First Genomic Characterization of Escherichia ruysiae Isolated from a Five-Year-Old Patient with Gastroenteritis and Bloody Diarrhea in Togo

1. Introduction

The Escherichia genus is one of the most extensively studied bacterial groups in microbiology, largely due to its most prominent species, Escherichia coli. It has served as a model organism for bacterial physiology, genetics, and host–microbe interactions for decades. Importantly, this genus encompasses both commensal and pathogenic members that colonize the gastrointestinal tract of humans and animals [1,2,3,4]. Within the E. coli species itself, a remarkable diversity exists ranging from harmless gut symbionts to strains responsible for severe gastrointestinal and extraintestinal infections, such as enterotoxigenic E. coli (ETEC), enterohemorrhagic E. coli (EHEC), enteroinvasive E. coli (EIEC), enteropathogenic E. coli (EPEC), enteroaggregative E. coli (EAEC), and diffusely adherent E. coli (DAEC). Additionally, uropathogenic E. coli (UPEC) are a major cause of urinary tract infections, and meningitis-associated E. coli (MNEC) cause sepsis and meningitis. This duality makes the Escherichia genus especially compelling from both medical, veterinary and ecological standpoints.

Six species have been described so far, including Escherichia coli, Escherichia hermannii, Escherichia fergusonii, Escherichia albertii, Escherichia marmotae, Escherichia whittamii, and recently E. ruysiae [3,4,5,6,7,8,9,10]. All known Escherichia species are ubiquitous and potentially associated with animal [6] and human disease [5,6,7,8,9,10]. E. ruysiae was first isolated from the faecal sample of an international traveller, and all so far known to belong to two phylogroups (Cryptic Clade III and Clade IV) [9]. Several studies have shown that the phenotypic identification of E. ruysiae is confused with E. coli. This is because there is no standardized protocol in clinical microbiology to phenotypically differentiate E. ruysiae from the other members of the genus. Like other species of the genus Escherichia, E. ruysiae shows a genomic and proteomic signature very close to that of E. coli, but with certain differences that can be subtle and difficult to distinguish without a well-maintained database (for example, the Bruker Biotyper database, VITEK MS, or others) [6]. E ruysiae was spotted in a prevalence study of Escherichia spp., with zoonotic significance in urban crows in Japan, posing a significant risk for their potential transmission to humans [11]. The two cryptic clades described so far, Clades III and IV, are commonly found in animals and the environment. Studies have investigated the ubiquity of E. ruysiae and commensalism in the human intestinal tract [6]. Recently, E. ruysiae was isolated in 2024 from wild boars in Italy, therefore, its prevalence in livestock needs to be further investigated [12]. Also, antimicrobial resistance in E. ruysiae has been shown to be associated with conjugative plasmids carrying clinically relevant antibiotic resistance genes, and it is thought to serve as a reservoir of resistance genes globally [13]. To date, the pathogenicity of E. ruysiae is not known and hasn’t been demonstrated in any clinical or veterinary cases. Also, this bacterium is generally overlooked in clinical and veterinary diagnostics globally because it lacks a standardized differential phenotypic assay in microbiology to delineate it from E. coli. In addition, its spectrum is still not in any of the MALDITOF databases [14]. This makes the majority of E. ruysiae identified as E. coli with a comfortable score.

In Togo, the microbiological diagnosis is usually limited to phenotypic identification assays after the growth and isolation of the microorganism in pure colonies. Using these conventional techniques, a case of gastroenteritis was reported to be caused by Escherichia coli from a bloody stool sample collected from a five-year-old child hospitalized in a district medical center in the maritime region of Togo. For the first time, we performed a thorough genomic description of the strain and demonstrated that this was E. ruysiae instead of E. coli. In this study, we aim to describe the phenotypic and genomic characteristics of this strain and highlight specific genomic characteristics associated with the virulence and antimicrobial resistance.

2. Materials and Methods

2.1. Sample Collection

Two swabs were taken from the stool sample collected in clean containers, equipped with a tight-fitting airtight lid. Each swab was introduced into Cary-Blair medium (OXOID, UK), previously cooled with ice for one to two hours before stool sample collection. The swabs, loaded with stool (by rotating the swab into the stool or anus), were introduced into the transport medium for transportation. The upper part of the stem, touched by fingers, was cut off and discarded. The caps of the Cary-Blair tubes and the lids of the stool containers were screwed back on and the tubes labeled following the protocol described by March and Ratnam [15]. The samples were transported to the laboratory in a refrigerated cooler box to be processed. The remaining waste was destroyed following the guidelines by standards of biohazard type 2 [16,17].

