A novel DNA chromatography method to discriminate Mycobacterium abscessus subspecies and macrolide susceptibility

Abstract

Background: The clinical impact of infection with Mycobacterium (M.) abscessus complex (MABC), a group of emerging non-tuberculosis mycobacteria (NTM), is increasing. M. abscessus subsp. abscessus/bolletii frequently shows natural resistance to macrolide antibiotics, whereas M. abscessus subsp. massiliense is generally susceptible. Therefore, rapid and accurate discrimination of macrolide-susceptible MABC subgroups is required for effective clinical decisions about macrolide treatments for MABC infection. We aimed to develop a simple and rapid diagnostic that can identify MABC isolates showing macrolide susceptibility.

Methods: Whole genome sequencing (WGS) was performed for 148 clinical or environmental MABC isolates from Japan to identify genetic markers that can discriminate three MABC subspecies and the macrolide-susceptible erm(41) T28C sequevar. Using the identified genetic markers, we established PCR based- or DNA chromatography-based assays. Validation testing was performed using MABC isolates from Taiwan.

Finding: We identified unique sequence regions that could be used to differentiate the three subspecies. Our WGS-based phylogenetic analysis indicated that M. abscessus carrying the macrolide-susceptible erm(41) T28C sequevar were tightly clustered, and identified 11 genes that were significantly associated with the lineage for use as genetic markers. To detect these genetic markers and the erm(41) locus, we developed a DNA chromatography method that identified three subspecies, the erm(41) T28C sequevar and intact erm(41) for MABC in a single assay within one hour. The agreement rate between the DNA chromatography-based and WGS-based identification was 99·7%.

Interpretation: We developed a novel, rapid and simple DNA chromatography method for identification of MABC macrolide susceptibility with high accuracy.

Funding: AMED, JSPS KAKENHI.

Keywords: ATS/ERS/ESCMID/IDSA guideline; DNA chromatography; Drug susceptibility; Non-tuberculosis mycobacteria; Subspecies discrimination; Whole genome sequencing.

Fig. 1

Maximum likelihood core-gene phylogeny of 148 clinical and environmental isolates of MABC. Core genome alignment of 148 isolates and three reference strains (M. abscessus ATCC19977, M. massiliense JCM 15300, and M. bolletii BD) of MABC was generated by Roary . An alignment containing 62,196 variable positions was used with RAxML to construct a maximum likelihood tree having 300 bootstrap replicates. Bootstrap values > 90% for the major nodes are shown. Scale bar indicates the mean number of nucleotide substitutions per site (Snps/Site) on the respective branch. Samples are highlighted based on inclusion in three major clusters corresponding to MABC subspecies. Arrowheads indicated isolates that were used to construct primers for MABC subspecies identification (listed in Table 1).

Fig. 2

Example of indels among MABC subspecies. (a) progressiveMauve alignment of the three reference strains of MABC is shown. Each genome is laid out in a horizontal track and white boxes indicate coding sequences annotated by dfast_core . A colored similarity plot is shown for each genome; the height is proportional to the sequence identity in that region. F and R indicate primer position of MAB2613F and MAB2613R (listed in Table 1), respectively, in each reference strain genome. (b) Representative multiplex PCR results for reference strains and clinical isolates amplified with primer pair MAB2613F and MAB2613R and primer pair MAB_1655F and MAB_1655R. Types 1, 1a, 1b, 2, 2a, 2b, 2c, 2d, 3, 3a, 3b, 3c, ds4 and ds5 are the sub-groupings (sequevars) of the clinical isolates based on their sequences [Table 3 of the previous article [31]]. Numerals below the sequevars correspond to reference strains (ab^T, ma^T, or bo^T) or the strain numbers of clinical isolates described in the previous article . Colored circles correspond to each member of MABC determined by WGS-based analysis. Lanes: M, DNA marker (100 bp ladder); ab^T, M. abscessus (ATCC19977); ma, M. massiliense (JCM15300); bo, M. bolletii (BD). The PCR reaction products were electrophoresed on 2% agarose gels.

Fig. 3

Macrolide susceptibility-associated genotypes of 148 MABC isolates. Maximum likelihood core-gene phylogeny of (a) M. abscessus, (b) M. massisliense, (c) M. bolletii correspond to those depicted in Fig. 1. The presence (black) and absence (gray) of macrolide resistance-associated mutations is indicated. The presence of a T-to-C substitution in position 28 or a truncation of the erm(41) gene, which are both associated with inducible resistance to macrolides, was detected. Substitutions or truncations with asterisks indicate non-synonymous mutations. The lineage to which all M. abscessus erm(41) T28C mutants belong is highlighted in magenta. The maximum likelihood trees, bootstrap values and scale bars correspond to those depicted Fig. 1.

Fig. 4

Visualization of lineage-specific genomic loci. (a) A progressiveMauve alignment of three clinical isolates carrying the erm(41) T28C mutation and the reference strains is shown. Boxes indicate coding sequences annotated by dfast_core . Red and blue boxes indicate genes that are significantly associated with the lineage to which all M. abscessus erm(41) T28C mutants belong (see Methods and Table S3). F1, R1, F2, R2, F3, and R3 indicate primer position of MAB18036_2551F, MAB18036_2551R, MAB18036_2558F, MAB18036_2558R, MAB18036_2560F, and MAB18036_2560R (listed in Table 2), respectively, in each genome of the presented clinical isolates. A similarity plot for each genome is colored as described for Fig. 2a. (b) Representative multiplex PCR results to identify inducible macrolide resistance in MABC. PCR was performed for the reference strains and clinical isolates were amplified with primer pairs MAB18036_2558F and MAB18036_2558R (listed in Table 2) and primer pair ermF (gaccggggccttcttcgtgatc) and ermR (agcttccccgcaccgattcca) . Colored circles correspond to each member of MABC determined by WGS-based analyses: red, M. abscessus; blue, M. massiliense; green, M. bolletii. The presence (+) or absence (-) of genotypes associated with the inducible macrolide resistance are shown. Lanes: M, DNA marker (100 bp ladder); ab^T, M. abscessus (ATCC19977); ma^T, M. massiliense (JCM15300); bo^T, M. bolletii (BD). The PCR reaction products were electrophoresed on 2% agarose gels. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 5

DNA chromatography to differentiate subspecies and macrolide susceptibility of MABC. DNA chromatography results for reference strains (ab^T, ma^T, or bo^T) or representative clinical isolates are shown. Colored circles above the strain names correspond to each member of MABC determined by WGS-based analyses: red, M. abscessus; blue, M. massiliense; green, M. bolletii, respectively. The presence (+) or absence (-) of genotypes associated with inducible macrolide resistance are shown. Bands: C, inner (negative) control; T1, erm(41) T28C polymorphism; T2, intact erm(41) genes; T3, M. bolletii; T4, M. abscessus; T5, M. massiliense. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)