Inhibitor tolerance and bioethanol fermentability of levoglucosan-utilizing Escherichia coli were enhanced by overexpression of stress-responsive gene ycfR: The proteomics-guided metabolic engineering

Abstract

Pretreatment of lignocellulosic biomass is crucial for the release of biofermentable sugars for biofuels production, which could greatly alleviate the burgeoning environment and energy crisis caused by the massive usage of traditional fossil fuels. Pyrolysis is a cost-saving pretreatment process that can readily decompose biomass into levoglucosan, a promising anhydrosugar; however, many undesired toxic compounds inhibitory to downstream microbial fermentation are also generated during the pyrolysis, immensely impeding the bioconversion of levoglucosan-containing pyrolysate. Here, we took the first insight into the proteomic responses of a levoglucosan-utilizing and ethanol-producing Escherichia coli to three representative biomass-derived inhibitors, identifying large amounts of differentially expressed proteins (DEPs) that could guide the downstream metabolic engineering for the development of inhibitor-resistant strains. Fifteen up- and eight down-regulated DEPs were further identified as the biomarker stress-responsive proteins candidate for cellular tolerance to multiple inhibitors. Among these biomarker proteins, YcfR exhibiting the highest expression fold-change level was chosen as the target of overexpression to validate proteomics results and develop robust strains with enhanced inhibitor tolerance and fermentation performance. Finally, based on four plasmid-borne genes encoding the levoglucosan kinase, pyruvate decarboxylase, alcohol dehydrogenase, and protein YcfR, a new recombinant strain E. coli LGE-ycfR was successfully created, showing much higher acetic acid-, furfural-, and phenol-tolerance levels compared to the control without overexpression of ycfR. The specific growth rate, final cell density, ethanol concentration, ethanol productivity, and levoglucosan consumption rate of the recombinant were also remarkably improved. From the proteomics-guided metabolic engineering and phenotypic observations, we for the first time corroborated that YcfR is a stress-induced protein responsive to multiple biomass-derived inhibitors, and also developed an inhibitors-resistant strain that could produce bioethanol from levoglucosan in the presence of inhibitors of relatively high concentration. The newly developed E. coli LGE-ycfR strain that could eliminate the commonly-used costly detoxicification processes, is of great potential for the in situ cost-effective bioethanol production from the biomass-derived pyrolytic substrates.

Keywords: Bioethanol; Gene overexpression; Inhibitor; Levoglucosan; Lignocellulosic biomass; Proteomics.

Graphical abstract

Fig. 1

Schematic diagram of the sampling and analysis experiments adopted in this work.

Fig. 2

Inhibitory profiles of individual and combined inhibitors on the cell growth of Escherichia coli LGE2. Panel A1 Cells challenged by acetic acid in minimal media. Panel A2 Cells challenged by acetic acid in LB media. Panel B1 Cells challenged by furfural in minimal media. Panel B2 Cells challenged by furfural in LB media. Panel C1 Cells challenged by phenol in minimal media. Panel C2 Cells challenged by phenol in LB media. Panel D Cells challenged by combined inhibitors in minimal media; the labels CK, C1, C2, C3, C4, C5, and C6 at horizontal axis represent respective combination of 0%, 25%, 35%. 45%, 55%, 60%, and 65% of the individual MIC (minimum inhibitory concentration) of each inhibitor. Panel E Cells challenged by individual and combined inhibitors at their IC₅₀ levels for 48-h incubation in minimal media; A_IC₅₀, F_IC₅₀, P_IC₅₀, and C_IC₅₀ denote the concentrations of acetic acid, furfural, phenol, and combined inhibitors inhibiting 50% of the cell growth with 48-h incubation time, respectively; dark dotted line denotes the sampling point (OD₆₀₀ = 0.68 ± 0.05) chosen for further proteomics analysis.

Fig. 3

Numbers of the unique and shared differentially expressed proteins (DEPs) of all the inhibitor-treated groups. Panel A shows the numbers of unique and shared up-regulated proteins, and panel B shows the numbers of unique and shared down-regulated proteins. These protein numbers are shown in an UpSet diagram, allowing for a clearer plotting of large data sets compared to the Venn diagram. Dark circles (●) connected with a line indicate that the proteins are only differentially expressed in the corresponding group(s) labelled with ●, but not in the other groups labelled with ●. For example, the label ① denotes the proteins are only differentially expressed in acetic acid treated group, while not differentially expressed in the other groups, that is, these proteins are unique in acetic acid treated group; ② denotes the proteins are differentially expressed in all the groups, that is, these proteins are shared by all the groups. Abbreviations A, F, P, and C represent cells treated by acetic acid, furfural, phenol, and combined inhibitors of acetic acid, furfural, and phenol, respectively.

Fig. 4

Heat map (A) and documented interactions (B) of the DEPs shared by all the inhibitor-treated groups. Panel A Rows are colored by the log2 fold-changes (FC) of the proteins in inhibitor-treated groups relative to the corresponding proteins in control. Groups A, F, P, and C represent cells treated by acetic acid, furfural, phenol, and combined inhibitors, respectively. avg represents the average fold-change of the four groups. The darker the red color, the greater the fold-change of up-regulation. The darker the cyan color, the greater the fold-change of down-regulation. Panel B Interactions were obtained based on String 11.0 database. A greater number of lines associated with the connection, indicates a greater level of confidence in the association. The network nodes are gene names for the corresponding biomarker proteins. The edges represent the predicted functional associations. An edge may be drawn with up to 7 differently colored lines - these lines represent the existence of the seven types of evidence used in predicting the associations. A red line indicates the presence of fusion evidence; a green line - neighborhood evidence; a blue line - coocurrence evidence; a purple line - experimental evidence; a yellow line - textmining evidence; a light blue line - database evidence; a dark gray line - coexpression evidence. The black boxes connected with the nodes by black lines exhibited the annotated functional categories of the proteins; the others are unannotated proteins in the database.

Fig. 5

The time-course changes of cell growth, levoglucosan consumption, ethanol production, and inhibitors concentration in the fermentation tests. E. coli LGE-ycfR was grown under the IC₅₀ of individual and combined inhibitors identified for the non-overexpression strain E. coli LGE2 (defined as CK). IC₅₀ denotes the inhibitor concentration inhibiting 50% of the cell growth obtained by cultures in the absence of inhibitors. Panel A cells were grown in the presence of acetic acid; Panel B cells were grown in the presence of furfural; Panel C cells were grown in the presence of phenol; Panel D cells were grown in the presence of combined inhibitors; Panel E inhibitors' concentrations before and after the fermentation. Cell densities _YcfR, LG _YcfR, and EtOH _YcfR, denote the cell density, levoglucosan concentration, and ethanol concentration exhibited by E. coli LGE-ycfR, while Cell densities _CK, LG _CK, and EtOH _CK, denote those values exhibited by the non-overexpression strain E. coli LGE2.

Fig. 6

Overview of the molecular functions of the YcfR-interacting proteins in E. coli against acetic acid, furfural, and phenol stresses. Red (green) arrows/nodes denote YcfR has negative (positive) interactions with the node protein, which mean that separate gene mutations/deletions combined in a same cell would result in a more (less) severe fitness defect or lethality than each gene did alone. Genetic/protein interactions were considered significant if they had an S-score≧3.08809 for positive interactions and S-score ≦ -3.38787 for negative interactions. The descriptions in the colored frames are the protein functions.