Targeted Chromosomal Barcoding Establishes Direct Genotype-Phenotype Associations for Antibiotic Resistance in Mycobacterium abscessus

ABSTRACT A bedaquiline-resistant Mycobacterium abscessus isolate was sequenced, and a candidate mutation in the atpE gene was identified as responsible for the antibiotic resistance phenotype. To establish a direct genotype-phenotype relationship of this mutation which results in a Asp-to-Ala change at position 29 (D29A), we developed a recombineering-based method consisting of the specific replacement of the desired mutation in the bacterial chromosome. As surrogate bacteria, we used two M. abscessus bedaquiline-susceptible strains: ATCC 19977 and the SL541 clinical isolate. The allelic exchange substrates used in recombineering carried either the sole D29A mutation or a genetic barcode of silent mutations in codons flanking the D29A mutation. After selection of bedaquiline-resistant M. abscessus colonies transformed with both substrates, we obtained equivalent numbers of recombinants. These resistant colonies were analyzed by allele-specific PCR and Sanger sequencing, and we demonstrated that the presence of the genetic barcode was linked to the targeted incorporation of the desired mutation in its chromosomal location. All recombinants displayed the same MIC to bedaquiline as the original isolate, from which the D29A mutation was identified. Finally, to demonstrate the broad applicability of this method, we confirmed the association of bedaquiline resistance with the atpE A64P mutation in analysis performed in independent M. abscessus strains and by independent researchers. IMPORTANCE Antimicrobial resistance (AMR) threatens the effective prevention and treatment of an ever-increasing range of infections caused by microorganisms. On the other hand, infections caused by Mycobacterium abscessus affect people with chronic lung diseases, and their incidence has grown alarmingly in recent years. Further, these bacteria are known to easily develop AMR to the few therapeutic options available, making their treatment long-lasting and challenging. The recent introduction of new antibiotics against M. abscessus, such as bedaquiline, makes us anticipate a future when a plethora of antibiotic-resistant strains will be isolated and sequenced. However, in the era of whole-genome sequencing, one of the challenges is to unequivocally assign a biological function to each identified polymorphism. Thus, in this study, we developed a fast, robust, and reliable method to assign genotype-phenotype associations for putative antibiotic-resistant polymorphisms in M. abscessus.

Bedaquiline (BDQ) has recently been shown to be active against M. abscessus. In this study, Calvet-Seral et al. developed a genetic strategy to establish a direct genotype-phenotype association for resistance to BDQ in M. abscessus. This method is based on the recombineering strategy and using a genetic barcode of silent mutations in the codons flanking the D29A and A64P mutations (identified earlier as key mutations in resistance to BDQ in M. abscessus). The authors clearly showed that the genetic barcode was associated to the target incorporation of the D29A and A64P mutations, in both the reference strain ATCC19977 as well as the clinical strain SL541. From a biological perspective, this work only confirms already characterized mutations involved in BDQ resistance in M. abscessus and does not provide new insights into the mechanisms of drug resistance. The major take-home message relies on the successful demonstration of the use of barcoded oligo-mediated targeted mutagenesis in the genome of M. abscessus, although the use of the recombineering strategy is not innovative as it has been widely used in different mycobacterial species, including M. abscessus. In general, the manuscript is very well written and the conclusions raised are supported by the data. However, there are a few points/controls that need to be clarified to improve the quality of the manuscript.
Major points: -line 151: the authors mention that the pJV53 was maintained in the presence of 50µg/ml kanamycin. Usually, the MIC of kanamycin is much higher that 50µg/ml in M. abscessus and very high concentrations are needed for the selection of a plasmid. Selection on low concentrations is often associated to false-positive clones growing on kanamycin but without containing the plasmid of interest. Therefore, I am not convinced that the abscessus strains used in this study contain the pJV53. This needs to checked either by PCR and or Western blotting using anti-gp61 antibodies. Since M. abscessus is able to perform homologous recombination, it is still possible the double recombination event required for the allelic exchange occurred even in the absence of pJV53. This needs to be clarified experimentally.
-the authors indicate at several places in the manuscript (as well as in Figure 1), that results can be obtain in a reasonable short time, mostly limited to the growth of drug resistant colonies. While this is probably true, they did not mention that prior to electroporation of the single stranded allelic exchange substrates, the strain containing pJV53 needs to be generated first, which can be challenging in the case of a clinical strain as mentioned above. This may add an extra 2 weeks to the overall timescale, and should be highlighted in the text as well as in Figure 1.
