OUP user menu

Neuraminidase (sialidase) activity of Haemophilus parasuis

Carol A Lichtensteiger, Eric R Vimr
DOI: http://dx.doi.org/10.1111/j.1574-6968.1997.tb10438.x 269-274 First published online: 1 July 1997


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.

  • Haemophilus parasuis
  • Haemophilus
  • Neuraminidase
  • Sialidase
  • Pasteurellaceae
  • Pasteurella
  • Actinobacillus

1 Introduction

Pasteurellaceae are obligate parasites of the mucus membranes of humans and animals and usually coexist as commensals in their hosts. Occasionally, these bacteria breach naso-pharyngeal barriers to cause potentially life-threatening diseases. To propagate in their hosts, bacteria synthesize a variety of enzymes and other proteins, many of which function in metabolic processes that are not necessarily unique to the host-parasite interaction. One approach to analyze molecular requirements for pathogenesis is to investigate bacterial enzymes whose substrates are unique components of host cells. Sialic acids are such substrates [13].

Sialic acids are a family of nine carbon sugar acids found in all animals of the complex metazoan lineage that includes humans, but are not found in plants, lower metazoans, or most microorganisms. Therefore, the ability of parasitic bacteria to use sialic acids as carbon, nitrogen, or energy sources may reflect an innate mechanism for growth and persistence in the host. At least three polypeptides may be involved in sialic acid catabolism: (i) neuraminidase (nanH) cleaves sialic acid from polysaccharides, glycoproteins, and glycolipids; (ii) permease (nanT) transports sialic acid through the cell membrane; and (iii) aldolase (nanA) splits sialic acid into pyruvic acid and the aldose sugar N-acetylmannosamine [2, 4, 5]. Neuraminidase may contribute to bacterial virulence by scavenging carbohydrates from host cells for nutrition via subsequent action of permease and aldolase. In addition to a nutritional role, neuraminidase removal of sialic acid residues from host glycoconjugates could (i) unmask receptors needed to colonize or invade host cells or (ii) interfere with the host defense system by decreasing viscosity of mucin or altering the functions of immune and inflammatory mediators and cells [1, 2, 6].

Neuraminidase activity has been investigated in a few species of Pasteurella, Haemophilus, and Actinobacillus, which represent the three genera of the Pasteurellaceae family. The enzyme has been confirmed in P. multocida and P. haemolytica[79] where it may contribute to the pathogenesis of pneumonia[10]. An early report based on an indirect assay concluded that some H. influenzae and H. parainfluenzae isolates produce neuraminidase[11]. H. aphrophilus and A. actinomycetemcomitans are reported to lack neuraminidase[12].

We screened five pathogenic Pasteurellaceae species for neuraminidase: H. influenzae, H. parasuis, H. somnus, H. paragallinarum, and A. pleuropneumoniae. Neuraminidase was found only in H. parasuis, an upper respiratory inhabitant that causes polyserositis, synovitis, meningitis, and pneumonia in young pigs[13].

2 Materials and methods

2.1 Screening Pasteurellaceae for neuraminidase

Bacteria were reference isolates from the American Type Culture Collection (ATCC; H. parasuis, 19417; H. influenzae, 10211; H. somnus, 43626; H. paragallinarum, 29976; and A. pleuropneumonia, 27082) or were diagnostic isolates recovered from clinical specimens submitted to Illinois state diagnostic laboratories. P. multocida is the swine isolate M33 we described in a previous study. Bacteria were grown on chocolate agar or in haemophilus test medium (Remmel, Lenexa, KS) at 37°C, usually with 5% CO2 enrichment. Nurse streaks of Staphylococcus sp. were used as needed.

Bacterial isolates, at 24–72 h growth, were screened for neuraminidase activity in broth-spot or colony-spot fluorescent assays [5, 12, 14]. The fluorescent substrate was 2′-(4-methylumbelliferyl)-α-d-N-acetylneuraminic acid (4MU-Neu5Ac) reconstituted at 2 mg ml−1 in 100 mM sodium acetate, 308 mM NaCl, and 8 mM CaCl2, pH 5.5. Specificity for neuraminidase was demonstrated by inhibiting hydrolysis with 2,3-dehydro-2-deoxy-N-acetylneuraminic acid (NeuAc2en). P. multocida, a species that produces neuraminidase, was the positive control for both assays [7, 8].

