Development of lamprey-derived antibodies against human blood group antigens

Abstract

A major challenge in the glycosciences is the scarcity of sensitive and specific glycan-binding reagents, such as monoclonal antibodies, for detecting and isolating glycans. Here we report the development and characterization of new monoclonal antibodies (mAbs) that bind carbohydrate-based red blood cell (RBC) antigens including the ABO(H) antigens. This approach exploits the immune system of the sea lamprey (Petromyzon marinus), which strongly responds to human glycans to enable the generation of high affinity antibodies. To develop these mAbs, we immunized the lamprey with RBCs and designed a targeted antibody enrichment and screening process using intact RBCs and a custom microarray displaying blood group antigens. Through multiple rounds of enrichment and testing we identified two mAbs; A_25 and A_39. Glycan binding analysis of the mAbs using glycan microarrays, the Luminex platform and western blot analysis revealed their binding to H antigens and terminal N-acetyllactosamine Galβ1-4GlcNAc (LacNAc, a type 2 sequence). Mechanistic insights into antigen specificity were gained through glycan inhibition assays, sequence homology analysis, and nanomolar-range affinity measurements. mAb binding to RBCs was determined using flow cytometry. Both mAbs bound RBCs of all ABO blood groups, whereas strongest binding was observed for blood group O RBCs. Our findings highlight the efficacy of the lamprey system to develop glycan-specific mAbs. These reagents allow investigation of expression of the H antigen and LacNAc-containing glycans in human tissues. In the future, they could also be modified using molecular engineering techniques to generate mAbs specific to other understudied blood group antigens.

Introduction

Glycans play crucial roles in biological processes, including cell–cell, cell-matrix, and cell-molecular ligand interactions, making them essential for cellular communication, migration, and adhesion. Additionally, they contribute to protein folding and structural stability (Fuster and Esko 2005; Gagneux et al. 2022; Stanley et al. 2022). In line with their physiological functions, glycans play an important role in the field of hematology. The ABO(H), Lewis, and I blood group systems are three well-characterized glycan-based blood groups. Of these, the ABO(H) antigens are the most extensively studied (Jajosky et al. 2023). Each individual has a specific ABO blood group i.e. A, B, AB, or O, with the rare exception of humans with the Bombay phenotype (Scharberg et al. 2016). The different blood groups are determined by the presence or absence of glycosyltransferases that modify a core glycan, the H antigen (Fucα1–2Gal). Two enzymes can form the H antigen: fucosyltransferase 1 (FUT1), which fucosylates type 2 chain Galβ1-4GlcNAcβ-R (LacNAc) to generate H-type 2 antigen, and fucosyltransferase 2 (FUT2), the secretor enzyme, which fucosylates type 1 chain Galβ1–3GlcNAcβ-R to generate H-type 1 antigen. The A-transferase adds N-acetylgalactosamine (GalNAc) to the either of these H antigens, forming the A antigen (GalNAcα1–3(Fucα1–2)Galβ-R), and the B-transferase attaches a Gal to the H antigen, forming the B antigen (Galα1–3(Fucα1–2)Galβ-R). The H antigen of blood group O individuals remains unmodified, as these individuals express non-functional A- and B-transferases (Jajosky et al. 2023; Yamamoto et al. 1990). Humans with blood group AB express both the A- and B-transferase and carry A and B antigens on their red blood cells (RBCs). Since A- and B-transferases are not fully efficient, RBCs of all ABO blood groups express H antigens on their surface (Curtis et al. 2000; O'Donnell et al. 2002). The ABO(H) antigens on type 2 LacNAc chains (Galβ1-4GlcNAc-R) are present on RBC glycosphingolipids and glycoproteins (Hakomori and Siddiqui 1974; Finne et al. 1978; Karhi and Gahmberg 1980; Morgan and Watkins 2000).

In addition to being expressed on RBCs, the ABO antigens are also expressed in tissues. Recently, we reported the expression of H antigens in human tissues and found they are abundantly expressed in the pancreas, stomach, and kidney among other tissues (Tulin et al. 2025). Matching ABO blood group antigens between donors and recipients is critical to prevent hyperacute rejection in organ transplantation. For example, ABO incompatible transplants have resulted in graft rejection after kidney transplantation (Williams et al. 1968; Böhmig et al. 2015). Thus, in addition to knowing ABO antigen expression in RBCs, mapping their precise expression in tissues, identifying pathogenic subtypes, and carrier proteins are important for transplantation medicine, all of which are currently understudied.

Conventional reagents used in the clinic to study ABO antigens include gel cards containing murine anti-A and anti-B antigen IgM antibodies (for example, Grifols, Cat# 210384). Although the pentameric structure of IgMs (high avidity) favors their use in diagnostics, their antigen affinity is relatively low (Goulet and Atkins 2020). Therefore, these high avidity, low affinity reagents are not suitable to precisely study the expression of the blood group antigens in more nuanced scenarios such as tissues. Historically, plant lectins such as Ulex europaeus agglutinin I (UEA-I), Anguilla anguilla agglutinin (AAA), Lotus tetragonolobus agglutinin (LTA), and Dolichos biflorus agglutinin (DBA) have been used to study H antigen expression. Using these lectins, Watkins and Morgan identified the ABO(H) blood group antigens as glycans and determined their structure (Watkins and Morgan 1952; Morgan and Watkins 2000; Cummings 2009). However, these reagents are unable to accurately distinguish between various ABO antigens because the minimum epitope requirements for binding are usually monosaccharides, such as L-fucose in the case of LTA or α-galactose in the case of DBA, and thus can bind to several structures that may not be ABO antigens (Bojar et al. 2022). To improve this situation, we aimed to develop monoclonal antibodies (mAbs) targeting blood group carbohydrate antigens.

We have previously reported the use of the sea lamprey (Petromyzon marinus) to develop specific anti-glycan mAbs through our smart anti-glycan reagent pipeline (SAGR) (McKitrick et al. 2020). In this approach, immunization of sea lamprey with human cells leads to the expression of glycan-specific variable lymphocyte receptor B (VLRB). These VLRBs are then expressed as yeast surface display (YSD) library which allows for enrichment of specific yeast clones that express a VLRB that binds to the target glycan (Fig. 1). Recently, we reported two new mAbs that bind to various forms of the H antigen and mapped the expression of H antigens in several human tissues and tumors (Tulin et al. 2025). Here, we aimed to identify new anti-ABO antibodies and designed a targeted version of the SAGR method specific to RBCs. We report two new anti-glycan mAbs; A_25 and A_39, which bind to carbohydrate antigens on RBCs. Using this optimized approach and new set of antibodies, we plan to expand the repertoire of tools dedicated to studying ABO(H) carbohydrate antigens.

Fig. 1.

Pipeline for the development of novel monoclonal antibodies binding carbohydrate-based blood group antigens. This procedure makes use of the unique immune system of the sea lamprey (Petromyzon marinus) and includes four main stages: (1) immunization of sea lamprey with red blood cells and generation of yeast surface display (YSD) library, (2) enrichment of YSD library for yeast expressing variable lymphocyte receptor B (VLRB) against blood group antigens, (3) discovery of yeast clones that bind to blood group antigens, and (4) expression and characterization of glycan-binding monoclonal antibodies. Created in BioRender. Cummings, R. (2025) https://BioRender.com/7lmtgfz.

