Mab2780c, a TetV-like efflux pump, confers high-level Spectinomycin resistance in Mycobacterium abscessus

Abstract

Mycobacterium abscessus is highly resistant to spectinomycin (SPC) thereby making it unavailable for therapeutic use. Sublethal exposure to SPC strongly induces whiB7 and its regulon, and a ΔMab_whiB7 strain is SPC sensitive suggesting that the determinants of SPC resistance are included within its regulon. In the present study we have determined the transcriptomic changes that occur in M. abscessus upon SPC exposure and have evaluated the involvement of 11 genes, that are both strongly SPC induced and whiB7 dependent, in SPC resistance. Of these we show that MAB_2780c can complement SPC sensitivity of ΔMab_whiB7 and that a ΔMab_2780c strain is ~150 fold more SPC sensitive than wildtype bacteria, but not to tetracycline (TET) or other aminoglycosides. This is in contrast to its homologues, TetV from M. smegmatis and Tap from M. tuberculosis, that confer low-level resistance to TET, SPC and other aminoglycosides. We also show that the addition of the efflux pump inhibitor (EPI), verapamil results in >100-fold decrease in MIC of SPC in bacteria expressing Mab2780c to the levels observed for ΔMab_2780c; moreover a deletion of MAB_2780c results in a decreased efflux of the drug into the cell supernatant. Together our data suggest that Mab2780c is an SPC antiporter. Finally, molecular docking of SPC and TET on models of TetV_Ms and Mab2780c confirmed our antibacterial susceptibility findings that the Mab2780c pump preferentially effluxes SPC over TET. To our knowledge, this is the first report of an efflux pump that confers high-level drug resistance in M. abscessus. The identification of Mab2780c in SPC resistance opens up prospects for repurposing this relatively well-tolerated antibiotic as a combination therapy with verapamil or its analogs against M. abscessus infections.

Introduction

Mycobacterium abscessus, a rapidly-growing non-tuberculous mycobacterium (NTM), causing severe respiratory and extrapulmonary infections, is intrinsically resistant to the anti-tuberculosis drugs, rifampicin, isoniazid, ethambutol and pyrazinamide as well as to most approved antibiotics (1–3). Pulmonary infections of M. abscessus are typically treated with a combination of an oral macrolide, injectable amikacin, and one or more of the injectables- cefoxitin, imipenem or tigecycline, for a period of 9–12 months (4–6). Despite an aggressive therapy lasting 6–9 months, clearance rates of bacteria are poor ( ~45%) (7, 8). Over the past decade a number of effectors that contribute to the intrinsic resistance of M. abscessus have been elucidated including Erm41, HflX, ABCFs in macrolide/ lincosamide resistance, MabArr and HelR in rifamycin resistance, MabBla in β-lactam resistance, Eis2 and MabAPH in aminoglycoside resistance and MabTetX in tetracycline (TET) resistance. The vast majority of these above-described mechanisms involve drug modification or target modification/protection and are induced via WhiB7 dependent and independent pathways (9–17). MAB_whiB7 encodes a transcriptional activator and is one of the earliest genes expressed in response to most ribosome-targeting antibiotics such as erythromycin (ERT), clarithromycin (CLA), amikacin (AMK), streptomycin (STR), tetracycline (TET) and spectinomycin (SPC) (15) and a ΔMAB_whiB7 is highly sensitive to most of the antibiotics that induce its expression.

