First comprehensive whole genome sequencing report of a multidrug-resistant Serratia nematodiphila strain from Bangladesh

Abstract

Background

Serratia spp. have emerged as one of the most prevalent causative agents of nosocomial infections. This study aims to study the whole genome sequence of a multidrug-resistant Serratia nematodiphila strain as well as characterize the prevalence of multidrug-resistant (MDR) Serratia spp., in hospital environments in Bangladesh.

Methods

A total of 78 samples comprising hospital sewage and soil were collected from seven renowned hospitals across Dhaka city. Next, the samples tested positive for Serratia spp. were subjected to further analysis. Antibiotic susceptibility test and biofilm formation assay were carried out for these isolates, and the most antimicrobial resistant Serratia isolate, along with moderate biofilm-forming ability, was subjected to whole genome sequencing (WGS). Subsequently, the WGS as well as pan-genomic and pan-resistome analysis were performed.

Results

A total of 23 Serratia spp. isolates along with 47.8% of MDR properties were identified. WGS analysis identified the highly MDR isolate, KS1, as Serratia nematodiphila. Further analysis revealed the presence of numerous antimicrobial resistance and efflux pump genes, a variety of mobile genetic elements, virulence factor genes, and prodigiosin producing biosynthetic gene cluster. These features are predicted to be responsible for its multidrug-resistant properties and high virulence. Furthermore, the lower core/pan ratio of this pangenome indicated the likelihood to acquire more antimicrobial resistance and virulence genes.

Conclusion

To date, this is the first detailed genomic study of Serratia nematodiphila in Bangladesh. Further extensive researches are required to find out the prevalence and clinical significance of this particular organism.

1. Introduction

Nosocomial infections, also denoted as healthcare-associated infections (HAIs), are infections that one acquires during the course of medical treatment in healthcare settings. These infections are typically caused by antibiotic-resistant bacteria, ultimately leading to increased morbidity, mortality, and healthcare costs. The most common nosocomial infections include urinary tract infections, surgical site infections, and pneumonia, particularly ventilator-associated pneumonia [1]. These infections are caused by a variety of organisms including Staphylococcus aureus (especially MRSA), Escherichia coli, Pseudomonas aeruginosa, and Clostridium difficile [2]. Recently, multidrug-resistant Serratia spp. have emerged as a significant cause of nosocomial infections, particularly in hospital settings where vulnerable patients are at an increased risk. These bacteria are resistant to multiple classes of antibiotics, for instance, beta-lactams and aminoglycosides, making infections difficult to treat. Serratia spp. can persist in hospital environments because of their ability of catheter colonization through forming biofilms [3]. According to the European Centre for Disease Prevention and Control (ECDC) 2017 report, Serratia spp. were identified as the sixth and ninth most common pathogen causing ICU-acquired pneumonia and bloodstream infections, respectively, in Europe [[4], [5], [6]]. In 2022, a nosocomial outbreak of bloodstream infection, caused by Serratia marcescens, was reported in an ICU unit of a Hungarian hospital. In this outbreak, Serratia marcescens isolates were resistant to a disinfectant, quaternary ammonium compound and was assumed to be implicated with a particular sink of that ICU unit [7].

Serratia spp. have some significant adaptive mechanisms for adaptability that also contribute to its role as opportunistic pathogens. For instance, Serratia marcescens can survive at a wider range of temperature (5°C-40 °C) and pH (5-9). Moreover, this bacterium can survive not only in moist environments, for example, basins, taps, air conditioning systems etc. but also on dry surfaces for around 60 days [7,8]. One striking feature that most of Serratia sp. exhibited is the production of red pigment called “Prodigiosin”. This pigment is also well-known for antimicrobial activity and produced from the two precursors 2-methyl-3-n-amyl-pyrrole (MAP) and 4-methoxy-2,2′-bipyrrole-5-carbaldehyde (MBC) and these precursors are encoded by 2 different combinations of pig genes [9,10]

Carbapenems and colistin are considered last-resort antibiotics for treating infections caused by multidrug-resistant (MDR) gram-negative bacteria. However, the combination of intrinsic colistin resistance and acquired carbapenem resistance in Serratia spp. have significantly reduced available treatment options, increasing the risk of treatment failure and associated morbidity and mortality. Globally, Serratia marcescens has also been reported as a major nosocomial pathogen with evidence of outbreaks due to multidrug-resistant strains. In China, two carbapenem-resistant Serratia marcescens strains producing KPC-2 were identified that were associated with bloodstream infection [11]. Also, a review mentioned that resistance of Serratia marcescens to numerous antibiotics, including carbapenems and colistin, making treatment challenging [12]. Recent studies have also identified Serratia nematodiphila (S. nematodiphila) in clinical settings, especially in bloodstream infections and highlighted the increasing potential of S. nematodiphila as an emerging human pathogen in the near future [5].

Bangladesh is no exception to the growing number of reports of Serratia spp. infections in recent years. In 2024, a clinical isolate of Serratia marcescens, isolated from a hospital of Dhaka city, was found to possess bla_NDM-7 [13]. Besides, another multidrug-resistant Serratia marcescens strain, isolated from tracheal aspirates, has been reported in this year [14]. Moreover, Serratia species were also isolated from environmental reservoirs or food in Bangladesh. Serratia marcescens was found in retailed shrimp, P. monodon. [15]. Surprisingly, another study identified S. nematodiphila from street vended foods [16].

