[117] Sialidase from Clostridium perfringens

A clostridial endo-beta-galactosidase that cleaves both blood group A and B glycotopes: the first member of a new glycoside hydrolase family, GH98

We have isolated an endo-beta-galactosidase designated E-ABase from Clostridium perfringens ATCC 10543 capable of liberating both the A trisaccharide (A-Tri; GalNAcalpha1-->3(Fucalpha1-->2)Gal) and B trisaccharide (B-Tri; Galalpha1-->3(Fucalpha1-->2)Gal) from glycoconjugates containing blood group A and B glycotopes, respectively. We have subsequently cloned the gene (eabC) that encodes E-ABase from this organism. This gene was found to be identical to the CPE0329 gene of C. perfringens strain 13, whose product was labeled as a hypothetical protein (Shimizu, T., Ohtani, K., Hirakawa, H., Ohshima, K., Yamashita, A., Shiba, T., Ogasawara, N., Hattori, M., Kuhara, S., and Hayashi, H. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 996-1001). Since the amino acid sequence of E-ABase does not bear detectable similarity to any of the 97 existing families of glycoside hydrolases, we have proposed to assign this unusual enzyme to a new family, GH98. We also expressed eabC in Escherichia coli BL21(DE3) and obtained 27 mg of fully active recombinant E-ABase from 1 liter of culture. Recombinant E-ABase not only destroyed the blood group A and B antigenicity of human type A and B erythrocytes, but also released A-Tri and B-Tri from blood group A(+)- and B(+)- containing glycoconjugates. The structures of A-Tri and B-Tri liberated from A(+) porcine gastric mucin and B(+) human ovarian cyst glycoprotein were established by NMR spectroscopy. The unique specificity of E-ABase should make it useful for studying the structure and function of blood group A- and B-containing glycoconju-gates as well as for identifying other glycosidases belonging to the new GH98 family.

Comparative analysis of cytotoxin, hemolysin, hemagglutinin and exocellular enzymes among clinical and environmental isolates of vibrio cholerae O139 and non-O1, non-O139

The presence of three major virulence genes toxR, tcpA and ctxA as well as expression of several putative virulence factors were compared in 12 Vibrio cholerae O139 and non-O1,non-O139 strains of clinical and environmental origin. All the strains possessed the gene encoding the regulatory protein TOXR. None of the non-O1, non-O139 strains as well as one of the O139 environmental strains carried the genes for ctxA and tcpA. Statistically significant differences in hemagglutinin and hemolysin production were observed amongst the strains depending on the source of their isolation. Expression of extracellular enzymes such as protease, elastase, neuraminidase, phospholipase A and phospholipase C, however, did not vary significantly from the groups of strains isolated from different sources.

X-Neu5Ac: a novel substrate for chromogenic assay of neuraminidase activity in bacterial expression systems

A chromogenic substrate 1, 5-bromo-4-chloroindol-3-yl 5-acetamido-3,5-dideoxy-alpha-D-glycero-D-galacto-2-nonulopyranosidon ic acid (X-Neu5Ac), has been synthesized to facilitate the screening of bacterial colonies or plaques for the detection of either natural or mutant neuraminidase activity. Substrate 1 was hydrolyzed by neuraminidase isolated from Clostridium perfringens to release a halogenated indol-3-ol 2 that undergoes rapid aerobic oxidation to form the dark blue pigment, 5,5'-dibromo-4,-4'-dichloroindigo 3. Preliminary kinetic studies indicate that this compound is a good substrate (Km 0.89 x 10(-3) M) for neuraminidase and is quite stable under identical conditions in the absence of enzyme. These results suggest that X-Neu5Ac 1 can be useful to screen for bacterially-encoded enzyme production directly on agar plates.

