Neutralization receptor-based immunoassay for detection of neutralizing antibodies to Escherichia coli verocytotoxin 1

Gene Mutations Versus Clinically Relevant Phenotypes

Background—

Currently, no method is available to identify α-galactosidase A (agalA) mutations determining clinically relevant Fabry disease. In our largest European Fabry cohort, we investigated whether a biomarker, specific for the defect, could stratify persons at risk.

Methods and Results—

A total of 124 individuals with agalA mutations were investigated with a comprehensive clinical workup, genetic analysis, and laboratory testing, including measurements of agalA activity and lyso-Gb3 (degradation product of the accumulating Gb3). Additionally, an extensive family screening with a clinical workup of relatives was performed. The patient population was divided into 2 samples: previously described mutations (n=72) and novel mutations (n=52). The patients with previously described mutations were subdivided into 2 groups: classical mutations, which were known to cause the classic type of Fabry disease with specific symptoms and a high risk for major events in all 3 main organs (heart, kidney, and central nervous system), and atypical mutations without the typical presentation. All patients with atypical mutations (n=17) had lower lyso-Gb3 levels than any of the patients with classical Fabry disease (n=55). A cutoff value of 2.7 ng/mL separated the 2 groups. Six out of 52 patients with novel mutations showed a lyso-Gb3 level <2.7 ng/mL. Clinical investigation, blinded to lyso-Gb3 results, revealed no classic organ involvement in these patients or their relatives. In contrast, the characterization of patients with lyso-Gb3≥2.7ng/mL suggested classical Fabry mutations in most of the patients (93%).

Conclusions—

Our data show that the biomarker lyso-Gb3 may identify the clinically relevant agalA mutations leading to Fabry disease.

Shiga toxin binding to glycolipids and glycans

BACKGROUND

Immunologically distinct forms of Shiga toxin (Stx1 and Stx2) display different potencies and disease outcomes, likely due to differences in host cell binding. The glycolipid globotriaosylceramide (Gb3) has been reported to be the receptor for both toxins. While there is considerable data to suggest that Gb3 can bind Stx1, binding of Stx2 to Gb3 is variable.

METHODOLOGY

We used isothermal titration calorimetry (ITC) and enzyme-linked immunosorbent assay (ELISA) to examine binding of Stx1 and Stx2 to various glycans, glycosphingolipids, and glycosphingolipid mixtures in the presence or absence of membrane components, phosphatidylcholine, and cholesterol. We have also assessed the ability of glycolipids mixtures to neutralize Stx-mediated inhibition of protein synthesis in Vero kidney cells.

RESULTS

By ITC, Stx1 bound both Pk (the trisaccharide on Gb3) and P (the tetrasaccharide on globotetraosylceramide, Gb4), while Stx2 did not bind to either glycan. Binding to neutral glycolipids individually and in combination was assessed by ELISA. Stx1 bound to glycolipids Gb3 and Gb4, and Gb3 mixed with other neural glycolipids, while Stx2 only bound to Gb3 mixtures. In the presence of phosphatidylcholine and cholesterol, both Stx1 and Stx2 bound well to Gb3 or Gb4 alone or mixed with other neutral glycolipids. Pre-incubation with Gb3 in the presence of phosphatidylcholine and cholesterol neutralized Stx1, but not Stx2 toxicity to Vero cells.

CONCLUSIONS

Stx1 binds primarily to the glycan, but Stx2 binding is influenced by residues in the ceramide portion of Gb3 and the lipid environment. Nanomolar affinities were obtained for both toxins to immobilized glycolipids mixtures, while the effective dose for 50% inhibition (ED(50)) of protein synthesis was about 10(-11) M. The failure of preincubation with Gb3 to protect cells from Stx2 suggests that in addition to glycolipid expression, other cellular components contribute to toxin potency.

Escherichia coli Shiga Toxin Mechanisms of Action in Renal Disease

Shiga toxin-producing Escherichia coli is a contaminant of food and water that in humans causes a diarrheal prodrome followed by more severe disease of the kidneys and an array of symptoms of the central nervous system. The systemic disease is a complex referred to as diarrhea-associated hemolytic uremic syndrome (D⁺HUS). D⁺HUS is characterized by thrombocytopenia, microangiopathic hemolytic anemia, and acute renal failure. This review focuses on the renal aspects of D⁺HUS. Current knowledge of this renal disease is derived from a combination of human samples, animal models of D⁺HUS, and interaction of Shiga toxin with isolated renal cell types. Shiga toxin is a multi-subunit protein complex that binds to a glycosphingolipid receptor, Gb3, on select eukaryotic cell types. Location of Gb3 in the kidney is predictive of the sites of action of Shiga toxin. However, the toxin is cytotoxic to some, but not all cell types that express Gb3. It also can cause apoptosis or generate an inflammatory response in some cells. Together, this myriad of results is responsible for D⁺HUS disease.

