Efflux Pump Activity and Mutations Driving Multidrug Resistance in Acinetobacter baumannii at a Tertiary Hospital in Pretoria, South Africa

Acinetobacter baumannii (A. baumannii) has developed several resistance mechanisms. The bacteria have been reported as origin of multiple outbreaks. This study aims to investigate the use of efflux pumps and quinolone resistance-associated genotypic mutations as mechanisms of resistance in A. baumannii isolates at a tertiary hospital. A total number of 103 A. baumannii isolates were investigated after identification and antimicrobial susceptibility testing by VITEK2 followed by PCR amplification of blaOXA-51. Conventional PCR amplification of the AdeABC efflux pump (adeB, adeS, and adeR) and quinolone (parC and gyrA) resistance genes were performed, followed by quantitative real-time PCR of AdeABC efflux pump genes. Phenotypic evaluation of efflux pump expression was performed by determining the difference between the MIC of tigecycline before and after exposure to an efflux pump inhibitor. The Sanger sequencing method was used to sequence the parC and gyrA amplicons. A phylogenetic tree was drawn using MEGA 4.0 to evaluate evolutionary relatedness of the strains. All the collected isolates were blaOXA-51-positive. High resistance to almost all the tested antibiotics was observed. Efflux pump was found in 75% of isolates as a mechanism of resistance. The study detected parC gene mutation in 60% and gyrA gene mutation in 85%, while 37% of isolates had mutations on both genes. A minimal evolutionary distance between the isolates was reported. The use of the AdeABC efflux pump system as an active mechanism of resistance combined with point mutation mainly in gyrA was shown to contribute to broaden the resistance spectrum of A. baumannii isolates.


Background
Acinetobacter baumannii (A. baumannii) is one of the nosocomial pathogens that exhibits high level of resistance to antibiotics [1,2].
Efflux pumps are the main mediators of resistance mechanisms against many antibiotic classes [21]. rough this mechanism, the bacteria avoid accumulation of drugs at the targeted site within it, leading to decreased susceptibility to antibiotics [17,21].
ree resistance-nodulation-cells division encoded in A. baumannii genome is reported to contribute to antibiotic resistance in A. baumannii clinical isolates [22]. e overexpression of efflux pumps in A. baumannii has been associated with an increased resistance to antibiotics such as tigecycline [14,23], carbapenems [24], minocycline, gentamycin, doxycycline, and tetracycline [25]. Several researchers support that among the resistancenodulation-cell division superfamily, the adverse effect of ATP-binding cassette (AdeABC) is the most associated with resistance in A. baumannii [26,27]. e system generally consists of three parts: multidrug transporter adeB that captures antibiotics in the inner membrane of phospholipids bilayer or cytoplasm, membrane fusion protein adeA that acts as membrane fusion protein, and finally, the outer membrane protein adeC that is a membrane channel protein, used by adeB to transport out the substrates [26]. is whole mechanism is regulated by adeR and adeS, a twocomponent system [26,27]. e working mechanism of this system suggests that adeB and the regulatory genes adeR and adeS are the main role players within the AdeABC efflux pump system [28]. is mechanism of resistance reduces the susceptibility of A. baumannii to multiple class of antibiotics [23,[28][29][30] and has been reported to be associated with resistance to newly developed drugs [30].
Alteration of target site or cellular function due to mutations has enabled emergence of resistance to antibiotics within the A. baumannii species [15,31]. is mechanism of resistance is used by A. baumannii to resist fluoroquinolones [32], colistin [33], novel synthetic beta-lactamase ceftazidime-avibactam [34], and several newly developed antibiotics [35]. Even though literature reports on the plasmid mediated quinolone resistance genes qnrA, qnrB, and qnrS as one of the resistance mechanisms to reduce susceptibility to quinolones [36,37], A. baumannii are mainly resistant to quinolones through chromosomal gene mutation in parC and gyrA [38,39] and other molecular mechanisms [37]. Literature supports that a single mutation in gyrA inducing an amino acid change from serine to leucine in position 83 (serine 83) reduces the susceptibility of A. baumannii clinical isolates to fluoroquinolones [40]. Two mutations in parC in position 80 (serine 80) and 84 (Glu-84) inducing change from serine to isoleucine and glutamic acid into valine, respectively, lead to resistance to fluoroquinolones [40]. e combined effect of active use of efflux pumps and alteration of target site due to mutations contribute to broaden the resistance spectrum of A. baumannii strains and pose serious therapeutic challenges to clinicians in establishment of effective treatment regime. Multiple outbreaks of A. baumannii infections have been observed globally with an increased resistance to antimicrobial drugs [41,42]. In this study, we focused on the use of AdeABC efflux pumps, parC and gyrA mutations, contributing to reduce A. baumannii clinical isolates susceptibility to antibiotics at an academic hospital in Pretoria. e aim was to investigate the use of the efflux pump and quinolone resistance-associated genes in the rise of multidrug resistant A. baumannii strains.

