Acinetobacter baumannii ATCC^® 17978^TM (ATCC^® 17978), ATCC^®19606^TM (ATCC^® 19606^T) and ATCC^® 17961^TM (ATCC^® 17961) were used for expression assays in order to cover different genetic backgrounds. Escherichia coli TOP10 was used as an intermediate host for all plasmid constructs. E. coli ATCC^® 25922^TM, Staphylococcus aureus ATCC^® 29213^TM and Pseudomonas aeruginosa ATCC^® 27853^TM were used as control strains in minimum inhibitory concentration (MIC) experiments according to guidelines by Clinical and Laboratory–Standards Institute for antimicrobials (CLSI, 2017). All bacterial strains were grown routinely at 37°C in Luria-Bertani broth (LB) or on LB agar plates (Condalab) and supplemented with gentamicin at an appropriate concentration (10–100 μg/ml) for transformants selection. Mueller–Hinton broth II (MHB-II, Sigma) was used for growth in both MIC and luminescence assays. For long-term storage, bacterial cultures were grown on LB supplemented with 20% glycerol and stored at −80°C for further use.

Minimum inhibitory concentration values were determined for a representative compound from different families of antimicrobials and disinfectants, covering a variety of mechanisms of action, and that included: colistin (COL), meropenem (MEM), rifampicin (RIF), tigecycline (TGC), chlorhexidine (CLX), and ethanol (EtOH). Product references and concentrations used are provided in Supplementary Table 1. Overnight cultures of each strain were adjusted to OD₆₀₀ = 0.1 in MHB-II and the MIC was determined according to the recommendations of the Clinical and Laboratory Standards Institute for antimicrobials and the Sociedad Española de Enfermedades Infecciosas y Microbiología Clínica (Picazo, 2000; CLSI, 2017). For each strain, MIC values were determined at least by triplicate on three different days.

Competent E. coli TOP10 cells were prepared for electrotransformation as follows. E. coli TOP10 cells were grown in LB at 37°C for 16 h. Cultures were refreshed by dilution 1:100 in 100 mL of LB and incubated for 3–4 h at 37°C with shaking at 200 rpm, until the OD₆₀₀ reached ∼0.5. Cells were harvested by centrifugation (6,000 rpm for 10 min), washed four times with cold 10% glycerol, and suspended in 2 mL of 10% glycerol. One hundred-microliter aliquots of competent cells were stored at −80°C until further use.

A. baumannii electrocompetent cells were prepared as reported elsewhere (Lucidi et al., 2019) with minor modifications. Briefly, bacteria were grown in 10 mL of LB at 37°C for 16 h. Bacterial cultures were diluted 1:100 into 100 mL of LB and the cells were grown for 24 h at 37°C with shaking at 200 rpm. Cells were harvested and washed as indicated above. Eighty-microliter aliquots of competent cells were stored at −80°C until further use.

Electroporation was performed using one aliquot of competent cells and 100–250 ng of plasmid in electroporation cuvettes of 0.1 cm electrode gap (Bio-Rad). After pulsing with Bacteria default settings (1.8 kv, ≥ 3.9 mS), the cells were recovered in 1 mL of LB for 1 h at 37°C with shaking at 200 rpm and then transformants were selected on LB agar plates with the appropriate gentamicin concentration.

To carry out gene expression experiments, the operons corresponding to the adeABC, adeFGH, adeIJK efflux pumps and the adeRS control system were located in the genome of the A. baumannii ATCC 19606^T strain (Accession number: CP046654). The 500 bp upstream of the start codon of the first gene in the operon were selected. Primers (Table 1) were designed from the first and last 20 nucleotides of the 500 bp sequence, and the target sequences of BamHI and NotI were included at the 5′ end of the forward and reverse primers, respectively. As the promoter regions of the four genes were highly similar among strains (≥98%, Supplementary Figure 1) one set of primers was used to amplify the 500 bp fragment in all the A. baumannii strains studied. All primers were obtained with desalted purification (Sigma).

Genomic DNA was extracted with the freeze/thaw method and further used for promoter amplification through conventional PCR. The reaction mixture was composed of: 1 × EconoTaq PLUS 2X Master Mix (Sigma-Aldrich), 1 μM of forward primer, 1 μM of reverse primer and 1 μL of genomic DNA extract as template, in a total volume of 50 μL.

