Background

Acinetobacter baumannii (A. baumannii) is a Gram-negative opportunistic pathogen, which can cause pneumonia, meningitis, wound infection, urinary tract infection, bacteremia and so on [1]. In all kinds of hospital-acquired infections caused by Gram-negative bacteria, A. baumannii accounted for 2%-10% [2, 3]. According to the declaration from the World Health Organization (WHO), A. baumannii has become one of the most threatening pathogens and an important focus of public health field and clinical research [4].

In recent years, the infection and drug resistance rates of A. baumannii have been increasing due to various factors, including the widespread use of antibiotics, the treatment of related immunosuppressants and invasive intervention. Some studies show that A. baumannii has the highest multidrug resistance (MDR) and extensive drug resistance (XDR) spectrum [5, 6]. According to the latest data report of China Antimicrobial Surveillance Network (CHINET, http://www.chinets.com/) in 2022, A. baumannii accounted for 7.5% of clinical isolates in China, ranking fifth among them [7]. The management of infected patients is difficult because of its complex drug-resistance mechanisms, the most worrying of which is carbapenem-resistant A. baumannii (CRAB). It is worth noting that the drug resistance rate of A. baumannii to antibiotics commonly used is basically increasing in China year by year. For example, the carbapenem-resistance rate of A. baumannii isolates such as imipenem and meropenem increased from 32.9% and 41.3% in 2005 to 77.7% and 79% in 2019, respectively. After 2019, the carbapenem-resistance rate decreased slightly, but remained above 70% [7].

According to the 2021 report from China Antimicrobial Resistance Surveillance System (CARSS, http://www.carss.cn/Report), he resistance rate of A. baumannii to carbapenem (imipenem or meropenem) was 54.3% on average, which was 0.6% higher than that in 2020 [8]. There are some differences in drug resistance rates among different regions. The average carbapenem-resistance rate of A. baumannii in Gansu province is 48.9% ranking in the middle level in China, which has also attracted our attention.

With the wide application and development of sequencing technology, many studies have focused on the genome-wide characteristics of A. baumannii isolates, including not only drug resistance and drug resistance genes, but also sequence types, plasmids, phylogenetic relationships and so on [9,10,11,12]. However, there are different characteristics in A. baumannii isolates from different hospitals.

Our study first analyzed the distribution and drug resistance characteristics, then clarified genomic characteristics of A. baumannii isolates from a hospital in Gansu province through whole-genome sequencing (WGS) technology, which could help guide infection control measures, prevention, and targeted antimicrobial therapy.

Method

Bacterial classification and antimicrobial susceptibility testing

Sixty-two non-repeated isolates were collected from Qingyang people's hospital in Gansu province, China, from August 2015 to November 2021. Bacterial isolates were initially subjected to identification using the VITEK-2 Compact system. Subsequently, the phenotypic identification was corroborated by whole-genome sequencing (WGS), where the genomic data obtained were compared against reference databases to validate and refine the taxonomic classification at the species or subspecies level. Antimicrobial susceptibility test to common clinical antibiotics was also detected by VITEK-2 compact system. Except the minimum inhibitory concentration (MIC) value of polymyxin was determined by the broth microdilution method. Escherichia coli ATCC25922 was used as a quality control strain. Seventeen common antibiotics include ticarcillin/clavulanic acid, piperacillin/tazobactam, ceftazidime, cefoperazone/sulbactam, cefepime, aztreonam, imipenem, meropenem, amikacin, tobramycin, ciprofloxacin, levofloxacin, doxycycline, minocycline, tigecycline, polymyxin, compound sulfamethoxazole.

The antimicrobial susceptibility test results were interpreted according to the Clinical and Laboratory Standards Institute (CLSI, 2021 version, https://clsi.org) unified protocol. Defining MDR, XDR and Pandrug resistant (PDR) criteria according to an international expert proposal for interim standard definitions for acquired resistance [13].

Whole-genome sequencing

Genomic DNA was extracted from bacterial cultures using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. DNA concentration and purity were measured using Qubit 4.0 (Thermo Fisher Scientific) and Nanodrop One (Thermo Fisher Scientific) at the same time. Sequencing libraries were generated using NEB Next Ultra DNA Library Prep Kit for Illumina (New England Biolabs) following the manufacturer's recommendations. The library quality was assessed on the Qubit 4.0 Fluorometer (Life Technologies) and Qsep400 High-Throughput Nucleic Acid Protein Analysis System (Houze Biological Technology Co.). DNA libraries were constructed with 150-bp paired-end fragments and sequenced on the HiSeq 2500 sequencer at Novogene Company (Beijing, China).

