**Understanding (***R***) Specific Carbonyl Reductase from** *Candida parapsilosis* **ATCC 7330 [CpCR]: Substrate Scope, Kinetic Studies and the Role of Zinc**

## **Vinay Kumar Karanam 1, Debayan Chaudhury <sup>1</sup> and Anju Chadha 1,2,\***


Received: 26 July 2019; Accepted: 18 August 2019; Published: 21 August 2019

**Abstract:** CpCR, an (*R*) specific carbonyl reductase, so named because it gave (*R*)-alcohols on asymmetric reduction of ketones and ketoesters, is a recombinantly expressed enzyme from *Candida parapsilosis* ATCC 7330. It turns out to be a better aldehyde reductase and catalyses cofactor (NADPH) specific reduction of aliphatic and aromatic aldehydes. Kinetics studies against benzaldehyde and 2,4-dichlorobenzaldehyde show that the enzyme affinity and rate of reaction change significantly upon substitution on the benzene ring of benzaldehyde. CpCR, an MDR (medium chain reductase/dehydrogenase) containing both structural and catalytic Zn atoms, exists as a dimer, unlike the (*S*) specific reductase (SRED) from the same yeast which can exist in both dimeric and tetrameric forms. Divalent metal salts inhibit the enzyme even at nanomolar concentrations. EDTA chelation decreases CpCR activity. However, chelation done after the enzyme is pre-incubated with the NADPH retains most of the activity implying that Zn removal is largely prevented by the formation of the enzyme-cofactor complex.

**Keywords:** MDR—medium-chain reductase/dehydrogenase; ADH—alcohol dehydrogenase; enzyme kinetics; EDTA (Ethylenediaminetetraacetic acid) chelation; ultrafiltration

## **1. Introduction**

The MDR superfamily is a part of the oxidoreductase class and contains a family of zinc-dependent alcohol dehydrogenases [1]. MDRs are hypothesized to have evolved from SDR (short-chain reductases/dehydrogenases) superfamily and later incorporated zinc atoms within themselves to facilitate divergence in catalytic abilities [2]. CpCR belongs to MDR superfamily and is reported to catalyse reductions of ketoesters, ketones and aldehydes leading to the production of some important pharmaceutical precursors [3]. It is one of the important enzymes present in *Candida parapsilosis* ATCC 7330, which is a well-known whole-cell biocatalyst [4]. CpCR, a heterodimer (PDB: 4OAQ), has two different Zn atoms viz. catalytic Zn and the structural Zn. The former is coordinated to two Cys, His and a water while the latter is coordinated to four Cys residues and lies away from the active site. Aldehyde reduction by various ADHs from horse liver, human liver and *Saccharomyces* sp. is well established [5–7]. CpCR reduces aliphatic and aromatic aldehydes with higher activity compared to other carbonyl substrates.

Even though a lot of literature on understanding the role of Zn in MDR superfamily exists [8–17], still there is some ambiguity in the function of structural Zn [8,9,16,17]. Chelation studies with multi-dentate ligands, like EDTA and 1,10-phenanthroline on ADHs, indicate that they significantly affect the activity by chelating one of the Zn atoms [9,17,18]. Dithiothreitol (DTT) at higher concentrations is known to cause heat lability of yeast ADH (YADH) by changing the Zn stoichiometry in the enzyme [9]. Cofactor binding to the liver ADH (LADH) induces a large conformational change where the two domains (catalytic and cofactor binding domains) rotate around 10 degrees to close the active site cleft [19]. A similar observation was made in alcohol dehydrogenase from *Arabidopsis thaliana* but the mechanism is different from that of LADH [20]. It is also established that the conformational change induced by cofactor binding requires the presence of the nicotinamide part of NAD(P)H, while the binding of ADP-ribose does not induce such a change [21]. Recently, cofactor binding to various ADHs was studied using circular dichroism wherein the orientation of nicotinamide ring of the cofactor at the active site could be observed [22]. Another study on cofactor binding shows that NAD<sup>+</sup> and NADH adopt different structures in water, but both fit in the enzyme's active site in a semi-extended conformation [23]. These studies are essential in understanding the initial step (binding of the cofactor to the active site) of the reactions catalysed by NAD(P)H-dependent ADHs. Cofactor switching is also an important aspect in obtaining enzymes with better catalytic ability and applications in metabolic engineering [24–27]. However, the effect of this changed configuration upon cofactor binding on enzyme activity has not been probed systematically to date. This is of importance because in nature most enzymes exist bound to their natural cofactor as evidenced by typically low Kd values of the cofactor [22].

