Background

Acinetobacter species are a heterogeneous group of strictly aerobic, non-motile gram-negative, non-fermenting encapsulated coccobacilli that can be found in the environment [1, 2]. Among them, Acinetobacter baumannii has become a major cause of nosocomial infections in recent years. A. baumannii is a conditional pathogen, especially when the resistance of inpatients declines; A. baumannii can invade the human body in various ways and cause fatal infection [3, 4]. Meanwhile, infection caused by Acinetobacter species other than A. baumannii has caused concern recently [5]. Moreover, the developed resistance of Acinetobacter to a variety of commonly used antimicrobials has led to ineffective antibiotics and increased mortality following infection [6]. Carbapenems are commonly used to treat infections caused by Acinetobacter species; an outbreak of carbapenem-resistant Acinetobacter species can complicate therapy and may lead to treatment failure [7, 8].

There are three main mechanisms that Acinetobacter species use to resist carbapenems: production of enzymes (e.g., carbapenemase) that are capable of hydrolyzing carbapenems; altered function of membrane-associated proteins such as porins; and activation of drug efflux pumps [9]. Carbapenemases that have been reported in Acinetobacter species include Ambler class B metallo-β-lactamases, such as IMP and VIM, and the recently described NDM; Ambler class D oxacillinases (OXAs); and Ambler class A β-lactamases [10, 11]. The OXA carbapenemase genes of Acinetobacter species are divided into some phylogenetic subgroups: blaOXA-23-like, blaOXA-24/40-like, blaOXA-51-like, blaOXA-58-like, blaOXA-143-like, and blaOXA-235-like. Although OXA carbapenemase weakly hydrolyze carbapenems, they can confer higher resistance when blaOXA genes are overexpressed using a strong promoter with mobile insertion elements such as ISAba1 [12, 13]. Low outer membrane permeability, due to changes in the primary structure or loss of carbapenem-associated outer membrane protein (carO), a 25/29-kDa outer-membrane protein, is one of the best-characterized mechanisms of intrinsic carbapenem resistance in Acinetobacter species [14, 15]. Drug efflux pumps are most often associated with carbapenem resistance in Acinetobacter species when overexpressed; the resistance-nodulation-division (RND)-type family is of particular clinical importance [16]. To date, five RND efflux pumps have been described in Acinetobacter species: AdeABC, AdeDE, AdeIJK, AdeXYZ, and AdeFGH [17].

Colistin has been used in veterinary and human medicine for over 50 years. Colistin has broad-spectrum activities against gram-negative bacteria, but is associated with both nephrotoxicity and neurotoxicity, limiting its clinical application. However, colistin has now emerged as an effective therapeutic against carbapenem-resistant bacteria. Now, over-reliance on colistin to treat these diseases has led to the development of resistance to colistin [18]; thus, exploring the resistance mechanism of Acinetobacter species will be important.

Mcr, a plasmid-mediated colistin resistance gene first reported in 2015 in China [19], has increased our knowledge of colistin-resistance mechanisms other than chromosomal mechanisms. The plasmid-borne colistin-resistance gene mcr can also be transmitted and transferred by mobile genetic components [20].

Mobile genetic elements, such as integrons, transposons, and insertion sequences, provide bacteria with a powerful genetic toolbox, which could lead to multidrug-resistant (MDR) strains. Transfer of mobile genetic elements between different isolates, even in different species, has also been described in Acinetobacter [21, 22].

Various bacteria, including MDR bacteria, are contained in the gut microbiota, which serves as an important and large reservoir of resistance genes. Fecal samples might therefore be the ideal specimen for studying and detecting antimicrobial-resistant genes. Several studies have reported high rates of fecal carriage for Acinetobacter, making the digestive tract a potential reservoir for nosocomial infections and outbreaks [23]. The aim of this study was to analyze the mechanism of resistance and perform genotyping of carbapenem-resistant Acinetobacter from clinical infection and fecal survey samples in Southern China.

Methods

Bacterial isolates and species identification

Between January 2014 and December 2015, 170 non-duplicate carbapenem-resistant Acinetobacter isolates were obtained from clinical infection samples (including sputum, puncture fluid, urine, instrument, blood) which were collected from Nanfang Hospital, a large general and teaching hospital in Guangzhou, China.