2.2. Isolation and Identification

The physical and microscopic examination of stools was carried out following the standard operational procedure (SOP) of the clinical microbiology laboratory [18,19]. A loopful suspension of the stool sample was inoculated into an enrichment broth (peptone broth) in a 1:10 sample-to-broth ratio and incubated at 37°C overnight. Subsequently, a 10 µl loop-full of the overnight enrichment broth was plated onto SMAC medium (Sorbitol-MacConkey) and incubated at 37°C for 24 h [19]. Pink colonies were further plated on Eosin Methylene Blue (EMB) agar and incubated at 37°C for 24 h [20]. Colony exhibiting a metallic sheen, suggestive of Escherichia coli, were confirmed with phenotypic assay using Api20E read by commercially available Microbact GNB 24E (Oxoid, UK) according to the manufacturer’s instructions [21]. The confirmation Api20E assay was carried on the strain FK53-34 at Institut Pasteur (Paris) using a Biomerieux kit. A 100 µl of 0.5 McFarlan bacterial suspension was inoculated into each well, and the strips were then incubated. After 2 hours of incubation at 37°C, a change in the media (or after adding revelers) was used to identify the profile of the bacterium [15]. In addition, we used MALDI-TOF MS (Matrix Assisted Laser Desorption Ionization – Time Of Flight Mass Spectrometry) to confirm or identify pure growth isolates [22,23].

2.3. Susceptibility Testing

From a suspension of 0.5 McFarland, prepared from 2 to 3 pure colonies of FK53-34 on Mueller Hinton medium [24], we carried out a susceptibility test according to Comite d’antibiogramme de la Societe francaise de microbiologie guideline CA-SFM) V1.0 [25]. Discs of 5µg of Ciprofloxacin(CIP), 30µg of Cefepim (FEP), 30 µg of Aztreonam (ATM), 30µg of Cefotaxim (CTX), 75 µg of Ticarcillin (TIC), 75µg/10 µg of Ticarcillin + clavulanic acid (TCC), 20 µg of Amoxicillin (AMX), 30µg of Cefoxitin (FOX), 30 µg of Cefuroxim (CXM), 10µg of Ceftazidim (CAZ) and 30µg of Nalidixic Acid (NAL) were tested. . The inhibition diameters are interpreted as resistant, intermediate and susceptible according to reference values (diameters in mm) for disc susceptibility testing (Kirby-Bauer method) recommendations [26].

2.4. DNA Extraction and Sequencing

We extracted nucleic acid of FK53-34 at Institut Pasteur from 24-hour culture growth on petri dishes using the KingFisher technique (Themo Scientific , OXQ32853, Oxoid brand, Germany), according to the manufacturer's instructions [21]. After quality control and DNA quantification by Nanodrop spectrophotometry and Qubit fluorometer [27], we denatured. By adding 500 μl prepared 1.4 pM libraries into the reservoir, we diluted and prepared DNA fragment ends for adapter ligation and attached Illumina sequencing adapters to the DNA fragments. We amplified adapter-ligated fragments to enrich the library[28]. Sequencing libraries were constructed, and the libraries were quantified using the 2100 Bioanalyzer System (Agilent) and Kapa Sybr Fast qPCR Kit. We carry on the sequencing on an Illumina Miseq V3 platform using a 2 × 300 paired-end approach (Illumina Inc., San Diego, CA, USA) [29].

2.5. Assembly and Annotation

We assessed the quality of the raw reads of FK53-34 using FastQC v0.11.5 [30] and filtered them using Trimmomatic v3 with cut-off Q28 [29], redundancies or over-represented reads were reduced, sequencing errors corrected, and reads with nucleobase identification quality measures (Phred scores > 28). The high-quality reads of FK53-34 were assembled with a denovo assembly approach using Spades V3.13.1 [31] and checked the final assembly scaffolds using Checkm2 v1.0.2 [32] to provide accurate estimates of genome completeness, contamination and heterogeneity. The taxonomy classification was carried out using GTDB-Tk v2.4.0 (RefDB version r220) [33]. A contig rearrangement was carried on using Mauve progressive alignment and E. ruysiae (strain OPT1704, AB136)). We performed the annotation of the obtained scaffolds using Prokka1.14.5 [32]. We used Get Homologues [35] to compute the coding sequence (CDS) ANI of FK53-34 using OrthoMCL approach with close E. ruysiae genomes available on NCBI. Furthermore, we performed a core alignment of the all available E. ruysiae including FK53-34 using Scapper (https://github.com/tseemann/scapper.git), cloned on March 29th 2025, a reference-based alignment tool composed of MuMmer 4 [36] , and TrimAL [37] . To support the ANI cladogram, we constructed a core genome maximum-likelihood phylogenetic tree using raxml-ng [38] and used FigTree [39] to visualize the tree. Proksee [40] was used to visualize the assembled, the annotation, the virulome and resistome of our genome.

