Comparative pathogenomics and in silico analysis of energy metabolism in Acinetobacter baumannii ST195 and novel Turkish isolates encoding blaOXA-23, blaOXA-66, and blaOXA-852

Abstract

Acinetobacter baumannii is a significant hospital-acquired pathogen recognized for its antibiotic resistance and environmental durability. The present study investigates the genomic and metabolic characteristics of Turkish A. baumannii isolates (ST195 and novel sequence type) using whole-genome sequencing, comparative pathogenomics, and phenotypic assays. Genomic analyses demonstrated significant horizontal gene transfer, phage integration (Salmon_SSU5, Acinet_Bphi_B1251), and genomic islands enriched with resistance and virulence genes. ST195 (T3) exhibited meropenem susceptibility (MIC ≤ 0.125 µg/mL) despite harboring blaOXA-23, linked to adeN efflux regulator loss. In contrast, colistin resistance in T3 was correlated with putative lpxA/C/D mutations causing LPS deficiency. Virulence profiling identified conserved systems for adherence (OmpA), biofilm formation (bap, csuABCDE), and iron acquisition (acinetobactin), while capsule heterogeneity appeared to affect immune evasion. Metabolic reconstruction highlighted nitrogen and sulfur assimilation, as well as ethanol catabolism, which facilitate survival under host stress. Resistome analysis linked blaOXA-852, adeFGH, and armA to resistance against carbapenems and aminoglycosides, while transposase-mediated frameshift mutations accounted for amikacin susceptibility in T3 despite the presence of APH(3’)-VIa. The open pangenome (6,402 genes, 41.9% core) reflected adaptive genomic plasticity. This study highlights the importance of genomic diversity, metabolic flexibility, and regulatory mutations in influencing A. baumannii resistance and virulence. It also identifies potential metabolic and virulence-related features that may guide future therapeutic and anti-virulence strategies.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10123-026-00828-x.

Introduction

Acinetobacter baumannii is an opportunistic, aerobic, non-motile, and catalase-negative, gram-negative bacterium that has emerged as a significant pathogen, particularly in hospital environments (Baumann 1968; Baumann et al. 1968; Dijkshoorn et al. 2007). It is recognized for its ability to survive under various conditions, including on both dry and damp hospital surfaces, making it a persistent and challenging pathogen in nosocomial infections (Lee et al. 2017; Morris et al. 2019). A. baumannii can affect a wide range of anatomical sites, such as the respiratory tract, skin, soft tissues, and urinary tract, with respiratory infections, particularly in ventilator-dependent patients, being especially common (Dijkshoorn et al. 2007; Ren and Palmer, 2023). The high prevalence of infections caused by this bacterium and its ability to adapt and persist in hospital environments make it an important target for medical research (Lee et al. 2017).

A critical aspect of A. baumannii pathogenicity is its remarkable ability to develop resistance to a broad spectrum of antibiotics. Recent trends indicate a rapid increase in pan-drug-resistant (PDR) and extensively drug-resistant (XDR) strains, which are resistant to all antibiotic classes, thus complicating treatment options (Kyriakidis et al. 2021). These strains often exhibit multidrug resistance (MDR) patterns, further complicating infection control (Blair et al. 2015). The rise of MDR, XDR, and PDR A. baumannii strains is a growing global concern in clinical environments, highlighting the need for in-depth studies to understand their genetic basis, virulence mechanisms, and metabolic flexibility, which enable them to thrive under extreme conditions (Sun et al. 2016; Moubareck and Halat 2020).

Genomic plasticity, through horizontal gene transfer (HGT), plays a key role in A. baumannii’s rapid adaptation and acquisition of resistance traits (Karampatakis et al. 2024). This flexibility enables the bacterium to survive in diverse environments, from hospital surfaces to the human microbiota (Guimaraes et al. 2015; Sonnenberg et al. 2020). Recent research has shown that A. baumannii can utilize various carbon and nitrogen sources (including sugars, organic acids, amino acids, and ethanol), demonstrating its metabolic versatility (Ren and Palmer, 2023). Notably, the use of nitrate and nitrite under anaerobic conditions offers a unique survival advantage, especially in oxygen-limited environments such as biofilms and infected tissues (Farrugia et al. 2013; Nemec et al. 2016; Dekic et al. 2019).

This study investigates the genomic characteristics, energy metabolism pathways, and virulence factors of Turkish A. baumannii isolates, with a focus on sequence type ST195 and novel sequence types. Through whole-genome sequencing and comprehensive bioinformatic analysis, we aim to elucidate the genetic diversity, antibiotic resistance profiles, and virulence mechanisms that contribute to the pathogenicity and survival of these multidrug-resistant strains. Our findings provide insights into the genomic structure and functional aspects of A. baumannii, emphasizing the potential of metabolic pathways as targets for novel therapeutic strategies. Understanding the A. baumannii’s energy metabolism, especially its use of alternative electron acceptors and diverse carbon sources, may reveal potential targets for combating infections caused by this highly adaptable and resilient pathogen.

Materials and methods

Bacterial isolates

This study examined 26 clinical isolates of Acinetobacter baumannii collected from the central laboratory at Erciyes University Hospital. These samples originated from diverse sources, including urine, burn wounds, diabetic foot ulcers, blood, pleural fluid, sputum, vascular catheters, bronchoalveolar lavage, and wound swabs (Data not shown). The isolates were initially identified by partial rpoB gene sequencing before proceeding with MLST characterization.

Preliminary identification of A. baumannii isolates by rpoB gene sequencing

Preliminary identification of the Turkish isolates as Acinetobacter baumannii was performed by partial rpoB gene sequencing. Compared with 16 S rRNA gene analysis, rpoB sequencing provides greater discriminatory power for differentiating members of the Acinetobacter calcoaceticus–baumannii complex and is considered a reliable approach for species-level identification of clinically important Acinetobacter isolates (La Scola et al. 2006; Gundi et al. 2009). Genomic DNA was extracted from pure overnight cultures as described in the next section. A partial fragment of the rpoB gene was amplified by PCR using the primer pair Ac696F (5′-TAYCGYAAAGAYTTGAAAGAAG-3′) and Ac1093R (5′-CMACACCYTTGTTMCCRTGA-3′), as previously described for molecular identification of Acinetobacter species (La Scola et al. 2006). Each PCR reaction was carried out in a final volume of 25 µL containing 1× PCR buffer, 1.5–2.5 mM MgCl₂, 200 µM of each deoxynucleoside triphosphate, 0.4 µM of each primer, 1 U EasyTaq DNA polymerase (TransGen Biotech, Beijing, China), and approximately 20–50 ng of template DNA. Amplification was performed using thermal cycler (ABI Veriti 96, Thermo-Fisher Scientific, USA) under the following conditions: initial denaturation at 94 °C for 5 min; 35 cycles of denaturation at 94 °C for 30 s, annealing at 54 °C for 30 s, and extension at 72 °C for 45 s; followed by a final extension at 72 °C for 5 min. The PCR products were verified by electrophoresis on 1.5% agarose gels and visualized under UV illumination after staining with an appropriate nucleic acid dye. Amplicons of the expected size were purified and bidirectionally Sanger-sequenced (Marcogen-Europe, Netherlands) using the same primers. The resulting partial rpoB sequences were compared with validated reference sequences available in GenBank using BLASTn. Isolates showing highest sequence similarity to A. baumannii reference strains in the informative rpoB region were considered preliminarily identified as A. baumannii. This sequence-based approach was used as an initial identification step prior to MLST characterization and whole-genome-based downstream analyses, consistent with previous studies validating partial rpoB sequencing for identification of clinically relevant and emerging Acinetobacter species (La Scola et al. 2006; Gundi et al. 2009).

Growth conditions and DNA extraction, and strain selection

Isolates were cultured on MacConkey agar (BD BBL, Franklin Lakes, New Jersey, USA) for 18–24 h. Colony purification was performed twice for all isolates. Isolates were cryopreserved at − 80 °C in tryptic soy broth supplemented with 15% glycerol. Genomic DNA was extracted from a single loopful of harvested bacterial colonies using the PureLink Genomic DNA Mini Kit (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA), following the manufacturer’s protocol for Gram-negative bacteria. A. baumannii strains selected for this study were characterized by multilocus sequence typing (MLST), using a panel of seven housekeeping genes (gltA, gyrB, gdhB, recA, cpn60, gpi, rpoD) recommended by the Institut Pasteur (Jolley et al. 2018). Whole-genome sequencing was performed on distinct or unknown MLST types (T3, T14, and T24) identified by the MLST results.

Determination of antimicrobial susceptibility

Antimicrobial susceptibility testing for T3, T14, and T24 strains was performed using an automated system (VITEK 2, Vitek Systems, BioMérieux, Marcy-l’Étoile, France). Data analysis and interpretation were conducted in accordance with the Clinical and Laboratory Standards Institute (CLSI) guidelines (CLSI, 2020). Antibiotic susceptibility was assessed using a ready-to-use test panel (BioMérieux, Marcy l’Étoile, France), which included the following antibiotics: amikacin, amoxicillin/clavulanic acid, ampicillin/sulbactam, ampicillin, cefazolin, cefepime, ceftazidime, ceftolozane-tazobactam, ciprofloxacin, ceftriaxone, cefuroxime, gentamicin, imipenem, meropenem, levofloxacin, tigecycline, trimethoprim/sulfamethoxazole, piperacillin/tazobactam, colistin, ertapenem, fosfomycin with G6P, nitrofurantoin, and tobramycin.

