Design, construction and optimization of formaldehyde growth biosensors with broad application in biotechnology

Abstract

Formaldehyde is a key metabolite in natural and synthetic one-carbon metabolism. To facilitate the engineering of formaldehyde-producing enzymes, the development of sensitive, user-friendly, and cost-effective detection methods is required. In this study, we engineered Escherichia coli to serve as a cellular biosensor capable of detecting a broad range of formaldehyde concentrations. Using both natural and promiscuous formaldehyde assimilation enzymes, we designed three distinct E. coli growth biosensor strains that depend on formaldehyde for cell growth. These strains were engineered to be auxotrophic for one or several essential metabolites that could be produced through formaldehyde assimilation. The respective assimilating enzyme was expressed from the genome to compensate the auxotrophy in the presence of formaldehyde. We first predicted the formaldehyde dependency of the biosensors by flux balance analysis and then analysed it experimentally. Subsequent to strain engineering, we enhanced the formaldehyde sensitivity of two biosensors either through adaptive laboratory evolution or modifications at metabolic branch points. The final set of biosensors demonstrated the ability to detect formaldehyde concentrations ranging approximately from 30 μM to 13 mM. We demonstrated the application of the biosensors by assaying the in vivo activity of different methanol dehydrogenases in the most sensitive strain. The fully genomic nature of the biosensors allows them to be deployed as "plug-and-play" devices for high-throughput screenings of extensive enzyme libraries. The formaldehyde growth biosensors developed in this study hold significant promise for advancing the field of enzyme engineering, thereby supporting the establishment of a sustainable one-carbon bioeconomy.

FIGURE 1

Formaldehyde detection via three different E. coli growth biosensors. Formaldehyde (circled in green) is an important metabolite and a widely used commodity chemical. For its detection, three E. coli biosensors depending on formaldehyde assimilation for cell growth were designed. Each strain is based on a different set of assimilating enzymes. (I) The RuMP biosensor is based on the RuMP cycle enzymes 3‐hexulose‐6‐phosphate synthase (HPS) and 6‐phospho‐3‐hexuloisomerase (PHI) to convert formaldehyde and ribulose 5‐phosphate (Ru5P) into the essential metabolites fructose‐6‐phosphate (F6P) and glucose 6‐phosphate (G6P). (II) The LtaE biosensor utilizes the promiscuous serine aldolase activity of low‐specificity L‐threonine aldolase (LtaE) to condense formaldehyde and glycine into the essential amino acid serine. (III) The HOB biosensor is based on the promiscuous 4‐hydroxy‐2‐oxobutanoate (HOB) aldolase (HAL) reaction of the E. coli enzyme 2‐keto‐3‐deoxy‐L‐rhamnonate aldolase (encoded by rhmA) followed by a HOB transaminase (HAT) reaction which assimilate formaldehyde and pyruvate into homoserine. Homoserine is the precursor of the essential amino acids threonine, isoleucine and methionine. The formaldehyde dependency of each strain (mmol/gCDW) was calculated by flux balance analysis based on the E. coli genome‐scale metabolic model and is indicated in the bottom of the figure.

FIGURE 2

The RuMP biosensor detects formaldehyde in the millimolar range. (A) Metabolic scheme of the RuMP biosensor. The strain uses the RuMP cycle enzymes HPS and PHI for the assimilation of formaldehyde and ribulose 5‐phosphate into fructose 6‐phosphate (F6P) which is further converted to glucose 6‐phosphate (G6P). Deletions of native enzymes generating F6P and G6P as well as of enzymes draining the intermediates were performed to generate the auxotrophy. hps and phi from Bacillus methanolicus were expressed from the genome with either expression regulated by a medium (rbsC) or strong ribosome binding site (rbsA). Additionally, SoxA was expressed from a plasmid to allow sarcosine conversion into formaldehyde as a stand‐in for extracellular formaldehyde supplementation. (B) The RuMP biosensors with rbsC (RuMP_C) or rbsA (RuMP_A) were cultivated in minimal containing all relevant carbon sources with and without the addition of 5 mM sarcosine. Only when sarcosine was added to the medium, the strains grew, indicating their dependency on formaldehyde. (C) To characterize the RuMP biosensor in more detail, the RuMP_A strain was cultivated with a sarcosine gradient ranging from 0 to 20 mM. The gradient revealed a correlation of growth and sarcosine concentration between 1.1 and 13.3 mM. DHAP, dihydroxyacetone phosphate; E4P, erythrose 4‐phosphate; FBP, fructose 1,6‐bisphosphate; GAP, glyceraldehyde 3‐phosphate; R5P, ribose 5‐phosphate; S7P, sedoheptulose 7‐phosphate; X5P, xylulose 5‐phosphate; fbp, fructose 1,6‐bisphosphatase 1; glpX, fructose 1,6‐bisphosphatase 2; tktAB, transketolase 1 and 2; frmRAB, formaldehyde detoxification system; zwf, glucose 6‐phosphate dehydrogenase; SoxA, sarcosine oxidase; HPS, 3‐hexulose‐6‐phosphate synthase; PHI, 6‐phospho‐3‐hexuloisomerase. Figure elements: Yellow background: Auxotrophy; blue circle: Substrate; red cross: Deletion causing auxotrophy; cyan cross: Non‐essential deletion; green arrow: Formaldehyde assimilation flux; dashed arrow: Multi enzymatic reaction.

