Overexpression of bla OXA-58 Gene Driven by ISAba3 Is Associated with Imipenem Resistance in a Clinical Acinetobacter baumannii Isolate from Vietnam.

The aim of this study was to investigate genetic structures and expression of bla OXA-58 gene in five Acinetobacter baumannii clinical isolates recovered from two hospitals in southern Vietnam during 2012-2014. A. baumannii isolates were identified by automated microbiology systems and confirmed by PCR. All isolates were characterized as multidrug resistant by antimicrobial testing using the disk diffusion method. Four imipenem susceptible and one nonsusceptible isolates (MIC > 32 μg·ml-1) were identified by E-test. PCR amplification of bla OXA-58 gene upstream and downstream sequences revealed the presence of ISAba3 at both locations in one multidrug-resistant isolate. Semiquantitation of bla OXA-51 and bla OXA-58 gene expression was performed by the 2-ΔΔCt method. The bla OXA-51 gene expression of five isolates showed little difference, but the isolate bearing ISAba3-bla OXA-58-ISAba3 exhibited significantly higher bla OXA-58 mRNA level. Higher β-lactamases activity in periplasmic than cytoplasmic fraction was found in most isolates. The isolate overexpressing bla OXA-58 gene possessed very high periplasmic enzyme activity. In conclusion, the A. baumannii isolate bearing ISAba3-bla OXA-58 gene exhibited high resistance to imipenem, corresponding to an overexpression of bla OXA-58 gene and very high periplasmic β-lactamase activity.

1. Introduction

Multidrug resistant A. baumannii constitutes a serious threat for nosocomial infection control [1]. Carbapenems are currently the antibiotics of choice against multidrug-resistant Acinetobacter infections [2], but an increasing rate of resistance to carbapenems was reported worldwide, seriously limiting therapeutic options [3]. Carbapenem-resistant A. baumannii has become an alarming health care problem, mainly in developing countries [4]. As a result, carbapenem-resistant A. baumannii is classified into the critical priority group according to the urgency of need for new antibiotic treatment and the level of reported antibiotic resistance by the World Health Organization [5].

Multiple mechanisms of carbapenem resistance have been identified in A. baumannii including low membrane permeability, mutation in its chromosome genes, overexpression of efflux pumps, and acquisition of mobile resistance genes [6]. However, the production of carbapenemases is considered the principal resistance mechanism [7, 8]. The most frequent ones are carbapenem-hydrolyzing class D β-lactamases (CHDLs) and secondly metalloenzymes (MBL) such as blaNDM [9]. In addition, class A β-lactamases such as blaKPC gene has been recently also detected in A. baumannii [10], presenting a serious threat of expanding resistance spectrum in the bacteria.

Currently, six main groups of CHDLs found in A. baumannii include blaOXA-51-like, blaOXA-23-like, blaOXA-24-like, blaOXA-58-like, blaOXA-143-like, and blaOXA-235-like genes [2, 11, 12]. CHDLs exhibit weak carbapenem hydrolysis; however, they can confer resistance mediated by the combination of natural low permeability and ISAba elements located upstream of the gene possibly leading to the gene's overexpression [2]. Overexpression of blaOXA genes usually corresponds to resistance phenotypes [13-15]. Overproduction of oxacillinases, including blaOXA-58 enzyme, results from the presence of insertion sequences such as ISAba1, ISAba2, ISAba3, or IS18 which provide strong promoters for gene expression [13, 16].

In Vietnam, blaOXA-23 is the most widely disseminated class D-carbapenemase in carbapenem-resistant Acinetobacter baumannii while blaOXA-24 is not detected [17]. Even though there is not any information of blaOXA-143 and blaOXA-235 in Vietnam up to now, these genes are believed to emerge in other parts of the world [18, 19]. During 2003-2014, the majority of A. baumannii clinical isolates recovered harbored blaOXA-51 and blaOXA-23 genes. The blaOXA-58 gene was only detected in isolates recovered from 2010, after the introduction of imipenem in 2008-2009 into hospitals in Vietnam [17, 20]. The blaOXA-58-positive isolates investigated in the present study probably emerged at the same time. This recent emergence was in contrast with the striking replacement of blaOXA-58 by blaOXA-23 reported in Italy and China for the same period [21, 22]. Furthermore, isolates bearing blaOXA-58-like gene were recovered from different countries during outbreaks and showed remarkable conserved gene sequence [23-25]. The aim of this study was to investigate genetic structures and relative expression of blaOXA-58 gene, which lead to imipenem nonsusceptibility in clinical isolates recovered from two Vietnamese hospitals during 2012-2014.

