The structures of glycophorin C N-glycans, a putative component of the GPC receptor site for Plasmodium falciparum EBA-140 ligand

Abstract

Glycophorins C and D are highly glycosylated integral sialoglycoproteins of human red blood cell membranes carrying the Gerbich blood group antigens. The O- and N-glycosidic chains of the major erythrocyte glycoprotein (Lisowska E. 2001, Antigenic properties of human glycophorins – an update. Adv Exp Med Biol, 491:155–169; Tomita M and Marchesi VT. 1975, Amino-acid sequence and oligosaccharide attachment sites of human erythrocyte glycophorin. Proc Natl Acad Sci USA, 72:2964–2968.) are well characterized but the structure of GPC N-glycans has remained unknown. This problem became important since it was reported that GPC N-glycans play an essential role in the interaction with Plasmodium falciparum EBA-140 merozoite ligand. The elucidation of these structures seems essential for full characterization of the GPC binding site for the EBA-140 ligand. We have employed detailed structural analysis using sequential mass spectrometry to show that many GPC N-glycans contain H2 antigen structures and several contain polylactosamine structures capped with fucose. The results obtained indicate structural heterogeneity of the GPC N-glycans and show the existence of structural elements not found in glycophorin A N-glycans. Our results also open a possibility of new interpretation of the data concerning the binding of P. falciparum EBA-140 ligand to GPC. We hypothesize that preferable terminal fucosylation of N-glycosidic chains containing repeating lactosamine units of the GPC Gerbich variant could be an explanation for why the EBA-140 ligand does not react with GPC Gerbich and an indication that the EBA-140 interaction with GPC is distinctly dependent on the GPC N-glycan structure.

Introduction

Glycophorins C (GPC) and D (GPD), carrying the Gerbich (Ge) blood group antigens (Reid and Spring 1994;Schawalder et al. 2004; Walker and Reid 2010), are minor but important integral sialoglycoproteins of human red blood cell (RBC) membranes involved in the regulation of membrane mechanical properties (Cartron et al. 1993; Salomao et al. 2008). GPC, composed of 128 amino acid residues, is encoded by four exons (Dahr et al. 1982; Colin et al. 1986; Blanchard et al. 1987; High and Tanner 1987). GPD is a truncated form of GPC which lacks the first 21 amino acid residues (Le Van Kim et al. 1996). GPC and GPD are produced by the same mRNA by the alternative use of two in-frame initiation codons. There are also two known rare RBC phenotypes, Yus- and Gerbich-negative, characterized by the presence of natural GPC deletion variants lacking amino acid residues 17–35 (encoded by exon 2) or 36–63 (encoded by exon 3), respectively (Colin et al. 1989; Johnson and Daniels 1997).

GPC and GPD are highly glycosylated. GPC contains∼12 O-linked glycans and one N-glycan linked to Asn8, and GPD contains only 6–8 O-glycans (Dahr et al. 1982; Colin et al. 1986; Blanchard et al. 1987; High and Tanner 1987). The high level of glycosylation of GPC and GPD modulates their antigenic properties (Lisowska 2001; Jaskiewicz, Czerwinski, Colin, et al. 2002; Jaskiewicz, Czerwinski, Uchikawa, et al. 2002; Schawalder et al. 2004). While the O- and N-glycosidic chains of glycophorin A (GPA, the major RBC sialoglycoprotein (Tomita and Marchesi 1975; Lisowska 2001)) are well characterized (Thomas and Winzler 1969; Yoshima et al. 1980; Lisowska 2001); the structure of GPC glycans was not studied due to difficulties in GPC purification and separation from more abundant GPA. There are several lines of indirect evidence that all glycophorins carry the same O-chains, sialylated Galβ1-3GalNAc-units (Lisowska 2001), but the structure of GPC N-glycans has remained unknown. This problem became important in view of recent studies on malaria infection. It has been shown that Plasmodium falciparum EBA-140 ligand binds to GPC (Mayer et al. 2001; Lobo et al. 2003; Maier et al. 2003). This binding is dependent on GPC sialylation and it was reported that GPC N-glycan plays an essential role in the interaction with the EBA-140 ligand (Mayer et al. 2006). However, the mechanism of GPC-EBA-140 interaction is far from being clear. In this context, the elucidation of the structure of GPC N-glycans seemed to be the necessary condition for full characterization of the GPC binding site for the EBA-140 ligand.

