|Home | About | Journals | Submit | Contact Us | Français|
LT-IIb, a type II heat-labile enterotoxin of Escherichia coli, is a potent immunologic adjuvant with high affinity binding for ganglioside GD1a. Earlier study suggested that LT-IIb bound preferentially to the terminal sugar sequence NeuAcα2-3Galβ1-3GalNAc. However, studies in our laboratory suggested a less restrictive binding epitope. LT-IIb(T13I), an LT-IIb variant, engineered by a single isoleucine-threonine substitution, retains biological activity, but with less robust inflammatory effects. We theorized that LT-IIb has a less restrictive binding epitope than previously proposed and that immunologic differences between LT-IIb and LT-IIb (T13I) correlate with subtle ganglioside binding differences. Ganglioside binding epitopes, determined by affinity overlay immunoblotting and enzymatic degradation of ganglioside components of RAW264.7 macrophages, indicated that LT-IIb bound to a broader array of gangliosides than previously recognized. Each possessed NeuAcα2-3Galβ1-3GalNAc, although not necessarily as a terminal sequence. Rather, each had a requisite terminal or penultimate single sialic acid and binding was independent of ceramide composition. RAW264.7 enterotoxin-binding and non-binding ganglioside epitopes were definitively identified as GD1a and GM1a, respectively, by enzymatic degradation and mass spectroscopy. Affinity overlay immunoblots, constructed to the diverse array of known ganglioside structures of murine peritoneal macrophages, established that LT-IIb bound NeuAc- and NeuGc-gangliosides with nearly equal affinity. However, LT-IIb(T13I) exhibited enhanced affinity for NeuGc-gangliosides and more restrictive binding. These studies further elucidate the binding epitope for LT-IIb and suggest that the diminished inflammatory activity of LT-IIb(T13I) is mediated by a subtle shift in ganglioside binding. These studies underscore the high degree of specificity required for ganglioside–protein interactions.
The type II heat-labile enterotoxin LT-IIb of Escherichia coli belongs to a superfamily of bacterial toxins that also includes cholera toxin (CT) and heat-labile toxins LT-I and LT-IIa of E. coli. Each acts by binding to ganglioside surface receptors on susceptible host cells. Cholera toxin binds primarily to GM1, although it also exhibits weaker affinity binding to GM2, GD1a, GM3, GT1b, and GD1b (Bennett and Cuatrecasas 1975). LT-I also binds primarily to GM1, but has some affinity for one or more glycoproteins (Orlandi et al. 1994). Binding of LT-IIa has been well characterized and involves GD1b, GM1, GT1b, GQ1b, GM2, GD1a, and GM3 (Fukuta et al. 1988).
The ganglioside binding characteristics of LT-IIb are less well characterized, but include high affinity binding to GD1a (Fukuta et al. 1988). Although GD1a and GM1a both have a gangliotetraose core structure, LT-IIb exhibits no detectable binding affinity for GM1a (Fukuta et al. 1988). Earlier investigations indicated that LT-IIb bound preferentially to gangliosides with terminal sugar sequences of NeuAcα2-3Galβ1-3GalNAc. However, LT-IIb also bound to GM3, which lacks this terminal sugar sequence, suggesting a less restrictive binding epitope for LT-IIb (Fukuta et al. 1988). Furthermore, previous characterization of LT-IIb-ganglioside interactions was limited to commercial gangliosides that are predominantly sialylated with N-acetyl neuraminic acid (NeuAc) and lack some structural attributes of gangliosides of immunologic cells that can be critical to immune interactions. For example, the respiratory pathogen nontypeable Haemophilus influenzae binds to human gangliosides with a neolacto carbohydrate core structure, that is present in human macrophages and respiratory epithelial cells, but is absent from most commercial ganglioside preparations (Fakih et al. 1997; Berenson et al. 2005).
The capacity to bind to membrane gangliosides has been firmly correlated with the potent immunoregulatory properties of LT-IIb. Intranasal co-administration of LT-IIb with a weak antigen greatly enhanced the immune response to that antigen in an established murine mucosal immunization model (Nawar et al. 2005; van Ginkel et al. 2005). However, the inherent toxicity of LT-IIb for most human cells has limited its practical use as an adjuvant. To attempt to diminish toxicity, but retain adjuvant immunologic effects, an LT-IIb variant, LT-IIb (T13I), was engineered, by genetically substituting isoleucine for threonine at position 13 in the B polypeptide of the enterotoxin (Connell and Holmes 1995). The effects of this substitution were investigated in a murine mucosal immunization model, where LT-IIb(T13I) retained full adjuvant properties of LT-IIb. Ganglioside-specific ELISA showed that LT-IIb(T13I) had no detectable binding activity to gangliosides bound by wild-type LT-IIb and the toxicity of LT-IIb(T13I) was dramatically reduced in comparison with the wild-type LT-IIb in an in vitro bioassay (Nawar et al. 2005). While the amino acid substitution altered toxicity of the enterotoxin, it did not appear to alter the overall pattern of the physical interaction with immunocompetent cells. Flow cytometric analyses revealed that both LT-IIb and LT-IIb(T13I) bound to macrophages, dendritic cells, T cells, and B cells from the cervical lymph nodes of immunized mice (Nawar et al. 2005). Collectively, these results suggest that the biological activities of LT-IIb(T13I) are mediated through binding to glycosphingolipid receptors distinct from those of LT-IIb.