2. Materials and Methods

2.1. Study Design

The study focused on A. baumannii isolates containing blaOXA-58 gene with the purpose of determining imipenem-resistance mechanism related to the gene.

2.2. Bacterial Isolates, Microbial Identification, and Antimicrobial Susceptibility Testing

Five A. baumannii isolates were chosen from a total of 252 nonduplicate Acinetobacter spp. isolates recovered from patients admitted to hospitals in southern Vietnam during 2012-2014 and were named DN and TN based on their source hospitals [17]. Microbial isolation and identification in source laboratories were performed using the Phoenix System (BD) and the API 20NE system (bioMérieux). Identification of A. baumannii isolates was confirmed by PCR amplification and sequencing of 16S-23S intergenic spacer (ITS) regions. The sequences were deposited in GenBank under accession numbers KY659325, KY659326, KY659327, KY659328, and KY659329. Antimicrobial susceptibility testing was performed by the disk diffusion method and interpreted according to the Clinical and Laboratory Standards Institute guidelines (CLSI, 2014). Tested antimicrobials included ceftazidime, cefotaxime, ceftriaxone, cefpodoxime, cefepime, piperacillin, ampicillin/sulbactam, piperacillin/tazobactam, ticarcillin/clavulanic acid, and meropenem, as well as others not belonging to β-lactams such as amikacin, gentamicin, ankamycin, netilmicin, ciprofloxacin, and levofloxacin. MIC values of imipenem were determined by the E-test (bioMérieux); the CLSI-approved breakpoints for imipenem ≥ 8 μg·ml−1 and ≤2 μg·ml−1 were considered resistant and susceptible, respectively.

2.3. Detection of blaOXA, blaNDM, and blaKPC Genes and Insertion Sequences

Amplification of blaOXA genes including blaOXA-51, blaOXA-23, blaOXA-24, and blaOXA-58 genes were performed and published in the previous study [17]. blaNDM and blaKPC genes were amplified in this study as previously reported [26]. The presence of ISAba1, ISAba2, ISAba3, ISAba4, and IS18 was detected as previously described [13, 27]. The sequence of all primers is shown in Table 1.

Table 1

Primers and probes used for PCR amplification and sequencing of antimicrobial resistance genes and related genetic elements.

