Development of a Rapid Multiplex PCR Assay To Genotype Pasteurella multocida Strains by Use of the Lipopolysaccharide Outer Core Biosynthesis Locus
ABSTRACT
Pasteurella multocida is a Gram-negative bacterial pathogen that is the causative agent of a wide range of diseases in many animal species, including humans. A widely used method for differentiation of P. multocida strains involves the Heddleston serotyping scheme. This scheme was developed in the early 1970s and classifies P. multocida strains into 16 somatic or lipopolysaccharide (LPS) serovars using an agar gel diffusion precipitin test. However, this gel diffusion assay is problematic, with difficulties reported in accuracy, reproducibility, and the sourcing of quality serovar-specific antisera. Using our knowledge of the genetics of LPS biosynthesis in P. multocida, we have developed a multiplex PCR (mPCR) that is able to differentiate strains based on the genetic organization of the LPS outer core biosynthesis loci. The accuracy of the LPS-mPCR was compared with classical Heddleston serotyping using LPS compositional data as the “gold standard.” The LPS-mPCR correctly typed 57 of 58 isolates; Heddleston serotyping was able to correctly and unambiguously type only 20 of the 58 isolates. We conclude that our LPS-mPCR is a highly accurate LPS genotyping method that should replace the Heddleston serotyping scheme for the classification of P. multocida strains.
INTRODUCTION
Pasteurella multocida is the primary causative agent of a wide range of economically important diseases, including hemorrhagic septicemia in ungulates, atrophic rhinitis in pigs, fowl cholera in birds, snuffles in rabbits, and enzootic pneumonia and shipping fever in cattle, sheep, and pigs (1). P. multocida also causes opportunistic infections in humans, often following cat or dog bites, and plays a contributory role, together with other pathogens, in a range of lower respiratory tract infections and sporadic septicemias in ungulates (1).
P. multocida strains have classically been differentiated using serological techniques. Strains can be classified into five capsular serogroups (A, B, D, E, and F) using an indirect hemagglutination test (2) and into 16 somatic or lipopolysaccharide (LPS) serovars (serotypes) using the Heddleston gel diffusion precipitin test (3). Both of these schemes have been widely used. Isolates are commonly assigned a combined designation, such as A:1 (capsular serogroup A and LPS serovar 1) or B:2 (capsular serogroup B and LPS serovar 2).
P. multocida LPS is an immunodominant antigen critical for homologous protection stimulated by bacterin (killed-cell) vaccines (4). Furthermore, in the P. multocida strain VP161, a full-length LPS molecule is essential for the ability to cause acute disease (5, 6). Heddleston serotyping is currently the only method used to differentiate P. multocida strains on the basis of LPS type. However, the accuracy of Heddleston serotyping has never been objectively tested, as the precise LPS structures produced by different strains have not been known. Indeed, there have been many informal as well as formal reports that the Heddleston system fails to type many isolates and lacks accuracy and reproducibility (7, 8). Furthermore, Heddleston serotyping is time-consuming and requires access to good-quality, serovar-specific antisera.
We have recently carried out a comprehensive analysis of the LPS structures expressed by the 16 Heddleston type strains and identified the genes required for LPS assembly in each strain (9–16). These combined analyses showed that the LPS produced by all strains consisted of a highly conserved inner core and a variable outer core and revealed that each of the 16 Heddleston type strains expresses structurally distinct LPS. Importantly, these analyses also showed that only eight unique LPS outer core biosynthesis loci are found in the 16 Heddleston type strains (Fig. 1). We have designated these genetic loci L1 through to L8. The type strains of Heddleston serovars 1, 2, 3, 5, 6, 8, 9, 12, and 16 express full-length or “parent” LPS structures, and the type strains of Heddleston serovars 4, 7, 10, 11, 13, 14, and 15 express truncated LPS—the result of mutations within the LPS outer core biosynthesis loci.
FIG 1
The partial differentiation of P. multocida strains on the basis of LPS biosynthesis genes has been reported previously (17). Using PCR-restriction fragment length polymorphism (RFLP) analysis, P. multocida strains were grouped into 5 PCR-RFLP types. However, only 11 of the 16 Heddleston serovars were included in the study. Here we report the development and testing of a multiplex PCR (mPCR) using the full set of Heddleston serovar type strains, which can accurately differentiate P. multocida strains into one of the eight distinct LPS genotypes. By comparing the results of Heddleston serotyping and the LPS-mPCR to the LPS structures predicted from LPS compositional analysis by mass spectrometry (MS), we have determined the accuracy of this mPCR and Heddleston serotyping for predicting LPS type. The LPS-mPCR gave a result that was indicative of LPS genotype >98% of the time, and we propose that the LPS-mPCR assay should be used to differentiate strains into their appropriate LPS genotype and, together with the cap mPCR, form a new molecularly based typing system for accurate strain differentiation of P. multocida.
MATERIALS AND METHODS
Strains used.
All P. multocida strains were grown at 37°C in heart infusion (HI) (Oxoid, Basingstoke, United Kingdom) liquid broth with shaking or on solid HI medium containing 1.5% agar. The P. multocida strains used in the study are described in Table 1. All field isolates were confirmed as P. multocida by use of one of two P. multocida-specific PCR assays (18, 19).
