Isolation, Identification and Pathogenicity Analysis of Proteus mirabilis in Cynomolgus Monkey From Yunnan, China

ABSTRACT

Background

Proteus mirabilis (P. mirabilis) is an opportunistic zoonotic pathogen that can be mutually transmitted with humans (Homo sapiens); knowledge on P. mirabilis infection in cynomolgus monkeys (Macaca fascicularis) remains limited.

Objectives

To isolate and characterize the diarrhoeal pathogen from a cynomolgus monkey; a bacterial strain MF103012 was isolated from the faeces of a diarrhoeal cynomolgus monkey.

Methods

P. mirabilis was identified via morphology, biochemical profiling and 16S rRNA sequencing. Polymerase chain reaction (PCR) was used to detect key virulence genes. Antibiotic susceptibility was determined by Kirby–Bauer disc diffusion using 24 antibiotics of P. mirabilis. Pathogenicity was assessed by intraperitoneal injection in mice.

Results

In our study, we successfully isolated and identified P. mirabilis from the faeces of diarrhoeal cynomolgus monkey. The isolated strain carried ureC, zapA, rsmA, hpmA, fliL, ucaA and atfA virulence genes. Strain MF103012 was susceptible to 19 antibiotics but resistant to vancomycin and erythromycin, and it induced morbidity and one case of mortality in mice.

Conclusions

Endowed with the trait of cross‐species transmission, P. mirabilis has become an opportunistic pathogen that requires focused attention in both the prevention and control of zoonoses and clinical anti‐infective therapy. The results of this study improve our understanding of the isolated and identified P. mirabilis from the intestinal tract of diarrhoeal cynomolgus monkeys, and also lays groundwork for diagnosis and treatment of P. mirabilis in primate colonies.

This study demonstrated that the pathogen responsible for diarrhoea in a cynomolgus monkey was identified as Proteus mirabilis (P. mirabilis) through methods including bacterial isolation and culture, morphological examination, biochemical profiling, 16S rRNA sequencing, analysis of key virulence genes and bacterial artificial infection tests. The isolated strains exhibited susceptibility to 19 antibiotics. These findings enhance our understanding of P. mirabilis isolated from the faeces of diarrhoeal cynomolgus monkeys.

1. Introduction

Proteus mirabilis is a member of the genus Proteus in the Enterobacteriaceae (Shelenkov et al. 2020) and is an opportunistic zoonotic pathogen, capable of causing severe diseases when host immunity is compromised or environmental conditions change (Cohen‐Nahum et al. 2009). P. mirabilis has been documented to infect various mammals, including humans (Homo sapiens) (Chakkour et al. 2024), rhesus monkeys (Wu et al. 2024), pigs (Sus) (Qu et al. 2022), sheep (Ovis aries) (Abdollahi et al. 2022), chicken (Gallus domesticus) and cattle (Bos primigenius taurus) (Sanches et al. 2021), among others, leading to gastroenteritis (Mobley 2019), haemorrhagic meningoencephalitis (Coskun et al. 2016), endocarditis (Liu et al. 2015) and arthritis (Sbiti et al. 2017). Additionally, it may progress to bacteraemia and sepsis, potentially resulting in death (Mobley 2019; Armbruster et al. 2018).

Cynomolgus monkeys exhibit significant similarities to humans in terms of heredity, physiology, tissue structure, immune metabolism and other aspects (Yan et al. 2011). Additionally, cynomolgus monkeys possess advantages including genetic and physiological similarities to humans, established breeding protocols and well‐characterized immune responses. Thus, they are frequently employed as experimental animal models in biomedical research (Jia et al. 2022; Barbier and Bélanger 2003). The composition of gut microbiota and host immune response in cynomolgus monkeys is more similar to that in humans. Compared with rodents, the cynomolgus monkeys can more accurately simulate the physiological functions and pathological effects of P. mirabilis in humans. However, cynomolgus monkeys raised in open‐cage environments are often susceptible to pathogenic microorganisms that can lead to disease development.

Diarrhoea frequently affects colony health and experimental outcomes in cynomolgus monkeys (Li et al. 2021). Bacterial diarrhoea specifically results from excessive proliferation of pathogenic bacteria, which damage the biological barrier of intestinal mucosa, and cause endogenous infection and/or inflammatory bowel diseases (Qin et al. 2010). Diarrhoea in animals can result in metabolic disorders and nutritional deficiencies that may even culminate in mortality, which not only poses a serious health risk for the animals but also significantly impacts the quality of the experimental animals. Furthermore, it raises the risk of zoonotic disease transmission (Liu et al. 2025).

In recent years, numerous cases have emerged involving human and animal diseases caused by P. mirabilis. However, to our knowledge, there exists limited documentation concerning P. mirabilis infection in cynomolgus monkeys. This study aims to isolate, identify and assess virulence and antibiotic susceptibility of P. mirabilis from a diarrhoeal cynomolgus monkey to inform disease control in primate facilities.