Detergent-resistant globotriaosyl ceramide may define verotoxin/glomeruli-restricted hemolytic uremic syndrome pathology

Verotoxin binding to its receptor, globotriaosyl ceramide(Gb(3)) mediates the glomerular pathology of hemolytic uremic syndrome, but Gb(3) is expressed in both tubular and glomerular cells. Gb(3) within detergent-resistant membranes, an index of glycolipid-cholesterol enriched lipid rafts, is required for in vitro cytotoxicity. We found that verotoxin 1 and 2 binding to human adult renal glomeruli is detergent resistant, whereas the strong verotoxin binding to renal tubules is detergent sensitive. Verotoxin binding to pediatric glomeruli was detergent resistant but binding to adult glomeruli was enhanced, remarkably for some samples, by detergent extraction. Detergent-sensitive glomerular components may provide age-related protection against verotoxin glomerular binding. Mouse glomeruli remained verotoxin unreactive after detergent extraction, whereas tubular binding was lost. Cholesterol extraction induced strong verotoxin binding in poorly reactive adult glomeruli, suggesting cholesterol can mask Gb(3) in glomerular lipid rafts. Binding of the human immunodeficiency virus (HIV) adhesin, gp120 (another Gb(3) ligand) was detergent sensitive, tubule-restricted, and inhibited by verotoxin B subunit pretreatment, and may relate to HIV nephropathy. Our study shows that differential membrane Gb(3) organization in glomeruli and tubules provides a basis for the age- and glomerular-restricted pathology of hemolytic uremic syndrome.

Differential intracellular transport and binding of verotoxin 1 and verotoxin 2 to globotriaosylceramide-containing lipid assemblies

Although verotoxin-1 (VT1) and verotoxin-2 (VT2) share a common receptor, globotriaosyl ceramide (Gb(3)), VT2 induces distinct animal pathology and is preferentially associated with human disease. Moreover VT2 cytotoxicity in vitro is less than VT1. We therefore investigated whether these toxins similarly traffic within cells via similar Gb(3) assemblies. At 4 degrees C, fluorescent-VT1 and VT2 bound both coincident and distinct punctate surface Gb(3) microdomains. After 10 min at 37 degrees C, similar distinct/coincident micropunctate intracellular localization was observed. Most internalized VT2, but not VT1, colocalized with transferrin. After 1 h, VT1 and VT2 coalesced during retrograde transport to the Golgi. During prolonged incubation (3-6 h), VT1, and VT2 (more slowly), exited the Golgi to reach the ER/nuclear envelope. At this time, VT2 induced a previously unreported, retrograde transport-dependent vacuolation. Cell surface and intracellular VT1 showed greater detergent resistance than VT2, suggesting differential 'raft' association. >90% (125)I-VT1 cell surface bound, or added to detergent-resistant cell membrane extracts (DRM), was in the Gb(3)-containing sucrose gradient 'insoluble' fraction, whereas only 30% (125)I-VT2 was similarly DRM-associated. VT1 bound more efficiently to Gb(3)/cholesterol DRMs generated in vitro. Only VT1 binding was inhibited by high cholesterol/Gb(3) ratios. VT2 competed less effectively for (125)I-VT1/Gb(3) DRM-binding but only VT2-Gb(3)/cholesterol DRM-binding was augmented by sphingomyelin. Differential VT1/VT2 Gb(3) raft-binding may mediate differential cell binding/intracellular trafficking and cytopathology.

Role of multiple drug resistance protein 1 in neutral but not acidic glycosphingolipid biosynthesis

Transfection studies have implicated the multiple drug resistance pump, MDR1, as a glucosyl ceramide translocase within the Golgi complex (Lala, P., Ito, S., and Lingwood, C. A. (2000) J. Biol. Chem. 275, 6246-6251). We now show that MDR1 inhibitors, cyclosporin A or ketoconazole, inhibit neutral glycosphingolipid biosynthesis in 11 of 12 cell lines tested. The exception, HeLa cells, do not express MDR1. Microsomal lactosyl ceramide and globotriaosyl ceramide synthesis from endogenous or exogenously added liposomal glucosyl ceramide was inhibited by cyclosporin A, consistent with a direct role for MDR1/glucosyl ceramide translocase activity in their synthesis. In contrast, cellular ganglioside synthesis in the same cells, was unaffected by MDR1 inhibition, suggesting neutral and acid glycosphingolipids are synthesized from distinct precursor glycosphingolipid pools. Metabolic labeling in wild type and knock-out (MDR1a, 1b, MRP1) mouse fibroblasts showed the same loss of neutral glycosphingolipid (glucosyl ceramide, lactosyl ceramide) but not ganglioside (GM3) synthesis, confirming the proposed role for MDR1 translocase activity. Cryo-immunoelectron microscopy showed MDR1 was predominantly intracellular, largely in rab6-containing Golgi vesicles and Golgi cisternae, the site of glycosphingolipid synthesis. These studies identify MDR1 as the major glucosyl ceramide flippase required for neutral glycosphingolipid anabolism and demonstrate a previously unappreciated dichotomy between neutral and acid glycosphingolipid synthesis.