Moreover, 5000 fecal survey samples were collected from July 2014 to June 2015, followed by incubation on MacConkey agar plates with 2 μg/mL meropenem. A total of 434 g-negative carbapenem-resistant isolates were collected. The recA [24] gene was amplified from non-fermenting isolates and used as a control for the genus Acinetobacter. Finally, 74 carbapenem-resistant Acinetobacter species were obtained. The study was approved by the Medical Ethics Committee of Nanfang Hospital Southern Medical University and conducted in compliance with the Declaration of Helsinki.

All isolates were confirmed as Acinetobacter species by biochemical methods, the recA gene, and analysis of 16S ribosomal RNA (rRNA) sequences. The Acinetobacter calcoaceticus–A. baumannii complex was confirmed by 16–23S rRNA and internal transcribed spacer sequences [25].

Antimicrobial susceptibility testing

Analysis of susceptibility to antimicrobial compounds was performed using the automated BD Phoenix 100 Microbiology System (Becton Dickinson and Co., Franklin Lakes, NJ, USA) according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI) M100-S27 criteria. The antibiotics tested were ampicillin, ampicillin-sulbactam, amoxicillin-clavulanic acid, piperacillin, piperacillin-tazobactam (4 mg/L tazobactam), ceftazidime, cefotaxime, cefazolin, cefepime, imipenem, meropenem, aztreonam, chloramphenicol, levofloxacin, ciprofloxacin, gentamicin, tetracycline, colistin, trimethoprim-sulfamethoxazole, and amikacin. Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were used as control strains for antimicrobial susceptibility testing.

Molecular detection of resistance genes

DNA was extracted using the AccuPrep® genomic DNA extraction kit according to the manufacturer’s instructions. PCR was performed for antibiotic resistance-related genes described below.

The following carbapenemase-encoding genes were detected: class A β-lactamases genes: blaKPC [26]; class B metallo-β-lactamases genes: blaIMP [26], blaVIM [26], blaSPM [26], and blaNDM [26]; and class D oxacillinases genes: blaOXA-23-like [27], blaOXA-24/40-like [27], blaOXA-51-like [27], and blaOXA-58-like [27]. The colistin-related resistance gene: mcr-1 [28], mcr-2 [28], mcr-3 [28], mcr-4 [28], and mcr-5 [28]. The protein-related gene carO [29] (A. baumannii outer membrane protein involved in carbapenem resistance) and the drug efflux pump component-related genes adeA [30], adeB [31], adeC [31], adeI [31], adeJ [31], and adeK [31] were also detected.

Database searches for DNA sequence similarities were performed using BLAST (https://blast.ncbi.nlm.nih.gov). The construction of a carO phylogenetic tree was obtained by MEGA using the neighbor-joining method. A model of the p-distance was used to estimate distances for nucleotide sequences. The significance of the groups observed in constructed trees was determined by bootstrap analysis with 1000 replicates.

Mobile genetic element-related resistance genes, including integrons (intI1 [32, 33], intI2 [32, 34], intI3 [32]), transposons (tnpU [35] and tnp513 [35]), the variable region of class 1 integron and class 2 integron, and insertion sequences (IS26 [36], ISAba1 [37], and ISAba125 [37]) were also analyzed. The PCR cycle consisted of denaturation at 94 °C for 5 min, followed by 35 cycles of 30 s at 94 °C, annealing for 40 s at optimal annealing temperature, and extension at 72 °C for 40 s for the genes and 4 min for the variable region of intI1 and intI2.

The primers used for PCR are shown in Table 1. Primer synthesis and sequencing of PCR products were performed by Sangon Biotech Co., Ltd. (Shanghai, China).

Table 1 Gene primers used for PCR amplification of genes associated with antimicrobial resistance
Full size table

Eric-PCR

Enterobacterial Repetitive Intergenic Consensus (ERIC-PCR) was performed on all isolates depending on the different species [38]. Band comparisons were performed by clustering analysis based on the unweighted pair-group method with arithmetic mean of isolates using Quantity One-v4.6.7.

