Internal transcribed spacer (ITS) sequence-based characterization of fungal isolates from multiple yogurt facilities-A case study.

Fungal spoilage remains a significant issue in dairy product quality, especially for cultured dairy products such as yogurt formulated without preservatives such as potassium sorbate. Fungal contamination can occur throughout the processing continuum, from the dairy farm environment to the finished product processing environment. As molecular characterization of fungal isolates is used more frequently, we obtained fungal isolates obtained in 2 yogurt processing facilities as part of routine fungal testing of raw materials (e.g., fruit preparations, added ingredients), in-process product samples, environmental samples (e.g., air plates, equipment surfaces such as valves, face plates, air nozzles), and finished product samples, to determine whether internal transcribed spacer (ITS) barcoding data would be helpful to support source tracking of fungal contamination issues. Internal transcribed spacer PCR amplification and sequencing allowed us to classify the 852 isolates from these 2 facilities into 200 unique ITS allelic types (AT), representing the phyla Ascomycota (743 isolates), Basidiomycota (97 isolates), and Mucoromycota (12 isolates). Thirty ITS AT were isolated from both facilities; 62 and 108 ITS AT were isolated from only facility A or only facility B, respectively. Nine ITS AT were each represented by more than 20 isolates; these AT comprised 53% of the 852 isolates. The considerable diversity of fungal isolates even within a single facility illustrates the challenge associated with controlling fungal contamination of dairy products. The ITS barcoding technique, however, did show promise for facilitating the source tracking of fungal contamination, particularly for ITS AT over-represented in a given facility. For example, we found evidence for equipment-specific reservoirs for 2 AT (14 and 219) in facility B. Our data suggest that despite its limited discriminatory power, ITS sequencing can provide initial information that can help trace fungal contamination along the processing continuum. However, development and implementation of discriminatory subtyping methods will be needed to further improve the ability to identify sources of fungal contamination in dairy facilities. Developing and implementing sampling plans that comprehensively capture yeast and mold diversity in a given processing facility remain a considerable challenge.

Evaluation of biopreservatives in Greek yogurt to inhibit yeast and mold spoilage and development of a yogurt spoilage predictive model.

Dairy products, including cultured dairy products such as cheese and yogurt, are susceptible to fungal spoilage. Traditionally, additives such as potassium sorbate have been used to control fungal spoilage; however, with consumer demand for clean-label products, other strategies to control fungal spoilage (e.g., biopreservatives) are increasingly being used in dairy formulations. In order to help the dairy industry better evaluate biopreservatives for control of fungal spoilage, we developed a challenge study protocol, which was applied to evaluate 2 commercially available protective cultures for their ability to control yeast and mold spoilage of Greek yogurt. Greek yogurt formulated with and without protective cultures was inoculated with a cocktail consisting of 5 yeasts and 1 mold to yield inoculum levels of 101 and 103 cfu/g of yogurt. The inoculated yogurts were stored at 7°C and fungal counts as well as time to visible growth, on the yogurt surface, of mycelium mold colonies or yeast were determined over shelf-life. Whereas fungal concentrations increased to spoilage levels (≥105 cfu/g) in all yogurt formulations at both inoculum levels by d 23 of storage at 7°C, no surface mold was observed over 76 d in any of the products formulated with protective cultures. Control yogurts without biopreservatives all showed surface mold by d 23. In order to allow industry to better evaluate the business effects of improved control of surface mold growth that can be achieved with protective cultures, we developed a Monte Carlo simulation model to estimate consumer exposure to visible mold growth in yogurt formulated without fungal inhibitors. Our model showed that initial mold contamination rate has the largest effect on the model outcome, indicating that accurate data on contamination rates are important for use of these models. When air plates were used, in a proof-of-concept approach, to estimate initial contamination rates in a small yogurt manufacturing operation, our model predicted that 550 ± 25.2 consumers (±standard deviation) would be exposed to visible mold growth for every 1 million cups of yogurt produced. With initial contamination rate data for individual facilities, this model could be used by industry to estimate the number of consumers exposed to visible mold spoilage and could allow industry to better assess the value of mold-control strategies.

Psychrotolerant spore-former growth characterization for the development of a dairy spoilage predictive model.

Psychrotolerant spore-forming bacteria represent a major challenge regarding microbial spoilage of fluid milk. These organisms can survive most conventional pasteurization regimens and subsequently germinate and grow to spoilage levels during refrigerated storage. To improve predictions of fluid milk shelf life and assess different approaches to control psychrotolerant spore-forming bacteria in the fluid milk production and processing continuum, we developed a predictive model of spoilage of fluid milk due to germination and growth of psychrotolerant spore-forming bacteria. We characterized 14 psychrotolerant spore-formers, representing the most common Bacillales subtypes isolated from raw and pasteurized milk, for ability to germinate from spores and grow in skim milk broth at 6°C. Complete growth curves were obtained by determining total bacterial count and spore count every 24 h for 30 d. Based on growth curves at 6°C, probability distributions of initial spore counts in bulk tank raw milk, and subtype frequency in bulk tank raw milk, a Monte Carlo simulation model was created to predict spoilage patterns in high temperature, short time-pasteurized fluid milk. Monte Carlo simulations predicted that 66% of half-gallons (1,900 mL) of high temperature, short time fluid milk would reach a cell density greater than 20,000 cfu/mL after 21 d of storage at 6°C, consistent with current spoilage patterns observed in commercial products. Our model also predicted that an intervention that reduces initial spore loads by 2.2 Log10 most probable number/mL (e.g., microfiltration) can extend fluid milk shelf life by 4 d (end of shelf life was defined here as the first day when the mean total bacterial count exceeded 20,000 cfu/mL). This study not only provides a baseline understanding of the growth rates of psychrotolerant spore-formers in fluid milk, it also provides a stochastic model of spoilage by these organisms over the shelf life of fluid milk, which will ultimately allow for the assessment of different approaches to reduce fluid milk spoilage.

Internal transcribed spacer (ITS) sequencing reveals considerable fungal diversity in dairy products.

Fungi are important spoilage organisms in dairy products. However, little is known about the diversity of naturally occurring spoilage fungi in raw milk and processed dairy products, due at least in part to the fact that classical fungal identification methods require considerable expertise. To gain further insight into the fungal diversity in the dairy system, we isolated fungi from raw milk, raw and pasteurized milk cheese, and yogurt using the selective dichloran rose bengal chloramphenicol agar. In total, 361 fungal isolates were obtained and further characterized by DNA sequencing of the internal transcribed spacer (ITS) region and the nuclear ribosomal large subunit (LSU) rRNA gene if needed. We conducted BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi) searches of the ITS region sequences against the UNITE Database (https://unite.ut.ee/analysis.php), and selected other databases if needed, which allowed identification to the species level of 183 isolates and to the genus level of 107 of the 346 isolates that were successfully ITS sequenced. The isolates characterized represented 3 phyla and 19 genera; the most common genera isolated were Penicillium (25% of isolates), Debaryomyces (18%), and Candida (9%). This study not only provides, by using modern molecular tools, a baseline understanding of the types of fungi in dairy products, but also confirms that ITS sequencing is a useful approach for identification of fungal organisms found in the dairy food chain.