- Research article
- Open Access
Epidemiological and genomic characteristics of Acinetobacter baumannii from different infection sites using comparative genomics
BMC Genomics volume 22, Article number: 530 (2021)
Acinetobacter baumannii is a common nosocomial pathogen that poses a huge threat to global health. Owing to the severity of A. baumannii infections, it became necessary to investigate the epidemiological characteristics of A. baumannii in Chinese hospitals and find the reasons for the high antibiotic resistance rate and mortality. This study aimed to investigate the epidemiologic and genetic characteristics of A. baumannii isolated from patients with hospital acquired pneumonia (HAP), bloodstream infection (BSI) and urinary tract infection (UTI) in China and uncover potential mechanisms for multi-drug resistance and virulence characteristics of A. baumannii isolates.
All isolates were classified into two primary clades in core gene-based phylogenetic relationship. Clonal complex 208 (CC208) mainly consisted of ST195 (32 %) and ST208 (24.6 %). CC208 and non-CC208 isolates had carbapenem resistance rates of 96.2 and 9.1 %, respectively. Core genes were enriched in ‘Amino acid transport and metabolism’, ‘Translation’, ‘Energy production and conversion’, ‘Transcription’, ‘Inorganic ion transport and metabolism’ and ‘Cell wall/membrane/envelope synthesis’. Most isolates possessed virulence factors related to polysaccharide biosynthesis, capsular polysaccharide synthesis and motility. Eleven isolates belong to ST369 or ST191 (oxford scheme) all had the virulence factor cap8E and it had a higher positive rate in UTI (35.3 %) than in BSI (18.9 %) and HAP (12.9 %). ABGRI1 antibiotic resistance islands were responsible for streptomycin, tetracycline and sulfonate resistance. The blaOXA−23 gene was the most probable cause for carbapenem resistance, although the blaOXA−66 gene with nonsynonymous SNPs (F82L, I129L) was not.
A. baumannii is a genomically variable pathogen that has the potential to cause a range of infectious diseases. There is high proportion of carbapenem resistance in isolates from all three infection sites (HAP, BSI and UTI), which can be attributed to the blaOXA−23 gene. CC208 is the predominant clone in blaOXA−23-carrying A. baumannii that should be monitored. Virulence factors involving bacteria motility and polysaccharide biosynthesis which are widespread in clinical A. baumannii strains deserve our attention.
Acinetobacter baumannii has emerged as a dominant opportunistic Gram-negative bacterium causing a wide range of nosocomial infections . The carbapenem resistance rate of A. baumannii reached nearly 75 % in response to treatment with imipenem or meropenem in China during 2019 (http://www.chinets.com/). Of particular concern is the mortality rate of ventilator associated pneumonia (VAP) in intensive care units (ICU), which ranges from 45.6 to 60.9 % and even reaches 84.3 % when VAP is caused by extensively drug-resistant A. baumannii (XDRAB) . Owing to the severe consequences of A. baumannii infection, it is necessary to investigate its epidemiological characteristics and explore potential causes for the high antibiotic resistance rate and severe infections.
Several studies have focused on the molecular epidemiology of clinical A. baumannii. In northern China, ST191 and ST195 are the most common sequence types (STs) belonging to clonal complex 92 (also known as CC208). All of the A. baumannii isolates of these two STs contained carbapenem resistance gene blaOXA−23 , while ST208 and ST191 are likely the most common STs in southern China [4, 5].
Several studies investigated the molecular epidemiology of A. baumannii, focusing on its carbapenem resistance and virulence [6,7,8]. Few studies pay attention to the whole genome characteristics of clinical isolates from diverse infection sites. Investigating the phenotype and genotype characteristics of pathogens (i.e. phylogenetic relationship, sequence types, resistance and virulence genes) from different infections may provide meaningful information for empirical clinical medications, thus reducing the resistance transmission and mortality caused by hypervirulent strains. In our study, differential genetic characteristics were analysed among 64 isolates from hospital acquired pneumonia (HAP), bloodstream infection (BSI), or urinary tract infection (UTI). The epidemiological characteristics, resistance and virulence mechanisms were investigated according to draft genomes. Comparative genomic analysis was conducted in order to identify the structural and functional genomic relationship among these isolates.
In vitro susceptibility of A. baumannii to antimicrobial agents
The resistance rates of clinical isolates of A. baumannii against antimicrobial agents are shown in Fig. 1. No polymyxins-resistant isolates were detected. Divided by infection types, isolates from HAP exhibited the highest resistance rates to other antimicrobial agents, while isolates from UTIs showed the lowest. The specific MIC values are presented in Table S4.
Functional annotation of the genomic sequence of 64 A. baumanii isolates
Table S1 provides the quality information of sequencing data. Among 2403 core genes, 2086 (86.8 %) in total were annotated into 21 COG terms (Fig. 2). Most of them were annotated into ‘Amino acid transport and metabolism’, ‘Translation’, ‘Energy production and conversion’, ‘Transcription’, ‘Inorganic ion transport and metabolism’ and ‘Cell wall/membrane/envelope synthesis’.