In this study we used various concentrations of EDTA for chelation studies against CpCR and employed ultrafiltration for rapid removal of EDTA. To the best of our knowledge this is the first report to elucidate the kinetic characteristics of cofactor-enzyme complex.

## **2. Results and Discussion**

## *2.1. Purification of CpCR, Expanding Its Substrate Scope and Kinetic Studies*

The purification protocol was modified, keeping in mind the yield and the stability of the enzyme. Compared to the previous protocol [28], the modified protocol increased the yield ten times and fold purification by 3.7 times. CpCR, an MDR, is a Zn-containing enzyme and it is very important that the Zn coordination stays unaffected by the buffer conditions in which it is purified/stored. Earlier purifications of CpCR had DTT in the buffers to maintain a reducing environment for the four free cysteine residues present in the enzyme. DTT, a reducing agent is known to reduce the Cys residues coordinated to the zinc and release it [9]. CpCR and YADH belong to the same MDR superfamily. Thus, the enzyme was purified without the addition of DTT in the buffering system and the effects were clear with the increase in the activity by more than three times. The presence of MgCl2 in the storage buffer was also omitted as the Mg ion does not have any significant interactions with the protein surface (PDB: 4OAQ). HEPES replacing Tris-HCl buffer was based on the fact that the pH of the Tris buffer is sensitive to changes in temperature.

Our previous study reported an asymmetric reduction of ketones and ketoesters by CpCR, but the activity with aldehydes was better [3]. Thus, in this work, a detailed study of aldehydes, i.e., aliphatic and various substituents of benzaldehyde as the substrates for CpCR, was carried out. Aldehyde reduction is NADPH-specific. Aliphatic aldehydes show 60–70% activity as that of benzaldehyde (Table 1). Any substitution on any position of benzaldehyde decreases the activity, due to electronic and steric effects. 2, 4-dichlorobenzaldehyde (substrate 15) shows the least activity. Among the *ortho* and *para* substituted benzaldehydes, electron withdrawing groups like NO2 and CN (substrates 13 & 12) show less activity as they can destabilise the benzene ring. Bromo and floro substitutions along with electron donating groups such as CH3 and O-CH3 at *ortho* and *para* positions (substrates 4, 9, 10 and 11) show comparatively better activity than substrates which contain electron-withdrawing groups (substrates 6, 12 and 13). It is expected that substitutions on *ortho* and *para* positions behave similarly but in the case of *ortho* bromo (substrate 5), steric effects dominate. Substrate 9 shows comparatively better activity than substrates 10 and 11 due to the presence of the smaller methyl group.


**Table 1.** The specific activity of CpCR against various aldehydes that were not reported earlier.

 One unit of the enzyme activity is defined as the amount of enzyme that oxidizes 1μmol of NADPH per minute at ◦C. <sup>2</sup> Previously reported [3]. Put here for comparative analysis with other substrates.

The kinetic parameters, i.e., enzyme affinity and the catalytic rates of CpCR with two different substrates (benzaldehyde and 2, 4-dichlorobenzaldehyde) show drastic differences in activity (Table 2). The dramatic decrease in activity with 2,4-dichlorobenzaldehyde could be because of its poor fit in the substrate cleft as reflected in >20 fold higher Km value as compared to that of benzaldehyde. A similar decreased activity can be seen in case of the *ortho* and *para* substituted benzaldehydes (Table 1). The presence of two Cl substitutions (inductive electron withdrawing groups) in 2,4-dichlorobenzaldehyde cause significant instability of the benzene ring which does not favour the reduction of the carbonyl group. Thus, both electronic and steric factors can explain the low activity of compound seen in entry 15, Table 1. Overall, the affinity of CpCR towards benzaldehyde decreases with increase in substitution on the benzene ring.


## *2.2. Oligomeric State of CpCR*

The crystal structure of CpCR (PDB: 4OAQ) shows that the enzyme is a hetero dimer. SDS-PAGE indicates the presence of a 40 kDa sub unit in the buffer [19]. Gel filtration chromatography data indicates that CpCR is indeed a dimer of 80 kDa (Figure 1A) and remains so even at higher concentrations (Figure 1B). SRED from *Candida parapsilosis* ATCC 7330 exists in other oligomeric states unlike CpCR [29].