In the context of studying pathogenic bacteria, the technique of whole genome sequencing (WGS) has gained significant traction in research, clinical diagnostics, and public health laboratories due to a recent advancement [17,18]. In addition to identifying bacteria at various taxonomic levels, WGS offers the opportunity to detect various genetic markers, such as virulence factors or antibiotic resistance-associated genes, and analyze unculturable isolates [19,20]. A thorough analysis of the complete genome of a pathogen reveals an extensive array of attributes pertaining to its virulence factors, pathogenesis, and underlying mechanisms. This comprehensive analysis makes a substantial contribution to the understanding of the pathogen's mode of action and offers valuable insights that can aid in the investigation of various alternative strategies designed to counteract its current treatment dominance [21]. Antimicrobial resistance, which has emerged as a significant peril to present and future human health, can be assessed and predicted through the utilization of whole genome sequencing. This involves the identification of emerging resistance patterns, novel antibiotic discovery, surveillance, diagnostic method development, and predicting control measures among clinically important microorganisms, which offers numerous benefits and remarkable precision in comparison to conventional phenotypic techniques [22].

In Bangladesh, there is only limited detailed research or reports on the occurrence and characterization of multidrug-resistant (MDR) Serratia spp. in hospital sewage and soil. Few studies have been published regarding the presence of this potential threat in hospital environments, where it plays a significant role as a nosocomial pathogen. The inadequate attention to this infectious opportunistic pathogen emphasizes the need for further studies on Serratia spp., which are intrinsically resistant to multiple classes of antibiotics, in Bangladesh.

2. Materials and methods

2.1. Sample collection, processing and presumptive identification

Hospital sewage and soil were collected from seven renowned hospitals in Dhaka city from 23 February, 2025 to 29 February, 2025. The number of samples per site was proportionate to the area size of hospitals as well as number of drainage lines and the information regarding sample collection is described in Supplementary Table 1. Both sewage and soil samples were transported to lab in 15 mL of sterile falcon tubes. For processing the soil sample, 1 g of soil was mixed with 9 ml of normal saline by vortexing to obtain a 10-fold dilution whereas no dilution was performed on the sewage sample. Due to intrinsic resistance to colistin, 1 ml of processed sample was added to 9 ml of nutrient broth supplemented with colistin sulphate to selectively enrich the growth of Serratia spp. (intrinsically resistant to colistin sulphate) and incubated overnight at 37 °C. In the next day, one loopful of broth was streaked on cetrimide agar and nutrient agar and incubated at 25 °C in the incubator. After incubation, single colonies were selected for further identification using biochemical tests and molecular confirmation targeting quorum-sensing gene luxS [23]. The sequences of primer pairs used to detect luxS gene in Serratia isolates are provided in Supplementary Table 12.

2.2. Antimicrobial susceptibility test and biofilm formation assay

After presumptive identification, the isolates were tested for antimicrobial susceptibility following Kirby Bauer method [24] to 15 antibiotics belonging to 12 different classes on Mueller Hinton agar in accordance with the Clinical and Laboratory Standards Institute (CLSI) guidelines (Clinical and Laboratory Standards Institute,2021). Biofilm formation assay was performed and evaluated by following a modified protocol of crystal violet assay [25]. Briefly, 1:100 diluted overnight cultures in LB were incubated statically at 37 °C for 48 h, followed by washing and staining with 0.1% crystal violet for 12 min. Next, the bound dye was solubilized with 33% acetic acid. Absorbance at 600 nm was measured, and isolates were classified as non-, weak, moderate, or strong biofilm-formers based on optical density (OD). The results of antibiotic resistance and biofilm formation assay for 23 Serratia isolates, obtained in this study, are summarized in Table 3.

Table 3.

Summarized version of Kirbey-Bauer assay and biofilm formation assay results in this study.

Sample ID	Hospital Origin	Type of Sample	MEM	IPM	CAZ	CPM	FOX	CIP	SXT	AMP	C	FOS	TE	GEN	AK	ATM	TZP	No. of Resistant Classesa	Classification of Resistance	Biofilm
SSO3	SSMC	Soil	R	S	S	S	R	R	R	R	S	R	R	S	S	S	S	5	MDR	MBF
SS1	SSMC	Sewage	R	S	R	R	R	I	S	R	S	R	S	I	R	R	R	6	MDR	MBF
SS2	SSMC	Sewage	I	S	S	S	R	R	S	R	S	S	S	S	S	S	S	1	Non-MDR	WBF
SSb2	SSMC	Soil	R	S	R	I	R	R	S	R	S	I	I	I	I	I	R	4	MDR	WBF
KS1	SKH	Sewage	R	R	R	R	R	R	S	R	S	R	I	R	I	R	R	7	MDR	MBF
KS2	SKH	Sewage	R	I	R	R	R	I	S	R	S	R	S	I	R	R	R	6	MDR	MBF
SHSO3	ShSMC	Soil	I	I	R	I	R	R	S	R	S	S	R	S	S	S	R	4	MDR	WBF
SHSe1	ShSMC	Sewage	S	S	R	I	R	I	S	R	S	S	R	S	S	S	I	2	Non-MDR	MBF
SHSe2	ShSMC	Sewage	S	I	R	I	R	I	S	R	R	I	R	S	S	S	I	3	MDR	WBF
SHSe3	ShSMC	Sewage	S	S	I	I	R	S	S	R	S	I	R	S	S	S	I	1	Non-MDR	MBF
M'S1	MuMC	Sewage	R	S	S	S	R	S	S	R	S	S	S	S	S	S	R	2	Non-MDR	MBF
M'S2	MuMC	Sewage	R	S	R	R	R	R	I	R	S	S	R	R	R	S	S	5	MDR	WBF
M'S6	MuMC	Sewage	R	R	R	R	R	I	R	R	R	S	S	R	I	R	R	7	MDR	WBF
M'S7	MuMC	Sewage	R	S	S	S	R	S	S	R	S	S	S	S	S	S	I	1	Non-MDR	WBF
M'S8	MuMC	Sewage	R	S	S	I	R	S	S	R	S	S	S	S	S	S	S	1	Non-MDR	MBF
M'So7	MuMC	Soil	I	S	S	S	R	I	S	R	S	S	S	S	S	S	S	0	Non-MDR	WBF
MS1	MCHTI	Sewage	R	R	R	R	R	R	S	R	S	I	S	S	S	R	I	4	MDR	MBF
MS2	MCHTI	Sewage	R	I	R	R	R	R	S	R	S	R	S	S	S	R	I	5	MDR	MBF
DAC	DMC	Sewage	R	S	S	S	R	S	S	R	S	S	S	S	S	S	S	1	Non-MDR	WBF
PSO3′	BMU	Soil	R	I	I	I	R	S	S	R	S	S	R	S	S	S	S	2	Non-MDR	WBF
PS2	BMU	Sewage	R	S	S	S	R	S	S	R	S	S	R	S	S	S	I	2	Non-MDR	MBF
PS5	BMU	Sewage	R	S	I	I	R	I	S	R	S	S	S	S	S	S	I	1	Non-MDR	WBF
PS7	BMU	Sewage	I	S	I	R	R	I	S	R	S	S	R	S	S	S	I	2	Non-MDR	NBF

