Attenuation mechanisms and vaccine potential of the serial passage–derived Pasteurella multocida strain PMZ8 in ducks

Abstract

Fowl cholera, caused by Pasteurella multocida, causes major losses in poultry worldwide, but its molecular pathogenesis remains poorly understood, limiting rational control. This study characterized an attenuated strain, PMZ8, obtained after 50 rounds of serial passaging from a highly virulent duck-derived isolate. Long-read whole-genome sequencing identified three fixed nonsynonymous mutations, in which Fis (R86S) and WaaF (G7R) contribute to virulence attenuation. PMZ8 exhibited an acapsular phenotype, a truncated LPS core, enhanced biofilm formation, defective envelope integrity, and increased intracellular ROS levels. Compared with the WT strain, PMZ8 presented at least a 10⁶-fold increase in the LD₅₀ via both intramuscular and oral routes. High-dose inoculation was well tolerated, indicating a favorable safety profile in ducklings. Vaccinated ducks developed significantly elevated anti–P. multocida IgG with strong bactericidal activity and serum IFN-γ production. Upon lethal homologous challenge, oral or intramuscular immunization with PMZ8 markedly reduced bacterial burdens and tissue lesions and conferred 85% or 80% protection, respectively. Collectively, our findings elucidate the capsule- and LPS-associated mechanisms underlying PMZ8 attenuation and support its further development as a live attenuated vaccine for the control of fowl cholera in ducks.

Introduction

Fowl cholera is a contagious and often acute disease of poultry caused by Pasteurella multocida (P. multocida), leading to septicemia with high morbidity and mortality (Wilson and Ho, 2013). The virulence of P. multocida is closely associated with its surface antigenic architecture, with the capsule and lipopolysaccharide (LPS) being the two principal determinants (Harper and Boyce, 2017). The typing of P. multocida involves two schemes: five capsule types (A, B, D, E, and F) defined by capsular polysaccharides and the 16-serovar Heddleston system based on LPS antigens (Heddleston et al., 1972; Peng et al., 2019). Fowl cholera is dominated by capsular type A strains, particularly serovars 1 and 3, and outbreaks result in significant economic losses in the poultry industry worldwide (Carpenter et al., 1988; Blackall et al., 1995; Kardos and Kiss, 2005). Effective vaccination is critical for fowl cholera control because antibiotic therapy is of limited use in fulminant cases and is increasingly constrained by antibiotic resistance and residue concerns (Ben et al., 2019). Currently available vaccines include adjuvanted killed bacterins and a naturally attenuated live vaccine (the Clemson University (CU) strain) (Layton, 1984). However, these traditional vaccines have notable limitations. Bacterins usually confer protection only against homologous serotypes, require parenteral injection, and can cause local injection-site reactions without inducing robust mucosal immunity. The live CU vaccine induces better respiratory mucosal and humoral immunity and can provide long-term protection, but there have been cases of incomplete protection and even disease outbreaks in vaccinated flocks (Ahmed et al., 1974). These drawbacks underscore the need for improved vaccines that are both safe and broadly protective (Mostaan et al., 2020).

The central roles of the capsule and LPS in virulence make them prime targets for rational vaccine design. The capsular polysaccharide enables the bacterium to evade phagocytosis and complement-mediated killing, and loss of the capsule in P. multocida serogroup A results in virtually complete abrogation of in vivo pathogenicity (Steen et al., 2010). Chung et al. (2001) analyzed the nucleotide sequence of the capsular biosynthesis locus and inactivated the gene encoding the capsular polysaccharide export protein hexA by inserting a tetracycline resistance cassette, thereby generating a serotype A:1 ΔhexA mutant. This mutant was avirulent in chickens, and intramuscular immunization of the natural host with the ΔhexA strain elicited high-level homologous protection (Chung et al., 2001, 2005). Similarly, previous work demonstrated that concurrent deletion of gatA and hptE yields the duck-derived attenuated strain PMZ2 with an outer-core–truncated LPS (Zhao et al., 2022). PMZ2 results in a 10⁵-fold increase in the LD₅₀ relative to that of the wild type, and following oral or intranasal immunization, it induces high IgG/IgA titers and potent serum bactericidal activity, resulting in 100% homologous protection (Zhao et al., 2022). These studies confirm that targeted knockout of capsule or LPS biosynthesis genes is a rational strategy for generating attenuated vaccine candidates (Harper et al., 2011).

While targeted genetic modification provides a rational approach, conventional attenuation through serial passaging of virulent strains under laboratory conditions remains a classical and effective method. This approach is well documented to generate genetically stable, attenuated strains of P. multocida suitable for vaccine development (Steen et al., 2010; Jiang et al., 2025).

In this study, we generated a markedly attenuated derivative of P. multocida, designated PMZ8, by performing 50 rounds of single-colony serial passaging from the WT virulent strain. Whole-genome sequencing identified three fixed mutations in PMZ8 relative to the parent: Fis R86S, the hypothetical protein T149S, and WaaF G7R. To assess its suitability as a live-attenuated vaccine, we systematically evaluated its safety profile and homologous protective efficacy in a duck model and investigated attenuation mechanisms through phenotypic and molecular assays. These findings support PMZ8 as a safe, immunogenic candidate for fowl cholera control and provide additional insight into the pathogenesis of P. multocida.

Materials and methods

Bacterial strains and growth conditions

All bacterial strains and plasmids used in this study are summarized in Table 1. The mutant strains were derived from the wild-type (WT) P. multocida PM0818, which belongs to the LPS genotype L1 (Zhao et al., 2021a). P. multocida was routinely maintained at 37°C in brain–heart infusion (BHI) broth or on tryptic soy agar (TSA, Difco Laboratories, USA) supplemented with 5% defibrinated sheep blood . The Escherichia coli (E. coli) strains used for plasmid construction were cultured in Luria–Bertani (LB) broth or on LB agar (Coolaber, Beijing, China). For enumeration of P. multocida viable counts, TSA supplemented with 5% defibrinated sheep blood was employed. When antibiotic selection was needed, the media were supplemented with kanamycin at 50 µg/mL or chloramphenicol at 25 µg/mL. Unless otherwise indicated, the cultures were incubated at 37°C under standard aerobic conditions.

Table 1.

Bacterial strains and plasmids used in this study.

