Introduction

The pathogen Haemophilus parasuis (HPS) is among the most commonly identified Gram-negative bacteria mainly causing serofibrinous polyserositis and arthritis which leads to major economic losses in the swine industry worldwide1,2,3. Of the 15 serovars, serovars 4 and 5 are widely associated with epidemics and serovar 5 is particularly highly virulent in China4. Various antimicrobial agents, including macrolides, β-lactams, phenicols, potentiated sulfonamides and tetracyclines, have been administered for the treatment and prevention of respiratory infections caused by HPS 5,6,7. Antimicrobials were thought to be the most powerful and typical way to combat HPS invasion8. However, the prolonged exposure of pathogens to drugs can induce resistance9, 10. In recent years, clinical isolates resistant to antimicrobials have been reported in Switzerland, the United Kingdom and Spain. It was found that clinical HPS exhibited high and extensive resistance to enrofloxacin, trimethoprim, sulfamethoxazole, tilmicosin and tulathromycin7, 11.

Tildipirosin, a new 16-membered ring macrolide, is a semisynthetic tylosin developed to treat respiratory pathogens. However, the resistance of Pasteurella multocida (PM) to macrolides including tildipirosin, tilmicosin and gamithromycin has previously been reported. Several resistant genes have been identified, such as msr(E), mph(E) and erm(42) 12, 13. HPS, belongs to the order Pasteurellales of family Pasteurellaceae which is made up of at least 15 genera and over 70 species14, has also been isolated from diseased swine and identified with different levels of sensitivity (MIC, minimal inhibitory concentration) to tildipirosin15. The resistance characteristics of HPS to different antibiotics including fluoroquinolone, marcolides, tetracycline and beta-lactam has been investigated in previously described reports and some classical resistant genes such as acrAB, Tet B, Tet A, ErmB, etc16,17,18,19. The resistance mechanism of HPS to macrolides has been associated with pathways of the amino acid ATP-binding cassette (ABC) transport system (HAPS_2069) and the metabolite transporter superfamily (HAPS_2067, HAPS_2068). However, no studies have been conducted on the mechanisms of tildipirosin resistance in HPS. In the current study, several resistance HPS strains were isolated in diseased swine and induced in lab, and a transcriptomic approach was applied to achieve a genetically tildipirosin-resistant characteristic and revealed promising therapeutic targets to combat resistance20.

Transcriptional profiling is a useful tool for rapidly and simultaneously identifying large numbers of genetic determinants. Transcriptional profiling analysis provides distinct and detailed genomic-level information related to specific pathogenic mechanisms involving virulence factors and resistance genes8, 21. The extent of bacterial mechanistic response to antibiotic invasion has been revealed to be time- or dose-dependent in previous reports22, 23. Thus, a systematic approach of transcriptional profiling may aid the discovery of the resistance mechanisms of HPS to tildipirosin.

The objective of this study was therefore to use an RNA sequence method to systematically analyse the altered response of the tildipirosin-resistant strain’s (JS32) transcriptome and morphological characteristics compared to JS0135. These findings will help us to better understand the tildipirosin resistance mechanism in HPS which could then contribute to reasonable administration of tildipirosin and the development of methods used to prevent or reduce resistance in HPS.

Results

Minimal inhibitory concentration (MIC) determination, growth comparison and transmission electron microscope (TEM) analysis

JS32 is a tildipirosin-resistant strain which was obtained after exposure to progressive concentrations of tildipirosin as described in detail in the experimental procedures. HB32 was obtained from clinical isolation. The MICs of JS0135, JS32 and HB32 were 0.125, 32 and 32 μg/ml respectively, determined with broth microdilution assays. When JS0135 was exposed to tildipirosin, it exhibited increased resistance (MIC ≥ 128 μg/ml). However, the high level of resistance was not maintained after a single passage of cells in growth medium without tildipirosin. JS32 kept stable resistance (MIC = 32 μg/ml). The serovars of JS0135, JS32 and HB32 were amplified by PCR with the appropriate primers listed in Table 1 and were identified as serovars 4, 4 (320 bp) and 13 (840 bp), respectively (Supplementary Figure S1).

