
==== Front
Appl Environ Microbiol
Appl Environ Microbiol
aem
Applied and Environmental Microbiology
0099-2240
1098-5336
American Society for Microbiology 1752 N St., N.W., Washington, DC

39194220
aem01472-24
10.1128/aem.01472-24
aem.01472-24
Enzymology and Protein Engineering
applied-and-industrial-microbiologyApplied and Industrial MicrobiologyA growth-based screening strategy for engineering the catalytic activity of an oxygen-sensitive formate dehydrogenase
Li Feilong 1 2
https://orcid.org/0000-0002-0667-9224
Scheller Silvan 1
https://orcid.org/0000-0001-8977-8887
Lienemann Michael 2 Michael.lienemann@vtt.fi

1 Department of Bioproducts and Biosystems, Aalto University , Espoo, Finland
2 VTT Technical Research Centre of Finland Ltd. , Espoo, Finland
Editor Lotti Marina University of Milano-Bicocca , Milan, Italy

Address correspondence to Michael Lienemann, Michael.lienemann@vtt.fi
The authors declare no conflict of interest.

9 2024
28 8 2024
28 8 2024
90 9 e01472-2425 7 2024
27 7 2024
Copyright © 2024 Li et al.
2024
Li et al.
https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license.

ABSTRACT

Enzyme engineering is a powerful tool for improving or altering the properties of biocatalysts for industrial, research, and therapeutic applications. Fast and accurate screening of variant libraries is often the bottleneck of enzyme engineering and may be overcome by growth-based screening strategies with simple processes to enable high throughput. The currently available growth-based screening strategies have been widely employed for enzymes but not yet for catalytically potent and oxygen-sensitive metalloenzymes. Here, we present a screening system that couples the activity of an oxygen-sensitive formate dehydrogenase to the growth of Escherichia coli. This system relies on the complementation of the E. coli formate hydrogenlyase (FHL) complex by Mo-dependent formate dehydrogenase H (EcFDH-H). Using an EcFDH-H-deficient strain, we demonstrate that growth inhibition by acidic glucose fermentation products can be alleviated by FHL complementation. This allows the identification of catalytically active EcFDH-H variants at a readily measurable cell density readout, reduced handling efforts, and a low risk of oxygen contamination. Furthermore, a good correlation between cell density and formate oxidation activity was established using EcFDH-H variants with variable catalytic activities. As proof of concept, the growth assay was employed to screen a library of 1,032 EcFDH-H variants and reduced the library size to 96 clones. During the subsequent colorimetric screening of these clones, the variant A12G exhibiting an 82.4% enhanced formate oxidation rate was identified. Since many metal-dependent formate dehydrogenases and hydrogenases form functional complexes resembling E. coli FHL, the demonstrated growth-based screening strategy may be adapted to components of such electron-transferring complexes.

IMPORTANCE

Oxygen-sensitive metalloenzymes are highly potent catalysts that allow the reduction of chemically inert substrates such as CO2 and N2 at ambient pressure and temperature and have, therefore, been considered for the sustainable production of biofuels and commodity chemicals such as ammonia, formic acid, and glycine. A proven method to optimize natural enzymes for such applications is enzyme engineering using high-throughput variant library screening. However, most screening methods are incompatible with the oxygen sensitivity of these metalloenzymes and thereby limit their relevance for the development of biosynthetic production processes. A microtiter plate-based assay was developed for the screening of metal-dependent formate dehydrogenase that links the activity of the tested enzyme variant to the growth of the anaerobically grown host cell. The presented work extends the application range of growth-based screening to metalloenzymes and is thereby expected to advance their adoption to biosynthesis applications.

KEYWORDS

enzyme engineering
growth-based screening
metalloenzyme
formate dehydrogenase
oxygen sensitivity
formate hydrogenlyase
Research Council of Finland (AKA) 321723 Lienemann Michael Research Council of Finland (AKA) 329510 Scheller Silvan cover-dateSeptember 2024
==== Body
pmcINTRODUCTION

Enzyme engineering has been instrumental in transforming wild-type enzymes into potent catalysts for the synthesis of commodity chemicals from highly concentrated non-natural substrates at elevated temperatures and in the presence of organic solvents (1). The methods employed for such enzyme customizations are either of a rational nature or introduce random mutations in the gene encoding the protein of interest. When the latter approach is combined with a selection step and iterated over several rounds, the process mimics natural selection and is, therefore, called directed evolution (2). The advantage of this method is that it requires little knowledge of the enzyme structure and its catalytic mechanism but depends on high-throughput screening (HTS) methods for identifying enzyme variants with improved properties (3). The development of these screening techniques has been the primary challenge in most enzyme engineering campaigns and is limited by the low amount of assay principles that are compatible with automated sample processing in the HTS format, i.e., being chiefly fluorometric and colorimetric (3, 4). A wider application of HTS methods for enzyme variant screening is hindered by the requirement for elaborate sample handling (e.g., induction of gene expression, cell fragmentation, and analysis of enzymatic activity) and expensive microfluidic devices or robotic platforms (5).

Growth-based screening is a powerful strategy for screening large variant libraries with easily measurable readout, high throughput, and minimal equipment dependence (2, 6, 7). The design of growth-based screening approaches is challenging because it requires a strong link between the fitness of a host organism and the mutations introduced into a gene of interest (6, 7). A widely employed growth-based screening method is based on auxotrophic host organisms where the introduction of catalytically active enzyme variants complements the disrupted metabolic functions and can be readily identified by measuring the growth of their host cells (6). This approach has proven useful for improving natural amino acid production [e.g., chorismate mutases and amino acid racemases (8, 9)], switching cofactor dependency [e.g., metal-free formate dehydrogenases (10) and NAD(P)H oxidases (11)], and increasing the activity for the synthesis of non-natural chiral amines [e.g., hydrolases (12, 13), amine transaminases, ammonia lyase, and monoamine oxidase (7)]. Noteworthy, other growth-based screening methods employ, e.g., the production of variants of detoxifying enzymes or the inclusion of fitness-enhancing reporter genes that are transcribed following the binding of a complex between a gene regulator and a molecule synthesized by the enzyme of interest (e.g., β-lactam hydrolases and benzaldehyde dehydrogenases) (14–16). An underrepresented group of biocatalysts in enzyme engineering campaigns are oxygen-sensitive enzymes, such as metal-dependent formate dehydrogenases, hydrogenases, and nitrogenases. These enzymes are of high interest to industrial applications owing to their bound soft, electron-rich metal ions allowing the transfer of electrons to chemically inert electron acceptor molecules such as CO2 and N2 (17). Relevant fields of active research include the hydrogen and formate economy as well as green nitrogen fixation and production of chemical feedstocks (18–20). The metal centers employed by these redox-active enzymes are easily inactivated by molecular oxygen. Therefore, screening of the coordinating enzymes can only be performed in formats that allow maintenance of strictly anoxic assay conditions and, at the same time, are robust and scalable. Growth-based screening methods fully comply with these requirements and are, therefore, ideally suited for the engineering of oxygen-sensitive metalloenzymes. So far, no screening strategy has been reported that links cell growth to the catalytic function of such enzymes and, thereby, prevents their engineering in a HTS format.

Here, we present a growth-based screening method for the identification of oxygen-sensitive variants of Escherichia coli formate dehydrogenase H (EcFDH-H) with increased catalytic activity. The screening system relies on the complementation of the membrane-bound formate hydrogenlyase (FHL) complex employing the EcFDH-H enzyme as an electron input module to couple intracellular formate oxidation to proton reduction (Fig. 1). The production of functional EcFDH-H variants alleviates growth inhibition by formate and proton accumulation during fermentative growth and enabled the screening of an enzyme variant library with cell density as a simple readout. This screening method significantly reduced the library size and was combined with a colorimetric assay in order to identify variants with improved catalytic activity.

