Biofilm formation in Acinetobacter baumannii was inhibited by PAβN while it had no association with antibiotic resistance

Abstract This study was conducted to investigate the relationship between Acinetobacter baumannii biofilm formation and antibiotic resistance. Furthermore, the effects of PAβN, a potential efflux pump inhibitor, on A. baumannii biofilm formation and dispersion were tested, and the gene expression levels of efflux pumps were determined to study the mechanisms. A total of 92 A. baumannii isolates from infected patients were collected and identified by multiplex PCR. The antimicrobial susceptibility of A. baumannii clinical isolates was tested by VITEK 2 COMPACT®. Genotypes were determined by ERIC‐2 PCR. Biofilm formation and dispersion were detected by crystal violet staining. The presence and mRNA expression of efflux pump genes were analyzed by conventional PCR and real‐time PCR, respectively. More than 50% of the A. baumannii strains formed biofilm and were divided into different groups according to their biofilm‐forming ability. Antibiotic resistance rates among most groups did not significantly differ. There were 7 clonal groups in 92 strains of A. baumannii and no dominant clones among the different biofilm‐forming groups. PAβN inhibited A. baumannii biofilm formation and enhanced its dispersion, whereas adeB, adeJ, and adeG and the mRNA expression of adeB, abeM, and amvA showed no differences in the different biofilm‐forming groups. In conclusion, there was no clear relationship between biofilm formation and antibiotic resistance in A. baumannii. The effects of PAβN on A. baumannii biofilm formation and dispersion were independent of the efflux pumps.

Although infections caused by A. baumannii have attracted the attention of researchers, its virulence factors, including biofilm, are not well-understood (Dahdouh, Hajjar, Suarez, & Daoud, 2016). Compared to its corresponding planktonic bacteria, biofilm-forming bacteria exhibit a modified phenotype, showing differences in gene transcription and interactions with other bacteria (Steenackers, Parijs, Foster, & Vanderleyden, 2016).
Biofilm plays an important role in persistent infections caused by pathogenic microorganisms. It has been reported that 65% of human infections are related to biofilm, which show up to 10-to 1,000-fold higher resistance compared to their corresponding planktonic cells (Cerqueira & Peleg, 2011;Sirijant, Sermswan, & Wongratanacheewin, 2016). Therefore, biofilm-related infections are more difficult to clear and more prone to relapse (Koo, Allan, Howlin, Stoodley, & Hall-Stoodley, 2017). Notably, some studies have demonstrated that biofilm formation is increased in response to subinhibitory concentrations of antibiotics, which is typically a direct consequence of low-dose therapy, indicating that biofilm regulation is involved in the global response to external stresses, such as antibiotics (Kaplan, 2011;Zaborskyte, Andersen, Kragh, & Ciofu, 2017). However, the correlation between the biofilm formation ability of A. baumannii and antibiotic resistance remains unclear. Over the past two decades, numerous studies have shown conflicting results (Perez, 2015;Sanchez et al., 2013;Wang et al., 2018), and thus further, studies are needed to resolve these issues.
The mechanisms of antibiotic resistance in A. baumannii are very complex, mainly involving the expression of active efflux pumps; changes in antibiotics targets; production of inactive enzymes; and reduction, deficiency, and mutation of outer membrane proteins (Lee et al., 2017). Overexpression of efflux pumps can decrease the accumulation of drugs and may be among the main mechanisms inducing A. baumannii multidrug resistance (Lin, Lin, Tu, & Lan, 2017;Yoon et al., 2015). To date, three Acinetobacter drug efflux (Ade) resistance-nodulation-cell division (RND) systems, AdeABC (Magnet, Courvalin, & Lambert, 2001), AdeFGH (Coyne, Rosenfeld, Lambert, Courvalin, & Périchon, 2010), and AdeIJK (Damier-Piolle et al., 2008), have been characterized in A. baumannii. In a study of 53 tigecycline-susceptible A. baumannii isolates, the relative expression levels of adeB and adeJ were significantly increased in 52 clinical tigecycline-non-susceptible isolates (Li et al., 2015). Another study showed that highly multidrug-resistant A. baumannii were positive for adeABC and adeIJK (Lin, Ling, & Li, 2009). AbeM, a type of efflux pump, was reported to play an important role in the resistance to imipenem (Hou, Chen, Yan, Wang, & Ying, 2012). AmvA is also efflux pump and belongs to the major facilitator superfamily and mediates the resistance of a clinical multidrug-resistant A. baumannii isolate. The expression of AmvA was found to be higher in clinical isolates that exhibited very high MICs of carbapenems, cephalosporins, aminoglycosides, and fluoroquinolones (Rajamohan, Srinivasan, & Gebreyes, 2010). Most studies of efflux pumps have focused on their roles in increasing antibiotic resistance. The relationship between these pumps and biofilm formation requires further analysis.
Phenylalanine-arginine beta-naphthylamide (PAβN) is a kind of efflux pump inhibitor. PAβN has been proposed to be an RND competitive substrate (Kourtesi et al., 2013). RND-type multidrug efflux pumps are a critical contributor to antibiotic resistance in A. baumannii (Leus et al., 2018). The effects of the PAβN on biofilm formation and dispersion of A. baumannii remain unknown.
In this study, the associations among the biofilm formation in 92 nonduplicated A. baumannii strains isolated from infected patients and antibiotic resistance were investigated. Additionally, the effects and possible mechanisms of PAβN on A. baumannii biofilm formation and dispersion were examined.