In this study, for the first time, we have used purified GPC preparations (free of GPA) for structural analysis of N-glycans released by hydrazinolysis and examined by mass spectrometry. General aspects of the MSⁿ structural analysis strategy have been described (Ashline et al. 2005, 2007; Hanneman et al. 2006; Prien et al. 2008; Stumpo and Reinhold 2010; Ashline, Hanneman, et al. 2014; Ashline, Yu, et al. 2014). In many biological samples, a given oligosaccharide composition can be made up of several isomers. Strategies relying only on intact mass information will empirically determine composition, but mixtures with multiple isomers will be transparent to this approach. Disassembly in an ion trap mass analyzer provides the opportunity to isolate diagnostic fragments, yielding direct empirical data about topology and substructures. Coupled with library matching of standard substructure spectra, this strategy also allows for assessment of particular epitopes of interest, such as H antigen and Lewis structures. From a de novo structural analysis perspective, sequential disassembly enables determination of new structures which may be previously unreported. In the current study, we examine the N-glycans of GPC, with a particular emphasis on fucosylated and poly-LacNAc structures, providing both a list of glycans and a blueprint for a strategic approach to other such problems.

The results obtained indicate structural heterogeneity of the GPC N-glycans and show the existence of structural elements not found in GPA N-glycans.

Results

To study the structure of GPC N-glycan, a fundamental problem was to obtain purified GPC. This is a difficult problem due to the abundant amount of GPA and the low abundance of GPC in erythrocyte membranes, as well as the strong aggregation of these sialoglycoproteins. We were successful in obtaining the GPC preparation free of GPA (Figure 1). The co-purification of a small amount of GPB is not important, since GPB does not contain N-glycans.

Fig. 1.

Western blotting of purified GPC+GPB sample used for N-glycan MS analysis detected with MoAb anti-GPC (clone NaM57-1F6) and anti-GPA+GPB (clone NaM26-3F4). The sample consists of GPC and GPB (which does not contain N-glycan) and GPA is not present. MW, protein molecular-weight standards; GPC+GPB sample; GP, crude glycophorins.

The N-linked glycan structures isolated from glycophorin C were predominantly complex biantennary structures containing bisecting HexNAc, with variable numbers of sialic acid and fucose residues. Fucosylation occurred both on the core and on the antennae. Small amounts of poly-LacNAc structures were also observed. Figure 2 shows a MALDI-Tof spectrum of the reduced and permethylated N-linked glycans. Due to the complex mixtures of isomers in the sample, peaks are annotated with masses only. Table I illustrates, in a condensed form, the various structures, including isomers, detected on GPC. Values for relative amounts were obtained from the MALDI-Tof data by normalizing peak intensity data. The remaining figures highlight the supporting MSⁿ data for selected compositions, including strategies for empirically localizing fucose residues, including on the extended antennae.

Fig. 2.

MALDI mass spectrum of reduced and permethylated N-linked glycans from GPC. Masses are labeled; structures and relative amounts are enumerated in Table I.

Table I.

Masses, m/z values and likely structure determined from purified GPC protein

Measured mass [M+Na]+ (MALDI-Tof)	Theoretical mass and m/z [M+Na]+ [M+2Na]2+	Normalized relative abundance (%)	Structures
1595.3	1595.8	2.5	High Mannose-Man5
1799.5	1799.9	2.3	High Mannose-Man6
2003.7	2004.0	3.1	High Mannose-Man7
2085.7	2086.1	1.7
2259.9	2260.2	2.2
2301.0	2301.2	3.2
2411.9	2412.2	7.9	High Mannose-Man9
2505.1	2505.31+ 1264.22+	13.6
2622.4	2621.31+ 1322.22+	2.2
2679.2	2679.41+ 1351.22+	12.5
2809.6	2808.41+ 1415.72+	2.2
2853.1	2853.51+ 1438.32+	5.5
2866.3	2866.51+ 1444.82+	11.3
2954.4	2954.51+ 1488.82+	2.3
3040.4	3040.51+ 1531.82+	3.7
3128.4	3128.6	4.3
3227.6	3227.61+ 1625.32+	5.6
3302.4	3302.71+ 1662.92+	4.5
3489.6	3489.81+ 1756.42+	1.9
3577.6	3577.81+ 1207.93+	1.9
3751.4	3751.91+ 1887.52+	2.0
3765.0	3764.91+ 1270.33+	1.1
3925.2	39261+ 1324.03+	1.1
3938.8	3939.01+ 1328.33+	1.1
4126.4	4126.11+ 1390.73+	<1.0
4200.3	4201.1 1415.73+	<1.0
4300.7	4300.2	<1.0

Normalized relative abundance was determined by dividing the MALDI-Tof peak intensity by the sum of all peak intensities multiplied by 100.

