Riparin-B as a Potential Inhibitor of AdeABC Efflux System from Acinetobacter baumannii

Acinetobacter baumannii is an important opportunistic pathogen that causes serious health-related infections, especially in intensive care units. The present study aimed to investigate the antimicrobial activity of Riparin-B (Rip-B) alone and in association with norfloxacin against multidrug-resistant clinical isolates of A. baumannii. For this, the minimum inhibitory concentrations were determined by the microdilution method. For the evaluation of resistance-modulating activity, MIC values for antibiotics were determined in the presence or absence of subinhibitory concentrations of Rip-B or chlorpromazine (CPZ). The AdeABC-AdeRS efflux system genes from these isolates were detected by PCR. Docking studies were also carried out to evaluate the interaction of Riparin-B and the AdeABC-AdeRS efflux system. The study was conducted from 2017 to 2019. The results showed that Rip-B showed weak intrinsic activity against the strains tested. On the other hand, Rip-B was able to modulate norfloxacin's response against A. baumannii strains that express efflux pump-mediated resistance. Docking studies provided projections of the interaction between Rip-B and EtBr with the AdeB protein, suggesting that Rip-B acts by competitive inhibition with the drug. Results found by in vitro and in silico assays suggest that Rip‐B, in combination with norfloxacin, has the potential to treat infections caused by multidrug-resistant A. baumanni with efflux pump resistance.


Introduction
Infections caused by multidrug-resistant micro-organisms are a current problem of public health all over the world and they have increased the morbidity and mortality rates in the houses, and among patients whose care demands the use of ventilators and intravenous catheters [3].
Te A. baumannii is an aerobic, nonfermented, Gramnegative, coccobacillus labeled as an opportunistic pathogen widely associated with infection outbreaks related to health assistance, verifed in intensive center unity environments [4]. Infections caused by that pathogen are intricately linked to pneumonia, septicemia, meningitis, urinary tract infections, surgical wounds, previous antibiotic therapy, burns, and immunosuppression diagnoses [5].
A. baumannii shows unconventional resistance to multiple drugs, and it may also bear countless protection mechanisms against most regular antibiotic substances used in its treatment [6]. Tis resistance might come from uncounted mechanisms, intrinsic or acquired because of an overproduction of efux pumps, permeability reduction of the outer membrane, amendment of the target site, and production of beta-lactamases [7]. Similar mechanisms are also verifed for resistance to biocides acquired by clinical isolates of A. baumannii [8].
Te search for alternative therapies to resistant microorganisms inspires research on the microbicidal potential of some plants and its metabolites [9,10] such as the substance alcamida, separated from the unripe fruit of Aniba riparia (Nees) Mez belonging to Lauraceae family. Riparins I, II, and III are substances that occur naturally isolated from A. riparia. Tese chemical compounds derived from Riparins showed diverse biological activities. In addition to the natural occurrence of molecules originating in A. riparia, synthetic Riparins such as Riparin-B (Rip-B) was obtained by the Schotten-Bauman reaction [11].
Previous research carried out by our group pointed out that Rip-B was enhanced the action of norfoxacin and ciprofoxacin by inhibition of NorA efux pump of Staphylococcus aureus [12]. Tis result motivated us to investigate if Rip-B could modulate the resistance of clinical isolates of A. baumanni carrying the AdeABC-AdeRS efux system.

Intrinsic Antimicrobial Activity of Riparin-B.
In the present study, the synthetic derivative Rip-B (Figure 1) was investigated for its potential antimicrobial activity against diferent A. baumannii strains isolated from an urgency hospital of Teresina, Piauí, Brazil. Rip-B is a nonheterocyclic alkaloid, more specifcally, a natural alkamide. It is formed from the union of tyramine, a phenylalanine and benzoic acid. It adds a substitution in the ring of benzoic acid, adding a hydroxyl, and the tyramine ring presents a methyl linked to oxygen forming an ether function.
Our results showed that Rip-B did not have intrinsic antimicrobial activity against multidrug-resistant A. baumannii strains tested (Table 1).