Isolation from opossum serum of a metalloproteinase inhibitor homologous to human alpha 1B-glycoprotein

Fractionation of opossum (Didelphis virginiana) serum with (NH4)2SO4, followed by chromatography on DEAE-Sepharose, phenyl-Sepharose, and Mono Q HR 5/5, has resulted in the isolation in homogeneous condition of a metalloproteinase inhibitor designated oprin (opossum proteinase inhibitor). Oprin is a single-chain glycoprotein (26% carbohydrate) with an estimated Mr = 52,000, pI = 3.5, and E(1%/1 cm) = 11. Oprin inhibited snake venom metalloproteinases, but showed no activity on venom serine proteinases or on bacterial metalloproteinases. Incubation of Crotalus atrox alpha-proteinase (EC 3.4.24.1) with oprin, and analysis of the reaction products by chromatography on Mono Q HR 5/5 and by electrophoresis under nondenaturing conditions, indicated formation of an inactive enzyme/inhibitor complex. The complex dissociated during SDS/polyacrylamide gel electrophoresis. An opossum liver cDNA library was immunoscreened, and clones containing cDNA encoding for part of the open reading frame for oprin were isolated. The cDNA inserts contained nucleotide sequences corresponding to two internal amino acid sequences of oprin which had been separately determined by protein sequence analysis. Protein database screening using a 211 amino acid sequence deduced from one of the cDNA inserts showed no significant homology to known proteinase inhibitors. There was, however, a 36% identity with human alpha 1B-glycoprotein, a plasma protein of unknown function related to the immunoglobulin supergene family. In addition, the amino-terminal sequence of oprin showed 46% identity with human alpha 1B-glycoprotein in a 26 amino acid residue overlap.(ABSTRACT TRUNCATED AT 250 WORDS)

Ethanol effects on total sialic acid of various brain regions and visceral organs

These experiments were designed to extend our earlier observations on acute ethanol effects on brain sialic acid (SA), an acidic sugar component of membrane gangliosides and many glycoproteins. Here, we tested for differential effects of ethanol on total sialic acid in various brain regions and for effects on other organs, such as liver, kidney, heart, and skeletal muscle. Subjects were young adult, male rats. The first experiment compared two commonly accepted analytical methods for brain SA on peripheral tissues. Consistently higher levels were evident with the resorcinol method in all tissues, especially in liver, compared with the thiobarbiturate method. Resorcinol-measured levels in the liver and kidney were on the order of 350 micrograms/g, wet weight. In the brain of controls, the resorcinol method revealed total SA levels to be on the order of 550-650 micrograms/g in the basal ganglia, cerebral cortex, cerebellum, and hippocampus. Brainstem levels were significantly lower. Ethanol, given IP at 2 g/kg, seemed to decrease total SA in all brain regions at four hours after injection, with statistically significant decreases in the hippocampus and brainstem. With 3 g/kg, only the cerebellum showed a significant decrease at four hours, compared to saline-injected controls, but the decrease was large (25%). Analysis of the other organs showed that, compared to saline-injected controls, ethanol decreased SA in the liver. There was a small, but significant, decrease in heart SA at four hours after 3 g/kg. In skeletal muscle, ethanol significantly increased total SA at 2 g/kg, but not at 3 g/kg.(ABSTRACT TRUNCATED AT 250 WORDS)

Number and topography of epitopes of human chorionic gonadotropin (hCG) are shared by desialylated and deglycosylated hCG

A previously established map of the surface epitopes of human chorionic gonadotropin (hCG) served as template for the present study in which we investigated the antigenic surfaces of two glycosylation variants of hCG, i.e. desialylated hCG (asialo-hCG) and deglycosylated hCG (degly-hCG). This map allocates five epitopes to the alpha subunit, five to the beta subunit and four alpha beta epitopes to structures formed only by the alpha/beta heterodimer holo-hCG (Schwarz et al. (1986) Endocrinology 118, 189-197; Berger et al. (1990) J. Endocrinol. 125, 301-309). Here it is described that both variants complied with this template: each of the 14 distinct monoclonal antibodies with which the epitopes of hCG were defined reacted with radiolabeled asialo-hCG and degly-hCG as well and generally bound degly-hCG with greater affinity than hCG. Moreover, every combination of capture and radiolabeled detection antibody that was either compatible or incompatible on unlabeled hCG was so also on unlabeled asialo-hCG and degly-hCG. It thus appears that alterations of the carbohydrate structure of hCG can be associated with a change in affinity between some antibodies and their respective epitopes but not with a loss of an epitope or with a change in the topographical relationships of the 14 epitopes.