Intrinsic antibiotic resistance by drug efflux pumps belonging to all four superfamilies, MFS, ABC SMR and RND have been characterized in mycobacteria, primarily M. tuberculosis and M. smegmatis, and typically confer low levels of drug resistance either to a single antibiotic or a variety of structurally unrelated antibiotics (18). The M. abscessus genome encodes a large number of genes with homology to transporters and efflux pumps several of which are strongly upregulated in M. abscessus upon drug exposure (19). Furthermore, previous studies show that treatment of M. abscessus with efflux pump inhibitors such as verapamil (VMP) and CCCP increases its susceptibility to antimicrobials suggesting that drug efflux contributes to intrinsic antibiotic tolerance in M. abscessus (20). Nonetheless a definite role of these efflux pumps has not been described and Mab2355c and Mab1846, which were previously thought to be involved in drug efflux due to the presence of an ABC domain, were subsequently shown to function in target protection (19, 21). To date the MmpS-MmpL genes encoded by MAB_2300–2301 and MAB_1135c-1134c are the only well characterized efflux pumps described in clofazimine and bedaquiline resistance in M. abscessus (22).

Aminoglycosides comprise a group of broad-spectrum antibiotics that bind at different locations on the 16S rRNA and function by either lowering the fidelity of incorporation of A-site tRNA, tRNA selection or ribosome translocation. The most commonly encountered mechanism of aminoglycoside resistance involves drug modification by three groups of enzymes encoded by resistant bacteria-aminoglycoside acetyltransferases (AACs), aminoglycoside phosphotransferases (APHs) and aminoglycoside nucleotidlytransferases (ANTs) (23). The aminoglycoside amikacin constitutes a drug of choice in the treatment of M. abscessus infections despite its potential for renal toxicity and ototoxicity. Comparatively, SPC, an aminoinositol, closely related in structure to the aminoglycosides and historically used to treat gonorrheal infections, has only minor side effects when administered at high doses (24). M. abscessus however displays a high intrinsic resistance against SPC (MIC >1000µg/mL), making it unavailable for therapeutic use. In the present study we investigate the molecular basis of this SPC resistance.

RESULTS & DISCUSSION

MAB_2780c can complement the SPC sensitivity of ΔwhiB7

Exposure of M. abscessus to subinhibitory levels of SPC leads to an upregulation of a large number of genes, including those within the WhiB7 regulon (Figure1a). A ΔMAB_whiB7 strain is highly sensitive to SPC, suggesting that the effector of resistance is included within the WhiB7 regulon (15). A subset of 11 WhiB7 dependent genes encoding putative acetyltransferases and transporters/efflux pumps were screened for their ability to restore the SPC sensitivity of ΔMAB_whiB7 to that of wildtype bacteria (Figure 1b). For this, complementing strains were constructed by transforming the ΔMAB_whiB7 strain with the integrating vector pMH94 expressing each of these genes from the constitutive promoter, P_hsp60. Only MAB_2780c was capable of complementing the SPC sensitivity of ΔMab_whiB7 (Figure 1c).

Figure 1. Identification of MAB_2780c in SPC resistance:

a) Volcano plot of differentially expressed genes in M. abscessus ATCC19977 upon exposure to SPC (256 μg/mL for 2 hs). Genes with a log₂ fold change >4 are shown in red. Genes with a log₂ fold change >4 and that are either annotated as efflux pumps or acetyltransferases are shown in blue. Location of whiB7, the most induced gene is indicated. b) List of acetyltransferases and efflux genes that are highly induced upon SPC exposure and evaluated for their ability to restore the SPC sensitivity of ΔMAB_whiB7. c) Growth of ten-fold serial dilutions of M. abscessus ATCC 19977, ΔMab_whiB7 and ΔMab_whiB7 :: phspMAB_2780c on Middlebrook 7H10 OADC containing indicated concentrations of SPC. Data is representative of >3 independent experiments.