Study Design, Settings, and Samples Collection.
Isolates for this study were collected between February 2018 and February 2020 at Dr. George Mukhari Tertiary Laboratory (DGMTL), a unit of the National Health Laboratory Services (NHLS) of South Africa. DGMTL is a level 3 clinical laboratory where routine laboratory diagnostics for patients presenting at Dr. George Mukhari Academic Hospital (DGMAH), and 3 district hospitals and surrounding clinics are performed. DGMTL is coupled with the Department of Microbiological Pathology of Sefako Makgatho Health Sciences University (SMU). Ethical approval to conduct this research was granted by Sefako Makgatho Health Sciences University Research Ethics Committee (SMUREC). A total number of 103 A. baumannii isolates were collected from DGMTL and stored at −70 o C until use.

Isolate's Identification and Antimicrobial Susceptibility
Profiles. Collected isolates were identified using a phenotypic and genotypic method, VITEK2 automated system (bioMerieux, France) and polymerase chain reaction (PCR) amplification of bla OXA-51 gene.

Phenotypic Evaluation of AdeABC Efflux Pump adeB, adeS, and adeR Gene Expression.
A functional AdeABC efflux system was assessed by evaluating the difference between the minimal inhibitory concentrations (MICs) for tigecycline (TGC) using the gradient diffusion method (tigecycline, MIC Test Strip, Liofilchem® Srl, Roseto d'Abruzzi, Italy) before and after exposure to an efflux pump inhibitor (EPI) carbonyl cyanide 3-chlorophenylhydrazone (CCCP) (Sigma-Aldrich, Dorset, United Kingdom) [46]. Briefly, CCCP was added to one of two MH agar plates at the final concentration of 100 µg/mL. From a fresh overnight culture of A. baumannii, a 0.5 Mac Farhland turbid sample was made (Densichek, BioMerieux DensiCHEK Plus, USA). A swab was then used to spread A. baumannii on each agar plate, followed by placement of a 0.016-256 mg/L TGC E-strip test in the middle of each agar plate followed by an overnight incubation at 37°C. E. coli strain ATCC 85218 was used as a control.

Sequencing of Quinolone Resistance-Associated Genes (parC and gyrA).
e Sanger sequencing method (ABI3500XL; Applied Biosymptoms, United States) at Inqaba Biotec (Pretoria) was used to sequence the parC and gyrA amplicons. Sequences were edited using ChromasPro software (version 2.0) and then aligned together with wildtype sequences from GenBank using BioEdit (http://www. mbio.ncsu.edu/BioEdit/BioEdit.html). Serine 83 of gyrA gene and serine 80 and serine 84 of parC gene were investigated for mutation. A phylogenetic tree was drawn using molecular evolutionary genetic analysis (MEGA) 4.0 [47] to evaluate evolutionary relatedness of the study strains.