Fig 1 Mixed-acid fermentation yielding succinate, ethanol, acetate, formate, and lactate. The shown reactions are catalyzed by phosphoenolpyruvate carboxylase (PEPC), pyruvate kinase (PYK), lactate dehydrogenase (LDH), pyruvate formate lyase (PFL), phosphotransacetylase (PTA), acetate kinase (ACK), acetate dehydrogenase (ADH), and acetaldehyde dehydrogenase (ALDH). Formate is disproportionated into CO2 and H+ by the formate hydrogenlyase complex. Here, two electrons are released at the formate dehydrogenase H subunit (FdhF, referred to as EcFDH-H in the main text), passed on to the hydrogenase 3 (HycE) subunit via an array of eight iron–sulfur clusters, and ultimately transferred onto two protons to yield H2. The membrane-bound glucose-specific phosphotransferase system (PtsG) and formate channel (FocA) facilitate the passage of glucose and formate, respectively, across the cytoplasmic membrane.

RESULTS AND DISCUSSION

Correlating FDH-H activity to fermentative metabolism of E. coli

Effect of EcFDH-H removal on cell viability

The presented screening strategy relies on E. coli growth during glucose fermentation as a measure of the activity of a formate dehydrogenase bound to HycB as an electron-donating component of the FHL complex (Fig. 1). In order to establish the importance of the functional FHL complex for fermentative growth, the EcFDH-H-producing and -deficient E. coli strains JG-X and FL004, respectively (Table 1), were cultivated under anaerobic atmosphere at varying glucose concentrations. The strain JG-X was included as a reference, owing to its ΔfdhF and ΔiscR genotype. The fdhF deletion eliminates the EcFDH-H background activity of the E. coli MC4100 wild-type strain while the production of the other two formate dehydrogenases EcFDH-O and EcFDH-N is maintained at natural levels. The iscR gene encodes an iron–sulfur cluster regulator that represses the transcription of the iscRSUA operon (21) and was deleted in the EcFDH-H production strains in order to increase the cofactor occupancy and production level of Fe-S cluster-containing proteins as reported for the production of 4-hydroxy-3-methylbut-2-enyl diphosphate reductase in E. coli (22, 23). As evident from Fig. 2A, strain JG-X reached a final cell density (OD600) of 1.25 ± 0.05 while strain FL004 did not exceed an OD600 value of 0.357 ± 0.006 revealing a strong growth retardation effect of the EcFDH-H deletion. The retarded fermentative growth of strain FL004 may be explained by the compromised function of its FHL complex, i.e., remediation of the cytosol acidification and induction of cellular stress resulting from the accumulation of weak organic acids during mixed acid fermentation (Fig. 1) (24, 25). Notably, EcFDH-H deletion did not result in growth retardation under aerobic conditions (Fig. S1) because these catabolic intermediates are not formed if oxygen is used as the exclusive terminal electron acceptor. Formate is among these fermentation products and can enter the cell via membrane-embedded transporters such as FocA (26). It also induces the synthesis of the membrane-bound FHL complex that catalyzes formate disproportionation into CO2 and H2 (27, 28). Possible formate cytotoxicity was investigated by the exposure of the E. coli strains FL004 and JG-X to varying concentrations of formate. Here, strain FL004 was found to be more sensitive to formate when compared to the EcFDH-H-producing control JG-X (Fig. S2A). This result is consistent with the faster growth of the parental strain JG-X during glucose fermentation (Fig. 2A) and primary formate degradation by the FHL complex (Fig. 1). The extent to which EcFDH-H contributes to deacidification during fermentative growth was assessed by monitoring the pH of anaerobic E. coli cultures and evolution of hydrogen gas. The culture broth pH of the EcFDH-H-deficient strain FL004 decreased from 7.0 to below 6.1 within 4 h while the strain JG-X maintained the culture medium pH above 6.1 and grew exponentially for up to 10 h (Fig. 2B). These results are in line with a previous study by Beyer et al. who reported retarded fermentative growth and similar pH profiles for a selC deletion strain of E. coli (29). In contrast, a more recent study by Gevorgyan et al. found E. coli fdhF deletion to decrease the final cell density during glucose fermentation by merely ~10% (30). We attribute this to the presence of 100-mM phosphate buffer in the growth media, which maintained the external media pH at ≥6.4 during their examined 72-h cultivation period. Therefore, the accelerated growth medium acidification of the EcFDH-H-deficient strain observed in the current study can be explained by the interruption of electron flow through the FHL complex to the HycE subunit catalyzing proton reduction to molecular hydrogen.

TABLE 1 Derivatives of Escherichia coli strain JG333 (MC4100 ΔfdhF, ΔiscR) used for the production of EcFDH-H and its variantsb

Strain	Plasmid-encoded FDH	Plasmids	VTT-CC ID	
JG-Xa	EcFDH-H	pSU21-selC, pTrc99a-fdhF	E-223614	
FL003	–c	pSU21-selC	E-233617	
FL004	–c	pSU21-selC, pTrc99a	E-233618	
FL005	EcFDH-H-A12G	pSU21-selC, pFL003	E-233619	
FL006	EcFDH-H-D179E	pSU21-selC, pFL004	E-233620	
FL007	EcFDH-H-D179L	pSU21-selC, pFL005	E-233621	
FL008	EcFDH-H-K44R	pSU21-selC, pFL006	E-233622	
FL009	EcFDH-H-K44A	pSU21-selC, pFL007	E-233623	
a Source: Prof. Shelley Minteer (Utah State University, Utah) (31).

b VTT Culture Collection strain identifiers (VTT-CC IDs) are indicated.

c –, neither the native nor the mutated fdhF gene is contained in the harbored plasmid(s).

Fig 2 Effects of FHL complementation by EcFDH-H on anaerobic growth of E. coli with glucose measured as (A) final cell density at varying glucose concentrations, (B) time-dependent change of cell density and culture medium pH at cGlucose = 4 g/L, and (C) gas production at cGlucose = 4 g/L. (D) Correlation of final host cell density reached after 24-h anaerobic fermentation of glucose and in vitro formate oxidation activity of EcFDH-H. The activities of EcFDH-H enzymes (wild type, D179E, and D179L: 1.5 µg; K44R and K44A: 7.5 µg) were determined spectroscopically at λ = 555 nm as BV2+ reduction using a phosphate-buffered reaction mix (pH 7.5) containing the substrates formate and BV2+ at a concentration of 10 and 2 mM, respectively. Error bars represent the standard deviation of three independent experiments.