Strains or plasmids	Description	Source
Plasmids
pMC-Express	A broad host-range shuttle vector derived from pMIDG100, sodC promoter, Cmr	(Bossé et al., 2009)
pCZb5	pMC-Express derivative with a 250 bp fragment containing the tpiA promoter	(Zhao et al., 2022)
pCZb5-waaF	Insertion of complete waaF into pCZb5	This work
pRE112	sacB mobRP4 R6K ori Cmr	(Edwards et al., 1998)
Strains
PM0818	P. multocida 0818, Wild-type and virulent, LPS genotype L1	(Zhao et al., 2022)
SM10 λ pir	E. coli thi thr-1 leu6 proA2 his-4 arg E2 lacY1 galK2, ara14xyl5 supE44, λpir	(Rubirés et al., 1997)
PMZ8	Derived from PM0818 by 50 rounds of single-colony serial passage	This work
PMZ9	PM0818 ΔwaaF::kanR	This work
PMZ9-pCZb5-waaF	PMZ9 harboring the pCZb5-waaF	This work
PMZ10	PM0818 ΔA0R64_04620::kanR	This work

Long-read WGS, hybrid assembly, and whole-genome variant analysis

High-molecular-weight genomic DNA from PMZ8 was sequenced with Oxford Nanopore long reads and matched short reads from a DNBSEQ platform by BGI Genomics (Shenzhen, China). Long reads were adapter-trimmed and quality-filtered; short reads were cleaned to remove adapters/low-quality bases. The genomes were assembled via a long-read–first strategy and polished with short reads to single-base accuracy; the assembly quality was verified via read-mapping and continuity metrics. The polished PMZ8 assembly was then compared with the WT genome for single nucleotide polymorphism (SNPs)/indels discovery by whole-genome alignment, with variants curated by read evidence and annotated on predicted coding sequences.

Genetic stability assessment by serial passaging and in vivo back-passaging

To evaluate the genetic stability of PMZ8 and the risk of reversion, the strain was subjected to seven serial passages in vitro and three consecutive back-passages in ducks. After each passage, the isolates were recovered by plating, and a single colony was selected for genomic DNA extraction. Briefly, locus-specific primers flanking three SNPs were used to amplify the corresponding fragments, and the resulting PCR products were sent to Sangon Biotech (Shanghai, China) for bidirectional Sanger sequencing. Sequencing chromatograms were inspected and aligned to the WT reference sequences.

Plasmid and mutant strain construction

For derivation of PMZ8 by serial single-colony passaging, the attenuated derivative PMZ8 was obtained from the virulent WT strain PM0818 by fifty consecutive rounds of single-colony transfer on TSA supplemented with 5% defibrinated sheep blood at 37°C under aerobic conditions. During each cycle, one well-isolated colony with a typical morphology was restreaked onto fresh plates and incubated for 18–24 h; glycerol stocks (15%, v/v) were archived every 10 passages. After passage 50, a single colony was purified twice and designated PMZ8.

To construct the PMZ9 mutant, an allelic-exchange method employing the sacB suicide vector pRE112(Edwards et al., 1998) was used. The primer sequences and oligonucleotide details are provided in Table S1. The upstream (485 bp) and downstream (468 bp) amplified from the PM0818 chromosome with a high-fidelity DNA polymerase were subsequently cloned with the primer pair waaF-Up-F/R or waaF-Down-F/R. The kanamycin resistance cassette (kanR, ∼0.85 kb; amplified from pET28a) was assembled between waaF-Up and waaF-Down by seamless cloning via a Seamless Cloning Kit (Sangon Biotech, Shanghai, China), yielding a contiguous Up-kanR-Down fragment. This fragment was inserted into pRE112 linearized by seamless cloning to produce pRE112-ΔwaaF::kanR. The construct was propagated in E. coli SM10 λpir (Rubirés et al., 1997) and transferred into P. multocida by conjugation. The mutant strain designated PMZ9 was selected on TSA supplemented with 5% defibrinated sheep blood containing kanamycin. For complementation, the waaF gene was PCR-amplified with the primers waaF-F/R and inserted into pCZb5, yielding pCZb5-waaF. The recombinant plasmid was then introduced into the PMZ9 background to generate the complemented strain (PMZ9-pCZb5-waaF). The same cloning and exchange framework was applied to construct other gene deletions and their complements.

Quantitative real-time RT-PCR (qRT-PCR)

The primers used are listed in Table S2. Total RNA was extracted from bacterial pellets, treated with RNase-free DNase I, and reverse-transcribed via a HiScript II kit (Vazyme, Nanjing, China). qPCR was performed with SYBR Green master mix in 20 µL reactions (2 µL cDNA; 50 nM each primer), run in triplicate. gyrB served as the internal reference. Specificity was verified by single-peak melting curves. Relative transcript levels were calculated by the 2^–ΔΔCt method with PM0818 used as the calibrator.

Capsule quantification

Capsule-associated dye binding was assessed on BHI agar supplemented with 0.005% Congo red. Mid-log cultures of WT and PMZ8 were normalized to OD₆₀₀, spotted or spread onto plates, and incubated overnight at 37°C. The colony biomass was harvested into preweighed microtubes containing 1 mL of PBS, briefly pelleted, and the wet mass was recorded. The bound dye was extracted by resuspending the pellets in 1% SDS in PBS (1 mL, 10 min at room temperature), and the supernatant absorbance was read at 490 nm in a microplate reader. Congo red absorption was expressed as A₄₉₀ per unit wet mass (A₄₉₀/g). Capsular uronic acid was quantified via the sulfuric acid–carbazole method following Method 1 of a published protocol (Bitter and Muir, 1962), with minor adaptations. Briefly, mid-log cultures of WT and PMZ8 plants were normalized to the OD₆₀₀ of the corresponding cultures, which were then pelleted and resuspended in PBS. Capsules were released by treating the suspensions with hyaluronidase and then clarified by brief centrifugation. For the assay, chilled concentrated H₂SO₄ was added to 1 mL of supernatant (or glucuronic acid standards), and the mixtures were gently mixed. The mixtures were heated in a boiling water bath for ∼10 min, cooled to room temperature, supplemented with carbazole reagent, reheated for ∼15 min, and cooled again. The absorbance was read at 530 nm in a microplate spectrophotometer. Uronic acid concentrations were interpolated from the standard curve and normalized to biomass.

