Simultaneous Formate and Syngas Conversion Boosts Growth and Product Formation by Clostridium ragsdalei

Electrocatalytic CO2 reduction to CO and formate can be coupled to gas fermentation with anaerobic microorganisms. In combination with a competing hydrogen evolution reaction in the cathode in aqueous medium, the in situ, electrocatalytic produced syngas components can be converted by an acetogenic bacterium, such as Clostridium ragsdalei, into acetate, ethanol, and 2,3-butanediol. In order to study the simultaneous conversion of CO, CO2, and formate together with H2 with C. ragsdalei, fed-batch processes were conducted with continuous gassing using a fully controlled stirred tank bioreactor. Formate was added continuously, and various initial CO partial pressures (pCO0) were applied. C. ragsdalei utilized CO as the favored substrate for growth and product formation, but below a partial pressure of 30 mbar CO in the bioreactor, a simultaneous CO2/H2 conversion was observed. Formate supplementation enabled 20–50% higher growth rates independent of the partial pressure of CO and improved the acetate and 2,3-butanediol production. Finally, the reaction conditions were identified, allowing the parallel CO, CO2, formate, and H2 consumption with C. ragsdalei at a limiting CO partial pressure below 30 mbar, pH 5.5, n = 1200 min−1, and T = 32 °C. Thus, improved carbon and electron conversion is possible to establish efficient and sustainable processes with acetogenic bacteria, as shown in the example of C. ragsdalei.


Introduction
The steady increase in greenhouse gas emissions in the atmosphere is a leading global issue that has received more attention and awareness in recent years [1][2][3].In regards to climate change, establishing a circular carbon economy with CO 2 emission reduction is an objective in the chemical industry since fossil raw material still serves for the production of carbonaceous chemicals [4,5].Syngas fermentation using acetogenic bacteria can be a biotechnological alternative to substitute fossil carbon-based chemical synthesis processes [6,7].In this case, acetogenic microorganisms can use CO, CO 2 , and H 2 as carbon and electron sources for growth and product formation, such as organic acids, their corresponding alcohols, or other value-added organic chemicals.For example, acetate, ethanol, and 2,3-butanediol can be produced autotrophically by the genetically closely related acetogens Clostridium ljungdahlii [8,9], Clostridium autoethanogenum [10,11], and Clostridium ragsdalei [11,12].Ethanol production by C. autoethanogenum using syngas from steel mills has already been established at an industrial scale (https://lanzatech.com/,accessed on 18 April 2024).Other acetogens, for example, Clostridium carboxidivorans [13] or Butyribacterium methylotrophicum [14], are known to produce butyrate and butanol from syngas.
Medium-chain organic acids, like butyrate, hexanoate, and octanoate, can be produced by chain elongation with anaerobic chain-elongating bacteria, like Clostridium kluyveri, through the reverse β-oxidation [15,16] from the main products of the microbial syngas (WLP), which is the carbon fixation pathway in acetogenic bacteria (see Figure 1).As a first step, a formate dehydrogenase catalyzes the reduction of CO 2 to formate.Next, two sequential reactions are catalyzed by formate-tetrahydrofolate ligase (FTL) to synthesize formyl-THF under the consumption of one ATP.This reaction is followed by three further reduction reactions to methenyl-THF, methylene-THF, and finally to methyl-THF.Together with a carbonyl group delivered from CO 2 reduction in the carbonyl branch of the WLP, the main central intermediate acetyl-CoA is formed.Acetobacterium woodii is an example of growth with formate using this pathway [39,40].
Molecules 2024, 29, x FOR PEER REVIEW 3 of 19 The sequential reduction of formate to methylene-THF is also present in the methyl branch of the reductive acetyl-CoA pathway, also known as Wood-Ljungdahl pathway (WLP), which is the carbon fixation pathway in acetogenic bacteria (see Figure 1).As a first step, a formate dehydrogenase catalyzes the reduction of CO2 to formate.Next, two sequential reactions are catalyzed by formate-tetrahydrofolate ligase (FTL) to synthesize formyl-THF under the consumption of one ATP.This reaction is followed by three further reduction reactions to methenyl-THF, methylene-THF, and finally to methyl-THF.Together with a carbonyl group delivered from CO2 reduction in the carbonyl branch of the WLP, the main central intermediate acetyl-CoA is formed.Acetobacterium woodii is an example of growth with formate using this pathway [39,40].In gas fermentation, the growth and product formation rates of autotrophic acetogenic bacteria are often limited by the gas-liquid mass transfer limitations within the bioreactor.Therefore, acetogenic bacteria could utilize formate, which has a good water solubility, more easily as a carbon source since the low solubility of the gaseous substrates limits their availability [41].Furthermore, the reduction degree of carbon in formate is higher than that in CO 2 .Instead of spending reduction equivalents on the CO 2 reduction to formate in the methyl branch of the WLP, this reducing power could be saved by formate usage.The generation of the more reduced [CO] in the carbonyl branch from CO 2 reduction also costs considerable biochemical energy equivalents due to the very low redox potential of CO 2 /CO [42,43].Therefore, bacteria able to use CO as a sole carbon source usually prefer it over CO 2 /H 2 for thermodynamic reasons.Nevertheless, CO 2 /H 2 growth cultures can really profit from an additional formate or CO supply.It was shown that feeding small amounts of CO to C. autoethanogenum allowed a higher biomass concentration and improved bacterial fitness [44].In another study, A. woodi achieved higher biomass yields when formate or CO were consumed in addition to CO 2 /H 2 [39].
A. woodii has already proven its flexibility towards C1 substrate utilization, but other acetogenic bacteria are attractive in regards of more value-added natural products than acetate.In this case, the bacterium C. ragsdalei portrays as an opportunity to produce ethanol and 2,3-butanediol from syngas.It is known that the highest growth rates of C. ragsdalei are achieved at 37 • C and pH 6.3 [12].Providing additional sulfur by sulfide feeding resulted in the highest autotrophic ethanol and 2,3-butanediol production at pH 5.5 in stirred-tank bioreactors with continuous gassing [45].Furthermore, C. ragsdalei can reduce propionic and butyric acid to their corresponding alcohols 1-propanol and 1-butanol [46].
C. ragsdalei prefers CO as a sole carbon and electron source [12].The parallel utilization of CO, CO 2 , H 2 , and formate has not been studied to date.Therefore, in the first part of our study, the goal was to find the initial CO partial pressure (pCO 0 ) necessary to allow additional CO 2 /H 2 consumption by C. ragsdalei in continuously gassed stirred-tank batch processes.The pCO was lowered stepwise in successive individual batch processes.In regards to a bio-electrochemical application where CO and formate are produced at a cathode, we focused in the second part of our study on the simultaneous formate conversion as an additional carbon and electron source.Formate was added continuously at various pCO values to study the simultaneous conversion of the C1 substrates (CO, CO 2 , and formate) and H 2 by C. ragsdalei in fed-batch-operated stirred-tank bioreactors with continuous gassing.It should be answered how its consumption affects C. ragsdalei's growth behavior and how the product formation alters.