-provide the MICs of the ATTCC strain in Table 2, so that the reader can compare the drug susceptibility profiles of the two strains used in this study. Indicate also the MIC of BDQ and kanamycin for both strains.
Minor points: -The section in the Discussion describing drug resistance mechanisms in M. tuberculosis is out of the scope of this study. Therefore, I propose the authors to deleted this section (lines 370 to 377). -line 150 "Kanamycin" instead of "Kanamicyn" -line 229: "establish" instead of "stablish" -line 297: "flanked" instead of "flaked" Reviewer #2 (Comments for the Author): Drug resistance is a growing problem in the control of mycobacterial infections. Various strategies have been used to establish a functional link between a SNP and the resistance to a drug. Most available methods have limitations that are mostly associated with unwanted side effects of genetic complementation of a SNP either by chromosomal exchange or episomal introduction of a wt copy of the resistance gene. To overcome these disadvantages, the authors have employed an elegant recombineering system developed by the Hartfull lab, that enables chromosomal point mutagenesis. Using a BDQ resistant clinical isolate of M. abscessus, they identified a SNP in the atpE gene, which is used to showcase their methodology. In addition, the authors use "chromosomal barcoding" (=silent SNPs near the resistance conferring mutation) to distinguish intended mutagenesis from random (natural) mutagenesis events. Performance of the new genotype-phenotype association method introduced here was verified by Sanger sequencing, allele specific PCR, qPCR and drug susceptibility testing of mutant M. abscessus colonies. Experimental data shown is of good scientific rigor and supports the conclusions made. Experiments were carefully designed and executed. The mostly well-written manuscript has very minor deficiencies in language -some are included below -others will be easily addressed by the copyeditors. The method of genotype-phenotype association is of interest to the field of drug development/resistance and certainly to readers of Spectrum as well. The discussion could be condensed, e.g. it is not overly important for the presented data to discuss M. tuberculosis drug resistance in detail (1st paragraph). The focus of the discussion should be the limitations (see below) and applications of the new methodology introduced by the authors. The points below should be considered for improvement of this overall exciting work.
Major points: 1. Fig. 3C and S2: negative PCR reactions are in general less appreciated. Amplification of a housekeeping gene should be included to verify that the absence of a band is not the result of a non-working PCR reaction. Fig. 3C: Why did two of the clones shown in the right panel fail to produce the expected amplicon? Did the PCR reaction not work, or have they gained BDQ resistance through mechanisms not involving atpE D29A? This needs to be explained.
2. The Discussion could be condensed, and further possible limitations of the introduced method should be included: (1) Impact of high in vitro M. abscessus mutation frequency of a study drug on the application of the new methodology. (2) How useful is the author's methodology to study drugs having multiple targets or multiple modes of resistance?
Staff Comments:

Preparing Revision Guidelines
To submit your modified manuscript, log onto the eJP submission site at https://spectrum.msubmit.net/cgi-bin/main.plex. Go to Author Tasks and click the appropriate manuscript title to begin the revision process. The information that you entered when you first submitted the paper will be displayed. Please update the information as necessary. Here are a few examples of required updates that authors must address: • Point-by-point responses to the issues raised by the reviewers in a file named "Response to Reviewers," NOT IN YOUR COVER LETTER.
• Upload a compare copy of the manuscript (without figures) as a "Marked-Up Manuscript" file. • Each figure must be uploaded as a separate file, and any multipanel figures must be assembled into one file. For complete guidelines on revision requirements, please see the journal Submission and Review Process requirements at https://journals.asm.org/journal/Spectrum/submission-review-process. Submissions of a paper that does not conform to Microbiology Spectrum guidelines will delay acceptance of your manuscript. " Please return the manuscript within 60 days; if you cannot complete the modification within this time period, please contact me. If you do not wish to modify the manuscript and prefer to submit it to another journal, please notify me of your decision immediately so that the manuscript may be formally withdrawn from consideration by Microbiology Spectrum.
If your manuscript is accepted for publication, you will be contacted separately about payment when the proofs are issued; please follow the instructions in that e-mail. Arrangements for payment must be made before your article is published. For a complete list of Publication Fees, including supplemental material costs, please visit our website.
Corresponding authors may join or renew ASM membership to obtain discounts on publication fees. Need to upgrade your membership level? Please contact Customer Service at Service@asmusa.org.
Thank you for submitting your paper to Microbiology Spectrum. Overall, we consider that the revised manuscript has significantly improved its quality after the reviewer's recommendations and we hope that it is suitable for publication in Microbiology Spectrum.