Table 1 Primers of RT-qPCR and serotype.
Full size table

The growth characteristics of JS0135, JS32 and HB32 were compared by measuring OD600nm at different time points. No differences were observed between JS0135 and HB32, but the growth rate of JS32 was the fastest (Fig. 1). JS32, the induced tildipirosin-resistant strain, achieved logarithmic phase growth at 8 h, while JS0135 and HB32 did so at 12 h. Although the three strains entered into stationary phase at 18 h, the total bacteria count of JS32 was significantly less than JS0135 and HB32 which was similar to previous research in response to tilmicosin20.

Figure 1
figure 1

Growth curves of JS0135, JS32 and HB32. *Presents statistically significant p ≤ 0.05, **presents extremely significant p ≤ 0.01.

Full size image

TEM was used to investigate morphologic diversity between sensitive cells (JS0135) and resistant cells (JS32 and HB32). Three samples were collected at 12 h (exponential phase of growth), based on the growth curves. The TEM results showed that the membranes of induced (JS32) and wild-type (HB32) resistant bacteria had smoother margins than the control sensitive bacteria (JS0135), and the membrane of JS32 was the smoothest among the three bacteria (Fig. 2). Similar changes between resistant and sensitive HPS were reported in previous research8, 24.

Figure 2
figure 2

Comparison of transmission electron microscope: (A) presents JS0135, (B) presents JS32, (C) presents HB32. Red arrow pointed to the membrane of strains.

Full size image

Transcriptome sequencing annotation

A total of 18,620,015 ± 158,693 raw reads and 32,093,782 ± 791,754 reads with Q20 values of 93.46% ± 0.004 and 94.90% ± 0.011 in control (JS0135) and treatment groups (JS32), respectively; 15,966,164 ± 201,137 and 27,829,816 ± 1065685 (means ± SD) high-quality mapped reads were obtained in the control and treatment groups, respectively, and mapping ratios of 95.21% ± 0.001 and 96.14% ± 0.005 were obtained after filtering adapters and trimming ambiguous results (Table 2). Compared to the control group, the treatment group (JS32) had a significantly different increase (p ≤ 0.01) in raw reads, clean reads, all reads and mapped reads, but no differences in Q20 value and mapping ratio.

Table 2 Statistical summary of RNA–seq datasets in JS32 and JS0135.
Full size table

Differential expression and functional analysis of genes

Differential analysis of the transcript expression profiles revealed that 349 genes, including 41 novel genes, were upregulated (FC ≥ 2); 113 genes, including 10 novel genes, were dwonregulated (FC ≤ 0.5); and as a whole the treatment group (JS32) were more responsive than the control group (JS0135) (Supplementary Figure S2). The full list of DE transcripts can be seen in Supplementary File 1. GO classification and Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway analysis were performed as bioinformatics tools to explore the potential roles of DE genes in the resistance mechanism. Of 462 DE genes, 321 (69.7%) were assigned GO categories, and were further classified into three types: cellular component, biological process and molecular function (Fig. 3a). Within the biological process group, the most abundant categories were cellular process, metabolic process and single-organism process; other appealing categories included biological regulation, locomotion and signalling. In the cellular component group, cell part, cell membrane and membrane part were the most highly described subcategories. From those three groups, 17 subcategories were in the biological process group, 11 subcategories were in the cellular component group, and 9 subcategories were in the molecular function group, and up- and downregulation were shown in the subcategories (Fig. 3b and Supplementary File 2).

Figure 3
figure 3

GO functional categories analysis (A), and up, down regulation of DE genes in subcategories statistics (B). A, the top groups in the three main categories: biological process (17), cellular component (11), molecular function (9) are summarized. The x-axis presents the categories, and the y-axis presents the number of genes in the categories. B, the number of up and down regulation genes are summarized in the subcategories belonging to the categories of A. Pink in X axis label represented biological process; green in the X axis label represented cellular component; blue in the X axis label represented molecular function.