Primers/probes	Sequence (5′→3′)	Length (bp)	Tm (°C)	Product (bp)	Ref.
OXA-23-F	CACTAGGAGAAGCCATGAAGC	21	55.0	114	Nguyen et al., 2017
OXA-23-R	CAGCATTACCGAAACCAATACG	22	55.0
OXA-24-F	GCTAAATGCTTTAATCGGGCTAG	24	55.0	141	Nguyen et al., 2017
OXA-24-R	ACTGGAACTGCTGACAATGC	20	55.0
OXA-51-F	GAAGTGAAGCGTGTTGGTTATG	22	55.0	148	Nguyen et al., 2017
OXA-51-R	GCCTCTTGCTGAGGAGTAAT	20	55.0
OXA-51-P	FAM-CGACTTGGGTACCGATATCTGCATTGC-BHQ1	27	61.3		This study
OXA-58-F	ATATTTAAGTGGGATGGAAAGCC	23	55.0	110	Nguyen et al., 2017
OXA-58-R	CGTGCCAATTCTTGATATACAGG	23	55.0
OXA-58-P	FAM-TTTACTTTGGGCGAAGCCATGCAAG-BHQ1	25	60.6		This study
16S-rRNA-F	CCAGTGACAAACTGGAGGAAG	21	55.5	199	This study
16S-rRNA-R	GCTGTGTAGCAACCCTTTGTA	21	55.2
16S-rRNA-P	HEX-ACGTCAAGTCATCATGGCCCTTACG-BHQ1	25	61.5
HRF/ISAba1	CACGAATGCAGAAGTTG	17	56.0	520	Segal et al., 2005
HRR/ISAba1	CGACGAATACTATGACAC	18	56.0
ISAba2A	AATCCGAGATAGAGCGGTTC	20	54.0	1200	Poirel et al., 2006
ISAba2B	TGACACATAACCTAGTGCAC	20	52.1
ISAba3A	CAATCAAATGTCCAACCTGC	20	52.3	200	Poirel et al., 2006
ISAba3C	AGCAATATCTCGTATACCGC	20	51.8
ISAba4A	ATTTGAACCCATCTATTGGC	20	50.6	612	Brown et al., 2007
ISAba4B	ACTCTCATATTTTTTCTTGG	20	45.3
IS18A	CACCCAACTTTCTCAAGATG	20	51.2	925	Poirel et al., 2006
IS18B	ACCAGCCATAACTTCACTCG	20	54.7
1512F/ITS	GTCGTAACAAGGTAGCCGTA	20	54.1	607	Chang et al., 2005
6R/ITS	GGGTTYCCCCRTTCRGAAAT	20	56.5
NDM-F	GACCGCCCAGATCCTCAA	18	55.4	52	Yong et al., 2009; CDC 2011
NDM-R	CGCGACCGGCAGGTT	15	57.0
NDM-P	HEX-TGGATCAAGCAGGAGAT-ZEN/IBFQ	17	48.3
KPC-F	GGCCGCCGTGCAATAC	16	56.0	61	Garcia et al., 2010; CDC 2011
KPC-R	GCCGCCCAACTCCTTCA	17	56.5
KPC-P	6FAM-TGATAACGCCGCCGCCAATTTGT-ZEN/IBFQ	23	62.2

2.4. PCR Mapping of blaOXA-58 and blaOXA-51 Genes

PCR mapping of blaOXA genes upstream regions was carried out using combinations of insertion sequence-specific forward primers and blaOXA-51 and blaOXA-58 gene-specific reverse primers (Table 1). The presence of ISAba3 downstream of blaOXA-58 was determined by a long-range PCR containing 1X PrimeSTAR GXL Buffer, 0.2 mmol dNTPs, 500 nmol OXA-58-F, 500 nmol ISAba3C, and 0.5 U PrimeSTAR GXL DNA polymerase (Takara). PCR products were sent to 1-BASE (https://order.base-asia.com/) for purification and sequencing. The sequences were analysed by BioEdit 7.0.9.0. (http://www.mbio.ncsu.edu/BioEdit/bioedit.html), and sequence similarity was assessed using the BLAST program (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The sequence of blaOXA-58 and its surrounding ISAba3 was deposited in GenBank under accession number KY660721.

2.5. Analysis of blaOXA-58 and blaOXA-51 Gene Expression by Real-Time RT-PCR

The midlog phase of bacterial cultures was treated with 1 μmol·ml−1 oxacillin for 24 h and was subsequently used for RNA extraction [28]. Treatment with RNAse-free DNAse I (Sigma) was performed at 37°C for 2 h. The concentration and DNase-free quality of RNA samples were spectrophotometrically assessed and confirmed by the amplification of chromosomal blaOXA-51 and 16S rRNA. Fifteen microliters of each RNA sample was reverse-transcribed in a final volume of 20 μl containing random hexamers, MMLV reverse transcriptase (Agilent) at 42o C for 45 min.

Amplification of blaOXA-51, blaOXA-58, and 16S rRNA was performed in a final volume of 25 μl containing 5 μl cDNA, 3 mmol MgCl2, 200 nmol dNTPs, 2 U h-Taq DNA polymerase (Solgent), 300 nmol of OXA-51/58-F/R primers, 150 nmol of OXA-51/58-P probes, 200 nmol of 16S-F/R primers, and 100 nmol of 16S-P probe (IDT). Primer and probe sequences are given in Table 1. Each real-time PCR was performed in triplicate on the Stratagene Mx3005P real-time PCR system (Agilent). The reaction mixture was incubated for 15 min at 95°C, followed by 40 cycles of 10 s at 95°C and 20 s at 60°C. Normalized expression of blaOXA-51 and blaOXA-58 genes was calculated relatively to the 16S rRNA reference gene according to the 2- method [29].