Studies on the Simultaneous Conversion of CO, CO 2 , and H 2 by C. ragsdalei
Since pH 5.5 and 32 • C supported ethanol and 2,3-butanediol formation by C. ragsdalei [45], these reaction conditions were applied for all batch processes in a stirred-tank bioreactor with continuous gassing.Starting from 600 mbar CO in the synthetic syngas, pCOs of 200 mbar and 50 mbar were tested and compared to 0 mbar CO in individual autotrophic batch processes, keeping the partial pressures of H 2 and CO 2 (pH 2 and pCO 2 , respectively) constant at 200 mbar by substituting CO with N 2 gas.The product formation kinetics measured in the individual batch processes are illustrated in Figure 2.
At the initial 50-600 mbar CO, C. ragsdalei started to utilize CO as the only carbon source (see Figure 2).When a pCO of 30 mbar and a cell dried weight (CDW) concentration of 0.25 g L −1 were reached after a process time of 1.2 d, a sudden increase in the H 2 uptake rate of up to 5 mM h −1 was observed (Figure 2).Meanwhile a maximal CO uptake rate of 8 mM h −1 was achieved and stayed constant for around 12 h, indicating CO limitation.Under excessive CO conditions (at 200 and 600 mbar), a higher CO uptake rate of 11 mM h −1 was measured without any hydrogen consumption (Figure 2).At the initial 50-600 mbar CO, C. ragsdalei started to utilize CO as the only carbon source (see Figure 2).When a pCO of 30 mbar and a cell dried weight (CDW) concentration of 0.25 g L −1 were reached after a process time of 1.2 d, a sudden increase in the H2 uptake rate of up to 5 mM h −1 was observed (Figure 2).Meanwhile a maximal CO uptake rate of 8 mM h −1 was achieved and stayed constant for around 12 h, indicating CO limitation.Under excessive CO conditions (at 200 and 600 mbar), a higher CO uptake rate of 11 mM h −1 was measured without any hydrogen consumption (Figure 2).
The growth of C. ragsdalei was independent of the initial pCO0 within the estimation error (Figure 2), but was very much improved compared to the reference process without CO resulting in a final CDW concentration of around 0.55 g L −1 compared to 0.18 g L −1 after 6 days.
Product formation varies strongly with the initial pCO0 (Figure 2): Acetate formation was the highest at a pCO0 of 50 mbar.In particular, during H2 uptake, the acetate production rate increased considerably at a process time of 1.2-1.8d.The formation of alcohols was the lowest at 50 mbar CO, with 1.5 g L −1 ethanol and 0.1 g L −1 2,3-butanediol, after 5 days.Only in the reference process without CO, fewer alcohols were formed (1 g L −1 ethanol and no 2,3-butanediol).The growth of C. ragsdalei was independent of the initial pCO 0 within the estimation error (Figure 2), but was very much improved compared to the reference process without CO resulting in a final CDW concentration of around 0.55 g L −1 compared to 0.18 g L −1 after 6 days.
Product formation varies strongly with the initial pCO 0 (Figure 2): Acetate formation was the highest at a pCO 0 of 50 mbar.In particular, during H 2 uptake, the acetate production rate increased considerably at a process time of 1.2-1.8d.The formation of alcohols was the lowest at 50 mbar CO, with 1.5 g L −1 ethanol and 0.1 g L −1 2,3-butanediol, after 5 days.Only in the reference process without CO, fewer alcohols were formed (1 g L −1 ethanol and no 2,3-butanediol).
At a pCO 0 of 200-600 mbar, the final 2,3-butanediol concentrations (0.63-0.98 g L −1 ) did not differ significantly, but variations were observed in the ethanol and acetate production.At the pCO 0 of 200 mbar, C. ragsdalei started to produce acetate, but after three days, the acetate concentration was kept constant.With final 4.2-4.8g L −1 ethanol and 2.1-3.08 g L −1 acetate, an ethanol-to-acetate ratio of 1.6-2.0g EtOH g acetate −1 was achieved.In comparison, the acetate production continued after three process days at a pCO 0 of Molecules 2024, 29, 2661 6 of 18 600 mbar to 5.7-6.2g L −1 acetate, whereas the final ethanol concentration was reduced to 2.7-4.0 g L −1 , resulting in an ethanol-to-acetate ratio of 0.5-0.6 g EtOH g acetate −1 .Due to the low gas solubility of CO and H 2 , a high stirrer speed of 1200 min −1 was applied to allow a high mass transfer from the gas to the liquid phase by the strong power input (not determined).Another option for higher dissolved gas concentrations is the application of overpressure in general.In industry, the gas fermentation process is usually conducted in 20-30 m high bubble column reactors, where through hydrostatic pressure, high gas partial pressures are achieved at the bottom.Since the lab-scale reaction vessel in this study was not designed to withstand overpressure, a high stirrer speed was chosen.