TABLE 1
| Strain no. | Heddleston serovar designation(s)b | Isolation date | Host species |
|---|---|---|---|
| X73 | H1 type strain | Prior to 1943 | Chicken |
| M1404 | H2 type strain | Prior to 1943 | Bison |
| P1059 | H3 type strain | Prior to 1943 | Turkey |
| P1662 | H4 type strain | 1968 | Turkey |
| P1702 | H5 type strain | 1971 | Turkey |
| P2192 | H6 type strain | 1971 | Chicken |
| P1997 | H7 type strain | 1971–1973 | Herring gull |
| P1581 | H8 type strain | 1971–1973 | Pine siskin |
| P2095 | H9 type strain | 1971–1973 | Turkey |
| P2100 | H10 type strain | 1971–1973 | Turkey |
| P903 | H11 type strain | 1971–1973 | Pig |
| P1573 | H12 type strain | 1971–1973 | Human |
| P1591 | H13 type strain | 1971–1973 | Human |
| P2225 | H14 type strain | 1971–1973 | Cattle |
| P2237 | H15 type strain | 1971–1973 | Turkey |
| P2723 | H16 type strain | 1974 | Turkey |
| PM1 | H3 (H3, H4) | 1993 | Turkey |
| PM3 | H15 (H4, H10, H15) | 1993 | Turkey |
| PM8 | H10 (H10) | 1993 | Turkey |
| PM18 | NT (H3) | 1986 | Chicken |
| PM19 | H13 (H3) | 1986 | Chicken |
| PM36 | H14 (H14) | 1985 | Unknown |
| PM37 | H3 (H3) | 1988 | Chicken |
| PM45 | NT (H3, H4) | 1986 | Chicken |
| PM46 | H6 (H6) | 1992 | Chicken |
| PM48 | H3 (H3, H4) | 1983 | Chicken |
| PM49 | NT (H1, H15) | 1984 | Chicken |
| PM51 | H9 (H4, H12) | 1984 | Chicken |
| PM64 | NT (H3) | 1979 | Chicken |
| PM67 | H3 (H3, H12) | 1969 | Turkey |
| PM72 | NT (H3, H14) | 1977 | Chicken |
| PM120 | H12 (H12) | 1993 | Chicken |
| PM135 | H8, H13 (H13) | Unknown | Turkey |
| PM140 | NT (H13, H14, H15) | 1994 | Chicken |
| PM147 | H7 (H7) | 1993 | Chicken |
| PM878 | H1, H4 | 2001 | Chicken |
| PM993 | H8 | 2002 | Duck |
| PM995 | H3 | 2002 | Chicken |
| PM1075 | H16 | 2004 | Chicken |
| PM1098 | H15 | 2004 | Unknown |
| PM1099 | H10 | 2004 | Unknown |
| PM1103 | H10 | 2004 | Unknown |
| PM1113 | NT | 2004 | Avian |
| PM1120 | NT | 2005 | Chicken |
| PM1124 | H1, H4, H12 | 2005 | Unknown |
| PM1128 | H10 | 2005 | Bovine |
| PM1132 | H1, H3, H4, H10, H14 | 2005 | Pig |
| PM1153 | H1, H3, H7 | 2005 | Avian |
| PM1165 | H1 | 2006 | Duck |
| PM1193 | H3 | 2006 | Duck |
| PM1205 | H1 | 2007 | Emu |
| PM1258 | NT | 2010 | Chicken |
| PM1268 | NT | 2010 | Chicken |
| PM1300 | H4 | 2009 | Turkey |
| PM1304 | H1 | 2009 | Chicken |
| PM1315 | H1 | 2009 | Chicken |
| PM1316 | H4 | 2009 | Unknown |
| PM1317 | H3 | 2009 | Unknown |
| PM1320 | H10, H13, H14 | 2010 | Chicken |
| PM1369 | H1 | 2010 | Chicken |
| PM1396 | H1, H3 | 2010 | Unknown |
| PM1398 | H1 | 2010 | Chicken |
| PM1405 | NT | 2010 | Chicken |
| PM1417 | H4 | 2010 | Chicken |
| PM1434 | NT | 2010 | Chicken |
| PM1435 | NT | 2010 | Chicken |
| PM1439 | NT | 2010 | Chicken |
| PM1441 | H2 | 2010 | Turkey |
| PM1455 | H1 | 2011 | Chicken |
| PM1456 | H14 | 2011 | Chicken |
| PM1457 | NT | 2011 | Chicken |
| PM1458 | H14 | 2011 | Chicken |
| PM1470 | H1 | 2011 | Turkey |
| PM1474 | H12 | 2011 | Duck |
a
Included are the Heddleston serovar type strains and details on the Australian P. multocida field isolates, including Heddleston serotyping results, isolation date, and host species.
b
The format in which multiple numbers are separated by a comma indicates that a precipitin line was observed with more than one type of serum. The presence of results in parentheses indicates that two distinct and separate serotyping assays were performed: the result in the parentheses is the first result with this isolate. NT, nontypeable by Heddleston serotyping.
Serotyping.
Each isolate of P. multocida was serotyped via the Heddleston method as described previously (3).
Molecular biology techniques.
Genomic DNA was purified from 1 ml of P. multocida overnight culture using the RBC genomic DNA purification kit (RBC, Taiwan). Each of the final LPS-mPCRs (50-μl final volume) was performed in 1× Taq polymerase buffer (10 mM Tris-HCl, 1.5 mM MgCl2, 50 mM KCl [Roche Diagnostic GmbH, Mannheim, Germany]) containing 0.4 μM each primer (Table 2), 0.2 mM deoxynucleoside triphosphates (dNTPs), and 1.7 U Taq polymerase (Roche Diagnostic GmbH, Mannheim, Germany). For each colony PCR, material from 2 to 3 well-isolated P. multocida colonies (obtained from overnight growth of each isolate at 37°C on an HI–1.5% agar plate) was collected using a sterile tip on a 20-μl micropipette (volume set at 20 μl) inserted into the middle of each colony. The collected material was then added to a 50-μl PCR mixture and mixed thoroughly by pipetting. For PCR using genomic DNA, approximately 50 ng of column-purified DNA was added to each PCR mixture. All reaction mixtures were mixed briefly then centrifuged (10 s, 13,000 × g). All PCRs were performed in an Eppendorf Mastercycler. For colony PCR, the cycling conditions were 96°C for 10 min, followed by 30 cycles of 96°C for 30 s, 52°C for 30 s, and 72°C for 2.5 min, with a final extension at 72°C for 5 min. For PCR using genomic DNA as the template, the cycling conditions were identical to those in the colony PCR, except that the initial denaturation step at 96°C was reduced to 5 min.
