Homologs of the Acinetobacter baumannii AceI Transporter Represent a New Family of Bacterial Multidrug Efflux Systems

BTP protein gene cloning and expression.

At the time of writing, the Pfam database (version 27.0) listed close to 800 proteins that contain BTP domains from more than 600 bacterial species (2). The majority (95%) of these proteins were predicted to have the same tandem BTP domain architecture as AceI and were encoded by Proteobacteria, particularly the gamma, beta, and alpha subdivisions (although this may be biased by the species for which genome sequence data are available). In addition to Proteobacteria, the genomes of several Veillonella and Micrococcus species (Firmicutes and Actinobacteria, respectively) also carried BTP domain protein genes. We have not, however, detected these genes in the genomes of archaeal or eukaryotic organisms.

In this study, we examined the drug resistance/transport capabilities of 24 BTP domain proteins, including AceI. These proteins were selected to encompass the full spectrum of currently sampled phylogenetic diversity within this group (Fig. 1) and included 18 gammaproteobacterial proteins, 3 betaproteobacterial proteins, and 1 representative protein from each of Alphaproteobacteria, Firmicutes, and Actinobacteria (Table 1). Seven of the genes encoding these proteins were previously cloned into the E. coli pTTQ18 expression vector via conventional methods (1). The remaining 17 were synthetically designed E. coli codon-optimized sequences and were synthesized in single gBlock gene fragments (Integrated DNA Technologies) and then cloned into the pTTQ18 plasmid vector (6). With the exception of Vpar_0264, ROS217_23162 and MHA_0890, the proteins under investigation were expressed at levels detectable in Western blot assays of whole-cell lysates (see Fig. S1 in the supplemental material). This level of expression success (87.5%) is in line with our previous experiences using this expression system for the heterologous production of transport systems (7–9).

Chlorhexidine resistance mediated by AceI homologs.

Previously, we demonstrated that BTP domain proteins homologous to the A. baumannii AceI protein from A. baylyi ADP1 (ACIAD1978), P. aeruginosa PA14 (PA14_26850), P. protegens Pf-5 (PFL_4558), and B. cenocepacia HI2424 (Bcen2424_2356) were able to confer resistance to chlorhexidine, whereas the phylogenetically distinct (Fig. 1) proteins from A. baumannii (A1S_1503) and P. protegens (PFL_4585) did not confer resistance (1). Chlorhexidine MIC analyses were conducted as previously described (8) to gauge the level of resistance provided by the 17 newly cloned BTP domain protein homologs. These assays were conducted in medium containing 0.05 mM isopropyl-β-d-thiogalactopyranoside (IPTG) to induce basal levels of expression. Among the 24 cloned genes, we saw reproducible increases in chlorhexidine resistance from 12 genes (Table 1). This result confirmed that chlorhexidine is a common substrate of this group of transporters.

BTP domain proteins are multidrug efflux systems.

To explore the possibility that BTP domain proteins represent a new family of multidrug efflux transporters, we tested their capacities to confer resistance to a range of additional antimicrobial compounds in MIC analyses. These compounds included: the biocides dequalinium, tetraphenylphosphonium, and benzalkonium, the antibiotics tetracycline and chloramphenicol, and a number of fluorescent antimicrobial dyes that are common substrates of multidrug efflux systems, such as proflavine, acriflavine, ethidium, Hoechst 33342, pyronin Y, acridine yellow, and 4′,6-diamidino-2-phenylindole (Table 1).

Resistance to a variety of antimicrobials was common, particularly among the Pseudomonas and betaproteobacterial genes (Table 1). Six BTP domain protein genes conferred reproducible resistance to the biocide benzalkonium. Additionally, two Pseudomonas proteins, PA14_26850 and PFL_4585, mediated resistance to dequalinium. Notably, PFL_4585 is a paralog of the chlorhexidine resistance protein PFL_4558, demonstrating that both BTP domain proteins encoded by P. protegens Pf-5 mediate resistance to selected biocides. Nine of the BTP domain protein genes also conferred resistance to one or both of the DNA-intercalating antimicrobial dyes acriflavine and proflavine (Table 1). Only 6 genes out of the 24 examined in this study did not provide resistance to any of the antimicrobials tested. Of these, three were not expressed at levels detectable by Western blotting (see Fig. S1 in the supplemental material). The V. parahaemolyticus gene VP1155 provided the highest and most consistent increases in resistance to the greatest number of compounds, with at least 4-fold increases in resistance to chlorhexidine, benzalkonium, proflavine, and acriflavine. To determine whether resistance mediated by these proteins was dependent on a TolC-like outer membrane, we examined resistance mediated by AceI and VP1155 in a TolC-inactivated background (10). Both of these proteins mediated resistance in this mutant strain.