To check the possibility of enlarging the process time for simultaneous CO and H 2 /CO 2 conversion in autotrophic batch processes with C. ragsdalei, an additional batch process was performed at the pCO 0 of 50 mbar CO with a reduced stirrer speed (800 min −1 instead of 1200 min −1 ).Reducing the mechanical power input in the stirred-tank bioreactors resulted in reduced gas transfer rates and, thus, in reduced solved gas concentrations in the liquid phase at the same microbial gas consumption rates.The measured partial pressures of H 2 , CO 2 , and CO as well as the estimated gas uptake rates are compared in Figure 3 for the batch processes with the varying stirrer speeds.
At a pCO0 of 200-600 mbar, the final 2,3-butanediol concentrations (0.63-0.98 g L −1 ) did not differ significantly, but variations were observed in the ethanol and acetate production.At the pCO0 of 200 mbar, C. ragsdalei started to produce acetate, but after three days, the acetate concentration was kept constant.With final 4.2-4.8g L −1 ethanol and 2.1-3.08 g L −1 acetate, an ethanol-to-acetate ratio of 1.6-2.0gEtOH gacetate −1 was achieved.In comparison, the acetate production continued after three process days at a pCO0 of 600 mbar to 5.7-6.2g L −1 acetate, whereas the final ethanol concentration was reduced to 2.7-4.0 g L −1 , resulting in an ethanol-to-acetate ratio of 0.5-0.6 gEtOH gacetate −1 .
Due to the low gas solubility of CO and H2, a high stirrer speed of 1200 min −1 was applied to allow a high mass transfer from the gas to the liquid phase by the strong power input (not determined).Another option for higher dissolved gas concentrations is the application of overpressure in general.In industry, the gas fermentation process is usually conducted in 20-30 m high bubble column reactors, where through hydrostatic pressure, high gas partial pressures are achieved at the bottom.Since the lab-scale reaction vessel in this study was not designed to withstand overpressure, a high stirrer speed was chosen.
To check the possibility of enlarging the process time for simultaneous CO and H2/CO2 conversion in autotrophic batch processes with C. ragsdalei, an additional batch process was performed at the pCO0 of 50 mbar CO with a reduced stirrer speed (800 min −1 instead of 1200 min −1 ).Reducing the mechanical power input in the stirred-tank bioreactors resulted in reduced gas transfer rates and, thus, in reduced solved gas concentrations in the liquid phase at the same microbial gas consumption rates.The measured partial pressures of H2, CO2, and CO as well as the estimated gas uptake rates are compared in Figure 3 for the batch processes with the varying stirrer speeds.CO consumption rates are very similar, but hydrogen uptake was observed until the end of the batch process with 800 min −1 .The characteristic increase in the hydrogen uptake rate after 1.2 d was observed as well, but was less pronounced with a reduced mechanical power input (maximal H2 uptake rate of 4 mM h −1 compared to 5 mM h −1 ).In total, more CO consumption rates are very similar, but hydrogen uptake was observed until the end of the batch process with 800 min −1 .The characteristic increase in the hydrogen uptake rate after 1.2 d was observed as well, but was less pronounced with a reduced mechanical power input (maximal H 2 uptake rate of 4 mM h −1 compared to 5 mM h −1 ).In total, more hydrogen was consumed at a reduced stirrer speed, and consequently less CO 2 was produced by C. ragsdalei in the batch process with continuous gassing (Table 1).The reduced power input did not affect the biomass or product formation (see Table 1).A direct comparison of the measured product concentrations of these two processes is presented in the Supplementary Material (see Figure S1).From the first part of this study, two main conclusions can be drawn.For C. ragsdalei, the pCO 0 had a great influence on the product distribution in autotrophic batch processes with continuous gassing.In terms of the best ethanol-to-acetate ratio, a pCO 0 of 200 mbar CO was determined to be best, because either a higher pCO 0 (600 mbar) or a lower pCO 0 (50 mbar) lead to enhanced acetate formation.Further, limiting CO concentrations in the stirred tank bioreactor were necessary to allow additional CO 2 /H 2 uptake.Looking to the future bio-electrochemical applications where CO 2 is reduced to mainly CO and formate as by-products, the next part of our study focused on the parallel conversion of the gaseous substrates together with formate.

