Comparison of the Acinetobacter baumannii Reference Strains ATCC 17978 and ATCC 19606 in Antimicrobial Resistance Mediated by the AdeABC Efflux Pump

ABSTRACT

The Acinetobacter baumannii RND efflux pump AdeABC is regulated by the 2-component regulator AdeRS. In this study, we compared the regulation and expression of AdeABC of the reference strains ATCC 17978 and ATCC 19606. A clearly stronger efflux activity was demonstrated for ATCC 19606. An amino acid substitution at residue 172 of adeS was identified as a potential cause for differential expression of the pump. Therefore, we recommend caution with exclusively using single reference strains for research.

INTRODUCTION

Infections with multidrug-resistant Acinetobacter baumannii in health care facilities are challenging due to limited treatment options (1). Antimicrobial resistance genes acquired by horizontal gene transfer are the main route toward multidrug resistance, and these genes are predominantly specific to single agents or classes (2). Contrary to this, intrinsic efflux mechanisms may contribute to reduced susceptibility to numerous antimicrobial classes simultaneously. The most common broad-spectrum efflux pumps in Gram-negative bacteria belong to the resistance-nodulation-division (RND) family. Besides their ability to pump out antimicrobials of different classes, they have been shown to affect susceptibility to diverse compounds used in medical practice, such as antiseptics, disinfectants, and detergents (3–6).

In A. baumannii, the RND efflux pump AdeABC, encoded on an operon and regulated by the two-component system AdeRS, has been shown to affect antimicrobial susceptibility and contribute to tigecycline resistance (7, 8). It was shown that single amino acid substitutions in AdeRS induce overexpression of AdeABC and are associated with antimicrobial resistance (9–11).

Bacteriological research into fundamental mechanisms of growth, metabolism, and antimicrobial resistance is mainly performed using a limited number of reference strains (12–14). For A. baumannii RND efflux investigations, the reference strains ATCC 17978 and ATCC 19606 are commonly used (7, 15). In this study, we compare and characterize AdeABC in terms of its regulation, expression, efflux activity, and role in antimicrobial susceptibility in these two A. baumannii reference strains.

We deleted adeRS in A. baumannii reference strains ATCC 17978 and ATCC 19606 according to the method of Stahl et al. (16). Briefly, the upstream and downstream regions of adeRS were integrated into pBIISK::sacB_kanR using the In-Fusion HD EcoDry cloning kit (TaKaRa Clontech, Saint-Germain-en-Laye, France) and primer pairs L49/L50 and L51/L52 or M66/M67 and M68/M69, respectively (Table S2 in the supplemental material). Transformed strains were selected overnight in 10 ml Luria-Bertani (LB) medium supplemented with 10% sucrose. Single colonies were obtained by plating appropriate dilutions onto LB agar with 10% sucrose. Colonies were tested for kanamycin sensitivity by replica plating and subjected to PCR to confirm the deletion of adeRS. To measure the expression of adeB, semiquantitative reverse transcription-PCR (qRT-PCR) was performed as described previously (17). Strains were grown in LB until mid-log phase. Samples were treated with RNAprotect (Qiagen, Hilden, Germany), followed by RNA isolation using the RNeasy minikit (Qiagen). cDNA was synthesized using the QuantiTect reverse transcription kit (Qiagen). Finally, RT-PCR was performed in a LightCycler (Roche, Mannheim, Germany) with QuantiFast SYBR green (Qiagen) in triplicates using freshly prepared cDNA. rpoB was used as the reference gene and was analyzed alongside adeB. Statistical analysis was performed using the unpaired t test. Strains and primers are listed in Tables S1 and S2.

We found that in ATCC 19606, adeRS deletion inhibited the expression of adeB (P < 0.03). In contrast, adeB expression in ATCC 17978 was already low in the wild-type strain, and there was no significant difference in expression levels (P > 0.5) compared to its adeRS knockout derivate (Fig. 1).

FIG 1.