We theorized that LT-IIb has a less restrictive binding epitope than previously reported and also that the differences in immunologic effects between LT-IIb and LT-IIb (T13I) correlate with subtle differences in ganglioside binding and designed experiments to test this hypothesis.
To confirm specificity of LT-IIb and LT-IIb(T13I) binding to gangliosides on thin-layer chromatograms and for comparison of chromatographic binding to previous ELISA binding results (Connell and Holmes 1995; Nawar et al. 2005), each was bound by TLC affinity overlay to commercial gangliosides (Figure (Figure1).1). Two separate glycolipid preparations were used. LT-IIb bound to GD1a, GT1b, GM2, and GM3 (Figure (Figure1,1, top panel, lanes numbered 1). In contrast, LT-IIb (T13I) at the same concentration bound only to GD1a (Figure (Figure1,1, bottom panel, lanes numbered 1) and failed to bind to GM3, GM2, and GT1b. In all experiments, neither enterotoxin bound to GM1a. Control lanes, consisting of buffer alone, confirmed the absence of nonspecific binding of buffer or blocking components to gangliosides (Figure (Figure1,1, top and bottom panels, lane 2).
To optimize detection of binding of enterotoxins to specific gangliosides, optimal concentrations of gangliosides for TLC affinity overlay assays had to be determined. To establish optimal concentrations and conditions for TLC affinity overlay binding, LT-IIb and LT-IIb(T13I) were bound in increasing concentrations to commercial ganglioside preparations on chromatograms (Figure (Figure2).2). LT-IIb bound to GD1a at all concentrations (0.05–0.5 μg/mL). However, lower affinity binding to GM3 and GM2, and even weaker binding to GD3, was detected at higher concentrations (Figure (Figure2,2, top panel).
In contrast, LT-IIb(T13I) bound with relatively weaker affinity to GD1a, compared with LT-IIb binding (Figure (Figure2,2, bottom panel). At higher concentrations, weak affinity binding to GM3 was seen. These experiments established optimal dose-concentration conditions for further TLC affinity overlay studies.
Flow cytometric assays using splenic cells or cells from cervical lymph nodes showed that LT-IIb binds to murine macrophages (Nawar et al. 2005). To establish the role of macrophage gangliosides on LT-IIb and LT-IIb(T13I) binding, RAW264.7 cells were treated by separate enzymatic degradations to remove surface moieties.
To investigate whether sialic acid plays a role in macrophage binding of LT-IIb(T13I), RAW264.7 macrophages were treated with Clostridium perfringens sialidase, which hydrolyzes α-(2,3), α-(2,6), and α-(2,8) anomeric linkages, but may omit internalized sialic acids (Yohe et al. 2001; Berenson et al. 2005). Following sialidase treatment, LT-IIb, LT-IIb(T13I), and cholera toxin (CT) were bound to treated and untreated cells. Sialidase treatment diminished binding of both LT-IIb and of LT-IIb(T13I) compared with buffer-treated controls (Figure (Figure3,3, Panel A). To confirm results, experiments were repeated with a different sialidase from Arthrobacter ureafaciens (Yohe et al. 2001; Berenson et al. 2005). Treatment with A. ureafaciens sialidase also diminished binding of LT-IIb and of LT-IIb(T13I) to RAW264.7 cells, compared with buffer-treated controls (Fig- ure ure3,3, Panel B). Treatment with either sialidase reduces more heavily sialylated gangliosides to GM1, the main receptor for CT. Thus, as expected, sialidase treatment of RAW264.7 macrophage cells enhanced binding of CT. These data support sialylated glycoconjugates as receptors for LT-IIb and LT-IIb(T13I).
To further confirm the role of gangliosides in LT-IIb(T13I) binding, RAW264.7 cells were deglycosylated with ceramide glycanase, which cleaves the carbohydrate moiety of glycosphingolipids (Li et al. 1986). Enzymatic deglycosylation of RAW264.7 cells significantly decreased binding of all three toxins, including LT-IIb(T13I), to RAW264.7 cells, compared with buffer-treated controls (P < 0.01, Figure Figure3,3, Panel C). These data further support glycosphingolipids as receptors for LT-IIb and LT-IIb(T13I).
