Characterization of ferredoxins from the thermophilic, acetogenic bacterium Thermoanaerobacter kivui

A major electron carrier involved in energy and carbon metabolism in the acetogenic model organism Thermoanaerobacter kivui is ferredoxin, an iron–sulfur‐containing, electron‐transferring protein. Here, we show that the genome of T. kivui encodes four putative ferredoxin‐like proteins (TKV_c09620, TKV_c16450, TKV_c10420 and TKV_c19530). All four genes were cloned, a His‐tag encoding sequence was added and the proteins were produced from a plasmid in T. kivui. The purified proteins had an absorption peak at 430 nm typical for ferredoxins. The determined iron–sulfur content is consistent with the presence of two predicted [4Fe4S] clusters in TKV_c09620 and TKV_c19530 or one predicted [4Fe4S] cluster in TKV_c16450 and TKV_c10420 respectively. The reduction potential (Em) for TKV_c09620, TKV_c16450, TKV_c10420 and TKV_c19530 was determined to be −386 ± 4 mV, −386 ± 2 mV, −559 ± 10 mV and −557 ± 3 mV, respectively. TKV_c09620 and TKV_c16450 served as electron carriers for different oxidoreductases from T. kivui. Deletion of the ferredoxin genes led to only a slight reduction of growth on pyruvate or autotrophically on H2 + CO2. Transcriptional analysis revealed that TKV_c09620 was upregulated in a ΔTKV_c16450 mutant and vice versa TKV_c16450 in a ΔTKV_c09620 mutant, indicating that TKV_c09620 and TKV_c16450 can replace each other. In sum, our data are consistent with the hypothesis that TKV_c09620 and TKV_c16450 are ferredoxins involved in autotrophic and heterotrophic metabolism of T. kivui.


Introduction
Acetogenic bacteria are one of the primordial life forms on Earth which fix CO 2 by the Wood-Ljungdahl pathway (WLP) [1,2]. In the WLP, two mol of CO 2 are reduced via the carbonyl and methyl branches leading to the production of acetyl-CoA [3]. In the anabolic route, acetyl-CoA is further carboxylated to pyruvate and from there, carbon flows in the usual biosynthetic routes for the production of cell material [4,5]. In the catabolic route, acetyl-CoA is converted to acetate via acetyl phosphate, giving rise to one mol ATP per mol of acetate [6]. Since one ATP is consumed in the first step of the methyl branch [7], the overall ATP balance by substrate level phosphorylation is zero. During heterotrophic growth on sugars, additional ATP is gained during glycolysis, but during autotrophic growth with hydrogen as electron donor for CO 2 reduction, additional energy is generated by a chemiosmotic mechanism. Up to date, two respiratory enzymes are known in acetogens, the ferredoxin:NAD + (Rnf) [8,9] or the ferredoxin:H + (Ech) oxidoreductase [10,11]. Noteworthy, both enzymes use low potential, reduced ferredoxin (Fd) as reductant.
Any ferredoxin-containing organism tends to contain multiple copies of ferredoxins, although the function of these copies is often unclear [13,25,26]. Thermoanaerobacter kivui is a thermophilic, acetogenic model organism that can sustain a heterotrophic and lithoautotrophic lifestyle using different sugars, H 2 + CO 2 , CO or mixtures of both (syngas) as carbon and energy source [17,[27][28][29][30][31]. The decryption of its genome sequence led to the identification of four possible ferredoxin-encoding genes (TKV_c09620, TKV_c16450, TKV_c10420 and TKV_c19530) with so far unknown function. Here, we provide evidence by biochemical and genetic analysis that TKV_c09620 and TKV_c16450 are ferredoxins involved in autotrophic and heterotrophic metabolism of T. kivui.

Results
Potential ferredoxin-encoding genes in T. kivui, their genetic organization and properties of deduced gene products Analyses of the genome sequence of T. kivui revealed four possible ferredoxin-like genes: TKV_c09620, TKV_c16450, TKV_c10420 and TKV_c19530. TKV_ c09620 is preceded by a gene transcribed in the opposite direction encoding a putative tRNA. Downstream of TKV_c09620 is an aminotransferase apoenzyme gene, followed by three genes putatively coding for an alanyl-tRNA synthetase (Fig. 1A). Inspection of the second gene TKV_c16450 showed that this ferredoxinlike gene is preceded by a putative regulator gene and is flanked upstream by genes coding for a putative spore protein, protease or an unknown reductase (Fig. 1B). The third gene, TKV_c10420, is surrounded by a gene that putatively encodes an endopeptidase and three genes whose predictions share high sequence similarities to a putative NADH:flavin oxidoreductases and a putative hydrolase (Fig. 1C). Lastly, TKV_ c19530 is preceded by three genes whose deduced protein sequences share high sequence similarities to a putative, membrane-anchored sulfite transporter, followed by a putative transcriptional regulator and a gene transcribed in the opposite direction encoding a protein with so far unknown function (Fig. 1D).
