A tale of tails: Sialidase is key to success in a model of phage therapy against K1-capsulated Escherichia coli.

Prior studies treating mice infected with Escherichia coli O18:K1:H7 observed that phages requiring the K1 capsule for infection (K1-dep) were superior to capsule-independent (K1-ind) phages. We show that three K1-ind phages all have low fitness when grown on cells in serum whereas fitnesses of four K1-dep phages were high. The difference is serum-specific, as fitnesses in broth overlapped. Sialidase activity was associated with all K1-dep virions tested but no K1-ind virions, a phenotype supported by sequence analyses. Adding endosialidase to cells infected with K1-ind phage increased fitness in serum by enhancing productive infection after adsorption. We propose that virion sialidase activity is the primary determinant of high fitness on cells grown in serum, and thus in a mammalian host. Although the benefit of sialidase is specific to K1-capsulated bacteria, this study may provide a scientific rationale for selecting phages for therapeutic use in many systemic infections.

Redirection of sialic acid metabolism in genetically engineered Escherichia coli.

Most microorganisms do not produce sialic acid (sialate), and those that do appear to use a biosynthetic mechanism distinct from mammals. Genetic hybrids of nonpathogenic, sialate-negative laboratory Escherichia coli K-12 strains designed for the de novo synthesis of the polysialic acid capsule from E. coli K1 proved useful in elucidating the genetics and biochemistry of capsule biosynthesis. In this article we propose a dynamic model of sialometabolism to investigate the effects of biosynthetic neu (N-acetylneuraminic acid) and catabolic nan (N-acylneuraminate) mutations on the flux of intermediates through the sialate synthetic pathway. Intracellular sialate concentrations were determined by high pH anion exchange chromatography with pulsed amperometric detection. The results indicated that a strain carrying a null defect in the gene encoding polysialyltransferase (neuS) accumulated > 50 times more CMP-sialic acid than the wild type when strains were grown in a minimal medium supplemented with glucose and casamino acids. Metabolic accumulation of CMP-sialic acid depended on a functional sialic acid synthase (neuB), as shown by the inability of a strain lacking this enzyme to accumulate a detectable endogenous sialate pool. The neuB mutant concentrated trace sialate from the medium, indicating its potential value for quantitative analysis of free sialic acids in complex biological samples. The function of the sialate aldolase (encoded by nanA) in limiting intermediate flux through the synthetic pathway was determined by analyzing free sialate accumulation in neuA (CMP-sialic acid synthetase) nanA double mutants. The combined results demonstrate how E. coli avoids a futile cycle in which biosynthetic sialate induces the system for its own degradation and indicate the feasibility of generating sialooligosaccharide precursors through targeted manipulation of sialate metabolism.

Adaptation of signature-tagged mutagenesis to Escherichia coli K1 and the infant-rat model of invasive disease.

With the exception of the polysialic acid capsule (K1 antigen), little is known about other virulence factors needed for systemic infection by Escherichia coli K1, the leading cause of Gram-negative neonatal meningitis in humans. In this work, the functional genomics method of signature-tagged mutagenesis (STM) was adapted to E. coli K1 and the infant-rat model to identify non-capsule virulence genes. Validation of the method was demonstrated by the failure to recover a reconstructed acapsular mutant from bacterial pools used to systemically infect 5-day-old rats. Three new genes required for systemic disease were identified from a total of 192 mutants screened by STM (1.56% hit rate). Gut colonization, Southern blot hybridization, mixed-challenge infection, and DNA sequence analyses showed that the attenuating defects in the mutants were associated with transposon insertions in rfaL (O antigen ligase), dsbA (thiol:disulfide oxidoreductase), and a new gene, puvA (previously unidentified virulence gene A), with no known homologues. The results indicate the ability of STM to identify novel systemic virulence factors in E. coli K1.

Strain-specific iron-dependent virulence in Escherichia coli.