Studies on the Simultaneous Conversion of Syngas and Formate by C. ragsdalei
Continuous formate feeding of 1.2-1.4mM h −1 was applied at varying pCO 0 values in the syngas (50-600 mbar CO) in the stirred-tank reactors.The process performances were compared to the corresponding batch processes without formate supply (Figures 4-6).The displayed concentration data represent the mean values of two individual processes with continuous gassing at the same process conditions, with the exception of the processes at pCO 0 of 50 mbar CO.The gas uptake data are depicted from one of the duplicated processes with and without formate feeding.Due to the constant formate feeding rate, formate accumulated at the beginning at low cell densities.When a CDW concentration of about 0.05 g L −1 was reached, all the formate in the liquid medium was taken up by C. ragsdalei.Until the end of the fed-batch processes, no formate was detected by HPLC anymore.
At a pCO 0 of 50 mbar CO and a reduced stirrer speed of 800 min −1 , parallel formate, CO, CO 2 , and H 2 conversion was observed (Figure 4).Hydrogen consumption was initiated earlier at a process time of 0.6 days with additional formate feeding.CO consumption as well as the growth of C. ragsdalei and acetate formation were clearly improved with formate feeding, whereas no differences in process performances were measured with respect to microbial ethanol and 2,3-butanediol production.
Compared to the autotrophic batch process, additional formate feeding raised both the total amount of consumed CO from 401 mmol to 508 mmol and produced CO 2 from 179 mmol to 285 mmol.Carbon balances and electron balances were closed within the estimation error as before, with a carbon recovery of 104-105% and an electron recovery of 109-113% (Table 2).Compared to the autotrophic batch process, additional formate feeding raised the total amount of consumed CO from 401 mmol to 508 mmol and produced CO2 179 mmol to 285 mmol.Carbon balances and electron balances were closed withi estimation error as before, with a carbon recovery of 104-105% and an electron rec of 109-113% (Table 2).Table 2. Integral carbon and electron balances of CO-limited autotrophic batch processes with and without formate feeding at an initial 50 mbar CO with C. ragsdalei (batch-operated stirred-tank bioreactors with continuous gassing: 32 the same effects as before (Figure 5): increased CO consumption, increased growth rate of C. ragsdalei,, enhancedacetate formation and no differences in ethanol formation compared to the corresponding autotrophic batch process without formate feeding.In contrast to the autotrophic fed-batch process with 50 mbar CO shown before, the microbial 2,3-butanediol production was clearly improved with formate feeding (1.15 ± 0.02 g L −1 compared to 0.76 ± 0.13 g L −1 ).Compared to the autotrophic batch process at a pCO0 of 200 mbar CO, additional formate feeding raised again both the total amount of consumed CO from 710 ± 21 mmol to 990 ± 1 mmol and produced CO2 from 455 ± 18 mmol to 636 ± 3 mmol (Table 3).Carbon balances and electron balances were closed within the estimation error as before, with a carbon recovery of 95-97% and an electron recovery of 91-94%.Integrating hydrogen gas consumption over the whole process time resulted in an estimated total uptake of small Formate was continuously fed with 1.3 mM h −1 .All processes were reproduced.The mean with min-max values are shown for the concentration data, whereas the gas uptake rates r i are presented only for one of the duplicated processes.
Applying a higher pCO 0 of 200 mbar CO and continuous formate feeding showed the same effects as before (Figure 5): increased CO consumption, increased growth rate of C. ragsdalei, enhancedacetate formation and no differences in ethanol formation compared to the corresponding autotrophic batch process without formate feeding.In contrast to the autotrophic fed-batch process with 50 mbar CO shown before, the microbial 2,3-butanediol production was clearly improved with formate feeding (1.15 ± 0.02 g L −1 compared to 0.76 ± 0.13 g L −1 ).
Compared to the autotrophic batch process at a pCO 0 of 200 mbar CO, additional formate feeding raised again both the total amount of consumed CO from 710 ± 21 mmol to 990 ± 1 mmol and produced CO 2 from 455 ± 18 mmol to 636 ± 3 mmol (Table 3).Carbon balances and electron balances were closed within the estimation error as before, with a carbon recovery of 95-97% and an electron recovery of 91-94%.Integrating hydrogen gas consumption over the whole process time resulted in an estimated total uptake of small amounts of 94-136 mmol H 2 within 6 days, whereas no net H 2 uptake was observed in the batch process without formate feeding.Summarizing the results from these investigations, simultaneous formate consumption is possible in stirred-tank bioreactors with continuous syngas gassing at the rates studied independently of the pCO0 tested.At the lower pCO0 of 50-200 mbar CO, formate addition boosted growth and product formation with C. ragsdalei, whereas no significant effects were observed with 600 mbar CO in the syngas (Figure 7).In contrast to the data shown before, no significant differences in process performances were observed at a pCO 0 of 600 mbar CO with and without formate feeding (Figure 6).Though the mean of the concentration data shows slightly increased product concentrations, the min-max values of the duplicated processes were two high to indicate significance.Carbon balances and electron balances were closed within the estimation error as before, with a carbon recovery of 97-103% and an electron recovery of 106-110% (Table 3).The share of carbon fed as formate and consumed by C. ragsdalei in relation to total carbon conversion was 12-16% at 600 mbar CO.As this was not significantly different from the respective carbon share of 11-13% at a pCO 0 of 200 mbar, we assumed that the improved availability of CO for C. ragsdalei cells at 600 mbar CO may be the reason for the nonsignificant effects of formate feeding with 1.3 mM h −1 .
Summarizing the results from these investigations, simultaneous formate consumption is possible in stirred-tank bioreactors with continuous syngas gassing at the rates studied independently of the pCO 0 tested.At the lower pCO 0 of 50-200 mbar CO, formate addition boosted growth and product formation with C. ragsdalei, whereas no significant effects were observed with 600 mbar CO in the syngas (Figure 7).

