Signature-Tagged Mutagenesis of Pasteurella Multocida Identifies

Signature-Tagged Mutagenesis of Pasteurella multocida Identiﬁes Mutants Displaying Differential Virulence Characteristics in Mice and Chickens Marina Harper,1 John D. Boyce,1 Ian W. Wilkie,2 and Ben Adler1* Bacterial Pathogenesis Research Group, Department of Microbiology, Monash University, Clayton, Victoria 3800,1 and Veterinary Pathology and Anatomy, University of Queensland, Brisbane, Queensland 4072,2 Australia

Received 14 April 2003/Returned for modiﬁcation 30 May 2003/Accepted 16 June 2003

Pasteurella multocida is the causative agent of fowl cholera in birds. Signature-tagged mutagenesis (STM) was used to identify potential virulence factors in a mouse septicemia disease model and a chicken fowl cholera model. A library of P. multocida mutants was constructed with a modified Tn916 and screened for attenuation in both animal models. Mutants identified by the STM screening were confirmed as attenuated by competitive growth assays in both chickens and mice. Of the 15 mutants identified in the chicken model, only 5 were also attenuated in mice, showing for the first time the presence of host-specific virulence factors and indicating the importance of screening for attenuation in the natural host.

Pasteurella multocida is a capsulated, gram-negative cocco- P. multocida mutants in two animal models to identify, for the bacillus and is the causative agent of avian fowl cholera, bovine first time, potential host-specific virulence factors. We screened hemorrhagic septicemia, atrophic rhinitis in pigs, and snuffles the mutant library in mice, the most widely used P. multocida in rabbits (30). To date, few virulence factors of P. multocida animal model, and in chickens, the natural host. have been identified; they include the capsule in serogroups A Modification of Tn916 for use in STM of P. multocida. The and B (7, 10), PMT toxin in strains causing atrophic rhinitis transposon Tn916 has been shown to transpose in a quasiran- (16), putative filamentous hemagglutinins PfhB1 and PfhB2 dom fashion into the P. multocida genome (14). In addition, (17), and several iron acquisition proteins, such as TonB, tetracycline resistance was stable in the absence of selection ExbD, and ExbB (6, 17, 31). over many generations, and the transposon could be intro- Recently, several studies have been undertaken to identify duced into P. multocida by electroporation rather than conju- the genes involved in the pathogenesis of pasteurellosis, in- gation, avoiding the need for a recipient strain carrying a coun- cluding in vivo expression technology (22), signature-tagged terselectable marker (14). However, in our study, the inherent mutagenesis (STM) (17), and whole-genome expression pro- instability of the transposon in the base plasmid pAM120 (18) filing (8). These genomic-scale methods have identified some resulted in an inadequate number of tagged transposons being true virulence factors and virulence-associated genes, including generated for library construction. To overcome this problem, those involved in iron transport and metabolism as well as we constructed a smaller version of the transposon, designated nucleotide and amino acid biosynthesis. However, many genes Tn916E⌬C, that was stable for manipulation in Escherichia coli identified by these analyses have no known function, and nei- and, when required, could be restored to an active transposon ther in vivo expression technology nor whole-genome expres- by removal of the erm(B) cassette from the int gene (Fig. 1). sion profiling gives direct information about the importance of Briefly, two PCR products were generated from pAM120. genes in virulence. The first was generated with primers BAP1127 (AGCAGTTC STM allows the large-scale screening of mutants for reduced TAGATGATGATACTGTCCC) and BAP1126 (TCGCTGCT survival in vivo, colonization and adhesion defects, and de- CGAGTTTATCCTCGCCAG) and included the right and left creased invasive ability (3, 9, 20). The method has been applied junctions of the transposon, the 3Ј end of the int gene, and the to many bacterial species, including a previous study on a entire base plasmid (Fig. 1). An XhoI site and XbaI site were bovine isolate of P. multocida using a tagged mini-Tn10 (17). introduced near the right junction of the transposon and the 3Ј This previous study used a septicemic mouse model of infec- end of the int gene, respectively. The second fragment, ampli- tion and identified 25 P. multocida genes that were important fied with BAP1128 (TCATCATCTAGAACTGCTTTCAG for survival and growth in vivo. However, one-third of the 62 TTG) and BAP1129 (TTTGGTACTCGAGAAGAACGGG mutants identified in that study contained multiple transposon AG), included the 5Ј end of the int gene (and an XbaI site) and insertions, and in these instances, it was not possible to deter- the entire tetracycline resistance gene, tet(M). An XhoI site mine the gene responsible for the attenuated phenotype (17). was also generated at the end of this fragment upstream of the In this study, we report the screening of an STM bank of tet(M) gene. A third fragment of DNA containing an erm(B) gene conferring erythromycin resistance was purified after digestion of the plasmid pJIR599 (4) with XbaI and inserted into * Corresponding author. Mailing address: Bacterial Pathogenesis the XbaI site engineered within the int gene. All three frag- Research Group, Department of Microbiology, Monash University, Clayton, Victoria 3800, Australia. Phone: 61 03 9905 4815. Fax: 61 03 ments were then ligated, and the subsequent E. coli transfor- 9905 4811. E-mail: [email protected]. mants were screened for the correct plasmid profile. The