Results

Magnetic sorting enriched for VLRB-yeast clones that bind to red blood cells

To identify antibodies that bind to carbohydrate blood group antigens, we investigated antibodies generated by sea lampreys after immunization with RBCs. Previously, we reported the generation of YSD libraries presenting antibodies derived from lampreys that were immunized with RBCs from blood group A donors (Type A YSD library), blood group B donors (Type B YSD library), and blood group AB donors (Type AB YSD library) (McKitrick et al. 2020). We wanted to enrich these libraries for VLRB-expressing yeast clones that bind to RBCs using magnetic sorting on biotinylated RBCs (Fig. 2). We confirmed the presence of ABO(H) carbohydrate antigens on RBCs using a hemagglutination assay (Fig. S1A) with a panel of lectins and a monoclonal antibody to the H antigen. RBCs identified to be blood group A or B were biotinylated using N-hydroxysuccinimide (NHS)-Biotin (Fig. S1B and C). The YSD libraries were expanded and VLRB expression was induced by addition of galactose. Successful induction of VLRB expression was assessed using an anti-Myc tag AlexaFluor488-labeled antibody and flow cytometry (Fig. S2). All three YSD libraries were MACS enriched twice using biotinylated RBCs. The type A YSD library was enriched with blood group A RBCs while Type B and AB YSD libraries were enriched with blood group B RBCs. After each enrichment step, the percentage of yeast expressing VLRBs was determined. MACS-enrichments of the Type A YSD library led to an increase in the percentage of yeast expressing VLRBs from 43.7% to 57.1% (Fig. 2A), whereas only a minor increase in yeast expressing VLRBs was determined in Type B and Type AB YSD libraries (Fig. 2B and C). We then assessed if these enriched libraries would bind to glycans on a microarray (Fig. 2D–F) and found that only the MACS-enriched Type A YSD library (Fig. 2D) bound to the array. Glycan binding of unenriched YSD libraries were undetectable by the microarray scanner. As a control of positive binding, we also ran the OmcFL library (Fig. 2G), which was previously reported to carry VLRB-expressing yeast clones that bind to the blood group antigen H (Tulin et al. 2025). Based on these results, we focused our work on the enriched Type A YSD library.

Fig. 2.

Magnetic sorting enriched for VLRB-yeast clones that bind to red blood cells. A–C) scatter plots showing the progression of yeast expressing VLRBs over two rounds of MACS enrichment for the A) type A, B) type B, and C) type AB YSD libraries. The percentage of yeast expressing VLRBs (“induced yeast”) is indicated in each plot. D–G) scans of glycan microarray containing ABO(H) blood group antigens incubated with the 2x MACS-enriched D) type A, E) type B, and F) type AB YSD libraries, G) as well as a control YSD library containing H antigen-binding yeast, referred to as the OmcFL YSD library (Tulin et al. 2025). D, G) Type A and OmcFL YSD libraries bound antigens printed on the glycan microarray in quadruplicate (four dots in a row), framed by dashed rectangular boxes.

Glycan microarrays further enriched for VLRB-yeast clones that bind to major ABO(H) blood group antigens

As our goal was to find novel anti-glycan antibodies, we further enriched for VLRB-yeast clones that bind to blood group antigens. After enriching for VLRB-yeast clones that bind to RBCs, we wanted to decrease library diversity and focus on clones that bind to glycan targets. The YSD library was incubated on the ABO blood group glycan microarray slide, washed thoroughly, and replica plated on a solid media plate. This procedure allowed to isolate yeast which express VLRBs that bind ABO(H) blood group antigens. The additional enrichment steps resulted in a major increase in the percentage of yeast expressing VLRBs, yielding an enriched YSD library with 71.9% of yeast expressing VLRBs (Fig. 3A). Glycan binding was assessed by screening on the ABO blood group glycan microarray. Results showed that the RBC/glycan-enriched Type A YSD library bound two major blood group antigens, the H antigen and the A antigen (Fig. 3B). Binding of Type A YSD library was compared to the OmcFL library which showed a different binding pattern, supporting the possible identification of novel antibodies from this newly enriched Type A YSD library.

Fig. 3.

Glycan microarrays further enriched for VLRB-yeast clones that bind to major ABO(H) blood group antigens. A) MACS and glycan microarray enrichment steps increased the percentage of yeast expressing VLRBs (scatter plots) and resulted in two distinct populations of yeast, one expressing and the other not expressing VLRBs (histograms). B) Glycan binding of type A (see legend above) and the control OmcFL YSD library (Tulin et al. 2025) on glycan microarray containing the ABO(H) blood group antigens showing binding to blood group antigens H, A, and Lewis X. Each bar represents the relative fluorescence units (RFU) for a specific glycan. Error bars represent standard deviation from the quadruplicate measurements.

Discovery of two novel VLRBs that bind to ABO(H) blood group antigens

The enriched Type A YSD library was serially diluted, plated, and individual yeast clones were sequenced to identify monoclonal yeast that contain a novel VLRB DNA sequence. A total of 48 candidate clones were screened, 25 colonies contained a VLRB DNA, and 13 VLRB DNA sequences were unique (Fig. S3). Unique VLRB DNA sequences varied in size, between 600 and 800 bp, and amino acid sequence homology (Figs. S3 and S4).

To identify monoclonal yeast expressing VLRBs specific to blood group antigens, all 13 yeast clones with unique VLRB DNA sequences were screened on glycan microarrays containing the ABO(H) blood group antigens. Out of 13 yeast clones, two bound blood group antigens, namely clone A_25 and clone A_39 (Fig. 4). We compared the sequences with OmcFL3–02, a mAb previously reported to bind specifically to the H antigen (Fucα1–2Gal) (Tulin et al. 2025), and found high sequence homology in the C-terminal region, and differences mainly in the N-terminal region (Fig. 4A). Of interest were the amino acid motifs highlighted in red; aspartic acid (D), glutamine (Q), and tryptophan (W), which were present in all three VLRBs. Collins et al. have shown that these DQ and W motifs are essential for the interaction of a VLRB with the H-antigen. They further demonstrated that VLRB sequence differences in the N-terminal part of H antigen binding VLRBs are responsible for recognizing LNnT (Lacto-N-neotetraose) (Collins et al. 2017). This suggests that yeast clones A_25 and A_39 bind to H antigens but possibly also to other glycan structures. Indeed, clones A_25 and A_39 strongly interacted with H antigens but also with the Lewis Y antigen, A antigen and B antigen, which was not previously observed for OmcFL3–02 (Fig. 4B).

Fig. 4.

Novel VLRB-carrying yeast clones bind to blood group antigens. A) VLRB polypeptide sequence alignment of yeast clones a_25, a_39, and the H antigen-binding VLRB-rIgG OmcFL3–02.(13) amino acid differences are marked in bold black, and residues reported to be essential for H antigen binding are marked in bold red (Collins et al. 2017). B, C) comparison of the glycan binding profiles of yeast clones A_25 and A_39 (see legend above) on B) a glycan microarray containing the ABO(H) blood group antigens and C) an NCFG glycan microarray. Binding to ABO(H) blood group antigens, Lewis Y antigen, and terminal N-acetyllactosamine (LacNAc) on poly-N-acetyllactosamine (PolyLacNAc) chains was observed. Each bar represents the relative fluorescence units (RFU) for a specific glycan. Error bars represent the standard deviation from the quadruplicate measurements.

Furthermore, clones A_25 and A_39 were screened on an NCFG glycan microarray containing a collection of glycans, including the ABO(H) blood group antigens, N-glycans, O-glycans, and LNnT (Fig. 4C). Clone A_25 consistently showed strong binding to H antigens and Lewis Y with weaker binding to A antigens, B antigens, and LacNAc. Importantly, no binding was observed to the glycans containing the type 1 LacNAc sequence Galβ1–3GlcNAc. Interestingly, clone A_39 not only interacted strongly with H antigens and Lewis Y antigen, but also with PolyLacNAc chains with terminal LacNAcs. Taken together, clones A_25 and A_39 express VLRBs with unique sequences that bind to carbohydrate blood group antigens such as the ABO(H) antigens, LacNAc and the Lewis structures.