**Figure 1.** Oligomeric state of CpCR. (**A**) A plot of Log Mol. Wt. vs. Kav shows the calibration curve of the known standards including and the molecular weight of CpCR; (**B**) A plot of velocity of reaction vs. micro molar concentration of CpCR.

## *2.3. E*ff*ect of Chelating Agent EDTA and Divalent Metal Salts on Activity of CpCR*

## 2.3.1. Effect of Time on Chelation

Addition of EDTA to the enzyme solution in a 1:3 enzyme: EDTA mole ratio, resulted in a drop of the specific activity of the enzyme immediately by 30% (from around 36 U/mg to 25 U/mg). Prolonged incubation of the mixture for up to 120 min showed that the specific activity remained around 25 U/mg for the entire duration. Furthermore, this experiment when repeated with a four-fold increased EDTA concentration (1:12—enzyme: EDTA mole ratio) over a lesser duration of 30 min showed the exact same trend (Figure S1). This indicates that EDTA binding to zinc in the enzyme is a fast process, and since the equilibrium is established, prolonged incubation is unnecessary.

## 2.3.2. Removal of EDTA

Different amounts of EDTA were incubated with a fixed concentration of enzyme in order to determine the effective concentration of EDTA necessary to remove zinc from the enzyme. However, it was found that despite the wide range of mole ratio tested, the specific activity of the protein samples containing EDTA remained the same (Figure S2). This was indicative of the fact that although EDTA binds to the enzyme quickly, it has to be removed in order to remove the zinc. This is consistent with what has been reported in literature [11,17,18].

The specific activities obtained after EDTA removal from samples by dialysis shows a decreasing trend in specific activity with increasing EDTA content (results not shown), indicating that the EDTA removal is necessary for zinc removal. A similar trend was observed in CPCR2 where the loss of activity of the enzyme is a function of loss of catalytic Zn [17]. However, the dialysis method was time consuming (12 h) and not feasible for this enzyme as it is not very stable. The specific activities obtained after EDTA removal from samples by ultrafiltration shows a similar decreasing trend using dialysis (Figure 2A). This method takes 3 h, and is likely to be associated with a gradual loss of specific activity which has to be taken into account. The protein recovery from this method is very high and, therefore, this method was optimized for use in future experiments.

**Figure 2.** Activity of CpCR after chelation with EDTA (**A**) Effect of EDTA on CpCR activity at various concentrations; (**B**) Change in specific activity of CpCR before and after incubation at 4 ◦C.

Prolonged incubation during the EDTA removal on the enzyme sample was also studied. Two controls were designed to check spontaneous loss of activity—a temperature control that was placed at 4 ◦C for three hours and a centrifugal control that was placed in the filtration unit without EDTA treatment. The specific activity values were measured for all the controls and a sample treated with 10:1 enzyme:EDTA (Figure 2B). It was seen that there is a decrease of specific activity from 35 U/mg to 30 U/mg due to the three-hour incubation. However, the loss due to EDTA treatment was much more significant.

#### 2.3.3. Inhibition of CpCR by Divalent Metal Salts

Attempts were made to restore specific activity of the enzyme samples treated with EDTA by addition of zinc chloride. Surprisingly, it was found that the addition of external zinc decreased the specific activity of not only the EDTA treated sample but also of the control sample without any EDTA (Figure 3A). This has been previously reported for carboxypeptidase-A and could be caused

by bridging of the water molecule bound to the catalytic zinc to the external zinc thereby preventing substrate entry [30]. It was confirmed that this phenomenon is not just specific to zinc but to divalent ions of size similar to zinc as shown in Figure 3B. Reports on inhibition of ADHs by divalent metal ions suggest that the inhibition can be pH dependent and the mechanism mainly involves the replacement of native metal ion present in the catalytic site or by the coordination of added divalent metal ions with the sulphydryl groups of the enzyme. The inhibition can be reversed by addition of EDTA to remove the excess Zn ions [31–34]. Currently we are investigating the mechanism of inhibition of CpCR by such divalent metal salts.

**Figure 3.** Effect of Zn2<sup>+</sup> and other divalent metals on activity of CpCR. (**A**) Effect of Zn2<sup>+</sup> (ZnSO4) on CpCR activity at various concentrations; (**B**) Specific activity of CpCR influenced by various divalent metal ions.