Effect of EcFDH-H removal on intracellular proton reduction

FHL-dependent hydrogen production was investigated by the collection and analysis of the gas that was produced by E. coli strains JG-X and FL004 during 24 h of fermentative growth. The strain JG-X produced CO2 and H2 at a molar ratio that was slightly different from 1:1 (Fig. 2C; Table S1) and indicated that the produced gas did not exclusively originate from the coupled formate oxidation and proton reduction at the FHL complex. A similar excess of released CO2 has been reported earlier for glucose fermentation in a minimal growth medium by the E. coli strain MG1655 (32). The excess of CO2 may result from varying activities of CO2-releasing and consuming enzymes, such as the periplasmic FDHs EcFDH-O and EcFDH-N (28) as well as phosphoenolpyruvate carboxylase (PEPC) (33), respectively (Fig. 1). In the present study, cultivations were performed in a complex growth medium that contains amino acids, and their enzymatic decarboxylation may contribute to the observed CO2 liberation. The E. coli strain JG-X produced 1.9 mmol OD600−1 CO2 more than H2, and the EcFDH-H deletion practically abolished H2 production and increased the CO2 excess to 4.0 mmol OD600−1. This finding may be linked to FHL inactivation as it may lead to periplasmic formate accumulation and thereby boost formate oxidation by EcFDH-O and EcFDH-N. As opposed to FHL, these enzymes are not electrically linked to a hydrogenase, but electrons are instead transferred to the cytochrome bd oxidase I and nitrate reductase via the quinone pool of the cytoplasmic membrane giving rise to water and nitrite as electron transfer products, respectively, instead of H2 (34, 35). While an active PEPC may decrease the amount of released CO2 (Fig. 1), an apparent lack of succinate in the product spectrum of glucose-fermenting E. coli indicates that CO2 consumption via this metabolic route is negligible (32). EcFDH-H deletion reduced the volume of produced gas from 761 ± 8 mL (JG-X) to 34.7 ± 8.3 mL (FL004) and diminished the hydrogen fraction from 10.8 ± 0.2 mmol OD600−1 to 57.5 ± 0.6 µmol OD600−1 (Fig. 2C). This corresponds to a reduction of the detected amount of hydrogen down to 0.53% in the case of the EcFDH-H-deficient strain and is suspected to be not a product of microbial metabolism but rather a result from the exposure of the cell culture to a hydrogen-containing atmosphere in the anaerobic cabinet that could not be completely removed by purging with nitrogen gas. Importantly, the complete abolition of hydrogen production upon FHL complex disruption is in line with earlier studies according to which the majority of hydrogen produced during fermentative E. coli growth is produced by the HycE hydrogenase subunit of the FHL complex that depends on EcFDH-H as the source of electrons (36–43). In addition, the presented results suggest that recombinantly produced EcFDH-H can effectively complement an incomplete FHL complex and that EcFDH-H extension with a C-terminal His6-tag does not disturb electron transfer within the FHL complex. The latter finding is consistent with the FHL complex structure in which the C-terminus of the EcFDH-H subunit and the HycB–EcFDH-H interface are separated 23 Å from each other [Fig. S3 (44)] allowing for correct complex assembly.

Correlation of cell density and EcFDH-H activity

Inspired by the distinctive growth characteristics of E. coli strains JG-X and FL004, we hypothesized that the activity of an EcFDH-H variant may be estimated using the final cell density reached by EcFDH-H variant-producing E. coli cells during 24 h of glucose fermentation. To test this idea, the residues K44 and D179 were selected as substitution targets based on their respective degree of conservation among metal-dependent FDHs. Specifically, K44 is highly conserved and its substitution was expected to result in a large activity loss, while the aspartate at position 179 is less conserved and its substitution may be expected to result in minor enzyme activity changes. A correlation analysis was performed with EcFDH-H variants -K44A, -K44R, -D179L, and -D179E exhibiting formate oxidation activities below the wild-type level of 199 s−1 and between 0.4 s−1 and 136 s−1. A good correlation between enzyme activity and final cell density was apparent for all tested EcFDH-H variants with the exception of D179E, which slightly deviated from the overall trend toward a decreased enzyme activity or higher cell density (Fig. 2D). The reasons for this are not obvious but may include an increased expression level of the gene encoding EcFDH-H-D179E gene compared to the wild-type fdhF gene and an altered electron transfer from the variant to BV2+ that may decrease the apparent enzyme activity in the cell-free assay. Similar deviations from a linear relationship between cell growth and in vitro enzyme activity have been reported for an NADPH-dependent d-lactate dehydrogenase (45).

Screening of an EcFDH-H variant library by FHL complex complementation

Screening assay development and validation

Next, a screening assay was designed to screen an enzyme variant library in a platform that is presented schematically in Fig. 3. It consists of an initial pre-screening step that exploited the dependency of fermentative growth of E. coli on the catalytic activity of EcFDH-H. In the case of impaired EcFDH-H activity, also, the in vivo activity of the FHL is decreased resulting in a lower optical density during glucose fermentation. Clones reaching OD600 levels that were close to that of the parental strain were subjected to a consecutive colorimetric rescreening by a previously described enzyme assay (46, 47). In this second screening step, EcFDH-H enzyme variants are directly assayed in vitro and improvements in catalytic activity can be resolved. Importantly, both screening stages can be performed in a 96-well plate format. Data variability in a screen is an important assay characteristic and should be minimized in order to be able to detect differences between assayed enzyme variants reliably. For the optimization of the screening assay, these parameters were determined as the coefficient of variance (CV) and Z′-factor (48) from additional growth trials performed with the two reference strains JG-X (fdhF) and FL004 (ΔfdhF). The optimal glucose concentration for the measurement of anaerobic cell growth as OD600 was identified as 4 g/L (Fig. S2B; Table S2). In order to assess the variability of the re- and pre-screening methods, the performance of both clones in the screen was analyzed by 96 replicate measurements. Growth experiments performed with E. coli JG-X indicated a CV of 0.046 (Fig. 4A), which is low in comparison to the 2.5-fold increase of the reached cell density from the control strain level and allows the reliable identification of actively growing E. coli strains for subjection to rescreening. Here, we determined the final OD600 of 1.01 as the lower threshold (χ0) based on the average final OD600 (µ) and standard deviation (σ) of the wild type (WT) according to the equation χ0 = µ – 3 · σ. This limit allowed us to exclude clones with an average final OD600 <0.74 and a CV of 0.083 from the second screening step at a low likelihood of false positives (P = 1.24 · 10−5; determined by the probability density function as described in Materials and Methods). The relation between the data variability and the overall signal range of an assay may be quantified using the Z′-factor that should ideally lie in the range of 0.5–1.0 (48). The Z′-factor for the investigated pre-screening assay was 0.65 indicating an ideal separation of OD600 data of the positive and negative control groups, i.e., E. coli strains JG-X and FL004, respectively.

Fig 3 Schematic of a two-step anaerobic screening platform for the engineering of EcFDH-H. Firstly, a clone library of E. coli strains encoding different variants of EcFDH-H is prepared aerobically by focused or random mutagenesis and distributed into wells of library master plates. The screening process is performed under anaerobic conditions in a 96-well plate format and initiated by pre-screening the entire clone library using the final optical density (OD600) as a selection criterion. Clones exceeding an OD600 threshold (dashed line) calculated from the EcFDH-H-producing strain JG-X are subjected to FDH activity-based rescreening, and the clones exhibiting improved enzyme activity can be selected for an optional additional round of gene mutagenesis and two-step screening.

Fig 4 Signal distribution of sample clones during pre- and rescreening (A and B, respectively) of E. coli JG-X producing EcFDH-H (WT) (black) and EcFDH-H-deficient strain FL004 (red). Pre-screening of clones was performed by OD600 measurements in flat-bottom microtiter plates (MTPs) prior to which the bacterial cells were cultivated for 24 h at 30°C under an anaerobic atmosphere in a 250-µL glucose-supplemented growth medium in wells of a 500-µL deep-well plate. Sample preparation for the rescreening step included the collection of the cells contained in the pre-screened cell suspensions by centrifugation and their lysis by the addition of B-PER reagent. The formate dehydrogenase activity in the cell lysate was determined at λ = 555 nm using flat-bottom MTPs containing 5 µL of the cell lysate as well as formate and BV2+ dissolved in 50-mM phosphate buffer (pH 7.5) at concentrations of 10 and 2 mM, respectively. The average value (µ) and the set threshold (χ0 = µ ± 3 · σ) are displayed as solid and dashed lines, respectively.