TABLE 2
| Locus | Primer | Sequence | Location | Product size (bp) |
|---|---|---|---|---|
| Oligonucleotides used in final LPS-mPCR typing assay | ||||
| L1 | BAP6119 | ACATTCCAGATAATACACCCG | Forward primer in pcgD | 1,307 |
| BAP6120 | ATTGGAGCACCTAGTAACCC | Reverse primer in pcgB | ||
| L2 | BAP6121 | CTTAAAGTAACACTCGCTATTGC | Forward primer in nctA | 810 |
| BAP6122 | TTTGATTTCCCTTGGGATAGC | Reverse primer in nctA | ||
| L3 | BAP7213 | TGCAGGCGAGAGTTGATAAACCATC | Forward primer in gatF | 474 |
| BAP7214 | CAAAGATTGGTTCCAAATCTGAATGGA | Reverse primer in gatF | ||
| L4 | BAP6125 | TTTCCATAGATTAGCAATGCCG | Reverse primer in latB | 550 |
| BAP6126 | CTTTATTTGGTCTTTATATATACC | Forward primer in latB | ||
| L5 | BAP6129 | AGATTGCATGGCGAAATGGC | Forward primer in rmlA | 1,175 |
| BAP6130 | CAATCCTCGTAAGACCCCC | Reverse primer in rmlC | ||
| L6 | BAP7292 | TCTTTATAATTATACTCTCCCAAGG | Forward primer in nctB | 668 |
| BAP7293 | AATGAAGGTTTAAAAGAGATAGCTGGAG | Reverse primer in nctB | ||
| L7 | BAP6127 | CCTATATTTATATCTCCTCCCC | Forward primer in ppgB | 931 |
| BAP6128 | CTAATATATAAACCATCCAACGC | Reverse primer in ppgB | ||
| L8 | BAP6133 | GAGAGTTACAAAAATGATCGGC | Forward primer in natG | 255 |
| BAP6134 | TCCTGGTTCATATATAGGTAGG | Reverse primer in natG | ||
| Oligonucleotides used only in LPS-mPCRv1a | ||||
| L3 | BAP6123 | TCCTTATCTGACATTGAAATCG | Forward primer in gatG | 415 |
| BAP6124 | CTAGACATCTGGTGGTTGCG | Reverse primer in gatG | ||
| L6 | BAP7039 | AATATCTTTATAATTATACTCTCCC | Forward primer in nctB | 668 |
| BAP6132 | AATGAAGGTTTAAAAGAGATAGC | Reverse primer in nctB |
a
These L3 and L6 primer sets were used in the initial LPS-mPCRv1 but were replaced in the final LPS-mPCR.
The PCR products generated from the LPS-mPCR were analyzed by gel electrophoresis using 2% agarose–Tris-acetate-EDTA (TAE) gel in 1× TAE buffer for 90 min at constant voltage (70 V).
For the initial LPS-mPCR (LPS-mPCR version 1 [LPS-mPCRv1]), all primers were used at a concentration of 0.3 μM, except for the L6 primers, which were used at 0.5 μM. The cycling conditions for the LPS-mPCRv1 using bacterial colony material as the template were 95°C for 10 min, followed by 30 cycles of 95°C for 30 s, 54°C for 30 s, and 72°C for 2.5 min, with a final extension at 72°C for 2 min.
Nucleotide sequences were determined by direct sequencing from genomic DNA and/or by sequencing of amplified PCR fragments as described previously (10). Sequencing reactions were analyzed using the Applied Biosystems 3730S genetic analyzer, and sequencing chromatograms were analyzed and the LPS loci assembled using Vector NTI Advance 11 (Invitrogen). Bioinformatic analyses, including amino acid sequence alignments, were conducted using BLAST and ClustalW2.
LPS sugar compositional analyses.
For compositional analysis of LPS produced by the Australian field isolates, small quantities of LPS were isolated from plate-grown cells as described previously (20). O-deacylated LPS (LPS-OH), core oligosaccharide (OS), and completely deacylated LPS were all isolated and purified from LPS as described previously (21). The sugar composition of the LPS from selected strains was determined by mass spectrometry as previously described (22). The predicted LPS structures produced by the P. multocida isolates were determined using MS compositional analysis and comparison with the known compositions and structures of the 16 Heddleston serovar type strains (9–16).
RESULTS
Design of a first-generation mPCR capable of differentiating P. multocida strains based on the genetics of LPS biosynthesis.
We have shown previously that the 16 unique LPS outer core structures produced by the P. multocida Heddleston type strains are generated from only eight distinct genetic loci, which we have named L1 to L8 (Fig. 1) (9–15, 20). We have therefore designated the following Heddleston serovar type strains as the LPS genotype type strains: X73 (L1), P1702 (L2), P1059 (L3), P2192 (L4), P2095 (L5), P1573 (L6), P1581 (L7), and P2723 (L8). Within each LPS genotype, strains displaying variation and/or truncation of the LPS structure can arise from random point mutations or deletions, in almost all cases, within the LPS outer core biosynthesis genes. These mutations can result in a change of function (sugar or donor specificity) or a total loss of function resulting in early termination of LPS assembly (10, 11). Given the random nature of LPS mutations, it was concluded that designing a PCR specific for each precise LPS structure identified was not possible. Thus, an mPCR assay was designed that was capable of differentiating the eight different LPS genotypes.