Full size image

According to the KEGG analysis, 116 DE genes were found to be classified into four parts and involved in 64 different pathways. From those four groups, one categories was in the cellular processes group, four categories were in the environmental information processing group, nine categories were in the genetic information processing group, 49 categories were in the metabolism group (Fig. 4a ), and up- and downregulation were in the subcategories (Fig. 4b and Supplementary File 3). The most abundant pathways in the KEGG analysis were metabolic pathways, biosynthesis of secondary metabolites, ribosomes, ABC transporters, biosynthesis of antibiotics, purine metabolism, microbial metabolism in diverse environments, quorum sensing and glycerophospholipid metabolism; other appealing pathways included aminoacyl-tRNA biosynthesis and cell cycle. Furthermore, the DE transcripts related to the GO and KEGG pathway results on resistance were involved in metabolism, ribosome, ABC transporters, metabolic pathways, the phosphotransferase system (PTS) and cationic antimicrobial peptide (CAMP) resistance. RNA-seq was displayed in Supplement File 1. In the total gene expression comparison of JS32 and JS0135, we selected resistance related genes with the value FC ≥ 2 or ≤ 0.5 (Tables 3 and 4).

Figure 4
figure 4

KEGG pathway classification analysis (A), and up, down regulation of DE genes in subcategories statistics (B). (A) The DE genes in the four pathways processes: metabolism (64), genetic information processes (27), environment information processes (27), cellular processes (1), are summarized. The x-axis presents categories pathways, and the y-axis presents the number genes in categories pathway. (B) the numbers of up and down regulation genes are summarized in the subcategories pathways belonging to the categories of A. Purple in the X axis label represented metabolism; blue in the X axis label represented genetic information processing; yellow in the X axis label represented environmental information processing; pink in the X axis label represented cellular processes.

Full size image
Table 3 The important up regulation genes of JS32 compared to JS0135 grouped by GO and KEGG pathways of interest.
Full size table
Table 4 The important down regulation genes of JS32 compared to JS0135 grouped by GO and KEGG pathways of interest.
Full size table

Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) analysis of the relationships between DE genes of the main pathways

STRING is a web-based interface which can predict protein associations – direct physical binding and indirect interaction – such as participation in the same metabolic pathway or cellular process on the basis of genomic context, co-expression and data from reported literature (https://string.embl.de/)25,26,27. DE genes were analysed with STRING to predict the network of proteins encoded by DE genes. DE genes involved in the three main pathways (metabolic pathway, ABC transporters, ribosomes) related to resistance were selected for STRING analysis, using the Sus scrofa database. The network of predicted associations for all upregulated and downregulated DE genes encoding proteins and string symbols were shown in Supplementary Files 4–6. The detailed fold changes of major upregulated and downregulated DE genes (FC ≥ 2 or FC ≤ 0.5) of JS32 compared to JS0135 were also shown in Tables 3–4. Among these DE genes, most molecules were key molecules that link to each others, while several encoded proteins which were not linked to each other, indicating that their functions were unrelated or unknown according to the STRING analysis results. As shown in Figs 5–7 (FC ≥ 2 or FC ≤ 0.5), the DE genes of the three main resistance-related pathways encoded proteins which were associated with each other contributing to the resistance of HPS to tildipirosin together. The 40 DE genes from the Tables 3 and 4 encoded proteins associated with the metabolic pathway including 30 upregulated and 10 downregulated genes were selected for STRING analysis. Of the 40 genes, 4 DE genes were not found in the STRING database and the other 36 genes were shown in the Fig. 5. Among the 16 DE genes encoded ABC transporter proteins including 11 upregulated and 5 downregulated genes, 4 DE gene were not found in the STRING database, and the network of the other 12 genes were shown in the Fig. 6. The network of 15 DE genes encoded ribosome proteins including 11 upregulated and 4 downregulated genes were shown in the Fig. 7. All of them were linked with each other to regulate the resistance of HPS to tildipirosin.