5440 VOL. 71, 2003 NOTES 5441

FIG. 1. Schematic representation of the strategy used to construct Tn916E⌬C. (A) Linear diagram of pAM120. Bold lines represent the transposon Tn916, and the fine line represents the base plasmid. Only genes of relevance are shown. Arrows above the line indicate the direction of amplification from each oligonucleotide (not to scale). tet(M), tetracycline resistance gene; int, integrase gene required for excision and integration of the transposon; RJ, right junction of the transposon; LJ, left junction. (B) Fragments generated for construction of Tn916E⌬C. The erythromycin gene [erm(B)] was obtained by digestion of the plasmid pJIR599 with XbaI. The remaining fragments and the restriction sites XhoI and XbaI were generated by PCR amplification of pAM120 with the oligonucleotides shown in panel A. (C) Diagrammatic representation of the plasmid pPBA1638 containing the transposon Tn916E⌬C. The vertical arrow labeled TAG indicates the site of insertion (BspE1) of the variable DNA tags.

erm(B) insertion resulted in the inactivation of transposon products were digested with BspEI to digest any plasmid not excision during DNA manipulations, and the XbaI site located containing a DNA tag. at either end of the erm(B) gene allowed the removal of the A library of 42 uniquely tagged Tn916E⌬C transposons in cassette when transposition was required. The modified trans- the plasmid pPBA1638 was constructed by selecting tagged poson was designated Tn916E⌬C, as it conferred erythromycin transposons that would act as efficient templates for PCR and resistance (E) and no longer contained the conjugation genes hybridize specifically in DNA dot blot hybridizations. Prior to (⌬C). Tn916E⌬C in pAM120 was designated pPBA1638, which transposition into P. multocida, the erm(B) cassette was re- is a suicide vector in P. multocida. Prior to transposition into moved from the plasmids by digestion with XbaI and self- P. multocida, the erm(B)cassette was removed by enzyme di- ligation. Each plasmid harboring a uniquely tagged transposon gestion and religation. Tn916E⌬C was shown to transpose in a was then introduced separately into P. multocida strain VP161 quasirandom fashion in P. multocida, and most mutants had by electroporation. Mutant pools, each containing 42 mutants, only a single transposon insertion (data not shown). Construction and cloning of DNA tags. The variable oligonucleotide used for construction of the tagged Tn916E⌬C was essentially as described previously (20). To ensure tag fidelity, we constructed the tag from three oligonucleotides (Fig. 2) The first, designated BAP1298 [GTACAACCTCAAGCTT