To analyse the transcriptional organization of the ferredoxin-like genes, RNA was isolated from T. kivui cells grown on glucose and complementary DNA (cDNA) was synthesized. Bridging PCR analyses revealed that TKV_c09620 and TKV_c16450 are monocistronic (Fig. 1E,F), whereas TKV_c10420 is part of a transcriptional unit including the NADH:flavin oxidoreductases and hydrolase-like genes (Fig. 1G) and TKV_c19530 forms a transcriptional unit with the genes coding for a putative membrane-anchored sulfite transporter and a putative transcriptional regulator (Fig. 1H).  Construction and characterization of T. kivui DTKV_c09620, DTKV_c16450, DTKV_c10420 and DTKV_c19530 mutant strains To address the physiological role of the ferredoxin-like genes in T. kivui, the genes were deleted individually (Fig. 2). The mutagenesis was performed in a T. kivui pyrE deletion mutant by double homologous recombination using suicide plasmids [36] (Fig. S1).
T. kivui DpyrE lacks the orotate phosphoribosyltransferase (pyrE) and is uracil-auxotrophic [36]. Since the plasmids contain homology arms with the upstream and downstream regions of TKV_c09620, TKV_c16450, TKV_c10420, TKV_c19530 and pyrE as a selection marker, the first round of selection was performed on Ferredoxins TKV_c09620 TKV_c10420 TKV_c16450 TKV_c19530 Table 1. Comparison of amino acid sequences from TKV_c09620, TKV_c16450, TKV_c10420 and TKV_c19530 with ferredoxins from archaea or bacteria. Fig. 2. Deletion of TKV_c09620, TKV_c16450, TKV_c10420 and TKV_c19530. Strategy for deletion of ferredoxin-like genes using plasmid pMJ007 via two independent homologous recombination events (A). To verify the deletion progress, T. kivui colonies were picked and the gene loci TKV_c09620 (B), TKV_c16450 (C), TKV_c10420 (D) and TKV_c19530 (E) were checked by using corresponding primers (21-36; Table S1) binding inside (1, 2, 3) or outside (4, 5, 6) of each gene locus. The resulting sizes for the "outside" PCR of undeleted and deleted agar plates in minimal media without uracil using glucose as a substrate to select for transformants with plasmid integration ( Fig. 2A). Further, these transformants were subjected to a second round of selection in minimal media supplemented with glucose, uracil and 5fluoroorotic acid to select the cells that had lost the plasmid ( Fig. 2A). The deletion sites were confirmed by flanking PCRs (Fig. 2B-E) using primer pairs listed in Table S1 and Sanger sequencing. The generated deletion mutants T. kivui DTKV_ c09620, DTKV_c16450, DTKV_c10420 and DTKV_ c19530 were tested for growth on different substrates (Fig. 3). Surprisingly, the mutants grew in minimal media on glucose, mannitol or H 2 + CO 2 as substrate with growth rates, final yields and product formations indistinguishable from T. kivui DpyrE (Fig. 3A,B,D). When cultivated on glucose, all T. kivui strains grew with a doubling time of 2.1 h and reached a final OD 600 of 1.5 after 13.0 h (Fig. 3A). When mannitol was used as carbon and energy source cells started to grow after a lag phase of 10.5 h (DpyrE), 17 h (DTKV_c09620), 10.5 h (DTKV_c16450), 10.5 h (DTKV_c10420) or 15 h (DTKV_c19530) respectively (Fig. 3B). After the lag phase, all strains reached a doubling time of % 3.8 h and a final OD 600 of 1.3 (Fig. 3B). When cultivated autotrophically on H 2 + CO 2 T. kivui DpyrE started to grow after a lag phase of around 28 h with a doubling time of 47.8 h and reached a final OD 600 of 0.33 after 150 h (Fig. 3D). However, the mutants had a slightly higher doubling time of around 53.2 h (DTKV_c09620), 53.8 h (DTKV_c10420), 56.3 h (DTKV_c16450) or 53.5 h (DTKV_c19530) and reached a final OD 600 of 0.27 (DTKV_c09620, DTKV_c10420, DTKV_c19530) or 0.24 (DTKV_c16450) after 150 h (Fig. 3D). The