For reasons unknown, certain Escherichia coli strains become highly virulent when injected with hemoglobin or other soluble iron sources. Two clinical isolates (virulent and nonvirulent) showed equivalent hemoglobin-mediated growth acceleration in vitro. However, when injected intraperitoneally into mice without hemoglobin, the virulent strain was cleared more slowly (t(1/2), >4 h vs. <30 min). The virulent E. coli strain had a polysialic acid-containing capsule, whereas the nonvirulent strain did not. Virulent E. coli grown at 20 degrees C (which blocks polysialylation) were cleared as rapidly as nonvirulent organisms. In another virulent E. coli strain having abundant outer membrane polysialic acid, targeted deletion of the polysialyltransferase accelerated host clearance and blocked iron-dependent virulence. The iron-dependent virulence of certain E. coli strains may represent the combined effect of slow in vivo clearance-associated, in this case, with outer membrane polysialylation coupled with accelerated growth permitted by iron compounds.

Sialic acid metabolism's dual function in Haemophilus influenzae.

Many bacterial commensals and pathogens use the sialic acids as carbon and nitrogen sources. In Escherichia coli, the breakdown of these sugars is catalysed by gene products of the nan (Nacylneuraminate) operon; other microorganisms may use a similar catabolic strategy. Despite the known ligand and antirecognition functions of the sialic acids, the contribution of their catabolism to infection or host colonization has never been directly investigated. We addressed these questions with Haemophilus influenzae type b, which metabolizes relatively few carbohydrates, using the infant-rat infection model. The predicted H. influenzae homologue (HI0142) of the E. coli sialic acid aldolase structural gene, nanA, was subcloned and mutagenized by insertion of a kanamycin resistance cassette. Phenotypic investigation of the resulting H. influenzae aldolase mutants showed that: (i) HI0142 is essential for sialic acid degradation; (ii) the products of the open reading frames (ORFs) flanking HI0142 (HI0140, 41, 44 and 45) are likely to have the same functions as those of their counterparts in E. coli; (iii) sialylation of the lipooligosaccharide (LOS) epitope recognized by monoclonal antibody 3F11 is dependent on an environmental source of sialic acid; (iv) a nanA mutant hypersialylates its LOS sialyl acceptor, corresponding to an apparent increased fitness of the mutant in the infant-rat model; and (v) expression of the LOS sialyl acceptor is altered in cells grown without exogenous sialic acid, indicating the direct or indirect effect of sialic acid metabolism on LOS antigenicity. Taken together the data show the dual role of sialic acid catabolism in nutrition and cell surface modulation.

Convergent pathways for utilization of the amino sugars N-acetylglucosamine, N-acetylmannosamine, and N-acetylneuraminic acid by Escherichia coli.

N-Acetylglucosamine (GlcNAc) and N-acetylneuraminic acid (NANA) are good carbon sources for Escherichia coli K-12, whereas N-acetylmannosamine (ManNAc) is metabolized very slowly. The isolation of regulatory mutations which enhanced utilization of ManNAc allowed us to elucidate the pathway of its degradation. ManNAc is transported by the manXYZ-encoded phosphoenolpyruvate-dependent phosphotransferase system (PTS) transporter producing intracellular ManNAc-6-P. This phosphorylated hexosamine is subsequently converted to GlcNAc-6-P, which is further metabolized by the nagBA-encoded deacetylase and deaminase of the GlcNAc-6-P degradation pathway. Two independent mutations are necessary for good growth on ManNAc. One mutation maps to mlc, and mutations in this gene are known to enhance the expression of manXYZ. The second regulatory mutation was mapped to the nanAT operon, which encodes the NANA transporter and NANA lyase. The combined action of the nanAT gene products converts extracellular NANA to intracellular ManNAc. The second regulatory mutation defines an open reading frame (ORF), called yhcK, as the gene for the repressor of the nan operon (nanR). Mutations in the repressor enhance expression of the nanAT genes and, presumably, three distal, previously unidentified genes, yhcJIH. Expression of just one of these downstream ORFs, yhcJ, is necessary for growth on ManNAc in the presence of an mlc mutation. The yhcJ gene appears to encode a ManNAc-6-P-to-GlcNAc-6-P epimerase (nanE). Another putative gene in the nan operon, yhcI, likely encodes ManNAc kinase (nanK), which should phosphorylate the ManNAc liberated from NANA by the NanA protein. Use of NANA as carbon source by E. coli also requires the nagBA gene products. The existence of a ManNAc kinase and epimerase within the nan operon allows us to propose that the pathways for dissimilation of the three amino sugars GlcNAc, ManNAc, and NANA, all converge at the step of GlcNAc-6-P.

