Abstract
Background
Accurate ABO blood group typing is essential for transfusion safety. However, ABO subtypes resulting from genetic variation can complicate this process. In contrast to the extensively studied exonic variants in the ABO gene, splice-site variants are rarely reported, and the mechanisms underlying their contribution to ABO subtypes remain poorly characterized.
Methods
Four unrelated, healthy blood donors exhibiting mixed-field agglutination with anti-A reagents were subjected to Sanger sequencing. Long-range PCR coupled with nanopore sequencing was further performed to resolve the phased haplotype across the entire ABO locus. The impact of the identified variant on splicing was predicted in silico and validated by an in vitro minigene assay; its effects on protein structure were assessed using bioinformatic tools. Short tandem repeat (STR) analysis was also conducted to exclude blood group chimerism.
Results
Four unrelated Chinese blood donors exhibiting mixed-field agglutination with anti-A reagents were identified. Sanger sequencing and nanopore sequencing revealed that all four probands harbored a novel c.239 + 6T>C variant in the ABO*A1.01 allele, with short tandem repeat (STR) analysis ruling out chimerism as the cause of the mixed-field agglutination. In silico analysis predicted activation of a cryptic splice site, leading to retention of a 55-bp intronic fragment in the mRNA transcript. Further, the minigene splicing assay validated this predicted aberrant splicing pattern and additionally revealed trace production of the normally spliced transcript. Secondary and three-dimensional structural modeling further predicted that the resulting aberrant transcript encodes a severely truncated and catalytically compromised glycosyltransferase A, providing a molecular basis for the observed A antigen expression defect.
Conclusion
In conclusion, a novel ABO variant (c.239 + 6T>C) responsible for the A3 phenotype was identified, and the mechanism underlying this phenotype was elucidated. By elucidating how a splice site variant disrupts glycosyltransferase function and alters antigen expression, this study contributes to a deeper understanding of the immunogenetic basis of ABO variation, with implications for transfusion safety and personalized immunohematology.
Keywords: A3 phenotype, ABO, immunogenetic, splice site variant, transfusion
Introduction
Since its discovery in 1900, the ABO blood group system has played an irreplaceable role in transfusion medicine and organ transplantation (Wang et al., 2022; Hu and Zhang, 2026). The ABO gene, located on chromosome 9 (9q34.2), encodes glycosyltransferases A (GTA) and B (GTB), which determine ABO phenotypes by catalyzing the addition of specific sugar residues to the H antigen acceptor substrate on red blood cells, resulting in the formation of A and B antigens, respectively (Paulson and Colley, 1989). Variations in this gene can alter the activity or specificity of these enzymes, giving rise to ABO blood group subtypes. Among these, the A3 phenotype is serologically defined by a characteristic mixed-field agglutination pattern with anti-A reagents, manifested as coexisting agglutinated red cell clumps and free red cells (Barjas-Castro et al., 2000). This mixed-field agglutination arises from ABO gene variations present in a homogeneous genetic background, instead of from chimerism, which involves the coexistence of two genetically distinct cell populations (Yamamoto et al., 1993). To date, at least 13 A3 alleles have been identified, harboring variations in exonic regions, intronic regulatory elements, or splice-site regions (Red blood histo). While previous molecular studies have largely focused on exonic and intronic regulatory variants (Li et al., 2007; Svensson et al., 2005; Thun et al., 2024), the role of splice-site variations in the etiology of the A3 phenotype remains poorly understood, despite their well-documented involvement in other weak subgroups such as Ael, and B3 (Yu et al., 2002; Chen et al., 2015). Thus, the present study aimed to address this gap by characterizing the functional impact of the c.239 + 6T>C variant in A3.
Faithful pre-mRNA splicing relies on accurate recognition of exon-intron boundaries. Splice-site variants can disrupt splicing through multiple mechanisms, including compromising canonical splice sites, activating cryptic splice sites, or altering the function of splicing regulatory elements. These disruptions can lead to exon skipping, intron retention, or the generation of aberrant splice variants, ultimately producing abnormal mRNA transcripts (Wilkinson et al., 2020; Xie et al., 2025; Anna and Monika, 2018). Such transcripts often harbor frameshifts, premature termination codons, or deletions of critical functional domains, ultimately yielding non-functional or severely impaired enzymatic proteins that cannot efficiently transfer N-acetylgalactosamine or galactose to the H antigen, leading to reduced A or B antigen expression (Huang et al., 2019). This advance provides a crucial foundation for systematically uncovering previously elusive non-coding variants (Gueuning et al., 2023). Furthermore, minigene splicing assays serve as a valuable tool for functional validation of these variants, thereby clarifying the molecular mechanisms by which splice-site variations lead to aberrant transcript formation (Sanoguera-Miralles et al., 2024).
In this study, we employed an integrated approach combining serological phenotyping, third-generation sequencing, in silico splicing prediction, in vitro minigene assays, and protein structure analysis to investigate a novel ABO variant (c.239 + 6T>C) identified in four unrelated individuals presenting with an A3 phenotype. By establishing a comprehensive genotype-splicing-phenotype correlation, we aim to elucidate how this intronic variant disrupts normal mRNA splicing, resulting in reduced A antigen expression. This work will not only clarify the immunogenetic basis of this recurrent A3-associated variant but also provide a valuable framework for the precise molecular diagnosis and investigation of ABO blood group subtypes in clinical practice.
Materials and methods
Blood sample collection
Four unrelated, healthy blood donors exhibiting mixed-field agglutination were included in this study. None had a history of blood transfusion or familial genetic disorders. After providing informed consent, 2 mL of peripheral blood was collected from each participant into EDTA-coated tubes.