Isolate Identification and Antimicrobial Susceptibility
Profile. All A. baumannii isolates identified by VITEK2 were PCR-positive for bla  . e antibiotic susceptibility of the collected isolates was tested against 10 commercially available antibiotics. e study isolates were 89% resistant to cefotaxime and 85% resistant to both ceftazidime and cefepime (Figure 1). e isolates were 76% resistant to piperacillin + tazobactam, a combination of penicillin and beta-lactamase inhibitor (Figure 1). Eighty-three percent (83%) of the isolates were resistant to imipenem and meropenem (Figure 1), while resistance to gentamycin, trimethoprim-sulfamethoxazole, and ciprofloxacin was 81%, 82%, and 83%, respectively (Figure 1). Antibiotic susceptibility testing for tigecycline showed that 87% of isolates were susceptible, while 3% demonstrated intermediate susceptibility ( Figure 1).

Molecular and Phenotypic Evaluation of the Active AdeABC Efflux Pump System as the Resistance Mechanism.
e evaluation of an active AdeABC efflux pump system as a mechanism of resistance was conducted using a combination of genetic and phenotypic tests targeting adeB, adeR, and adeS genes. A positive PCR and qRT-PCR of targeted genes suggest that the genes of interest are present and that the related proteins are actively produced. e phenotypic evaluation confirms the actual use of the efflux pump system as the resistance mechanism at the phenotypic level. e positive PCR amplification of targeted genes was 100% for adeB and 99% for both adeR and adeS genes ( Table 1; Supplementary 3). e qRT-PCR was 100% positive for adeB and 99% and 98.1% for adeR and adeS, respectively. A total of 100 isolates (97%) were positive for PCR and qRT-PCR targeting adeB, adeR, and adeS genes. Among the 100 isolates, 75% of these isolates phenotypically demonstrated the active use of efflux pumps as a drug resistance mechanism (Table 1; Supplementary 3) and 25% did not show any phenotypic level of efflux pump involvement in resistance to tigecycline. Of the 25% that did not demonstrate phenotypic expression of efflux pump use, 88% (22 isolates) had the complete required set of genes for an active efflux pump (Table 1; Supplementary 3). ere was a statistically significant association between positive conventional PCR and qRT-PCR amplification of structural adeB and regulatory adeR genes and active phenotypic expression of the efflux pump (p value <0.05).

Investigation of Point Mutations in parC and gyrA.
e point mutations in parC and gyrA genes were investigated following Sanger sequencing of PCR products. All study isolates (100%) were PCR-positive for gyrA and 99% positive for parC (Table 2; Figure 2; Supplementary 4), while 2 samples failed quality control for gyrA Sanger sequencing and could not be sequenced. e sequence analysis of gyrA revealed that 89% of the isolates showed a point mutation in serine 83 inducing a change in amino acid from serine to leucine (Table 2; Figure 3; Supplementary 4), while 11% did not have this mutation. All the isolates (100%) were negative for point mutation on serine 84 of parC gene; while on serine 80, 39% of isolates had a point mutation inducing change in amino acid from serine to leucine. is point mutation was not observed in 61% of the isolate sequences that were analysed (Table 2; Figure 3; Supplementary 4). e analysis of the two gene sequences revealed that 36% (37 isolates) had both point mutations. Serine 83 for gyrA and serine 80 for parC and 10 isolates did not have any of the mutations ( Table 2