Virulence genes and infection sites
Multiple virulence factors were identified in A. baumannii such as tviB (polysaccharide biosynthesis), cap8E (capsular polysaccharide synthesis), pilT, pilG and pilU which are twitching motility proteins. Almost every isolate had at least one twitching motility protein which played a vital role in bacterial invasiveness and colonization. Twitching motility protein encoded gene pilU was widespread in all infection sites while pilT and pilG were both primarily identified in BSI and UTI (Figure S1). The gene cap8E involved in capsular polysaccharide synthesis had a higher positive rate in UTI (35.3 %) than in BSI (18.9 %) and HAP (12.9 %) although there was no significant difference (p = 0.18, chi-square test). The virulence factor bplB which encoded probable acetyltransferase was mostly identified in isolates belonging to ST195 (19/21, Fig. 3).
MLST and eBURST analysis
Sixty-four A. baumannii isolates were divided into 11 STs according to the Oxford scheme. ST195 (21/64, 32.8 %) was the dominant sequence type followed by ST208 (16/64, 25.0 %), ST369 (6/64, 9.38 %), ST191 (5/64 7.81 %), ST540 (3/64, 4.69 %) and ST218 (2, 3.13 %). Each of the remaining sequence types (ST761, ST429, ST852, ST605 and ST373) had only one representative isolate. The six new STs were assigned sequence types from ST2242 ~ ST2245 and two strains were named STnew01 and STnew02 because of only 6/7 host genes. Figure 4 A displays the distribution of STs across each infection site. The HAP group was mainly composed of ST195 and ST208, while ST208 accounted for most of the BSI group. Isolates from the UTI group were more genetically diverse with 5 new STs. When using the Pasteur scheme, ST2 (53/64, 82.8 %) belonging to global clone II (GC2) accounted for the most of the sequence types.
The eBURST analysis depending on Oxford results showed that ST195 was the primary founder (Fig. 4B). ST195, ST208, ST369, ST191, ST540 and ST218 (53/64, 82.8 % in total) all belonged to clonal complex 208 (CC208, corresponding to GC2). The carbapenem resistance rates of CC208 and non-CC208 isolates were 96.2 and 9.1 %, respectively, suggesting that CC208 is a major epidemic clonal complex of carbapenem-resistant A. baumannii.
The core genome-based phylogenetic tree is presented in Fig. 5. It showed that clinical isolates from different infection sites were interspersed in the core-gene based phylogenetic tree rather than clustered together. All our clinical isolates were classified into two primary clades: CC208 and non-CC208. In the non-CC208 group, 9 of 11 were UTI isolates and 10 of 11 were carbapenem-susceptible. All of the CC208 isolates were carbapenem-resistant with blaOXA−23 gene, except HAP-isolates 080311 and 130,811 which demonstrated susceptibility to carbapenems without blaOXA−23. Intriguingly, one resistant isolate 172,315 from UTI carried carbapenemase gene blaOXA−58 instead of blaOXA−23. The positive rate of the tetracycline resistance gene tet (A) was 71.9 %, while that of blaTEM−1 responsible for monobactam resistance was 73.4 % (Fig. 5). A. baumannii isolates from public database were closer to our UTI isolates and A. pitti was far away from all the isolates.
In the cladogram, 69 A. baumannii isolates were separated into several clusters. Strictly speaking, these clusters were not well divided by infection sites or sequence types, however, strains with the same sequence type tend to cluster in near branches. Strikingly, virulence factor cap8E which encode capsular polysaccharide synthesis enzyme were positive in all isolates belonging to ST191 and ST369.
The SNP (Figure S2) and core-gene (Fig. 5) based phylogenetic tree display high consistency. Specifically, 53 isolates were clustered into CC-208 group with close genetic distance while the remaining 11 isolates belong to another cluster.
AbGRI1 antibiotic resistance islands
Intact genomic islands were confirmed in 35 of 64 isolates (Table S2). The smallest island was 9.11 kb and the largest reached 37.8 kb. The GC contents in the islands ranged from 33.7 to 46.8 % (median 40.9 %), compared to 38.9 % of the whole genomes of 64 isolates.
Among the 16 isolates from ST195 (n = 4) and ST208 (n = 12), 14 aside from 160058 and 130811 had identical Tn6022△ structure carrying genes tniA (transposase), tniB (NTP-binding protein), uspA (universal stress protein) and sup (sulphate permease) (Fig. 6). The truncated Tn5393△ structure with strA and strB (streptomycin phosphotransferase) was detected in all of the isolates near the 5’ end of the comM gene. The other genes, such as tet(B), tetR and sul2, were located in the remaining regions conferring tetracycline and sulfonamide resistance. The genetic structures of the ABGRI1 resistance islands of the other 19 isolates are provided in Table S3.
The genetic context of the genomic islands was the same, indicating that the acquisition was ancestral to these 16 strains. The upstream genes encoded proteins included DNA modification methylase, diaminohydroxyphosphoribosylaminopyrimidine deaminase, ammonium transporter and ribonucleotide reductase transcriptional regulator NrdR. The downstream genes were involved in dihydrodipicolinate synthase family PA0223, ribulose-5-phosphate 4-epimerase and transcriptional regulator, GntR family domain.