The colorimetric rescreening was performed using E. coli cell lysates that were transferred into microtiter plate (MTP) wells containing the FDH substrates formate and benzyl viologen (BV2+). Enzymatic activity was monitored as the production of BV+ at λ = 555 nm. Statistical analysis of the obtained data (Fig. 4B) determined the CV value to be 0.66 for which samples with an average final A555/OD600 < 2.9 and a CV of 0.117 can be deselected at a low likelihood of false positives (P = 2.75 · 10−5; determined by the probability density function). To further complete the statistical analysis of the rescreening assay, the Z′-factor was determined as 0.51 validating the overall screening method. Overall, the throughput capacity of this assay is limited by the requirement of anaerobic conditions and was enhanced by the demonstrated screen robustness and effectiveness of the pre-screening stage for reducing the clone number subjected to rescreening. Based on the presented results, we estimate the upper limit of assay throughput to be 104 E. coli clones per round without automation.

Semi-rational engineering of EcFDH-H for increased catalytic activity

For the purpose of testing the screening system, we prepared a library that probed the role of five amino acids in loop 1 because it is positioned between the Mo-cofactor of the EcFDH-H active site and the [4Fe-4S] cluster a region (Fig. 5A) and may, therefore, provide interactions to neighboring enzyme regions involved in substrate binding and electron transfer. In order to minimize the risk of enzyme inactivation, residues with low conservation degrees were selected for substitution. These substitution targets were P9, A12, and S13 in the vicinity of the [4Fe-4S] cluster as well as V7 and K16 below the enzyme surface (Fig. 5A) and were investigated using a library comprised of 1,032 EcFDH-H variants where each of the five positions was mutated individually. The size of this library was considered sufficiently large to reach an acceptable genetic variability based on a study by Reetz et al. who determined that a 95% coverage can be obtained with NNK codon degeneracy if a minimum of 94 clones are included for a single position (49). The growth-based pre-screening significantly reduced the library size from 1,032 to 96 clones and revealed that, among the tested positions, only the glycine substitution variant of the A12X variant library reached the set lower OD600 threshold of 0.95 (Fig. 5B). These clones clustered around an OD600 of 1.1 while most of the remaining clones exhibited retarded growth characterized by a low final OD600 of ~0.39. Notably, a few clones were scattered between these two extremes and identified by gene sequencing as variants A12V and A12S as well as a single false-negative clone producing EcFDH-H variant A12G. The apparent partial retention of EcFDH-H function in the case of A12V and A12S may be explained by the chemical similarity between the substituting side chains and the methyl group at the Ala α-carbon. This notion is supported by the fact that both, valine and serine, are common in homologous metal-dependent formate dehydrogenases or nitrate reductases at positions corresponding to residue A12 in EcFDH-H (Fig. 5A). Notably, six out of the seven identified A12G clones reached OD600 values >χ0 indicating a high recognition probability for clones producing active EcFDH-H variants.

Fig 5 (A) Multi-sequence alignment of EcFDH-H and bacterial metal-dependent FDHs [PDB IDs: 1FDO (EcFDH-H), 6SDR (DvFDH), 1H0H (DgFDH), 6TG9 (RcFDH), 7BKB (MhFDH), and 5T5I (MwFDH)] with decreasing degrees of homology indicated by yellow, green, and blue and residue positions in EcFDH-H stated above the aligned sequences. The structure of the EcFDH-H active-site region is displayed with side chains of substituted residues as well as the [4Fe-4S] cluster and molybdopterin guanine dinucleotide (MGD) cofactors in stick representation. The substrate tunnel providing substrate access to the EcFDH-H active site (magenta filled-sphere array) was calculated and visualized using the CAVER PyMOL plugin (50) and the PyMOL Molecular Graphics System, version 1.7.2 Schrödinger, LLC. The results of pre- and rescreening of a variant library obtained by saturation mutagenesis of the codon encoding EcFDH-H residue A12 (marked with a red box) are shown in panels (B and C), respectively. The E. coli strains JG-X and EcFDH-H, respectively, are included for reference as clones 1–3. Only E. coli clones exceeding the pre-screening threshold (χ0, dashed line) were subjected to rescreening. The WT average (µ) is indicated as a solid line. The χ0 of the re- and pre-screening assay were calculated as χ0 = µ ± 3 · σ. Unidentified clones performing below χ0 are shown as gray squares.

The ability of selected clones to oxidize formate was assessed by the formate oxidation assay and identified the EcFDH-H variant A12G as the only enzyme with an improved activity exceeding the EcFDH-H (WT) conversion rate by a factor of 0.81 ± 0.10 (Fig. 5C). The purified EcFDH-H-A12G oxidized formate at 534 ± 10 s−1 at saturating formate concentration and 2 mM BV2+, corresponding to an 82% increase compared to the wild-type enzyme (Fig. S4A), and bound formate more tightly than the native enzyme, which was apparent as change of KmFormate from 2.36 ± 0.20 to 1.28 ± 0.08 mM. The binding affinity of the A12G variant for the soluble electron acceptor BV2+, on the other hand, was indistinguishable from the wild-type enzyme (Fig. S4B). Comparative molecular docking analysis suggested an alternative binding of formate in the active center via additional hydrogen bonds originating from residues H141 and R333 as a possible cause for the observed increased substrate binding affinity of the A12G variant (Fig. S5). Despite the A12G variant being more active than EcFDH-H (WT) in the colorimetric rescreening step, their host E. coli cells reached a similar final cell density in the growth-based pre-screening step. The reason for this different trend could be that the supplied glucose was limiting cell growth during the pre-screening and thereby restricted the final observable cell density. Nevertheless, screening at an increased glucose titer is unlikely to increase the quality of the pre-screening assay because the final OD600 did not significantly change at glucose concentrations ≥ 4g/L (Fig. 2A). Another possible cause for the apparent upper detection limit of the pre-screening assay is the catalytic coupling between EcFDH-H and HycE within the FHL complex. The conversion rate of its least active component limits catalysis of the overall reaction HCOO− + H+ → CO2 + H2. This notion is supported by comparable gas production by strains JG-X [EcFDH-H (WT)] and FL005 (EcFDH-H-A12G) during anaerobic cultivation, being 761 ± 8 and 758 ± 32 mL, respectively. As there were no significant differences between the final cell densities reached by clones producing the catalytically enhanced A12G variant and the wild-type enzyme, a consecutive activity-based rescreening step is required to reliably identify the enzyme variants with improved activity.

Conclusions

In summary, we present a robust growth-based screening strategy for the semi-rational engineering of EcFDH-H that was validated through the identification of variant A12G exhibiting an 82% increased formate oxidation activity. The presented screening method relies on the growth-inhibiting effect of glucose fermentation products that rapidly accumulate in EcFDH-H-deficient cells and are removed by the FHL. Therefore, the growth benefit provided by the FHL during fermentative growth may be employed to identify FHL components that completely or partially restore the activity of either of the two catalytic FHL centers and/or the iron–sulfur-cluster-mediated electron transfer between these active sites. In line with this mechanism, we envision the adaptation of this screening system to different FHL components and other electron-transferring enzyme complexes affecting microbial host fitness (51).