## *2.4. Cofactor Pre-Treatment Prevents CpCR Activity Loss*

Pre-treatment of enzyme samples with cofactors NADPH/NADP<sup>+</sup> prevented the loss of specific activity upon EDTA treatment in a concentration-dependent manner (Figure 4A). This is also reported with 1,10-phenanthroline when YADH is pre-incubated with the cofactor [18]. It was seen that pre-treatment with 1 mM NADP<sup>+</sup> gave 50% more specific activity as compared to the EDTA treated sample without cofactor treatment, while 2 mM NADPH retained the entire specific activity of the initial control. This may be due to retaining zinc and not allowing EDTA to access it possibly a result of structural changes due to cofactor binding.

**Figure 4.** Effect of NADPH incubated CpCR on its activity. (**A**) Effect of incubation of different concentrations of NADPH on CpCR activity; (**B**) Kinetics of CpCR, NADPH bound vs. unbound.

The kinetics of the cofactor-bound enzyme were determined by the enzyme obtained from the modified purification protocol presented in this study with changes in the kinetic parameters duly noted. It was observed that the Km value for both cofactor bound and unbound CpCR remained

more or less constant at 0.23 mM, implying that substrate binding remained largely unchanged in the cofactor-bound enzyme. However, there was a two-fold decrease in Vmax of cofactor-bound enzyme (Figure 4B). The structure of CpCR is significantly similar to LADH (PDB: 1HLD) with a *p*-value of 1.41 <sup>×</sup> 10−<sup>13</sup> [35]. The decreased Vmax value may also be attributed to the structural changes in the cofactor bound enzyme resulting in narrowing of the catalytic cleft and hindering the entry of the substrate [36]. Overall, the conversion of the *apo* enzyme to the *holo* form seems to affect the rate of the reaction significantly even though the enzyme affinity towards the substrate is retained.

#### **3. Materials and Methods**

#### *3.1. Chemicals and Media*

All the chemicals and media were purchased from SRL, Chennai, India and HiMedia, Mumbai, India. AKTA protein purification system and GST affinity column were purchased from GE Healthcare Life Sciences, Bangalore, India. The ultra-centrifugal filters were obtained from Merck, Mumbai, India.

## *3.2. Enzyme Expression and Purification*

The overexpression and purification of CpCR were performed as per the reported methodology using GST affinity chromatography [28]. For all the experiments, except cofactor binding studies, the enzyme obtained from this protocol was used. Slight modifications to the protocol were done to obtain better yield and stability of the enzyme. They include: 1. Removal of DTT and MgCl2 from all the buffers and replacing the Tris HCl with HEPES buffer. 2. Cell disruption was done using a 150 W ultra-sonicator. 3. The cleared lysate was loaded onto the GST column at a flow rate of 1 ml min−1. 4. The use of a superdex column was skipped. The yield and fold purifications were obtained by checking the specific activity of CpCR against benzaldehyde.

The composition of the buffers used are as follows: Equilibration and wash buffer: 50 mM HEPES pH 7.5, 500 mM NaCl, 2.5% glycerol; Elution buffer: 50 mM HEPES pH 7.5, 20 mM Glutathione; Desalting buffer: 20 mM HEPES pH 7.5, 50 mM NaCl.

#### *3.3. Specific Activity, Substrate Scope and Kinetic Studies of CpCR*

The protocol used for determining the specific activity of the enzyme against different substrates was the same as reported previously [3]. The activity of CpCR against benzaldehyde substituents and aliphatic aldehydes was checked. Substrate concentrations varying from 0–4 mM were used to determine the specific activity of CpCR and the Lineweaver–Burk plot gave the Km and Vmax of CpCR against the specific substrates.

## *3.4. Oligomeric State of CpCR*

Gel filtration chromatography was done to find out the oligomeric state of CpCR [37]. Mixture of standard proteins containing ribonuclease (13.7 kDa), chymotrypsin (25 kDa), ovalbumin (43 kDa), albumin (67 kDa), aldolase (158 kDa), catalase (232 kDa), ferritin (440 kDa) and thyroglobulin (669 kDa) were used to calibrate the Sephadex 200 HR column. A plot of Log Mol. Wt. and Kav was made to calculate the molecular weight of CpCR.

A plot of velocity (μmoles min<sup>−</sup>1) vs. concentration of CpCR (μM) was made to find the presence of higher oligomeric state of CpCR at its higher concentrations of up to 500 μM.

