Background

Glaesserella parasuis is a gram-negative bacterium belonging to the Pasteurellaceae family [1]. This organism is known to cause Glässer’s disease, which is characterized by a constellation of symptoms, including fibrinous polyserositis, arthritis, meningitis, acute pneumonia, and septicemia [2,3,4]. Glässer’s disease is a significant bacterial affliction of swine and has been reported on swine farms worldwide [5,6,7]. In China, the isolation ratio of G. parasuis was found to be as high as 22.6% in weaned swine and 9.7% in healthy swine herds [5, 8]. Therefore, gaining insight into the pathogenesis of Glässer’s disease and understanding the interaction between G. parasuis and its host are crucial. To achieve this objective, it is essential to develop a marker-free gene knockout system.

Most strains of G. parasuis have been observed to possess highly stringent restriction-modification systems that impede the uptake of heterogeneous plasmids, particularly through electrotransformation [9, 10]. Consequently, allelic exchange knockout mutants of G. parasuis strains are frequently constructed using natural transformation with suicide plasmids. The plasmids contain antibiotic-resistance genes that are flanked by homologous regions upstream and downstream of the target gene that is to be deleted. Consequently, the resulting knockout mutants are marked with antibiotic-resistance [9, 11,12,13].

While marked mutants are valuable for investigating the molecular pathogenesis of G. parasuis, they may also exhibit antibiotic-resistance gene drift and other unpredictable polar effects. Moreover, these mutants are not suitable for use in commercial vaccines due to the presence of exogenous antibiotic-resistance genes. In our preliminary research, we established a markerless gene knockout system in G. parasuis based on the Flp-FRT system; however, it still leaves FRT sites on the genome [14]. Thus, a marker-free knockout approach is necessary for studying G. parasuis pathogenesis and developing genetically engineered vaccines. A sacB-based counterselectable marker system has been applied to generate marker-free knockout mutants in G. parasuis; however, this method requires two natural transformations of two donor plasmids and is restricted to G. parasuis strains that lack spontaneous sucrose resistance [15].

In the present study, transformants carrying temperature-sensitive (TS) donor plasmids were cultured alternately at 37 °C and 30 °C to ensure genomic DNA replication and plasmid maintenance. Our results demonstrated successful knockout of target genes in different G. parasuis strains, leading to the establishment of a comprehensive marker-free gene knockout system for this organism.

Methods

Bacterial strains and growth conditions

The bacterial strains used in the study are listed in Table 1. Escherichia coli strain DH5α was grown in Luria–Bertani broth (LB; Difco, Detroit, MI, USA) or on LB agar. G. parasuis strains were grown in tryptic soy broth (TSB; Difco) or on tryptic soy agar (TSA; Difco) supplemented with 10 µg/mL NAD and 5% bovine serum. When needed, media were supplemented with kanamycin (50 µg/mL) or gentamicin (20 µg/mL). G. parasuis strains JS0135 (CCTCC AB 2021004) and CF7066 (CCTCC AB 2021003) were isolated in China and stored in our laboratory and at the China Center for Type Culture Collection. The original transformed recipient strain JS0135ΔnanH::KanR was generated through natural transformation with pK18-nHUD according to a previously published method [9]. The control mutant CF7066ΔnanH-N was constructed using a modified NgAgo system [16] (Table 1).

Cell culture

Porcine kidney 15 (PK-15) cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, USA) at 37 °C with 5% CO2.

Cloning and genetic manipulation

The plasmids used in this study are listed in Table 1. Primers for plasmid construction and mutant identification were designed based on the sequence of G. parasuis SH0165 (CP001321.1; Table S1). To construct the TS vector pSHG5, a fragment containing pUC ori and pA13 oriTS was amplified from pSHK3TS, and the gentamicin cassette was amplified from pSHK3-Gm [17, 18]. The resulting fragments were fused using a one-step cloning kit (Vazyme Biotech, Nanjing, China). For pSHG5-ΔnanH, the flanking homology arms of the deleted nanH allele of JS0135 (MT625975) were amplified using pK18-nHUD as a template and then ligated into pSHG5. For pSHG5-Δapd, the flanking homology fragments of the deleted apd allele were obtained from G. parasuis CF7066 (MK617354) and fused to the pSHG5 vector using overlapping polymerase chain reaction (PCR) and a one-step cloning kit. Genetic constructs and PCR fragments were sequenced by Sanger sequencing (TsingKe Biotech, Beijing, China; Table 1, Table S1, and Fig. S1).