Evaluation of the Drug-Resistance Modulation.
In the tests where Rip-B was added at subinhibitory concentration there was a signifcant reduction in the MIC values for norfoxacin, as it happens when a known efux pump inhibitor (CPZ) is put in combination with norfoxacin ( Figure 2). Tese results indicate that Rip-B is a modulating agent of the drug resistance in A. baumannii strains.
Ethidium bromide (EtBr) is a genotoxic dye that intercalates into DNA double-helix leading, to DNA damage and cell death, and the only known mechanism of resistance to EtBr in bacteria is mediated by the efux pump [13]. Tus, this intercalating agent has been used as an indicator of the resistance mediated by efux pumps [14]. In the present study, the modulating efect of Rip-B on the resistance to EtBr was verifed only in the strains expressing resistance mediated by efux pump phenotype ( Figure 3). Tese results are strong evidence that modulating of the antibiotic resistance by Rip-B could be due to efux pump inhibition. Table 2 shows the presence of genes from the AdeABC-AdeRS efux system among the resistant strains of A. baumannii multidrug (MDR) used in this study. It was observed that the strains tested have genes from the AdeABC-AdeRS efux system. Te three strains also amplifed the oxa-51 gene, confrming that it is A. baumannii. Table 2 also provides information on the resistance profle of the clinical isolates of A. baumanii. All A. baumannii clinical isolates tested were multidrugresistant (MDR) and were resistant to norfoxacin.

Docking.
To the best of our knowledge, there are no docking studies involving Rip-B on the AdeABC efux pump. Tus, to check if this substance could indeed act as an AdeABC inhibitor and to get some insights on the inhibition mechanism, we have carried out the docking of Riparin-B and the known substrate EtBr with the AdeABC structure as a target. Te AdeABC efux system is comprised of three parts: the AdeA periplasmic protein, the AdeB efux pump protein, and the AdeC outer membrane protein. AdeB has a 3-fold symmetrical structure with two periplasmic  , and is then delivered to the distal site for extrusion [15]. Tese sites are shown in Figure 4, labeled P (proximal) and D (distal). Figure 4 displays the best poses of both EtBr and Rip-B docked on the binding site of the AdeABC efux pump. A 2D protein-ligand diagram for EtBr is provided in Figure 5 and 2D protein-ligand diagram for Riparin-B is provided in Figure 6.

Discussion
Due to the high prevalence of infections caused by multidrug-resistant A. baumannii, mainly in hospital environments, the research of new antimicrobial agents effective against this pathogen has been considered as a priority. To contribute to this context, we analyzed the intrinsic antimicrobial activity of Rip-B against A. baumannii clinical isolates, as well as its modulating activity on the resistance to norfoxacin. Te molecular and susceptibility profle characterization of the clinical samples used in this study shows a multidrug-resistance profle and confrms the presence of genes from the AdeABC efux system. Our previous study reported a high percentage of extremely drug-resistant (XDR) strains (81.1%) that was verifed among A. baumannii clinical isolates at a northeast Brazilian Emergency Hospital, with greater prevalence of resistance to gentamicin (98.0%), ceftriaxone (94.3%), ceftazidime (92.0%), ciprofoxacin (90.5%), and levofoxacin (90.5%) [16]. Several studies report the prevalence of multiresistant and extremely resistant clinical isolates among strains of hospital origin, showing a great diversity of mobile genetic elements and the ability to acquire and expand expressed of the antimicrobial resistance factors [17][18][19][20][21]. Rip-B did not show intrinsic antimicrobial activity; however, enhanced the activity of norfoxacin against clinic strains of A. baumannii, showing that it could be useful as an adjuvant of norfoxacin in the treatment of infections caused by multidrug-resistant A. baumannii. Te antibacterial activity of natural Riparins I, III, and XII isolated from A. riparia against multidrug-resistant S. aureus and E. coli strains has already previously reported [22]. Moreover, while the synthetic analog Rip-E showed a good antibacterial activity against S. aureus strains, the synthetic derivative Rip-B was inactive against S. aureus [11]. In the present study, we also verifed that Rip-B was inactive against all A. baumannii strains tested once MIC values were found higher than 1000 μg/mL (Table 1) which is considered as clinically insignifcant [23]. Tese results show that Rip-B it is not recommended for use as an antimicrobial agent.
Te natural Rip-III isolated from A. riparia was able to induce plasmid elimination in S. aureus strains changing the Penicillin resistance phenotype to sensitive [24]. Furthermore, Rip-B was reported as being an inhibitor of the NorA, an efux pump of S. aureus belonging to the Major

