TYPE Original Research

PUBLISHED 07 March 2024

DOI 10.3389/fmicb.2024.1329620

Nanopore sequencing for

OPEN ACCESS identification and
characterization of
EDITED BY
Lucilla Iacumin,
University of Udine, Italy

REVIEWED BY
Keshani Bhushan,
antimicrobial-resistant
Punjab Agricultural University, India
Manyun Yang,
Harvard University, United States
Escherichia coli and Salmonella
*CORRESPONDENCE
Shafiq Rheman
spp. from tilapia and shrimp sold
[email protected]
Jérôme Delamare-Deboutteville
[email protected]
at wet markets in Dhaka,
†
These authors have contributed equally to
this work
Bangladesh
RECEIVED 29 October 2023
ACCEPTED 09 February 2024 Shafiq Rheman 1*†, Sabrina Hossain 1†, Md Samun Sarker 2,
PUBLISHED 07 March 2024
Farhana Akter 1, Laura Khor 3, Han Ming Gan 4, Andy Powell 5,6,
CITATION
Rheman S, Hossain S, Sarker MS, Akter F, Roderick M. Card 7, Yaovi Mahuton Gildas Hounmanou 8,
Khor L, Gan HM, Powell A, Card RM, Anders Dalsgaard 8, Chadag Vishnumurthy Mohan 3,
Hounmanou YMG, Dalsgaard A, Mohan CV,
Bupasha ZB, Samad MA, Zamila Bueaza Bupasha 2, Mohammed A. Samad 2,
Verner-Jeffreys DW and
Delamare-Deboutteville J (2024) Nanopore David W. Verner-Jeffreys 5,6 and
sequencing for identification and Jérôme Delamare-Deboutteville 3*
characterization of antimicrobial-resistant
Escherichia coli and Salmonella spp. from 1
Laboratory Department of Sustainable Aquaculture, WorldFish, Dhaka, Bangladesh, 2 Antimicrobial
tilapia and shrimp sold at wet markets in Resistance Action Center (ARAC), Animal Health Research Division, Bangladesh Livestock Research
Dhaka, Bangladesh. Institute, Savar, Bangladesh, 3 Department of Sustainable Aquaculture, WorldFish, Penang, Malaysia,
Front. Microbiol. 15:1329620. 4
Patriot Biotech Sdn Bhd, Bandar Sunway, Malaysia, 5 Weymouth Laboratory, Cefas: Centre for
doi: 10.3389/fmicb.2024.1329620 Environment Fisheries and Aquaculture Science, Weymouth, United Kingdom, 6 Veterinary Medicines
COPYRIGHT
Directorate FAO Reference Centre for Antimicrobial Resistance, Weybridge, United Kingdom,
7
Bacteriology Department, Animal Plant Health Agency, Weybridge, United Kingdom, 8 Department of
© 2024 Rheman, Hossain, Sarker, Akter, Khor,
Veterinary and Animal Sciences, Faculty of Health and Medical Sciences, University of Copenhagen,
Gan, Powell, Card, Hounmanou, Dalsgaard,
Frederiksberg, Denmark
Mohan, Bupasha, Samad, Verner-Jeffreys and
Delamare-Deboutteville. This is an open-
access article distributed under the terms of
the Creative Commons Attribution License Wet markets in low-and middle-income countries are often reported to
(CC BY). The use, distribution or reproduction have inadequate sanitation resulting in fecal contamination of sold produce.
in other forums is permitted, provided the
original author(s) and the copyright owner(s)
Consumption of contaminated wet market-sourced foods has been linked
are credited and that the original publication to individual illness and disease outbreaks. This pilot study, conducted in two
in this journal is cited, in accordance with major wet markets in Dhaka city, Bangladesh during a 4-month period in 2021
accepted academic practice. No use,
distribution or reproduction is permitted
aimed to assess the occurrence and characteristics of Escherichia coli and non-
which does not comply with these terms. typhoidal Salmonella spp. (NTS) from tilapia (Oreochromis niloticus) and shrimp
(Penaeus monodon). Fifty-four individuals of each species were collected. The
identity of the bacterial isolates was confirmed by PCR and their susceptibility
toward 15 antimicrobials was tested by disk diffusion. The whole genome of
15 E. coli and nine Salmonella spp. were sequenced using Oxford Nanopore
Technology. E. coli was present in 60–74% of tilapia muscle tissue and 41–
44% of shrimp muscle tissue. Salmonella spp. was found in skin (29%) and gills
(26%) of tilapia, and occasionally in muscle and intestinal samples of shrimp.
The E. coli had several Multilocus sequence typing and serotypes and limited
antimicrobial resistance (AMR) determinants, such as point mutations on glpT
and pmrB. One E. coli (BD17) from tilapia carried resistance genes for beta-
lactams, quinolones, and tetracycline. All the E. coli belonged to commensal
phylogroups B1 and A and showed no Shiga-toxin and other virulence genes,

Frontiers in Microbiology 01 frontiersin.org

Rheman et al. 10.3389/fmicb.2024.1329620

confirming their commensal non-pathogenic status. Among the Salmonella

isolates, five belonged to Kentucky serovar and had similar AMR genes and
phenotypic resistance patterns. Three strains of this serovar were ST198, often
associated with human disease, carried the same resistance genes, and were
genetically related to strains from the region. The two undetermined sequence
types of S. Kentucky were distantly related and positioned in a separate
phylogenetic clade. Two Brunei serovar isolates, one Augustenborg isolate, and
one Hartford isolate showed different resistance profiles. This study revealed
high fecal contamination levels in tilapia and shrimp sold at two main wet
markets in Dhaka. Together with the occurrence of Salmonella spp., including S.
Kentucky ST198, a well-known human pathogen, these results stress the need
to improve hygienic practices and sanitation standards at markets to improve
food safety and protect consumer health.

KEYWORDS

Salmonella, Escherichia coli, tilapia, shrimp, antimicrobial resistance, whole genome

sequencing, food safety, Bangladesh

Introduction and their presence in aquatic products has been suggested as an

important source of AMR (Khan et al., 2022; Samad et al., 2023).
Aquaculture production of tilapia and shrimp in Bangladesh has The origin of AMR in aquatic products can be attributed to
experienced significant growth in recent years, driven by increasing various factors, including AMU in aquaculture, contamination with
demands from domestic consumers and export markets (Haque and human and animal fecal matter containing AMR bacteria, and cross-
Belton, 2021; Debnath et al., 2023; El-Sayed and Fitzsimmons, 2023). contamination with AMR bacteria during processing and handling
However, this expansion has raised concerns regarding antimicrobial (Shikongo-Nambambi et al., 2012; Neela et al., 2015; Watts et al.,
use (AMU) practices and food safety aspects within the industry. 2017). The presence of AMR bacteria and associated genes in aquatic
Several studies highlighted the widespread and indiscriminate use of products raises food safety concerns, especially since aquatic products
antimicrobials in tilapia and shrimp farms, with a lack of veterinary are often consumed with minimal cooking or even in their raw state.
supervision and adherence to proper dosage and withdrawal periods Resistance genes can subsequently be transferred among bacterial
(Thornber et al., 2020; Dewi et al., 2022). This unregulated AMU in populations in the human intestine, leading to the spread of AMR and
aquaculture settings has been linked to the emergence and spread of potential treatment failure of bacterial infections in humans (Okon
antimicrobial resistance (AMR) in bacterial populations, including et al., 2022).
those present in aquatic environments and seafood products, which E. coli and Salmonella spp. are two major pathogens causing
can enter the food chain and potentially represent a risk to humans serious infections in humans (Havelaar et al., 2015). E. coli causes
(Lulijwa et al., 2020; Preena et al., 2020; Thornber et al., 2020). simple gastrointestinal diseases to more severe diseases, while
Furthermore, studies have reported the presence of AMR bacteria, Salmonella spp. are responsible for illnesses, such as typhoid/
such as extended-spectrum β-lactamase-producing Escherichia coli, paratyphoid fever, and non-typhoidal Salmonella spp. (NTS) cause
Salmonella spp., and multidrug-resistant Vibrio spp., in aquaculture food poisoning. The transmission route of E. coli and Salmonella spp.
systems and associated products, posing potential risks to human to humans is mainly via fecally contaminated water and food. E. coli
health (Hamilton et al., 2018; Dewi et al., 2022). Therefore, and Salmonella spp. can develop resistance by selective pressures from
implementing robust food safety measures from culture ponds to antimicrobials and horizontally through the transmission of different
markets is crucial to ensure the sustainability of tilapia and shrimp mobile elements like plasmids and transposons (Butaye et al., 2006).
production and public health safety in Bangladesh. In this study, we aimed to develop a pilot genomic surveillance for
Wet markets are important points of sales of aquatic food products AMR in E. coli and NTS isolated from tilapia and shrimp products
in Bangladesh and, therefore, play a crucial role in food safety. In purchased at wet markets in Dhaka, Bangladesh. Growth-based
Bangladesh, hygiene issues in these markets have always been a major methods and PCR were used for bacterial genus identification and
concern. Different factors, including lack of proper sanitation and their antimicrobial susceptibility was tested by the disk diffusion
hygiene practices, inadequate storage and handling, lack of quality method. Oxford nanopore sequencing was used to characterize the
control, and extensive use of chemicals contribute to the challenges. genomes of E. coli and Salmonella spp. to confirm their serovars,
Studies have shown that wet markets often have poor hygiene practices virulence factors, and the genetic basis of resistance. The findings of
and facilities (Alam et al., 2016; Sobuj et al., 2022). In addition, the study provide important insight into the occurrence and
contaminated water may be used to process aquatic produce, leading transmission of antimicrobial-resistant E. coli and Salmonella from
to the potential contamination of such products by fecal organisms, tilapia and shrimp sold at wet markets. These findings can contribute
such as E. coli and Salmonella spp. (Rocha Rdos et al., 2014; Fernandes to initiatives aimed at minimizing the risk of transmitting fecal
et al., 2018). These bacteria can cause foodborne illnesses in humans, bacterial pathogens and AMR along the food chain.

