Genomic relatedness and dissemination of bla_NDM-5 among Acinetobacter baumannii isolated from hospital environments and clinical specimens in Thailand

Samples

Clinical and environmental A. baumannii isolates were collected from Naresuan University Hospital between December 2020 and April 2021. Naresuan University is a level III hospital with 400 beds located in the lower northern region of Thailand. Hospital environment and clinical isolates were collected from five wards, which were two medical wards, Medicine-man (MED-1) and Medicine-woman (MED-2), and three intensive care units, the ICU Cardio-Vascular-Thoracic Surgery (ICU-1), ICU Surgery (ICU-2), and ICU Medicine (ICU-MED). The sources of the samples included staff contact samples, which included samples collected from stethoscopes (n = 15), charts (n = 15), computers/keyboards (n = 15), nurse station counters (n = 15), medical lab coats (n = 15), restroom door handles (n = 15), telephones (n = 15), and dressing trolleys (n = 15). Patient contact samples were collected from bedrails (n = 15), bedsheets (n = 15), suction tubes (n = 15), patient tables (n = 15), curtains (n = 15), humidifiers (n = 15), intravenous (IV) stands (n = 15), ventilators (n = 15), ventilator monitors (n = 9), water from ventilators (n = 9), suction tubes (n = 9), and AMBU bags (n = 9). Other environmental samples were collected from the air (n = 15), sinks (n = 15), and water from sinks (n = 15). The protocol was approved by the Naresuan University Institutional Biosafety Committee, and the project number was NUIBC MI62-09-42.

Isolation and identification of A. baumannii from hospital environments

The air samples were collected using Leeds Acinetobacter Medium (LAM) (Hi-media, Mumbai, Maharashtra, India) in 9 cm diameter Petri dishes. Petri dishes were exposed for 24 h. The other samples from environmental surfaces were collected using cotton swabs soaked in 0.85% normal saline and then placed in transfer media. The swab samples were enriched in Luria-Bertani broth (LB) (HiMedia, Mumbai, Maharashtra, India) by shaking at 160 rpm at 37 °C for 24 h and then cultured in Leed Acinetobacter Media (LAM) at 37 °C for 24 h. Cultures with pink colonies were selected for further evaluation using Gram’s stain and biochemical tests (catalase, oxidase, TSI, citrate). Molecular identification of the bacterial isolates was confirmed by polymerase chain reaction (PCR) using 16S rRNA, and bla_OXA-51 primers (Table S1).

Determination of antibiotic susceptibility

Antibiotic susceptibility testing was performed according to the disk diffusion method using 12 antibiotics: piperacillin/tazobactam (100 and 10 μg), ceftazidime (30 μg), cefepime (30 μg), cefotaxime (30 μg), ceftriaxone (30 μg), imipenem (10 μg), meropenem (10 μg), gentamicin (10 μg), amikacin (30 μg), tetracycline (30 μg), ciprofloxacin (5 μg), and trimethoprim/sulfamethoxazole (1.25 and 23.75 μg). The plates were incubated at 37 °C for 24 h. The zones of inhibition determined whether the microorganism was susceptible, intermediate, or resistant to each antibiotic according to the Clinical and Laboratory Standards Institute (CLSI) guidelines (2022). All isolates were interpreted as non-drug-resistant A. baumannii (NRAB) and carbapenem-resistant A. baumannii (CRAB). In addition, MDRAB was defined when A. baumannii resistant to three or more antibiotic classes and XDRAB was defined when A. baumannii was resistant to all antimicrobial agents except polymyxins (Magiorakos et al., 2012).

PCR amplification of antibiotic resistance genes and rep-PCR typing

As mentioned earlier, PCR assays to detect bla_OXA-23, bla_OXA-24, bla_OXA-58, and bla_NDM were performed using the primers shown in Table S1. The genomic DNA of each isolate was extracted from the overnight cultures using a PureDirex Genomics DNA Isolation Kit (Bio-Helix, New Taipei City, Taiwan). Rep-PCR was performed by using genomic DNA as a template for PCR amplification with the ERIC-2 primer (Table S1) with the conditions described by Leungtongkam et al. (2018). PCR-banding patterns and rep-PCR types were analyzed and interpreted as previously described.