Novel anti-glycan monoclonal antibodies bind to ABO(H) blood group antigens and terminal N-acetyllactosamine

After the successful isolation and characterization of two yeast clones that bind blood group antigens, their VLRB DNA sequence was used to generate monoclonal VLRB-rIgG chimeric proteins and screened on various glycan microarrays. The VLRB DNA was cloned into a pcDNA expression plasmid containing the rat IgG Fc with a N183A mutation at the N-glycosylation site. The VLRB-rIgG chimeric proteins that are non-glycosylated were expressed by transiently transfecting ExpiCHO cells and purified from media using a HisPur™ Ni-NTA resin. Purified antibodies were analyzed using both Coomassie staining and western blot under reducing and non-reducing conditions, confirming that they were > 98% pure (Fig. S5).

The newly developed and produced mAbs, A_25 and A_39, were screened on three different glycan microarrays for a comprehensive analysis of their glycan binding profile their apparent affinities were determined using the Luminex bead platform (Figs 5–7). First, the two mAbs were screened on the ABO(H) blood group glycan microarray (Fig. 5A and B). Both antibodies interacted with H antigens and weakly with A and Lewis Y antigens (Fig. 5C). Additionally, the glycan-binding specificity of the monoclonal antibodies developed from the YSD clone were demonstrated to be equivalent to those of the yeast-displayed antibody fragment (Fig. S6A). This equivalence confirms that the functional characteristics of the antibody are preserved during coupling to rFc, expression in CHO cells, secretion and purification, validating the YSD screening process as a reliable predictor of soluble antibody behavior.

Fig. 5.

The newly developed monoclonal antibodies (mAbs) bound blood group antigens and terminal N-acetyllactosamine (LacNAc)-containing glycans. A, B) scans of glycan microarray containing ABO(H) blood group antigens showing binding of A) a_25 mAb and B) a_39 mAb to the antigens in quadruplicate. C) Direct comparison of the glycan binding profiles of the mAbs A_25 and A_39 (see legend above) on the ABO(H) blood group antigen microarray showing binding to H antigens, A antigens, and to the Lewis Y antigen. D, E) scans of the NCFG glycan microarray showing binding of D) A_25 mAb and E) A_39 mAb to glycan antigens. F) Direct comparison of the glycan binding profiles of the mAbs A_25 and A_39 (see legend above) showing binding to H antigens, Lewis Y antigen, and to terminal LacNAc on PolyLacNAc chains and complex N-glycans. Each bar represents the relative fluorescence units (RFU) for a specific glycan. Error bars represent standard deviation from the quadruplicate measurements.

Fig. 7.

Luminex beads-based glycan binding analysis revealed binding affinities of monoclonal antibodies to H antigens and lacto-N-neotretraose (LNnT) in the low nM-range. A, B) the binding of A) A_25 mAb and B) A_39 mAb to 25 different glycans was assessed using a Luminex bead-based glycan array. C, D) apparent affinity measurement of C) A_25 mAb and D) A_39 mAb showing apparent binding affinities in the low nM-range. Each bar or point represents the average net median fluorescence intensity (Avg net MFI) for a specific glycan. Error bars represent standard deviation from the 100 beads measurement (trimmed standard deviation). These standard deviation data were proceeded by the FLEXMAP 3D instrument. Legend: MTZ = linker (control), A2 = A type 2, H2 = H type 2, H5 = H type 5 (or 2’-FL), LNnT = lacto-N-neotetraose.

Next, the mAbs were screened on an NCFG glycan microarray slide, which revealed differences in glycan binding between A_25 mAb and A_39 mAb (Fig. 5D and E). A_25 mAb primarily bound to the H antigen and showed weak binding to the Lewis Y antigen while A_39 mAb bound the H antigen, Lewis Y antigen, and also terminal LacNAc on PolyLacNAc chains and complex N-glycans (Fig. 5F). Consistent with the ABO(H) blood group antigen microarray results described above, binding of the newly developed mAbs and of the yeast clones were also virtually identical on the NCFG glycan microarray (Fig. S6B).

The VLRB-rIgG chimeras were then screened on a CFG glycan microarray containing over 500 different mammalian-type N- and O-glycans for comprehensive epitope mapping of their glycan binding properties. Again, there were differences in glycan binding between the mAbs A_25 and A_39. Both newly developed mAbs bound to H antigens, 6-O sulfated H antigens, and complex N-glycans with terminal H antigens (Fig. 6A and B). In addition, A_39 mAb bound to several structures with terminal LacNAc on complex N-glycans or PolyLacNAc chains (Fig. 6B). To further test the potential of A_39 mAb to bind LacNAc sequences, we then also performed western blotting on two Chinese hamster ovary (CHO) cell lines; Pro5 and Lec1 (Fig. 6C, Fig. S7). Pro5 CHO cells carry complex N-glycans with terminal LacNAc that may be sialylated, whereas Lec1 cells predominantly expresses oligomannose glycans (Man5) due to a mutation in the Mgat1 gene resulting in the loss of GlcNAcT-I activity (North et al. 2010). Immunoblotting on Pro5 CHO cell lysates showed positive reactivity, whereas binding to Lec1 CHO cell lysates was weak. Furthermore, treatment with neuraminidase which cleaves terminal sialic acid, exposing LacNAc extensions, led to a significant increase in binding to glycoproteins from Pro5-CHO cells; this binding was largely to N-glycans, as treatment with PNGase F causes a loss of binding. These results indicate that A_39 mAb binds to complex N-glycans with terminal LacNAc extensions. Residual binding to Lec1 CHO cells is probably due to a small contamination of serum glycoproteins from the FBS culture media, which carry complex N-glycans.

Fig. 6.

CFG glycan microarray and western blotting revealed binding to different versions of the H antigen and terminal N-acetyllactosamine (LacNAc) containing glycans. (A, B) glycan binding profiles of A) a_25 mAb and B) a_39 mAb on CFG glycan microarray. Both mAbs bound to H antigens (grey), 6-O sulfated H antigens (yellow), and terminal H antigens presented on complex N-glycans (green). A_39 mAb additionally interacted with terminal LacNAc on complex N-glycans (purple) and PolyLacNAc chains (red). Each bar represents the relative fluorescence units (RFU) for a specific glycan. Error bars represent standard deviation from the quadruplicate measurement. C) Western blotting of Chinese hamster ovary (CHO) cell mutants with monoclonal antibody A_39 with and without glycosidase treatment. Pro5 CHO cells (similar to wildtype glycosylation) and Lec1 CHO cells (unable to produce extended N-glycans) were treated with NeuA, a sialidase which cleaves α2–3,6,8,9 sialic acid, and PNGase F which cleaves N-glycans. Binding to Pro5 CHO cells increased after NeuA treatment and decreased or no binding was observed in Lec1 and PNGase F treated Pro5 cells. Total protein was used as loading control and the experiment was performed in triplicate.

Lastly, binding of the mAbs to the Luminex beads derivatized with ABO(H) antigens was assessed to determine apparent binding affinities (K_d.app) to different glycan epitopes (Fig. 7,  Fig. S8). Binding of mAbs were specific and consistent with the results from the ABO(H) blood group antigen microarray (Fig. 7A and B). A_25 mAb bound to H type II (H2) with a K_d.app of 3.05 nM and to H type 5 (H5) with a K_d.app of 5.28 nM (Fig. 7C). Furthermore, A_39 mAb bound to H2 with a higher affinity, K_d.app of 2.38 nM, compared to lacto-N-neotetraose (LNnT) with a K_d.app of 5.29 nM (Fig. 7D).