2. Materials and Methods

2.1. Samples

The cynomolgus monkey (14 years old, female) employed in this study originated from the State Key Laboratory of Primate Biomedical Research, Kunming University of Science and Technology, China (SYXK [Yunnan] K2022‐0001). The animal was kept in an environment with a 12 h light–dark cycle, temperatures between 18°C and 26°C and humidity ranging from 40% to 70%. They were provided with commercially available monkey maintenance pellet feed and a sufficient number of fruits daily to meet their nutritional needs. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Kunming University of Science and Technology (approval code: KUST202404003) on 12 April 2024. The observed clinical symptoms included watery stools, lethargy, dehydration and decreased appetite. Following disinfection of the anus orifice and surrounding skin with 0.5% povidone‐iodine, faeces samples were collected into a sterile 2 mL centrifuge tube using a sterile disposable sampling swab. The sample was transferred to the laboratory within 30 min for parasites examination and bacterial culture.

2.2. Bacterial Isolation and Morphological Observation

Anal swab samples were thoroughly mixed in sterile saline solution. The liquid was streaked onto SS agar plate medium (catalogue: 20250529) and Colombian blood agar plate medium (catalogue: 20250428) (Hopebio, Qingdao, China) using an inoculation loop. The plates were incubated at 37°C for 24 h. Single colonies about 1–2 mm in size that were round, thin and translucent with a black dot at their centre were selected and subcultured by streaking onto SS agar plates (Hopebio, Qingdao, China), which were then cultured at 37°C for 18–24 h. Once purified single colonies were obtained, they were evenly spread on glass slides containing saline solution, air‐dried and subjected to Gram staining using a Gram stain kit (Solarbio, Beijing, China). The slide was placed on the stage of an optical microscope (Nikon, Tokyo, Japan), and the low‐power lens was used to find the observation target. Subsequently, the high‐power lens was utilized to finely adjust the focal length for achieving a clear field of view. A drop of cedarwood oil was then placed at the centre of this field, allowing for the examination of bacterial morphology, arrangement and staining characteristics using an oil immersion lens.

2.3. Physiological and Biochemical Identification

Selected colonies for detection were suspended in normal saline to prepare bacterial suspension to the 0.5 McFarland standard (≈1 × 10⁸ CFU/mL). Under sterile conditions, 100 µL of this bacterial suspension was transferred with an inoculation rod into the biochemically encoded identification tube CYZ‐15E of Enterobacteriaceae kit (CYZ‐15E Enterobacteriaceae Identification Kit, Binhe Microorganism Reagent, Hangzhou, China; cat. no. 250106) which were then sealed appropriately. According to the kit's instructions, these inoculated identification tubes were incubated at 37°C for 18–24 h before results were observed. The physiological and biochemical characteristics of the isolated strains were compared with the standard strains results in the Manual of Identification of Common Bacterial Systems (Garrity 2005) as well as those provided by the kit's instructions to preliminarily identify species.

2.4. Amplification and Sequencing of 16S rRNA of the Isolated Strains

The 16S rRNA gene was amplified using specific primers (27F, 5′‐AGAGTTTGATCCTGGCTCAG‐3′ and 1492R, 5′‐CGGCTACCTTGTTACGACTT‐3′) (Heuer et al. 1997). The polymerase chain reaction (PCR) mixture consisted of 12.5 µL of 2× Taq PCR Master Mix (Tiangen, Beijing, China), the upper and lower primers (10 µM) were 1 µL separately, and supplemented with distilled water up to a total volume of 25 µL. A small quantity of bacteria collected from a single colony on Columbia blood agar plate was utilized as the PCR template. The PCR amplification protocol included an initial denaturation at 95°C for 3 min, followed by 35 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 30 s and extension at 72°C for 1 min. A final extension step occurred at 72°C for 5 min. Subsequently, 5 µL PCR product underwent electrophoresis on 1.5% (w/v) agarose gel (the composition of TAE electrophoresis buffer: 2 mol/L Tris base, 1 mol/L acetic acid, 0.05 mol/L EDTA with pH 8.3) for 30 min at 120 V. Following this process, the agarose gel was stained with gene green nucleic acid dye (TianGen Biochemical Technology Co., Ltd, Beijing, China, cat. no. RT210) and screened using a UV illuminator (Pleasanton, California, USA). Target DNA fragment was purified utilizing a DNA gel extraction kit (Trelief, Beijing, China) for sequencing by Beijing Tsingke Biotech Co., Ltd. in Beijing, China.