Whole genome sequencing and assembly

Following the DNA extraction protocol, the purity and concentration of genomic DNA was evaluated using a Qubit 3.0 fluorometer (Invitrogen, Thermo-Fisher Scientific) and a 1.5% agarose gel. The genome sequencing libraries were prepared using the Nextera XT DNA Library Preparation Kit (Illumina, San Diego, CA, USA). Preparation Kit (Illumina, San Diego, California, USA). The sequencing procedure was performed on the Illumina NovaSeq platform using paired-end (PE) 2 × 150 base reads (Yetiman 2025). Low-quality reads were filtered and assembled using the SPAdes assembler within the BV-BRC 3.40.2 genome assembly service.

Bioinformatic analyses

Genome annotations were performed using the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) and the BV-BRC platform (v.3.31.12) and compared with each other (Tatusova et al. 2016; Olson et al. 2023). BRIG v.0.95 was used to perform BLAST ring alignment and phage mapping on the genomes analyzed in this study (Alikhan et al. 2011). Average nucleotide identity (ANI) values for the A. baumannii genomes were calculated using the FastANI tool (Jain et al. 2018). Pangenome analysis was conducted with the PanACoTA program via the PanExplorer web server (Perrin and Rocha, 2021; Dereeper et al. 2022). For pangenome analysis and functional annotation, well-known A. baumannii reference genomes from NCBI were utilized, which can be sorted as follows: Ab04-mff (CP012006.1), ab736 (CP015121.1), AR_0063 (NZ_CP026711.1), ATCC 17,978 (CP018664.1), AYE (CU459141.1), K09-14 (CP043953.1), and KAB01 (CP017642.1). The genome-wide functional annotation of each A. baumannii genome was performed using the EggNOG-Mapper (v.5.0) web server with the Prodigal prediction method (Huerta-Cepas et al. 2017). Resistome screening was performed by comparing the complete genome sequences of T3, T14, and T24 genomes against the CARD database-RGI 6.0.5 (Alcock et al. 2023). The analysis of virulence factors in the target genomes was conducted using the Virulence Factor Database (VFDB-2022) (Liu et al. 2022). The BacWGSTdb 2.0 (http://bacdb.cn/BacWGSTdb/) and pubMLST web servers were used for 7-gene multi-locus sequence typing (MLST), core genome MLST, and single-nucleotide polymorphism (SNP) comparisons of the studied genomes (Jolley et al. 2018; Feng et al. 2021). Prophage regions in the target genomes were detected and annotated using the PHASTER (Phage Search Tool Enhanced Release) platform (Arndt et al. 2016; Wishart et al. 2023). Prediction and comparison of genomic islands were performed via the online application Island Compare v.1.0 (Bertelli et al. 2019, 2022). Pathogenicity islands of A. baumanii strains T3, T14, and T24 were predicted using the GIPSy genomic island search tool (Soares et al. 2016). The RStudio environment was used for data visualization, including the generation of heatmaps, bar plots, and UpSet plots (R-Core-Team 2021). BlastKOALA and RASTtk were employed to predict energy metabolism pathways in Turkish A. baumanii isolates (Brettin et al. 2015; Kanehisa et al. 2016). The complete genome sequences and plasmids of strains T3, T14, and T24 were deposited in NCBI under the accession numbers CP171394-CP171395, CP171396-CP171397, and CP171398-CP171399, respectively. Additionally, to investigate genotype–phenotype discrepancies, resistance-associated loci were analyzed through comparative sequence analysis against reference gene sequences and published resistance determinants using the T-coffee web server (https://tcoffee.crg.eu/). Genes of interest were examined for deletions, insertion sequence interruptions, frameshifts, premature stop codons, and amino acid substitutions that could be linked to changes in antimicrobial susceptibility. Functional interpretations were based on predicted gene disruptions and literature-supported genotype–phenotype relationships. No complementation assays, expression analyses, or other functional validation experiments were conducted in this study.

Results and discussion

Taxogenomics of Turkish A. baumannii isolates

First, we performed next-generation sequencing, draft genome assembly, and annotation for all three strains to assess their genetic diversity, determine their phylogenetic relationships, and identify the genetic basis underlying their varied antibiotic resistance profiles. All assembled genomes had a CheckM completeness score of 100%, indicating high genome quality. The other genome quality metrics and descriptive features are shown in Table 1. The GC content of the sequenced A. baumanni genomes ranged from 38.94% to 39.03%. Moreover, each of these strains harbored a single plasmid. The Fig. 1 illustrates the BLAST ring alignment of the A. baumannii T3, T14, and T24 genomes compared to other reference A. baumannii genomes (Table 2). Genomic differences among them are observable due to sequence variation. At first glance, genomic islands might be identified by decreases in GC content and by their presence or absence in other strains, as indicated by gaps in the BLAST ring alignment (Juhas et al. 2009). These regions often contain integrases, insertion sequences, and transposases, which are characteristic of horizontally acquired elements (Juhas et al. 2009; Bertelli et al. 2022).

Table 1.

Summary and quality metrics for the genomes and plasmids of Acinetobacter baumannii T3, T14, and T24

Feature	T3	T14	T24
Accession	CP171394	CP171396	CP171398
Length (bp)	4,042,586	3,856,874	3,856,788
GC content (%)	38,94	39,02	39,03
Genes (total)	4,063	3,795	3,800
CDSs (total)	3,992	3,723	3,728
CDSs (with protein)	3,876	3,631	3,631
Genes (RNA)	71	72	72
rRNAs (5 S, 16 S, 23 S)	1, 1, 1	1, 1, 1	1, 1, 1
tRNAs	64	65	65
ncRNAs	4	4	4
Pseudo Genes (total)	116	92	97
Pseudo Genes (ambiguous residues)	0 of 116	0 of 92	0 of 97
Pseudo Genes (frameshifted)	17 of 116	29 of 92	30 of 97
Pseudo Genes (incomplete)	86 of 116	54 of 92	59 of 97
Pseudo Genes (internal stop)	16 of 116	19 of 92	19 of 97
Pseudo Genes (multiple problems)	3 of 116	10 of 92	11 of 97
CRISPR Arrays	-	2	2
Hypothetical CDS Ratio	0,35340247	0,31627908	0,3155446
Genome Quality Metrics
Average short read coverage	131.546	130.894	130.727
N50	162,174	463,263	348,364
L50	10	3	4
Coarse Consistency	99,5	99,4	99,4
Fine Consistency	98,7	98,9	98,9
CheckM Completeness	100	100	100
Plasmids	1	1	1
Accession (Plasmid)	CP171395	CP171397	CP171399
Plasmid length (bp)	78,284	114,671	114,671
GC content (%)	33,55	41,51	41,51
CDSs (total)	114	132	132
Hypothetical CDS Ratio	0,7719298	0,77272725	0,780303
Plasmid Quality Metrics
Average short read coverage	133.974	131.917	134.362
N50	46,781	114,671	114,671
L50	1	1	1
Coarse Consistency	93.8	93.8	93.8
Fine Consistency	93.8	93.8	93.8
CheckM Completeness	Insufficient length to interpret each plasmid.

Fig. 1.

BLAST ring alignment of Acinetobacter baumannii genomes, including Turkish isolates T3, T14, and T24. The genomes are aligned from the blue ring to the red ring are ordered as follows: T3, T14, T24, KAB01, K09-14, AYE, ATCC 17,978, AR_0063, ab736, and Ab04-mmf. The GC content is depicted in the third inner circle before the T3 genome. The GC skew (+/-) of T3’s genome is also shown in the second inner circle as a reference, and the genome size is displayed in the first inner circle. Meanwhile, the intact prophage regions of the analyzed genomes are indicated by navy blue arcs. Genome codes in brackets on the outer circle specify which phages belong their respective genomes

Table 2.

Genomic features of the reference genomes used in this study were obtained from NCBI

Feature	Ab04-mff	ab736	AR_0063	ATCC 17,978	AYE	K09-14	KAB01
Accession	CP012006.1	CP015121.1	NZ_CP026711.1	CP018664.1	CU459141.1	CP043953.1	CP017642.1
Length (bp)	3,935,688	3,980,886	4,028,761	4,004,792	3.936.291	3,972,439	3,963,217
GC content (%)	38,92	39,16	38,90	39	39,5	39	39
Genes (total)	4,067	3,785	3,953	3,848	3,902	3,760	3,885
CDSs (total)	3,908	3,690	3,858	3,754	3,808	3,666	3,790
rRNAs (5 S, 16 S, 23 S)	6, 6, 6	6, 6, 6	6, 6, 6	6, 6, 6	6, 6, 6	6, 6, 6	6, 6, 6
tRNAs	73	73	72	72	72	72	73
ncRNAs	1	4	5	4	4	4	4
Pseudo Genes (total)	67	19	90	79	112	58	143
Isolation Country	Canada	USA	USA	USA	France	Malaysia	South Korea

Regarding average nucleotide identity (ANI), the T14 and T24 genomes shared 99.974%, while the T3 and KAB01 genomes shared 99.613%. The ANI similarity of the T3 genome with the T14 and T24 genomes was 97.802% and 97.805%, respectively. Compared with other reference genomes (Table 2), these strains exhibited ANI values ranging from 97.5% to 97.9% (Fig. 2). Variations among these genomes may arise from phage and mobile genetic elements, as well as horizontal gene transfer, influenced by their evolutionary trajectories and the environments they inhabit (Yetiman et al. 2024). Furthermore, ANI analysis highlighted genomic relatedness among the analyzed strains, while complementary analyses of genomic islands, prophages, and insertion sequences suggested a possible contribution of horizontal gene transfer to genome diversification (Almeida et al. 2021). Additionally, the genomes analyzed in this study were subjected to 7-gene MLST using the BacWGSTdb 2.0 and pubMLST databases. Seven-gene MLST results indicated that the T3 isolate was ST195, whereas the T14 and T24 isolates were determined to be novel (Table 3). Likewise, whole-genome screening using pubMLST revealed that 85.7% of the loci in the novel types matched with ST755. The SNP distribution matrix of A. baumannii genomes is shown in Fig. 3. The number of single-nucleotide polymorphisms (SNPs) observed among the genomes analyzed ranges from 124 to 47,422. Genomes sharing the same sequence type exhibit fewer SNP variations. The T3 genome differs from the KAB01 by 4382 SNPs, which are the closest ANI-exhibiting A. baumannii strains. The SNP difference between the T3 and KAB01 genomes, demonstrating the highest ANI similarity, was 4382. The other SNP differences observed are consistent with the ANI values reported above.

