Crystal structures exploring the origins of the broader specificity of escherichia coli heat-labile enterotoxin compared to cholera toxin

Cholera toxin (CT) and Escherichia coli heat-labile enterotoxin (LT) are structurally and functionally related and share the same primary receptor, the GM1 ganglioside. Despite their extensive similarities, these two toxins exhibit distinct ligand specificities, with LT being more promiscuous than CT. Here, we have attempted to rationalize the broader binding specificity of LT and the subtle differences between the binding characteristics of LTs from human and porcine origins (mediated by their B subunit pentamers, hLTB and pLTB, respectively). The analysis is based on two crystal structures of pLTB in complexes with the pentasaccharide of its primary ligand, GM1, and with neolactotetraose, the carbohydrate determinant of a typical secondary ligand of LTs, respectively. Important molecular determinants underlying the different binding specificities of LTB and CTB are found to be contributed by Ser95, Tyr18 and Thr4 (or Ser4 of hLTB), which together prestabilize the binding site by positioning Lys91, Glu51 and the adjacent loop region (50-61) containing Ile58 for ligand binding. Glu7 and Ala1 may also play an important role. Many of these residues are closely connected with a recently identified second binding site, and there appears to be cross-talk between the two sites. Binding to N-acetyllactosamine-terminated receptors is further augmented by Arg13 (present in pLT and some hLT variants), as previously predicted.     

Structure-based exploration of the ganglioside GM1 binding sites of Escherichia coli heat-labile enterotoxin and cholera toxin for the discovery of receptor antagonists

Ganglioside GM1 is the natural receptor for cholera toxin (CT) and heat-labile enterotoxin (LT), which are the causative agents of cholera and traveler's diarrhea, respectively. This observation suggests that small molecules interfering with this recognition process may prevent entry of the toxins into intestinal cells, thereby averting their devastating effects. Here, the terminal sugar of ganglioside GM1, galactose, was chosen as a lead in designing such receptor antagonists. Guided by the experimentally determined binding mode of galactose, we selected a "substructure" for searching the Available Chemicals Database, which led to the purchase of 35 galactose derivatives. Initial screening of these compounds in an LT ELISA revealed that 22 of them have a higher affinity for LT than galactose itself. A structurally diverse subset of these galactose derivatives was selected for determination of IC50 values in the LT ELISA and IC50 values in a CT assay, as well as for the determination of Kd's using the intrinsic fluorescence of LT. The best receptor antagonist found in this study was m-nitrophenyl alpha-galactoside with an IC50 of 0.6 (2) mM in the LT ELISA and 0.72 (4) mM in the CT assay, 100-fold lower than both IC50 values of galactose. Careful analysis of our binding data and comparison with crystal structures led to the derivation of correlations between the structure and affinity of the galactose derivatives. These characteristics will be used in the design of a second round of LT and CT receptor antagonists.     

Common antigenic determinant in the receptor binding sites of Escherichia coli enterotoxin and cholera toxin

Abstract A mutant (TUH No. 9) of a porcine strain of enterotoxigenic Escherichia coli (ETEC) produces as abnormal B subunit (B') of heat-labile enterotoxin (LT), which has aspartate instead of glycine at residue 33 from the N-terminus and does not bind to the receptor, GM1 ganglioside. The antigenicities of the receptor-binding site of LT were analyzed.
The antibody, which could not bind to the B' subunit in the anti-B subunit of porcine LT(LTp)-serum, could bind to cholera toxin (CT), LTp and LT produced by a human ETEC strain (LTh), suggesting that it recognizes a common epitope of LTp, LTh and CT. Thus glycine at residue 33 from the N-terminus in the B subunit of CT, LTh and LTp may be related to the common epitope of these three toxins. The bindings of CT, LTh and LTp to the antibody were inhibited by the GM1 ganglioside.
These data indicate that the antibody recognizes a common epitope in the receptor (GM1 ganglioside)-binding site of CT, LTh and LTp.     

Using a galactose library for exploration of a novel hydrophobic pocket in the receptor binding site of the Escherichia coli heat-labile enterotoxin

The binding of the B subunits of Escherichia coli heat-labile enterotoxin (LT) to epithelial cells lining the intestines is a critical step for the toxin to invade the host. This mechanism suggests that molecules which possess high affinity to the receptor binding site of the toxin would be good leads for the development of therapeutics against LT. The natural receptor for LT is the complex ganglioside GM1, which has galactose as its terminal sugar. A chemical library targeting a novel hydrophobic pocket in the receptor binding site of LT was constructed based on galactose derivatives and screened for high affinity to the receptor binding site of LT. This screening identified compounds that have 2-3 orders of magnitude higher affinity toward the receptor binding site of LT than the parent compound, galactose. The present findings will pave the way for developing simple and easily synthesizable molecules, instead of complex oligosaccharides, as drugs and/or prophylactics against LT-caused disease.     

Comparison of the glycolipid-binding specificities of cholera toxin and porcineEscherichia coli heat-labile enterotoxin: identification of a receptor-active non-ganglioside glycolipid for the heat-labile toxin in infant rabbit small intestine