2.5. Sequence Alignment Analysis and Phylogenetic Tree Construction

For sequence alignment analysis and phylogenetic tree construction purposes, the obtained gene sequences underwent homology comparison via BLAST (gap penalty: 15.00; gap length penalty: 6.66; delay divergent seps: 30%; DNA transition weight: 0.50) search against related sequences available in NCBI GenBank (MegAlign 7.0.26). Subsequently, MEGA11 software (number of trails: 1000; random seed: 111) facilitated the construction of a phylogenetic tree aimed at analysing and comparing the genetic relationships among P. mirabilis strains derived from various sources (accession numbers of nine reference sequences: KF769533, MT276300, MZ067152, OK271889, OL629191, OL629214, OL629230, PV354943, PV366835).

2.6. PCR Amplification of Virulence Genes of Isolated Bacteria

Although reverse transcription quantitative polymerase chain reaction (RT‐qPCR) and proteomics methods are mature approaches for quantifying gene transcription and protein expression, respectively, and are of great value in correlating the presence of genes with functional virulence expression, their implementation in this study was hindered by technical and resource limitations. Furthermore, the expression of virulence genes of P. mirabilis is strictly regulated by host environmental signals such as iron limitation, nutrient availability and host tissue adhesion (Cheng et al. 2019). In the absence of dedicated cell culture or animal models that accurately represent the primary clinical niche of this pathogen, simulating these signals in vitro poses significant challenges. Even if suitable models can be obtained, the inherent heterogeneity of P. mirabilis gene expression across different isolates necessitates a substantial number of impractical biological replicates to yield meaningful transcriptomic or proteomic data, thereby exacerbating resource limitations. Previous studies have demonstrated that conventional PCR detection of virulence genes including mrpA, rsbA and ureC revealed high pathogenicity confirmed through mouse pathogenicity tests; specifically, there was a concordance rate of 91.3% between PCR results and actual virulence phenotypes (Yi et al. 2020). Similarly, Nissanka et al. (2025) found that routine PCR detection indicated a relatively high prevalence (> 84%–92%) of key virulence genes (ureC, mrpA, speA, rsbA) in P. mirabilis isolates; quantitative polymerase chain reaction (qPCR) verification further illustrated variability in gene expression based on isolate origin without negating the functional correlation between gene presence and both virulence as well as antimicrobial resistance (AMR) phenotypes (Nissanka et al. 2025). These findings suggest that even without RT‐qPCR or proteomics validation methods employed for P. mirabilis virulence genes detection via PCR still possess interpretive value. Consequently, this study utilized PCR methodology to assess the virulence potential of isolated strains from P. mirabilis with notable biological significance. The P. mirabilis virulence genes were synthesized according to the reference Sun et al. (2020) (Table 1). The bacterial colony served as a template for amplification according to the described PCR system and cycling conditions in Section 2.4, utilizing 2× Taq PCR Master Mix (Tiangen, Beijing, China) and primers listed in Table 1. A total of nine virulence genes (ureC, zapA, rsmA, hpmA, mrpA, fliL, ucaA, pmfA and atfA genes) of the tested strains were analysed through electrophoresis.

TABLE 1.

Primers of virulence gene primers of isolated bacteria.

Gene	Prime sequence (5'→3')	Annealing temperatures/°C	Size/bp
ureC	F:GTTATTCGTGATGGTATGGG R:ATAAAGGTGGTTACGCCAGA	53	540
zapA	F:ACCGCAGGAAAACATATAGCCC R:GCGACTATCTTCCGCATAATCA	49	562
rsmA	F:TAGCGAGTGTTGACGAGTGG R:AGCGAGGTGAAGAACGAGAA	49	654
hpmA	F:CCAGTGAATTAACGGCAGGT R:CGTGCCCAGTAATGGCTAAT	40	550
mrpA	F:ACACCTGCCCATATGGAAGATACTGGTACA R:AAGTGATGAAGCTTAGTGATGGTGATGGTGATGAGAGTAAGTCACC	40	770
fliL	F:CTCTGCTCGTGGTGGTGTCG R:GCGTCGTCACCTGATGTGTC	50	560
ucaA	F:GTAAAGTTGTTGCGCAAAC R:TTGAGCCACTGTGGATACA	54	618
pmfA	F:CAAATTAATCTAGAACCACTC R:ATTATAGAGGATCCCTTGAAGGTA	50	382
atfA	F:CATAATTTCTAGACCTGCCCTAGCA R:CTGCTTGGATCCGTAATTTTTAACG	52	317

2.7. Antibiotic Sensitivity Testing

Twenty‐four common antibiotic sensitivity test paper discs (BKMAM, Changde, China, catalogue: 250523) were selected. The isolated strains of P. mirabilis were subjected to antibiotic sensitivity testing by the standard Kirby–Bauer paper disc diffusion method. The sensitivity, intermediacy and resistance of the isolated strains were assessed based on the size of the inhibition zones, perform the experiment in triplicate and calculate the mean value of the results, and the respective standards (Clinical and Laboratory Standards Institute (CLSI) 2025) for each antibiotic.

