B. melitensis has an adaptive response and is resistant to low levels of rifampicin in the early stages of growth

The results MIC of rifampicin for B. melitensis bv.3 Ether was 0.5 μg/mL; based on this value and CLSI rifampicin breakpoints for slow-growing bacteria (MIC value ≥ 2 μg/mL), this study chose 1 μg/mL rifampicin for culture and transcriptome sequencing analysis. The growth characteristics of Brucella strains were determined in Brucella broth media without rifampicin and with 1 μg/mL rifampicin. Significant variations were observed between different media (Fig 1A). In the presence of rifampicin, the Brucella strains grew slowly for the first 24 h. At this time, the Brucella strains under normal medium were in the logarithmic phase, and those under rifampicin had barely replicated. At 24 h post-infection, the strains grown with rifampicin began to replicate, whereas the strains grown in normal medium were in the late logarithmic phase. After 96 h, the strains reached the plateau phase at low-level rifampicin. These results suggested that low-level rifampicin inhibited the in vitro growth of Brucella spp. at the early stages.

Fig 1

Growth characteristics and variations in the rpoB gene under normal and low-level rifampicin medium.

(A) shows the growth characteristics of all samples. For the curve, BM3 was the parent strain grown without rifampicin and BM3-RIF was the parent strain grown with rifampicin. (B) shows the variation of parent strain grown with rifampicin in rpoB by sequencing at different time point.

This study also detected the variation in the rpoB gene by sequencing at different times during growth. The results showed that no variations in the rpoB gene were observed during the first 48 h, while variation was observed in the rpoB gene at 60 h, and this variation was located at nucleic acid position 1606 (C→T) (Fig 1B).

Compared with the downloaded genome sequences, the Brucella strain in this study had 11 SNPs by whole genome sequencing. There were 8 SNPs before variations found in the presence of 1 μg/mL rifampicin, while 11 SNPs were found after variations were observed (Table 1). Seven SNPs were common among all samples. Compared with the downloaded genome sequences, the SNP at position 140902 of chromosome 1 was detected in the normal medium but not in the medium with rifampicin. This study also detected the SNP at position 832186 of chromosome 2 in the medium with rifampicin but not in the normal medium. Importantly, the SNP at position 187170 of chromosome 1 was only identified in the sample in the medium supplemented with rifampicin when the variation in the rpoB gene was observed. The above results showed that B. melitensis had an adaptive resistance response to low levels of rifampicin in the early stages of growth. At the same time, the SNPs in the rpoB gene changed in the late stages of growth when incubated at 37°C with shaking.

Table 1

SNPs in B. melitensis under low-level rifampicin and normal media.

Chromosome	BM3T	BM31 (no variations)	BM31 (with variations)	Annotation
Etherchromosome1	127451	127451	127451	intergenic region#
Etherchromosome1	140902	/	/	intergenic region
Etherchromosome1	172495	172495	172495	intergenic region
Etherchromosome1	/	/	187170*	missense variation
Etherchromosome1	815415	815415	815415	intergenic region
Etherchromosome1	998263	/	/	synonymous variation
Etherchromosome1	1089097	/	1089097	missense variation
Etherchromosome1	1380326	/	1380326	missense variation
Etherchromosome2	84544	84544	84544	intergenic region
Etherchromosome2	117849	117849	117849	intergenic region
Etherchromosome2	531185	531185	531185	intergenic region
Etherchromosome2	/	832186	832186	frameshift variation
Etherchromosome2	1079291	1079291	1079291	intergenic region

Note: All samples were compared with B. melitensis bv.3 Ether (reference genome). BM31 (no variations) is no variation had been observed in rpoB gene, and BM31 (no variations) is variations had been observed in rpoB gene.

* This SNP was the variation of rpoB gene located at nucleic acid position 1606.

# Intergenic region shows the SNP was located between genes; missense/ synonymous/ frameshift variation shows the SNP was located in the gene.