Relationship between single nucleotide polymorphisms (SNPs) and drug resistance
All identified SNPs were gene mutations with no insertion or deletion detected. Table S4 enumerates all the ns-SNPs that occurred in 64 A. baumannii. All of the ns-SNPs in the polymyxin resistance genes are listed in Table 1. The most frequently-occurring ns-SNPs had two amino acid substitutions in the histidine kinase gene pmrB which were V9I and I216T. The AA substitution also occurred in functional regions such as I216T in the HisK domain and Q344P in the HATPaseC domain. For the response regulator gene pmrA, mutations were discovered in the predicted receiver domain (A39T) and the unknown functional domain (S119T). Lipid A synthesis genes lpxA and lpxC had one (H131Y) and four (D159N, H149Y, D287N, M115I) amino acid substitutions, respectively, although no polymyxin resistance was detected with MICs ≤ 1 mg/L.
No nucleotide mutation was found in the carbapenemase gene blaOXA−23. Eleven of twelve non-CC208 isolates had the ns-SNPs in blaOXA−66 gene, although these 11 isolates exhibited susceptibility to carbapenems. It is noteworthy that three CC208 isolates belonging to ST369 which carries blaOXA−23 had the same non-synonymous mutation (F82L, I129L) in blaOXA−66 and displayed resistance to both meropenem and doripenem. They were identified as gene blaOXA−83 with single mutation (I129L) and blaOXA−425 with two mutations (F82L, I129L) in blaOXA−66 [9, 10]. The corresponding base change were 244T→C and 385 A→C. Nevertheless, in our cloning experiments, the MICs of E.coli top10 with mutation 244T→C remained constant or had an ≥ 2-fold decrease compared to wild-type E.coli top10. A 2-fold increase of the MICs was observed for E.coli top10 with mutation 385 A→C. For transformants with both mutations, the MICs remained the same or doubled (Table 2).
In the MIC tests, two tigecycline resistant isolates (MIC = 8 mg/L) were identified. One ns-SNP of I100N in efflux pump gene adeS may explain the tigecycline-resistance of 160098 while for the remaining isolates (172315, 131211, 020411, 162456) that were non-susceptible to tigecycline, no AA substitution were identified in genes adeR and adeS.
A. baumannii is a common nosocomial pathogen, mostly causing hospital acquired pneumonia, bloodstream infection and urinary tract infection. However, the genomic characteristics, structural and functional relationship of A. baumannii from the three infection sites are still unknown. Investigating the epidemiological and genomic characteristics of these isolates provides reference for monitoring the antimicrobial resistance and virulence, thus reducing epidemiology of pathogenetic strains. In our study, we investigated the resistance and virulence characteristics of A. baumannii from HAP, BSI and UTI, analysing the probable resistance mechanisms and phylogenetic relationship of A. baumannii.
Previous studies have shown that A. baumannii belonging to ST195, ST208, ST191 and ST365 are the most widely spread in Chinese hospitals [4, 11]. The dominant types in our study were ST195, ST208, ST218 and ST191. Li  collected 52 clinical isolates of A. baumannii mainly from sputum as well as bloodstream. The proportion of multidrug and pan-drug resistant isolates in the CC208 group was not significantly different from that in the non-CC208 group. In our study, however, the carbapenem resistance rate of CC208 isolates was significantly higher than that of the non-CC208 group. Most studies have shown that A. baumannii CC208 outbreaks are highly correlated with the presence of the blaOXA−23 gene . This was also confirmed in our study, in which 51 carbapenem-resistant isolates from the CC208 group (n = 53) were positive for the blaOXA−23 gene. CC208/GC2 was the largest clonal complex in A. baumannii which can often carry carbapenemase genes like blaOXA−23−like, blaOXA−40−like and blaOXA−58−like. Compared to CC109/GC1 and CC187/GC3, CC208 was the only complex carrying three types of hydrolases; this could explain its wide spread to some extent . It needs to be mentioned that MLST schemes are probably not the best ways to type A. baumannii due to its high genome variations. In order to evaluate the genetic relationship more accurately, we established phylogenetic relationship between strains of A. baumannii based on core genes, non-core genes, as well as SNPs. Strains are not well clustered according to infection sites, indicating that no specific core or accessary genes are responsible for different infections. In contrast, Liu  and Zhao et al.  claimed that unique genes may be closely associated with the ability of the bacteria to adapt to challenging niches. Our phylogenetic analysis also demonstrated that A. baumannii is a genomically variable pathogen that has the potential to cause a range of infectious diseases . Considering the importance of phylogenetic relationship and sequence type, BacWGSTdb 2.0 (http://bacdb.cn/BacWGSTdb/)  provides a platform for rapid typing, phylogenetic relatedness linked to antibiotic resistance genes and clinical data, which is useful for further investigations.
