All samples were obtained from routine diagnostic postmortem examinations conducted at the IFER, Errachidia, Morocco. No live birds were handled or experimentally manipulated. Adult females, chicks, and juveniles died naturally or were submitted for necropsy as part of standard veterinary investigations, while egg samples originated from routine incubation monitoring. As no additional interventions were performed, formal ethical approval was not required. Institutional management provided a written exemption, and all procedures complied with national regulations and institutional animal welfare guidelines.

This study was conducted during the 2025 production season (encompassing the breeding and rearing cycle from approximately January to June, aligned with the species’ natural reproductive period under captive management) at the IFER captive breeding facility in Errachidia, Morocco. The facility is located in a semi-arid region (coordinates approximately 31.93°N, 4.42°W) and serves as a dedicated center for the conservation and reintroduction of the African Houbara bustard. All sampling, necropsies, and laboratory analyses were performed on-site or at affiliated laboratories within the IFER premises to ensure biosecurity and compliance with conservation protocols.

This study was designed as a cross-sectional diagnostic investigation conducted during the 2025 production season. A purposive sampling strategy was used, in which all cases meeting the inclusion criteria were included in the study. In total, 110 samples were collected and categorized into distinct stages of the reproductive and early life cycle of C. undulata: early embryonic death (n = 50), chicks (n = 25), juveniles (n = 25), and adult females (n = 10). The selected sample size reflected the range of age categories encountered and corresponded to routine diagnostic submissions received during the study period to support bacteriological and molecular analyses.

A total of 110 cases were examined, comprising 10 adult females, 25 chicks, 25 juveniles, and 50 early embryonic death samples. Among adult females, six cases of sudden death were recorded, while four additional birds exhibited lethargy and anorexia prior to death, including two cases complicated by dystocia and oviductal impaction. Of the chicks examined, 15 died suddenly, and 10 displayed clinical signs of anorexia and lethargy before death. In the juvenile group, 12 birds died suddenly without overt clinical signs, whereas 13 showed systemic illness characterized by ruffled feathers, marked lethargy, and anorexia. Fifty egg samples exhibiting early embryonic mortality at 5–7 days of incubation were also included.

Necropsies were performed by licensed veterinarians following established standard protocols [27]. Postmortem examination revealed macroscopic lesions consistent with E. coli septicemia, including fibrinous pericarditis, perihepatitis, airsacculitis, exudate accumulation, peritonitis, and visceral adhesions.

All samples were collected aseptically from internal organs, including the heart, liver, lung, and spleen, as well as oviduct and ovary tissues from adult females when indicated by clinical and pathological findings. Eggshell surfaces were decontaminated by immersion in 70% ethanol for 5–10 s, air-dried, aseptically cracked, and the internal contents pooled [28]. Samples were transported to the IFER laboratory in sterile stomacher bags under cooled conditions. Tissue samples were pooled and processed for bacterial screening and isolation.

Isolation and identification of E. coli were performed following previously described methods [29]. Briefly, pooled organ samples were pre-enriched in peptone water and incubated aerobically at 37°C for 24 h, followed by streaking onto MacConkey agar and EMB agar (Oxoid, Hampshire, UK) and incubation at 37°C for 24–48 h. Presumptive colonies were subcultured to obtain pure isolates. Biochemical identification was carried out using the VITEK®2 system (bioMérieux SA, Marcy-l’Étoile, France). Bacterial suspensions were adjusted to 0.5 McFarland, loaded into VITEK®2 GN cards (Ref. No. 21341), and processed according to the manufacturer’s instructions to obtain species-level identification within the Enterobacteriaceae family.

Antimicrobial susceptibility testing was performed on a representative subset of 32 isolates using VITEK®2 GN97 cards (Ref. No. 42000, bioMérieux SA, Marcy-l’Étoile, France) to assess susceptibility to 18 antimicrobial agents and screen for ESBL production. The disk diffusion method was additionally applied following EUCAST guidelines and the latest VITEK®2 clinical breakpoint tables (version 14.0, 2024) [30]. E. coli ATCC 25922 was used as the quality control strain. Resistance patterns were classified according to international definitions [31] as MDR (resistance to ≥1 agent in ≥3 antimicrobial classes) and XDR (resistance to ≥1 agent in all but one or two classes).