Membrane permeability assays

Each strain was grown in BHI at 37°C, diluted 1:100, and harvested at mid-log; the cells were subsequently washed and normalized to the OD₆₀₀. N-phenyl-1-naphthylamine (NPN) uptake was measured as previously described with minor modifications (Ma et al., 2020): cells were resuspended in 5 mM HEPES/5 mM glucose to an OD₆₀₀ of approximately 0.2, NPN was added to 10 µM, the mixture was incubated for 15 min in the dark, and fluorescence was recorded (Ex 350 nm/Em 420 nm) in black plates. Propidium iodide (PI) uptake was assessed by incubating PBS-suspended cells (OD₆₀₀≈0.2) with PI (5 µg/mL, 15 min, dark) and reading the wavelength at Ex 535 nm/Em 617 nm. Extracellular nucleic acids were quantified by pelleting cultures (5,000 × g, 10 min), filtering supernatants (0.22 µm), and reading A₂₆₀ in UV-transparent 96-well plates with medium blanks.

LPS phenotyping by Tricine–SDS–PAGE and ammonium silver staining

The LPS profiles of P. multocida strains (WT, PMZ8, and PMZ9; complemented derivatives where indicated) were analyzed via a tricine–SDS–PAGE system followed by ammonium silver staining, with slight modifications to previously described procedures (Kittelberger and Hilbink, 1993; Zhao et al., 2022). Briefly, cultures grown in BHI at 37°C to mid-log phase were washed in PBS and boiled in SDS sample buffer, and equal bacterial equivalents were loaded onto 15% polyacrylamide Tricine gels. After electrophoresis, the gels were silver-stained with an ammonium silver reagent and developed in citrate–formaldehyde, then rinsed and imaged under identical settings.

Determination of the minimum inhibitory concentration (MIC) of polymyxin B

The MICs of polymyxin B were determined via a broth microdilution assay. Briefly, overnight cultures were diluted in BHI broth and inoculated into 96-well plates containing twofold serial dilutions of polymyxin B. Plates were incubated at 37°C for 16–18 h, and the MIC was defined as the lowest concentration that prevented visible growth.

Transmission electron microscopy (TEM)

Single colonies of WT and PMZ8 were grown overnight in BHI at 37°C, diluted 1:100 into fresh BHI, and cultured to the mid-log phase. The cells were washed twice in sterile PBS, and fixed in 2.5% glutaraldehyde for 2 h at 4°C. After three rinses with PBS, the samples were fixed in 1% osmium tetroxide for 1 h, dehydrated through a graded ethanol (or acetone) series, infiltrated with Epon 812 resin, and embedded. Ultrathin sections (∼70 nm) were cut with a diamond knife, stained with uranyl acetate followed by lead citrate, and examined via a JEM-1400-FLASH TEM (JEOL, Japan) at 80 kV.

Liquid chromatography–mass spectrometry (LC-MS)

Strains were grown in BHI at 37°C to an OD₆₀₀ of∼0.8, rapidly chilled, pelleted, and quenched in liquid N₂. Metabolites were extracted from pellets with H₂O:acetonitrile:isopropanol (1:1:1, v/v/v), sonicated for 30 min at 4°C, and clarified (12,000 g, 20 min, 4°C). The supernatants were analyzed via LC–HRMS on a Q Exactive HFX (Thermo, USA) with HESI in full MS–ddMS² mode (m/z 70–1050; MS¹ resolution 70,000; MS² 17,500; typical source: sheath 40, aux 10, +3.0/−2.8 kV, 350°C/320°C). Raw files (Xcalibur 4.1) were processed in Progenesis QI for peak picking/alignment and normalization; features were exported for statistical analysis (see Statistics). Metabolites were annotated by accurate mass and MS/MS spectral matching against reference databases.

Measurement of intracellular reactive oxygen species (ROS)

The WT and PMZ8 strains were cultured overnight. One milliliter of fresh bacterial suspension was prepared according to the instructions of an ROS detection kit (BB-46111, BBcell, Shanghai, China). The bacterial cells were harvested via centrifugation, washed 2-3 times with PBS, and resuspended in 500-1000 μL of the staining solution. The suspension was incubated at 37°C in the dark for 30–45 minutes. After incubation, the bacterial cells were collected by centrifugation, washed once with PBS, and resuspended in 500 μL of PBS. 200 μL of the bacterial suspension was transferred to a black 96-well plate, and the fluorescence intensity was measured via a microplate reader (excitation wavelength: 488 nm; emission wavelength: 526 nm).

Virulence and bacterial colonization of P. multocida strains in ducks

To compare virulence among PMZ8, PMZ9 and PMZ10, groups of 7-day-old ducklings (n = 10 per group; sourced from Grimaud Breeding Co., Ltd. (Chengdu, China)) were intramuscularly challenged with the indicated strain. The animals were monitored at least twice daily for 7 days; mortality and humane-endpoint criteria were recorded to generate survival curves. In parallel, the LD₅₀ values for WT and PMZ8 were determined by challenging separate cohorts with tenfold serial dilutions via the oral or intramuscular route and calculating the LD₅₀ via the Reed–Muench method as previously described (Zhao et al., 2021b) over 14 days. For bacterial burdens, additional ducks were challenged orally with 10¹⁰ colony-forming units (CFUs) or intramuscularly with 10⁷ CFUs and the blood, liver, and spleen were sampled at 24 h for CFU enumeration. To determine when PMZ8 was completely cleared, additional PMZ8-inoculated ducklings were sampled on successive days after inoculation, and blood, liver, and spleen samples were collected and plated for CFU counting until no bacteria were detected.

Safety assessment in ducklings

Seven-day-old ducklings were randomly assigned to receive PMZ8 or PBS via one of two routes to evaluate tolerability at high inoculation doses. Ducklings in the oral group were fed 10¹¹ CFU PMZ8; ducklings in the intramuscular group were injected with 1 × 10⁸ CFU PMZ8 into the thigh muscle. Route-matched PBS recipients served as controls. Clinical signs, rectal temperature, and body weight were recorded once daily for 14 days at a consistent time.