^aExcluding those classes of antibiotics to which Providencia spp. are intrinsically resistant.

2.3. Whole genome sequencing, quality control, assembly and identification

The whole genome sequencing of the target isolate was performed on the Illumina platform using the NovaSeq 6000 sequencing system at the Genomic center, International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b). The generated FASTQ file was evaluated for quality through FastQC (v0.11) [26]. Trimmomatic v0.39 was used to perform adapter trimming and quality filtering (minimum read length of 50 bp; minimum average quality score of 20) [27]. Using k-mers ranging from 21 to 121, high-quality reads were de novo assembled using SPAdes v3.15.4 [28]. The careful modes were not utilized during assembly. After assembly, scaffolding was carried out through the Multi-CSAR tool using multiple reference genomes. The completeness of the scaffolded genomes was checked using CheckM [29]. Isolate identification was conducted using LINbase [30], KmerFinder [31], and NCBI BLASTn [32]. For further validation, Type Strain Genome Server (TYGS) was used for dDDH (digital DNA-DNA hybridization) and phylogenetic analysis [33]. ANI (Average Nucleotide Identity) was calculated using ANI Calculator of Ezbiocloud.net [34].

2.4. Genome annotation, pathogenicity detection and genome organization

The assembled sequence for the isolate underwent annotation through Prokka [35] and Patric server [36]. Prediction of pathogenicity of the bacterial isolate towards human hosts was carried out through the PathogenFinder (v2) [37], a webtool of the Center for Genomic Epidemiology. Besides, 18 available genomes of S. nematodiphila were downloaded in FASTA format from Patric Server (Supplementary Table 11) and their pathogenicity score were also determined using the same tool. Next, their relative pathogenicity towards human hosts were also compared. The graphical maps of the circular genomes were generated through Proksee Server (Previously CGviewer) [38] which is an expert system for genome assembly, annotation and visualization.

2.5. Investigation of antibiotic resistance genes (ARGs), virulence factor genes (VFGs), heavy metal resistance genes and subsystems profiling

Antibiotic resistance genes (ARGs) were investigated through the ResFinder tool [[39], [40], [41]] of the Center for Genomic Epidemiology. The obtained result was further verified by analyzing the ARGs through CARD under Proksee server [42] and CARD-RGI (Resistance Gene Identifier) web platform [43]. Mutation in outer membrane porin protein, Omp, was also detected by using CARD-RGI. The three-dimensional structure of Omp was predicted through Alphafold3 [44] and visualized by using PyMOL v 3.1.6.1 [45]. The virulence factor genes (VFGs) were explored with Victors [46], which is a novel, manually curated, web-based integrative knowledge base and analysis resource for VFs of pathogens that cause infectious diseases in humans and animals. Afterwards, the VFs were also investigated with PATRIC_VF [36] and VFDB [47]. Virulence factors were also represented in the genome map of KS1 using Proksee. Heavy metal resistance genes were determined by using BacMet [48]. The subsystems of the genomes were predicted through the Patric Server analysis [36].

2.6. Mobile genetic elements (MGEs), phage and CRISPR/Cas system investigation

The MGEs from the whole genome of the isolate was investigated through the mobileOG-db (betarix 1.6) under Proksee tool [49]. Phage integration inside the bacterial genome as well as in the plasmids was inspected through the PHASTEST webserver [50] and VirSorter [51] tools. CRISPR/Cas system in the bacterial whole genome was investigated through the CRISPR/Cas Finder algorithm [52].