(NK)20GAATTCGGTTAGAATG], was a 72-bp oligonucleotide, which included the 40-bp variable region. The remaining bases required for cloning into Tn916E⌬C were synthesized on separate oligonucleotides BAP1299 (ATAAACCCGGGTAC AACCTCAAGCT) and BAP1300 (GCATCCCGGGCATTC TAACCGAATTC), and each included an XmaI site for insertion into the BspEI site of Tn916. Double-stranded tags were generated by annealing BAP1298 and BAP1300 and allowing T4 polymerase to fill the 5Ј single-stranded extensions. The resulting double-stranded molecule was further extended by PCR amplification using oligonucleotides BAP1299 and BAP1300 (Fig. 2). FIG. 2. Construction of the variable DNA tags. The figure is not Generation of a P. multocida signature-tagged Tn916E⌬C drawn to scale. The oligonucleotide BAP1298, consisting of 72 bp, library. For construction of the tagged transposon library, the including the 40-bp variable region, was used as a template for the plasmid pPBA1638 was digested with BspE1 (Fig. 1) and li- full-length tag construction. Double-stranded tags were generated by annealing the oligonucleotides BAP1298 and BAP1300 and allowing gated to XmaI-digested DNA tags, generating a plasmid which T4 polymerase to fill the 5Ј single-stranded extensions. The resulting could no longer be digested by either enzyme. Prior to elec- double-stranded molecule was further extended by PCR amplification troporation into competent E. coli DH5␣ cells, the ligated with oligonucleotides BAP1299 and BAP1300. 5442 NOTES INFECT.IMMUN. were assembled by selecting a single P. multocida colony from by injection of approximately 105 CFU of the pool into the each transformation plate, and the procedure was repeated breast muscle. Infections were allowed to proceed until the until 10 pools were assembled. onset of terminal signs of infection (12 to 16 h), whereupon Other studies have indicated that the animal model chosen blood was recovered and the birds were euthanized in accor- for STM studies can influence the genes identified, as some dance with animal ethics requirements. An output probe was virulence genes will be host specific or involved in host adap- generated from the bacteria recovered from each bird and tation (47). We therefore chose to use two animal models, used separately in dot blot hybridizations to determine the namely, the mouse, which has been used regularly as a model relative abundance of each tagged mutant. for fowl cholera, and the chicken, a natural host for P. multo- Thirty-five mutants were identified as having reduced output cida. hybridization signals and subsequently tested individually for Screening of the P. multocida mutant library in vivo. For reduced growth in vivo with competitive growth assays. each input pool, the mutants were grown separately overnight, Identification of attenuated strains in both animal hosts. and an aliquot from each culture was transferred to a fresh Competitive growth assays were used to quantify the relative tube. A sample of the mix was then spread onto brain heart growth rates of putative attenuated mutants. Log-phase cul- infusion (BHI) plates with tetracycline and grown overnight. tures of mutant and wild-type were mixed at a ratio of 1:1, and Colonies were washed from the plate and used to prepare serial 10-fold dilutions were prepared. To determine the input genomic DNA representing the input pool. ratio of wild-type to mutant organisms, 100 ␮l of the appro- For screening in mice, female, inbred BALB/c mice, aged 8 priately diluted culture was plated immediately onto NB agar to 10 weeks, were injected intraperitoneally with approxi- in order to obtain single colonies representing the input pool. 5 ϫ 2 mately 10 CFU (approximately 2 10 CFU/mutant) from For the in vivo growth assays, animals were infected with ap- the input pool to ensure that all mutants were well repre- proximately 105 CFU, and blood samples were obtained at the sented. This dose gave reproducible disease progression, and terminal stages of the disease (6 h for mice or 12 to 18 h for all mice showed terminal signs of disease within 6 h, when chickens). Blood was diluted twofold in BHI containing hep- blood was recovered and the animals were euthanized in ac- arin and plated onto NB agar. For the in vitro growth assay, a cordance with animal ethics requirements. After infection, 100-fold dilution of the mixed bacteria was grown in BHI at blood was diluted fivefold in BHI and spread onto BHI plates 37°C with shaking for the same length of time as the animal containing tetracycline. A minimum of 5,000 colonies were infections, diluted appropriately, and spread onto NB agar. washed from these plates and used to prepare genomic DNA After growth on nonselective plates, a minimum of 100 in- representing the output pool. dividual colonies was patched onto NB agar with or without Each pool of mutants was tested in two mice, and the rela- tetracycline. Mutants were identified as attenuated if the pro- tive abundance of each tag in the input and output pools was portion of mutants obtained after in vivo growth was signifi- determined by dot blot hybridization. Genomic DNA from cantly less than the proportion obtained after in vitro growth. bacteria isolated from both animals was pooled prior to gen- Significance was determined by calculating the P value by an eration of the output probe and therefore represented the approximate z test for the difference in two proportions (Table average output profile from two animals. For preparation of the labeled probe, genomic DNA iso- 1) The relative competitive index (rCI) was determined by lated from each pool was used as a template for PCR ampli- dividing the percentage of tetracycline-resistant colonies fication with the primers BAP1129 and BAP1414 (GTGCAT (transposon mutants) obtained in vivo by the percentage of GAAATAATATACGAGT), generating an 800-bp fragment tetracycline-resistant colonies obtained in vitro (Table 1). that contained the unique tag region. The amplified products Of the 15 mutants originally identified in mice, only 5 were were digested with HindIII and EcoRI (located at the end of significantly attenuated, as determined by the competitive the variable tag region) followed by Klenow end filling. The growth assay in mice (Table 1). Subsequent competitive growth digested DNA was then separated on a 20% polyacrylamide assays in chickens showed that all five mutants were also at- gel, eluted overnight (2), and then labeled with the digoxigenin tenuated in chickens, although the level of attenuation varied oligonucleotide 3Ј end labeling kit (Roche Diagnostics). For between the two hosts. preparation of the target DNA for high-stringency dot blot Fifteen of the 35 mutants identified in the STM screening in hybridizations, each tag was amplified separately by PCR with chickens showed significantly reduced growth in competitive the primers BAP1414 and BAP1129, and 1 ng of amplified growth assays in chickens. Each of these was also tested in product was spotted onto nylon filters. competitive growth assays in mice to assess whether there was Mutants were identified as displaying reduced growth in vivo a difference in the level of attenuation between the two animal if the hybridization with the corresponding target DNA was models (Table 1). Only 5 of the 15 mutants were also attenu- stronger with the input probe than with the output probe. ated in mice; these were the same mutants identified in the Mutants identified in the first screening were collected into a mouse STM screening. new pool, injected into a mouse, and used in a second STM Southern hybridization confirmed single transposon inser- screen. After the second STM screening, 15 mutants were tions in all mutants. To identify the exact site of transposon identified that were potentially attenuated. Each mutant was insertion, inverse PCR was performed using Sau3A-digested then tested in a competitive growth assay together with wild- and ligated genomic DNA as the template and oligonucleo- type strain VP161. tides BAP1128 and BAP2252 (ACATAGAATAAGGCTTTA For screening of each STM pool in chickens, duplicate CGAGC) (36). The amplified products were sequenced di- chickens (commercially obtained Leghorn cross) were infected rectly by using BAP2252, and P. multocida DNA sequence VOL. 71, 2003 NOTES 5443

TABLE 1. Characterization of the P. multocida mutants

Competitive growth assay results in:

Group and mutant Disrupted gene or region Predicted function or homologue Mouse Chicken

rCIa P valueb rCIa P valueb Attenuated in chickens and mice AL249 dcaA Phosphoethanolamine transferase 0.33 Ͻ0.01 0.34 Ͻ0.01 0.41 Ͻ0.01 0.66 Ͻ0.01 0.12 Ͻ0.01 0.35 Ͻ0.01

AL250 pm0855 (ﬂp1) Surface adhesion 0.62 Ͻ0.01 0.39 Ͻ0.01 0.67 Ͻ0.01 0.49 Ͻ0.01 0.50 Ͻ0.01 0.89 0.18 0.79 0.02 0.71 0.01

AL251 pm1294 Heptosyl transferase 0.62 Ͻ0.01 Ͻ0.01 Ͻ0.01 0.44 Ͻ0.01 0.01 Ͻ0.01 0.77 0.03 0.01 Ͻ0.01 0.64 Ͻ0.01