Taken together, glycan binding assays based on glycan microarrays and the Luminex platform showed that the two newly developed mAbs differ in their binding specificities. Whereas A_39 showed strong binding to poly-LacNAc and complex-type N-glycans, both A_39 and A_25 bound to carbohydrate blood group antigens such as different versions of the H antigen, the A antigen, Lewis Y antigens, and terminal LacNAc.

2’-Fucosyllactose inhibits binding of A_39 mAb to terminal N-acetyllactosamine residues

To further define the unusual ability of these antibodies to bind glycans with different modifications of LacNAc termini, e.g. lacking modification or having H or Lewis antigens, we further analyzed their binding in presence of specific small oligosaccharide inhibitors. These assays were aimed at providing insight on whether the mAbs might have two distinct antigen-binding sites within their antigen-binding domain, or a single binding site with varying affinities for different antigens, and to identify the minimal glycan motif required for binding. The mAbs were incubated on an NCFG glycan microarray with three different glycans: 2′-FL (2′-Fucosyllactose; Fucα1–2Galβ1-4Glc), Lac (Lactose; Galβ1-4Glc), and LacNAc (N-acetyllactosamine; Galβ1-4GlcNAc). None of the three glycans had any effect on A_25 mAb binding to the NCFG glycan microarray (Fig. S9). By contrast, A_39 mAb binding to terminal LacNAc on PolyLacNAc chains and complex N-glycans, and to the Lewis Y antigen, but not the H antigen, was strongly inhibited by 2’-FL (Fig. 8A). Neither Lac nor LacNAc had any inhibitory effect on the glycan binding of A_39 mAb (Fig. 8B and C). This unusual and specific inhibition to some glycans and not others suggests that A_39 mAb contains two binding sites, one that recognizes the H type II trisaccharide for binding and is not inhibited by any of the competing sugars, and another that accommodates both extended LacNAc chains and the H antigen (Fucα1–2Gal), with stronger affinity for the H-antigen and which explains the competitive inhibition by 2’-FL.

Fig. 8.

2’-Fucosyllactose inhibited binding of A_39 mAb to terminal N-acetyllactosamine. A–C) Comparison of A_39 mAb binding on the NCFG glycan microarray in the absence and presence (see legend above) of A) 2′-fucosyllactose (2’-FL), B) lactose (lac), and C) N-acetyllactosamine (LacNAc). Each bar represents the relative fluorescence units (RFU) for a specific glycan. Error bars represent standard deviation from the quadruplicate measurements.

Novel anti-glycan monoclonal antibodies bind to red blood cells of all blood types

Finally, to assess the potential clinical applications of these novel mAbs as well as the validity of our approach to generate antibodies that bind to RBCs via carbohydrate antigens, we tested these mAb binding to RBCs from individuals of four ABO blood groups (O, A, B, and AB) using flow cytometry. A fluorescently labeled secondary antibody was used to detect the binding of the mAbs A_25 and A_39 to RBCs. While all RBCs were typed by the blood bank center, we additionally performed qualitative assessment of ABO(H) blood group antigens using plant lectins, namely UEA-I (binding the H antigen) and DBA (binding the A antigen) (Watanabe and Gipson 1994). UEA-I strongly bound to blood group O RBCs and weakly interacted with RBCs from other ABO blood groups (Fig. 9), indicating the presence of the H antigen on RBCs of all ABO blood groups, but foremost on blood group O RBCs. DBA bound to blood group A RBCs, more weakly to blood group AB RBCs, and did not bind to RBCs of either blood group O or B (Fig. 9), indicating the presence of A antigens on blood group A and AB RBCs.

Fig. 9.

The newly developed monoclonal antibodies (mAbs) bound red blood cells (RBCs) from all ABO blood groups. A) Flow cytometry analysis of RBC staining across all four ABO blood groups. Histograms (half-offset) represent fluorescence intensity (arbitrary units) for seven staining conditions per blood group, normalized to the modal cell count for comparison. B) Mean fluorescence intensity of the fluorescence distribution curve (arbitrary unit) with standard deviation (of three data points per condition), comparing mAb and lectin binding to RBCs of blood group O, A, B, and AB (see legend above).

The newly developed mAbs bound to RBCs of all ABO blood groups, with the strongest binding to blood group O RBCs (Fig. 9A). For all four ABO blood groups, binding of A_39 mAb resulted in higher fluorescence intensities than binding of A_25 mAb, suggesting a higher affinity for carbohydrate blood group antigens (Fig. 9B). To assess the utility of these antibodies to bind more carbohydrate blood group antigens, we tested binding with our previous anti-H antigen antibodies and found that the new mAbs have observably stronger binding to RBCs from blood group A, B, and AB compared to the anti-H antibodies which primarily bind blood group O RBCs (Fig. S10).

Taken together, using our novel pipeline we have successfully developed new mAbs targeting carbohydrate antigens of RBCs from blood groups A, B, AB, and O. These mAbs will propel the development of more defined reagents for various subtypes of ABO(H) antigens, Lewis antigens, and terminal LacNAc.

Discussion

Here, we report the generation of two mAbs against carbohydrate antigens present on RBCs. The present paper describes three main aspects of the workflow for developing what we term smart anti-glycan reagents (SAGRs): 1) optimizing a protocol to develop antibodies against ABO antigens using the sea lamprey; 2) enrichment of a YSD library that carry several antibodies against various subtypes of ABO antigens; and 3) characterizing the specificity of two new anti-glycan mAbs.

The strength of this approach stems from its RBC-centric strategy, wherein lampreys were immunized with human A, B, or AB blood type RBCs to generate glycan-specific antibodies, as previously described (Fig. 1) (McKitrick et al. 2020). Subsequent enrichment of lamprey antibodies using biotinylated type A or B RBCs (Fig. 2,  Fig. S1), coupled with high-throughput screening on an ABO(H) blood group glycan microarray comprising >70 distinct ABO antigens (Fig. 3,  Table S2), enabled precise isolation of glycan-binding variable lymphocyte receptor (VLR) antibodies. Notably, this approach demonstrates that lampreys—when challenged with human cells—produce antibodies targeting conserved mammalian glycans, which can then be mapped and refined using the cognate glycans expressed on the immunizing cells. This work validates the utility of the SAGR pipeline as a robust platform for discovering antibodies against mammalian glycan epitopes, highlighting its potential for translational applications in blood typing, glycan biomarker discovery, and therapeutic antibody development.

VLRB amino acid sequence alignments of yeast clones A_25 and A_39 allowed us to compare the VLRBs on a molecular level (Fig. 4). This comparison was based on the work of Collins et al. (Collins et al. 2017), who published the VLRB amino acid sequences of three H antigen-binding VLRB-rIgG chimeras (namely: O13, Tn4–22, and RBC36). Using crystal structure analysis, they showed that two amino acid motifs, DQ (aspartic acid at position 176 and glutamine at position 177) and W (tryptophan at position 228), are critical for the interaction of a VLRB with the H antigen (Collins et al. 2017). These residues are also present in the VLRB DNA sequence of A_25 and A_39 (Fig. 4A; highlighted in red). Furthermore, their data identified VLRBs that recognize the H antigen but they often interact with a second ligand on microarrays, such as LNnT containing terminal LacNAc. Crystal structure analysis of such a VLRB (i.e. O13) (Collins et al. 2017) in complex with LNnT indicates that the same Q (glutamine at position 177) and W (tryptophan at position 228) motifs required for H antigen binding are essential, but not sufficient, for the recognition of LNnT. Additional amino acid sequence differences in the N-terminal region of the VLRBs (lysine at position 36 and aspartic acid at position 58) were predicted to be essential for LNnT interaction (Collins et al. 2017). Differences in the N-terminal region of A_25, A_39 and OmcFL3–02 support the observed differences in antibody recognition of glycans.

