Molecular characterization and susceptibility of foodborne Escherichia coli to conventional antibiotics and natural extracts of Solanum palinacanthum and Siparuna guianensis

TL;DR

This study found that many E. coli strains from food are resistant to multiple antibiotics and not affected by natural plant extracts, raising concerns for food safety.

Contribution

The study reports high multidrug resistance in foodborne E. coli and lack of susceptibility to two natural plant extracts.

Findings

55.3% of E. coli isolates were multidrug-resistant.

Most E. coli strains belonged to phylogenetic groups B1, A, or C.

No susceptibility was observed to the tested natural plant extracts.

Abstract

The present study aimed to evaluate the antimicrobial susceptibility profile and molecular characteristics of Escherichia coli isolated from various foods, including ground beef, pork, Minas fresh cheese, lettuce, and chicken. A total of 150 E. coli strains were recovered and tested against conventional antimicrobials and crude extracts of Solanum palinacanthum and Siparuna guianensis. Susceptibility testing was conducted using disk diffusion and broth microdilution methods. Additionally, phylogenetic characterization and pathotype identification were performed via polymerase chain reaction (PCR). Results indicated that all E. coli isolates were resistant to at least two antimicrobial classes, with 55.3% (n = 83) classified as multidrug-resistant (MDR). Most strains belonged to phylogenetic groups B1, A, or C. Regarding the natural extracts, the isolates were not sensitive to either S. palinacanthum or S. guianensis. In conclusion, the high prevalence of MDR E. coli across diverse food groups represents a critical concern for food safety and public health. Furthermore, the lack of susceptibility to the tested plant extracts highlights the challenge of identifying alternative natural antimicrobials, underscoring the urgent need for stricter control measures to mitigate the spread of resistant foodborne pathogens.

The online version contains supplementary material available at 10.1007/s42770-026-01898-9.

Figures

Percentage of multidrug-resistant *E. coli* strains present in food

Percentage of E. coli strains resistant to antimicrobial groups present in food. ß-lac - ß-lactams; Qui - quinolones; Tetra - tetracyclines; Ami - Aminoglycosides; Amp - amphenicols; Sulf - sulfonamides; Lin - lincosamides; Pol - polymyxins; Mac – macrolides

Full text

Introduction

The World Health Organization (WHO) estimates that each year, around 600 million people worldwide become ill from eating contaminated food, resulting in 420,000 deaths. In Brazil, between 2014 and 2023, Escherichia coli was the primary etiological agent identified in foodborne disease outbreaks, accounting for 34.8% of cases [1]. Beyond its role as a primary pathogen, E. coli is a major opportunistic agent and a critical indicator of fecal contamination and antimicrobial pressure in the environment. Its genomic plasticity allows it to act as a central reservoir for the transmission of antimicrobial resistance (AMR) genes, particularly through horizontal gene transfer, which poses a severe threat to public health by complicating the treatment of common infections [2].

E. coli is a commensal bacterium of the human and animal intestinal microbiota. However, some strains acquire virulence factors, becoming major causes of intestinal infections [3]. The rising global threat of AMR has turned foodborne E. coli into a global concern, as the food chain serves as a bridge between animal reservoirs and human populations. Consequently, continuous monitoring of food isolates is essential to track the evolution of resistance patterns and to inform surveillance programs under the “One Health” approach [4].

The indiscriminate use of antibiotics in the agricultural sector contributes to selection pressure on E. coli, allowing them to acquire resistance genes that infiltrate the food chain [5]. Strains recovered from food show high resistance rates to critically important antimicrobials, such as quinolones, cephalosporins, and tetracyclines [6]. Given this escalating resistance crisis, there is an urgent need to explore alternative therapeutic and sanitizing strategies. Plant-derived biocides have emerged as promising alternatives due to their complex chemical compositions and ability to exert antibacterial effects through multiple mechanisms, which may reduce the likelihood of developing further resistance [7].