Genomic DNA was extracted from 20 overnight bacterial cultures grown to an optical density of 1.0–1.2 at OD600. Cell pellets were obtained by centrifugation at 8,000 × g for 5 min and processed using the Kylt® DNA/RNA Kit (Kybio Co., Ltd., Shenzhen, China) following the manufacturer’s instructions. DNA concentration and purity were assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), with acceptable purity defined as an A260/A280 ratio of 1.8–2.0.

Virulence-associated genes were detected using the Kylt® APEC qPCR Kit (Kybio Co., Ltd., Shenzhen, China), targeting 15 APEC-related genes, including adhesins (papC, tsh), invasion (ibeA), iron acquisition (iucD, iutA, irp2, iroN), toxins (astA, vat, F11, hlyF), protectins (sitA, cvi/cva, iss), and outer membrane protease (ompT). Four multiplex reactions (APEC1–APEC4) were prepared in a final volume of 20 µL containing 16 µL master mix and 4 µL template DNA. Amplification was performed using a QuantStudio™ 5 Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA) with cycling conditions of 95°C for 10 min, followed by 42 cycles of 95°C for 15 s and 60°C for 60 s. Samples with Ct ≤ 30 were considered positive. Negative controls were included in all runs. Targeted genes are summarized in Table 1.

Data were analyzed using Microsoft Excel® and SPSS® version 26 (IBM Corp., Armonk, NY, USA). Descriptive statistics were used to calculate frequencies and percentages of E. coli isolates and virulence genes. Fisher’s exact test was applied to compare prevalence among groups due to small sample sizes. Statistical significance was set at p ≤ 0.05. Confidence intervals are presented in Table 2.

E. coli was isolated from 69 of 110 cases (62.7%). The highest isolation rate was observed in juveniles (84%), followed by chicks (80%), adult females (60%), and early embryonic death (44%). Sixty-two clinical isolates were identified as E. coli using the VITEK®2 system, with identification confidence ranging from 96% to 99%. Minor deviations in expected biochemical reactions (PHOS-81 and dTAG-22) were detected in a limited number of isolates; however, overall identification confidence remained high. Confidence intervals were calculated for each group.The prevalence of E. coli isolates across different life stages of C. undulata is summarized in Table 2.

Two clustering patterns were identified: phenotypic clustering based on antimicrobial susceptibility profiles and genotypic clustering based on virulence gene distribution. Figure 1 presents a heat map illustrating resistance percentages among E. coli isolates recovered from early embryonic death, chicks, juveniles, and adult females.

Heatmap showing the percentage resistance of Escherichia coli isolates to individual antibiotics, stratified by sample source (early embryonic death, chicks, juveniles, and adult females) in captive African Houbara bustards.

Isolates originated from eggs (n = 8), chicks (n = 7), juveniles (n = 12), and adult females (n = 5). No resistance was detected to fosfomycin, amikacin, gentamicin, neomycin, or chloramphenicol. In contrast, all isolates exhibited resistance to enrofloxacin (ENR) and ampicillin (AMP).

The distribution of phenotypic antibiotic resistance patterns is shown in Table 3. A total of 25 distinct multidrug-resistant (MDR) patterns were identified, ranging from resistance to 3 to 15 antibiotics. The most frequent resistance profile consisted of AMP, marbofloxacin (MAR), pradofloxacin (PRA), doxycycline (DOX), tetracycline (TET), nitrofurantoin (NIT), and trimethoprim–sulfamethoxazole (TMP-SMX) and was observed in three isolates. Several additional resistance profiles involved combinations of 8–10 antibiotics. One isolate demonstrated resistance to 15 antibiotics, indicating an extensively drug-resistant (XDR) phenotype. AMP, fluoroquinolones, TET, NIT, and TMP-SMX were the most frequently represented antimicrobial classes across resistance patterns. The distribution of MDR among sample groups is presented in Figure 2. Of the 32 E. coli isolates examined, 31 (96.9%) were classified as MDR. All isolates from chicks, juveniles, and adult females exhibited 100% MDR. Only one isolate from the early embryonic death group was non-MDR, resulting in an MDR prevalence of 87.5% in that group.

Distribution of multidrug-resistant (MDR) Escherichia coli isolates across sample groups (early embryonic death, chicks, juveniles, and adult females).

Figure 3 illustrates the distribution and frequency of MDR patterns across life stages. The most predominant resistance combination consisted of β-lactams, fluoroquinolones, TET, NIT, and sulfonamides in chicks (71.5%) and eggs (75%). Adult females exhibited a higher proportion (60%) of resistance to β-lactams, fluoroquinolones, TET, and NIT. Juvenile isolates showed multiple resistance combinations. These findings indicate substantial variation in MDR patterns among life stages, although overall MDR prevalence remained consistently high (96.9%) across all groups.

Statistical analysis of resistance rates by antimicrobial class is presented in Table 4. Due to small sample sizes in certain groups, Fisher’s exact test was applied instead of the chi-square test. No statistically significant differences in resistance rates were detected among early embryonic death, chicks, juveniles, and adult females (p > 0.05), largely due to uniform resistance and susceptibility patterns across most antimicrobial classes.

The prevalence of extended-spectrum β-lactamase–producing E. coli across life stages is shown in Figure 4. ESBL-producing isolates were detected in 14% of juvenile samples and 9% of egg samples, indicating stage–specific occurrence and highlighting the need for targeted monitoring during early developmental stages.

The prevalence of avian pathogenic E. coli–associated virulence genes is summarized in Table 5. Most isolates harbored irp2, iss, iucD, iutA, ompT, and sitA, which were consistently detected across isolates from chicks, early embryonic death, juveniles, and adult females. In contrast, papC and tsh were absent in all isolates. Greater virulence gene diversity was observed in chick and egg isolates compared with juvenile isolates.

Gene prevalence and patterns of gene absence by isolate group (n = 20) are illustrated in Figure 5. A conserved core gene set, including protectin- and invasion-associated genes, was detected in 100% (20/20) of isolates, indicating their essential role in bacterial survival and pathogenicity. Iron acquisition genes exhibited the highest collective prevalence (17/20–20/20), emphasizing their importance in nutrient acquisition. Conversely, adhesin and toxin genes were detected at low frequencies (0/20–3/20), reflecting heterogeneity in pathogenic potential.

Gene prevalence and absence patterns for 15 avian pathogenic Escherichia coli-associated virulence genes among 20 selected E. coli isolates, grouped by source (n = 20). Prevalence numbers (x/20) = Total number of positive isolates (isolates possessing the gene). The presence of a group letter next to a gene indicates absence of that gene in the corresponding group (CH = chicks, EED = early embryonic death, F = adult female, and J = juvenile).

The mean virulence gene burden across life stages of C. undulata is shown in Figure 6. Juvenile isolates exhibited the lowest standard deviation, indicating consistent gene burden values within this group, whereas chick isolates showed the highest standard deviation, reflecting greater variability in virulence gene distribution.

Mean number of virulence genes (gene burden) carried by Escherichia coli isolates across different life stages of Houbara bustards. EED = Early embryonic death.

Only limited literature has addressed MDR E. coli and APEC-associated virulence factors in C. undulata, resulting in a substantial gap in understanding the pathogenic potential and epidemiology of E. coli in endangered avian species [8, 32, 33]. Consequently, the present findings represent one of the first comprehensive investigations integrating phenotypic MDR profiling with genotypic virulence characterization, including the detection of XDR isolates circulating within captive flocks [34, 35]. The combined phenotypic–genotypic framework applied in this study enabled detailed assessment of pathogen dynamics across life stages, supported by profiling of 15 virulence-associated genes [36, 37]. Isolates recovered from breeders, early embryonic death, chicks, and juveniles were associated with elevated mortality in captive African Houbara bustards (C. undulata).