Electrotransformation of G. Parasuis

Electrotransformation of G. parasuis was performed according to a modified method of Lancashire et al. [19]. In brief, G. parasuis strains were grown on TSA plates at 37 °C for 12–14 h. Bacterial cells were then harvested in sucrose/glycerol (SG) buffer (15% [v/v] glycerol, 272 mM sucrose), washed three times, and resuspended in 0.12 mL SG buffer per sample. After electroporation (using a Bio-Rad MicroPulser Electroporator [Bio-Rad Laboratories, Hercules, CA, USA]; 2.5 kV, 0.2 cm cuvettes) with 5 µg plasmid DNA, cells were suspended in 1 mL TSB medium, grown at 37 °C for 2 h, and then spread on TSA plates supplemented with 20 µg/mL gentamicin.

Construction of marker-free knockout mutants in G. Parasuis

To generate crossover mutants, recipient strains were electroporated with donor plasmids and grown on gentamicin plates at alternating temperatures of 37 °C and 30 °C every 12 h. After several passages, single or double crossover colonies were screened by PCR using primers targeting the flanking regions of upstream and downstream homologous arms in the genome. Once a colony with a double crossover event was identified, it was cultured at 42 °C to cure the plasmid. To generate JS0135ΔnanH and CF7066ΔnanH, the double-crossover colony was further cultured at 42 °C until the optical density at 600 nm (OD600) reached 0.5 and then spread on gentamicin-free TSA plates to isolate cells of different genotypes. For CF7066Δapd, double-crossover colonies were directly streaked on TSA plates at 42 °C overnight and subsequently identified by PCR.

Detection of NanH and Apd using Western blotting

The expression of NanH and Apd in the wild-type and deletion strains was assessed by Western blotting using anti-NanH [20] and anti-Apd [21] mouse sera. Newly harvested bacterial cells were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to nitrocellulose membranes. After blocking and washing, the membranes were incubated with 1:5000 antisera specific for recombinant Apd and NanH in TBST buffer (20 mM Tris-HCl, 150 mM NaCl, 0.05% Tween 20, pH 7.4) containing 5% skim milk. The membranes were then exposed to horseradish peroxidase-labeled goat anti-mouse IgG (ABclonal, Inc., Wuhan, China) at a 1:5000 dilution in TBST buffer containing 5% skim milk, and the bound antibodies were visualized using Clarity Western ECL Substrate (Bio-Rad Laboratories).

Bacterial growth curve assays

The strains were activated on TSA plates for three generations, then single colonies were inoculated into 5 mL of TSB media, cultured with shaking at 37 °C/200 rpm for 12 h, and transferred to fresh TSB media at a 1:100 dilution. Next, 200 µL of bacterial solution was added to each well of the special detection plate suitable for the Bioscreen automatic growth curve analyzer. The OD600 of the bacterial solution was measured every 30 min, and the growth curve was recorded and plotted.

PK-15 cell adhesion and invasion assay

G. parasuis CF7066, JS0135 and the mutants were streaked on TSA plates and cultured at 37 °C for 16 h. The bacterial cells were then harvested on TSB media and the OD600 was adjusted to 1.0. PK-15 cells were grown to confluent monolayers in 24-well plates and infected with G. parasuis at a multiplicity of infection (MOI) of 0.02. The plates were centrifuged at 800 × g for 10 min to transport bacteria to the surface of the monolayer and incubated at 37 °C for 2 h. Thereafter, 100 µL medium and 900 µL ice water were added to each well and the plates were incubated at 4 °C for 15 min. The cells were then vigorously washed three times with PBS to eliminate nonspecific bacterial attachment and incubated for 15 min at 4 °C. The cells were then disrupted by repeated pipetting to liberate bacteria. Serial dilutions of the cell lysate were plated on TSA and incubated at 37 °C for 48 h. Adhesion rate = (Cell-associated CFU/Total CFU) × 100. All assays were repeated at least three times in independent experiments.