Some plants, such as Siparuna guianensis and Solanum palinacanthum, contain diverse secondary metabolites, including tannins, alkaloids, and flavonoids, whose synergistic presence is associated with potent antibacterial activity [8–10]. These botanical compounds represent a sustainable frontier for the development of new antimicrobials capable of targeting resistant strains where conventional drugs fail. In this sense, considering the importance of monitoring E. coli in food and the search for control alternatives, this study aimed to evaluate the molecular characteristics of E. coli isolated from food and its susceptibility to conventional antimicrobials and crude extracts of S. guianensis and S. palinacanthum.

Materials and methods

Experimental design

The current study used 150 bacterial strains of E. coli isolated from various foods as part of research carried out between 2013 and 2019, belonging to the bacteriotheque of the Federal University of Jataí (UFJ). Among them, 41 were isolated from chicken meat, 41 from ground beef, and 21 from pork, 26 from lettuce, and 21 from fresh Minas cheese.

Microbiological recovery of strains

From the different food samples, 25-gram aliquots were taken from an initial 200-gram sample and added to 225 mL of sterile buffered peptone water. The mixtures were homogenized in a stomacher for 60 s and subsequently incubated at 36 °C for 18 h in a shaker at 60 RPM. The presence of the bacteria was confirmed by streak-plating onto Harlequin^®^ Chromogenic Agar E. coli/Coliform, where purple colonies were considered positive for E. coli after incubation at 36 °C for 24 h in a bacteriological incubator. Typical E. coli colonies were further confirmed by phenotypic biochemical tests, including indole, methyl red, citrate utilization, and Voges-Proskauer (IMViC tests). Quality control for the antimicrobial susceptibility tests was ensured by using the reference strain Escherichia coli ATCC 25,922, in accordance with the CLSI M100 35th edition (2025) guidelines [11].

Antimicrobial susceptibility testing by disk diffusion

Previously recovered E. coli strains were tested for susceptibility to 17 antimicrobials using the disk diffusion method on Mueller-Hinton agar (Oxoid Ltd., Hampshire, United Kingdom). Bacterial suspensions were adjusted to a 0.5 McFarland standard (1.5 × 10⁸ CFU/mL) and inoculated onto Mueller-Hinton agar plates. After incubation at 35 °C for 18–24 h, inhibition zones were measured with a digital caliper. The results for clinical antibiotics were interpreted as Susceptible (S), Intermediate (I), or Resistant (R) according to CLSI M100 35th ed. guidelines [11]. For veterinary-specific drugs or those without established CLSI breakpoints for Enterobacteriaceae (Penicillin, Tilmicosin, and Lincospectin), susceptibility was evaluated based on manufacturer instructions or epidemiological cut-offs to maintain consistency with the previously generated profiles. The following antimicrobials were evaluated: Aminoglycosides [Gentamicin (10 µg), Neomycin (30 µg)], β-lactams [Amoxicillin-clavulanic acid (20/10 µg), Ampicillin (10 µg), Ceftiofur (30 µg), Penicillin (30 µg)], Quinolones [Norfloxacin (10 µg), Ciprofloxacin (5 µg), Enrofloxacin (5 µg)], Macrolides [Tilmicosin (15 µg)], Sulfonamides [Trimethoprim-Sulfamethoxazole (25 µg)], Phosphonics [Fosfomycin (200 µg)], Tetracyclines [Oxytetracycline (30 µg), Doxycycline (30 µg), Tetracycline (30 µg)], Lincosamides [Lincospectin (109 µg)], and Polymyxin [Colistin (10 µg)]. Isolates were classified as multidrug-resistant (MDR) when they showed resistance to three or more antimicrobial classes [12].

Obtaining botanical material and identifying species

The botanical material of the native Cerrado plants S. palinacanthum and S. guianensis were obtained in the municipality of Jataí, Goiás, Brazil, located in the southwest region of the state of Goiás (GO). Branches containing green leaves and free of pathogens were used. Botanical identification of the species was performed visually, accompanied by literary material from the Department of Botanical Biology of UFJ.

Obtaining the ethanolic extracts

The extracts were obtained from the Organic and Inorganic Chemistry Laboratory of UFJ, following the methodology described by Carvalho et al. [13]. First, healthy leaves were separated, washed in running water, and placed in plastic trays for drying at room temperature and in a ventilated place for seven days.

