Production of a P. multocida strain M1404 Himar1 mutant library, systemic infections in mice, and TraDIS library production

We previously constructed a Himar1 transposon mutant library in the P. multocida strain VP161 (capsule type A and LPS type L1) [31], which was also used in this study. For TraDIS analysis of strain M1404 (capsule type B and LPS type L2), we generated a comparable Himar1 transposon mutant library. A total of ~1.12 x 10⁶ M1404 Himar1 transposon mutants were pooled to produce the final library. To generate a high-quality M1404 reference genome for TraDIS analysis, we used a combination of Nanopore and Illumina sequencing that allowed for the assembly of a 2.42 Mbp genome that contained two contigs (see S1 Text for full assembly methodology and S3 Text for full annotation).

The M1404 and VP161 Himar1 mutant libraries were each used to infect BALB/c mice at a dose of ~2 x 10⁷ CFU via intraperitoneal injection (project overview described in S1 Fig, infectious dose given in S1 Table). All mice succumbed to systemic infection, reaching humane endpoints at five hours (S2A Fig). Surviving P. multocida mutants were recovered from the blood, liver, and spleen by plating in vivo-derived samples onto rich media. Greater than 4 x 10⁶ CFU of P. multocida were recovered from each site for all infections (S2B Fig). Four TraDIS library replicates were generated using genomic DNA isolated from the recovered mutants representing each strain (VP161 and M1404) and condition (rich media input culture, recovered mutants from each tissue following infection in mice). Each library was sequenced on an Illumina NextSeq and analysed using the Bio-TraDIS pipeline and related scripts (full TraDIS data in S1 Data). Combined, the four TraDIS library replicates generated for each strain and condition had high transposon-DNA junction read coverage across the genome and a small bp distance between unique insertion sites (UIS) (S3 Fig); the VP161 libraries had between 31 and 33 bp distance between each UIS, and the M1404 libraries had between 40 and 43 bp between each UIS (S2 Table). Principal component analysis using UIS or reads per gene showed that in all conditions the rich media TraDIS libraries clustered together and separated away from the mouse TraDIS libraries (S4 Fig).

Identification of core genes essential for growth in rich media

We used the input mutant libraries, grown on solid heart infusion media overnight, to identify genes essential for growth in rich media for both P. multocida strain VP161 and M1404. Essential genes were identified by comparing normalised UIS counts per gene (Figs 1A and S5; see S2 Data for the list of all essential genes for growth on rich media), identifying genes that resulted in lethality or a severe fitness defect when insertionally inactivated. There were 585 genes, of which 542 were coding sequences (CDS), identified as essential for strain VP161. There were also 585 genes, of which 540 were CDS, essential for strain M1404. Although a similar number of protein encoding genes were identified as essential in both strains, only 461 were homologs (Fig 1B), as determined using eggNOG or BlastP. The majority of these 461 shared essential proteins were involved in housekeeping processes. Analysis using the STRING database identified several processes and pathways that were enriched in the essential protein set, including: DNA replication (false discovery rate [FDR] = 4 x 10^-4), transcription (FDR = 0.047), translation (FDR = 2.04 x 10^-16) including ribosomal structural proteins (FDR = 2.7 x 10^-5) and tRNA amino acylation (FDR = 7.83 x 10^-5), cell division and regulation of cell shape (FDR = 0.015), fatty acid biosynthesis (FDR = 4 x 10^-6), peptidoglycan (FDR = 0.004) and lipid A biosynthesis (FDR = 0.019), glycolysis (FDR = 0.026), purine nucleoside biosynthesis (FDR = 0.0026), heme/porphyrin biosynthesis (FDR = 0.027), isoprenoid biosynthesis (FDR = 0.027), and ATP production (FDR = 0.0033).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Identification and comparison of genes essential for P. multocida strains VP161 and M1404 to grow in rich media.