Evidence-Based Complementary and Alternative Medicine
Facilitator Family [12]. Tese studies motivated us to investigate if Rip-B were able to potentiate the norfoxacin activity against A. baumannii strains expressing-resistance mediated by efux pump phenotype, with the goal of suggesting a possible use of Rip-B as an adjuvant of this antibiotic in the treatment of A. baumannii infections. Te results in Figure 2 suggest that Rip-B inhibits the efux pump mechanism by decreasing MIC values for norfoxacin, modulating the action of this drug in the clinical strains of A. baumannii. Tis hypothesis was corroborated by results obtained for the known NorA inhibitor CPZ [25] that also modulated the antibiotic resistance in the A. baumanii strains expressing the AdeABC-AdeRS genes.
In A. baumannii, resistance to Norfoxacin can be mediated by the multidrug efux pump AdeABC, a proton-    motive force dependent efux pump belongs to the Resistance and Nodulation Cell Division (RND) [27][28][29][30][31]. It is observed that in the strain HUT90 the presence of all genes of the AdeABC-AdeRS complex was detected, while in strains HUT 89 and 105 the presence of the AdeB gene was detected, however, there was no amplifcation of the regulatory AdeR and AdeS genes (  [33]. A previous study also point out that the expression of the AdeA, AdeB, and AdeC genes are inconsistent and that despite variations in detection rates, there is a predominance of amplifcation of the AdeB gene in clinical isolates of A. baumannii [28].

Evidence-Based Complementary and Alternative Medicine
Another important point to consider is that functional mutations have been found in conserved domains of AdeRS in all strains that overexpress AdeABC [28,34]. Furthermore, it is known that functional mutations in the insertion sequence (IS) in AdeRS can also afect the regulation of the AdeABC system, leading to overexpression of the system [35]. Te way in which the AdeR and AdeS genes regulate the overexpression of the AdeABC system is not fully understood. Te analysis of the interaction between AdeR and AdeABC by electrophoresis mobility shift rehearsal and found that the promoters AdeR and AdeABC did not interact [36]. Even if AdeS was present, AdeR was not found to link to the promoting region of AdeABC. In this way, it is known that AdeRS that regulates the expression of AdeABC is defned, but the mode of action is still difcult to specify. In addition, the regulation of AdeABC gene expression is complex. Under some conditions, the insertion of ISAbaI does not lead to overexpression of this pump [37], indicating that other regulators may be involved. Lin et al. [38] also showed that the other two-component system, BaeSR, can regulate AdeA and AdeB. Tus, in strains 89 and 105, regulation of the overexpression of the AdeABC genes may be occurring by mutation in the amplifcation regions or by another mechanism not yet known. Multidrug-resistant strains overexpressing efux pumps are able to extrude this antibiotic reducing its intracellular concentrations. Tus, it is possible that the strains tested in the present study, even with diferent genetic profles, could overexpress AdeABC that were inhibited by Rip-B. In its turn, inhibition of efux pumps by Rip-B could lead to a higher antibiotic accumulation in periplasm or cytoplasm where antibiotic targets are located enhancing their activity.
Previous study documented that both Holarrhena antidysenterica extract and conessine, a steroidal alkaloid compound, could restore antibiotic activity due to interference with the AdeIJK pump in A. baumannii, not interfering with the AdeABC pump [39]. However, the presence of AdeABC efux pump genes in the strains of clinical origin in the study had not been confrmed, unlike what was done in this study. Te same alkaloid, conessine, was shown to be efective as an efux pump inhibitor in Pseudomonas aeruginosa [40]. Both studies indicated the potential of an alkaloid product as an inhibitor of the homologous resistance-nodulation-splitting (RND) family.
Docking studies of Rip-B on the NorA efux pump of Staphylococcus aureus were already reported by Costa et al. [12]. Figure 4 shows the possible binding sites of a drug with the AdeB protein, highlighting the proximal and distal sites. As reported by Su et al. [15], residues Phe179, Phe277, Ile607, and Trp610 create a hydrophobic patch that could be connected to the stabilization of substrates. Te best pose of EtBr (with a binding energy of − 7.3 kcal/mol) makes close contact with Trp610 and to some residues of the G-loop, such as Gly611, Phe612, and Gly614. It also interacts through a hydrogen bond (2.24Å) with Asp83. Rip-B also bind to this region (with binding energy of − 7.3 kcal/mol), in-between both binding sites, interacting with residues 610 through 615 of the G-loop. Tere is also a hydrogen bond with Gly615, and close contacts with several residues, with two in particular: Tr91 and Ser134. It is reported that the antibiotic gentamicin interacts with these residues and ciprofoxacin binding site overlaps with that of gentamicin (10.1128/ mBio.01295-19). Interestingly, Tyr77, Tr91, and Ser134 are conserved residues between AdeB and the MexY pump. Tere's reason to believe, therefore, that these residues play a signifcant role in drug recognition. As both EtBr and Rip-B bind to the same region of the binding site one could argue that Rip-B could act as a competitive inhibitor, preventing the extrusion of drugs as norfoxacin and EtBr. Binding near the G-loop, Rip-B could, for instance, hinder the passing of drugs from the proximal to the distal site. Its preferred site also overlaps with that of antibiotics such as gentamicin. Tus, they could also be expelled in place of the antibiotic, acting as a competitive inhibitor.