2.8. Bacterial Artificial Infection Test

The isolated and purified strain of P. mirabilis (MF103012) was inoculated into Tryptone Soy Broth (TSB) liquid medium and incubated overnight at 37°C with shaking at 140 r/min. The bacterial concentration in the bacterial solution was determined using the plate counting method. The bacterial solution concentration (1.5 × 10⁹ CFU/mL) was prepared through 10‐fold dilution method using TSB liquid medium (Hopebio, Qingdao, China) as the diluent. Bacterial viability was determined by plating for CFU confirmation. Ten healthy specific pathogen‐free (SPF) C57BL/6 mice aged 9–10 weeks were selected. At the beginning of the experiment, the body weight of the mice was controlled at 18–22 g. All mice were purchased from Shanghai Nanfang Model Biotechnology Co., Ltd. The mice were kept in a barrier environment that met the Chinese national standards. The environmental conditions were strictly controlled; the temperature was maintained at 22 ± 2°C, the relative humidity was 40%–70% and the light cycle was 12 h of alternating light/dark. The mice had free access to radiation‐sterilized feed and sterile drinking water, and the bedding materials were autoclaving shavings and corn cobs. All operations were performed in a biosafety cabinet to maintain a sterile state. The mice were randomly divided into experimental group and blank control group by simple random grouping method with five mice in each group. Each mouse was intraperitoneally injected with 0.5 mL bacterial solution in experimental group, while the control group mice were administered the same volume of TSB liquid medium via intraperitoneal injection. The diseased tissues of mice were collected and selective plating and molecular identification were carried out using the methods in Sections 2.2–2.5. Mortality and clinical symptoms in mice were recorded over 1 week, and the mice were kept under normal feeding management for 7 days. Mice were evaluated for euthanasia if they lost more than 20% of their initial body weight or were unable to eat and drink independently.

3. Results

3.1. Bacterial Morphological Observation

Anal swabs were inoculated on Columbia blood agar and SS medium; both media yielded morphologically identical colonies. The isolated bacteria were cultured on SS medium for 24 h, colonies appeared 1–2 mm in size with round, thin, moderate translucency with a black dot (H₂S positive) at their centre (Figure 1A). When cultured on Columbia blood agar medium for 24 h, it resulted in migratory growth (swarming motility) phenomena; the swarm zone diameter is 1–2 cm (Figure 1B). Gram staining was negative, and the bacteria exhibited blunt ends without spores (Figure 1C).

FIGURE 1.

The morphology of P. mirabilis isolated in cynomolgus monkey. (A) Morphological observations of the P. mirabilis isolate on SS agar medium. (B) Bacterial migration phenomenon observed in Colombian blood agar medium. (C) Microscopic examination of Gram staining (100×).

3.2. Physiological and Biochemical Identification Results

Physiological and biochemical results (Table 2) indicated that the isolated strain demonstrated motility and produced H₂S, tested positive for xylose, but negative for mannitol, indigo substrate, lysine, ornithine and scutellarin, which was consistent with the physiological and biochemical characteristics of P. mirabilis.

TABLE 2.

Physiological and biochemical characteristics of strain MF103012.

Test items	Results	Reference strain (O'Hara et al. 2000)	Test items	Results	Reference strain (O'Hara et al. 2000)
Mannitol	−	−	Motility test (semi‐solid agar)	+	±
H2S	+	±	Glucose fermentation	+	+
Phenylalanine	+	±	Lysine	−	−
Indigo substrate	−	±	Ornithine	−	±
VP test	−	±	Raffinose	−	±
MR test	+	±	Sorbitol	−	−
Citrate	+	±	Urea	+	±
Scutellarin	−	−	Xylose	+	±

Note: +, positive; −, negative. (1) Carbohydrate fermentation tests (mannitol, scutellarin, glucose fermentation, raffinose, sorbitol, xylose) were based on the colour change threshold of bromocresol violet indicator; yellow culture medium (pH < 5.2) was judged as positive for acid production (+), purple culture medium (pH > 6.8) as negative (−) and light purple or yellow green was suspicious. (2) H₂S production was evaluated on Kligler Iron Agar, with distinct black ferrous sulphide precipitate at the bottom as positive (+), no colour change as negative (−) and minimal faint grey or scattered tiny black specks as suspicious. (3) Indigo substrate was assessed by surface colour reaction, with results read within 10 min to avoid false delayed colouration. A stable rose‐red ring at the meniscus was positive (+), no colour change was negative (−) and faint pink, pale orange or uneven light red discolouration was suspicious. (4) Methyl red (MR) test: Using methyl red as indicator, bright red (pH < 4.4) was positive (+), clear yellow (pH > 6.0) was negative (−) and intermediate orange or pale pinkish‐orange discolouration (pH 4.4–6.0) was suspicious. (5) Urea test: Phenol red was used as the indicator for urease testing; pink to deep red (pH > 8.4) indicated positive (+), stable orange‐yellow (pH 6.8) with no shift indicated negative (−) and faint light pink discolouration was suspicious. (6) For remaining tests (phenylalanine, citrate, lysine, ornithine, motility test, VP test), suspicious results were defined as non‐definitive colour changes, weak slant discolouration, borderline turbidity or indefinite reaction signs. All suspicious results were validated via duplicate retesting, and final results were only recorded after consistent confirmation (Garrity 2005; O'Hara et al. 2000).

3.3. rRNA Gene Sequencing, Homology Comparison and Phylogenetic Analysis

The amplified PCR products were analysed using 1.5% agarose gel electrophoresis to isolate the target gene band of the expected size (1000–2000 bp) (Figure 2). The 16S rRNA gene sequence obtained from the P. mirabilis strain isolated in a cynomolgus monkey was determined to be 1401 bp in length.

FIGURE 2.

Identification of PCR‐amplified P. mirabilis with 16S rRNA gene in cynomolgus monkey. 1, sample; M, DL‐2K Marker; N, negative control.

Sequencing of the obtained 16S rRNA gene (GenBank accession no.: PV654206) was performed using the Sanger dideoxy method and has been reported in NCBI. Alignment analysis revealed that the isolated bacteria MF103012 (GenBank accession no.: PV654206) shared 99.6% identity with various sources of P. mirabilis, which confirmed that the isolated bacteria MF103012 (GenBank accession no.: PV654206) was P. mirabilis (Figure 3).

FIGURE 3.

Comparison of P. mirabilis (GenBank accession no. include: KF769533, MT276300, MZ067152, OK271889, OL629191, OL629214, OL629230, PV354943, PV366835) similarity between cynomolgus monkey (MF103012) (GenBank accession no.: PV654206) and other strains. ‘▲’ is the isolated bacterial strain in the cynomolgus monkey in this study.

These results indicate that the bacterial strain (MF103012) (GenBank accession no.: PV654206) isolated from anal secretion in diarrhoeal cynomolgus monkey clusters in a well‐supported clade (bootstrap ≥ 70%) alongside mink (Mustela vison) (OL629230) within P. mirabilis phylogeny (Figure 4). Furthermore, it clusters together with strains derived from raccoon (Procyon lotor) (OL629214), Labeo rohita, Oreochromis (PV366835), shrimp (KF769533) and human (OK271889) strains within P. mirabilis phylogeny, while forming a separate branch from fox (Vulpes) (OL629191), sika deer (Cervus nippon) (PV354943), chicken (MZ067152) and cattle (Bos taurus) of P. mirabilis (Figure 4). Therefore, based on analyses conducted through examination of its 16S rRNA gene sequence, we conclude that the pathogenic bacterium isolated from the cynomolgus monkey is indeed identified as P. mirabilis. It was most closely related to the P. mirabilis from mink (OL629230), while it had a distant relationship with P. mirabilis of fox (OL629191), sika deer (PV354943), chicken (MZ067152) and cattle (MZ067152) (Figure 4).

FIGURE 4.

Phylogenetic tree of 16S rRNA sequences among different reported strains of P. mirabilis including cynomolgus monkey (MF103012) (GenBank accession no.: PV654206). ‘▲’ is the isolated bacterial strain in the cynomolgus monkey in this study.

3.4. Amplification Results of Virulence Genes

The DNA of the isolated strains was used as a template to detect the virulence genes carried by the isolated strains by PCR. The results showed that we obtained the target bands including 317 bp (ureC gene), 540 bp (zapA gene), 562 bp (rsmA gene), 654 bp (hpmA gene), 770 bp (flagella gene fliL), 560 bp (pilus gene ucaA) and 382 bp (pilus gene atfA), while the pilus structural subunit genes mrpA and pmfA were no amplification observed at 35 cycles (Figure 5).

FIGURE 5.

Amplification results of virulence genes. 1, ureC gene (317 bp); 2, zapA gene (540 bp); 3, rsmA gene (562 bp); 4, hpmA gene (654 bp); 5, mrpA gene (no amplification observed at 35 cycles); 6, fliL gene (770 bp); 7, ucaA gene (560 bp); 8, pmfA gene (no amplification observed at 35 cycles); 9, atfA gene (382 bp); 10, negative control; M, DL‐2K Marker.

3.5. Antimicrobial Susceptibility Testing

To evaluate antibiotic sensitivity profiles for strain (MF103012) against 24 different chemical agents, we employed Kirby–Bauer disc diffusion methodology. Results indicated that this isolated strain exhibited sensitivity towards ceftriaxone, amikacin, gentamicin, ampicillin, cefazolin, ceftazidime, levofloxacin, cotrimoxazole, ciprofloxacin, norfloxacin, rifampicin, amoxicillin, cefoperazone, cefadroxil, ofloxacin, bacitracin, imipenem, meropenem and azithromycin; moderate sensitivity towards furazolidone, lincomycin and polymyxin B; and resistance towards vancomycin and erythromycin. While the intermediate results for polymyxin B, furazolidone and lincomycin did not affect treatment recommendations (Table 3).

TABLE 3.

Results of drug sensitivity testing of P. mirabilis strains isolated in cynomolgus monkey.

Drugs	Dose (µg/disc)	IZD/mm (mean ± SD, n = 3)	Sensitivity	Judgment standard of inhibition zone diameter (mm)
				Resistant	Medium	Sensitivity
Ceftriaxone	30	36.67 ± 2.08	S	≤ 13	14–22	≥ 23
Amikacin	30	18.00 ± 2.00	S	≤ 14	15–16	≥ 17
Gentamicin	10 ± 2.5	24.67 ± 1.15	S	≤ 12	13–14	≥ 15
Ampicillin	10	30.33 ± 0.58	S	≤ 13	14–16	≥ 17
Cefazolin	30	27.67 ± 1.73	S	≤ 14	15–17	≥ 18
Ceftazidime	30	37.67 ± 4.87	S	≤ 17	18–20	≥ 21
Levofloxacin	5	38.33 ± 2.89	S	≤ 13	14–16	≥ 17
Vancomycin	30	7.67 ± 1.53	R	≤ 9	10–11	≥ 12
Cotrimoxazole	23.75	27.33 ± 0.58	S	≤ 10	11–15	≥ 16
Ciprofloxacin	5	34.00 ± 2.00	S	≤ 15	16–20	≥ 21
Norfloxacin	10	36.00 ± 1.00	S	≤ 12	13–16	≥ 17
Erythromycin	15	7.67 ± 2.08	R	≤ 13	14–22	≥ 23
Rifampicin	5	14.33 ± 1.53	S	≤ 6	7–9	≥ 10
Amoxicillin	10	36.00 ± 4.00	S	≤ 5	6–9	≥ 10
Furazolidone	300	7.33 ± 1.53	I	≤ 5	6–9	≥ 10
Lincomycin	2	6.67 ± 2.52	I	≤ 5	6–11	≥ 12
Polymyxin B	300	10.33 ± 1.29	I	≤ 8	9–11	≥ 12
Cefoperazone	75	37.33 ± 1.29	S	≤ 15	16–20	≥ 21
Cefadroxil	30	25.00 ± 1.00	S	≤ 13	14–16	≥ 17
Ofloxacin	5	36.33 ± 1.15	S	≤ 12	13–15	≥ 16
Bacitracin	0.04	16.00 ± 2.00	S	≤ 6	7–9	≥ 10
Imipenem	10	35.33 ± 1.53	S	≤ 19	20–22	≥ 23
Meropenem	10	36.33 ± 2.24	S	≤ 19	20–22	≥ 23
Azithromycin	15	16.33 ± 2.89	S	≤ 12	—	≥ 13

Note: Clinical and Laboratory Standards Institute (CLSI) 2025.

Abbreviations: I, intermediacy; R, resistance; S, sensitivity.

3.6. Bacterial Artificial Infection Test Results

Each mouse was intraperitoneally injected successively with 0.5 mL bacterial solution in the experimental group; the mice were recorded every 6 h for clinical symptoms. Following inoculation, one mouse in the experimental group died within 24 h, while other mice in the experimental group exhibited symptoms including lethargy, disheveled fur, reduced activity, reduced food intake and emaciation. The symptoms of the other mice in the experimental group gradually disappeared after 72 h and they returned to normal. However, no abnormal phenomena occurred for 7 days in the mice of the control group. The bacteria isolated from the tissues of dead mouse were identified as P. mirabilis by molecular biological analysis. Surviving mice were euthanized by intraperitoneal injection of sodium pentobarbital at 150 mg/kg. The results indicated that the isolated strain MF103012 elicited acute infection with one fatality and transient morbidity in the mice model; it also indicated the effect of the limited sample size on generalizability.