Figure 5
figure 5

STRING analysis of the relationship between main 40 DE genes in metabolic pathways. The downregulated genes were marked with red, and the others were upregulated genes.

Full size image
Figure 6
figure 6

STRING analysis of the relationship between 16 DE genes in ABC transporter. The downregulated genes were marked with red, and the others were upregulated genes.

Full size image
Figure 7
figure 7

STRING analysis of the relationship between 15 DE genes in ribosome. The downregulated genes were marked with red, and the others were upregulated genes.

Full size image

Validation by real-time quantitative PCR (RT-qPCR)

For verification of the RNA sequencing results, ten of the DE genes and three samples including JS0135, JS32 and HB32, were selected on the basis of their importance as resistance determinants. Among the ten tested genes, HAPS_RS09315, HAPS_RS09320, HAPS_RS11130, HAPS_RS06145, glmM, HAPS_RS04930, HAPS_RS03600, HAPS_RS03625, HAPS_RS07815 and HAPS_RS10945 of JS32 had fold changes of infinity (499,108), infinity (114,954), 1270, 98, 158, 676, 30, 683 and 460, respectively, when their expression levels were compared in the test and reference control. The fold changes of ten HB32 genes were similar to those of JS32.

Discussion

In the present study, JS0135 and HB32 were used to investigate the resistance mechanism to tildipirosin in HPS. JS32 was induced from JS0135 and could grow well on tryptone soy agar containing 256 MIC tildipirosin. The total bacterial count of JS32 was significantly (p ≤ 0.01) decreased compared to JS0135 and HB32, but attained logarithmic growth phase faster than the others; the growth curve of HB32 was similar to JS0135 (Fig. 1). The results of the current study were similar to those from the research reported by Chunmei Wang8. This variation might be associated with tildipirosin stimulation and the DNA replication pathway, which are involved in the downregulation of rnhB which expresses ribonuclease HII protein and is essential for growth according to previous reports (Table 4)28,29,30,31. The growth difference between tildipirosin-resistant and sensitive strains required further research. Three HPS serovars were indentified with a previously described multiplex PCR method which is faster, more sensitive and more specific than indirect hemagglutination (IHA)32. The results in Supplementary Figure S1 distinctly show that JS0135, JS32 and HB32 are serovars 4, 4 and 13, respectively.

According to a previous study by Chunmei Wang in 2014, and the significant KEGG membrane transport pathway analysis in Fig. 4, JS0135, JS32 and HB32 were selected to explore the resistance mechanism by observing membrane morphology diversity with SEM. The SEM results showed that the outer surfaces of induced and wild-type strains JS32 and HB32 were smoother than the control (JS0135), but no contrast was found between JS32 and HB32 (Fig. 2). Similar changes in the ultrastructure of CB-resistant HPS have been reported previously8, 24. The variance between resistant and sensitive HPS might be caused by membrane proteins including those encoded by the upregulated genes HASP_RS10075, HASP_RS11135, HASP_RS07320, HASP_RS03695, HASP_RS08120 and HASP_RS05335 (Supplementary File 1). The similarity of the JS32 and HB32 induced and wild-type tildipirosin-resistant bacteria suggest the same resistance mechanism from the morphology. A known membrane protein gene HAPS_RS01150 is related to resistance in Escherichia coli, encoding an outer membrane lipoprotein involved in copper homeostasis and adhesion; its overproduction was found to increase multidrug resistance and copper through activation of genes encoding the AcrD and Mdt ABC drug efflux pumps33, 34. HAPS_RS01150 (1.003 fold change) in JS32 did not show any upregulation in the present study, but other genes encoding proteins displayed up- and downregulation (FC ≥ 2 or FC ≤ 0.5), and it is necessary to study these genes further.

In previous reports, CAMPs were shown to play an important role in inhibiting colonization and clearance of infections; furthermore Gram-negative bacteria represent a major target for CAMPs. However, the development of CAMP resistance permits Gram-negative bacteria to avoid being killed by both the host immune system and antibiotics35, 36. CAMP resistance genes including HAPS_RS07240, HAPS_RS11325 and HAPS_RS06175 encoding relative resistance proteins exhibited upregulation of gene expression (≥twofold changes), shown in Table 3.