To further characterize the receptor for LT-IIb(T13I), RAW264.7 cell monolayers were pretreated with either sodium metaperiodate, to oxidize surface-accessible sugars, or with proteinase K, to cleave external protein residues, before incubating cells with LT-IIb, LT-IIb(T13I), or CT. Pretreatment with sodium periodate reduced binding of CT, LT-IIb, and LT-IIb(T13I), compared with untreated controls (Figure (Figure3,3, Panel D). In contrast, there was no statistically significant decrease in binding of CT, LT-IIb, or LT-IIb(T13I) after treatment with proteinase K (Figure (Figure3,3, Panel D). Separate cell monolayers were treated with proteinase K and stained with anti-CD11b as controls, to confirm proteolytic activity. Results of three separate experiments confirmed reduced anti-CD11b binding (mean ± SEM) for proteinase K-treated cells (12.3 ± 1.2 RFU) compared with control cells (318.9 ± 12.1 RFU) (P ≤ 0.001). The same experiments performed with LT-IIb, LT-IIb(T13I), and CT and J774.1 and M12 cells produced the same results (data not shown).
Collectively, results of enterotoxin binding to enzymatically modified macrophages specifically support gangliosides as receptors for each.
LT-IIb not only binds to RAW264.7 macrophages, but also induces cAMP production (Nawar et al. 2005). To precisely characterize the specific ganglioside receptors for LT-IIb present on RAW264.7 cells, each toxin was overlaid on purified RAW264.7 cell gangliosides on TLCs (Figure (Figure4).4). RAW264.7 cell gangliosides consisted of seven major groups, including two predominant spots that comigrated near, but not precisely with, GM1 and GD1a standards in both solvent dimensions. The latter two are identified in Figure Figure44 (Panel B) by dotted and solid arrows, respectively.
LT-IIb bound to most gangliosides of RAW264.7 cells (Figure (Figure4,4, Panel A), most notably to the major ganglioside nearest to the origin, whose position is indicated by a solid arrow. However, there was no binding to the other major ganglioside, whose position is indicated by a dotted arrow (Figure (Figure4,4, Panel A).
LT-IIb (T13I) bound to the same gangliosides as did LT-IIb, but with weaker overall affinities (Figure (Figure4,4, Panel C). Included were binding to the major ganglioside peak nearest to the origin (solid arrow) and the absence of binding to the other major ganglioside (dotted arrow).
The previous experiments established specific binding interactions of enterotoxins with the major gangliosides of RAW264.7 cells. To more precisely determine the structural characteristics of the major RAW264.7 cell gangliosides, purified gangliosides were extracted and were analyzed for their susceptibility to C. perfringens sialidase. Conditions were verified using commercial samples of known structures (Figure (Figure5,5, Panel C) and demonstrated that GM3, GD3, GD1a, and GD1b, each of which has an external sialic acid, were no longer resorcinol-positive after treatment with C. perfringens sialidase. GM1a and GM2, each of which possesses internal sialic acids, were sialidase resistant.
Under the same conditions, the two major gangliosides of RAW264.7 cells were evaluated for susceptibility to C. perfringens sialidase. The major enterotoxin-binding ganglioside peak of RAW264.7 cells (solid arrow) was no longer resorcinol-positive after enzymatic treatment and thus was sialidase-susceptible (Figure (Figure5,5, Panel B, solid arrow). However, the major ganglioside peak (dotted arrow) that failed to bind either LT-IIb or LT-IIb(T13I) (Figure (Figure5,5, Panel B, dotted arrow) remained resorcinol positive after enzymatic treatment and thus was sialidase resistant, consistent with structural characteristics of GM1a. Furthermore, the relative quantity of this peak (dotted arrow) increased with sialidase treatment, consistent with generation of GM1a from desialylation of more heavily sialylated gangliosides. Gangliosides of RAW264.7 cells treated with buffer alone are shown in Figure Figure5,5, Panel A, and were unchanged from untreated gangliosides.
The major singly charged components of the first peak that did not bind LT-IIb had a mass-to-charge ratio of m/z 1933 and 1821, and were fragmented to verify the glycan and ceramide structures (Figure (Figure6).6). The ion at m/z 1933 was selected for CID and the resulting MS2 spectrum is shown in Figure Figure6B.6B. The peak at m/z 1558 represents the loss of a single sialic acid (375 amu). The peak at m/z 1470 represents the loss of a terminal Hex-HexNAc disaccharide (463 amu). The peak at m/z 1273 represents the loss of the ceramide moiety (660 amu). The peak at m/z 1094 represents the loss of both a terminal Hex-HexNAc disaccharide (463 amu) and a sialic acid. The ceramide moiety is observed as a minor peak at m/z 660. The glycan moiety at m/z 1273 was fragmented, yielding the subsequent MS3 spectrum (Figure (Figure6C),6C), suggesting that the major structure has Neu5Ac on the second sugar. The large peak at m/z 898 corresponds to loss of the sialic acid. In addition, an isomer of GM1a is present. Terminal sialic acid, which yields a peak at m/z 847 (Neu5Ac-Gal-GalNAc), is detectable in the MS3 spectrum, as is an internal Hex-HexNac disaccharide at m/z 472.