MATERIALS AND METHODS

Plasmids and strains

E. coli strain JG-X (MC4100, ∆fdhF, ∆iscR) harboring plasmids pTrc99a-fdhF (AmpR) and pSU21-selC (ChlR) was kindly provided to us by Prof. Shelley Minteer (Utah State University, Utah) (31). The strain JG-X was originally prepared by members of the Golbeck group (Pennsylvania State University, Pennsylvania) who have tested the MC4100 strain as well as the iscR and fdhF deletions and selC overexpression for the high-level production of EcFDH-H as described in their Air Force Research Laboratory project report AFRL-OSR-VA-TR-2014-0088 (52). The E. coli strain MC4100 is a K-12 derivative with the genotype [araD139]B/r Δ[argF-lac]169 λ− e14- flhD5301 Δ(fruK-yeiR)725(fruA25) relA1 rpsL150(strR) rbsR22 Δ(fimB-fimE)632(::IS1) deoC1, and its genome sequence has been published by Peters et al. (53). The selC gene of vector pSU-selC encodes the tRNA from which selenocysteyl-tRNASec (54, 55) is synthesized and overexpressed to support the high-level production of Sec-containing EcFDH-H. The vectors pFL004–pFL007 were used for EcFDH-H variant production (D179E, D179L, K44R, and K44A, respectively) and were prepared by site-directed mutagenesis of the fdhF gene in vector pTrc99a-fdhF using a previously described method (46, 56) and complementary DNA primer pairs (Table S3). For EcFDH-H variant production, pTrc99a-fdhF was eliminated from E. coli strain JG-X by repeated strain transfer (12 times) onto selective lysogeny broth (LB) agar (50 µg/mL chloramphenicol) yielding the strain FL003. The plasmid removal and absence of growth-inhibiting mutations were confirmed by replica-plating onto a double-selective solid medium containing ampicillin at 50 µg/mL as an additional antibiotic and comparative growth trials with the original and reconstituted parental strain JG-X (Fig. S6), respectively. The EcFDH-H-deficient control strain FL004 and EcFDH-H variant production strains FL005–FL009 were prepared by the transformation of electrocompetent cells of strain FL003 (V = 90 µL) with 1 µL of plasmid solution (c ≈ 50 ng/µL) of either insert-free pTrc99a (Nova Lifetech Ltd., Hong Kong) or pTrc99a-fdhF derivatives (Table 1), respectively. Electroporation of E. coli was performed using a MicroPulser electroporator (Bio-Rad, California) according to the manufacturer’s protocol and followed by a recovery step for which the cell suspension was supplemented with 800-µL SOC medium and incubated at 37°C for 1 h.

fdhf mutant library construction

Site saturation mutagenesis of the fdhF gene was performed according to the previously described QuikChange method (57). In brief, the codon encoding the amino acid of interest on each strand was mutated in two parallel PCR reactions during a few cycles (typically five) using only the forward or reverse primer of a primer set specific for the chosen DNA locus. In the second stage, the two PCR reactions are combined, and the entire plasmid was amplified in the presence of the complete complementary primer pair during 20–30 additional thermocycles. The codons encoding EcFDH-H residues V7, P9, A12, S13, and K16 were substituted with nucleotides encoding one of the other 19 common proteogenic amino acids. The used PCR reaction mix (V = 25 µL, Phusion PCR buffer) contained each dNTP at 0.1 mM, 200 nM degenerate primers (Table S3), 20 ng pTrc99a-fdhF template, and 0.02 U Phusion Hot Start II DNA polymerase (ThermoFisher Scientific, Massachusetts) and was subjected to five thermocycles including a 5-min primer extension step at 72°C. Subsequently, both reactions were combined, 0.02 U DNA polymerase was added, 20 thermocycles were performed, and the vector template was removed by incubation at 37°C for 2–3 h with 10 U added DpnI restriction enzyme. TOP 10 electrocompetent cells were transformed with the pTrc99a-fdhF derivatives by combining a 100-µL cell suspension with 1 µL of the PCR product followed by electroporation, as described above, followed by the addition of 900-µL SOC medium and incubation at 37°C for 2 h for cell recovery. A 50-µL cell culture aliquot was spread on a selective LB agar medium and cultivated overnight at 37°C after which typically >400 colonies had formed. The cells were resuspended in 1-mL sterile ddH2O, and contained plasmids were isolated using the NucleoSpin plasmid isolation kit (Macherey-Nagel, Germany). The purified plasmids were used to transform E. coli FL003 cells (Table 1) by electroporation in a reaction mix consisting of 1-µL plasmid solution (~50 ng/µL) and 90-µL cell suspension, and cell recovery was performed by the addition of a 800-µL SOC medium followed by cultivation at 37°C for 1 h. Then, the cell suspension was diluted with ddH2O by a factor of 30–50, spread onto a solid selective LB medium, and incubated overnight at 37°C. In order to optimally represent the mutational diversity of the generated pTrc99a-fdhF derivatives in the clone library, colonies derived from a single mutation site were distributed into individual 2.0-mL wells of two 96-well deep-well MTPs by resuspension in 400-µL selective LB. Next, the wells were sealed with gas-permeable film and incubated at 37°C, 70% humidity, and 600 rpm for 24 h. These master plates were stored at −80°C following the addition of sterile glycerol solution to a final concentration of 23% (vol/vol). The coverage of the mutational diversity was assessed by sequencing of plasmids (Microsynth, Germany) that were purified from clones contained in nine randomly selected wells of each master plate using the E-Z 96 FastFilter Plasmid DNA kit (Omega Bio-Tek, Georgia).

Bacterial cultivation in microtiter format for monitoring bacterial growth

Glycerol stocks of either E. coli JG-X or FL004 were back-diluted 1:50 in a 250-µL selective LB medium and incubated in 500-µL deep-well plates with conical bottoms (Eppendorf, Germany) for 24 h at 37°C and 600 rpm orbital shaking. The pre-culture was back-diluted 1:50 with a selective LB medium supplemented with 1 mM Na2MoO4 and 10 µM Na2SeO3, and EcFDH-H production was induced by the addition of isopropyl-β-D-1-thiogalactopyranoside (IPTG) (Sigma Aldrich, Germany) to 100 µM according to a published protocol (31). Then, the deep-well plates were sealed with Easyseal transparent film (676001; Greiner Bio-one, Germany) and placed inside a Ziploc bag along with anaerobic indicator strips (ThermoFisher Scientific, Massachusetts) for transfer into a custom-made container (Fig. S7) containing palladium catalysts (Advanced Instruments, Massachusetts). The container was hermetically sealed, filled with anaerobic gas mix (95% N2, 5% H2), and incubated for 24 h at 30°C and 200 rpm. Microbial growth was monitored in an anaerobic cabinet as optical density at 600 nm in flat-bottom MTPs (ThermoFisher Scientific, Massachusetts) containing 200-µL culture samples using a Multiskan SkyHigh microplate reader (ThermoFisher Scientific, Massachusetts). The OD600 in MTP wells with an unknown pathlength was determined based on the absorbance measured at λ = 600 nm using the equation OD600 nm = A600 nm/(1 cm · α). Here, α is the ratio between 1 cm and the length of the path that light travels through the 200-µL liquid sample in a MTP well and was determined empirically to be 3.57.

Bacterial cultivation in 10-mL serum bottles for monitoring of growth medium acidification

Pre-cultivation of E. coli from glycerol stocks was carried out as described above. Cultivation trials were performed by first sealing sterile 10-mL serum bottles with butyl rubber stoppers in an anaerobic cabinet. The pre-culture was then back-diluted at a ratio of 1:50 in a 5-mL selective LB medium supplemented with 1 mM Na2MoO4, 10 µM Na2SeO3, 4 g/L glucose, and 100 µM IPTG and incubated for 18 h at 30°C and 200 rpm. During cultivation, serum bottles were removed from the incubator in 2-h intervals and sampled for pH determination using an AE150 Benchtop pH Meter (Fisher Scientific, Finland) and for spectroscopic analysis (OD600) in flat-bottom MTPs (200-µL sample volume) using a Multiskan SkyHigh microplate reader. Serum bottles sampled at timepoints 20, 22, and 24 h were inoculated 14 h past inoculation of the remaining serum bottles (sampled at 2–18 h).