Both the VP161 and M1404 Himar1 mutant libraries were grown to mid-exponential phase in heart infusion (HI) broth, before being plated on HI agar. Viable mutants were recovered, used to generate TraDIS libraries that were sequenced by Illumina sequencing, and the data analysed using Bio-TraDIS and related scripts. A. Insertion indexes (number of unique Himar1 transposon insertion sites divided by gene length) for all genes following growth in HI broth. The cut-off for essentiality is marked by a grey line for each dataset. B. Comparison of essential protein encoding genes present in both strain VP161 and M1404, or only one of the strains. Protein homologs between strains were identified using eggNOG matches, or by having > 90% amino acid identity. A 100% P. multocida core-genome was identified in strains VP161 and M1404 along with 167 complete, publicly available P. multocida genomes using Roary, and compared using a > 90% amino acid identity cut-off to the essential proteins identified in both strains VP161 and M1404.

https://doi.org/10.1371/journal.ppat.1013398.g001

We then assessed if these shared essential genes were also conserved across the entire P. multocida species. We analysed 167 publicly available complete P. multocida genomes, together with VP161 and M1404, and identified a total of 1,564 genes present in all strains (the 100% core-genome). A core-genome phylogeny generated from this set of 169 complete genomes was highly similar to a previously constructed P. multocida core-genome phylogeny generated using 402 P. multocida strains that included complete and incomplete genomes (S6 Fig, see S2 Text for newick format file) [7], indicating that this subset of 169 strains is representative of the entire P. multocida species. Of the 461 essential genes identified in both strain VP161 and strain M1404, 447 (97%) were present in the 100% core-genome (Fig 1B). We also compared the 100% core-genome and conserved rich media essential genes to our curated P. multocida-specific virulence factor and antibiotic resistance database (PastyVRDB) [7]. This analysis identified 53 genes associated with virulence that are present in all P. multocida strains, and included the genes hpbA, oma87, spoT, fur, Pm0442, and kdtA, which were all essential in both strains (S3 Table).

There were 75 genes essential in VP161 that had a non-essential homolog in M1404, and 65 genes essential in M1404 that had a non-essential homolog in VP161 (Fig 1B). Genes identified as essential in VP161 but not M1404 included crp, crr and relA, and genes identified as essential in M1404 but not VP161 included several genes involved in the Tol-Pal system (tolBQR and pal), DNA repair (ruvBC and aroQ), and sugar transport (ptsH). There were also six essential VP161 genes that had no homolog in M1404, and 14 essential genes in M1404 that had no homolog in VP161 (Fig 1B). Most of the unique essential M1404 genes encoded hypothetical proteins; however, M1404_00858 and M1404_00859 contained antitoxin-like domains, and M1404_01327 contained a putative peptidase toxin domain.

P. multocida strains VP161 and M1404 have both shared and unique genes required for virulence in mice

To identify genes that were important for P. multocida fitness during murine systemic infections, we compared the read count per gene between the mouse and rich media TraDIS libraries. Any gene with a four-fold decrease in Himar1 insertions and a q-value < 0.01 was identified as resulting in decreased in vivo fitness when disrupted, and any gene with a four-fold increase in Himar1 insertions and a q-value < 0.01 was identified as showing increased in vivo fitness when disrupted (see S3 and S4 Data for full fitness-associated data). Overall, 134 genes were in vivo fitness-associated in at least one type of tissue and one strain (Table 1 and Fig 2A). There were 63 and 94 genes that resulted in decreased in vivo fitness of strain VP161 and M1404, respectively (Fig 2B and S4 Table). For both VP161 and M1404, most of these genes were identified as important for in vivo fitness in all three tissues (Fig 2B). The genes were then compared to identify homologs between the two strains. Of the genes that had a homolog in the other strain, 35 were important for fitness in both strains, 23 were important for fitness only in VP161, and 47 were important for fitness only in M1404 (Figs 2C and 3). Furthermore, there were five genes important for fitness in VP161 that did not have a homolog in M1404, and 12 genes important for fitness in M1404 that did not have a homolog in VP161 (Figs 2C and 3). In total, there were seven genes that increased fitness when disrupted in either VP161 or M1404, with only two genes, yeeZ and speB, common to both strains (Fig 3 and S5 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Number of genes identified in P. multocida strain VP161 or M1404 that when disrupted by a Himar1 transposon insertion resulted in decreased or increased in vivo fitness.

https://doi.org/10.1371/journal.ppat.1013398.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Identification and comparison of genes associated with in vivo fitness for P. multocida strain VP161 and M1404 during systemic infections in mice.