#### *3.5. Treatment of CpCR with EDTA, Divalent Metal Salts*

Benzaldehyde was used as the substrate to check the activity of CpCR against the effects of EDTA and other metal salts. The mixture containing the ratio of 1:3 and 1:12 (number of moles of CpCR to the number of moles of EDTA) was checked for the activity instantaneously and compared to the untreated CpCR. Another experiment with the same mixtures incubated for up to two hours at 4 ◦C with the activity checked every 15 min from the start of incubation was also done. Additionally, different ratios

of the number of moles of CpCR to EDTA were tried to see the instantaneous effect on the activity of CpCR.

Ultrafiltration was done to remove the low Mol. Wt. EDTA from the above mixtures to later determine the change in the activity of CpCR. The chelator treated protein (100–500 μg) in a volume < 500 μL was placed in 0.5 mL filter (Amicon, 10 kDa mol. wt. cutoff) and the volume was made up to 500 μL using a desalting buffer. The samples were then concentrated at 14,000 *g* at 4◦ C for 30 min, following which the volume was again made up to 500 μL. This process was repeated for a total of six times over three hours. At the end of the process, the enzyme was recovered and its concentration was determined by Bradford's method [38] followed by its activity assay.

The addition of micro- and nanomolar concentrations of ZnSO4 to the EDTA treated CpCR was also studied. The metal salt in question was directly added to the assay solution from a 1 M stock to obtain the desired concentration after addition of substrate and enzyme to determine the activity spectrophotometrically. Effects of divalent metal salts like CoCl2, NiCl2, MgCl2, CuCl2, MnCl2 and ZnCl2 at nanomolar concentrations on the CpCR activity were noted.

## *3.6. CpCR—Cofactor Binding Studies*

For this study, the modified protocol for purification was used. NADPH was added to 100 μg of the protein from 10 mM and 20 mM stock of cofactor to a final concentration of 1 mM and 2 mM respectively, and the solution was incubated at 4 ◦C for one hour. Following this, the sample was treated with EDTA maintaining a 1:1000 enzyme: EDTA mole ratio. The EDTA was removed by ultra-centrifugation, and the specific activity of the sample was determined. Kinetic studies of CpCR against benzaldehyde were compared with the CpCR pre-incubated with 2 mM NADPH.

## **4. Conclusions**

CpCR substrates include substituted benzaldehydes and aliphatic aldehydes. Substituted benzaldehydes showed lower activity as compared to benzaldehyde. The oligomeric state of the enzyme was confirmed to be dimeric at all concentrations, in agreement with the crystal structure. In chelation studies with EDTA, a decrease in enzyme activity with an increase in EDTA concentration is seen after removal of the EDTA from the solution by ultrafiltration. All the divalent metal ions inhibit the activity of CpCR even at nanomolar concentrations. The protocol for CpCR purification was modified to obtain 10 times more yield and > 3-fold purification and used in studying the cofactor binding studies. The pre-incubation of CpCR with cofactor makes the enzyme resist the Zn removal by EDTA chelation and retains activity. The *apo* and *holo* forms of CpCR do not differ in their affinity towards benzaldehyde but differ in their reaction rates.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4344/9/9/702/s1; Figure S1. Time study of chelation; Figure S2. Concentration based chelation without removing the chelator.

**Author Contributions:** Conceptualization, A.C., V.K.K. and D.C.; Experiments carried out by, V.K.K. and D.C.; Formal analysis, A.C., V.K.K. and D.C.; Writing—original draft preparation, V.K.K.; Writing—review and editing, A.C.; Supervision, A.C.

**Funding:** This research received no external funding.

**Acknowledgments:** One of the authors, Vinay Kumar Karanam, expresses his gratitude to IIT Madras, India, for the fellowship.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**

1. Persson, B.; Hedlund, J.; Jörnvall, H. Medium- and short-chain dehydrogenase/reductase gene and protein families: The MDR superfamily. *Cell. Mol. Life Sci.* **2008**, *65*, 3879–3894. [CrossRef] [PubMed]


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Accelerated H2 Evolution during Microbial Electrosynthesis with** *Sporomusa ovata*