Multiple factors contribute to the virulence of A. baumannii such as biofilm formation, motility, glycosylation and the micronutrient acquisition system . In this study, motility proteins (pilT, pilU, pilG) and a polysaccharide biosynthesis protein (tviB) seemed to take a leading role in the fitness and colonization success of A. baumannii. Subashchandrabose and Wang et al. investigated about the genes necessary for persistence in the lung and for bacterial survival in bloodstream infection. Seven fitness genes were identified in these two studies, suggesting the presence of a core set of fitness genes irrespective of the site of infection [19, 20]. In the present study, the gene pilU was widespread in three infection sites while pilT and pilG genes tended to exist in BSI and UTI. It was noteworthy that virulence factors were likely associated with sequence types. The gene cap8E encoding capsular polysaccharide synthesis enzyme was identified in all isolates classified as ST191 and ST369. It has been recently reported that one of seven hypervirulent A. baumannii was identified as ST369 . Further investigations are warranted to confirm the relationship between virulence genes and sequence types and monitor the key virulence factors associated with hypervirulence.
The multi-drug resistance of A. baumannii is a huge threat for clinical treatment and patient health. ABGRI1 resistance islands are a class of vital mobile genetic elements known to be involved in multiple antimicrobial resistance in A. baumannii GC2 . Tn6022 and Tn6022△ were the most common transposons in AbGRI1. Tn6022 consists of 7 known functional genes and 2 open reading frames and carries no resistance genes [22, 23]. In our study, the ABGRI1 resistance islands inserted in the comM gene shared similar backbones. Tn6022△ consists of tniA, tniB, uspA and sup with tniC, tniD and tniE deleted compared to Tn6022. Tn6022 has been shown to sometimes acquire the OXA-23 carbapenem resistance transposon Tn2006 , though in our study blaOXA−23 may be located on mobile genetic elements like Tn2009 or Tn2006 which are the most common carriers and it is always plasmid-carried which enhances the spread of resistance . Gene blaOXA−23 was not detected in ABGRI1 of our clinical isolates but it has been identified in Tn2006 in AbaR4 . Genes associated with streptomycin, tetracycline and sulfonamide resistance were located on the genomic islands which indicates that ABGRI1 resistance island was not the only contributor to the MDR or carbapenem-resistant phenotype.
Oxacillinases are major causes of carbapenem resistance in A. baumannii. Within the CC208 group, the carbapenem-resistant isolates were all positive for blaOXA−23. For the three carbapenem-resistant isolates belonging to ST369 (050711, 051211, 130911), two non-synonymous mutations were both discovered in blaOXA−66 gene which referred to F82L and I129L. Previous studies have shown that enzyme OXA-66 can be converted to OXA-83, another subtype in the OXA-51 family, after the substitution of I129L. OXA-83 was first detected in two meropenem-resistant A. baumannii strains in the United Kingdom but the values of MIC were both 4 mg/L for imipenem . In terms of tertiary protein structure, Ile-129 was close to the active site Ser-80 and the δ carbon of this isoleucine would cause a steric clash with the hydroxyethyl group of carbapenems  that was adverse to substrate binding. I129L relieved this clash, thus promoting carbapenem binding. This has been confirmed by molecular dynamics simulations . When F82L and I129L substitutions both occurred in OXA-66, the subtype was specified as OXA-425. The OXA-425-positive and carbapenem resistant strain was first isolated in Beijing . To the best of our knowledge, the influence of F82L and I129L on carbapenem resistance has not been confirmed by separate cloning experiments. Our study verified that both of these substitutions failed to cause carbapenem resistance. This suggests that the blaOXA−23 gene should be the most likely factor for carbapenem resistance in our A. baumannii isolates and mutational blaOXA−66 was not.
The resistance mechanism of A. baumannii to polymyxins is mainly regulated by two pathways. One is point mutations of the lipid A synthesis-related genes lpxA, lpxC and lpxD, which inhibits the synthesis of lipid A . Another mechanism is regulated by the two-component system of pmrAB. Studies have shown that point mutations in pmrB and the subsequent upregulation of pmrAB are critical for polymyxin resistance . Resistance-related point mutations are mainly located in the histidine kinase domain (HisK, AA 216 ~ 276) and the ATP binding domain (HATPaseC, AA 331 ~ 419) of pmrB . Resistance caused by point mutations of pmrA has also been reported , especially in the signal receiver domain (Rec). In this study, several non-synonymous mutations occurred in the pmrAB functional domain demonstrating that not all non-synonymous mutations in the pmrAB functional region cause resistance. Amino acid changes in the PmrAB two-component system have been suggested not essential for A. baumannii colistin resistance . Meanwhile, no resistance occurred in the mutants with ns-SNPs in the lpxA and lpxC genes.
Over-expression of AdeABC efflux pump stimulated continuously by the mutated AdeRS two component system has been found to result in tigecycline resistance. The AA substitution in gene adeS (I100N) is probably responsible for higher MIC of isolate 160098. For remaining tigecycline non-susceptible isolates, mutations in regulatory genes of resistance-nodulation cell division efflux pumps such as adeN, adeJ and ISaba1 insertion into genes adeN and adeRS may explain the tigecycline resistance [32,33,34].
In summary, our study sheds new light on the epidemiological characteristics and phylogenetic relationship of clinical A. baumannii across China and uncovers the possible molecular mechanisms of multi-drug resistance and virulence. The blaOXA−23 gene is probably responsible for high proportion of carbapenem resistance. CC208 was the predominant clone in blaOXA−23-carrying A. baumannii. Several key virulence factors such as cap8E also deserves attention. Multiple phylogenetic analysis indicates that A. baumannii is a genomically variable pathogen that has the potential to cause a range of infectious diseases. All of the evidence indicates that the resistance and virulence should be monitored to reduce the resistance transmission and mortality caused by probably hypervirulent A. baumannii.