Neuraminidase (sialidase) activity of Haemophilus parasuis.

Neuraminidase (sialidase), a potential virulence factor in bacteria, was demonstrated in Haemophilus parasuis, an invasive swine pathogen, but not in four other pathogens of the Pasteurellaceae family: H. influenzae, H. somnus, H. paragallinarum, or Actinobacillus pleuropneumoniae. H. parasuis neuraminidase had an acidic pH optimum and a specificity for several substrates also cleaved by other bacterial neuraminidases. Similar to the neuraminidase of Pasteurella multocida, H. parasuis neuraminidase was cell associated and did not require divalent cations for activity. Exogenous sialic acid added to growth medium of H. parasuis was cleared after a lag of about 10 h and these cultures grew to a greater final density than cultures without added sialic acid, indicating that exogenous sialic acid is metabolized. The role of sialidase in providing nutrients to H. parasuis may be an important factor in its obligate parasitism.

Purification and characterization of the Escherichia coli K1 neuB gene product N-acetylneuraminic acid synthetase.

Escherichia coli K1 produces a capsular polysaccharide of alpha(2-8) poly-N-acetylneuraminic acid. This polysaccharide is an essential virulence factor of these neuropathogenic bacteria. The genes necessary for the synthesis of neuNAc were localized to a plasmid containing the neuBAC genes of the K1 gene cluster. Cells harboring the neuB+ allele in an aldolase (nanA-) negative background produce neuNAc in vivo. Enzymatic synthesis of neuNAc could be demonstrated in extracts of cells harboring an expression plasmid (pNEUB) containing the neuB gene alone. NeuNAc synthetase was purified to homogeneity from extracts of cells harboring pNEUB. The molecular weight of the purified enzyme is 40 kDa, similar to that predicted by the nucleotide sequence of the neuB gene. The amino terminal sequence of the purified protein matches that predicted by the nucleotide sequence of the neuB gene. NeuNAc synthetase catalyzes the formation of neuNAc as indicated by its coupling to the CMP-neuNAc synthetase reaction. The enzyme condenses manNAc and PEP with the release of phosphate. The E. coli neuNAc synthetase is specific for manNAc and PEP, unlike rat liver enzyme that utilizes N-acetylmannosamine-6-phosphate to form neuNAc-9-PO4. This represents the first report of a purification of a sialic acid synthetase from either a eukaryotic or prokaryotic source to homogeneity. These experiments clearly demonstrate an aldolase-independent sialic acid synthetase activity in E. coli K1.

Reduced polysialic acid capsule expression in Escherichia coli K1 mutants with chromosomal defects in kpsF.

