Optimization of Culture Conditions for Oxygen-Tolerant Regulatory [NiFe]-Hydrogenase Production from Ralstonia eutropha H16 in Escherichia coli

Abstract

Hydrogenases are abundant metalloenzymes that catalyze the reversible conversion of molecular H₂ into protons and electrons. Important achievements have been made over the past two decades in the understanding of these highly complex enzymes. However, most hydrogenases have low production yields requiring many efforts and high costs for cultivation limiting their investigation. Heterologous production of these hydrogenases in a robust and genetically tractable expression host is an attractive strategy to make these enzymes more accessible. In the present study, we chose the oxygen-tolerant H₂-sensing regulatory [NiFe]-hydrogenase (RH) from Ralstonia eutropha H16 owing to its relatively simple architecture compared to other [NiFe]-hydrogenases as a model to develop a heterologous hydrogenase production system in Escherichia coli. Using screening experiments in 24 deep-well plates with 3 mL working volume, we investigated relevant cultivation parameters, including inducer concentration, expression temperature, and expression time. The RH yield could be increased from 14 mg/L up to >250 mg/L by switching from a batch to an EnPresso B-based fed-batch like cultivation in shake flasks. This yield exceeds the amount of RH purified from the homologous host R. eutropha by several 100-fold. Additionally, we report the successful overproduction of the RH single subunits HoxB and HoxC, suitable for biochemical and spectroscopic investigations. Even though both RH and HoxC proteins were isolated in an inactive, cofactor free apo-form, the proposed strategy may powerfully accelerate bioprocess development and structural studies for both basic research and applied studies. These results are discussed in the context of the regulation mechanisms governing the assembly of large and small hydrogenase subunits.

Keywords: Escherichia coli; Ralstonia eutropha; [NiFe]-hydrogenase; cofactor assembly; difficult-to-express protein; heterologous protein production.

Figure 1

Schematic representation of the heterotetrameric native RH hydrogenase associated with the histidine kinase HoxJ (left) and the truncated heterodimeric RH_stop used in this study (right) adapted from the review of Lenz et al. [10]. The NiFe(CN)₂(CO) active site is bound to the large subunit (blue) via four cysteine residues, while the small subunit (green) hosts three [4Fe–4S] clusters.

Figure 2

Heterologous overproduction of Hox proteins from plasmids pQF4, pQF5, or pQF8 in E. coli BL21-Gold (BQF4B, BQF5C, or BQF8RH). Cells were cultivated in TB medium in deep-well plates at 37 °C as described in Materials and Methods. RH production was induced with varying IPTG concentrations. HoxB production was analyzed by Western blotting (normalized to OD₆₀₀) with Strep-tag antibodies (right part) and subsequent quantification of the bands using ImageJ. For each strain, protein amounts were calculated relative to the amount of protein at a maximum inducer concentration of 1 mM IPTG.

Figure 3

RH production in E. coli BQF8RH cultivated in 50 mL TB medium in UYF at 37 °C and 30 °C. RH production was induced with IPTG as indicated for 5 h. Soluble RH was purified by affinity chromatography and subsequently analyzed by SDS-PAGE and Western blotting. (A) Calculated amounts of purified RH after quantification of the stained gels in Figure S3 with ImageJ; (B) Western blot of total protein and soluble protein with antibodies against the Strep-tag at HoxB.

Figure 4

Time-dependent overproduction of HoxB_Strep, HoxC_Strep and RH. Strains BQF4B, BQF5C, and BQF8RH were cultivated in 50 mL TB medium at 30 ^°C in UYF. Samples normalized to OD₆₀₀ of 5 were taken at the indicated time points after the addition of 50 µM IPTG (t = 0 h). Production of Hox proteins was analyzed by Western blotting. Growth curves of the different strains and relative Hox protein levels (left) as quantified from the Western blot (right). The individual protein levels after 23 h of induction were set to 1.

Figure 5

Heterologous production of RH by strain BQF8RH grown in EnPresso B medium with or without addition of booster. Cells were cultivated in deepwell plates at 30 °C as described in Materials and Methods. RH production was induced with varying IPTG concentrations. (A) Analysis of HoxB production by Western blotting with anti-Strep-tag antibodies. (B) Calculation of the specific HoxB yield normalized to the yield in non-boosted cultures induced with 1 mM IPTG. (C) Growth curve of representative cultures. The black arrow indicates the induction point. (D) Calculation of the volumetric HoxB yield normalized as in B. (E) Growth curves of strain BQF8RH in 50 mL EnPresso medium with or without booster in UYF. (F) Calculation of specific and volumetric RH yield from strain BQF8RH as grown in E.

Figure 6

Up-scaling of RH production to 500 mL culture volume in 2.5 L UYF (A) growth curve of strain BQF8RH in complex TB or boosted EnPresso. (B) Calculation of specific and volumetric RH yield from strain BQF8RH as grown in A.

Figure 7

Comparative spectroscopic characterization of RH and isolated HoxC purified from R. eutropha and E. coli. (A) Comparison of IR spectra of as-isolated HoxC produced from E. coli (EcHoxC, red line) and R. eutropha (ReHoxC, black line). (B) Comparison of IR spectra of as-isolated RH_stop produced from E. coli (dark yellow line) and R. eutropha (black line). (C) Comparison of UV-vis absorption spectra of 20 µM as-isolated HoxC produced from E. coli (red line) and R. eutropha (black line). The spectrum of E. coli-derived HoxC shows only minor absorptions around 420 nm, probably related to heme-containing protein contaminants copurified with the large subunit. HoxC from R. eutropha shows active site contributions characterized by two absorption bands at 370 and 390 nm [14]. (D) EPR spectra recorded at 10 K with a microwave power of 1 mW of as-isolated RH_stop (0.2 mM) from E. coli (top trace), containing minor contributions of a [3Fe–4S]⁺ cluster species, in line with data recorded of native RH [34]. The bottom trace represents reduced RH_stop treated with an excess of sodium dithionite. The spectrum contains rhombic signals attributed to a [4Fe–4S]⁺ cluster. The asterisk (*) denotes a weak signal, deriving presumably from a [2Fe–2S] cluster subspecies originating either from partial [4Fe–4S] cluster degradation or from protein contaminants. Spectra in A, B are normalized to the amide II band intensity at 1548/cm.