AL252 purN De novo purine biosynthesis 0.20 Ͻ0.01 0.41 Ͻ0.01 0.16 Ͻ0.01 0.83 0.15 0.40 Ͻ0.01 0.28 Ͻ0.01 0.39 Ͻ0.01

AL396 Upstream of pm0529 Sodium/alanine symport protein 0.42 Ͻ0.01 0.16 Ͻ0.01 0.13 Ͻ0.01 0.20 Ͻ0.01

Attenuated only in chickens AL383 Between rpl20 and pm0606 Unknown 1.13 0.90 0.65 0.01

AL386 pm1069 Homology with H. inﬂuenzae P1 0.86 0.15 0.26 Ͻ0.01 surface protein 1.29 0.89

AL387 deoC Deoxyribose synthesis 1.02 0.56 0.69 0.01

AL388 pyrF Pyrimidine biosynthesis 1.48 0.99 0.58 Ͻ0.01 0.90 0.19

AL389 ponC Penicillin binding, murein synthesis 0.97 0.44 0.57 Ͻ0.01

AL390 pm1294 Heptosyl transferase; transposon in 0.72 0.04 0.47 Ͻ0.01 opposite orientation to AL251 0.83 0.15 1.13 0.20

AL391 pm1294 Heptosyl transferase; transposon in 0.85 0.08 Ͻ0.01 Ͻ0.01 same orientation as AL251 1.08 0.69 0.04 Ͻ0.01 1.07 0.24

AL392 pm0081 Homology with HI0902; probable 1.18 0.84 0.42 Ͻ0.01 integral membrane protein

AL393 Between purD and glyA Probable effect on glyA, which encodes 1.00 0.50 0.73 Ͻ0.01 glycine hydroxymethyltransferase

AL407 295 bp upstream of tfoX Between tfoX and a 16S RNA gene 3.15 Ͻ0.99 0.33 Ͻ0.01