2.6. Multiple-Locus Variable Number Tandem Repeat Analysis

Multiple-locus variable number tandem repeat analysis (MLVA) as previously described [17, 30, 31] was used to profiling the A. baumannii isolates in the study. The method works on eight variable number tandem repeat (VNTR) loci, namely, 3468, 1988, 3002, 845, 2396, 5350, 826, and 2240 to determine relatedness among the A. baumannii isolates.

2.7. β-Lactamase Extraction and Quantitation

Isolates were grown on LB medium supplemented with 1 μmol·ml−1 oxacillin for 18-24 h at 37o C in a shaking incubator. The supernatants (extracellular fraction) were collected after centrifugation of bacterial cultures and precipitated with absolute ethanol (1 : 4) in 20 min at -20°C [32]. Periplasmic fractions were recovered from cell pellets [33]. Protein concentration was determined by the Bradford method [34].

β-Lactamase activity was determined based on nitrocefin hydrolysis [35, 36]. Briefly, 1-5 μl extracellular and periplasmic fractions obtained from each isolate were incubated with 40 nmol nitrocefin dissolved in 0.1 M phosphate buffer, pH 7.0 in a total volume of 100 μl. Samples were loaded onto microtiter plates, and the absorbance at 482 nm was measured kinetically at room temperature for 2-30 minutes using an ELISA spectrophotometer. The specific β-lactamase activity was calculated and expressed as mU·mg−1 of protein based on the quotient of β-lactamase activity (mU·ml−1) and protein concentration (mg·ml−1).

2.8. Statistical Analysis

The analysis of variance (ANOVA) was used to analyse the difference among β-lactamase activity means of isolates. A t-test was used to determine the significant difference of extracellular and periplasmic β-lactamase activity. A p value < 0.05 was considered significant.

3. Results and Discussion

bla OXAs are prevalent in A. baumannii. We had previously performed blaOXA identification in A. baumannii isolates from three hospitals in Southern Vietnam and found blaOXA-23 was dominant [17]. Even though blaOXA-58 existed with a small number in Vietnamese population, the exact genetic context involving antimicrobial resistance elements remained unknown. Here, we uncovered the imipenem-resistance mechanism of blaOXA-58-positive A. baumannii isolates. The overexpression of blaOXA-58 gene has been seen in the isolate with high-resistance phenotype through relative quantification of mRNA of the corresponding gene. The specific possible-intact ISAba3 sequence upstream of blaOXA-58 gene could be the key factor for the high expression. In addition, the high β-lactamase activity in the periplasmic space observed in the study could be the outcome of the phenomenon.

3.1. Antimicrobial Susceptibility Testing

All five isolates (DN050, TN078, DN014, TN341, and TN345) were classified as multidrug resistant (MDR) since they were nonsusceptible to at least one agent in three or more antimicrobial categories including aminoglycosides, antipseudomonal carbapenems, antipseudomonal fluoroquinolones, antipseudomonal penicillins and β-lactamase inhibitors, extended-spectrum cephalosporins, folate pathway inhibitors, penicillins and β-lactamase inhibitors, polymyxins, and tetracyclines [37]. In this study, although several antimicrobials were not tested because of their availableness at different times and hospitals, all isolates satisfied the definition to be defined as MDR. Isolate DNA050 was nonsusceptible to all antimicrobials tested. The other four were all susceptible to imipenem (there were three isolates nonsusceptible to meropenem as hospitals reported), but for other antimicrobials, their susceptibility varied. Isolate TN078 and DN014 were nonsusceptible to three categories while isolates TN341 and TN345 were nonsusceptible to five categories (Table 2).

Table 2

β-Lactamase susceptibility profiles and genotypes of five A. baumannii isolates.

