Glaesserella parasuis was cultured in Tryptone Soy Agar (TSA) or Tryptic Soy Broth (TSB) supplemented with 10 μg/mL NAD (Sangon Biotech) and 3% bovine serum at 37°C, and E. coli DH5α was cultured in Luria-Bertani broth (LB) at 37°C. A final of 30 μg/mL of kanamycin was added to the medium as needed. ZJ1208 is a serovar 13 G. parasuis clinical isolate with high natural transformation competency (12). The cloning vector is pEASY (Beijing TransGen Biotech), and pET-28a is a plasmid stored in the laboratory.

The primers used in this study are listed in Table 1 and synthesized by Sangon Biotech. Using the genomic DNA of ZJ1208 strain as template, homologous arm fragments of upstream (729 bp) and downstream (721 bp) were amplified. Using pET-28a as template, the 848 bp KanR cassette was amplified with L32-KF and L32-KR primers. The three fragments were ligated by overlapping PCR with primers L32-UF and L32-DR, and the PCR products were ligated with cloning vector pEASY to obtain the recombinant plasmid pEASY-L32 (Figure 1), the restriction sites BamH I and Xho I on the vector were used for double enzyme digestion. The recombinant plasmid was further confirmed by gene sequencing.

The recombinant plasmid was transformed into G. parasuis using the natural transformation method as described (12). Briefly, 100 ng of plasmid was added to 300 μL of starvation medium induced competent G. parasuis, and the mixture was incubated at 37°C for 30 min. Then 200 μL of 80% glycerol solution was added and incubated at 25°C for 10 min, following by incubation at 37°C for 100 min in a horizontal shaker at 200 rpm. Finally, the cell culture was spread on TSA (supplemented with 30 μg/mL of kanamycin) and incubated at 37°C for 48 h. The primers listed above were then used to identify the insertion of KanR cassette and the deletion of L32 gene.

To check the possible polar effect of L32 deletion, the gene expression level of the upstream gene YceD and the downstream gene plsX of L32 gene were determined by real-time fluorescence quantitative PCR (RT-qPCR). The 16 s rRNA gene was selected as a reference housekeeping gene in PCR amplification. The primers, as shown in Table 2, were synthesized by Sangon Biotech.

ΔL32 and ZJ1208 were cultured overnight in TSB medium supplemented with NAD and serum until the optical density at 600 nm reached approximately 0.7, the medium was removed, washed twice with PBS, and RNA was extracted using a combination of plasmid small amount extraction kit (Beijing Solarbio Science & Technology Co., Ltd.) and RNA extraction kit (Takara Bio Inc.). Finally, RNA was collected using 50 μL of eluent. The concentration of extracted RNA was measured using NanoVue Plus spectrophotometer (GE Healthcare, United States). Then store it at −70°C. The extracted RNA also needs to undergo gDNA removal. Finally, 900 ng of RNA is reverse transcripted into cDNA before undergoing qPCR.

ΔL32 and ZJ1208 were cultured overnight in TSB medium supplemented with NAD and serum until the optical density at 600 nm reached approximately 0.5. Subsequently, they were inoculated into fresh TSB medium at a ratio of 1:50 and incubated at 37°C. MicroScreen high-throughput real-time microbial growth analysis system (MicroScreen HT, Gering) was employed to measure the OD₆₀₀ every hour throughout the incubation period. This assay was repeated three times, and subsequently, a growth curve was constructed using the incubation time (h) as the x-axis and the average OD₆₀₀ values as the y-axis.

ZJ1208 and ΔL32 were inoculated on TSA plates, and incubated at 37°C for 48 h, and colony morphology was observed under Chemiluminescence imaging system (Bio-Rad ChemiDoc XRS+).

The colonies of solid medium TSA and the logarithmic phase bacteria of liquid medium TSB were diluted with deionized water, and subsequently deposited onto polyvinyl formal-carbon-coated grids. Following natural drying, a brief 8-s staining with 2% phosphotungstic acid enabled direct observation of the bacteria under transmission electron microscopy (H-7650, Hitachi).

Stress resistance assays were referred to the method described previously with slight modification (11, 14). G. parasuis ZJ1208 and ΔL32 were cultured overnight, and the OD₆₀₀ was adjusted to 0.65.

For the osmotic tolerance assay, 1 mL bacteria are resuspended in equal volume of TSB containing different concentrations of KCl (1, 2, and 3 M) and incubated at 37°C for 2 h.

For the oxidative stress assay, 1 mL bacteria were resuspended in equal volume of TSB containing different concentrations (10, 50, and 150 mM) H₂O₂ and incubated at 37°C for 1 h.

For heat shock assay, the bacteria culture was incubated in water bath at different temperatures (42, 50, and 56°C) for 10 min.

All samples for stress test were serially diluted and spread on TSA plates.

The minimum inhibitory concentration (MIC) was determined in 96-well plates using commercial kit (Biofosun Diagnostics) as protocol. The freshly grown bacteria on TSA plates were picked into sterile normal saline, and calibrated to OD₆₀₀ = 0.1 in nutrient broth medium (containing serum and NAD). Then 100 μL of the calibrated bacteria suspension was added to each well, and incubated in a constant temperature incubator at 37°C for 48 h. A total of 16 antibacterial drugs are used: Ampicillin (0.25–512 μg/mL); amoxicillin/clavulanate (0.25/0.12–512/256 μg/mL); gentamicin (0.25–512 μg/mL); spectinomycin (0.25–512 μg/mL); tetracycline (0.25–512 μg/mL); florfenicol (0.25–512 μg/mL); Sulfafurazole (0.25–512 μg/mL); trimethoprim/sulfamethoxazole (0.06/1.2–32/608 μg/mL); Ceftiofur (0.12–256 μg/mL); ceftazidime (0.12–256 μg/mL); enrofloxacin (0.015–32 μg/mL); ofloxacin (0.03–64 μg/mL); meropenem (0.008–16 μg/mL); Apramycin (0.06–128 μg/mL), Colistin (0.12–256 μg/mL); acemequine (1–512 μg/mL).