Differential tissue targeting and pathogenesis of verotoxins 1 and 2 in the mouse animal model

BACKGROUND

Both verotoxin (VT)1 and VT2 share the same receptor, globotriaosyl ceramide (Gb(3)). Although VT1 is slightly more cytotoxic in vitro and binds Gb(3) with higher affinity, VT2 is more toxic in mice and may be associated with greater pathology in human infections. In this study we have compared the biodistribution of iodine 125 ((125)I)-VT1 and (125)I-VT2 versus pathology in the mouse.

METHODS

(125)I-VT1 whole-body autoradiography defined the tissues targeted. VT1 and VT2 tissue distribution, clearance, and tissue binding sites were compared. The effect of a soluble receptor analogue, adamantylGb(3), on VT2/Gb3 binding and in vivo pathology was assessed.

RESULTS

(125)I-VT1 autoradiography identified the lungs and nasal turbinates as major, previously unrecognized, targets, while kidney cortex and the bone marrow of the spine, long bones, and ribs were also significant targets. VT2 did not target the lung, but accumulated in the kidney to a greater extent than VT1. The serum half-life of VT1 was 2.7 minutes with 90% clearance at 5 minutes, while that of VT2 was 3.9 minutes with only 40% clearance at 5 minutes. The extensive binding of VT1, but not VT2, within the lung correlated with induced lung disease. Extensive hemorrhage into alveoli, edema, alveolitis and neutrophil margination was seen only after VT1 treatment. VT1 targeted lung capillary endothelial cells. Identical tissue binding sites (subsets of proximal/distal tubules and collecting ducts) for VT1 and VT2 were detected by toxin overlay of serial frozen kidney sections. Glucosuria was found to be a new marker of VT1- and VT2-induced renal pathology and positive predictor of outcome in the mouse, consistent with VT-staining of proximal tubules. Lung Gb3 migrated on thin-layer chromatography (TLC) faster than kidney Gb(3), suggesting a different lipid composition. AdamantylGb(3), a soluble Gb(3) analogue, competed effectively for Gb3 binding by VT1 and VT2 in vitro. However, the effect in the mouse model (only measured against VT2, due to the lower LD(50), a concentration required for 50% lethality) was to increase, rather than reduce, pathology and further reduce the VT2 serum clearance rate. Additional renal pathology was seen in VT2 + adamantylGb(3)-treated mice.

CONCLUSIONS

The lung is a preferential (Gb(3)) "sink" for VT1, which explains the relatively slower clearance of VT2 and subsequent increased VT2 renal targeting and VT2 mortality in this animal model.

Effect of globotriaosyl ceramide fatty acid alpha-hydroxylation on the binding by verotoxin 1 and verotoxin 2

Variation in the lipid moiety of the verotoxin (VT) receptor glycosphingolipid, globotriaosyl ceramide (Gb3) can modulate toxin binding. The binding of VT1 and VT2 to C18 and C22 alpha hydroxy and nonhydroxy fatty acid isoforms of Gb3 were compared using a receptor ELISA and a 125I-labeled toxin/glycolipid microtitre plate direct binding assay. Increased binding to the hydroxylated species, particularly C220H, was observed for both toxins. Increased RELISA binding at low glycolipid concentrations only, suggested the binding affinity is increased following Gb3 fatty acid hydroxylation. Nonlinear regression analysis of direct binding assay to these Gb3 isoforms confirmed the increased affinity of both toxins for the C22 hydroxylated Gb3. The capacity was also significantly increased. The increased binding of VTs for hydroxylated fatty acid Gb3 isoforms may be a factor in the selective renal pathology which can follow systemic verotoxemia, particularly in the mouse model. The more pronounced effect at lower glycolipid concentrations prompted investigation of VT1 binding affinity at different Gb3 concentrations. Unexpectedly, the VT1 Kd for Gb3 was found to decrease as an inverse function of the Gb3 concentration. This shows that glycolipids have "nonclassical" receptor properties.