Genome-wide identification of B. melitensis genes that respond to rifampicin conditions before the variation in rpoB was observed

This study analyzed the effects of rifampicin on gene expression in B. melitensis without variation in the rpoB gene using transcriptome sequencing. The total clean bases of the control (BM3T) and rifampicin-challenged (BM31) samples were 6.66 and 6.49 Gb, respectively. The RNA-seq results showed that exposure to rifampicin led to significantly changed transcript levels of genes with various functions. Overall, 122 genes were up- or downregulated (Fig 2) (S4 Table).

Fig 2

Differentially expressed genes identified by transcriptome analysis of low-level rifampicin.

Different gene expression categories and pathway analysis

Biological process, molecular function, and cellular component were the three categories of gene ontology (GO) analysis. The gene expression analysis showed that the gene functions of the significantly upregulated gene (21 genes) were enriched in molecular function and cellular component, especially in metabolic process (category P), catalytic activity (category F) and membrane and membrane part (category C) (Fig 3A). The gene expression analysis showed that the gene functions of the significantly downregulated gene (20 genes) were also enriched in molecular function and cellular component, especially in multi-organism process (category P) and membrane and membrane part (category C) (Fig 3B).

Fig 3

GO annotation of differentially expressed genes assigned to three GO categories.

(A) shows the categories of upregulated genes, (B) shows the categories of downregulated genes. Blue color indicated the biological process function, including metabolic process, localization, biological regulation, response to stimulus, signaling, multi-organism process, cellular process; red color indicated the cellular component function, including macromolecular complex, cell, membrane part, membrane; green color indicated molecular function, including binding, catalytic activity, transporter activity, molecular transducer activity.

KEGG pathway analysis demonstrated that the significantly upregulated genes were enriched in benzoate degradation, pyrimidine metabolism, quorum sensing, microbial metabolism in diverse environments and biosynthesis of antibiotics (Fig 4A), whereas the significantly downregulated genes were primarily involved in the bacterial secretion system (Fig 4B).

Fig 4

KEGG distribution pattern of differentially expressed genes.

(A) shows the distribution of upregulated genes, and (B) shows the downregulated genes. The ordinate denotes the pathway name, the abscissa denotes the rich factor, the size of the dot indicates the number of differentially expressed genes in this pathway, and the color of the dot corresponds to different Q-value ranges.

Cell membrane components of B. melitensis play essential roles in resistance to low rifampicin concentrations before the observed variation in the rpoB gene

The most exciting results were that 12 genes associated with the type IV secretion system (T4SS) were enriched and downregulated by rifampicin without variations in rpoB, including the VirB1-12 gene. Among the functions of cell membrane components, the quorum-sensing protein potA, β-resistance DacC, and ABC transporter AapJMQ genes were upregulated. In the cytoplasm, the T4SS regulation factor vjbR, HutC, BvrR and RpoH-like heat shock sigma factor were downregulated; while the DNA repair protein DpoI and DNA replication protein DnaG were upregulated. Based on the transcriptome results in this study, the following low rifampicin resistance mechanism of Brucella spp. was suggested: 1) Rifampicin exposure stimulates the outer membrane of cells to produce GABA, leading to an increase in the expression of quorum-sensing protein potA, which activates the quorum-sensing system. Furthermore, potA may regulate the physiological characteristics of Brucella spp. 2) the ABC transporter AapJMQ genes were upregulated, and their role may be to excrete rifampicin. 3) The expression of the T4SS regulation factor vjbR [26], HutC [27], BvrR [28] and RpoH-like heat shock sigma factor [29] were down-regulated, while the expression of the VirB operon was also downregulated, which led to a reduction in the secretory system. These changes may lead to less rifampicin intake. 4) Other known resistance-related genes such as DacC were upregulated for rifampicin resistance. 5) The expression of the DNA repair protein DpoI and DNA replication protein DnaG in the cytoplasm was upregulated, and these proteins may be used for rpoB variation or self-repair (Fig 5).

Fig 5

Hypothesis model of the rifampicin resistance mechanism in B.

melitensis.