A Limiting pCO Induced Simultaneous Conversion of CO, CO2, and H2 by C. ragsdalei
With the aim of complete carbon conversion, the parallel consumption of all components of a synthesis gas mixture is crucial.Therefore, autotrophic batch processes were conducted with C. ragsdalei to identify the process conditions for the simultaneous conversion of CO, CO2, and H2 by varying the pCO0 in the inlet gas of a stirred-tank bioreactor with continuous gassing.
When CO was limited at around 30 mbar in the reactor, additional CO2 and H2 uptake

A Limiting pCO Induced Simultaneous Conversion of CO, CO 2 , and H 2 by C. ragsdalei
With the aim of complete carbon conversion, the parallel consumption of all components of a synthesis gas mixture is crucial.Therefore, autotrophic batch processes were conducted with C. ragsdalei to identify the process conditions for the simultaneous conversion of CO, CO 2 , and H 2 by varying the pCO 0 in the inlet gas of a stirred-tank bioreactor with continuous gassing.
When CO was limited at around 30 mbar in the reactor, additional CO 2 and H 2 uptake satisfied the demand for gaseous carbon and electrons to achieve the maximal biomass concentration in batch processes.The CO limitation was reflected by the reduced CO uptake rate in comparison to the rates achieved at 200 and 600 mbar CO, and the rate was constant for half a day during CO 2 /H 2 uptake.C. ragsdalei is known to prefer CO as a carbon and electron source over CO 2 /H 2 [12].In addition, CO using acetogenic microorganisms usually favor CO over CO 2 /H 2 as a carbon and electron source for thermodynamic reasons [47].Missing Fd red as a reducing equivalent that is mainly generated from the CO to CO 2 oxidation (see Figure 1) causes ATP limitation since the ATPase activity is dependent on the Fd red pool [48].Besides, a strong acetate production decouples the proton motive force, resulting in additional ATP costs [49].Coincidently with H 2 uptake, the acetate production was raised to produce more ATP, but missing reducing equivalents restricted the further conversion to ethanol and the production of 2,3-butanediol.
In the study of Heffernan et al. (2020), CO was supplemented in a CO 2 /H 2 grown C. autoethanogenum culture with a share of 2% that was lower than that in our study (5%), and it was shown that the CO supplementation boosted the biomass production, enabled a higher ethanol-to-acetate conversion, and demonstrated parallel CO and CO 2 conversion [44].In C. autoethanogenum, the high ATP demand was satisfied by an increased CO oxidation to elevate the Fd red and consequently the ATP pool [49].For C. ragsdalei under limiting CO conditions, this mechanism could not be employed, resulting in an aggravated ethanol-to-acetate ratio.However, the simultaneous conversion of CO, CO 2 , and H 2 by C. ragsdalei was proven in this study.

The pCO Affected the Product Distribution
The batch studies performed with C. ragsdalei in a fully controlled stirred-tank bioreactor with continuous gassing showed that the acetate, ethanol, and 2,3-butanediol production rates of this bacterium were highly affected by the pCO value (Figure 7).
C. ragsdalei grew very slowly without CO, but this might be improved by a pH increase, since its optimum value for growth is pH 6.3 [12], and throughout this study, a pH of 5.5 was applied.Product formation from CO 2 /H 2 did not enable 2,3-butanediol production, but the ethanol-to-acetate ratio (0.7 g EtOH g acetate −1 ) was higher than in the process conducted at 50 mbar CO (0.2 g EtOH g acetate −1 ), where C. ragsdalei grew solely with CO first and started to take up CO 2 /H 2 later.Comparing the growth without CO to C. ljungdahlii, like in the study from Hermann et al. (2020), mainly biomass and acetate were produced at pH 5.9 as well, and ethanol production even started decoupled from growth only at the end of the batch process [50].
When C. ragsdalei grew solely with CO, a pCO of 200 mbar resulted in the highest ethanol-to-acetate ratio, as a direct conversion of acetate into ethanol was observed.This is possible via acetaldehyde by the aldehyde ferredoxin oxidoreductase (AOR, see Figure 1).That ethanol is rather synthesized via this route and not directly from acetyl-CoA was also reported for C. autoethanogenum [51] and for C. ljungdahlii [52] as overflow metabolism, in the latter case explained by an excess of acetyl-CoA and reduction equivalents, such as ferredoxin, which cannot be used for biomass formation but are forwarded to alcohol production.For C. ragsdalei, the reduction from pH 6.0 to pH 5.5 confirmed the beneficial influence on ethanol and 2,3-butanediol production [45], but in our study, an optimal pCO was found to maximize the alcohol production.It should be mentioned that, in the study from Oliveira et al. [45], the acetate-to-ethanol conversion was observed at a pCO of 600 mbar with a reduced stirrer speed (800 min −1 instead of 1200 min −1 in our study), affecting therefore the real (not measured) dissolved CO concentration in the liquid phase.
Remarkably, at 600 mbar CO, little hydrogen production was observed throughout the process, indicating a disposal of excess reducing equivalents instead of redirecting the reducing power towards ethanol and 2,3-butanediol in terms of redox homeostasis.
3.3.Parallel Utilization of Formate, CO, CO 2 , and H 2 Formate as a stable C1 product from electrocatalytical CO 2 reduction gained considerable attention.Acetogenic bacteria natively dispose one of the most straightforward carbon fixation pathways, the reductive Acetyl-CoA pathway, which is energetically very efficient at the same time and allows the direct utilization of formate as a further C1 carbon and electron source.Therefore, electrocatalytical formate formation was simulated physically by a continuous formate feed.A parallel CO and CO 2 /H 2 uptake had been already demonstrated at an initial pCO 0 of 50 mbar, but a parallel formate consumption at various pCO (50,200, and 600 mbar) was the next research question to be investigated at moderate formate supply rates (1.3-1.4 mM h −1 ).
The simultaneous conversion of CO, CO 2 , H 2 , and formate was observed only at a pCO 0 of 50 mbar.The formate supply did not affect the H 2 uptake, but an increased total CO uptake was measured.Carbon recoveries proved that formate was utilized as an additional carbon and electron source and did not substitute part of the CO or CO 2 /H 2 .An increased flux through the methyl branch in the WLP might explain the enhanced CO uptake to deliver the needed carbonyl group to produce 1 acetyl-CoA (see Figure 1).The enhanced flux enabled a higher growth rate by 20-50%, independent on the initial pCO 0 .A deficient alcohol production under CO limitation was shown to be one characteristic of C. ragsdalei, and additional formate feeding increased the synthesis of the less attractive fermentation product acetate and worsened the 2,3-butanediol production.However, a stronger acetate production was observed independent of the incoming pCO 0 .The enhanced ATP demand of the ATP-consuming formate fixation step in the methyl branch was compensated by acetate production, impairing the ethanol-to-acetate ratio (see Figure 1).
Remarkably, at a pCO 0 of 200 mbar, 2,3-butanediol production was boosted by formate feeding (Figure 5).Considering the carbon in formate, the carbon enters the methyl branch in a more reduced state compared to CO 2 .Accordingly, C. ragsdalei can save reduction equivalents (2H + + 2e − ) through preventing the required reduction of CO 2 to formate by the formate dehydrogenase, in theory.If ethanol is mainly synthesized through the acetate route, these energy savings were useful in the production of 2,3-butanediol, which needs 2 NADH as reducing equivalents (see Figure 1).But the improved 2,3-butanediol synthesis was dependent on a high CO availability since at limiting CO concentrations, C. ragsdalei mainly synthesized acetate.
Among different microorganisms, only few can utilize formate.For example, Acetobacterium woodii is known for its good CO 2 /H 2 utilization as the main carbon and electron sources, but can be very flexible towards autotrophic and heterotrophic substrates [39].An A. woodii culture, grown with CO 2 /H 2 , was supplemented with formate.Although the reduced gas uptake rates indicated that formate substituted a part of the total consumed carbon, higher biomass yields could be achieved [39].Interestingly, independent of the used carbon and electron sources, the final acetate yields based on the total carbon consumed did not differ significantly between the A. woodi cultures grown with CO 2 /H 2 or CO 2 /H 2 /formate or CO 2 /H 2 /formate/CO, showing that the additional energy delivered by CO was rather used for biomass production.In the end, the co-utilization of CO, formate, and CO 2 /H 2 by A. woodii resulted in higher biomass and acetate concentrations, similar to C. ragsdalei, although A. woodi is not an alcohol producer.