Results

Identification of isolates

During the study period, 244 (170 from clinical infection samples and 74 from fecal survey samples) carbapenem-resistant Acinetobacter species were collected. PCR amplification of the recA gene of all strains was positive. Further identification revealed 181 A. baumannii and 63 non-baumannii Acinetobacter species. Remarkably, clinical infection samples and fecal survey samples exhibited large differences in species (Table 2). Acinetobacter baumannii was the predominant clinical carbapenem-resistant bacteria (97.1%). However, 58 carbapenem-resistant non-baumannii Acinetobacter were isolated in fecal survey samples, and 77.59% were further identified as A. junni.

Table 2 Numbers of isolated Acinetobacter strains
Full size table

Antimicrobial susceptibility testing

According to the minimal inhibitory concentration (MIC) values of tested antibiotics (Table 3 and Additional file 1: Table S1), the MIC breakpoints of A. baumannii complex to colistin refer to the CLSI standard, and the MIC breakpoints of non-baumannii Acinetobacter to colistin refer to the EUCAST standard, because the CLSI standard was applicable to A. baumannii complex only. All isolates were resistant to carbapenems and cephems, including cefazolin, ceftazidime, and cefepime, and sensitivity to amikacin and colistin was observed in 32.80 and 97.10% of the isolates, respectively. However, non-baumannii Acinetobacter was more sensitive to amikacin than Acinetobacter baumannii. Six non-baumannii Acinetobacter and one Acinetobacter baumannii were resistant to colistin. Notably, all of the carbapenem-resistant Acinetobacter isolates were MDR (Table 3) [39].

Table 3 Susceptibilities of Acinetobacter species and the distribution of resistant genes among these species
Full size table

Molecular detection of resistance genes

Carbapenemase-encoding genes

The prevalence of blaOXA-23-like, blaOXA-24/40-like, blaOXA-51-like, and blaOXA-58-like gene occurrences were 71.31, 2.45, 75.00, and 12.30%, respectively. Co-existence of the blaOXA-23-like and blaOXA-51-like genes was detected in 171 strains. It is notable that three strains simultaneously carried the genes blaOXA-23-like, blaOXA-51-like, and blaOXA-58-like. With regard to the 181 isolates of A. baumannii, 179 (98.90%) strains carried the blaOXA-51-like gene, which is characteristic of A. baumannii. For other carbapenemase genes in all strains, 2 (0.82%) were positive for the blaIMP gene, 4 (1.64%) for the blaVIM gene, and 57 (23.36%) for the blaNDM gene. It is remarkable that all blaNDM-positive strains were isolated from fecal survey samples, while 41 strains possessed blaNDM-1, 12 strains possessed blaNDM-5, and 4 strains possessed blaNDM-4. The genes blaKPC and blaSPM were not identified.

Colistin-related resistant genes

Six (8.70%) non-baumannii Acinetobacter and one (0.55%) A. baumannii isolate was resistant to colistin, while five non-baumannii Acinetobacter strains from fecal samples tested positive for the mcr-1 gene.

Protein-related genes

The carO gene was detected in 164 (67.21%) strains, and an analysis of 164 complete carO porin nucleotide sequences was performed. All strains showed amplified products of approximately 1200 bp, with the exception of A. baumannii 100588A and 513,007, in which a band of approximately 2000 bp was amplified and the distal part of the insertion (from 1066 to 2259 of the sequence) showed 100% identity with a fragment from A. baumannii BJAB0715 (GenBank accession number CP003847.1, position 2960028 to 2961220). Acinetobacter sp. 1029087 (GenBank accession number KX517470.1) contained an insertion of 2 bp (starting at +230 of the carO gene), which generated a stop codon (TAA) at position 247–249 in the nucleotide sequence of the carO gene.

Sequencing of carO genes revealed the presence of 11 distinct variants; 10 novel sequences were discovered, which were designated as different if they had at least one nucleotide change. The dominant sequence was 99–100% identical to A. baumannii strain 3027STDY5784958 (GenBank accession number LT594095.1) and the ten novel sequences exhibited 92–99% identity with sequences in the database. The novel sequences were a result of deletions, insertions, or point mutations. Phylogenetic analysis of the detected carO gene showed that all sequences were classified into two groups (Fig. 1).