Bacterial isolates and antimicrobial susceptibility testing
In this study, clinical A. baumannii isolates from BSI (n = 17) and UTI (n = 16) were collected from Huashan Hospital in Shanghai between 2016 and 2017. Isolates from HAP (n = 31) were from a domestic thirteen-centre clinical study on colistin methanesulfonate (registration number: NCT01940731) including Hunan People’s Hospital, Shanghai 10th People’s Hospital, The Second Hospital of Jilin University, West China Hospital of Sichuan University, etc. across China. Colistin sulfate (lot number SLBD8306V; Sigma-Aldrich, St Louis, MO), polymyxin B (lot number R046V0; USP), minocycline, doxycycline, sulbactam, meropenem, aztreonam (lot numbers: 130514–200401, 130485–201703, 130430–201408, 130506–201403 and 130507–201303, respectively; National Institutes for Food and Drug Control, Beijing, China), tigecycline (lot number: 10-MWC-62-1; USP) and doripenem (lot number 0379; Shionogi & Co Ltd) were used in this study. The minimum inhibitory concentrations (MICs) of 64 isolates were determined using the microbroth dilution method for three replicates in one batch with Escherichia coli ATCC25922 as the quality control. A total of nine antimicrobials were included and the results were interpreted referring to the CLSI where possible (CLSI M100 2020). CLSI does not currently provide breakpoints for tigecycline, sulbactam and aztreonam. FDA-recommended criteria for Enterobacteriaceae were used for tigecycline susceptibility (≤ 2, 4, ≥ 8 mg/L) .
DNA extraction and whole genome sequencing
The genome DNA of 64 isolates were extracted according to the Takara DNA Extraction Kit protocol. The whole genome was sequenced using the Illumina Hiseq X10 platform, with the 2*150 bp paired-end sequencing strategy . The raw read data were assembled de novo using Velvet software .
Functional annotation of resistance and virulence genes
For functional classification of the predicted core genes, BLASTP was used to align amino acids of predicted genes against the Clusters of Orthologous Groups (COG) database with an expected threshold of 1E− 3 using the Conserved Domains Database (CDD) . We also performed sequence alignment of the amino acid sequences to the NCBI non-redundant (NR) database (E-value ≤ 1E− 3).
In order to identify antibiotic resistance genes, the protein-coding sequences were aligned against the Comprehensive Antibiotic Resistance Database (CARD, https://card.mcmaster.ca/) . The similarity and length consistency were required to be over 80 %. To search possible virulence factors, the Virulence Factors Database (VFDB, http://www.mgc.ac.cn/VFs/)  was aligned to the ORF protein sequences, using the same thresholds as alignment of resistance genes. Single nucleotide polymorphisms in genes were investigated including lpxA, lpxC, pmrA, pmrB (polymyxin resistance), blaOXA−23, blaOXA−66 (carbapenem resistance), carO, oprD (efflux pump), blaTem−1 (monobactam, penem), tet (A), adeR, adeS (tetracycline, glycylcycline). Only non-synonymous SNPs (ns-SNPs) were included in the analysis.
Multilocus sequence typing (MLST) and eBURST analysis
To determine the sequence types, multi-locus sequence typing (MLST) was performed according to both the Oxford and Pasteur schemes. Sequences were compared to the PubMLST database for A. baumannii (http://pubmlst.org/abaumannii/) and then assigned to the appropriate sequence types. The genome submission to PubMLST was completed for all the 64 isolates (submission ID: BIGSdb_20201227120226_136766_49027). The following four isolates were assigned new STs: 162912 (ST2242), 171938 (ST2243), 172830 (ST2244) and 173876 (ST2245). Another two isolates 170876 and 173011 had only 6/7 host genes, they can not be assigned new STs in PubMLST. Thus, they were named STnew01 and STnew02.
The eBURST analysis of the isolates was performed to determine their homology . Isolates sharing 6/7 alleles were considered to be a single clonal complex (CC) group.
Phylogenetic trees were constructed using three strategies: gene-by-gene genomic analysis (cgMLST), non-core gene-based strategy (cladogram) and a reference genome-based single nucleotide polymorphism (SNP) strategy. Four A. baumannii isolates (ATCC17978, AYE, ACICU, AB5075-UW) and one A. pitti isolate (XJ88) from public database were included in the analysis. The detailed information about these isolates is provided in Table S5. Raw sequence data from all clinical isolates were independently mapped to the reference isolate A. baumannii AC30 genome sequence (accession number: CP007577.1) using bowtie2 (http://bowtie-bio.sourceforge.net/bowtie2/index.shtml). A. baumannii AC30 belongs to ST 195 which is the most epidemic sequence type in our study. GATK (https://gatk.broadinstitute.org/hc/en-us) was utilized to identify mutations such as SNPs and INDEL. High-quality single nucleotide polymorphisms (SNPs) were selected with a mapping coverage of > 10 and a frequency of > 70 %. The multiple sequence alignment was performed using MAFFT and phylogenetic trees were constructed with FastTree following the maximum likelihood method. The cladogram was constructed according to the presence /absence of non-core genes based on the pangenome analysis using the software ROARY. The iTOL v5 (https://itol.embl.de/) was utilized for integrating the phylogenetic tree with sequence types, infection sites, resistance phenotype and genotype and virulence factors.
