**Background:** Cronobacter spp. is a foodborne pathogen that causes meningitis, sepsis, and necrotizing enterocolitis in infants, with mortality rates of 40–80%. Powdered infant formula (PIF) and its processing environment are major contamination routes. Antibiotic resistance in Cronobacter is increasing, but resistance mechanisms remain poorly understood. This study aimed to identify, type, and characterize antibiotic resistance in Cronobacter isolates from PIF and processing environments, and to explore resistance mechanisms using transcriptomics.
**Methods:** Thirty-five Cronobacter strains isolated from PIF and processing environments across different regions of China were studied, along with eight type strains and two reference strains. Strains were identified by 16S rRNA sequencing and typed by multilocus sequence typing (MLST) using seven housekeeping genes (atpD, fusA, glnS, gltB, gyrB, infB, pps). Phylogenetic trees were constructed using MEGA7.0 with neighbor-joining analysis (1000 bootstrap replicates). Antimicrobial susceptibility was tested against 20 antibiotics from 9 categories following CLSI M100 standards. For transcriptomics, representative resistant strain CS14 (ST42) was cultured with and without 4 μg/mL erythromycin; RNA was extracted and sequenced. Differentially expressed genes were identified and validated by qRT-PCR on 10 selected genes in resistant (CS14, CS17) and susceptible (CS6, CS9) strains.
**Key Results:** 16S rRNA sequencing identified 27 C. sakazakii (77.14%), 5 C. malonaticus (14.29%), 2 C. turicensis (5.71%), and 1 C. dublinensis (2.86%). MLST yielded 35 sequence types (STs) distributed across 15 homologous complexes. Three new STs were discovered: C. dublinensis CD31 (ST788), C. sakazakii CS32 (ST789), and CS33 (ST790). Dominant STs were ST1 (14.42%), ST4 (18.27%), and ST64 (11.54%). All 35 isolates (100%) were resistant to erythromycin; 45.71% were resistant to sulfamethoxazole/trimethoprim; resistance to other antibiotics ranged from 2.86% (ofloxacin, chloramphenicol) to 37.14% (neomycin). All isolates were sensitive to ciprofloxacin. Multidrug-resistant strains (resistant to ≥3 antibiotic classes) accounted for 68.57% (24/35). The most resistant strain (CS14, ST42) showed 13-fold multidrug resistance. Transcriptomic analysis identified 2,324 differentially expressed genes (1,154 upregulated, 1,170 downregulated) under erythromycin stress. Of 94 drug resistance-related genes, 77 were differentially expressed (35 upregulated, 42 downregulated). Key findings included upregulation of methyl-accepting chemotaxis proteins (2.06–2.66-fold), flagellar motor switch genes fliG (1.76-fold) and fliM (1.83-fold), multidrug efflux system genes mdlA (2.05-fold) and yojI (1.96-fold), and spermidine transport genes potA–potD (2.06–3.10-fold). qRT-PCR confirmed transcriptome results: in resistant strains CS14 and CS17, 8 of 10 genes were significantly upregulated (2^−ΔΔCt ≥ 2, p < 0.05) and 2 significantly downregulated (2^−ΔΔCt ≤ 0.5, p < 0.05); in susceptible strains CS6 and CS9, most genes were downregulated with no significant upregulation.
**Clinical Implications:** The high prevalence of multidrug resistance (68.57%) and universal erythromycin resistance among Cronobacter isolates from PIF underscores the need for ongoing surveillance and rational antibiotic selection. Fluoroquinolones (especially ciprofloxacin) remain effective treatment options. The identification of three novel STs expands the known genetic diversity of Cronobacter. The transcriptomic evidence that Cronobacter activates multidrug efflux systems via chemotaxis gene regulation under antibiotic stress provides a mechanistic basis for resistance that could inform development of new antibacterial strategies. These findings have direct public health significance for controlling Cronobacter contamination in PIF production and treating infections in infants.