2.2 Characterizing H. parasuis neuraminidase

Fluorescence in 4MU-Neu5Ac reactions was measured with a fluorometer and rates (fluorescence units [fu] min−1) were calculated [5, 14, 15] to determine (i) enzyme specificity and cation requirements (preincubation of bacteria with NeuAc2en, EDTA, or EGTA for 10 min at 37°C) or (ii) cell association of activity of H. parasuis neuraminidase. For these studies, H. parasuis and P. multocida were collected and washed from overnight broth cultures and recombinant V. cholera neuraminidase[16] was used for control reactions. For the cell association study, supernatant and cell pellets of H. parasuis and P. multocida were collected sequentially from low speed centrifugation (13 000×g, 3 min) and ultracentrifugation (280 000×g, 20 min) and assayed for neuraminidase.

Sialic acid released by H. parasuis from five substrates was measured by the periodate-thiobarbituric acid (TBA) assay with dimethyl sulfoxide extraction[17]. Pelleted, washed H. parasuis cells (stored at −20°C if needed) were incubated with 2.5 mg ml−1 fetuin, α1-acid glycoprotein (bovine), N-acetylneuramin-lactose (bovine), submaxillary gland mucin (bovine), or colominic acid (poly-α2,8-sialic acid) in a 55.5 mM sodium acetate buffer at 37°C; free sialic acid was measured with the TBA assay at two time points between 3 and 15 min. To determine the optimal pH, reactions against fetuin were done in the acetate buffer adjusted from pH 2 to pH 9. The remaining substrates were tested at the pH optimum of 4.5; mucin and colominic acid were also tested at pH 6.0. Substrates were tested in duplicate or triplicate samples in most runs. Aliquots of substrate and bacteria alone were incubated in parallel to determine background hydrolysis. To be considered positive for activity, the rate of sialic acid release from a given substrate had to increase with time over background rates, which were negligible for all substrates except mucin. Specificity of hydrolysis was confirmed by adding the competitive inhibitor NeuAc2en at 5 mM. Samples of pure sialic acid (N-acetylneuraminic acid) were run in parallel at 0–100 μg ml−1; R2 of the standard curves was greater than 0.99.

3 Results and discussion

3.1 Neuraminidase activity of Pasteurellaceae

Forty-three Pasteurellaceae isolates were screened for neuraminidase activity with the fluorescent broth- (Fig. 1) or colony-spot assays. The background control reactions and the inhibition of the fluorescent response with NeuAc2en indicate that the assays specifically detected neuraminidase. Neuraminidase activity was readily detectable in P. multocida, the positive control species [7, 8], and 30 of 32 isolates of H. parasuis; neuraminidase activity was not seen in the six H. somnus isolates, the three A. pleuropneumoniae isolates, the H. paragallinarum ATCC isolate, or the H. influenzae ATCC isolate.

Figure 1

Representative broth-spot assay for neuraminidase activity of six Pasteurellaceae isolates against the fluorescent substrate, 4MU-Neu5Ac. Images were collected electronically promptly after adding bacteria to the reaction. Row 1, bacteria and substrate; row 2, bacteria alone; row 3, bacteria, substrate, and neuraminidase inhibitor (11 mM NeuAc2en); row 4, substrate alone (spotted in first column only).

Among the above hemophili, only H. parasuis exhibited neuraminidase activity, suggesting it may be unique in the genus because the other three species we tested and H. aphrophilus[12] did not have activity. The lack of neuraminidase activity in H. influenzae differs from the result of a previous study that used an indirect assay of neuraminidase[11]. The H. influenzae genome (isolate Rd) has recently been sequenced (GenBank, L42023) and we found no evidence of a neuraminidase gene by searching for homology (relaxed stringency) with other bacterial neuraminidase genes (nanH). Screening of more H. influenzae isolates may indicate that the ATCC and Rd isolates are atypical, that H. influenzae may be similar to Salmonella spp. in that nanH is irregularly distributed in the genus[3], or that the earlier results are incorrect. Our A. pleuropneumoniae isolates lacked neuraminidase activity, a result that matches the lack of neuraminidase in another actinobacillus species, A. actinomycetemcomitans[12]. Two of the 32 H. parasuis isolates had no detectable neuraminidase activity, which is similar to the variability in other biochemical reactions in bacterial species and to neuraminidase activity in occasional isolates of P. multocida and P. haemolytica[8, 9]. Because these two neuraminidase negative H. parasuis isolates are diagnostic isolates from diseased pigs, either neuraminidase is not a required virulence factor or these two isolates were from pigs ill due to some other agent.

The taxonomy of the Pasteurellaceae family is not perfect [18, 19]. The presence of neuraminidase activity in H. parasuis and P. multocida but not in H. somnus, H. influenzae, and A. pleuropneumoniae correlates with a dendrogram based on homology of 16S RNA in which H. parasuis and P. multocida cluster closer to each other than any of the other three species[18]. The possibility should be considered that H. parasuis is a Pasteurella sp. with a growth requirement for factor V, one of the original classifying characteristics of hemophili.