2.6. Resistome, Virulome Genome Sequence Typing

We characterized the resistome and the virulome of FK53-34 using Abricate v1.9.8 [41]. The resistance genes were searched against the Resfinder 4.1 database (367 resistance genes) [42], NCBI (6570 sequences downloaded on June 23, 2023), CARD V3.2.4 (6627 ontology terms, 5010 reference sequences, 1933 mutations, 3004 publications and 5057 AMR detection model [43] , ARG-ANNOT V6 (1749 sequences) [44] and MEGAres V3.0 (9,000 resistance genes antimicrobials) [45]. We identified virulence genes by mapping the genome sequences against the virulence database VFDB V1.0 using default parameters [46]. We performed in silico typing using Ectyper V0.8.1 [47] with default parameters, and sequence types (ST) were determined using MLST 2.23.0 [48] based on seven housekeeping genes (adk , fumC , gryB , icd , mdh , purA and recA ). We finally used the phylogroup classification tool, ClermonTyping [49] to identify the phylogroup of our strain.

3. Results

3.1. Clinical Case Description

A 5-year-old, 18 kg female was examined by a general practitioner for abdominal pain with vomiting and bloody diarrhea without fever. The diagnosis was “gastroenteritis”. The present clinical case was found at Centre Medico-Social de Togblekope (altitude: 71 m above sea level, coordinates: 6°16'24" N and 1°12'49" E), located in the suburban region of Lomé. Clinical symptoms are dominated by vomiting, with bloody diarrhea and absence of fever and abdominal cramps. The patient was first treated with Motilium® (domperidone) oral suspension (2/day), ceftriaxone (500mg) injectable (2/day) and Actapulgite® (activated Mormoiron attapulgite) powder in oral suspension (2 days). The vomiting stopped on the third day; however, the bloody diarrhea and abdominal cramps persisted. On the fourth day, symptoms reappear with a thermal surge up to 40°C in the afternoon. Most of the blood cells counts are within normal range (red count = 4.22x10⁶/µl (4.8-6.2.10⁶), white count = 6200 cells/µl, hemoglobin = 13.6 g/dl (11-15g/dl), hematocrit =43.8 (40-54), platelet = 244,000/µl (150 000-400 000/µl)). On the seventh day, The results of microbiology test and mass spectrometry (MALDI TOF) of FK53-34 showed Escherichia coli with a score of 2.2 and sensitive to all antibiotics tested. The patient was later referred to a gastroenterologist around 10 a.m., but she couldn’t survive. She died after 8 days of hospitalization, despite the care provided. The autopsy concluded at 2 p.m. a massive dehydration from infectious gastroenteritis and weight loss estimated at nearly 14% of total weight.

3.2. Phenotypic, Biochemical Characterization and Resistance Profile

A culture growth of the strain FK53-34 at 37°C in an aerobic atmosphere for 24 hours on the TSA agar as transparent, refractive and translucent colonies of various sizes. The colonies appear creamy to white, with circular shapes and edges ranging from smooth to slightly convex. They vary in size, from 1 to 3 mm in diameter. FK53-34 is Gram-negative (Suppl.Data: Bacteria Cultivation). Biochemical test performed on the API 20E assay showed results in Table 1, the strain FK53-34 shows similar phenotypic metabolism profile with that of the reference AB134 with the exception that the strain FK53-34 metabolise citrate (CIT), tryptophane (TDA) and melibiose (MEL). Compared to E. coli, the latest was positive to Alanine dehydrogenase (ADH), lysine dehydrogenase (LDC) and saccharose (SAC) while E. ruysiae was negative.

FK53-34 is phenotypically susceptible to all antibiotics tested according to reference values (diameters in mm) for disc susceptibility testing (Kirby-Bauer method) recommendations (Table 2) [26].