Neuroinvasive Escherichia coli K1 synthesizes and assembles a polysialic acid capsule virulence factor on the external leaflet of the outer membrane. This capsule functions in pathogenesis by blocking non-immune host defence mechanisms and acting as a relatively non-immunogenic molecular mimic of the polysialic acid chains found in high concentrations on neural cell adhesion molecules of the human embryo and neonate. The synthetic, regulatory and export components for capsule expression are encoded in three functionally distinct gene blocks or regions of the 20 kb kps 'pathogenicity island'. These regions are organized as two convergently transcribed operons inserted into the monocistronic tRNA gene, pheV. The six genes of the so-called region 1 operon are transcribed in the same direction as pheV, and at least four of these genes are required for polysialic acid export. Expression of this operon is thermoregulated by transcriptional control of its first gene, kpsF. To investigate the function of region 1 further, two independent chromosomal disruptions were engineered by inserting promoterless, terminatorless kanamycin or chloramphenicol resistance cassettes into the HindIII site of the kpsF coding sequence. The chromosomal insertions were regulated by temperature in the same way as the wild-type operon, demonstrating that this control mechanism remained intact in these mutants. Chemical, immunological and ultrastructural microscopical methods demonstrated that full-length polysialic acid chains were synthesized but not exported by the kpsF mutants. This phenotype was correlated with decreased plaque diameter when the mutants were infected with the capsule-specific bacteriophage K1F. The export defect could not be complemented in trans with kpsF+ containing its cis-regulatory region because of titration of an apparent positive regulator of region 1 expression, whereas complementation was observed with a plasmid expressing kpsF from a physiologically irrelevant promoter. An N-terminal polyhistidine peptide was attached to KpsF and used to purify the overproduced polypeptide. Antibodies raised against KpsF identified at least one of its paralogues in E. coli, GutQ, suggesting that KpsF and its homologues are membrane associated. The results indicate the requirement for a precise balance between region 1 components of the capsule export machinery, and that KpsF plays a positive role in the assembly, operation or regulation of this apparatus.

Direct PCR analysis for toxigenic Pasteurella multocida.

A more rapid, accurate method to detect toxigenic Pasteurella multocida is needed for improved clinical diagnosis, farm biosecurity, and epidemiological studies. Toxigenic and nontoxigenic P. multocida isolates cannot be differentiated by morphology or standard biochemical reactions. The feasibility of using PCR for accurate, rapid detection of toxigenic P. multocida from swabs was investigated. A PCR protocol which results in amplification of an 846-nucleotide segment of the toxA gene was developed. The PCR amplification protocol is specific for toxigenic P. multocida and can detect fewer than 100 bacteria. There was concordance of PCR results with (i) detection of toxA gene with colony blot hybridization, (ii) detection of ToxA protein with colony immunoblot analysis, and (iii) lethal toxicity of sonicate in mice in a test set of 40 swine diagnostic isolates. Results of an enzyme-linked immunosorbent assay for ToxA agreed with the other assays except for a negative reaction in one of the 19 isolates that the other assays identified as toxigenic. In addition to accuracy, as required for a rapid direct specimen assay, toxigenic P. multocida was recovered efficiently from inoculated swabs without inhibition of the PCR. The results show that PCR detection of toxigenic P. multocida directly from clinical swab specimens should be feasible.

Thermoregulation of kpsF, the first region 1 gene in the kps locus for polysialic acid biosynthesis in Escherichia coli K1.

The kps locus for biosynthesis of the capsular polysialic acid virulence factor in Escherichia coli K1 contains at least two convergently transcribed operons, designated region 1 and regions 2 plus 3. On the basis of DNA sequence analysis, kpsF appeared to be a good candidate for the first gene of region 1 (M. J. Cieslewicz, S. M. Steenbergen, and E. R. Vimr, J. Bacteriol. 175:8018-8023, 1993). A preliminary indication that kpsF is required for capsule production is the capsule-negative phenotype of an aph T insertion in the chromosomal copy of kpsF. The present communication describes the isolation and phenotypic characterization of this mutant. Although transcription through kpsF was required for capsule production, complementation analysis failed to indicate a clear requirement for the KpsF polypeptide. However, since E. coli contains at least two other open reading frames that could code for homologs of KpsF, the apparent dispensability of KpsF remains provisional. DNA sequence analysis of 1,100 bp upstream from the kpsF translational start site did not reveal any open reading frames longer than 174 nucleotides, consistent with kpsF being the first gene of region 1. Since kpsF appeared to be the first gene of a region whose gene products are required for polysialic acid transport and because capsule production is known to be thermoregulated, primer extension analyses were carried out with total RNA isolated from cells grown at permissive (37 degrees C) and nonpermissive (20 degrees C) temperatures. The results revealed a potentially complex kpsF promoter-like region that was transcriptionally silent at the nonpermissive temperature, suggesting that thermoregulation of region 1 may be exerted through variations in kpsF expression. Additional evidence supporting this conclusion was obtained by demonstrating the effects of temperature on expression of the gene kpsE, immediately downstream of kpsF. Chloramphenicol acetyltransferase assays were carried out with constructs containing the kpsF 5' untranslated region fused to a promoterless cat cassette, providing further evidence that kpsF is thermoregulated. Although the function of KpsF is unclear, primary structure analysis indicated two motifs commonly observed in regulatory proteins and homology with glucosamine synthase from Rhizobium meliloti.