Short-term environmental viability and −80°C recovery assay

WT and PMZ8 were incubated at 4°C, 25°C, or 37°C. At 12, 24, 36, and 48 h, the samples were serially diluted and plated to determine viable counts (CFUs). For −80°C storage, bacterial aliquots were frozen at −80°C and then thawed; viable counts were determined by plating and compared between strains to assess recovery.

Immunization and challenge

PMZ8 vaccine inoculates were prepared by overnight static culture at 37°C in BHI, followed by 1:100 subculture in fresh BHI to the mid-logarithmic phase. Bacteria were pelleted, washed, and resuspended in sterile PBS to the target dose; inoculum concentrations were verified via plate counts. Seven-day-old ducklings were randomly assigned (32 ducks per group) to receive PMZ8 either by oral 10¹⁰ CFU or intramuscular injection 10⁷ CFU on day 1, with an identical booster given on day 14. Route-matched PBS recipients served as negative controls. Fourteen days after boosting (day 28), six ducks per group were bled via the wing vein; blood was centrifuged at 1,500 × g for 10 min at 4°C, and serum was stored for antibody assays. The remaining ducks were challenged on day 28 with the WT strain administered intramuscularly at 100 × LD₅₀. At 24 h post-challenge, six ducks per group were euthanized for blood, spleen, and liver sampling to quantify bacterial loads and assess histopathology, while the remaining 20 ducks per group were monitored for survival for 14 days.

Enzyme‐linked immunosorbent assay (ELISA)

Antibody responses to P. multocida were quantified via indirect ELISA with minor adaptations (Xu et al., 2025). Heat-killed PM0818 whole cells (10⁵ CFU/mL in carbonate–bicarbonate buffer) served as coating antigens (100 µL/well, high-binding plates, 4°C overnight). The plates were washed with PBS-0.05% Tween-20 (Amresco, USA), blocked with 5% BSA in PBS, and then incubated with sera diluted 1:200 in 1% BSA in PBS. HRP-conjugated anti-duck IgG (KPL,USA) diluted 1:10000 was applied, TMB was developed for 10 min at room temperature, the reactions were stopped with 2 M H₂SO₄, and the OD₄₅₀ was recorded. Serum interferon-gamma (IFN-γ) was quantified via a commercial duck IFN-γ ELISA kit (Shfksc, Shanghai, China) following the manufacturer’s instructions with a standard curve.

Serum bactericidal assay

Serum bactericidal activity against P. multocida was evaluated with minor adaptations to published methods (Zhao et al., 2017). The duck sera were clarified and either kept complement-active or heat-inactivated at 56°C for 30 min. The mid-log PM0818 was washed and adjusted to ∼10³ CFU/mL in PBS. Mixtures containing 90% serum and 10% bacterial suspension were incubated at 37°C for 2 h with gentle mixing, then serially diluted and plated on TSA overnight incubation at 37°C to enumerate the CFUs. Relative survival was expressed as the number of CFUs recovered after incubation in complement-active serum relative to the paired heat-inactivated serum. For serum sensitivity, mid-log cultures of the indicated strains were incubated with normal duck serum (complement-active) in parallel with heat-inactivated serum, and survival was calculated as described above.

Histopathological lesions and scoring criteria after challenge

At 24 h post-challenge, ducks assigned to pathology were humanely euthanized according to institutional guidelines. The liver and spleen were aseptically excised, trimmed (≤5 mm thickness), and fixed in 4% paraformaldehyde at 4°C for 24–48 h. After rinsing, the tissues were processed through graded ethanol, cleared in xylene, infiltrated with paraffin, and embedded. Sections (5 µm) were cut on a rotary microtome (Leica RM2235, Nussloch, Germany), floated in a warm water bath, mounted on glass slides, and dried. The slides were deparaffinized, rehydrated, stained with hematoxylin–eosin (H&E), dehydrated, cleared, and cover slipped with neutral resin.

Lesions were scored via a 0–4 semi-quantitative scale (0 = none/trace, 1 = mild, 2 = mild–moderate, 3 = moderate, 4 = severe), which was applied to predefined domains. Liver domains: inflammatory infiltration, necrosis, congestion/hemorrhage, fibrin/thrombi, and lobular injury. Spleen domains: white-pulp depletion, inflammatory infiltration, congestion/hemorrhage, necrosis, and fibrin/thrombi. For each section, the entire slide was surveyed at low power to identify the most affected areas; domain scores were assigned according to the most severe lesion observed. A composite score per organ was defined as the maximum of the domain scores. The data are presented as the median (IQR) per group.

Ethics statement

All procedures complied with the Guide for the Care and Use of Agricultural Animals in Research and Teaching and the Ministry of Science and Technology of China guidelines and were approved by the Animal Ethics Committee of Sichuan Agricultural University and the Sichuan Administration Committee of Laboratory Animals (permit SYXK2019-187). For terminal sampling, ducks were humanely euthanized by intravenous barbiturate overdose in accordance with regulations. The animals were maintained under standard husbandry, randomly assigned, and monitored daily, with efforts to minimize numbers and distress.

Statistical analysis

All the statistical analyses were performed in GraphPad Prism (GraphPad Software, CA, USA). Unless otherwise stated, values are reported as the mean ± standard deviation (SD) from independent biological replicates. Two-group comparisons were performed via two-tailed unpaired Student’s t tests; multiple-group comparisons were performed via one-way ANOVA with Tukey’s post hoc test. Survival data were analyzed via the log-rank test. Two-sided P values < 0.05 were considered statistically significant.

Results

Serial passage drives capsule attenuation and a biofilm-prone phenotype in PMZ8

After fifty rounds of single-colony serial passaging of the WT strain, the colony morphology shifted from smooth/mucoid to dry and granular (Fig. 1A), suggesting capsule loss (Carter and Bigland, 1953). Long-read whole-genome sequencing revealed three missense substitutions relative to the parent protein—Fis R86S, the hypothetical protein (A0R64_04620) T149S, and WaaF G7R (Table 2). To assess genetic stability and the risk of reversion, PMZ8 was further passaged seven times in vitro and back-passaged three consecutive times in ducks; sequencing of these SNP loci in recovered isolates revealed no reversion or additional mutations (Fig. S1). The WT strain PM0818 belongs to serogroup A, whose capsule is composed of hyaluronic acid (Guan et al., 2020). Because the capsular polysaccharide impedes the binding of hydrophobic dyes, we quantified the capsule by measuring hyaluronic acid (indexed by glucuronic acid) and by a Congo red binding assay. PMZ8 markedly reduced the glucuronic acid content (Fig. 1B), and increased Congo red uptake (Fig. 1C), indicating substantially diminished capsule formation. Consistently, qRT-PCR of the A-capsule loci (phyAB, hyaBCD, and hexABCD) revealed downregulation of key transcripts (Fig. 1D), which is in line with the reported Fis-dependent control of capsule biosynthesis (Steen et al., 2010). In addition, as capsular polysaccharides can antagonize biofilm formation in P. multocida (Petruzzi et al., 2017), crystal violet assays demonstrated significantly enhanced biofilm formation by PMZ8 compared with the WT strain (Fig. 1E). Overall, these data indicate that capsule biosynthesis is suppressed in PMZ8 and that the strain adopts a more biofilm-prone phenotype.