The ceramide moiety was characterized by fragmenting the m/z 660 ion to generate the MS3 spectrum (Figure (Figure6D).6D). Fragmentation of the ceramide shows the sphingosine chain at m/z 310, which infers a C24 fatty acid chain. To determine whether the mass difference between m/z 1933 and m/z 1822 is due to a variation in the glycan or the ceramide, m/z 1822 was fragmented (data not shown) and in the MS2 spectrum of m/z 1822, the glycan did not vary (m/z 1273 ion is present). The mass fragments that contain the ceramide ion vary by 112 amu. Fragmentation of the ceramide at m/z 548, a mass difference of 112 amu, provides the sphingosine fragmentations at m/z 278 and m/z 310. The variation, it is inferred, must lie in the fatty acid chain, consistent with a C16 fatty acid chain. The mass difference between the two is 112 amu, correlating to four units of 28 amu inferring a variation in the ceramide moiety. The peak at m/z 1933 indicates the presence of a sialylglycosphingolipid with a C24 fatty acid chain. The peak at m/z 1822 indicates a C16 fatty acid chain. The groups of smaller peaks surrounding these major peaks were undermethylated sample and not variations of these structures. Thus, serial collisions identify this peak as species of GM1a (II3NeuAc-GgOse4Cer), whose respective ceramide structures each have C18 sphingosine, but differ in fatty acyl chains lengths of C16 and C24 (Figure (Figure66).
The second ganglioside peak comprised the major LT-IIb-binding ganglioside of RAW264.7 cells (Figure (Figure7).7). The larger mass of GD1a gangliosides primarily results in doubly charged ions in the mass spectrum. The m/z 1158.7 ion (doubly charged) was selected for fragmentation and the resulting MS2 spectrum is shown in Figure Figure7B.7B. Both singly- and doubly charged fragment ions are apparent. As with the GM1a sample, a singly charged ion at m/z 660 is the putative ceramide fragment. The doubly charged m/z 828.9 ion is the complete glycan moiety. The singly charged m/z 1919.3 (m/z 971.2 doubly charged) ion is formed by loss of a single sialic acid. The m/z 1544 is formed by loss of two sialic acids. The m/z 1470 is formed by loss of a Neu5Ac-Gal-GalNAc trisaccharide fragment.
The complete oligosaccharide fragment, m/z 828 (doubly-charged), was isolated and fragmented to obtain the MS3 spectrum (Figure (Figure7C).7C). The relevant fragmentation assignments are shown in Figure Figure7E.7E. Thus, serial collisions identify this peak as species of GD1a (IV3NeuAc,II3NeuAc-GgOse4Cer), differing in their respective ceramide structures (Figure (Figure7).7). The ceramide moiety (m/z 660) was fragmented and the resulting MS3 spectrum shows ceramides comprising C16 and C24 fatty acids, identical to those of the GM1a structure (Figure (Figure66).
In summary, the major species of the first (LT-IIb-nonbinding) and second (LT-IIb-binding) ganglioside peaks are GM1a and GD1a, respectively. Each contained heterogeneous ceramides with a C18 sphingosine base and C16 and C24 fatty acid chains.
LT-IIb interacts with murine macrophages to mediate TLR2-dependent NF-κB activation and cytokine release (Hajishengallis et al. 2005). Gangliosides of thioglycollate-elicited murine peritoneal macrophages are complex and offer a diverse array of gangliosides, beyond those available from commercial sources and whose individual structures have been identified (Yohe et al. 1991, 1997) (Figure (Figure8,8, top panel). These include not only diversity in carbohydrate structure, but also diversity in ganglioside sialylation with both N-acetyl (NeuAc) and N-glycolyl (NeuGc) neuraminic acid. Each is identified by ellipses in the schematic diagram in Figure Figure88 (bottom panel), which correspond to gangliosides of the TLC plate (Figure (Figure8,8, top panel).
To determine specificity of binding to the array of structures in murine peritoneal macrophage gangliosides, each enterotoxin was overlaid on TLC preparations (Figure (Figure9).9). LT-IIb bound to most gangliosides, and as expected from our earlier studies, had high affinity binding for all species of GD1a. LT-IIb also bound to GM2 and GM3 (Figure (Figure9,9, top panel). However, previously unrecognized LT-IIb-binding gangliosides included GM1b and GD1alpha. Although LT-IIb bound to GM1b, it demonstrated no binding for either NeuAc or NeuGc glycoconjugates of GM1a. In contrast, LT-IIb(T13I) bound species of GD1a, GM2, and GM3, but with weaker affinity.
To more closely analyze differences in ganglioside binding, three major regions of the TLC were evaluated in greater detail by scanning densitometry. Densitometric data shown in Table TableII correspond to images in Figure Figure8.8. Gangliosides in each region contain NeuAc species, denoted by a solid arrow, and NeuGc species, denoted by a dotted arrow. Region 1 comprises species of GM3. In Region 1, LT-IIb bound to GM3-NeuAc and to GM3-NeuGc in nearly equal proportions (0.8:1.0) (Figure (Figure9,9, top panel, left inset and Table TableI).I). However, LT-IIb(T13I) bound with relative greater affinity for GM3-NeuGc (NeuAc:NeuGc – 0.4:1.0) (Figure (Figure9,9, top panel, lower inset and Table TableII).