ABGRI1 resistance island analysis
The ABGRI1 inserted in the comM gene was extracted and analysed. An intact ABGRI1 could be split on different contigs in draft genomes; thus only intact ABGRI1 islands were chosen for further comparative analysis. The annotations were performed with RAST (https://rast.nmpdr.org/) and genomic island structure was constructed using R 3.5.0 for isolates belonging to ST195 and ST208.
Cloning experiments were conducted to determine the effects of non-synonymous mutations in blaOXA− 66 gene on carbapenem susceptibility. In order to clone blaOXA− 66 into pHSG398 in the proper orientation, restriction sites BamHI and PstI in the multiple cloning sites downstream of the pLac promoter of pHSG398 were chosen to insert the 5’-end and 3’-end of blaOXA− 66, respectively. Primers were designed according to the principles listed in the In-Fusion® HD Cloning Kit User Manual (Takara Bio USA, Inc): adding 15 to 20 bp homologous sequences of linearized vector to the 5’-ends of both forward and reverse primers to make the ends of amplified inserts and linearized vectors identical to each other (Table 3). Thus, primers BamHI-blaOXA− 66-FW and PstI-blaOXA− 66-RW were used to amplify the blaOXA− 66 gene. Purified PCR product of blaOXA− 66 gene and the linearized pHSG398 vector, which were digested by BamHI and PstI enzymes, were mixed at an appropriate ratio and incubated with 5× In-Fusion HD Enzyme Premix at 50℃ for 15 min. Then the recombination product was transformed into competent cell E.coli Top10 by heat-shock and screened on plates containing 50 mg/L chloramphenicol. The clones with blaOXA− 66 were further confirmed by PCR and sequencing.
For the blaOXA−66 alleles with T244C mutation, similar cloning experiment was performed except that the insert gene was amplified with two pairs of PCR primers: BamHI-blaOXA−66-FW and T244C-RW for the first 264 bp of blaOXA−66 gene with T244C mutation, and T244C-FIW and PstI-blaOXA−66-RW for the last 581 bp of blaOXA−66 gene. Both the two purified PCR fragments were mixed with linearized pHSG398 vector at an appropriate ratio together with 5×In-Fusion HD Enzyme Premix for recombination reaction.
As for the blaOXA−66 allele with T244C and A385C mutations, three pairs of primers were used for amplification: BamHI-blaOXA−66-FW and T244C-RW, T244C-FIW and A385C-RW, and A385C-FIW and PstI-blaOXA−66-RW.
The bases in the box were designed for the mutation of T244C and A385C.
Accession numbers of 64 A. baumannii are provided in Table S4.
Availability of data and materials
The datasets analysed in this study are all available in NCBI (BioProject number PRJNA633416, https://www.ncbi.nlm.nih.gov/bioproject/PRJNA633416).
Hospital acquired pneumonia
Urinary tract infection
Extensively drug-resistant A. baumannii
Multilocus sequence typing
Single nucleotide polymorphisms
Minimum inhibitory concentration
Mahamat A, Bertrand X, Moreau B, Hommel D, Couppie P, Simonnet C, et al. Clinical epidemiology and resistance mechanisms of carbapenem-resistant Acinetobacter baumannii, French Guiana, 2008–2014. Int J Antimicrob Agents. 2016;48:51–5.
Inchai J, Pothirat C, Bumroongkit C, Limsukon A, Khositsakulchai W, Liwsrisakun C. Prognostic factors associated with mortality of drug-resistant Acinetobacter baumannii ventilator-associated pneumonia. J Intensive Care. 2015;3:9.
Ning NZ, Liu X, Bao CM, Chen SM, Cui EB, Zhang JL, et al. Molecular epidemiology of blaOXA–23 -producing carbapenem-resistant Acinetobacter baumannii in a single institution over a 65-month period in north China. BMC Infect Dis. 2017;17:14.
Ying J, Lu J, Zong L, Li A, Pan R, Cheng C, et al. Molecular epidemiology and characterization of genotypes of Acinetobacter baumannii isolates from regions of south China. JPN J Infect Dis. 2016;69:180–5.
Li YJ, Pan CZ, Fang CQ, Zhao ZX, Chen HL, Guo PH, et al. Pneumonia caused by extensive drug-resistant Acinetobacter baumannii among hospitalized patients: genetic relationships, risk factors and mortality. BMC Infect Dis. 2017;17:371.
Ruan Z, Chen Y, Jiang Y, Zhou H, Zhou Z, Fu Y, et al. Wide distribution of CC92 carbapenem-resistant and OXA-23-producing Acinetobacter baumannii in multiple provinces of China. Int J Antimicrob Agents. 2013;42:322–8.