After drying, the leaves were crushed in a blender to obtain a powder. This material was then subjected to the extraction process using ethanol as a diluent. The mixture was left at room temperature, with stirring, for at least 12 h. Three consecutive extractions of the same powder were performed to remove as many active biological compounds as possible.

After extraction, the material was filtered using a funnel and filter paper, in addition to a vacuum pump to remove the solid part, leaving only the crude ethanol extract. The extract was then placed in a rotary evaporator to evaporate the solvent at a temperature of 55 °C, leaving only the crude extract. The material was weighed and stored in properly identified and packaged glass containers. The material was left in the exhaust hood to completely eliminate the ethanol. After being weighed three time consecutively, on alternate days, the process was completed and the extract was stored in a freezer at −80 °C for later analysis.

Identification of secondary metabolites present in crude extracts of S. palinacanthum and S. guianensis

The phytochemical screening of the ethanolic extracts was performed according to the methodology described by Barbosa et al. [14], with modifications. For the tests of alkaloids, phenols, and tannins, general flavonoids, leucoanthocyanidins, catechins, steroids and triterpenoids and flavones, and saponins, the extracts were diluted in ethanol 92.8° NIWM.

Alkaloids

A total of 100µL of the extract diluted in ethanol (10 mg/mL) were transferred to a contact plate and the medium was acidified with 3 drops of 1% HCl solution (v/v). Then, 4 drops of Dragendorff reagent were added and the color change for orange or reddish-orange in samples positive for the presence of alkaloids was observed.

Steroids and triterpenoids

In total, 10 mg of the extracts were diluted in 1000µL of chloroform. They were then filtered and transferred to a test tube to which 100µL of acetic anhydride and 3 drops of concentrated H_2_SO_4_ were added. After the procedure, a change in color was observed in the positive samples. The presence of steroids is confirmed by a blue-green coloration, while triterpenoids are indicated by a brown to red coloration.

Phenols and tannins

On a contact plate, 2 drops of 1% (m/v) FeCl_3_ solution were added to 100µL of extract diluted in ethanol (10 mg/mL). As an indicator, a blank containing ethanol and 1% (m/v) FeCl_3_ was used. The presence of phenols or tannins results in a coloration that varies between blue-green, dark blue or black. Any change in color or formation of a precipitate is considered positive when compared with the blank test.

Flavonoids

A total of 100µL of extract diluted in ethanol (10 mg/mL) were pipetted into three wells on the contact plate, acidifying one medium with pH 3, while the remaining two were alkalinized to pH 8.5 and 11. The results were then observed. Initially, the extract is treated with an alkaline solution, such as sodium hydroxide, which can intensify the characteristic yellow color of flavonoids. Then, the addition of an acid, such as hydrochloric acid, can cause this color to disappear or change, indicating the presence of flavonoids.

Leucoanthocyanidins, catechins, and flavones

A total of 500µL of the extract diluted in ethanol (10 mg/mL) were transferred to two numbered test tubes. Tube 1 was acidified with 4 drops of 1% HCl (v/v). In tube 2, 4 drops of 1% NaOH solution (v/v) were added. The solutions were heated in a water bath. Changes in color indicate a positive result for the test. Leucoanthocyanidins may present a reddish or purple coloration after acidification and heating, due to the formation of anthocyanidins. Catechins may exhibit a greenish or brown coloration after alkalinization and heating, due to oxidation. Flavones may change color in an alkaline medium, exhibiting yellowish or orange tones.

Saponins

A total of 20 mg of the extracts were diluted in 2mL of ethanol and transferred to a test tube, to which 5mL of distilled water and 5mL of hot water were added. The solution was stirred manually for about two minutes and then the result was observed. Saponins are substances with surface-active properties, forming stable foam when agitated in water.

Sensitivity test of crude extracts of S. palinacanthum and S. guianensis by disk diffusion

The sensitivity test for the crude extracts of S. palinacanthum and S. guianensis was performed in triplicate, using the disk diffusion method on filter paper, as adapted from the Kirby-Bauer method [15] using the bacterial strains that showed resistance or multiresistance to the antimicrobials tested.