Region 2 contains GM1b gangliosides. LT-IIb bound to GM1b-NeuAc with slightly greater affinity than to GM1b-NeuGc (NeuAc:NeuGc-1.3:1.0) (Figure (Figure9,9, top panel, lower inset and Table TableI).I). However, LT-IIb(T13I) bound preferentially to GM1b-NeuGc (NeuAc:NeuGc – 0.5:1.0) (Figure (Figure9,9, bottom panel, lower inset).
Region 3 consists of GD1a gangliosides. LT-IIb bound to GD1a-NeuAc and GD1a-NeuGc in nearly equal proportions (NeuAc:NeuGc – 0.8:1.0) (Figure (Figure9,9, top panel, right inset and Table TableI).I). LT-IIb(T13I) again bound with relatively higher affinity to GD1a-NeuGc compared with NeuAc-GD1a (NeuAc:NeuGc – 0.4:1.0) (Figure (Figure9,9, bottom panel, right inset).
Collectively, these data indicated that, compared to LT-IIb, LT-IIb(T13I) bound preferentially to NeuGc-gangliosides of murine macrophages.
Gangliosides function as receptors for a wide variety of bacteria and bacterial products (Hakomori 2000). Notable among bacterial products are Staphylococcus aureus alpha toxin (Kato and Naiki 1976), Clostridium botulinum toxins (Simpson and Rapport 1971a, 1971b), tetanus toxin (Stoeckel et al. 1977), and cholera toxin (Bennett and Cuatrecasas 1975). While best established as receptors for cholera toxin, gangliosides are also critical ligands for E. coli heat labile toxins, including LT-I and LT-IIb (Bennett and Cuatrecasas 1975; Fukuta et al. 1988). Each bacterial product has adapted to bind to specific glycosphingolipid molecules. For example, cholera toxin binds with highest affinity to GM1a, while LT-IIb binds to GD1a, but not to GM1a (Bennett and Cuatrecasas 1975; Fukuta et al. 1988). Gangliosides have long been recognized as having key interactions with the innate immune system. TLR4-deficient mice have dramatic shifts in LPS-induced macrophage ganglioside expression and surface accessibility, attributable to alteration of downstream events dependent upon TLR4 cell activation (Yohe et al. 1991; Macala and Yohe 1995). Thus, one focus of our study was the investigation of LT-IIb binding to ganglioside structures that exist in immunological tissues, in addition to those included in earlier studies.
Results of this investigation revealed a broader array of LT-IIb-binding gangliosides than was previously known (Fukuta et al. 1988) and more detail of the LT-IIb ganglioside-binding epitope, by several independent methods, including affinity chromatography, enzymatic degradation, and mass spectroscopy (Figure (Figure10).10). Previous study indicated that, aside from GD1a, LT-IIb also bound to GT1b, suggesting a terminal sugar epitope of NeuAcα2-3Galβ1-3GalNAc. In fact, in the current study, LT-IIb binding to GT1b appeared comparable to GD1a, suggesting GT1b as an alternative receptor. In both studies, weaker binding to GM3 was also detected in both studies indicating that LT-IIb has the ability to bind weakly to a truncated epitope. Not previously recognized was LT-IIb binding to GM2, whose structure includes the trisaccharide epitope common to other LT-IIb binding gangliosides, although with GalNAc at its terminus. Although LT-IIb does not bind to GM1a, it does bind to GM1b, whose structure also includes NeuAcα2-3Galβ1-3GalNAc. Also newly recognized was binding to GD1alpha, whose structure includes an additional sialic acid attached to the second sugar (GalNAc) by an alpha 2,6 anomeric linkage. Thus, while inclusion of an additional sialic acid on the third sugar, as occurs with GM1a, is prohibitive, inclusion of an additional sialic acid on the second sugar is permissive for LT-IIb-ganglioside binding. Each LT-IIb-binding ganglioside possesses NeuAcα2-3Galβ1-3GalNAc, although not necessarily as a terminal sequence. Rather, each has a requisite terminal or penultimate single sialic acid. Figure Figure1010 summarizes LT-IIb-binding and nonbinding gangliosides.
While study of ganglioside binding has traditionally focused on adherence via the oligosaccharide portions (Schengrund 1995), heterogeneity in ceramide fatty acyl chain length is a potential contributor, not only to glycosphingolipid–protein interactions (Kronke 1999), but also to membrane organization and function of neutrophils (Iwabuchi et al. 2008; Sonnino et al. 2009). Our own experience confirms a vital role for long chain fatty acids of ceramide of human macrophage gangliosides in protein interactions (Berenson et al. 2002). Therefore, it was critical to the current study to determine that heterogeneity of RAW cell ganglioside ceramides was not a determinant of either LT-IIb or LT-IIb (T13I) interactions with either GM1a or GD1a and that LT-IIb-ganglioside interactions are primarily dependent on carbohydrate composition of ganglioside receptors.