## **Pier-Luc Tremblay 1,2,†, Neda Faraghiparapari 3,† and Tian Zhang 1,2,3,\***


Received: 14 January 2019; Accepted: 1 February 2019; Published: 8 February 2019

**Abstract:** Microbial electrosynthesis (MES) is a process where bacteria acquire electrons from a cathode to convert CO2 into multicarbon compounds or methane. In MES with *Sporomusa ovata* as the microbial catalyst, cathode potential has often been used as a benchmark to determine whether electron uptake is hydrogen-dependent. In this study, H2 was detected by a microsensor in proximity to the cathode. With a sterile fresh medium, H2 was produced at a potential of −700 mV versus Ag/AgCl, whereas H2 was detected at −500 mV versus Ag/AgCl with cell-free spent medium from a *S. ovata* culture. Furthermore, H2 evolution rates were increased with potentials lower than −500 mV in the presence of cell-free spent medium in the cathode chamber. Nickel and cobalt were detected at the cathode surface after exposure to the spent medium, suggesting a possible participation of these catalytic metals in the observed faster hydrogen evolution. The results presented here show that *S. ovata*-induced alterations of the cathodic electrolytes of a MES reactor reduced the electrical energy required for hydrogen evolution. These observations also indicated that, even at higher cathode potentials, at least a part of the electrons coming from the electrode are transferred to *S. ovata* via H2 during MES.

**Keywords:** industrial biotechnology; electrochemistry; biohydrogen; biocatalysis; process development; bacteria

## **1. Introduction**

Reductive bioelectrochemical processes rely on the transfer of electrons from a cathode to a microbial catalyst for the reduction of a substrate with protons coming from an anodic reaction [1,2]. The substrate can be inorganic carbon molecules like CO2 that will be reduced to multicarbon compounds or CH4 via microbial electrosynthesis (MES) [3–10]. Organic carbon compounds can also be converted into commodity chemicals via electrofermentation [11,12] or electrorespiration [13].

In reductive bioelectrochemical systems (BES), the electrons are thought to be transferred directly via physical contact between the microbes and the cathode or indirectly via an electron shuttle such as H2 [14–16]. Experimental evidences suggest that H2 evolution from a graphite electrode often used in reductive BES starts happening only at potentials below −800 mV vs. Ag/AgCl in batch experiments [17]. Thus, it has been proposed that when the cathode potential is set higher than −800 mV, electrons are transferred via a H2-independent mechanism that could possibly involve the direct acquisition of electrons by components of the bacterium cell wall [3,18].

A recent study indicated that in the presence of cell-free spent medium from the electroactive acetogen *Sporomusa sphaeroides*, a significant quantity of H2 is produced in a BES with a cathode potential set at −710 mV vs. Ag/AgCl [19]. Furthermore, the same study showed that in the presence of cell-free spent medium from the electroactive methanogen *Methanococcus maripaludis*, H2, as well as formate, are formed in a BES at higher cathode potential compared to sterile fresh medium control and in sufficient quantities to account for all the methane produced from CO2. The authors suggested a novel electron transfer mechanism in which hydrogenases and formate dehydrogenase released in the medium from microbial cells would interact with a cathode set at a potential above −800 mV vs. Ag/AgCl to catalyze the formation of soluble electron shuttles. Alternatively, other groups have proposed that copper, nickel, iron, or vanadium deposited at the surface of a cathode via microbial activity could be responsible for the increase of bioelectrochemical hydrogen production observed at different potentials after exposure of the cathode to microbial catalysts [20–22].

Cathode materials [23], reactor designs [24–27], and operating modes [28,29] are all parameters that can positively affect H2 evolution. The chemistry of the solution filling the cathodic reactor also has an impact on the relation between the cathode potential and H2 evolution through changes in H2 initial concentration [22,30], changes in buffer composition [31–33], and through the presence of weak acids [34,35]. During MES, the microbial catalyst will alter the chemical environment surrounding the cathode by releasing metabolic wastes, products, or diverse cell components or debris in the cathodic solution [19,21,36,37]. To study the possible correlation between microbial alterations and H2 evolution in a MES system, a hydrogen microsensor was placed in close proximity to the surface of a cathode set at different potentials to measure H2 evolution in situ in the presence of sterile medium, bacterial culture or cell-free filtrate. *Sporomusa ovata*, one of the best pure culture MES catalysts for the production of acetate from CO2 [38,39] is used here as a model because of its capacity to perform MES over a large range of cathode potential [3,40–46]. In order to further understand H2 evolution and electron uptake during MES, other variables were investigated, including the presence of metals at the cathode surface as well as the presence of hydrogenases and other enzymes in the cell-free spent medium.