Quantitative reverse-transcription PCR (qRT‒PCR)

Total RNA was extracted from bacterial cells using a HiPure Bacterial RNA Kit (Magen, Guangzhou, China), and 1 µg of RNA was reverse transcribed into cDNA using a HiScript II Q RT SuperMix Kit (Vazyme Biotech, Nanjing, China) according to the manufacturer’s protocol. qRT-PCR was performed using a ChamQ Universal SYBR qPCR Master Mix Kit (Vazyme Biotech, Nanjing, China) with the primers listed in Supplementary Table S1, and the results were normalized to those for GAPDH. The G. parasuis harboring plasmid pSHG5-Δapd was streaked on TSA plates, divided into two groups and cultured at 37 °C or 30 °C. Bacterial cells were harvested and the OD600 was adjusted to 1.0. For each bacterial solution group, 1 µL of the sample was used for qRT-PCR. All assays were repeated at least three times in independent experiments.

Table 1 Strains and plasmids used in this study
Full size table

Results

Knockout of the KanR cassette in JS0135ΔnanH::KanR

Given the potential resistance to endogenous gene knockout, we initially targeted the exogenous KanR gene in JS0135ΔnanH::KanR for knockout. The resulting KanR knockout strain did not exhibit growth defects compared to the parental strain JS0135ΔnanH::KanR, facilitating subsequent mutant screening. Therefore, we used the natural transformation method as previously described to replace the nanH gene with KanR in the G. parasuis strain JS0135, generating the marked mutant JS0135ΔnanH::KanR.

To knock out the KanR gene, the donor plasmid pSHG5-ΔnanH carrying the homologous arms of nanH was electroporated into JS0135ΔnanH::KanR. However, due to the abundant restriction-modification systems in JS0135, the strain exhibited strong resistance to electroporation, resulting in the generation of only one transformant without recombination after multiple attempts (Fig. 1a and b). Notably, the colonies grew better at 37 °C, but the temperature-sensitive plasmid pSHG5-ΔnanH could only replicate properly at or below 30 °C. To accommodate different temperatures requirements, the transformants were incubated for 2 days on gentamicin TSA plates at 37 °C, then streaked on fresh gentamicin TSA plates and cultured alternately at 37 °C and 30 °C for one passage (24–48 h). Newly grown single colonies were picked and cultured in the same manner for another generation. After five generations, we found that most colonies exhibited multiple genotypes (Fig. 1c). The bands observed in the gel correspond to four genotypes of JS0135 mutants: the 3239 bp band represents the parent strain genotype (ΔnanH::KanR) without recombination; the 2330 bp band represents the knockout mutant genotype (ΔnanH) resulting from double crossover with the parent strain; and the 6650/10,061 bp bands represent the single/twice crossover genotypes where the plasmid integrated into the parent strain genome via the upstream or downstream homologous arms (Fig. 1a and c). Interestingly, a faint band near the single crossover band (6650 bp) was observed after inverting the image colors, speculated to correspond to plasmid integration into the knockout mutant genome (Fig. S2).

To further screen for the knockout mutant, we selected colony No. 4 (Fig. 1c), which exhibited a relatively bright band for the double-crossover genotype. After inoculating this colony into gentamicin-free TSB and culturing it to the logarithmic growth phase, the cells were spread on gentamicin-free TSA plates at 42 °C. This approach successfully isolated KanR knockout colonies, with 90% of the colonies displaying the KanR cassette knockout genotype as the only genotype (Fig. 1d). The double and single crossover PCR fragments were then purified and confirmed through Sanger sequencing (Fig. 1e and Fig. S3).