Fig. 1
figure 1

Phylogenetic inferences of the carO gene among Acinetobacter species. A phylogenetic tree of the carO gene was constructed by MEGA software using the neighbor-joining method. The numbers in parentheses correspond to the number of isolates within the genotype. The black block represents the variants detected in this study. The nucleotide sequences of the structure A to K in this work have been submitted to the GenBank database and assigned the following accession numbers: KX517486, LT594095, KX517462, KX517464, KX517472, KX517465, KX517458, KX517469, KX517470, KX517467, and KX517489

Full size image

Drug efflux pump component-related genes

Efflux system genes were present in approximately 70% of isolates, and adeABC and adeIJK were observed in 76.23 and 72.13%, respectively.

Mobile genetic element-related resistance genes

Multidrug resistance to aminoglycosides, quinolones, tetracycline, and trimethoprim-sulfamethoxazole antimicrobial agents might be a result of the presence of a high rate of mobile genetic elements, including IS26 (97.54%), ISAba1 (90.16%), ISAba125 (93.44%), tnpU (63.52%), and tnp513 (68.03%). Class 1 and 2 integrons were detected in 180 (73.77%) and 5 (2.05%) strains, respectively, while class 3 integrons were absent. The amplicons ranged in size from 1.5 to 2.6 kb. A single band was obtained in 177 isolates and two bands were obtained in three isolates. Sequencing results revealed that the variant region of class 1 integron has four distinct gene cassettes (aadA1-catB8-aacA4, aacC1-OrfA-OrfB-aadA1, dfrA17-aadA5, aadA2-orfF-dfrA12), and the variant region of the class 2 integron has one gene cassette (dfrA1-sat2-aadA1-orfX) (Table 4).

Table 4 Characterization of integrons in Acinetobacter species
Full size table

Eric-PCR

Genotyping by ERIC-PCR demonstrated a high genetic diversity of non-baumannii Acinetobacter in different samples, and one was selected as a representative of the same band for dendrogram cluster analysis (Fig. 2). However, A. baumannii detected in clinical samples had greater than 90% similarity, suggesting they may have originated from the same clone.

Fig. 2
figure 2

Dendrogram from ERIC-PCR analysis of 63 Acinetobacter isolates. In total, 62 distinct genotypes were identified based on 80% similarity. The scale bar indicates the similarity values. ERIC, enterobacterial repetitive intergenic consensus

Full size image

Discussion

The results of the carbapenem resistance gene sequencing showed that the blaNDM gene was detected in 57 isolates and only present in the fecal survey samples. Fu et al. [40, 41] described that among Acinetobacter species, the blaNDM gene is widely disseminated in China. It is possible that the spread of New Delhi metallo-beta-lactamase-1(NDM-1) carbapenemases may occur rapidly through Acinetobacter spp. because it may become markedly more difficult to eradicate [42] and has begun to steadily express NDM [43].

We showed that 98.90% of the A. baumannii isolates contained the blaOXA-51-like gene. The results of Lowings et al. [44] and Correa et al. [45] indicated that 99 and 97.5% of the A. baumannii isolates contained this gene. Zhao et al. [46] showed 91.7% of the A. baumannii isolates contained this gene. Zander et al. [47] reported the identification of three A. baumannii isolates that did not harbor the predicted blaOXA-51-like gene. Further study of the genomic environment of these isolates revealed changes in the blaOXA-51-like genetic structure, resulting in variants blaOXA-66 or blaOXA-78, which were interrupted by the insertion sequence. An investigation conducted in 27 hospitals located in 14 European and Mediterranean countries during 2009–2011 showed that 277(67.4%) were positive for blaOXA-23-like [48]. Correa et al. showed that 97.5% were positive for blaOXA-23-like in Colombia (2008–2010) [45]. Moreover, 107(91.5%) isolates were positive for blaOXA-23-like in the countries of the Gulf Cooperation Council [49], Carvalho et al. reported that 96 of 110 (87%) carried the blaOXA-23-like gene in Rio de Janeiro [10]; however, in northern Croatia and Istria during 2009–2010, only 3.24% (6/185) of carbapenem-resistant A. baumannii isolates possessed the blaOXA-23-like gene [50], while 25.7 and 28.3% were positive for the blaOXA-24/40-like and blaOXA-58-like genes, respectively, 152 (100%) isolates were negative for blaOXA-23-like in Mexico [51], these data indicated that the prevalence or predominance of some OXA subgroups depends on geographic/regional variations.