In all of the P. multocida strains so far examined, the genes required for synthesis of the LPS outer core are located between the conserved non-LPS genes priA and fpg (23). Each LPS outer core biosynthesis locus contains between 5 and 13 genes, including the highly conserved rpL31_2 gene, encoding ribosomal protein L31, which is not involved in LPS assembly (Fig. 1). To design an mPCR specific for the LPS outer core biosynthesis loci, a bioinformatic comparison was first performed using all the predicted protein sequences from each of the eight LPS outer core biosynthesis loci. One unique (or least similar) protein sequence was selected, and the corresponding nucleotide sequence was compared with the entire nucleotide sequence of each of the eight LPS loci. Where some similarity was observed in the selected region with the nucleotide sequence from another locus, nucleotide alignments were generated and the alignments visually inspected to identify the most divergent DNA sections. This information was used to design a set of eight primer pairs specific for each of the eight LPS biosynthetic loci (Table 2). Each primer pair was also designed to generate a distinct amplicon size for optimal electrophoretic separation on 2% agarose gels.
Testing of the LPS-mPCRv1.
All primer sets were initially tested in separate PCRs using either genomic DNA or colony-derived cells from each of the Heddleston type strains as the template. All PCRs amplified a product of the correct size when the appropriate type strains were used as the template (e.g., when the L1 primers were used against the L1 strains X73 and P2225). The different primer pairs were then combined and used in a single mPCR using either genomic DNA or colony-derived cells from each of the Heddleston type strains as the template. Following PCR optimization, a reproducible mPCR (LPS-mPCR version 1 [LPS-mPCRv1]) was developed that generated a single product of the expected size for all strains (Fig. 2A).
FIG 2
Testing of the LPS-mPCRv1 against Australian P. multocida field isolates.
To test the reproducibility and accuracy of LPS-mPCRv1, 58 P. multocida field isolates were typed using both LPS-mPCRv1 and classical Heddleston serotyping. The 58 field isolates included strains obtained from a range of Australian poultry farms and other sources between 1977 and 2011. In total, 33 of the strains were recorded as being isolated from chickens, 8 from turkeys, 4 from ducks, and 1 from an emu. The host species was not recorded for 10 of the isolates. One bovine isolate and one porcine isolate were also included (Table 1).
Historical strains were serologically typed using the Heddleston serotyping system when they were first received at the Australian reference laboratory (Agri-Science Queensland), and the typing was then repeated again for this study. For some strains, the serovar determined when the Heddleston serotyping was repeated was not in agreement with the initial serovar typing result (Table 1). Of the 58 strains, 32 gave an unambiguous Heddleston serovar result (55%) (Table 1), 17 gave an ambiguous result of two or more possible serovars, and 9 were nontypeable (no precipitin line observed). The most common serovars identified unambiguously were serovars 1 and 3.
All strains were then LPS genotyped using the LPS-mPCRv1. An example of an LPS-mPCRv1 result is shown in Fig. 2B. Of the 58 strains tested, the LPS-mPCRv1 gave an unambiguous LPS genotype for 48 of the strains (Table 3), but no PCR product could be generated for 10 strains (nontypeable). A comparison of the LPS-mPCRv1 genotype and Heddleston serovar designations of each strain (Table 3) revealed that there was complete agreement between the typing methods for only 16 of the 58 strains. Partial agreement was obtained for a further 11 strains (where serotyping gave an ambiguous result and the LPS-mPCRv1 result was in agreement with one of the serotyping results). For 15 strains, the LPS-mPCRv1 gave a locus designation that was incompatible with the serovar designation (Table 3). These data indicate that there were clear discrepancies between Heddleston serotyping and LPS genotype, as determined by the LPS-mPCRv1. Importantly, the LPS-mPCRv1 consistently assigned strains to a single genotype, whereas serotyping frequently assigned strains to multiple Heddleston serovars.
