A total of 1071 RTE rice product samples were collected from 11 cities covering the whole Zhejiang Province, Eastern China, during 2017−2019. The samples included 756 boiled rice dishes, 91 boiled rice noodles, 105 fried rice noodles, 33 sticky rice rolls, 64 boiled sticky rice dishes, and 22 fried rice cakes. Quantitative detection of B. cereus in each sample was performed by using the direct plating method (Liu et al., 2021; Yue et al., 2021; Anwar et al., 2022). Briefly, Twenty-five grams of each sample was suspended in 225 mL of PBS and subsequently homogenized for 2 mins. The homogenate was 10-fold serially diluted in sterilized normal saline. The dilutions were spread on Mannitol-Egg-Yolk-Polymyxin (MYP) agars separately in duplicate. Plates were incubated at 30°C for 24 h. Five presumptive colonies with typical morphology on each plate were selected for further identification. Suspected colonies were then identified using Gram staining and the VITEK2 compact system (BioMerieux, France), followed by rhizoid growth and parasporal crystal formation tests to differentiate B. cereus from Bacillus thuringiensis and Bacillus mycoides. GB/T 4789.14-2014 (Ministry of Health of the People’s Republic of China, 2014) was used to calculate B. cereus numbers. One B. cereus isolate from each positive sample was stored for further characterization.

Genomic DNA was extracted from B. cereus by using a bacterial DNA extraction Kit (Omega, United States), according to manufacturer’s instructions. The primers and PCR protocol for eleven virulence genes were used as previously described ces (Ehling-Schulz et al., 2005), hblA (Zhou et al., 2008), hblC, hblD, nheA, nheB, nheC (Melnick et al., 2012), bceT (in’t Veld et al., 2001), cytK1, cytK2 (Guinebretiere et al., 2006), and entFM (Ngamwongsatit et al., 2008). Individual PCR reaction (25 μL) contain 50 ng of DNA template, 0.5 μL of each primer (10 μM), 0.125 U of Taq polymerase (TaKaRa, Japan), 2.5 μL of 10 × PCR buffer (Mg²⁺ free), 1.5 μL of MgCl₂ (25 mM), and 2 μL of dNTP Mixture (2.5 mM). The amplicon was analyzed with 1% agarose gel. The gels were visualized by a UV Imaging System. A 100 bp DNA ladder (TaKaRa, Japan) was used as a DNA marker.

Antimicrobial susceptibility assay of B. cereus isolates was tested by using the broth micro-dilution minimum inhibitory concentrations (MICs) method according to the standard Clinical and Laboratory Standard Institute (CLSI) guidelines (Clinical and Laboratory Standards Institute, 2015). After 18 h of cultivation on nutrient agar at 37°C, B. cereus was suspended in 0.85 per cent (w/v) NaCl solution to 1.0 MCF, followed by dilution with Mueller-Hinton broth to the final concentration of 1∼2 × 10⁵ CFU/mL. Each 100 μL of the bacterial inoculum was added to 96-well plates containing antibiotics and incubated at 37°C for 20 h. Twelve antimicrobials from different classes were employed, including imipenem (1−64 μg/mL), penicillin (0.06−8 μg/mL), chloramphenicol (2−128 μg/mL), ceftriaxone (4−128 μg/mL), vancomycin (1−128 μg/mL), amikacin (8−128 μg/mL), erythrocin (0.25−32 μg/mL), tetracycline (2−32 μg/mL), ciprofloxacin (0.5−16 μg/mL), clindamycin (0.12−16 μg/mL), trimethoprim/sulfamethoxazole (0.5/9.5−16/304 μg/mL), and rifampin (0.5−8 μg/mL). The MIC results were analyzed based on the breakpoints for Bacillus species as per CLSI guidelines (Clinical and Laboratory Standards Institute, 2015). The breakpoint for ceftriaxone was from CLSI documents M45-A2 (Clinical and Laboratory Standards Institute, 2010). The isolates resistant to three or more types of antimicrobial classified into different antimicrobial categories were defined as multi-drug resistant (Li et al., 2022b). Staphylococcus aureus ATCC 29213 was used as a positive control.