2.7. Pangenome & pan resistome analysis

Pangenome and comparative genome analysis were carried out in order to find out the diversity of the target strains and to reveal gene or gene family presence absence variations (PAVs) among strains from different origins and countries. A total of eighteen whole genome sequences of S. nematodiphila reported from different countries were downloaded from the Patric database. Annotation of the genomes was carried out using Prokka. The gff3 files were subjected to being prepared for pangenome analysis. Pangenome analysis of total 18 isolates (including the whole genome sequenced isolate of this study) was carried out using Roary [53], a tool that rapidly builds large-scale pangenomes identifying the core and accessory gene. During the analysis, minimum percentage identity for blastp was set to 95, maximum number of clusters was set to 50000 and percentage of isolates a gene must be in to be core was considered to be 99.0. Next the pangenome exhibiting the presence absence of genes were plotted using roary_plots.py function (https://github.com/sanger-pathogens/Roary/tree/master/contrib/roary_plots). All the genes in each genome were compared based on the pangenome analysis. In addition, the genome-wide resistance pattern of all available S. nematodiphila genomes and the sequenced genome of this study were analyzed and plotted using PRAP (Pan Resistome Analysis Pipeline) tool [54]. The PRAP analysis was performed using a k-mer size of 25, two search kernels, and a depth threshold of 20, with a minimum area score of 100 and at least 90% coverage by length. For BLASTn and BLASTp searches within PRAP, identity thresholds were set at 95% and 98%, respectively, with a minimum 90% coverage requirement for both.

3. Results

3.1. Isolation, preliminary identification and prevalence of Serratia spp. from hospital environment

Serratia spp. were identified based on their distinct red pigmented, medium sized, circular shaped colonies on cetrimide agar plates (Supplementary Fig. 1). Among 78 samples collected from different hospitals across Dhaka city (Supplementary Table 1), 23 samples tested positive for Serratia spp., indicating an overall prevalence of approximately 29%. The reported 23 isolates represent one isolate recovered per positive sample. The prevalence of Serratia spp. in seven hospitals each is depicted in Supplementary Fig. 2. The highest prevalence of Serratia spp. was experienced in MCHTI (66.67%), followed by SSMC (36.36%). Five (21.74%) of the 23 Serratia positive samples were collected from hospital soil while the remaining 18 (78.26%) isolates were obtained from hospital sewage showing its prevalence in hospital associated environmental settings. Besides, the results of several biochemical tests predicted that these isolates belong to the genus Serratia spp. The identification results obtained from biochemical tests were further validated through molecular detection by means of an assay targeting the luxS gene involved in AI-2-dependent quorum sensing (Supplementary Fig. 3).

3.2. Interpretation of antibiotic susceptibility test and biofilm forming capacity

The antimicrobial resistance pattern of 15 antibiotics belonging to 12 different classes were determined for all the isolates. Isolates that showed resistance to at least one agent from at least three antimicrobial groups, exempting the classes of antibiotics to which Serratia spp. is intrinsically resistant, are regarded as multidrug-resistant (MDR) isolates[12]. In this study, 11 of the 23 isolates (47.83%) fell under this category (Supplementary Figs. 4 and 5). On the contrary, biofilm formation assay revealed that out of the 23 isolates examined, one (4.35%) isolate was non-biofilm former (NBF), while 11(47.83%) isolates were classified as weak biofilm former (WBF) and the remaining 11(47.83%) as moderate biofilm former (MBF).

3.3. Whole genome sequence-based identification, sequence annotation and mapping

KmerFinder, LINbase and SpeciesFinder, used for whole genome sequence-based identification, revealed the isolate to be Serratia nematodiphila (S. nematodiphila). Further, dDDH (digital DNA-DNA hybridization) (Supplementary Table 2) and phylogenetic analysis (Fig. 1) and ANI (Average Nucleotide Identity) (Supplementary Table 3) validated the identification. The scaffolded genome of the isolate was annotated using Prokka, and Patric server (Table 1). From the annotated data, the genome size, GC content, total CDS, rRNA, tRNA, genome completeness and more features of the genome was revealed. The genome size was 4994812 bp and draft genome annotation revealed 369 contigs with 5478 CDS. The completeness of the genome, coarse and fine consistency and low contamination indicated the good quality of the genome. The genome map of S. nematodiphila KS1 is depicted in Fig. 2.

Fig. 1.

Phylogenomic relationship of S. nematodiphila KS1 with closely related Serratia species based on whole-genome analysis.

Table 1.

General features of assembled whole genome of Serratia nematodiphila KS1.

Patric annotation	Serratia nematodiphila KS1	Prokka annotation	Serratia nematodiphila KS1
Coarse consistency (%)	98.7	Bases	4994812
Completeness (%)	99.4	Contigs	630
Contig count	630	CDS	4737
DNA size (bp)	4994812	rRNA	4
CDS	5478	tRNA	69
tRNA	76	tmRNA	1
rRNA	3	Repeat region	2
Contigs N50 (bp)	16916
Contigs L50	89
Predicted Roles	1498
Completeness Roles	321
Total Distinct Roles	3675
Protein-Encoding Genes with Functional Assignment	4016
Protein-Encoding Genes without Functional Assignment	1462
% Protein-Encoding Feature Coverage	109.67
% Features that are Hypothetical	26.69
% Features that are in Local Protein Families	95.58
#The variance is due to different ORF-calling algorithms (RAST/GLIMMER vs. Prodigal)

Fig. 2.

Genome map of Serratia nematodiphila KS1.

Annotation revealed the presence of pigA, pigC, pigD, pigE, pigF, pigG, pigH, pigI, pigJ and pigM genes along with regulatory gene cueR that are responsible for the synthesis of Prodigiosin in the genome of S. nematodiphila KS1 (Fig. 3).

Fig. 3.

Position of pig genes and cluster of pig genes involved in MBC biosynthetic pathway regulated by cueR.