Bioinformatics analysis

Sequencing reads were quality filtered using the FastQC v0.11.8 software [14], adapters and low-quality reads were removed and filtered out using Trimmomatic with default parameters [15]. The A. baumannii strain XH731 (NZ_CP021321.1) was selected as the best-matching chromosomal reference by the genomic distance estimation tool Mash v2.3 [16]. Assembly of sequencing reads was carried out with Unicycler v0.4.8 in normal mode [17]. Then the quality of the acquired genome was assessed by QUAST v5.2.0 [18]. Genome assemblies was annotated by Prokka v1.12 [19]. Then a core-genome alignment was derived by Panaroo v1.2.10 [20]. GTR + I + G4 was selected as the best evolutionary model by using Modeltest-ng v0.1.7 [21]. The maximum likelihood phylogenetic tree was constructed by IQ-TREE v2.2.2 with 1000 bootstrap replicates [22]. The tree was midpoint-rooted and visualized by Figtree v1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/).

Sequence types (STs) and unknown STs were verified by MLST v2.23.0 [23] and MLST server v2.0 via the online service of the Center for Genomic Epidemiology of the Danish University of Technology (CGE) [24] respectively. The MLST analysis of these two software was based on the Institute Pasteur MLST schemes. The capsular polysaccharides loci (KL) and lipooligosaccharides outer core loci (OCL) were detected using the A. baumannii KL and OCL reference sequence databases by Kaptive v2.0.6 [25]. Resistance genes was identified with Abricate v0.8 (https://github.com/tseemann/abricate) using resfinder database [26]. The genotypic structure surrounding the special resistance gene was annotated by RAST (http://rast.nmpdr.org/) [27] and Easyfig v2.2.5 [28] was used to visualize the multiple sequence alignment. As a set of modular tools, the MOB-suite was used to reconstruct and type plasmids [29].

Results

General features of the patients and A. baumannii isolates

From August 2015 to November 2021, A. baumannii isolates was obtained from 62 samples including catheter (1/62), cerebrospinal fluid (1/62), urine (1/62), skin (1/62), wound secretion (3/62), sputum (54/62) and pleural effusion (1/62). Isolates were obtained from 47 (75.8%) men between 36 and 81 years old and 15 (24.2%) women between 12 days and 70 years old. The wards mainly include the intensive care unit (ICU), infection ward, orthopedic ward, respiratory ward, surgery ward, neonatal ward and traditional Chinese medicine ward. The detailed clinical information from patients with A. baumannii isolates was shown in Supplementary Table S1.

Antimicrobial susceptibility testing

As shown in Table 1, 62 A. baumannii isolates were resistant to aztreonam (100%). Most isolates (56/62) were resistant to multiple drugs. Most isolates were resistant to piperacillin/tazobactam (88.71%), ceftazidime (88.71%), imipenem (85.48%), meropenem (85.48%), ciprofloxacin (82.26%), ticarcillin/clavulanic acid (80.65%), cefepime (79.03%), doxycycline (77.42%) and tobramycin (74.19%), but most isolates were sensitive to tigecycline (96.87%), polymyxin (95.16%) and minocycline (79.03%). Among 62 A. baumannii isolates, 55 MDR and 1 XDR A. baumannii isolates were identified, respectively. The MIC values of the 62 isolates were provided in Supplementary Table S2.

Table 1 Antimicrobial susceptibility of Acinetobacter baumannii Isolates
Full size table

Genomic characterization of A. baumannii

The genome sequences of 62 A. baumannii were obtained by whole-genome sequencing (Supplementary Table S3). Reads mapping to A. baumannii reference genome XH731 showed an average of 97.8% coverage among all 62 genomes. Seven sequence types were identified (Fig. 2 and Supplementary Table S1). Most of these A. baumannii isolates belonged to ST2 (52/62), which mainly isolated from ICU (27/52) and neurosurgery ward (12/52). ST2 A. baumannii isolates isolated in 2015, 2016 and 2021, with the largest number in 2016 (29/52). Others belonged to ST34 (1/62), ST40 (3/62), ST104 (1/62), ST132 (2/62) ST1108 (1/62), and ST1376 (2/62).

Thirteen distinct KL (KL3, KL161, KL230, KL47, KL61, KL72, KL106, KL160, KL22, KL33, KL34, KL45, KL9) configurations and five OCL (OCL1, OCL6, OCL21, OCL15, OCL18) configurations were detected among the isolates (Fig. 2 and Supplementary Table S1). The most common KL loci and OCL loci were KL3 (38/62) and OCL1(54/62), respectively.