The binding specificities of cholera toxin andEscherichia coli heat-labile enterotoxin were investigated by binding of¹²⁵I-labelled toxins to reference glycosphingolipids separated on thin-layer chromatograms and coated in microtitre wells. The binding of cholera toxin was restricted to the GM1 ganglioside. The heat-labile toxin showed the highest affinity for GM1 but also bound, though less strongly, to the GM2, GD2 and GD1b gangliosides and to the non-acid glycosphingolipids gangliotetraosylceramide and lactoneotetraosylceramide. The infant rabbit small intestine, a model system for diarrhoea induced by the toxins, was shown to contain two receptor-active glycosphingolipids for the heat-labile toxin, GM1 ganglioside and lactoneotetraosylceramide, whereas only the GM1 ganglioside was receptor-active for cholera toxin. Preliminary evidence was obtained, indicating that epithelial cells of human small intestine also contain lactoneotetraosylceramide and similar sequences. By computer-based molecular modelling, lactoneotetraosylceramide was docked into the active site of the heat-labile toxin, using the known crystal structure of the toxin in complex with lactose. Interactions which may explain the relatively high toxin affinity for this receptor were found.Abbreviations CT cholera toxin - CT-B B-subunits of cholera toxin - LT Escherichia coli heat-labile enterotoxin - hLT humanEscherichia coli heat-labile enterotoxin - pLT porcineEscherichia coli heat-labile enterotoxin - EI electron ionization     

Novel carbohydrate binding site recognizing blood group A and B determinants in a hybrid of cholera toxin and Escherichia coli heat-labile enterotoxin B-subunits

The B-subunits of cholera toxin (CTB) and Escherichia coli heat-labile enterotoxin (LTB) are structurally and functionally related. However, the carbohydrate binding specificities of the two proteins differ. While both CTB and LTB bind to the GM1 ganglioside, LTB also binds to N-acetyllactosamine-terminated glycoconjugates. The structural basis of the differences in carbohydrate recognition has been investigated by a systematic exchange of amino acids between LTB and CTB. Thereby, a CTB/LTB hybrid with a gain-of-function mutation resulting in recognition of blood group A and B determinants was obtained. Glycosphingolipid binding assays showed a specific binding of this hybrid B-subunit, but not CTB or LTB, to slowly migrating non-acid glycosphingolipids of human and animal small intestinal epithelium. A binding-active glycosphingolipid isolated from cat intestinal epithelium was characterized by mass spectrometry and proton NMR as GalNAcalpha3(Fucalpha2)Galbeta4(Fucalpha3)Glc NAcbeta3Galbeta4Glc NAcbeta3Galbeta4Glcbeta1Cer. Comparison with reference glycosphingolipids showed that the minimum binding epitope recognized by the CTB/LTB hybrid was Galalpha3(Fucalpha2)Galbeta4(Fucalpha3)GlcNAc beta. The blood group A and B determinants bind to a novel carbohydrate binding site located at the top of the B-subunit interfaces, distinct from the GM1 binding site, as found by docking and molecular dynamics simulations.     

The role of gangliosides in the interaction of human chorionic gonadotropin and cholera toxin with murine Leydig tumor cells

A clonal line of murine Leydig tumor cells (MLTC-1) bound both human chorionic gonadotropin (hCG) and cholera toxin (CT) with high affinity and accumulated cyclic AMP in response to either effector. The major cellular ganglioside was GM3 with small amounts of GM2, GM1, and GD1a. The gangliosides became labeled when the cells were grown in medium containing [3H] galactose or were exposed to galactose oxidase or NaIO4 followed by NaB3H4. CT specifically protected GM1 from surface labeling whereas hCG did not protect any gangliosides from being labeled. When the cells were exposed to sialidase, surface GD1a was eliminated, and GM1 increased with a corresponding increase in CT binding. When sialidase-treated cells were first incubated with the B component of CT, binding and action of CT was blocked. The cells, however, retained their ability to bind and respond to hCG. Addition of purified gangliosides to the medium effectively inhibited the binding and action of CT but not hCG. The cells incorporated the exogenous gangliosides and exhibited increased binding of and responsiveness to CT but not hCG. Both hCG- and CT-receptor complexes were extracted from the cells with nonionic detergent and analyzed by sucrose gradient centrifugation. The hCG-receptor complex had an apparent molecular weight of 190,000 whereas the CT-receptor complex sedimented only slightly faster than CT itself. MLTC-1 gangliosides were separated on thin layer chromatograms which were overlayed with either iodinated CT or hCG. The toxin bound to a ganglioside corresponding to GM1 whereas the hormone did not bind to any of the gangliosides. When the cells were incubated overnight with hCG, they lost their hCG receptors but exhibited an increase in CT binding and gangliosides. Our results indicate that GM1 is the specific receptor for CT whereas gangliosides are not involved in the binding and action of hCG.     

Structural basis for differential receptor binding of cholera and Escherichia coli heat-labile toxins: influence of heterologous amino acid substitutions in the cholera B-subunit

The closely related B-subunits of cholera toxin (CTB) and Escherichia coli heat-labile enterotoxin (LTB) both bind strongly to GM1 ganglioside receptors but LTB can also bind to additional glycolipids and glycoproteins. A number of mutant CT B-subunits were generated by substituting CTB amino acids with those at the corresponding positions in LTB. These were used to investigate the influence of specific residues on receptor-binding specificity. A mutated CTB protein containing the first 25 residues of LTB in combination with LTB residues at positions 94 and 95, bound to the same extent as native LTB to both delipidized rabbit intestinal cell membranes, complex glycosphingolipids (polyglycosylceramides) and neolactotetraosylceramide, but not to non-GM1 intestinal glycosphingolipids. In contrast, when LTB amino acid substitutions in the 1–25 region were combined with those in the 75–83 region, a binding as strong as that of LTB to intestinal glycosphingolipids was observed. In addition, a mutant LTB with a single Gly-33→Asp substitution that completely lacked affinity for both GM1 and non-GM1 glycosphingolipids could still bind to receptors in the intestinal cell membranes and to polyglycosylceramides. We conclude that the extra, non-GM1 receptors for LTB consist of both sialylated and non-sialylated glycoconjugates, and that the binding to either class of receptors is influenced by different amino acid residues within the protein.     