The P. multocida strain VP161 and M1404 Himar1 libraries were individually used in systemic infections in BALB/c mice. Surviving mutants were recovered from the bloodstream, liver and spleen, and then plated on heart infusion agar. Genomic DNA was extracted from surviving mutants and used to generate TraDIS libraries. TraDIS libraries were sequenced via Illumina sequencing and the data analysed using Bio-TraDIS and related scripts. A. Volcano plots showing the log₂ fold-change in read count per gene and -log₁₀ q-value for all genes in the bloodstream (red), liver (blue), and spleen (gold) TraDIS libraries compared to the rich media TraDIS library. Genes were identified as in vivo fitness-associated if they had a four-fold increase or decrease in read count, with a q-value of < 0.01 (cut-offs shown as grey lines). B. Venn diagrams showing the number of genes identified as important for in vivo fitness that were shared between the bloodstream, liver, and spleen TraDIS libraries. C. Comparison of in vivo fitness genes between strain VP161 and M1404. Protein homologs between strains were identified using matching eggNOG IDs, or by having > 90% amino acid identity and > 90% coverage.

https://doi.org/10.1371/journal.ppat.1013398.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Heat maps showing the log₂ fold-change in reads for in vivo fitness-associated genes.

The log₂ fold-change in reads per gene was identified using the fitness.comparison R-script, using the rich media TraDIS data as the control, and each of the bloodstream, liver, and spleen TraDIS libraries as the condition. Genes were included in the heatmap if they were identified as in vivo fitness-associated in any one of the TraDIS libraries. The gene clcA, which is known not to impact in vivo fitness [70], was included as a control. A. In vivo fitness-associated genes that are present in both strains. For genes that are annotated with different names or are identified only by different locus tags, both names or locus tags are given, with the VP161 name or locus tag first followed by the M1404 name or locus tag. For the locus tags, the prefix (PmVP161_ or M1404_) has been omitted. B. In vivo fitness-associated genes in M1404 with no homolog in VP161. C. In vivo fitness-associated genes in VP161 with no homolog in M1404.

https://doi.org/10.1371/journal.ppat.1013398.g003

Common pathways and processes were identified as important for both VP161 and M1404 during infection in mice (Figs 3A, S7 and S8). In both strains, disruption of almost all capsule biosynthesis genes had a severe fitness-cost. Both fis and hfq, which encode transcriptional regulators known to positively regulate capsule biosynthesis genes in strain VP161, were important for fitness in VP161 but not in M1404, suggesting these strains used different regulatory mechanisms. Several genes encoding proteins involved in LPS core-polysaccharide synthesis were identified as important for in vivo fitness. Genes from the ADP-L-glycero-β-D-manno-heptose synthesis pathway (hldDE-gmhAB), that encode proteins that synthesise a modified heptose that is added to the LPS core-polysaccharide, were either important for in vivo fitness or essential in both strains. The gene hptA, which encodes an LPS inner-core heptosyltransferase, was identified as important for fitness in both strains. Additionally, genes encoding the LPS inner-core heptosyltransferase genes hptC and hptD, the LPS inner-core glucosyltransferase gtcD, and the LPS outer-core galactosyltransferase gatD were important for fitness of strain M1404. Several genes identified as important for in vivo fitness in both strains encoded proteins involved in amino acid synthesis and transport, or protein translation (Fig 3). These included an aminotransferase involved in aspartate, phenylalanine, and tyrosine synthesis (aspC), an asparagine synthase (asnA) and regulator of asparagine synthase expression (asnC), a putative alanine transporter (alsT_1), and proteins involved in translation efficiency (prmC and epmA) (Fig 3). Two genes in the L-serine biosynthesis pathway (serAB) and a putative tyrosine transporter (tyrP_1) were important for the in vivo fitness of VP161 but not M1404. Conversely, elongation factor Tu subunit genes (tufB_1 and tufB_2) were important for the in vivo fitness of strain M1404 but not strain VP161. Other systems important for in vivo fitness in both strains included the purine biosynthesis pathway, the Ton-B energy transduction system (tonB-exbBD), methionine uptake (plpB-metPN), and aroA from the chorismate pathway.