Frontiers in Microbiology 02 frontiersin.org

Rheman et al. 10.3389/fmicb.2024.1329620

Materials and methods a sterile falcon tube. After removing the shell, a meat sample for each
shrimp (approximately 15 g) was collected. The different sample types
Selection of fish and shrimp wet markets were placed in separate sterile containers or tubes and labeled with
unique designation numbers. All collected tissues from each specimen
For this pilot surveillance study of AMR, tilapia and shrimp were processed and analyzed separately. Collected tilapia and shrimp
specimens were purchased from two wet markets, Karwan Bazar specimens were stored at 4°C within two hours upon arrival at
wholesale market (KBW), and Karwan Bazar retail market (KBR), the laboratory.
both located in the middle of Dhaka city. These markets are among
the largest wholesale and retail wet markets in Bangladesh.
Additionally, KBW is the main supplier to a range of different retail Bacterial culture and species identification
markets in Dhaka.
Each sample type was analyzed for E. coli and Salmonella spp.
All the samples were enriched in sterile trypticase soy broth (TSB,
Collection and processing of fish and Oxoid, Hampshire, UK) following a 1:10 ratio. From each tilapia,
shrimp samples approximately 25 g of muscles and 5 g of gill tissues were used and
added to 225 mL and 45 mL of TSB, respectively, in separate
The target species were tilapia (Oreochromis niloticus) and tiger stomacher bags. Approximately, 1 g of tilapia gut consisting of a
shrimp (Penaeus monodon), which are commonly sold at the markets. mixture of two pieces of foregut and hindgut was added to 9 mL TSB
Tilapia, ranging between 250 and 300 g, were mostly sold fresh or live in a 15 mL Falcon tube and the skin swab was placed into a 15 mL
whereas shrimp, 25–30 g each, were displayed on trays occasionally on Falcon tube containing 5 mL of TSB. Ten grams of shrimp muscle
ice by the traders at the markets. A total of 3 separate visits were made tissue were added to 90 mL TSB in a stomacher bag and the whole
to KBR to collect 27 tilapia. During each visit, nine fresh tilapia were intestine of the shrimp was added into a 15 mL Falcon tube
collected from one vendor; however, the vendor changed between the containing 3 mL of TSB. All the samples were incubated for 18 ± 2 h
visits. Similarly, four separate visits were made to KBW to collect 27 at 37 ± 1°C.
fresh tilapia. For shrimp sample collection, two visits were made to A loopful of enriched broth culture was streaked onto different
KBR and two visits were made to KBW to collect 54 shrimp samples selective and indicative media as follows. E. coli was isolated on eosin
(27 from KBR and 27 from KBW). During each sampling, samples methylene blue (EMB; Oxoid, Hampshire, UK) agar plates, which
were collected from one single vendor; however, vendors were were incubated at 35–37°C for 18–24 h. Salmonella spp. was isolated
different at every visit. Sample numbers included: KBW: tilapia on xylose lysine deoxycholate agar (XLD; Oxoid, Hampshire, UK)
(n = 27), shrimp (n = 27); KBR: tilapia (n = 27), shrimp (n = 27). The plates incubated at 35–37°C for 18–24 h.
number of shrimps purchased from each vendor between the visits Presumptive E. coli colonies on EMB and Salmonella spp. colonies
was not the same. A total of 18 and 9 shrimp samples were collected on XLD agar were streaked onto nutrient agar (NA, Oxoid, Hampshire,
during visit 1 and visit 2 at KBW, respectively. In a similar way, the UK) to test and ensure purity. The indole test for E. coli and the triple
same number of shrimp samples (18 and 9 samples) were collected sugar iron (TSI) test for Salmonella spp. were done for biochemical
during visit 1 and visit 2 at KBR market. validation. Isolates were stored at-70°C in 30% glycerol with brain
Trained personnel wearing sterile plastic gloves collected the heart infusion (BHI; Oxoid, Hampshire, UK) broth. In case the DNA
samples at the markets, which were individually placed in sterile extraction from the bacterial isolates was not done on the same day, a
labeled zipper plastic bags kept in insulated boxes containing ice 10 μL loop full of bacterial suspension for each isolate was preserved/
packs. Samples were transported to Savar, Dhaka, for further fixed in a 1.5 mL tube containing 500 μL of 100% ethanol with the
processing and analysis in the Animal Health laboratory of the tubes stored at-20°C for future DNA extraction. E. coli ATCC 25922
Bangladesh Livestock Research Institute (BLRI). The duration time and S. enteritidis ATCC 4931 were used as positive controls when
from the point of sampling at the market until returning to BLRI to subculturing on selective agar media and for PCR. All selected isolates
start processing the samples was never more than 3 h. were Gram-stained following standard procedures. The identity of the
Surface skin swabs, gill, muscle, and intestinal samples were presumptive E. coli and Salmonella spp. was confirmed by PCR
obtained from tilapia using standardized methods. A sterile cotton (Table 1).
swab was used to make three long body swipes while twisting the
swab to maximize mucus collection along the body (head to tail).
The outer surface of the fish was then sterilized by wiping a gauze Antimicrobial susceptibility testing (AST)
pad soaked with 70% ethyl alcohol. Using a sterile scalpel and a
pair of forceps, the skin was removed and approximately 50 g of AST was done using the disk diffusion method following the
muscle and 10 g of gills were collected (in excess) from both sides Clinical & Laboratory Standards Institute (CLSI) guideline
of the fish. The liver and other organs were carefully removed, and VET01S-Ed5 (CLSI, 2020). Bacterial suspensions were prepared
the gut was obtained by cutting it from both extremities from the in phosphate-buffered saline solution to a 0.5 McFarland standard
pyloric caeca to the anus. and spread onto Mueller Hinton agar (MH; Oxoid, Hampshire,
The outer surface of the shrimp was disinfected as done for tilapia. UK) plates. Antimicrobial disks were placed on the surface of the
Using sterile scissors or scalpel, the intestine from the hepatopancreas MH agar plates, which were incubated at 35–37°C for 18–24 h.
to the anus was removed and the content was carefully extruded into Disks with the following antimicrobials were used: ampicillin

Frontiers in Microbiology 03 frontiersin.org

Rheman et al. 10.3389/fmicb.2024.1329620

TABLE 1 PCR primers used for the identification of E. coli and Salmonella spp.