The role of the VirB operon in low rifampicin resistance

The expression levels of genes in the VirB operon were evaluated under low rifampicin conditions (Fig 6A). The results showed that the expression of VirB1-6 and VirB12 was induced by low rifampicin, while the expression of VirB7-11 was inhibited in the samples treated with low rifampicin compared with the controls.

Fig 6

MIC values of rifampicin for the gene deletion and parent strains cultured for 10 d.

(A) shows the expression levels of the VirB genes by RT-PCR, (B) shows the MIC values of the gene deletion and parent strains.

VirB7-11 gene deletion strains named BMΔVirB7- BMΔVirB11 were constructed in this study. The growth characteristics of the gene deletion and parent strains were determined in Brucella broth medium in 96-well plates. No significant variations were observed. The gene deletion and parent strains were cultured under different rifampicin concentrations. Under rifampicin conditions, all MIC values of the strains were not different. After 3 d, the MIC values of the gene deletion strains were higher than those of the parent strain (Fig 6B). The There was no variation in rpoB during the culture by sequencing. We also constructed the virB1-6 and virB 12 genes deletion mutants and named BMΔVirB1- BMΔVirB6 and BMΔVirB12. The growth characteristics of the gene deletion and parent strains were determined and no significant variations were observed. Under rifampicin conditions, all MIC values of the strains were not different compared with the parent strain.

The complemented gene deletion mutants named BMΔVirB7/p-VirB7- BMΔVirB11/p-VirB11. There was no significant difference in growth characteristics between the complemented gene deletion mutants and parent strains. Meanwhile, although the complemented gene deletion mutants showed variations in MIC values of rifampicin compared with parent strains, it also demonstrated lower MIC values compared with the gene deletion strains at the same time points. It was also found that all MIC values of the strains were not different after 10 days. These results showed that VirB7-11 played important roles in rifampicin adaptive resistance in B. melitensis.

Brucella strain and genome in this study

The genome, amino acid, and nucleic acid sequences of B. melitensis bv.3 Ether were downloaded from the NCBI RefSeq database (ftp://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/740/355/GCF_000740355.1_ASM74035v1). A cultured and inactivated sample of B. melitensis bv.3 Ether were obtained in a BSL-3 laboratory of National Institute for Communicable Disease Control and Prevention of the Chinese Center for Disease Control and Prevention.

Rifampicin susceptibility testing

The susceptibility of a Brucella sp. strain to rifampicin susceptibility was tested by the double dilution method according to CLSI guidelines. Rifampin was double diluted from 0.125–256 μg / mL in 96-well plates. Strains were suspended in saline water to a 0.5 McF turbidity, and suspended in Brucella broth adjusted to pH 7.1 ± 0.1 (BD, USA). The plate was cultured at 35 ± 2°C with 5% CO₂ for 48 h. On the same plate, the negative and positive controls were set. The quality control strain was Streptococcus pneumoniae ATCC 49619 (refer to CISL_M45 (2016)).

In vitro growth characteristics in Brucella broth medium

For the growth analysis of the Brucella strain in presence of different rifampicin concentrations in vitro, one colony of the Brucella strain was inoculated into 3 mL of Brucella broth medium and cultured for 24–48 h at 37°C in an incubator shaking at 200 rpm. Subsequently, the cultures were adjusted to the same concentration (OD₆₀₀ ≈ 1.0) and used for growth curve analysis. A 50 μL sample of the Brucella culture was inoculated into 5 mL of Brucella broth medium with different concentrations of rifampicin. The cultures were incubated at 37°C with shaking at 200 rpm, and the OD₆₀₀ value was determined at different time points. These strains were tested in triplicate in two independent experiments.