High-resolution crystal structures of Erythrina cristagalli lectin in complex with lactose and 2'-alpha-L-fucosyllactose and correlation with thermodynamic binding data

The primary sequence of Erythrina cristagalli lectin (ECL) was mapped by mass spectrometry, and the crystal structures of the lectin in complex with lactose and 2'-alpha-L-fucosyllactose were determined at 1.6A and 1.7A resolution, respectively. The two complexes were compared with the crystal structure of the closely related Erythrina corallodendron lectin (ECorL) in complex with lactose, with the crystal structure of the Ulex europaeus lectin II in complex with 2'-alpha-L-fucosyllactose, and with two modeled complexes of ECorL with 2'-alpha-L-fucosyl-N-acetyllactosamine. The molecular models are very similar to the crystal structure of ECL in complex with 2'-alpha-L-fucosyllactose with respect to the overall mode of binding, with the L-fucose fitting snugly into the cavity surrounded by Tyr106, Tyr108, Trp135 and Pro134 adjoining the primary combining site of the lectin. Marked differences were however noted between the models and the experimental structure in the network of hydrogen bonds and hydrophobic interactions holding the L-fucose in the combining site of the lectin, pointing to limitations of the modeling approach. In addition to the structural characterization of the ECL complexes, an effort was undertaken to correlate the structural data with thermodynamic data obtained from microcalorimetry, revealing the importance of the water network in the lectin combining site for carbohydrate binding.     

Characterisation of the binding sites for Escherichia coli heat-labile toxin type I in intestinal brush borders

Intestinal brush borders from Wistar rats contained a total of 20-30-times more binding sites for Escherichia coli heat-labile enterotoxin (LT-1) than for cholera toxin (CT). The results suggest that LT-1 binds to sites in addition to ganglioside GM1, the binding site for CT. Brush border proteins were separated by SDS-PAGE, blotted to nitrocellulose and the filters incubated with 125I-labeled toxins. [125I]LT-1 was shown to bind to a series of brush border galactoproteins ranging in size from 130-140 kDa. Binding was inhibited by unlabeled LT-1 (but not CT), and by ricin and free galactose. A number of brush border enzymes are large glycoproteins which can be solubilised by papain. The papain-solubilised sucrase-isomaltase complex was purified by affinity chromatography and shown to bind LT-1, as did the proteins in fractions enriched in maltase activity. However, such brush border galactoproteins do not account for all of the additional LT-1 binding sites. Thus, brush borders prepared from 1-15-day-old rabbits contained many more binding sites for LT-1 than CT despite the absence of any sucrase-isomaltase activity, and no [125I]LT-1 binding proteins could be detected by blotting. There was a marked variation in the number of LT-1 binding sites in different strains of rat, and between different species.     

Little Heterogeneity among Genes Encoding Heat-Labile and Heat-Stable Toxins of Enterotoxigenic Escherichia coli Strains Isolated from Diarrheal Pigs