Isolate	ID	DN050	TN078	DN014	TN341	TN345
Phenotype	Imipenem	+	−	−	−	−
	Meropenem	+	N/A	+	+	+
	Ceftazidime	+	+	−	+	+
	Cefotaxime	+	+	+	N/A	+
	Ceftriaxone	+	+	+	+	+
	Cefpodoxime	N/A	+	N/A	+	+
	Cefepime	+	−	+	+	+
	Piperacillin	+	N/A	−	N/A	N/A
	Ampicillin/sulbactam	+	−	−	+	+
	Piperacillin/tazobactam	+	−	−	+	+
	Ticarcillin/clavulanic acid	N/A	+	N/A	+	+
	Amikacin	+	+	N/A	N/A	N/A
	Gentamicin	+	+	N/A	+	+
	Ankamycin	N/A	N/A	+	+	−
	Netilmicin	N/A	N/A	−	+	+
	Ciprofloxacin	+	−	+	+	+
	Levofloxacin	+	−	−	N/A	N/A
Genotype	bla OXA-51	+	+	+	+	+
	bla OXA-23	−	−	−	−	−
	bla OXA-24	−	−	−	−	−
	bla OXA-58	+	+	+	+	+
	ISAba1	+	+	−	+	−
	ISAba2	+	+	+	−	−
	ISAba3	+	+	+	+	+
	ISAba4	−	−	−	−	−
	IS18	−	−	−	−	−
	ISAba3_blaOXA-58	+	−	−	−	−
	NDM	+	+	−	−	−
	KPC	−	−	−	−	−
	MLVA profile∗	6-0-7-1-17-5-0-3	6-0-7-14-17-6-15-3	9-0-7-1-7-5-14-3	9-0-5-15-15-6-0-2	9-0-5-15-15-6-0-2

N/A: not determined; −: assay negative (susceptible/absence); +: assay positive (resistant/presence). ∗MLVA profile according to the surveyed loci: 3468-1988-3002-845-2396-5350-826-2240.

3.2. Isolate Genotyping and Profiling

All isolates were identified as A. baumannii based on 16S-23S intergenic spacer (ITS) region sequencing. Based on MLVA profiling, four different MLVA types within the five isolates reflected substantial genetic diversity in the sampled Vietnamese A. baumannii isolates, as previously described [17].

No isolate with blaKPC gene was detected, while two isolates contained blaNDM gene (DN050 and TN078). Even though the two isolates were singletons (based on MLVA types from previous study [17]) with different phenotypes, they had close relatedness with just difference in 3/8 loci surveyed and very similar resistance determinants, especially the blaNDM gene. Therefore, the difference in resistance phenotype was mostly because of the distinguished genotype with ISAba3_blaOXA-58 in isolate DN050, compared to isolate TN078. It might be necessary for blaNDM gene located in a specific genetic context to be expressed as one of the important and strong resistance determinants. The mechanism should be explored further.

Regardless of the genetic diversity of the isolates, the blaOXA-58 gene sequence analysis (data not shown) of all isolates was identical with the reported blaOXA-58 gene [38]. This was in agreement with a previous work showing a lack of diversity in this gene, probably due to its recent acquisition by A. baumannii from other species [3].

All isolates were blaOXA-58- and blaOXA-51-positive and blaOXA-23- and blaOXA-24-negative (Table 2). The analysis of insertion sequences revealed the presence of ISAba1 and ISAba2, but they were not located upstream of blaOXA-51 nor blaOXA-58 genes in all isolates. ISAba4 and IS18 were not detected. ISAba3 was detected in all isolates (Table 2). However, only isolate DN050 possessed a blaOXA-58 gene bracketed by two ISAba3 elements (Figures 1 and 2). The promoter region of blaOXA-58 gene in this isolate (Figure 2) was similar to sequences described by Poirel and Nordmann [38]. The genetic structure of blaOXA-58 upstream sequences which led to overexpression of this gene displayed a remarkable variability [38-40]. Hybrid promoters constituting an ISAba3 sequence truncated by other insertion sequences were generally considered strong promoters [22, 41]. However, in this study, isolate DN050 bearing possible-intact ISAba3 sequence upstream of blaOXA-58 gene was not interrupted by inserted sequences, provided -35 and -10 promoter sequences as already described [38]. This structure probably drove high level carbapenemase production. The acquisition of insertion sequences by an imipenem-susceptible blaOXA-58 harboring isolate can lead to carbapenem resistance in A. baumannii [38]. Our results highlighted the threat of undetected reservoirs of carbapenem-resistant determinants and mechanisms in Vietnamese A. baumannii isolates.

Figure 1

Genetic structures identified in the ISAba3-blaOXA-58-positive A. baumannii isolate, DN050. ISAba3 and blaOXA-58 genes were indicated by horizontal bold arrows. Horizontal dash lines indicated sequences separating ISAba3 and blaOXA-58. Vertical arrows were for the truncated sites previously reported that did not exist in this isolate. Positions of primers were indicated as referred to Table 1 with short thin arrows. The figure is not to scale.

Figure 2

Promoter structure of blaOXA-58 gene from isolate DN050. The -35 and -10 putative promoter sequences and the +1 transcription initiation site within ISAba3 are boxed. The blaOXA-58 start and stop codons, ATG (M) and TAA (∗), respectively, are underlined. Upstream ISAba3/blaOXA-58 sequences and downstream ISAba3/blaOXA-58 gene junctions are indicated by arrows. Full sequences obtained are deposited in GenBank (accession number KY660721).

3.3. Relative Quantitation of blaOXA-58 and blaOXA-51 mRNA Level

We chose three isolates (DN050, TN341, and TN345) to study the relative expression of blaOXA-51 and blaOXA-58 under condition with oxacillin as an inducer and without oxacillin induction. They all had high β-lactamase activity in periplasmic fractions as shown in the following experiment (Table 3). The mRNA level of blaOXA-58 and blaOXA-51 genes in all isolates was determined by quantitative real-time RT-PCR. Under oxacillin induction, DN050 showed a significantly higher level of blaOXA-58 mRNA expression than isolates TN341 and TN345 (Figure 3). blaOXA-51 expression was also upregulated, but not comparable to that of blaOXA-58. Interestingly, the high expression level of blaOXA-58 from DN050 could be associated with the presence of an upstream ISAba3 sequence as previously suggested [38]. Furthermore, in this study, the possible intact ISAba3 sequence might be customized to blaOXA-58 gene to drive a very strong gene expression, as seen in ISAba1 for blaOXA-23 and AmpC genes [42]. The other isolates lacked upstream ISAba3 sequence.

Table 3

Relative quantitation of blaOXA-51 and blaOXA-58 mRNA level and β-lactamase activity in five A. baumannii isolates.

Isolate		DN050	TN078	DN014	TN341	TN345
MIC imipenem (μg·ml−1)		≥32	0.5	0.19	0.75	0.5
Relative expression of blaOXA-51	ΔCt	8.87 ± 0.39	12.87 ± 0.30	7.54 ± 0.66	9.64 ± 0.71	8.80 ± 0.60
	Expression (time)	1.00(0.76-1.31)	0.06(0.05-0.08)	2.51(1.58-3.98)	0.59(0.36-0.96)	1.04(0.69-1.59)
Relative expression of blaOXA-58	ΔCt	4.18 ± 1.18	8.64 ± 0.48	7.99 ± 2.23	8.45 ± 0.53	7.28 ± 0.24
	Expression (time)	25.69(11.3-58.37)	1.17(0.84-1.64)	1.84(0.39-8.61)	1.33(0.93-1.93)	3.01(2.55-3.56)
Total β-lactamase activity (U·mg−1)	Extracellular	10.8 ± 3.3	17.7 ± 5.2	13.9 ± 4.1	10.8 ± 3.3	12.9 ± 3.8
	Periplasmic	44.7 ± 12.8	15.3 ± 4.3	28.5 ± 8.2	49.6 ± 16.0	27.4 ± 10.0

Figure 3

Duplex real-time RT-PCR analysis of the blaOXA-51 and blaOXA-58 mRNA relative expression in three A. baumannii isolates. The error bars represent the deviation for the normalized fold expression of blaOXA-51 and blaOXA-58 in three isolates which were positive or negative for the ISAba3 upstream of the blaOXA-58 gene. −: not induced; +: induced.