Several pathways were identified as important for in vivo fitness of M1404 but not VP161 (Fig 3A). This included genes encoding two components of the sialic acid uptake system (nanP and nanU), sialic acid metabolism (nanE) and the N-acetylneuraminate CMP transferase (neuA_1), a sensor histidine kinase (cpxA), the sorbitol-specific EIIA (srlB), and genes encoding the Mla phospholipid transport system (mlaBDEF). The genes encoding adenylate cyclase (cyaA) and cAMP receptor protein (crp) were also identified as important for in vivo fitness of M1404 but not VP161; however, crp was identified as essential for strain VP161 to grow in rich media. The gene M1404_01811, which encodes a P. multocida capsule type B- and E-specific surface lipoprotein (SLP) PmSLP-3 [32], and the adjacent gene M1404_01812 that encodes a surface lipoprotein assembly modulator (Slam) required for PmSLP-3 export, were both identified as important for M1404 in vivo fitness. Comparatively, other putative SLPs (PlpE and PmVP161_1666 in strain VP161, M1404_01814 in strain M1404), and the Slam exporter in VP161 (PmVP161_1663) were not important for in vivo fitness. Most of the genes important for in vivo fitness of M1404 that did not have a homolog in VP161 were required for production of the type B capsule polymer (Fig 3B). Genes required for in vivo fitness of VP161 but not for M1404 included genes encoding components of the ubiquinone biosynthesis pathway (ubiGHE_3X), a heme utilisation protein (hutZ_1), components of the aminodeoxychorismate synthase complex (pabAB), a putative sodium proton antiporter (PmVP161_0183), a putative tyrosine importer (tyrP_1), and aroE from the chorismate pathway (Fig 3A). All of the genes important for in vivo fitness in VP161 that did not have a homolog in M1404 were required for production of the type A capsule polymer (Fig 3C).

The genes alsT_1, crp and cyaA are important for strain M1404 fitness in vivo

We performed time to euthanasia experiments with individual P. multocida directed mutants to confirm that a sub-set of genes identified by TraDIS were individually important for in vivo survival. ClosTron mutagenesis was used to generate individual alsT_1, crp and cyaA mutants in strain M1404. Wild-type M1404 and each mutant strain was used to inoculate separate groups of BALB/c mice via intraperitoneal injection on two different days (see S1 Table for infectious doses). Signs of systemic infection were monitored and used to determine time to euthanasia. Mice infected with wild-type M1404 reached the humane endpoint on average 6.5 hours post infection (hpi) (Fig 4). Mice infected with the crp or cyaA mutants reached the humane endpoint at an average 11 and 10.3 hpi, respectively (Fig 4). Mice infected with the alsT_1 mutants did not reach the humane endpoint and showed low total clinical scores over the 12-hour infection period (Fig 4). All strains had a significant increase in time to euthanasia when compared to wild-type M1404, agreeing with the TraDIS data that alsT_1, crp and cyaA are important for in vivo fitness of strain M1404.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Kaplan-Meier survival curves for mice infected with P. multocida wild-type M1404 or an M1404 directed mutant.

Groups of six BALB/c mice (three male and three female) were inoculated via intraperitoneal injection with wild-type strain M1404, the M1404 alsT_1 mutant, the M1404 crp mutant, or the M1404 cyaA mutant (colony forming units ranged between 2 x 10⁴ and 2 x 10⁵). Statistical differences in survival between strains was determined using a log-rank Mantel-Cox test; *** P < 0.001.

https://doi.org/10.1371/journal.ppat.1013398.g004

The carbon catabolite repression system controls capsule production differently in P. multocida strains VP161 and M1404

Carbon catabolite repression was important for P. multocida fitness and is controlled by the cyclic-adenosine monophosphate (cAMP) receptor protein (CRP, also called catabolite activator protein (CAP) or virulence factor regulator (VFR) in other species). CRP was important for in vivo fitness of strain M1404, and essential for strain VP161 to grow in rich media. While CRP has not been well characterised in P. multocida, in other bacterial species CRP family proteins are global transcriptional regulators that have increased affinity for DNA targets when bound to cAMP [33]. Thus, CRP activity is modulated by the intracellular levels of cAMP. The level of intracellular cAMP is controlled by the adenylate cyclase CyaA, which synthesises cAMP from ATP [34], and the cAMP phosphodiesterase CpdA, which degrades cAMP to AMP [35]. As shown above, both crp and cyaA were identified as important for in vivo fitness of M1404 by TraDIS analysis and individual mutants in these genes showed reduced virulence in mice. Furthermore, the TraDIS data showed that inactivation of cpdA in M1404 slightly improved fitness during in vivo growth (average log₂ fold-change of between 0.93 and 1.34, see S1 Data). Collectively, these data suggest that cAMP-dependent activation of CRP is important for M1404 in vivo fitness. CRP family proteins are known to regulate a wide range of targets in different bacterial species [36–38], and transcriptomic analysis of a P. multocida strain 0818 crp mutant identified reduced capsule biosynthesis gene expression [39]. To determine if inactivating the catabolite repression system affected capsule production in M1404, capsule production was assessed using cells grown overnight on solid media. Both the M1404 crp and M1404 cyaA mutants harbouring empty vector produced barely detectable absorbance readings, indicating that these strains are essentially acapsular (Fig 5A). Providing the crp mutant with a wild-type copy of crp in trans led to a modest, but significant increase in absorbance readings, compared to the crp mutant harbouring empty vector (Fig 5A). Providing the cyaA mutant with the cyaA complementation plasmid restored absorbance readings, and therefore capsule production, to wild-type levels (Fig 5A). These data confirm that adenylate cyclase is required for capsule production in M1404 and strongly suggest that this is a CRP-dependent mechanism.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Capsule produced by VP161 and M1404 wild-type and mutant strains.