To examine whether the heat-labile enterotoxin gene in porcine enterotoxigenic Escherichia coli (ETEC) strains is as divergent as in human ETEC strains, we sequenced the heat-labile and heat-stable toxin genes from 52 and 33 porcine ETEC strains, respectively. We found that the STa gene is identical, that the LT gene has only two mutations in 4 (of 52) strains, and that both mutations cause a reduction in GM1 binding and toxicity.Enterotoxigenic Escherichia coli (ETEC) strains that colonize small intestines and produce enterotoxins are the major cause of diarrheal disease in humans and animals (8, 16, 18, 21). The key virulence factors of ETEC in diarrhea include enterotoxins and colonization factors or adhesins. Colonization factors or adhesins mediate the attachment of bacteria to host epithelium cells and facilitate bacterial colonization. Enterotoxins disrupt fluid homeostasis and stimulate fluid hyper-secretion in the intestinal epithelial cells that results in diarrhea. Heat-labile toxin (LT) and heat-stable toxin (ST) are the main enterotoxins associated with diarrhea in humans and farm animals, but different LT and ST are produced by human and animal ETEC strains (9, 16).The LT produced by porcine ETEC strains (pLT) or human ETEC strains (hLT) is a holotoxin-structured protein that has one LT_A subunit and five LT_B subunits. Although pLT and hLT are highly homologous in structure and function, these two proteins differ antigenetically (9). Sequence comparative studies showed that the following seven amino acids are different between pLT and hLT: the 4th, 213th, and 237th amino acids of the A subunits and the 4th, 13th, 46th, and 102nd amino acids of the B subunits (6, 7). Similarly, STa (ST type 1) carried by human and porcine ETEC strains is also different. The STa associated with porcine diarrhea (pSTa) is a peptide of 18 amino acids, whereas the STa produced by human ETEC strains (hSTa) is 19 amino acids in length (5, 19). Despite the fact that ETEC constructs expressing pLT or hLT, and pSTa or hSTa, are equivalently virulent in causing diarrhea in gnotobiotic pigs (25), pLT and pSTa are typically expressed by porcine ETEC strains that only cause diarrhea in pigs, whereas hLT and hSTa are exclusively produced by human ETEC strains associated with diarrhea in humans. Although pLT and STb, another porcine-specific ST, were occasionally detected in ETEC strains isolated from human diarrheal patients (3), only infections with hSTa⁺, hLT⁺, or hSTa⁺/hLT⁺ ETEC strains cause diarrhea in humans (17).Interspecies LT have been intensively compared for molecular and immunological characteristics (4, 10, 20, 23). In contrast, intraspecies LT has not been studied much. For a long time, both pLT and hLT were assumed to be highly conserved. However, a very recent study showed that the hLT gene carried by human ETEC strains is considerably divergent (12). After restriction fragment length polymorphism analysis and DNA sequencing of 51 human ETEC strains, Lasaro et al. reported that the human LT gene had seven polymorphic restriction fragment length polymorphism types and 30 nucleotide polymorphic sites and recognized 16 different hLT types (12). To examine whether the LT gene carried by porcine ETEC strains has a similar heterogeneity, we PCR amplified and DNA sequenced the LT genes and also the STa genes of various ETEC strains isolated from diarrheal pigs and analyzed gene sequence conformity.Fifty-two porcine ETEC strains that express LT alone or LT together with other toxins (LT⁺/STb⁺, LT⁺/STb⁺/STa⁺, LT⁺/STb⁺/EAST1⁺, and LT⁺/STa⁺/STb⁺/EAST1⁺) and K88ac or F18 fimbria were selected for the sequencing of the LT gene. Those porcine ETEC strains were isolated from pigs with postweaning diarrhea at different farms in South Dakota, Iowa, Minnesota, Nebraska, and North Dakota. The eltAB gene encoding LT from these 52 strains was PCR amplified with primers pLT-F (5′-ATCCTCGCTAGCATGTTTTAT-3′) and pLT-R (5′-CCCCTCCGGCCGAGCTTAGTT-3′) (25). PCRs were performed in an MJ PT-100 thermocycler (Bio-Rad, Hercules, CA) in a reaction of 50 μl containing 1× Taq DNA polymerase buffer (with Mg²⁺), 0.2 mM deoxynucleoside triphosphate, 0.5 μM each forward and reverse primers, 100 ng of total genomic DNA, and 1 unit Taq DNA polymerase (Applied Biosystems, Foster City, CA). The PCR program contained one cycle of 2 min at 94°C; 30 cycles of 35 s at 94°C, 35 s at 52°C, and 2 min at 72°C; and an extension of 6 min at 72°C. The amplified PCR products were separated on 1% agarose gels (FMC Bioproducts, Rockland, MA) by electrophoresis and purified using a QIAquick gel extraction kit according to the manufacturer''s instructions (Qiagen, Valencia, CA). A mixture of purified PCR product (100 to 150 ng) and 10 pmol primer was sent to the Nevada Genomic Center at the University of Nevada for sequencing. Three primers, pLT-F, LT₁₉₂-F (5′-GATTCATCAAGAACAATCCACAGGTG-3′), and LT₁₉₂-R (5′-CCTGTGATTGTTCTTGATGAATC-3′), were used for sequencing the entire eltAB gene.The sequences of the eltAB gene from all 52 porcine ETEC strains were aligned and visually examined. We found that the eltAB gene was nearly identical among the sequenced porcine ETEC strains. Forty-eight (of 52) ETEC strains had identical gene sequences, and only four strains showed heterogeneity. The pathotypes of these four strains were K88/LT/STb, K88/LT/STb/STa, K88/LT/STb/EAST1, and F18/LT/STa/STb/Stx2e. Furthermore, only nucleotides coding two amino acids, the 44th (S44N) and the 60th (S60T) of the eltB gene encoding the B subunit, differed among these four strains. To our surprise, neither of these two substitutions were homologous to the hLT gene nor to any of the hLT types recognized by Lasaro et al. (12). Lasaro et al. showed that 11 of the 15 different hLT types shared some homology with pLT, and some hLT types had as many as four amino acids (K4R and K213E of LT_A and S4T, R13H, or A46E of LT_B; out of seven heterogeneous amino acids) homologous to pLT. Indeed, the hLT6 type differed from the LT of human ETEC prototype {"type":"entrez-nucleotide","attrs":{"text":"H10407","term_id":"875229","term_text":"H10407"}}H10407 in four amino acids (K4R and K213E of LT_A and S4T and A46E of LT_B) (12), but all four of these heterogeneous amino acids were homologous to pLT. Similarly, four of the five amino acids that differed from the prototype hLT in the hLT4 type were identical to pLT. That means that the hLT4 and hLT6 types had only three amino acids heterogeneous to pLT but four different residues compared to the hLT prototype. It seems that hLT4 and hLT6 are more likely pLT rather than hLT. Given that the divergence of the pLT and hLT genes is assumed to be a very recent evolutionary event that occurred 0.9 million years ago (23), it is likely that the hLT gene retains some pLT gene characters (amino acids) that could be of their common ancestor. However, a high homology in the pLT gene certainly seems unparallel to the evolution of the hLT gene. Our further sequence comparison indicated that S44N-substituted pLT_B [pLT_B(S44N)] is homologous to cholera toxin (CT). It has been suggested that the CT and LT genes were derived from the same ancestor but diverged to two lineages about 130 million years ago (23). Then, it is more likely that this pLT_B(S44N) represents a plesiomorphic character, meaning a primitive character that belongs to the common ancestor of CT and LT. The retention of this primitive pLT_B(S44N) by some porcine ETEC strains suggests that the pLT gene could have evolved at a relatively lower rate. Whether such a lower substitution rate of the LT gene in porcine ETEC strains is associated with a lower host exchange rate or a limited travel range in pigs is unclear to us. However, future studies to determine whether an increase in sampling sizes, by including porcine ETEC strains from a greater geographic coverage, could reveal a higher heterogeneity or a greater evolution rate in the pLT gene will be worthwhile.To examine whether the heterogeneity of pS44N and pS60T at the B subunit could affect the biological function of pLT, we cloned the native pLT gene into vector pBR322 (p8458), performed site-directed mutation of the eltAB gene for a substitution of S44N or S60T, and tested these two mutated LT proteins for their binding capability to GM1 receptors and their enterotoxic activity in stimulating intracellular cyclic AMP (cAMP) in cells. Primers pBRNheI-F2 (5′-CAGCATCGCCATTCACTATG-3′) and pBREagI-R (5′-AGATGACGACCATCAGGGAC-3′) were designed to amplify the porcine eltAB gene. The amplified eltAB gene products and vector pBR322 were digested with NheI and EagI (New England Biolabs, Beverly, MA), separated by gel electrophoresis, purified with the QIAquick gel extraction kit, and then ligated with T4 DNA ligase (Promega, Madison, WI). Two microliters of the T4-ligated products were introduced into 25 μl of TOPO cells (Invitrogen, Valencia, CA) in a standard electroporation. Antibiotic-selected colonies were initially screened by PCR, and positive colonies were sequenced to ensure that the cloned gene was in the reading frame. The