The structures of Salmonella typhimurium LT2 neuraminidase and its complexes with three inhibitors at high resolution.

The structure of Salmonella typhimurium LT2 neuraminidase (STNA) is reported here to a resolution of 1.6 angstroms together with the structures of three complexes of STNA with different inhibitors. The first is 2-deoxy-2,3-dehydro-N-acetyl-neuraminic acid (Neu5Ac2en or DANA), the second and third are phosphonate derivatives of N-acetyl-neuraminic acid (NANA) which have phosphonate groups at the C2 position equatorial (ePANA) and axial (aPANA) to the plane of the sugar ring. The complex structures are at resolutions of 1.6 angstroms, 1.6 angstroms and 1.9 angstroms, respectively. These analyses show the STNA active site to be topologically inflexible and the interactions to be dominated by the arginine triad, with the pyranose rings of the inhibitors undergoing distortion to occupy the space available. Solvent structure differs only around the third phosphonate oxygen, which attracts a potassium ion. The STNA structure is topologically identical to the previously reported influenza virus neuraminidase structures, although very different in detail; the root-mean-square (r.m.s) deviation for 210 C alpha positions considered equivalent is 2.28 angstroms (out of a total of 390 residues in influenza and 381 in STNA). The active site residues are more highly conserved, in that both the viral and bacterial structures contain an arginine triad, a hydrophobic pocket, a tyrosine and glutamic acid residue at the base of the site and a potential proton-donating aspartic acid. However, differences in binding to O4 and to the glycerol side-chain may reflect the different kinetics employed by the two enzymes.

In vivo enzymatic removal of alpha 2-->6-linked sialic acid from the glomerular filtration barrier results in podocyte charge alteration and glomerular injury.

The epithelial polyanion (podocalyxin) on the foot processes (pedicels) of podocytes plays a pivotal role in maintaining slit pore integrity and excluding proteins from the glomerular filtrate. Chromatographically purified recombinant sialidase from Vibrio cholerae, a corresponding heat-inactivated enzyme, truncated enzyme (missing the last 17 amino acids from the carboxyl terminus), and the sialidase from Salmonella typhimurium strain LT2 were inoculated intraperitoneally into mice, and the resultant renal alterations were documented by a variety of functional, morphologic, and histochemical techniques. Proteinuria and renal failure developed in a dose-dependent manner after a single inoculation of sialidase from Vibrio cholerae, but not with the corresponding heat-inactivated enzyme, truncated enzyme, or the sialidase from Salmonella typhimurium strain LT2. Biotinylated lectins of known sialyl linkage specificity demonstrated that Vibrio cholerae sialidase primarily removed alpha 2-->6-linked sialic acids from the glomerulus. Furthermore, the use of a poly-L-lysine cationic gold ultrastructural probe confirmed a transient loss of charge from the endothelium and epithelium of the glomerular filtration barrier. Loss of the epithelial polyanion was accompanied by the effacement of pedicels and the apparent formation of tight junctions between adjacent podocytes. The anionic charge returned to endothelial and epithelial sites within 2 days of sialidase inoculation, but the foot process loss remained. This animal model, in addition to providing an opportunity to study basic mechanisms of renal physiology, seems to mimic minimal change disease in children, diabetic nephropathy, and the renal effects of some bacterial infections.