a Values represent the rCI calculated for individual animals, determined by dividing the in vivo competitive index (CI) by the in vitro CI. Each CI value was determined by dividing the output mutant/wild-type ratio by the input mutant/wild-type ratio. b P values were calculated by using an approximate z test for the difference in two proportions. The test was performed to determine if the proportion of mutants in vivo was significantly less than the proportion obtained in vitro. adjoining the Tn916 sequence was searched against the Gen- pyrF, and deoC. Previous global screenings in bacteria such as Bank database with the BLASTX algorithm (1). Staphylococcus aureus, Streptococcus pneumoniae, and Yersinia Attenuated strains with insertions in metabolic genes. We enterocolitica also identified many biosynthetic genes which identified a number of mutants that had transposon insertions were important for growth in vivo (11, 13, 26). The mutant affecting genes encoding biosynthetic enzymes, including purN, AL252 had an insertion in the purN gene and was attenuated 5444 NOTES INFECT.IMMUN. in both mice and chickens (Table 1). This gene encodes the integrity of the cell wall, making the mutant more susceptible enzyme 5Ј-phosphoribosylglycinamide transformalase N, which to osmotic stress in vivo. catalyzes the fourth step of de novo purine biosynthesis (48). Three mutants, namely, AL251, AL390, and AL391, were The ability to synthesize purine nucleotides from simple pre- identified with insertions in the gene pm1294. This gene is cursors has been identified as being important for bacterial predicted to encode a heptosyl transferase, as the amino acid growth in vivo, and it has been reported recently that several of sequence displayed 58% identity to the heptosyltransferases II the genes involved in purine biosynthesis in P. multocida are and III from Haemophilus ducreyi. These H. ducreyi proteins up-regulated in bacteria harvested from the blood of infected are glycosyltransferases responsible for the addition of heptose chickens (8). In addition, in a previous STM study on a bovine to lipooligosaccharide (15). The three P. multocida mutants isolate of P. multocida in mice, a purF mutant was identified, had the transposon inserted at the same site, but in one, the product of which is also involved in the de novo purine AL390, the transposon was in the opposite orientation. Inter- biosynthesis pathway (17). The mutant AL388 (Table 1) was estingly, each mutant showed different levels of attenuation in attenuated only in chickens and had an insertion in the pyrF mice and chickens. Strain AL251 was significantly attenuated gene, which encodes an orotidine 5Ј-monophosphate decar- in mice, whereas AL390 and AL391 were not. In chickens, all boxylase that in E. coli is involved in pyrimidine biosynthesis three strains were attenuated, but AL251 and AL391 were (34). more so than AL390. Although the mutant AL251 was signif- The mutant AL387 had an insertion upstream of deoC and icantly attenuated in mice, with an average rCI value of 0.62, it deoR, and this strain was attenuated in chickens but not in was more highly attenuated in chickens, with an average rCI of mice. The E. coli homologs of these genes encode a deoxyri- less than 0.01. Interestingly, the strain AL391 had the same bose phosphate aldolase and a repressor protein, respectively, insertion site and transposon orientation as AL251 and was and are predicted to be part of the nucleoside catabolism also highly attenuated in chickens, but in contrast, it showed no pathway. In E. coli, purine and pyrimidine nucleosides and attenuation in mice. Preliminary data on the virulence of deoxynucleosides can be utilized in the nucleoside catabolic strains AL251 and AL391 indicate that they are both unable to pathway to extract pentose or deoxypentose for use as a carbon cause disease in chickens. The mutant AL249 had an insertion in the gene annotated and energy source. DeoC converts a deoxyribose intermediate as dcaA, whose protein’s amino acid sequence has a high de- to glylceraldehyde-3-P and acetaldehyde, which can be utilized gree of identity with LptA and Lpt3 from Neisseria meningitidis in glycolysis and the TCA cycle, respectively. In E. coli, four (12,28,31).InN.meningitidis,lpt-3andlptAencodephosphoetha- genes encoding the deoxynucleoside enzymes are located in nolamine transferases which are required for the addition of two adjacent operons, and the DeoR repressor protein, en- phosphoethanolamine (PEtn) to specific sites within the inner coded elsewhere, binds to two DNA binding sites upstream of core of lipopolysaccharide (LPS) (12, 28). The inactivation of the deoC,-A,-B, and -D genes, thereby generating a loop in the lpt-3 in N. meningitidis resulted in bacteria that were lacking DNA and preventing initiation of transcription (32). In P. PEtn on the LPS structure, and the mutants were less resistant multocida, the gene arrangement is different, with only deoC to bactericidal killing and in vitro opsonophagocytosis (28). and deoR being adjacent to each other. Without a functional The dcaA gene has also been identified as being significantly deoxyribose aldolase enzyme, the mutant will be unable to up-regulated in P. multocida during infections in chickens (8). convert the deoxyribose intermediates to the final products in Preliminary structural information on the P. multocida LPS the pathway. In a rich in vitro medium, this pathway may not be purified from the wild-type VP161 indicates that a PEtn mol- essential for the mutant to grow normally, but in chickens, the ecule is attached to the LPS (data not shown). Moreover, bacteria may need to utilize nucleosides as an alternative based on the positions of PEtn in N. meningitidis (28) and source of carbon and energy. It is also possible that in the Haemophilus influenzae (38, 39), it is unlikely that the lack of mutant, the transposon insertion upstream of deoC and deoR the PEtn in the mutant would have any significant effect on the affects the transcriptional regulation of the deo genes, as the LPS structure that would result in major structural changes in DeoR repressor protein may no longer be synthesized. the outer membrane. Therefore, the attenuation observed in Mutants with altered cell envelopes. Three genes predicted the P. multocida mutant AL249 is likely to be a direct effect of to be involved in cell envelope biosynthesis were identified: an altered LPS. Taken together, these data are consistent with ponC, pm1294, and dcaA. The mutant AL389 had a transposon a more important role for LPS in the progression of fowl insertion in ponC, a homologue of the E. coli gene pbpC, which cholera in chickens than in the progression of pasteurellosis in encodes the protein PBP1C that has a penicillin-binding do- mice. Currently, studies are being undertaken to analyze the main and functions as a transglycosylase for murein polymer- LPS composition of the mutant AL251 compared to wild-type ization (45). Murein is required for the biosynthesis of the LPS. murein sacculus, the scaffolding structure of the bacterial cell Notably, in the previous STM study of a bovine pneumonia wall. Mutants in the pbpC gene in E. coli are viable in vitro, but isolate of P. multocida in a septicemic mouse model, another the cross-linkage of the murein structure is different from that LPS mutant was identified with an insertion in lgtC, encoding of the wild type, with mainly nascent murein being used for the a galactosyltransferase, the homologue of which in H. influen- biosynthesis of the sacculus instead of a combination of pre- zae is responsible for the addition of galactose to the LPS existing murein strands with newly synthesized