Relative adeB expression levels of A. baumannii ATCC 19606 and ATCC 17978 and the derived adeRS deletion strains determined by qRT-PCR. Results are represented as mean values ± standard errors of the means. Statistical analysis was done by using the unpaired t test on the absolute values. *, P < 0.03.

These findings were confirmed by accumulation studies using the AdeABC substrate ethidium bromide (18). Mid-log-phase cells were washed in potassium phosphate buffer (50 mM potassium phosphate buffer, 1 mM MgSO₄ [pH 7.4]), adjusted to an optical density of 20 at 600 nm, and finally supplemented with 0.2% (wt/vol) glucose (Sigma-Aldrich, Steinheim, Germany). Cells were kept on ice during washing. Ethidium bromide (Merck, Darmstadt, Germany) was applied at a final concentration of 10 μM. Fluorescence (excitation at 530 nm and emission at 600 nm) was measured in a 96-well Nunclon delta surface plate (Thermo Scientific) using an Infinite M1000 pro plate reader (Tecan, Crailsheim, Germany). Measurements were performed in triplicates, and mean values were analyzed for statistical significance using the unpaired t test. Accumulation of ethidium is facilitated by reduced expression of AdeABC, resulting in enhanced DNA binding of ethidium, which is revealed by increased fluorescence (9, 18, 19). Whereas ATCC 19606 ΔadeRS revealed higher fluorescence rates, indicating significantly reduced (P < 0.0001) efflux compared to its parental strain (Fig. 2), this effect of ΔadeRS was not detected for ATCC 17978 (P < 0.05) (Fig. 3).

FIG 2.

Ethidium bromide accumulation of ATCC 19606 and ATCC 19606 ΔadeRS (A) and ATCC 17978 and ATCC 17978 ΔadeRS (B). Fluorescence was measured every 10 s over 30 min. Data were collected from three independent experiments and are presented as mean values ± standard errors of the means.

FIG 3.

Relative adeB expression levels of A. baumannii ATCC 17978 ΔadeRS transformants determined by qRT-PCR. Results are represented as mean values ± standard errors of the means. Statistical analysis was done by using the unpaired t test on the absolute values (P = 0.0699).

Furthermore, deletion of adeRS in ATCC 19606 caused increased susceptibility to different antimicrobial classes, as determined by agar dilution or broth microdilution following the current CLSI guidelines (Table 1) (20). The biggest impact of adeRS deletion in ATCC 19606 was on azithromycin resistance, with a 16-fold-reduced MIC, emphasizing the affinity of azithromycin as a substrate of AdeABC. In accordance with our analysis of adeABC expression and efflux activity, the deletion of adeRS in ATCC 17978 had no detectable effect on antimicrobial susceptibility to the antimicrobials tested. All these findings illustrate the strain-dependent role of AdeABC expression, which should be considered in future efflux studies based on findings in a single reference or laboratory strain.

TABLE 1.

MICs of antimicrobials from different classes against ATCC 19606, ATCC 17978, and derived adeRS deletion strains

Strain	MIC (mg/liter) ofa:
	AMK	GEN	AZM	ERY	CIP	LVX	CHL	MEM	RIF	MIN	TET	TGCb
ATCC 19606	8	16	32	32	1	0.5	>128	0.5	4	0.25	8	1
ATCC 19606 ΔadeRS	2	8	2	16	0.5	0.25	128	0.25	4	≤0.125	4	≤0.125
ATCC 17978 wild type	2	1	2	16	0.25	≤0.125	64	0.5	4	≤0.125	1	0.25
ATCC 17978 ΔadeRS	2	1	2	16	0.25	≤0.125	64	0.5	4	≤0.125	1	≤0.125
ATCC 17978 ΔadeRS(pJN17/04)	2	0.5	2	8	≤0.125	≤0.125	64	0.5	4	≤0.125	1	≤0.125
ATCC 17978 ΔadeRS(pJN17/04::adeRS17978)	2	1	2	8	≤0.125	≤0.125	64	0.5	4	≤0.125	1	≤0.125
ATCC 17978 ΔadeRS[pJN17/04::adeR(L172P)S17978]	2	2	4	16	0.25	≤0.125	64	0.5	4	≤0.125	2	≤0.125

^aMICs were determined by agar dilution unless otherwise noted. AMK, amikacin; AZM, azithromycin; CHL, chloramphenicol; CIP, ciprofloxacin; ERY, erythromycin; GEN, gentamicin; MEM, meropenem; MIN, minocycline; LVX, levofloxacin; RIF, rifampicin; TET, tetracycline; TGC, tigecycline.