TABLE 3
| Parameter and strain no. | Serotypinga,b | LPS-mPCRv1 | Final LPS-mPCR (Heddleston serovars within each genotype)c | LPS compositiond |
|---|---|---|---|---|
| LPS-mPCR and serotyping in agreement | ||||
| PM36 | H14 (H14) | L1 | L1 (H1, H14) | No LPS analysis |
| PM37 | H3 (H3) | L3 | L3 (H3, H4) | No LPS analysis |
| PM45 | NT (H3, H4) | L3 | L3 (H3, H4) | No LPS analysis |
| PM46 | H6 (H6) | L4 | L4 (H6, H7) | 3HexNAc, 2Hex (H6) |
| PM120 | H12 (H12) | L6 | L6 (H10, H11, H12, H15) | 2Hex, 1Hep |
| PM1165 | H1 | L1 | L1 (H1, H14) | No LPS analysis |
| PM1193 | H3 | L3 | L3 (H3, H4) | 4Hex, Hep |
| PM1300 | H4 | L3 | L3 (H3, H4) | No LPS analysis |
| PM1304 | H1 | L1 | L1 (H1, H14) | No LPS analysis |
| PM1315 | H1 | L1 | L1 (H1, H14) | No LPS analysis |
| PM1316 | H4 | L3 | L3 (H3, H4) | No LPS analysis |
| PM1317 | H3 | L3 | L3 (H3, H4) | No LPS analysis |
| PM1398 | H1 | L1 | L1 (H1, H14) | 2PCho, 2Hex, Hep (H1) |
| PM1417 | H4 | L3 | L3 (H3, H4) | No LPS analysis |
| PM1455 | H1 | L1 | L1 (H1, H14) | 2PCho, 2Hex, Hep (H1) |
| PM1458 | H14 | L1 | L1 (H1, H14) | No LPS analysis |
| LPS-mPCR and serotyping in partial agreement | ||||
| PM3 | H15 (H4, H10, H15) | L3 | L3 (H3, H4) | 1Hex, Hep/2Hex, Hep |
| PM19 | H13 (H3) | L3 | L3 (H3, H4) | No LPS analysis |
| PM49 | NT (H1, H15) | L1 | L1 (H1, H14) | 2PCho, 2Hex, Hep (H1) |
| PM51 | H9 (H4, H12) | L6 | L6 (H10, H11, H12, H15) | 2Hex, Hep |
| PM67 | H3 (H3, H12) | L6 | L6 (H10, H11, H12, H15) | 1HexNAc, 3Hex, Hep (H12) |
| PM72 | NT (H3, H14) | L3 | L3 (H3, H4) | 3Hex, Hep (H4)/4Hex, Hep/1HexNAc, 4Hex, Hep (H3)/2HexNAc, 4Hex, Hep |
| PM140 | NT (H13, H14, H15) | L1 | L1 (H1, H14) | 1 Hex, Hep (H14) |
| PM878 | H1, H4 | L1 | L1 (H1, H14) | No LPS analysis |
| PM1124 | H1, H4, H12 | L1 | L1 (H1, H14) | 2PCho, 2Hex, Hep (H1) |
| PM1132 | H1, H3, H4, H10, H14 | L6 | L6 (H10, H11, H12, H15) | 3Hex, Hep/HexNAc, 3Hex, Hep (H12) |
| PM1396 | H1, H3 | L1 | L1 (H1, H14) | 2PCho, 2Hex, Hep (H1) |
| LPS-mPCR and serotyping not in agreement | ||||
| PM8 | H10 | L3 | L3 (H3, H4) | 2Hex, Hep/Hex, Hep/Hep |
| PM64 | NT (H3) | L6 | L6 (H10, H11, H12, H15) | 1HexNAc, 3Hex, Hep (H12) |
| PM147 | H7 (H7) | L3 | L3 (H3, H4) | 2Hex, Hep |
| PM993 | H8 | L6 | L6 (H10, H11, H12, H15) | No outer core (H10) |
| PM995 | H3 | L6 | L6 (H10, H11, H12, H15) | 1HexNAc, 3Hex, Hep (H12) |
| PM1098 | H15 | L3 | L3 (H3, H4) | 3Hex, Hep (H4) |
| PM1099 | H10 | L3 | L3 (H3, H4) | 3Hex, Hep (H4)/4Hex, Hep/1HexNAc, 4Hex, Hep (H3) |
| PM1103 | H10 | L3 | L3 (H3, H4) | 4Hex, Hep/1HexNAc, 4Hex, Hep (H3) |
| PM1128 | H10 | L3 | L3 (H3, H4) | 1HexNAc, 4Hex, Hep (H3) |
| PM1205 | H1 | L3 | L3 (H3, H4) | 3Hex, Hep (H4) |
| PM1320 | H10, H13, H14 | L3 | L3 (H3, H4) | Hex, Hep/2Hex, Hep (H4) |
| PM1441 | H2 | L3 | L3 (H3, H4) | 3Hex, Hep (H4)/4Hex, Hep/1HexNAc, 4Hex, Hep (H3) |
| PM1456 | H14 | L4 | L4 (H6, H7) | 1Hex (H7) |
| PM1470 | H1 | L3 | L3 (H3, H4) | 1HexNAc, 4Hex, Hep (H3)/2HexNAc, 4Hex, Hep |
| PM1474 | H12 | L3 | L3 (H3, H4) | 3Hex, Hep (H4), 4Hex, Hep |
| Nontypeable using Heddleston serotyping | ||||
| PM1113 | NT | L4 | L4 (H6, H7) | 1Hex (H7) |
| PM1268 | NT | L3 | L3 (H3, H4) | 2Hex, Hep, 3Hex, Hep (H4) |
| PM1405 | NT | L1 | L1 (H1, H14) | 2PCho, 2Hex, Hep (H1) |
| PM1435 | NT | L1 | L1 (H1, H14) | 2PCho, 2Hex, Hep (H1) |
| PM1439 | NT | L3 | L3 (H3, H4) | 3Hex, Hep (H4) |
| PM1457 | NT | L4 | L4 (H6, H7) | 3HexNAc, 1Hex/3HexNAc, 2Hex (H6) |
| Nontypeable using initial LPS-mPCRv1e | ||||
| PM1 | H3 (H3, H4) | NT | L3 (H3, H4) | 3Hex, Hep (H4)/4Hex, Hep/1HexNAc, 4Hex, Hep (H3) |
| PM18 | NT (H3) | NT | L3 (H3, H4) | 2Hex, Hep |
| PM48 | H3 (H3, H4) | NT | L3 (H3, H4) | 3Hex, Hep (H4)/4Hex, Hep/1HexNAc, 4Hex, Hep (H3) |
| PM135 | H8, H13 (H13) | NT | NT, L7 sequencef | 1HexNAc, 2Hex, Hep (H13) |
| PM1075 | H16 | NT | L3 (H3, H4) | No outer core |
| PM1120 | NT | NT | L3 (H3, H4) | No outer core |
| PM1153 | H1, H3, H7 | NT | L3 (H3, H4) | 3Hex, Hep (H4)/4Hex, Hep/1HexNAc, 4Hex, Hep (H3)/2HexNAc, 4Hex, Hep |
| PM1258 | NT | NT | L3 (H3, H4) | No outer core |
| PM1369 | H1 | NT | L3 (H3, H4) | 3Hex, Hep (H4)/4Hex, Hep |
| PM1434 | NT | NT | L3 (H3, H4) | 1HexNAc, 4Hex, Hep (H3) |
a
The format in which multiple numbers are separated by a comma indicates that a precipitin line was observed with more than one type serum. Results in parentheses indicate that two distinct and separate serotyping assays were performed: the result in parentheses is the first result with this isolate. NT, not able to be typed by this method (i.e., no precipitin line was observed using serotyping, or no amplicon was produced by mPCR).