Although both LT-IIb and LT-IIb(T13I) bind to macrophage gangliosides, our studies delineate a distinct shift in sialic acid preferential binding by each. LT-IIb (T13I) differs from LT-IIb by a single amino acid, yet evokes less robust biological responses (Nawar et al. 2005; Liang et al. 2007). Close inspection of affinity overlay immunoblots indicated minor binding differences within small groups of gangliosides. To better evaluate, we used a ganglioside template that offered greater diversity of gangliosides, containing known structures that are absent from commercial gangliosides. Thioglycollate-elicited murine peritoneal macrophages have an array of 28 distinct ganglioside peaks on 2-D TLC (Yohe et al. 1991, 1997) with broad diversity in carbohydrate and ceramide structures and in sialic acid composition. While LT-IIb and LT-IIb(T13I) both bound to the same general groups of gangliosides, a distinct shift of relative binding of LT-IIb(T13I) toward NeuGc-gangliosides was evident in each chromatographic region. By comparison, binding preferences of LT-IIb were more evenly distributed between NeuAc and NeuGc components of each groups of gangliosides. In addition, LT-IIb(T13I) did not bind to GT1b, indicating that the amino acid alteration in the B subunit of LT-IIb inhibits binding to more heavily sialylated gangliosides that still retain the NeuAcα2-3Galβ1-3GalNAc epitope.
NeuGc differs from NeuAc by a single hydroxylation, catalyzed by CMP-hydroxylase (Varki 2009). NeuGc is prominent on cell surface glycoconjugates in most primates, among whom the notable exception is humans (Varki 2009). An inactivating mutation in the CMAH gene, responsible for synthesis of NeuGc, is estimated to have occurred in hominid ancestors of humans 2–3 million years ago, eliminating the capability of human cells to make NeuGc-glycoconjugates, with important consequences for human immune interactions with pathogens. LT-IIb and other AB5 toxins comprise an A subunit and pentameric B subunits, and the ability to distinguish between NeuAc and NeuGc glycoconjugates may be a feature of other AB5 toxins. Among AB5 toxins, subtilase cytotoxin of E. coli contains a B subunit that preferentially binds NeuGc glycans units (Byres et al. 2008). Thus, it is intriguing that wild-type LT-IIb, also an AB5 toxin, binds NeuGc and NeuAc gangliosides almost equally, while a shift to preferential binding of NeuGc gangliosides is brought about by the B subunit amino acid alteration in LT-IIb(T13I). Differences in recognition of sialic acids, associated with preferential binding of antigens, are posited as evidence for species-specific susceptibility to some micro-organisms (Martin et al. 2005) and may also confer species-susceptibility to specific enterotoxins. Diversity in cell surface sialic acid presentations also plays key roles in microbial regulation of immune responses (Crocker and Varki 2001; Varki and Varki 2007). Therefore, it is not surprising that the shift in sialic acid preference toward NeuGc-gangliosides, and a more restrictive ganglioside binding preference of LT-IIb(T13I) compared with LT-IIb, accompanies a change in immune regulation. One limitation of the current study is the focus on binding to NeuAc- and NeuGc-gangliosides. It is conceivable that either toxin may bind preferentially to gangliosides bearing other sialic acids, such as O-acetylated sialic acids, that were cleaved as a result of methodology.
The specificity of structure, including that of individual sialic acids, has long been recognized as a critical component in ganglioside–protein associations (Osborne et al. 1982; Campanero-Rhodes et al. 2007). Our findings of a less restrictive binding epitope on gangliosides for LT-IIb and of a shift in ganglioside recognition exhibited by single amino acid substitution of LT-IIb(T13I) exemplify the relationship of specificity of structure to ganglioside–protein interactions.
Eagle's minimum essential medium (MEM), fetal bovine serum (FBS), and trypsin-EDTA were purchased from Gibco (Grand Island, NY). Purified bovine brain gangliosides were purchased from Matreya LLC, Pleasant Gap, PA.
All organic solvents were of standard analytical high-performance liquid chromatographic grade (Baker Chemical Co., Phillipsburg, NJ). All chemicals were standard analytical reagent quality. High-performance silica gel 60 thin-layer chromatography (TLC) plates were purchased from E. Merck, Darmstadt, Germany.
RAW264.7 cells were purchased from the American Type Culture Collection (Rockville, MD). Murine peritoneal gangliosides were obtained from 8-week-old C3H/HeN mice as previously described (Fakih et al. 1997). Mice were euthanatized 4 days after intraperitoneal injection of 1 mL of 10% thioglycollate broth per mouse. Cells were recovered by peritoneal lavage with Hanks’ balanced salt solution. Cell suspensions were centrifuged at 10°C at 1000 RPM (200 × g) for 10 min and then resuspended. Approximately 2.5 × 107 cells in 10 mL MEM/10% FBS were incubated on glass petri dishes for 90 min at 37°C with 5% CO2 to permit adherence of macrophages. Adherent cells were washed with phosphate buffered saline (PBS) to remove nonadherent cells and then rinsed with 0.31 M pentaerythritol to remove excess salts. Gangliosides were extracted as described below.