^bTGC MICs were determined by broth microdilution.

Previous studies have shown that mutational hot spots in RND-type efflux regulatory genes identified in clinical isolates cause differential expression of the corresponding RND efflux pumps and thus contribute to antimicrobial resistance (9, 11, 21). To identify the cause of the different roles of AdeRS for ATCC 19606 and ATCC 17978, the amino acid sequences were aligned and compared (22). AdeR of ATCC 17978 revealed only one amino acid substitution, a change of L to P at position 241 (P241L), which was predicted by the PROVEAN online tool to be neutral (23). AdeS of ATCC 17978, however, revealed a P172L amino acid substitution, which was classified as deleterious. Furthermore, previous studies have shown that this region is a hot spot for resistance mutations in Acinetobacter baumannii (10, 11, 24). Residue 172 is part of the dimerization and histidine-containing phosphotransfer domain, which includes the phosphorylation residue H149 and might therefore affect the activity of the sensor kinase. Moreover, analysis of available genome sequences revealed that clinical isolates predominantly harbor the P172 configuration, which is also present in other commonly used laboratory strains like ACICU and AYE (25).

Subsequently, ATCC 17978 ΔadeRS was either recomplemented with wild-type adeRS of ATCC 17978 or adeRS subjected to Q5 site-directed mutagenesis (New England Biolabs) to obtain the nucleotide triplet coding for proline at residue 172 cloned into shuttle vector pJN17/04. Sanger sequencing (LGC Genomics GmbH Berlin, Germany) was used to confirm the corresponding nucleotide exchange. This single amino acid substitution caused an approximately 5-fold increase in the expression of adeB as determined by qRT-PCR performed in triplicates (Fig. 4).

FIG 4.

Ethidium bromide accumulation in untreated (A) and CCCP-treated (500 μm) (B) ATCC 17978 ΔadeRS transformants. Fluorescence was measured every 10 s over 30 min. Data were collected from three independent experiments and are presented as mean values ± standard errors of the means.

Analysis of ethidium accumulation revealed significantly lower accumulation levels (P < 0.0001) for ATCC 17978 ΔadeRS[pJN17/04::adeRS(L172P)¹⁷⁹⁷⁸] than for the strain recomplemented with the ATCC 17978 wild-type adeRS (Fig. 4). To show that the difference in accumulation levels was induced by altered efflux activity, ethidium accumulation was measured while the strains were subjected to the proton motive force uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP) (Sigma-Aldrich) at a final concentration of 500 μM. In this way, all proton-mediated efflux was inhibited, allowing the determination of the contribution of the amino acid substitution to efflux. The application of CCCP led to an alignment of the accumulation levels of both strains, indicating that the L172P amino acid substitution in AdeS caused increased efflux activity (Fig. 4).

Finally, the effect of the L172P amino acid substitution on antimicrobial susceptibility was determined, revealing a 2-fold increase in the MICs for gentamicin, azithromycin, erythromycin, ciprofloxacin, and tetracycline. These results confirm the increased efflux activity induced by the L172P amino acid substitution.

Overall, our results show that ATCC 17978 is comparatively weak in terms of AdeABC efflux because of its AdeS configuration, indicating that it is a rather inappropriate strain to study AdeABC efflux. Furthermore, it has to be considered that ATCC 17978 is derived from an isolate first described in 1951 (26) and therefore might not represent current A. baumannii isolates persisting in the hospital environment. This also applies for ATCC 19606, which was isolated in 1948.

Taking all these data into account, our results emphasize that focusing on single reference strains to characterize antibacterial resistance mechanisms can be problematic because of phenotypic variation within bacterial species based on genetic inconsistency.