verified clone was selected as a pLT recombinant strain and designated strain 8458. To construct mutant strains, two pairs of primers, LT_B44-F (5′-ATCATTACATTTAAGAACGGCGAA-3′) and LT_B44-R (5′-TTCGCCGTTCTTAAATGTAATGAT-3′) and LT_B60-F (5′-CAACATATAGACACCCAGAAAAAAGCC-3′) and LT_B60-R (5′-GGCTTTTTTCTGGGTGTCTATATGTTG-3′), were used for site-directed mutation at nucleotides coding the 44th and 60th amino acids of the LT_B subunit, respectively. Briefly, the amplified products from two separate PCRs, one using pBRNheI-F2 with LT_B44-R or LT_B60-R and the other using pBREagI-R with LT_B44-F or LT_B60-F, with recombinant pLT plasmid p8458 as the DNA template, were overlapped in a third splicing overlap extension PCR to produce mutated pLT genes. The splicing overlap extension PCR products were digested with NheI and EagI restriction enzymes and ligated into vector pBR322 for the p8647 (S44N) and p8649 (S60T) plasmids. Plasmids p8647 and p8649 were separately introduced into TOPO 10 E. coli cells (Invitrogen) for mutant strains 8647 (S44N) and 8649 (S60T).Equivalent amounts of cells from overnight-grown cultures of the recombinant (8458) and two mutant (8647 and 8649) strains were used for total protein preparation by using bacterial protein extraction reagent (B-PER in phosphate buffer; Pierce, Rockford, IL). Both pelleted protein samples (periplasmic proteins) and culture supernatant samples (outer-membrane secreted proteins) were used in a GM1 enzyme-linked immunosorbent assay (ELISA) to examine whether a substitution at the 44th or 60th amino acid would affect the binding of LT to GM1 receptors. Anti-CT rabbit antiserum (1:5,000; Sigma) and horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (1:5,000; Sigma, St. Louis, MO) were used as the primary and secondary antibodies as described previously (2, 14, 24). GM1 ELISA data indicated optical density (OD) values from the pellet samples of strains 8548, 8647, and 8649 and phosphate-buffered saline of 0.677 ± 0.004, 0.616 ± 0.001, 0.647 ± 0.004, and 0.006 ± 0, whereas the OD values of the supernatant samples which were vacuum concentrated were 0.949 ± 0.008, 0.726 ± 0.004, 0.660 ± 0.005, and 0.05 ± 0.002, respectively (Fig. ​(Fig.1).1). Statistical analysis using the Student t test with two-tailed distribution indicated that the binding of the pellet samples from the native and the mutated LT to GM1 was not significantly different (P = 0.10 and P = 0.45, respectively). However, the GM1 binding from the supernatant samples of the LT mutant strains was significantly lower than that of the LT recombinant strain (P < 0.01 and P < 0.01, respectively).Open in a separate windowFIG. 1.GM1 ELISA to detect LT proteins expressed by the pLT recombinant (8458) and mutant [8647(S44N) and 8649(S60T)] strains. Protein samples from the pellet and vacuum-concentrated supernatants of overnight-grown cultures were used in the GM1 ELISA. Each sample was assayed in triplicate to calculate OD means and standard deviations. Anti-CT serum (1:5,000) was used as the primary antibody and goat anti-rabbit horseradish peroxidase-conjugated immunoglobulin G (1:5,000) was used as the secondary antibody. OD values were measured after a 20-min reaction with peroxidase substrates (KPL, Gaithersburg, MD) at a wavelength of 405 nm.Our GM1 ELISA data indicated that the supernatant sample of the recombinant strain expressing a native LT had a greater GM1 binding activity. This could suggest that the recombinant strain had more LT protein crossing the outer membrane and being secreted in the supernatant than either mutant strain or that mutations at the B subunit negatively affected the binding of LT proteins to GM1 receptors. It has been reported that a single amino acid mutation of the LT_B or CT_B subunit resulted in lower GM1 binding activity, especially mutations of residues from the binding pocket (13, 15, 22). When amino acid 33 or 88 of the CT_B subunit was replaced, both mutants failed to bind or bound poorly to GM1 (22), and when a substitution at residue 57 of its B subunit occurred, this CT mutant showed 1.5-log-lower GM1 binding than the native CT (1, 13). Similarly, when amino acid 46 or 47 of the B subunit was replaced, both LT mutants exhibited lower GM1 binding activity than the wild-type LT strain (13). However, in contrast to our observation that our 8647 and 8649 mutant strains showed lower GM1 binding activity in the supernatant, Mudrak et al. indicated that the T47A mutant strain had more LT protein detected in the supernatant than the wild-type strain (13). Whether and how a mutation at amino acid 44 or 60 of the B subunit affects the formation, stability, or secretion of the mutant LT proteins will be studied in the future.To examine whether the lower GM1 binding activity of the supernatant samples from the mutant strains was caused by a lower LT production, we conducted an ELISA by directly coating an ELISA plate with total proteins from the pellet and supernatant samples of each strain (without GM1) and by using anti-CT antiserum to quantify the LT protein. ELISA data showed that the OD values of strains 8458, 8647, and 8649 were 0.209 ± 0.005, 0.225 ± 0.009, and 0.21 ± 0 in the supernatant samples and 0.571 ± 0.025, 0.614 ± 0.060, and 0.616 ± 0.026 in the pellet samples, respectively. A Student t test indicated that there were no significant differences between the recombinant strain and the mutant strains in the OD values for the pellet and supernatant protein samples (P = 0.26 and P = 0.84, respectively, for the supernatant samples; P = 0.34 and P = 0.10, respectively, for the pellet samples). These data suggested that a similar amount of LT proteins was produced among these three strains.A single amino acid substitution of the B subunit can result in a reduction in not only GM1 binding but also toxicity for the mutated LT proteins (11, 13, 22). To study whether the mutation of S44N or S60T at the B subunit affected pLT toxicity, we measured the recombinant and mutant strains for their stimulation of intracellular cAMP levels in T-84 cells by using a cAMP competitive enzyme immunoassay (EIA) kit (Invitrogen) by following the manufacturer''s instructions. Briefly, 1 × 10⁵ T-84 cells were seeded in each well of a 24-well plate. After removing the Dulbecco''s modified Eagle medium (DMEM/F12; Gibco/Invitrogen, Grand Island, NY), 75 μl of overnight-grown (in 4AA medium) supernatant of the recombinant or each mutant strain (in triplicate) was added to each well. The cells were lysed with 100 μl of 0.1 M HCl after 2 h of incubation and then neutralized. A total of 100 μl of lysis supernatant was mixed with kit-supplied conjugates and antibody reagents, and the mixture was added to each well of the supplied EIA plate. After incubation on a shaker at 500 rpm at room temperature for 2 h, the plate was washed and dried by blotting, and p-nitrophenyl phosphate substrate solution was added. The OD was measured at 405 nm after 20 min of development. Data from the cAMP ELISA indicated that cAMP levels in T-84 cells incubated with supernatant samples from strains 8458, 8647, and 8649 (from equivalent amounts of cells) were 2.3 ± 0.1, 0.46 ± 0.05, and 0.35 ± 0.01 pmol/ml, respectively (Fig. ​(Fig.2).2). Data clearly indicated that the mutations of S44N and S60T reduced the LT toxic activity. Knowing that it is the A subunit that determines the toxicity of LT and CT, whereas the LT_B and CT_B subunits mediate the binding of the toxin to the host GM1 receptors, we thought that substitution at the B subunits would not affect toxicity. However, we believe that mutations at the B subunits could alter LT protein structure and reduce the binding of the holotoxin to the host GM1 receptors, thus resulting in the reduction of toxic activity.Open in a separate windowFIG. 2.Intracellular cAMP ELISA to detect the toxicity of native LT and mutated LT proteins. Supernatants (in 4AA medium) of overnight-grown cultures from the 8458 (recombinant), 8647 (S44N), and 8649 (S60T) strains were used to stimulate an increase in intracellular cAMP levels in T-84 cells by using a cyclic GMP EIA kit (Invitrogen).The estA gene encoding STa from 33 STa-positive porcine ETEC strains was also sequenced for conformity. This porcine estA gene was PCR amplified using primers pSTaSfcI-F2 and STaEagI-R under conditions described previously (25). The PCR products were purified and sequenced with pSTaSfcI-F2 primer. The sequencing data showed that all sampled STa genes were identical and of porcine origin.Sequence data from our study clearly indicated that both LT and STa expressed by porcine ETEC strains are porcine specific. The LT gene of porcine ETEC strains showed little heterogeneity, and the STa gene is identical. Information from this study will be helpful for a prevalence study of toxin genes among porcine ETEC strains and toxin gene evolution and possibly instructive in antitoxin vaccine development. However, future studies with increasing sampling sizes and a greater geographic coverage will be helpful to understand divergence in the LT and STa genes among porcine ETEC strains.     