The GO and KEGG analysis results are shown in Tables 3, 4; molecular function, biological process, cellular component, integral component of membrane, plasma membrane, transport, transposase activity and DNA-mediated transposition were the most abundant GO classification terms. Metabolic pathways, biosynthesis of secondary metabolites, ribosome, ABC transporters, biosynthesis of antibiotics, purine metabolism, microbial metabolism in diverse environments, quorum sensing and glycerophospholipid metabolism were the most abundant KEGG classification pathways. In these results, increased DE in the treatment group was focused on metabolic pathways, ABC transporters and ribosomes, while decreased DE was focused on metabolic pathways, PTS, ABC transporters and ribosomes (Fig. 4b). These profiles of major upregulated and downregulated genes in GO and KEGG analysis in the Tables 3, 4 have enabled us for the first time to systematically elucidate the resistance of HPS to tildipirosin. The following paragraphs analysed the possible mechanisms of HPS resistance to tildipirosin from metabolic, PTS, ABC transporters and ribosome pathways.

The genes involved in metabolic pathways, HAPS_RS09315, HAPS_RS09320, HAPS_RS08960, HAPS_RS08955 and HAPS_RS08950, encoding restriction endonuclease subunit M, DNA cytosine methyltransferase, hydroxyethylthiazole kinase, hydroxymethylpyrimidine kinase and thiamine phosphate synthase, respectively, were infinitely upregulated (Table 3); this was verified by RT-qPCR, which indicates that the RNA sequence results were reliable (Fig. 8). Among these genes, DNA cytosine methyltransferase is a key factor as a marker for the presence of a family of phage-like elements, which confer macrolide resistance in streptococci and resistance to target site methylation in PM 13, 37. Moreover, nucleotide methylation can also offer antibiotic resistance, such as 16S rRNA methyltransferase in Enterobacteriaceae38. It has been previously been reported that the upregulation of thiamine phosphate synthase can cause an increase in resistance to multiple stresses in Schizosaccharomyces, and thiamine supplementation might also contribute to chemotherapy resistance in cancer cells39, 40. Another key upregulated gene glmM (2.6-fold change), encoding phosphoglucosamine mutase, has been demonstrated to contribute to the resistance of Streptococcus, and is the drug target for regulating resistance. In addition, glmM is directly upstream of a multiple repeat polypeptide essential for the expression of methicillin resistance in Staphylococcus aureus 41,42,43,44. The other upregulated genes have not been reported, but also might contribute to regulate metabolic pathways related to bacterial resistance to tildipirosin which were in need of verification in the future.

Figure 8
figure 8

The differential expression on relative mRNA abundance of Ten genes in JS32 and HB32 compared with JS0135. Control, the value = 1, Values are mean ± SD. *Presents statistically significant p ≤ 0.05, **presents extremely significant p ≤ 0.01.