RNA extraction and transcriptome

Before extraction, two volumes of RNAprotect Bacteria Reagent (Qiagen, DEU) were added directly to one volume of culture sample. RNA was extracted from the Brucella strains under different concentrations of rifampicin using TRIzol (Invitrogen, USA) according to the manufacturer’s instructions and treated with DNase (TaKaRa Bio, JPN) before reverse transcription to remove DNA contamination. Total RNA was dissolved in diethylpyrocarbonate (DEPC)-treated water, and the concentration and purity of the total RNA were estimated by NanoDrop and Aligilent 2100. The samples were sent to Huada Gene Co., Ltd. (China) for sequencing and transcriptome analysis.

To obtain more accurate and reliable results in subsequent bioinformatics analysis, the raw data were treated in the following manner: 1) reads with a certain proportion of low quality(≤20) bases(40% as default) were removed; 2) reads with a certain proportion of Ns (40% as default) were removed;3) adapter contaminations were removed; 4) duplication contaminations were removed to obtain clean data. Bowtie2[40] was used to align the reads with the reference genome. RSEM[41] software was used to calculate gene expression levels. Then, cluster and functional analyses were performed based on the different gene expression levels.

qRT-PCR

cDNAs were synthesized by extracting total RNA using the PrimeScript RT Reagent Kit (TaKaRa Bio, JPN.) according to the manufacturer’s instructions. The reverse transcription product was stored at -20°C. PCR was performed with the primers shown in S1 Table to evaluate gene expression. The value obtained by the relative quantitative method (2^-△△Ct) was used to compare the expression level of the reference gene at different rifampicin concentrations, and the 16S rRNA expression level in B. melitensis was used as a reference to normalize all values. Three replicate wells for each gene were evaluated, and the results presented in this paper represent the averages from at least three separate experiments.

Whole genome sequencing

Bacterial genomic DNA was extracted using the Wizard Genomic DNA Purification kit (Promega, USA) according to the manufacturer’s instructions. Library preparation was performed using the Nextera XT Library Prep kit (Illumina, USA) according to the manufacturer’s manual. The libraries were sequenced using an Illumina/Solexa sequencing analyzer to 100-fold (100×) genome coverage at the Beijing Genomics Institute (BGI) (Shenzhen, China).

SNPs identification and annotation

SNPs and indels identification and annotation were performed as described in a previous study [42] with minor modifications. We removed the SNPs with a quality score of less than 1000. All SNPs were identified by PCR amplification and sequencing.

PCR assay of the rpoB gene

The highly variable region and whole sequence of the rpoB gene were amplified using primers designed by Vector NTI (www.invitrogen.com/VectorNTI). Sequencing and data analysis were conducted as previously described[43] with some modifications. The liquid culture was streaked on a plate, which was incubated at 37°C for 5 days and then 10 colonies were examined at each time points.

Construction of gene deletion mutants

Gene deletion mutants were constructed using a recombination system with pBluescript_II phagemid vectors. The primers are shown in S2 Table. First, approximately 1050 bp of the chloromycetin resistance gene was amplified from plasmid pKD3, including the flippase recognition target (FRT). Second, the 500 bp flanking either side of the gene were amplified from B. melitensis bv.3 E ther genomic DNA with primers, gel-purified, and used as templates for a second round of overlapping PCR. Then, the fragments were cloned into pBluescript_II to generate the recombinant plasmid. Finally, the recombinant plasmid was introduced into B. melitensis bv.3 Ether via electroporation. Colonies that were sensitive to ampicillin (Amp) but not sensitive to chloromycetin (CM) were selected. The gene deletion mutants were then verified by PCR and sequencing analysis.

Construction of the complemented gene deletion mutants

The complete gene and its operon promoter sequences were amplified (S3 Table) in order to construct the complemented strain. The fragment was recombined into the pBBR1mcs-2 plasmid. After confirmed by sequencing, the recombinant plasmid was subsequently electroporated into the gene deletion mutant. The complemented gene deletion mutant strains were selected on the Brucella agar medium containing kanamycin. Finally, the selected complemented strains were confirmed through PCR amplification.

Statistical analysis

Statistical analyses were performed using Excel and SPSS. A P value < 0.05 was considered significant when using a one-way analysis of variance (ANOVA). The heatmap and other figures were drawn by R software.