Phylogenetic analysis

In order to explore the relationship between the A. baumannii isolates collected in this study and those previously studied, genome sequences of 567 A. baumannii isolates (Supplementary Table S4) were downloaded from the Bacterial and Viral Bioinformatics Resource Center (BV-BRC, https://www.bv-brc.org/), which were combined with 62 isolates in this study for the phylogenetic analysis. Figure 1 showed that the 62 isolates in this study did not cluster completely by source, but were dispersed among multiple clades in the common genome. Among the 52 ST2 strains collected in this study, phylogenetic analysis revealed that despite being isolated from the same hospital, they did not cluster on the same branch. Instead, they formed two main branches, comprising 38 and 9 isolates respectively, with the remaining 5 isolates scattered across different branches.

Fig. 1
figure 1

Phylogenetic analysis map of 629 A. baumannii isolates. The phylogenetic tree contains 567 A. baumannii isolates (ab_number) from BV-BRC and 62 A. baumannii isolates in this study (gs_number), annotating the isolation time, country, and host origin of the A. baumannii isolates from the inner ring to the outer ring. The black block means NA (Not available). The blue blocks on the tree branches signifies 62 A. baumannii isolates in this study

Full size image

Antibiotic resistance gene profiles of A. baumannii

The distribution of antibiotic resistance genes (ARGs) of different STs A. baumannii isolates was shown in Fig. 2. A total 27 types of ARGs were identified, conferring resistance to 7 classes of antibiotic including aminoglycosides, beta-lactams, lincomycin, sulphonamide, phenicols, macrolides, tetracycline. Among them, blaTEM-1D (51/61) and blaADC-25 (62/62) genes were the most prevalent on the putative plasmid and chromosomal contigs, respectively. The blaADC-25 gene only on the chromosomal contigs was further explored carried by all strains.

Fig. 2
figure 2

The phylogenetic tree and distribution of STs, KL types, OCL types, plasmids and ARGs among 62 A. baumannii isolates. On the left half, a phylogenetic tree was generated using IQ-TREE v2.2.2 software and midpoint-rooted for improved visualization using Figtree v1.4.4. Isolates are color-coded according to their STs, capsular polysaccharide (KL) types, and lipooligosaccharides outer core loci (OCL). Fifty-two ST2 and one ST1108 isolates were clustered into two monophyletic clades, referred to as Clade1 and Clade2. On the right, four frames provide information about plasmid carriage and distribution of Antibiotic Resistance Genes (ARGs). Frame1 indicates the distribution of plasmids; Frame 2 highlights ARGs found only on plasmid contigs; Frame 3 focuses on ARGs present on both plasmid and chromosome contigs; Frame 4 shows ARGs exclusive to chromosome contigs, the light blue squares represent the absence of plasmid or ARGs, dark blue and purple squares represent one and two plasmids or ARGs, respectively

Full size image

In order to understand the surrounding environment of blaADC-25 gene, we randomly selected a representative isolate of each ST type. Then the blaADC-25-containing contig sequences were extracted from the genome data, among which seven contigs from isolates gs18, 28, 29, 45, 55, 58 and 65. However, no mobile genetic element, such as a transposon or insertion sequence (IS), was found surrounding the blaADC-25, as shown in Fig. 3.

Fig. 3
figure 3

Analysis of blaADC-25 surrounding environment. Each arrow represents an open reading frame (ORF). The red arrows represent blaADC-25 gene, the orange arrows represent putative outer membrane protein and the gray arrows represent hypothetical proteins and others. The grey shaded areas represent highly homologous regions between contigs. The two overly long contigs of gs58 and gs28 have been truncated to improve the figure visualization

Full size image

Plasmid carriage profile

Plasmids were detected in 61 of 62 A. baumannii isolates (98.4%), totaling 110 putative plasmids (median: 2 per isolate; range: 1–5). The putative plasmids were grouped into 11 clusters by mob suite, including rep_cluster_586, rep_cluster_639, rep_cluster_650, rep_cluster_734, rep_cluster_718, rep_cluster_1118, rep_cluster_1172, rep_cluster_1226, rep_cluster_1254, rep_cluster_1259, rep_cluster_1281. A. baumannii ST2 isolates mainly carried rep_cluster_1172 (40/52) and rep_cluster_734 (20/52). In addition, no virulence genes were found in the 110 putative plasmid contigs.

Discussion

Phylogenetic analyses combined with BV-BRC and the data in this study showed that the 629 isolates were derived from 34 countries, with most of them being from Asia. A review summarizes Acinetobacter accounts for approximately 2% of nosocomial infections in the USA but these rates are doubled in Asia and the Middle East [30]. A. baumannii infection in Asia is a considerable challenge to healthcare professionals in hospitals.