Fig. 2.

Average Nucleotide Identity (ANI) phylogram of Acinetobacter baumannii strains T3, T14, and T24, and other well-known Acinetobacter baumannii core genomes

Table 3.

Results of 7-gene MLST analysis for Turkish A. baumannii genomes and additional core genomes from BacWGSTdb 2.0

Strain	ST*	gltA	gyrB	gdhB	recA	cpn60	gpi	rpoD
Ab04-mff	447	1	15	13	12	4	106	2
ab736	931	1	10	8	6	1	110	14
AR_0063	205	1	17	8	10	28	99	32
ATCC17978	112	1	12	56	36	1	61	26
AYE	231	10	12	4	11	4	98	5
K09-14	46	31	33	211	40	16	142	7
KAB01	451	1	3	3	2	2	142	3
T14	NA**	1	114	62	54	4	165	72
T24	NA**	1	114	62	54	4	165	72
T3	195	1	3	3	2	2	96	3

ST Sequence type, NA Not available

Fig. 3.

Single nucleotide polymorphism (SNP) distribution matrix of Acinetobacter baumannii genomes examined in the present study

Comparative genomics and functional annotation

In addition to taxogenomic analyses, a pangenome analysis was conducted using all A. baumannii genomes included in this study (Tables 1 and 2). The pangenome consists of an estimated 6,402 genes, comprising 2,053 strain-specific genes (singletons), 1666 dispensable genes, and 2,683 core genes (Fig. 4A). The clustered accessory presence/absence matrix is shown in Fig. 5. The core genome represents 41.9% of the total pangenome and is mainly associated with essential metabolic processes and housekeeping functions (Fig. 6). On top of that, it may harbor genes that distinguish the species from others within the genus (Mosquera-Rendon et al. 2016). The pangenome curve followed a logarithmic trend and did not reach a plateau. The alpha value was calculated as 0.601 (Fig. 4B). According to Heap’s law, as described by Tettelin et al. (2008), the pangenome is classified as open, indicated by an alpha value less than 1 and the observed logarithmic curve trend (Liu et al. 2015). Similarly, Acinetobacter baumannii has been reported to possess an open pangenome, reflecting extensive genomic plasticity and ongoing acquisition of accessory genes through horizontal gene transfer. This dynamic genome structure contributes to its remarkable adaptability, persistence, and diversification, particularly with respect to antimicrobial resistance and virulence-associated traits (Imperi et al. 2011; Touchon et al. 2014; Chan et al. 2015). The distribution of dispensable genes is presented in the Figure. 6. Among the 2,053 strain-specific genes, strain T3 contained 213 unique genes, while T14 and T24 harbored 12 and 10 strain-specific genes, respectively (Supplementary File 1: Tables S1-S3). Additionally, the genomes of KAB01 (122), ATCC17978 (340), AR0063 (326), AYE (281), ab736 (274), K09-14 (191), and Ab04-mff (284) each contain more than one hundred strain-specific genes (Fig. 6). Consistent with previous studies, many of the strain-specific genes identified here encoded hypothetical proteins and transposon- or phage-related elements (Liu et al. 2015; Almeida et al. 2021). However, direct comparison with previously published strains should be made cautiously, since our isolates originated from the same or nearby locations and over a short sampling interval; therefore, observed similarities within our collection and differences relative to other studies may partly reflect temporal and geographic relatedness rather than intrinsic species-wide patterns alone. The substantial number of singleton genes highlights A. baumannii’s propensity to acquire exogenous genetic material and expand its dispensable genome, suggesting a rapid increase in the size of the A. baumannii pan-genome as new genome sequences become available (Imperi et al. 2011).

Fig. 4.

Pangenome metrics of analyzed genomes in this study. (A) Distribution of the core and accessory genomes. (B) Rarefaction curve showing the cumulative count of gene clusters (pangenome) observed with an increasing number of genomes analyzed

Fig. 5.

Clustering of dispensable genes in the pangenome using a phylogram based on their presence/absence

Fig. 6.

Upset plot showing the number of core-genome genes, strain-specific singletons, and other dispensable genes. Furthermore, it displays clusters of orthologous groups (COGs) linked to the core genome

Figure 6 also illustrates the distribution of COG categories within the core genome. The ten most represented COG categories in the core genome, in descending order, are: 354 genes for Amino acid transport and metabolism [E], 268 genes for Replication, recombination, and repair [L], 228 genes for Translation, ribosomal structure, and biogenesis [J], 217 genes for General function prediction only [R], 207 genes for Energy production and conversion [C], 165 genes for Transcription [K], 162 genes for Coenzyme transport and metabolism [H], 158 genes for Lipid transport and metabolism [I], 154 genes for Inorganic ion transport and metabolism [P], and 139 genes for Posttranslational modification, protein turnover, chaperones [O]. The frequency of COG categories across both the core genome and pangenome is shown in Fig. 7. The top ten COG categories for pangenomes can be arranged in ascending order as follows: 339 genes for Lipid transport and metabolism [I], 339 genes for Coenzyme transport and metabolism [H], 388 genes for Cell wall/membrane/envelope biogenesis [M], 401 genes for Inorganic ion transport and metabolism [P], 440 genes for Transcription [K], 471 genes for Translation, ribosomal structure, and biogenesis [J], 474 genes for Energy production and conversion [C], 609 genes for General function prediction only [R], 783 genes for Amino acid transport and metabolism [E], and 1251 genes for Replication, recombination, and repair [L]. On the other hand, the minimum number of genes associated with a COG category was observed in Cell motility [N], with 14 genes in the core genome and 39 genes in the pangenome. The number of Clusters of Orthologous Groups (COGs) for protein-coding sequences is shown in Fig. 8. Variations in gene counts within identical COG categories across different genomes of the same species are often driven by adaptive and defensive responses to environmental factors (Zhaxybayeva et al. 2004; Wiezer and Merkl, 2005; Fuchsman et al. 2017). The locations of genomic islands, groups of genes that are likely of horizontal origin, were identified by comparing the genomes of the strains with those of their closest phylogenetic relatives (Fig. 9).

Fig. 7.

Distribution of COG categories between the core genome (green) and pangenome (blue) in this study

Fig. 8.

Comparison of clusters of orthologous groups (COG) among Acinetobacter baumannii genomes analyzed in the present study

Fig. 9.

Predicted genomic islands from Turkish Acinetobacter baumannii isolates and other core genomes. Grey areas indicate regions of sequence similarity among the aligned genome groups. Genomic islands are colored based on sequence uniformity across the genomes

Phage structures, related horizontal gene transfer, mobilome, and pathogenicity islands

Phage screening identified 15 potential phage regions across the genomes: T3 harbored three intact, one incomplete, and three questionable regions; T14 contained two intact and two incomplete regions; and T24 had two intact and two incomplete regions (Table 4). In the PHASTEST database, putative phage sites are categorized and ranked as intact (score > 90), incomplete (scoring 70–90), and questionable (score < 70) based on the number of matching phage proteins (Wishart et al. 2023). Only intact phages identified in the aforementioned genomes were mapped onto the BLAST ring alignment (Fig. 1). All A. baumannii isolates contain the intact phages Salmon_SSU5_NC_018843 and Acinet_Bphi_B1251_NC_019541 within their genomes. The GC content of Acinet_Bphi_B1251_NC_019541 phages identified in the Turkish A. baumannii genomes ranges from 38.96% to 40.53%. Furthermore, this phage region has been reported in A. baumannii genomes by multiple authors (Leal et al. 2020; Vijayakumar et al. 2022). The GC content of Salmon_SSU5_NC_018843 in these genomes ranges from 42.30% to 42.64%. This phage has previously been identified in Cronobacter sakazakii isolates from infant formula processing facilities and in Pseudomonas strains isolated from coastal waters in Michigan, USA (Batrich et al. 2019; Mousavi et al. 2023). It has been reported that this phage promotes siderophore synthesis and confers resistance to heavy metals such as mercury and copper (Batrich et al. 2019). However, no evidence of these features was found in the phage genomes from the Turkish isolates. Moreover, among the identified intact phages, only the Salmon_SSU5_NC_018843 phage sites contain putative phage lysozyme-encoding genes classified as endolysins within the previously mentioned genomes. Lysozymes, also known as N-acetylmuramidases, are phage-encoded proteins that rapidly degrade the bacterial cell wall, particularly in Gram-positive bacteria, thereby releasing new viral particles (Vachher et al. 2021). These findings suggest that the studied A. baumannii isolates may have acquired lysozyme-encoding phages from Gram-positive bacteria. Among all identified intact phage structures, only the Acinet_Bphi_B1251_NC_019541 phage within the T3 genome, which is 37.5 Kb in size, contains phage integrase. Conversely, it has been observed that all intact Salmon_SSU5_NC_018843 phages contain phage integrases. Phage integrases facilitate recombination between attachment sites on phage and bacterial genomes, namely attL and attR, respectively (Groth et al. 2000). No attachment sites have been detected in any of the identified phage sites. The presence of integrase in a bacterial genome is a reliable indicator of pathogenicity islands, phages, and integrative plasmids (Juhas et al. 2009; Yetiman et al. 2022). Genomic islands can be distinguished from the rest of the bacterial genome using nucleotide statistics such as cumulative GC skew, GC content, codon usage, and tetranucleotide frequencies (Yetiman et al. 2024). In this study, it was also observed that the intact phage regions identified in the BLAST ring alignment are located in or near genomic islands. Additionally, all intact Salmon_SSU5_NC_018843 phages were found to contain phage-encoded thioredoxin A protein (trxA), which acts as a virulence factor with various functions, including the reduction of protein disulfides (Shrihari et al. 2022). The trxA is a small, ubiquitous redox protein essential for reducing protein disulfide bonds and maintaining protein functionality. It plays a role in several cellular processes, including transcription regulation, energy transduction, cell division, and multiple biosynthetic pathways (Kumar et al. 2004). The trxA has been reported as a significant virulence factor in the colonization of the gastrointestinal tract by A. baumannii, contributing to subsequent pneumonic infections that can result in lethal sepsis (Ketter et al. 2018). The trxA promotes the virulence of A. baumannii by modulating bacterial surface hydrophobicity through disulfide bond reduction and regulation of type IV pili (May et al. 2019). Nevertheless, this possibility remains speculative in the present study, since no functional assays were performed to determine the contribution of prophage-associated trxA to the phenotypes of the Turkish isolates. Further genetic and biochemical studies will therefore be required to evaluate whether trxA is a tractable anti-virulence target. Overall, most intact phages were found to contain all packaging, head/tail gene clusters, and DNA packaging genes. Supplemental File 2 (Excel sheets) provides further information regarding phage components. Understanding the structures of prophages in the genomes of A. baumannii strains is important for developing new therapeutic strategies as alternatives to antibiotics.

