Long Range Order and Short Range Disorder in Saccharomyces cerevisiae Biofilm

Abstract

Biofilm, a colony forming cooperative response of microorganisms under environmental stress, is a major concern for food safety, water safety and drug resistance. Most current works focus on controlling biofilm growth by targeting single genes. Here, we investigated transcriptome-wide expressions of the biofilm yeast Saccharomyces cerevisiae in wildtype, and 6 previously identified biofilm regulating overexpression strains (DIG1, SAN1, TOS8, ROF1, SFL1, HEK2). When tested across various statistical distributions, all transcriptome-wide data fitted well with lognormal distribution above TPM value of 5. Using this threshold as a low expression filter, Pearson auto-and cross-correlation reveal a strong transcriptome-wide invariance among all genotypes, which is also reflected by the random selection of 50 gene expressions. Focusing on the 50 highly expressed genes, however, they differ significantly between the genotypes. Principal components analysis (PCA) shows global similarity between DIG1, SAN1, ROF1, SFL1 and HEK2. Thus, although single overexpression strains may show significant favourable local and acute expression changes (short range disorder), the almost unperturbed global and collective structure between the genotypes indicate gradual adaptive response converging to original stable biofilm states (long range order). Hierarchical clustering and Gene Ontology show 11 groups of local (e.g. mitochondria processes, amine & nucleotide metabolic processes) and 6 groups of global (e.g. transcription, translation & cell cycle) processes for all genotypes. These data indicate that there is a strong global regulatory structure that keeps the overall biofilm stable in all investigated strains.

Large‐scale‐free network organisation is likely key for biofilm phase transition

“…That is, out of over 11,000 transcript expressions, we found that the top 50–100 expressed genes were poorly correlated with the different strains. This result was confirmed by comparing the correlation with the random extraction of 50–100 genes from the whole transcriptome [27]. Furthermore, removing the top 50 or 500 expressed genes from the data still showed the correlation close to unity (Fig.…”

“…More recently, we investigated transcriptome‐wide expression data of S. cerevisiae biofilm, in wild type and six different single‐gene over‐expressed strain ( SAN 1, TOS 8, ROF 1, SFL 1, HEK 2, and DIG 1), related to biofilm regulation using various statistical methods [27]. These strains were previously studied due to their biofilm morphological regulation [28].…”

Connect and Conquer: Collectivized Behavior of Mitochondria and Bacteria

“…These advantageous properties, which ultimately might directly influence higher-order behavior [e.g., between bacteria and their hosts (Hughes and Sperandio, 2008; Carding et al, 2015)], outweigh the high energetic costs of biofilm formation (Lyons and Kolter, 2015) (see section “The Benefits of the Impermanent Interactions”). Collectivization in bacterial biofilms thus represents exemplary self-organization and circular causality, i.e., bacteria generate their own local microenvironments (e.g., nutrient and oxygen gradients) that elicits responses on different temporal scales (e.g., by differential gene expression) which subsequently modulates localized conditions thereby tuning their ensuing behavior within the population etc., (Klauck et al, 2018; Piras et al, 2018).…”