Relationship of orientation with affinity and activity of receptor-bound glycosylation variants of human chorionic gonadotropin (hCG), as visualized by monoclonal antibodies (MCA)

When hCG was receptor-bound, only 2 epitopes (i.e. beta 3 and beta 5) out of its previously mapped total of 14 surface epitopes (10, 11, 12) could be detected by the respective 125I-MCA. Clearly, this indicates a non-random orientation of hCG in this particular state (1). Now we report that on receptor-bound desialylated (asialo-hCG) as well as on receptor-bound deglycosylated hCG (degly-hCG), the beta 3 and beta 5 epitopes were inaccessible for 125I-MCA as were the remaining epitopes, although both variants, when not receptor-bound, were indistinguishable from native hCG with respect to number and topography of epitopes. Thus, the carbohydrate (CHO) units of hCG neither seem to be part of these 14 antigenic sites nor to contribute to the affinity of receptor binding: both variants had even higher affinities than native hCG. However, since the CHO are known to be obligatory for triggering the postreceptor responses of hCG-stimulated target cells, as seen by the 50% reduced or totally abolished biological potency of asialo-hCG and degly-hCG, respectively, the here demonstrated clear-cut differences in epitope accessibility can be related to differences in receptor-bound orientations which reflect signal transduction-competent and incompetent modes of interaction with the receptor. We believe that the CHO, while not contributing to receptor binding per se, function to assure correct positioning of hCG in the ligand recognition domain to allow for proper protein-protein interactions between ligand and receptor and thus optimal activation and hence transduction of the external signal to the cell's interior. In addition, the fact that most of the surface epitopes were masked on receptor-bound hCG, represents the first experimental support for the sofar unproven hypothesis that this unusually long (i.e. 341 amino acid residues) extracellular N' terminal domain of the recently cloned hCG receptor (hCG-R) (23,24), is indeed involved in ligand recognition.

Effect of acute injections of ethanol on lipid and protein-bound sialic acid in mice of different ages

Mice of different age groups (weanling, young adult and aged) were tested for changes in brain lipid- and protein-bound sialic acid (SA) 2 h after ethanol (2 g/kg, i.p.), either as a single dose or after binge dosing of five repeated doses of ethanol spaced 2 h apart. The results clarify our earlier demonstrations that acute ethanol can reduce whole brain SA. Ethanol generally decreased SA of both gangliosidic and glycoprotein origin, with the effect varying with number of doses and mouse age. Single-dose ethanol decreased both lipid-bound and protein-bound SA in young adults and decreased lipid-bound SA in aged mice. There was no effect on lipid-bound SA in weanlings, but weanlings did have a 72% decline in protein-bound SA. Repeated injections in young adults did not cause the SA decrease seen with acute injection. In both weanling and aged mice, however, repeated injections did cause large decreases in both lipid- and protein-bound SA. Small, but statistically significant, changes also occurred in free SA. Ethanol increased free SA in singly-dosed young adults and in multiply-dosed aged adults, while causing a distinct decrease in singly-dosed weanlings.