We confirmed the binding of the new mAbs to carbohydrate-based blood group antigens using various approaches, such as glycan microarrays, western blotting and the Luminex platform – we also computed apparent affinities in the nM-range (Figs 5–7). Furthermore, when performing competitive inhibition experiments, we found that incubation with 2′-fucosyllactose inhibits binding of A_39 mAb to LacNAc and Lewis Y but does not affect H antigen binding (Fig. 8) suggesting the unusual presence of two carbohydrate binding sites. Though unclear at present, such dual binding capacity of VLRBs-rIgG chimeras could potentially make them interesting reagents in the future to study multi-faceted glycan interactions, where more than one glycan may be important for biological activity. Crystal structure experiments are planned to analyze carbohydrate binding sites of the mAbs. Understanding which amino acids are important for glycan binding will provide key information to perform molecular modeling and engineer these binding sites to create new antibodies with enhanced or unique binding specificities such as those that bind exclusively to the A antigen, H antigen, or even unique subtypes of the ABO(H) antigens. These antibodies could also serve as a foundation for the development of new mAbs that bind to polylactosamine, which are vital in the study of galectins, innate lectins which bind to polylactosamine (Povey et al. 1990; Hirabayashi et al. 2002) and their roles in immunity and hematology (Stowell et al. 2010). A similar approach was done by Collins et al. (Collins et al. 2017) where they used site-directed mutagenesis to change the dual binding specificity of VLRB-rIgG chimera O13 by creating O13 N81H-N82Q double mutant which removed binding to LNnT while retaining binding to the H type 2 trisaccharide; however, the inverse was not performed where a mutant antibody that only binds LNnT was produced.

It is important to note that while the monoclonal yeast clones of A_25 and A_39 had considerable binding to the A and B antigens (Fig. 4), there was significantly less binding of the VLRBs in the VLRB-rIgG format, specifically for these two antigens when tested on the same array and no binding when testing for A and B antigens printed on a different array (Figs 5 and 6). One possible explanation for different binding signals when testing on different microarray is the glycan presentation. Glycan presentation on glycan microarrays (and on Luminex beads) affects the glycan recognition. Glycans are linked to the glycan microarray slide using bifunctional linkers, and Grant et al. (Grant et al. 2014) have shown that linkers impact the glycan presentation. Using computational modeling, they studied the glycan presentation when linked to different bifunctional linkers. They found that the 3D structure of the linker itself, as well as how the linker binds the glycan, influences the 3D presentation of the glycans (Grant et al. 2014). Therefore, the binding of VLRB-rIgG chimeras or other glycan-binding proteins to glycan microarrays is linker-dependent.

On yeast clones vs. VLRB-rIgG chimeras, the difference in binding may be attributed to avidity effects: Herrin et al. have shown that individual VLRB monomers have relatively low antigen-binding affinity, and that the overall avidity of a decavalent sea lamprey VLRB antibody is significantly higher (Herrin et al. 2008). Transformed yeast express many VLRBs on their surface, making them a large multivalent antigen-binding structure or particle. VLRB-rIgG chimeras, on the other hand, are small divalent glycan binding reagents. Hence, yeast expressing VLRBs with low affinity for the A or B antigen could interact with the antigens due to the sum of multiple low affinity binding interactions (avidity effect). As the VLRB-rIgG chimeras are much smaller in size and of divalent structure, interaction with more than two antigens is not possible. Therefore, we may hypothesize, that due to low antigen affinity, the VLRB-rIgG chimeras interact weakly with the A and B blood group antigens. Thus, glycan microarray or Luminex binding must be assumed to be incomplete, or to any single array format as incomplete, as VLRB-rIgG chimeras might bind to the glycan presented on one linker but not on another due to effects on glycan presentation or avidity. Despite these limitations, high throughput tools such as glycan arrays and Luminex beads are invaluable tools to develop and characterize VLRB-rIgG chimeras especially when used in combination taking into account linker and valency issues (Gillmann et al. 2023; Jia et al. 2025).

Finally, the newly developed VLRB-rIgG chimeras were tested on biological samples, demonstrating binding to RBCs across all ABO blood groups (Fig. 9). Both mAbs strongly interacted with blood group O RBCs, which aligns with the results from the glycan microarrays, where strongest binding to the H antigen was observed. Additionally, the mAbs bound RBCs of blood groups A, B, and AB. When compared to other antibodies we have identified that bind specifically to the H antigens, binding to type A, B, and AB were significantly higher (Fig. S9), confirming binding to additional carbohydrates on the RBC that are not the H antigen. These additional glycans are within members of the ABO antigens, based on weak binding to these sugars on the blood group array, and also since there is a difference in binding signal for A and B antigens when comparing both antibodies. While these antibodies currently cannot be used to distinguish blood types, and we do not know specifically which ABO antigens they are recognizing in the RBC due to broad binding specificity to multiple ABO, Lewis, and I antigens, it is promising that we can use these reagents to bind RBCs in flow cytometry. Our next steps will be to engineer these antibodies to bind specifically to A type 2, A type 4, Lewis Y, sulfated H, H type 4, or LacNAc only—based on evidence of binding to these carbohydrate antigens present in RBCs (Figs 5–7).

Here, we demonstrate the utility of our glycan-binding mAbs for western blotting and flow cytometry. Such reagents can also be used in histology and FACS experiments. When using these mAbs, it is important to add appropriate controls such as enzyme treatment with N-glycosidase (PNGase F, P0704, NEB) to confirm binding to N-glycans, or α1,3-galactosidase (P0747, NEB) and α1–2 fucosidase (G1-FM1–020, Genovis) to narrow down binding to A, B, or H antigens. Hapten inhibition with sugars can also be performed, or the development of a VLRB-rIgG null which has no glycan binding can serve as an isotype control. Other avenues of future work include the use of this SAGR pipeline to immunize lamprey with RBCs from unique or diseased patients such as RBCs from Bombay phenotype donors, sickle cell patients, or leukemia patients. This may give rise to the discovery of novel antibodies that can be used to study human health.

In conclusion, the present study resulted in the development of two monoclonal VLRB-rIgG chimeras that interact with carbohydrate-based blood group antigens. The reported mAbs will serve as foundation for studying the expression of ABO antigens and terminal LacNAc-containing glycans (e.g. blood group antigen I) on human RBCs and tissues.

Materials and methods

YSD library preparation

YSD libraries from sea lampreys (P. marinus) immunized with RBCs from blood group A, B, and AB were prepared as previously described (McKitrick et al. 2020). In brief, sea lamprey larvae (n = 3) were immunized with 10⁹ human blood type RBCs. Injections (n = 3) were given at two-week intervals, and two weeks after the final injections, animals were sacrificed and exsanguinated. Total lamprey blood was collected into 2/3 PBS in water with 30 mM EDTA, subsequently layered onto 55% Percoll, and centrifuged at 500 × g for 20 min. Total leukocytes were isolated, washed, and stored in RNAlater at −20 °C. Total RNA was extracted from lamprey leukocytes using the RNeasy Plant Mini Kit (Qiagen) and a cDNA library was generated using Superscript III First-Strand Synthesis system (ThermoFisher Scientific) with oligo dTTT primers. VLRB-specific cDNA libraries were then PCR amplified (2 μg total) from total leukocyte cDNA and electroporated into EBY100 strain of Saccharomyces cerevisiae with 1 μg of the pCT-BDNF-ESO expression vector. YSD libraries were grown with SD-CAA selection media and expression of VLRBs were induced using SG-CAA induction media. Successful inductions were determined by incubating the YSD library with anti-Myc 488 mAb and monitored by flow cytometry.