Strains and Chemicals.
Evaluation of the Rip-B antimicrobial activity was performed against multidrugresistant A. baumannii strains isolated from patients attended in an urgency hospital from Teresina, Piauí, Brazil. Te isolates were collected from cultures of respiratory tract specimens (HUT 89 e HUT 90) and spinal cerebral fuid aspirate (HUT 105). Bacterial strain isolation was performed in blood agar followed by subculture in MacConkey Agar (Sigma-Aldrich) using duplicates. Bacterial strain identifcation was performed using the BD PHOENIX 5.1 automation method (Becton Dickison Sparks, MD 21152, USA) and confrmed on blaOXA-51 polymerase chain reaction (PCR), as described in item 4.6.
Assays for evaluation of the modulating efect of Rip-B on the antibiotic resistance were performed with multidrugresistant A. baumannii (according to item 4.2) previously screened to resistance mediated by efux pump phenotype with carbonylcyanide m-chlorophenylhydrazone (CCCP, Sigma-Aldrich), as described in item 4.3. For a comparison, also were conducted assays with strains expressing resistance not mediated by efux pumps. Bacterial strains were maintained on Brain Heart Infusion Agar (BHIA, Himedia, India) slants at 4°C, and prior to the assay the cells were grown overnight at 37°C in Brain Heart Infusion (BHI, Himedia, India). Norfoxacin (Nor), ethidium bromide (EtBr), and chlorpromazine (CPZ) were obtained from Sigma Chemical Corp., St. Louis. Nor was dissolved in a mixture of 1 M NaOH and sterile distilled water (1 : 9 proportion). EtBr and CPZ were dissolved in sterile water. N-[2-(3,4-dimethoxyphenyl)ethyl]-benzamide (Riparin-B, Brazil) was prepared in dimethyl sulfoxide (DMSO, Merck) and then diluted with sterile water.

Determination of the Occurrence of Resistance Mediated by Efux Pump Phenotype.
To verify the occurrence of efux pump-mediated resistance, MIC values of amikacin, ceftazidime or norfoxacin were determined in the presence or absence of a carbonyl cyanide m-chlorophenylhydrazone (CCCP, Sigma-Aldrich) solution at the subinhibitory concentration. CCCP is a decoupler from oxidative phosphorylation that interrupts the proton gradient of membranes. Microtitration plates were incubated at 37°C for 24 hours, and following this time 20 μl of a 0.01% (w/v) aqueous Resazurin sodium (Sigma-Aldrich) solution was added to each well. Tese plates were incubated for 1 hour at room temperature, where following this period a reading was performed taking into account that a change in coloration from blue to pink indicated the occurrence of bacterial growth due to resazurin reduction [42,43]. As a criterion for classifying the occurrence of resistance mediated by the efux pump phenotype, a minimum of a 2-fold antibiotic MIC reduction in the presence of CCCP was necessary [22].