A ΔMAB_2780c strain is hypersensitive to SPC

Sequence analysis using BlastP revealed that Mab2780c shows ~40 % and 20% sequence identity to the M. smegmatis TetV (TetV_Ms) and M. tuberculosis/M. fortuitum Tap (Rv1258c) efflux pumps respectively, that confer low levels of TET and aminoglycoside resistance (Figure 2a). A model of Mab2780c predicts the existence of 12 transmembrane domains characteristic of MFS transporters that are also observed in TetV (Figure 2b). The Tap proteins from M. tuberculosis and M. fortuitum, despite sharing 66% amino acid sequence identity, display differences in their resistance spectra; Tap_Mfor is known to mediate resistance to TET, gentamycin and STR, whereas Tap_Mtb confers resistance to TET, but not to STR and gentamycin (25). TetV_Ms has been shown to be involved in resistance to TET, but not to STR(26). While Tap_Mtb also confers SPC resistance, the involvement of TetV_Ms and Tap_Mfor in SPC resistance is unknown (27, 28).

Figure 2. Similarity of MAB_2780c to related TET/SPC efflux pumps.

a) Sequence alignment of MAB_2780c, TetV_Ms, Tap_Mtb and Tap_Mfor generated using Clustal Omega. Location of conserved motifs, motif A- found in all MFS proteins, motif C- found in MFS proteins with 12 or 14 transmembrane helices and motif G- found in MFS proteins with 12 transmembrane helices are indicated (48). b) An overlay of the TetV_Ms (orange) and Mab_2780c (purple) homology models generated using AlphaFold-Colab and rendered as cartoon ribbons.

To determine the role of MAB_2780c in SPC resistance and its spectrum of activity, we created an unmarked, isogenic deletion of MAB_2780c in M. abscessus ATCC 19977 using phage recombineering (15, 29). A complementing strain was also constructed by transforming the deletion mutant with phspMAB_2780c, in which MAB_2780c is expressed from the constitutive promoter, P_hsp60 from a chromosomal location. Wildtype M. abscessus, ΔMAB_2780c and the complemented strain were tested for their susceptibility to 9 ribosome targeting antibiotics (Figure 3a-c). Figure 3a and Table 1 show that the ΔMAB_2780c is ~ 120 times more sensitive to SPC than wildtype M. abscessus, and its susceptibility is restored to wildtype levels in the complementing strain. The susceptibility of ΔMAB_2780c to the other 8 antibiotics tested, including TET, remained unchanged (Figure 3b-c).

Figure 3. Deletion of MAB_2780c results in SPC hypersensitivity :

(a-b)Ten-fold serial dilutions of M. abscessus ATCC19977, ΔMAB_ 2780c and the complementing strains were grown to A₆₀₀ of 0.7 and spotted on Middlebrook 7H10 OADC containing indicated concentrations of antibiotics targeting the 30S ribosomal subunit - spectinomycin (SPC), amikacin (AMK), tetracycline (TET), tigecycline (TIG) and streptomycin (STR). c) Ten-fold serial dilutions of M. abscessus ATCC1997 and ΔMAB_ 2780c were grown to A₆₀₀ of 0.7 and spotted on Middlebrook 7H10 OADC containing indicated concentrations of antibiotics that target the 50S ribosome subunit - clarithromycin (CLA), clindamycin (CLIN), chloramphenicol (CM) and erythromycin (ERT). d) Ten-fold serial dilutions of M. abscessus ATCC19977, ΔMabTetX as well as the indicated complementing strains were grown to A₆₀₀ of 0.7 and spotted on Middlebrook 7H10 OADC containing 5μg/mL of tetracycline (TET). The ΔMAB_ 2780c mutant is hypersensitive to only spectinomycin.

Table 1: MICs of Wild-type and mutant M. abscessus strains using a macrobroth dilution assay.

Survival of wild type M. abscessus ATCC19977, ΔMAB_2780c, ΔMAB_2780c::phspMab_2780c and ΔMAB_2780c::phspMstetV in a 2-fold macrobroth dilution series of SPC and TET in Middlebrook 7H9/OADC medium. The minimum concentration of antibiotic required to inhibit 99 % of growth after 72 hours is shown. Data is representative of 3 biological replicates.