mutants also had a phenotype when growing on pyruvate (Fig. 3C). When cultivated on 100 mM pyruvate, T. kivui DpyrE grew with a doubling time of 2.9 h and reached a final OD 600 of 1.2 after 17.3 h (Fig. 3C). In contrast, the doubling time of T. kivui DTKV_c09620, DTKV_c16450, DTKV_c10420 and DTKV_c19530 was reduced by 2.1-, 2.4-, 1.9-or 2.9-fold (to 6.2, 7.3, 5.9 or 8.9 h) respectively (Fig. 3C). Not only the growth rate but also the final yield was affected. T. kivui DTKV_c09620, DTKV_c16450, DTKV_c10420 and DTKV_c19530 reached a final OD 600 of 1.0, 0.8, 1.1 or 0.8 after 30, 36, 36 or 39 h respectively (Fig. 3C). However, all substrates were always completely consumed under all growth conditions tested and acetate was the only metabolic product observed in the exponential as well as in the stationary growth phase (data not shown). Since all the mutants were viable at all conditions tested, TKV_c09620, TKV_c16450, TKV_ c10420 and TKV_c19530 are not essential for catabolic and anabolic processes in T. kivui.

TKV_c09620 and TKV_c16450 are highly regulated
To shed some light on the involvement of the four ferredoxin-like proteins under different physiological conditions, we subsequently analysed whether the expression level of the ferredoxin-like genes was affected in the mutant strains during growth on glucose or H 2 + CO 2 . Thus, cDNA was prepared using RNA isolated from T. kivui DpyrE and mutant cultures in the mid-exponential growth phase (Fig. 4). Interestingly, when T. kivui DTKV_c09620 cells were grown on glucose or H 2 + CO 2 , the relative transcript level of TKV_c16450 was highly increased by a factor (log2-fold change) of 7.4 AE 0.7 or 4.9 AE 1.1 compared to the DpyrE strain (Fig. 4A,B). Vice versa, TKV_c09620 was 4.9-(AE 0.5) or 6.2-(AE 0.4) log2-fold upregulated compared to DpyrE in the DTKV_c16450 cells (Fig. 4A,B). In contrast, the relative transcript level of TKV_c10420 and TKV_c19530 did not change significantly in any of the mutants (Fig. 5A,B). In sum, TKV_c09620 and TKV_c16450 seem to replace each other.
Therefore, we aimed to delete both, TKV_c09620 and TKV_c16450. Although the technique of markerless mutagenesis to get multiple deletions in one strain is well established for T. kivui [37], we did not obtain Fig. 4. Expression analyses of the ferredoxin-like genes in T. kivui ΔpyrE and deletion mutants. T. kivui cells were grown on minimal media at 66°C with 25 mM glucose (A) or H 2 + CO 2 (2 9 10 5 Pa) (B) as substrate to mid-exponential growth phase. Cells were harvested and mRNA was isolated. DNase-digested mRNA was transcribed into complementary DNA (cDNA). cDNA was used as a template for qRT-PCR analysis as described in Material and Methods. cDNA was prepared from two biological replicates. All data points are mean AE SEM; n = 3 independent experiments. a DTKV_c09620/TKV_c16450 double mutant. We tried to get the second deletion either in the TKV_c09620 or the TKV_c16450 background. After transforming cells with the suicide plasmid, they were plated for the first round of selection on minimal media without uracil containing either glucose, pyruvate, formate, H 2 + CO 2 or the before-mentioned substrates under a hydrogen atmosphere. Further, these transformants were subjected to second round of selection in minimal media supplemented with uracil, 5-fluoroorotic acid and before-mentioned substrates to select the isolates with the loss of plasmid. We also tried substrate combinations such as glucose and formate. The same substrates or substrate combinations were also tested on iron-depleted media (% 0.11 lM FeSO 4 ) in the hope that ferredoxins may not be required, but this was also without success. Despite all attempts, a double deletion mutant was not obtained.