PCR conditions consisted of 10 min at 95°C for initial denaturation, followed by 30 cycles of 30 s at 95°C, 30 s at 58°C, and 30 s at 72°C and then a final extension of 10 min at 72°C. Products were visualized under UV light after gel electrophoresis on 1% (wt/vol) agarose gels in 1 × TAE buffer (Tris-acetate-EDTA, pH 8.0) stained with 1 × Gel Red 10,000 × (Biotium). The bands corresponding to the promoters were excised and the DNA was purified using the NucleoSpin Gel and PCR Clean-up kit (Macherey-Nagel) according to manufacturer recommendations.

The pLPV1Z shuttle-vector, which includes the luciferase reporter systems (Lucidi et al., 2019), was used to evaluate changes in gene expression. A total of 12 plasmids were constructed including the promoters of the four operons, for the three A. baumannii strains of study. Both, the plasmid and the purified 500 bp amplicon including the promoters were double-digested with BamHI and NotI (New England Biolabs) for 30 min and 37°C. After purification using the NucleoSpin Gel and PCR Clean-up kit (Macherey-Nagel) according to manufacturer recommendations, promoter and vector were ligated with T4 DNA ligase (New England Biolabs) at a plasmid:insert ratio of 1:20–1:50 at 16°C for 20 h. The final plasmid constructions were used to transform electrocompetent E. coli TOP10 cells. Transformants were selected by plating on LB agar supplemented with gentamicin (50 μg/ml) and screened by conventional PCR with T3 and RBS primers (Table 1), performed as described above.

Verified plasmid constructs were extracted and purified from a 10 mL E. coli TOP10 culture in LB supplemented with gentamicin 50 μg/mL using the NucleSpin Plasmid kit (Macherey-Nagel) following the protocol recommended by the manufacturer, and further confirmed by sequencing. Subsequently, constructs were introduced into electrocompetent cells of the corresponding strain of A. baumannii by electroporation. The clones incorporating the plasmid were selected by gentamicin resistance (10–100 μg/mL) according to the requirements of each strain (Table 2). Transformants were screened by conventional PCR with T3 and RBS primers as indicated previously, and also by measuring bioluminescence emission of the luciferase-luciferin system in the Orion II Microplate Luminometer (Berthold).

Agar diffusion assays were used to visually assess the effect of subinhibitory concentrations of antimicrobials on the expression of the operons. Briefly, cultures were prepared as described above, with a culture adjusted to OD₆₀₀ = 0.1 to inoculate Mueller-Hinton agar plates (Oxoid) by swabbing. Up to 50 μL of the appropriate antimicrobial solution ranging from 25 × to 40,000 × of MIC (Supplementary Table 1) was placed in the center of the plate and incubated overnight at 37°C. A control with water was included as growth control and to assess the basal expression of the promoters in the absence of antimicrobials. Parental strains without any construct were also tested with all the antimicrobials as negative controls for bioluminescence. Bioluminescence was visualized using an IVIS Lumina XRMS series II in vivo Imaging System (Perkin Elmer) setting 1 s as exposure time for image capture.

Initial experiments were carried out to optimize inoculum and antimicrobial concentration, as well as the most appropriate time for measurement (data not shown). The optimized protocol was used in the subsequent experiments to establish promoter activity. Briefly, A. baumannii strains carrying their respective construct of pLPV1Z:PadeA, pLPV1Z:PadeF, pLPV1Z:PadeR and pLPV1Z:PadeI as well as the parental strain without plasmid were grown overnight at 37°C on LB agar with the appropriate gentamicin concentration. A subculture in LB maintaining the previous gentamicin concentration was incubated at 37°C for 20 h. These cultures were adjusted to an OD₆₀₀ = 0.1 and then diluted 100-fold to inoculate 5 mL of MHB-II supplemented with the appropriate antimicrobial, at a concentration corresponding to 1/4 of the MIC (Supplementary Table 1). A growth control containing media without antimicrobial was included in all the experiments and was used for subsequent data normalization. To calculate background, a sample consisting of LB supplemented with antimicrobial was also included. The OD₆₀₀ value and bioluminescence emission were recorded after incubation for 6 h (exponential phase) and 24 h (stationary phase) at 37°C and continuous shaking at 200 rpm. Cultures were grown in an incubator, and a 100 μl sample was taken at those time intervals to quantify OD₆₀₀ on the one hand and bioluminescence on the other hand, using the Tecan Sunrise Absorbance Reader Analyser (Tecan) and an Orion II Microplate Luminometer (Berthold) reader, respectively. All experiments were performed at least in quadruplicate on four different days.