Yours faithfully, Jesús Gonzalo-Asensio, PhD; on behalf of the co-authors of this work Reviewer #1 (Comments for the Author): We acknowledge the overall positive impression of the reviewer. Please find below a point-by-point answers to your comments which has also resulted in the inclusion of new figures/tables in this revised manuscript.
Major points: -line 151: the authors mention that the pJV53 was maintained in the presence of 50µg/ml kanamycin. Usually, the MIC of kanamycin is much higher that 50µg/ml in M. abscessus and very high concentrations are needed for the selection of a plasmid. Selection on low concentrations is often associated to false-positive clones growing on kanamycin but without containing the plasmid of interest. Therefore, I am not convinced that the abscessus strains used in this study contain the pJV53. This needs to checked either by PCR and or Western blotting using anti-gp61 antibodies.
We thank the reviewer for noticing this discrepancy, which in our opinion comes from the various kanamycin concentrations used in the literature to select M. abscessus recombinants. While some studies use high kanamycin concentrations (i. e. 250 µg/mL), other studies use lower concentrations (in the range of 25-100 µg/mL) of this antibiotic to select either mutant, or plasmid-carrying, bacteria. Since we routinely work in mycobacterial genetics, we are aware of the diverse levels of antibiotic susceptibility displayed by laboratory and clinical strains. Accordingly, prior to introduce pJV53 into M. abscessus, we checked the antibiotic susceptibility against kanamycin in both, the ATCC19977, and the SL541, strains of M. abscessus.

µg/mL
The picture above shows MICs against kanamycin and gentamycin of the aforementioned strains, showing that MIC against kanamycin lies between 12.5 and 6.25 µg/mL in the SL541 strain, and between 3.12 and 1.56 µg/mL in the ATCC19977 strain. Thus, both M. abscessus strains used in this study are susceptible to kanamycin at 50 µg/mL. Therefore, selection of M. abscessus pJV53 transformants on plates containing 50 µg/mL results in an efficient recovery of colonies as shown in the picture above. Note that non-transformed bacteria are unable to grow on 50 µg/mL kanamycin plates, which minimizes the chance of recovering spontaneous mutants against this antibiotic.
Nevertheless, to molecularly confirm the presence of pJV53 into M. abscessus colonies grown on 50 µg/mL kanamycin plates, we checked the presence of the plasmid by PCR amplification of the gp60-gp61 recombineering genes contained in the plasmid. Results confirmed the presence of pJV53 in most colonies analyzed. We carefully selected pJV53-bearing colonies for our downstream experiments.
Finally, prior to perform recombineering experiments, we checked again the presence of pJV53 into M. abscessus by assessing the kanamycin resistance profile of our recombineering strains. As shown above, an SL541 strain carrying pJV53 exhibits a kanamycin resistance profile, in contrast to the wild type SL541, or the bedaquiline resistant mutant of this strain. Since M. abscessus is able to perform homologous recombination, it is still possible the double recombination event required for the allelic exchange occurred even in the absence of recombination, it is still possible the double recombination event required for the allelic exchange occurred even in the absence of pJV53. This needs to be clarified experimentally.
This is a very relevant observation which we overlooked in the initial submission. We repeated the experimental flow described in the original manuscript with M. abscessus SL541 and ATCC19977 strains, with or without pJV53. All strains were grown in parallel, and 600 ng of each AES (atpE wt as negative control, atpE D29A with or without barcode, and atpE A64P with barcode), were electroporated into 200 µL of electrocompetent cells. 100 µL of transformed bacteria were plated at the indicated dilution in presence or absence of Bedaquiline. Results showed absence of bacterial growth in wild type bacteria not transformed with pJV53. In contrast, pJV53-transformed bacteria resulted in increased CFU numbers, which were significantly higher in transformants with AES conferring bedaquiline resistance (atpE D29A with or without barcode, and atpE A64P with barcode), when compared with the atpE wt AES. Differences in CFU numbers are not related to differences in bacteria plated as observed in the CFU grown in the absence of bedaquiline.
Together, we consider that these results unequivocally demonstrate that a recombineering system is required for the allelic exchanges reported in the manuscript, and we sincerely acknowledge this reviewer's clarification. A new Supplementary Figure 4 including these results in the SL541 and ATCC19977 strains is provided, and additional sentences clarifying this suggestion have been included in the manuscript (lines 350-359, 754-763) -the authors indicate at several places in the manuscript (as well as in Figure 1), that results can be obtain in a reasonable short time, mostly limited to the growth of drug resistant colonies. While this is probably true, they did not mention that prior to electroporation of the single stranded allelic exchange substrates, the strain containing pJV53 needs to be generated first, which can be challenging in the case of a clinical strain as mentioned above. This may add an extra 2 weeks to the overall timescale, and should be highlighted in the text as well as in Figure 1.