Whole-genome sequencing and bioinformatics analysis

Eight representative A. baumannii strains from four wards, four from hospital environments (AE17, AE30, AE73, AE106) and four from clinical isolates (AC02, AC09, AC23, and AC59) were analyzed. We selected two A. baumannii strains from each ward that were isolated from the same time frame and showed similar antibiotic susceptibility profiles and ARG patterns. All strains were cultured onto an LB agar plate and incubated overnight at 37 °C. Genomic DNA was extracted using a PureDire Genomics DNA Isolation Kit (Bio-Helix, New Taipei City, Taiwan). The extracted DNA was quantified by a nanodrop (Hercuvan, Cambridge, UK). The purified genomic DNA was used to construct libraries followed by sequencing with the Illumina HiSeq 2500-PE125 platform at Macrogen, Korea. The nucleotide sequences of the eight A. baumannii strains have been deposited in NCBI’s database under Sequence Read Archive (SRA) with Bioproject PRJNA862456. The genome of A. baumannii ATCC17978 (CP000521) was used as a reference strain for comparison with the eight A. baumannii strains.

Genome assembly and annotation

Raw sequencing reads were trimmed by using Trim Galore v0.6.7 with default settings and by using Unicycler v0.4.8 with default parameters prior to assembly (Krueger, 2012; Wick et al., 2017). The assembled contigs that were larger than 300 bp in length were selected and subjected to further bioinformatic analysis. The remaining contigs were annotated by using Prokka v1.14.6 with default options (Seemann, 2014).

Identification of MLST, antimicrobial resistance, and virulence genes

The remaining contigs were subjected to detection of drug-resistance and virulence genes by using Abricate v1.0.1 with default settings (Seemann, 2016) against the comprehensive antibiotic resistance database (CARD) and virulence factor database (VFDB) (Alcock et al., 2020; Liu et al., 2022). Multilocus sequence typing (MLST) were performed by using MLST v2.0, which is accessible from the Center for Genomic Epidemiology (www.genomicepidemiology.org). The gene arrangement analysis of bla_NDM-5 was performed using Easyfig version 2.1 (Sullivan, Petty & Beatson, 2011).

Phylogenomic relationships

The selected genomes of eight A. baumannii were subjected to Roary v3.13.0 with the default parameters to identify pan- and core genes (Page et al., 2015). The resultant core genes among the eight genomes were concatenated prior to the construction of a pangenome tree in the CSI phylogeny, which is accessible from the Center for Genomic Epidemiology (www.genomicepidemiology.org) (Kaas et al., 2014). A core-genome tree was constructed based on the presence/absence of identified core-genes and visualized in FigTree v1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/). The SNP count matrix of all selected genomes was calculated in snp-dists v0.6.3 with default settings (Seemann, 2019).

Statistical analysis

Statistical analysis were performed using Stata (Stata 12.0 Corporation). The comparisons of the proportions of antibiotic resistance between A. baumannii obtained from the two different origins were analyzed by using the Z-test. The comparisons of antibiotic resistance among A. baumannii collected from the five hospital wards were analyzed by using the chi-square test. P values < 0.05 were considered to be a statistically significant difference.

A. baumannii strains isolated from the hospital environments and clinical isolates

A total of 106 A. baumannii isolates were obtained from 312 hospital environmental samples (33.97%). The isolates associated with patient contact from AMBU bags, bedrails, suction tubes, water from ventilators, bedsheets, and IV stands were found in 77.9%, 66.7%, 66.7%, 55.6%, 53.3%, and 13.3% of the samples, respectively. We also found 33.3% of the samples from patient tables, humidifiers, ventilators, and curtains were A. baumannii colonization (Table S2). The isolates associated with staff contact and other environments from the air, keyboards, counters, medical lab coats, and dressing trolleys were found in 60.0%, 53.3%, 46.7%, 42.9%, and 33.3% of the samples, respectively (Table S2). The colonization rate of the samples from charts and stethoscopes was 26.7%, while 6.7% of the samples from restroom door handles, and telephones were A. baumannii colonization. However, we did not find A. baumannii isolates on sinks, water from sinks, or ventilator monitors (Table S2). Of the 312 environmental samples collected from each ward, we found the highest A. baumannii contamination in the samples obtained from ICU Surgery, with a rate of 52.9% (36/38), followed by those obtained from the Medicine-woman (40.7%; 22/54), ICU Medicine (38.2%; 26/68), Medicine-man (27.8%; 5/54), and ICU Cardiovascular-Thoracic Surgery (10.3%; 7/68) wards (Table S2).