Concentrations were prepared from the crude extract at 50% (500ul/mL) and diluted to 10% (100ul/mL), 5% (50ul/mL), 2.5% (25ul/mL), 1% (10ul/mL), and 0.1% (1ul/mL) [16]. Discs of 6 mm in diameter were used, soaked with 10µL of each extract under study.

Bacterial suspensions were prepared in 0.9% sterile saline to reach a turbidity equivalent to a 0.5 McFarland standard (approximately 1,5 × 10⁸ CFU/mL). This standardized inoculum was then evenly spread across the surface of Mueller-Hinton agar plates (4 mm depth). Subsequently, disks impregnated with the plant extracts at different concentrations were added, along with negative control disks (solvents and distilled water) and positive control disks (fosfomycin 30 µg). Following a 24-hour incubation at 36º +/- 1 °C, the diameters of the bacterial growth inhibition zones were measured in millimeters (mm) [17, 18].

Broth microdilution test of crude S. palinacanthum and S. guianensis

The strains that showed resistance and multiresistance to the antimicrobials used in the study were tested using the microdilution technique in 96-well plates containing BHI broth and ethanol extract of S. palinacanthum and S. guianensis, which were diluted in distilled water, previously autoclaved in an amber bottle, obtaining a concentration of 10 mg/mL. In total, 200µL of the extract (10 mg/mL) and 100µL of the BHI broth were used, following the serial dilution principle (½), obtaining the following extract concentrations: 10 mg/mL, 5 mg/mL, 2.5 mg/mL, 1.25 mg/mL, 0.625 mg/mL, and 0.313 mg/mL. After the dilution procedure, 10µL of the bacterial solution adjusted to the 0.5 MacFarland scale were added in triplicate. To validate the microdilution test, a positive control containing BHI broth and the bacterial inoculum of the tested strain was considered, as well as a negative control, containing only the BHI broth.

Molecular characterization of E. coli strains isolated from food

The DNA of all 150 E. coli strains was extracted by the boiling method, according to Olsvik and Strockbine [19]. The DNA was transferred to a 0.5mL Eppendorf tube, free of DNase and RNase, previously sterilized and kept at −20 °C until the time of analysis.

Phylogenetic classification

The determination of the phylogenetic group was performed by means of simultaneous amplification assays (PCR-quadruplex), aiming at the sequential detection of DNA from certain genes, as described by Clermont et al. [20] (Supplementary Table 1). The reactions were performed with a final volume of 20 µl, containing PCR Buffer 1x, 3mM MgCl2, 0.2mM dNTP, 2 units of Taq DNA polymerase, 3 µl of DNA, and the appropriate primers. In total, 0.2 µM of each primer were used, except for AceK.f, which had a concentration in the reaction of 0.4 µM. All reactions were performed in the Veriti thermocycler (Applied Biosystems) programmed for initial denaturation at 94 °C (4 min), followed by 30 cycles of 94 °C (5 s), 57 °C (group E) or 59 °C (quadruplex and group C) (20 s) and 72 °C (10 s). After the cycles, the final extension was performed at 72 °C (5 min).

The PCR amplified products were detected by 2.0% agarose gel electrophoresis (BioAmerica Biotech^®^) using Tris/Borate/EDTA - TBE X buffer (0.89 M Tris; 0.02 M EDTA; 0.89 M boric acid) and the 100 bp Ladder plus molecular weight standard (Sinapse Inc.). Electrophoretic separation was performed at 4 Volts/cm for a period of approximately 1 h. After electrophoresis, the gels were stained in 0.5 mg/mL ethidium bromide solution and visualized under an ultraviolet transilluminator.

For the classification of phylogroups, the combination of the presence and absence of primers is as follows: A: arpA or arpA and yjaA; B1: arpA and TspE4.C2; C: arpA and yjaA; E: arpA and chuA or arpA and chuA and yjaA or arpA and chuA and TspE4.C2; D: arpA and chuA or arpA and chuA and TspE4.C2; F: chuA; B2: chuA and TspE4.C2 or chuA and yjaA andTspE4.C2 or chuA andTspE4.C2: E. clades: yjaA or arpA and chuA and yjA [20].

