Development of a genetically programed vanillin-sensing bacterium for high-throughput screening of lignin-degrading enzyme libraries

Lignin can potentially be converted to valuable aromatics such as vanillin, by controlled enzymatic catalysis. While natural enzymes have great potential, their performance could be further improved using directed evolution approaches. In addition to other enzymes, several natural and engineered laccases and peroxidases have been studied for enzymatic conversion of lignin to vanillin [11, 14, 41]. In spite of vast research in this area, no single enzyme has been reported to convert actual lignin substrates to their monomeric phenolic subunits. Remarkably, there are very few reports of selecting engineered enzymes using genuine lignin substrates. This is partially due to unavailability of high-throughput screening tools to detect lignin degradation and also because lignins have extremely heterogeneous structures with versatile chemical linkages that are least likely to be cleaved by the action of a single enzyme [13, 42]. Considering the complexity of lignin structures, future research may be directed towards simultaneous evolution of multiple enzymes or a multi-enzyme pathway using high-throughput screening tools like the live cell vanillin sensor described here. In this respect, the multi-enzyme system described by Reiter et al. [15] is particularly applicable. It was possible to release a small amount of lignin monomers from complex lignin structures using a combination of C?-dehydrogenase, ?-etherase and glutathione lyase enzymes. Salvachua et al. [22] reported lignin depolymerization by fungal secretomes containing a high level of laccase and peroxidase enzymes.

The VSC biosensor described here will be a useful tool in selecting vanillin-synthesizing enzymes from both metagenomic and mutant libraries. Developing enzymes for the conversion of lignin to vanillin would be of particular interest as vanillin is the most important lignin degradation product due to its large-scale use in the food, flavour and cosmetic industries. Induction of the putative E. coli promoter used in our biosensor is vanillin specific and no vanillin analogue or non-specific toxic chemical (like acrylic acid) can induce the expression of the vGFP gene under the control of this promoter, which is particularly interesting considering the absence of a known vanillin metabolism pathway in native E. coli [43]. However, vanillin’s mode of antimicrobial activity may explain this ambiguity. This comes mainly from its ability to damage the plasma membrane of the microbial cells through interaction with the lipids or proteins, which cause subsequent loss of the ionic gradient across the membrane and inhibition of bacterial respiration [35, 44]. A study using propidium iodide staining suggests that a significant proportion of E. coli cells remain alive even after treatment with 50 mM vanillin, although vanillin can completely arrest E. coli growth at a concentration of 15 mM, indicating that microbial growth inhibition by vanillin is bacteriostatic in nature rather than bactericidal [35]. This report also showed that E. coli can maintain partial potassium gradients after exposure to 50 mM vanillin for 40 min; vanillin treatment in this condition completely dissipates potassium ion gradients of Lactobacillus plantarum. Collectively, these observations suggest that the extent of E. coli membrane damage caused by vanillin is relatively less severe, and that when exposed to sublethal concentrations of vanillin, E. coli may cope with the stress by reestablishing ion gradients by alternative means, without vanillin metabolism. Although we cannot establish any functional group in the up-regulated genes identified by RNAseq analysis of vanillin-treated E. coli cells, the functions of the top seven up-regulated genes imply association with osmoprotection, metal ion transport and heavy metal toxicity (Additional file 1: Table S1). The up-regulated gene ydcI encodes a putative LysR-type DNA-binding transcriptional regulator. The exact function of ydcI protein is not known yet but other members of LysR-type transcriptional regulators are involved in the expression of various unrelated proteins including sodium–hydrogen antiporter and proteins involved in zinc homeostasis and oxidative stress defence [45, 46]. The proteins encoded by the other two up-regulated genes yeiW and sodC also play some role in metal ion detoxification and oxidative stress defence. The fourth highest up-regulated gene proA encodes a subunit of glutamate-5-semialdehyde dehydrogenase and gamma-glutamyl kinase-GP-reductase multi-enzyme complex that catalyses the first step in the synthesis of the osmoprotective amino acid proline [47, 48].

The VSC biosensor responds in a dose-dependent manner and upon induction with 5.0 mM vanillin fluorescence of the sensor is increased more than 4-fold but further increase of signal is not possible as the E. coli cannot grow at higher vanillin concentrations. Detectability within a relatively narrow range of vanillin concentration may be a limitation in selecting for mutants that can produce very low (0.5 mM) or very high (5 mM) concentrations of vanillin. Transposing the genetic sensing construct into a vanillin-tolerant microorganism could potentially address toxicity issues. In this respect, the top seven up-regulated genes upon vanillin exposure of E. coli cells are conserved within members of the Enterobacteriaceae family including several strains from the genus Escherichia, Shigella, Salmonella and Enterobacter. The large molecular weight of lignin precludes ready uptake into microbial cells. The VSC biosensor will likely find optimal use in screening extracellular enzymes that convert lignin to vanillin (Fig. 1). Selection experiments could be carried out on lignin-containing agar plates or using emulsion encapsulation methodology [49]. Additionally, this system could find use in screening enzymes that produce vanillin from smaller cell-permeable precursors. Examples include vanillyl alcohol oxidase and carboxylic acid reductase that produce vanillin from precursors like creosol, vanillylamine and vanillic acid [50, 51]. Versatility of live cell biosensors is also restricted by relatively short cellular lifespans, requirement of specific conditions for growth and survival of microbial systems and restriction of using engineered or live microorganisms in various end products. However, these issues will not create any major challenge in using this biosensor as a host cell for screening lignin-degrading enzyme libraries.

Inducible regulator-based whole-cell sensing systems are established as valuable analytical tools, which allow highly specific detection of target chemicals using fluorescent or bioluminescent reporters [32, 52–56]. While most of these biosensors were developed for environmental microbiology or bioremediation applications, inducible promoter-based sensing systems could be extremely useful in metabolic engineering applications including high-throughput screening of engineered enzyme libraries. Despite their vast potential, only a handful of small-molecule-inducible regulators (e.g. LacI) are well characterized and repeatedly used for a diverse range of applications. Development of additional inducible regulators can potentially provide newer biotechnology tools with innovative applications. The product inhibition approach used here could be potentially applied to identify putative promoter regions and develop biosensors for any chemical that exhibits partial toxicity to a microorganism with known genome sequence. Using a similar approach, Rogers et al. [32] developed four genetically encoded biosensors that respond to acrylate, glucarate, erythromycin and naringenin, and demonstrated the usage of glucarate biosensor for selecting superior enzyme variants from the glucarate biosynthesis pathway.