Table 4.

Summary of predicted phage regions within the genomes of Acinetobacter baumannii T3, T14, and T24

No	Origin	Region	Completeness	Score	Total Proteins	Region Position	Most Common Phage*	GC%
1	T3	32.1Kb	questionable	90	44	2,811,795–2,843,957	Salmon_SJ46_NC_031129(4)	36.00%
2	T3	20Kb	questionable	80	23	3,934,374–3,954,383	Acinet_vB_AbaS_TRS1_NC_031098(11)	37.97%
3	T3	27.9Kb	incomplete	60	14	3,352,592–3,380,543	Acinet_Bphi_B1251_NC_019541(7)	38.18%
4	T3	34Kb	intact	150	53	3,782,165–3,816,252	Acinet_Bphi_B1251_NC_019541(18)	38.96%
5	T3	29.8Kb	questionable	80	25	3,960,241–3,990,138	Acinet_Bphi_B1251_NC_019541(18)	39.07%
6	T3	37.5Kb	intact	140	45	4,004,979–4,042,515	Acinet_Bphi_B1251_NC_019541(17)	40.53%
7	T3	95.5Kb	intact	150	88	2,487,007–2,582,584	Salmon_SSU5_NC_018843(34)	42.30%
8	T24	35.1Kb	incomplete	60	28	825,231–860,383	Acinet_Bphi_B1251_NC_019541(11)	36.98%
9	T24	20.8Kb	incomplete	60	34	2,965,659–2,986,530	Acinet_Bphi_B1251_NC_019541(13)	37.28%
10	T24	29.1Kb	intact	150	46	3,827,281–3,856,470	Acinet_Bphi_B1251_NC_019541(18)	40.19%
11	T24	85.4Kb	intact	150	89	3,371,456–3,456,936	Salmon_SSU5_NC_018843(36)	42.64%
12	T14	35.1Kb	incomplete	60	28	825,231–860,383	Acinet_Bphi_B1251_NC_019541(11)	36.98%
13	T14	20.8Kb	incomplete	60	34	3,077,498–3,098,378	Acinet_Bphi_B1251_NC_019541(13)	37.31%
14	T14	27.9Kb	intact	150	44	3,828,840–3,856,829	Acinet_Bphi_B1251_NC_019541(18)	40.14%
15	T14	85.4Kb	intact	150	89	3,465,094–3,550,574	Salmon_SSU5_NC_018843(36)	42.64%

* The numbers in parentheses refer to the total number of matching proteins in the PHASTEST database

Horizontal gene transfer (HGT) facilitates the transmission of genetic material within a single generation rather than between parent and offspring, serving as a primary mechanism for bacterial adaptation and evolution (Averhoff et al. 2021). HGT mechanisms play a critical role in the dissemination of antimicrobial resistance. The main mechanisms of HGT include conjugation, transduction (via phage infections), natural transformation, and membrane vesicle fusion. All four mechanisms have been previously validated in A. baumannii, particularly regarding the spread of antimicrobial resistance (Da Silva and Domingues, 2016). Protein BLAST results indicate that 2.04% (83 genes) of the T3 genome, 1.11% (42 genes) of the T14 genome, and 1.13% (43 genes) of the T24 genome have been acquired through phage-associated horizontal gene transfer from various bacteria, including the Acinetobacter calcoaceticus/baumannii complex, Acinetobacter sp., Pseudomonadota, Gammaproteobacteria, Klebsiella pneumoniae, Escherichia marmotae, and Acinetobacter pittii (Supplemental File 2). Among the elements identified in all phage sites, phage-encoded transposases (T3; PP_02315, PP_03575, PP_03781, PP_03782, PP_03837, PP_03844, PP_03853, PP_03861/T14; PP_00786, PP_02848/T24; PP_00785, PP_02726) were also observed, potentially acquired through recombination, replication, or repair processes (Yetiman et al. 2024). Finally, a whole-genome-based transposase search using IS Finder identified transposases from the IS256, IS3, IS30, IS4, IS5, IS6, IS66, ISCR, ISL3, and Tn3 families in the T3 genome (E < 0.05). Conversely, only transposases from the IS3 and IS5 families were identified in the T14 and T24 genomes (E < 0.05). Moreover, based on ISFinder screening results, the majority of these transposases were detected to originate from sources other than A. baumannii. The identified IS elements are summarized in Supplemental File 2.

Pathogenicity islands (PIs) in the T3, T14, and T24 genomes were predicted using the GIPSy bioinformatic tool, which analyzes deviations in G + C content and codon usage, together with the presence of virulence factors, hypothetical proteins, and transposon-associated elements related to the mobilome (Soares et al. 2016). KAB01 was used as the reference genome for pathogenicity island screening because it showed the highest ANI similarity to A. baumannii T3, T14, and T24. PIs with strong or moderate prediction scores were selected from each Turkish A. baumannii genome for comparative analysis. All identified PIs are summarized in Table S4 (Supplemental File 1). The identified genomic islands demonstrate significant variation in G + C content and codon usage patterns, suggesting potential acquisition from external sources via horizontal gene transfer (Juhas et al. 2009). Several islands harbor genes that encode virulence factors, antibiotic resistance elements, and mobile genetic elements, indicating their possible involvement in pathogenicity and adaptation (Schmidt and Hensel, 2004). Strong prediction scores were observed for the PIs identified in strains T3, T14, T24, and KAB01, supporting their classification as robust candidate pathogenicity islands, although their specific functional roles would require further experimental validation. The presence of hypothetical proteins within these genomic islands indicates that these regions harbor uncharacterized genetic elements whose roles may involve diverse biological processes and require further functional investigation. The T3, T14, and T24 genomes differed from their neighbors in both the quantity and distribution of these genomic islands. These findings suggest that genomic islands play a significant role in shaping the unique functional and adaptive characteristics of the Turkish isolates.

Energy metabolism

Acinetobacter species are completely aerobic and non-fermentative (Juni 1978). Previous studies indicate that A. baumannii can survive for at least 28 days in the absence of oxygen, implying mechanisms for persistence in anaerobic environments (Dekic et al. 2019). Additionally, earlier studies showed that strong oxygenation, a lower pH of 5.5, and the use of nitrate as a nitrogen source promoted the enrichment of Acinetobacter species (Baumann 1968). Moreover, Acinetobacter species can utilize ammonium, nitrite, and nitrate as inorganic nitrogen sources (Baumann et al. 1968). Most Acinetobacter species are prototrophs, capable of growth in minimal medium without requiring exogenous nutrients such as amino acids, vitamins, or nucleotides (Warskow and Juni, 1972). These strains also metabolize a diverse range of carbon sources, including sugars, organic acids, amino acids, and ethanol. However, there are significant variations in the specific carbon preference (Farrugia et al. 2013; Nemec et al. 2016).

