Genomic and transcriptomic variation in Bordetella spp. following induction of erythromycin resistance

Abstract

Background: The emergence of macrolide resistance in Bordetella pertussis, the causative agent of pertussis, due to mutations in the 23S rRNA gene has been recently recognized. However, resistance mechanisms to macrolides in Bordetella parapertussis and Bordetella holmesii remain unknown.

Objectives: This study investigated genomic changes induced by in vitro exposure to erythromycin in these three main pathogens responsible for pertussis-like disease.

Methods: A set of 10 clinical and reference strains of B. pertussis, B. parapertussis and B. holmesii was exposed to erythromycin for 15 weeks or 30 subculture passages. Antibiotic pressure was achieved by growth on the selective media with erythromycin Etest strips or impregnated discs. Genome polymorphisms and transcriptomic profiles were examined by short- and long-read sequencing of passaged isolates.

Results: B. parapertussis and B. holmesii isolates developed significant in vitro resistance to erythromycin (MIC >256 mg/L) within 2 to 7 weeks and at 5 to 12 weeks, respectively. B. pertussis remained phenotypically susceptible to the antibiotic following 15 weeks of exposure, with the MIC between 0.032 to 0.38 mg/L. Genomic analysis revealed that B. holmesii developed resistance due to mutations in the 23S rRNA gene. The resistance mechanism in B. parapertussis was hypothesized as being due to upregulation of an efflux pump mechanism.

Conclusions: These findings indicate that both B. holmesii and B. parapertussis can be more prone to induced resistance following exposure to treatment with erythromycin than B. pertussis. The surveillance of macrolide resistance in Bordetella isolates recovered from patients with pertussis, especially persistent disease, is warranted.

Figure 1.

Average erythromycin MIC of Bordetella spp. across all 15 weeks. B. parapertussis became resistant in 2 weeks (average 4.5 weeks), while B. holmesii took on average 6 weeks. However, the erythromycin MIC of B. pertussis remained persistently low. Error bars represent the standard deviation of the MIC at the timepoint. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

Figure 2.

Comparison (BLASTn) of the mexAB-oprM operon ±3 flanking genes of Pseudomonas aeruginosa PA01 (NC_002516.2), Bordetella bronchiseptica 253 (NC_019382.1), B. parapertussis CIDM-BPP2 and CIDM-BPP2R, B. pertussis Tohama I (NC_002929.2), and B. holmesii CIDM-BH3. The primary pathogens from the Bordetella spp. carry a >71% orthologue by BLASTn of the mexAB-oprM operon, however, there is a gene deletion present in the mexA and oprM genes in B. pertussis. Nucleotide sequence similarity is scaled according to the scale bar. This image was generated using EasyFig. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

Figure 3.

Expression profile of the mexAB-oprM orthologue in B. parapertussis CIDM-BPP2 calculated based on genomic position and CPM. (a and b) Comparison of BPP2OG (erythromycin susceptible) with BPP2RES (resistant without antibiotic pressure), demonstrating a large proportion of reads encompassing the mexA gene. (a and c) Comparison of BPP2OG (susceptible) with BPP2RAB (resistant with antibiotic pressure), shows a similar outcome as BPP2RES. (d) Box plot of gene expression between conditions. For genes within mexAB-oprM and its transcriptional regulator (dmlR) all were significantly upregulated in the RES and RAB conditions. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

Figure 4.

Expression profile of the mexAB-oprM equivalent efflux pump in B. holmesii. (a and b) Comparison of BH3OG (susceptible) with BH3RES (resistant without antibiotic pressure), demonstrating relatively even distribution of reads across the gene locus. (a and c) Comparison of BPP2OG (susceptible) with BPP2RAB (resistant with antibiotic pressure), showed a similar outcome as BPP2RES. (d) Box plot of gene expression in isolates under different conditions. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.