FIGURE 3

The LtaE biosensor detects formaldehyde in the milli‐ to micromolar range. (A) Metabolic scheme of the LtaE biosensor. The strain is based on the promiscuous activity of the LtaE enzyme that condenses formaldehyde and glycine into serine. Therefore, a serine auxotrophy was established by deleting essential genes for serine biosynthesis. To assimilate formaldehyde and glycine into serine, ltaE was expressed from the genome either with a medium or strong ribosome binding site (rbsC or rbsA, respectively). To generate intracellular formaldehyde from sarcosine, SoxA was expressed from a plasmid. (B) The LtaE biosensors with either rbsC or rbsA (LtaE_C and LtaE_A) were cultivated in a minimal medium containing all relevant carbon sources with and without the addition of 1 mM sarcosine. The strains only grew in the presence of sarcosine, indicating their dependency on formaldehyde. (C) An amino acid exchange (LtaE_C188Y) observed in an ALE experiment was introduced into the LtaE_A biosensor, creating the LtaE*_A biosensor. A comparison of the LtaE_A and the LtaE*_A biosensor showed that the LtaE*_A showed a 2‐fold faster growth when cultivated with 0.8 mM sarcosine. (D) For detailed characterization of the formaldehyde sensitivity, the LtaE*_A biosensor was cultivated in a gradient of sarcosine concentrations ranging from 0 to 6 mM. The gradient revealed a correlation between growth and sarcosine concentration between 0.2 and 6 mM. GlyA, serine hydroxymethyl transferase; serA, phosphoglycerate dehydrogenase; frmRAB, formaldehyde detoxification system; SoxA, sarcosine oxidase; LtaE, threonine aldolase; GCV, glycine cleavage system. Figure elements: Yellow background: Auxotrophy; blue circle: Substrate; red cross: Deletion causing auxotrophy; cyan cross: Non‐essential deletion; green arrow: Formaldehyde assimilation flux; dashed arrow: Multi enzymatic reaction.

FIGURE 4

The HOB biosensor detects formaldehyde in the low micromolar range. (A) Metabolic scheme of the HOB biosensor. The strain utilizes HAL and HAT activities to convert formaldehyde and pyruvate into homoserine. In the HOB biosensor, the native homoserine biosynthesis pathway is knocked out making it auxotrophic for threonine and methionine. These auxotrophies can be relieved by the assimilation of intracellular formaldehyde via HAL and HAT. To allow formaldehyde assimilation, the HAL gene was expressed from the genome using a strong constitutive promoter and either a medium or strong ribosome binding site (rbsC or rbsA, respectively). SoxA was expressed from a plasmid to allow sarcosine oxidation to formaldehyde. (B) The HOB_C and HOB_A biosensors were cultivated in a minimal medium containing all relevant carbon sources with and without the addition of 2 mM sarcosine. (C) As the HOB_A biosensor was shown to be more sensitive, it was used for detailed characterization of its formaldehyde sensitivity. The strain was cultivated with a gradient of sarcosine concentrations and showed growth correlation in a range from 125 to 2000 μM. (D) When 1 mM threonine was supplemented to reduce the formaldehyde dependency of the strain, the HOB_A biosensor was capable of growing on sarcosine concentrations as low as 32 μM. Asd, aspartate semialdehyde dehydrogenase; frmRAB, formaldehyde detoxification system; SoxA, sarcosine oxidase; HAL, HOB aldolase; HAT, HOB aminotransferase. Figure elements: Yellow background: Auxotrophy; blue circle: Substrate; red cross: Deletion causing auxotrophy; cyan cross: Non‐essential deletion; green arrow: Formaldehyde assimilation flux; dashed arrow: Multi enzymatic reaction.

FIGURE 5

Comparison of formaldehyde sensitivities reveals the operational range of the biosensors. (A) To create a sensitivity plot that allows a direct comparison of all formaldehyde biosensors created in this work, the mean maximum OD₆₀₀ was plotted against sarcosine concentrations. (B) Together, the HOB_A biosensor (with the addition of threonine), the LtaE*_A biosensor, and the RuMP_A biosensor present an operational range from 0.03 to 10 mM, spanning almost 3 orders of magnitude. For each concentration of sarcosine, the average maximum OD₆₀₀ of three independent growth experiments is shown and standard deviations are represented by error bars.

FIGURE 6

Assessment of in vivo activities of different methanol dehydrogenases in the HOB biosensor. (A) MDH expression in the HOB_A biosensor allows methanol oxidation to formaldehyde and its subsequent assimilation into homoserine. Different biotechnologically relevant MDHs were expressed from a plasmid in the HOB_A biosensor. (B–D) The resulting strains were cultivated in minimal media containing all relevant carbon sources and 10 to 80 mM methanol. (B) With 20 mM methanol, the difference between the MDH variants became most obvious. (C) While all MDH variants rescued growth at 80 mM methanol, only BsMDH was capable of rescuing growth at 10 mM methanol. Furthermore, BsMDH resulted in the highest max. OD₆₀₀ with all tested methanol concentrations. (D) The doubling time (DT) of the different strains increased with lower methanol concentration. At 20 mM methanol, the difference in DT is most pronounced. With 10 mM methanol, only the DT for BsMDH is shown as it is the only enzyme that enabled growth at this concentration. Asd, aspartate semialdehyde dehydrogenase; frmRAB, formaldehyde detoxification system; BsMDH, Bacillus stearothermophilus MDH; BmMDH, engineered Bacillus methanolicus MDH; CgMDH, Corynebacterium glutamicum MDH; CnMDH, engineered Cupriavidus necator N‐1 MDH; HAL, HOB aldolase; HAT, HOB aminotransferase. Figure elements: Yellow circle, auxotrophy; blue circle, substrate; red cross, deletion causing auxotrophy; cyan cross, non‐essential deletion; dashed arrow, multi enzymatic reaction.