## **2. Results and Discussion**

## *2.1. H2 Evolution in an Abiotic MES Reactor*

A hydrogen microsensor with a sensitivity of ≥ 0.1 μM was inserted into an abiotic MES reactor to monitor H2 evolution with a cathode set at potential ranging from −900 to −400 mV vs. Ag/AgCl (Figure S1). H2 concentration was measured in close proximity to the cathode surface where microbial catalysts in operating MES reactor [3,38,41] are likely to oxidize large fraction of H2, if any is produced, before it can diffuse away in the medium and in the reactor gas phase. The initial evaluation of H2 evolution was conducted in an abiotic MES reactor. Both cathodic and anodic chambers were filled with sterile 311 medium at pH 6.8, which is the growth medium as well as the electrolyte solution normally used in *S. ovata*-driven MES reactors [3]. Under these experimental conditions, the highest cathode potential at which H2 evolution was observed was −700 mV vs. Ag/AgCl (Table 1). As expected, more current was drawn at lower potentials versus higher potentials because H2 evolved faster at lower potentials (Figure 1, Table 1, Figure S2).

**Table 1.** H2 evolution, current density and electrons recovery in MES reactors with sterile fresh 311 medium a.


<sup>a</sup> Each result is the mean and standard deviation of three replicates. <sup>b</sup> Not detected. <sup>c</sup> Not applicable.

**Figure 1.** Current draw in a MES reactor filled with sterile fresh 311 medium at different cathode potentials. Results shown are from a representative example of three replicate.

## *2.2. H2 Evolution in the Presence of a S. ovata Cell Suspension*

*S. ovata* is an efficient acetogenic microbial catalyst for the production of acetate from CO2 by MES capable of acquiring electrons from the cathode at potentials as high as −600 mV vs. Ag/AgCl [1,3,41]. To determine the impact of *S. ovata* on H2 accumulation in a MES system, a cell suspension was added to a cathode chamber equipped with a H2 microsensor. At the tested potentials higher than −900 mV, no H2 was detected (Table 2). At −900 mV, 0.11 ± 0.02 <sup>μ</sup>M min−<sup>1</sup> (n = 3) of H2 accumulated, and the electron recovery from current to H2 was only 8.5 ± 1.5% (Table 2, Figure 2, Figure S3), indicating that *S. ovata* cells probably quickly consumed most of the H2 generated at the cathode. In the meantime, ca. 6.6 ± 0.2 μM of acetate was produced by *S. ovata* cell suspension from CO2 in the cathode chamber of the MES system over a period of 25 minutes, and the coulombic efficiency was 75 ± 3%.

**Table 2.** H2 evolution, current density and electrons recovery in MES reactors with *S. ovata* cell suspension a.


<sup>a</sup> Each result is the mean and standard deviation of three replicates. <sup>b</sup> Not detected. <sup>c</sup> Not applicable.

**Figure 2.** Current draw in a MES reactor containing a *S. ovata* cell suspension at a cathode potential of −900 mV vs. Ag/AgCl. Results shown are from a representative example of three replicate.

## *2.3. H2 Evolution Shifting in the Presence of a Cell-Free Filtrate from S. ovata Culture*

To recreate the chemical environment of an operating MES reactor in the absence of the H2-oxidizing microbial catalyst, a cell-free filtrate from a *S. ovata* culture grown beforehand for four weeks under electrosynthetic condition was employed as the cathodic solution. Compared to a *S. ovata* cell suspension, H2 was present in significant quantities at −900 mV vs. Ag/AgCl with simultaneous current draw indicating that the cell-free filtrate was unable to oxidize detectable amount of H2 (Table 3, Figure 3). Furthermore, higher H2 evolution rates in the MES reactor were detected with the cell-free filtrate than with the sterile fresh medium at cathode potential ranging from −900 mV to −700 mV vs. Ag/AgCl (Figure 4, Figure S4). In addition, the detectable H2 evolution in the presence of cell-free filtrate was shifted by +200 mV compared to sterile medium control. H2 started to accumulate at −500 mV with *S. ovata* cell-free filtrate, whereas H2 evolution was detected at −700 mV with sterile fresh 311 medium. These results suggested that secreted metabolites, secreted cell components or chemical characteristics of the *S. ovata* cell-free filtrate enabled H2 evolution at higher cathode potentials and accelerated it. Furthermore, it seems to indicate that electron transfer in MES driven by *S. ovata* at high cathode potentials involved H2 as an electron shuttle. However, further characterization is required to determine whether all the electrons required for the acetate production by MES at high cathode potentials are transferred via H2, or if a significant fraction of the electrons comes from an alternative H2-independent route such as direct electron transfer.