Capsule absorbance assays were performed by mixing capsule extracted from cells with stains-all solution and measuring the absorbance at 640 nm. A. Capsule measured from wild-type M1404 harbouring empty vector, the M1404 crp mutant and M1404 cyaA mutant harbouring either empty vector or the appropriate complementation plasmid. Capsule was extracted from cells recovered from overnight growth on heart infusion agar. B. Capsule measured from wild-type VP161 harbouring empty vector, and the VP161 cyaA mutant harbouring empty vector or an expression plasmid containing an intact copy of the VP161 cyaA gene. Capsule was extracted from cells grown in heart infusion broth to early stationary phase (7 h), and compared to known amounts of purified hyaluronic acid. For both assays, absorbance or hyaluronic acid was standardized to value per 1 mL of culture at an OD₆₀₀ of 1. Error bars represent mean ± standard error of the mean, and statistical differences in capsule production between samples were determined using Mann-Whitney U-tests; * P < 0.05.

https://doi.org/10.1371/journal.ppat.1013398.g005

The role of active CRP and catabolite repression in VP161 was different from that observed in M1404. While crp was essential for strain VP161 growth in rich media, the TraDIS data showed that inactivation of cyaA had no impact on in vivo fitness of VP161, and inactivation of cpdA slightly reduced fitness (average log₂ fold-change of between -1.75 and -1.93, see S1 Data). In our previous TraDISort study, cpdA was identified as important for in vitro capsule production in strain VP161 [31]. Collectively, these data suggest that active CRP-cAMP may act to repress capsule production in strain VP161. As CRP was essential in VP161, we attempted to inactivate cpdA and cyaA using ClosTron mutagenesis; however, only a cyaA mutant was successfully generated. Capsule absorbance assays were performed using wild-type VP161 and the cyaA mutant harbouring empty vector, and the complemented cyaA mutant. Capsule was extracted from cells during early stationary phase, when cAMP levels would be expected to be high. The VP161 cyaA mutant produced significantly more capsule than wild-type VP161, with complementation restoring normal levels of capsule production (Fig 5B). These data indicate that inactivation of adenylate cyclase, which should reduce the level of cAMP and therefore CRP activity, de-represses capsule production in VP161. This strongly suggests that an active CRP-cAMP represses capsule in strain VP161, which is the opposite of the role of CRP-cAMP in M1404.

Ethics statement

Animal experiments were performed according to Victorian State Government regulations, approved by the Monash University Animal Ethics Committee (Monash University AEC no. MARP/2023/36738).

Bacterial strains and growth conditions

The bacterial strains and plasmids used in this study are listed in S6 Table. E. coli strains were cultured in lysogeny broth (LB), and P. multocida strains were cultured in heart infusion (HI) broth (Oxoid). Solid media was generated by adding 1.5% agar (w/v) to liquid media. When required, antibiotics were added to the media at the following concentrations: kanamycin 50 µg/mL, spectinomycin 50 µg/mL, tetracycline 1.25 µg/mL, ampicillin 100 µg/mL. When required, diaminopimelic acid (DAP) was added to media at a concentration of 100 µM.