Primer name Sequence (5′ – 3′) Amplicon size (bp) Reference

E. coli 166 (Heijnen and Medema, 2006)

uidA (F) TAT GGA ATT TCG CCG ATT TT

uidA (R) TGT TTG CCT CCC TGC TGC GG

Salmonella enterica spp. 429 (Malkawi, 2003)

ST11 AGC CAA CCA TTG CTA AAT TGG CGC A

ST15 GGT AGA AAT TCC CAG CGG GTA CTG

(AMP; 10 μg), azithromycin (AZM; 15 μg), chloramphenicol Nanopore library preparation and
(CHL; 30 μg), cefepime (CEP; 30 μg), cefoxitin (FOX; 30 μg), sequencing
ceftriaxone (CRO; 30 μg), cefuroxime sodium (CXM; 30 μg),
ciprofloxacin (CIP; 5 μg), gentamicin (GEN; 10 μg), levofloxacin Libraries were prepared using the following two rapid barcoding
(LVX; 5 μg), meropenem (MEM; 10 μg), nalidixic acid (NAL; sequencing kits. The first kit, SQK-RBK004, was used for low
30 μg), nitrofurantoin (NIT; 100 μg), norfloxacin (NOR; 10 μg), throughput, up to 24 samples per library. Approximately 400 ng of
and trimethoprim-sulfamethoxazole (SXT; 1.25/23.75 μg) (Oxoid, DNA as measured by Qubit was fragmented according to the
Hampshire, UK). Inhibition zones were measured in millimeters manufacturer’s instructions (Oxford Nanopore, UK). The barcoded
using the scan inhibition zone reader (Interscience Scan 4,000, samples were pooled in equal ratios to get a final 400 ng of genomic
Interscience, France). Interpretation of the data was done using DNA and purified using SPRI Bead, followed by ligation of the rapid
CLSI breakpoints for all antimicrobials except for nitrofurantoin, adapter (RAP). The second kit, SQK-RBK110.96, for high throughput,
of which the zone of inhibition was measured in accordance with 96 samples per library, used approximately 50 ng of DNA per sample,
the European Committee on Antimicrobial Susceptibility Testing and then samples were pooled, purified, and ligated with RAP. The
(EUCAST) (CLSI, 2020; EUCAST, 2023). Whole genome final pooled library was sequenced on a Flongle flow cell for at least
sequencing and bioinformatic analyses. 12 h to assess index distribution. Base calling and demultiplexing of
the fast5 files were both performed with Guppy v.5.0.7. Based on the
index percentage, the tagmented products were re-pooled or
DNA extraction and purification re-prepped for large-scale sequencing on the MinION flow cell and
base calling (super accuracy mode) and demultiplexing performed as
A total of 16 E. coli and 14 Salmonella spp. isolates from tilapia and before to generate the final fastq files for each of the samples prepared.
shrimp samples were selected for whole genome sequencing and further
analysis. Bacterial resuspension was pelleted, followed by the removal of
supernatant (ethanol) via decantation. DNA extraction was performed as De novo assembly
per the method of (Sokolov, 2000) with some modifications. The pellet
was resuspended in 500 μL of lysis buffer (50 mM NaCl, 50 mM Tris–HCL Raw nanopore reads were adapter trimmed with a qscore of 9 or
pH8, 50 mM EDTA, 2% SDS) and incubated at 60°C for 30 min. Then, higher using porechop (Wick et al., 2017) and then filtered to retain
3 μL RNAse A (10 mg/mL) was added to the lysate, followed by incubation reads longer than 2,000 bp using NanoFilt v2.6.0 (De Coster et al.,
at room temperature for 10 min. Salting out was performed via the 2018). The filtered Nanopore reads were assembled de novo using Flye
addition of 333 μL (2/3 vol) saturated NaCl at 4°C for 5 min. The lysate v2.9 (Kolmogorov et al., 2019), followed by one round of polishing
was centrifuged to remove precipitated proteins and lipids and the with racon (Vaser et al., 2017) and another round of polishing with
aqueous layer containing the DNA was transferred to a new tube and medaka v1.4.41. Genome assembly statistics were generated using
mixed with 100 μL of isopropanol via inversion. Fifteen μl of magnetic QUAST (Gurevich et al., 2013). Assessment of the genome
beads were added and the solution was mixed by inversion, then completeness was done using BUSCO5 (Simão et al., 2015) and
incubated at room temperature for 10 min to promote the binding of identified conserved microbial single-copy genes as listed in the
nucleic acid to the beads’ surface. Beads and DNA were separated from bacteria_odb10 database.
the remaining lysate on a magnetic rack and washed twice with 75%
ethanol. DNA elution from the bead was performed by the resuspension
of the beads with 50 μL of TE buffer followed by incubation at 50°C Bacterial species identification and
for 5 min. Salmonella serotyping
A subsample of the extracted gDNA was visualized by 2% agarose
gel electrophoresis to assess overall DNA integrity and yield. DNA was Whole genome sequence identification was performed using
purified and size-selected using 0.5X vol of magnetic beads to remove kraken2 to confirm the bacterial species of the sequenced strains
small molecular weight DNA as well as any other carry-over impurities
from the initial extraction. The DNA was subsequently measured
using a Denovix High sensitivity kit (Denovix, Wilmington, DE, USA)
and normalized to approximately 25 ng/μl. 1 https://github.com/nanoporetech/medaka

Frontiers in Microbiology 04 frontiersin.org

Rheman et al. 10.3389/fmicb.2024.1329620

TABLE 2 Occurrence of Salmonella spp. in tilapia and shrimp.

strain 201,001,922 (accession number CP028357). For this analysis,
Tilapia Shrimp single-nucleotide variants were called using Snippy version 4.6.05
under the following parameters: mapping quality of 60, a minimum
KBW* KBR** KBW* KBR**
base quality of 13, a minimum read coverage of 4, and a 75%
Skin 2 (7%) 8 (29%) NA*** NA*** concordance at a locus. We aligned core genome single-nucleotide
Gill 14 (52%) 7 (26%) NA*** NA*** variants by using Snippy Core version 4.1.0 for phylogeny inference
Muscle 0 (0%) 0 (0%) 0 (0%) 4 (15%) and detected masked putative recombinogenic regions by using
Gubbins version 2.4.1. A maximum-likelihood phylogenetic tree was
Intestine 1 (3%) 0 (0%) 0 (0%) 2 (7%)
built using RAxML version/8.2.12 and the generalized time-reversible
Sub-total 16 (14%) 16 (14%) 0 (0%) 6 (11%)
model with 200 bootstraps. The final tree was rooted on the reference
Total 38/324 (11%) genome CP028357 and visualized with Microreact.6
* KBW, Karwan Bazar wholesale market, ** KBR, Karwan Bazar retail market, *** NA, Not
applicable.

Results
(Wood et al., 2019). Out of the 16 E. coli and 14 Salmonella spp.
Isolates from tilapia and shrimp samples that were sequenced, only 15 Occurrences of E. coli and Salmonella spp.
E. coli and 9 Salmonella spp. Genomes were further analyzed based on in tilapia and shrimp
DNA quality control (e.g., high genomic completeness) (Tables 2, 3).
The genomes confirmed as Salmonella spp. were further analyzed Surface skin swab, gill, muscle, and intestinal samples from tilapia
with sistr_cmd and seqsero for confirmation of the serovar (Zhang obtained at the KBW and KBR contained E. coli as confirmed by PCR
et al., 2015; Yoshida et al., 2016). Where a discordant serovar was in between 30 and 78% of samples analyzed. There was no apparent
identified by these two tools, the KMER-based taxonomic assignment association between the different sample types and the different visits
of kraken2 was used as the third comparison point where the majority to the markets. It is noted that despite muscle tissue samples being
was taken to decide on the serovar. obtained following disinfection of the skin surface, 60 to 74% of such
samples contained E. coli indicating cross-contamination during
sample processing as meat samples normally would be expected to
In-silico identification of multilocus be sterile.
sequence typing (MLST), phylogroup E. coli confirmed by PCR was found in 41 and 44% of muscle
analysis, AMR genes and virulence factors tissue and 22 and 33% of intestinal samples from shrimp at the KBW
and KBR markets, respectively. There were no apparent differences in
In-silico MLST was performed using the PubMLST database2 findings during the different visits to the markets.
developed by Keith Jolley (Jolley and Maiden, 2010) and the open- Salmonella spp. confirmed by PCR was found in 29% of skin
source software (mlst tool) used to interrogate the database3 (Jolley swabs and 26% of gill samples, but not in muscle tissue and intestinal
et al., 2018). AMR genes were determined in the assembled genomes samples, from 27 tilapia fish collected at KBR (Table 2). Salmonella
using Resfinder4.1 (Bortolaia et al., 2020) and AMRFinderPlus spp. was isolated in 7% of skin swabs, 52% of gill samples, and one
(Feldgarden et al., 2021). The virulence determinants of E. coli were intestinal sample of 27 tilapia fish analyzed from the KBW. No
determined using a virulence factor database (Chen et al., 2005) and muscle tissue and intestinal samples from shrimp obtained at KBW
the curated VirulenceFinder database for E. coli (Malberg Tetzschner contained Salmonella spp. as confirmed by PCR whereas Salmonella
et al., 2020), while Salmonella virulence was determined by looking spp. was isolated and confirmed by PCR in 4/27 (15%) muscle tissue
for their Salmonella Pathogenicity Islands (SPIs) (Roer et al., 2016). and 2/27 (7%) intestinal samples in shrimp obtained at KBR
Moreover, E. coli genomes were subjected to phylogroup analysis (Table 2).
using pathogenwatch4 and clermontyping tools (Beghain et al,. 2018).