Microorganism and Media
The acetogen Clostridium ragsdalei, obtained from the German Collection of Microorganism and Cell Cultures GmbH (DSM No. 15248, DSMZ GmbH, Braunschweig, Germany) as a freeze-dried sample, was grown in a fermentation medium and then stored at −80 • C with 10% glycerol (v/v).The fermentation medium for pre-cultivation and batch processes in the stirred-tank bioreactor was composed according to Doll et al. [53], and the ingredients are listed in the Supplementary Material (Table S1).For pre-cultivation, 15 g L −1 2-(Nmorpholino)ethanesulfonic acid (MES) was added as a buffer.A total of 2 L of the freshly prepared fermentation medium for pre-cultivation was boiled for 20 min and strongly purged with N 2 for at least 20 min for anaerobization.100 mL of the medium was aliquoted to 0.5 L anaerobic bottles in an anaerobic glove box and then autoclaved.Chemical stocks (fructose, cysteine-hydrochloride, and PBS buffer) were made with anaerobic water and autoclaved for sterilization.

Pre-Cultivation and Inoculum Preparation
The pre-cultures were incubated at 37 • C and 100 rpm (WiseCube WIS-20, Witeg Labortechnik GmbH, Wertheim, Germany).The first heterotrophic pre-culture (7 g L −1 fructose) grew over 70 h and was used to inoculate a second autotrophic pre-culture with 10 mL (gas mixture: 1.2 bar CO, 0.4 bar CO 2 , and 0.4 bar H 2 ).The second pre-culture contained 0.4 g L −1 of cysteine hydrochloride as a reducing agent and served as the inoculum for the batch process.The stirred-tank bioreactor was always inoculated in a way to achieve a starting CDW concentration of around 0.03 g L −1 .For the main inoculum, the second grown culture was harvested by centrifugation at 3600 rcf (Rotixa 50 RS, Hettich GmbH & Co., KG, Tuttlingen, Germany), and the cell pellet was resuspended in 10 mL of anaerobic PBS buffer.

