Identification of a Distinct, Cryptic Heparosan Synthase from Pasteurella multocida Types A, D, and F

Abstract

The extracellular polysaccharide capsules of Pasteurella multocida types A, D, and F are composed of hyaluronan, N-acetylheparosan (heparosan or unsulfated, unepimerized heparin), and unsulfated chondroitin, respectively. Previously, a type D heparosan synthase, a glycosyltransferase that forms the repeating disaccharide heparosan backbone, was identified. Here, a ∼73% identical gene product that is encoded outside of the capsule biosynthesis locus was also shown to be a functional heparosan synthase. Unlike PmHS1, the PmHS2 enzyme was not stimulated greatly by the addition of an exogenous polymer acceptor and yielded smaller- molecular-weight-product size distributions. Virtually identical hssB genes are found in most type A, D, and F isolates. The occurrence of multiple polysaccharide synthases in a single strain invokes the potential for capsular variation.

Glycosaminoglycans (GAGs) are long linear polysaccharides consisting of disaccharide repeats that contain an amino sugar. Three prevalent and essential GAGs found in higher animals are heparin or heparan (β4GlcUA-α4GlcNAc)_n, chondroitin (β4GlcUA-β3GalNAc)_n, and hyaluronan (β4GlcUA-β3GlcNAc)_n (where GlcUA is glucuronic acid, GlcNAc is N-acetylglucosamine, and GalNAc is N-acetylgalactosamine). GAGs are also produced by certain pathogenic bacteria in the form of extracellular polysaccharide coatings, called capsules, which serve as virulence factors (3). The capsule is classically thought to assist in the evasion of host defenses such as phagocytosis and complement, but there is some evidence that the pathogen also commandeers host GAG recognition or signaling systems. As the microbial polysaccharide is identical or very similar to the host GAG, the antibody response is either very limited or nonexistent.

Many Pasteurella multocida isolates produce GAGs (12). P. multocida Carter type A (fowl cholera and pasteurellosis pathogen) makes a hyaluronan (HA) capsule (11). The capsules of type D (swine atrophic rhinitis pathogen) and type F (minor fowl cholera pathogen) are composed of unsulfated N-acetylheparosan (heparosan) or unsulfated chondroitin, respectively (5). Dual-action synthases were identified for type A (PmHAS), D (PmHS), and F (PmCS) strains that polymerize the GAG chain disaccharide repeat by transferring both GlcUA and hexosamine (either GlcNAc or GalNAc, depending on the GAG) monosaccharides derived from uridine diphospho (UDP) precursors (3, 6-8). A capsule biosynthesis locus that encodes the synthase, the sugar nucleotide precursor-forming enzymes, and the polymer transport proteins (15) was found in all three capsular types.

hssA and hssB nomenclature.

A deduced gene, which is called pglA (GenBank accession number AAK02498) and which encodes 651 amino acids, similar to PmHS (∼73% identical in the major overlapping region), was uncovered by the University of Minnesota in their type A P. multocida genome project (10). However, the pglA gene is not located in the putative capsule locus. The function of pglA was not established, but this name has been used to describe another product involved in bacterial protein glycosylation from Campylobacter jejuni. The evidence presented here indicates that the Pasteurella enzyme designated PglA is actually a functional heparosan synthase; thus, PmHS2 is a more appropriate nomenclature. Thus, the original PmHS enzyme (8) should now be called PmHS1. To conform to standard nomenclature, the genes encoding PmHS1 and PmHS2 are called hssA and hssB, respectively (heparosan synthase).

Expression and characterization of recombinant heparosan synthases.

Recombinant PmHS1 and PmHS2 were prepared in an Escherichia coli host that does not normally produce GAG polymers. The sources of the genes were wild-type encapsulated P. multocida isolates from the U.S. Department of Agriculture (Ames, Iowa) or the American Type Culture Collection (Rockville, Md.). The 651-amino-acid PmHS2 open reading frame (ORF) that was predicted from a deposited type A genome (10) was amplified from a type A (P-1059; ATCC 15742), type D (P-3881), or type F (P-4218) genomic DNA template by 18 cycles of PCR (94°C, 30 s; 52°C, 30 s; 72°C, 2.5 min) with Taq DNA polymerase (Fisher, Fair Lawn, N.J.), sense primer ATGAAGAGAAAAAAAGAGATGACTC (the deduced start codon is underlined), and antisense primer ATCATTATAAAAAATAAAAAGGTAAACAGG encoding a stop codon.