Bacterial cultivation in 1-L laboratory bottles for headspace analysis

Bacterial cultures were inoculated in an anaerobic cabinet by first transferring a 500-mL selective LB medium (50 µg/mL ampicillin, 50 µg/mL chloramphenicol) supplemented with 1 mM Na2MoO4, 10 µM Na2SeO3, 4 g/L glucose, and 100 µM IPTG into 1-L laboratory bottles (GL45; Schott, Germany) and adding 2.5 mL saturated E. coli pre-culture (preparation described above) followed by sealing with butyl rubber stoppers. The headspace of cultivation bottles was flushed with a stream of nitrogen gas for 10 min to remove hydrogen gas originating from the anaerobic cabinet. Cells were cultivated for 24 h at 30°C and 150 rpm followed by the discharging of headspace gas into a 1-L FlexFoil PLUS sample bag (SKC, UK). Gas chromatography was employed for the determination of the gas composition using an Agilent 490 Micro gas chromatograph (Agilent, California) equipped with two 10-m-long porous-layer open-tubular columns (0.32 mm inner diam.) coated either with a 30-µm-thick film of 5 Å molecular sieves (CP7535, Agilent, California) or a 10-µm-thick polar porous polymer (CP7580, Agilent, California). Gas sample components were separated in a stream of helium carrier gas and quantified with a micromachined thermal conductivity detector (Agilent, California) following calibration with standard gas mixtures. The total gas sample volume was determined by batch-wise removal of the sample bag contents with a graduated syringe.

Screening system evaluation

The Z′-factor and CV of the MTP-based screening system were used to assess its quality and calculated as described previously using equations 1 and 2, respectively (48). Here, the mean μ and the standard deviation σ were determined from the cell densities reached by the positive and negative control strains being E. coli JG-X and FL004, respectively:

(1) Z′ factor=1−[3⋅(σpos+σneg)μpos−μneg]

(2) CV=σμ

Growth-based pre- and colorimetric rescreening

Growth-based pre-screening was performed using the above-mentioned method for bacterial cultivation in microtiter format. Clones displaying higher final OD600 than the set threshold of the parental strain E. coli JG-X: χ0 = µ – 3 · σ were selected as active variants and analyzed further by the colorimetric rescreening as follows. Cells were cultured in 2.5-mL deep-well plates anaerobically as described above and sealed with Easyseal transparent film (676001; Greiner Bio-one, Germany) inside an anaerobic cabinet. Bacterial cells were harvested by centrifugation for 10 min at 4°C and 4,000 × g. The plates were then transferred into an anaerobic chamber, where the supernatant was removed and the cell pellet was resuspended in the 50-µL degassed B-PER cell lysis reagent containing 100 µg/mL of lysozyme (ThermoFisher Scientific, Massachusetts). The lysis mix was agitated for 30 min at room temperature and sampled for colorimetric rescreening. Here, a 5-µL cell lysate sample was transferred into a flat-bottom MTP well along with 185 µL of EcFDH-H reaction mix containing the EcFDH-H substrates formate and BV2+ dissolved in 50-mM potassium phosphate buffer (pH 7.5) at a concentration of 10 and 2 mM, respectively. The chromogenic reaction mixture was incubated for 5 min at 25°C after which its absorbance was recorded at 555 nm (A555) using a Multiskan SkyHigh microplate reader (ThermoFisher Scientific, Massachusetts). Clones exhibiting higher specific formate oxidation activity than the set activity threshold of E. coli JG-X (χ0 = µ + 3 · σ), as apparent from the A555/OD600 ratio, were selected for fdhF gene sequencing to identify the encoded EcFDH-H variant.

Statistical analysis

To assess the quality of the pre- and rescreening methods, the probability of selecting false-positive clones was determined at a standard distribution of the measured parameter (final OD600 reached during fermentative growth or cell-mass specific FDH enzyme activity) at a given hypothetical average value. Here, false positives are colonies belonging to a clone that perform above the set threshold in the used assay and whose average of all colonies belonging to the same clone is below the threshold. This analysis was based on a hypothetical clone library with a given average value of the monitored parameter (µ) around which the data of all samples of the same clone library are distributed according to a standard distribution. The CV value of this sample was assumed to be identical to the one determined for the strain FL004 analyzed by either the pre- or re-screening method. The probability of identifying false-positive clones at a given clone-characteristic average value was obtained using the probability density function (equation 3):

(3) f(x)=12πσe−12(χ0−μσ)2

Alignment of EcFDH-H homologs

Amino acid sequences of proteins with >30% sequence homology with EcFDH-H (PDB ID: 1FDO) were identified in the RCSB Protein Data Bank and aligned to EcFDH-H using the Clustal Omega tool (58) included in the bioinformatics software Geneious Prime (Biomatters, New Zealand).

ACKNOWLEDGMENTS

The authors wish to thank Prof. Shelley Minteer affiliated with the Utah State University (USA) for providing the E. coli strain JG-X and Niko Vuorio from VTT Technical Research Centre of Finland for gas chromatography analysis of headspace samples. The authors acknowledge the Aalto University Raw Materials Research Infrastructures and Bioeconomy facilities (Finland) and computational resources provided by CSC–IT Center for Science (Finland).

This research has received financial support from the Research Council of Finland in the form of an Academy Research Fellowship grant awarded to M.L. and funding for the ExtremoForm project granted to S.S. (Decision Nos. 321723 and 329510, respectively).

F.L. performed the conceptualization, methodology, investigation, formal analysis, writing (original draft), and writing (review and editing). S.S. performed the project administration, funding acquisition, resources, and writing (review and editing). M.L. performed the conceptualization, project administration, funding acquisition, resources, supervision, visualization, and writing (review and editing).

AFTER EPUB

[This article was published on 28 August 2024 with an error in equation 2. The equation was corrected in the current version, posted on 30 August 2024.]

DATA AVAILABILITY

Data used for the figures and supplemental tables are available upon request.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.01472-24.

10.1128/aem.01472-24.SuF1 Supplemental material aem.01472-24-s0001.pdf

Tables S1 to S3; Figures S1 to S7.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.
==== Refs
REFERENCES