3.2 Characterization of H. parasuis neuraminidase activity

Reaction conditions of the H. parasuis neuraminidase were investigated. Neither EDTA or EGTA inhibited neuraminidase activity of H. parasuis indicating that cations are not required (Fig. 2). In parallel reactions, the cation requiring Vibrio cholerae neuraminidase[15] was inhibited whereas the apparently cation-independent P. multocida neuraminidase[8] was not affected by the presence of either cation chelator. As expected, the specific inhibitor NeuAc2en inhibited the activity of all three bacterial preparations. The pH required for optimum hydrolysis of fetuin by H. parasuis was 4.5; about 50% of the activity was still detectable at pH 4 and pH 6 (Fig. 3). This acid pH optimum and the shape of the curve are typical of bacterial neuraminidases[1].

Figure 2

Neuraminidase activity in the presence of potential inhibitors, cation chelators (EDTA and EGTA) and NeuAc2en. The bacterial source of neuraminidase activity is indicated on the X axis. In all reactions, the substrate was 4MU-Neu5Ac in 8 mM CaCl2, pH 5.5, alone or with the indicated inhibitor. The graph is an example of duplicate trials with similar results.

Figure 3

The effect of pH on H. parasuis neuraminidase activity against fetuin. The rate of free sialic acid released from fetuin by H. parasuis at pH 2–9 was measured with the TBA assay. The reactions were run in duplicate after two exploratory runs to determine the pH range. Mean±S.E.M.

H. parasuis neuraminidase released sialic acid from fetuin, α-acid glycoprotein, and N-acetylneuramin-lactose but not from mucin or colominic acid. This substrate preference differs from that of P. multocida which recognizes all five substrates, albeit activity against mucin is weaker and four of 16 serotypes do not recognize mucin[7].

Neuraminidase activity of H. parasuis (and of P. multocida run in parallel) against 4MU-Neu5Ac was found only in the cell pellet (cell associated) as long as the supernatant was carefully separated from the low speed centrifugation pellet (cells). If this precaution was ignored, up to 10% of the total activity was in the supernatant; after ultracentrifugation of the low centrifugation supernatant, the pellet (cell membrane fragments) contained activity, but not the ultrasupernatant. The neuraminidase of P. multocida is cell associated[8], although some investigators use the supernatant of P. multocida cultures as a source of enzyme[7], which may reflect sloughing of membrane blebs as the source of enzyme activity. The cell associated neuraminidase of H. parasuis and P. multocida contrasts with that of many neuraminidase-producing bacteria which synthesize and often secrete soluble enzymes [13].

3.3 Metabolism of sialic acid by H. parasuis

After a lag of about 10 h, exogenous sialic acid (N-acetylneuraminic acid) gradually disappeared from the culture supernatant of H. parasuis (Fig. 4A). The two control reactions demonstrate that the loss was related to H. parasuis: (i) the background reading in the TBA assay was low in supernatant of cultures without exogenous sialic acid and (ii) the sialic acid was stable in medium incubated without H. parasuis. H. parasuis cultured with exogenous sialic acid had a delayed growth advantage (Fig. 4B). The delay before the loss of the sialic acid and the growth advantage suggest that enzymes for metabolism were produced after the exhaustion of some limiting nutrient(s), analogous to sialic acid inducing aldolase and permease production in Escherichia coli after glucose is depleted[4]. However, the catabolism of sialic acid by H. parasuis was not likely regulated by glucose consumption because it was not a specific component of the medium.

Figure 4

H. parasuis cultures with 100 μg ml−1 exogenous sialic acid. A: Persistence of sialic acid in culture supernatant, measured by the TBA assay. B: Effect on H. parasuis growth, measured by absorbance at 600 nm. The cultures were run in triplicate and the sialic acid alone samples were incubated in duplicate in this experiment after three exploratory runs in duplicate to identify the time range. Means±S.E.M.

Sialic acid induced growth advantage of H. parasuis may also occur in host pigs during infection and disease. Further work with H. parasuis neuraminidase is needed to understand its regulation and its role in vivo.


We thank Jennifer Herring for laboratory assistance and we thank Cynthia Brown, Illinois Animal Disease Laboratory, Galesburg, IL, and Sheila Davis, Laboratories of Veterinary Diagnostic Medicine, Urbana, IL for the diagnostic isolates. This project was partially funded by the Illinois Pork Producers Association.


View Abstract