b
A Heddleston serovar shown in boldface correlates with both LPS composition and LPS genotype.
c
An LPS genotype shown in boldface correlates with LPS composition.
d
Outer core LPS sugar composition as predicted by MS/MS compositional analysis. The Heddleston serovar within the designated genotype that matches the LPS outer core composition is shown in boldface and in parentheses. Those compositions without a Heddleston serovar designation shown in parentheses do not precisely match any of the Heddleston type strain LPS structures. Multiple LPS glycoforms when detected have been separated by “/.” Hex, hexose (glucose or galactose); HexNAc, N-acetyl hexosamine [N-acetylglucosamine, N-acetyl galactosamine, or (1S)-2-acetamido-2-deoxy-d-galactose]; Hep, heptose; PCho, phosphocholine. Nonstoichiometric phosphoethanolamine additions to the outer core have not been determined.
e
Sequencing of the LPS outer core biosynthesis locus revealed significant nucleotide differences where the LPS-mPCRv1 primers were located.
f
Sequencing of the PM135 LPS outer core biosynthesis locus revealed a large deletion in the region where the L7 LPS-mPCR primers were located.
In order to determine whether Heddleston serotyping or the LPS-mPCRv1 gave a more accurate representation of the LPS produced by each strain, the LPS composition from a set of selected strains was analyzed by mass spectrometry. The strains examined included five strains from the agreement group, all strains from the nonagreement group, 9 of the 11 strains from the partial agreement group (where Heddleston serotyping gave ambiguous results), and all strains that remained nontypeable in one or both typing systems. As we have reported previously, these analyses identified a number of strains belonging to the L3 genotype, which expressed multiple LPS glycoforms (10). When interpreting the typing results, the LPS glycoform that contained the largest number of sugars/residues was deemed to be representative of the most extended LPS structure produced by the strain, and any additional glycoforms observed that contained fewer sugars (but common to the largest glycoform) were considered truncated variants.
As expected, the five strains analyzed within the agreement group (Table 3) gave LPS-mPCRv1 designations that were in agreement with both LPS composition and with serotyping. The LPS composition of the nine strains examined within the partial agreement group correlated always with the genotype designation but correlated with one of the multiple Heddleston serovar designations in only 6 of the 9 strains (Table 3). The LPS composition of two strains in the agreement group (PM120 and PM1193) and two in the partial agreement group (PM3 and PM51) did not correlate precisely with the serovar-specific LPS structures within the assigned LPS genotype. However, in each case the LPS composition did correlate with a truncated version of the LPS structure specific to the assigned LPS genotype (10, 11).
For the strains where serotyping and PCR were in nonagreement, the LPS compositional analysis was always compatible with the LPS genotype assigned using the LPS-mPCRv1. In contrast, the Heddleston serotyping designation did not correlate with the predicted LPS composition for any strain in this group, clearly showing that Heddleston serotyping is unreliable for prediction of LPS composition. Importantly, the LPS-mPCRv1 gave unambiguous LPS genotyping results (producing only a single amplicon) that always correlated with LPS composition (Table 3).
Redesign of the LPS-mPCR to increase coverage.
The LPS-mPCRv1 gave an unambiguous LPS genotype for 48 of 58 field strains (Table 3) but failed to amplify a product from 10 isolates. PCR and nucleotide sequence analyses of the LPS outer core biosynthesis locus in each strain revealed that nine of the untypeable strains contained an L3 LPS locus but with significant nucleotide differences within gatG, where the L3 primers were located. The tenth strain, PM135, contained an L7 LPS locus but with a major deletion of 2,210 nucleotides (14) that included ppgB, where the L7 primers for the LPS-mPCRv1 were located.
To improve the strain coverage of the LPS-mPCRv1, the nucleotide sequence of the L3 type strain (P1059) was used to design new primers in a region within gatF that shared 100% identity with the nine L3 strains that were nontypeable using the LPS-mPCRv1 (Fig. 1). Substitution of the gatF L3 primers in the multiplex PCR resulted in amplification of all locus-specific products from the appropriate templates, except for locus 6, where only weak amplification of the product was observed for some strains (data not shown). To improve the L6 amplicon yield, the L6 primers were slightly modified. This final typing PCR was designated the LPS-mPCR. The full set of the final LPS-mPCR primers and amplicon sizes is shown in Table 2.
The final LPS-mPCR was used to genotype the 16 Heddleston type strains and was able to accurately differentiate all of these strains into the eight LPS genotypes (Fig. 3). The LPS-mPCR was then used to genotype the 58 field isolates. This final LPS-mPCR gave a single reproducible amplification product for 57 of 58 strains (strain PM135 was nontypeable), including the nine L3 strains that were nontypeable using the LPS-mPCRv1 (data not shown). All positive LPS-mPCR results were compatible with the LPS compositions that were determined (Table 3).