To account for differences in growth between strains and conditions, bioluminescence was normalized to the OD₆₀₀ of the culture prior to fold-change calculation. To calculate the activity of the promoters, the normalized data was divided by that obtained from the corresponding growth control (culture without antimicrobial) in order to obtain the fold-change. For fold-change < 1 the inverse was calculated to obtain the real magnitude of the change and was expressed as negative growth in order to indicate decreased expression. In some cases, it was not possible to calculate the fold change of the strains because it did not grow sufficiently (OD₆₀₀ < 0.02), and/or no growth was recorded after 24 h of incubation.

First, we determined if the fold-change calculated was different from 1 (as we established 1 as no changes in expression) by a one sample T-test. Secondly, the Kruskal–Wallis test was used to calculate differences in expression between strains and compounds. Further pairwise comparisons by subcategories of these variables were analyzed using the Dunn’s multiple comparison test. In this case, non-parametric statistical tests were used given the non-normal distribution of the data and the lack of homoscedasticity as assessed by Kolmogorov–Smirnov and Levene tests, respectively. Finally, correlation between expression was assessed by strain by calculating the Pearson correlation coefficient. In this case, the mean values of fold-changes of the four replicates performed in each antimicrobial treatment were used. Graphpad (Prism) and SPSS (IBM) were used for data plotting and statistical analysis. A p-value ≤ 0.05 was considered significant.

All three strains of A. baumannii were sensitive to the antimicrobials used according to CLSI and EUCAST breakpoint tables for interpretation of MIC values (CLSI, 2012; The European Committee on Antimicrobial Susceptibility Testing, 2021). However, each strain featured a different susceptibility profile (Table 2). A. baumannii ATCC 17961 had the lowest MIC for COL, whereas strain ATCC 17878 was the most sensitive to MEM. The A. baumannii ATCC 19606^T MIC of RIF was the lowest among the three strains tested. MIC values ranged between 0.25–16 μg/mL, and variation in the range of two- to fourfold were observed between strains. With regard to disinfectants, A. baumannii ATCC 17978 demonstrated a higher MIC for EtOH, which was twofold that of the other strains (Table 2). In contrast, the CLX MIC was similar for all the A. baumannii strains tested.

We first aimed to characterize the bioluminescence produced by the different plasmid constructs when subjected to a concentration gradient of the different compounds to determine if there were qualitative differences in the expression from the RND promoters at subinhibitory concentrations of antimicrobials (Figures 1A,B).

In strains containing the adeABC promoter construct, a halo of bioluminescence was observed at subinhibitory concentrations of all the antimicrobials tested in strains ATCC 17978 and ATCC 19606^T. In contrast, for strain ATCC 17961 the halo of bioluminescence was only observed in the presence of CLX and MEM.

In the case of the adeFGH promoter, all strains featured a rather similar pattern. No bioluminescence was observed for any of the antimicrobials.

With regard to expression of adeIJK promoter, assays with water revealed a low basal expression in the case of ATCC 17978 and ATCC 17961 strains but not in ATCC 19606^T. A halo of increased bioluminescence was observed at subinhibitory concentrations in the presence of all the antimicrobials for ATCC 17978. This general phenomenon was also observed in the other two strains, but no bioluminescence was detected for strain ATCC 19606^T in the presence of CLX. For strain ATCC 17961, in assays with CLX only baseline levels of bioluminescence were observed. Altogether this result suggests a strain-specific response of this promoter to certain compounds.

Constitutive basal expression of adeRS promoters was also revealed in the absence of antimicrobials, especially for ATCC 19606^T and ATCC 17961. Nonetheless, halos of increased bioluminescence were observed with subinhibitory concentrations of all antimicrobials in all strains. This was especially clear in the presence of MEM in ATCC 19606^T. These antibiotics also induced differences in expression in adeABC, suggesting that they may affect the expression of both the efflux pump and the regulatory system.

To further study the effect of subinhibitory concentrations of antimicrobials on the expression of RND family members, quantitative assays were carried out by determining luciferase activity of the cultures treated with the different antimicrobials at 1/4 of the MIC. This allowed us to calculate fold changes as well as to determine whether there was overexpression or decreased expression of the promoter compared to growth in absence of antimicrobials.