Discussion
Acinetobacter baumannii increasing the spectrum of resistance to available antibiotics is of a public health concern [1]. Similar to previous studies, PCR amplification of bla OXA-51 gene was used as a genotypic identification and confirmatory method for A. baumannii strains previously identified by VITEK2 (bioMerieux, France) [49][50][51]. Even though it has been reported in literature that some of A. baumannii strains do not harbour bla OXA-51 [53], the gene has been reported intrinsic to the species by several authors [49][50][51]53]. e use of bla OXA-51 has been recommended as a simple and reliable identification method for A. baumannii strains [49][50][51].
is study results reveal a high resistance to most of available antibiotics used by clinicians in management of A. baumannii infections (cefotaxime 89%, ceftazidime 85%, cefepime 85%, piperacillin + tazobactam 76%, imipenem 83%, meropenem 83%, ciprofloxacin 83%, gentamycin 81%, and trimethoprim-sulfamethoxazole 82%) (Figure 1). e findings are similar to several reports made by different researchers within the region [54][55][56] and globally [58]. In their published report on the district of Oliver Reginald Tambo in the Eastern Cape Province of South Africa, Anane et al. [55] highlighted that A. baumannii strains showed resistance rates above 80% against the same antibiotics tested in this current study. Similarly, Lowe et al. [57] in Tshwane district in Gauteng province of South Africa reported high prevalence of resistance (69-90%) of A. baumannii strains to the same antibiotics tested in the current study.
An increasing number of isolates with an intermediate susceptibility to tigecycline (3%) is of concern (Figure 1). A raise of resistance to last resort antibiotics against A. baumannii suggests that soon there will be no antibiotics with effective antibacterial action against locally circulating strains of A. baumannii.
is report is another call to support cautious prescription and use of antibiotics in accordance with local and international guidelines [59] and stress the need to develop a new antimicrobial alternative against this bacterium. e evaluation of the AdeABC efflux pump system as a mechanism of resistance used by A. baumannii strains isolated at DGMTL revealed a high prevalence of adeB (100%), adeR (99%), and adeS (99%). is is in agreement with other studies that reported the AdeABC efflux pump system as the most prevalent and with the highest detection rate in clinical isolates compared to other efflux pump genes [26]. is may be explained by the physiologic role of the efflux pumps which are proteins on the bacterial cell membrane that regulate the movement of substances from the internal to the external cell environment [60]. However, a high prevalence of the AdeABC efflux pump in bacteria clinical isolates may be explained by the involvement of the pump in driving resistance against antibiotics as hypothesized by Ranjbar and his colleagues [61]. A study conducted in China demonstrated that the AdeABC efflux pump is responsible for an increase in multidrug-resistant A. baumannii strains in their paediatric intensive care unit [42]. Similar findings were reported in Iraq where the AdeABC efflux pump system was 96% prevalent in clinical isolates of A. baumannii [61]. Mahmoudi et al. [62] also mentioned that AdeABC efflux pump system genes were present in more than 90% of A. baumannii clinical isolates in their study. e positive statistical association of structural adeB and regulatory adeR genes may be explained by the molecular structural organisation of the AdeABC operon system regulated by adeRS. e adeR gene has been identified as a recognition response factor that acts as a transcriptional activator under the influence of adeS [26]. However, exploring the interaction between adeR and adeABC, Chang et al. [63] reported that adeR action can be independent of adeS influence. eir study revealed that adeR gene may be activated as a result of an amino acid substitutions in a repeat motif region between adeR and adeABC leading to adeABC overexpression and increase tolerance to antibiotic action. e complete understanding of the AdeABC efflux pump system working and regulation mechanism is still limited. Even though additional research is needed to explore the regulation of this mechanism of resistance, this report is inclined to support the finding by Chang et al. [63], whereby, despite the presence of adeS; some AdeABC efflux pumps system in the study isolates did not demonstrate phenotypic resistance to tigecycline (Table 1; Supplementary 3). is study data revealed that 97% (100/103) of the isolates had the required set of genes (adeB, adeR, and adeS) for expression of an active efflux pump (Table 1;   ctx  ptz  caz  fep  imp  mem  cn10  cip  tg  sxt   r  89  76  85  85  83  83  81  83  0     International Journal of Microbiology Supplementary 3). Of these isolates, 75% (75/100) demonstrated an active use of the AdeABC efflux pump as a mechanism of