Structure determination of Discoidin II from Dictyostelium discoideum and carbohydrate binding properties of the lectin domain

The social amoeba Dictyostelium discoideum adopts a cohesive stage upon starvation and then produces Discoidin I and II, two proteins able to bind galactose and N-acetyl-galactosamine. The N-terminal domain or discoidin domain (DS) is widely distributed in eukaryotes where it plays a role in extracellular matrix binding while the C-terminal domain displays sequence similarities to invertebrate lectins. We present the first X-ray structures of the wild-type and recombinant Discoidin II in unliganded state and in complex with monosaccharides. The protein forms a homotrimer which presents two binding surfaces situated on the opposite boundaries of the structure. The binding sites of the N-terminal domain contain PEG molecules that could mimics binding of natural ligand. The C-terminal lectin domain interactions with N-acetyl-D-galactosamine and methyl-beta-galactoside are described. The carbohydrate binding sites are located at the interface between monomers. Specificity for galacto configuration can be rationalized since the axial O4 hydroxyl group is involved in several hydrogen bonds with protein side chains. Titration microcalorimetry allowed characterization of affinity and demonstrated the enthalpy-driven character of the interaction. Those results highlight the structural differentiation of the DS domain involved in many cell-adhesion processes from the lectin activity of Dictyostelium discoidins.     

High-resolution crystal structures of the lectin-like xylan binding domain from Streptomyces lividans xylanase 10A with bound substrates reveal a novel mode of xylan binding

Carbohydrate-binding module (CBM) family 13 includes the "R-type" or "ricin superfamily" beta-trefoil lectins. The C-terminal CBM, CBM13, of xylanase 10A from Streptomyces lividans is a family 13 CBM that is not only structurally similar to the "R-type" lectins but also somewhat functionally similar. The primary function of CBM13 is to bind the polysaccharide xylan, but it retains the ability of the R-type lectins to bind small sugars such as lactose and galactose. The association of CBM13 with xylan appears to involve cooperative and additive participation of three binding pockets in each of the three trefoil domains of CBM13, suggesting a novel mechanism of CBM-xylan interaction. Thus, the interaction of CBM13 with sugars displays considerable plasticity for which we provide a structural rationale. The high-resolution crystal structure of CBM13 was determined by multiple anomalous dispersion from a complex of CBM13 with a brominated ligand. Crystal structures of CBM13 in complex with lactose and xylopentaose revealed two distinct mechanisms of ligand binding. CBM13 has retained its specificity for lactose via Ricin-like binding in all of the three classic trefoil binding pockets. However, CBM13 has the ability to bind either the nonreducing galactosyl moiety or the reducing glucosyl moiety of lactose. The mode of xylopentaose binding suggests adaptive mutations in the trefoil sugar binding scaffold to accommodate internal binding on helical polymers of xylose.     