Most of the structures found are fucosylated with bisecting GlcNAc. Localization of the fucose and the bisecting GlcNAc residues were determined empirically. The data set obtained through the disassembly of the doubly charged m/z 1351 ion, composition GlcNAc₅Gal₂Man₃Fuc₂ is shown in Figure 3. The MS² spectrum contains several informative fragments. Reduction prior to permethylation introduces a mass difference at the reducing end GlcNAc residue, producing a distinct fragment and neutral loss complement. The m/z 490 fragment is characteristic of a fucosylated reducing end GlcNAc, thus localizing one fucose to the reducing end. The fragment at m/z 660 is characteristic of a B-type terminal fucosylated lactosamine fragment. It should be pointed out that the mass of this ion empirically gives composition; determination of the structure requires that this ion be disassembled further. This spectrum is shown in Figure 3F; employing a spectral matching approach, this spectrum is characteristic of H2 antigen (Ashline, Hanneman, et al. 2014; Ashline, Yu, et al. 2014). Presence of bisecting GlcNAc is empirically determined by sequentially isolating and dissociating the trimannosyl core fragments. For a permethylated sample, the number of substitutions on the core can be determined by the mass of the trimannosyl-containing fragment after loss of all substituents. The MS⁴ spectrum of the m/z 990 precursor contains a singly charged ion at m/z 852. This mass is typical of the B/Y/Y/Y core fragment (trimannose and one GlcNAc of the chitobiose core). The loss of the reducing end GlcNAc leaves a B type scar and the loss of non-reducing end substituents typically leaves Y-type (hydroxyl) scars. Thus, the m/z 852 ion is indicative of a structure with three substituents, one of which may or may not be a bisecting GlcNAc. To determine the presence of a bisecting GlcNAc, this ion is dissociated to reveal losses of internal mannose residues and produces an m/z 444 B/Y/Y/Y-type Man-GlcNAc fragment. The relevant spectra are shown in Figure 3D and E, where the cartoon graphic shows the structure for these fragments, and the presence of the m/z 444 fragment, with no detectable m/z 458 fragment. An m/z 458 fragment would be expected if there were no bisecting GlcNAc. Figure 3I shows the structure of the m/z 852 fragment for a structure containing a bisecting GlcNAc. A structure without a bisecting GlcNAc would typically produce an m/z 458 B/Y/Y-type Man-GlcNAc ion.

Fig. 3.

MSⁿ spectra of the m/z 1351²⁺, composition Hex₅HexNAc₅dHex₂, shows a biantennary structure with one core fucose and one antennal fucose. (A) The MS² spectrum of the doubly-charged m/z 1351. The structure is shown in (G). (B) The MS³ of the doubly-charged m/z 1221. (C) The MS⁴ of the doubly-charged m/z 990. This structure is shown in (H), reflecting the loss of bisecting GlcNAc and terminal LacNAc. Fragmentation of the core m/z 852 fragment in (D) and (E) is indicative of a structure having a bisecting GlcNAc as shown in (I). The fucosylated antenna fragment was isolated as the B-type ion, m/z 660, shown in (F). This fragmentation pattern is consistent with an H2 structure (J), with no trace of Lewis X. Normalization level (NL) and signal averaging time are shown for each spectrum.

Figure 4 shows the disassembly of a biantennary, monosialylated, and monofucosylated composition. For this composition, there seem to be two isomers differing in fucose localization; one isomer with core fucose and the second with antennal fucose. Panels A–C show several stage of disassembly, showing the core fucosylated fragment at m/z 490 and the antennal fragment at m/z 660. Figure 4D shows the MS³ spectrum of the terminal fucosylated lactosamine fragment, which matches that of an H2 standard (Ashline, Hanneman, et al. 2014; Ashline, Yu, et al. 2014). Figure 4E and F shows the major fragments of the two isomers. Figure 4G shows the fragments of the B-type H2 structural fragment, corresponding to the Figure 4D spectrum.

Fig. 4.