Isolation and blood typing of red blood cells

Blood samples were collected from the blood donation center at Boston Children's Hospital. Blood (V ≈ 10 to 15 mL) was transferred into a 50 mL Falcon tube and diluted with 1× PBS in a 1:1 ratio. The diluted blood suspension was slowly pipetted to a 50 mL Falcon tube containing 15 mL Ficoll-Paque PLUS (GE Healthcare). The tube was spun down at 800 × g at 20 °C for 20 min (acceleration = 5, break = 2) using the Eppendorf Centrifuge 5810R. The RBCs were collected for future experiments and stored at 4 °C. Presence of blood group antigens A, B and H was assessed using four biotinylated plant lectins and one monoclonal antibody; D. biflorus agglutinin (DBA, Vector Laboratories), Griffonia simplicifolia lectin I (GSL I, Vector Laboratories), Griffonia simplicifolia lectin I Isolectin B4 (GSL I-B₄, Vector Laboratories), U. europaeus agglutinin I (UEA-I, Vector Laboratories), and Tn4-31 L (Tulin et al. 2025). RBCs (4 μL RBCs in 200 μL total solution) were incubated with lectins or antibodies prepared in PBS buffer. DBA, GSL-I, and GSL I-B₄ were used at 10 and 20 μg/mL, UEA-I at 5 and 10 μg/mL, and Tn4-31 L at 5 μg/mL. The RBC suspensions were thoroughly mixed by pipetting up and down, and the plate was incubated at 37 °C for 30 min. After incubation the plate was checked for hemagglutination.

Biotinylation of red blood cells

RBCs that carry blood group A and B antigens as assessed by lectins and antibodies, were biotinylated for subsequent magnetic activated cell sorting (MACS). Fifty (50) μL of RBCs was mixed with 450 μL cold 1x PBS (pH 7.4), centrifuged at 400 × g for 5 min (Centrifuge 5424, Eppendorf) and the RBCs were washed 3x with 500 μL cold PBS. After washing, the RBC suspension was diluted at a ratio of 1:19 with cold 1× PBS to a volume of 1 mL (≈2.5 × 10⁷ RBCs/mL) and 1 mg EZ-link Sulfo-NHS-LC–LC-Biotin (ThermoFisher, Cat# 21338) was added, mixed by inverting the tube, and incubated for 30 min at room temperature (RT). The RBCs were spun down at 400 × g for 5 min and washed 3x with 1 mL ice-cold PBS containing 0.1 M glycine. After glycine quenching, the cells were pelleted at 400 × g for 5 min, resuspended in 1 mL cold PBS, and successful biotinylation was assessed using flow cytometry. In brief, 100 μL of biotinylated and non-biotinylated RBCs were transferred to 1.7 mL tubes (Eppendorf) and centrifuged at 400 × g for 5 min. The supernatant was discarded, and the cells were resuspended in 50 μL FACS buffer (PBS with 0.2% BSA) containing Alexa Fluor 647-conjugated Streptavidin (2 μg/mL, Jackson ImmunoResearch, Cat# 016–600-084) and incubated for 30 min at RT in the dark. Samples were washed twice with FACS buffer before analyzing on flow cytometry using standard methods on an Attune NxT acoustic focusing cytometer (Invitrogen). Negative controls were performed by analyzing unstained cells. The data was examined using the FlowJo software (v10, BD Bioscience).

Red blood cell specific VLRB-YSD library enrichment strategy

The induced YSD libraries were enriched for yeast clones expressing VLRBs specific for antigens present on RBCs using MACS. The Type B and Type AB YSD libraries were enriched on blood group B RBCs, while the Type A YSD library was enriched on blood group A RBCs following the method by McKitrick et al. (McKitrick et al. 2022). YSD libraries were resuspended in 4 mL PBSM buffer (PBS with 0.5% BSA and 2.5 mM EDTA) to a cell density of 1.5 × 10⁷ yeast cells/mL and incubated with biotinylated RBCs (6 × 10⁶ RBCs) for 60 min at 20 °C, shaking at 100 rpm in an almost horizonal orientation (30-degree angle). After incubation the cells were put on ice for 10 min. Next, 10 μL Streptavidin-MicroBeads (Miltenyi Biotec, Cat# 130–048-101) were added and incubated with the cells for 10 min on ice with gentle shaking every 2 minutes. MACS enrichment was performed using LS columns (Miltenyi Biotec, Cat# 130–042-401) following standard protocols. The enriched yeast libraries were then expanded, induced and enriched for the second time on MACS using a lower volume of RBCs (3 × 10⁶ RBCs) and microbeads (5 uL beads) to enhance the selection specificity and minimize weaker non-specific interactions. After each enrichment step successful enrichment of YSD libraries was assessed using flow cytometry. In brief, 100 μL of induced YSD library were transferred to 1.7 mL tubes. The tubes were centrifuged at 1000 × g for 3 min, the supernatant was removed, and the cells were washed with 200 μL PBS-T (PBS containing 0.05% Tween®20). The cells were spun down at 1′000 × g for 3 min and the supernatant was removed. Thereafter, the cells were resuspended in 50 μL PBS-T 0.05% containing anti-Myc tag mAb Alexa Fluor 488 (1:100, clone 4A5, EMD Millipore Corp.) and incubated for 30 min in the dark. Samples were washed thrice with PBS-T 0.05% and diluted with FACS buffer (1:1 ratio) before analyzing on flow cytometry using standard methods on an Attune NxT acoustic focusing cytometer (Invitrogen). Negative controls were performed by analyzing unstained and uninduced yeast cells. The data was examined using the FlowJo software (v10, BD Bioscience).

Blood group antigen specific VLRB-YSD library enrichment strategy

YSD libraries were further enriched using glycan microarray slides to select for RBC-binding yeast clones that bind to carbohydrate antigens. Glycan microarray slides were removed from TSM-PB (TSM with 0.05% Tween®20 and 1% BSA) rinsed with ddH₂O, dried using a table centrifuge, and a one-chamber ProPlate adapter was mounted onto the slide. Then, 2.4 mL PBS-T (PBS with 0.05% Tween®20) were added and the slide was put on a shaker for 10 min to equilibrate. During equilibration, the OD₆₀₀ of the induced YSD library was measured and the suspension was diluted to an OD₆₀₀ of 0.5 with yeast incubation buffer (PBS with 0.1% Tween®20 and 0.2% BSA). After equilibration, the slide was 3x washed with 2.4 mL sterile PBS-T and 1.8 mL diluted yeast suspension (OD₆₀₀ of 0.5) was added. The slide was enclosed in a sterile Petri dish and incubated for 4 h at 4 °C on a rocker (25 rpm) in a horizontal orientation. After incubation, the chamber was washed 4× with 2 mL sterile PBS-T. The one-chamber ProPlate adapter was removed, and the slide was put into a 50 mL Falcon tube containing sterile PBS-T. Unbound yeast was removed by gently dipping the slide in solution. After washing, excess liquid was removed by tapping the edges of the slide on tissue paper. The slide was replica plated on a SD-CAA agar plate (glycans with bound yeast facing the agar) and incubated for 48 h at 30 °C. The glycan microarray slides were re-used by washing the slides using the Microwave Assisted Wet-Erase (MAWE) method, described by Mehta et al. (Mehta et al. 2024). After incubation, replica plated yeast colonies were detached and transferred to culture tubes (Globe scientific Inc.) containing 5 mL SD-CAA medium. Then, the resulting YSD library was expanded and induced to assess successful enrichment. Additionally, the enriched YSD library was diluted 1:20′000 with sterile PBS-T and 100 μL was plated on a SD-CAA agar plate to grow monoclonal yeast. The plated yeast was incubated for 48 h at 30 °C. Colonies were picked and replica plated for subsequent colony PCR.