Differences in susceptibility of rat liver and brain sialidases to ethanol and gangliosides

Based on reports that ethanol can decrease the level of sialic acid (SA) (neuraminic acid) in several tissues, we tested the hypothesis that ethanol promotes SA cleavage by enhancing the activity of sialidases (neuraminidases). We also investigated whether brain and liver sialidases have the same response to ethanol and gangliosides, especially since our prior studies have demonstrated that gangliosides could antagonize ethanol-induced behavior. Experiments were conducted on homogenates of brain and liver and of liver slices of adult rats. In liver slices, cleavage of SA did not fall in proportion to the ethanol-induced inhibition of sialidase; in fact, at 0.1 M ethanol, free SA increased, even though sialidase was inhibited. Brain sialidase activity on endogenous sialoglycoconjugates was much more resistant to ethanol than liver sialidase and was fully active even in concentrations as high as 1 M. When gangliosides were incubated with liver slices in the absence of ethanol, sialidase was markedly stimulated. The ethanol-induced inhibition of sialdase in liver slices was mimicked by sorbitol, suggesting that the inhibition may be caused by a shift in redox state as a result of increased NADH. The ethanol metabolite, acetaldehyde, does not seem to be a factor, because sialidase inhibition still occurred when slices were incubated with ethanol containing pyrazole. The results indicate that ethanol promotes the accumulation of free SA in liver without stimulating sialdase; our other work suggests that the cause is an increase in accessibility to sialoglycoconjugates rather than decreased utilization of SA. Brain and liver sialidases clearly respond differently to both ethanol and gangliosides.

Ethanol promotes hydrolysis of 3H-labeled sialoconjugates from brain of mice in vitro

Acute administration of ethanol reportedly decreases total sialic acid in brain. Here, we tested the hypothesis in brain and liver that the decrement is due to increased hydrolysis of sialoglycoconjugates. Mouse tissue slices were pulse-labeled with N-[3H]acetyl-D-mannosamine, the precursor of sialic acid. Incorporation was linear for up to 4 hr of incubation. When the labeled slices were incubated with three concentrations of ethanol (0.1, 0.5, and 1 M) for 5 hr, labeled liver sialoconjugates were significantly affected only at 0.5 and 1 M ethanol, whereas labeled brain sialoconjugates were markedly decreased even at 100 mM ethanol. Sialidase activity decreased steadily with increasing concentration of ethanol, indicating that the increased hydrolysis was not attributable to an enhanced sialidase activity. n-Propanol and t-butanol had the same degradative effect as ethanol on sialocompounds; and 3 mM pyrazole, an inhibitor of alcohol dehydrogenase (ADH), had no effect on ethanol-induced degradation of sialocompounds. The protein/DNA ratio in liver showed a steady decrease with increasing ethanol. The data thus confirm the in vivo reports of ethanol-enhanced cleavage and rule out any increase in sialidase activity as a major cause.

The anionic sites at luminal surface of peritubular capillaries in rats

Anionic sites have been demonstrated in the basement membranes of peritubular capillaries. The anionic barrier function of peritubular capillary wall has been ascribed to these sites. Fenestrated capillaries in other organs have anionic sites in the endothelial cell glycocalyx and at the luminal surface of the fenestral diaphragms. The purpose of this study was to map anionic sites at the luminal surface of peritubular capillaries and to assess whether a concentration gradient for albumin exists across the endothelium. Partial chemical characterization of these anionic sites was done by in vivo enzymatic degradation. The difference in distribution of albumin following enzyme digestion was also studied. The binding of cationized ferritin to the luminal surface indicated that the rat peritubular capillaries have anionic sites along the entire luminal surface of the endothelial cell, including the fenestral diaphragms. Partial biochemical characterization of these sites shows that the sites in the glycocalyx are mainly from neuraminic acid, while the fenestral diaphragms have mainly heparan sulfate proteoglycans. Intravascular albumin extended to the endothelial luminal plasmalemma and to the luminal surface of fenestral diaphragms. Digestion with heparitinase was associated with the leakage of albumin outside the capillary wall. These findings suggest that the anionic surface of fenestrae constitutes a charge barrier of the peritubular capillaries.