Surface glycoconjugates of RAW264.7 cell monolayers were desialylated with C. perfringens sialidase (Sigma Chemical Co., St. Louis, MO) or with A. ureafaciens sialidase (EY Laboratories Inc., San Matteo, CA). Monolayers were incubated with either enzyme (0.1 U) for 90 min at 37°C in the 50 mM sodium acetate buffer (pH 5.5) before measuring binding of enterotoxins (Berenson et al. 2002). In all of the following experiments, binding of enterotoxins was measured as previously detailed (Nawar et al. 2005). Briefly, enterotoxins CT, LT-IIb, or LT-IIb(T13I) were incubated (100 μL of 1 μg/mL) in each well. After removal of unbound enterotoxins, bound enterotoxins were measured by adding 100 μL of rabbit anti-CT or LT-IIb (1:5000 in 1% BSA/PBS) followed by alkaline phosphatase-conjugated goat anti-rabbit IgG (1:4000 in 1% BSA/PBS, Southern Biotech, Birmingham. AL). After washing, nitrophenyl phosphate (Amresco Inc., Solon, OH) diluted in the diethanolamine buffer was added and color reactions were terminated with 2 M NaOH (100 μL). Optical density was measured at 405 nm (Versamax microplate reader, Molecular Devices, Sunnyvale, CA).
Cell surface glycosphingolipids of RAW264.7 cells were deglycosylated with Macrobdella decora ceramide glycanase (V-Labs, Covington, LA). Monolayers were incubated with ceramide glycanase (0.2 U) for 2 h at 37°C in the buffer containing 50 mM sodium acetate and 10 mM sodium cholate (pH 4.5) (Berenson et al. 2002), before measuring binding of enterotoxins, as detailed above (Nawar et al. 2005).
Surface carbohydrates were disrupted by oxidation after incubation for 1 h at room temperature in the dark with 10 mM sodium metaperiodate (Sigma, St. Louis, MO) in the 50 mM sodium acetate buffer (pH 4.5). Monolayers were washed three times with PBS. The oxidized sugars were reduced by treatment of the cells for 30 min with 50 mM sodium borohydride (Sigma Chemical Co., St. Louis, MO) in PBS (Virkola et al. 1993).
To eliminate surface proteins, monolayers were treated for 15 min at 37°C with 0.1 mg/mL proteinase K (Sigma) in PBS (Virkola et al. 1993).
After treatment, cell monolayers were washed with PBS before the addition of 100 μL of 1.0 μg/mL of CT, LT-IIb, or LT-IIb(T13I) to each well. Plates were incubated at room temperature for 2 h to allow binding of the enterotoxins to the cells. After washing with PBS, 100 μL of rabbit anti-CT or anti-LT-IIb antiserum (diluted 1:5000 in PBS, 1% bovine serum albumin (BSA), and 0.25% Triton X-100) was added to the wells. Plates were developed, as previously described (Fakih et al. 1997). Optical density was measured at 405 nm (Versamax microplate reader, Molecular Devices, Sunnyvale, CA). In all cases, control monolayers of cells were incubated with the relevant buffer without bioactive reagents.
In all instances, separate cell monolayers were treated with proteinase K and stained with anti-CD11b, as controls, to confirm proteolytic activity. Following proteinase K treatment, cell monolayers were washed with PBS before adding 1% BSA/PBS (100 μL) and 0.1 μg of FITC-conjugated anti-mouse CD11b (clone M1/70, BD Biosciences, Franklin Lakes, NJ). Plates were incubated at 4°C for 30 min. After washing with PBS, 1% BSA/PBS (100 μL) was added and anti-CD11b binding was measured (excitation wavelength – 490 nm; emission wavelength – 530 nm) with a Spectramax Gemini XS microplate spectrofluorometer (Molecular Devices, Sunnyvale, CA). Results are expressed as relative fluorescence units (RFU) at 530 nm.
Total lipids were extracted from cells with chloroform:methanol (1:1-v/v) and sonication. Gangliosides were purified from the total lipid extract as previously described (Yohe and Ryan 1986; Berenson et al. 1989). In brief, ganglioside-containing acidic lipids were eluted through a 3 mL column of DEAE-Sephadex A-25 (Sigma Chemical Co., St. Louis, MO), dried by rotary evaporation and hydrolyzed with 0.1 N NaOH at 37°C for 90 min. Samples were neutralized with 0.1 N HCl and desalted on a reverse-phase silica gel column (SepPak, Waters Assoc., Waltham MA). Samples were applied to a 2–3 mL column of Iatrobeads 6RS-8060 (Iatron Laboratories, Tokyo, Japan) in chloroform–methanol (85:15, v/v). After elution of less polar lipids, the total ganglioside fraction was eluted with chloroform–methanol (1:2, v/v) and dried by rotary evaporation.