Production of TKV_c09620, TKV_c16450, TKV_c10420 and TKV_c19530 To elucidate the biochemistry of the four putative ferredoxins, we purified each ferredoxin-like protein using a plasmid-based production system in T. kivui [15]. Therefore, we cloned TKV_c09620, TKV_c16450, TKV_c10420 or TKV_c19530 together with a DNA sequence coding for a 109 His-tag into the expression vector pMU131 (Fig. S2). The expression of each ferredoxin-like gene was under the control of the constitutively expressed S-layer promotor [15,36]. Naturally competent cells of T. kivui were transformed with the plasmids pMU131_His-TKV_c09620, pMU131_ His-TKV_c16450, pMU131_His-TKV_c10420 or pMU 131_His-TKV_c19530 and the encoded proteins containing a genetically engineered His-tag were purified. Therefore, crude extract of the strains grown on glucose was prepared and applied to a Ni 2+ -NTA column as described in Material and Methods. His-TKV_c09620 and His-TKV_c16450 were purified by affinity chromatography followed by size exclusion chromatography on Superdex 200 to apparent homogeneity (Fig. 5A,B). Surprisingly, His-TKV_c10420 and His-TKV_c19530 could not be purified using this method (Fig. S3). Therefore, we changed the strategy and decided to produce the proteins in E. coli BL21 (DE3) ΔiscR [38]. We cloned TKV_c10420 and TKV_c19530 together with a DNA sequence coding for a 109 His-tag into the expression vector pET21a and transformed the plasmids pET21a_His-TKV_ c10420 and pET21a_His-TKV_c19530 into E. coli BL21 (DE3) ΔiscR (Fig. S4). Using this expression system, His-TKV_c10420 and His-TKV_c19530 could be purified from the crude extract by affinity chromatography followed by size exclusion chromatography on Superdex 200 to apparent homogeneity (Fig. 5C,D). His-TKV_c09620, His-TKV_c16450, His-TKV_ c10420 and His-TKV_c19530 were purified with yields of 0.6, 0.3, 0.5 or 0.4 mg per g wet cells respectively. Analyses of the purified proteins separated on a 14% SDS-polyacrylamide gel revealed one protein with apparent molecular masses of % 20 (  (Fig. 5D). These molecular masses do not correspond with the expected sizes for TKV_c09620 (5.9 kDa), TKV_c16450 (6.9 kDa), TKV_c10420 (7.9 kDa) and TKV_c19530 (12.6 kDa) of T. kivui, due to the fact that the migration of ferredoxin-like proteins in PAGE behaves differently [39,40]. However, the identity of the four proteins was verified by peptide mass fingerprinting. Chemical analysis revealed 8.6 AE 1.3, 8.4 AE 0.8, 3.4 AE 0.6 or 3.0 AE 0.3 mol of iron per mol of protein, which matches the prediction that TKV_c09620 and TKV_c19530 contain two [4Fe4S]-and TKV_c16450 and TKV_c10420 one [4Fe4S]-cluster respectively.
Next, we analysed the electrochemical properties of the purified, ferredoxin-like proteins ( Fig. 6; Table 2). TKV_c09620 and TKV_c16450 were both observed to exhibit one pronounced redox feature at À386 mV vs. SHE, which in both cases closely resembled a 1e À redox couple (adsorbed fit analysis performed using QSOAS [41]) (Fig. 6A,B). In contrast, TKV_c10420 and TKV_c19530 both exhibited two 1e À redox species ranging from À350 to À559 mV vs. SHE. In the case of TKV_c19530, the peak separation between the oxidation and reduction peaks was found to be between 40 and 50 mV (Fig. 6C,D); this could be due to non-ideal protein adsorption to the electrode and/or poor interfacial electron transfer between the electrode and the redox cofactor of the protein, or due to a dispersion of ferredoxin orientations across the electrode surface.
Moreover, we obtained the UV-visible absorption spectra of the purified ferredoxin-like proteins reduced by the purified CODH/ACS from T. kivui [23]. When His-TKV_c09620 and His-TKV_c16450 were incubated under an atmosphere of CO in the presence of the purified bi-functional CODH (CODH/ACS) from T. kivui [23], a typical decrease in absorption at 430 nm was obtained (Fig. S7A,B). In contrast, when the experiments were repeated using His-TKV_c10420 and His-TKV_c19530, no reduction at 430 nm could be observed, indicating that these [4Fe4S]-like proteins could not accept electrons from the CODH/ACS (Fig. S7C,D).