3.4. Antimicrobial resistance genes, mobile genetic elements and prophage investigation

Investigation of antimicrobial resistance genes of the isolate S. nematodiphila KS1 revealed a total of 17 genes responsible for conferring resistance to different antibiotics (Supplementary Table 4). Among them 7 genes (adeF, KpnH, KpnF, emrR, msbA, qacG, tetA(41)) encode efflux pumps. Analysis through CARD-RGI demonstrated the presence of G133D mutation in outer membrane porin protein, OmpC, that is predicted to be responsible for conferring resistance to carbapenems, cephalosporins, β-lactams and monobactams. Structural analysis revealed the presence of G133D mutation in the loop region of porin protein (Supplementary Fig. 6). The rest of the genes are responsible for the resistance to different antibiotic classes that is exhibited in Supplementary Table 5. The mapping of antimicrobial resistance genes detected from the whole genome sequenced isolate using CARD (Comprehensive Antibiotic Resistance Database) under Proksee server is depicted in Fig. 4a. The investigation of mobile genetic element by mobileOG-db identified a total of 267 features related to mobile genetic elements and their functions. There was a significant link between mobile genetic elements and antibiotic resistance genes, emphasizing the risk of their dissemination through horizontal gene transfer. A focused annotation view using MobileOG-db along with CARD showed that GlpT mutant gene conferring resistance to fosfomycin and FosA8 were in close proximity to integrase XerD and several recombination-repair genes (uvrD, recQ) indicating a significant potential of horizontal gene transfer (Fig. 4d). The association of aminoglycoside resistant gene, AAC(6′)-Ic, and recombination/repair related genes dnaG and mug indicated its recombination-driven dissemination (Fig. 4e). The β-lactamase gene, bla_SRT-2 resided closely with recombination related mobile genetic element, sbcB, suggesting acquired resistance through horizontal gene transfer (Fig. 4h). Analysis by VirSorter under Proksee server revealed the presence of 4 prophage sequences in the genome of S. nematodiphila KS1 and no antibiotic resistance gene was present inside the prophage sequences (Supplementary Fig. 7). One incomplete and another questionable prophage region were found out using PHASTEST tool (Supplementary Table 6) which also annotated viral protein (tail protein, fiber protein etc.) encoding genes and identified most common phages based on sequence similarity. The mapping of prophage sequence is depicted in Supplementary Fig. 8.

Fig. 4.

Distribution of antimicrobial resistance genes, mobile genetic elements and prophage integration in the genome of Serratia nematodiphila KS1. (a) Antimicrobial resistance genes mapping. (b) The organization of mobile genetic elements. (c) Mapping of mobile genetic elements associated with antimicrobial resistance genes. (d,e,f,g,h) The potential for horizontal transfer of antibiotic resistance gene based on their genomic proximity to mobile genetic element.

3.5. Pathogenicity, virulence factor genes and heavy metal resistance genes analysis

The pathogenicity determination of the KS1 strain using PathogenFinder tool predicted it to be human pathogen exhibiting the pathogenicity score of 0.972. The virulence factors investigation revealed a plethora of virulence genes which play role in the pathogenicity of S. nematodiphila (Supplementary Table 7). Genes associated with motility and adherence including fliA, fliQ, fliR and fliM were identified. Efflux and transport-related genes (nqrF, carB, tktA, and znuB) and metabolism-associated genes (purD and aroA) were also detected. Besides, iron acquisition genes, including znuB and yfeA, were found in the genome of S. nematodiphila. The presence of stress response genes including arcA, dnaK, and ruvB, and regulatory genes rpoS and phoU were also confirmed. In addition, virulence-associated genes focA and slyD, invasion genes vacJ and clpP, and immune evasion genes cheY and hns were identified. Secretion system components vacJ and prfC were also detected. The virulence genes of S. nematodiphila are mapped in Supplementary Fig. 9.

The BLAST search against the experimentally confirmed database and predicted database of BacMet detected a total of 23 heavy metal resistance genes along with 23 biocide resistance genes in the genome of S. nematodiphila KS1. The collection of metal and biocide resistance genes along with corresponding metal or biocide of resistance is provided in Supplementary Tables 8 and 9 pstA, pstB, pstC, pstS and glpF genes were detected conferring resistance to arsenic. Besides, cadmium (dsbA, yhcN), iron (fetB/ybbM, yfeD, yfeC, yfeB, yfeA), manganese (corA, yfeD, yfeC, yfeB, yfeA) and zinc (zur/yjbK, zntR/YhdM, mdtB, dsbA) resistant genes were also prevalent. Besides, there is also a noticeable fact that molybdenum, tungsten (modB) and selenium (sodA) resistant genes were also observed in S. nematodiphila KS1 indicating its high ecological adaptability.

3.6. Genome wide pathogenicity profiling

The pathogenicity scores against human host of 19 S. nematodiphila isolates including our sequenced isolate of interest were determined and plotted in Fig. 5. According to the figure, the highest pathogenic potential was observed in S. nematodiphila strain CGMCC_1.6853 exhibiting the pathogenicity score of 0.9778, followed by S. nematodiphila strain S8 (0.9768 and S. nematodiphila strain MB307 (0.9764). The isolate of our study, KS1, displayed moderate pathogenicity (0.972).

Fig. 5.

Genome wide Pathogenicity score of 19 Serratia nematodiphila isolates.

3.7. Subsystems and CRISPR/Cas system investigation

Subsystem analysis showed that number of genes are involved in different subsystems of the isolate. Subsystem analysis of S. nematodiphila KS1 reveals that “Metabolism” superclass constitutes the largest functional category (1032 genes), followed by “Energy production” superclass (331 genes) and “Stress response, defense & virulence” superclass (222 genes) (Supplementary Table 10). The total gene count of every subsystem class is depicted in Fig. 6. In our study, CRISPR–Cas Finder identified multiple Cas gene clusters and CRISPR arrays distributed across the genome, consistent with a Type I-F CRISPR–Cas system and the prevalence of Cas3 clusters and csy genes (Supplementary Figs. 10 and 11).

