B4GALNT2 and B4GALNT2-like modifications were introduced into a GGTA1 KO background expressing hCD55 to generate quadruple-modified (4GM) pigs by using the CRISPR/Cas9 system. GGTA1 KO targeted (Tg)-hCD55 fetal fibroblast cells (2×10⁵ cells) were co-transfected via electroporation with gRNA/Cas9 plasmids targeting both B4GALNT2 and B4GALNT2-like genes (gRNA oligonucleotide 1 forward: 5’ - CACCGTGAGGATCGACAGACATCTA −3’, reverse: 5’- AAACTAGATGTCTGTCGATCCTCAC- 3’; gRNA oligonucleotide 2 forward [7]: 5’ CACCGTGTATCGAGGAACACGCTT-3’, reverse: 5’- AAACAAGCGTGTTCCTCGATACAC-3’). Two hundred cells were seeded per 100 mm dish for colony formation (~200 cells/plate). Following culture for 10 days, individual colonies were collected for genotyping (1/3 of the cells) and the remaining (2/3) were expanded to be cryopreserved for long-term preservation [17].

Primers were designed to amplify homologous regions between B4GALNT2 and B4GALNT2-like. These assays encompassed unique nucleotide signatures, allowing us to identify all four alleles. PCR amplicons were cloned and Sanger sequenced, and colonies showing bi-allelic modifications for B4GALNT2 and B4GALNT2-like were used for somatic cell nuclear transfer (SCNT) [18]. Six embryo transfer were performed, resulting in two pregnant recipient gilts that gave rise to female and male cloned founders (NSRRC:0086 line). For phenotypic analysis, liver and kidney sections from NSRRC:0086 and a wild-type (WT) pig were stained with isolectin B4 (IB4)-FITC (Sigma-Aldrich), which binds to Gal residues, and rhodamine labeled Dolichos biflorus agglutinin (DBA) (Vector Laboratories), which recognizes GalNAc epitopes as presented in the Sd(a) glycan. PCR diagnostic assays to confirm GGTA1 homozygous deletion and hCD55 integration were performed (data not shown). Additionally, we isolated primary porcine kidney (PKC) cells as previously described [19] from two of the euthanized NSRRC:0086 piglets and from a WT piglet to confirm the results observed in the ex vivo experiment.

A breeding colony was established by crossing NSRRC:0086 founder (F0) clones with B4GALNT2/B4GALNT2-like KO pigs (R614 line, produced via zygote injection) and non-cloned GGTA1 KO/hCD55 pigs. Genotyping and phenotyping confirmed successful transmission of genetic modifications to F1 progeny. To generate GGTA1/B4GALNT2/B4GALNT2-like KO piglets, F1 individuals were bred with each other and with F0 founders. Subsequently, F2 pigs with the desired modifications were bred with WT pigs to enhance genetic diversity.

Plasmids expressing the most common transcript of B4GALNT2 (Accession No: NM_001244330) or B4GALNT2-like (Accession No: XM_021067117.1) were generated commercially (GenScript) by inserting synthesized coding sequence between the XhoI/XbaI restriction sites of pcDNA3.1/Hygro (+) (GenScript, SC1317).

To evaluate the potential involvement of B4GALNT2-like in the synthesis of the Sd(a) antigen, cells were transfected with plasmids expressing either B4GALNT2 or B4GALNT2-like. WT and the NSRRC:0086 line, which did not stain for DBA, served as negative control. WT porcine kidney cells were used as a positive control, as DBA binding had been successfully confirmed previously.

Porcine tail, kidney, and fetal fibroblasts from WT and NSRRC:0086 lines, isolated as previously described [19,20,21], were cultured in Dulbecco Modified Eagle medium (DMEM; Gibco, Cat #11885084) supplemented with 0.5% Glutamax (Gibco, Cat #35050061) and 12% FBS (Gibco, Grand Island, NY). These cells were incubated at 38.5°C in a humidified atmosphere of 5% O₂ and 5% CO₂ and transfected with 2 μg of B4GALNT2 or B4GALNT2-like pcDNA3.1/Hygro (+)-expressing plasmid. WT and NSRRC:0086 tail and fetal fibroblast cells were transfected via electroporation [17], while WT and NSRRC:0086 PKC were transfected by using Jet-OPTIMUS transfection reagent (Polyplus #101000025) and seeded in a 6 well-plate. Non-transfected WT fibroblast cells and NSRRC:0086 line served as negative controls. WT PKC were used as positive controls. Forty-eight hours post-transfection, the cells were incubated in the dark with DBA (1:400; Vector Laboratories) or PBS for 20 minutes at 4°C.