Discussion
In this work, we purified and characterized four ferredoxin-like proteins from the anaerobic acetogenic bacterium T. kivui. Moreover, we developed an improved and simple purification protocol for thermophilic ferredoxins, based on homologous overproduction in an acetogen, a method that may be transferred to other acetogenic ferredoxins as well. Since oxidoreductases from thermophilic organisms require temperature-stable ferredoxins as electron transfer proteins, the purified ferredoxins from T. kivui can be used in vitro for biochemical assays at high temperatures. The genome of T. kivui encodes four putative ferredoxin-like proteins TKV_c09620, TKV_c16450, TKV_c10420 and TKV_c19530. However, only TKV_c09620 and TKV_c16450 share high sequence identities with the ferredoxins of C. pasteurianum (Fd_CLOPA) [32,33], T. maritima (TM0927) [34] or P. furiosus (PFC_07725) [35] (Table 1). TKV_c09620 and TKV_c16450 not only share high sequence identities but also have similar redox potentials as postulated for ferredoxins from archaea or bacteria (Table 3). We clearly demonstrate the capability of TKV_c16450 and TKV_c09620 to support several enzymatic reactions that required the delivery of electrons such as PFOR [15], Ech [10,11], HydABC [17], MetFV [17], CODH/ACS or CooS [23]. Moreover, our results clearly demonstrate that TKV_c09620 and TKV_c16450 can replace each other in vivo and in vitro. The sequence identity and similar redox potentials of TKV_c09620 and TKV_c16450 support these findings (Tables 1 and 3). In contrast, our data also demonstrate that TKV_c10420 and TKV_c19530 are iron-sulfur-containing proteins of unknown physiological function. TKV_c10420 could be an electron transfer subunit from an electron-bifurcating enzyme that may consist of the NADH:flavin oxidoreductase and a hydrolase. Similar subunit compositions in electron-bifurcating enzymes are typical for bacteria and archaea [44,45]. TKV_c19530 is probably part of a sulfite transporter, similar to the iron-uptake transporter NFeoB/FeoC in Klebsiella pneumoniae [46]. How electron transport is mediated in these transporters and if they are linked to the respiratory chain is not known.
Interestingly, only TKV_c09620, TKV_c16450 as well as the ferredoxin from C. pasteurianum served as electron acceptor for the pyruvate:Fd oxidoreductase (PFOR) [15], electron-bifurcating hydrogenase (HydABC) [17], monomeric CODH (CooS) [23], bifunctional CODH (CODH/ACS) [23], methylene-THF reductase (MetFV) [17] and energy-converting hydrogenase (Ech2) [11] from T. kivui. In addition, a lot of enzymes from acetogens and other anaerobes use the ferredoxin isolated from C. pasteurianum (Fd_CLOPA) [9,11,15,17,18,23]. Although ferredoxins from different species are unrelated, they generally can substitute for each other. This does not mean that there is no selectivity at all for a specific ferredoxin. Indeed, there is always some preference [47][48][49]. For example, the catalytic efficiency of the ferredoxin: NADP + oxidoreductase (FNR) from chloroplasts is much higher with the ferredoxin from the same plant than with the ferredoxin of other organisms [50]. Regardless of the fact that the specific role of TKV_c09620 or TKV_c16450 remains elusive, our results clearly show a higher preference of TKV_c09620 for Ech2 and TKV_c16450 for HydABC, CooS and CODH/ACS. The compatibility of Fd_CLOPA with different Fddependent enzymes of acetogens can be explained structurally. While the NMR structure of Fd_CLOPA was already available [51], AlphaFold2 was used to predict an overall fold for the other four ferredoxins (Fig. 8). TKV_c09620 and TKV_c16450 have high structural identities compared to Fd_CLOPA (Fig. 8). However, the Fd-binding site of enzymes is not yet demonstrated, the identical structural design of Fd_CLOPA, TKV_c09620 and TKV_c16450 could explain the functionality across the species. Nevertheless, there must be a difference in electron transfer mechanism of TKV_c09620 and TKV_c16450 regarding the number of their [4Fe4S] clusters. All [4Fe4S]clusters can transfer only one electron at a time. Probably, the intended target for ferredoxins with two or more [4Fe4S] clusters requires a quick succession of two electrons, as shown, for example, for FNR [52], but there is no reason why TKV_c09620 could not also transfer a single electron in two independent steps. In contrast, TKV_c16450 is an one-electron transferring protein. It is believed that among [FeS] types, [4Fe4S] clusters were the first that have evolved at the early stage of life [53,54]. Studies indicated that the [4Fe-4S] cluster is sensitive to oxygen compared to [2Fe2S] clusters [55] and thus, it is predicted that after the Great Oxidation Event, organisms preferred [2Fe2S] clusters due to their oxygen tolerance [55].