This study examined the energy metabolism of the Turkish isolates (T3, T14, and T24) to clarify their energy pathways (Fig. 10). Initially, the genomes of the Turkish isolates were found to contain a fructose-specific phosphotransferase system with only fruA/B and fruB/ptsH genes. Additionally, KEGG analysis confirmed the presence of several ABC transporters for sulfate/thiosulfate, taurine, alkanesulfonate, molybdate, iron (III), phospholipids, phosphate, lysine/arginine/ornithine, histidine, glutamate/aspartate, D-methionine, zinc, lipoproteins, and lipopolysaccharides. The limited use of carbohydrates as carbon sources suggests that these isolates primarily generate energy by degrading amino acids into pyruvate, succinyl-CoA, or acetyl-CoA. Through beta-oxidation, fatty acids can be broken down into acetyl-CoA, which is then sent into the TCA cycle to produce energy. Besides, this can assist in combating polyunsaturated fatty acids (PUFAs) toxicity, which is generally released by immune cells like neutrophils during the inflammatory bursts (Ren and Palmer, 2023). Furthermore, glycolysis bypasses the initial glucose-related enzyme step and converts fructose to fructose-1P, allowing the subsequent steps of glycolysis to proceed normally. This may be of particular concern for individuals with diabetes, who have been reported to exhibit elevated fructose levels in both blood and urine (Kawasaki et al. 2002). KEGG annotation also revealed that acetate, acetaldehyde, and ethanol can be converted into pyruvate to provide energy when needed. The absence of a dedicated pentose sugar transporter, along with the non-production of acids from pentoses, suggests that glyceraldehyde-3P from glycolysis can be utilized in the pentose phosphate pathway for purine and pyrimidine synthesis when required (Yetiman and Ortakci 2023). Based on KEGG analysis results, several components of the electron transport chain (ETC) were identified, including succinate dehydrogenase (sdhABCD), NADH-quinone oxidoreductase (nuoABCDEFGHIJKLMN), cytochrome bd ubiquinol oxidase (cydABX), cytochrome o ubiquinol oxidase (cyoABCDE), NADH: quinone reductase (non-electrogenic, ndh), F-type H+-transporting ATPase (atpABCDEFGH), polyphosphate kinase (ppk1), and inorganic pyrophosphatase (ppa).

Fig. 10.

Putative summary of energy metabolism in Turkish Acinetobacter baumanni isolates, based on KEGG prediction via BlastKOALA. The numbers indicate the enzymes responsible for the metabolic pathways as follows: 1: fructose PTS system (fruA/B), 2: 1-phosphofructokinase (fruK), 3: fructose-bisphosphate aldolase, class I (ALDO), 4: glyceraldehyde 3-phosphate dehydrogenase (gapA), 5: phosphoglycerate kinase (pgk), 6: 2,3-bisphosphoglycerate-independent phosphoglycerate mutase (gpmI), 7: enolase (eno), 8: pyruvate, water dikinase (ppsA), 9: phosphate acetyltransferase (pta), 10: acetyl-CoA synthetase (acs), 11: acetate kinase (ackA), 12: fructose-bisphosphate aldolase, class I (ALDO), 13: fructose-1,6-bisphosphatase I (fbp), 14: transketolase (tktA/B), 15: ribulose-phosphate 3-epimerase (rpe), 16: ribose 5-phosphate isomerase A (rpiA), 17: transketolase (tktA/B), 18: transaldolase (TALDO1), 19: transketolase (tktA/B), 20: ribose-phosphate pyrophosphokinase (prsA), 21: citrate synthase(gltA)/2-methylcitrate synthase(prpC), 22: aconitate hydratase(ACO), 23: isocitrate dehydrogenase(icd), 24: 2-oxoglutarate dehydrogenase(sucA/B), 25: succinyl-CoA: acetate CoA-transferase (aarC)/succinyl-CoA synthetase (sucD), 26: succinate dehydrogenase (ubiquinone) flavoprotein subunit (SDHA), 27: fumarate hydratase, class I(fumAB), 28: malate dehydrogenase (MDH1)/malate dehydrogenase [quinone] (mqo), 29: aspartate aminotransferase (GOT1), 30: aspartate ammonia-lyase (aspA), 31: argininosuccinate synthase (argG), 32: argininosuccinate lyase (argH), 33: adenylosuccinate synthase (purA), 34: adenylosuccinate lyase (purB), 35: succinate-semialdehyde dehydrogenase/glutarate-semialdehyde dehydrogenase (gabD), 36: 4-aminobutyrate aminotransferase (ABAT), 37/38: L-serine/L-threonine ammonia-lyase (SDS/SDH), 39: D-serine ammonia-lyase (DSD1), 40: alanine transaminase(ALT), 41:D-3-phosphoglycerate dehydrogenase/2-oxoglutarate reductase (serA/PHGDH), 42: phosphoserine aminotransferase (serC), 43: phosphoserine phosphatase (serB), 44: aldehyde dehydrogenase (aldB), 45: alcohol dehydrogenase [NADP+], (adh)/alcohol dehydrogenase 1/7 (K13951), 46: nitrate reductase/nitrite oxidoreductase (narGZ/nxrA), 47: nitrite reductase [NADH], (nirB), 48: glutamine synthetase (glnA), 49: glutamate synthase [NADPH] (gltB), 50/51: glutamate dehydrogenase [NADP+] (gdhA), 52: thiosulfate/3-mercaptopyruvate sulfurtransferase (TST), 53: taurine dioxygenase (tauD), 54: alkanesulfonate monooxygenase (ssuD), 55: sulfite reductase (NADPH) flavoprotein alpha-component (cysJ), 56: cysteine synthase (cysK)

Moreover, nitrate/nitrite reduction genes, including Nrt2 (nitrate/nitrite transporter) and NarG (nitrate reductase/nitrite oxidoreductase), were identified in the Turkish isolates. This finding also supports the central role of nitrogen metabolism in energy dynamics. The nitrate reductase/nitrite oxidoreductase enzyme facilitates the reduction of nitrate to ammonia, generating proton gradients and supporting ATP synthesis, thereby enabling survival in oxygen-depleted environments such as biofilms and deep tissues (Vazquez-Torres and Baumler, 2016; Ren and Palmer, 2023). Thus, glutamine synthetase (glnA) and glutamate synthetase (gltB) mediate ATP-dependent reactions converting ammonia into glutamate and glutamine (Sonawane and Rohm, 2004). Figure 10 illustrates nitrogen metabolism pathways, highlighting the integration of nitrate and nitrite utilization into complex energy networks.

Sulfur metabolism is also tightly integrated with energy pathways. Turkish isolates of A. baumannii were observed to possess enzyme systems that convert thiosulfate, taurine, and alkanesulfonate to sulfite. However, they lack the enzymes required to convert sulfate to sulfite during assimilatory sulfate reduction. Subsequently, they can produce hydrogen sulfide for incorporation into cysteine biosynthesis via sulfite reductase (cysJ) and serve as a proton donor (via NADPH) during anaerobic respiration. This demonstrates A. baumannii’s adaptation capacity to sulfur-rich niches (Ostrowski et al. 1989; Gebhardt et al. 2015). Apart from cysteine, methionine, another sulfur-containing amino acid, is linked to sulfur assimilation pathways involved in energy metabolism (Stipanuk and Ueki, 2011). Additionally, the use of host-derived sulfur sources, such as sulfate (plasma: 215.5 µM), thiosulfate (serum: 100.8 µM), taurine (plasma: 55.5 µM), and methionine (serum: 39.8 µM), poses a significant threat to host health by contributing to severe infections (Lensmire and Hammer, 2019).

In silico metabolic reconstruction suggested that these strains harbor pathways potentially involved in nitrate/nitrite utilization under oxygen-limited conditions, as well as the metabolism of substrates such as ethanol and amino acids. However, these results reflect predicted metabolic potential derived from genomic annotation and curated metabolic databases, rather than experimentally confirmed physiological activity. Therefore, the inferred energy metabolism profiles should be interpreted cautiously, and future phenotypic and biochemical studies will be necessary to validate the functional importance of these pathways under relevant environmental or infection-associated conditions.

Comparative pathogenomics

A. baumannii possesses various virulence factors that enhance its pathogenicity and enable its survival under adverse conditions. These factors in A. baumannii play pivotal roles in its ability to colonize, evade immune responses, and thrive in hostile environments. Analysis of the Turkish isolates (T3, T14, and T24) and comparisons with other strains revealed variations in virulence factor copy numbers and their potential functional implications, as summarized in Table 5. Key adhesion-related factors, such as the outer membrane protein OmpA, were universally present across all isolates, including T3, T14, T24, and reference strains like KAB01 and ATCC 17,978. Instead of other orthologs like ompF, oomp 33–36, and omp22, all of the A. baumannii genomes used in this study were found to possess the ompA gene. The ompA gene, which encodes a β-barrel porin, significantly contributes to the adhesion and invasion of epithelial cells through its interaction with fibronectin (Gaddy et al. 2012; Smani et al. 2012). The OmpA protein attaches to factor H in human serum, which may enable A. baumannii to evade complement-mediated killing (Kim et al. 2009). It has been reported that A. baumannii mutants expressing the OmpA protein induce apoptosis in human epithelial cells (Choi et al. 2005). Additionally, previous studies revealed that OmpA enhances the survival and persistence of A. baumannii by promoting surface mobility and biofilm formation (Clemmer et al. 2011; Gaddy et al. 2012; McConnell et al. 2013).

Table 5.