Characterization of diarrheagenic E. coli pathotypes

DNA samples obtained from the 150 E. coli isolates were grouped into pools of a maximum of 5 samples for the characterization of Shiga toxin-producing E. coli (STEC), enteropathogenic E. coli (EPEC), and enterohemorrhagic E. coli (EHEC), as described by OH et al. [21], with modifications. Two PCRs were performed: a triplex PCR, allowing the identification of the virulence factors stx1 and stx2, in addition to the E16S gene (used as an internal control for E. coli), and the second PCR, performed to identify the virulence factor eaeA. The sequences of the primers used are described in supplementary Table 2.

Triplex PCR was performed in a final volume of 25 µL, containing 10.7 µL of nuclease-free ultrapure water, 1x PCR buffer, 3 mM MgCl2, 0.2 mM dNTPs, 0.4 µM each of the primers stx1.F, stx1.R, stx2.F, and stx2.R, 0.1 µM of primers E16S.F and E16S.R, 1.5 units of Taq DNA Polymerase (Ludwig Biotec), and 3 µL of DNA. Amplification was performed in a Veriti thermocycler (Applied Biosystems) with initial denaturation at 95˚C for 1 min, followed by 35 cycles at 95˚C for 1 min, 60˚C for 1.5 min, and 72˚C for 1.5 min, and a final extension at 72˚C for 7 min.

For PCR of the eaeA gene, a final volume of 20 µL was used, containing 10.8 µL of ultrapure water, 1x PCR buffer, 2 mM MgCl2, 0.2 mM dNTPs, 0.3 µM of each of the primers eaeA200.F and eaeA200.R, 1 unit of Taq DNA Polymerase (Ludwig Biotec), and 3 µL of DNA. DNA amplification was performed in a Veriti thermocycler (Applied Biosystems) with initial denaturation at 95˚C for 5 min, followed by 35 cycles at 95˚C for 30 s, 55˚C for 30 s, and 72˚C for 30 s, and a final extension at 72˚C for 7 min.

The PCR amplified products were detected by 2% agarose gel electrophoresis (BioAmerica Biotech^®^) using Tris/Borate/EDTA - TBE 1X buffer (0.89 M Tris; 0.02 M EDTA; 0.89 M boric acid) and 100 bp Ladder plus molecular weight standard (Sinapse Inc.). Electrophoretic separation was performed at 4 V/cm for approximately 40 min. The gels were stained in 0.5 mg/mL ethidium bromide solution and visualized under an ultraviolet transilluminator.

Results

Susceptibility profile of E. coli strains to antimicrobials

The 150 E. coli strains were recovered and tested for susceptibility to 17 antimicrobials, as shown in Table 1. These antimicrobials were grouped according to their mechanism of action, resulting in 10 distinct groups: ß-lactams, quinolones, tetracyclines, aminoglycosides, amphenicols, sulfonamides, lincosamides, polymyxins, macrolides, and fosfomycins. In the current study, all E. coli strains presented resistance to at least 2 groups of antimicrobials tested and 55.3% (n = 83) of the samples were classified as multiresistant. Lettuce was the food with the highest number of multiresistant strains, followed by Minas fresh cheese, pork, ground beef, and chicken (Fig. 1).Table 1. Antimicrobial susceptibility profile for 150 E. coli isolates from foodPharmacological groupAntimicrobialSusceptibleIntermediateResistantn (%)n (%)n (%)ß-lactamsAmoxicillin-clavulanic acid (AMC; 20 µg)102 (68.0)14 (9.3)34 (22.7)Ampicillin (AMP; 10 µg)101 (67.3)1 (0.7)48 (32.0)Penicillin (P; 30 µg)0 (0)0 (0)150 (100)Ceftiofur (EFX;30 µg)77 (51.3)26 (17.3)47 (31.3)LincosamidesLincospectin (LS; 109 µg)143 (95.3)0 (0)7 (4.7)AminoglycosidesGentamicin (CN; 10 µg)143 (95.3)4 (2.7)3 (2.0)Neomycin (N; 30 µg)130 (86.7)12 (8.0)8 (5.3)QuinolonesNorfloxacin (NOR; 10 µg)143 (95.3)6 (4.0)1 (0.7)Ciprofloxacin (CIP; 5 µg)131 (87.3)15 (10.0)4 (2.7)Enrofloxacin (ENR; 05 µg)100 (67.0)40 (26.0)10 (7.0)TetracyclinesOxytetracycline (OX; 30 µg),111 (74.0)3 (2.0)36 (24.0)Doxicycline (DOX; 30 µg),112 (74.7)7 (4.7)31 (20.6)Tetracycline (TRA; 30 µg)83 (55.3)4 (2.7)63 (42.0)PolymyxinColistin (CL; 10 µg)147 (98.0)0 (0)3 (2.0)MacrolidesTilmicosin (TIL; 15 µg)6 (4.0)0 (0)144 (96.0)SulfanamidesTrimethoprim-Sulfamethoxazole (SXT; 25 µg)133 (88.7)11 (7.3)6 (4.0)PhosphonicsFosfomycin (FOS; 200 µg)149 (99.3)1 (0.7)0 (0)Fig. 1. Percentage of multidrug-resistant E. coli strains present in food

