine
P
,
Celano
M
,
Lange
S
,
Stroup
M
.
The influence of gene interaction on dosage effects with complete anti-D sera
.
Vox Sang
.
1959
;
4
(
1
):
33
9
.
53.
Ceppellini
R
,
Dunn
LC
,
Turri
M
.
An interaction between alleles at the Rh locus in man which weakens the reactivity of the Rh0 factor (Du)
.
Proc Natl Acad Sci USA
.
1955
;
41
(
5
):
283
8
.
54.
Christophersen
MK
,
Jöud
M
,
Ajore
R
,
Vege
S
,
Ljungdahl
KW
,
Westhoff
CM
, et al
.
SMIM1 variants rs1175550 and rs143702418 independently modulate Vel blood group antigen expression
.
Sci Rep
.
2017
;
7
(
1
):
40451
.
55.
Westman
JS
,
Stenfelt
L
,
Vidovic
K
,
Möller
M
,
Hellberg
Å
,
Kjellström
S
, et al
.
Allele-selective RUNX1 binding regulates P1 blood group status by transcriptional control of A4GALT
.
Blood
.
2018
;
131
(
14
):
1611
6
.
56.
Yeh
CC
,
Chang
CJ
,
Twu
YC
,
Chu
CC
,
Liu
BS
,
Huang
JT
, et al
.
The molecular genetic background leading to the formation of the human erythroid-specific Xga/CD99 blood groups
.
Blood Adv
.
2018
;
2
(
15
):
1854
64
.
57.
Bannister
AJ
,
Kouzarides
T
.
Regulation of chromatin by histone modifications
.
Cell Res
.
2011
;
21
(
3
):
381
95
.
58.
Rowley
MJ
,
Corces
VG
.
Organizational principles of 3D genome architecture
.
Nat Rev Genet
.
2018
;
19
(
12
):
789
800
.