All 91 B. cereus isolates were genotyped by ERIC-PCR using the following primers ERIC-1: 5′-ATGTAAGCTCCTGGGGATTCAC-3′ and ERIC-2: 5′-AAGTAAGTGACTGGGGTGAGCG-3′ (Dorneles et al., 2012; Dorneles et al., 2014). The PCR mixture (25 μL) was 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 4 mM MgCl₂, 0.3 mM of each dNTP, 1 U of Taq DNA polymerase (Takara, Dalian, China), 0.4 μM of each primer and 75 ng of DNA template. PCR reaction was carried out as follows: 95°C for 3 min, 35 cycles of 94°C for 30 s, 46°C for 40 s, 72°C for 3 min and a final incubation at 72°C for 10 min. Amplicons size was analyzed by 2.0% agarose gel. The gels were visualized by a UV Imaging System. A 100 bp DNA ladder (TaKaRa, Japan) was used as a marker. A 100% of similarity in bands pattern was defined as an ERIC-PCR genotype according to previous report (Magyar et al., 2019).

The software BioNumerics 7.6 (Applied Maths, Sint-Martens-Latem, Belgium) was applied to estimate the band size of ERIC-PCR amplicons and analyze the genotypes. Clustering analysis was based on the Dice similarity coefficient and the unweighted pair group method with arithmetic mean (UPGMA). The Hunter and Gaston Diversity Index (HGDI) was calculated to evaluate the discriminatory capability of ERIC-PCR (Shi et al., 2021). Isolates that share 100% similarity of amplicon bands pattern were grouped into one genotype.

Chi-square analysis was performed using the SPSS v 21.0 software package to determine if a significant difference existed in the prevalence distribution of B. cereus in different RTE rice products. The p-Value of <0.05 was used as a significance level. Relationships between genotype groups and virulence gene distribution, and antibiotic resistance profiles were analyzed by carrying out Pearson’s chi-square test and Fisher’s exact test with the Bonferroni correction.

The prevalence of B. cereus in 1071 RTE rice product samples examined in this study was described in the Table 1. B. cereus was detected in 8.49% (91/1076) of all samples collected, out of which 65/91 (71.4%) were from boiled rice dishes, 6/91 (6.6%) were from boiled rice noodles, 13/91 (14.3%) were from fried rice noodles, 2/91 (2.2%) were from sticky rice roll, 3/91 (3.3%) were from boiled sticky rice dishes and 2/91 (2.2%) were from fried rice cake. According to previously published data, there are significant variances in the detection rate of B. cereus in various types of food samples from different regions of the world (Wang et al., 2019; Xu et al., 2020; Qiu et al., 2021; Shi et al., 2021; Wu et al., 2021). The total occurrence rate of B. cereus in our study was similar to a previous study in which B. cereus was isolated from dairy products, rice and flour products in China (Zhao et al., 2020). The prevalence of B. cerous in our study was lower than in a previous study isolated from artisan cheeses made in Mexico and powdered food products in Switzerland (Heini et al., 2018; Adame-Gomez et al., 2020). There was no statistically significant difference (p > 0.05) in the prevalence of B. cereus across the six types of rice products in our study. B. cereus is an opportunistic pathogen found in food. Ingestion of 10⁵−10⁸ vegetative cells or 8 μg of emetic toxin per kg of body weight may lead to gastroenteritis or/and vomiting syndrome in adults (Paananen et al., 2002; Schoeni and Wong, 2005). According to our findings, the number of B. cereus detected in 1.0 % fried rice noodles and 0.8% boiled rice meal samples varied from 10³ to 10⁶ CFU/g. Food poisoning can occur after consuming a specific amount of these highly contaminated meals (Zeng et al., 2021, 2022).

A number of safety criteria for B. cereus in RTE meals have been developed. In Canada and the United Kingdom, an acceptable threshold of 10⁴ CFU/g is recommended. In South Korea, Australia, and New Zealand, a lower permissible threshold (10³ CFU/g) is adopted (Nsw Food Authority, 2009; Health Canada, 2010; Chon et al., 2015). In our study, 91.5% of the samples had less than 10³ CFU/g of B. cereus. However, 0.5% of the samples of RTE rice products had more than 10³ CFU/g of B. cereus, which is more than the acceptable level in some countries. Although B. cereus in 91.5% of the samples was <10 CFU/g in our study, B. cereus in 0.5 % of RTE rice product samples outnumbered 10³ CFU/g that could exceed the acceptable level of some countries.

For many years, scientists have been studying the molecular mechanisms of B. cereus virulence. The diarrheal and emetic syndromes have been linked to several virulence factors including, secreted hemolysin BL (Hbl), necrotic enterotoxin (CytK), non-hemolytic enterotoxin (Nhe), enterotoxin FM (EntFM), BceT, and emetic toxin cereulide (Granum and Lund, 1997; Ehling-Schulz et al., 2005; Schoeni and Wong, 2005; Senesi and Ghelardi, 2010). Hbl or Nhe can promote fluid accumulation in ligated rabbit ileal loops due to their hemolytic, dermonecrotic, and vascular permeability activities (Schoeni and Wong, 2005; Griffiths and Schraft, 2017). Both of these enterotoxins comprise the tripartite complex. Three components are required for their maximal biological activity: proteins B, L1 and L2 in Hbl, and proteins A, B, and C in Nhe. Toxin activity has not been detected in any individual components (Arora, 2021; Sornchuer et al., 2022). The genes encoding Hbl and Nhe components are hblA, hblC, hblD, as well as nheA, nheB, and nheC, are located on two different operons (Sastalla et al., 2013). BceT, EntFM, CytK are all single-protein enterotoxins. BceT has cytotoxic, vascular permeability activities and can cause fluid accumulation in ligated mouse ileal loops (Agata et al., 1995). The necrotic enterotoxin CytK, which presents highly cytotoxic, necrotic and hemolytic activities, was initially incriminated in a severe gastroenteritis outbreak causing three patients’ death in France (Lund et al., 2000; Alonzo et al., 2015).

In this study, the distribution of associated encoding genes of the above toxins in B. cereus isolates was investigated. 84.6% of the B. cereus isolates had at least one enterotoxin or emetic toxin gene. A total of 31 distribution patterns of virulence genes were determined in our study (Table 2). The predominant one was XXV, 9.9% of isolates belonged to it and possessed only one enterotoxin gene entFM. The nheABC genes were present in 47.3% of the isolates, this frequency was lower than in B. cereus isolates from various food source samples and clinical isolates associated with foodborne outbreaks in previous studies (Kim et al., 2011; Glasset et al., 2016; Zhang et al., 2017). The occurrence rate of hblACD was 36.3%, which is similar to the previous reports isolated from milk products (Hwang and Park, 2015; Zhang et al., 2017), and it is lower than that of ready-to-eat foods, including vegetables, infant rice flour, rice, and grain-based foods (Chon et al., 2015; Hwang and Park, 2015; Zhang et al., 2017). The coexistence of hblACD and nheABC was found in 24/91 (26.4%) isolates. Six isolates (6.6%) were found to possess all enterotoxin encoding genes detected in this study.

Two distinct variants of CytK have been reported: CytK1 and CytK2. CytK1 is more harmful than CytK2. Although CytK2 proteins are hemolytic and toxic to Vero cells and human intestinal Caco-2 cells, their toxicity was only around 20% CytK1 (Fagerlund et al., 2004). Furthermore, CytK1 has been linked to major B. cereus outbreaks (Fagerlund et al., 2004; Guinebretiere et al., 2006). According to our findings, 33.0% of B. cereus isolates had either cytK1 or cytK2. CytK1 was found in one strain, accounting for 3.3% of all cytK-positive isolates. A previous study also observed this significant variation in cytK1 and cytK2 detection rates in B. cereus isolates from Chinese infant meals (Zhang et al., 2017). Foodborne B. cereus isolates may be slightly mild when producing diarrhea, according to the researchers. Meanwhile, in China, there was a risk of a B. cereus outbreak driven by a cytK1-positive strain. The ces gene for emetic toxin cereulide production was found in 4.4% of B. cereus isolates, consistent with the fact that B. cereus with ces was rarely isolated from food and environmental materials (Arslan et al., 2014; Chon et al., 2015; Zhang et al., 2017).