Isoelectric focusing patterns of urinary kallikrein in Dahl salt-hypertension susceptible and resistant rats

Dahl salt-sensitive (S) rats which are susceptible to hypertension have lower urinary kallikrein excretion than salt-resistant (R) rats which are not susceptible. Some physicochemical characteristics of partially purified urinary kallikrein were compared between the S and R strains. The isoelectric focusing pattern of S kallikrein was shifted so that a higher proportion of enzyme was present in isoelectric forms that had higher pI values compared to the pattern for R kallikrein. This strain difference was unique to urinary kallikrein; it was not seen in kallikrein extracted from salivary glands. The isoelectric focusing pattern for R urinary kallikrein could be converted to an S-type pattern by treatment with neuraminidase, which suggests that the differing isoelectric focusing patterns arose from differences in the sialic acid content of the kallikrein. The S kallikrein was slightly more heat-labile than R kallikrein, which was also compatible with the lower sialic acid content of the S enzyme. Tests involving the active site of the enzyme (Km values, pH curves, and heat of activation) were identical for the S and R strains. It was concluded that the structural differences observed in urinary kallikrein between S and R strains were compatible with strain-specific posttranslational processing of the enzyme.

Modulation of neuraminidase activity by the physical state of phospholipid bilayers containing gangliosides Gd1a and Gt1b

The thermotropic behavior of large unilamellar dipalmitoylphosphatidylcholine vesicles containing the disialoganglioside Gd1a and the trisialoganglioside Gt1b on their outer surface has been studied as a function of the ganglioside molar fraction and Ca2+ concentration by using high-sensitivity differential scanning calorimetry and steady-state fluorescence spectroscopy. These studies indicate that both gangliosides have an ordering effect on the hydrocarbon region of the bilayer and that this effect is enhanced by the presence of Ca2+ ions. The calorimetric experiments also indicate that ganglioside Gt1b has an intrinsic tendency to phase separate into compositional-rich ganglioside domains even in the absence of Ca2+. Ganglioside Gd1a, on the other hand, only phase separates at Ca2+ concentrations equal to or higher than 10 mM. These studies have allowed us to identify and evaluate the factors affecting the rates of hydrolysis of gangliosides by the soluble neuraminidase from Clostridium perfringens. The data presented in this paper indicate that the rates of hydrolysis of membrane-bound gangliosides are correlated to the physical state of the membrane and the state of aggregation of the ganglioside molecules within the lipid bilayer. For membrane-bound gangliosides, maximal activation energies were found at temperatures slightly below the lipid phase transition temperature. The rates of hydrolysis of the soluble substrate sialyllactose or that of the micellar ganglioside is independent of Ca2+ concentration, whereas the rates of hydrolysis of membrane-bound ganglioside are inhibited by Ca2+ especially under conditions in which the clustering effect of Ca2+ is maximal. These studies suggest that the soluble neuraminidases from Clostridium perfringens prefer ganglioside substrates that are dispersed within the membrane and not forming part of largely aggregated clusters.

The carbohydrate on human chorionic gonadotropin produced by cancer cells

Using the methods described, it is not possible to determine the number of N- and O-linked oligosaccharides on ectopic hCG beta. On standard hCG beta there are two NeuAc residues on each N- and O-linked oligosaccharide, so that the number of NeuAc residues is proportional to the number of oligosaccharides. Ectopic hCG beta and desialylated ectopic hCG beta are of similar molecular size to the standard preparations (gel filtration and RIA with anti-CTP antisera, data not presented). This suggests that ectopic hCG beta is sialylated to a similar extent as standard hCG beta, so the number of oligosaccharides on ectopic hCG beta could be similar to the number on standard hCG beta. There is a Fuc attached to the N-linked oligosaccharides of standard hCG beta (Fig. 3). Using the methods described, it was not possible to determine if this residue is also found on the N-linked oligosaccharides of ectopic hCG beta. Recently, a second form of ectopic hCG beta was identified (22). This form lacks the characteristic hCG beta carboxyterminal peptide, and as such is unrecognized by the RIA used in this study. Like the ectopic hCG beta described herein, and that produced by other cancers, this molecule only partially binds to Con A, and binds to Ricinus communis-120 following neuraminidase digestion. Intact hCG and free hCG subunits, which only partially bind to Con A, are found in cancer tissues, cancer sera, and the medium of cultured trophoblastic and nontrophoblastic cancer cells (Table 1). Our studies with DoT cancer of the cervix cells clearly indicate that the partial binding could be the consequence of the linkage of extra beta G1cNAc residues.