Fig. 6.

Total gene counts of every subsystem class.

3.8. Pangenome and comparative genome analysis

For pangenome and comparative genome analysis, a total of 19 genomes including the whole genome sequenced isolate analyzed in this study were selected. Other genomes of S. nematodiphila, selected for the pangenome analysis, were reported from various global locations. The accession IDs of 18 S. nematodiphila reference genomes are provided in Supplementary Table 11. While initial screening via Patric identified 5478 CDS, the pangenome and resistome analyses were based on the 4737 CDS predicted by Prokka to maintain consistency across the 19-strain dataset. Roary pangenome analysis exhibited a matrix of 13665 gene clusters in which there are 2871 core genes and 2634 shell genes. No softcore genes were observed and 8160 gene clusters belonged to cloud genes (Table 2). The core/pan ratio is 0.21 indicating the pangenome is open and carrying out the phenomenon of gene acquisition that ultimately contribute to antimicrobial resistance and virulence pattern.

Table 2.

Pangenome analysis of Serratia nematodiphila KS1.

Criteria	Number of Genes
Core genes (99%≤strains≤100%)	2659
Soft core genes (95%≤strains < 99%)	0
Shell genes (15%≤strains < 95%)	2794
Cloud genes (0%≤strains < 15%)	8283
Total genes (0%≤strains≤100%)	13736

Pangenome analysis depicted in Fig. 7 stated that S. nematodiphila strain KS1 is more closely related to the S. nematodiphila strain TM-17, reported from India, forming a distinct clade at the top of the tree. Besides, the repertoire of accessory genes in S. nematodiphila strain KS1 was lower compared to other global strains representing limited genomic variability.

Fig. 7.

Pangenome and comparative genome analysis of 19 different strains of Serratia nematodiphila worldwide including the test isolate of this study.

3.9. Investigation of genome wide antimicrobial resistance pattern using PRAP

The overall resistome and acquired resistome of 19 S. nematodiphila strains was analyzed by using the tool, PRAP (Pan Resistome Analysis Pipeline), based on the annotation of antimicrobial resistance genes by CARD and ResFinder each. ResFinder represents only acquired resistance genes associated with horizontal gene transfer, transposition etc. whereas CARD database is comprised of all possible genes and mutations conferring resistance, efflux pump genes etc. [54] Fig. 8 exhibited each type of resistome across all the genomes of S. nematodiphila, delineated by PRAP. The deeper the color of the block in the heatmap, the more antibiotic resistance genes of corresponding class present in that particular strain. S. nematodiphila strain JBAAAEI-19-0004 has higher resistance profile than other strains. It is noticeable from these plots that β-lactamases genes were acquired through horizontal gene transfer or other mechanisms in S. nematodiphila strains. Among β-lactamase genes, bla_SST was most prevalent in S. nematodiphila strains (Fig. 9). Moreover, the distribution of resistance genes related to each antibiotic class, and resistance gene heatmaps regarding other antibiotics, according to CARD and ResFinder database, are provided in the Supplementary Figs. 12–17 and 18-20 respectively.

Fig. 8.

Distribution of antibiotic resistance genes across different classes of antibiotics among analyzed strains. (A) Acquired ARGs Resistome in Serratia nematodiphila genomes; (B) Overall Resistome of Serratia nematodiphila genomes based on CARD. Here the color bar represents the “Gene copy number” at antibiotic resistance level.

Fig. 9.

Distribution of ARGs associated with carbapenem and β-lactam antibiotics. (A) Carbapenem resistance across Serratia nematodiphila genomes; (B) β-lactam resistance across Serratia nematodiphila genomes.

4. Discussion

Throughout the study, among a total of 78 samples collected from different hospitals across Dhaka city, 23 samples were tested positive for Serratia spp. based on the growth and characteristic pattern of the isolates on cetrimide agar. These isolates produced distinct red pigmented colonies on cetrimide agar and were primarily selected for the study. Through biochemical and cultural approaches, the isolates were presumptively identified as Serratia spp. followed by molecular confirmation via genus-specific PCR targeting quorum-sensing gene luxS.

Antimicrobial susceptibility test revealed the presence of 11 multi-drug resistant (MDR) isolates (47.8%) that show resistance to at least three antimicrobial classes. The growing prevalence of MDR Serratia spp. poses a significant public health challenge as these bacteria are increasingly being implicated in nosocomial infections, especially in intensive care units (ICUs) where patients are already immunocompromised [7]. Carbapenem resistance was more prevalent with 16 isolates (including 9 MDR) resistant to meropenem and 3 MDR isolates resistant to imipenem. As carbapenems are widely considered as the last line of defense against Gram-negative infections, and resistance to this class is particularly worrisome in hospital environments.

A total of 22 isolates were biofilm former with the majority of MDR strains exhibiting moderate biofilm-forming capacity. Biofilms help bacteria persist and cause chronic infections by shielding them from host defenses and antimicrobial treatments. This finding is consistent with previous reports that link biofilm formation with increased antibiotic resistance in Serratia marcescens, particularly to β-lactams, aminoglycosides, and fluoroquinolones [55,56].