Fig. 1.

Serial passage drives capsule attenuation and a biofilm-prone phenotype in PMZ8. (A) Colony morphology on TSA supplemented with 5% defibrinated sheep blood after 50 single-colony passages. (B) Capsular hyaluronic acid quantified by the sulfuric acid–carbazole uronic acid assay from hyaluronidase-treated extracts (glucuronic acid standard). (C) Congo red binding to BHI + 0.005% Congo red, dye extracted with 1% SDS; uptake reported as A₄₉₀ per g wet biomass. (D) qRT-PCR of capsule loci. Expression was normalized to that of gyrB and calculated via the 2^–ΔΔCt method relative to that of the WT. (E) Biofilm biomass determined via 24-h crystal-violet staining in 96-well plates. ** P < 0.01, *** P < 0.001, **** P < 0.0001.

Table 2.

Nonsynonymous SNPs identified in PMZ8 by long-read whole-genome sequencing relative to the WT strain.

CHROM	POS	TYPE	REF	ALT	EVIDENCE	FTYPE	NT_POS	AA_POS	EFFECT	LOCUS_TAG	GENE	PRODUCT
CP059702	67683	SNP	C	A	A:20 C:0	CDS	256/300	86/99	Arg86Ser	A0R64_00370	fis	DNA-binding transcriptional regulator Fis
CP059702	1020142	SNP	C	G	G:20 C:0	CDS	446/1272	149/423	Thr149Ser	A0R64_04620		hypothetical protein
CP059702	1511898	SNP	G	A	A:20 G:0	CDS	19/1047	7/348	Gly7Arg	A0R64_06900	waaF	lipopolysaccharide heptosyltransferase II

WaaF-dependent LPS core integrity contributes to serum resistance and virulence

Tricine–SDS–PAGE silver staining revealed an LPS core-truncation pattern in PMZ8 that mirrors that of PMZ9 (ΔwaaF) (Fig. 2A). The WT showed three dominant bands consistent with a mature core LPS, whereas PMZ8 carrying the WaaF G7R substitution and PMZ9 (ΔwaaF) each displayed faster-migrating bands, indicative of accumulated truncated core glycoforms (Fig. 2A). The complementation of ΔwaaF restored the WT-like pattern (Fig. 2A). These data indicate that the LPS changes in PMZ8 are caused mainly by WaaF G7R mutation. The ducklings infected with the WT strain died by day 2, whereas those challenged with PMZ8 or PMZ9 (ΔwaaF) showed 100% or 80% survival at 7 days (Fig. 2B). Additionally, PMZ10 (ΔA0R64_04620) presented a survival phenotype similar to that of WT strain, suggesting that the attenuation of PMZ8 is attributable to the combined effects of the fis and waaF lesions rather than the deletion of A0R64_04620.

Fig. 2.

WaaF is required to maintain LPS core integrity in PMZ8 and contributes to virulence. (A) Tricine–SDS–PAGE followed by ammonium silver staining of LPS from lane 1, WT (PM0818); lane 2, PMZ8 (WaaF G7R); lane 3, PMZ9 (ΔwaaF); and lane 4, PMZ9-pCZb5-waaF (complemented ΔwaaF). (B) Survival rates of ducklings (n = 10/group) challenged with WT, PMZ8, PMZ9 or PMZ10 for 7 days. (C) Mid-log bacteria were incubated with 90% duck serum at 37°C for 2 h. Survival is expressed as the number of CFUs recovered after incubation in complement-active serum relative to the paired heat-inactivated serum. * P < 0.05, ** P < 0.01,*** P < 0.001, **** P < 0.0001.

To assess susceptibility to cationic peptide stress and duck serum–mediated innate killing, we measured the polymyxin B MIC and performed a duck serum killing assay. PMZ8 presented a markedly lower polymyxin B MIC than did the WT, and PMZ9 (ΔwaaF) also displayed increased sensitivity, whereas waaF complementation restored resistance to WT levels (Table 3). In addition, compared with those in WT duck serum, the survival of PMZ8 and PMZ9 (ΔwaaF) in normal duck serum was reduced, whereas heat-inactivated serum abolished killing, and waaF complementation rescued survival (Fig. 2C). Collectively, these results indicate that WaaF-dependent LPS core integrity is important for resistance to polymyxin B and for duck serum–mediated killing.

Table 3.

Polymyxin B MICs of P. multocida strains.

Strain	Drug (µg/mL)
	Polymyxin B
WT	64
PMZ8	8
PMZ9 (ΔwaaF)	16
PMZ9-pCZb5-waaF	64

PMZ8 exhibits increased envelope permeability

Truncation of the LPS core and loss of the capsule are expected to compromise the bacterial envelope, increasing membrane permeability. Accordingly, we examined ultrastructure by TEM and quantified permeability via NPN uptake, PI uptake, and extracellular nucleic acid release. TEM revealed discontinuities in the outer envelope with surface irregularities and cytoplasmic leakage in PMZ8 relative to the WT strain (Fig. 3A). Consistently, NPN uptake increased, indicating greater outer-membrane permeability (Fig. 3B); extracellular nucleic acid levels in the cell-free supernatants increased, which was consistent with leakage (Fig. 3C); and PI uptake increased, indicating compromised inner-membrane integrity (Fig. 3D). These results demonstrate a marked increase in the envelope permeability of PMZ8 compared with that of WT strain.

Fig. 3.