FIG 3
DISCUSSION
In this study, a multiplex PCR was designed to differentiate P. multocida strains on the basis of the LPS genotype. The final LPS-mPCR was able to unambiguously type the 16 Heddleston type strains and 57 of the 58 field isolates. However, strain PM135 remained untypeable, due to a large deletion in the region where the L7 LPS-mPCR primers were located (14). The failure of the LPS-mPCR due to large deletions within the LPS loci where primers are located cannot be avoided. However, our previous analyses of the Heddleston type strains and field isolates containing LPS gene mutations indicate that such large deletions are rare; most mutations within the P. multocida LPS outer core biosynthesis loci involve single point mutations (9–11) and would be unlikely to compromise PCR amplification.
During the testing of the LPS-mPCR for differentiation of P. multocida isolates, mass spectrometry analysis was used as the “gold standard” to assess the composition of the LPS produced by individual strains. This method of LPS analysis identifies sugar type and overall sugar content, which are then compared to the LPS compositions of the fully elucidated Heddleston LPS structures to predict the LPS glycoforms expressed by any particular strain.
Our previous analyses of the LPS genetics and structure in P. multocida showed that strains with the same serological designation can produce structurally distinct LPS (9–16). These analyses also showed that the LPS structures produced by many P. multocida strains are truncated variants of the full-length or “parent” LPS structure and that there is significantly more LPS diversity in the field than is represented by the current 16 type strains (9–12, 14). Interestingly, many P. multocida field isolates belonging to the L3, L4, and L6 genotypes produce a wide range of LPS glycoforms, including some which have a significantly truncated LPS outer core or no outer core at all. Many of these strains were isolated from poultry exhibiting clear signs of fowl cholera (data not shown), indicating that strains belonging to these genotypes do not require a full-length LPS molecule to cause disease. Our studies on the L1/serovar 1 strain VP161 have shown that any shortening of the LPS outer core structure in this strain results in attenuation of virulence in chickens (24), but it is possible that isolates expressing truncated LPS may be able to persist in some host niches but are not as virulent as parent strains expressing full-length LPS. Indeed, infection of chickens with a VP161 hptE LPS mutant (which produces a highly truncated outer core) showed that this mutant could persist at the site of muscle injection but could not be recovered from the blood, thus supporting this hypothesis (20). Importantly, many of the structures expressed by L3 and L6 genotype strains mimic host glycosphingolipids, and this may allow the bacteria to avoid recognition by the components of the innate immune system (10, 11).
Our comparison of LPS composition, Heddleston serotyping, and LPS-mPCR typing showed clearly that the Heddleston serovar designation frequently failed to correlate with the composition of the LPS produced by each strain. In contrast, the LPS-mPCR assay, specific for the identification of the LPS outer core biosynthesis loci, always correlated with LPS composition. Knowledge of the LPS genotype allows for the identification of the LPS type and possible range of LPS structures that strains can produce. However, the LPS-mPCR cannot predict the precise LPS structures produced by individual strains as random mutations within the LPS locus often lead to changes in LPS structure. We propose that the high diversity in numbers and types of LPS molecules produced by P. multocida strains indicates that a typing system exactly predictive of LPS structure is not feasible. If knowledge of the precise LPS structure is important for diagnosis and the control of outbreaks, then following mPCR analysis, further experiments would need to be conducted, such as carbohydrate-specific silver staining of cell lysates to assess the relative size of the LPS produced. Alternatively, for more detailed analysis, nucleotide sequencing of the LPS biosynthesis locus to identify the specific LPS gene mutations combined with carbohydrate mass spectrometry could be employed. However, these detailed analyses are beyond the scope of diagnostic laboratories.
The LPS-mPCR developed here is a highly reproducible typing system for differentiating P. multocida strains. We have also shown that many field isolates produce multiple LPS glycoforms simultaneously; these naturally occurring “multivalent” strains could be excellent candidates for killed-cell vaccines as they may show broader protective efficacy than strains expressing single LPS molecules. We are currently assessing this possibility.
ACKNOWLEDGMENTS
This research was partly conducted within the Poultry CRC, established and supported under the Australian Government's Cooperative Research Centres, and was also funded in part by the Australian Research Council, Canberra, Australia.
We thank Perry Fleming for bacterial growths and Jacek Stupak for capillary electrophoresis (CE)-MS.
REFERENCES
1.
Wilkie IW, Harper M, Boyce JD, Adler B. 2012. Pasteurella multocida: diseases and pathogenesis. Curr Top Microbiol Immunol 361:1–22.
2.
Carter GR. 1952. The type specific capsular antigen of Pasteurella multocida. Can J Med Sci 30:48–53.
3.
Heddleston KL, Gallagher JE, Rebers PA. 1972. Fowl cholera: gel diffusion precipitin test for serotyping Pasteurella multocida from avian species. Avian Dis 16:925–936.
4.
Glisson JR, Hofacre CL, Christensen JP. 2008. Fowl cholera, p 739–758. In Swayne DE (ed), Diseases of poultry, 12th ed. Blackwell Publishing, Ames, IA.
5.
Harper M, Boyce JD, Adler B. 2012. The key surface components of Pasteurella multocida: capsule and lipopolysaccharide. Curr Top Microbiol Immunol 361:39–51.
6.
Harper M, Cox AD, St. Michael F, Wilkie IW, Boyce JD, Adler B. 2004. A heptosyltransferase mutant of Pasteurella multocida produces a truncated lipopolysaccharide structure and is attenuated in virulence. Infect Immun 72:3436–3443.
7.
Singh R, Blackall PJ, Remington B, Turni C. 2013. Studies on the presence and persistence of Pasteurella multocida serovars and genotypes in fowl cholera outbreaks. Avian Pathol 42:581–585.
8.
Wilson MA, Morgan MJ, Barger GE. 1993. Comparison of DNA fingerprinting and serotyping for identification of avian Pasteurella multocida isolates. J Clin Microbiol 31:255–259.