3.3. Genomics Features and Challenges in the Strain Identification

We assembled the genomes of FK53-34 into 4785617 bp, comprising 136 contigs with a GC content of 50.55% with an average length of the contigs ≈ 35,190 bp. The N50 is 94629 with a coding density of 0,87. There were 4464 coding sequences (CDS), 4558 predicted genes, 83 tRNA, 1 repeat regions, 1 mmRNA, 10rRNA, 1173 markers and no heterogeneity (Suppl.Data, Genome quality). The ANI (Average Nucleotide Identity) gives an average similarity of nucleotide sequences of 98.78 for Escherichia ruysiae. The MultiLocus Sequence Typing (MLST) match with the following allele: adk(356), gyr(73), icd(670), mdh(56), purA(322), recA(454) and no best match with fumC(~1709). The final ST ID as well as the antigenic profile which could not identify alleles for the genotype ST and OH antigens respectively. However, the serogroup of the strain FK53-34 was identified as Clade IV. The strain FK53-34 is close to the strain C14-1 and S1-IND-07-A with ANI 99.04% and 98.91% respectively (Figure 1). We found two clusters in the ANI dendrogram which was confirmed by the core genome Maximum likelihood phylogenetic tree grouped in Clade III and Clade IV (Figure 2)

3.4. Resistome, Virulome and Mobilome of FK53-34

E. ruysiae have been found to harbor antimicrobial resistance genes, including those encoding extended-spectrum beta-lactamases (ESBLs). FK53-34 carries genes associated with the production of beta-lactamases BlaEC, CPXAR, (Bla)CTX, (Bla)ampH, (Bla)AmpC1, (Bla)AmpC2, (Bla)Penicillin_Binding_Protein, (Bla)ampH. It carries the eptA and pmrF gene, conferring resistance to cationic antimicrobial peptides, such as colistin and multidrug resistance gene (Emr, Mdt, Acr,) (Table 3). Several virulence factor genes were also identified, including those encoding flagellar protein (flgABCDE, filABC, flhABC, motAB), colonization factors (fimA, fimB, fimC, fimD, fimE, fimF, fimH, fimI), fitness (iutA and kpsMII), heat-stable enterotoxin (astA, senB), chemotaxis regulatory protein ( CheR, CheB, CheY et CheZ), AidA-I adhesin-like protein (ehaAB, hofB, hofC, hofq), ompT gene for outer membrane protease and Plasmid R100 complement resistance gene (traJ, traT). We also note the presence of the operon (aatPABC), adaptation gene (cfa), adhesion and biofilm formation gene (csgABCDEFG, agn43) and blood-brain barrier invasion (ibeBC) (Table virulence factor). The virulome shows the absence of the wzx and wzy genes. We noted also the presence of specific adhesion genes (EC958_4611, EC958_4614), immune evasion (ECS88_3547), motility and biofilm (fliC-H56), and pathogenicity islands (EC55989_3335 and Z1307) in the FK53-34 isolate. (Table 4)

4. Discussion

The successful isolation and culturing of Escherichia ruysiae from the fecal sample of a five-year-old patient with gastroenteritis and bloody stools in Togo significantly expands our understanding of its ecology and its pathogenic potential. Previously, E. ruysiae has been detected in diverse hosts and locations, including humans in Europe [50], an urban wild crow in Asia [11], a domestic chicken in North America [51], and domestic dogs in Oakland, USA [52], highlighting its broad ecological niches. The presence of E. ruysiae in humans, animals and the environment confirms its contribution to the One Health concept that recognizes the interconnectedness of human, animal, and environmental health. These findings emphasize the need to investigate the phylogenomic of E. ruysiae to better understand the spray and map the evolutionary relationships as well as the global distribution of this emerging pathogen. Many studies have noted the challenge of correctly identifying E. ruysiae, often misclassifying it as E. coli, particularly when using MALDI-TOF MS [52,53]. This misidentification stems from the lack of E. ruysiae specific spectra in current databases, underscoring the urgent need to update reference libraries. Our phylogenomic analysis of E. ruysiae from various environments reveals notable genetic diversity divided into two branches associated with the Clade III [6] and the clade IV [13]. The FK53-34 strain belongs to the clade IV clusters and some of the strains in that cluster have been described to harbored antimicrobial resistance genes and found in diverse environments. For instance, it shares a close phylogenetic relationship with C14-1, a strain previously isolated from a healthy domestic hen in the UK [51], reinforcing the cross-host adaptability of E. ruysiae. Unlike E. coli, which can typically be classified into phylogroups (A, B1, B2, D, etc.) [51], E. ruysiae does not fit these categories. Furthermore, our isolate could not be assigned to a known sequence type (ST), suggesting either a novel lineage or gaps in current MLST databases for this species. This lack of classification can complicate the epidemiological tracking and the understanding of E. ruysiae's biology and clinical impact. Whether these strains are primarily commensal or pathogenic remains unclear, but their genetic diversity implies complex evolutionary pressures and niche-specific adaptation.