The strains ΔL32 and ZJ1208 were inoculated in 5 mL of TSB medium, and cultured overnight at 37°C, 200 rpm. The cell density was adjusted to OD₆₀₀ = 0.65 (ΔL32 2.1 × 10⁹ CFU/mL, ZJ1208 2.7 × 10⁹ CFU/mL). Subsequently, 100 μL of the suspension was dispensed into a 96-well plate (Corning, 3,599) with 8 wells allocated for each strain. A negative control group with no bacteria in the medium was also included. The plate was then incubated at 37°C for 48 h. Following incubation, the medium was removed, and each well underwent two gentle washes with ultrapure water (200 μL per well). After drying, 100 μL of 1% crystal violet solution was added to each well and allowed to stain for 2 min before being washed three times with ultrapure water. Finally, 100 μL ethanol (70%, v/v) was added to each well, and the optical density at wavelength of 590 nm (OD₅₉₀) was measured using a microplate reader (SpectraMax M5, Molecular Devices).

The experimental animals used in this study were 4-week-old male BALB/c mice, and were obtained from Shanghai Slake Laboratory Animal Co, Ltd. The mice were randomly assigned into three groups of 10 each, and intraperitoneally injected with G. parasuis (3.2 × 10⁹ CFU/mL) resuspended in 0.5 mL saline; the control group received saline injection only. Survival rates were monitored every 12 h throughout the experiment period. All mice were provided with ad libitum access to food and water, and humane practices were employed to minimize their suffering. At the end of the experiment, all mice were euthanized. This study was approved by the Animal Ethics Committee of Zhejiang Academy of Agricultural Sciences (Approval Number: 2023ZAASLA86).

Statistical analysis was conducted using GraphPad Prism software (one-way ANOVA; Tukey’s post hoc test, ***p < 0.001, **p < 0.01, *p < 0.05). The results were presented as mean ± standard deviation (SD).

The upstream and downstream homology arms and KanR cassette fragments were obtained through amplification (Figure 2A). Subsequently, the upstream and downstream homologous arms along with the KanR cassette fragment were fused via overlapping PCR to generate a fragment of 2,298 bp (Figure 2B). Then recombinant plasmids were validated by restriction endonuclease digestion (Figure 2C), and the accuracy of the inserted sequence was further validated by sequencing (data not shown). Following natural transformation, the colonies cultivated on kanamycin supplemented plates were selected to ascertain the knockout of the L32 gene and insertion of the KanR cassette (Figure 2D). Colony present with KanR cassette and absent with L32 gene was selected, and was designated as ΔL32.

The relative expression levels of the upstream gene YceD and downstream gene plsX of the L32 gene is represented by 2^{−(∆∆ CT)}. The results showed that L32 gene deletion did not cause significant differences in gene expression levels of both upstream and downstream genes (Figure 3). Hence, deletion of L32 did not cause polar effect.

The growth curve of ΔL32 is clearly different compared with that of ZJ1208, and the lag phase of ΔL32 was extended by about 8 h (Figure 4). ZJ1208 achieves its stationary phase with OD₆₀₀ value of 0.65 after 7 h of culture, whereas ΔL32 reaches its stationary phase with OD₆₀₀ value of 0.5 after 18 h.

Translucent, smooth and rounded small colonies (approximately 3 mm in diameter) were seen after 48 h at 37°C for ZJ1208 (Figure 5A), while it took 48 h for ΔL32 to reach a diameter of about 1 mm (Figure 5B).

The electron microscopy results revealed a higher abundance of outer membrane vesicles (OMVs) when cultured with liquid medium compared to TSA plates either for ZJ1208 or ΔL32. Interestingly, regardless of the growth condition, ΔL32 secreted more OMVs which showed an increased pleomorphism, including bead-string shaped, mallet-shaped and long tubular shape etc. (Figure 6). In contrast, ZJ1208 produced fewer and smaller OMVs with uniform shape and size, mostly appearing as regular near-round structures.

The results showed that the ΔL32 was significantly more susceptible to osmotic pressure under the condition of 0.5 M and 1 M KCl, and the resistance to osmotic pressure of ZJ1208 was decreased dramatically when KCl concentration was increased to 2 M (Figure 7A). There is no significant difference in oxidation tolerance under low concentration of H₂O₂, while ΔL32 was significantly more susceptible to 150 mM H₂O₂ compared with ZJ1208 (Figure 7B). Similarly, L32 was more susceptible to high temperature of 50°C, while both strains could not survive when the temperature raised to 56°C (Figure 7C). This indicates that the absence of L32 gene affects the stress tolerance of G. parasuis.

The minimum inhibitory concentration (MIC) of ΔL32 decreased in respect to all aminoglycoside antibiotics used in this study, including spectinomycin, Apramycin, gentamycin. Higher susceptibility was also observed for ΔL32 to sulfonamide antibiotics (sulfafurazole) in this study (Table 3). The mutant shown no difference in resistance to other antibiotics used in this study compared to ZJ1208. This suggests that L32 plays an important role in resistance to these two types of antibiotics.

On the first day of challenge, four mice succumbed to infection in the ZJ1208 group, followed by an additional death on the second day. In contrast, only one mouse died on the second day in the ΔL32 group. The ΔL32 exhibited a significant reduction of mortality by 40% compared to that of ZJ1208 (Figure 8). No mortality was observed in the negative control group (normal saline).