The antimicrobial classes with the highest resistance frequencies were β-lactams, followed by macrolides. Regarding the susceptibility profile, all E. coli strains were susceptible to fosfomycin (100%; n = 150). Figure 2 shows the percentages of E. coli strains resistant to antimicrobials present in food.Fig. 2. Percentage of E. coli strains resistant to antimicrobial groups present in food. ß-lac - ß-lactams; Qui - quinolones; Tetra - tetracyclines; Ami - Aminoglycosides; Amp - amphenicols; Sulf - sulfonamides; Lin - lincosamides; Pol - polymyxins; Mac – macrolides

Phytochemical characterization and susceptibility profile of E. coli strains to ethanolic extracts of S. palinacanthum and S. guianensis

The present study observed the presence of secondary metabolites in the ethanolic extracts of S. guianensis and S. palinacanthum. The extract of S. guianensis was positive for the presence of phenols and tannins, flavonoids, leucoanthocyanidins, catechins, flavones, and saponins. The extract of S. palinacanthum presented the following compounds as secondary metabolites: steroids and triterpenoids, phenols and tannins, flavonoids, leucoanthocyanidins, catechins, and flavones (Table 2).Table 2. Secondary metabolites found in ethanol extracts of S. guianensis and S. palinacanthumCompounds analyzedEthanolic extract of leavesS. guianensisS. palinacanthumAlkaloids--Steroids and triterpenoids-+Phenols and tannins++Flavonoids*++Leucoanthocyanidins, catechins and flavones++Saponins+-Legend: + positive; - negative; *yellow coloration at pH 11, indicative of flavones, flavonols and xanthones

Although both extracts present secondary metabolites with antimicrobial action, such as tannins and flavonoids, the present study did not observe such an effect in E. coli isolates by the disk diffusion and microdilution methods.

Molecular characterization of E. coli strains isolated from food

From the determination of phylogenetic groups performed by quadruplex PCR, the present study observed that the majority of the analyzed samples (62.6%; n = 94) belonged to phylogenetic group B1, followed by group A or C (18.0%; n = 27). The phylogenetic groups with the lowest frequency were: D or E (6.0%; n = 9), F (4.0%; n = 6), Clade I or II (2.7%; n = 4), and B2 (0.7%; n = 1). A total of 6.0% (n = 9) of the analyzed samples were classified as undetermined (U) (Table 3).Table 3. Distribution of phylogenetic groups of E. coli isolated from foodPhylogenetic GroupGround beefPorkChickenCheeseLettuceTotal by Group% by GroupA or C769232718.0%B130122017159462.6%B2--1--10.7%Clade I or II-2-2-42.6%D or E2-1-696.0%F--6--64.0%U214-296.0%ALL4121412126150100%

Regarding the presence of STEC, EHEC, and EPEC pathotypes in the analyzed samples, no isolate presented positive amplification results for the virulence genes stx1, stx2, and eaeA. However, the gene for 16 S rRNA (E16S) was amplified in all E. coli pools used in the study, demonstrating the presence, physical quality, and purity of the DNA.