Fig. 1
figure 1

Knockout of the KanR cassette from the marked mutant JS0135ΔnanH::KanR. (a) Schematic diagram illustrating the process for generating the marker-free mutant JS0135ΔnanH from the marked mutant JS0135ΔnanH::KanR. SCO, the single crossover genotype with one entire plasmid integrated into the genome. (b) Only one transformant grew, and PCR analysis using primers nanH-F1/R1 flanking the upstream and downstream homologous arms of nanH (3239 bp) showed that this transformant had no genotype change compared to JS0135ΔnanH::KanR. The nanH knockout mutant, CF7066ΔnanH-N, generated by the NgAgo gene deletion system, served as the positive control. C, negative control. (c) Multiple genotypes (2330, 3239, 6650, and 10061 bp) were observed in progeny colonies after culturing alternately at 30 °C and 37 °C for five generations and were then screened by colony PCR analysis. Colony No. 4 was selected for culture at 42 °C for two passages using the process described above. (d) The marker-free nanH knockout mutants (2330 bp) were detected in 90.0% (18/20) of colonies. (e) Sanger sequencing was conducted to confirm the elimination of the antibiotic-resistance marker

Full size image

Knockout of the nanH gene in wild-type CF7066

A previous study by our laboratory failed to screen mutants at 30 °C using a traditional temperature-sensitive plasmid method (unpublished). However, in this study, we successfully screened mutants and observed numerous recombinant genotypes. We speculated that the success of this method was due to the alternating culture temperatures of the transformants. To further confirm that alternating culture temperatures at 37 °C and 30 °C contributed to the accumulation of recombinants, we electroporated CF7066 with the pSHG5-ΔnanH plasmid and cultured it on gentamicin plates with alternating temperatures every 12 h. Among 21 randomly picked transformants, colonies No. 1 and 14 exhibited a single crossover genotype (5929 bp), and the other transformants were wild-type (2518 bp) (Fig. 2a and b).

As anticipated, colony No. 1 (Fig. 2b), which underwent alternating streaking at 37 °C and 30 °C, yielded the nanH knockout genotype (1610 bp) after eight passages (Fig. 2c). Colony No. 21 (Fig. 2c) showed the brightest band corresponding to the ΔnanH genotype. Further culturing of this colony allowed successful isolation of nanH knockout colonies using the described method (Fig. 2d). Consequently, the nanH knockout genotype was present in 7.6% (8/105) of the colonies (Fig. 2d and Fig. S4). Sanger sequencing confirmed the successful knockout of nanH (Fig. 2e).

In contrast to the JS0135ΔnanH generation, where most colonies exhibited the knockout genotype (Fig. 1d), most colonies in the CF7066ΔnanH screen restored the wild-type genotype (Fig. 2d and Fig. S4). This difference may be attributed to the growth advantage of the dominant genotype. Specifically, JS0135ΔnanH outperformed JS0135ΔnanH::KanR because the expression of KanR did not affect growth on non-selective plates but imposed an additional growth burden. Similarly, the absence of nanH in CF7066ΔnanH might have impaired strain metabolism, giving CF7066 a growth advantage over CF7066ΔnanH. This growth advantage likely intensified during the subculture process.

Fig. 2
figure 2

Generation of marker-free in-frame knockout in G. parasuis CF7066. (a) Schematic diagram illustrating the process for generating an in-frame nanH knockout mutant. WT, wild-type G. parasuis CF7066. (b) The genotypes of transformants were amplified using primers nanH-F2/R2, which flank the upstream and downstream homologous arms of nanH. Two transformants contained a single crossover genotype (5929 bp), while the others retained the wild-type genotype (2518 bp). CF7066ΔnanH-N, the nanH knockout mutant generated by the NgAgo deletion system. (c) Three genotypes were observed in progeny colonies after culturing colony No. 1 alternately at 37 °C and 30 °C for eight passages. Subsequently, colony No. 21, which exhibited the brightest knockout band (1620 bp), was selected and cultured at 42 °C for two passages. (d) Two colonies exhibited the nanH knockout genotype (1620 bp). (e) Sanger sequencing was conducted to confirm the marker-free in-frame nanH knockout in CF7066