A verotoxin 1 B subunit-lambda CRO chimeric protein specifically binds both DNA and globotriaosylceramide (Gb(3)) to effect nuclear targeting of exogenous DNA in Gb(3) positive cells

Inefficient nuclear incorporation of foreign DNA remains a critical roadblock in the development of effective nonviral gene delivery systems. DNA delivered by traditional protocols remains within endosomal/lysosomal vesicles, or is rapidly degraded in the cytoplasm. Verotoxin I (VT), an AB(5) subunit toxin produced by enterohaemorrhagic Escherichia coli, binds to the cell surface glycolipid, globotriaosylceramide (Gb(3)) and is internalized into preendosomes. VT is then retrograde transported to the Golgi, endoplasmic reticulum (ER), and nucleus of highly VT-sensitive cells. We have utilized this nuclear targeting of VT to design a unique delivery system which transports exogenous DNA via vesicular traffic to the nucleus. The nontoxic VT binding subunit (VTB) was fused to the lambda Cro DNA-binding repressor, generating a 14-kDa VTB-Cro chimera. VTB-Cro binds specifically via the Cro domain to a 25-bp DNA fragment containing the consensus Cro operator. VTB-Cro demonstrates simultaneous specific binding to Gb(3). Treatment of Vero cells with fluorescent-labeled Cro operator DNA in the presence of VTB-Cro, results in DNA internalization to the Golgi, ER, and nucleus, whereas fluorescent DNA alone is incorporated poorly and randomly within the cytoplasm. VTB-Cro mediated nuclear DNA transport is prevented by brefeldin A, consistent with Golgi/ER intracellular routing. Pretreatment with filipin had no effect, indicating that caveoli are not involved. This novel VTB-Cro shuttle protein may find practical applications in the fields of intracellular targeting, gene delivery, and gene therapy.

Oxidation of aglycone of glycosphingolipids: serine and ceramide acid precursors for soluble glycoconjugates

A new oxidation protocol for the cleavage of sphingosine double bonds is described. The procedure is applicable to both natural and deacyl glycolipids and can be applied to microgram quantities of precursors. Under neutral conditions, glycosyl ceramide acids are obtained and under basic conditions glycosyl serine acids are obtained. The glycosyl ceramide acid-based glycoconjugates--BSA-neoglycoprotein and adamantyl-neohydrocarbon--demonstrate the importance that an aglycone can play in carbohydrate-protein interaction. Studies with HIV coat protein gp120 and BSA-neoglycoprotein conjugates derived from galactosylceramide (GalC) showed that binding affinities of the conjugates depend on the manner in which the glycosyl unit is coupled to the protein. Deacyl-GalC conjugates, in which the glycosyl unit is coupled via the amine of the sphingosine, showed significantly lower affinity as compared to glycosylceramide acid conjugates. In the case of Gb3-VT1 binding, it was found that ceramide acid conjugates bound to VT1 better than the serine acid conjugates. These studies show that the aglycone organization, particularly the region adjacent to the carbohydrate region (or in a membrane environment, the aglycone-glycone interface) modulate carbohydrate presentation. It is possible that in each of the conjugates described above, the interface region could have different hydrogen-bonding networks (see Scheme 4.) This, in turn, could influence the solvation and/or conformation of this region and thereby influence ligand binding.

Effect of a soluble pseudo-receptor on verotoxin 2-induced toxicity

To neutralize the toxicity of verotoxin (VT) produced by Escherichia coli type O-157, a soluble pseudo-receptor (Lyso Gb3) was synthesized with the deacylated form of the natural receptor, globotrisylceramide (Galalpha 1-4Galbeta1-4-glucosylceramide; Gb3). In this study, we evaluated the characteristics and pharmacological effects of Lyso Gb3, using VT2. It was confirmed that Lyso Gb3 specifically recognized VT2. Lyso Gb3 itself had little influence on the in-vitro growth of Vero cells, but markedly augmented VT2-induced cytotoxicity. In addition, the VT2-induced killing of mice was not decreased, but was, rather, increased by Lyso Gb3. These results indicate that the soluble pseudo-receptor Lyso Gb3 recognized VT2. However, it did not reduce, but, rather, enhanced, VT2-induced toxicity in the presence of the natural receptor, although Lyso Gb3 alone had no toxicity.

Oxidation of glycosphingolipids under basic conditions: synthesis of glycosyl "serine acids" as opposed to "ceramide acids". Precursors for neoglycoconjugates with increased ligand binding affinity