Strain	Spectinomycin(μg/mL)	Tetracycline(μg/mL)
WT	1024	8
Δ2780c	8	8
Δ2780c::phspMab2780c	1024	8
Δ2780c::phspMs5187	32	8

Since the Tap and TetV homologues of MAB_2780c are associated with TET resistance, it is possible that MAB_2780c also confers low-level TET resistance which is masked in the presence of MAB_tetX that confers high-level TET resistance in M. abscessus (30). To explore this possibility, we constitutively expressed MAB_2780c from a chromosomal location in a ΔMAB_tetX strain that is highly sensitive to TET. As seen in Figure 3d, the expression of MAB_2780c does not result in an increase in the tolerance of ΔMAB_tetX to TET suggesting that MAB_2780c is not involved in TET resistance.

Although TetV_Ms has been shown to not be involved in resistance to several aminoglycosides in M. smegmatis, its involvement in SPC resistance has not been investigated (26). Constitutive expression of TetV_Ms in ΔMAB_2780c strain results in an increase in the MIC of SPC, but to a much lower extent as compared to that by MAB_2780c (Figure 3a and Table 1), indicating that TetV_Ms is capable of mediating low levels of SPC resistance.

Mab2780c is involved in SPC efflux

The SPC sensitivity of ΔMAB_2780c as well as the homology of Mab2780c with previously established efflux pumps, TetV_Ms and Tap_Mtb, suggests that Mab2780c similarly functions as a protonmotive- force-dependent efflux pump which can be disrupted by the use of a membrane energy uncoupler such as verapamil (VMP). We therefore compared the fractional inhibitory concentration index (FICI) of M. abscessus strains in checkerboard assays with VMP (Figure 4a-c). The WT strain displayed synergy (FICI 0.28125–0.3125) in combination with VMP (Figure 4a). However, this synergistic interaction was abolished in the ΔMAB_2780c strain (FICI 1–1.5; Figure 4b) and subsequently recapitulated in the MAB_2780c :: phsp MAB_2780c strain (FICI 0.28125; Figure 4c). Specifically at 128 μg/mL of VMP, SPC susceptibility in the WT and MAB_2780c :: phsp MAB_2780c strains increased (SPC MIC 32 μg/mL), whereas SPC susceptibility in ΔMAB_2780c was unaltered by VMP at the same concentration (Figure 4a-c; Table 2). These results indicate that the function of MAB_2780c is dependent upon the proton gradient.

Figure 4. The efflux inhibitor VMP strongly inhibits Mab2780c.

Heatmap of checkerboard assay of SPC and VMP in (a) M. abscessus ATCC 19977, (b) M. abscessus ∆2780c, and (c) M. abscessus ∆2780c::phsp2780c. Percent viability was determined by resazurin conversion over the final 24 hours of a 72-hour incubation at 37oC with darker blue indicating higher reduction potential (surrogate for cell viability). Median values from triplicate data were plotted. Fractional inhibitory concentration index (FICI) scores were calculated for each replicate and interpreted as synergistic (≤0.5), indifferent (>0.5–2) or antagonistic (>2).

Table 2: Effect of verapamil (VMP) on the MICs of M. abscessus strains using a macrobroth dilution assay.

Survival of wild type M. abscessus ATCC19977, ΔMAB_2780c and ΔMAB_2780c::phspMab_2780c in a 2-fold macrobroth dilution series of SPC in Middlebrook 7H9/OADC medium containing either 128 or 256 μg/mL VMP. The minimum concentration of antibiotic required to inhibit 99 % of growth after 72 hours is shown. Data is representative of 3 biological replicates.