4. Discussion

Studies have shown that P. mirabilis is associated with a variety of digestive diseases, and it can spread among a variety of primates, including humans. Gong et al. (2019) demonstrated that a strain of P. mirabilis from human (Homo sapiens) could cause damage to various intestinal segments in mice. Povar (1965) reported that rhesus monkey infected with P. mirabilis exhibited clinical symptoms consistent with acute and chronic gastroenteritis. Hu et al. (2020) found that a strain of P. mirabilis derived from a dog caused severe intestinal haemorrhage, diarrhoea and death in mice. Hamilton et al. (2018) and Drzewiecka (2016) also noted the association between P. mirabilis and human Crohn's disease, suggesting it as a potential pathogen responsible for gastrointestinal disease. In this study, P. mirabilis was isolated from the intestinal tract of diarrhoeal cynomolgus monkey (its pathogenic characteristics are similar to those reported) and identified it as the primary pathogen responsible for diarrhoeal disease in cynomolgus monkey. Notably, mink (Mustela vison) and human (Homo sapiens) are significantly different taxonomically. However, the isolated strain MF103012 obtained from cynomolgus monkey clusters with P. mirabilis isolates derived from mink (Mustela vison) to form a monophyletic branch (bootstrap value = 61), indicating a close evolutionary relationship. This suggests that this bacterium may possess the ability to adapt to hosts across different taxonomic orders. Consequently, it implies that mink (Mustela vison) could serve as significant natural reservoir of P. mirabilis in nature or share the same infection source with non‐human primates (NHPs). Furthermore, P. mirabilis is recognized as an important opportunistic pathogen capable of causing diseases such as urinary tract infections in humans (Hassuna et al. 2025). The potential for cross‐species transmission between primates and mink indicates that these animal‐derived strains may be transmitted to humans through direct contact or via the food chain (e.g., consumption of animal‐derived foods), thereby posing a potential public health threat.

The identified virulence gene repertoire of MF103012 reflects key pathogenic mechanisms that significantly contribute to its pathogenicity (Debnath et al. 2018; Pathirana et al. 2018), thereby heightening its potential threat to public health. It has been reported that virulence factors of P. mirabilis include pilus, flagella, urease, haemolysin, protease, biofilm and quorum sensing (Armbruster et al. 2018). Pilus genes including ucaA, pmfA and atfA of P. mirabilis are critical virulence factors in the early stage of infection, which can make the bacteria adhere to host cells while facilitating invasion through multiple pathways leading to toxin release; concurrently, virulence genes regulate other bacterial factors promoting host infection spread and immune evasion (Sun et al. (2020; Jansen et al. 2004; Hussein et al. 2020). Pathogenic genes including ureC, rsmA, mrpA and atfA play important roles in the process of bacterial adhesion and aggregation (Shi et al. 2016; Barbour et al. 2013; Scavone et al. 2016; Elhoshi et al. 2023). Studies have indicated that the proportions of rsmA, mrpA and atfA in the isolates were 81.25%, 65.91%–100% and 64.77%–100%, respectively (Sun et al. 2020; Elhoshi et al. 2023; Danilo de Oliveira et al. 2021), and ureC is the most frequently identified virulence gene of P. mirabilis isolated from the faeces of diarrhoeal animals; the isolation rate was 90.91%–100% (Sun et al. 2020; Elhoshi et al. 2023). Therefore, ureC is often used as a target gene for the identification of P. mirabilis. The metalloproteinase zapA can degrade defensins in the immune system, resulting in a decrease in immunoglobulin levels which may lead to disease manifestations; zapA was detected in 85.8%–100% of isolates (Sun et al. 2020; Elhoshi et al. 2023; Danilo de Oliveira et al. 2021). Pathogenic gene fliL serves as a flagellar matrix protein influencing bacterial motility; fliL was detected in 56.82% of isolates (Belas and Suvanasuthi 2005; Sun et al. 2020). hpmA is an extracellular protein exhibiting haemolytic activity by integrating into cell membranes causing perforation, leading to tissue damage, exfoliation and nutrient release, and it is essential for tissue destruction and bacterial transmission (Mobley et al. 1991; Cestari et al. 2013); hpmA was detected in 90%–100% of isolates (Elhoshi et al. 2023; Danilo de Oliveira et al. 2021). In this study, nine virulence genes of P. mirabilis including ureC, zapA and rsmA were amplified via PCR. It demonstrated that seven reported virulence genes including ureC, zapA, rsmA, hpmA, fliL, ucaA and atfA of P. mirabilis were present in the isolated strain MF103012. Coupled with the results of animal pathogenicity test results, we speculate this might be an important reason for the MF103012 strain to exhibit strong pathogenicity to mice and induce severe diarrhoea in cynomolgus monkey. In this study, the associated genes mrpA and pmfA were not detected in P. mirabilis strain MF103012. These genes encode fimbrial structural proteins that play a crucial role in bacterial adhesion, aggregation and biofilm formation, which are closely linked to host colonization (Jansen et al. 2004; Scavone et al. 2016). The absence of these genes may lead to reduced biofilm formation and diminished recognition by the host's innate immune system, thereby alleviating immune clearance (Armbruster et al. 2018; Jansen et al. 2004; Pearson et al. 2008). However, it may also result in decreased mucosal adhesion and colonization efficiency, exerting dual effects on host adaptability. Such variations in the virulence gene profile may reflect the adaptive evolution of the strain to the microenvironment of cynomolgus monkeys. By downregulating genes associated with broad‐spectrum colonization, the strain achieves host‐specific adaptation to cynomolgus monkeys, which may partially explain its preference for the particular host.