Screening of enriched VLRB-YSD libraries on glycan microarray

An eight-subarray glycan microarray containing multiple versions of the ABO(H) blood group antigens printed in quadruplicate (PR-124, Table S1) was used to screen the enriched YSD libraries for binding to blood group antigens.

The slide was removed from TSM-PB, rinsed with ddH₂O, dried using a table centrifuge, and an eight-chamber ProPlate adapter was mounted onto the slide. Then, 250 μL PBS-T were added per chamber and the slide was put on a shaker for 10 min to equilibrate. During equilibration, the OD₆₀₀ of the induced YSD libraries as well as induced (positive control) and uninduced (negative control) OmcFL YSD library (MACS- and FACS-enriched YSD library containing yeast binding to blood group antigen H) (Tulin et al. 2025) was measured. The suspensions were diluted to an OD₆₀₀ of 0.5 with yeast incubation buffer (PBS with 0.1% Tween®20 and 0.2% BSA). After equilibration, the slide was 3x washed with 300 μL sterile PBS-T and 200 μL of diluted yeast suspension were added according to a predefined slide design. The slide was enclosed in a sterile Petri dish and incubated for 4 h at 4 °C on a rocker (25 rpm) in horizontal position. After incubation, the chambers were washed 4x with 300 μL sterile PBS-T. Then, 150 μL PBS-T containing anti-Myc Tag Alexa Fluor488 antibody (1:250, clone 4A5, EMD Millipore Corp.) were added per chamber and the slide was incubated in the dark for 30 min at RT, on a shaker. The slide was washed 3x with 300 μL sterile PBS-T and scanned using ImageXpress® Pico (Molecular Devices) at 488 nm and brightfield. The slide was washed using MAWE and stored in TSM-PB at 4 °C. The scans were analyzed using GenePix (Molecular Devices) software and GLAD (GLycan Array Dashboard) software (Mehta and Cummings 2019).

Discovery of yeast clones that bind to blood group antigens

VLRB sequences from monoclonal yeast were screened using colony PCR. PCR master mix per reaction was prepared as follows: 5 μL 10× KOD PCR buffer (Novagen), 3 μL MgSO₄ (25 mM, Novagen), 5 μL dNTP mix (2 mM each, Novagen), 1.5 μL forward primer (10 μM, pCT_seq_fwd: 5′-ACGACGTTCCAGACTACG), 1.5 μL reverse primer (10 μM, pCT_seq_rev: 5′-TACAGTGGGAACAAAGTCG), 0.5 μL Zymolase (4 U/μL, zymoresearch), 32.5 μL nuclease free water and 1 μL KOD hot-start high-fidelity DNA polymerase (Novagen, Cat# 71086). Thermal cycler was programmed for 30 min at 37 °C for yeast cell wall degradation by Zymolase, followed by 5 min at 95 °C for initial DNA denaturation. Next, 40 cycles of 30 s at 95 °C (denaturation), 30 s at 58 °C (annealing), and 30 s at 72 °C (extension) were carried out followed by final extension at 72 °C for 5 min. PCR products were run on 2% (w/v) agarose gel in 1× TAE buffer. Gel electrophoresis was run at 80 V for 50 min. After electrophoresis, the gels were examined using the iBright1500 imaging system (Invitrogen). VLRB amplified bands (between 500 to 800 bp) were excised, purified using NucleoSpin Gel and PCR Clean-up kit (Macherey-Nagel, Cat# 740609.50) then sent for Sanger sequencing. Multiple sequence alignment was performed using Clustal Omega (https://www.ebi.ac.uk/jdispatcher/msa/clustalo). Yeast clones with unique VLRB sequence were subsequently screened in glycan microarrays.

Generation of VLRB-rIgG chimeric proteins

VLRB sequences were cloned into a pCDNA mammalian expression vector in frame with an N-terminal Interleukin-2 signal peptide and a C-terminal rat IgG2b Fc chain with a C-terminal 6xHis tag for ease in expression and purification of monoclonal VLRB-rat IgG Fc (VLRB-rIgG) chimeras. The following primers were used for cloning via Gibson assembly (VLRB_GA_F: 5′-CACTAAGTCTTGCACTTGTCACGAATTCGG-CTAGCTGTCCCTCGCAGTG-3′, VLRB_at_Fc_GA_R: 5′-ATTTGTGACATGTAGGGCATGTAGGGGATCCCGTGGT-CGTAGCAACG-3′, pcDNA_Rat_Fc_GA_F: 5′- CCAGGCTACG-TTGCTACGACCACGGGATCCCCTACATGCC-3′, pcDNA_GA_R: 5′- CACGAACACTGCGAGGG-ACAGCTAGCCGAATTCGTGAC-3′). Thereafter, monoclonal VLRB-rIgG chimeric proteins were expressed by transiently transfecting into ExpiCHO cells (ThermoFisher, Cat# A29133) following manufacturer’s protocols. The antibodies were purified from the media using a HisPur™ Ni-NTA resin (ThermoFisher, Cat# 88222) following standard protocols. Monoclonal antibody concentrations were determined using the Pierce BCA Protein Assay Kit (Thermo Scientific, Cat# 23225) and the purity of the antibodies were verified by reducing and non-reducing sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) on 4–12% gels stained using Coomassie and Western blotting to nitrocellulose membranes followed by development using IRDye 800CW goat anti-rat IgG (LI-COR, Cat# 926–32219).

Analysis of antibody specificity to glycan microarrays

Antigen binding was determined by screening the produced mAbs on three different glycan microarrays; the Consortium for Functional Glycomics (CFG), which is the most complete glycan microarray available with over 600 immobilized diverse mammalian-type glycans with N- and O-glycans, NCFG glycan microarray PR-164, with unique glycan structures including some ABO blood group antigens, and PR181 which is an ABO(H) blood group antigen glycan microarray. The list of glycans printed on the CFG and NCFG array is available at the NCFG website (https://research.bidmc.org/ncfg/microarrays), or published by Mehta et al. (Mehta et al. 2024), while the glycans printed on the PR181 array can be found in Table S2. Each glycan on the microarrays was printed in quadruplicate to ensure reproducibility and reliability of the binding data. VLRB-rIgG (50 μg/mL) diluted in TSM-BB (TSM with 0.05% Tween®20 and 1% BSA) were added on the slides. Then, slides were incubated for 1 h at RT on a shaker. Next, the chamber/slide was washed 4x with TSM-W (TSM with 0.05% Tween®20) and 4× with TSM before incubating with biotinylated goat-anti-rat IgG secondary antibody (5 μg/mL, Jackson ImmunoResearch, Cat# 112–065-167) for 1 h at RT with shaking. Slides were washed 4× with TSM-W and 4x with TSM, then incubated with Streptavidin-Cy5 (0.5 μg/mL, Invitrogen, Cat# 434316), 1 h at RT with shaking. Slides were washed and then scanned at 647 nm using ImageXpress® Pico. The data was analyzed using GenePix (Molecular Devices) and GLAD software (Mehta and Cummings 2019).

Generation of ABO blood group microarray

The ABO blood group microarray was prepared as previously described (Wei et al. 2019). Briefly, 3-(methoxyamino) propylamine (MAPA) and 2-amino-N-(2-amino-ethyl)-benzamide (AEAB) linked blood group glycans were dissolved to a final concentration of 50 mM in 100 mM sodium phosphate buffer, pH 8.5. Printing was performed using Scienion sciFLEXARRAYER SX (Berlin, Germany) onto Nexterion H NHS functionalized slides from Schott (Louisville, KY), where each probe was printed as 4 spots within a single array, and each slide contained 8 arrays. The average volume per spot was 330 pL. The slides were incubated at room temperature on the deck of the printer overnight at 70% relative humidity. The next day, the slides were blocked with 50 mM ethanolamine in 100 mM sodium tetraborate buffer (pH 8.5) for 1 h. Following this, the slides were dip-washed 10 times in 1 x phosphate buffered saline (PBS) containing 0.05% Tween-20 and then dip-washed 10 times in Milli-Q water. The slides were then dried in a slide centrifuge and stored at −20 °C until use.