Uterine cross sections from WT and B4GALNT2/B4GALNT2-like KO (derived from NSRRC:0086 × R614 breeding) females were fixed in 10% buffered formalin for 24–48 h and paraffin embedded using standard procedures. For fluorescence localization, sections were deparaffinized in xylene and rehydrated to water through a graded alcohol series. Sections were submerged into citrate solution and then blocked in SuperBlock^™ (Thermo Scientific) solution. Sections were incubated with DBA lectin–FITC conjugate (30 μg/mL) in PBS solution overnight at 4C. Sections were lightly stained with hematoxylin and overlaid with a coverslip and antifade mounting reagent (VectaSheild, Vector Co, USA).

Cells and slides were examined under a fluorescence stereo microscope (M165FC, Leica Microsystems AG, Heerbrugg, Switzerland) to confirm binding to Rhodamine or FITC-labeled Dolichos biflorus agglutinin (DBA) lectin. All digital fluorescence and brightfield images were recorded by using the Leica LAS X Software.

By using the annotated pig genome assembly Sus scrofa 11.1 from the NCBI database, we compared the B4GALNT2 and B4GALNT2-like regions across four species. We identified conserved flanking sequences, which corresponded to the IGF2BP1-ZNF652 interval, and then searched for this region in the human (GRCh38.p14), mouse (GRCm39), and bovine (ARS-UCD2.0) genomes. Genomic view images from the NCBI database were included in the Supplementary Data (Figure S4–S7).

To determine the conservation of the B4GALNT2 gene across species, nucleotide and protein sequences for porcine, human, bovine, and murine orthologs were retrieved from the NCBI RefSeq database. Multiple sequence alignment (MSA) was performed using Clustal Omega (v1.2.4) via the EMBL-EBI web portal using default parameters. Percent identity scores were subsequently calculated based on the resulting pairwise identity matrix to quantify the genetic relatedness between the porcine B4GALNT2 isoforms and their mammalian counterparts.

PCR diagnostic assays and Sanger sequencing confirmed the absence of WT alleles in the genetically modified lines (Figure S1). Consistent with these genotypic results, no fluorescence signal was detected when organs or primary kidney cells (PKCs) from founder pigs were stained with Rhodamine-labeled DBA lectin to detect B4GALNT2 and B4GALNT2-like activity (Figure S2), or with FITC-conjugated isolectin B4 to detect GGTA1 expression Additionally, uterine tissues from a B4GALNT2/B4GALNT2-like null female did not stain for DBA when compared to a WT female (Figure S3).

Male and female founder cloned lines (NSRRC:0086) were bred; however, they were unable to establish pregnancies. Thus, we developed a breeding plan by using different combinations of genotypes, as shown in Figure 1. We produced 13 independent pregnancies that reached full term, resulting in viable null offspring. These B4GALNT2/B4GALNT2-like knockout piglets were derived from founder lines NSRRC:0086 and R614. The consistent birth and survival of these animals provide definitive evidence that the loss of these enzymes does not impair porcine fetal development or parturition.

As summarized in Figure 1, three specific breeding strategies generated a total of 14 B4GALNT2/B4GALNT2-like null piglets. Notably, most of these offspring also carried knockout alleles for the GGTA1 gene and the human transgene hCD55. To our knowledge this is the first report showing the creation of a domestic line of GGTA1^−/−, B4GALNT2 ^−/−, B4GALNT2-like^{− /−}, Tg-hCD55 piglets by breeding.

Edited pig cells lacking B4GALNT2 and B4GALNT2-like expression were stained with Rhodamine-DBA lectin after transfection with either pcDNA3.1/B4GALNT2 or pcDNA3.1/B4GALNT2-like expressing vectors (Figure 2). Similarly, WT tail and fetus cell lines that previously did not stain for DBA fluoresced red after transfection and Rhodamine-DBA staining (Figure 2). Non-transfected (Figure 2) or PBS-treated (data not shown) fetal and tail lines did not react with DBA lectin. Transfected and non-transfected WT PKC cells were positive for DBA staining (Figure 2).

We found that the IGF2BP1-ZNF652 interval is conserved among the four species (Figure 3A and 3C). However, in the pig genome we observed a duplication of the genes Phospho1, ABI3, GNGT2 and B4GALNT2 (Figure 3C). We propose that a chromosomal rearrangement occurred involving three distinct events: duplication, inversion, and reinsertion. Specifically, we hypothesize that this rearrangement was triggered by a double-strand break (DSB) repaired via a template-switching mechanism involving the homologous chromosome (Figures 3B and 3C).

Interestingly, the porcine tRNA region appears to be the site of reinsertion. Since the tRNA was not duplicated and now resides between the rearranged segments, we suspect that the duplicated genes replaced a portion of the original tRNA region (Figure 3C).

Multiple sequence alignment revealed that, at nucleotide level, B4GALNT2-like exhibits greater similarity to bovine, murine, and human B4GALNT2 orthologs than does the annotated porcine B4GALNT2 (Tables S1 and S2). Consequently, we hypothesize that B4GALNT2-like represents the ancestral locus, while the current porcine B4GALNT2 is the duplicated copy (Figure 3D).