Studies have shown that P. mirabilis represents prevalent multi‐drug resistant bacterium (MDRO) (Lei et al. 2016; Li et al. 2016). With the emergence of MRDO, the resistance to β‐lactams, aminoglycosides, tetracyclines and sulphonamides has increased, as well as the biofilm produced by the bacteria has enhanced resistance to antibiotics and inflammatory response; the difficulty of disease treatment has increased and the harm to the aquaculture industry and public health safety has gradually increased (Schaffer and Pearson 2015; Algammal et al. 2021; Fusco et al. 2017). The Kirby–Bauer disk diffusion method presents certain clinical limitations in the antimicrobial susceptibility testing of P. mirabilis. This bacterium exhibits swarming growth, which can significantly interfere with the interpretation of inhibition zones and may result in missed detection of low‐level resistance. Furthermore, this method does not directly provide the minimum inhibitory concentration (MIC), making it difficult to accurately guide dose optimization for severe infections. According to the CLSI M100 standard (Clinical and Laboratory Standards Institute (CLSI) 2025), the disk diffusion method categorizes strains as susceptible (S), intermediate (I) or resistant (R) based on inhibition zone diameters, which pertains to categorical resistance interpretation. In contrast, MIC determined by reference methods such as broth microdilution serves as a quantitative indicator of resistance that directly reflects bacterial resistance levels and drug exposure. Therefore, in future clinical practice, the Kirby–Bauer disk diffusion method may be utilized for preliminary screening of clinical antimicrobial agents, while quantitative MIC is more appropriate for predicting therapeutic efficacy and resistance surveillance. In this study, the strain isolated from the faeces of diarrhoeal cynomolgus monkey was identified followed by detection of bacterial virulence genes and antimicrobial susceptibility testing. These findings have important implications for preventing bacterial diarrhoea and managing antimicrobials in primates. Firstly, establishing regular microbiota monitoring programmes is recommended for long‐term colony surveillance; these programmes should involve periodic faecal sampling along with 16S rRNA gene sequencing to track shifts in commensal microbiota composition‐enabling early detection of dysbiosis that may predispose primates to diarrhoea and emergence of AMR. Secondly, selective antibiotic use policies should be formally implemented based on the documented resistance profile in this study; vancomycin and erythromycin should be strictly avoided in life‐threatening infections due to their high resistance to the isolated pathogens, while alternative antimicrobial agents with demonstrated susceptibility (e.g., third‐generation cephalosporins or fluoroquinolones, where appropriate) should be prioritized to minimize selection pressure on resistant strains. Thirdly, rigorous infection control procedures tailored towards preventing faecal‐oral transmission are essential; these include enhanced environmental disinfection protocols within primate enclosures using chlorine‐based disinfectants effective against enteric pathogens alongside strict hand hygiene practices for animal care staff, and separation of diarrhoeal monkeys from healthy individuals to prevent cross‐contamination. Lastly, probiotic interventions should be considered as an adjunctive strategy aimed at restoring commensal microbiota diversity in diarrhoeal monkeys; select probiotic strains such as Bifidobacterium, with proven efficacy in modulating gut microbiota composition and reducing diarrhoea severity in primates, should be administered under veterinary supervision alongside necessary antimicrobial therapy to mitigate dysbiotic effects caused by antibiotics (He et al. 2016). In conclusion, these targeted measures not only address immediate challenges related to diarrhoea management and AMR in primate colonies but also contribute positively towards long‐term colony health, zoonotic disease prevention and rational use of antimicrobials in laboratory and captive primate settings.

5. Conclusions

In this study, a potent pathogenic bacterium was successfully isolated from faeces of diarrhoea cynomolgus monkey. Through morphological, physiological and biochemical tests and molecular biological identification, it was proved that the isolated strain was P. mirabilis. The strain carried ureC, zapA, rsmA, hpmA, fliL, ucaA and atfA virulence genes. Drug susceptibility testing indicated that the strain was sensitive to 19 antibiotics including ceftriaxone, amikacin and gentamicin, while resistant to vancomycin and erythromycin. Results herein offer valuable insights into comprehensive prevention and control strategies of clinical bacterial diarrhoea in cynomolgus monkeys.

Author Contributions

Heling Li: methodology, writing – original draft. Ziyao Qian: methodology, writing – original draft. Lan Luo: methodology, writing – original draft. Xinglong Chen: resources. Baohong Tian: resources. Zhigang Chen: resources. Hong Wang: project administration, supervision, writing – review and editing.

This study was supported by the Natural Science Foundation of Yunnan Province (202102AA100053).

Ethics Statement

The authors confirm that the ethical policies of the journal, as noted on the journal's author guidelines page, have been adhered to and the appropriate ethical review committee approval has been received.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

All data in this study are available from the corresponding author upon reasonable request.