During the investigation of the prevalence of A. baumannii isolates from the hospital environments of various wards, we found the highest rate of A. baumannii in the ICU Surgery ward (33.9%), followed by the ICU Medicine (24.5%), Medicine-woman (20.8%), Medicine-man (14.2%), and ICU Cardio-Vascular-Thoracic surgery (6.6%) wards (Table 1). Sixty-one A. baumannii isolates were collected from patient specimens. A. baumannii isolates were found in the patient specimens collected from the ICU Medicine (24.6%), Medicine-man (24.6%), ICU Surgery (19.7%), Medicine-woman (16.4%), and ICU Cardio-Vascular-Thoracic surgery (14.8%) wards (Table 1).

Antibiotic susceptibility patterns of A. baumannii isolates

All A. baumannii isolates were subjected to antimicrobial susceptibility testing, and the results are shown in Table 2. A. baumannii isolates from hospital environments were highly resistant to meropenem (100%), cefotaxime (100%), ceftazidime (100%), and ceftriaxone (100%), while the A. baumannii clinical isolates were highly resistant to ceftazidime (90.2%) and ceftriaxone (90.2%). NRAB was detected in only 16.39% of A. baumannii clinical isolates. A high prevalence of MDRAB and CRAB was detected in A. baumannii isolated from hospital environment (ABHE) (93.4% and 100%) and clinical isolates (82.0% and 92.0%) with p value < 0.05, as shown in Table 3. The prevalence of XDRAB in A. baumannii isolates from hospital environments and clinical isolates was 44.7% and 55.7%, respectively (Table 3). Among the five wards, a high prevalence of XDRAB was detected in A. baumannii isolates from ICU Surgery (Table 4).

Antibiotic resistance genes and rep-PCR typing

The 16S rRNA gene was detected in all A. baumannii isolates. The intrinsic bla_OXA-51 gene was detected in all ABHE and 96.7% (59/61) of clinical isolates. The oxacillinase gene, bla_OXA-23 was the most frequently detected gene at 80.20% (85/106) in ABHE and 80.33% (49/61) in clinical isolates (Table 3). The bla_OXA-58 gene was detected in one ABHE (0.94%) and one clinical isolate (1.64%). The bla_NDM gene was detected in 78.3% (83/106) of ABHE (p value < 0.05) compared to 55.74% (34/61) of clinical isolates. The bla_OXA-24 gene was not detected in any of the isolates. Among the five wards, a high prevalence of bla_OXA-23 was detected in ICU Cardio-Vascular-Thoracic Surgery, and a high prevalence of bla_NDM was detected in ICU Surgery (p value < 0.05) (Table 4).

Rep-PCR typing was performed, and fingerprinting represented 33 different DNA patterns consisting of amplicon sizes ranging from 500 to 4,000 bp. The genotypes were named T1 to T33. The major genotype of ABHE was T30 at 21.7% (23/106), followed by T23 at 17% (18/106) and T2 at 15% (15/106). The major genotype of the A. baumannii clinical isolates was T4 at 34.4% (21/61), followed by T23 at 29.5% (18/61).

Heatmaps representing the antibiotic susceptibility patterns, antimicrobial resistance genes, and rep-PCR typing from the five wards is shown in Figs. S1–S5. We found genetic similarity between ABHE and A. baumannii clinical isolates in each ward with antibiotic susceptibility patterns and antimicrobial resistance genes since most A. baumannii strains in the same ward showed similar profiles. No association was found between rep-PCR typing of ABHE and A. baumannii clinical isolates (Figs. S1–S5). Eight strains of A. baumannii with similar profiles from four wards were selected for genome sequencing.