The amplicons were cloned using the pETBlue-1 Acceptor system (Novagen, Madison, Wis.) according to the manufacturer's instructions. The ORFs in the pETBlue-1 vector were sequenced in their entirety in both directions by automated sequencing (Oklahoma State University Sequencing Facility). The actual DNA sequences corresponding to the expression construct oligonucleotide primer sites in the native P. multocida chromosome were obtained by sequencing total PCR products derived from genomic DNA and pairs of primers that flanked the 5′ or the 3′ termini of the ORF.

Membranes from E. coli Tuner cells with plasmids containing the PmHS2 ORF or the PmHS1 ORF were prepared (8). The recombinant PmHS1 (type D; P-4058) and PmHS2 (type D; P-3881) polypeptides were analyzed by using standard 8% polyacrylamide-sodium dodecyl sulfate gels and Western blotting utilizing a monospecific antibody directed against a synthetic peptide (acetyl-KGDIIFFQDSDDVCHHERIER-amide) corresponding to residues 173 to 193 of PmHS1 or residues 207 to 227 of PmHS2 using a colorimetric detection method (9). As predicted from the deduced ORF sequence, the larger PmHS2 polypeptide migrates slower than PmHS1 (Fig. 1).

FIG. 1.

Western blot analysis of PmHS1 and PmHS2. Membrane preparations from recombinant E. coli cells with plasmids containing hssA, hssB, or no insert (vec) were analyzed with an antipeptide antibody. The relevant region spanning the 95.5- to 55-kDa standards is shown. As predicted from the deduced sequence, the larger PmHS2 polypeptide migrates slower than PmHS1.

Our attempts to visualize the native PmHS2 protein from several P. multocida isolates in various media (including defined or complex media supplemented with chicken tissue or sera) by Western blotting were unsuccessful. Native PmHS1 from type D strains, however, was easily detected in parallel tests; the protein migrated identically to recombinant PmHS1 (data not shown). Preliminary reverse transcriptase PCR results suggest that the hssB transcript is present at ∼10-fold-lower levels than the hssA transcript in log-phase P. multocida cells (data not shown); thus, the hssB promoter appears functional. The 5′ regions (75 bp upstream of start codon) of the hssA and hssB ORFs are not very similar. The predicted hssA message contains a typical gram-negative ribosomal binding site sequence (AAGGAGATATAGATTATG; the start codon is in bold), but the site is absent in the putative hssB message (AAACTCATTTATTGAAATG), which may explain the inability to observe the PmHS2 protein.

The incorporation of radiolabeled monosaccharides from UDP-[¹⁴C]GlcUA and/or UDP-[³H]GlcNAc precursors (Perkin-Elmer NEN, Boston, Mass.) was used to monitor recombinant heparosan synthase activity. Samples were usually assayed in a buffer containing 50 mM Tris (pH 7.2), 10 mM MgCl₂, 10 mM MnCl₂, 0.5 mM UDP-GlcUA, and 0.5 mM UDP-GlcNAc at 30°C and analyzed by either descending paper or gel filtration chromatography (8). The recombinant Pasteurella synthases will initiate sugar chains in vitro, but the addition of an appropriate exogenous sugar acceptor bypasses the slow initiation phase of polymerization; the fast elongation phase predominates the reaction, thereby stimulating the sugar incorporation rate (5, 8, 9). Depending on the experiment, a ∼55-kDa type D acceptor polymer processed by extended ultrasonication of a capsular polysaccharide preparation (8) was also added to the reaction mixture. The data from the recombinant construct containing the hssB gene from the type D strain P-3881 are presented, but the results were similar to those from experiments with constructs derived from the type A strain P-1059.