Full size image

Other upregulated ribosome and ABC transporter pathway genes encoding ribosomal proteins, transporter permeases and membrane proteins, including HAPS_RS07815, HAPS_RS07810, rpsJ, rplD, HAPS_RS07825, HAPS_RS07790, HAPS_RS07805 and HAPS_RS07780; and HAPS_RS10945, HAPS_RS03625, HAPS_RS05335, HAPS_RS03630, HAPS_RS00310, HAPS_RS05165 and HAPS_RS00315, respectively, were found to have a significant effect on the treatment group. Previous reports have stated that tigecycline resistance is associated with mutations in rpsJ in Klebsiella pneumoniae. RpsJ acts as general target of tigecycline adaption and a marker for alterations in antibiotic resistance in bacteria; the V57L mutation in rpsJ might cause weaker binding of tigecycline to 16S rRNA, leading to tigecycline resistance45,46,47,48. In RplD, encoding the ribosomal protein L4, it has also been found that the A2059G mutation confers resistance to macrolides and lincosamides12, 49,50,51. Other genes relative to ribosomes, encoding ribosomal proteins, are also concerned with resistance. Dennis conducted a study on the E. coli response to chloramphenicol52; when Gram-negative bacteria, such as HPS, experience low levels of translation inhibition, a compensatory mechanism might be triggered in which the synthesis of ribosomal proteins is initially upregulated, but as the inhibition stress increases this compensation fails to keep pace and the cells succumb to antibiotic killing20, 52. There were 11 significantly upregulated genes related to the ABC transport system in the treatment group (Table 3). The bacterial cell envelope is a target of many antibiotics, and disruption of its structure inhibits transmembrane transport functions and impairs normal physiological functions. The key transport systems critical for bacterial viability and survival are the ABC transporter pathways53. ABC transporters play a significant role in bacteria, conferring multidrug resistance (MDR) through overexpression as described in previous reports54. Moreover, the active movement of compounds across membranes carried out by ABC transporters can cause drug resistance in anti-infective therapies55. Resistance against antimicrobial peptides in many firmicutes bacteria is mediated by an ABC transporter56. ABC transporters are involved in secretion of the antibiotic through the cell membrane and also contribute to acquisition of antibiotic resistance. ABC transporters were the first proteins to be implicated in the mechanism of resistance to macrolides, as described in antibiotic-producing actinomycetes57, 58. The variation between treatment and control groups was also caused by the ABC transporter cell membrane proteins expressed, as described in Fig. 6. Although DE genes in the ABC transporter pathway have not been reported in resistance, these are novel genes related to the resistance mechanism, worth exploring further.

Other downregulated genes belonging to the PTS, metabolism, ribosome and ABC transport pathways are shown in Table 4. The PTS system is responsible for the transport of a variety of carbohydrates in prokaryotes. PTS components participate in signal transduction, chemotaxis and the regulation of essential physiological processes59, 60. As for downregulation, reduced expression of ABC transporter genes (ABC subfamily) is tightly linked to Cry1Ac resistance in Plutella xylostella 61. All downregulated genes in these pathways, such as metQ, MetN, metF and rpmE, contributed to the regulation of resistance to tildipirosin in this study, shown in Table 4. Meanwhile, the STRING analysis indicated that the main up and down-regulated DE genes encoded proteins which could interact with the metabolic pathway (Fig. 5), ABC transporters (Fig. 6), ribosomes (Fig. 7) and PTS, regulating these genes or other cells to facilitate the resistance of tildipirosin in HPS. HAPS_RS08950, HAPS_RS08955 and HAPS_RS08960 which were part of metabolic pathway encoded thiamine phosphate synthase, hydroxymethylpyrimidine and hydroxyethylthiazole kinase, respectively were associated with each other immediately whose upregulated fold changes were infinite in the Fig. 5 and Table 3. Meanwhile, downregulated genes of metN, HAPS_RS02205 and MetQ belonging to ABC transporters pathway in the Fig. 6 were linked with each other, and all up and down regulated genes of ribosome pathway were connected with each other closely in the Fig. 7. All of these key genes regulated and controlled the resistance of HPS to tildipirosin together, especially for the upregulated genes of metabolic pathway who may contributed to resistance of HPS crucially.

Ten selected genes from the transcriptome profiling in Table 3 were selected for RNA sequence validation by RT-qPCR. There was the same trend of upregulation, but a difference in fold changes in these genes between transcriptome and RT-qPCR analysis, shown in Fig. 8. The main reasons were different batches of samples resulting in fold change variation.

Concluding our findings, the data obtained from transcriptional profiling of JS32 and JS0135 provide new sights into the complex mechanisms underlying the general response to tildipirosin treatment. In addition, distinctive DE genes in the treatment group indicate that more attention should be paid to a new resistance factor metabolic pathway, particularly related to the upregulated genes (HAPS_RS09315, HAPS_RS09320, HAPS_RS08950 and HAPS_RS08955) which are overexpressed infinitely. The other new genes HAPS_RS03625 and HAPS_RS04930 (fold changes > 500, Fig. 8) involved in ribosomes, ABC transport and CAMP, which are interrelated closely as shown in Figs 5–7, are also worthy of future study. The new tildipirosin resistance mechanisms in HPS are complex, and this study provides a new perspective to study macrolide resistance. More attention to study at the protein level is needed to investigate the expression of resistance genes.