Full size image

Open reading frame (ORF) knockout contributes to the screening of mutants

To further confirm our hypothesis, we attempted to knock out the apd gene of CF7066 using the same method. To address potential difficulties in screening knockout colonies due to growth disadvantages, we improved the design of the donor plasmid pSHG5-Δapd by placing the upstream and downstream homologous arms outside the ORF of apd. This design ensures that even if the transformant undergoes gene recombination resulting in the knockout of apd in the genome, the Apd protein can still be expressed from the plasmid.

The plasmid pSHG5-Δapd was electroporated into CF7066 and cultured on gentamicin plates at 37 °C. Among the 22 transformants tested by colony PCR analysis, 21 produced a band corresponding to the wild-type genome (3958 bp). However, colony No. 17 produced a band indicative of the knockout genotype (1448 bp; Fig. 3a and b). After streaking this transformant on gentamicin-free plates at 42 °C for one passage, 7 progeny colonies were identified as the apd knockout genotype (Fig. 3c). Subsequently, colony No. 8 was streaked on gentamicin-free plates at 42 °C for a second passage, and the apd knockout genotype appeared in 100% of the progeny colonies tested (Fig. 3d and Fig. S5). Sanger sequencing was performed to further confirm the deletion of apd (Fig. 3e).

During the process, the plasmid genotypes of the transformants were also detected by PCR using primers targeting the plasmid backbone that flanked the upstream and downstream homologous arms of apd (Fig. 3a). Of the 22 transformants tested, 14 contained pSHG5-Δapd carrying the deleted apd allele (1130 bp), while 6 possessed the plasmid carrying the wild-type apd allele (3660 bp) (Fig. 3b). This result suggested that it might be necessary to examine the homology arms on the plasmid in the primary transformants to ensure that they had not been swapped with the genome. After colony No. 8 was cultured on gentamicin-free plates for one passage, 72.7% of the progeny colonies were cured (Fig. 3c and d). Compared to the previous two mutants JS0135ΔnanH and CF7066ΔnanH, the screening of CF7066Δapd required only two passages, which considerably reduced the time required for mutant construction.

Fig. 3
figure 3

Generation of an ORF knockout in G. parasuis CF7066. (a) Schematic diagram illustrating the process for generating the apd ORF knockout mutant. The plasmid pSHG5-Δapd harbored the upstream and downstream homologous arms of apd, with the promoter and stop codon respectively. Three patterns of plasmid and genome genotype combinations in one colony are shown, with the proportion of each pattern among the 22 colonies indicated below the patterns. (b) After culturing the transformed bacteria alternately at 37 °C and 30 °C for one passage, the knockout genotype (1448 bp) appeared in the bacterial genome of colony No. 17, and the wild-type apd gene (3660 bp) was replaced by the Δapd gene in the plasmids. (c) Colony No. 17 was streaked on gentamicin-free plates for one passage, and 7 out of 22 progeny colonies possessed the apd knockout genotype, while plasmids with the wild-type apd allele appeared in all colonies. (d) A second streaking at 42 °C eliminated plasmids. (e) Sanger sequencing was conducted to confirm the apd ORF knockout

Full size image

Biological characteristics of the three mutants

Target gene knockout was confirmed through Western blotting. In the wild-type strains JS0135 and CF7066, NanH expression was detectable, but absent in the JS0135ΔnanH and CF7066ΔnanH knockout mutants (Fig. 4a and b). Similarly, Apd was expressed in the wild-type CF7066 strain but not in the mutant CF7066Δapd strain (Fig. 4c). These results provide further evidence that the TS plasmid-based procedure successfully generated marker-free deletion mutants of G. parasuis.