Genomic characterization and

Phylogenetic analysis antimicrobial resistance of E. coli

Salmonella Kentucky was the main serovar determined in the The genomic characterization (i.e., size, MLST, serovar.,
analysis. Strains with this serovar were added to publicly available phylogroup) and antimicrobial resistance patterns and genes of the 15
S. Kentucky genomes from migratory birds (Card et al., 2023), E. coli isolates obtained from shrimp and tilapia samples and that
humans, and poultry in Bangladesh and its neighboring countries, yielded quality genomes are shown in Table 3. The E. coli isolates from
such as India (human and sesame seeds), Nepal (human), Pakistan shrimp had total genome lengths ranging from 4,725,736 to 4,897,945
(chili), and Myanmar (human) (Supplementary Table S4). The base pairs, whereas the E. coli isolates from tilapia exhibited total
phylogenetic analysis was run using as reference genome S. Kentucky genome lengths ranging from 3,968,775 to 5,077,851 base pairs (see

2 https://pubmlst.org/
3 https://github.com/tseemann/mlst 5 https://github.com/tseemann/snippy
4 https://pathogen.watch/ 6 https://microreact.org/

Frontiers in Microbiology 05 frontiersin.org

Rheman et al. 10.3389/fmicb.2024.1329620

TABLE 3 Genomic characterization and antimicrobial resistance of E. coli.

Isolate ID Date of Sample MLST Serovar Phylo Resistance patterns* AMR genes**
isolation source group
BD08 19/9/2021 Shrimp/muscle 1,662 H16-O103 B1 AMP(I); CXM(I) glpT; pmrB

BD09 19/9/2021 Shrimp/muscle 48 H9-O8 A AMP(R); CXM(I) glpT; pmrB

BD14 19/9/2021 Shrimp/muscle 196 H16-O8/O160 B1 AMP(I); CXM(I) glpT; pmrB

BD01 31/10/2021 Shrimp/muscle 3,640 H49 B1 AMP(I); NAL(I); LVX(I); CXM(I) glpT; pmrB

BD03 31/10/2021 Shrimp/muscle 155 H19 B1 AMP(R); FOX(R); CXM(I) glpT; pmrB

BD13 31/10/2021 Shrimp/muscle 398 H21-O159 A CXM(I) glpT; pmrB

BD23 31/10/2021 Shrimp/muscle 3,501 H16 B1 CXM(R); CRO(I) glpT; pmrB

BD12 12/9/2021 Tilapia/skin 8,349 H7-O18 B1 NIT(R); AMP(I); CXM(I) glpT; pmrB

BD05 17/10/2021 Tilapia/skin und H7-O83 B1 FOX(R); AMP(I); CXM(I) glpT; pmrB

BD19 17/10/2021 Tilapia/skin und H19-O8 B1 AMP(I); CXM(I) glpT; pmrB

BD21 17/10/2021 Tilapia/skin 155 H12-O28ab B1 - glpT; pmrB

BD22 17/10/2021 Tilapia/skin und H19-O8 B1 NIT(R); CRO(R); NAL(I), CXM(I) glpT; pmrB

BD11 12/9/2021 Tilapia/muscle 164 H9-O49 B1 AMP(R); MEM(R); GEN(R); CXM(I) glpT; pmrB

BD17 17/10/2021 Tilapia/muscle 2,165 H7 B1 AMP(R); CXM(I); NAL(I) glpT; pmrB; blaTEM-1B;
qnrS13; tet(A)

BD20 10/10/2021 Tilapia/muscle und H26-O8/O80 B1 NIT (R), CIP (I), CXM (I) glpT; pmrB
*Antimicrobials; AMP, ampicillin (10 μg); FOX, cefoxitin (30 μg); CRO, ceftriaxone (30 μg); CXM, cefuroxime sodium (30 μg); CIP, ciprofloxacin (5 μg); GEN, gentamicin (10 μg); LVX,
levofloxacin (5 μg); MEM, meropenem (10 μg); NAL, nalidixic acid (30 μg); and NIT, nitrofurantoin (100 μg); ** glpT (glycerol-3-phosphate) mutations encode resistance to fosfomycin; pmrB
(polymyxin B) mutations encode resistance to polymyxins; und: undetermined.

further details in Supplementary Table S1). MLST analysis of the One E. coli strain was resistant to ceftriaxone. Further details on
isolates from shrimp revealed the presence of 7 sequence types (STs), the phenotypic resistance are provided in Supplementary Table S5.
including 3,640, 3,501, 1,662, 398, 196, 155, and 48. Serovar
classification identified different serovars, such as H49, H19,
H16-O103, H9-O8, H21-O159, H16-O8/O160, and H16 (Table 3). Genomic characterization and
The tilapia isolates also display several different MLSTs. Serovar antimicrobial resistance of Salmonella spp.
classifications included H7, H9-O49, H7-O18, H19-O8, H26-O8/O80,
and H12-O28ab (Table 3). The genomic characteristics and serovars of the nine Salmonella
The phylogenetic analysis of isolates from both shrimp and spp. isolates obtained from tilapia samples are shown in Table 4 (see
tilapia revealed that they belonged predominantly to the further details in Supplementary Table S1). These strains yielded
commensal phylogroups B1 and A (Table 3). The commensal quality genomes and were isolated between 12 September and 31
nature of these isolates in warm-blooded animals and humans, but December 2021. All strains were Salmonella enterica
not in fish, was further underlined by the absence of major subspecies enterica.
virulence factors displayed by the known E. coli pathotypes in the Isolate BD25 exhibited a total genome length of 4,774,230 bp and
genomes (Supplementary Table S2). As commensal strains, the belonged to the Augustenborg serovar and the AMR genes detected
genome sequences did not contain any clinically important AMR were aac(6′)-Iaa and fosA7, as well as parC:p.T57S mutation. The
genes except for point mutations on the glycerol-3-phosphate isolate showed intermediate resistance to CXM. Notably, the
transporter (glpT) and the polymyxin resistance gene B (pmrB), pathogenicity islands SPI-1, SPI-13, SPI-14, SPI-2, SPI-3, and C63PI
known to be associated with potential resistance to fosfomycin were present, while no plasmids were detected.
and colistin, respectively (Table 3). The isolate BD17 from tilapia Isolate BD35, belonging to the Brunei serovar., had a total genome
harbored the resistance genes, blaTEM-1B, qnrS13, and tet(A) length of 4,832,090 bp. MLST analysis revealed ST1794 and the
encoding resistance to beta-lactams, quinolones, and tetracycline, presence of the aac(6′)-Iaa gene encoding resistance to
respectively. None of the E. coli strains contained any plasmids. aminoglycosides. The isolate was resistant to AMP, FOX, LVX, and
This isolate displayed phenotypic resistance to ampicillin and CRO. SPI-1, SPI-13, SPI-2, SPI-3, and C63PI were identified, along
intermediate resistance to nalidixic acid and cefuroxime sodium with the IncFII plasmid. BD37 had also the Brunei serovar. The
(Table 3). Several strains showed phenotypic resistance to aminoglycoside resistance gene aac(6′)-Iaa was also detected together
ampicillin (4 strains) and nitrofurantoin (3 strains), and several with SPI-1, SPI-13, SPI-14, SPI-2, and C63PI. This isolate was resistant
strains showed intermediate resistance to some antimicrobial to AMP, GEN, and FOX and showed intermediate resistance to other
classes, e.g., 13/14 E. coli strains were intermediately resistant to antimicrobials. Both isolates contained the IncFII plasmid.
cefuroxime sodium. However, no genes associated with such Isolate BD41 belonged to the Hartford serovar. The AMR genes,
resistance and intermediate resistance were identified (Table 3). aac(6′)-Iaa and fosA7 (fosfomycin) were present, along with C63PI,

Frontiers in Microbiology 06 frontiersin.org

Frontiers in Microbiology

Rheman et al.
TABLE 4 Genomic characterization of Salmonella enterica isolates from tilapia sold at wet markets in Bangladesh.