Two types of oxidative cleavage of the double bond of glycosphingolipids (GSLs) are described. Oxidation of peracetylated GSL precursors with stoichiometric proportions of KMnO4 and an excess of NaIO4, in a neutral aqueous tert-butanol solvent system, gave nearly quantitative yields of the glycosyl ceramide acid, 2-hydroxy-3-(N-acyl)-4-(O-glycosyl)oxybutyric acid [Mylvaganam, M., and Lingwood, C. A. (1999) J. Biol. Chem. 274, 20725-20732]. However, if the reaction medium was made alkaline, the hydroxyallylic function of the sphingolipid, as a whole, was oxidized and the glycosyl serine acid, 2-(N-acyl)-3-(O-glycosyl)oxypropionic acid, was obtained in good yield. This represents a new type of oxidation reaction. Optimized conditions gave glycosyl ceramide or serine acids with greater than 90% selectivity and in good yields (90%). Oxidation of dGSLs gave serine and ceramide oligosaccharides, devoid of hydrocarbon chains. An intriguing glycosyl species containing 5-hydroxy-4-oxo-3-hydroxy-2-(N-acyl)sphingosine (hydroxy-acyl intermediate) was identified via ESMS analyses. We propose that further oxidation of this intermediate is pH-dependent and will be oxidized to either serine or ceramide acids. On the basis of MS-MS analysis of specific homologues of serine and ceramide acids, two types of collision-induced dissociation (CID) patterns have been established. These CID patterns were then used in the identification of serine and ceramide acids synthesized from natural GSL samples. Also, on a qualitative basis, this oxidation protocol, in conjunction with ESMS, provides a novel method for characterizing the aglycone composition (acyl chain length, unsaturation position, dihydrosphingosine content, etc.) of natural GSLs. A novel class of neohydrocarbon conjugates were synthesized by coupling the acids to rigid hydrocarbon frames such as 2-aminoadamantane. Preliminary studies with conjugates derived from globotriaosyl ceramide (Gb3C), lactosyl ceramide (LC), and galactosyl ceramide (GalC) bound verotoxin with the expected specificity but with affinities much greater than that of the natural glycolipid. Also, the ceramide acid-based conjugates were better ligands than serine acid conjugates.

The identification of three biologically relevant globotriaosyl ceramide receptor binding sites on the Verotoxin 1 B subunit

The Verotoxin 1 (VT1) B subunit binds to the glycosphingolipid receptor globotriaosylceramide (Gb3). Receptor-binding specificity is associated with the terminally linked Galalpha(1-4) Galbeta disaccharide sequence of the receptor. Recently, three globotriose (Galalpha[1-4] Galbeta [1-4] Glcbeta) binding sites per B-subunit monomer were identified by crystallography. Two of these sites (sites I and II) are located adjacent to phenylalanine-30. Site I was originally predicted as a potential Gb3 binding site on the basis of sequence conservation, and site II was additionally predicted based on computer modelling and receptor docking. The third (site III) was also identified by crystallography and is located at the N-terminal end of the alpha-helix. To determine the biological significance of sites II and III, and to support our previous findings of the significance of site I, we examined the binding properties and cytotoxicity of VT1 mutants designed to block Gb3 binding at each site selectively. The Scatchard analysis of saturation-binding data for each mutant revealed that only the amino acid substitutions predicted to affect site I (D-17E) or site II (G-62T) caused reductions in the binding affinity and capacity of VT1 for Gb3. Similarly, those mutations at sites I and II also caused significant reductions in both Vero and MRC-5 cell cytotoxicity (by seven and five logs, respectively, for G-62T and by four and two logs, respectively, for D-17E). In contrast, the substitution of alanine for W-34 at site III did not reduce the high-affinity binding of the B subunit, despite causing a fourfold reduction in the receptor-binding capacity. The corresponding mutant W-34A holotoxin had a two-log reduction in cytotoxicity on Vero cells and no statistically significant reduction on MRC-5 cells. We conclude that the high-affinity receptor binding most relevant for cell cytotoxicity occurs at sites I and II. In contrast, site III appears to mediate the recognition of additional Gb3 receptor epitopes but with lower affinity. Our results support the significance of the indole ring of W-34 for binding at this site.

Adamantyl globotriaosyl ceramide: a monovalent soluble mimic which inhibits verotoxin binding to its glycolipid receptor

The globotriaosylceramide (Gb3) verotoxin (VT) interaction is one of several examples of glycolipid receptors where the ceramide (or lipid) free oligosaccharides fail to show the expected binding parameters. We present a novel, yet simple strategy to synthesize monovalent, water soluble glycosphingolipid mimics which retain receptor function. Replacing the fatty acid chain with rigid, three dimensional hydrocarbon frames, such as adamantane, gives a novel class of neohydrocarbon glycoconjugates. Such adamantyl conjugates derived from Gb3 showed significantly enhanced solubility in water compared to natural Gb3. Adamantyl-Gb3 showed a thousand fold enhanced inhibitory activity (IC50 = 1 microM) for VT-Gb3 binding as compared to a lipid free Gb3 oligosaccharide derivative, alphaGal1-4betaGal1-4betaGlc1-O-CH2CH(CH2SO2C 4H9)2 (IC50 > 2 mM). This represents a new approach to the generation of antagonists of glycolipid receptors.