PMZ8 exhibits increased envelope permeability. (A) TEM image of mid-log cells; scale bars are as indicated. (B) NPN uptake: cells were washed in 5 mM HEPES/5 mM glucose, adjusted to an OD₆₀₀≈0.2, incubated with NPN and read at Ex 350 nm/Em 420 nm. (C) Extracellular nucleic acids: cultures were pelleted (5,000 × g, 10 min), 0.22-µm-filtered supernatants and A₂₆₀ measured in UV-transparent plates. (D) PI uptake: cells in PBS (OD₆₀₀≈0.2) were incubated with propidium iodide (5 µg/mL, 15 min, dark) and read at Ex 535 nm/Em 617 nm. ** P < 0.01, **** P < 0.0001.

Oxidative stress and metabolic alterations in PMZ8

Untargeted LC-MS revealed 357 metabolites in PMZ8, with increased levels of oxidized glutathione (GSSG) and ophthalmate (Fig. 4A), indicative of oxidative stress (Schomakers et al., 2024). qRT-PCR revealed downregulation of the expression of antioxidant genes (gshAB, glutathione reductase, ahpC, kat, and sodA), confirming impaired antioxidant capacity (Fig. 4B). Compared with those in the WT strain, the ROS levels in the PMZ8 strain were significantly greater (Fig. 4C). UDP-d-glucuronate, a precursor of the A-type capsule, was markedly reduced, reflecting capsule attenuation (Fig. S2). Despite elevated serA/serB transcript levels, the levels of serine and its derivatives were notably reduced (Fig. 4D), suggesting altered metabolic flux. Overall, PMZ8 exhibits reprogramming associated with oxidative stress, capsule precursor depletion, and disrupted serine metabolism.

Fig. 4.

PMZ8 shows redox imbalance with suppression of serine metabolism. (A) Volcano plot of LC–MS features with differentially abundant metabolites defined as those with a fold change (FC) > 2 and p < 0.05. (B) and (D) qRT-PCR analysis of genes involved in oxidative-defense and serine pathway genes. Expression was normalized to gyrB and calculated via the 2^–ΔΔCt method relative to WT. (C) Intracellular ROS measurement * P < 0.05, *** P < 0.001.

Measurements of the bacterial loads of PMZ8 in ducklings

Route-specific LD₅₀ values of PMZ8 for ducklings were estimated via dose–response analysis. After intramuscular challenge, the WT strain had an LD₅₀ < 8 CFU, whereas the LD₅₀ of PMZ8 was > 8.5 × 10⁷ CFU. After oral challenge, the LD₅₀ values were < 2.6 × 10⁴ CFU for PM0818 and > 8.5 × 10¹⁰ CFU for PMZ8, indicating ≥10⁶-fold attenuation via both routes (Table 4). To quantify the in vivo bacterial burdens, ducklings were inoculated intramuscularly (10⁷ CFU) or orally (10¹⁰ CFU) and sampled at 24 h. In the blood, liver, and spleen, PMZ8 yielded markedly lower bacterial burdens than did the WT strain via both routes, with a consistent reduction in all tissues (Fig. 5A and 5B). Notably, after oral infection, PMZ8 was only detectable at trace levels in the blood, and following intramuscular challenge it persisted only at low levels (Fig. 5A and 5B). Bacterial persistence in hepatic and splenic tissues was observed, suggesting that PMZ8 remains present in vivo enough to drive protective immunogenicity while remaining highly attenuated.

Table 4.

Determination of the LD₅₀ of the WT and PMZ8.

Route	Strains	Challenge dose (CFU) and survival											LD50 (CFU)
		<10	101	102	103	104	105	106	107	108	109	1010
Intramuscular	WT	0/10	0/10	0/10	0/10	0/10							<8
	PMZ8						10/10	10/10	10/10				>8.5 × 107
	PBS	10/10
Oral	WT					4/10	2/10	0/10	0/10	0/10	0/10		<2.6 × 104
	PMZ8								10/10	10/10	10/10	10/10	>8.5 × 1010
	PBS	10/10

Fig. 5.

Measurements of bacterial loads of PMZ8 in ducklings. Seven-day-old ducklings were challenged with WT PM0818 or PMZ8 either orally at 10¹⁰ CFU (A) or intramuscularly at 10⁷ CFU(B). At 24 h post-infection, the liver and spleen were homogenized, and blood was collected for CFU enumeration via TSA. The data are shown as log10 CFU/g (tissues) or CFU/mL (blood). ** P < 0.01, *** P < 0.001, **** P < 0.0001. Clearance of PMZ8 from ducklings after oral (C) or intramuscular (D) inoculation.

To examine how long PMZ8 persists in vivo, we tracked bacterial counts in the blood, liver, and spleen over time. After oral inoculation, the PMZ8 counts decreased rapidly and were no longer detected in the sampled tissues by days 7–8 (Fig. 5C). After intramuscular inoculation, clearance was slightly delayed, with PMZ8 becoming undetectable by days 8–9, and the liver and spleen remained positive longer than the blood (Fig. 5D).These findings indicate that PMZ8 is cleared within ∼1 week after inoculation.

Safety evaluation of PMZ8

To assess the biosafety of PMZ8, 7-day-old ducklings were immunized with PMZ8 either orally (1 × 10¹¹ CFU) or intramuscularly (1 × 10⁸ CFU), with PBS as a control, and monitored for 14 days. A transient, low-grade rise in body temperature was noted on days 1–3 (<0.5°C versus PBS), and the body temperature returned to baseline by day 4; thereafter, the temperature trajectories overlapped with those of the controls (Fig. 6A). No abnormal clinical signs or deaths were observed. Body weight increased similarly in both groups through day 9 (from ∼0.2 to ∼0.4 kg), after which the PMZ8 group tended to be slightly heavier (Fig. 6B). These data indicate a favorable safety profile for PMZ8 via both oral and intramuscular routes, even at high inoculation doses. Despite the elevated intracellular ROS level and compromised envelope integrity observed in PMZ8, its short-term environmental viability was comparable to that of the parental WT strain. Specifically, the CFU trajectories of PMZ8 were similar to those of the WT at 4°C, 25°C, and 37°C over 12–48 h, and comparable recoveries were observed after storage at −80°C (Fig. 6C).

Fig. 6.