| Bacterial isolates and data collection
The clinical isolates used in this study were collected from June 2014 to January 2015 as different kinds of specimen sample types, including sputum, urine, abscess secretion, blood, hydrothorax, and as-

| DNA extraction by boiling and freezethawing method
Acinetobacter baumannii genomic DNA was extracted without using chemical reagents and DNA purification. The frozen strain was transferred to a blood agar plate and incubated aerobically at 37°C overnight after which it was transferred to another blood agar plate to obtain the third generation. The third generation was used to extract genomic DNA. Approximately 10 single colonies of bacteria were selected and transferred into a microtube containing 400 μl ddH 2 O. The microtube was incubated in boiling water for 10 min, immediately cooled on ice, and frozen at −20°C for 20 min. The samples were then thawed at room temperature (approximately 25°C) and homogenized by vortex mixing for 10 s. Finally, the microtubes were centrifuged at 13,362 g for 15 min at 4°C. The upper aqueous layer containing the DNA was carefully transferred to a sterile microtube. The DNA samples were stored at −80°C until use.

| Identification of A. baumannii by multiplex polymerase chain reaction
We collected the strains first identified as A. calcoaceticus-A. baumannii complex with the VITEK 2 COMPACT ® automated microbiology system. One-tube multiplex polymerase chain reaction (PCR) was performed as described previously to isolate A. baumannii (Chen et al., 2007). A pair of primers was used to amplify an internal 208base pair (bp) fragment from the internal transcribed spacer (ITS) region of A. baumannii and the second pair of primers to amplify a highly conserved 425-bp region of the recA gene of Acinetobacter spp. PCR was performed in a total reaction volume of 25 μl containing 12.5 μl 2 × Taq Master Mix, 8.5 μl ddH 2 O, 1 μl each primer for the ITS (10 μM), and 0.5 μl recA (10 μM), with 1 μl DNA template.
The PCR conditions were as follows: 94°C for 3 min, followed by 30 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C 30 s, with a final extension at 72°C for 3 min. Amplicons were analyzed by 1.5% agarose gel electrophoresis in 0.5× Tris/boric acid/EDTA buffer run at 100 V for 35 min.