Derived structure of the putative sialic acid transporter from Escherichia coli predicts a novel sugar permease domain.

Catabolism of sialic acids by Escherichia coli requires the genes nanA and nanT, which were previously mapped between argG and rpoN (E.R. Vimr and F.A. Troy, J. Bacteriol. 164:845-853, 1985). This organization is confirmed and extended by physical mapping techniques. An open reading frame beginning 135 bp from the nanA translational stop codon could code for a 53,547-Da hydrophobic polypeptide predicted to contain 14 transmembrane segments. Complementation analysis confirmed that nanT is required for sialic acid uptake when expressed in trans. NanT is homologous to a putative permease encoded by open reading frame 425, which maps between leuX and fecE in the E. coli chromosome. However, unlike this hypothetical permease or previously reported monosaccharide transporters, NanT contains a centrally located domain with two additional potential membrane-spanning segments plus one amphiphilic alpha-helix that may be important for the structure and function of sialic acid-permease.

Biosynthesis of the polysialic acid capsule in Escherichia coli K1.

The extracellular polysaccharides elaborated by most or all bacterial species function in cell-to-cell and cell-substratum adhesion, cell signaling, and avoidance or inhibition of noxious agents in animal hosts or free-living environments. Recent advances in our understanding of exopolysaccharide synthesis have been facilitated by comparative approaches in both plant and animal pathogens, as well as in microorganisms of industrial importance. One of the best understood of these systems is the kps locus for polysialic acid synthesis in Escherichia coli K1. The genes for sialic acid synthesis, activation, polymerization and translocation have been identified and assigned at least tentative functions in the synthetic and export pathways. Initial studies of kps thermoregulation suggest that genetic control mechanisms will be involved which are distinct from those already described for several other exopolysaccharides. Information about the common as well as unique features of polysialic acid biosynthesis will increase our knowledge of microbial cell surfaces which in turn may suggest novel targets for therapeutic or industrial interventions.

Microbial sialidases: does bigger always mean better?

Sialidases are a superfamily of N-acylneuraminate-releasing (sialic-acid-releasing) exoglycosidases found mainly in higher eukaryotes and in some, mostly pathogenic, viruses, bacteria and protozoans. The functions of sialidases are poorly understood and, until recently, their biochemical and evolutionary relationships were unclear. A comparative approach has demonstrated the remarkable similarities and differences between nonviral sialidases, and is providing clues about their functions.

Crystal structure of Vibrio cholerae neuraminidase reveals dual lectin-like domains in addition to the catalytic domain.

BACKGROUND: Vibrio cholerae neuraminidase is part of a mucinase complex which may function in pathogenesis by degrading the mucin layer of the gastrointestinal tract. The neuraminidase, which has been the target of extensive inhibitor studies, plays a subtle role in the pathology of the bacterium, by processing higher order gangliosides to GM1, the receptor for cholera toxin. RESULTS: We report here the X-ray crystal structure of V. cholerae neuraminidase at 2.3 A resolution. The 83 kDa enzyme folds into three distinct domains. The central catalytic domain has the canonical neuraminidase beta-propeller fold, and is flanked by two domains which possess identical legume lectin-like topologies but without the usual metal-binding loops. The active site has many features in common with other viral and bacterial neuraminidases but, uniquely, has an essential Ca2+ ion which plays a crucial structural role. CONCLUSIONS: The environment of the small intestine requires V. cholerae to secrete several adhesins, and it is known that its neuraminidase can bind to cell surfaces, and remain active. The unexpected lectin-like domains possibly mediate this attachment. These bacterial lectin folds represent additional members of a growing lectin superfamily.

Cloning, sequencing, expression, and complementation analysis of the Escherichia coli K1 kps region 1 gene, kpsE, and identification of an upstream open reading frame encoding a protein with homology to GutQ.