Discussion

Although most strains of E. coli present in the food chain are harmless, some strains have the ability to acquire virulence genes and are considered one of the main causes of foodborne illnesses associated with intestinal diseases [22]. An additional concern is the increasing levels of antibiotic-resistant E. coli. The misuse and overuse of antimicrobial agents in food production puts selection pressure on bacteria, allowing them to acquire resistance genes [23]. This phenomenon represents a global threat to public health, as the food chain acts as a primary bridge for the dissemination of MDR bacteria from animal reservoirs to human populations.

E. coli strains have been found in different foods, such as pasteurized milk, ground beef, cheese, vegetables (carrots, spinach, and lettuce), chicken, turkey, and pork, among others. These products can serve as a source of contamination and transmission of resistant strains that pose a risk to human health [24–26]. The present study observed that all E. coli strains recovered from food showed resistance to at least 2 classes of antimicrobials tested and that 55.3% were classified as multidrug-resistant. This high prevalence of MDR strains in diverse food groups underscores a critical risk; these findings suggest that foodborne E. coli is not merely a contaminant but a significant reservoir for antimicrobial resistance genes that can be horizontally transferred to human pathogens during consumption or through environmental contact.

High rates of antimicrobial-resistant E. coli present in food are described in the literature. According to Baloyi et al. [27], 85.71% of E. coli isolates from spinach, apples, carrots, cabbage, and tomatoes were multidrug-resistant, and the antibiotic classes with the highest resistance rate were: aminoglycosides (94.81%), cephalosporins (93.51%), β-lactams (93.51%), and chloramphenicol (87.01%). Igbinosa et al. [28] identified that 85.9% of the isolates were resistant to ≥ 3 and ≤ 7 antimicrobial classes. Even in broader screenings, such as the study by Guragain et al. [29] with 3,367 samples, 17.3% were resistant to at least three antibiotics.

This scenario highlights the necessity of the “One Health” approach, recognizing that the resistance found in animal products and vegetables, such as lettuce, is part of an interconnected cycle of environmental and human exposure. According to Popli et al. [30], approximately 73% of antimicrobials sold globally are used in livestock, which explains the high selection pressure observed in our animal-origin samples. Furthermore, the detection of resistance in lettuce suggests a bidirectional flow: contamination via irrigation water containing animal waste or urban effluents allows ARGs to migrate from the agricultural environment directly to the consumer’s Table [31].

Despite presenting high resistance rates, the E. coli samples recovered in the study belonged mostly to phylogenetic group B1, followed by group A or C. Jalil et al. [32] observed similar results, where the most prevalent phylogenetic group was B1 (57.0%), followed by A (33.0%), D (6.0%), and B2 (4.0%). While groups B1 and A are typically classified as commensals, their high rate of multiresistance in this study is a significant public health concern. It indicates that even non-pathogenic E. coli lineages are successfully adapting to antimicrobial pressure, serving as “silent” environmental reservoirs capable of transferring resistance traits to pathogenic strains within the human gut [4].

Regarding the presence of STEC, EHEC, and EPEC pathotypes in the analyzed samples, no isolate presented positive amplification results for the virulence genes stx1, stx2, and eaeA. This result corroborates the phylogenetic data found, since 80.0% of the samples in the present study belong to groups B1, A, or C, that is, they are commensal bacteria. Amézquita-Montes et al. [24] also observed a low prevalence of pathogenic strains. Isolates carrying virulence genes exclusive to EPEC, STEC, and ETEC were identified in 8 (2.0%) of the 380 food samples analyzed. Although the rates found were relatively low, the presence of diarrheagenic pathotypes such as STEC, EHEC, and EPEC are responsible for large outbreaks of foodborne illnesses.

The presence of virulence factors, combined with the high prevalence of multiresistant E. coli strains found in food and the ease of transmission of antimicrobial resistance genes, have sparked researchers’ interest in developing new methods of controlling microorganisms, such as the use of plant-based products [33, 34]. S. palinacanthum and S. guianensis have been highlighted for their inhibitory actions on different microorganisms due to the presence of alkaloids, tannins, and flavonoids [35–37].