9.
Harper M, St. Michael F, John M, Vinogradov E, Adler B, Boyce JD, Cox AD. 2011. Pasteurella multocida Heddleston serovars 1 and 14 express different lipopolysaccharide structures but share the same lipopolysaccharide biosynthesis outer core locus. Vet Microbiol 150:289–296.
10.
Harper M, St. Michael F, John M, Vinogradov E, Steen JA, van Dorsten L, Steen JA, Turni C, Blackall PJ, Adler B, Cox AD, Boyce JD. 2013. Pasteurella multocida Heddleston serovar 3 and 4 strains share a common lipopolysaccharide biosynthesis locus but display both inter- and intrastrain lipopolysaccharide heterogeneity. J Bacteriol 195:4854–4864.
11.
Harper M, St. Michael F, Steen JA, John M, Van Dorsten L, Parnas H, Vinogradov E, Adler B, Cox A, Boyce JD. 2014. Structural analysis of lipopolysaccharide produced by Heddleston serovars 10, 11, 12 and 15 and the identification of a new Pasteurella multocida LPS outer core biosynthesis locus, L6. Glycobiology 24:649–659.
12.
Harper M, St. Michael F, Steen JA, John M, Wright CL, Van Dorsten L, Vinogradov E, Adler B, Cox A, Boyce JD. 7 October 2014. Characterisation of the lipopolysaccharide produced by Pasteurella multocida serovars 6, 7 and 16; identification of lipopolysaccharide genotypes L4 and L8. Glycobiology
13.
Harper M, St. Michael F, Vinogradov E, John M, Boyce JD, Adler B, Cox AD. 2012. Characterization of the lipopolysaccharide from Pasteurella multocida Heddleston serovar 9; identification of a proposed bi-functional dTDP-3-acetamido-3,6-dideoxy-α-d-glucose biosynthesis enzyme. Glycobiology 22:332–344.
14.
Harper M, St. Michael F, Vinogradov E, John M, Steen JA, van Dorsten L, Boyce JD, Adler B, Cox AD. 2013. Structure and biosynthetic locus of the lipopolysaccharide produced by Pasteurella multocida serovars 8 and 13 and the identification of a novel phospho-glycero moiety. Glycobiology 23:286–294.
15.
St. Michael F, Harper M, Parnas H, John M, Stupak J, Vinogradov E, Adler B, Boyce JD, Cox AD. 2009. Structural and genetic basis for the serological differentiation of Pasteurella multocida Heddleston serotypes 2 and 5. J Bacteriol 191:6950–6959.
16.
St. Michael F, Li J, Cox AD. 2005. Structural analysis of the core oligosaccharide from Pasteurella multocida strain X73. Carbohydr Res 340:1253–1257.
17.
Tsai YC, Shien JH, Wu JR, Shieh HK, Chang PC. 2011. Polymerase chain reaction-restriction fragment length polymorphism analysis of the genes involved in the biosynthesis of the lipopolysaccharide of Pasteurella multocida. J Vet Diagn Invest 23:543–546.
18.
Miflin JK, Blackall PJ. 2001. Development of a 23S rRNA-based PCR assay for the identification of Pasteurella multocida. Lett Appl Microbiol 33:216–221.
19.
Townsend KM, Frost AJ, Lee CW, Papadimitriou JM, Dawkins HJ. 1998. Development of PCR assays for species- and type-specific identification of Pasteurella multocida isolates. J Clin Microbiol 36:1096–1100.
20.
Boyce JD, Harper M, St. Michael F, John M, Aubry A, Parnas H, Logan SM, Wilkie IW, Ford M, Cox AD, Adler B. 2009. Identification of novel glycosyltransferases required for assembly of the Pasteurella multocida A:1 lipopolysaccharide and their involvement in virulence. Infect Immun 77:1532–1542.
21.
St. Michael F, Li J, Vinogradov E, Larocque S, Harper M, Cox AD. 2005. Structural analysis of the lipopolysaccharide of Pasteurella multocida strain VP161: identification of both Kdo-P and Kdo-Kdo species in the lipopolysaccharide. Carbohydr Res 340:59–68.
22.
Harper M, Boyce JD, Cox AD, St. Michael F, Wilkie I, Blackall P, Adler B. 2007. Pasteurella multocida expresses two LPS glycoforms simultaneously both in vitro and in vivo but only a single form is required for virulence: identification of two acceptor specific heptosyl I transferases. Infect Immun 75:3885–3893.
23.
Harper M, Cox AD, Adler B, Boyce JD. 2011. Pasteurella multocida lipopolysaccharide: the long and the short of it. Vet Microbiol 153:109–115.
24.
Harper M, Cox AD, St. Michael F, Parnas H, Wilkie I, Blackall PJ, Adler B, Boyce JD. 2007. Decoration of Pasteurella multocida LPS with phosphocholine is important for virulence. J Bacteriol 189:7384–7391.
Information & Contributors
Information
Published In
Journal of Clinical Microbiology
Volume 53 • Number 2 • February 2015
Pages: 477 - 485
Editor: B. W. Fenwick
PubMed: 25428149
Copyright
© 2015 American Society for Microbiology.
History
Received: 30 September 2014
Returned for modification: 5 November 2014
Accepted: 16 November 2014
Published online: 23 January 2015
Contributors
Editor
B. W. Fenwick
Editor
Metrics & Citations
Metrics
Note: There is a 3- to 4-day delay in article usage, so article usage will not appear immediately after publication.
Citation counts come from the Crossref Cited by service.
Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
View Options
View options
PDF/ePub
PDF/ePubGet Access
Purchase Options
Subscribe and get full access to this article