ones (37, 45). oligosaccharide (17, 21). The P. multocida ponC mutant was also viable in vitro but Mutants with insertions in genes encoding outer membrane showed significant attenuation in chickens, perhaps indicating proteins and surface structures. Fimbriae have previously that the changed structure of the murein sacculus affects the been identified on the surface of P. multocida cells and genes VOL. 71, 2003 NOTES 5445 encoding putative fimbriae, and fibrils are present in the P. biogroup aegyptius demonstrated that although both the wild multocida genome (23, 31, 42, 44). However, it has not been type and mutant grew at the same rate under anaerobic con- determined which of the fimbrial proteins are present within ditions, the expression of P1 was up-regulated in the wild-type the observed structures and what role, if any, they play in the strain (46). The P. multocida mutant AL386 was attenuated pathogenesis of disease. In this study, we identified one atten- only in chickens, suggesting that the uptake of long-chain fatty uated mutant, namely, AL250, with the transposon inserted acids as an energy source is required only during infections in between the start codon and the putative promoter of the gene chickens and not during in vitro growth or during infections in pm0855. The deduced amino acid sequence of pm0855 has mice. 43% identity with Flp1 from Actinobacillus actinomycetem- Concluding remarks. We have screened 420 P. multocida comitans, and the gene is located at the start of a potential STM mutants in both mice and chickens and identified 15 operon containing 14 genes. This putative P. multocida operon attenuated mutants and 13 disrupted genes or regions. Ten of also contains other genes with similarities to those from the flp the insertions caused attenuation in chickens but not in mice, operon of A. actinomycetemcomitans, including genes encoding indicating for the first time P. multocida virulence genes that tight adherence proteins (Tad) that are thought to be involved may be host specific. Five mutants were attenuated in both in the synthesis of long filamentous fimbriae (24). A second flp hosts, and no mutants were identified that were attenuated in homologue, flp2, was identified adjacent to flp1, but this puta- mice but not in chickens. tive gene has not been annotated in the published P. multocida Significantly, in this study, we did not identify any of the genome sequence (31). These two P. multocida genes contain genes identified in a previous STM study that used a bovine an Flp motif that has been predicted to be characteristic of a pneumonia isolate of P. multocida in a mouse infection model. new pilin gene subfamily (25). However, several genes were identified that encoded proteins In a previous STM study on P. multocida,atadD mutant was involved in similar pathways (17). These differences are unsur- shown to be attenuated. The tadD gene is positioned down- prising, as the studies utilized different transposons, the P. stream of the flp genes but potentially within the same operon multocida strains were isolated from two very different diseases (17). To determine if the tadD gene was affected in the flp (fowl cholera and bovine pneumonia), and each strain exhib- mutant AL250, we measured the levels of RNA expressed ited different infection kinetics in the mouse model. Further- from selected genes within the putative operon. Quantitative more, it is likely that neither study approached saturation cov- analysis of mRNA levels indicated that only the flp1 and, to a erage of the genome. lesser extent, flp2 mRNAs had reduced levels in the mutant, Mutant pools were introduced into the intraperitoneal cavity while the mRNA expression levels of genes downstream of flp2 of mice and via the breast muscle of chickens. While these were identical to those of the wild type (data not shown). This inoculation routes reliably cause disease in both animal mod- finding indicates that there may be a promoter located down- els, these infection routes would not allow the detection of stream of flp2. Given that the transposon was located between bacteria with mutations in genes involved in the initial attach- the putative promoter and the start codon, it was surprising ment and colonization of mucosal surfaces. Previous studies that there was any mRNA expression from the flp1 gene in the have indicated that birds can be infected via the respiratory mutant. tract, and it has been demonstrated that P. multocida can In competitive growth assays, the flp mutant was consistently adhere to turkey air sac macrophages (40, 41, 43). Future attenuated in mice. However, in chickens, there was variation studies will aim to detect mutants unable to attach and colo- in the level of attenuation between individual birds, with the flp nize by using intratracheal inoculation of the STM pools into mutant being able to grow at levels similar to that of the wild chickens. type in one of the four chickens tested. This may indicate real variability in the growth of the strain in different birds or may This work was funded in part by grants from the Australian Re- indicate that there is some instability of the mutant. search Council, the Australian Centre for Agricultural Research, and Finally, a mutant was identified with an insertion in pm1069, the Rural Industries Research and Development Corporation. We thank the staff at Veterinary Pathology, University of Queens- the amino acid sequence of which has significant similarity to land, for their valuable assistance with the chicken experiments and the E. coli protein FadL and the P1 outer membrane protein Vicki Vallance and Ian McPherson for their excellent technical assis- from H. influenzae.InE. coli, FadL has been identified as an tance. We also thank Aidan Sudbury for valuable statistical advice and outer membrane porin required for the uptake of long-chain Trudi Bannam for critical reading of the manuscript. fatty acids (5, 29). Mutants lacking FadL are unable to grow on medium containing long-chain fatty acids as a sole carbon REFERENCES source (35). A search of the PM70 genome sequence showed 1. Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. J. Mol. Biol. 215:403–410. that homologues of two other proteins required for the uptake 2. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. of long-chain fatty acids, namely, FadR and FadD, are also Smith, and K. Struhl (ed.). 1995. Current protocols in molecular biology, vol. present. The H. influenzae homologue P1 has been considered 1. John Wiley & Sons, Inc., New York, N.Y. 3. Badger, J. L., C. A. Wass, S. J. Weissman, and K. S. Kim. 2000. Application a potential vaccine candidate for both type b and nontypeable of signature-tagged mutagenesis for identification of Escherichia coli K1 strains of H. influenzae, as both polyclonal and monoclonal genes that contribute to invasion of human brain microvascular endothelial cells. Infect. Immun. 68:5056–5061. antibodies raised against P1 protect against bacteremia in the 4. Berryman, D. I., and J. I. Rood. 1995. The closely related ermB-ermAM infant rat model (19, 27). In H. influenzae type b, a P1 mutant genes from Clostridium perfringens, Enterococcus faecalis (pAM␤1), and was shown to grow at the same rate as wild-type H. influenzae Streptococcus agalactiae (pIP501) are flanked by variants of a directly repeated sequence. Antimicrob. Agents Chemother. 39:1830–1834. in vitro and could still induce bacteremia in the infant rat 5. Black, P. N., B. Said, C. R. Ghosn, J. V. Beach, and W. D. Nunn. 1987. model (33). However, analysis of a P1 mutant in H. influenzae Purification and characterization of an outer membrane-bound protein in- 5446 NOTES INFECT.IMMUN.