Strain	SPC (μg/mL)	SPC (μg/mL) + VMP (256 μg/mL)	SPC (μg/mL) + VMP (128 μg/mL)
WT	1024	8	32
Δ2780c	8	8	8
Δ2780c::phspMab2780c	1024	8	32

Next, we measured the MAB_2780c dependent efflux of SPC in a wildtype and ΔMAB_2780c strain. For this, the cells were exposed to SPC, washed to remove unincorporated drug and the amount of SPC effluxed into the supernatant was determined after 4 hours (Figure 5a). Antibiotic uptake was found to be indistinguishable between the two strains (not shown). However, significantly lower quantities of SPC were recovered in the supernatant of the MAB_∆2780c strain compared to wildtype (p-value 0.014) after 4 hours (Figure 5b) supporting the role of MAB_2780c as an SPC efflux transporter.

Figure 5. Efflux of SPC by M. abscessus wildtype and ΔMAB_ 2780c strains:

a) Schematic representation of sample preparation for determination of drug efflux by LC-MS/ MS created using BioRender. b) The amount of SPC recovered in the supernatant was measured at time 0 and after 4 hours of incubation in wildtype and ΔMAB_ 2780c bacteria.

Molecular docking studies reveal the proposed SPC and TET binding poses on TetV_Ms and Mab2780c

Despite a 62% sequence similarity (~40% identity) between TetV_Ms and Mab2780c, our data demonstrate differences in substrate specificity between the pumps. TetV_Ms confers resistance to both SPC and TET; in contrast, Mab2780c preferentially pumps out SPC with no specificity for TET. Thus, we performed molecular docking to determine the proposed binding modes and rationalize the observed selectivity. Using AlphaFold-Colab to generate their respective protein structures, both TetV_Ms and Mab2780c formed 12 transmembrane structures consistent with other Major Facilitator Superfamily (MFS) efflux pumps, with both C- and N-terminal ends located in the cytosolic compartment (Figure 6a-b). Both SPC and TET dock to the TetV_Ms pump in neighboring binding regions with similar docking scores (Table 3 and Figure 6a). The highest-scored SPC proposed binding pose, −6.13 kcal mol⁻¹, is driven by hydrogen bonding interactions and salt bridges between D361 and SPC. (Figure 6a). Like SPC, the TET binding pose, −6.19 kcal mol⁻¹, is driven by hydrogen bonding and salt bridging interactions between R157, Q161, R162, E251, and D334 (Figure 6a). The similar docking scores and the proposed binding site interactions suggest that TetV_Ms recognizes both SPC and TET as substrates for efflux.

Figure 6. The molecular modeling of spectinomycin (SPC) and tetracycline (TET) on the (a) M. smegmatis efflux pump TetV_Ms and (b) M. abscessus efflux pump Mab_2780c.

The binding mode of the highest glide docking score of SPC (cyan) and TET (green) are shown. The proteins are cartoon ribbons with residues interacting with the ligand labeled and represented as sticks and color-coded to match the ligand. The dashed lines represent hydrogen bonding and salt bridging interactions.

Table 3. The Glide docking scores of Spectinomycin and Tetracycline using Schrdinger.

The highest scored SPC and TET molecules docked onto the TetV_MS (Ms5187) and Mab2780c AlphaFold-Colab homology models using Glide in Schrödinger Release 2022–3. The Glide docking scores (g_score) are reported in kcal mol⁻¹

Receptor	SPC gscore (kcal mol−1)	TET gscore (kcal mol−1)
TetVMs(Ms5187)	−6.13	−6.19
Mab2780c	−6.11	−4.65

As anticipated from the antibacterial susceptibility testing, the molecular docking reveals that Mab2780c has more favorable interactions with SPC than TET (Table 3 and Figure 6b). For SPC, the highest-scored pose makes several hydrogen bonding contacts with W61, T62, and E278 with a docking score of −6.11 kcal mol⁻¹ (Table 3 and Figure 6b). In contrast to TetV_Ms, the highest-scored TET pose overlays with the SPC site with a modest docking score, −4.65 kcal mol⁻¹, suggesting a poor binding pose on the receptor. The proposed binding mode of TET makes hydrogen bonding interactions with E149, Q184, and E278 (Figure 6b). Despite hydrogen bonding interactions, TET has a lower docking score than SPC on Mab2780c, suggesting an unfavorable docked pose on the receptor that would result in a lower binding affinity. From the molecular docking, we can hypothesize that the preferential selectivity of SPC by Mab2780c compared to the TetV_Ms pump is due to the number of protein-ligand interactions driving SPC binding. Molecular dynamics followed by energy decomposition of the proposed binding models would shed light on the binding affinities of SPC on Mab2780c and TetV_Ms and are under investigation for future studies.