Comprehensive insights on the resistance and virulence profile of the isolate S. nematodiphila KS1 was obtained through whole genome sequencing. Genome identification through KmerFinder, LINbase and SpeciesFinder revealed the isolate to be S. nematodiphila. All three tools independently identified isolate KS1 as S. nematodiphila ensuring the reliability of the results. Using complex genetic algorithms, these tools are able to cross check the isolate's genetics with reference databases, identifying the species conclusively. Genome annotation through Prokka and Patric server provided information about genome size, GC content, total CDS, rRNA, tRNA, genome completeness and more features of the genome.

Prodigiosin in Serratia spp. is mainly produced from the condensation of two precursors, 2- methyl-3-n-amyl-pyrrole (MAP) and 4-methoxy-2,2′-bipyrrole-5-carbal-dehyde (MBC). Among the pig genes identified in the genome of S. nematodiphila KS1, pigA, pigF, pigG, pigH, pigI and pigM were MBC encoding genes whereas pigD and pigE were involved in the expression of MAP. Besides, condensing enzyme encoding pigC gene was also observed [57].

Numerous antimicrobial resistance (AMR) genes were found in the genome, such as β-lactamases, aminoglycoside resistance gene (AAC(6′)-Ic), and Fosfomycin resistance genes. Furthermore, the identification of genes linked to the efflux pump (adeF, KpnF, and KpnH) emphasizes the significance of efflux-mediated resistance. These findings collectively corroborate the multidrug-resistant (MDR) nature of the isolate. In addition, it is also noticeable that CARD database and Prokka annotation identified the presence of arnT gene arnBCADTEF operon in Serratia nematodiphila KS1 which is predicted to be responsible for intrinsic resistance to colistin. The arnT gene encodes lipid A modifying enzyme through the addition of a cationic sugar, 4-amino-4-deoxy-L-arabinose (L-Ara4N) that ultimately confers colistin resistance [58].

An internal substitution, G133D mutation, was found in outer membrane porin protein, OmpC. It is anticipated that Gly to Asp mutation in the loop region of porin protein incorporates negative charge in place of glycine residue that may cause significant changes in porin properties such as, lower conductance and restricted opening [59]. As a result, the uptake of antibiotics, particularly cephalosporins is inhibited. This phenomenon may also contribute to the resistance to carbapenems. Similar phenomenon has also been observed in two strains of Enterobacter aerogenes due to the Gly to Asp mutation in L3 porin domain of Omp36 [60]. For confirmation, further expression and detailed experimental study is required to find out the exact reasons behind the carbapenem resistance of this strain.

Among other antibiotic resistance genes, FosA8 provides Fosfomycin resistance through enzymatic modification of the antibiotic [61] whereas the vanG gene confers vancomycin resistance by altering peptidoglycan precursors [62]. In reaction to environmental cues, regulatory genes like rsmA alter the expression of virulence and resistance determinants [63]. The identification of AdeF (AdeFGH system) demonstrated efflux-mediated resistance to ciprofloxacin, chloramphenicol and tetracycline ([64]. Tetracycline resistance through efflux pump mediated activity was contributed by tetA(41) which was also reported in Serratia marcescens before [65,66]. Resistance to aminoglycosides is conferred by aminoglycoside 6′-N-acetyltransferase enzyme encoding gene AAC(6′)-Ic which encodes enzyme that catalyze the acetylation of -NH2 group in aminoglycoside [67].

Mobile genetic element investigation reveals the presence of several MGE-associated genes that act as ‘critical dirvers” in the dissemination of antimicrobial resistance genes [49]. Proximity analysis showed that specific resistance genes (vanG, aac(6')-Ic, bla_SRT-2, msbA etc.) were located near integrases and other MGE-associated genes that strongly suggest their potential for dissemination through horizontal gene transfer.

The identification of several virulence factors in S. nematodiphila KS1 demonstrates its strong pathogenic potential. The presence of genes related to adherence, efflux pumps, secretion systems and iron uptake indicates that the isolate has the ability to persist in challenging environments and create a favorable environment for evading the host, combating the host's immune systems and drugs and promoting pathogenicity. Genes responsible for invasion and immune evasion were also found in its genome that helps the bacteria to enter into the host, evade the host immune systems and simultaneously carry the progression of pathogenicity.

The diverse array of heavy metal resistance genes was observed in the genome of S. nematodiphila KS1 representing its adaptability in wide range of environments and hospital settings. Genes such as corA, corC, corD, zntR/yhdM, cueR/ybbI, modB, fetB/ybbM, and dsbA/B indicated the resistance to transition metals, such as, cobalt, zinc, copper, molybdenum, mercury and cadmium, which are enormously used in industrial processes and hospital infrastructure. Among them, copper and zinc resistance systems are very concerning for public health. Nowadays metallic Cu is used as an antimicrobial surface on medical devices and so Cu-resistant bacteria are a threat for the hospitalized patients [68]. Besides, the association between zinc/cadmium resistance and macrolides along with aminoglycosides was also brought to light [69].

The analysis of subsystems and metabolic pathways revealed a significant number of subsystems and associated genes, along with various key metabolic pathways. The subsystem superclass “Metabolism”, with 1032 genes, was the most prominent, highlighting its critical role in energy production and bacterial growth. “Membrane transport” (200 genes) and “Cellular processes” (161 genes) are also essential for maintaining cell integrity and nutrient exchange. The presence of 222 stress response, defense and virulence-related genes illustrates how the bacterium adapts to adverse conditions and its potential virulence.

In our study, type I-F CRISPR/Cas system was detected in the genome of S. nematodiphila KS1. This is a form of adaptive immune system that guards the bacterium against foreign genetic material such as bacteriophages and plasmids, enabling them to fend off infections, uphold the integrity of their genome and limit the effectiveness of specific phage therapy [70,71].