MSⁿ spectra of m/z 1444²⁺, composition Hex₅HexNAc₅dHexNeuAc, showing a biantennary structure with two isomers, differing in fucose location. (A) The MS² spectrum of the m/z 1444 precursor. (B) The MS³ (m/z 1444 → 1257) spectrum of the fragment formed by loss of NeuAc. (C) The MS⁴ (m/z 1444 → 1257 → 1127) spectrum of the fragment formed by successive losses of NeuAc and terminal HexNAc. These spectra show isomers with both core Fuc-GlcNAc (m/z 490) and antennal Fuc (m/z 660). (D) and (E) The core fragments indicating a structure containing a bisecting GlcNAc. (F) and (G) The putative structures of these two isomers. (H) The structure of the isolated H2 B-type ion with fragment assignments. (I) The MS³ m/z 660, a terminal fucosylated lactosamine ion, with fragments consistent with H2, based on similarity to standards and de novo fragment assignment.

Interrogation of the m/z 1663²⁺ ion, composition Hex₆HexNAc₆dHex₃, revealed two predominating isomers, both with bisecting HexNAc, as shown in Figure 5. One isomer was a triantennary structure and the other was a biantennary with an extended antenna. Fucosylation occurred on both the core and the antennae. MS^2–4 spectra revealed an intense fragment ion at m/z 660, consistent with a B-type fucosylated lactosamine structure. CID of this fragment revealed the typical fragmentation of the H2 epitope. Discernment of the isomeric bi- and triantennary structures required a deeper interrogation. As is typical of collision-induced dissociation, earlier stage spectra are dominated by cleavage of GlcNAc-Man and GlcNAc-GlcNAc bonds, producing a multitude of B- and Y-type fragments. Following the selected fragmentation pathway through MS⁶, the losses are consistent with two-terminal fucosylated lactosamines, a HexNAc (most likely the bisecting GlcNAc) and the fucosylated GlcNAc of the reducing end. The resulting fragment culminates in the MS⁶ spectrum of the m/z 1301 precursor. Of interest here are the fragments at m/z 852 and m/z 838, formed by neutral losses of 449 (B/Y-type LacNAc) and 463 (B-type LacNAc), respectively. These are the fragments that are consistent with a biantennary structure (m/z 852) and a triantennary structure (m/z 838), both with bisecting GlcNAc. In addition to the aforementioned core fragmentation, the extended antennal fragment can be isolated as the B-type fragment at m/z 1109, as shown in Figure 6. This mass is consistent with a B-type monofucosylated di-LacNAc structure. Dissociation of this unit indicates clearly that the predominant isomer has the fucose located on the distal LacNAc (indicated by the pair of fragments at m/z 472 and 660), though a small amount of an isomer with the fucose on the proximal LacNAc is also detectable (indicated by the pair of fragments at m/z 486 and 646). Isolation of the m/z 660 fragment from the extended antenna fragment reveals, again, the H2 epitope. The unsubstituted LacNAc is clearly four-linked, as indicated by the m/z 315 fragment, a ^3,5A-type cross-ring cleavage, in the m/z 472 CID spectrum. This motif was found on several GPC N-glycans, as shown in Table I, with additional mass spectral data shown in the supplemental information.

Fig. 5.

Partial MSⁿ data set for m/z 1663²⁺, composition Hex₆HexNAc₆dHex₃. Disassembly reveals at least two isomers, differing in antenna number and length, but both having bisecting GlcNAc. Differing antennal numbers are verified by the presence of both m/z 852 and 838 fragment ions, indicative of biantennary (with bisecting GlcNAc) and triantennary (with bisecting GlcNAc). As before, the m/z 660 ion and its CID spectrum are consistent with H2 epitope.

Fig. 6.

Disassembly of the extended antenna from Hex₆HexNAc₆dHex₃, m/z 1663²⁺. The singly charged fragment at m/z 1109 is consistent with a B-type monofucosylated diLacNAc structure, composition Hex₂HexNAc₂dHex₁.

For many of the larger compositions interrogated, we found that tri-LacNAc substructures, with and without sialylation and/or fucosylation, were frequently branched, with detectable amounts of the linear topology. This was determined empirically, with the relevant data shown in the supplement for those structures.

All of the fucosylated lactosamine units detected and analyzed were consistent with H2 structures. There was no indication of Lewis structures, either terminal or internal. Multiply fucosylated structures showed no evidence of Lewis Y structures; in fact, there appeared to be no doubly fucosylated lactosamine structures.

