Microbial community structure of plant-based meat alternatives

Abstract

A reduction in animal-based diets has driven market demand for alternative meat products, currently raising a new generation of plant-based meat alternatives (PBMAs). It remains unclear whether these substitutes are a short-lived trend or become established in the long term. Over the last few years, the trend of increasing sales and diversifying product range has continued, but publication activities in this field are currently limited mainly to market research and food technology topics. As their popularity increases, questions emerge about the safety and nutritional risks of these novel products. Even though all the examined products must be heated before consumption, consumers lack experience with this type of product and thus further research into product safety, is desirable. To consider these issues, we examined 32 PBMAs from Austrian supermarkets. Based on 16S rRNA gene amplicon sequencing, the majority of the products were dominated by lactic acid bacteria (either Leuconostoc or Latilactobacillus), and generally had low alpha diversity. Pseudomonadota (like Pseudomonas and Shewanella) dominated the other part of the products. In addition to LABs, a high diversity of different Bacillus, but also some Enterobacteriaceae and potentially pathogenic species were isolated with the culturing approach. We assume that especially the dominance of heterofermentative LABs has high relevance for the product stability and quality with the potential to increase shelf life of the products. The number of isolated Enterobacteriaceae and potential pathogens were low, but they still demonstrated that these products are suitable for their presence.

Fig. 1. Present isolates per sample, group or manufacturer are represented by dots.

The 16S rRNA gene sequences of isolates belonging to a particular genus were aligned to each other, and were clustered based on their sequence similarity. Consequently, different clusters represents different strains or species. The higher the number of clusters within a genus, the larger the plotted dot in the figure. The surrounding area is shaded according to the relative abundances in the 16S rRNA gene amplicon high-throughput sequencing (for the group and manufacturer summary, the mean relative abundance of each included sample is used). Additionally, genera, that were highly abundant in some samples, but had no corresponding isolates, were added to the figure denoted capitalized letters.

Fig. 2. Taxonomy plot based on amplicon sequencing, showing the relative abundances on genus level.

The underlying relative frequencies of the ASVs are recognizable as pale yellow lines within a genus. Genera with a maximum value of 10% across all samples were subsumed by color in the next higher taxonomic level. The samples are ordered by main protein source, texture and manufacturer. ASVs with matching isolates ( > 99% identity), were outlined in red.

Fig. 3. Differences in community profiles.

a tSNE plot clustered samples to different profiles based on Bray-Curtis dissimilarity (Similar clusters were also found with other distance matrices - see Supplementary Fig. 2). b Hill-Shannon diversity and c: Hill-Simpson diversity comparing these profiles. Box plots show the median with hinges that correspond to the 25th and 75th percentiles. The whiskers extend from the hinge to the largest and smallest value no further than 1.5 multiplied by the inter-quartile range. Individual data points are overlaid.

Fig. 4. Linear discriminant analysis Effect Size (LEfSe) per group identified several predictive characteristic genomic features for product categories with a log₁₀ LDA score > 4.0.

Additionally illustrated are the medians of the relative abundances as a point, with lines representing the first and third quartiles. Additionally, LEfSe per community profiles are displayed in Supplementary Fig. 3.