Zhou K, Tang X, Wang L, Guo Z, Xiao S, Wang Q, et al. An emerging clone (ST457) of Acinetobacter baumannii Clonal Complex 92 with enhanced virulence and increasing endemicity in south China. Clin Infect Dis. 2018;67:179-S88.
Kim MH, Jeong H, Sim YM, Lee S, Yong D, Ryu CM, et al. Using comparative genomics to understand molecular features of carbapenem-resistant Acinetobacter baumannii from South Korea causing invasive infections and their clinical implications. PLoS One. 2020;15:e0229416.
Turton JF, Ward ME, Woodford N, Kaufmann ME, Pike R, Livermore DM, et al. The role of ISAba1 in expression of OXA carbapenemase genes in Acinetobacter baumannii. FEMS microbiol Lett. 2006;258:72–7.
Li H, Liu F, Zhang Y, Wang X, Zhao C, Chen H, et al. Evolution of carbapenem-resistant Acinetobacter baumannii revealed through whole-genome sequencing and comparative genomic analysis. Antimicrob Agents Chemother. 2015;59:1168–76.
Zhou Y, Wu X, Zhang X, Hu Y, Yang X, Yang Z, et al. Genetic Characterization of ST195 and ST365 carbapenem-resistant Acinetobacter baumannii harboring blaOXA–23 in Guangzhou, China. Microb Drug Resist. 2015;21:386–90.
He C, Xie Y, Fan H, Kang M, Tao C, Zhang R, et al. Spread of imipenem-resistant Acinetobacter baumannii of European clone II in Western China. Int J Antimicrob Agents. 2011;38:257–60.
Evans BA, Amyes SG. OXA beta-lactamases. Clin Microbiol Rev. 2014;27:241–63.
Fei Liu YZ, Yi Y, Lu N, Zhu B, Yongfei, Hu. Comparative genomic analysis of Acinetobacter baumannii clinical isolates reveals extensive genomic variation and diverse antibiotic resistance determinants. BMC Genomics. 2014;15:1163.
Zhao Y, Wu J, Yang J, Sun S, Xiao J, Yu J. PGAP: pan-genomes analysis pipeline. Bioinformatics. 2012;28:416–8.
Jason W, Sahl JKJ, Anthony D, Harris AM, Phillippy WW, Hsiao KA, Thom. David A Rasko. Genomic comparison of multi-drug resistant invasive and colonizing Acinetobacter baumannii isolated from diverse human body sites reveals genomic plasticity. BMC Genomics. 2011;12:291.
Feng Y, Zou S, Chen H, Yu Y, Ruan Z. BacWGSTdb 2.0: a one-stop repository for bacterial whole-genome sequence typing and source tracking. Nucleic Acids Res. 2021;49:D644-D50.
Harding CM, Hennon SW, Feldman MF. Uncovering the mechanisms of Acinetobacter baumannii virulence. Nat Rev Microbiol. 2018;16:91–102.
Subashchandrabose S, Smith S, DeOrnellas V, Crepin S, Kole M, Zahdeh C, et al. Acinetobacter baumannii genes required for bacterial survival during bloodstream infection. mSphere. 2016;1:e00013–15.
Wang N, Ozer EA, Mandel MJ, Hauser AR. Genome-wide identification of Acinetobacter baumannii genes necessary for persistence in the lung. mBio. 2014;5:e01163-14.
Li J, Yu T, Luo Y, Peng JY, Li YJ, Tao XY, et al. Characterization of carbapenem-resistant hypervirulent Acinetobacter baumannii strains isolated from hospitalized patients in the mid-south region of China. BMC Microbiol. 2020;20:281.
Nigro SJ, Hall RM. Antibiotic resistance islands in A320 (RUH134), the reference strain for Acinetobacter baumannii global clone 2. J Antimicrob Chemother. 2012;67:335–8.
Yoon EJ, Kim JO, Yang JW, Kim HS, Lee KJ, Jeong SH, et al. The blaOXA–23-associated transposons in the genome of Acinetobacter spp. represent an epidemiological situation of the species encountering carbapenems. J Antimicrob Chemother. 2017;72:2708–14.
Hamidian M, Kenyon JJ, Holt KE, Pickard D, Hall RM. A conjugative plasmid carrying the carbapenem resistance gene blaOXA–23 in AbaR4 in an extensively resistant GC1 Acinetobacter baumannii isolate. J Antimicrob Chemother. 2014;69:2625–8.
Smith CA, Antunes NT, Stewart NK, Frase H, Toth M, Kantardjieff KA, et al. Structural basis for enhancement of carbapenemase activity in the OXA-51 family of class D beta-lactamases. ACS Chem Bio. 2015;10:1791–6.
June CM, Muckenthaler TJ, Schroder EC, Klamer ZL, Wawrzak Z, Powers RA, et al. The structure of a doripenem-bound OXA-51 class D beta-lactamase variant with enhanced carbapenemase activity. Protein Sci: a publication of the Protein Society. 2016;25:2152–63.
Moffatt JH, Harper M, Harrison P, Hale JD, Vinogradov E, Seemann T, et al. Colistin resistance in Acinetobacter baumannii is mediated by complete loss of lipopolysaccharide production. Antimicrob Agents Chemother. 2010;54:4971–7.