Discussion

Our results demonstrate differences between the structures of GPA and GPC N-glycans. The major N-glycan of GPA is a biantennary sialylated complex chain containing bisecting GlcNAc and fucosylated proximal GlcNAc residue, and no elongated chains have been found (Yoshima et al. 1980; Lisowska 2001). Although the same structure was identified in GPC-derived N-glycan pool, it was accompanied by a number of structures with elongated antennae, containing two repeating N-acetyllactosamine (LacNAc) units and terminated with α1,2-linked fucose residue instead of sialic acid. The terminal Fucα1-2Gal-unit (blood group H determinant) occurred rarely in chains with a single LacNAc residue. It is in agreement with the finding that structures with blood group H determinant (and A or B, dependently on the blood group of the donor) identified as variant forms of the GPA N- and O-glycans constituted only their minor portion (Podbielska, et al. 2004; Fredriksson et al. 2010).

The preference for biosynthesis of polylactosaminoglycans is not fully understood and several models have been proposed (Merkle and Cummings 1987; Fukuda et al. 1988). It may be dependent on the way of protein anchoring in the membrane, a close proximity of the N-glycosylation site to the membrane-bound glycosyltransferases, or a longer residence time in the Golgi compartment (Wang et al. 1991). A possibility that polylactosamine biosynthesis is determined by the physical distance of N-glycosylation site from the membrane was suggested for two polylactosaminoglycans-containing erythrocyte proteins, the anion transporter (band 3) (Kopito and Lodish 1985) and glucose transporter (band 4.5) (Mueckler et al. 1985). Also, the amino acid sequence adjacent to the acceptor site may have an effect. The homology found between band 3 and band 4.5 is Trp-Val sequence near the glycosylation site. GPC does not contain this sequence, but there is 1Met-Trp-Ser3 sequence in proximity to N-glycosylated Asn8 residue. GPA and GPC have polypeptide chain of similar length (131 and 128 aa residues, respectively), but the extracellular portion of GPA is longer (72 aa residues) than that of GPC (57 aa residues). On the other hand, the distance of GPA N-glycosylation site (Asn26) to the membrane is slightly shorter than in GPC (Asn8).

We show here that in GPC N-glycan the terminal α1,2-linked fucose residue replaced sialic acid preferentially in chains containing two LacNAc units. This finding is in agreement with the literature data. In the fucosyltransferase (α1,2FT) transfected CHO cells, the preferential fucosylation of polylactosaminoglycans was observed (Prieto et al. 1997). Moreover, the presence of terminal fucosylation in these CHO cells occurred at the expense of sialylation, indicating a competition between α1,2-fucosylation and sialylation. The α1,2-fucosylation of galactose forms the blood group H determinant and is a necessary step for biosynthesis of blood group A and B antigens. The major glycoprotein carriers of ABH antigens in human RBCs are band 3 and band 4.5 proteins containing polylactosaminoglycans (Kopito and Lodish 1985; Mueckler et al. 1985), while these antigens are only minor components of GPA which does not contain elongated N-glycans (Podbielska et al. 2004; Fredriksson et al. 2010). All these data show that α1,2-fucosylation successfully competes with sialylation in terminating polylactosamine chains, while for shorter structures terminating by sialic acid is preferable. Our present results showed that the GPC N-glycosylation pattern is a new natural example to support these conclusions.

Our results also open a possibility of new interpretation of the data concerning the binding of P. falciparum EBA-140 ligand to GPC (Mayer et al. 2001; Lobo et al. 2003; Maier et al. 2003). The engagement of GPC N-glycan in this interaction was shown by the Miller group (Mayer et al. 2006). They found that GPC de-N-glycosylated with endoglycosidase F lost the ability to inhibit the binding of the EBA-140 ligand to human erythrocytes. Moreover, the EBA-140 interaction with GPC is distinctly dependent on GPC sialylation (Mayer et al. 2001, 2006; Lobo et al. 2003; Maier et al. 2003). Therefore, it was proposed that a receptor for the EBA-140 ligand might be a conformationally arranged cluster of sialic acid residues attached to the N- and O-linked oligosaccharide chains on the GPC molecule (Mayer et al. 2006; Lin et al. 2012). This suggestion has been recently supported by finding that the crystal form of the EBA-140 binding region (Region II) has two glycan-binding pockets interacting with sialic acid residues (Malpede et al. 2013). However, this view is complicated by the lack of the EBA-140 binding to the Gerbich-negative RBCs containing the truncated variant of GPC (GPC-Ge), lacking the nonglycosylated (or containing one O-glycan only) aa residues 36–63 of GPC polypeptide chain (Mayer et al. 2001,  2006). The GPC-Ge migrates as a broad band in sodium dodecylsulfate (SDS)–polyacrylamide gel electrophoresis that indicates its increased and more heterogeneous glycosylation, comparing with normal GPC. The structure of the N-glycans of GPC Gerbich-type has not been studied by instrumental methods, but there is some indirect evidence based on enzymatic studies that this chain is a polylactosaminoglycan which is more heterogenous and larger than N-glycan of normal GPC (Reid et al. 1987; Jaskiewicz, Czerwinski, Colin, et al. 2002; Mayer et al. 2006). These data and our present results with GPC N-glycan, indicating preferable terminal fucosylation of chains containing repeating lactosamine units suggests that elongated antennae of polylactosaminoglycans of the GPC-Ge variant may be terminating only by fucose, and not by sialic acid. The altered structure of GPC-Ge N-glycan and the lack of its functional sialic acid residues could be a reasonable explanation why the EBA-140 ligand does not react with GPC-Ge. In the present state of knowledge, this suggestion is largely hypothetical, but interesting enough to deserve an elucidation.