Isolate ID Date Sample ST Serovar SPI* Plasmids Resistance patterns*** AMR genes
isolation source
BD40 12/9/2021 Tilapia/gills 198 Kentucky C63PI; SPI-1; SPI-2; SPI-3 IncQ1 CIP(R); AMP(R); NAL(R); NOR(R); aac(6′)-Iaa; blaTEM-1B; gyrA (p.S83F);
GEN(R); LVX(R); CXM (I) sul1; tet(A); aac(3)-Id; aadA7; qacE

BD41 12/9/2021 Tilapia/gills und Hartford C63PI; SPI-1; SPI-2; SPI-5; SPI-13; SPI-14 none - aac(6′)-Iaa; fosA7

BD43 12/9/2021 Tilapia/gills und Kentucky C63PI; SPI-1; SPI-2 IncQ1 CIP(R); AMP(R); NAL(R); NOR(R); aac(6′)-Iaa; blaTEM-1B; gyrA(p.S83F);
GEN(R); CRO(I) sul1; tet(A); aac(3)-Id; aadA7; qacE

BD25 10/10/2021 Tilapia/gills und Augustenborg C63PI; SPI-1; SPI-2; SPI-3; SPI-13; SPI-14 none CXM(I) aac(6′)-Iaa; fosA7; parC:p.T57S

BD35 ** 28/11/21 Tilapia/gills 1794 Brunei C63PI; SPI-1; SPI-2; SPI-3; SPI-13 IncFII AMP(R); FOX(R); LVX(R); CRO(R) aac(6′)-Iaa
07

BD37** 28/11/21 Tilapia/gills und Brunei C63PI; SPI-1; SPI-2; SPI-13; SPI-14 IncFII AMP(R); FOX(R); GEN(R); NAL(I); aac(6′)-Iaa
CIP(I); CXM(I)

BD42 12/9/2021 Tilapia/skin und Kentucky SPI-1; SPI-2; SPI-5; SPI-13 IncQ1 CIP(R); AMP(R); NAL(R); FOX(R); aac(6′)-Iaa; blaTEM-1B; gyrA(p.S83F);
GEN(R); LVX(R); CXM(I) sul1; tet(A); aac(3)-Id; aadA7; qacE

BD45 12/9/2021 Tilapia/skin 198 Kentucky C63PI; SPI-1; SPI-2, SPI-3 IncQ1 CIP(R); AMP(R); NAL(R); NOR(R); aac(6′)-Iaa; blaTEM-1B; gyrA(p.S83F);
GEN(R); LVX(R); CXM(I) sul1; tet(A); aac(3)-Id; aadA7; qacE

BD46 12/9/2021 Tilapia/skin 198 Kentucky C63PI; SPI-1; SPI-2; SPI-3 IncQ1 CIP(R); AMP(R); NAL(R); NOR(R); aac(6′)-Iaa; blaTEM-1B; gyrA (p.D87Y);
GEN(R); LVX(R); CXM(I) sul1; tet(A); aac(3)-Id; aadA7; qacE
* SPI: Salmonella Pathogenicity Islands; ** from wholesale market (all others from retail market).
***Antimicrobials; AMP, ampicillin (10 μg); FOX, cefoxitin (30 μg); CRO, ceftriaxone (30 μg); CXM, cefuroxime sodium (30 μg); CIP, ciprofloxacin (5 μg); GEN, gentamicin (10 μg); LVX, levofloxacin (5 μg); NAL, nalidixic acid (30 μg) and NOR, norfloxacin (10 μg).
und: undetermined.

10.3389/fmicb.2024.1329620
frontiersin.org
Rheman et al. 10.3389/fmicb.2024.1329620

FIGURE 1
S. Kentucky phylogeny (A) including publicly available genome characteristics of S. Kentucky (B) originating from Bangladesh and surrounding
countries. The geographical location of strains included from each country is shown on the map (C).

SPI-1, SPI-13, SPI-14, SPI-2, and SPI-5. No plasmids were observed. remaining 10 migratory birds’ isolates (see further details on pairwise
The isolate was fully susceptible to all antimicrobials tested. matrix on S. Kentucky Supplementary Table S4).
Isolates BD42, BD40, BD43, BD45, and BD46, were all the
Kentucky serovar and exhibited total genome lengths ranging from
4,887,734 to 4,920,732 bp. These isolates shared similar AMR genes Discussion
including blaTEM-1B, gyrA mutations, sul1, tet(A), aac(3)-Id, and aadA7.
The isolates showed similar phenotypic resistance patterns, i.e., to In this study, we assessed the occurrence of E. coli and NTS and
AMP, CIP, GEN, LVX, NAL, and NOR. Some isolates did also show the AMR of such isolates obtained from tilapia and shrimp products
intermediate resistance to other antimicrobials (further details on the at retail, to assess the potential food safety and health risk
phenotypic resistance are provided in Supplementary Table S6). SPI-1, to consumers.
SPI-2, and SPI-5 were identified, along with the IncQ1 plasmid. Three The high occurrences of E. coli in skin swabs, gill, muscle, and
isolates were ST198. They all have the K variant of SGI-1 carrying the intestinal samples from tilapia and shrimp sold at two main wet
same resistance genes all on the same contig in each genome. It should markets in Dhaka, Bangladesh indicates that fecal pollution and cross-
be noted that the plasmids were detected based on their replication contamination were common at the markets. The findings of E. coli in
proteins not by reconstructing the entire plasmids. intestinal samples suggest that the tilapia and shrimp were raised in
The phylogenetic analysis of the genomes of the five S. Kentucky aquatic systems with fecal pollution, as E. coli is not part of the normal
isolates from this study along with publicly available genomes of intestinal flora in fish and shrimp (Mohamed Hatha et al., 2003;
S. Kentucky from Bangladesh and neighboring countries revealed Duran and Marshall, 2005; Vu et al., 2018), which contrasts with
different levels of genetic relatedness among the strains circulating in livestock such as in poultry and pigs (Dang and Dalsgaard, 2012;
the region and most of the isolates in our study (Figure 1, AbuOun et al., 2020; Islam et al., 2023). E. coli in muscle tissue may
Supplementary Table S3). Out of all the tilapia isolates, BD40 had the indicate cross-contamination, e.g., from the fish surface, during
least number of SNPs compared with the non-tilapia isolates (e.g., 30 collection of the muscle tissue with other tissues before analysis. E. coli
SNPs with a migratory bird isolate). BD42 shows the greatest number may also have entered the muscle tissue, e.g., from the fish gut and
of SNPs compared with BD43 (5,306 SNPs) and all remaining isolates. surface; however, this seems only to occur under highly stressed
This was followed by BD43. BD42 and BD43 (non-ST198) clustered environmental aquatic conditions (Dang and Dalsgaard, 2012). It
together. The Salmonella isolates from tilapia isolates evolved in one should be noted that the likelihood and levels of fecal contamination
phylogenetic cluster from a common ancestor with the migratory bird at markets may show seasonal variations, e.g., as a result of flooding
isolates (SRR24520380/64), which have then further diverged with the and decreased hygiene conditions at markets.