The literature highlights the inhibitory actions of S. palinacanthum extract on different microorganisms, such as Staphylococcus aureus, Pseudomonas aeruginosa, and Salmonella typhimurium [35]. Lobô et al. [36] also demonstrated the antimicrobial activity of the plant against bacteria such as E. coli, Staphylococcus aureus, and Pseudomonas aeruginosa. This activity was related to the presence of large amounts of alkaloids and tannins.

The phytochemistry of the crude leaf extract of S. guianensis suggests antimicrobial potential [37]. Studies carried out by Alves et al. [38] and Pereira et al. [39] demonstrated that S. palinacanthum extract can inhibit the growth of several bacteria in vitro, including E. coli. Oliveira et al. [40] detected bactericidal activity of S. guianensis essential oil in vitro against Streptococcus mutans, Enterococcus faecalis and E. coli bacteria, as well as antifungal action against Candida albicans. Moura et al. [41] also detected bactericidal activity of S. guianensis against E. coli, Pseudomonas aeruginosa, Staphylococcus aureus and Streptococcus pyogenes.

According to literature data, both plants present large groups of secondary metabolites that give them antimicrobial action [42], which were identified in the present study by phytochemical screening. However, it was not possible to determine the potential for bacteriostatic or bactericidal action by the proposed methods, since they were not able to inhibit bacterial growth. Other studies also did not observe an antimicrobial effect of S. guianensis and S. palinacanthum extracts against strains of E. coli [35, 43].

The lack of susceptibility of E. coli to the tested extracts can be largely attributed to the intrinsic structural resistance of Gram-negative bacteria. Unlike Gram-positive organisms, E. coli possesses a complex outer membrane containing lipopolysaccharides (LPS), which creates a robust hydrophobic barrier that limits the diffusion of many plant-derived bioactive compounds [44, 45]. This architecture effectively restricts the access of larger or more polar molecules, such as certain tannins and flavonoids identified in our phytochemical screening, to their internal cellular targets [46]. This phenomenon is clearly reflected in our results, where 100% resistance was observed for Penicillin, Tilmicosin, and Lincospectin—drugs to which E. coli is naturally resistant due to membrane impermeability. Furthermore, the presence of specialized porins and active efflux pumps in these wild MDR isolates may further diminish the intracellular concentration of any internalized phytochemicals, rendering the crude extracts ineffective at the concentrations tested [47]. This structural complexity explains why botanical extracts often show more potent activity against Gram-positive than Gram-negative pathogens.

This structural defense is likely amplified in our study by the extreme resistance profile of the foodborne isolates. Most previous studies reporting success, such as Alves et al. [38] and Oliveira et al. [40], often utilize standard laboratory strains (ATCC), which are significantly more susceptible than the wild, MDR isolates characterized here. This highlights a ‘resistance gap’, suggesting that the robust resistance mechanisms of these environmental strains exceeded the inhibitory capacity of the crude extracts. Consequently, while our findings provide a crucial baseline, they emphasize that future research must move beyond crude extracts toward purified fractions or synergistic combinations to overcome the high biological fitness and membrane complexity of MDR foodborne pathogens [48, 49].

Conclusion

The study demonstrates a high prevalence of MDR E. coli across diverse food groups, including animal-derived products and fresh produce. Although the isolates predominantly belong to commensal phylogenetic groups (B1 and A) and lack the tested virulence genes, their robust resistance profile poses a significant threat to public health. These strains act as silent reservoirs of resistance genes that can be disseminated through the food chain, emphasizing the urgent need for integrated surveillance under the “One Health” framework.

Furthermore, while the crude extracts of S. palinacanthum and S. guianensis did not show inhibitory activity against the MDR foodborne isolates, the phytochemical screening confirms the presence of bioactive compounds. This lack of susceptibility highlights the high biological fitness of wild MDR strains compared to laboratory lineages. Therefore, future studies should explore purified fractions or synergistic applications of these botanical extracts with conventional antibiotics to overcome bacterial resistance mechanisms and ensure food safety.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 (DOCX 18.4 KB)