The antimicrobial susceptibility profile of the B. cereus isolates is shown in Table 3. Various susceptibility patterns against 12 types of antibiotics were exhibited. All isolates were susceptible to vancomycin, amikacin, imipenem, and rifampin. 90.1%, 78.0%, 95.6%, and 97.8% of the isolates showed susceptibility to chloramphenicol, erythromycin, tetracycline, and ciprofloxacin, respectively. 97.8% of isolates were resistant to penicillin, consistent with published reports that B. cereus isolates from either clinical or food sources were mostly resistant to penicillins (Park et al., 2009; Merzougui et al., 2014; Zhang et al., 2017). A high rate of antimicrobial resistance (84.6% isolates) was also detected against trimethoprim/sulfamethoxazole. All B. cereus isolates were classified into eight antibiotic resistance patterns (Figure 1). Resistance to penicillin/trimethoprim/sulfamethoxazole was the most common in our study. 21 isolates (23.1%) were multidrug resistance, with 85.7% were resistant to penicillin/ceftriaxone/trimethoprim/sulfamethoxazole.

All 91 B. cereus isolates were fingerprinted and assigned genotypes by using ERIC-PCR. The size of amplicons bands ranged approximately from 100 bp to 2000 bp. Each isolate produces 3-12 DNA bands. Considering 100% similarity in band pattern as a cut-off criteria, 89 genotypes were obtained, as a PCR-mediated fingerprinting typing approach. ERIC-PCR is more straightforward and rapid than PFGE and ribotyping (Dorneles et al., 2014; Magyar et al., 2019). It was initially applied to B. cereus for genetic discrimination by PO-REN Hsueh et al. (1999). However, a low number of B. cereus strains and low genetic diversity made it insufficient to evaluate the discriminatory capability of ERIC-PCR for B. cereus. In subsequent reports, ERIC-PCR was utilized to distinguish the strains of different species in the Bacillus genus (Shangkuan et al., 2000). According to our results, the calculated Hunter Gaston Discriminatory Index (HGDI) of ERIC-PCR on B. cereus genotyping attained 0.9995 using the optimal PCR system, demonstrating a high discriminatory capability.

Cluster analysis was performed based on UPGMA (Figure 1). Two major genotype groups (Group A and Group B) were defined in our study. 49.5% (45/91) of the isolates belonged to Group A, and 50.5% of the isolates belonged to Group B. Relationships analysis results between genotype groups and virulence gene distribution, antimicrobial resistance profiles demonstrated that there was no association between genotype groups and nheC (χ² = 3.167, p = 0.075), bceT (χ² = 3.167, p = 0.075), cytK1 (χ² = 0.000, p = 0.987), cytK2 (χ² = 0.088, p = 0.767), and ces (χ² = 0.239, p = 1.000). Meanwhile, Group A B. cereus tended to carry hblA (χ² = 6.018, p = 0.014), hblC (χ² = 4.09, p = 0.043), hblD (χ² = 4.967, p = 0.026), nheA (χ² = 3.963, p = 0.046), and nheB (χ² = 3.957, p = 0.047), and to be resistant to penicillin/trimethoprim/sulfamethoxazole (χ² = 4.643, p = 0.031). Considering a limitation of ERIC-PCR as to the repetitive capabilities, genotypic diversity analysis of B. cereus based on the more reproducible methods, i.e., MLST and genomic sequencing (Nguyen and Tallent, 2019; Shen et al., 2021; Zhang et al., 2020) might be more informative in the future study.