Significant overexpression of adeABC was observed for EtOH (p = 0.023), CLX (p = 0.0001) and RIF (p = 0.046) in strain ATCC 17978 (Figure 2). It is remarkable that CLX caused over a 100-fold-change at the exponential phase of growth. This effect was still significant, although smaller, during the stationary phase (p = 0.039). CLX also caused a 45-fold increase in expression of this promoter in strain ATCC 17961 during the stationary phase (p = 0.047). None of the compounds tested had a significant effect on the expression of the adeABC promoter in strain ATCC 19606 during exponential and stationary growth.

For adeFGH promoter expression, almost all fold-change values were around 1 or below (Figure 3). This suggested no overexpression from these promoters at subinhibitory concentrations of the antimicrobials and was in line with the results observed in the qualitative assay showing no bioluminescence. Only EtOH (p = 0.001) showed significantly decreased expression during the stationary phase in strain ATCC 17978. For strains ATCC 19606^T and ATCC 17961, significant effect of the antimicrobials was observed only during exponential phase. EtOH (p = 0.0001) and MEM (p = 0.026) significantly decreased the expression of adeFGH in strain ATCC 19606^T. For strain ATCC 17961, MEM and of TGC had a significant effect on the expression from this promoter, however, they had opposite effects on promoter activity. Whereas MEM caused a 1.125-fold overexpression (p = 0.034), TGC caused a 1.25-fold decrease in expression (p = 0.0001).

With regards to expression from the adeIJK promoter (Figure 4), the effect on expression was mainly observed during the exponential phase. In strain ATCC 19606^T, CLX, COL and MEM significantly decreased the expression of this promoter (p ≤ 0.0001). For strain ATCC 17978, EtOH (p = 0.0001), CLX (p = 0.006) and TGC (p = 0.0001) also showed a significant downregulation. Interestingly, RIF caused a significant overexpression of this pump in all strains (p ≤ 0.025), ranging between 1.44 to 2-fold increase in expression. In the stationary phase, only a significant decreased expression was observed for TGC in strain ATCC 17978 (p = 0.0001).

For the adeRS promoter (Figure 5), EtOH, CLX, and MEM significantly decreased the expression from this promoter (p ≤ 0.026) during the exponential phase for strain ATCC 19606^T. The same effect was observed for CLX in strain ATCC 17978 (p = 0.032). No significant effect on the stationary phase was observed for any of the antimicrobials in the other two strains.

It is important to point out that the apparent inconsistencies between qualitative (Figure 1) and quantitative assays (Figures 2–5) are due to technical limitations of the qualitative assay, related to the detection limit. This assay represents an end-point picture where overexpression of promoters can be observed through increases in luminescent intensity. In these images, no changes in expression or reduced expression would be observed as no bioluminescence. In contrast, in the quantitative assay, by comparison with the control without antimicrobial, a decrease in expression can also be calculated. This is the case of adeFGH, for example, where quantitative analysis demonstrated that almost all fold-change values were around 1 or below, and therefore not detected in the qualitative assay.

Data from the quantitative expression assays were used to assess if the tested compounds affected the expression of the four promoters similarly in the three A. baumannii strains used in this study. Altogether, the result of this analysis showed differences among strains in promotors expression under exposure to certain compounds and also depending on growth phase of the culture (Figure 6).

For the adeABC promoter, significant differences between strains were observed in four out of the six antimicrobials tested. During the exponential growth phase, differences in the effect of EtOH and CLX were observed (p ≤ 0.03). In the stationary phase, significant differences were also obtained for CLX (p = 0.015) and MEM (p = 0.04). Two antimicrobials affected adeFGH promoter differently between strains. In this case, during exponential phase significant differences were found for CLX and MEM (p = 0.02). For the adeIJK promoter, no statistically significant differences were observed between strains with any of the compounds tested, regardless of growth phase. For the adeRS promoter, significant differences were observed with CLX and MEM treatments (p ≤ 0.03) in the exponential phase. In the stationary phase there were also significant differences after treatment with CLX and RIF (p ≤ 0.024). Overall, CLX was the compound that featured more differences in expression among strains, regardless of the growth phase.

To estimate if the expression of the different RND family members were related, based on the strain tested, for each strain we calculated pairwise correlations of the mean fold-changes obtained with all the antimicrobials tested together (Table 3). In general, few significant correlations were observed. During the exponential phase, negative correlation between adeABC and adeRS was found in strain ATCC 17978 whereas positive correlation between adeFGH and adeRS was found in strain ATCC 19606^T. In the stationary phase, positive correlation between adeIJK with adeABC and adeRS were found in strains ATCC 19606^T and ATCC 17961, respectively.