Purified gangliosides were run in two dimensions on TLC plates in chloroform–methanol–0.25% KCl (50:45:10) (Solvent 1) and, after rotating the TLC plate 90° counterclockwise, in chloroform–methanol–0.25% KCl in 2.5 N NH4 in 0.25% aqueous KCl (50:40:10) (Solvent 2). Distinct ganglioside peaks were visualized by heating the TLC plates to 92–94°C after spraying with the resorcinol reagent (Yohe and Ryan 1986; Berenson et al. 1989).
Gangliosides containing 5–10 μg sialic acid were incubated with C. perfringens sialidase (2 U/mL) (Sigma Chemical Co., St. Louis, MO) in 0.5 mL of 50 mM sodium citrate-phosphate buffer, pH 5.5, at 37°C, for 2 h, as previously described (Berenson et al. 2002). A duplicate sample of equal quantity of gangliosides was incubated in the buffer alone. Reactions were terminated by the addition of 0.1 M NaOH and neutralized with 0.1 M HCl. Solutions were desalted on SepPak (Waters Assoc., Milford, MA) columns and evaluated on TLC for hydrolytic products. TLCs were run in two dimensions and were sprayed with resorcinol and heated (92–94°C). Resorcinol-positive intensity was quantitated by scanning densitometry and desialylation was determined by loss of resorcinol positivity, compared with untreated samples. The presence of resorcinol-negative spots on sialidase-treated samples was confirmed on TLCs, by reversible staining with iodine vapor, prior to resorcinol spraying (Berenson et al. 2002).
To verify efficacy of enzymatic activity, positive and negative control gangliosides were concomitantly treated with sialidase, including gangliosides with sialidase-susceptible external sialic acid residues (GM3) and gangliosides with sialidase-resistant internal sialic acid residues (GM1a, GM2), respectively (Matreya, Inc., Pleasant Gap, PA).
Semi-quantitative binding of toxins to gangliosides on TLC plates was determined by a modified immunoblotting method (Magnani et al. 1980). Briefly, TLC chromatograms were immersed for 1 min in 0.1% polyisobutylmethacrylate dissolved in hexane, blocked with 1% BSA in PBS for 30 min, and overlaid with 0.5 μg/mL of LT-IIb or LT-IIb(T13I) in PBS for 1 h. After removing unbound enterotoxin with serial rinses of PBS, plates were overlaid with polyclonal rabbit anti-LT-IIb antiserum (1:5000) and with horseradish peroxidase-conjugated goat anti-rabbit IgG antiserum (1:8000) (Southern Biotech, Birmingham, AL) sequentially, each for 1 h, and then placed on the XAR-5 X-ray film (Kodak, Rochester, NY). Film was developed using enhanced chemiluminescence (Perkin-Elmer Life Sciences, Boston, MA) as previously described (Arnsmeier and Paller 1995).
Densitometric analyses of TLC immu- noblot overlays of LT-IIb and LT-IIb(T13I) binding to gangliosides were performed by two-dimensional analytical scanning (Amersham/Molecular Dynamics, Sunnyvale, CA) (Berenson et al. 2001). To account for possible slight variations in quantities of gangliosides or of enterotoxins on each chromatogram, the relative volume of each individual peak on TLC overlays was determined as a percentage of the total peaks of each region of the TLC.
Each of the two major ganglioside peaks of RAW264.7 cells were separated by preparative two-dimensional TLC. Discrete peaks were recovered from silica scrapings and pooled from several TLC preparations. The first peak comprised the major ganglioside that failed to bind LT-IIb; the second peak included the major LT-IIb-binding ganglioside. Samples were permethylated with iodomethane in dimethyl sulfoxide with powdered sodium hydroxide (Ciucanu and Kerek 1984; Ciucanu and Costello 2003). Permethylated products were extracted with dichloromethane and evaporated under nitrogen stream. Samples were dissolved in methanol for subsequent mass spectrometric analysis, as described previously for murine macrophage (Reinhold et al. 1994; Yohe et al. 1997) and human macrophage glycolipids (Yohe et al. 2001).
Sequential mass spectrometry was performed on an ion trap mass spectrometer (LTQ, ThermoElectron, San Jose, CA) with a nanoelectrospray source. Samples were infused at flow rates of 0.30 μL/min. MSn spectra were collected using Xcalibur 1.4 software (Thermo Electron, San Jose, CA). Signal averaging was accomplished by increasing the number of microscans averaged for each scan and averaging over multiple scans (30–100). Collision-induced dissociation (CID) parameters were left at default values, normalized collision energy set to 35%, activation Q at 0.25 and activation time at 30 ms.
National Institutes of Health (1RO1HL6654901 to C.S.B., DE13833 to T.D.C., and DE015254 and DE017138 to G.H.) and research support of the Department of Veteran's Affairs (to C.S.B.).
The authors are grateful for the technical assistance of Ms. Kate E. Jank and Ms. Mary Alice Garlipp in performing these studies and for the assistance of Timothy F. Murphy, M.D. for critically reading this manuscript.
University at Buffalo, South Campus
Buffalo, NY 14215