ABO blood grouping
ABO blood grouping was performed using both the standard tube test and gel column agglutination. The standard tube test used anti-A, anti-B, anti-H, and anti-A1 reagents (Shanghai Hemo-pharmaceutical Biological Company) for forward typing, and A1, B, O, and A2 cells (Shanghai Hemo-pharmaceutical Biological Company) for plasma antibody identification in reverse typing. Gel column agglutination testing was performed with blood group gel cards (BIO-RAD, Switzerland) according to the manufacturer’s instructions, with reverse typing using A1 and B standard red blood cells (Shanghai Hemo-pharmaceutical Biological Company).
DNA extraction and quantification
Genomic DNA was isolated from EDTA-anticoagulated peripheral blood using TIANGEN Blood DNA Kit (for whole blood samples, TIANGEN Biotech, China). DNA concentration and quality were measured by NanoDrop 2000/2000c spectrophotomer (Thermo Fisher Scientific, Waltham, MA, United States).
Short tandem repeat (STR) analysis
Short tandem repeat (STR) genotyping was performed using the PowerPlex® 21 System (Promega Corporation, United States) according to the manufacturer’s instructions. A total of 20 autosomal STR loci together with the amelogenin (AMEL) locus were amplified using peripheral blood genomic DNA as the template. PCR products were separated by capillary electrophoresis on an ABI 3500 Genetic Analyzer (Applied Biosystems, United States), and allele calling was performed using GeneMapper™ ID-X software (Applied Biosystems, United States).
Sanger sequencing
The PCR reaction was performed in a total volume of 40 μL containing 1× DNA polymerase mix (Qingke, Beijing, China) with 1 U of high-fidelity DNA polymerase, 200 μM of each dNTP, and 2.0 mM Mg2+, as well as 0.2 μM of each primer and approximately 100 ng of genomic DNA. The primer sequences for amplification and sequencing, as well as the PCR amplification parameters, were based on protocols described in a previous study (Zhu et al., 2010). The resulting sequences were analyzed using SeqMan software (Lasergene package, DNASTAR Inc., Madison, WI, United States). The ABO*A1.01 allele sequence was used as the reference. Differences from this reference sequence are denoted with arrows.
Full-length ABO gene third-generation sequencing (three-fragment long-range PCR method)
To cover the full-length ABO gene, including approximately 5 kb upstream of exon 1 and 3 kb downstream of exon 7, three pairs of overlapping primers (LR1, LR2, LR3) (Supplementary Table S1) were designed to generate amplicons of 9.0 kb, 9.6 kb, and 12.0 kb, respectively. Adjacent amplicons overlapped by approximately 1.3 kb, enabling the discrimination of the two complete haplotypes based on the SNPs present in the overlapping regions (Figure 1). The reaction mixture and cycling conditions are detailed in Supplementary Tables S2, S3, respectively.
-
Library Preparation and ONT Sequencing
PCR amplicons were purified with AMPure XP beads (0.9:1 ratio). DNA end repair was performed using the NEBNext Ultra II End Repair/dA-Tailing module (30 µL), followed by 1:1 bead purification and barcode ligation (EXP-NBD114, Oxford Nanopore Technologies, ONT) with Blunt/TA Ligase Master Mix (NEB). Barcode-indexed libraries were purified (1:1 ratio), pooled according to concentrations measured by the Qubit dsDNA HS assay, and ligated with Adapter Mix (AMII, ONT) in Ligation Buffer (LNB) using NEBNext Quick T4 DNA Ligase. After final purification (0.4:1 ratio) and Long Fragment Buffer washes, the library was eluted into 15 µL of Elution Buffer. A median of 302 ng (range: 84–552 ng) of library was loaded onto R10.4 flow cells (FLO-PRO114M, ONT) and sequenced for 48 h using the SQK-NBD114-96 protocol. Sequencing was controlled by MinKNOW v24.02.16 on a 16-core Intel Xeon Silver 4314 CPU. Basecalling and demultiplexing were performed with Dorado v7.3.11 (super-accurate model at 400 bps) using a 24-thread, 64 GB RAM system with an NVIDIA RTX 4090 GPU.
-
Data Analysis and DNA Variant Calling
Raw reads were demultiplexed, and barcodes were automatically identified. For ABO genotyping, reads were aligned to the T2T-CHM13v2.0 genome and corrected using DAFEI Biotechnology software. FASTQ reads for each ABO allele were extracted with bedtools (v2.30.0), and a custom Python workflow converted them into BAM file clusters. Reads were re-aligned to the ISBT ABO reference gene using minimap2. Variants, including SNVs and small indels, were identified with DeepVariant v1.2.0. Alignments and genotypes were visualized in the Integrative Genomics Viewer. Final haplotypes, including crossover regions, were generated by DAFEI Biotechnology (Guangzhou, China). Results were compared to ISBT ABO alleles v4.1 across different serological phenotypes.
-
Sequencing and variant analysis
All four samples were successfully sequenced with a mean depth of 450× across the full-length ABO gene (range 210–680×). After applying the filtering criteria (depth ≥10, Phred quality ≥20, both strands supported), no additional pathogenic variants or structural rearrangements were detected apart from c.239 + 6T>C. Low-confidence calls that were excluded predominantly consisted of short insertions or deletions within homopolymeric tracts (≥4 identical nucleotides), a well-known systematic error mode of nanopore sequencing. These artifactual calls were manually inspected using IGV and removed when they showed strand bias or lacked clear signal transitions. No other unexpected or off-target variants were confirmed.
FIGURE 1.
Schematic representation of the three-fragment amplification strategy for the ABO gene. The ABO gene spans approximately 24 kb. Blue rectangles represent the coding regions (exons 1–7), and red rectangles represent the 5′and 3′untranslated regions (UTRs). The three overlapping amplicons (9.0 kb, 9.6 kb, and 12.0 kb) are indicated by horizontal bars within the gene structure.