The maximal activity of PmHS2 was observed in reactions that contained the Mn²⁺ ion, but Mg²⁺ and Co²⁺ also supported incorporation (∼25 to 30% of the level with Mn²⁺). On the other hand, the levels of PmHS1 activity were very similar in the presence of Mn²⁺ or Mg²⁺ (8). Membrane extracts derived from E. coli Tuner cells containing the plasmid encoding PmHS2, but not samples from cells with the vector alone, synthesized polymer in vitro only when supplied with both UDP-GlcUA and UDP-GlcNAc simultaneously (Table 1). The identity of the PmHS2-derived polymer as heparosan was verified by its sensitivity to Flavobacterium heparin lyase III (99.5% of the polymer was destroyed) and its resistance to the action of Streptomyces HA lyase. The addition of the heparosan polymer acceptor increased PmHS2-catalyzed sugar incorporation by only about 2.5-fold (data not shown). In contrast, PmHS1 was stimulated at least 7- to 25-fold in comparison to parallel reactions without acceptor (8), similar to observations of PmHAS (4) and PmCS (7).

TABLE 1.

Sugar nucleotide transfer specificity of recombinant PmHS2

Substrate and second sugar	Substrate incorporation (pmol)
	PmHS2a	Vector
[3H]GlcNAc
None	300	26
UDP-GlcUA	4,100	23
UDP-GalUA	230	25
UDP-Glc	240	23
[3H]GlcUA
None	13	4
UDP-GlcNAc	3,100	4
UDP-GalNAc	12	5
UDP-Glc	50	4

^aEnzyme preparations from a type D PmHS2 construct or from vector alone (0.36 mg of total protein) were incubated with one radiolabeled authentic sugar nucleotide substrate (either UDP-[³H]GlcNAc or UDP-[³H]GlcUA; 0.4 μCi/50-μl reaction mixture) in the presence of the indicated second unlabeled sugar nucleotide. After 120 min, the incorporation into polymer was assessed by paper chromatography. Results from a representative experiment with averaged duplicate assay points are shown.

Gel filtration chromatography indicated that the recombinant PmHS2 enzyme made shorter chains (∼100 kDa based on hyaluronan standards or ∼500 monosaccharides) (Fig. 2A) than recombinant PmHS1 (∼500 to 1,000 kDa) (Fig. 2B) under identical conditions without acceptor in vitro. If acceptor polymer was supplied to PmHS1, then high levels of sugar incorporation were observed as short chains added to the acceptor (resulting in a major peak at ∼12.5 min). PmHS2 also catalyzed the extension of acceptor chains, as seen by the increased amplitude of the peak at ∼12.5 min, but the effect was not as dramatic as that with PmHS1.

FIG. 2.

Gel filtration analysis of polysaccharide products of recombinant PmHS1 and PmHS2. The crude membranes containing PmHS2 (panel A) or PmHS1 (panel B) (0.36 mg of total protein) were incubated with UDP-[¹⁴C]GlcUA (0.15 μCi/reaction) and UDP-GlcNAc in parallel reactions (75 μl) either in the presence (thick line) or the absence (thin line) of type D acceptor polymer (0.4 μg of uronic acid; 12.8-min elution time). After 60 min, the samples were analyzed on the PolySep 4000 column; a representative profile is depicted. The elution times of standards are marked with arrows (from left to right: void volume, 112-kDa hyaluronan, and 55-kDa heparosan) (totally included volume, 16.7 min). The background signal for experiments using vector control membranes (data not shown) was <6 disintegrations per second (DPS) throughout the relevant polymer region. In the absence of acceptor, the PmHS2 product has a lower-molecular-weight size distribution than the PmHS1 product. In the presence of exogenous acceptor molecules, as seen by the comparison of the new peaks in the region of ∼12.5 min, the incorporation by PmHS1 is stimulated more than that of PmHS2.

Overall, PmHS1 and PmHS2 are both heparosan synthases with similar amino acid sequences, but their metal cofactor specificity, acceptor usage, and polymer product size distributions are distinct.