Continuously Gassed Stirred-Tank Bioreactor
Batch processes were performed in a stirred-tank bioreactor.The cylindrical reaction vessel was designed previously for bio-electrochemical applications [27].The processes were conducted with a 1 L working volume under fully controlled conditions.The temperature and pH were adjusted to 32 • C and pH 5.5, respectively, according to Oliveira et al. [45], using a control station (Labfors 2, Infors AG, Bottmingen, Switzerland).The pH was measured by a sterizable sensor (405-DPAS-SC-K8s/120, Mettler Toledo, Giesen, Germany).The bioreactor was continuously gassed with a defined gas mixture at ambient pressure.Each gas volume flow was controlled by an independent mass flow controller (F-201CV-500 RGD-33-V, Bronhorst High Tech B.V., Ruurlo, the Netherlands) and mixed before reaching the inlet of the bioreactor.The gas composition always comprised 20% CO 2 and 20% H 2 .The CO share was varied in the range of 5-60%, whereas at lower partial pressures than 600 mbar CO, the mixture was complemented with N 2 to achieve a constant gas volume flow of 5 NL h −1 .The stirred-tank bioreactor was equipped with baffles and two Rushton turbines.A stirrer speed of n = 1200 min −1 was applied.Both the bioreactor and fermentation medium without vitamins were autoclaved separately.The medium was transferred to the bioreactor under sterile conditions.Sterile filtrated vitamins (0.2 µm pore size) were added to the medium that was anaerobized with 5 NL h −1 N 2 for around 2 h.Afterwards, the bioreactor was calibrated with the defined gas mixture over night.Before the inoculation with C. ragsdalei, 0.4 g of L −1 cysteine hydrochloride was added aseptically as a reducing agent to generate a redox potential below −100 mV.
To simulate electrocatalytical formate production physically, a continuous formate feed was applied using a 2.5 M sodium formate solution (pH 5.5) with a feed rate of 0.5 mL h −1 .The feed was started concurrently with the inoculation.The weight of the stock solution was also monitored and allowed for the determination of the real applied formate feed rate by linear regression at the end.Due to sampling and the very slow feeding rate, dilution was neglectable.

Analytical Methods
During the batch processes, samples were collected regularly from the bioreactor and analyzed for biomass and product concentrations.The bacterial growth was monitored through optical density measurements at 600 nm (OD 600 ).Afterwards, a previously determined correlation between the OD 600 and CDW concentration was used to calculate the CDW for the corresponding measurements.The samples were filtered (0.2 µm size exclusion), and the supernatant was analyzed for organic acids and alcohols by HPLC (LC2030C-Plus, Shimadzu, Kyoto, Japan) using a carbohydrate analysis column (Aminex HPX-87H, Biorad, Munich, Germany) at 60 • C. The analytes were eluted isocratically using 5 mM of H 2 SO 4 as the mobile phase at a flow rate of 0.6 mL min −1 .Detection was achieved by a refractive index detector (RID-20A, Shimadzu, Kyoto, Japan), and an external standard with defined analyte concentrations allowed for their quantification.The lowest concentrations in the calibration curves were 0.1 g L −1 of ethanol, 0.04 g L −1 of formate, 0.2 g L −1 of acetate, and 0.04 g L −1 of 2,3-butanediol.Through sigmoidal model fittings by non-linear regression using the Gompertz curve [54], time-dependent product formation rates were estimated.To calculate the maximal specific growth rate, the CDW concentration curve was fitted to the logistic growth function by Richards F.J. [55].The derivative of the model was related to the initial CDW concentration to calculate the growth rate at each time point.
The off-gas of the bioreactor was continuously measured by a mass flow meter (F-111B-500-RGD-33-E, Bronkhorst High Tech B.V., Ruurlo, The Netherlands) and analyzed online by gaschromatography every 10 min.The micro-gaschromatograph (590 Micro GC, Agilent Technologies, Waldbronn, Germany) was equipped with a COX column for the retention of H 2 , CO, and CO 2 , which were detected by thermal conductivity.External calibration with the defined gas mixtures allowed for the determination of the respective partial pressures or gas shares in the exhaust gas.N 2 could not be measured and was calculated by subtracting the determined partial pressures from the absolute pressure.The off-gas mass flow was corrected by the mixing gas conversion factor that included the gas shares and the individual linear calibration parameters of each analyzed gas.Then, the corrected volume flow was multiplied by the gas shares to estimate the individual gas volume flows for each GC measurement.Subtracting the measured volume flow from the initial volume flow allowed for the calculation of the gas uptake rates.Further integration over time, using a 10 min increment, delivered the total gas consumption in mmol.

Carbon and Electron Recoveries
The measured final product concentrations in [g L −1 ] were converted to [mmol C L −1 ].The total amount of products (CDW, ethanol, acetate, 2,3-butanediol, and CO 2 ) was then related to the total consumed carbon (CO, yeast extract, and formate) to calculate the carbon recoveries.For determining the electron recovery, the respective reduction degrees of each molecule were used and H 2 was also considered.

Conclusions
In electrochemical systems, the main goal is to achieve high faradaic and energy efficiencies.Biotransformation processes using microorganisms can allow the direct conversion of electrochemical products into valuable molecules.In case of the electrocatalytical CO 2 reduction reaction (CO 2 RR) in a bioreactor, formate can be synthesized as a by-product beside CO, and H 2 gas usually evolves from the competing proton reduction reaction in the aqueous system.Depending on the applied cell potential, the electrochemical product distribution can vary.The evidence of simultaneous CO, CO 2 , H 2 , and formate utilization by C. ragsdalei was provided.In regards of a bio-electrocatalytical system in which all these C1 substrates and H 2 are accessible in various stoichiometries, a parallel conversion will be necessary for high carbon and energy efficiency.Using electric energy, generated by solar panels or windmills, and CO 2 derived from fossil fuels, a sustainable process can be developed in the future.In this case, C. ragsdalei is a candidate for bioelectrocatalytical acetate and ethanol production from CO, formate, CO 2 , and H 2 .In industrial-scale gas fermentation without electrochemical application, the process with C. ragsdalei can be conducted in high bubble column reactors.At the bottom of the column, a high CO partial pressure enables a high CO availability, whereas at the top, a low CO concentration will allow parallel CO, CO 2 , and H 2 conversion.Co-cultivation with chain-elongating anaerobic microorganisms, like C. kluyveri, in a bio-electrochemical system may allow the assimilation of acetate and ethanol to produce longer-chained organic acids and alcohols and should be investigated in future studies.