volved in long-chain fatty acid transport in Escherichia coli. J. Biol. Chem. 27. Loeb, M. R. 1987. Protection of infant rats from Haemophilus influenzae type 262:1412–1419. b infection by antiserum to purified outer membrane protein a. Infect Im- 6. Bosch, M., E. Garrido, M. Llagostera, A. M. P. de Rozas, I. Badiola, and J. mun. 55:2612–2618. Barbe. 2002. Pasteurella multocida exbB, exbD and tonB genes are physically 28. Mackinnon, F. G., A. D. Cox, J. S. Plested, C. M. Tang, K. Makepeace, P. A. linked but independently transcribed. FEMS Microbiol. Lett. 210:201–208. Coull, J. C. Wright, R. Chalmers, D. W. Hood, J. C. Richards, and E. R. 7. Boyce, J. D., and B. Adler. 2000. The capsule is a virulence determinant in Moxon. 2002. Identification of a gene (lpt-3) required for the addition of the pathogenesis of Pasteurella multocida M1404 (B:2). Infect. Immun. 68: phosphoethanolamine to the lipopolysaccharide inner core of Neisseria men- 3463–3468. ingitidis and its role in mediating susceptibility to bactericidal killing and 8. Boyce, J. D., I. Wilkie, M. Harper, M. L. Paustian, V. Kapur, and B. Adler. opsonophagocytosis. Mol. Microbiol. 43:931–943. 2002. Genomic scale analysis of Pasteurella multocida gene expression during 29. Maloy, S. R., C. L. Ginsburgh, R. W. Simons, and W. D. Nunn. 1981. growth within the natural chicken host. Infect. Immun. 70:6871–6879. Transport of long and medium chain fatty acids by Escherichia coli K12. 9. Chiang, S. L., and J. J. Mekalanos. 1998. Use of signature-tagged transposon J. Biol. Chem. 256:3735–3742. mutagenesis to identify Vibrio cholerae genes critical for colonization. Mol. 30. Mannheim, W. 1984. Family III. Pasteurellaceae,p.550–575. In N. R. Krieg Microbiol. 27:797–805. and J. G. Kolt (ed.), Bergey’s manual of systematic bacteriology, vol. 1. 10. Chung, J. Y., I. Wilkie, J. D. Boyce, K. M. Townsend, A. J. Frost, M. Williams & Wilkins, Baltimore, Md. Ghoddusi, and B. Adler. 2001. Role of capsule in the pathogenesis of fowl 31. May, B. J., Q. Zhang, L. L. Li, M. L. Paustian, T. S. Whittam, and V. Kapur. cholera caused by Pasteurella multocida serogroup A. Infect. Immun. 69: 2001. Complete genomic sequence of Pasteurella multocida, Pm70. Proc. 2487–2492. Natl. Acad. Sci. USA 98:3460–3465. 11. Coulter, S. N., W. R. Schwan, E. Y. Ng, M. H. Langhorne, H. D. Ritchie, S. 32. Mortensen, L., G. Dandanell, and K. Hammer. 1989. Purification and char- Westbrock-Wadman, W. O. Hufnagle, K. R. Folger, A. S. Bayer, and C. K. acterization of the DeoR repressor of Escherichia coli. EMBO J. 8:325–331. Stover. 1998. Staphylococcus aureus genetic loci impacting growth and sur- 33. Munson, R., Jr., and A. Hunt. 1989. Isolation and characterization of a vival in multiple infection environments. Mol. Microbiol. 30:393–404. mutant of Haemophilus influenzae type b deficient in outer membrane pro- 12. Cox, A. D., J. C. Wright, J. Li, D. W. Hood, E. R. Moxon, and J. C. Richards. tein P1. Infect. Immun. 57:1002–1004. 2003. Phosphorylation of the lipid a region of meningococcal lipopolysac- 34. Neuhard, J., and R. A. Kellin. 1996. Biosynthesis and conversions of pyrim- charide: identification of a family of transferases that add phosphoethanol- idines, p. 580–599. In F. C. Neidhardt, R. Curtiss III, J. L. Ingraham, E. C. C. amine to lipopolysaccharide. J. Bacteriol. 185:3270–3277. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and 13. Darwin, A. J., and V. L. Miller. 1999. Identification of Yersinia enterocolitica H. E. Umbarger (ed.), Escherichia coli and Salmonella: cellular and molec- genes affecting survival in an animal host using signature-tagged transposon ular biology, 2nd ed., vol. 1. ASM Press, Washington D.C. mutagenesis. Mol. Microbiol. 32:51–62. 35. Nunn, W. D., and R. W. Simons. 1978. Transport of long-chain fatty acids by 14. DeAngelis, P. L. 1998. Transposon Tn916 insertional mutagenesis of Pasteu- Escherichia coli: mapping and characterization of mutants in the fadL gene. rella multocida and direct sequencing of disruption site. Microb. Pathog. Proc. Natl. Acad. Sci. USA 75:3377–3381. 24:203–209. 36. Ochman, H., A. S. Gerber, and D. L. Hartl. 1988. Genetic applications of an 15. Filiatrault, M. J., B. W. Gibson, B. Schilling, S. H. Sun, R. S. Munson, and inverse polymerase chain reaction. Genetics 120:621–623. A. A. Campagnari. 2000. Construction and characterization of Haemophilus 37. Park, J. T. 1996. The murein sacculus, p. 48–57. In F. C. Neidhardt, R. ducreyi lipooligosaccharide (LOS) mutants defective in expression of hepto- Curtiss III, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. syltransferase III and ␤1, 4-glucosyltransferase: identification of LOS glyco- Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.), Escherichia forms containing lactosamine repeats. Infect. Immun. 