Comparison of virulence factor copy numbers among A. baumannii genomes employed in the present study

VFclass	Virulence factors	Related genes	T3	T14	T24	KAB01	AR_0063	K09_14	ab04_mff	ab736	ATCC 17,978	AYE
Adherence	Outer membrane protein	ompA	1	1	1	1	1	1	1	1	1	1
	LPS O-antigen (P. aeruginosa)	lpsO	-	-	-	-	-	-	1	-	-	-
	Type IV pili (Neisseria)	pilE	-	-	-	-	-	-	1	1	-	-
Biofilm formation	AdeFGH efflux pump/transport autoinducer	adeF	1	1	1	1	1	1	1	1	1	1
		adeG	1	1	1	1	1	1	1	1	1	1
		adeH	1	1	1	1	1	1	1	1	1	1
	Biofilm-associated protein	bap	3	2	2	2	-	5	1	-	1	1
	Csu pili	csuA/B	1	1	1	2	-	1	-	1	1	1
		csuA	1	1	1	1	1	1	-	1	1	1
		csuB	1	1	1	1	1	1	-	1	1	1
		csuC	1	1	1	1	1	1	-	1	1	1
		csuD	1	1	1	1	1	1	-	1	1	1
		csuE	1	1	1	1	1	1	-	1	1	1
	PNAG (Polysaccharide poly-N-acetylglucosamine)	pgaA	1	1	1	1	1	1	1	1	1	1
		pgaB	1	1	1	1	1	1	1	1	1	1
		pgaC	1	1	1	1	1	1	1	1	1	1
		pgaD	1	1	1	1	1	1	1	1	-	1
Enzyme	Phospholipase C	plc	2	2	2	2	2	2	3	2	2	2
	Phospholipase D	plcD	1	1	1	1	1	1	1	1	1	1
Immune evasion	Capsule	Undetermined	18	12	12	49	23	22	23	18	18	18
	LPS	lpsB	1	1	1	1	1	1	1	1	1	1
		lpxA	1	1	1	1	1	1	1	1	1	1
		lpxB	1	1	1	1	1	1	1	1	1	1
		lpxC	1	1	1	1	1	1	1	1	1	1
		lpxD	1	1	1	1	1	1	1	1	1	1
		lpxL	1	1	1	1	1	1	1	1	1	1
		lpxM	1	1	1	1	1	1	1	1	1	1
Iron uptake	Acinetobactin	barA	1	1	1	1	1	1	1	1	1	1
		barB	1	1	1	1	1	1	1	1	1	1
		basA	1	1	1	1	1	1	1	1	1	1
		basB	1	1	1	1	1	1	1	1	1	1
		basC	1	1	1	1	1	1	1	1	1	1
		basD	1	1	1	1	1	1	1	1	1	1
		basF	1	1	1	1	1	1	1	1	1	1
		basG	1	1	1	1	1	1	1	1	1	1
		basH	1	1	1	1	1	1	1	1	1	1
		basI	1	1	1	1	1	1	1	1	1	1
		basJ	1	1	1	1	1	1	1	1	1	1
		bauA	1	1	1	1	1	1	1	1	1	1
		bauB	1	1	1	1	1	1	1	1	1	1
		bauC	1	1	1	1	1	1	1	1	1	1
		bauD	1	1	1	1	1	1	1	1	1	1
		bauE	1	1	1	1	1	1	1	1	1	1
		bauF	1	1	1	1	1	1	1	1	1	1
		entE	1	1	1	1	1	1	1	1	1	1
	Heme utilization	Undetermined	1	1	1	1	1	1	1	1	1	1
		Undetermined	1	-	-	1	1	1	1	-	-	-
		Undetermined	1	-	-	1	1	1	1	-	-	-
		Undetermined	1	-	-	1	1	1	1	-	-	-
		Undetermined	1	-	-	1	1	1	1	-	-	-
		Undetermined	1	-	-	1	1	1	1	-	-	-
		Undetermined	1	-	-	1	1	1	1	-	-	-
		Undetermined	1	-	-	1	1	1	1	-	-	-
		Undetermined	1	1	1	1	1	1	1	1	-	1
		hemO	1	-	-	1	1	1	1	-	-	-
Regulation	Quorum sensing	abaI	1	1	1	1	1	1	-	1	1	1
		abaR	1	1	1	1	1	1	-	1	1	1
	Two-component system	bfmR	1	1	1	1	1	1	1	1	1	1
		bfmS	1	1	1	1	1	1	1	1	1	1
Serum resistance	PbpG	pbpG	1	1	1	1	1	1	1	1	1	1
	-	rmlD	-	1	1	-	-	-	-	-	-	-
		wbbL	-	1	1	-	-	-	-	-	-	-
Antiphagocytosis	Capsular polysaccharide (Vibrio)	wbjD/wecB	-	-	-	-	-	-	1	-	-	-
		rmlC	-	1	1	-	-	-	-	-	-	-
Stress adaptation	Catalase (Neisseria)	katA	1	-	-	1	1	-	-	-	-	-
		katG	1	1	1	1	1	1	1	1	1	1

The findings indicate that the OmpA protein represents a promising target for the development of new antibiotics and preventive measures against A. baumannii. Additionally, the presence of genes associated with biofilm formation, including bap (biofilm-associated protein [BAP]), adeFGH (efflux pump), csuA/B, csuABCDE (csu pili genes), and pgaABCD operon (proteins involved in poly-N-acetylglucosamine [PNAG] synthesis), was confirmed in Turkish isolates of A. baumannii. BAP is associated with biofilm formation and adhesion to eukaryotic host cells. It consists of a large cell-surface protein composed of multiple copies of repeat elements (Upmanyu et al. 2022). These proteins primarily comprise arrays of 80–110 amino acid modules that exhibit immunoglobulin-like motifs (De Gregorio et al. 2015). As a hydrophobic cell-surface protein, BAP not only facilitates biofilm formation on human cell surfaces but also on medically relevant materials such as polystyrene, polypropylene, and titanium (Loehfelm et al. 2008; Brossard and Campagnari, 2012). The bap gene copy number varied among the studied genomes, with T3 harboring three copies, compared to two in T14 and T24.

This discrepancy likely reflects a higher biofilm-forming capacity in T3. All examined genomes contained the adeFGH efflux pump operon, which is classified as a resistance-nodulation-cell division (RND)-type efflux system. Reports suggest that AdeFGH efflux pumps are synergistically linked to tigecycline resistance (Damier-Piolle et al. 2008). The expression of AdeFGH is regulated by the LysR-type transcriptional regulator adeL and the TetR-type transcriptional regulator adeN (Lee et al. 2017). Previous studies have shown that low-dose antimicrobial treatment for A. baumannii increases the expression of the adeFGH operon. This leads to the production and transfer of autoinducer molecules, thereby strongly enhancing biofilm formation (Kaplan 2011; He et al. 2015). CSU pili, essential for surface adherence and biofilm formation, were uniformly encoded across genomes used in this study, except for ab04_mff. However, slight variations in the expression of the csuA/B operon genes may contribute to phenotypic differences in biofilm structure and robustness among the strains (Law and Tan, 2022; Ahmad et al. 2023). Ahmad et al. (2023) suggested that the CSU pilus is a functional element of mature A. baumannii biofilms and a significant virulence factor that facilitates disease progression by promoting bacterial adherence to epithelial cells.

Another element responsible for biofilm formation is PNAG, produced by the PGA machinery, comprising the pgaABCD operon-encoded proteins (Choi et al. 2009). The distribution of these genes across all compared genomes was homogeneous, except for the absence of the pgaD gene in ATCC 17,978. Briefly, the PGA complex likely spans both the inner and outer bacterial membranes. The PgaA functions as a porin in the outer membrane and may interact with PgaB, a periplasmic lipoprotein anchored to the inner part of the outer membrane (Skurnik et al. 2016). The PgaB deacetylates a small portion of the NAG units within the polymer. The PgaC and PgaD form a complex embedded in the inner membrane that extends across both sides (Whitfield et al. 2015). PgaC is a glycosyltransferase that functions catalytically and is predicted to possess 4–5 transmembrane helices (TMHs). The PgaD likely comprises two transmembrane helices, featuring a short periplasmic loop and a small cytosolic C-terminus (Choi et al. 2009; Wang et al. 2016). The production of extracellular poly-beta-1,6-N-acetylglucosamine by the pgaABCD machinery is linked to biofilm formation (Choi et al. 2009; Bi et al. 2015).

This study confirmed the presence of phospholipase C enzyme-encoding genes plc1 and plc2, as well as the phospholipase D enzyme-encoding gene plcD, in all genomes analyzed through VFDB (Table 5). Phospholipases are lipolytic enzymes that break down phospholipids and serve as virulence factors in various bacteria, including C. perfringens, A. baumannii, and P. aeruginosa (Lee et al. 2017). These enzymes contribute to pathogenesis by breaking down phospholipids in host cell membranes and at mucosal barriers, making it easier for bacteria to enter (Jacobs et al. 2010).

Capsule synthesis genes showed notable variability. Strain T3 had a higher number of capsule-related genes than strains T14 and T24, potentially contributing to enhanced immune evasion. Capsular structures in A. baumannii play a crucial role in protecting the bacteria against the host’s innate immune response (Russo et al. 2010). The capsule gene cluster exhibits significant variability across A. baumannii genomes depending on sequence type. The K locus cluster typically contains either capsule export genes (wza, wzb, and wzc) or genes involved in simple sugar synthesis (galU, ugd, gpi, gne1, and pgm). The genes are positioned adjacent to a central variable region essential for the synthesis of a specific monosaccharide (Shan et al. 2022). Genes associated with lipopolysaccharide (LPS) production, namely lpxD, lpsB, lpxA, lpxB, lpxL, lpxM, and lpxC, are present in all genomes. LPS is the primary component of the outer leaflet of the outer membrane in most Gram-negative bacteria and acts as an immunostimulatory protein that stimulates the release of tumor necrosis factor and interleukin-8 from macrophages via a Toll-like receptor-4 (TLR4)-dependent mechanism (Lee et al. 2017). LPS functions to evade the host immune response and trigger an inflammatory response, which is strongly associated with morbidity and mortality in infected patients (Luke et al. 2010).

Iron acquisition is vital for bacterial survival and virulence, especially in iron-limited host environments (Mey et al. 2021). All isolates, including T3, T14, and T24, displayed a complete acinetobactin biosynthesis cluster (barA, basA-J, bauA-F), consistent with their ability to overcome host nutritional immunity. Acinetobactin is a high-affinity catechol-hydroxamate siderophore that competes with host cells for iron uptake (Fiester and Actis, 2013). Conversely, it was observed that half of the analyzed genomes lacked eight genes, including heme oxygenase (hemO), which are associated with the heme uptake system, except for T3, KAB01, AR_0063, K09-14, and ab04-mff. Previous studies have shown that most hypervirulent strains of A. baumannii possess a heme uptake system, potentially including hemO gene, thereby confirming that T3 exhibits hypervirulent activity as well (Giardina et al. 2019).

