Six splice site variations, three of them novel, in the ABO gene occurring in nine individuals with ABO subtypes

Nucleotide mutations in the ABO gene may reduce the activity of glycosyltransferase, resulting in lower levels of A or B antigen expression in red blood cells. Six known splice sites have been identified according to the database of red cell immunogenetics and the blood group terminology of the International Society of Blood Transfusion. Here, we describe six distinct splice site variants in individuals with ABO subtypes. The ABO phenotype was examined using a conventional serological method. A polymerase chain reaction sequence-based typing method was used to examine the whole coding sequence of the ABO gene. The ABO gene haplotypes were studied using allele-specific primer amplification or cloning technology. In silico analytic tools were used to assess the functional effect of splice site variations. Six distinct variants in the ABO gene splice sites were identified in nine individuals with ABO subtypes, including c.28 + 1_2delGT, c.28 + 5G > A, c.28 + 5G > C, c.155 + 5G > A, c.204-1G > A and c.374 + 5G > A. c.28 + 1_2delGT was detected in an Aw individual, while c.28 + 5G > A, c.28 + 5G > C, and c.204-1G > A were detected in Bel individuals. c.155 + 5G > A was detected in one B3 and two AB3 individuals, whereas c.374 + 5G > A was identified in two Ael individuals. Three novel splice site variants (c.28 + 1_2delGT, c.28 + 5G > A and c.28 + 5G > C) in the ABO gene were discovered, all of which resulted in low antigen expression. In silico analysis revealed that all variants had the potential to alter splice transcripts. Three novel splice site variations in the ABO gene were identified in Chinese individuals, resulting in decreased A or B antigen expression and the formation of ABO subtypes.


Background
Karl Landsteiner discovered the ABO blood group system in 1900, and it was of great clinical significance in blood transfusion and organ transplantation [1]; additionally, it is important for studying the development of numerous human diseases [2][3][4]. Incompatibility in the ABO blood group may result in severe haemolytic reactions during transfusion and neonatal haemolytic disease [5,6].
Although the ABO blood group system is composed of four phenotypes, including A, B, O, and AB, the distribution of ABO phenotypes varies across populations and regions [9]. Additionally, some subtypes of ABO phenotypes have been identified in populations, which often exhibit differences in forward and reverse ABO typing due to reduced antigen and/or antibody expression [9,11,12]. Variations in the ABO gene may affect the activity and/or specificity of glycosyltransferases, resulting in the formation of ABO subtypes. According to the database of names for ABO blood group alleles (v1.1) by the red cell immunogenetics and blood group terminology of international society of Blood Transfusion (ISBT), five splice sites with Aweak or Ael and one splice site with B3 were found in populations from around the world. In this study, we described six different splice site variants of the ABO gene in individuals with ABO subtypes.

Study specimens
Individuals with ABO subtypes were either blood donors or patients. All specimens were obtained after the individuals provided informed consent. This study was approved by the ethics committee of the Blood Center of Zhejiang Province, China. The difference in these specimens was discovered during routine ABO blood group typing. The specimens placed in tubes with or without EDTA anticoagulant were sent to the immunohaematology reference laboratory in the Blood Center of Zhejiang Province for further analysis.

Serological tests
A, B, and H antigens, as well as anti-A and anti-B antibodies, were detected using a conventional serological method [13,14]. Anti-A, anti-A 1 , anti-B, anti-AB, and anti-H antibody reagents were used (Shanghai Blood Biotechnology Co., Ltd., Shanghai, China). Red blood cells (RBCs) in groups A, B, and O were prepared in the laboratory using fresh blood from three donors of the same type at random.

ABO gene full exon sequencing analysis
According to our previous reports, we used the polymerase chain reaction sequence-based typing (PCR-SBT) technique to analyse the entire CDS of the ABO gene [9,13,14]. Three sets of primers were used to amplify all exons of the ABO gene. The amplicons were purified, followed by sequencing and analysis using an ABI 3730 sequencer (Applied Biosystems, Foster City, CA, USA). SeqScape v2.5 software (Applied Biosystems) was used to evaluate the sequencing data. The ABO gene reference sequence was obtained from GenBank (ID number NG_006669.2), and the ABO genotype was assigned based on nucleotide polymorphism. The ABO allele was nominated in accordance with the ISBT guidelines for red cell immunogenetics and blood group terminology [15].

ABO gene sequence analysis using NGS
Sequences from the start codon to the stop codon of the ABO gene were analysed using next-generation sequencing (NGS). First, the ABO gene sequence was amplified using two pairs of primers. In the first pair, the forward and reverse primer sequences were 5'GCG CCG TCC CTT CCT AGC AG 3' and 5' AGC CAC CAA CTT CCC CTA GT3' . The primer sequences in the second pair were 5'TAC TCA CCT ATT ATT GGC CTT TGG TT3' and 5'TAG GCT TCA GTT ACT CAC AAC AGG AC3' . The expected lengths of the amplicons were approximately 12,763 and 7250 bp, respectively. The total volume for each PCR amplification reaction was 25 μL, which included 5 × GLX PCR buffer 5 μL (Takara Bio Company, Dalian, China), 200 μmol/L dNTP concentration, 0.2 μmol/L primer concentration, 0.625 U GLX Taq enzyme (Takara Bio Company) and 2.5 μL DNA sample. Amplification was performed on an ABI PCR 9700 instrument (Applied Biosystems). The following conditions were used for PCR amplification: predenaturation at 94 ℃ for 1 min, denaturation at 98℃ for 10 s, annealing at 68℃ for 10 min, 30 cycles, and extension at 68℃ for 10 min. The amplicons were digested with Tn5 transposase, and the index was added to construct the library using the Trans NGS Tn5 DNA library prep kit for Illumina (Transgene, Beijing, China). All procedures were carried out strictly according to the manufacturer's instructions. Following the qualification of the library's quality, the sequences were detected on an Illumina MiSeq Sequencer using the MiSeq sequencing reagent kit (V2, 300 cycles, Illumina Inc., San Diego, CA, USA). The sequencing data were analysed using the ABO reference sequence (GenBank ID number NG_006669.2 for genomic, NM_020469.2 for transcript) and CLC main workbench 12.0 software (Qiagen company, Hilden, Germany), and all polymorphism nucleotides were recorded and analysed.

Analysis of the ABO gene haplotype
Allele-specific primer amplification sequencing or cloning technology was used to haplotype the ABO gene [9,13]. For allele-specific primer amplification (specimen ID numbers 4 to 9), specific primers for the A, B, and O alleles were used to amplify the corresponding alleles, and the amplicons were then sequenced and analysed as previously reported [9,13]. For cloning technology (specimen ID numbers 1 to 3), the PCR-SBT amplicon was ligated with the pCR4 @ TOPO plasmid vector according to the manufacturer's instructions and transfected into competent cells to grow. As previously described, plasmid DNA was extracted as a template for sequencing analysis [9,13].

In silico splicing transcript analysis
Alamut ® software v2.10 was used in conjunction with four splice site prediction tools, SpliceSite Finder-like, MaxEntScan, NNSPLICE, and GeneSplicer, to predict the effects of these splice site variations (www. inter active-bioso ftware. com/ doc/ alamut-visual/ 2.6/ splic ing. html) [16][17][18]. The Berkeley Drosophila Genome Project Searches Splice Site Prediction software and Net-Gene2-2.42 software were also used to predict potential splicing transcripts [19,20]. A splice site score calculator was used to assess the strength of the constitutive and cryptic acceptor splicing sites. A ≥ 10% change in the splice site signal in at least two algorithms was considered to have an effect on splicing [21].

The ABO subtypes' phenotypes
Nine Chinese individuals with ABO blood group typing inconsistencies were studied. Four were blood donors, while the rest were patients. Table 1 shows the agglutination reaction states of these individuals' RBCs with anti-A, anti-A 1 , anti-B, anti-AB, anti-H, and serum containing known A, B, O group RBCs. All individual RBCs exhibited 3 + or 4 + strength agglutination with anti-H. In the absorption and elution test, B antigen was expressed in ID numbers 2, 3, and 7 specimens, whereas A antigen was positive in ID numbers 8 and 9 specimens. Individuals with ID numbers 4, 5, and 6 had mixed-field agglutination. These individuals were classified as ABO subtypes based on serological characteristics (Table 1), with 3 individuals belonging to subtype A and 6 individuals belonging to subtype B.

Analysis of the ABO gene's CDS region
All nucleotides for full exons of the ABO gene in nine individuals were sequenced, but no variation was observed in the CDS regions. Table 1 shows the ABO genotypes of the individuals based on the sequences of all exons. Further sequence analysis of exon and intron splicing acceptor/donor sites in the nine individuals showed six distinct heterozygotes at positions c.28 + 1_2, c.28 + 5, c.155 + 5, c.204-1, and c.374 + 5 (Additional file 1: Figure S1).

ABO gene sequence analysis using NGS
Additional file 2: Table S1 lists 166 polymorphic nucleotides of these ABO subtypes (ID numbers 3 to 8) compared to ABO gene sequences in the GenBank database (NG_006669.2 for genomic, NM_020469.2 for transcript) (sequence between start codon and stop codon in the ABO gene). The ID numbers 1, 2, and 9 specimens were not analysed using the NGS method. Except for the splicing acceptor/donor sites, which were consistent with the PCR-SBT, no variation was observed in the NGS method.

In silico predictions to assess the functional implications of splice site variations
In the in silico analysis using the Berkeley Drosophila Genome Project Searches Splice Site Prediction software, several new splice sites were predicted in the c.155 + 5G > A, c.204-1G > A, and c.374 + 5G > A variants. Some new donor sites were predicted at c.155 + 507 (acataagGTaggagg) with a score of 0.95 and c.374 + 840 (ctccttaGTaagagg) with a score of 0.51. One of the new acceptor sites was predicted to be located at position c.204-224 (ctcttgccAGtttgtaag) with a score of 0.84. However, some additional spliceosomes resulting from variations in the splice sites might generate partial functional transferases as a result of the RBCs of the probands.

Discussion
The common ABO subtypes are A 3 , A x , A el , B 3 , B x , B el , B m , B(A), cisAB, etc. Seltsam A et al. reported that the B W phenotype is caused by variations in the CCAATbinding factor/NF-Y enhancer region of the ABO gene [22]. Sano R demonstrated for the first time that deletion of the ABO gene's erythroid cell-specific regulatory element could downregulate transcription in the B(m) allele [23]. Numerous ABO variations have been identified in individuals with ABO subtypes to date [24][25][26][27][28][29]. These variants of the ABO gene are located in the CDS region, intron 1 erythroid-specific regulatory element region, splice site, promoter, cis-or trans-regulatory element, etc. Kronstein-Wiedemann R et al. found that miR-331-3p and miR-1908-5p directly target the mRNA of GTA and GTB and that overexpression of these miR-NAs in haematopoietic stem cells may result in a significant reduction in the expression of A antigens [30]. Some variations in the splice sites of the ABO gene are associated with some ABO subtypes. Chen DP et al. reported c.155 + 5G > A (IVS3 + 5G > A) in a B3 individual and c.374 + 5G > A (IVS6 + 5G > A) in an Ael individual31, 32. In theory, changes in the ABO gene splice site result in the formation of new RNA splice sites and therefore novel versions of ABO mRNA.
The Chinese population has a high prevalence of ABO subtypes [13,14]. In our research, we routinely analysed the ABO gene's full CDS and the sequence of the erythroid cell-specific regulatory element region for ABO subtypes using the PCR-SBT technique. We discovered over 50 novel alleles from ABO subtypes [9,13,33]. In this study, six distinct splicing site variants in the ABO gene were identified in nine individuals with ABO subtypes. Between 2015 and 2019, our laboratory screened and obtained specimens from 369 individuals with suspected ABO subtypes using a combination of serological and molecular methods.
Multiple ABO mRNA forms were detected in the normal ABO phenotype by RT-PCR, the majority of which lacked exon 6 [34,35]. However, in some individuals with ABO subtypes, RNA splicing of the ABO gene was detected [31,32,36,37]. An ABO* A1-like allele with a 4 bp deletion (c.236-239delCGTG) in exon 5 and a 20 bp downstream deletion in intron 5 affected the donor splice site [36]. c.28G > A in exon 1 is associated with the weak B subtype via its effect on the ABO gene's RNA splicing [37]. Previously, c.155 + 5G > A was discovered in B3 individuals [31]. At least 7 distinct types of splicing transcripts were identified in B3 individuals [31]. While it is possible to generate a mRNA without the matching exon