The pangenome analysis of our whole genome sequenced isolate of interest with other 19 isolates of S. nematodiphila from different sources and countries revealed a significant genomic diversity and depicted a comparative scenario of the isolates of S. nematodiphila worldwide. The pangenome based phylogenetic analysis revealed the close relation of S. nematodiphila KS1 with the genome of S. nematodiphila strain TM-17 reported from India. The analysis also revealed that the pangenome is still growing and the genomic diversity of the isolates of S. nematodiphila are quite distinct and diverse as well.

Genome wide in-silico pathogenicity analysis also warned that S. nematodiphila could be a threat for immunocompromised patients in future as 12 isolates demonstrated their pathogenicity score above 0.97 indicating a rich set of virulence genes and an elevated probability to cause infection in humans as opportunistic pathogens. The finding demonstrates a hospital environmental reservoir that could lead to infections, but this study does not demonstrate active clinical cases.

PRAP analysis demonstrated the presence of both acquired and intrinsic resistance to aminoglycosides in S. nematodiphila KS1 and other isolates. Two different aminoglycoside resistance genes, aad and Hangzhou variant of aac(6′)-Ib, was detected in S. nematodiphila strain JBEAAAI-19-0004. In case of resistance to cephalosporins, S. nematodiphila KS1 and other 12 isolates contained bla_SRT-2 gene, encoding AmpC type β-lactamases. Another MDR isolate, S. nematodiphila strain GN-140, was isolated from the blood sample of a hospitalized neonate exhibiting higher pathogenicity score [72]. Besides, its pattern of resistance genes is higher than the pattern of KS1 in PRAP analysis.

To summarize the findings of this study, the presence of multidrug-resistant (MDR) Serratia spp. in hospital settings is a growing concern, especially among immunocompromised individuals. This bacterium exhibits intrinsic resistance to colistin and is increasingly acquiring resistance to carbapenems, which are considered last-resort antibiotics. This development heightens the risk of treatment failure, as well as associated morbidity and mortality. The underlying cause of the alarming prevalence of MDR Serratia spp. in hospital environments could be the extensive use of antibiotics to which Serratia spp. already exhibit intrinsic resistance, enabling the organism to thrive in these settings. This study also highlighted a critical concern by identifying S. nematodiphila for the first time in hospital environment in Bangladesh.

5. Conclusion

Nosocomial infections caused by multidrug-resistant (MDR) Serratia spp. are a significant public health concern due to the high morbidity, mortality, and expense of treating these infections. This study provides a pioneering report on the identification, characterization, and occurrence of MDR Serratia spp. in hospital sewage and soil in Bangladesh, utilizing cultural, molecular, and whole genome sequence analysis approaches. The findings highlight the epidemiology and genetic features of these resistant strains in the local context. Intrinsic resistance of this bacterium to colistin as well as acquired resistance to carbapenems through the expression of efflux pumps signify a concerning situation as these antibiotics are considered as last line resort to combat this type of pathogen. The detection of various antibiotic and antiseptic-resistance genes also points out the potential for elevated resistance against antibiotic treatment and general sanitation practices. This complicates treatment and prevention strategies for infection by these bacteria. In addition, these isolates manifested a wide variety of virulence determinants, including biofilm formation, hemolysis, iron uptake, secretion systems, and efflux pumps, indicating their high potential as human pathogens. The presence of mobile genetic elements and associated AMR genes indicates an augmented threat for the effective transmission and uptake of resistance and virulent determinants among the isolates. The occurrence of CRISPR/Cas systems in the bacterial genome suggests a sophisticated defense mechanism exploited by the isolates against phage therapy. For these reasons, further investigation and comprehensive analysis are necessary to examine the frequency of Serratia spp. in Dhaka, Bangladesh, elucidate their pathogenicity and resistance mechanisms, and uncover potentially enhanced and alternative methods of therapy.

CRediT authorship contribution statement

Inshira Afrose Setu: Writing – original draft, Visualization, Validation, Software, Methodology, Investigation, Formal analysis, Data curation. Atiq Abrar Rahman: Writing – original draft, Visualization, Validation, Software, Methodology, Investigation, Formal analysis, Data curation. Mohammad Azizul Hoque: Writing – review & editing, Validation, Software, Methodology, Investigation, Formal analysis. Md Mehedi Morsalin: Writing – review & editing, Validation, Methodology, Investigation, Data curation. Sikder Sajjad Hossain Tushar: Writing – review & editing, Validation, Software, Methodology, Investigation. Spencer Mark Mondol: Writing – review & editing, Visualization, Validation, Software, Project administration, Methodology, Formal analysis, Conceptualization. Md Rafiul Islam Ranga: Writing – review & editing, Validation, Software, Resources, Project administration, Investigation, Conceptualization. Hussain Md Shahjalal: Writing – review & editing, Validation, Software. Donald James Gomes: Writing – review & editing, Validation, Resources, Conceptualization. Md Mizanur Rahaman: Writing – review & editing, Validation, Supervision, Software, Resources, Project administration, Investigation, Funding acquisition, Conceptualization.

Funding

This work has received partial funding from the University Grants Commission in collaboration with University of Dhaka.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

The draft genome sequences of the isolate were deposited in DDBJ/ENA/GenBank of NCBI under the BioProject PRJNA1356648 and BioSample Accession SAMN53065043. Besides, the other genomes used for secondary analysis were retrieved from NCBI (https://www.ncbi.nlm.nih.gov/) and BV-BRC (https://www.bv-brc.org/).