Characterizing microbiome ecology and developing targeted interventions for Salmonella control in low-moisture wheat systems

Abstract

Low-moisture foods (LMFs), including wheat and wheat-based products, have been considered "microbiologically safe" due to their low water activity (aw < 0.85). However, wheat grains and their products have emerged as a novel vehicle for the transmission of foodborne illnesses to humans. Current knowledge of microbiological hazards in this system relies heavily on culture-dependent, pathogen-specific methods that provide a very narrow view of the broader microbial ecology of the wheat milling system. Culture-independent approaches, such as 16S rRNA sequencing, can provide a more comprehensive characterization of microbial communities and can offer predictive knowledge on the wheat microbiome, which can aid in the development of mechanistically informed intervention strategies. However, the lack of systematic benchmarking in the LMF environment leaves significant uncertainty regarding the optimal bioinformatics workflow. To address these gaps, this research benchmarked a 16S rRNA sequencing protocol for microbiome analysis in wheat milling environment (Chapter 3); Characterize the wheat milling microbiome through taxonomic profiling, diversity analysis, and functional pathway prediction across different milling stages (Chapter 4); Determine the efficacy of sodium bisulfate (SBS) (Chapter 5); acidic tempering (Chapter 6) and non-thermal techniques (cold plasma, ultraviolet, and pulsed light) in reducing Salmonella in wheat during tempering and assess their impact on flour functionality, and baking quality (Chapter 7). Chapter 3 systematically evaluated 32 bioinformatics workflows (eight 16S rRNA pipeline X four taxonomic database) using wheat milling samples (n=160). At the family level, diversity metrics were conserved across pipeline-database combinations (Chao1: 22.97±2.20-24.92±2.04; Shannon: 2.59±0.19-2.74±0.18; InvSimpson: 10.63±1.25-11.27±1.06; Bray-Curtis: 0.790.82; Jaccard: 0.79-0.80), whereas at the genus level both alpha and beta diversity exhibited wider ranges and larger dispersion (Chao1: 45.27±5.68-50.20±5.64; Shannon: 2.54±0.24-2.73±0.22; InvSimpson: 10.37±1.4-11.03±1.05; Bray-Curtis: 0.79-0.82; Jaccard: 0.79-0.80). Furthermore, ASVs and OTUs were comparable across metrics; however, ASVs showed numerically higher values for some genus-level measures than OTUs due to the fact that they can resolve variation down to the single-nucleotide level, thereby retaining low-abundance features important for LMF safety. This work paves the way toward using bioinformatics and the 16S pipeline to characterize sparse, low-density, and uneven samples in low-moisture environments. Building on this, Chapter 4 validated the DADA2_SILVA (ASV) workflow for profiling bacterial communities across different wheat milling stages: before milling (BM), after milling (AM), and the milled fraction (MF). 22 families and 132 genera were identified. Genus-level richness (alpha diversity) decreased by 41.5% as a result of the BM to AM transition. Similarly, beta-diversity analysis indicated significant differences in community composition across milling stages, with MF consistently forming a compositionally distinct cluster. Furthermore, PICRUSt2-based KEGG module predicted 20 KEGG pathways, enriched for stress- and biofilm-related pathways, indicating persistence. Overall, our findings reveal that wheat milling drives stage-dependent restructuring and selective redistribution of microbial communities with direct implications for food safety, contamination control, and risk management. To address Salmonella contamination in wheat, Chapter 5 evaluated the effect of sodium bisulfate (SBS) tempering (17% moisture) on 4 different Salmonella enterica serovars. SBS tempering (1.5% SBS, w/v) reduced the Salmonella load by 4.3 log CFU/g after 24 h of tempering. Flour functionality and baking characteristics were comparable to those of the control (p>0.05), indicating that SBS is a viable antibacterial tempering agent. Chapter 6 employed a hurdle approach combining acidic water tempering with heat treatment (55 °C). Tempering alone with SBS, lactic acid (LA), and citric acid (CA) at 15% w/v reduced Salmonella load by 3.15, 3.23, and 2.91 log CFU/g, respectively (p<0.001). Heating alone reduced Salmonella load by 4.1 lg CFU/g after 24 h. Combining both tempering and heat treatment, however, resulted in greater reduction as non-detectable levels (<2 log CFU/g) of Salmonella in wheat were obtained after 12 h of tempering with LA (15%) + heat, while CA (15%) + heat and SBS (15%) + heat after 18 h of tempering. Flour functionality and baking characteristics remained comparable to the control across all treatments (p>0.05), supporting commercial viability. The final chapter 7 evaluated the antimicrobial potential of Pulsed light (PL) following a preliminary screening in which PL was the only non-thermal technology among cold plasma, ultraviolet, and PL to achieve ≥3 log CFU/g reduction under the evaluated conditions. PL at 0.329 J/cm2, achieved 4.22 log CFU/g on xylose-lysine deoxycholate (XLD) selective agar medium alone and 3.91 log CFU/g on XLD with a tryptic soy agar overlay (XLD+T), indicating a 7.2% sublethal injury associated with permeabilization after 18 h of tempering (p<0.0001). The Weibull model best described the non-first-order inactivation kinetics, predicting a 5-log reduction at 0.41-0.48 J/cm2 (99-118 s), consistent with the experimental data. Bacterial lysis was induced by oxidative stress, which damaged membranes, led to leakage, and compromised DNA integrity. PL at 0.329 J/cm² did not affect particle size or dough rheology (p>0.05); however, a significant difference in damaged starch content was observed (p<0.05), affecting baking and C-cell characteristics. Together, this work establishes a validated bioinformatic workflow for microbiome analysis in LMF environments, such as the wheat milling system, revealing a stage-dependent restructuring of the wheat microbiome. The study also demonstrated that acidic tempering, hurdle approaches, and non-thermal interventions can reduce wheat Salmonella load during tempering with minimal or no effect on wheat flour functionality and baking characteristics, providing a science-based foundation for improving food safety in commercial wheat mill settings.