Figure 2 .
Figure 2. Autotrophic product formation by C. ragsdalei at various CO partial pressures in the incoming gas (pCO 0 ).Batch processes were conducted in a fully controlled stirred-tank bioreactor (T = 32 • C, pH 5.5, n = 1200 min −1 ).The bioreactor was continuously gassed with 5 L h −1 of a defined gas mixture composed of 200 mbar CO 2 , 200 mbar H 2 , and 600 mbar CO/N 2 mixtures (50 mbar CO complemented with 550 mbar N 2 , 200 mbar CO complemented with 400 mbar N 2 , and 600 mbar CO , respectively).The reference process without CO was gassed with 200 mbar CO 2 and 800 mbar H 2 .

Figure 4 .
Figure 4. Autotrophic formate conversion by C. ragsdalei at the pCO 0 of 50 mbar CO with and without formate feeding.Processes were conducted in a fully controlled stirred-tank bioreactor (T = 32 • C, pH 5.5, n = 800 min −1 ).The bioreactor was continuously gassed with 5 L h −1 of a defined gas mixture composed of 200 mbar CO 2 , 200 mbar H 2 , 50 mbar CO, and 550 mbar N 2 .Formate was continuously fed with 1.4 mM h −1 .

Figure 5 .
Figure 5. Autotrophic formate conversion by C. ragsdalei at the pCO0 of 200 mbar CO with and without formate feeding.Processes were conducted in a fully controlled stirred-tank bioreactor (T = 32 °C, pH 5.5, n = 1200 min −1 ).The bioreactor was continuously gassed with 5 L h −1 of a defined gas mixture composed of 200 mbar CO2, 200 mbar H2, 200 mbar CO, and 400 mbar N2.Formate was continuously fed with 1.3 mM h −1 .All processes were reproduced.The mean with min-max values are shown for the concentration data, whereas the gas uptake rates ri are presented only for one of the duplicated processes.

Figure 5 .
Figure 5. Autotrophic formate conversion by C. ragsdalei at the pCO 0 of 200 mbar CO with and without formate feeding.Processes were conducted in a fully controlled stirred-tank bioreactor (T = 32 • C, pH 5.5, n = 1200 min −1 ).The bioreactor was continuously gassed with 5 L h −1 of a defined gas mixture composed of 200 mbar CO 2 , 200 mbar H 2 , 200 mbar CO, and 400 mbar N 2 .Formate was continuously fed with 1.3 mM h −1 .All processes were reproduced.The mean with min-max values are shown for the concentration data, whereas the gas uptake rates r i are presented only for one of the duplicated processes.

Figure 6 .
Figure 6.Autotrophic formate conversion by C. ragsdalei at the pCO0 of 600 mbar CO with and without formate feeding.Processes were conducted in a fully controlled stirred-tank bioreactor (T = 32 °C, pH 5.5, n = 1200 min −1 ).The bioreactor was continuously gassed with 5 L h −1 of a defined gas mixture composed of 200 mbar CO2, 200 mbar H2, and 600 mbar CO.Formate was continuously fed with 1.3 mM h −1 .All processes were reproduced.The mean with min-max values are shown for the concentration data, whereas the gas uptake rates ri are presented only for one of the duplicated processes.

Figure 6 .
Figure 6.Autotrophic formate conversion by C. ragsdalei at the pCO 0 of 600 mbar CO with and without formate feeding.Processes were conducted in a fully controlled stirred-tank bioreactor (T = 32 • C, pH 5.5, n = 1200 min −1 ).The bioreactor was continuously gassed with 5 L h −1 of a defined gas mixture composed of 200 mbar CO 2 , 200 mbar H 2 , and 600 mbar CO.Formate was continuously fed with 1.3 mM h −1 .All processes were reproduced.The mean with min-max values are shown for the concentration data, whereas the gas uptake rates r i are presented only for one of the duplicated processes.

Figure 7 .
Figure 7. Relative effect of formate feeding on CDWmax, μmax, the final ethanol-to-acetate ratio (EtOH/Acfinal), and the relative final product concentrations (acetate, ethanol, and 2,3-butanediol) at the initial 50, 200, and 600 mbar CO.Processes were conducted in a fully controlled stirred-tank bioreactor (T = 32 °C, pH 5.5, n = 1200 min −1 ).The formate feed rate was 1.4 mM h −1 at a pCO0 of 50 mbar and 1.3 mM h −1 at 200 and 600 mbar CO.The black lines present the data without formate feeding at the respective pCOs.

Figure 7 .
Figure 7. Relative effect of formate feeding on CDW max , µ max , the final ethanol-to-acetate ratio (EtOH/Ac final ), and the relative final product concentrations (acetate, ethanol, and 2,3-butanediol) at the initial 50, 200, and 600 mbar CO.Processes were conducted in a fully controlled stirred-tank bioreactor (T = 32 • C, pH 5.5, n = 1200 min −1 ).The formate feed rate was 1.4 mM h −1 at a pCO 0 of 50 mbar and 1.3 mM h −1 at 200 and 600 mbar CO.The black lines present the data without formate feeding at the respective pCOs.

Table 3 .
Integral carbon and electron balances of autotrophic batch processes with and without formate feeding at the initial 200 and 600 mbar CO with C. ragsdalei.Batch-operated stirred-tank bioreactors with continuous gassing: 32 • C, pH 5.5, and 5 NL h −1 syngas with 200 mbar CO 2 , 200 mbar H 2 , 200 (or 600) mbar CO, and 400 (or 0) mbar N 2 .The carbon and electron balances were calculated for each individual process until t = 4.4 d and allowed the presentation of minimal and maximal values.If a data point was not available, the value was calculated using a sigmoidal model (see the Methods Section).