Interestingly, many acetogenic genomes carry and express several genes for different ferredoxins [56].
There are three plausible arguments why acetogenic bacteria may utilize many copies of ferredoxins. First, primordial acetogenic bacteria colonized iron-rich environments (hydrothermal vents) [57][58][59] which lead to the evolution of several different iron-containing ferredoxin copies. Different copies of ferredoxins may have different functions or specificities in vivo [25]. This allows bacteria to use ferredoxins in a variety of metabolic pathways and environmental conditions. Second, acetogens need to maintain a balanced redox state within their cells to support their metabolic processes [17,60]. Multiple copies of ferredoxins can help to maintain this balance by providing a range of electron transfer capabilities. Third, having multiple copies of ferredoxins provides a backup system in case one copy is mutated. This redundancy can be important for the survival of the bacterium in changing or stressful environments [57,58].
Purification of His-TKV_c09620, His-TKV_c16450, His-TKV_c10420 and His-TKV_c19530 All purification steps were performed under strictly anoxic conditions at room temperature in an anoxic chamber (Coy Laboratory Products, Grass Lake, MI, USA) filled with 95-98% N 2 and 2-5% H 2 . All buffers used were prepared using the anaerobic techniques described previously [61,62].

Spectral analyses
Spectral analyses were performed under oxic or anoxic conditions at room temperature in 1.8 mL of cuvettes (Glasger€ atebau Ochs GmbH, Bovenden, Niedersachsen, Germany). Under oxic conditions, the assay mixture contained 30 lM TKV_c09620, TKV_c16450, TKV_c10420 or TKV_c19530 in buffer B. To reduce the proteins, 2 mM sodium dithionite was added. For the determination of the extinction coefficient, the ferredoxin concentrations ranged between 0 and 40 lM respectively. Under anoxic conditions, the assay mixture contained buffer C and 30 lM His-TKV_c09620, His-TKV_c16450, His-TKV_c10420 or His-TKV_c19530 respectively. To reduce the proteins, 10 lg CODH/ACS and a CO atmosphere (2 9 10 5 Pa) were added [23]. All measurements were carried out in a spectral photometer Specord S600 (Analytik Jena, Jena, Germany).

Electrochemical measurements and bioelectrode preparation
All electrochemical measurements were performed using a Metrohm Autolab potentiostat (PGSTAT101, Metrohm, Herisau, Appenzell Ausserrhoden, Switzerland) wired into an Ar-filled glovebox (O 2 < 0.8 ppm, Jacomex, Dagneux, Auvergne-Rhone-Alpes, France) with a potential window of À0.3 to À1 V vs. SCE, using a three-electrode configuration comprising a standard calomel electrode (SCE) as the reference electrode, a platinum-wire (Pt) as the counter electrode and a graphite rod working electrode (⌀ = 3.05 mm). All potentials were converted to the standard hydrogen electrode (SHE) by the relationship: E SHE = E SCE + 0.242 V [66]. Graphite rod working electrodes were fabricated by cutting graphite rods (Alfa Aesar product 40765) into small~5 cm length pencil electrodes and coating the carbon rod with heat-shrink tubing. Finally, the end of the electrode was cut straight using a blade and the resulting 3.05 mm electrode end was polished with abrasive paper and sonicated in MilliQ water (18.2 MΩ cm) before use. All electrochemical measurements were performed in 25 mM phosphate buffer (pH 7.5) containing 0.1 M NaCl and either 4 mM neomycin sulfate or 50 mM MgCl 2 to aid with the absorption of the protein on the working electrode surface [67]. All measurements were performed at 22 AE 1°C. All solutions were thoroughly deoxygenated within the Ar-filled glovebox before use. Bioelectrodes were prepared by drop-coating 5 lL of each ferredoxin stock solution (TKV_c09620 = 300 mM, TKV_c16450 = 200 mM, TKV_c10420 = 171 mM, and TKV_c19530 = 417 mM) onto the surfaces of individual graphite rod electrodes, which were allowed to dry for 10 min prior to electrochemical evaluation.