68:3352–3361. coli and Salmonella: cellular and molecular biology., 2nd ed., vol. 1. ASM 16. Foged, N. T., J. P. Nielsen, and S. E. Jorsal. 1989. Protection against pro- Press, Washington, D.C. gressive atrophic rhinitis by vaccination with Pasteurella multocida toxin 38. Phillips, N. J., M. A. Apicella, J. M. Griffiss, and B. W. Gibson. 1992. purified by monoclonal antibodies. Vet. Rec. 125:7–11. Structural characterization of the cell surface lipooligosaccharides from a 17. Fuller, T. E., M. J. Kennedy, and D. E. Lowery. 2000. Identification of nontypable strain of Haemophilus influenzae. Biochemistry 31:4515–4526. Pasteurella multocida virulence genes in a septicemic mouse model using 39. Phillips, N. J., M. A. Apicella, J. M. Griffiss, and B. W. Gibson. 1993. signature-tagged mutagenesis. Microb. Pathog. 29:25–38. Structural studies of the lipooligosaccharides from Haemophilus influenzae 18. Gawron-Burke, C., and D. B. Clewell. 1984. Regeneration of insertionally type b strain A2. Biochemistry 32:2003–2012. inactivated streptococcal DNA fragments after excision of transposon Tn916 40. Pruimboom, I. M., R. B. Rimler, and M. R. Ackermann. 1999. Enhanced in Escherichia coli: strategy for targeting and cloning of genes from gram- adhesion of Pasteurella multocida to cultured turkey peripheral blood mono- positive bacteria. J. Bacteriol. 159:214–221. cytes. Infect. Immun. 67:1292–1296. 19. Hansen, E. J., S. M. Robertson, P. A. Gulig, C. F. Frisch, and E. J. Haanes. 41. Pruimboom, I. M., R. B. Rimler, M. R. Ackermann, and K. A. Brogden. 1996. 1982. Immunoprotection of rats against Haemophilus influenzae type B dis- Capsular hyaluronic acid-mediated adhesion of Pasteurella multocida to tur- ease mediated by monoclonal antibody against a haemophilus outer-mem- key air sac macrophages. Avian Dis. 40:887–893. brane protein. Lancet i:366–368. 42. Rebers, P. A., A. E. Jensen, and G. A. Laird. 1988. Expression of pili and 20. Hensel, M., J. E. Shea, C. Gleeson, M. D. Jones, E. Dalton, and D. W. capsule by the avian strain P-1059 of Pasteurella multocida. Avian Dis. 32: Holden. 1995. Simultaneous identification of bacterial virulence genes by 313–318. negative selection. Science 269:400–403. 43. Rhoades, K. R., and R. B. Rimler. 1990. Pasteurella multocida colonization 21. Hood, D. W., M. E. Deadman, M. P. Jennings, M. Bisercic, R. D. Fleisch- and invasion in experimentally exposed turkey poults. Avian Dis. 34:381–383. mann, J. C. Venter, and E. R. Moxon. 1996. DNA repeats identify novel 44. Ruffolo, C. G., J. M. Tennent, W. P. Michalski, and B. Adler. 1997. Identi- virulence genes in Haemophilus influenzae. Proc. Natl. Acad. Sci. USA 93: fication, purification, and characterization of the type 4 fimbriae of Pasteu- 11121–11125. rella multocida. Infect. Immun. 65:339–343. 22. Hunt, M. L., D. J. Boucher, J. D. Boyce, and B. Adler. 2001. In vivo-expressed 45. Schiffer, G., and J. V. Holtje. 1999. Cloning and characterization of PBP1C, genes of Pasteurella multocida. Infect. Immun. 69:3004–3012. a third member of the multimodular class A penicillin-binding proteins of 23. Isaacson, R. E., and E. Trigo. 1995. Pili of Pasteurella multocida of porcine Escherichia coli. J. Biol. Chem. 274:32031–32039. origin. FEMS Microbiol. Lett. 132:247–251. 46. Segada, L. M., G. M. Carlone, L. L. Gheesling, and A. J. Lesse. 2000. 24. Kachlany, S. C., P. J. Planet, M. K. Bhattacharjee, E. Kollia, R. DeSalle, Characterization of P1-deficient isogenic mutant of Haemophilus influenzae D. H. Fine, and D. H. Figurski. 2000. Nonspecific adherence by Actinoba- biogroup aegyptius associated with Brazilian purpuric fever. Microb. Pathog. cillus actinomycetemcomitans requires genes widespread in bacteria and ar- 28:145–155. chaea. J. Bacteriol. 182:6169–6176. 47. Tsolis, R. M., S. M. Townsend, E. A. Miao, S. I. Miller, T. A. Ficht, L. G. 25. Kachlany, S. C., P. J. Planet, R. DeSalle, D. H. Fine, D. H. Figurski, and J. B. Adams, and A. J. Baumler. 1999. Identification of a putative Salmonella Kaplan. 2001. flp-1, the first representative of a new pilin gene subfamily, is enterica serotype Typhimurium host range factor with homology to IpaH and required for non-specific adherence of Actinobacillus actinomycetemcomi- YopM by signature-tagged mutagenesis. Infect. Immun. 67:6385–6393. tans. Mol. Microbiol. 40:542–554. 48. Zalkin, H., and P. Nygaard. 1996. Biosynthesis of purine nucleotides, p. 26. Lau, G. W., S. Haataja, M. Lonetto, S. E. Kensit, A. Marra, A. P. Bryant, D. 561–579. In F. C. Neidhardt, R. Curtiss III, J. L. Ingraham, E. C. C. Lin, K. B. McDevitt, D. A. Morrison, and D. W. Holden. 2001. A functional genomic Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. analysis of type 3 Streptococcus pneumoniae virulence. Mol. Microbiol. 40: Umbarger (ed.), Escherichia coli and Salmonella: cellular and molecular 555–571. biology, 2nd ed., vol. 1. ASM Press, Washington, D.C.

Editor: J. T. Barbieri