1 Sheldon RA, Woodley JM. 2018. Role of biocatalysis in sustainable chemistry. Chem Rev 118 :801–838. doi:10.1021/acs.chemrev.7b00203 28876904
2 Wang Y, Xue P, Cao M, Yu T, Lane ST, Zhao H. 2021. Directed evolution: methodologies and applications. Chem Rev 121 :12384–12444. doi:10.1021/acs.chemrev.1c00260 34297541
3 Xiao H, Bao Z, Zhao H. 2015. High throughput screening and selection methods for directed enzyme evolution. Ind Eng Chem Res 54 :4011–4020. doi:10.1021/ie503060a 26074668
4 Quaglia D, Ebert M, Mugford PF, Pelletier JN. 2017. Enzyme engineering: a synthetic biology approach for more effective library generation and automated high-throughput screening. PLoS ONE 12 :e0171741. doi:10.1371/journal.pone.0171741 28178357
5 Han L, Liu X, Cheng Z, Cui W, Guo J, Yin J, Zhou Z. 2022. Construction and application of a high-throughput in vivo screening platform for the evolution of nitrile metabolism-related enzymes based on a desensitized repressive biosensor. ACS Synth Biol 11 :1577–1587. doi:10.1021/acssynbio.1c00642 35266713
6 Li Z, Deng Y, Yang G-Y. 2023. Growth-coupled high throughput selection for directed enzyme evolution. Biotechnol Adv 68 :108238. doi:10.1016/j.biotechadv.2023.108238 37619825
7 Wu S, Xiang C, Zhou Y, Khan MSH, Liu W, Feiler CG, Wei R, Weber G, Höhne M, Bornscheuer UT. 2022. A growth selection system for the directed evolution of amine-forming or converting enzymes. Nat Commun 13 :7458. doi:10.1038/s41467-022-35228-y 36460668
8 Neuenschwander M, Butz M, Heintz C, Kast P, Hilvert D. 2007. A simple selection strategy for evolving highly efficient enzymes. Nat Biotechnol 25 :1145–1147. doi:10.1038/nbt1341 17873865
9 Femmer C, Bechtold M, Panke S. 2020. Semi-rational engineering of an amino acid racemase that is stabilized in aqueous/organic solvent mixtures. Biotechnol Bioeng 117 :2683–2693. doi:10.1002/bit.27449 32492177
10 Calzadiaz-Ramirez L, Calvó-Tusell C, Stoffel GMM, Lindner SN, Osuna S, Erb TJ, Garcia-Borràs M, Bar-Even A, Acevedo-Rocha CG. 2020. In vivo selection for formate dehydrogenases with high efficiency and specificity toward NADP+. ACS Catal 10 :7512–7525. doi:10.1021/acscatal.0c01487 32733773
11 King E, Maxel S, Zhang Y, Kenney KC, Cui Y, Luu E, Siegel JB, Weiss GA, Luo R, Li H. 2022. Orthogonal glycolytic pathway enables directed evolution of noncanonical cofactor oxidase. Nat Commun 13 :7282. doi:10.1038/s41467-022-35021-x 36435948
12 Fernández‐Álvaro E, Snajdrova R, Jochens H, Davids T, Böttcher D, Bornscheuer UT. 2011. A combination of in vivo selection and cell sorting for the identification of enantioselective biocatalysts. Angew Chem Int Ed 50 :8584–8587. doi:10.1002/anie.201102360
13 Reetz MT, Höbenreich H, Soni P, Fernández L. 2008. A genetic selection system for evolving enantioselectivity of enzymes. Chem Commun (Camb) Chem Commun :5502–5504. doi:10.1039/b814538e
14 Baker K, Bleczinski C, Lin H, Salazar-Jimenez G, Sengupta D, Krane S, Cornish VW. 2002. Chemical complementation: a reaction-independent genetic assay for enzyme catalysis. Proc Natl Acad Sci U S A 99 :16537–16542. doi:10.1073/pnas.262420099 12482929
15 van Sint Fiet S, van Beilen JB, Witholt B. 2006. Selection of biocatalysts for chemical synthesis. Proc Natl Acad Sci U S A 103 :1693–1698. doi:10.1073/pnas.0504733102 16446453
16 Nearmnala P, Thanaburakorn M, Panbangred W, Chaiyen P, Hongdilokkul N. 2021. An in vivo selection system with tightly regulated gene expression enables directed evolution of highly efficient enzymes. Sci Rep 11 :11669. doi:10.1038/s41598-021-91204-4 34083677
17 Stripp ST, Duffus BR, Fourmond V, Léger C, Leimkühler S, Hirota S, Hu Y, Jasniewski A, Ogata H, Ribbe MW. 2022. Second and outer coordination sphere effects in nitrogenase, hydrogenase, formate dehydrogenase, and CO dehydrogenase. Chem Rev 122 :11900–11973. doi:10.1021/acs.chemrev.1c00914 35849738
18 Yishai O, Lindner SN, Gonzalez de la Cruz J, Tenenboim H, Bar-Even A. 2016. The formate bio-economy. Curr Opin Chem Biol 35 :1–9. doi:10.1016/j.cbpa.2016.07.005 27459678
19 Oliveira AM, Beswick RR, Yan Y. 2021. A green hydrogen economy for a renewable energy society. Curr Opin Chem Eng 33 :100701. doi:10.1016/j.coche.2021.100701
20 Cherkasov N, Ibhadon AO, Fitzpatrick P. 2015. A review of the existing and alternative methods for greener nitrogen fixation. Chem Eng Process Intensif 90 :24–33. doi:10.1016/j.cep.2015.02.004
21 Schwartz CJ, Giel JL, Patschkowski T, Luther C, Ruzicka FJ, Beinert H, Kiley PJ. 2001. IscR, an Fe-S cluster-containing transcription factor, represses expression of Escherichia coli genes encoding Fe-S cluster assembly proteins. Proc Natl Acad Sci U S A 98 :14895–14900. doi:10.1073/pnas.251550898 11742080
22 Jansing M, Mielenbrink S, Rosenbach H, Metzger S, Span I. 2023. Maturation strategy influences expression levels and cofactor occupancy in Fe-S proteins. J Biol Inorg Chem 28 :187–204. doi:10.1007/s00775-022-01972-1 36527507
23 Gräwert T, Kaiser J, Zepeck F, Laupitz R, Hecht S, Amslinger S, Schramek N, Schleicher E, Weber S, Haslbeck M, Buchner J, Rieder C, Arigoni D, Bacher A, Eisenreich W, Rohdich F. 2004. IspH protein of Escherichia coli: studies on iron-sulfur cluster implementation and catalysis. J Am Chem Soc 126 :12847–12855. doi:10.1021/ja0471727 15469281
24 Kiefer D, Merkel M, Lilge L, Henkel M, Hausmann R. 2021. From acetate to bio-based products: underexploited potential for industrial biotechnology. Trends Biotechnol 39 :397–411. doi:10.1016/j.tibtech.2020.09.004 33036784
25 Kirkpatrick C, Maurer LM, Oyelakin NE, Yoncheva YN, Maurer R, Slonczewski JL. 2001. Acetate and formate stress: opposite responses in the proteome of Escherichia coli. J Bacteriol 183 :6466–6477. doi:10.1128/JB.183.21.6466-6477.2001 11591692
26 Wang Y, Huang Y, Wang J, Cheng C, Huang W, Lu P, Xu Y-N, Wang P, Yan N, Shi Y. 2009. Structure of the formate transporter FocA reveals a pentameric aquaporin-like channel. Nature New Biol 462 :467–472. doi:10.1038/nature08610
27 Rossmann R, Sawers G, Böck A. 1991. Mechanism of regulation of the formate-hydrogenlyase pathway by oxygen, nitrate, and pH: definition of the formate regulon. Mol Microbiol 5 :2807–2814. doi:10.1111/j.1365-2958.1991.tb01989.x 1779767
28 Sawers G. 1994. The hydrogenases and formate dehydrogenases of Escherichia coli. Antonie Van Leeuwenhoek 66 :57–88. doi:10.1007/BF00871633 7747941
29 Beyer L, Doberenz C, Falke D, Hunger D, Suppmann B, Sawers RG. 2013. Coordination of FocA and pyruvate formate-lyase synthesis in Escherichia coli demonstrates preferential translocation of formate over other mixed-acid fermentation products. J Bacteriol 195 :1428–1435. doi:10.1128/JB.02166-12 23335413