According to the source distribution of 62 A. baumannii isolates in this study, the main specimen type was sputum, which is consistent with the results of previous studies [31, 32]. The wards distribution showed that the isolates mainly from ICU and neurosurgery ward. Especially for patients in the ICU, the establishment of artificial airway operations such as ventilator invasion may form biofilms on the surfaces, thus increasing the risk of infection and antibiotic resistance of A. baumannii [2, 33,34,35]. Therefore, patients in the ICU and neurosurgery ward are more infected with A. baumannii and medical staff should pay attention to the standardization of invasive operations.

Although previous studies have shown that A. baumannii can produce wide-spectrum beta-lactamases, which affected the resistance of A. baumannii to β-lactam antibiotics by hydrolysis of beta-lactams [36, 37]. However, as wide-spectrum antibiotics, carbapenem antibiotics are widely used in clinic treatment, causing high drug resistance. In this study, A. baumannii isolates showed high-level resistance to carbapenem antibiotics and high-level sensitivity to tigecycline, polymyxin and minocycline, which was similar to the results from Italy [38] and China [39]. Therefore, carbapenem antibiotics in combination with polymyxin or tigecycline should be considered to in clinical treatment of MDR A. baumannii infection [40]. If there are co-infections, polymyxin will be used to treat patients [41, 42].

Seven STs in the study belonged to known A. baumannii international clones (ICs). The analysis of all publicly available genome sequences in 2019 indicated that ST2, ST1, ST79 and ST25 account for more than 71% of all sequenced genomes, with ST2 by far the most dominant type [43]. The results of MLST typing of 62 A. baumannii isolates showed that ST2 accounted for 83.9% (52/62), which was the dominant ST type in this study. The results are consistent with other related researches in China [44, 45]. The reason why ST2 isolates have become the dominant ST type may be due to its strong biofilm formation ability, high serum resistance, and high pathogenicity [45].

Several studies have proved that the prevalence of A. baumannii is associated with blaADC-25 [12, 44], blaOXA-23 and blaOXA-66 genes [46]. This has also been proved in our study, which exhibited all A. baumannii isolates have blaADC-25 gene, which are dominant in cephalosporins and carbapenems resistance in A. baumannii ST2 isolates. Insertion sequences (ISs) was responsible for the overexpression of the chromosomally encoded blaADC gene [47]. There are no ISs in the blaADC-25 surrounding environment in the selecting representative isolates per ST, suggesting that the gene is an inherent drug resistance gene located on chromosomes.

The MOB-suite identified contigs of plasmid origin showed both high sensitivity and specificity (95 and 88%, respectively) [27]. Therefore, we used the software for clustering, reconstruction and typing of plasmids from assemblies. Unfortunately, these rep_clusters are not identified to specific known replicators. But our focus is mainly to understand the plasmids sequences for ARGs-carrying contigs.

The phylogenetic analysis of the 62 strains revealed that among the 52 ST2 strains, two distinct monophyletic clades were formed on the phylogenetic tree. Clade1 predominantly harbored the rep_cluster_1172 plasmid along with the aph (3')-Ia resistance gene, whereas clade2 was characterized by the presence of the rep_cluster_1226 plasmid and lacked the aph (3')-Ia resistance gene. This finding illuminated the underlying relationship between different genetic lineages and their associations with specific plasmids as well as antibiotic resistance profiles.

There are some limitations in this study. The number of other types of specimens apart from sputum was very small. The number of specimens such as pleural effusion and urine can be increased in follow-up study. Although we used the cluster codes provided by MOB-cluster for description of plasmids that share significant sequences without the need for defined biomarkers. Only rep_clusters were obtained which cannot be identified to specific known replication factors. We mainly focused on studying resistance gene carrier, and virulence and detailed transmission routes of A. baumannii isolates need to be understand in the future.

To sum up, A. baumannii isolates in this study are resistant to most antibiotics and carries a variety of drug resistance genes, indicating the severe situation of drug resistance. The surveillance of drug resistance and active screening of A. baumannii should be strengthened for curbing the spread of A. baumannii in hospitals. There are few previous studies about phenotypic and molecular characteristics of A. baumannii in Gansu province. Our study provided implications to further explore the main genotypic and phenotypic characteristics of A. baumannii in Gansu province.

Conclusions

Our study reveals genotypic and phenotypic characteristics of A. baumannii isolates from a hospital in Gansu Province, China, offering reference into further exploring the main characteristics of A. baumannii.