Genomics plays a crucial role in modern clinical microbiology, particularly in taxonomic classification, profiling antimicrobial resistance (AMR) carriage and deciphering their mechanisms. Despite phenotypic susceptibility to all antibiotics tested, genomics analysis shows the presence of common multiple resistance associated genes such as blaPenicillin Binding Protein, blaampH, acrA, emrF, baeS et tolC and many other chromosome-mediated resistance genes (blaEC-15, eptA, pmrF). Among all compared strains only our strain E. ruysiae FK53-34 possessed the gene blaEC-15, a natural non-plasmid transferable chromosomal variant of AmpC-like. This might explain why the strain was sensitive to all cephalosporins tested including Cefepim, Cefotaxim, Cefoxitin, Cefuroxim and Ceftazidim. This has been previously shown with E. coli which harbored genes associated with resistance to antimicrobial but with a susceptible phenotypic profile when tested [54,55,56,57].

The mobilome of FK53-34 shows that this genome is rich in mobile genetic elements (phages, integrons, plasmids). It is a highly diverse and active mobilome, reflecting high genome plasticity. Previous studies have shown the role of mobile elements in the prediction and their contribution to bacterial virulence [58], evolution [59], adaptation [60,61], and transfer within bacteria [62]. Of all E. ruysiae published so far, only the strain FK53-34 carries specific virulence genes namely aap/aspU, aatABCP, senB, traJT. The gene aap/aspU also known as Dispersin is thought to neutralize the bacterial cell surface, repelling and projecting the positively charged AAFs (Aggregative adherence Fimbriae), thereby facilitating the dispersion of bacteria across the intestinal mucosa. It enables the bacteria to disperse after initial aggregation and spread across the intestinal mucosa. This process also supports biofilm formation and chronic colonization, contributing to the persistent diarrhea typical of EAEC infections. This dispersal may allow the bacteria to spread and establish new foci of infection [63]. The AatABCD export system is a critical exporter responsible for secreting specific proteins from the bacterial cytoplasm to the extracellular environment. Its most well-known role in EAEC is the export of the dispersin (Aap/AspU) protein [64]. The senB heat stable enterotoxin gene encodes an enterotoxin that increases fluid and electrolyte secretion in the intestinal mucosa. It contributes to watery diarrhea, complementing other virulence factors such as the type III secretion system and invasions [65]. All these genes, if expressed, may have exacerbated the diarrhea condition of the patient, extensively lost water and inevitably leading to death. In addition, the presence of ibeB and ibeC genes is known to confer to the pathogen the ability to cross the blood-brain barrier, producing the invasion of brain endothelial cells, and promoting neonatal meningitis [66,67]. Although, the gastroenteritis looked mild and should have been treated easily with common antibiotic therapy, the miss identification of the strains, the enclosed antimicrobial resistance profile added to the arsenal of virulence factors have contributed to the overlooking of the diagnosis which let unfortunately to the death of the patient.

5. Conclusions

This study reports the first clinical isolation and genomic characterization of Escherichia ruysiae in Togo, associated with a fatal case of gastroenteritis in a pediatric patient. Our findings reveal that E. ruysiae can harbor a large spectrum of antimicrobial resistance genes not detected in the susceptibility test and a virulome mainly found in pathogenic E. coli. These results underline the diagnostic limitations of conventional methods in accurately identifying cryptic Escherichia species and highlight the potential public health threat posed by overlooked pathogenic strains. The genomic features of FK53-34 reinforce the need for expanding surveillance, updated diagnostic tools, and incorporation of E. ruysiae into routine microbial monitoring. From a One Health perspective, its presence in both clinical, veterinary and environmental settings points to the urgent integrating approaches and controlling antimicrobial resistance of emerging pathogens.