Frontiers in Microbiology 08 frontiersin.org

Rheman et al. 10.3389/fmicb.2024.1329620

On the contrary, there is no apparent explanation of the frequent Conclusion

finding of Salmonella spp. in gill and skin samples and their isolation
in muscle tissue and intestinal samples, but the two latter samples are Our study showed a high level of fecal contamination with
less likely to be fecally contaminated, e.g., during handling of the common findings of commensal E. coli in different sample types of
tilapia at the markets as opposed to surface skin and gills. While fish tilapia and shrimp sold at two main wet markets in Dhaka. Together
do not normally have Salmonella as part of their natural flora, they can with the occurrence of Salmonella spp. in several products, e.g.,
become passive carriers (Fernandes et al., 2018). Reporting of S. Kentucky ST198, a well-known human pathogen, stresses the need
Salmonella spp. in different types of seafood is relatively common due to improve hygienic practices and sanitation standards at markets as
to fecal cross-contamination during handling and processing, but well as in people’s homes to improve food safety and protect
different serovars have also been found in the skin, gills, muscle, consumer health. Transmission of AMR bacteria to humans can
intestine, and feces of live fish (Fernandes et al., 2018; Baniga et al., occur directly through the consumption of or contact with
2019; Hounmanou et al., 2020, 2022). contaminated products, posing a significant public health concern.
The five Salmonella Kentucky strains analyzed in this study carried Further genomic epidemiological analysis and disease burden
the SGI1-K island; all had an In4-type class 1 integron that contained estimations are needed in Bangladesh to assess the contribution of
only one cassette array and an adjacent mercury resistance module. seafood to the overall occurrence of human salmonellosis as well as
Other studies have also found such a type of class 1 integrons in AMR problems in humans.
S. Kentucky as well as in other serovars (Levings et al., 2007; Doublet
et al., 2009). The SGI-K island harbored resistance genes such as
aac(6′)-Iaa, blaTEM-1B, sul1, tet(A), aac(3)-Id, and aadA7 that confer Scope statement
resistance to aminoglycosides, tetracycline, sulphonamide, and
narrow spectrum beta-lactam antibiotics. The set of genes present in Retail wet markets in low- and middle-income countries are often
these strains is, however, different from that reported in a previous reported to have inadequate sanitation resulting in fecal microbial
study (Hawkey et al., 2019); for instance, the presence of blaTEM1 on the contamination. Such contamination of sold produce with
SGI-K is uncommon. This may be attributed to the plasticity of the antimicrobial resistant (AMR) bacteria can be a potential risk to
genomes that allows for the loss or gain of segments by homologous public health. In this study, we conducted a pilot genomic surveillance
recombination and which is responsible for the development of for antimicrobial resistant E. coli and nontyphoidal Salmonella spp.
several variants of SGI1 in Salmonella (Hawkey et al., 2019; Jibril et al., isolated from tilapia and shrimp products purchased at two largest wet
2021). Moreover, on the SGI-K island of the strains, the resistance markets in Dhaka using Oxford nanopore sequencing. The study's
genes were located on the IS6 transposon as were also found in recent results provide important insights into the prevalence and
poultry isolates in Nigeria (Jibril et al., 2021), showing the potential transmission of antimicrobial-resistant E. coli and Salmonella in tilapia
distribution of resistance through these islands via various and shrimp sold at wet markets. This has the potential to play a crucial
mobile elements. role in efforts aimed at reducing the risk of transmitting fecal bacterial
The identification of SPI-1, SPI-2, and SPI-5 in S. Kentucky, as pathogens and antimicrobial resistance throughout the food chain.
well as an IncQ1 plasmid, further highlights the genetic diversity and
potential for horizontal gene transfer among these isolates. Notably,
three of the isolates were found to belong to ST198. Of particular Data availability statement
interest are the strains BD42 and BD43, which branched separately
on the phylogenetic tree. This finding suggests the presence of distinct The datasets generated and analyzed during the current study that
genetic lineages or recent evolutionary events contributing to the support our findings are available in the National Center for
diversification of S. Kentucky strains in the study area as can already Biotechnology Information (NCBI) repository at the following
be observed in the distinctive SPIs compared to the other persistent web links: demultiplexed FastQ files for all 24 strains
three strains. of E. coli and Salmonella spp. can be found under BioProject https://
When comparing our strains to publicly available genomes of dataview.ncbi.nlm.nih.gov/object/PRJNA974206 PRJNA974206
S. Kentucky from countries around Bangladesh, we observed different with the corresponding BioSample accession numbers from https://
levels of genetic relatedness among the strains circulating in the dataview.ncbi.nlm.nih.gov/object/SAMN35176355 SAMN35176355
region and most of the isolates in our study. This suggests a potential to https://dataview.ncbi.nlm.nih.gov/object/SAMN35176378
regional dissemination of S. Kentucky strains and emphasizes the SAMN35176378 and Sequence Read Archive (SRA) accession
importance of continued surveillance and molecular epidemiological numbers from https://dataview.ncbi.nlm.nih.gov/object/
studies to monitor the spread and evolution of this pathogen. Since SRR24673188 SRR24673188 to https://dataview.ncbi.nlm.nih.gov/
S. Kentucky is rarely isolated from fish, these findings support fecal object/SRR24673211 SRR24673211.
contamination from human or animal origin at the wet markets
where the samples were obtained. Nevertheless, fish can
be asymptomatic carriers, where Salmonella spp. have been isolated Ethics statement
from surface tissues, muscle, and intestine (Fernandes et al., 2018;
Baniga et al., 2019; Hounmanou et al., 2020, 2022). In Bangladesh, Ethical approval was not required for the study involving animals
S. Kentucky has also been reported in poultry and migratory birds in accordance with the local legislation and institutional requirements
(Card et al., 2023), which show close relatedness with the isolates because the authors confirm that the ethical policies of the journal, as
from this study. noted on the journal’s author guidelines page, have been adhered to.

Frontiers in Microbiology 09 frontiersin.org

Rheman et al. 10.3389/fmicb.2024.1329620

No ethical approval was required as no live animals were used in this Acknowledgments
study. The tissues were obtained from fish killed as part of routine
commercial food production (i.e., sourced from wet markets). This study was a collaboration between WorldFish, Centre for
Environment Fisheries and Aquaculture Science (CEFAS), and
Bangladesh Livestock Research Institute (BLRI) and was conducted as
Author contributions part of the CEFAS project “AMR: Facilitating future collaboration on
COVID-19 responses through capacity building on AMR livelihoods”
SR: Conceptualization, Formal analysis, Investigation, and the CGIAR Initiative on “Protecting human health through a One
Methodology, Validation, Writing – original draft, Writing – review Health approach.” WorldFish would like to thank its partners from the
& editing. SH: Conceptualization, Formal analysis, Investigation, Fleming Fund Country Grant, Patriot Biotech, the Department of
Methodology, Writing – original draft, Writing – review & editing. Veterinary and Animal Sciences, Faculty of Health and Medical
MSS: Methodology, Supervision, Writing – review & editing. FA: Sciences, University of Copenhagen, Animal Plant Health Agency
Investigation, Methodology, Writing – review & editing. LK: (APHA), UK, and Veterinary Medicines Directorate (VMD), UK FAO
Conceptualization, Writing – review & editing. HMG: Data curation, reference Centre for AMR.
Formal analysis, Writing – review & editing. AP: Conceptualization,
Methodology, Writing – review & editing. RM-C: Writing – review
& editing. YM-G-M: Data curation, Formal analysis, Writing – Conflict of interest
original draft, Writing – review & editing. AD: Formal analysis,
Visualization, Writing – original draft. CVM: Conceptualization, The authors declare that the research was conducted in the
Writing – review & editing. ZBB: Investigation, Writing – review & absence of any commercial or financial relationships that could
editing. MA-S: Conceptualization, Supervision, Writing – review & be construed as a potential conflict of interest.
editing. DV-J: Methodology, Writing – review & editing. JD-D:
Conceptualization, Data curation, Formal analysis, Investigation,
Methodology, Supervision, Writing – original draft, Writing – Publisher’s note
review & editing.
All claims expressed in this article are solely those of the
authors and do not necessarily represent those of their affiliated
Funding organizations, or those of the publisher, the editors and the
reviewers. Any product that may be evaluated in this article, or
The author(s) declare financial support was received for the research, claim that may be made by its manufacturer, is not guaranteed or
authorship, and/or publication of this article. This work was supported endorsed by the publisher.
by the UK FAO Reference Centre for Antimicrobial Resistance (which
receives funding from the Fleming Fund, UK Aid program, the
Department of Health and Social Care, Grant Ref: ODA/BANGLADESH) Supplementary material
and by the CGIAR Initiative on "Protecting Human Health Through a
One Health Approach" (which receives funding from funders who The Supplementary material for this article can be found online
support this research through their contributions to the CGIAR Trust at: https://www.frontiersin.org/articles/10.3389/fmicb.2024.1329620/
Fund: https://www.cgiar.org/funders/, Grant Number: INIT-07). full#supplementary-material