The overall isolation rate of E. coli (62.7%) in C. undulata was notably high. This contrasts with a previous report describing 100% prevalence in postmortem samples from Chlamydotis macqueenii [38]. The highest isolation rates were recorded in juveniles (84%) and chicks (80%), supporting the association between immature immune function and age-related susceptibility in avian species. Environmental stressors have been widely reported as predisposing factors for colibacillosis in young birds [39–41]. Isolation rates in adult females (60%) during the breeding season were consistent with previous findings identifying E. coli as the predominant pathogen in clinical reproductive disorders of captive Houbara bustards [6]. Detection of E. coli in early embryonic death samples (44%) supports the hypothesis of horizontal or vertical transmission in avian systems [42–45], although recent studies indicate that true vertical transmission in poultry remains limited, with environmental contamination playing a more dominant role [46, 47].

Marked phenotypic resistance was observed against AMP, ENR, MAR, TET, NIT, DOX, and PRA across all isolates, in agreement with previous reports from Houbara bustards and poultry production systems [38, 42, 48, 49]. Earlier studies documented low ENR resistance in Houbara bustards and other wild birds [38, 50]; therefore, the widespread ENR resistance observed here indicates a concerning shift in resistance dynamics. Notably, PRA resistance was identified for the first time in Houbara bustards, highlighting the emergence of novel resistant phenotypes.

In contrast, complete susceptibility was observed to fosfomycin, amikacin, gentamicin, and chloramphenicol, supporting their potential role as alternative therapeutic options. These findings align with previous observations in poultry, where universal susceptibility to fosfomycin was reported [51]. Low resistance levels against cefalotin, cefovecin, ceftiofur, and neomycin in juvenile isolates further suggest limited but promising treatment alternatives for APEC infections in captive bustards.

Alarmingly, one isolate exhibited resistance to 15 antibiotics, including imipenem, indicating the emergence of XDR strains (Table 3). Similar findings have been reported in poultry populations [35]. MDR patterns were consistently distributed across all life stages, predominantly involving β-lactams, fluoroquinolones, TET, NIT, and sulfonamides (40%–75%), corroborating earlier reports linking resistance gene accumulation with treatment failure risks [32, 33].

The indiscriminate use of antibacterial agents across human, veterinary, and agricultural sectors has been implicated in the global dissemination of ESBL-producing bacteria [52, 53]. Sub-therapeutic antibiotic exposure in poultry has been proposed as a driver for environmental dissemination of ESBL-producing E. coli, posing substantial public health risks [22, 54]. While transmission of ESBL-producing E. coli between poultry products and humans has been documented [55], data regarding transmission from captive Houbara bustards to personnel remain unavailable.

In the present study, ESBL-producing E. coli were detected primarily during early life stages, with positivity rates of 15% in early embryonic death and 10% in juveniles. These findings are consistent with previous observations suggesting early horizontal exposure or breeder-associated transmission [36]. Juveniles, due to immunological immaturity, may facilitate persistence and dissemination of resistant strains [32, 33]. The absence of statistically significant differences among age groups (p > 0.05) supports flock-level dissemination of MDR E. coli, reinforcing the need for stringent biosecurity during incubation and early rearing.

APEC isolates are characterized by diverse virulence determinants, including adhesins, toxins, iron acquisition systems, and invasion-associated factors [56, 57]. Distinct life stage–specific virulence patterns were identified in captive C. undulata. Early embryonic death and chick isolates exhibited the highest virulence gene burdens, with most isolates harboring ≥10 genes, including iss, iucD, iutA, ompT, hlyF, and sitA. Similar gene burden–based pathogenicity classifications have been reported in poultry [58], although others have questioned the discriminatory power of virulence gene profiling alone [59].

Juvenile isolates displayed intermediate virulence profiles, potentially reflecting a transitional phase with reduced pathogenic potential. Adult female isolates retained core virulence genes, such as iss, iutA, and ompT, but exhibited lower overall gene richness, consistent with previous reports in Houbara bustards [6]. Reduced virulence gene prevalence may be associated with enhanced immune competence or cumulative environmental exposure.

The absence of certain adhesion-associated genes further emphasizes that E. coli pathogenicity is multifactorial and influenced not only by virulence gene content but also by host immunity, susceptibility, and environmental and management conditions [8, 33, 60, 61].