Neuraminidase-activated attachment of Actinomyces naeslundii ATCC 12104 to human buccal epithelial cells

The attachment of Actinomyces naeslundii ATCC 12104 to human buccal epithelial cells pre-treated with neuraminidase (sialidase) was evaluated. Both commercial clostridial neuraminidase and neuraminidase preparations from the test strain of A. naeslundii enhanced attachment. The results suggest that the A. naeslundii beta-galactoside-seeking ligand involved in hemagglutination and interbacterial coaggregation also mediates one type of binding to human buccal epithelial cells.

Role of membrane integral proteins in the modulation of red cell shape by albumin, dinitrophenol and the glass effect

Defatted serum albumin is found to induce a cup shape in erythrocytes. At 40 mg/ml of albumin, approx. 80% of washed erythrocytes possess this morphology, which can be reversed to disc shape by dinitrophenol. Erythrocytes treated with trypsin, papain, pronase or neuraminidase show enhanced susceptibility to cup formation by albumin; however, chymotrypsinized erythrocytes exhibit a normal response. Red cells treated with concanavalin A (but not its succinylated derivative) show resistance to the cupping effect of albumin as well as the crenating effects of dinitrophenol and glass. The resistance develops after about 20 min following the exposure of cells to the lectin, and is rapidly abrogated on removal of the bound lectin by alpha-methylmannoside. Incubation of the concanavalin A-exposed cells at low temperature leads to prolongation of the time required to achieve the resistance. These results indicate an involvement of membrane integral proteins in mediating the shape modulating effects of albumin, dinitrophenol and exposure to glass.

Deficiency of neuraminidase in the sialidoses and the mucolipidoses

Neuraminidase activity in cultured fibroblasts from patients either with various forms of sialidosis or with I-cell disease (ICD) or mucolipidosis (ML) III has been determined by both a colorimetric and a fluorometric method. The former applied to frozen fibroblast pellets demonstrated a specific deficiency of neuraminidase in patients with the sialidoses. The enzyme was also deficient in I-cells, as were other lysosomal hydrolases. With the fluorogenic substrate these data could be confirmed and extended, and elementary kinetics of neuraminidase studied. In unfrozen freshly harvested fibroblasts, neuraminidase activity was severalfold that in frozen aliquots. A comparative and simultaneous study could not reveal substantial differences between the residual neuraminidase activity found in the various clinical forms of sialidosis. And, in fibroblasts from patients with ICD, also called ML II, the deficiency of this enzyme is quantitatively similar to that in the sialidoses, but the residual activity in ML III is three times higher. In both ML II and ML III the defect is probably secondary to the unknown metabolic error.

A 2 (N2) neuraminidase of the X-7 influenza virus recombinant: determination of molecular size and subunit composition of the active unit

Neuraminidase activity of influenza virus was directly seen on sodium dodecyl sulfate polyacrylamide gels with the aid of the synthetic substrate, methoxyphenol neuraminic acid. Neuraminidase (NA) appeared as a high-molecular-weight fraction with a size in the range of 220,000 to 250,000 daltons. Isolation of this fraction from the X-7 strain of influenza virus, dissociation with sodium dodecyl sulfate, and reduction showed the presence of two polypeptides of 66,000 (NA(1)) and 58,000 (NA(2)) molecular weights in equimolar concentration. We postulate that the minimum active unit for the viral A(2) neuraminidase is a tetramer composed of two NA(1) and two NA(2) subunits.