Summary

A number of antibiotic resistance determinants have been described over the past decade that shed light on the extreme drug resistance of M. abscessus. The vast majority of these determinants that confer high-level drug resistance do so by modification of the target or the antibiotic. While efflux pumps confer high-levels of resistance in Gram-negative bacteria, they typically only confer low-levels of resistance in mycobacteria and the roles of several putative efflux pumps remain largely unresolved. To our knowledge, this is the first demonstration of an M. abscessus efflux pump that confers high-level resistance, specifically to SPC. Furthermore, we demonstrate that SPC can function synergistically with the efflux pump inhibitor, VMP, reducing the MIC of SPC in wildtype M. abscessus >100 fold to levels similar to that of SPC alone in ΔMab_2780c bacteria. VMP, a clinically approved drug used for treatment of cardiovascular diseases, as well as its analog KSV21, have been previously shown to potentiate the bactericidal activity of anti-TB drugs against intracellular M. tuberculosis as well as in BALB/c mice making these adjuvants attractive for evaluation as combinatorial therapy with SPC against M. abscessus (31, 32). Additionally, the development of completive inhibitors that specifically target MAB2780c also remains a possibility. EPIs are nonetheless associated with adverse side effects at high concentration and cytotoxicity studies of verapamil and its derivatives would need to be investigated to optimize dosage. In contrast, semisynthetic SPC analogs that have previously been shown to evade MFS-based efflux mechanisms in M. tuberculosis offer a promising alternative option for evaluation and redevelopment for use in M. abscessus therapy (28, 33).

Materials and Methods

‘Media and Strains.

Mycobacterium abscessus was grown at 37°C in Middlebrook 7H9 (DIFCO) supplemented with 10% OADC and 0.05% Tween 80. Antibiotics were added as required to indicated amounts. An isogenic deletion in MAB_2780c was constructed using recombineering, followed by removal of the apramycin cassette by Cre mediated recombination at loxP sites as described previously (15). The mutant was confirmed by PCR primers flanking the deletion site as well as by sequencing. Complementing strains were generated by transforming ΔMAB_2780c with pHspMab_2780c or phspMs_tetV, ΔMAB_whiB7 with pHspMab_2780c and ΔMAB_tetX with pHspMab_2780c or phspMs_tetV, followed by selection on 7H10-kan and PCR confirmation.

RNA preparation, qPCR and RNA-Seq analysis.

Wild type M. abscessus ATCC 19977 was grown to exponential phase (0.7 OD) in Middlebrook 7H9-OADC and exposed to varying concentrations of SPC for varying periods of time (0–120mins) and evaluated for lethality. Total RNA was prepared from wild type strains exposed to 256 μg/ml of SPC for 2 hours, using the Qiagen RNA preparation kit followed by DNAse I treatment. Unexposed samples were used as controls. Approximately 5 μg total RNA samples were treated with the Ribo-Zero^™ rRNA removal procedure (Illumina) to enrich for mRNA. Approximately 500 ng of RNA was used for library preparation using the NEBNEXT Ultra II DNA library kit and high throughput sequencing on the Illumina NextSeq platform. The sequence data was analyzed using Rockhopper in which the data is normalized by upper quartile normalization and transcript abundance is reported as RPKM. Differential gene expression is tested for each transcript and q-values are then reported that control the false discovery rate (34, 35). All experiments were performed in biologic duplicates.