resistance, while 25% (25/100) did not demonstrate at phenotypic level expression of the AdeABC efflux pump as a mechanism of resistance. Literature supports that it is not just the presence of the AdeABC efflux pump-associated gene but their overexpression triggered by a mutation that induce resistance to antibiotics [15,63,64]. Furthermore, it is documented that A. baumannii can use other efflux pumpassociated genes [14,65] and superfamily [22] to develop a multidrug-resistance profile. In addition, researchers reported that the AdeABC efflux pump need, in some cases, a synergistic action with other resistance mechanisms to express a level of resistance to particular antibiotic classes [67]. All these may explain the phenotypic result of 22/25 isolates of the study that did not demonstrate the use of the AdeABC efflux pump as an active mechanism of resistance, while the required set of genes was present, and their mRNA produced. Quinolone resistance-determining regions are specific regions on the gyrA and parC genes that code for the amino acid mutations that give A. baumannii its ability to resist quinolones [38,39]. ese quinolone resistancedetermining regions are found on codon 80 (serine 80) and 84 (Glu-84) of the parC gene and codon 83 (serine 83) of the gyrA gene [40]. In this study, all the clinical isolates (100%) were PCR positive for gyrA and 99% positive for parC (Table 2; Figure 3; Supplementary 4). e sequencing analysis of gyrA revealed that 89% of isolates showed a point mutation in serine 83, 39.2% had mutation on serine 80, and none of isolates exhibits mutation on Glu-84 for parC-amplified gene. Several studies reported similar findings suggesting that amino acid substitutions on gyrA and parC genes are associated with quinolone resistance in A. baumannii [67,68]. e phenotypic susceptibility testing of isolated strains of A. baumannii to ciprofloxacin (83% resistance) correlates with fluoroquinolone resistance-associated genes sequence analysis investigated in this study. However, it is known that other amino acid substitutions in gyrA induce A. baumannii resistance to quinolone. A study reported quinolone resistance mutations at serine 81 on the gyrA gene inducing change from serine to leucine [31]. Isolates showing mutations from glycine 81 to valine and alanine 84 to proline on the gyrA gene have also been reported to contribute to resistance to quinolone in A. baumannii [69,70]. Of all the study isolates, 37 had mutations on both genes, while 10 did not have any mutations (Table 2; Figure 4; Supplementary 4). Presence of mutation in the both gyrA and parC gene (b) Figure 4: (a) Phylogenetic analysis of the gyrA gene sequences showing highly related strains. e evolutionary history was inferred using the neighbour-joining method [48]. (b) Phylogenetic analysis of the parC gene sequences showing highly related strains. e evolutionary history was inferred using the neighbour-joining method [48]. contributes to a higher level of ciprofloxacin resistance rather than a single mutation in the gyrA or parC gene as reported by Lee and his colleagues [72]. e results of this current study support the report by Lee et al. [72]. e 37 isolates with mutation on both genes revealed higher level resistance to tigecycline than the 10 isolates that did not have any (Table 2; Supplementary 4). ese results suggest that single or double mutation on parC and/or gyrA and their concomitant presence or absence on both genes affect the level of resistance to antibiotics. e phylogenetic relatedness of the isolates was investigated using the neighbour-joining method [48] on MEGA 4.0 (Figures 4(a) and 4(b)). e analysis showed very minimal evolutionary distance between the isolates from DGMAH. is observation is most likely because of the nature of the genes; parC and gyrA are housekeeping genes very essential for replication and thus highly conserved.
is study has also observed that mutations within the A. baumannii isolates have gradually increased over time from 2017 to 2020, implying that isolates at DGMAH are increasingly becoming resistant to drugs used to treat them. Other studies have shown a gradual increase in resistance in healthcare institutions [57,73,74].

Conclusion
Isolates of A. baumannii at DGMAH demonstrated a high resistance prevalence to available antibiotics. e use of the AdeABC efflux pump system as an active mechanism of resistance combined with point mutation mainly in gyrA contributes to broaden the resistance spectrum of A. baumannii isolates at DGMAH. is situation is particularly alarming as the locally isolated strains demonstrated an increase in resistance to tigecycline. Judicial use of antimicrobials supported by antibiotic susceptibility results should be instituted to control the rise of resistance in A. baumannii strains.

Data Availability
e data used to support the findings of this study are included within the Supplementary Materials.

Conflicts of Interest
e authors declare that they have no conflicts of interest.