The crystal structure of the family 6 carbohydrate binding module from Cellvibrio mixtus endoglucanase 5a in complex with oligosaccharides reveals two distinct binding sites with different ligand specificities

Glycoside hydrolases that release fixed carbon from the plant cell wall are of considerable biological and industrial importance. These hydrolases contain non-catalytic carbohydrate binding modules (CBMs) that, by bringing the appended catalytic domain into intimate association with its insoluble substrate, greatly potentiate catalysis. Family 6 CBMs (CBM6) are highly unusual because they contain two distinct clefts (cleft A and cleft B) that potentially can function as binding sites. Henshaw et al. (Henshaw, J., Bolam, D. N., Pires, V. M. R., Czjzek, M., Henrissat, B., Ferreira, L. M. A., Fontes, C. M. G. A., and Gilbert, H. J. (2003) J. Biol. Chem. 279, 21552-21559) show that CmCBM6 contains two binding sites that display both similarities and differences in their ligand specificity. Here we report the crystal structure of CmCBM6 in complex with a variety of ligands that reveals the structural basis for the ligand specificity displayed by this protein. In cleft A the two faces of the terminal sugars of beta-linked oligosaccharides stack against Trp-92 and Tyr-33, whereas the rest of the binding cleft is blocked by Glu-20 and Thr-23, residues that are not present in CBM6 proteins that bind to the internal regions of polysaccharides in cleft A. Cleft B is solvent-exposed and, therefore, able to bind ligands because the loop, which occludes this region in other CBM6 proteins, is much shorter and flexible (lacks a conserved proline) in CmCBM6. Subsites 2 and 3 of cleft B accommodate cellobiose (Glc-beta-1,4-Glc), subsite 4 will bind only to a beta-1,3-linked glucose, whereas subsite 1 can interact with either a beta-1,3- or beta-1,4-linked glucose. These different specificities of the subsites explain how cleft B can accommodate beta-1,4-beta-1,3- or beta-1,3-beta-1,4-linked gluco-configured ligands.     

Studies on the binding substances on human erythrocytes for the heat-labile enterotoxin isolated from chicken enterotoxigenic Escherichia coli

To study the predominant binding substance for the heat-labile enterotoxin (LTc) isolated from chicken enterotoxigenic Escherichia coli, competitive binding assays were performed with neuraminidase-treated human type B erythrocytes and 125I-labeled B subunit of LTc (LTc-B). Of all inhibitors used, the ganglioside GM1 was the most effective in inhibiting the binding of 125I-labeled LTc-B to the erythrocytes. The other gangliosides used as inhibitors, gangliosides GD1b, GD1a, GM2, GT1b and GM3, were about 24, 166, 250, 440 and at least 440 times less reactive than ganglioside GM1, respectively. With glycoproteins as inhibitors, on the other hand, hog A + H, porcine thyroglobulin and bovine salivary mucin were over 10(4) times less potent. No inhibition was obtained by other mono-, di- and polysaccharides at the highest concentrations used. These findings suggest that the predominant binding substance on neuraminidase-treated human type B erythrocytes for the LTc-B is ganglioside GM1 and that the combining site of LTc-B may be specific for the terminal disaccharide (galactose-N-acetyl-D-galactosamine)-linked portion of ganglioside GM1.     

Specificity of anion binding in the substrate pocket of bacteriorhodopsin

The structure of the D85S mutant of bacteriorhodopsin with a nitrate anion bound in the Schiff base binding site and the structure of the anion-free protein have been obtained in the same crystal form. Together with the previously solved structures of this anion pump, in both the anion-free state and bromide-bound state, these new structures provide insight into how this mutant of bacteriorhodopsin is able to bind a variety of different anions in the same binding pocket. The structural analysis reveals that the main structural change that accommodates different anions is the repositioning of the polar side chain of S85. On the basis of these X-ray crystal structures, the prediction is then made that the D85S/D212N double mutant might bind similar anions and do so over a broader pH range than does the single mutant. Experimental comparison of the dissociation constants, K(d), for a variety of anions confirms this prediction and demonstrates, in addition, that the binding affinity is dramatically improved by the D212N substitution.     

Spectrum of sialylated and nonsialylated fuco-oligosaccharides bound by the endothelial-leukocyte adhesion molecule E-selectin. Dependence of the carbohydrate binding activity on E-selectin density.

Carbohydrate recognition by the human endothelial-leukocyte adhesion molecule, E-selectin, has been investigated by binding studies using 3H-labeled Chinese hamster ovary cells expressing different levels of the transfected full-length adhesion molecule and a series of structurally defined oligosaccharides linked to the lipid phosphatidylethanolamine dipalmitoate (neoglycolipids) and synthetic glycolipids chromatographed on silica gel plates or immobilized on plastic wells. Evidence is presented for density-dependent binding of the membrane-associated E-selectin not only to 3'-sialyl-lacto-N-fucopentaose II (3'-S-LNFP-II) and 3'-sialyl-lacto-N-fucopentaose III (3'-S-LNFP-III) which express the sialyl Le(a) and sialyl Le(x) antigens, respectively, but also to the nonsialylated analogue LNFP-II; there is a threshold density of E-selectin required for binding to these sialylated sequences, and binding to the nonsialylated analogue is a property only of cells with the highest density of E-selectin expression. The presence of fucose linked to subterminal rather than to an internal N-acetylglucosamine is shown to be a requirement for E-selectin binding, and although the presence of sialic acid 3-linked to the terminal galactose of the LNFP-II or LNFP-III sequences substantially enhances E-selectin binding, the presence of 6-linked sialic acid abolishes binding. E-selectin binding is unaffected in the presence of the blood group H fucose (alpha 1-2 linked to galactose to form the Le(b) antigen). However, the binding is abolished when in addition alpha 1-3-linked N-acetylgalactosamine to the galactose (blood group A antigen) is present. These results indicate that some E-selectin-mediated adhesive events may be influenced by blood group status.     