References
AbuOun, M., O’Connor, H. M., Stubberfield, E. J., Nunez-Garcia, J., Sayers, E., recovered from migratory birds in Bangladesh. Front. Microbiol. 14:1162657. doi:
Crook, D. W., et al. (2020). Characterizing antimicrobial resistant Escherichia coli and 10.3389/fmicb.2023.1162657
associated risk factors in a cross-sectional study of pig farms in Great Britain. Front.
Chen, L., Yang, J., Yu, J., Yao, Z., Sun, L., Shen, Y., et al. (2005). VFDB: a reference database
Microbiol. 11:861. doi: 10.3389/fmicb.2020.00861
for bacterial virulence factors. Nucleic Acids Res. 33, D325–D328. doi: 10.1093/nar/gki008
Alam, M., Haque, M., and Shikha, F. (2016). Studies on public health and hygiene
condition of retailers at fish markets in south-Central Bangladesh. J. Bangladesh Agric. CLSI (2020). Performance standards for antimicrobial susceptibility testing. Thirtieth
Univ. 12, 411–418. doi: 10.3329/jbau.v12i2.28702 information supplement (M100). Wayne, PA: CLSI.

Baniga, Z., Mdegela, R. H., Lisa, B., Kusiluka, L. J. M., and Dalsgaard, A. (2019). Dang, S. T. T., and Dalsgaard, A. (2012). Escherichia coli contamination of fish raised
Prevalence and characterisation of Salmonella Waycross and Salmonella enterica subsp. in integrated pig-fish aquaculture systems in Vietnam. J. Food Prot. 75, 1317–1319. doi:
salamae in Nile perch (Lates niloticus) of Lake Victoria, Tanzania. Food Control 100, 10.4315/0362-028X.JFP-11-501
28–34. doi: 10.1016/j.foodcont.2019.01.006 De Coster, W., D’Hert, S., Schultz, D. T., Cruts, M., and Van Broeckhoven, C. (2018).
Beghain, J., Bridier-Nahmias, A., Le Nagard, H., Denamur, E., and Clermont, O. NanoPack: visualizing and processing long-read sequencing data. Bioinformatics 34,
(2018). ClermonTyping: An easy-to-use and accurate in silico method for Escherichia 2666–2669. doi: 10.1093/bioinformatics/bty149
genus strain phylotyping. Microbial Genomics 4, e000192. doi: 10.1099/mgen.0.000192
Debnath, S. C., McMurtrie, J., Temperton, B., Delamare-Deboutteville, J.,
Bortolaia, V., Kaas, R. S., Ruppe, E., Roberts, M. C., Schwarz, S., Cattoir, V., et al. Mohan, C. V., and Tyler, C. R. (2023). Tilapia aquaculture, emerging diseases, and the
(2020). ResFinder 4.0 for predictions of phenotypes from genotypes. J. Antimicrob. roles of the skin microbiomes in health and disease. Aquac. Int. 31, 2945–2976. doi:
Chemother. 75, 3491–3500. doi: 10.1093/jac/dkaa345 10.1007/s10499-023-01117-4
Butaye, P., Michael, G. B., Schwarz, S., Barrett, T. J., Brisabois, A., and White, D. G. Dewi, R. R., Hassan, L., Daud, H. M., Matori, M. F., Nordin, F., Ahmad, N. I., et al.
(2006). The clonal spread of multidrug-resistant non-typhi Salmonella serotypes. (2022). Prevalence and antimicrobial resistance of Escherichia coli, Salmonella and Vibrio
Microbes Infect. 8, 1891–1897. doi: 10.1016/j.micinf.2005.12.020 derived from farm-raised red hybrid tilapia (Oreochromis spp.) and Asian sea bass (Lates
Card, R. M., Chisnall, T., Begum, R., Sarker, M. S., Hossain, M. S., Sagor, M. S., et al. calcarifer, Bloch 1970) on the west coast of peninsular Malaysia. Antibiotics 11:136. doi:
(2023). Multidrug-resistant non-typhoidal Salmonella of public health significance 10.3390/antibiotics11020136