Oligonucleotides, DNA extraction, DNA manipulation and DNA sequencing

All oligonucleotides used in this study were synthesised by Sigma, and the sequences of all oligonucleotides are given in S7 Table. Genomic DNA was extracted using the HiYield Genomic DNA Mini Kit (Real Biotech Corporation) as per the manufacturer’s instructions for extraction from Gram-negative bacteria or extracted using Genomic Tip 100/G (QIAGEN) as per manufacturer’s instructions. Plasmids were extracted using the NucleoSpin Plasmid Mini Kit (Machery-Nagel) as per the manufacturer’s instructions. Polymerase chain reactions (PCR) were performed using either Taq DNA polymerase (Roche) or Phusion High-Fidelity DNA polymerase (New England Biolabs), as per the manufacturer’s instructions for each polymerase. DNA digests were performed using restriction enzymes sourced from New England Biolabs as per manufacturer’s instructions. Ligations were performed using T4 DNA ligase (Roche) as per manufacturer’s instructions. DNA was purified using the NucleoSpin Gel and PCR Clean-up Mini Kit (Machery-Nagel). Sanger sequencing reactions were performed as previously described [76]. For whole-genome sequencing of P. multocida strain M1404, genomic DNA was submitted to The Ramaciotti Centre for Genomics (University of New South Wales) and sequenced using both an Illumina MiSeq v3, and a Nanopore GridION X5 Mk1 using a GridION flow cell (FLO-MIN112) with DNA prepared for sequencing using a Native Barcoding Expansion Kit (SQK-NBD112.24). Additionally, M1404 genomic DNA was also submitted to Plasmidsaurus and sequenced using a Nanopore PromethION using a PromethION flow cell (FLO-PRO114M) with DNA prepared for sequencing using a Rapid Barcoding Kit (SQK-RBK114.96). The M1404 genome was assembled as previously described [7]; see S1 Text for additional information.

Transformation, site-directed mutagenesis using ClosTron and construction of complementation plasmids

Electrocompetent P. multocida cells were prepared and transformed as previously described [30]. P. multocida site-directed mutants were generated by a modified version of the ClosTron system (also called TargeTron) in the pAL953 plasmid as previously described [76,77] (see S1 Text for more detail on pAL953 modification). Intron insertions into each of the genes of interest were confirmed by PCR-amplification using primers that flanked the gene of interest, and by Sanger sequencing directly from genomic DNA extracted from putative mutants with oligonucleotide BAP6544, which binds within the intron and directs sequencing into the adjacent region of the genome. For complementation, the wild-type gene was amplified using Phusion polymerase and cloned into pAL99S under the control of the tpiA promoter. The sequence of the cloned gene and promoter were confirmed by Sanger sequencing. Complementation plasmids or pAL99S empty vector were used to transform the corresponding mutant strains to produce the complemented mutant and empty vector strains, respectively.

Conjugation conditions and generation of the P. multocida strain M1404 Himar1 transposon mutant library

Conjugation between an E. coli donor and P. multocida recipient was performed by filter mating as previously described [31], with minor modifications. The donor used in this study was E. coli strain JKE201 [78], a diaminopimelic acid (DAP) auxotrophic strain that cannot grow without the addition of DAP in the media. Filters containing the conjugation mix were incubated on HI agar plates containing DAP for one hour. The conjugation mix was recovered from the filter by vortexing the filter in 1 mL of HI broth. The conjugation mix was plated onto LB agar with ampicillin and DAP to recover donor E. coli, HI agar to recover all P. multocida recipients, and HI agar with spectinomycin to recover P. multocida Himar1 mutants. P. multocida Himar1 transconjugants were recovered from initial selection plates after overnight growth and replated on fresh HI agar containing spectinomycin to ensure selection against mutants with insertions in essential genomic regions, and to increase culture density for library storage. Following the second overnight growth, the P. multocida Himar1 mutants were recovered in glycerol broth (30% glycerol in HI broth) and stored at -80°C. A total of ~1.12 x 10⁶ M1404 Himar1 mutants were pooled to generate the final M1404 mutant library.