Western blotting on CHO cell lines

Wild-type CHO (Pro.5) and mutant (Lec1) cells were purchased from ATCC and cultured in Dulbecco’s modified eagle’s medium (DMEM) at 37 °C and 5% CO₂ in medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S). Cells were lysed in RIPA buffer with protease inhibitor (Roche, Cat# 46931320019) and dissociated using a hand-held motorized pestle (Kimble, Cat# 749540), followed by two brief pulses of sonication for 10 seconds with a microtip (Qsonica Q700). Protein concentration was analyzed using the Pierce BCA Protein Assay Kit (ThermoFisher Scientific, Cat# 23255). Samples were treated with 500 U of PNGaseF (NEB, Cat# P0704L) or 200 U of Neuraminidase A (NEB, Cat# P0722L) overnight at 37 °C. Lysates were prepared with 4X Sample Loading Buffer (Li-COR, Cat# 928–40,004) with 10% v/v β-mercaptoethanol, and denatured for 10 min at 95 °C. Twenty (20) μg protein was loaded per well. Gels were run using the MiniProtean Tetra Electrophoresis System (BioRad, Cat# 1658004) at 60 mV for 30 min followed by 110 mV for 1 h. Proteins were transferred to nitrocellulose membranes (BioRad, Cat# 1704158) using the Trans-Blot Turbo Transfer System (BioRad, Cat# 1704150). Total protein (Revert 700 Total Protein Stain, LiCOR, Cat# 926–11,011) was used as a loading control. Membranes were then incubated in 5% BSA in TBS-Tween 0.1% for 1 h, followed by incubation with A_39 mAb at 5 μg/mL in blocking buffer overnight at 4 °C on a rocking platform shaker. Membranes were washed three times in TBS-Tween 0.1% for 5 min, and then incubated with fluorescent conjugated anti-rat IRDye 800CW (LiCOR, Cat# 926–32,219) at 0.05 μg/mL dilution in 5% BSA in TBS-Tween 0.1% for 1 h protected from light. Membranes were again washed three times in TBS-Tween 0.1% for 5 min and imaged using a LiCOR Odyssey CLx Imaging System and analyzed using LiCOR Image Studio Software.

Glycan binding assay using Luminex platform

For glycan binding on the Luminex platform, the method by Jia et al. (Jia et al. 2025) was adapted. Briefly, 1000 beads/well for each glycan were mixed in a 1.7 mL microcentrifuge tube to create an array. Ten (10) μL mixed beads suspension was added to each well of a 96-well plate (white, round bottom), followed by the addition of 40 μL PBS-TBN containing serial dilution anti-glycan antibodies. The mixture was incubated on a shaker at RT for 1 h. The 96-well plate was placed on a magnetic separator and separation was allowed to occur for 3 minutes, the supernatant was removed and the wells were twice washed with PBS-TBN. After washing, 50 μL PBS-TBN containing 0.5 μg Alexa Flour 546 nm labeled goat-anti-rat IgG was added to each well and incubated for 1 h. The plate was washed twice with PBS-TBN, and beads were resuspended in 100 μL of PBS-TBN and the fluorescence intensities were measured on a Luminex FLEXMAP 3D instrument with minimum events 100/beads.

Competitive inhibition experiments

For competitive binding experiments, VLRB-rIgG was diluted to 20 μg/mL in TSM-BB with either 100 μM 2′-fucosyllactose (Sigma-Aldrich, Cat# SMB0093), 100 μM lactose (Fisher Chemical, Cat# L5–500), or 100 μM N-acetyllactosamine (Sigma-Aldrich, Cat# A7791). Antibody-sugar solutions were incubated on the NCFG glycan microarray for 1 h at RT, with shaking. After incubation, the chambers were washed 4× with TSM-W and 4× with TSM. Addition of secondary antibodies, and subsequent washing and scanning steps were performed following method above for glycan microarray testing.

Binding of novel antibodies to biological samples

RBCs from donors that are type A, B, AB, and O were stained in triplicates and analyzed using flow cytometry to check binding with novel anti-glycan antibodies. Typed RBCs (provided by Dr. Leon Zheng, Brigham and Women's Hospital, Harvard Medical School, Mass General Brigham Institutional Review Boards [IRB] Approved protocol #2023P0003321, to Stowell SR) were diluted 1:200 in 1× PBS and distributed across a 96-well plate (200 μL/well, V-bottom, Costar, Cat# 3897). The cells were spun down at 450 × g for 5 min at 8 °C. The supernatant was removed, and the cells were resuspended in 100 μL primary antibody or lectin diluted in FACS buffer; A_25 mAb and A_39 mAb (0.25 μg/mL), biotinylated UEA-I (0.5 μg/mL), and biotinylated DBA (10 μg/mL) were used. RBCs were incubated for 30 min at 4 °C. After incubation, cells were resuspended in 100 μL goat-anti-rat AlexaFluor647 (Jackson ImmunoResearch, Cat# 112–605-167) or 100 μL Streptavidin-Cy5 (Invitrogen, Cat# 434316) both diluted to 1 μg/mL in FACS buffer. Samples were washed twice with FACS buffer before analyzing on flow cytometry using standard methods on an Attune NxT acoustic focusing cytometer (Invitrogen). Negative controls were performed by analyzing unstained cells and cells stained with secondary antibodies only. The data was examined using the FlowJo software (v10, BD Bioscience).

Author contributions

RDC and EKCT conceptualized and designed the project. PBK performed enrichment, screening, expression and characterization of antibodies, and wrote the manuscript with EKCT. AYM performed the printing of the microarray slides for the NCFG and the ABO glycan microarrays. AYM and JHM oversaw all glycan microarray experiments and wrote related portions of the manuscript. TWJ performed antibody screening on Luminex glycan beads. VIO co-supervised PBK and helped conceptualize the project. SRS provided the typed red blood cells and useful discussions. RDC co-supervised PBK, EKCT, and TWJ, provided funding, wrote the manuscript, and coordinated collaborators. All authors contributed feedback and edits to the manuscript.

Pascal B Kunz (Formal analysis [equal], Methodology [equal], Writing—original draft [equal]), Ea Kristine Clarisse Tulin (Conceptualization [equal], Formal analysis [equal], Investigation [equal], Methodology [equal], Writing—original draft [equal]), Akul Yugesh Mehta (Formal analysis [equal], Supervision [equal], Visualization [equal], Writing—review & editing [equal]), Tianwei Jia (Formal analysis [equal], Investigation [equal], Writing—review & editing [equal]), Jamie Heimburg-Molinaro (Formal analysis [equal], Supervision [equal], Writing—original draft [equal], Writing—review & editing [equal]), Vivianne I Otto (Conceptualization [equal], Supervision [equal]), Sean R Stowell (Funding acquisition [equal], Resources [equal], Writing—review & editing [supporting]), Richard D Cummings (Conceptualization [equal], Methodology [equal], Project administration [equal], Resources [equal], Supervision [equal], Writing—original draft [equal], Writing—review & editing [equal]).

Funding

We acknowledge funding from the National Institutes of Health (grants P01HL171803, R24GM137763, U01CA199882, and R01GM140210 to RDC).We thank the Massachusetts Life Science Center Research Infrastructure program for support. PBK gratefully acknowledges the support of the Swiss European Mobility Program (SEMP), which supported his research stay in Boston.

Conflicts of interest

None declared.