The regulation of biofilm formation is mediated by quorum-sensing genes abaI and abaR, as well as the two-component system genes bfmR and bfmS, which displayed homogeneous distributions across all examined genomes, except for the loss of abaI and abaR genes in the ab04_mff genome. As mentioned earlier, the ability of A. baumannii to form biofilms begins with adhesion to surfaces, which is facilitated by the production of e pili. The csu operon in A. baumannii mediates pili expression, which is regulated by bfmRS genes (Tomaras et al. 2008). The removal of the response regulator bfmR in A. baumannii ATCC19606 has been reported to completely eliminate biofilm formation (De Silva and Kumar, 2019). The quorum-sensing system of A. baumannii is composed of two components: the abaI inducer and its corresponding receptor, abaR, resembling the traditional luxI/luxR system observed in other Gram-negative bacteria (Saipriya et al. 2020). AbaI is a sensor protein, encoded by the abaI gene, that acts as an auto-inducer synthase, producing acyl homoserine lactone (AHL) signal molecules necessary for biofilm formation in A. baumannii (Niu et al. 2008). The AbaR serves as a receptor protein for AHL, and its binding to AHL triggers a cascade of reactions related to the production of biofilm (Saipriya et al. 2020). These findings suggest that these regulatory genes represent promising therapeutic targets, novel drug candidates, and new treatment strategies.

Serum resistance–associated genes pbpG, rmlD, and wbbL were screened across all genomes utilized in this study. The pbpG gene showed a uniform distribution across all analyzed genomes. Penicillin-binding proteins (PBPs) encoded by pbpG gene are primarily linked to the attachment and inactivation of β-lactam antibiotics. PBPs are involved in the final stages of peptidoglycan biosynthesis, thereby contributing to bacterial cell stability (Russo et al. 2010). Notably, the rmlD and wbbL genes are present exclusively in the T14 and T24 genomes among all analyzed genomes. The functions of rmlD and wbbL genes from Serratia marcescens were tested by Rubires et al. (1997), who cloned these genes into Escherichia coli DH5α and demonstrated that both are essential for O4 antigen biosynthesis, which provides serum resistance. The findings verify that T14 and T24 strains may exhibit serum resistance due to the presence of the aforementioned genes.

The examined genomes revealed genes related to stress adaptation and resistance. The katA, encoding a catalase enzyme originally derived from Neisseria, is present only in the T3, KAB01, and AR_0063 genomes. On the other hand, all genomes were identified to include the katG gene, which encodes catalase-peroxidase [E.C. 1.11.1.21]. Phagocytes, including neutrophils, macrophages, and monocytes, in the host generate H₂O₂ and other reactive oxygen species (ROS) to eliminate bacteria, a process commonly known as the “respiratory burst” (Sato et al. 2019). The breakdown of H₂O₂ by catalase provides protection and enhances the survival of A. baumannii strains (Sun et al. 2016). The katG is identified as a major factor affecting the resistance to H₂O₂ in Acinetobacter species, whereas katA is reported to be ineffective against H₂O₂ (Sato et al. 2019). Although it is not possible to comment on the effect of the additional katA gene on A. baumannii virulence, it has been reported that wild-type A. baumannii strains lacking the katG gene are more virulent (Sun et al. 2016). Based on this information, the combined application of a katG inhibitor and H₂O₂ may serve as an effective antiseptic agent against A. baumannii and other nosocomial pathogens that express this enzyme.

These findings highlight the complex interplay of virulence factors that contribute to the pathogenicity and persistence of A. baumannii across diverse environments. The Turkish isolates exhibit a conserved core virulence gene repertoire; however, strain-specific variations in adhesion-related proteins, biofilm-forming components, capsule synthesis genes, and iron acquisition systems underscore the adaptive mechanisms these bacteria employ to enhance survival and immune evasion. The identification of unique gene variations, such as differences in bap gene copy numbers, phospholipase genes, and quorum-sensing regulators, suggests potential targets for novel therapeutic interventions. Moreover, the presence of serum-resistance genes in specific strains underscores the need for further research into host-pathogen interactions and immune-escape strategies. Understanding these virulence determinants not only advances our knowledge of A. baumannii pathogenicity but also provides a basis for the development of targeted antimicrobial therapies to combat multidrug-resistant infections.

Antibiotic resistance and the associated resistome

The multidrug resistance (MDR) observed in A. baumannii strains T3, T14, and T24 is supported by a complex resistome that includes both acquired and intrinsic antibiotic resistance genes (ARGs). Minimum inhibitory concentration (MIC) assays (Table 6) indicate that these genetic factors are associated with phenotypic resistance to a broad spectrum of antimicrobial agents. Resistome analysis conducted via the CARD and BV-BRC platforms (Alcock et al. 2023; Olson et al. 2023) highlights the significant role of antibiotic-inactivating enzymes, target modification, and efflux mechanisms in the phenotypic resistance profiles of these strains (Supplemental File 3).

Table 6.

Antibiotic susceptibility test results based on minimum inhibitory concentrations (MICs)

Class	Antibiotic	MICs (mg/L)
		T3	T14	T24
Aminoglycoside	Amikacin	<=8 (S)	> 32 (R)	> 32 (R)
	Gentamicin	> 8 (R)	> 8 (R)	> 8 (R)
	Tobramycin	-*	> 8 (R)	-*
Beta-Lactams	Amoxicillin-Clavunate (Combination of a penicillin and a beta-lactamase inhibitor)	> 16/2 (R)	> 32/2 (R)	> 16/2 (R)
	Ampicillin (Penicillin)	> 16 (R)	> 16 (R)	> 16 (R)
	Cefazolin (First-generation cephalosporin)	> 32 (R)	> 32 (R)	> 32 (R)
	Cefepime (Fourth-generation cephalosporin)	> 8 (R)	> 4 (R)	> 8 (R)
	Ceftazidime (Third-generation cephalosporin)	> 8 (R)	> 16 (R)	> 8 (R)
	Ceftolozane-Tazobactam (Combination of a cephalosporin and a beta-lactamase inhibitor)	> 4/4(R)	-*	> 4/4(R)
	Ceftriaxone (Third-generation cephalosporin)	> 4 (R)	> 4 (R)	> 4 (R)
	Ertapenem (Carbapenem)	1 (R)	> 2 (R)	> 1 (R)
	Imipenem (Carbapenem)	> 8 (R)	> 8 (R)	> 8 (R)
	Meropenem (Carbapenem)	<=0.125 (S)	> 8 (R)	> 8 (R)
	Piperacillin-Tazobactam (Combination of a penicillin and a beta-lactamase inhibitor)	> 16/4	> 32/4 (R)	> 16/4
Fluoroquinolone	Ciprofloxacin	> 1 (R)	> 1 (R)	> 1 (R)
	Levofloxacin	> 2 (R)	> 4 (R)	> 2 (R)
Folate pathway inhibitor	Trimethoprim-Sulfamethoxazole	> 8/152 (R)	> 8/152 (R)	> 8/152 (R)
Glycylcycline	Tigecycline	> 2(R)	4 (R)	> 2(R)
Nitrofuran	Nitrofurantoin	-*	> 128 (R)	-*
Phosphonic acid derivative	Fosfomycin w/G6P	-*	> 64 (R)	-*
Polymyxin	Colistin	> 4(R)	-*	<=1 (S)

* - Not tested

Beta-lactam resistance in these isolates is primarily mediated by the expression of class beta-lactamases. The presence of class D beta-lactamase (OXA family), specifically blaOXA-66 and blaOXA-23 in T3, and blaOXA-852 in T14 and T24, supports carbapenem resistance (Evans and Amyes, 2014). Notably, the presence of the blaOXA-852 carbapenem-specific resistance gene and the adeN resistance-nodulation-cell division (RND) antibiotic efflux pump gene in the T14 and T24 genomes confers meropenem resistance in these isolates. Therefore, it is possible to hypothesize that T3 retains sensitivity to meropenem (MIC ≤ 0.125 µg/mL), presumably because it lacks the aforementioned genes. The CARD database reports that blaOXA-852 is prevalent across sequenced genomes, plasmids, and whole-genome shotgun assemblies (Alcock et al. 2023). Additionally, the beta-lactamase database indicates that blaOXA-852 is naturally present in A. baumannii strains, a finding confirmed by our horizontal gene transfer test results (Naas et al. 2017). Moreover, isolates T14 and T24 encoding blaOXA-852, adeL, and adeN exhibited resistance to all tested carbapenems, including ertapenem, imipenem, and meropenem. The adeL regulates the adeFGH efflux pump associated with resistance to fluoroquinolones and tetracyclines. At the same time, the adeN controls the adeIJK efflux pump linked to resistance against macrolides, fluoroquinolones, lincosamides, carbapenems, cephalosporins, penicillin beta-lactams, tetracyclines, rifamycins, diaminopyrimidines, and phenicols in A. baumannii (Alcock et al. 2023; Wimalasekara et al. 2025). Moreover, mutations or deletions of these regulatory genes may lead to either overexpression of efflux pumps or loss of efflux pump activity (Rosenfeld et al. 2012; Yoon et al. 2013). The absence of the adeN regulator gene in the T3 genome may be another factor in the development of meropenem susceptibility. Genes encoding class C beta-lactamase (ADC family) were detected across all three strains, correlating with resistance to cephalosporins such as cefazolin, ceftriaxone, and cefepime (Table 6). According to CARD database analysis, only the blaADC-73 cephalosporinase was identified in the T3 genome, while blaADC-270 was detected exclusively in T14 and T24. Previous studies have reported blaADC-73 in A. baumannii isolates from Turkish hospitals (Gulbuz and Saral Sariyer, 2022). Acinetobacter-derived cephalosporinases (ADCs) confer resistance to cephamycin (including cefoxitin and cefotetan), cephalosporins, penicillins, and beta-lactamase inhibitor combinations. Interestingly, they remain unaffected by beta-lactamase inhibitors such as clavulanate and sulbactam (Moubareck and Halat 2020), which is confirmed by the MIC results obtained in this study.