Mouse systemic infections

Each P. multocida systemic infection was performed with an equal number of male and female 6–10-week-old BALB/c mice (housed separately) supplied by Monash Animal Research Platform (Monash University), with mice housed in the Animal Research Laboratory (Monash University). P. multocida strains used for systemic infections were grown in HI medium to early-exponential phase (OD₆₀₀ ~ 0.25). Mice were each injected intraperitoneally using a 27-gauge needle with a 200 µL total volume containing one of the following: ~ 1 x 10⁷ CFU of the P. multocida VP161 transposon mutant library, ~ 1 x 10⁷ CFU of the P. multocida strain M1404 transposon mutant library, ~ 5 x 10⁴ CFU of wild-type strain M1404, or ~1 x 10⁵ CFU of one of the directed M1404 mutants (see S1 Table). Mice were monitored using a clinical score monitoring system for systemic infections in mice [79]; mice were monitored hourly and reached humane endpoints when they had a clinical score of 4 in a single category, or a cumulative clinical score of 10 or more. Mice were humanely euthanised using CO₂ asphyxiation followed by cervical dislocation. Blood samples were collected via cardiac puncture, and liver and spleen samples were recovered and homogenized. Tissue and blood samples were plated at the appropriate dilutions onto HI agar to obtain single colonies and determine the bacterial burden in the bloodstream or tissue. For TraDIS analysis, to ensure there were enough cells for genomic DNA extraction, blood and homogenized tissue samples were plated neat onto heart infusion agar and grown overnight before bacteria were recovered. Samples of the VP161 and M1404 Himar1 mutant libraries used to inoculate the mice were also plated neat onto HI agar in the same way as the bacteria recovered from mice. For DNA isolation, bacterial samples recovered from solid media using 1 x PBS were centrifuged at 13,000 x g then the cell pellet was resuspended 1 x PBS to wash cells before being centrifuged again as above. The cell pellets generated were then used to extract genomic DNA.

Transposon-directed insertion site sequencing library production, Illumina sequencing and bioinformatic analysis of TraDIS libraries

TraDIS libraries were prepared for Illumina sequencing as previously described [31]. TraDIS libraries were sequenced by The Ramaciotti Centre for Genomics at the University of New South Wales using an Illumina NextSeq1000 P1 with 100 bp-single end reads. TraDIS data was analysed using the Bio-TraDIS toolkit and modified scripts, and replicates compared by principal component analysis, as previously described [31]. Briefly, genes essential for growth in rich media were identified by comparing normalized UIS per gene excluding insertions in the final 10% of the gene. Insertion indexes for all genes were plotted on a histogram that generates a bimodal distribution. Two normal curves are drawn from this histogram, and the intersect between the two normal curves is taken to be the cut-off to determine essentiality. Genes important for fitness were identified by comparison of read count per gene between each mouse TraDIS library and the input rich medium TraDIS library. Genes with log₂ fold-change < -2 and a q-value < 0.01 were called as having a fitness cost, and genes with a log₂ fold-change > 2 and a q-value < 0.01 as having a fitness benefit. Full data sheets are provided as S1–S4 Data.

To identify protein homologs between strains, all coding sequences from P. multocida strains VP161 and M1404 were analysed using eggNOG mapper 5.0 [80] and then matched by seed ortholog value; proteins with the same seed ortholog value were taken as homologs. Proteins without a match in eggNOG and non-coding sequences were compared using the Basic Local Alignment Search Tool (BLAST) [81], with homologs called when there was > 90% amino acid/nucleotide identity over >90% of the query. Venn diagrams were generated using https://eulerr.co/. Heatmaps were generated in R using the ggplot2 package [82]. String was used to identify genes involved in the same pathway or processes, and to identify pathways and processes enriched in each dataset [83]. Roary analysis was performed using the VP161 and M1404 genomes, and an additional 167 complete P. multocida genomes obtained from the BV-BRC database (in April 2024). Pan-genome analysis and maximum-likelihood core-genome phylogenies were generated using Roary and IQ-TREE as previously described [7].

Capsule absorbance assays

For capsule extractions from VP161 and derivative strains, cells were cultured in liquid broth and samples taken to assess capsule production at early stationary phase (7 h, OD₆₀₀ ~ 2.2). For capsule extractions from M1404 and derivative strains, cells were cultured overnight on heart infusion agar and then recovered by adding 1 x PBS to the plate and scraping (OD₆₀₀~ 15). As purified polysaccharide equivalent to the P. multocida type B capsule is not commercially available to use as a standard, the absorbance values were normalised against the optical density values of the culture used for each capsule extraction. Capsule absorbance assays were performed as previously described [7]. Briefly, 10 µL of extracted capsule was mixed with 90 µL stains-all solution, added to a 96-well plate and the optical density at 640 nm measured using a Tecan Infinite 200 Pro. A standard curve was generated using purified hyaluronic acid (Sigma catalogue number H5388), allowing hyaluronic acid amounts recovered from VP161 strains to be determined. For both VP161 and M1404, hyaluronic acid or absorbance values were normalised by the optical density of the input culture used to give hyaluronic acid or absorbance per 1 mL of culture at an OD₆₀₀ of 1. Statistical comparisons between strains were performed using a Mann-Whitney U-test.