Materials and Methods

Bacterial strains and antibiotics

HPS JS0135 was obtained from the State Key Laboratory of Microbiology at Huazhong Agricultural University; HB32 was isolated from the lung of a diseased piglet in Jiangsu and Hubei, China. They were identified as serovars 4 and 13, respectively, by PCR with a previously described method32, 62. The primers were designed as shown in Table 1. HPS was subcultured in tryptone soya agar (TSA) and tryptone soya broth (TSB) (Qingdao Hai Bo Biological Technology Co., Ltd., Shangdong, China) supplemented with 5% fetal bovine serum (Zhejiang Tianhang Biotechnology Co., Ltd., Zhejiang, China) and 10 μg/ml nicotinamide adenine dinucleotide (NAD) (Qingdao Hope Bio-Technology Co., Ltd., Shandong, China). Tildipirosin with >99.5% purity was used, donated from Hubei Huisheng Biological Technology Company (Hubei, China).

Determination of induced and natural resistance

The MICs of JS0135, JS32 and HB32 were determined with twofold broth dilution (0.0625–32 μg/ml) according to the CLSI M07-A9 standard. Enterococcus faecalis (ATCC 29212) was used as the quality control (QC) strain to detect the credibility of susceptibility testing63. JS32 was induced from JS0135 by incubation with increasing tildipirosin concentrations (from 0.0625 to 64 μg/ml)8. One colony of JS0135 (MIC = 0.125) was incubated into TSB with 0.5 MIC tildipirosin at 37 °C with shaking (220 rpm) for 12 h. When induced colonies had grown stable, cultures were inoculated into TSB with the next highest concentration of tildipirosin64. At last, one colony (MIC = 32) remained with high resistance stability, and was named JS32. HB32 (MIC = 32), a clinical isolate, is a naturally resistant strain. MICs for tildipirosin to HPS were determined by using agar dilution method as recommended by the Clinical and Laboratory Standards Institute (CLSI) M31-A3 guidelines. All experiments involved in MIC determination were preformed according to these guidelines.

Growth curve comparison

JS32, HB32 and JS0135 were inoculated into TSB cultures for more than three generations until stable growth was achieved. Then, 100 μl of the three bacterial cultures (1 × 106 CFU/ml) was selected to inoculate into new 100 ml TSB cultures. Each newly selected strain was incubated on a shaker at 220 rpm at 37 °C for 24 h. Growth curves were determined by measuring the optical density (600 nm) of the cultures every 2 h with a spectrophotometer (UV2100, Shanghai, China).

Transmission Electron Microscopy analysis

Bacteria (JS32, JS0135, HB32) were cultured in TSB to reach mid-logarithmic phase (12 h). Three cultures were centrifuged and washed with phosphate-buffered saline (PBS) twice. The washed bacterial sediment was fixed with 2.5% buffered glutaraldehyde for 1 h, and then fixed in 1% buffered osmium tetroxide for 1 h. The fixed samples were dehydrated through a graded ethanol series, and embedded in resin. The morphology of JS32, JS0135 and HB32 was observed using a Tecnai G2 20 S-TWIN transmission electron microscopy (TEM) (JSM-6390LV, NTC, Japan) at an acceleration voltage of 200 kv (FEI, Hillshoro, Oreg, USA).