pH-independent retrograde targeting of glycolipids to the Golgi complex

A small fraction of the molecules internalized by endocytosis reaches the Golgi complex through a retrograde pathway that is poorly understood. In the present work, we used bacterial toxins to study the retrograde pathway in Vero cells. The recombinant B subunit of verotoxin 1B (VT1B) was labeled with fluorescein to monitor its progress within the cell by confocal microscopy. This toxin, which binds specifically to the glycolipid globotriaosyl ceramide, entered endosomes by both clathrin-dependent and -independent pathways, reaching the Golgi complex. Once internalized, the toxin-receptor complex did not recycle back to the plasma membrane. The kinetics of internalization and the subcellular distribution of VT1B were virtually identical to those of another glycolipid-binding toxin, the B subunit of cholera toxin (CTB). Retrograde transport of VT1B and CTB was unaffected by addition of weak bases in combination with concanamycin, a vacuolar-type ATPase inhibitor. Ratio imaging confirmed that these agents neutralized the luminal pH of the compartments where the toxin was located. Therefore, the retrograde transport of glycolipids differs from that of proteins like furin and TGN38, which require an acidic luminal pH. Additional experiments indicated that the glycolipid receptors of VT1B and CTB are internalized independently and not as part of lipid "rafts" and that internalization is cytochalasin insensitive. We conclude that glycolipids utilize a unique, pH-independent retrograde pathway to reach compartments of the secretory system and that assembly of F-actin is not required for this process.

Translocation of verotoxin-1 across T84 monolayers: mechanism of bacterial toxin penetration of epithelium

Verotoxin-producing Escherichia coli (VTEC) are pathogenic bacteria associated with diarrhea, hemorrhagic colitis, and hemolytic uremic syndrome. Verotoxins (VTs) elaborated by these organisms produce cytopathic effects on a restricted number of cell types, including endothelial cells lining the microvasculature of the bowel and the kidney. Because human intestinal epithelial cells lack the globotriaosylceramide receptor for VT binding, it is unclear how the toxin moves across the intestinal mucosa to the systemic circulation. The aims of this study were to determine the effects of VT-1 on intestinal epithelial cell function and to characterize VT-1 translocation across monolayers of T84 cells, an intestinal epithelial cell line. VT-1 at concentrations up to 1 microgram/ml had no effect on the barrier function of T84 monolayers as assessed by measuring transmonolayer electrical resistance (102 +/- 8% of control monolayers). In contrast, both VT-positive and VT-negative VTEC bacterial strains lowered T84 transmonolayer resistance (45 +/- 7 and 38 +/- 6% of controls, respectively). Comparable amounts of toxin moved across monolayers of T84 cells, exhibiting high-resistance values, as monolayers with VTEC-induced decreases in barrier function, suggesting a transcellular mode of transport. Translocation of VT-1 across T84 monolayers paralleled the movement of a comparably sized protein, horseradish peroxidase. Immunoelectron microscopy confirmed transcellular transport of VT-1, since the toxin was observed within endosomes and associated with specific intracellular targets, including the Golgi network and endoplasmic reticulum. These data present a mode of VT-1 uptake by toxin-insensitive cells and suggest a general mechanism by which bacterial toxins lacking specific intestinal receptors can penetrate the intestinal epithelial barrier.

Localization of intravenously administered verocytotoxins (Shiga-like toxins) 1 and 2 in rabbits immunized with homologous and heterologous toxoids and toxin subunits

Rabbits challenged intravenously with Shiga toxin or with Escherichia coli verocytotoxin 1 or 2 (VT1 or VT2) are known to develop diarrhea, paralysis, and death, which can be prevented by immunization with a toxoid. The pathological effects of VT1 in the central nervous system and the gastrointestinal tract of unimmunized rabbits correlate with the localization of 125I-VT1 in these tissues, whereas in immunized animals, localization of 125I-VT1 in target tissues is inhibited and labeled toxin is cleared by the liver and spleen. By using the approach described above in this study, rabbits immunized with VT1 toxoid, VT2 toxoid, or with the A or B subunit of each toxin were challenged with intravenous 125I-VT1 or 125I-VT2. After 2 h, the animals were sacrificed, and selected tissues were analyzed for uptake of labeled toxin. It was found that animals immunized with either VT1 toxoid or VT2 toxoid were protected from target tissue uptake of both 125I-VT1 and 125I-VT2. Rabbits immunized with either the VT1 A or VT2 A subunit were also protected from target tissue uptake of both the homologous and heterologous 125I-labeled holotoxins. In contrast, in animals immunized with the toxin B subunits, protection extended only against challenge by the homologous toxin. These results provide evidence of VT1 and VT2 cross-neutralization in vivo in the rabbit model and indicate that the in vivo cross-neutralization is a function, mainly, of antibodies directed to the VT A subunits. This suggests that the VT1 A or VT2 A subunit may be a suitable immunogen for immunizing humans against systemic VT-mediated disease.