Assays for Evaluation of the Intrinsic Antimicrobial
Activity. Stock solutions of Rip-B or CPZ were prepared in DMSO (Merck), followed by dilution in sterile distilled water to a fnal concentration of 1024 μg·mL − 1 . Minimal inhibitory concentrations (MICs) were determined by microdilution assay in BHI broth with bacterial suspensions of approximately 105 CFU·mL − 1 and concentrations of Rip-B (or CPZ) solution ranging from 8 to 512 μg·mL − 1 . Microtiter plates were incubated at 37°C for 24 h, and then 20 μL of resazurin (0.01% w/v in sterile distilled water) was added to each well to detect bacterial growth by a visual color change from blue to pink as described above.

Assays for Evaluation of the Drug-Resistance Modulation.
To evaluate if Rip-B was able to modulate antibiotic resistance in A. baumannii strains expressing or not resistance mediated by efux pump phenotype, MIC value for norfoxacin was determined in the presence or absence of Rip-B solution at subinhibitory concentration (MIC 1/8). Antibiotic concentrations ranged from 0.125 to 128 μg·mL − 1 . Microtiter plates were incubated at 37°C for 24 h and readings were performed with resazurin as described above. To verify if the drug-resistance modulation occurred due to efux pump inhibition, modulation assays were performed replacing antibiotics by EtBr, which is a known substrate of efux pumps [13]. Control assays were also performed replacing Rip-B y CPZ (Sigma-Aldrich) which is a known efux pump inhibitor [19]. Microtiter plates were incubated at 37°C for 24 h, and then 20 μL of resazurin (0.01% w/v in sterile distilled water) was added to each well to detect bacterial growth by a visual color change from blue to pink as described above.

Docking Procedure.
Te structure for the AdeABC efux pump was downloaded from the RCSB.org [47] site (PDB-ID: 6OWS). Te structure was uploaded to the MolProbity Server for protonation [48]. Ligands were optimized using the Gaussian 09 Program [49]. Initial structures were created in the GaussView module and were then optimized using the PM6 forcefeld. Partial Gasteiger charges were added to the protein (although Autodock Vina ignores it) and to the ligand atoms using the AutoDock Tools interface [50], nonpolar hydrogen atoms were mixed while all other parameters were kept at their default values. Docking procedure was carried out using the Autodock Vina software [51]. Te docking poses were chosen based on the best binding score.

Statistical
Analysis. Experiments were performed in triplicate and results were normalized by calculation of geometric mean values. Te error deviation and standard deviation of the geometric mean were revealed. Statistical analyzes were performed using GraphPad Prism, version 5.02. Diferences between treatment with antibiotics (or EtBr) alone or associated with Rip-B or CPZ were examined using one-way analysis of variance (ANOVA). Te diferences mentioned above were analyzed by Bonferroni posttest and p < 0.05 were considered statistically signifcant.

Conclusions
Te combination of antibiotics with efux pump inhibitors may be a promising strategy for the treatment of infections caused by multidrug-resistant A. baumannii. Results obtained in the present study showed that Rip-B was inactive against A. baumannii strains. However, Rip-B was able to increase the activity of norfoxacin against strains of A. baumannii that exhibit efux pump-mediated resistance. Te results obtained through Rip-B/EtBr association tests suggested that the increase in antibiotic activity probably involves the inhibition of overexpressed efux pumps in the strains tested. Te clinical isolates of A. baumannii MDR had genes for the AdeABC-AdeRS efux system, with the AdeB gene present in all tested isolates. Docking studies show the interaction of Rip-B and EtBr with the proximal and distal sites of the AdeB protein, suggesting that Rip-B could act as a competitive inhibitor for both norfoxacin and EtBr, preventing the passage of these compounds from the proximal to the distal site. As a limitation, the present study did not investigate if Rip-B could inhibit the expression of the AdeABC-AdeRS genes. Tus, Rip-B could be applied as an adjuvant of norfoxacin in the treatment of infections caused by multi-resistant strains of A. baumannii with resistance by efux pumps. However, in vivo preclinical studies will be needed to verify if Rip-B could enhance the norfoxacin activity against A. baumannii strains expressing AdeABC-AdeRS genes genes in animal models experimentally infected.

Data Availability
Te data supporting the current study are given in the article.

Additional Points
Sample of the Riparin-B and clinical strains of the A. baumanii are available from the authors.

Conflicts of Interest
Te authors declare no conficts of interest.