Beceiro A, Llobet E, Aranda J, Bengoechea JA, Doumith M, Hornsey M, et al. Phosphoethanolamine modification of lipid A in colistin-resistant variants of Acinetobacter baumannii mediated by the PmrAB two-component regulatory system. Antimicrob Agents Chemother. 2011;55:3370–9.
Arroyo LA, Herrera CM, Fernandez L, Hankins JV, Trent MS, Hancock RE. The pmrCAB operon mediates polymyxin resistance in Acinetobacter baumannii ATCC 17978 and clinical isolates through phosphoethanolamine modification of lipid A. Antimicrob Agents Chemother. 2011;55:3743–51.
Adams MD, Nickel GC, Bajaksouzian S, Lavender H, Murthy AR, Jacobs MR, et al. Resistance to colistin in Acinetobacter baumannii associated with mutations in the PmrAB two-component system. Antimicrob Agents Chemother. 2009;53:3628–34.
Park YK, Choi JY, Shin D, Ko KS. Correlation between overexpression and amino acid substitution of the PmrAB locus and colistin resistance in Acinetobacter baumannii. Int J Antimicrob Agents. 2011;37:525–30.
Sun JR, Perng CL, Chan MC, Morita Y, Lin JC, Su CM, et al. A truncated AdeS kinase protein generated by ISAba1 insertion correlates with tigecycline resistance in Acinetobacter baumannii. PLoS One. 2012;7:e49534.
Grossman TH. Tetracycline antibiotics and resistance. Cold Spring Harb Perspect Med. 2016;6:a025387.
Gerson S, Nowak J, Zander E, Ertel J, Wen Y, Krut O, et al. Diversity of mutations in regulatory genes of resistance-nodulation-cell division efflux pumps in association with tigecycline resistance in Acinetobacter baumannii. J Antimicrob Chemother. 2018;73:1501–8.
Shankar C, Nabarro LEB, Anandan S, Veeraraghavan B. Minocycline and tigecycline: what is their role in the treatment of carbapenem-resistant Gram-negative organisms? Microb Drug Resist. 2017;23:437–46.
Chen Q, Zhou JW, Fan JZ, Wu SH, Xu LH, Jiang Y, et al. Simultaneous emergence and rapid spread of three OXA-23 producing Acinetobacter baumannii ST208 strains in intensive care units confirmed by whole genome sequencing. Infect Genet Evol. 2018;58:243–50.
Zerbino DR. Using the Velvet de novo assembler for short-read sequencing technologies. Curr Protoc Bioinformatics. 2010;Chap. 11:Unit 11 5.
Si-Tuan N, Ngoc HM, Hang PTT, Nguyen C, Van PH, Huong NT. New eight genes identified at the clinical multidrug-resistant Acinetobacter baumannii DMS06669 strain in a Vietnam hospital. Ann Clin Microbiol Antimicrob. 2017;16:74.
Jia B, Raphenya AR, Alcock B, Waglechner N, Guo P, Tsang KK, et al. CARD 2017: expansion and model-centric curation of the comprehensive antibiotic resistance database. Nucleic Acids Res. 2017;45:D566-D73.
Chen L, Yang J, Yu J, Yao Z, Sun L, Shen Y, et al. VFDB: a reference database for bacterial virulence factors. Nucleic Acids Res. 2005;33:D325-8.
Francisco AP, Bugalho M, Ramirez M, Carrico JA. Global optimal eBURST analysis of multilocus typing data using a graphic matroid approach. BMC Bioinformatics. 2009;10:152.
The authors thank Huashan hospital affiliated to Fudan University for providing clinical isolates. We would like to acknowledge Chinese National Human Genome Center for providing library preparation and whole genome sequencing services.
This research was supported by the Science and Technology Commission of Shanghai Municipality (grant number 19431903000 and 19411964900); Major Research and Development Project of Innovative Drugs, Ministry of Science and Technology of China (grant number 2017ZX09304005). Project of Shanghai Municipal Commission of Health and Family Planning (ZY (2018–2020)-FWTX-1108). Funding bodies played no role in the design of the study or analysis or interpretation of data or in writing the manuscript.
Ethics approval and consent to participate
The clinical trial (registration number: NCT01940731) has been approved by Huashan Institutional Review Board affiliated to Fudan University. Collecting samples were permitted. Samples used in this study all belong to The Culture Collection of Institute of Antibiotics, Huashan Hospital. The Informed Consent Form was waived by Huashan Institutional Review Board if using strains for further study from The Culture Collection. Personal privacy is not involved in this study.
Consent for publication
There is no conflict of interest.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The original version of this article was revised: figure 4 was mistakenly a duplicate of figure 5.
About this article
Cite this article
Bian, X., Liu, X., Zhang, X. et al. Epidemiological and genomic characteristics of Acinetobacter baumannii from different infection sites using comparative genomics. BMC Genomics 22, 530 (2021). https://doi.org/10.1186/s12864-021-07842-5
- Acinetobacter baumannii
- Whole genome sequencing
- Epidemiological characteristics
- Multi-drug resistance
- Comparative genomics