Transcriptome analysis

In this study, an RNA sequence analysis was prepared and submitted to Shanghai Biochip Corporation (Shanghai, China) for mRNA purification, library preparation and sequencing. In brief, bacterial cultures (JS32, JS0135, HB32) were centrifuged for 10 min (3000 g at 4 °C). Total RNA of bacterial samples was extracted and purified with RNAiso Plus Reagent (TaKaRa Biotechnology Co., Ltd, USA) and DNase (Qiagen, Germany) according to the manufacturer’s instructions20. The remaining DNA was removed by RNase-free DNase I (Ambion Inc., Texas, USA). RNA concentration and purity were evaluated by A260/A280 spectrophotometer readings (NanoDrop 2000, Thermo Fisher Scientific Inc., USA) and agarose gel electrophoresis, respectively. Ribosomal RNA was removed from the total RNA with Ribozero Kit was followed with the strand specific RNA-seq protocol on Illumina Hiseq. 2500 platform (paired-edn sequencing; 100 bp fragments) at Shanghai Biochip Corporation. Firstly, strand cDNA synthesis was conducted with using SuperScriptII (Invitrogen, Carlsbad, CA) in the presence of random hexamer primers. Secondly, another cDNA was synthesized before end-repair and dA-tailing. DNA fragment ligation was performed with TruSeq adapter and amplified with TruSeq PCR primers for sequencing. Reads longer than 35 nt and ≤2 N (ambiguous nucleotides) were retained. Meanwhile, paired reads that got mapped to sliva database (https://www.arb-silva.de/download/arb-files/) were removed.

Each gene expression in different samples were transformed to counts per gene (CPG) by DE sequence package with blind and fit-only parameter65. Mean and SD of CPG expression were calculated for JS32 and JS0135 from their respective repeats and compared to check the DE genes. Genes with a fold change ≥2 and q-value ≤ 0.05 were selected for analysis, since a 1.5-fold change in transcription level was regarded as biologically significant in previous studies66, 67. DE analysis of the transcripts was conducted with the R package DESEq68. A transcript was considered to have significant DE if the false discovery rate (FDR) was ≤0.05. The data had been deposited in Gene Expression Omnibus (GEO) and were accessible through accession number GSE42814 (https://www.ncbi.nlm.nih.gov/gds/?term=SH0165). GO, as an international standardized system for a functional classification of genes, provided an updated terminology and comprehensively described the properties of genes and their products in the organism. KEGG database (https://www.genome.jp/kegg) was utilized to find the linkage of the DE with different pathways. Functional classification of transcripts with significant DE was conducted with Blast2GO software and KEGG pathway analysis. Associations of the proteins encoded by DE genes were analysed with STRING (http://www.string-db.org/)25.

RT-qPCR analysis

Ten genes (HAPS_RS03625, HAPS_RS11130, glmM, HAPS_RS06145, HAPS_RS04930, HAPS_RS10945, HAPS_RS07815, HAPS_RS03600, HAPS_RS09315, HAPS_RS09320) encoding proteins related to the resistance mechanism of HPS were selected for validation of RNA sequence results with RT-qPCR (CFX 384, Bio-Rad). Total RNA was extracted from JS32, JS0135 and HB32. RT-qPCR was performed in triplicate as described previously65, 69. All primers were originally designed by the NCBI online primer-blast function, as shown in Table 1 (https://www.ncbi.nlm.nih.gov/). The thermal cycler conditions were as follows: denaturation at 95 °C for 10 s, annealing at 56 °C for 20 s and extension at 72 °C for 20 s. The 2−ΔΔCt method was used for quantification with 16S rRNA as a reference gene, and the relative abundance was normalized to the control. The fold changes were calculated by the 2−ΔΔCt formula70.

Statistical analysis

Statistical analysis were conducted with using SPSS version 22.0 (IBM Corp., Armonk, NY, USA). The two-tailed t-test was applied to estimate the mean ± standard deviation (SD) and significant difference of RNA-seq and RT-qPCR results. A p value of ≤0.05 was considered to indicate a statically significant result. *p ≤ 0.05 and **p ≤ 0.01.

Ethic Statement

The animals which were used to isolate HPS in this study were conducted according to relevant guidelines and regulations of Animal Care Center, Hubei Science and Technology Agency in China (SYXK 2013-0044) and animal housing care and experimental protocol were conducted according to the regulation of experimental animal usage in Hubei province of China. In addition, the protocol was approved by the Ethics Committee of Huazhong Agricultural University.