Antibiotic Susceptibility Assays.

Wild type and mutant strains of M. smegmatis and M. abscessus were grown to an A₆₀₀ of 0.6–0.7. Cells were tested for their drug susceptibility by spotting a 10-fold serial dilution on Middlebrook 7H10 (DIFCO) plates containing the indicated concentration of each drug. Minimum Inhibitory concentration was determined using the macrobroth dilution method in which a two-fold dilution series of each antibiotic in 2 ml Middlebrook 7H9 + 10% OADC +0.05% Tween-80was inoculated with a given strain at an initial A₆₀₀ of 0.0004 with or without the addition of 128/256 μg/ml of VMP. Previous experiments were performed that determined the MIC of VMP in M. abscessus as 512 μg/ml. The cultures were incubated at 37°C and the A₆₀₀ was measured after 72 hs. All experiments were performed in biological triplicate.

Checkerboard Assays (Methods)

Whole cell in vitro synergy assays were performed in triplicate using M. abscessus ATCC 19977, M. abscessus ∆2780c, and M. abscessus ∆2780c :: phspMab_2780c strains. Drug source plates were prepared in 96-well plates by two-fold serial dilution of SPC (Sigma; CAS 22189–32-8) in water (Range 102.4–0.1 mg/mL; highest to lowest concentration in columns 1 to 11, respectively, with no drug in column 12) and verapamil (VMP; Sigma, CAS 152–11-4) in DMSO (Range 51.2–0.05 mg/mL in rows A to G, respectively, with no drug in column H). Using a Biomek FXP liquid handling robot (Beckman Coulter, CA), 4 μL of each drug source plate was transferred to each 96-well assay plate containing 92 μL Middlebrook 7H9 + 10% OADC and 0.05% Tween-80 (2% DMSO; final concentration range of 2048–2 μg/mL and 1024–1 μg/mL for SPC and VMP, respectively). Each well of the assay plate was inoculated with bacteria diluted to an initial A₆₀₀ of 0.0005 in media and incubated for 48 hours at 37°C. At 48 hours, 20 μL of 0.77 M resazurin (resazurin sodium salt; Sigma) was added to each well via a Multidrop Combi liquid dispenser (Thermo) for a final resazurin concentration of 70 μM and incubated for an additional 24 hours at 37°C. After incubation, fluorescent intensity (540 nm excitation and 590 nm emission) was measured using a PHERAstar FS Multilabel reader (BMG, Cary, NC). Percent viability was estimated by normalizing data to wells containing the highest concentration of both compounds (well A1, 0%) and no drug (well H12, 100%) for each plate. The median % viability of three replicates was plotted in GraphPad Prism version 9.4.1 (GraphPad Software, La Jolla, CA). Fractional inhibitory concentration index (FICI) scores were calculated for each replicate as previously described and interpreted as synergistic (≤0.5), indifferent (>0.5–2), and antagonistic (>2) (36).

SPC Efflux Assay

Detection of SPC was carried out using modifications of a previously described assay (37). M. abscessus ATCC 19977 and M. abscessus ∆2780c were grown from single colonies in 150 mL of Middlebrook 7H9 containing 10% OADC and 0.05% Tween-80 at 37°C with shaking to an optical density (OD₆₀₀) of 0.6–0.8. The bacteria were harvested at 4,700 rpm for 10 minutes at 4°C and the supernatant was discarded. The cells were washed in 50 mL of buffer PBS+ buffer [PBS + 0.05% Tween-80 + 5% glucose (w/v)] and pelleted as before, and the supernatant was discarded. Glucose was added to energize efflux process in the minimal buffer. The pellets were then resuspended in 3 mL of PBS+ buffer and placed in a glass tube. The concentrated cell mixture was exposed to 250 μM SPC for 30 minutes at 37°C in a shaking incubator followed by centrifugation at 4,700 rpm for 10 minutes at 4°C. The cell pellet was washed twice in 50 mL of PBS+ buffer and resuspended in 3 mL PBS+ buffer. The number of colony-forming units (CFU) was determined by serial dilution. 800 μL of the washed concentrated cell mixture was layered on 700 μL of silicone oil and centrifuged through the oil at 14,500 rcf for 2 minutes at 4°C as described previously (38). After centrifugation through the oil, 400 μL of the supernatant above the oil was collected and mixed 1:1 with 5% TCA. The remaining 2.2 mL of concentrated cell mixture was placed back in the shaking incubator for 4 hours at 37°C. This process was repeated at the 4-hour time point. All samples were filtered using 1.5 mL centrifugal 0.22 μm filter tubes (Millipore) and analyzed by LC-MS/MS.