| Quantification of the biofilm-forming ability of A. baumannii
Overnight cultures were used to inoculate in 5 ml of LB broth (Qi et al., 2016;Rodríguez-Baño et al., 2008), which was grown again overnight. The cultures were centrifuged, and the supernatant was

| Homology analysis of A. baumannii by enterobacterial repetitive intergenic consensus sequence type 2-based PCR
The homology of the isolates was evaluated by enterobacterial repetitive intergenic consensus sequence type 2-based PCR (ERIC-2 PCR), which reveals strain-specific banding patterns obtained by amplifying multiple anonymous regions of the genome with the primer ERIC-2, as previously described by Versalovic, Koeuth, and Lupski (1991). Template DNA was extracted as described in section 2.2. The PCR system contained 12.5 μl 2× Taq Master Mix, 10.5 μl ddH 2 O, 1 μl ERIC-2 primer (10 μM), and 2 μl DNA template. The reaction mixtures were amplified as follows: initial denaturation at 94°C for 3 min, followed by 30 cycles of DNA denaturation at 94°C for 1 min, annealing at 54°C for 1 min, primer extension at 72°C for 1 min, and a final extension step at 72°C for 10 min. The amplification products were electrophoresed in 1.8% agarose gels, stained with Gelstain, and photographed under UV light. The results were analyzed by visual inspection. Isolates were considered as a single clonal group when they exhibited at least one high-intensity band difference according to visual inspection.

| Presence of efflux pump genes adeB, adeJ, and adeG analyzed by conventional PCR
Fifteen isolates were chosen from each group to detect the positive rates of the efflux pump genes adeB, adeJ, and adeG. DNA templates were extracted as described in section 2.2. PCR amplification was performed in a 25 μl reaction mixture, containing 12.5 μl 2× Taq Master Mix; 10.5 μl ddH 2 O; 0.5 μl each primer for adeB, adeJ, and adeG (Table 2); and 1 μl DNA template. The primers used for conventional PCR analysis are listed in Table 1. The PCR conditions were as follows: 94°C for 3 min, followed by 30 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s, with a final extension at 72°C for 3 min.
Amplicons were analyzed by electrophoresis in a 1.5% agarose gel in 0.5× Tris/boric acid/EDTA buffer at 100 V for 35 min.

| RNA extraction and cDNA synthesis
According to the results of homology analysis, 1-2 isolates were chosen from one clonal group in each biofilm formation ability group for RNA extraction and cDNA synthesis. A single colony-forming unit of A. baumannii was selected from the overnight cultures and inoculated into 5 ml LB and cultured at 37°C and 180 rpm overnight. Next, 1 ml of this culture was centrifuged. RNA extraction was performed using the E.Z.N.A Total RNA Kit (Omega Biotek) following the manufacturer's instructions. Total RNA (1 μg) was used to synthesize cDNA using the EasyScript First-Strand cDNA Synthesis SuperMix Kit (TransGen Biotech) following the manufacturer's instructions.
The synthesized cDNA was stored at −20°C.
The primers used for qRT-PCR analysis are listed in Table 1. Gene expression was determined by qPCR using the TransStart SYBR Green qPCR SuperMix UDG Kit (TransGen Biotech). A no-template control was included in the analysis. RpoB was used as the reference gene. Relative expression was determined using the 2 −ΔΔCt method with the Eppendorf Real-Time System. Reaction samples were prepared according to the manufacturer's instructions with a 10-fold dilution of cDNA. All reactions were carried out in triplicate with at least two biological replicates. Target gene expression was measured as the relative expression compared to that of rpoB.

| Assessment of the effects of PAβN on A. baumannii biofilm formation and dispersion
To determine the relationship between biofilm formation ability and efflux pump expression, the effects of PAβN, a potential efflux pump inhibitor, on A. baumannii biofilm formation and dispersion were examined. The method was adapted from a previously described study (Chen et al., 2016). PAβN at final concentrations of 0, 20, 40, 60, 80, or 100 μg/ml was cocultured with 10 strains of A. baumannii with biofilm formation ability in a 96-well cell culture microtiter plate for 24 hr. The cells were stained with crystal violet, and the OD 570 was measured. To evaluate the dispersion effect, the strains were cultured to primarily form biofilm in a 96-well cell culture microtiter plate for 24 hr. The plate was washed with phosphate-buffered saline, and 200 μl of media containing 0, 60, 80, or 100 μg/ml PAβN in LB was added to each well and incubated for an additional 24 hr.
The cells were washed as described above, and the biomass was determined by measuring the OD 570 with 1% (w/v) crystal violet solubilized in ethanol. Each assay was performed in triplicate.