The phenotype, growth and adhesion of the strains were also detected. The growth curve presented in Fig. 4d showed that the deletion of apd had no significant effect on the growth of CF7066, while the deletion of nanH significantly inhibited the growth of CF7066 and JS0135. Additionally, the deletion of apd led to a significant increase in the bacterial adhesion rate of CF7066 compared to the wild-type strain, whereas the absence of nanH in CF7066 and JS0135 did not result in any noticeable differences compared to their respective wild-type strains (Fig. 4e). The observations could be attributed to the distinct roles of nanH and apd in bacterial biological functions, which warranted further investigation. Moreover, variations in the growth curves of the mutants may provide an explanation for why apd was knocked out more rapidly and easily than nanH.

Fig. 4
figure 4

Characterization of three marker-free knockout mutants. (a, b) Western blot analysis revealed that NanH was detectable in wild-type JS0135 and CF7066 strains, but not in their respective knockout mutants. (c) Similarly, Apd was absent in CF7066Δapd, while it was expressed in wild-type CF7066. Red arrowheads indicate the presence of NanH or Apd in the wild-type strains. (d) The bacterial OD600 was measured every 30 min to draw the growth curve. Compared to the CF7066 and JS0135 respectively, CF7066Δapd had no effect on growth, but both JS0135ΔnanH and CF7066ΔnanH exhibited poor growth. (e) The adhesion rate of CF7066Δapd increased significantly in contrast to wild-type CF7066, while both CF7066ΔnanH and JS0135ΔnanH were similar to CF7066 and JS0135 respectively

Full size image

Plasmid copy number significantly affects the expression of recombination-related proteins

The optimal temperature for G. parasuis growth and DNA replication is approximately 37 °C, rather than 30 °C. Therefore, we speculated that 37 °C might be more conducive to homologous DNA recombination. We first evaluated the expression of recombinant genes (recA, recB, recC, recD, and ruvA) in wild-type CF7066 at 30 °C and 37 °C by RT‒qPCR. The results revealed minimal differences in the expression of each gene at different temperatures (Fig. 5a). However, in the transformant CF7066 (pSHG5-Δapd), the expression of the recombinant genes was significantly upregulated at 30 °C (Fig. 5b). Specifically, the expression of recC was the most obviously upregulated, being approximately 7 times greater than that at 37 °C (Fig. 5b). Moreover, the RT‒qPCR results of the bacterial solution indicated that the plasmid copy number at 30 °C was more than twice that at 37 °C (Fig. 5c). This indicated that the plasmid copy number was higher at 30 °C, whiched resulted in greater expression of the recombinant genes. However, the slow growth of colonies at 30 °C was not conducive to recombination, and our previous study had never successfully screened mutants at 30 °C. It can be proposed that within the temperature alternation window, a high plasmid copy number results in the high expression of recombination genes, while bacterial chromosomes maintain efficient replication. Therefore, it can be posited that DNA homologous recombination events are more likely to occur during the temperature alternation window.

Fig. 5
figure 5

Transcription levels of DNA recombination-related genes (recA, recB, recC, recD, ruvA) at different temperatures. (a) Transcription levels of various DNA recombination-related genes in the wild-type strain CF7066 at 30℃ and 37℃. (b) Transcription levels of various DNA recombination-related genes in the transformant CF7066 (pSHG5-Δapd) at 30℃ and 37℃. (c) Plasmid copy numbers of the transformant CF7066 (pSHG5-Δapd) at 30℃ and 37℃

Full size image

Discussion

In previous studies, we attempted to use the method described by Taturn and Briggs [24] to screen for double-crossover strains. In brief, the transformants were first subjected to selection at 42 °C to cure the plasmids, allowing only the single-crossover colonies to grow under selection pressure; then, the single-crossover colonies were propagated on gentamicin-free plates at 30 °C to screen for the knockout strains. However, this method failed in our study because homologous DNA recombination was infrequent during normal cellular metabolism, and target gene knockout conferred a growth disadvantage to the resulting mutants. As a result, the single-crossover strains were more inclined to revert to the wild-type. Moreover, single-crossover strains can only recombine once after the plasmids are cured, which presents a significant challenge in screening procedures.

To address these challenges and increase the likelihood of obtaining double-crossover genotypes, we developed a strategy for culturing transformants alternately at 37 °C and 30 °C. Our approach was based on the hypothesis that DNA damage occurs spontaneously during normal cellular metabolism and that double-strand DNA breaks (DSBs) are the most severe type of DNA damage. Homologous recombination is a well-known pathway for repairing DSBs [25, 26] and occurs mainly during DNA replication [27,28,29,30]. Therefore, we proposed that a greater copy number of plasmids in cells would lead to more frequent DSBs and consequently activate the repair mechanism more frequently, thus increasing the probability of homologous recombination. This hypothesis was supported by our observations, where the efficiency of gene knockouts improved significantly with the alternating temperature strategy.

Additionally, previous work has shown that recombination is partially dependent on DNA replication [31, 32] and that blocking DNA synthesis by high temperatures reduces the frequency of recombination [32]. We believe that DNA double-strand unwinding before DNA replication provides opportunities for homology searching and DNA strand invasion, which is a key reaction involved in homologous recombination [33]. Which means, in this study, the improvement of the frequency of homologous recombination is dependent upon the efficient replication of both the genome and the temperature-sensitive plasmids. Conversely, G. parasuis grew significantly more slowly at 30 °C, demonstrating that DNA synthesis decreases, and the TS plasmid failed to replicate at or above 42 °C in G. parasuis [34]. Consequently, culturing transformants alternately at 37 °C and 30 °C could ensure that both the genome and plasmids replicate efficiently, thereby improving the frequency of recombination.

Moreover, the increase in plasmid copy number at the lower temperature of 30 °C likely enhanced the availability of target sequences for homologous recombination. This phenomenon was observed in our study, where the expression of recombinant genes was significantly higher at 30 °C compared to 37 °C, suggesting that the elevated plasmid copy number facilitated higher recombination efficiency. Additionally, the higher plasmid copy number at 30 °C also implies that, at the molecular level, the likelihood of recombination between the homologous arms on the genome and the plasmid increases. This insight provides a deeper understanding of the molecular mechanisms underlying the improved success rate of our optimized knockout system.

Based on these findings, we speculate that the efficiency of homologous recombination in bacteria might peak during the temperature transition window. Additionally, homologous recombination is one of the adaptive mutation mechanisms that bacteria use to cope with environmental changes [35,36,37,38]. Frequent fluctuations in cultivation temperature may potentially stimulate the activation of homologous recombination. We believe that increasing the frequency of temperature transitions might further enhance the recombination efficiency, though this warrants further investigation.

Our findings also highlight the broader implications of this optimized system. By facilitating more efficient gene knockouts, researchers can more rapidly investigate gene function and pathogenic mechanisms in G. parasuis. This advancement not only accelerates basic research but also has potential applications in developing targeted treatments and interventions for Glässer’s disease. Furthermore, the principles of our approach could be adapted for use in other bacterial species, thereby contributing to the wider field of microbial genetics and pathogenesis research.

Overall, our optimized marker-free knockout system represents a significant advancement in genetic manipulation techniques for G. parasuis. By addressing the limitations of previous methods and leveraging the biological principles of homologous recombination, we have developed a more efficient and reliable method for creating gene knockouts. This system paves the way for future studies to explore the complex genetics of G. parasuis and develop novel strategies to combat Glässer’s disease.

Conclusion

In this study, we successfully established a marker-free knockout system for G. parasuis by electrotransforming TS donor plasmids. Our screening process revealed that culturing transformants alternately at 37 °C and 30 °C promoted homologous recombination, making the production and screening of double-crossover mutants more feasible. Our findings also showed that simultaneous single exchange events based on the two different homologous arms of the target gene could occur, resulting in the insertion of two plasmids into the genome. This marker-free knockout system is simple and efficient, requiring no high transformation efficiency, and the plasmid used for transformation is small and requires no additional elements, which is particularly important for G. parasuis due to its plentiful and stringent restriction-modification systems. Additionally, the promotion of crossover and recombination observed in the TS mutation system through alternate cultivation at 37 °C and 30 °C may have potential applications in other microorganisms.