30 Gevorgyan H, Vassilian A, Poladyan A, Trchounian K. 2022. The influence of hydrogen production on the formation of metabolic pathways and regulation of ΔpH in Escherichia coli. Int J Hydrogen Energy 47 :40264–40274. doi:10.1016/j.ijhydene.2022.07.227
31 Sahin S, Cai R, Milton RD, Abdellaoui S, Macazo FC, Minteer SD. 2018. Molybdenum-dependent formate dehydrogenase for formate bioelectrocatalysis in a formate/O2 enzymatic fuel cell . J Electrochem Soc 165 :H109–H113. doi:10.1149/2.0431803jes
32 Metcalfe GD, Sargent F, Hippler M. 2022. Hydrogen production in the presence of oxygen by Escherichia coli K-12. Microbiology (Reading) 168 :001167. doi:10.1099/mic.0.001167 35343886
33 Morikawa M, Izui K, Taguchi M, Katsuki H. 1980. Regulation of Escherichia coli phosphoenolpyruvate carboxylase by multiple effectors in vivo: I. Estimation of the activities in the cells grown on various compounds. J Biochem 87 :441–449. doi:10.1093/oxfordjournals.jbchem.a132764 6987214
34 Sousa PMF, Silva STN, Hood BL, Charro N, Carita JN, Vaz F, Penque D, Conrads TP, Melo AMP. 2011. Supramolecular organizations in the aerobic respiratory chain of Escherichia coli. Biochimie 93 :418–425. doi:10.1016/j.biochi.2010.10.014 21040753
35 Sparacino-Watkins C, Stolz JF, Basu P. 2014. Nitrate and periplasmic nitrate reductases. Chem Soc Rev 43 :676–706. doi:10.1039/c3cs60249d 24141308
36 Mnatsakanyan N, Bagramyan K, Trchounian A. 2004. Hydrogenase 3 but not hydrogenase 4 is major in hydrogen gas production by Escherichia coli formate hydrogenlyase at acidic pH and in the presence of external formate. Cell Biochem Biophys 41 :357–366. doi:10.1385/CBB:41:3:357 15509886
37 Maeda T, Sanchez-Torres V, Wood TK. 2012. Hydrogen production by recombinant Escherichia coli strains. Microb Biotechnol 5 :214–225. doi:10.1111/j.1751-7915.2011.00282.x 21895995
38 Redwood MD, Mikheenko IP, Sargent F, Macaskie LE. 2008. Dissecting the roles of Escherichia coli hydrogenases in biohydrogen production. FEMS Microbiol Lett 278 :48–55. doi:10.1111/j.1574-6968.2007.00966.x 17995952
39 Maeda T, Sanchez-Torres V, Wood TK. 2007. Escherichia coli hydrogenase 3 is a reversible enzyme possessing hydrogen uptake and synthesis activities. Appl Microbiol Biotechnol 76 :1035–1042. doi:10.1007/s00253-007-1086-6 17668201
40 Seol E, Jang Y, Kim S, Oh Y-K, Park S. 2012. Engineering of formate-hydrogen lyase gene cluster for improved hydrogen production in Escherichia coli. Int J Hydrogen Energy 37 :15045–15051. doi:10.1016/j.ijhydene.2012.07.095
41 Kim S, Seol E, Oh Y-K, Wang GY, Park S. 2009. Hydrogen production and metabolic flux analysis of metabolically engineered Escherichia coli strains. Int J Hydrogen Energy 34 :7417–7427. doi:10.1016/j.ijhydene.2009.05.053
42 McDowall JS, Murphy BJ, Haumann M, Palmer T, Armstrong FA, Sargent F. 2014. Bacterial formate hydrogenlyase complex. Proc Natl Acad Sci U S A 111 :E3948–56. doi:10.1073/pnas.1407927111 25157147
43 Pinske C, Sawers RG. 2016. Anaerobic formate and hydrogen metabolism. EcoSal Plus 7 . doi:10.1128/ecosalplus.ESP-0011-2016
44 Steinhilper R, Höff G, Heider J, Murphy BJ. 2022. Structure of the membrane-bound formate hydrogenlyase complex from Escherichia coli. Nat Commun 13 :5395. doi:10.1038/s41467-022-32831-x 36104349
45 Zhang L, King E, Luo R, Li H. 2018. Development of a high-throughput, in vivo selection platform for NADPH-dependent reactions based on redox balance principles. ACS Synth Biol 7 :1715–1721. doi:10.1021/acssynbio.8b00179 29939709
46 Li F, Scheller S, Lienemann M. 2023. Comparative analysis of CO2 reduction by soluble Escherichia coli formate dehydrogenase H and its selenocysteine-to-cysteine substitution variant. J CO2 Util 77 :102608. doi:10.1016/j.jcou.2023.102608
47 Axley MJ, Grahame DA, Stadtman TC. 1990. Escherichia coli formate-hydrogen lyase. purification and properties of the selenium-dependent formate dehydrogenase component. J Biol Chem 265 :18213–18218. doi:10.1016/S0021-9258(17)44740-5 2211698
48 Zhang J-H, Chung TDY, Oldenburg KR. 1999. A simple statistical parameter for use in evaluation and validation of high throughput screening assays. SLAS Discov 4 :67–73. doi:10.1177/108705719900400206
49 Reetz MT, Carballeira JD. 2007. Iterative saturation mutagenesis (ISM) for rapid directed evolution of functional enzymes. Nat Protoc 2 :891–903. doi:10.1038/nprot.2007.72 17446890
50 Chovancova E, Pavelka A, Benes P, Strnad O, Brezovsky J, Kozlikova B, Gora A, Sustr V, Klvana M, Medek P, Biedermannova L, Sochor J, Damborsky J. 2012. CAVER 3.0: a tool for the analysis of transport pathways in dynamic protein structures. PLoS Comput Biol 8 :e1002708. doi:10.1371/journal.pcbi.1002708 23093919
51 Lienemann M. 2021. Molecular mechanisms of electron transfer employed by native proteins and biological-inorganic hybrid systems. Comput Struct Biotechnol J 19 :206–213. doi:10.1016/j.csbj.2020.12.004 33425252
52 Golbeck JH, Bryant DA, Lubner CE, Gorka M, Applegate A, Barchinger S. 2014. Solar to liquid fuels production: light-driven reduction of carbon dioxide to formic acid. The Pennsylvania State University, Laboratory AFR, Arlington, Virginia.
53 Peters JE, Thate TE, Craig NL. 2003. Definition of the Escherichia coli MC4100 genome by use of a DNA array. J Bacteriol 185 :2017–2021. doi:10.1128/JB.185.6.2017-2021.2003 12618467
54 Leinfelder W, Zehelein E, Mandrand-Berthelot M-A, Bock A. 1988. Gene for a novel tRNA species that accepts L-serine and cotranslationally inserts selenocysteine. Nature 331 :723–725. doi:10.1038/331723a0 2963963
55 Baron C, Heider J, Böck A. 1990. Mutagenesis of selC, the gene for the selenocysteine-inserting tRNA-species in E. coli: effects on in vivo function. Nucleic Acids Res 18 :6761–6766. doi:10.1093/nar/18.23.6761 1702199
56 Liu H, Naismith JH. 2008. An efficient one-step site-directed deletion, insertion, single and multiple-site plasmid mutagenesis protocol. BMC Biotechnol 8 :91. doi:10.1186/1472-6750-8-91 19055817
57 Wang W, Malcolm BA. 1999. Two-stage PCR protocol allowing introduction of multiple mutations, deletions and insertions using QuikChange site-directed mutagenesis. Biotechniques 26 :680–682. doi:10.2144/99264st03 10343905
58 McWilliam H, Li W, Uludag M, Squizzato S, Park YM, Buso N, Cowley AP, Lopez R. 2013. Analysis tool web services from the EMBL-EBI. Nucleic Acids Res 41 :W597–W600. doi:10.1093/nar/gkt376 23671338
