Extracellular riboflavin induces anaerobic biofilm formation in Shewanella oneidensis

Abstract

Background: Some microorganisms can respire with extracellular electron acceptors using an extended electron transport chain to the cell surface. This process can be applied in bioelectrochemical systems in which the organisms produce an electrical current by respiring with an anode as electron acceptor. These organisms apply flavin molecules as cofactors to facilitate one-electron transfer catalyzed by the terminal reductases and in some cases as endogenous electron shuttles.

Results: In the model organism Shewanella oneidensis, riboflavin production and excretion trigger a specific biofilm formation response that is initiated at a specific threshold concentration, similar to canonical quorum-sensing molecules. Riboflavin-mediated messaging is based on the overexpression of the gene encoding the putrescine decarboxylase speC which leads to posttranscriptional overproduction of proteins involved in biofilm formation. Using a model of growth-dependent riboflavin production under batch and biofilm growth conditions, the number of cells necessary to produce the threshold concentration per time was deduced. Furthermore, our results indicate that specific retention of riboflavin in the biofilm matrix leads to localized concentrations, which by far exceed the necessary threshold value.

Conclusion: This study describes a new quorum-sensing mechanism in S. oneidensis. Biofilm formation of S. oneidensis is induced by low concentrations of riboflavin resulting in an upregulation of the ornithine-decarboxylase speC. The results can be applied for the development of strains catalyzing increased current densities in bioelectrochemical systems.

Keywords: Biofilm; Current density; Microbiology; Quorum sensing; Shewanella oneidensis.

Fig. 1

Impact of riboflavin and speC expression on current and biofilm formation on anode surfaces. Overexpression of speC resembles the phenotype after riboflavin addition, while deletion mutants seem to be blind for the riboflavin signal. The bar chart shows an increase in current density of 1.8-fold due to the addition of 37 nM riboflavin. Furthermore, the number of cells on the anode increases 2.4-fold. A very similar effect can be observed by the overexpression of speC. The addition of 37 nM riboflavin to a speC overexpressing strain leads to no increase in current density or in cell number. The deletion of speC leads to a slight decrease in current density, but the supply of 37 nM riboflavin does not show any significant effect on the speC knockout strain. Error bars represent the standard deviation from individual replicates (n = 3)

Fig. 2

Impact of srtA deletion on current generation and biofilm formation on anode surfaces. The bar chart shows that the knockout of srtA does not have any significant effect on the current density. Error bars represent the standard deviation from individual replicates (n = 3)

Fig. 3

Growth (A) and flavin secretion (B) of S. oneidensis wild type in a batch culture and results of the modeling attempt. Measured values are depicted as point squares or bars, while the modeling results are shown as solid lines. Error bars represent the standard deviation from 3 individual bacterial samples (n = 3)

Fig. 4

A speC expression after the addition of different concentrations of riboflavin relative to transcript abundance without riboflavin addition. The addition of up to 15 nM riboflavin does not have any significant effect on speC expression, while the addition of 18.5, 37 and 100 nM riboflavin leads to a 2- and 2.3-fold increase in speC expression, respectively. Error bars represent the standard deviation from individual replicates (n = 3). B speC expression after the addition of different concentrations of flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN) relative to cells without exogenous flavin addition. The addition of FAD and FMN in these concentrations does not have a significant effect on speC expression. Expression is normalized to the gene for the RNA polymerase subunit rpoA. Error bars represent the standard deviation from individual replicates (n = 3). Asterisks represent significant differences (unpaired t-test p < 0.05)

Fig. 5

Exemplary prediction of time needed to produce 18 nM riboflavin with different cell numbers. The delineation is based on the model described in the text and supplemental material

Fig. 6

Mean biofilm thickness, lactate and flavin concentration of S. oneidensis wild type grown in a PDMS chip over time. After 121 h there was no detectable amount of lactate in the effluent. The flavin content significantly increased over time. However, especially the riboflavin concentration is much higher within the biofilm compared to the effluent. Error bars represent the standard deviation from 3 individual bacterial samples (n = 3)