The reviewer is right. A new panel in Figure 1 has been added to indicate the experimental flow and timelines for the construction of a recombineering strain, if needed. The challenge to transform non-laboratory strains is also indicated elsewhere (lines 337-339, 675-677) -provide the MICs of the ATTCC strain in Table 2, so that the reader can compare the drug susceptibility profiles of the two strains used in this study. Indicate also the MIC of BDQ and kanamycin for both strains.
This is an important consideration for potential readers. Table 2 has been updated with data from the ATCC19977 strains, and including bedaquiline and kanamycin MICs for both strains (lines 664-666) Minor points: -The section in the Discussion describing drug resistance mechanisms in M. tuberculosis is out of the scope of this study. Therefore, I propose the authors to deleted this section (lines 370 to 377).
We agree with the reviewer, and this paragraph has been deleted to provide a more focused message. We thank the reviewer for the encouraging comments and we hope to have addressed her/his suggestions in the revised version of the text, and figures.
Major points: 1. Fig. 3C and S2: negative PCR reactions are in general less appreciated. Amplification of a housekeeping gene should be included to verify that the absence of a band is not the result of a non-working PCR reaction.
The reviewer is right and we greatly acknowledge his/her suggestion since it contributes to improve the overall reliability of the results. We repeated the PCR using DNA from the original SL541 colonies transformed with either the barcode or the non-barcode D29A oligonucleotide. In parallel, we amplified by PCR the coding region of the 16S rRNA housekeeping gene, using the same oligonucleotides used in the qRT-PCR reactions shown in Figure 3D. Results show amplification of the 16S gene in all colonies analyzed, irrespective of their transformation with barcoded or nonbarcoded oligonucleotide. Concerning amplification of the barcode region, this time we obtained amplification in 9/12 colonies, compared to the positive amplification of 10/12 colonies in the original submission. Colonies 1 and 5 failed to produce amplification in both cases. Colony 12 showed a slight PCR amplification in our original submission, but failed to produce a PCR band in this revised version. However, observing the PCR amplification of the 16S control gene in this colony, it is evident that the low amounts of template DNA might influence the outcome of the PCR reaction in this case. This same approach was performed with ATCC19977 D29A bedaquiline resistant colonies in the new Figure S3 (formerly S2). Together, results were highly reproducible, and the inclusion of an unrelated, positive control of amplification is useful to discriminate putative false positives due to spontaneous mutations. Panel C of Figure 3, and Figure S3 (formerly S2), have been updated and a description of the positive control is provided in lines 319-321, 699-700, and 750-751. Fig. 3C: Why did two of the clones shown in the right panel fail to produce the expected amplicon? Did the PCR reaction not work, or have they gained BDQ resistance through mechanisms not involving atpE D29A? This needs to be explained.
As explained in the previous point, the reviewer's suggestion to include an unrelated, positive control of PCR amplification is useful to discriminate between the desired mutations and false positives due to spontaneous mutations also conferring bedaquiline resistance. Thus, colonies 1 and 5 probably arose from a spontaneous mutation either in the atpE gene or elsewhere in the chromosome. Colony 12, as shown above, is a true D29A mutation selected by recombineering, albeit the low quatities of template DNA do not allow PCR amplification in this case. For the shake of clarity, and to provide a clear and straighforward message to the reader, these 3 colonies have been considered as false positives, and these points are summarized in lines 321-325.
2. The Discussion could be condensed, and further possible limitations of the introduced method should be included: (1) Impact of high in vitro M. abscessus mutation frequency of a study drug on the application of the new methodology.
(2) How useful is the author's methodology to study drugs having multiple targets or multiple modes of resistance?
In line with the suggestions of both reviewers, some paragraphs, including those describing antimicrobial resistance in M. tuberculosis, have been eliminated. These additional limitations of the technique are discussed in lines 448-458.
Corrected, line 155 4. L162: "plates were incubated": do the authors refer here to agar plates or multi-well plates with liquid cultures? This should be specified.
The reviewer is right. This methodology is confusing and not correctly explained in the initial submission, and we acknowledge this constructive criticism. We used multi-well cultures to quantify the MIC. These multi-well plates were incubated in either solid or liquid media to enumerate MICs by the resazurin, and the MTT methods, respectively. The complete method section has been rewritten (lines 157-168)