Safety evaluation of PMZ8. Seven-day-old ducklings were vaccinated with PMZ8 either orally or intramuscularly; route-matched PBS controls were included. Rectal temperature (A) and body weight (B) were measured daily for 14 days post-vaccination. Each point represents an individual animal. (C) Bacteria were washed and resuspended in PBS, then incubated at 4°C, 25°C, or 37°C, or stored at −80°C and subsequently allowed to recover. Viable counts were determined by CFU enumeration at 0–48 h.

PMZ8 immunization stimulates antigen-specific IgG and IFN-γ and enhances complement-dependent serum bactericidal activity

Seven-day-old ducklings were immunized with PMZ8 or PBS via a prime–boost regimen with an interval of 14 days. To assess post-vaccination immunity, booster sera against whole-bacteria antigens were analyzed by indirect ELISA; the amount of IFN-γ in the serum was also quantified. Compared with the PBS control, PMZ8 immunization—via either the oral or intramuscular route—produced significantly more anti-P. multocida IgG (Fig. 7A) and increased IFN-γ (Fig. 7B), indicating effective induction of humoral and cell-mediated responses. In a complement-dependent serum bactericidal assay, booster sera from PMZ8 orally or intramuscularly immunized ducks markedly reduced the survival of wild-type PM0818 compared with that of PBS-immunized ducks (Fig. 7C). These data show that PMZ8 elicits a strong antigen-specific antibody response with functional complement-mediated bactericidal activity, accompanied by increased IFN-γ production, indicating high immunogenicity of the PMZ8.

Fig. 7.

PMZ8 immunization elevates the levels of antigen-specific IgG and IFN-γ and enhances complement-dependent serum bactericidal activity. Seven-day-old ducklings were immunized with PMZ8 by oral gavage (p.o., 10¹⁰ CFU) or intramuscular injection (i.m., 10⁷ CFU) and boosted on day 14; PBS served as a control. Sera collected 14 days post-boost immunization (n = 6/group) were subjected to anti-P. multocida IgG response measurement (A) via indirect ELISA using heat-killed PM0818 whole cells as the coating antigen, IFN-γ production quantification (B) with a commercial duck IFN-γ ELISA kit, and a serum bactericidal assay (C) via survival calculation after incubation with active serum or heat-inactivated serum incubation for 2 h at 37°C. ** P < 0.01, *** P < 0.001, **** P < 0.0001.

PMZ8 elicits robust homologous protection against lethal challenge in ducks

To evaluate the protection efficacy of the PMZ8, orally or intramuscularly immunized ducks were challenged with a lethal dose of the WT strain 14 days after boost immunization. Tissue bacterial burdens (blood, liver, and spleen) were quantified at 24 h post-challenge, and survival was monitored for a period of 14 days. PMZ8 immunization markedly reduced the bacterial loads in the detected tissues (Fig. 8A). The bacterial counts in the liver, spleen and blood in the oral or intramuscular immunization groups were much lower than those in the PBS group (Fig. 8A). Even, the bacterial loads in the blood were below the limit of detection for the two PMZ8 immunization groups (Fig. 8A). Furthermore, animals in the PBS control group died within 2 days, whereas survival reached 85% in the orally immunized group and 80% in the intramuscularly immunized group (Fig. 8B). Thus, the delivery of PMZ8 orally or intramuscularly, limits early bacteremia and tissue bacterial burdens under lethal homologous challenge and confers robust protective efficacy.

Fig. 8.

PMZ8 elicits robust homologous protection against lethal challenge in ducks. The immunized ducks were challenged intramuscularly with the WT strain at a dose of 100 × LD₅₀ 14 days after the second immunization. The bacterial loads (A) in the liver, spleen, and blood of all the immunization groups were subsequently measured and calculated as log10 CFU/g or CFU/mL at 24 h post-challenge. Survival curves (B) of immunized ducks were also monitored for 14 days post-challenge (n = 20 per group).

PMZ8 immunization reduces hepatic and splenic histopathology following challenge

H&E-stained sections obtained 24 h post-challenge were scored on a 0–4 semi-quantitative scale. The PBS group presented multifocal heterophil, mononuclear/macrophage infiltration, scattered single-cell necrosis, and marked sinusoidal dilatation with erythrocyte extravasation in the liver (Fig. 9). High lobular-injury scores were consistent with architectural disruption. In contrast, livers from PMZ8-immunized ducks presented only mild scattered inflammation, mild sinusoidal dilatation, and no confluent necrosis (Fig. 9). The composite scores for liver injury (≈1) in the two PMZ8 groups were lower than those (≈3) in the PBS control group, indicating preservation of the lobular architecture. For the spleen, the PBS controls presented red-pulp–dominant fields with widespread sinusoidal congestion/erythrocyte extravasation, multifocal heterophil/mononuclear infiltrates, and reduced, poorly demarcated white pulp, yielding composite scores of approximately 3 (Fig. 9). The ducks of the two PMZ8 groups retained white pulp with patent sinusoids and only mild, scattered inflammation, with composite scores of 0–1 (Fig. 9). Overall, PMZ8 immunization mitigated acute hepatic and splenic pathology and maintained normal tissue organization after challenge.

Fig. 9.

PMZ8 immunization reduces hepatic and splenic histopathology following challenge. Livers and spleens from ducks immunized with PMZ8 by oral or intramuscular injection and PBS controls were collected 24 h post-challenge and stained with H&E. Scale bars, 20 µm.

Discussion

The attenuated strain PMZ8, derived by 50 serial passages, demonstrated a markedly diminished virulence and robust vaccine potential attributable to two key genetic mutations: Fis R86S and WaaF G7R. These mutations target two major virulence determinants of P. multocida, the capsule and the LPS, thereby explaining the strain’s profound attenuation (Harper et al., 2004, 2007; Steen et al., 2010; Harper and Boyce, 2017). The capsule is a critical protective armor that enables P. multocida to evade host innate immunity (Hansen and Hirsh, 1989; DeAngelis, 1999). Loss of the capsule virtually abolishes the ability of bacteria to cause disease in vivo (Chung et al., 2001). In PMZ8, the Fis R86S substitution is consistent with impaired Fis function, leading to reduced capsule biosynthesis and an acapsular phenotype. This finding aligns with previous findings that a functional Fis protein is essential for capsule production in P. multocida. Single point mutations in fis have been shown to eliminate capsule expression, and complementation with an intact fis restores the capsule (Steen et al., 2010). Consequently, the Fis R86S mutation in PMZ8 is a primary driver of attenuation by rendering the strain capsule-deficient, thereby not only reducing intrinsic virulence but also potentially enhancing vaccine efficacy through improved antigen exposure and immune recognition (Wang et al., 2025). The phenotypic traits reported by Wang et al. for a Δfis mutant—slower growth, reduced colony size, diminished capsule production, and enhanced biofilm formation (Wang et al., 2025)-further support the molecular alterations observed here. The increased biofilm formation in PMZ8 is likely an indirect consequence of reduced capsule production. However, whether Fis directly regulates biofilm-associated genes in P. multocida remains to be determined.