Bioinformatic analysis
We employed the Rare Disease Data Center (RDDC) online RNA Splicer Tool (https://rddc.tsinghua-gd.org/zh), SpliceRover (http://bioit2.irc.ugent.be/rover/splicerover), and SpliceAPP (https://bc.imb.sinica.edu.tw/SpliceAPP/prediction.html) to predict the splicing patterns of potential splice variants. Furthermore, secondary structures of the wild-type (WT) and mutant (MT) glycosyltransferase proteins were generated using the online tool NovoPro (https://www.novopro.cn/tools/secondary-structure-prediction.html). Three-dimensional (3D) structural models were subsequently constructed using the Swiss-Model server (https://swissmodel.expasy.org/).
Minigene functional validation
To determine the effect of the c.239 + 6T>C variant on ABO mRNA splicing, a minigene splicing reporter assay was performed. Briefly, a wild-type minigene construct spanning ABO exons 4-6 and containing flanking intronic sequences (including the splice donor site of intron 5) was generated. The variant was then introduced by site-directed mutagenesis, and the splicing patterns of the wild-type and mutant constructs were compared after transient transfection into HEK293T cells.
-
Construction of wild-type and mutant minigene plasmids
Construction of wild-type minigene plasmid: A genomic segment spanning ABO exons 4-6 with flanking intronic sequences (∼4.25 kb) was amplified from a healthy control individual using primers that incorporated restriction sites. The primer, reaction mixture and cycling conditions are detailed in Supplementary Tables S4-S6, respectively.
The PCR product was purified by 1.5% agarose gel electrophoresis followed by gel extraction using a Gel Extraction Kit (TIANGEN, DP209). The purified PCR product was double-digested with BamHI and XhoI and inserted into the pMini-CopGFP vector (Hitrobio) using a seamless cloning kit (Beyotime). Positive clones were confirmed by Sanger sequencing.
Construction of mutant minigene plasmid: The c.239 + 6T>C mutation was introduced into the wild-type plasmid via site-directed mutagenesis PCR. The primer, reaction mixture and cycling conditions are detailed in Supplementary Tables S7-S9, respectively.
The PCR product was purified by 1.5% agarose gel electrophoresis followed by gel extraction using a Gel Extraction Kit (TIANGEN, DP209). The purified linearized product was self-circularized using a seamless cloning kit (Beyotime) and transformed into E. coli Top10 competent cells. Positive clones were confirmed by Sanger sequencing. Figure 2 outlines the construction of the mutant minigene plasmid.
-
Cell transfection and RT-PCR
HEK293T cells were seeded at 2–3 × 105 cells per 35 mm dish. When cultures reached 50%–60% confluence on the following day, cells were transfected with 4 μg of wild-type or mutant minigene plasmid using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. At 48 h post-transfection, total RNA was extracted with TRIzol reagent and reverse-transcribed into cDNA using the HiScript II First Strand cDNA Synthesis Kit (Vazyme). RT-PCR was then performed. The primer, reaction mixture and cycling conditions are detailed in Supplementary Tables S10-S12, respectively.
-
Analysis of RT-PCR products
RT-PCR products were analyzed by two complementary methods. First, for routine analysis, products were resolved on a 1.5% agarose gel stained with GelRed at constant voltage for 30–40 min. Gel images were captured to assess amplification specificity, and bands of interest were excised, purified, and subjected to Sanger sequencing to confirm the identity of the splicing products at nucleotide resolution.
FIGURE 2.
Flowchart for the Construction of c.239 + 6T>C Mutant Minigene Plasmid.
Second, to precisely determine transcript lengths and relative abundances, the same RT-PCR products were analyzed by capillary electrophoresis. The forward primer used in the RT-PCR reaction was labeled with 6-FAM at its 5’ end. In brief, an aliquot of the RT-PCR product was mixed with an appropriate volume of loading buffer, denatured at 95 °C for 5 min, and immediately cooled on ice. Capillary electrophoresis was then performed using a genetic analyzer, and transcript sizes were determined by comparison with an internal size standard. The relative abundance of each splicing variant was calculated as the ratio of its peak area to the total peak area of all detected transcripts.
Results
Serological results
The ABO blood group serology results for the four subjects are presented in Table 1. By the tube method, forward typing revealed 2+ mixed-field agglutination with anti-A and no reaction with anti-B for all subjects; in reverse typing, all subjects showed 4+ agglutination with B cells and no agglutination with O or A2 cells. Reactivity with A1 cells, however, was variable: three subjects were non-reactive, while Subject 1 showed weak (1 + w) agglutination. Additionally, all subjects were non-reactive with anti-A1 lectin but exhibited strong (4+) reactions with anti-H. Consistent with these findings, gel column agglutination also showed double population (2+, ∼50% each)) reactions with anti-A and no reaction with anti-B in forward typing. In reverse typing, gel column results similarly demonstrated 4+ agglutination with B cells (Subjects 1, 3, and 4 showed 4 + w) and no agglutination with A1 cells or the control (Figure 3). Collectively, these serological patterns are characteristic of the A3 subgroup, with inconsistent anti-A1 production.
TABLE 1.
Results of ABO serological grouping by tube test and gel column agglutination technology.
| Sample | Forward typing (Tube method) | Forward typing (Gel column) | Reverse typing (Tube method) | Reverse typing (Gel column) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Anti-A | Anti-B | Anti-A1 | Anti-H | Anti-A | Anti-B | A1c | A2c | Bc | Oc | A1c | Bc | Oc | |
| Subject 1 | 2+mf | 0 | 0 | 4+ | 2+dcp | 0 | 1+W | 0 | 4+ | 0 | 0 | 4+W | 0 |
| Subject 2 | 2+mf | 0 | 0 | 4+ | 2+dcp | 0 | 0 | 0 | 4+ | 0 | 0 | 4+ | 0 |
| Subject 3 | 2+mf | 0 | 0 | 4+ | 2+dcp | 0 | 0 | 0 | 4+ | 0 | 0 | 4+W | 0 |
| Subject 4 | 2+mf | 0 | 0 | 4+ | 2+dcp | 0 | 0 | 0 | 4+ | 0 | 0 | 4+W | 0 |
mf, mixed-field agglutination; dcp, double population (indicating mixed-field reaction in gel column); w, weak reaction.
FIGURE 3.
Serological phenotyping by gel column agglutination. All four subjects showed consistent results: forward typing showed anti-A double population (2+, ∼50% each) and anti-B negative; reverse typing showed strong agglutination with B cells (range 4 + w ∼ 4+) but not with A1 cells.
Sanger sequencing results
Sanger sequencing was initially performed on exons 5-7 of the ABO gene in the four subjects. The results revealed no variants within exons 5, 6, or 7 in the context of an ABO*A1.01 allelic background. However, an identical c.239 + 6T>C variant within intron 5 was identified in all individuals (Figure 4).
FIGURE 4.
Sanger sequencing of ABO exons 5-7 in Subjects 1–4. All four subjects carried an identical c.239 + 6T>C variant.
Single-molecule nanopore sequencing results
Nanopore sequencing revealed that, across all four samples, no variants were identified in the exonic or other intronic regions of the ABO*A1.01 haplotype, with the sole exception of a single-base substitution located in the intronic region adjacent to the exon 5 splice site (c.239 + 6T>C). As summarized in Table 2, a novel ABO*A1.01-like allele was identified in all four subjects, defined by the c.239 + 6T>C variant. The other allele was ABO*O.01.01 in Subjects 1 and 2, and ABO*O.01.02 in Subjects 3 and 4.
TABLE 2.
Results of nanopore single-molecule sequencing of the ABO gene.
| Sample | Gene | Haplotype 1 | Haplotype 2 | Phenotype | ||||
|---|---|---|---|---|---|---|---|---|
| Pheno-type | Allele | Mutation | Pheno-type | Allele | Mutation | |||
| Subject 1 | ABO | 0 | ABO*O.01.01 | c.261delG; | A3 | ABO*A.NEW | c.239+6T>C | A3 |
| Subject 2 | ABO | 0 | ABO*O.01.01 | c.261delG; | A3 | ABO*A.NEW | c.239+6T>C | A3 |
| Subject 3 | ABO | 0 | ABO*O.01.02 | c.106G>T;c.188G>A; c.189C>T; c.220C>T; c.261delG;c.297A>G; c.646T>A;c.681G>A; c.771C>T; c.829G>A |
A3 | ABO*A.NEW | c.239+6T>C | A3 |
| Subject 4 | ABO | 0 | ABO*O.01.02 | c.106G>T; c.188G>A; c.189C>T; c.220C>T; c.261delG; c.297A>G; c.646T>A; c.681G>A; c.771C>T;c.829G>A |
A3 | ABO*A.NEW | c.239+6T>C | A3 |
Short tandem repeat analysis results
Due to sample depletion, STR analysis could not be performed for Subject 3. For the remaining three subjects (Subjects 1, 2, and 4), STR analysis was performed to exclude blood group chimerism as a cause of the mixed-field agglutination. All three STR profiles showed exclusively single (homozygous) or two peaks (heterozygous) at each of the 20 autosomal loci examined, with no locus exhibiting three or more peaks and no significant peak height or area imbalance across the profile (Figure 5)-findings consistent with the absence of chimerism.
FIGURE 5.
STR profiles of Subject 1, Subject 2, and Subject 4. All three subjects showed one or two allelic peaks at each of the 20 autosomal loci, with no locus displaying three or more alleles.
In silico splicing analysis
Splicing analysis using the RDDC RNA Splicer tool revealed that the novel ABO variant (c.239 + 6T>C) generates a cryptic donor site located 55 nucleotides downstream of the canonical 5′splice site of intron 5, which is predicted to result in aberrant splicing. Splice-altering potential scores (DanQ and SPTransformer) increased significantly following the variant: in intron 5, DanQ rose from 0.9868 to 0.9967, and SPTransformer from 0.3176 to 0.9884 (Figure 6).
FIGURE 6.
Splicing analysis using the RDDC RNA Splicer tool for the ABO c.239 + 6T>C variant: The variant generates a cryptic donor site 55 bp downstream of the intron 5 donor site, leading to a 55 bp insertion, frameshift mutation, and premature termination. DanQ and SPTransformer scores range from 0 to 1, with values closer to 1 indicating a higher likelihood of being a true splice site.
These findings were further supported by SpliceRover and SpliceAPP. SpliceRover scored the wild-type (WT) at 0.570276 and the mutant (MT) at 0.953626, suggesting activation of this cryptic donor site (Table 3). SpliceAPP also predicted a significant splicing alteration for this variant, with a score of 0.1903, which exceeds the official 5′ss model threshold of 0.13, indicating a high likelihood of disrupted canonical splicing (Table 4). Collectively, these in silico predictions are highly consistent and indicate that c.239 + 6T>C creates a cryptic donor site 55 nt downstream of the canonical 5′splice site of intron 5.
TABLE 3.
Prediction of wild-type (WT) and mutant (MT) splicing probability by SpliceRover.
| SpliceRover | Hit (with context) | Score | Prediction |
|---|---|---|---|
| WT | TGAAG|GTATTAG | 0.570276 | Baseline |
| MT | TGAAG|GTATTAG | 0.953626 | Cryptic donor activated |
Scores range from 0 to 1; values closer to 1 indicate a higher likelihood of being a true splice site; “|”: a cryptic donor site located 55 nucleotides downstream of the 5′ splice site of intron 5.
TABLE 4.
SpliceAPP prediction of splicing effects for the c.239+6T>C variant.
| Prediction tool | Variant | Model | Distance to 5'ss | Score | Effect |
|---|---|---|---|---|---|
| SpliceAPP | ABO:c.239 + 6T>C | 5' ss | 6 nt | 0.1903 | Significant |
Scores range from 0 to 1; Score > 0.13 (official threshold for the 5‘ ss model) indicates a significant splicing alteration.
Minigene splicing assay
We performed an in vitro minigene assay to definitively characterize the impact of the ABO c.239 + 6T>C variant on RNA splicing. RT-PCR gel electrophoresis results revealed a single 290 bp (vector: 71 bp, exons 4–6: 48 + 36+135 bp; total 290 bp) product from the wild-type (WT) construct, whereas the mutant (MT) produced a longer 345 bp band (Figure 7A). Capillary electrophoresis further resolved these products with higher precision: the WT construct yielded a normal transcript of 288.65 bp; the MT construct yielded a predominant aberrant transcript of 344.62 bp along with a minor normal transcript of 288.65 bp, indicating that the variant severely impairs but does not completely abolish canonical splicing (Figure 7B). Sanger sequencing confirmed normal splicing in the WT, while in the MT, it revealed that the aberrant product resulted from the retention of a 55-nucleotide intronic fragment (Figure 7D). Due to the extremely low abundance of the normal transcript in the MT sample, Sanger sequencing failed to detect it. Mechanistic analysis demonstrated that the c.239 + 6T>C variant generates two splicing patterns. The aberrant splicing pattern disrupts the canonical donor splice site at the 5′end of intron 5, activating a cryptic donor site 55 bp downstream and resulting in a 55 bp intronic insertion between exons 5 and 6, whereas the normal splicing pattern produces the correctly spliced transcript (Figure 7C). The aberrant 55 bp intronic insertion shifts the reading frame and introduces a premature termination codon (p.Cys80TrpfsTer2). Amino acid alignment reveals the resultant frameshift and early termination (Supplementary Figure S1A). Furthermore, structural modeling using the Swiss-Model server demonstrates that, in contrast to the fully folded wild-type (WT) enzyme, the mutant (MT) protein is drastically truncated and lacks essential functional domains (Supplementary Figure S1B).
FIGURE 7.
Validation of aberrant splicing for the ABO c.239+6T>C variant by minigene splicing assays. (A) RT-PCR gel electrophoresis results: WT shows a 290 bp band; MT shows a 345 bp band, indicating a 55 bp insertion. (B) Capillary electrophoresis results: WT shows a 288.65 bp fragment; MT shows a 344.62 bp fragment along with a small amount of the 288.65 bp normal transcript. A small amount of aberrant transcripts was detected in the wild-type sample, which is consistent with the expected background level. The x-axis shows product length (bp); the y-axis shows relative fluorescence units. (C) Splicing pattern of the c.239 + 6T>C variant. Aberrant pattern: c.239 6T>C disrupts the canonical donor splice site of intron 5, activating a cryptic donor site 55 bp downstream, resulting in a 55 bp intronic insertion between exons 5 and 6. Normal pattern: Normal splicing. (D) Sequencing results: WT shows the normal exon 5–6 junction; MT shows a 55 bp intronic retention (c.239 + 1_239 + 55).
Discussion
Accurate ABO blood typing is essential for transfusion compatibility, yet ABO subtypes often complicate identification. With the advent of third-generation sequencing technologies, splice-site variations have emerged as a key focus in ABO subtype research. Notably, these non-coding variants involve more complex mechanisms than coding variants, making their elucidation critical to understanding ABO subtype genesis (Yang et al., 2026; Hong et al., 2021). mRNA splicing is a critical step in eukaryotic gene expression. The spliceosome precisely excises non-coding regions and ligates coding regions to generate mature mRNA. This process depends on the recognition of consensus sequences, including the 5′splice site, 3′splice site, and branch point (Zhang et al., 2022). Variants that disrupt these sequences impair spliceosome recruitment, triggering aberrant splicing events such as exon skipping or intron retention (Abramowicz and Gos, 2019). Ultimately, these splicing defects compromise the production of authentic transcripts, impair glycosyltransferase synthesis, and disrupt antigen expression.
In this study, we identified four unrelated probands exhibiting mixed-field agglutination with anti-A, a hallmark of the A3 phenotype. Sanger sequencing and nanopore sequencing revealed that all four samples carried an ABO*A1.01 allele harboring a novel c.239 + 6T>C variant. The conserved GT-AG dinucleotide motif at splice sites is essential for accurate mRNA splicing (Lee and Rio, 2015). Although the c.239 + 6T>C variant does not alter the GT dinucleotide at the 5′splice site, previous studies have shown that substitutions at positions +5 and +6 can impair U1 snRNP binding to pre-mRNA, leading to aberrant splicing (Schmid et al., 2013). In two classic studies, Chen et al. dissected the c.155 + 5G>A (IVS3+5G>A) variant in a B3 individual and the c.374 + 5G>A (IVS6+5G>A) variant in an Ael individual, demonstrating that these variants drive aberrant splicing, predominantly generating transcripts with skipping of exons 3 and 6, respectively (Chen et al., 2015; Chen et al., 2012). Subsequently, Hong et al. identified a novel allele defining the Bel phenotype, harboring the c.28 + 5G>A variant (Hong et al., 2021). Shao et al. further demonstrated that the same variant, when placed in an ABO*A2.01 background, shifts the expected A2 phenotype to Ael (Shao et al., 2024). Both studies point to a common splicing defect characterized by partial intron 1 retention (Hong et al., 2021; Shao et al., 2024). Collectively, these studies establish that aberrant RNA splicing drives the genesis of these ABO subtypes.
Unlike the well-characterized splice-site variants in Ael, Bel, and B3 subtypes, those associated with the A3 phenotype have been rarely investigated. The novel c.239 + 6T>C variant represents the second splice-site variant linked to the A3 phenotype, following the c.98 + 3A>G variant previously reported by Yang et al. (Cai et al., 2013). To further characterize the c.239 + 6T>C variant, we conducted in silico splicing analyses using RDDC, SpliceRover, and SpliceAPP. All three tools predicted aberrant splicing. Critically, both RDDC (DanQ: 0.9967; SPTransformer: 0.9884) and SpliceRover (0.953626) indicated activation of a cryptic splice site 55 bp downstream of the canonical site, predicting the generation of an aberrantly elongated transcript retaining an intronic fragment. SpliceAPP yielded a score of 0.1903 (5’ss model, threshold >0.13), further supporting this prediction. Analysis of minigene splicing products by gel electrophoresis confirmed this prediction. The wild-type A allele transcript yielded a 290-bp band, whereas the c.239 + 6T>C variant produced a 345-bp band, consistent with the retention of a 55-bp intronic sequence. This aberrant splicing event introduces a premature stop codon, which is predicted to truncate the polypeptide to 80 amino acids. Similarly, the c.98 + 3A>G variant is predicted to result in exon skipping or intron retention, generating truncated polypeptides of 33 and 32 amino acids, respectively (Supplementary Figure S2). Importantly, transcripts bearing premature termination codons are typically recognized and degraded by the nonsense-mediated mRNA decay (NMD) pathway-a conserved eukaryotic mRNA surveillance mechanism (Behera et al., 2025). Consequently, these aberrant transcripts are unlikely to be translated. Even if a small fraction escape NMD, the resulting truncated GTA polypeptides lack the catalytic domain and are therefore predicted to be catalytically inactive.
However, the mixed-field agglutination observed in A3 red blood cells following incubation with anti-A reagent suggests reduced, rather than absent, A antigen expression on the cell surface. To investigate this discrepancy, we performed capillary electrophoresis on the minigene-derived RT-PCR products, a more sensitive method than agarose gel electrophoresis. Strikingly, this analysis uncovered a preponderance of aberrant transcripts with a 55 bp intronic retention, together with a scant population of correctly spliced normal transcripts. These correctly spliced normal transcripts are likely generated by residual activity of the canonical splice site and are expected to be translated into trace amounts of functional GTA (Bueno-Martínez et al., 2022). This minute quantity of GTA transfers A antigens onto a subset of red blood cells, thereby contributing to the A3 phenotype (Ogasawara et al., 2024; Liang et al., 2015). A similar mechanism has been described for the ABO*A2.01 allele carrying the c.28 + 5G>A variant. Minigene assay revealed that this variant generates not only an aberrant transcript (introducing a premature termination codon and yielding a nonfunctional GTA) but also a trace amount of a wild-type transcript that produces a functional GTA, thereby shifting the expected A2 phenotype to the weaker Ael phenotype (Shao et al., 2024). Collectively, these findings establish a refined mechanistic model for A3 phenotype generation: a splice-site variant that disrupts canonical splicing while permitting residual correct splicing, producing trace amounts of functional glycosyltransferase A sufficient for weak but detectable A antigen expression. This splicing paradigm may extend to other ABO subtypes and beyond.
Notably, anti-A1 antibodies were detected in one of the four A3 probands, possibly reflecting variations in immune response among individuals with the A3 phenotype (Daniels and Daniels, 2013). Regarding transfusion management, as the four subjects were blood donors, it was not possible to monitor the outcome of any transfusion. However, should future transfusion be required, it is advisable for these individuals to receive washed group O red blood cells, particularly for those who have developed anti-A1 antibodies, as transfusion of group A red blood cells should be avoided in these cases to minimize the risk of hemolytic transfusion reactions.
One limitation of this study is that, due to the unavailability of fresh blood samples, we were unable to directly examine the transcript status in vivo or validate GTA expression at the protein level.
Conclusion
In conclusion, we identified a novel ABO*A1.01 (c.239 + 6T>C) allele responsible for the A3 phenotype in four unrelated Chinese blood donors. The nucleotide sequence of this allele has been submitted to GenBank under accession no. PV862049. By combining Sanger and nanopore sequencing with minigene splicing assays, we demonstrated that the c.239 + 6T>C variant causes both aberrant splicing and residual correct splicing. The former produces a truncated, inactive GTA, while the latter generates trace amounts of functional GTA. These two mechanisms together explain the A3 phenotype. Together, these findings expand the variant spectrum of the ABO gene and highlight the diagnostic value of integrating sequencing-based and functional assays for deciphering blood group subtypes. This study contributes to a deeper understanding of the immunogenetic basis of ABO variation, with implications for transfusion safety and personalized immunohematology.
Acknowledgements
The authors thank the proband for his participation and cooperation in this study.
Funding Statement
The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the Chengdu Medical Research Fund (2025014).
Footnotes
Edited by: Faming Zhu, Blood Center of Zhejiang Province, China
Reviewed by: Eva Maria Matzhold, Medical University of Graz, Austria
Xiaohong Cai, Shanghai Jiao Tong University, China
Data availability statement
The data presented in this study have been deposited in the GenBank database (accession number PV862049; available at: https://www.ncbi.nlm.nih.gov/nuccore/PV862049).
Ethics statement
The studies involving humans were approved by Ethical Committee of the Chengdu Blood Center. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.