| Statistical analysis
The quantitative data are presented as the means ± standard deviations. Different groups were compared by one-way analysis of variance followed by Tukey-Kramer multiple comparison test. pvalues < .05 were considered as statistically significant. For the antibiotic susceptibility ratios and conventional PCR results, statistical analysis was performed using the chi-square test. Statistical analysis was conducted using SPSS 21.0 software (SPSS, Inc.).

| Acinetobacter baumannii identification by onetube multiplex PCR
A total of 121 isolates of A. calcoaceticus-A. baumannii (Ac-Ab) complex, which were identified phenotypically by the VITEK 2  Figure 1). The bacteria belonging to other genera produced no amplicons, including P. aeruginosa. We identified 92 isolates belonging to A. baumannii among the 121 isolates.

| Antibiotic susceptibility analysis among different biofilm-forming groups
Acinetobacter baumannii in each group showed high resistance rates (>70%) to the most widely used antibiotics in the clinic. Most of the studied antibiotics showed no significant differences in resistance rates among the different groups. Only tobramycin and gentamicin showed significant differences among the different groups, with the lowest resistance rate observed in the biofilm-forming positive group (Table 3).

| Homology analysis of A. baumannii by ERIC-2 PCR
The clonal relationships of 92 A. baumannii isolates were obtained by ERIC-2 PCR and found to have 7 different genotypes. Figure

| Presence and expressions of efflux genes in different biofilm-forming groups
The distribution of efflux genes adeB, abeJ, and adeG is shown in Table 4. These efflux genes were expressed in most clinic isolates.
However, there were no significant differences among the different biofilm-forming groups. To verify these results, we chose adeB F I G U R E 1 Examples of multiplex PCR products resolved by agarose gel electrophoresis. Note: PCR was performed using the specific Acinetobacter baumannii primers for the ITS gene (product: 208bp) and the internal control primers specific for the recA gene (product: 425bp) of all Acinetobacter spp. were selected from each genotype in each group to be tested. The results agreed with those of conventional PCR. The expression levels of adeB, abeM, and amvA did not differ among the four studied groups (Figure 3).