Brief heat shock treatment induces a long-lasting alteration in the glycolipid receptor binding specificity and growth rate of Haemophilus influenzae

After brief heat shock treatment, clinical strains of nontypeable Haemophilus influenzae show a long-lasting change in the binding specificity for glycolipids and a markedly increased growth rate in vitro. Non-heat-shocked H. influenzae specifically binds to phosphatidylethanolamine (PE), gangliotetraosylceramide (Gg4), and gangliotriosylceramide (Gg3) and binds minimally to sulfatoxygalactosylceramide (SGC; also called sulfatide). After a 5-min heat shock at 42 degrees C, strains of H. influenzae showed a marked increase in binding to SGC and acquired the ability to bind to sulfatoxygalactosylglycerol (SGG) in thin-layer chromatography overlays. Additionally, heat-shocked H. influenzae cells showed an increased growth rate (twofold). Increased sulfatide binding and growth rate were retained for approximately 60 generations, after which the heat-shocked organisms reverted to their original glycolipid binding pattern (i.e., PE, Gg3, and Gg4) and growth rate. Such organisms could then be reexposed to heat, and the heat shock phenotype would be reestablished. After exposure of the organisms to brief heat shock, Western blotting of a surface extract of H. influenzae with anti-bovine-brain hsp-70 monoclonal antibody showed an increase in two protein bands at 82 and 60 kDa. This antibody was a potent inhibitor of the binding of heat-shocked H. influenzae to SGC and SGG but had no effect on PE, Gg3, or Gg4 binding in vitro. In contrast, an antibody against an H. influenzae PE-Gg3-Gg4-binding adhesin that was recently identified (J. Busse, E. Hartmann, and C. A. Lingwood, J. Infect. Dis. 175:77-83, 1996) selectively inhibited the organism's binding to PE and Gg3. This indicates that cell surface hsp-70-related heat shock proteins can mediate H. influenzae attachment to sulfoglycolipids following heat shock. We suggest that such increased binding to sulfated glycolipids may be a response to fever following H. influenzae infection in humans.

Lipid modulation of glycolipid receptor function. Availability of Gal(alpha 1-4)Gal disaccharide for verotoxin binding in natural and synthetic glycolipids

Verotoxins bind to glycosphingolipids containing terminal Gal(alpha 1-4)Gal residues. Globotriaosylceramide is the most effective receptor for verotoxin-1 in vitro and is the functional plasma-membrane receptor which mediates cytopathology for most sensitive cells. Binding of verotoxin-1 to a series of galabiose-containing or globotriaose-containing synthetic glycolipids with monoalkylsulfides and bisalkylsulfides or sulfones as the lipid moiety, have been studied for toxin binding by TLC overlay and in solid phase in the presence of auxiliary lipids. The results demonstrate that for an identical carbohydrate, binding is dramatically altered according to the nature of the lipid moiety. The close proximity of the galabiose sequence and the hydrophobic species also compromised recognition. The lipid environment is also a major determinant of receptor function, since species that were effective, even preferred toxin receptors as monitored by TLC overlay, were not necessarily recognized in the presence of auxiliary lipids. Certain glycolipids, which were not recognized by TLC overlay, were nevertheless found to be effective receptors in an auxiliary lipid matrix. These results demonstrate the crucial role of the lipid moiety in verotoxin/glycolipid recognition and are discussed in relation to toxin pathogenesis and glycolipid receptor function.

Enzyme-linked immunosorbent assay for detection of immunoglobulin G antibodies to Escherichia coli Vero cytotoxin 1

The frequency of Vero cytotoxin 1 (VT1)-neutralizing antibody (NAb) in serum specimens from 790 age-stratified (0 to 70 years) control individuals from Toronto was 61 of 790 (7.7%), with a peak of 19% in the 20- to 30-year-old age group and a second peak of 16.7% in the 60- to 70-year-old age group. A total of 568 serum specimens, including 538 from the 790 Toronto control subjects, 21 from patients from three outbreaks of VT-producing Escherichia coli (VTEC) infection, and 9 known VT1-NAb-positive serum specimens from patients with hemolytic-uremic syndrome (HUS), were then tested for the presence of anti-VT1 immunoglobulin G (IgG) by an enzyme-linked immunosorbent assay (ELISA). The mean ELISA values of 522 VT1-NAb-negative serum specimens and 46 VT1-NAb-positive serum specimens were 0.09 +/- 0.06 (range, 0 to 0.56) and 0.78 +/- 0.66 (range, 0.16 to 2.91), respectively (P < 0.001; Student's t test). With a breakpoint of 0.21 (mean ELISA value of the VT1-NAb-negative sera + 2 standard deviations), the sensitivity, specificity, positive predictive value, and negative predictive value of the VT1 IgG ELISA compared with those of the VT1-NAb assay were, respectively, 95.7, 98.7, 86.3, and 99.6%. There were nine discrepant serum specimens, of which seven were anti-VT1 IgG positive and VT1-NAb negative and two were anti-VT1 IgG negative and VT1-NAb positive. The ELISA was also used for testing 238 control serum specimens from The Netherlands, Japan, and India and acute- and convalescent-phase serum specimens from 42 Toronto patients with HUS. The frequencies of anti-VT1 IgG (with VT1-NAb frequencies in parantheses) in control sera from the Netherlands, Japan, and India were 6% (3%), 1.1% (0%), and 12% (10%), respectively, with no age clustering. The frequencies of anti-VT1 IgG seropositivity in HUS patients were 5 of 14 (35.7%) in patients with unknown toxin exposure, 2 of 22 (9.1%) in individuals with known exposure to VT1 plus VT2 or VT1 alone, and 0 of 6 (0%) in patients exposed to only VT2. Development of serum anti-VT1 IgG response appears to be the exception rather than the rule in sporadic HUS patients infected with VTEC expressing VT1. However, in two family outbreaks associated with VTEC strains expressing VT1 alone and VT1 plus VT2, respectively, the presence of anti-VT1 IgG in virtually all exposed individuals who remained symptom free suggests that the presence of antibody was associated with protection.