Samples were analyzed using an Acquity UPLC (Waters) coupled with 6500 Triple Quad System (AB Sciex) with Acquity UPLC (Waters). 10 μL of extract was separated using an Acquity UPLC HSS T3 1.8µm, 2.1 × 50 mm column, with solvent A (0.1% Heptafluorobutyric acid (HFBA) in MilliQ Water) and solvent B (0.1%HFBA in Acetonitrile). The inlet method for these samples utilized a flow rate of 0.9 mL min−1 with the following gradient: 0−0.5 min (diverted to waste), 99.0% solvent A and 1% solvent B; 0.5–1.4 min, gradient of 95–5% solvent A and 5–95% solvent B; 1.4–1.8 min, hold on 5% solvent A and 95% solvent B; 1.80–1.82 min gradient of 5–99.0% solvent A and 95–1% solvent B; and 1.82–2 min, hold on 99.0% solvent A and 1% solvent B. Mass spectra were acquired positive electrospray ionization at the ion spray voltage of 5,500 V. The source temperature was 600°C. The curtain gas, ion source gas 1, and ion source gas 2 were 30, 60, and 60 respectively. Multiple reaction monitoring was used to quantify the SPC metabolite based on linear calibration curve (daughter peak: 207.1 Da).

All experiments were performed in biologic triplicate and data was plotted as mean and standard deviation in GraphPad Prism (v9.3.1).

Generating homology models of TetV_Ms and Mab2780c efflux pumps using AlphaFold-Colab

The three-dimensional (3D) Msmeg_5187 and Mab_2780c protein structures were generated using AlphaFold-Colab (39). AlphaFold-Colab is a rapid, open-sourced tool for protein folding by combining the computational power of Google Collaboratory with the artificial intelligence capabilities of AlphaFold 2.0 (40). Using AlphaFold-Colab in ChimeraX-1.4, the protein sequences of TetV_MS and Mab2780c were used to predict the 3D structure. The protein structures were exported as PDB files for molecular docking studies.

Molecular Docking of Spectinomycin and TET on TetV_Ms and Mab2780c

The preparation of ligands and receptors, the molecular docking, and visualization of docked poses were performed using Schrödinger Release 2022–3 (41). To prepare the ligands, the SMILES strings of spectinomycin (SPC) and TET (TET) were used to generate 3D structures. All possible ionization states at pH 7.0 were generated for each ligand and minimized using an OPLS4 force field (42). The 3D models of TetV_Ms and Mab_2780c from AlphaFold-Colab were prepped by protonating residues at pH 7.4 and minimizing the structure with an OPLS4 forcefield (42). With no crystal structure or experimental data pinpointing the binding sites, SiteMap in Schrödinger was used to identify potential binding surfaces on both TetV_Ms and Mab2780c models (43, 44). The predicted binding surfaces were then used to place the receptor grid for the docking studies. Using the docking grids, the prepped SPC and TET molecules were docked using Glide with default settings, except for the output, which was set to 100 poses per ligand (45–47). The docked poses were filtered by the docking score (g_score), and the highest-scored ligands were visualized to delineate protein-ligand interaction driving substrate specificity.