To explore further the substrate recognition profile of the VP1155 protein, we conducted Biolog OmniLog Phenotype microarray (PM) experiments (11). The resistance levels of E. coli BL21 carrying pTTQ18-VP1155 were compared to those of E. coli BL21 carrying pTTQ18 for 240 different antimicrobials in the PM11-20 plate series, as previously described (1, 11). These tests confirmed that VP1155 provides resistance to the substrates identified by conventional MIC assays: chlorhexidine, proflavine, and acriflavine (see Fig. S2 in the supplemental material; benzalkonium is not included in the PM11-20 panel of compounds). Furthermore, the Biolog PM tests indicated that VP1155 also provides resistance to 9-aminoacridine, domiphen bromide, 3,5-diamino-1,2,4-triazole (guanazole), and plumbagin (see Fig. S2).

Fluorimetric transport assays demonstrate efflux mediated by VP1155.

The observations of resistance to the DNA-intercalating antimicrobial dyes proflavine and acriflavine from several of the BTP domain protein genes presented the opportunity to assay directly the efflux of these substrates in real-time fluorimetric transport assays. We applied these assays to cells expressing VP1155, exploiting the capacity of this protein to mediate resistance to both dyes. The assays were conducted in the E. coli triple deletion mutant strain BW25113 (ΔacrB ΔemrE ΔmdfA::kan), which is defective in the three major E. coli multidrug efflux system genes (12) and provides a sensitive background for these assays. Cells carrying the pTTQ18-VP1155 expression plasmid were assayed both pre- and postinduction of VP1155 expression by using 0.2 mM IPTG (see Fig. S3 in the supplemental material). Cells carrying “empty” pTTQ18 treated with IPTG were also included as a negative control. The transport assays were conducted essentially as described previously (13), except that cells were grown in glycerol-supplemented medium and reenergized by using glycerol to initiate transport from substrate-loaded cells.

The fluorescence intensity of both proflavine and acriflavine is lower when intercalated into DNA, so efflux from the cell was characterized by an increase in fluorescence over time. In our transport experiments we observed a rapid increase in fluorescence in cells that expressed the VP1155 protein but not in control cells lacking this protein (Fig. 2; see also Fig. S3 in the supplemental material). These results provide additional evidence that efflux is the mechanism of resistance operating in this group of proteins.

Conclusions.

In this study, we examined a large panel of genes encoding proteins related to the AceI chlorhexidine efflux system for their ability to confer resistance to a set of 12 different biocides, antibiotics, and antimicrobial dyes. To facilitate this broad survey of phylogenetically diverse proteins, we adopted a synthetic cloning approach; the majority of genes were codon optimized for expression in E. coli and chemically synthesized for cloning into our expression system. Of the 24 transport proteins studied, 21 were expressed at levels detectable in Western blot assays of whole-cell lysates, and 18 conferred resistance to one or more antimicrobial compounds.

Our results demonstrate that this group of proteins is a new family of bacterial multidrug efflux systems, which we have designated the proteobacterial antimicrobial compound efflux (PACE) family. The PACE family is only the sixth family of bacterial multidrug efflux systems to have been described and the first new family for more than 15 years. Multidrug efflux systems encoded by nosocomial pathogens are particularly problematic. Antimicrobial pressures selecting for increased expression of multidrug exporters can promote resistance to not only the selecting compound but also to a swath of otherwise-effective compounds.

Using a radiolabeled substrate, we previously demonstrated that the AceI protein within the PACE family is able to mediate the active efflux of chlorhexidine. In this work, we have identified fluorescent substrates for several members of the PACE family that facilitate the development of rapid fluorimetric efflux assays. These assays will be highly valuable in future studies to define the molecular transport mechanism operating in members of this family, including the mode of energization, which is likely to involve an electrochemical gradient.

Notably, PACE family proteins are encoded in the core genomes of many proteobacterial species that are separated by hundreds of millions of years of evolution. Given that the substrates we have now defined for these efflux systems—chlorhexidine, dequalinium, benzalkonium, proflavine, and acriflavine—are synthetic biocides that have only been widely used within the last century, it seems unlikely that these compounds are the physiological substrates of these transporters. Nonetheless, as with other multidrug efflux systems, the substrate promiscuity of PACE efflux systems is likely to have enhanced the success of proteobacterial pathogens in clinical settings.