Author contributions
H-LG: Writing – original draft. JH: Writing – original draft. J-LG: Formal Analysis, Methodology, Writing – review and editing. JL: Investigation, Software, Writing – review and editing. X-MZ: Project administration, Resources, Writing – review and editing. XC: Funding acquisition, Writing – review and editing. X-MF: Writing – review and editing.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fgene.2026.1839848/full#supplementary-material
Structural analysis of wild-type and mutant GTA for the c.239 + 6T>C variant. (A) Secondary structures and amino acid sequences of wild-type and mutant. Helical regions are highlighted in red, and β-strand regions in blue. (B) Three-dimensional structural models of wild-type A glycosyltransferase and its truncated mutant.
Prediction of aberrant splicing for the ABO c.98+3A>G variant by in silico prediction and structural analysis of mutant GTA. (A) The c.98+3A>G variant exhibits two distinct splicing patterns. Pattern 1: Activation of a cryptic splice donor site within intron 2, resulting in a 152 bp insertion and a consequent frameshift with premature termination. Pattern 2: Skipping of exon 2, also resulting in a frameshift and premature termination. (B) Secondary structures and amino acid sequences of GTA resulting from the two aberrant splicing patterns. Secondary structure predictions were generated using the online tool (https://www.novopro.cn/tools/secondary-structure-prediction.html).
References
- Abramowicz A., Gos M. (2019). Correction to: splicing mutations in human genetic disorders: examples, detection, and confirmation. J. Applied Genetics. 60, 231. 10.1007/s13353-019-00493-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- Anna A., Monika G. (2018). Splicing mutations in human genetic disorders: examples, detection, and confirmation. J. Applied Genetics. 59, 253–268. 10.1007/s13353-018-0444-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Barjas-Castro M. L., Carvalho M. H., Locatelli M. F., Bordin S., Saad S. T. (2000). Molecular heterogeneity of the A3 subgroup. Clin. Laboratory Haematology 22, 73–78. 10.1046/j.1365-2257.2000.00289.x [DOI] [PubMed] [Google Scholar]
- Behera A., Panigrahi G. K., Sahoo A. (2025). Nonsense-mediated mRNA decay in human health and diseases: current understanding, regulatory mechanisms and future perspectives. Mol. Biotechnol. 67, 3374–3390. 10.1007/s12033-024-01267-7 [DOI] [PubMed] [Google Scholar]
- Bueno-Martínez E., Sanoguera-Miralles L., Valenzuela-Palomo A., Esteban-Sánchez A., Lorca V., Llinares-Burguet I., et al. (2022). Minigene-based splicing analysis and ACMG/AMP-based tentative classification of 56 ATM variants. J. Pathol. 258, 83–101. 10.1002/path.5979 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cai X., Jin S., Liu X., Fan L., Lu Q., Wang J., et al. (2013). Molecular genetic analysis of ABO blood group variations reveals 29 novel ABO subgroup alleles. Transfusion 53, 2910–2916. 10.1111/trf.12168 [DOI] [PubMed] [Google Scholar]
- Chen D. P., Tseng C. P., Wang W. T., Sun C. F. (2012). Genetic and mechanistic evaluation for the mixed-field agglutination in B3 blood type with IVS3+5G>A ABO gene mutation. PloS One 7, e37272. 10.1371/journal.pone.0037272 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen D. P., Sun C. F., Ning H. C., Peng C. T., Wang W. T., Tseng C. P. (2015). Genetic and mechanistic evaluation for the weak A phenotype in Ael blood type with IVS6 + 5G>A ABO gene mutation. Vox Sanguinis 108, 64–71. 10.1111/vox.12196 [DOI] [PubMed] [Google Scholar]
- Daniels G. (2013). “ABO, H, and lewis systems,” in Human Blood Groups, Editor Daniels G. 3rd ed. (West Sussex: Wiley-Blackwell; ), 11–95. [Google Scholar]
- Gueuning M., Thun G. A., Wittig M., Galati A. L., Meyer S., Trost N., et al. (2023). Haplotype sequence collection of ABO blood group alleles by long-read sequencing reveals putative A1-diagnostic variants. Blood Advances 7, 878–892. 10.1182/bloodadvances.2022007133 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hong X., Ying Y., Zhang J., Chen S., Xu X., He J., et al. (2021). Six splice site variations, three of them novel, in the ABO gene occurring in nine individuals with ABO subtypes. J. Translational Medicine 19, 470. 10.1186/s12967-021-03141-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hu P., Zhang X. (2026). Effects of ABO-incompatible blood transfusion on immune response and rejection after organ transplantation. Clin. Transplantation 40, e70469. 10.1111/ctr.70469 [DOI] [PubMed] [Google Scholar]
- Huang H., Jin S., Liu X., Wang Z., Lu Q., Fan L., et al. (2019). Molecular genetic analysis of weak ABO subgroups in the Chinese population reveals ten novel ABO subgroup alleles. Blood Transfus. = Trasfus. del Sangue 17, 217–222. 10.2450/2018.0091-18 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lee Y., Rio D. C. (2015). Mechanisms and regulation of alternative Pre-mRNA splicing. Annu. Review Biochemistry 84, 291–323. 10.1146/annurev-biochem-060614-034316 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li L., Yang M. H., Chak K. F., Lin P. H., Lai C. H., Lin K. T., et al. (2007). Three missense mutations, including a novel 860C>T transition, and allelic enhancement phenomenon associated with ABO blood subgroups A in Taiwan. Transfusion 47, 1014–1021. 10.1111/j.1537-2995.2007.01228 [DOI] [PubMed] [Google Scholar]