Studies on gangliosides with affinity for Helicobacter pylori: binding to natural and chemically modified structures

Helicobacter pylori, like many other microbes, has the ability to bind to carbohydrate epitopes. Several sugar sequences have been reported as active for the bacterium, including some neutral, sulfated, and sialylated structures. We investigated structural requirements for the sialic acid-dependent binding using a number of natural and chemically modified gangliosides. We have chosen for derivatization studies two kinds of binding-active glycolipids, the simple ganglioside S-3PG (Neu5Ac alpha 3Gal beta 4GlcNAc beta 3Gal beta 4Glc beta 1Cer, sialylparagloboside) and branched polyglycosylceramides (PGCs) of human origin. The modifications included oxidation of the sialic acid glycerol chain, reduction of the carboxyl group, amidation of the carboxyl group, and lactonization. Binding experiments confirmed a preference of H. pylori for 3-linked sialic acid and penultimate 4-linked galactose. As expected, neolacto gangliosides (with Gal beta 4GlcNAc in the core structure) were active in our assays, whereas gangliosides with lacto (Gal beta 3GlcNAc) and ganglio (Gal beta 3GalNAc) carbohydrate chains were not. Negative binding results were also obtained for disialylparagloboside (with terminal NeuAc alpha 8NeuAc) and NeuAc alpha 6-containing glycolipids. Chemical studies revealed dependence of the binding on Neu5Ac and its glycerol and carboxyl side chains. Most of the derivatizations performed on these groups abolished the binding; however, some of the amide forms turned out to be active, and one of them (octadecylamide) was found to be an excellent binder. The combined data from molecular dynamics simulations indicate that the binding-active configuration of the terminal disaccharide of S-3PG is with the sialic acid in the anticlinal conformation, whereas in branched PGCs the same structural element most likely assumes the synclinal presentation.     

Phase variation of a beta-1,3 galactosyltransferase involved in generation of the ganglioside GM1-like lipo-oligosaccharide of Campylobacter jejuni

Ganglioside mimicry by Campylobacter jejuni lipo-oligosaccharide (LOS) is thought to be a critical factor in the triggering of the Guillain-Barré and Miller-Fisher syndrome neuropathies after C. jejuni infection. The combination of a completed genome sequence and a ganglioside GM1-like LOS structure makes C. jejuni NCTC 11168 a useful model strain for the identification and characterization of the genes involved in the biosynthesis of ganglioside-mimicking LOS. Genome analysis identified a putative LOS biosynthetic cluster and, from this, we describe a putative gene (ORF Cj1139c), which we have termed wlaN, with a significant level of similarity to a number of bacterial glycosyltransferases. Mutation of this gene in C. jejuni NCTC 11168 resulted in a LOS molecule of increased electrophoretic mobility, which also failed to bind cholera toxin. Comparison of LOS structural data from wild type and the mutant strain indicated lack of a terminal beta-1,3-linked galactose residue in the latter. The wlaN gene product was demonstrated unambiguously as a beta-1,3 galactosyltransferase responsible for converting GM2-like LOS structures to GM1-like by in vitro expression. We also show that the presence of an intragenic homopolymeric tract renders the expression of a functional wlaN gene product phase variable, resulting in distinct C. jejuni NCTC 11168 cell populations with alternate GM1 or GM2 ganglioside-mimicking LOS structures. The distribution of wlaN among a number of C. jejuni strains with known LOS structure was determined and, for C. jejuni NCTC 12500, similar wlaN gene phase variation was shown to occur, so that this strain has the potential to synthesize a GM1-like LOS structure as well as the ganglioside GM2-like LOS structure proposed in the literature.     

The fine specificity of mannose-binding and galactose-binding lectins revealed using outlier motif analysis of glycan array data

Glycan-binding proteins are commonly used as analytical reagents to detect the levels of specific glycan structures in biological samples. A detailed knowledge of the specificities of glycan-binding proteins is required for properly interpreting their binding data. A powerful technology for characterizing glycan-binding specificity is the glycan array. However, the interpretation of glycan-array data can be difficult due to the complex fine specificities of certain glycan-binding proteins. We developed a systematic approach, called outlier-motif analysis, for extracting fine-specificity information from glycan-array data, and we applied the method to the study of four commonly used lectins: two mannose binders (concanavalin A and Lens culinaris) and two galactose binders (Bauhinia purpurea and peanut agglutinin). The study confirmed the known, primary specificity of each lectin and also revealed new insights into their binding preferences. Lens culinaris's main specificity may be non-terminal, α-linked mannose with a single linkage at its 2' carbon, which is more restricted than previous definitions. We found broader specificity for bauhinea purpurea (BPL) than previously reported, showing that BPL can bind terminal N-acetylgalactosamine (GalNAc) and penultimate β-linked galactose under certain limitations. Peanut agglutinin may bind terminal Galβ1,3Gal, a glycolipid motif, in addition to terminal Galβ1,3GalNAc, a common O-linked glycoprotein motif. These results could be used to more accurately interpret data obtained using these well-studied lectins. Furthermore, this study demonstrates a systematic and general approach for extracting fine-specificity information from glycan-array data.