The G7R mutation of WaaF in PMZ8 contributes further to attenuation by disrupting the assembly of the LPS outer core. As intact LPS is essential for resisting host innate immune defenses, core-defective strains typically exhibit reduced virulence and heightened sensitivity to antimicrobial peptides and complement (Fontana et al., 2016), which is consistent with our polymyxin B MIC and duck serum killing assays. In PMZ8, silver staining revealed a marked loss of high-molecular-weight LPS bands, which was consistent with outer-core truncation. To validate the contribution of waaF to attenuation, we constructed a PMZ9(ΔwaaF) strain, which similarly exhibited significantly reduced virulence in ducklings. These findings confirm the central role of waaF disruption in driving attenuation. Furthermore, truncated LPS may expose conserved outer membrane antigens that are normally masked, potentially enhancing cross-protective immune responses (Zhao et al., 2022). PMZ8 exhibits increased membrane permeability and endogenous ROS, likely due to LPS truncation and fis-associated metabolic shifts (Duprey et al., 2014). Notably, despite the upregulation of serine biosynthesis genes, the level of intracellular serine remained reduced, supporting the interpretation that serine depletion is a secondary metabolic consequence of elevated oxidative stress rather than a primary transcriptional block of serine biosynthesis. Together, Fis and WaaF mutations confer a high degree of attenuation (LD₅₀ increase >10⁶-fold) and reduce the risk of reversion, suggesting the genetic stability of this dual gene-targeted attenuation strategy. This is supported by the absence of reversion after seven in vitro passages and three consecutive back-passages in ducks.

PMZ8 exhibits good immunogenicity and strong protection against lethal homologous challenge after oral or intramuscular inoculation. Compared with traditional inactivated and CU live vaccines, PMZ8 offers several distinct advantages. Unlike inactivated bacterins, which are serotype-restricted and elicit poor mucosal immunity (Pace et al., 1998), PMZ8 can be administered via multiple routes and induces robust protection without adjuvants. While the CU strain remains the most widely used live vaccine, its limited cross-protection and occasional failure in the field are likely due to residual virulence and a serotype mismatch—CU is serotype A:3,4, whereas field outbreaks are predominantly caused by A:1 strains (Hopkins and Olson, 1997; Mostaan et al., 2020). In contrast, PMZ8 is derived from a virulent A:1 isolate, which aligns more closely with circulating field strains. Moreover, its defined multi-locus attenuation with a clear genetic background, including capsule loss and LPS truncation, not only reduces serotype dependency but also enhances safety and genetic stability. These attributes make PMZ8 a promising live vaccine candidate for preventing fowl cholera. However, protection was assessed only at 14 d after the booster; thus, the durability of the antibody response and long-term protective efficacy remain to be determined. Under a stringent challenge dose, PMZ8 conferred 80–85% protection, and further optimization of the dose and/or immunization schedule may improve its efficacy. In addition, we did not phenotype non-survivors at the individual level; future studies will pair individual serology with early post-challenge bacterial loads to clarify correlates of protection.

Recent advances in the gene-targeted attenuation of P. multocida have yielded promising vaccine candidates. Capsule-deficient strains and LPS-truncated mutants have shown safety and homologous protection in poultry, but their cross-protective capacity remains limited (Chung et al., 2001; Zhao et al., 2022). Broader-spectrum vaccines have been explored through regulatory or metabolic gene deletions. Notably, aroA and fis mutants provide cross-serotype protection in animal models, underscoring the advantages of live vaccines in stimulating mucosal and cellular immunity (Tabatabaei et al., 2002; Wang et al., 2025). In this study we evaluated homologous protection against the parental strain, and cross-protection against heterologous serotypes/Heddleston serovars remains to be determined. However, PMZ8 fits this paradigm by combining capsule and LPS attenuation through naturally acquired fis and waaF mutations.

The attenuated strain PMZ8, developed in this study, exhibited markedly reduced virulence due to synergistic mutations in fis (R86S) and waaF (G7R), while retaining strong immunogenicity. High-dose inoculation in the duck model confirmed its safety, with no signs of pathogenicity. Although further cross-serotype challenge studies in poultry are warranted, the current findings support PMZ8 as a promising live vaccine candidate for the prevention of fowl cholera. Future studies will evaluate heterologous challenge using representative field isolates to directly assess the breadth of protection.

Funding

This research was supported by the National Key Research and Development Program of China (2023YFD1800200), the Sichuan Veterinary Medicine and Drug Innovation Group of the China Agricultural Research System (SCCXTD-2021-18) and the Earmarked Fund for China Agriculture Research System (CARS-42-17).

CRediT authorship contribution statement

Xiangfei Ji: Writing – original draft, Investigation. Yao Meng: Investigation. Haijing Yang: Investigation. Xiaoyan Su: Investigation. Qiao Yang: Investigation. Juan Huang: Investigation. Xumin Ou: Investigation. Bin Tian: Validation. Yu He: Validation. Zhen Wu: Validation. Mingshu Wang: Formal analysis. Anchun Cheng: Writing – review & editing, Methodology, Formal analysis. Xinxin Zhao: Writing – review & editing, Project administration.

Disclosures

We, the undersigned authors of the manuscript entitled “Attenuation mechanisms and vaccine potential of the serial passage–derived Pasteurella multocida strain PMZ8 in ducks”, declare that we have no competing interests that could influence the objectivity of our research. No commercial associations, personal relationships, or other interests related to this study have occurred. The study was conducted independently, and all data and results presented in the manuscript are based on our true research findings.

We confirm that this statement has been read and approved by all named authors and that we understand our responsibility to report potential conflicts of interest.