Materials and methods

GPC purification

Sialoglycoproteins were prepared from outdated O,MN erythrocytes by phenol/saline extraction of membranes (Lisowska et al. 1987). The crude glycoproteins were fractionated on Sephadex G 200 equilibrated with 0.05 M pyridine-acetate buffer, pH 5.3, in the presence of 1% SDS (Waśniowska et al. 1985). The column fractions were monitored for absorbance at 280 nm, for neutral sugar content by the phenol/sulfuric acid method and stained by periodic acid-Shiff's reagent after SDS–polyacrylamide gel electrophoresis (Laemmli 1970). Selected fractions enriched in glycophorins B and C were pooled and rechromatographed on Sephadex G 200 under the same conditions. Fractions containing GPC and GPB but no GPA were pooled, dialyzed against 50% ethanol to remove SDS, dialyzed against distilled water and lyophilized. The obtained glycoprotein sample was assayed by immunoblotting with MoAb anti-GPC, clone NaM57-1F6 and MoAb anti-GPA+GPB, clone NaM26-3F4 (Reid et al. 1997; Wasniowska et al. 1997).

N-Glycan processing

N-Glycans were released by hydrazinolysis (Patel et al. 1993; Hanneman et al. 2006). Briefly, the 400 μL anhydrous hydrazine (Sigma, St. Louis, MO) was added to the dried glycoprotein and incubated at 100°C for 6 h. Hydrazine was removed via N₂ stream. Glycans were re-N-acetylated with acetic anhydride in saturated sodium bicarbonate. Any esters (O-acetylation) formed during this step were hydrolyzed with 0.01 M sodium hydroxide at room temperature. Cellulose column cleanup was performed as described (Shimizu et al. 2001); cellulose was equilibrated with 4 : 1 : 1 butanol/ethanol/water, and glycans were eluted with 50 : 50 ethanol/water. Glycans were reduced with ammonia/borane and permethylated using spin columns as described (Kang et al. 2005; Desantos-Garcia et al. 2011), except that a graphitized carbon solid-phase extraction was performed between the reduction and permethylation steps. Permethylated oligosaccharides were purified by liquid–liquid extraction with dichloromethane and 0.5 M sodium chloride.

Mass spectrometry

MALDI-Tof mass spectrometry was performed on a Shimadzu Kratos Axima-CFR in reflectron mode (Shimadzu Life Sciences, Columbia, MD). Ion trap mass spectrometry was performed either on an LTQ (ThermoFisher, San Jose, CA) equipped with a TriVersa Nanomate (Advion, Ithaca, NY) or a VelosPro equipped with a Nanospray Flex ion source (ThermoFisher, San Jose, CA) and a Harvard Apparatus 22 syringe pump (Harvard Apparatus, South Natick, MA). Most of the data were collected on the LTQ; the superior MSⁿ sensitivity of the VelosPro enabled collection of some higher stage MSⁿ data not obtainable on the LTQ. Samples were dissolved in 50/50 methanol/water for direct infusion. All ions were sodium adducts. Peak selection for MSⁿ disassembly was performed manually. Data collection and analysis were performed as described (Ashline et al. 2005,  2007; Ashline, Hanneman, et al. 2014; Ashline, Yu, et al. 2014).

Funding

This work was supported by the Ministry of Science and Higher Education of Poland (No N N302 281436).

Abbreviations

EBA-140, erythrocyte-binding antigen 140; Ge, Gerbich; GPC, Glycophorin C; GPA, Glycophorin A; RBCs, red blood cells (erythrocytes); SDS, sodium dodecylsulfate.