Frontiers in Microbiology 10 frontiersin.org

Rheman et al. 10.3389/fmicb.2024.1329620

Doublet, B., Praud, K., Weill, F.-X., and Cloeckaert, A. (2009). Association of IS26- Malberg Tetzschner, A. M., Johnson, J. R., Johnston, B. D., Lund, O., and Scheutz, F.
composite transposons and complex In4-type integrons generates novel multidrug (2020). In silico genotyping of Escherichia coli isolates for extraintestinal virulence genes
resistance loci in Salmonella genomic island 1. J. Antimicrob. Chemother. 63, 282–289. by use of whole-genome sequencing data. J. Clin. Microbiol. 58, e01269–e01220. doi:
doi: 10.1093/jac/dkn500 10.1128/JCM.01269-20
Duran, G. M., and Marshall, D. L. (2005). Ready-to-eat shrimp as an international Malkawi, H. I. (2003). Molecular identification of Salmonella isolates from poultry
vehicle of antibiotic-resistant bacteria. J. Food Prot. 68, 2395–2401. doi: and meat products in Irbid City, Jordan. World J. Microbiol. Biotechnol. 19, 455–459. doi:
10.4315/0362-028x-68.11.2395 10.1023/A:1025113912366
El-Sayed, A.-F. M., and Fitzsimmons, K. (2023). From Africa to the world—the Mohamed Hatha, A. A., Maqbool, T. K., and Suresh Kumar, S. (2003). Microbial
journey of Nile tilapia. Rev. Aquac. 15, 6–21. doi: 10.1111/raq.12738 quality of shrimp products of export trade produced from aquacultured shrimp. Int. J.
Food Microbiol. 82, 213–221. doi: 10.1016/s0168-1605(02)00306-9
EUCAST (2023). The European committee on antimicrobial susceptibility testing.
Breakpoint tables for interpretation of MICs and zone diameters. Available at: http:// Neela, F., Banu, M., Rahman, M., Rahman, M. A., and Alam, M. (2015). Occurrence
www.eucast.org/clinical_breakpoints/ of antibiotic resistant bacteria in pond water associated with integrated poultry-fish
farming in Bangladesh. Sains Malays. 44, 371–377. doi: 10.17576/jsm-2015-4403-08
Feldgarden, M., Brover, V., Gonzalez-Escalona, N., Frye, J. G., Haendiges, J.,
Haft, D. H., et al. (2021). AMRFinderPlus and the reference gene catalog facilitate Okon, E. M., Okocha, R. C., Adesina, B. T., Ehigie, J. O., Alabi, O. O., Bolanle, A. M.,
examination of the genomic links among antimicrobial resistance, stress response, and et al. (2022). Antimicrobial resistance in fish and poultry: public health implications for
virulence. Sci. Rep. 11:12728. doi: 10.1038/s41598-021-91456-0 animal source food production in Nigeria, Egypt, and South Africa. Front. Antibiot.
1:1043302. doi: 10.3389/frabi.2022.1043302
Fernandes, D. V. G. S., Castro, V. S., Cunha Neto, A. D., and Figueiredo, E. E. D. S.
(2018). Salmonella spp. in the fish production chain: a review. Cienc. Rural:e20180141:48. Preena, P. G., Swaminathan, T. R., Kumar, V. J. R., and Singh, I. S. B. (2020).
doi: 10.1590/0103-8478cr20180141 Unravelling the menace: detection of antimicrobial resistance in aquaculture. Lett. Appl.
Microbiol. 71, 26–38. doi: 10.1111/lam.13292
Gurevich, A., Saveliev, V., Vyahhi, N., and Tesler, G. (2013). QUAST: quality
assessment tool for genome assemblies. Bioinformatics 29, 1072–1075. doi: 10.1093/ Rocha Rdos, S., Leite, L. O., de Sousa, O. V., and Vieira, R. H. (2014). Antimicrobial
bioinformatics/btt086 susceptibility of Escherichia coli isolated from fresh-marketed Nile tilapia (Oreochromis
niloticus). J. Pathog. 2014:756539. doi: 10.1155/2014/756539
Hamilton, K., Chen, A., Owusu-Ansah, E., Gitter, A., Kozak, S., Niquice, C., et al.
(2018). Salmonella risks due to consumption of aquaculture-produced shrimp. Microb. Roer, L., Hendriksen, R. S., Leekitcharoenphon, P., Lukjancenko, O., Kaas, R. S.,
Risk Anal. 9, 22–32. doi: 10.1016/j.mran.2018.04.001 Hasman, H., et al. (2016). Is the evolution of Salmonella enterica subsp. enterica linked
to restriction-modification systems? mSystems 1:e00009-16. doi: 10.1128/
Haque, M., and Belton, B. (2021). Shrimp farming and research in Bangladesh adapts mSystems.00009-16
to COVID-19 pandemic. Available at:(https://digitalarchive.worldfishcenter.org/
handle/20.500.12348/4433). Samad, M. A., Eberson, L., Begum, R., Alam, M. G. S., Talukdar, F., Akter, R., et al.
(2023). Microbial contamination and antibiotic resistance in marketed food in
Havelaar, A. H., Kirk, M. D., Torgerson, P. R., Gibb, H. J., Hald, T., Lake, R. J., et al. Bangladesh: current situation and possible improvements. Antibiotics 12:555. doi:
(2015). World Health Organization global estimates and regional comparisons of the 10.3390/antibiotics12030555
burden of foodborne disease in 2010. PLoS Med. 12:e1001923. doi: 10.1371/journal.
pmed.1001923 Shikongo-Nambambi, M. N., Shoolongela, A. S., and Schneider, M. B. (2012). Control
of bacterial contamination during marine fish processing. Available at:http://www.
Hawkey, J., Le Hello, S., Doublet, B., Granier, S. A., Hendriksen, R. S., Fricke, W. F., macrothink.org/journal/index.php/jbls/article/view/1033
et al. (2019). Global phylogenomics of multidrug-resistant Salmonella enterica serotype
Kentucky ST198. Microb. Genom. 5:e000269. doi: 10.1099/mgen.0.000269 Simão, F. A., Waterhouse, R. M., Ioannidis, P., Kriventseva, E. V., and Zdobnov, E. M.
(2015). BUSCO: assessing genome assembly and annotation completeness with single-
Heijnen, L., and Medema, G. (2006). Quantitative detection of E. coli, E. coli O157 and copy orthologs. Bioinformatics 31, 3210–3212. doi: 10.1093/bioinformatics/btv351
other Shiga toxin producing E. coli in water samples using a culture method combined
with real-time PCR. J. Water Health 4, 487–498. doi: 10.2166/wh.2006.0032 Sobuj, M. K. A., Rabby, A. F., Rahman, S., Hasan, S. J., Bhowmik, S., Islam, M. A., et al.
(2022). Knowledge, attitudes, and practices on food safety and hygiene of wet and dry
Hounmanou, Y. M. G., Baniga, Z., García, V., and Dalsgaard, A. (2022). Salmonella fish handlers in Cox’s bazar, Bangladesh. Food Sci. Nutr. 10, 4139–4154. doi: 10.1002/
Salamae and S. Waycross isolated from Nile perch in Lake Victoria show limited human fsn3.3004
pathogenic potential. Sci. Rep. 12:4229. doi: 10.1038/s41598-022-08200-5
Sokolov, E. P. (2000). An improved method for DNA isolation from
Hounmanou, Y. M. G., Dalsgaard, A., Sopacua, T. F., Uddin, G. M. N., mucopolysaccharide-rich molluscan tissues. J. Molluscan Stud. 66, 573–575. doi:
Leekitcharoenphon, P., Hendriksen, R. S., et al. (2020). Molecular characteristics and 10.1093/mollus/66.4.573
zoonotic potential of Salmonella Weltevreden from cultured shrimp and tilapia in
Vietnam and China. Front. Microbiol. 11:1985. doi: 10.3389/fmicb.2020.01985 Thornber, K., Verner-Jeffreys, D., Hinchliffe, S., Rahman, M. M., Bass, D., and
Tyler, C. R. (2020). Evaluating antimicrobial resistance in the global shrimp industry.
Islam, M. S., Hossain, M. J., Sobur, M. A., Punom, S. A., Rahman, A. M. M. T., and Rev. Aquac. 12, 966–986. doi: 10.1111/raq.12367
Rahman, M. T. (2023). A systematic review on the occurrence of antimicrobial-resistant
Escherichia coli in poultry and poultry environments in Bangladesh between 2010 and Vaser, R., Sović, I., Nagarajan, N., and Šikić, M. (2017). Fast and accurate de novo
2021. Biomed. Res. Int. 2023:e2425564. doi: 10.1155/2023/2425564 genome assembly from long uncorrected reads. Genome Res. 27, 737–746. doi: 10.1101/
gr.214270.116
Jibril, A. H., Okeke, I. N., Dalsgaard, A., Menéndez, V. G., and Olsen, J. E. (2021).
Genomic analysis of antimicrobial resistance and resistance plasmids in Salmonella Vu, T. T. T., Alter, T., Roesler, U., Roschanski, N., and Huehn, S. (2018). Investigation
serovars from poultry in Nigeria. Antibiotics 10:99. doi: 10.3390/antibiotics10020099 of extended-spectrum and AmpC β-lactamase-producing Enterobacteriaceae from retail
seafood in Berlin, Germany. J. Food Prot. 81, 1079–1086. doi: 10.4315/0362-028X.
Jolley, K. A., Bray, J. E., and Maiden, M. C. J. (2018). Open-access bacterial population JFP-18-029
genomics: BIGSdb software, the PubMLST.org website and their applications. Wellcome
Open Res. 3:124. doi: 10.12688/wellcomeopenres.14826.1 Watts, J. E. M., Schreier, H. J., Lanska, L., and Hale, M. S. (2017). The rising tide of
antimicrobial resistance in aquaculture: sources, sinks and solutions. Mar. Drugs 15:158.
Jolley, K. A., and Maiden, M. C. (2010). BIGSdb: scalable analysis of bacterial genome doi: 10.3390/md15060158
variation at the population level. BMC Bioinformatics 11:595. doi:
Wick, R. R., Judd, L. M., Gorrie, C. L., and Holt, K. E. (2017). Completing bacterial
10.1186/1471-2105-11-595
genome assemblies with multiplex Minion sequencing. Microb. Genom. 3:e000132. doi:
Khan, M., Paul, S. I., Rahman, M. M., and Lively, J. A. (2022). Antimicrobial resistant 10.1099/mgen.0.000132
bacteria in shrimp and shrimp farms of Bangladesh. Water 14:3172. doi: 10.3390/w14193172
Wood, D. E., Lu, J., and Langmead, B. (2019). Improved metagenomic analysis with
Kolmogorov, M., Yuan, J., Lin, Y., and Pevzner, P. A. (2019). Assembly of long, error-prone kraken 2. Genome Biol. 20:257. doi: 10.1186/s13059-019-1891-0
reads using repeat graphs. Nat. Biotechnol. 37, 540–546. doi: 10.1038/s41587-019-0072-8
Yoshida, C. E., Kruczkiewicz, P., Laing, C. R., Lingohr, E. J., Gannon, V. P. J., Nash, J. H.
Levings, R. S., Partridge, S. R., Djordjevic, S. P., and Hall, R. M. (2007). SGI1-K, a variant E., et al. (2016). The Salmonella in silico typing resource (SISTR): an open web-accessible
of the SGI1 genomic island carrying a mercury resistance region, in Salmonella enterica tool for rapidly typing and subtyping draft Salmonella genome assemblies. PLoS One
serovar Kentucky. Antimicrob. Agents Chemother. 51, 317–323. doi: 10.1128/aac.01229-06 11:e0147101. doi: 10.1371/journal.pone.0147101
Lulijwa, R., Rupia, E. J., and Alfaro, A. C. (2020). Antibiotic use in aquaculture, Zhang, S., Yin, Y., Jones, M. B., Zhang, Z., Deatherage Kaiser, B. L., Dinsmore, B. A.,
policies and regulation, health and environmental risks: a review of the top 15 major et al. (2015). Salmonella serotype determination utilizing high-throughput genome
producers. Rev. Aquac. 12, 640–663. doi: 10.1111/raq.12344 sequencing data. J. Clin. Microbiol. 53, 1685–1692. doi: 10.1128/jcm.00323-15

Frontiers in Microbiology 11 frontiersin.org