- Liang W., Yang L., Mei C., Xu D., Deng G., He Y., et al. (2015). Study of molecular mechanism for a blood sample with A3 phenotype. Chin. J. Med. Genet. 32, 703–706. 10.3760/cma.j.issn.1003-9406.2015.05.021 [DOI] [PubMed] [Google Scholar]
- Ogasawara K., Sano R., Kominato Y. (2024). Review of ABO expression and variations based on transcriptional regulation of the ABO blood group gene. Transfus. Med. Hemother 51, 210–224. 10.1159/000536556 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Paulson J. C., Colley K. J. (1989). Glycosyltransferases. Structure, localization, and control of cell type-specific glycosylation. J. Biological Chemistry 264, 17615–17618. [PubMed] [Google Scholar]
- Sanoguera-Miralles L., Valenzuela-Palomo A., Bueno-Martínez E., Esteban-Sánchez A., Lorca V., Llinares-Burguet I., et al. (2024). Systematic minigene-based splicing analysis and tentative clinical classification of 52 CHEK2 splice-site variants. Clin. Chemistry 70, 319–338. 10.1093/clinchem/hvad125 [DOI] [PubMed] [Google Scholar]
- Schmid F., Hiller T., Korner G., Glaus E., Berger W., Neidhardt J. (2013). A gene therapeutic approach to correct splice defects with modified U1 and U6 snRNPs. Hum. Gene Therapy 24, 97–104. 10.1089/hum.2012.110 [DOI] [PubMed] [Google Scholar]
- Shao L. N., Song W. Q., Zhou L., Pan L. Z., Duan Y., Xiao N., et al. (2024). Characterization of a novel AEL allele harboring a c.28 + 5G>A mutation on the ABO*A2.01 background: a study utilizing PacBio third-generation sequencing and functional assays. Front. Immunology 15, 1396426. 10.3389/fimmu.2024.1396426 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Svensson L., Rydberg L., Hellberg A., Gilliver L. G., Olsson M. L., Henry S. M. (2005). Novel glycolipid variations revealed by monoclonal antibody immunochemical analysis of weak ABO subgroups of A. Vox Sanguinis 89, 27–38. 10.1111/j.1423-0410.2005.00642 [DOI] [PubMed] [Google Scholar]
- Thun G. A., Gueuning M., Sigurdardottir S., Meyer E., Gourri E., Schneider L., et al. (2024). Novel regulatory variant in ABO intronic RUNX1 binding site inducing A(3) phenotype. Vox Sanguinis 119, 377–382. 10.1111/vox.13580 [DOI] [PubMed] [Google Scholar]
- Wang Y., Mou Q., Lei H., Heililahong H., Zou W., Wang X., et al. (2022). Molecular biology analysis of ABO blood group variants caused by natural chimaerism. Vox Sanguinis 117, 1310–1317. 10.1111/vox.13356 [DOI] [PubMed] [Google Scholar]
- Wilkinson M. E., Charenton C., Nagai K. (2020). RNA splicing by the spliceosome. Annu. Review Biochemistry 89, 359–388. 10.1146/annurev-biochem-091719-064225 [DOI] [PubMed] [Google Scholar]
- Xie D., Peng Q., Tian Y., Han Y., Lu G. (2025). Protocol to study the effects of mutations near splicing sites on pre-mRNA splicing. STAR Protocols 6, 103673. 10.1016/j.xpro.2025.103673 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yamamoto F., McNeill P. D., Yamamoto M., Hakomori S., Harris T., Judd W. J., et al. (1993). Molecular genetic analysis of the ABO blood group system: 1. Weak subgroups: A3 and B3 alleles. Vox Sanguinis 64, 116–119. 10.1111/j.1423-0410.1993.tb02528.x [DOI] [PubMed] [Google Scholar]
- Yang H., Long D., Lei H., Guo L., Gao C., Hao H., et al. (2026). Functional analysis of four splice-site variants, including a novel variant, on antigen expression and ABO subgroup formation. J. Molecular Diagnostics JMD. 28, 170–178. 10.1016/j.jmoldx.2025.10.007 [DOI] [PubMed] [Google Scholar]
- Yu L. C., Twu Y. C., Chou M. L., Chang C. Y., Wu C. Y., Lin M. (2002). Molecular genetic analysis for the B(3) allele. Blood 100, 1490–1492. 10.1182/blood-2002-01-0188 [DOI] [PubMed] [Google Scholar]
- Zhang H., Chen C., Wu X., Lou C., Liang Q., Wu W., et al. (2022). Effects of 14 F9 synonymous codon variants on hemophilia B expression: alteration of splicing along with protein expression. Hum. Mutation 43, 928–939. 10.1002/humu.24377 [DOI] [PubMed] [Google Scholar]
- Zhu F., Tao S., Xu X., Ying Y., Hong X., Zhu H., et al. (2010). Distribution of ABO blood group allele and identification of three novel alleles in the Chinese Han population. Vox Sanguinis 98, 554–559. 10.1111/j.1423-0410.2009.01291.x [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Structural analysis of wild-type and mutant GTA for the c.239 + 6T>C variant. (A) Secondary structures and amino acid sequences of wild-type and mutant. Helical regions are highlighted in red, and β-strand regions in blue. (B) Three-dimensional structural models of wild-type A glycosyltransferase and its truncated mutant.
Prediction of aberrant splicing for the ABO c.98+3A>G variant by in silico prediction and structural analysis of mutant GTA. (A) The c.98+3A>G variant exhibits two distinct splicing patterns. Pattern 1: Activation of a cryptic splice donor site within intron 2, resulting in a 152 bp insertion and a consequent frameshift with premature termination. Pattern 2: Skipping of exon 2, also resulting in a frameshift and premature termination. (B) Secondary structures and amino acid sequences of GTA resulting from the two aberrant splicing patterns. Secondary structure predictions were generated using the online tool (https://www.novopro.cn/tools/secondary-structure-prediction.html).
Data Availability Statement
The data presented in this study have been deposited in the GenBank database (accession number PV862049; available at: https://www.ncbi.nlm.nih.gov/nuccore/PV862049).