Comparative genomic and phylogenomic analysis of A. baumannii from hospital environmental and clinical isolates

Eight strains of A. baumannii from clinical and environmental isolates were analyzed and compared with the genome of A. baumannii ATCC17978. The four ABHE were AE17 (patient table), AE30 (bedrail), AE73 (dressing trolley), and AE106 (AMBU bag). The four clinical isolates were AC02 (blood hemoculture), AC09 (sputum), AC23 (sputum), and AC59 (right hepatic drain). AC02 and AE03 were obtained from the Medicine-man ward. AC59 and AE17 were obtained from the Medicine-woman ward. AC09 and AE106 were derived from the ICU Cardio-Vascular-Thoracic Surgery ward. AC23 and AE73 were derived from the ICU Surgery ward. The genome characterization of the isolates is summarized in Table 5. The genome analysis revealed that AC02, AE30, AC09, AE106, AC23 and AE73 belong to ST2 based on the Pasteur MLST scheme. However, AC59 and AE17 belong to ST164. The predicted genome sizes of the eight A. baumannii strains ranged from 3.78 to 4.01 Mb compared to the genome of ATCC17978, which was 3.97 Mb.

ARGs and virulence genes of eight A. baumannii strains showed genetic similarity among A. baumannii hospital environments and clinical isolates but were slightly different from the genome of ATCC17978 (Figs. 1A and 1B). The ARGs detected in all eight A. baumannii strains as well as ATCC 17978 encoded macrolide resistance genes (amvA) and a number of genes encoding efflux pumps involved in resistance in glycylcycline/tetracycline (adeR, adeS, adeA, adeB), fluoroquinolone/tetracycline (adeF, adeG, adeH, adeL), fluoroquinolone (abaQ, abeM), fosfomycin (abaF), and multidrug resistance (adeI, adeJ, adeK, adeN, abeS). We identified 23 ARGs present in only some A. baumannii strains, which encoded the efflux pump (adeC) and genes involved in resistance to tetracycline (tet(39), tetB), cephalosporins (bla_ADC-10, bla_ADC-6, bla_ADC-73, bla_ADC-79, bla_TEM-1, bla_TEM-12), carbapenems (bla_OXA-23, bla_OXA-66, bla_OXA-91, bla_OXA-259), macrolide (mphE, msrE), aminoglycoside (aadA5, armA, aph(3′′)-Ib, aph(6)-Id), sulfonamide (sul1, sul2), and integron-encoded dihydrofolate reductase (dfrA17).

A class B β-lactamase gene, bla_NDM-5, that hydrolyzes virtually all β-lactam antibiotics, including carbapenems, was detected in six strains except ATCC17978, AE17 and AC59 (Figs. 1A and 1B). Genetic contexts of bla_NDM-5 revealed mobile genetic elements (MGEs), such as integron1 (intl1), IS91 family transposase, and transposase (ISAba125), along with other AGRs, ant(3″)-Ia, qacEΔ1, and sul1, located upstream and downstream of bla_NDM-5 (Fig. 1C).

Analysis of the virulence genes of eight A. baumannii strains and ATCC17978 revealed that the genes were involved in biofilm formation (adeF, adeG, adeH, bap, csuA/B, csuA, csuB, csuC, csuD, csuE, pgaA, pgaB, pgaC, pgaD), enzyme phospholipase (plcC, plcD), immune evasion (lpsB, lpxA, lpxB, lpxD, lpxL, lpxM), iron uptake (barA, barB, basA, basB, basC, basD, basF, basG, basI, basJ, bauA, bauB, bauC, bauD, bauE, bauF, entE), gene regulation (abal, abaR, bfmR, bfmS), serum resistance (pbpG), and host cell adherence (ompA) (Fig. 1B). The genes involved in capsule polysaccharide synthesis (weoB) and the gene encoding glycosyltransferase in lipopolysaccharide (LPS) biosynthesis (lpsB) were detected in only one strain, ATCC 17978 and AC09 (Fig. 1B).

The phylogenomic relationship of the core and pan genomes of eight A. baumannii and ATCC17978 strains shown in Figs. 2A and 2B revealed three major clades. The A. baumannii strains obtained from the ICU-1, ICU-2, and Med-1 wards were in the same clade, while the A. baumannii strains obtained from the Med-2 ward were in different clades. The genome of ATCC17978 showed different clades from all eight A. baumannii strains. The SNP count matrix of all selected genomes confirmed that the high number of SNPs of AC59 and AE17 derived from the Med-2 ward were comparable with other A. baumannii strains (Fig. 2C).