| Effects of PAβN on A. baumannii biofilm formation and dispersion
The

| D ISCUSS I ON
Acinetobacter baumannii is widely found on environmental surfaces and is likely important in disease transmission within healthcare settings (Chen et al., 2017). In recent years, A. baumannii has attracted attention because of its increased rate of causing serious TA B L E 3 The antibiotic resistance ratio in different biofilm formation ability groups  Smidt, Tjernberg, & Ursing, 1991). However, in the Ac-Ab complex, different genomic species show different antibiotic susceptibilities and epidemic potential. It is very important to rapidly and accurately distinguish A. baumannii from other species in the complex. In the present study, using the multi-PCR method, A. baumannii was identified quickly and accurately. Our results showed that in the clinic, the isolates of the Ac-Ab complex mostly contained A. baumannii (76.03%). Additionally, the antibiotic resistance rate was significantly higher than that of other species in the complex (data not shown).
In this study, among the detected 92 strains of A. baumannii, 50 strains formed biofilm. In these 50 strains, the biofilm formation abilities were weakly positive, positive, and strongly positive, with approximately 1/3 of the strains in each group. It must be pointed out that although the negative controls were uninoculated LB, its readings were a little high in this study. We hypothesize this was due to unavoidable dye residue rather than contamination, as the procedures for negative control wells were the same as samples, including incubation, washing, staining with crystal violet, and solubilization with ethanol. After all these procedures, the negative controls could not be colorless, which resulted in little high OD readings. The biofilm-forming abilities of A. baumannii strains were determined by the OD value differences between the sample wells and the negative control wells. Interestingly, we observed a relationship between the A. baumannii biofilm formation ability and its antibiotic resistance. For most antibiotics, the antibiotic resistance rates were lowest in the positive biofilm-forming group. However, overall, there was no clear association between biofilm-forming ability and antibiotic resistance. The association of biofilm formation ability and antibiotic resistance is controversial. Perez reported that in 116 strains of A. baumannii, the biofilm formation abilities of meropenem-resistant strains were significantly lower than those of meropenem susceptibility strains (p < .001) (Perez, 2015).
Cusumano also reported that multidrug-resistant Klebsiella pneumoniae isolates were more commonly weak than strong in biofilm formation ability (Cusumano et al., 2019). In contrast, another study showed that A. baumannii biofilm formation-positive strains more easily developed a multidrug-resistant phenotype. Compared to strains with weak or non-biofilm formation abilities, A. baumannii clinic isolates with strong biofilm formation abilities more easily developed resistance to aminoglycosides, carbapenems, tetracyclines, and sulfonamides (Sanchez et al., 2013). In previous studies, the biofilm formation abilities of the chosen strains were not subdivided, which may have led to inconsistent results.
Because biofilm formation and the efflux pump system are both related to A. baumannii resistance and survival in the hospital environment, we examined the relationship between biofilm and the efflux pump system. Studies of biofilm and the efflux pump system are limited and have shown controversial results. One study showed that as the A. baumannii biofilm matured, adeB and adeG expression was increased (He et al., 2015). In contrast, another study reported that in a total sequenced A. baumannii-sensitive strain BM4587, overexpressions of AdeABC and AdeIJK significantly reduced biofilm formation (Yoon et al., 2015). Our results revealed no significant differences in these efflux pump genes among strains with different biofilm formation abilities. In general, biofilm is considered as a self-protective mechanism, as the biofilm forms a barrier to resist various stresses. According to our results for biofilm formation and antibiotic susceptibility, these differences may be because the resistance rate and gene expression were tested in planktonic cultures in vitro but not in biofilm cultures. In the clinic, bacteria may show susceptibility to drugs in vitro but the curative effects are poor in vivo, which may be related to the bacteria's biofilm formation ability in vivo.
Although no efflux pump genes showed a direct relationship with the biofilm formation ability, we investigated the effect of PAβN, a universal efflux inhibitor, on biofilm formation and dispersion. As the ability of A. baumannii isolates to acquire drug resistance by the efflux pump mechanism is a concern, most studies of PAβN  (Fazli et al., 2014;Rasamiravaka, Labtani, Duez, & El Jaziri, 2015), and pelA is an important biofilm constituent (Marmont et al., 2017).
Whether the effects of PAβN on A. baumannii biofilm are related to the QS system requires further analysis.
The limitations of this study were its small sample size and retrospective nature. Thus, prospective studies of larger sample sizes are needed to confirm our conclusions. Another limitation is the methodology for biofilm quantification. Until now, there is still no universally recognized standard for biofilm quantification. One report used the reference strain (ATCC 19606) to categorize weak or strong biofilm formers (Qi et al., 2016). Other reports used OD 570 = 1 to categorize weak or strong/moderate biofilm formation (Beganovic, Luther, Daffinee, & LaPlante, 2019). Another report used the same standard as us (Amin et al., 2019;Zhang et al., 2016). The different standards make it very difficult to compare the studies.
In summary, we showed that A. baumannii had a strong biofilm formation ability. Biofilm formation by A. baumannii was not associated with antibiotic resistance and was inhibited by PAβN. The mechanisms of the effects of PAβN on A. baumannii biofilm formation and dispersion may be independent of the efflux pumps.

CO N FLI C T O F I NTE R E S T
None declared.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analyzed during this study are included in this published article.