Expression of A and B subunits of Shiga-like toxin II as fusions with glutathione S-transferase and their potential for use in seroepidemiology

We used the plasmid vector pGEX-2T for the expression of recombinant subunits of Shiga-like toxin II (SLT-II). The 5' terminus of the genes that code for either the SLT-IIA or SLT-IIB subunits was genetically fused to the 3' terminus of the gene coding for the enzyme glutathione S-transferase, which serves as a carrier in this expression system. The subunit genes were constructed synthetically by polymerase chain reaction, with appropriate restriction sites to permit in-frame downstream insertion of the genes. The resulting plasmids containing the A and B subunit genes were designated pFG1 and pFG2, respectively. Induction of Escherichia coli laboratory strains harboring pFG1 with isopropyl-beta-D-thiogalactopyranoside (IPTG) yielding only small quantities of SLT-IIA fusion proteins. Since IPTG induction was lethal for cells harboring pFG2, we constructed the recombinant plasmid pFG4, which contained a subgenic fragment of slt-IIB but without the 5' signal sequence. With this construct we were able to express very large quantities of a 33.5-kDa fusion protein, which was purified by affinity chromatography on immobilized glutathione and used as an antigen in immunoblot analysis. Rabbit serum against native SLT-II, as well as all of 12 serum samples with high neutralizing activity against SLT-II, reacted with SLT-IIB purified from an E. coli pFG4 expression system, whereas only 3 of 208 human serum samples with low neutralization titers and none of 54 serum samples with no SLT-II-neutralizing capability reacted. Failure of specific reactivity with the SLT-IIB fusion protein in the majority of human serum samples with low neutralizing activity suggests that serum factors other than immunoglobulins may be responsible for neutralizing activity in these cases. The immunoblot assay with recombinant SLT-IIB as the antigen can be recommended for use in a diagnostic setting as a simple and reliable approach to detect specific human serum antibodies to SLT-II.

Experimental verocytotoxemia in rabbits

The clinicopathologic effects of intravenously administered purified verocytotoxin 1 (VT1; Shiga-like toxin 1) in 2-kg male rabbits was studied. The 50% lethal dose was 0.2 micrograms of protein per kg of body weight (2 x 10(4) 50% cytotoxic doses per kg). The clinical features included nonbloody diarrhea and a progressive flaccid paresis, usually culminating in death. The histopathology was characterized by edema and hemorrhage in the mucosa and submucosa of the cecum and edema, hemorrhage, and neuronal necrosis in the brain and gray matter of the spinal cord. Thrombotic microangiopathy, the characteristic histopathologic renal lesion in the hemolytic-uremic syndrome, was also found to be the underlying lesion in verocytotoxemic rabbits. To determine the specific distribution of VT1 in rabbit tissues, purified 125I-labelled VT1 was administered intravenously to 20 rabbits (both immunologically naive and VT1-immune rabbits). The highest specific uptake of 125I-VT1 was in the spinal cord, brain, cecum, colon, and small bowel in unimmunized animals but in the liver, spleen, and lungs in immune animals. Immunofluorescent staining of cecal and spinal cord tissues after intravenous administration of VT1 showed evidence of specific vascular endothelial cell binding of the toxin. The striking correlation of the central nervous system and gastrointestinal localization of 125I-VT1 with the sites of known histopathology is consistent with direct toxin-mediated injury to these tissues, initiated by the specific binding of VT1 to the vascular endothelium. We conclude that the vascular damage induced by VT1 in affected rabbit tissues is similar to that seen in the kidneys and other tissues in patients with verocytotoxin-producing Escherichia coli-associated hemolytic-uremic syndrome. This suggests that although the rabbit model fails to replicate human hemolytic-uremic syndrome, it is useful for studying the pathogenesis of the vascular lesions in verocytotoxin-producing E. coli-associated diseases.