The kps locus for polysialic acid capsule expression in Escherichia coli K1 is composed of a central group of biosynthetic neu genes, designated region 2, flanked on either side by region 1 or region 3 kps genes with poorly defined functions. Chromosomal mutagenesis with MudJ and subsequent complementation analysis, maxicell and in vitro protein expression studies, and nucleotide sequencing identified the region 1 gene, kpsE, which encodes a 39-kDa polypeptide. Polarity of the kpsE::lacZ mutation suggests an operonic structure for region 1. KpsE is homologous to putative polysaccharide-translocation components previously identified in Haemophilus influenzae type b and Neisseria meningitidis group B. An open reading frame upstream of kpsE encodes a 35-kDa polypeptide with homology to GutQ, a putative ATP-binding protein of unknown function encoded by gutQ of the glucitol utilization operon. Whether expression of the gutQ homolog as the potential first gene of region 1 is required for polysialic acid synthesis or localization is presently unknown.

Crystal structure of a bacterial sialidase (from Salmonella typhimurium LT2) shows the same fold as an influenza virus neuraminidase.

Sialidases (EC 3.2.1.18 or neuraminidases) remove sialic acid from sialoglycoconjugates, are widely distributed in nature, and have been implicated in the pathogenesis of many diseases. The three-dimensional structure of influenza virus sialidase is known, and we now report the three-dimensional structure of a bacterial sialidase, from Salmonella typhimurium LT2, at 2.0-A resolution and the structure of its complex with the inhibitor 2-deoxy-2,3-dehydro-N-acetylneuraminic acid at 2.2-A resolution. The viral enzyme is a tetramer; the bacterial enzyme, a monomer. Although the monomers are of similar size (approximately 380 residues), the sequence similarity is low (approximately 15%). The viral enzyme contains at least eight disulfide bridges, conserved in all strains, and binds Ca2+, which enhances activity; the bacterial enzyme contains one disulfide and does not bind Ca2+. Comparison of the two structures shows a remarkable similarity both in the general fold and in the spatial arrangement of the catalytic residues. However, an rms fit of 3.1 A between 264 C alpha atoms of the S. typhimurium enzyme and those from an influenza A virus reflects some major differences in the fold. In common with the viral enzyme, the bacterial enzyme active site consists of an arginine triad, a hydrophobic pocket, and a key tyrosine and glutamic acid, but differences in the interactions with the O4 and glycerol groups of the inhibitor reflect differing kinetics and substrate preferences of the two enzymes. The repeating "Asp-box" motifs observed among the nonviral sialidase sequences occur at topologically equivalent positions on the outside of the structure. Implications of the structure for the catalytic mechanism, evolution, and secretion of the enzyme are discussed.

The sialidase superfamily and its spread by horizontal gene transfer.

Sialidases (neuraminidases, EC 3.2.1.18) belong to a class of glycosyl hydrolases that release terminal N-acylneuraminate (sialic acid) residues from glycoproteins, glycolipids, and polysaccharides. These enzymes are common in animals of the deuterostomate lineage (Echinodermata through Mammalia) and also in diverse microorganisms that mostly exist as animal commensals or pathogens. Sialidases, and their sialyl substrates, appear to be absent from plants and most other metazoans. Even among bacteria, sialidase is found irregularly so that related species or even strains of one species differ in this property. This unusual phylogenetic distribution makes sialidases interesting for evolutionary studies. The biochemical diversity among bacterial sialidases does not indicate close relationships. However, at the molecular level, homologies are detectable, supporting the hypothesis of a common sialidase origin and thus of a sialidase superfamily. Some findings indicate that sialidase genes were recently transferred via phages among bacteria. The proposal of a sialidase origin in higher animals is suggested by the presence of apparently homologous enzymes in this kingdom, supporting the idea that some microbes may have acquired the genetic information during association with their animal hosts.