In the present study, six (8.70%) non-baumannii Acinetobacter and one (0.55%) A. baumannii isolates were resistant to colistin. In 74 fecal survey samples, we detected that five colistin-resistant A. junii were positive for the mcr-1 gene, and the mcr gene in Acinetobacter isolated from human samples had not been reported in the literature thus far. Moreover, all three A. junii (C1121030, C1121090, C1121152) contained blaOXA-58-like and blaNDM, one of the three strains contained blaNDM, blaVIM, blaOXA-58-like, and mcr-1, while the remaining strains contained blaNDM, blaIMP, blaOXA-58-like, and mcr-1. This is the first report of more than two carbapenem resistance genes coexisting with the mcr-1 gene in Acinetobacter species.

In the 74 fecal survey samples, 57 blaNDM genes were detected; the mcr-1 gene was detected in five isolates, and co-existed with carbapenem resistance genes. In 174 clinical infection samples, both the blaNDM and mcr-1 genes were absent, and no more than two carbapenemase genes co-existed. Comparison of clinical infection samples and fecal survey samples revealed that the detection rate of carbapenem resistance and mcr-1 genes was higher in fecal survey samples, and the strains in which multiple MDR genes co-existed were all isolated from fecal samples, which may indicate that the intestinal tract is an important place for the transfer of bacterial resistance, reminding us of the importance to control nosocomial infections.

We observed a high mutation rate in the porin carO gene. The sequencing of carO genes revealed the presence of 11 distinct variants and 10 novel sequences. We analyzed the homology of carO gene in carbapenem-resistant Acinetobacter and carbapenem-sensitive reference strain ATCC 17978. The resulting nucleotide sequence and amino acid sequence exhibited 90 and 91% identity, respectively, with the carO sequence of ATCC 17978. When we compared these sequences with those available in different databanks (sequences from NCBI databanks), we identified a non-variable N-terminal domain (1–131) and two variable domains (132–162 and 200–238) by comparing amino acid sequences (Additional file 2: Figure S1), and two variable domains (397–480 and 588–677) by comparing nucleotide sequences (Additional file 3: Figure S2). We speculate that the structure and function of protein CarO will have considerable variation due to the presence of these two variable domains. The phenomenon of CarO participates in carbapenem influx has been reported, changes in the amino acid composition of CarO that could lead to altered porin may result in carbapenem resistance [17], although this will require further verification.

Many studies have described the harboring of the AdeABC efflux system in clinical isolates of A. baumannii [4]. Some studies have suggested that these genes are only present in MDR isolates while some studies have reported them in MDR and non-MDR isolates [52, 53]. Here, we report that the presence of adeABC and adeIJK in A. baumannii was 96.13 and 92.27%, but 12.77 and 2.13% in A. junii, respectively. Sirawit et al. [53] showed that the MDR phenotype in most A. baumannii isolates was associated with efflux pumps. AdeIJK expression was most common (97%), and 52% of the A. baumannii isolates simultaneously produced up to three RND-type efflux systems. Yoon et al. [54] detected the AdeB gene in 13 (92.86%) clinical isolates and AdeG and intrinsic AdeJ in all clinical isolates of A. baumannii. This is consistent with other studies [17], suggesting that efflux systems are likely species-specific. Overexpression of these pumps in response to antibiotic exposure leads to decreased susceptibility to the antibiotics [55].

A. baumannii was primarily isolated from clinical infection samples, and non-baumannii Acinetobacter was primarily isolated from fecal survey samples. ERIC-PCR demonstrated a high genetic diversity of non-baumannii Acinetobacter from different samples, but A. baumannii from clinical infection samples had greater than 90% similarity, suggesting that A. baumannii isolated from clinical infection samples may be from the same clone strain. However, all strains were isolated from samples collected from 2014 to 2015, thus indicating that there was no outbreak in the short term.

Conclusion

The data demonstrated that carbapenem resistance genes were widely disseminated among Acinetobacter species in southern China, and multiple mechanisms contribute to carbapenem resistance. Colistin is the last line of defense to treat infection with carbapenem-resistant gram-negative bacteria, but the emergence of the mcr gene has brought greater challenges to clinical treatment. The appearance of notable clinical resistance genes such as blaNDM and mcr-1, even in non-baumannii Acinetobacter gut colonizers, is important. According to the types and quantities of drug-resistant genes in Acinetobacter isolated from feces, the gut may be an important reservoir of resistant opportunistic bacteria.