Resistance to aminoglycosides such as amikacin and gentamicin is mediated by aminoglycoside-modifying enzymes. The ANT(3’’)-IIa gene, encoding aminoglycoside nucleotidyltransferase, was consistently identified in all strains and is responsible for resistance to streptomycin and spectinomycin (Kyriakidis et al. 2021). In contrast, the aminoglycoside resistance genes APH(3’’)-Ib (streptomycin), APH(3’)-VIa (amikacin, kanamycin), APH(6)-Id (streptomycin), and armA (gentamycin) were only detected in the T3 genome (Kyriakidis et al. 2021; Alcock et al. 2023). Based on the MIC results, all Turkish A. baumannii genomes were resistant to gentamicin (MIC > 8), while only the T3 isolate was sensitive to amikacin. Several factors may explain this issue. First, the resistance to gentamycin in the T3, T14, and T24 genomes can be attributed to the presence of the AmvA gene, which encodes a major facilitator superfamily (MFS) antibiotic efflux pump. However, despite the identification of APH(3’)-VIa and AmvA genes in the T3 genome, the observed sensitivity to amikacin may result from the IS4-like element ISAba1 family transposase (ACGIGR_19715) found upstream of the APH(3’)-VIa gene, that may responsible for a transposase-associated frameshift predicted to disrupt the coding sequence, which may explain the observed amikacin susceptibility despite the presence of this aminoglycoside resistance gene (Fig. S1). The genetic context, including the location of resistance genes on mobile elements such as plasmids or within the chromosome, and the presence of insertion sequences, can significantly influence the expression and mobilization of antibiotic resistance genes (Razavi et al. 2020; Udaondo et al. 2022).

Fluoroquinolone resistance, as evidenced by elevated MICs to ciprofloxacin and levofloxacin, is associated with efflux pumps such as AdeFGH, AdeIJK, and AbaQ. Notably, the CARD database suggests that AbaQ is primarily responsible for fluoroquinolone resistance in A. baumannii (Alcock et al. 2023). Moreover, Perez-Varela et al. (2018) used efflux pump inhibitor carbonyl cyanide 3-chlorophenylhydrazone (CCCP) to assess the effectiveness of abaQ versus ciprofloxacin and levofloxacin. The application of CCCP (20 mg/liter) resulted in a 32-fold reduction in the ciprofloxacin MIC and a 16-fold reduction in the levofloxacin MIC in abaQ-carrying cells, whereas it only caused a 4-fold reduction in MIC in cells lacking the abaQ gene.

High MIC values (> 8/152 µg/mL) for trimethoprim-sulfamethoxazole across all Turkish strains correlate with the presence of dfrA1 (dihydrofolate reductase) and the adeIJK efflux pump operon (Damier-Piolle et al. 2008; Brolund et al. 2010; Kyriakidis et al. 2021). Additionally, the sul2 gene, which encodes a sulfonamide-resistant dihydropteroate synthase, was detected in T3, T14, and T24 genomes (Kyriakidis et al. 2021). Interestingly, phenotypic susceptibility to colistin (MIC ≤ 1 µg/mL) in T24 aligns with the absence of plasmid-mediated mcr genes, which confer polymyxin resistance. Conversely, colistin resistance (MIC > 4 µg/mL) has been detected in the T3 isolate despite the absence of a colistin resistance gene in its genome. The resistance mechanism may be attributed to mutations in one of the initial three genes of the lipid A production pathway: lpxA, lpxC, or lpxD, leading to LPS loss in A. baumannii (Moffatt et al. 2010). The absence of LPS results in resistance by removing the lipid A that colistin targets. These mutations may lead to a persistent and irreversible resistance phenotype (Moffatt et al. 2010; Kyriakidis et al. 2021). We have aligned amino acid sequences of all LPS genes (lpsB, lpxA, lpxB, lpxC, lpxD, lpxL, and lpxM) from the T3, T14, and T24 genomes to verify this issue (Fig. S2). It has been confirmed that, except for the lpxA gene, all genes in the T3 genome have at least one amino acid mutation. These findings therefore provide a putative explanation for the observed phenotype, although direct functional validation was not performed here.

Tigecycline resistance in all strains (MIC > 2 µg/mL) may involve efflux pump overexpression rather than direct modification of the target ribosome. The adeFGH and adeIJK efflux pumps, present in all strains, are known to contribute to tigecycline resistance by expelling the antibiotic from the bacterial cell (Sun et al. 2023). Furthermore, it has been shown that a mutation in rpsJ (ribosomal protein S10p) alone can confer tigecycline non-susceptibility, and such mutations in S. aureus, A. baumannii, and E. coli frequently arise following tigecycline exposure (Beabout et al. 2015). Therefore, the rpsJ mutation, which induces conformational changes in the extended loop of the S10 protein adjacent to the 16 S rRNA tigecycline binding site, should not be ignored for its role in reducing sensitivity. The resistome screening also identified the abaF gene, which encodes a major facilitator superfamily (MFS) antibiotic efflux pump, in the strains. The MFS transporter protein AbaF has been shown to actively efflux fosfomycin, contributing to cell resistance, with its expression upregulated in response to fosfomycin exposure (Sharma et al. 2017). The AbaF also plays a role in the secretion of the biofilm matrix, which contributes to the pathogenicity of A. baumannii and may facilitate the expulsion of host defense molecules, thereby significantly affecting its virulence (Sharma et al. 2017; Kyriakidis et al. 2021). Furthermore, the presence of DNA gyrase subunit B (gyrB) in T24 may influence resistance to fluoroquinolones and fosfomycin, particularly when considered alongside mutations in DNA gyrase subunit A (gyrA) (Drlica and Zhao, 1997; Maruri et al. 2012; Sharma et al. 2017).

Despite extensive screening, no antibiotic resistance genes (ARGs) were identified in T3, T14, and T24 plasmids. This absence suggests that the antibiotic resistance observed in these strains is primarily due to chromosomal mutations. The lack of plasmid-borne ARGs may reflect evolutionary adaptations in these strains, which rely on intrinsic or chromosomally acquired mechanisms rather than horizontal gene transfer via plasmids (Kyriakidis et al. 2021; Cain and Hamidian, 2023; Mack et al. 2025). Additionally, the absence of plasmid-mediated resistance genes could reduce the likelihood of these resistance traits spreading across different bacterial populations, potentially reflecting localized genetic evolution (Lee et al. 2017; Kyriakidis et al. 2021). The resistome analysis highlights the significant role of antibiotic-inactivating enzymes, target modification, and efflux mechanisms in the phenotypic resistance profiles of these strains. For instance, T3’s susceptibility to meropenem contrasts with the carbapenem resistance of T14 and T24, which is due to differential expression of beta-lactamase genes. Moreover, the widespread occurrence of aminoglycoside-modifying enzymes underpins the phenotypic resistance observed across all strains. These findings highlight the genomic basis of antimicrobial resistance in A. baumannii and underscore the importance of integrative phenotypic and genotypic analyses to inform therapeutic strategies against MDR pathogens.

Conclusion

This study provides a comprehensive genomic and metabolic analysis of multidrug-resistant Turkish Acinetobacter baumannii isolates, revealing critical insights into their pathogenicity, resistance mechanisms, and survival strategies. The comparative pathogenomics of ST195 (T3) and novel sequence types (T14, T24) revealed substantial genomic diversity, influenced by horizontal gene transfer, phage integration, and the acquisition of genomic islands. Key resistance genes (blaOXA-23, blaOXA-66, blaOXA-852, armA, adeFGH) were correlated with phenotypic resistance profiles. Unexpected susceptibility to meropenem in T3 (MIC ≤ 0.125 µg/mL) and colistin in T24 (MIC ≤ 1 µg/mL) highlighted the significance of regulatory mutations (absence of adeN) and loss of LPS (mutations in lpxA/C/D), respectively. Virulence factors, including ompA, bap, and csu pili, were conserved across isolates, with strain-specific variations in capsule synthesis and phospholipase genes contributing to enhanced immune evasion and tissue invasion. Metabolic profiling revealed dependence on nitrogen and sulfur absorption pathways, facilitating survival under nutrient-deficient conditions. The open pangenome (α = 0.60) and extensive accessory genome highlighted A. baumannii’s adaptive plasticity. These findings emphasize the urgent need for targeted therapeutics that address metabolic vulnerabilities (e.g., the TCA cycle and nitrate reductase) and regulatory nodes (e.g., bfmRS and abaIR). Future studies should validate the transcriptomic profiles of resistance and virulence genes and develop combinatorial approaches to inhibit biofilm formation and efflux pump activity, offering potential solutions against this resilient pathogen.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

MSH. Performed the experiment. Analyzed the data. MA. Conception and design of the study. Wrote manuscript. AEY. Analyzed the data. Wrote the manuscript. PS. Provided strains. Contributed to writing.

Funding

Partial financial support was received from the Erciyes University Research Fund.

Data availability

The complete genome sequences and plasmids of strains T3, T14, and T24 were deposited in NCBI under the accession numbers CP171394-CP171395, CP171396-CP171397, and CP171398-CP171399, respectively. All data supporting the findings of this study are available within the paper and its Supplementary Information.

Declarations

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable.