The improved Frey culture medium was obtained from China Sea Biotechnology Co., LTD., Beijing, China. The bacterial genome DNA extraction kit was purchased from Tiangen Biochemical Technology Co., LTD., Beijing, China. Crystal violet was acquired from Chongming County Yuxi Reagent Factory. Ethanol (95%) and methanol were purchased from Guangnuo Chemical Technology Co., LTD., Shanghai, China. Glutaraldehyde fixative (2.5%, for electron microscopy) was from Proanti Biotechnology Development Co., LTD., Shaanxi, China. Ofloxacin, Ennofloxacin, Norfloxacin, Taymycin, Erythromycin, Talomycin, Temicoxacin, Spiramycin, Tetracycline Hydrochloride, Lincomycin, Gentamicin, Dakanamycin, Kanamycin, Streptomycin, Doxycycline, and Warnemylin were purchased from Golden Clonix Biotechnology Co., LTD., Beijing, China.

Ethics approval and specific permission were not required for the study, as all samples were collected by the authors during necropsies with the consent of the owners. Air sacs of chickens suspected to be infected with M. gallisepticum were collected from poultry farms in Ningxia (Felice et al., 2020; Yadav et al., 2022). They were cut and placed in modified Frey’s liquid medium overnight at 4°C under sterile conditions. After filtration, they were inoculated into Frey liquid medium and incubated in a biochemical incubator at 37°C. The MG were purified by Kahya et al. (2010) and Sprygin et al. (2010), the harvested isolates were coated in the modified Frey’s solid medium and placed in a 5% CO₂ incubator at 37°C for 3–7 days. Single colonies were selected and cultured in the modified Frey’s liquid medium, and the purified isolates were repeated three times to get the purified isolates. The colony morphology on the solid medium was observed with an inverted microscope.

Bacterial genomic DNA was extracted using the kit of Tiangen Biochemical Technology Company (Ahmad et al., 2020; Li L. et al., 2020). After the expanded culture, put 1 mL of fresh bacterial solution into a centrifuge tube and add 20 μL of protease K solution. Add 220 μL anhydrous ethanol and mix thoroughly. Then, the solution and precipitate are added to the adsorption Add 220 μL anhydrous ethanol and mix thoroughly. Then the solution and precipitate are added to the adsorption column, and 500 μL buffer GD is added to the column. Discard the waste liquid after centrifugation. Add 600 μL PW to the adsorption column, and the waste liquid was discarded after centrifugation. Add 50 μL eluent buffer TE in the middle of the adsorbed strain and let it stand at room temperature for 5 min. The solution was collected into the centrifuge tube to obtain the genome extract of MG strains. Meanwhile the isolated MG strain was amplified by PCR based on the OIE 16S rRNA and mgc2 primers to identify the MG strain in the bacterial solution (Table 1), nuclease-free water was used as a negative control in all PCR assays. Finally, the bacterial solution was mixed at a ratio of 1:1 to glycerol and stored at −80°C.

Modified tests using crystal violet staining were used to quantify biofilm formation (Hu et al., 2010). The suspension concentration of the prepared MG strains was adjusted to 1 × 10⁷ CCU/mL, and 200 μL was inoculated on sterile 96-cell plates. Six replicates of the bacterial culture were prepared and subsequently incubated at 37°C for 12, 24, 36, and 48 h. Meanwhile, the crystal violet staining method was employed to measure the OD₅₉₀ value of MG biofilm by enzyme marker, and the absorbance results were recorded. According to ODc value (ODc is equal to the average OD value of blank well), biofilm classification: OD ≤ ODc, no membrane capacity (−); ODc < OD ≤ 2ODc, weak biofilm forming ability (+); 2ODc < OD ≤ 4ODc, medium biofilm forming ability (++); and 4ODc < OD, strong biofilm forming ability (+++).

The growth curves of the biofilm matrix were measured by a semi-quantitative adhesion test (Liang et al., 2019). 200 μL of 1 × 10⁶ CCU/mL bacterial suspension was inoculated on 96-cell plates with six replicates per strain, after which the plates were then incubated at 37°C for 0, 3, 6, 9, 12, 24, 36, 48, and 72 h. In the corresponding period time, crystal violet staining was used to detect the OD₅₉₀ value of two MG biofilms. According to the absorbance results, the biofilm matrix growth curve was drawn.

Mycoplasma gallisepticum strains were inoculated into the liquid Frey medium for 12 h, and the suspension concentration was adjusted to 1 × 10⁴ CCU/mL. The sterile six-well plate was used to hold the 14 mm round slip-plate, with 5 mL of bacterial suspension added to each well. The culture was incubated at 37°C for 24 h, and 1 mL 2.5% glutaraldehyde fixed solution (EM Grade) was added to fixed at 4°C overnight. The gradient elution was performed on each well was gradient eluted with ethanol of different concentration (30, 50, 70, 80, 90, and 95%). The obtained sample was sliced and photographed using an ISI-SX-40 scanning electron microscope (Shatila et al., 2021).

The drug sensitivity of the isolated stains were determined by Hannan’s microdilution method (Hannan, 2000). The suspension concentration of MG strains was adjusted to 1 × 10⁴ CCU/mL. 40 μL 512 μg/mL of antibiotics were added to the first hole of the 96-well plate, and add 20 μL sterile PBS from well 2 to 12. Then 20 μL of liquid was discarded from well 1 to 2 and mixed, and the same operation was performed until well 12. Thus, the antibiotics were continuously diluted at multiple ratios. Finally, 180 μL diluted MG bacterial solution was added to well 1–12. Growth control wells (containing culture medium and bacterial solution) and negative control Wells (containing liquid culture medium) were also set up and each antibiotic were repeated three times, After being inoculated, the 96-well plates were transferred to a constant temperature incubator set at 37°C for a week. The color of the growth control well was yellow, and the negative control well was pink. The highest concentration of the drug with no color change in the test well was considered the lowest inhibitory concentration of the drug (Gautier-Bouchardon, 2018).

Sangon Bioengineering (Shanghai) Co., LTD. completed the genome-wide sequencing of the MG strains. MG complete genome sequences were obtained using Illumina HiSeq platform (Caporaso et al., 2012) for MG genome sequencing. The original image data file of the MG genome was processed by Base Calling and stored in FASTQ (Brown et al., 2017) file format. Next, the FastQC software was used to evaluate sequencing data quality. Using Trimmomatic (Bolger et al., 2014) was used to clear joint sequences and lower quality sequences to ensure that all the clean reads false discovery rate was <0.01.

To splice sequencing data, the SPAdes (Bankevich et al., 2012) was utilized (Table 2). In order to fill gaps in the resulting contigs, the GapFiller (Boetzer and Pirovano, 2012) was employed. The accuracy of the spliced sequences was improved using PrInSeS-G (Massouras et al., 2010) correction tool that can rectify splicing errors during editing and detect small insertion losses. The genetic elements were predicted using Prokka (Seemann, 2014), and repeated sequences in the genome were identified using RepeatMasker (Saha et al., 2008). The CRISPRs were predicted through CRT (Bland et al., 2007) analysis. To obtain functional annotation information, gene sequences were analyzed separately using NCBI NR (Yu and Zhang, 2013), COG (Tatusov et al., 2000), GO (The Gene Ontology, 2019), and KEGG (Kanehisa and Goto, 2000), VFDB (Chen et al., 2016), CARD (McArthur et al., 2013), and CAZy (Lombard et al., 2014), and compared using NCBI Blast + (Altschul et al., 1997).

The annotation data of the genome-wide of the locally isolated MG strains were analyzed, including total gene length, GC content, coding gene, etc. Molecular evolutionary tree were constructed using N-J method and MEGA6 software (Tamura et al., 2013). The whole-genome sequences of MG strain were compared with those of the five MG strains available in the NCBI database. Subsequently, the linear analysis of genome-wide sequences of MG strains were conducted using the MUMmer (Delcher et al., 2003; Table 3). Finally, a plot was created to visualize the region of genetic differences.

According to strain isolation and identification, five MG strains NX-01, NX-02, NX-03, NX-04, and NX-05 were isolated during this study (Figures 1, 2). The capacity of biofilm formation showed that all 5 MG strains could form biofilms, with the following formation ability: NX-01 > NX-03 > NX-05 > NX-04 > NX-02 (Figure 3A), whereas the M. gallisepticum strains NX-01 and NX-03 showed strong biofilm forming ability (+++). Meanwhile, M. gallisepticum strain NX-05 showed medium biofilm-forming ability (++), and NX-04 and NX-02 were weak biofilm formers (+). The OD₅₉₀ readings of the strong biofilm-forming strain NX-01 significantly differed from the medium and weak biofilm-forming strains at 12, 24, and 48 h of biofilm culture (p < 0.01). Additionally, at 36 h of biofilm culture, the OD₅₉₀ readings of the M. gallisepticum strain NX-01 were significantly different from the other four strains (p < 0.01; Figure 3B). The results showed that NX-01 strain was the strongest strain and NX-02 strain was the weakest among the five MG strains. Therefore, these two strains were selected as the subsequent test bacteria.

The matrix growth curve test of the biofilm model in vitro showed that the biofilm formation of the tested MG was divided into the initial adhesion stage, aggregation stage, and biofilm maturation stage. The two tested strains NX-01 and NX-02reached the initial adhesion stage after 12 h culture.

The growth rate of the NX-01 strain was the highest at 12–48 h, and the strain gathered to form microcolonies, corresponding to the aggregation stage of the biofilm. There was no significant change in the number of viable cells of stain at 48–72 h (p > 0.05), corresponding to the maturation stage of the biofilm, at which the microcolony continued to expand and formed a mature biofilm with a dense structure (Figure 3C).

The cells of the NX-02 strain adhered to each other and gathered to form microcolonies within 12–36 h, after which they rapidly multiplied and began to adhere to microplates. During the logarithmic phase of biofilm growth from 36 to 42 h, the formed strain microcolonies fused with each other and developed into biofilm with a mature three-dimensional structure. At 48–72 h, bacterial metabolic activity in the biofilm decreased (Figure 3D). The results showed that the membrane growth rate of strong membrane-forming MG strain NX-01 was faster than that of weak membrane-forming strain (p < 0.01), and the maintenance time of biofilm of NX-01 strain was longer than that of NX-02 strain (p < 0.05).

Scanning electron microscopy (SEM) was used to visually detect the morphology of the biofilm adhesion stage of NX-01 and NX-02 strains. Bacterial aggregates of NX-01 strain was detected across the entire visual field at varying magnifications (Figure 4A). They formed the mature biofilm, which is closely adhered together after being wrapped by the extracellular polysaccharide and other substances secreted by pathogenic bacteria. The three-dimensional mushroom-like structure can even be observed at 50,000× magnification (Figure 4E). The distribution of the NX-02 strain was relatively dispersed in the adhesion stage (Figure 4B), and its cell gathered in some areas forming tiny colonies and organized mature biofilms. The results showed that both NX-01 and NX-02 strains could form biofilms. The biofilm formed by NX-01 strain was multilayer and denser than that of NX-02 strain, which was consistent with the results of the biofilm formation ability test (Figures 4C,D).

The sensitivity results of two MG stains isolated from Ningxia to 17 kinds of antibiotics showed that the order of sensitivity of NX-01 and NX-02 to six antibiotics was similar, which from strong to weak is tetracyclines, quinolones, macrolides, pleulopleuloides, aminoglycosides, and lincomycin (Table 4). The minimum inhibitory concentration of the NX-01 strain was 2 μg/mL, and that of the NX-02 strain was 0.25 μg/mL. The NX-02 strain was more sensitive to antibiotics than the NX-01 strain.

DNA of NX-01 and NX-2 stains were, respectively, extracted and tested the integrity. All DNA sample for sequencing were not degrade in 0.7% nucleic acid gel electrophoresis.

The genome-wide of two MG strains was sequenced using the Illumina HiSeq platform, and the total bases count of NX-01 and NX-02 strains were 1,563,147,314 and 1,184,091,000 bp, and the genome coverage was above 1,000× (Table 5). Compare and evaluate the assembly results after the initial assembly using software SPAdes, NX-01, and NX-02 strains were assembled to obtain 40 and 33 nodes, respectively. The average length of assembled sequences of them is 24608.33 and 28679.33 bp. The maximum length of a single sequences of MG strains NX-01 and NX-02 strains is 147,070 and 17,76,641 bp, respectively. The N50 lengths of them are 67,869 and 101,937 bp. NX-01 and NX-02 strains has been deposited in NCBI (Bio project PRJNA972534 and PRJNA972540).

The genome-wide of the two MG strains showed that the genome size of NX-01 and NX-02 strains were 0.98 and 0.94 Mb, respectively (Table 6), and the GC content of their genome were 31.43 and 31.46%. The predicted CDS numbers in the genomes of NX-01 and NX-02 strains were 814 and 1,548 respectively, and predicted rRNAs were 2 and 3, respectively, and both had the tRNAs of 3.

Annotation results of NR database showed that more than 90% of the sequenced genes of the two isolates strains were derived from genome of reported MG, which proved that both of them were MG.

The functional annotation of COG genes showed (Figures 5A,B) that the number of functional genomes of NX-01 and NX-02 strains were 487 and 648, respectively. The predicted types of functional genes of these two MG strains were identical, whereas the number of certain functional genes differed significantly. The number of the two functional genes, including ribosome structure and biogenesis, and DNA replication and repair, was 40 less in the NX-01 strain than in the NX-02 strain. The number of cell wall or membrane or envelope-related genes of NX-01 strain was higher than that of NX-02 strain. The number of other functional genes of NX-01 was less or equal to that of NX-02. S-adenosylmethionine synthetase and methionyl tRNA synthetase were annotated in the genome-wide of NX-01 and NX-02 strains, both of which are biofilm-related enzymes. A biofilm-related protein, adhesin P1 precursor, was only noted in the genome-wide of the NX-01 strain.

GO database annotation showed (Figures 5C,D) that the functional genes of the two MG strains could be divided into biological process, molecular function, and cell component. NX-01 and NX-02 strains annotated 537 and 947 functional genes, respectively. The same gene may have multiple functions, the total number of genes of NX-01 and NX-02 strains involved in biological processes were 781 and 1,920, involved in molecular function were 1,127 and 1,431, involved in cell components were 1,133 and 1,785, respectively. The most annotated genes of two MG strains,in biological processes category belong to metabolic processes, including 367 genes in NX-01strain and 645 genes in NX-02 strain. The most annotated genes of two MG strains in the category of molecular functions belong to catalytic activity, including 327 genes in NX-01strain and 629 genes in NX-02 strain. The most annotated genes of two MG strains in the category of cell components belong to cells and cell parts, including 343 genes in NX-01 strain and 554 in NX-02 strain. These results indicated that the number of functional genes of NX-01 was significantly different than that of NX-02.

The results of KEGG database annotation results showed that the genes of two MG strains could be divided into five branches, including cell process, environmental information processing, genetic information processing, metabolism, and organism system (Figures 6A,B). The gene number of the NX-01 strain was 302, and that of the NX-02 strain was 502. Metabolism accounted for the highest proportion of genes in the two strains. Metabolism-related genes of NX-01 strain account for 62.3% of the total genes, including energy metabolism, cofactor, and vitamin metabolism, while which of NX-02 strain lack glycan biosynthesis and metabolism genes. The number of metabolism-related genes in NX-02 strain accounted for 55.9% of the total genes. The genes related to genetic information processing accounted for 26.1% in NX-01 strain compared 27.5% in NX-02 strain. Cell process genes account for the least proportion of total genes, with only 0.3% of genes in NX-01 and 0.1% in NX-02. Five genes, oppA, oppD, pdhA, eno, and msbA were annotated in the environmental information processing pathway of two MG strains. Seven genes, manB, pdhB, pdhC, pdhD, RelA, deoA, and gapA, were annotated in the metabolic pathway. One genes, ropS, was annotated in the basic human pathway. All of these genes are biofilm-related gene, and among them, PDH, eno, and gapA participate in multiple pathways at the same time.

The annotation of the VFDB database showed that NX-01 strain has 18 species virulence factors while NX-02 has 22 species (Table 7). NX-01 strain is characterized by Cytadherence organelle, MntA transporter, and RelA virulence factor. Seven toxic factors, including Clp-type ATPase chaperone protein, type III secretion system ATPase VscN, ATPase EscN, and K1 capsule, were unique to the NX-02 strain. Hemolysin factor was the most annotated MG factor in these two strains, which was cytotoxic to various cell types. From the number of noted virulence factors, it was found that the MG strain NX-02 had more virulence factors than the NX-01 strain, which may lead to stronger virulence of the MG strain NX-02 than NX-01.

The annotation of the CARD database showed seven kinds of drug-resistant genes in the M. gallisepticum strain NX-01 and 14 kinds in the MG strain NX-02 (Table 8). Quinolone resistance genes is the most frequent occurring drug resistance gene in two MG strains, which NX-01 strain had 20 quinolone resistance genes, and NX-02 strain had 28. Additional resistance genes were also detected, among which the immunogenic protein EF-Tu accounted for the largest proportion which range from 77.7% to 87.5%. Others genes including specific lincomycin resistance factor ImrD, macrolide resistance factor MacB, aminocoumarin resistance factor alaS, coumarin resistance factor novA, rifamycin resistance factor rpoB, and streptomycin resistance factor rpsL, occurs in the NX-02 strain. The results showed that NX-02 strain annotated more drug resistance factors than NX-01 strain, and they may have different degrees of resistance to antibiotics.

The annotation of the CAZy database showed (Figure 6C) that M. gallisepticum strain NX-01 was annotated to five carbohydrate enzymes (gene number 14), including glycoside hydrolase (GH), glycosyltransferase (GT), carbohydrate esterase (CE), helper activity enzyme (AA), and carbohydrate-binding module (CBM). NX-02 strain had one less carbohydrate-binding module (CBM) and six genes. The number of related enzymes annotated by NX-01 strain was higher than that of the NX-02 strain, especially the carbohydrate-binding module, which may indicate different metabolic capacities.

A phylogenetic tree based on genome-wide construction is shown in Figure 7A. NX-01 and NX-02 strains had the closest homology and were in the same group. They were slightly distant from other MG strains, such as S6, ts-11, 6/85, R(low), NCTC10115, and so on, and were on different evolutionary branches. They were related to M. gallisepticum strains mx-4 and CA06_2006-052-5-2p, and furthest related to M. gallisepticum strains KUVMG01 and faa lab. The results showed that the 12 MG strains formed three large branches without clear distinctions based on geographic location. The two Ningxia MG isolates were closely related to M. gallisepticum Russian strain S6, suggesting that the two Ningxia MG isolates and MG strain S6 may derive from a common ancestor.

The genome-wide circle map of the seven strains of MG (Figure 7B) showed that the genome-wide of NX-01 strain was significantly different from that of the other six strains. Mycoplasma gallisepticum strain NX-01 had fewer gene deletions and had the least difference from M. gallisepticum Russian strain S6. The MG gene differed less between NX-02 and the other five strains were less than NX-01. Compared with NX-01 strain, NX-02 strain had little genetic difference in MG with five strains from other regions. In addition, NX-02 strain has less genome-wide differences from F and R (low) strains in the United States. Five major differences were found in the comparative genome of seven MG strains, which were 490–500, 525–540, 600–615, 782–792, and 882–892 kb fragment regions. The results indicated that the genetic information of MG isolates from Ningxia had rich regional specificity.

The two MG were from Ningxia were different from the five endemic M. gallisepticum strains [6/85, F, MX-4, R(low), and S6] in five regions of the world, and the two MG isolates were different from M. gallisepticum strain F, but little from M. gallisepticum strains R(low) and S6.

In the fragment region of 490–500 kb (Figure 8A), the NX-01 and NX-02 strains showed more long fragment deletion than the other five strains. In addition, NX-01 and NX-02 strains showed translocation of MG recombinase Polymerase recA, rpmA, and nfo compared with other five MG strains. The direction of NX-01 and NX-02 strains were consistent with S6, R(low), and 6/85 gene fragments. Compared with MX-4 and F strains, NX-01 and NX-02 strains had similar inversions of larger gene segments.

In the 520–540 kb fragment (Figure 8B), the two MG strains still had longer fragment deletion than the other five MG strains, and the same genes, such as rsmH, scpB, and plsC, were displaced. Mycoplasma gallisepticum NX-02 strain-specific fssZ can cause host immunogenicity. Two MG strains were in the same direction as M. gallisepticum strains S6, R(low), and 6/85, and the whole segment was inverted with M. gallisepticum strain MX-4 and F.

The 600–615 kb sequences were compared (Figure 8C). Besides removing longer fragments in two of the MG strains compared to the other five, the NX-01 strain is distinguished by the absence of pepE, geoA, and pgcA genes as opposed to the M. gallisepticum strain NX-02. Notably, the pepE was associated with the adhesion process of mycoplasma. Mycoplasma gallisepticum strain NX-02 had more deoA, pgcA, and tpiA genes than the other six strains. The two MG and M. gallisepticum strain F showed an inversion of the entire segment, and the direction of the gene segments of the other four MG strains was the same.

The 782–792 kb sequences were compared (Figure 8D), which revealed found that NX-01 strain had more deletion of MG fragments than the other six strains. NX-01 strain showed more ecfA1 gene than the other five MG isolates from different regions, and the similar genes cdd and gpml were displaced. The ecfT, pgcA, and deoA genes were unique to the NX-02 strain, and the similar gene truA was translocated. The direction of the segments of two MG strains was consistent with that of 6/85, R(low), and S6 strains, and the whole segment was inverted with F strain.

In the 882–892 kb difference region (Figure 8E), the direction of MG gene fragments in two strains was consistent with that in the other five strains. However, fragment deletion was the largest among the five different regions. The same genes rplK, pth, and rpoE of the two MG strains and the other five MG strains were shifted to different degrees. The length of the tilS in M. gallisepticum strain NX-02 was shorter and more translocated than in the other five MG strains.