Improving 5-(hydroxymethyl)furfural (HMF) tolerance of Pseudomonas taiwanensis VLB120 by automated adaptive laboratory evolution (ALE)

The aldehyde 5-(hydroxymethyl)furfural (HMF) is of great importance for a circular bioeconomy. It is a renewable platform chemical that can be converted into a range of useful compounds to replace petroleum-based products such as the green plastic monomer 2,5-furandicarboxylic acid (FDCA). However, it also exhibits microbial toxicity for example hindering the efficient biotechnological valorization of lignocellulosic hydrolysates. Thus, there is an urgent need for tolerance-improved organisms applicable to whole-cell biocatalysis. Here, we engineer an oxidation-deficient derivative of the naturally robust and emerging biotechnological workhorse P. taiwanensis VLB120 by robotics-assisted adaptive laboratory evolution (ALE). The deletion of HMF-oxidizing enzymes enabled for the first time evolution under constant selection pressure by the aldehyde, yielding strains with consistently improved growth characteristics in presence of the toxicant. Genome sequencing of evolved clones revealed loss-of function mutations in the LysR-type transcriptional regulator-encoding mexT preventing expression of the associated efflux pump mexEF-oprN. This knowledge allowed reverse engineering of strains with enhanced aldehyde tolerance, even in a background of active or overexpressed HMF oxidation machinery, demonstrating a synergistic effect of two distinct tolerance mechanisms.


Figure S1 :
Figure S1: Evaluation of optimal culture conditions for an ALE with GRC1 ROX.All experiments were carried out in a 48-well FlowerPlate in a BioLector using MSM supplemented with 80 mM glycerol, 2 mM glucose, and varying concentrations of HMF (black: no HMF, red: 4 mM, green: 5 mM, orange: 6.5 mM, purple: 7.5 mM).Three buffer concentrations were tested: standard ((A), (D), (G)); two-fold ((B), (E), (H)); and four-fold ((C), (F), (I)).(A)-(C) Cell growth monitored by scattered light intensities.(D)-(F) pH values of the respective cultures.Graphs result from a second-order smoothing to the mean values obtained from three replicates.The dots represent the standard deviation.(G)-(I) HPLC analysis of final (55 h) HMF concentrations.The mean and standard deviation of three replicates is shown.

Figure S2 :
Figure S2: Analysis of all isolated clones from the ALE.Two-fold buffered MSM supplemented with 40 mM glycerol and 2 mM glucose as carbon sources was inoculated with the evolved strains (see legend for color coding), GRC1 ROX (black), and GRC1 (red) to an OD600 of 0.1.Cells were cultivated in a Growth Profiler in 96-well microtiter plates.The growth curves result from a second-order smoothing to the mean values obtained from three replicates.The dots represent the standard deviation.(A) No HMF added.(B) 8 mM HMF added.Because the preculture of clone C7.2 did not grow the respective isolate was not tested.

Figure S3 :
Figure S3: MexT deletion also confers a fitness advantage in strains with intact aldehyde oxidation machinery including the oxidation-optimized BOX strains (control experiments).Two-fold buffered MSM supplemented with 40 mM glycerol and 2 mM glucose as carbon sources in absence of HMF was inoculated with GRC1 or BOX derivatives (red) and the respective mexT deletion mutant (blue) to an OD600 of 0.1.Cells were cultivated in a Growth Profiler in 96-well microtiter plates.The growth curves result from a second-order smoothing to the mean values obtained from three replicates.The dots represent the standard deviation.

Figure S4 :
Figure S4: HMF conversion assays in 24-deepwell microplates (two-fold buffered MSM with 40 mM glycerol, 2 mM glucose, and 10 mM HMF) using whole-cells of GRC1 and derived BOX strains (increased expression of paoEFG and aldB-I) with deletion of mexT.For the determination of initial HMF depletion and HMFA formation rates a linear fit covering the first 2 h of each experiment was performed (shown as solid line).In cases HMF was fully converted faster this period was shortened accordingly.The OD600 was considered constant during that time and equaled the starting conditions.For each measurement, error bars represent the mean ± standard deviation of triplicates.(A) GRC1.(B) BOX-C2P2.(C) GRC1 ∆mexT.(D) BOX-C1 ∆mexT.(E) BOX-C2 ∆mexT.(F) BOX-P2 ∆mexT.(G) BOX-C1P2 ∆mexT.(H) BOX-C2P2 ∆mexT.(I) Initial HMF depletion and HMFA formation rates of all shown experiments.The error bars correspond to the standard error of the slope of the linear regression.MexT deletion does not alter HMF-oxidation properties of examined strains.

Figure S5 :
Figure S5: Genomic context of mexEF-oprN in P. taiwanensis VLB120 highlighting the nod-box-like sequence containing two putative MexT binding sites upstream of the operon.Annotations were made based on alignment with a previously determined consensus sequence ATCA(N)7CGAT identified in P. aeruginosa(Kim et al., 2019).

Table S3 : Oligonucleotides (name, sequence, and description) used as PCR primers for cloning procedures. Lower
case letters indicate overhangs, letters representing the binding sequence are capitalized, customly added spacers are italicized, and restriction sites are underlined.Oligonucleotides used for diagnostic PCRs and sequencing reactions are not included.

Table S4 : Overview of the growth parameters of all strains under the tested conditions in this work.
The end of the lag phase was estimated based on the first increase in g-value by more than 5% between two measurements.