Occurrence of hssB in many P. multocida isolates.

PCR analysis was employed to assess whether the hssB gene was widespread or rare in the P. multocida population represented by a variety of independent strains isolated from around the world (including the United States, United Kingdom, Belgium, and Malaysia) from various animal hosts (both domestic and wild animals) over a long period of time (1972 to 1998). Genomic DNA templates were subjected to PCR with three distinct primer pairs specific for various regions of the hssA gene. The presence of an amplicon of the appropriate size for each primer set was assessed by agarose gel electrophoresis and ethidium staining. All type A strains tested (3 of 3) as well as 67% of type D (6 of 9) and 55% of type F (6 of 11) strains were positive for the hssB-specific amplicons with all three primer sets. More type D and F isolates (78 and 82%, respectively) were positive with two primer sets.

We cloned a sampling of PmHS2 ORFs into an E. coli-based expression system and found that all of the tested enzymes were functional. However, at least one of the isolates (P-1059) had a frameshift mutation in the region of the second codon of the PmHS2 ORF; thus, we corrected the nonsense mutation during PCR and regained the function of the enzyme.

The hssB gene is not in the capsule locus.

In many gram-negative microbes, the genes essential for production of the extracellular capsule are usually grouped together on the chromosome (14, 16). Several capsule loci of different P. multocida types which resemble the prototypical E. coli group II capsule loci have been reported (15). Typically, the polysaccharide biosynthesis genes (e.g., glycosyltransferases and sugar nucleotide precursor production enzymes) are flanked by regions carrying export-related genes required to translocate the nascent polysaccharide chain across the two lipid membranes (e.g., ATP-binding cassette transporters and periplasmic proteins). The genes encoding PmHAS, PmHS1, or PmCS synthases of type A, D, or F strains, respectively, are adjacent to a gene encoding UDP-glucose dehydrogenase (enzyme that produces UDP-GlcUA) in the capsule locus. Genes associated with capsular transport, including phyA and hexD, flank this region (15). However, the hssB gene in the type A genome strain (10) resides between two metabolic genes not known to be directly involved in capsule polymer synthesis, the alanine racemase gene (alr) and the glucose-6-phosphate isomerase gene (pgi).

The hssB gene of several independent strains was also flanked by the same two genes (alr and pgi) as the genome strain, as assessed by sequencing PCR amplicons derived from genomic DNA. Southern blotting was employed to assess whether the hssB gene in other strains was also located outside of the capsule locus. Duplicate nitrocellulose replicas screened by hybridization with specific probes for either the hssA or the hssB gene indicated that the two genes were not linked in the tested type D strains (data not shown).

Potential role of hssB in modulation of capsular polysaccharide composition.

It was initially quite puzzling that the type A or type F strain would have a heparosan synthase as well as the known hyaluronan synthase or chondroitin synthase, respectively. An interesting hypothesis is that type A or F organisms utilize either the capsule locus synthase gene (encoding PmHAS or PmCS, respectively) or the hssB gene, depending on environmental conditions or stage of infection. Employing different capsular polymers may serve to enhance infection via changing the microbe's interaction with a subset of the host GAG-binding proteins or extracellular matrix. In another scenario, if hssB was expressed simultaneously with the genes encoding PmHAS or PmCS under some conditions, then the result would be a capsule composition with a mixture of two different types of GAG chains (e.g., HA and heparosan chains coexisting in the capsule of a type A microbe).

With respect to capsular polysaccharide variation in other bacterial species, two putative examples appear to change only the abundance of capsule polymer without switching its chemical composition (2, 13). On the other hand, phase variation changes in lipo-oligosaccharide composition have been reported (1). P. multocida produces HA, heparosan, or chondroitin depending on the Carter capsule type, but now the possibility of switching or altering molecular camouflage within one clone needs to be explored.

Nucleotide sequence accession numbers.

The sequences of an hssB clone from a type A and a type D strain were deposited in GenBank (AY292199 and AY292200, respectively) and are ∼99% identical to the deposited genome sequence.