3.4. Analysis of Periplasmic β-Lactamase Activity in Association with blaOXA-51/-58 Relative Expression

Under the condition of oxacillin induction, the blaOXA-58 expression of isolate DN050 (MICimipenem ≥ 32 μg·ml−1) was also significantly higher than the expression of other four isolates, TN078, DN014, TNA341, and TN345 with MICimipenem which were 0.5, 0.19, 0.75, and 0.5, respectively (Table 3). All isolates expressed a low level of blaOXA-51, confirming that the presence of blaOXA-51, without an upstream ISAba1, did not confer a resistance phenotype [16]. Furthermore, in variants harboring blaOXA-51 and blaOXA-58 genes, carbapenem resistance only correlated with blaOXA-58 [43], which is in agreement with the results of this study.

The enzyme activity of extracellular fractions was not significantly different (p = 0.2187) while one of the periplasmic fractions exhibited a significant difference among isolates (p < 0.0001). Extracellular fractions possessed lower enzyme activity than periplasmic fractions (p = 0.0355) in most cases. The periplasmic fraction recovered from all isolates exhibited variable β-lactamase activity, with very high activity corresponding to isolate DN050. Isolates TN341 displayed the highest β-lactamase activity though weakly expressed blaOXA-58 gene. This high enzyme activity probably corresponded to other β-lactamases responsible for the multidrug resistance phenotypes of the isolate, such as extended-spectrum AmpCs [44]. The presence of other β-lactamases could explain the high enzyme activity in periplasmic fractions of the other isolates. Particularly, blaNDM gene detected in both isolates DN050 and TN078, but the corresponding β-lactamase activities as well as the antimicrobial susceptibilities were different between the two isolates. The mechanism that a strain carrying a blaNDM gene is not resistant to carbapenems needs to be discovered further in A. baumannii. It might need a unique genetic structure for blaNDM gene to be expressed as seen in K. pneumoniae [45].

In a transformed A. baumannii strain with a blaOXA-58 plasmid-borne vector, this carbapenemase is selectively released via outer membrane vesicles (OMV) after periplasmic translocation through Sec-dependent system [32]. Furthermore, overexpression of blaOXA-58 gene increases its periplasmic enzyme concentration and extracellular release leading to efficient carbapenem hydrolysis [32]. The blaOXA-58 high mRNA level and high periplasmic β-lactamase activity of the DN050 isolate in this study suggested a similar overexpression, periplasmic translocation, and release mechanism of blaOXA-58 carbapenemase, even though our experimental work did not directly show the selection of OMV after being translocated to a periplasmic space. The high periplasmic β-lactamase activity of the isolates, especially TN341 in this study, also suggested a possible translocation and release of other β-lactamases with a mechanism similar to that identified with blaOXA-58. Further studies should be carried out to prove the suggested mechanism in clinical isolate similar to the transformed A. baumannii strain. To the best of our knowledge, our study is the first report on the overexpression of blaOXA-58 gene of A. baumannii clinical isolates from Vietnam.

This study had some limitations. The first limitation involved the small sample size due to the low prevalence of clinical isolates harboring blaOXA-58 gene in the population surveyed. The screening has been done in previous studies [17]. Secondly, we did not characterize other resistance mechanisms in A. baumannii such as the overexpression of efflux pump genes or existence of multicopy blaOXA-58 gene [7, 11, 46]. In addition, the presence of other β-lactamase genes such as blaIMP, blaVIM, blaGES, blaOXA-143, and blaOXA-235 was not excluded. Furthermore, we did not carry out an alternative experimental approach, such as western blotting against blaOXA-58 to unequivocally determine if the increase in β-lactamase activity is mainly due to this protein.

4. Conclusions

This study identified a mechanism of imipenem resistance related to the overexpression of blaOXA-58 gene preceded by ISAba3 and its corresponding periplasmic enzyme present at high concentration in a multidrug-resistant clinical isolate recovered from a hospital in Vietnam.