**Table 3.** H2 evolution, current density and electrons recovery in MES reactors with cell-free spent medium a.

<sup>a</sup> Each result is the mean and standard deviation of three replicates. <sup>b</sup> Not detected. <sup>c</sup> Not applicable.

**Figure 3.** Current draw in a MES reactor containing cell-free spent medium from an electrosynthetic *S. ovata* culture. Results shown are from a representative example of three replicate.

**Figure 4.** H2 evolution rate with fresh sterile medium, *S. ovata* cell suspension and cell-free spent medium. (**A**) Cathode potential at −900 mV vs. Ag/AgCl and (**B**) cathode potentials ranging from −850 to −500 mV. Above −500 mV vs. Ag/AgCl, no H2 evolution was detected under all tested conditions. No H2 evolution was detected with *S. ovata* cell suspension above −900 mV vs. Ag/AgCl. Each result is the mean and standard deviation of three replicates.

A possible explanation for the faster H2 evolution observed here would be the presence of acetic acid in the cell-free spent medium. Weak acids including acetic acid have been shown to have a catalytic effect on H2 evolution in BES at acidic pH as well as in abiotic electrochemical systems [34,35]. In the cell-free filtrate samples of *S. ovata* tested here, the concentration of acetate/acetic acid varied from 10.4 to 12.4 mM (Figure S5). This was generated by the filtered cells beforehand during the MES process. However, the pH of the cell-free filtrate maintained with a carbonate buffer was measured at ca. 6.8, which suggests that the acetic acid/acetate ratio is unlikely to be high enough to have a significant impact on the electrochemical generation of H2.

## *2.4. Metals at the Surface of the Cathode after Exposure to Cell-Free Spent Medium*

Energy-dispersive X-ray spectroscopy (EDS) was employed to examine the presence of elements on the surface of cathode electrodes exposed to electrosynthetic *S. ovata* cell-free spent medium or sterile 311 medium. As expected, C, O, Na, Mg, P, Cl, K and Ca were observed on spectra for each condition (Figure 5). Interestingly, two metal elements, i.e., cobalt and nickel, that may be involved in the acceleration of H2 evolution, were detected with the electrode samples exposed to cell-free spent medium but not with samples exposed to sterile 311 medium. When subtracting the background, the EDS signal corresponding to cobalt had an X-ray count of ca. 50 and 120 in cell-free spent

medium replicate A and B, respectively. For nickel, the X-ray count was ca. 60 for replicate A and 140 for replicate B. No signal corresponding to nickel or cobalt was detected with sterile 311 medium (Figure 5C).

**Figure 5.** EDS spectra (from 0.0 to 4 KeV) of the cathode electrode surface after exposure to cell-free spent medium from electrosynthetic *S. ovata* (**A**,**B**) or to sterile fresh 311 medium (**C**).

Multiple proteins and enzymes found in acetogens, such as NiFe hydrogenases, acetyl-CoA synthase, CO dehydrogenase, corrinoid iron-sulfur protein and cobalamin-dependent methyltransferases, have metal centers containing nickel or cobalt [47–49]. Evidence suggested that hydrogenases released in the medium by S. sphaeroides cells are responsible for improved H2 evolution from a cathode [19]. In this study, we were unsuccessful at detecting hydrogenase activity in the cell-free spent medium with a methyl-viologen based assay. Additionally, SDS-PAGE and mass spectrometry were applied to examine *S. ovata* cell-free spent medium and no intact hydrogenases were detected (Figure S6). Further research is warranted to establish if functional hydrogenases or other redox enzymes are released by *S. ovata* during MES, and if these enzymes participates actively to electron transfer from the cathode.

Three main visible bands were excised from the SDS-PAGE corresponding to an aldehyde oxidoreductase (Sov\_1c12660), a glutamate synthase (gltB2) and an ll-diaminopimelate aminotransferase (dapL2) (Figure S6). All three are predicted to be cytoplasmic proteins [50], which suggest that microbial cell content was spilled in the MES reactor medium due to lysis. This further increased the chemical complexity of the material surrounding the cathode. The presence of cytoplasmic enzymes in the cell-free spent medium suggests that the detected nickel and cobalt at

the surface of the cathode could come from three sources: intact enzymes, metal centers attached to damaged enzymes and free metal centers detached from apoenzymes. Furthermore, metal centers associated for instance with Sov\_1c12660 (Fe, Mo), as well as with other proteins possibly found in the cell-free spent medium of *S. ovata*, could also interact transiently with the cathode and facilitate hydrogen evolution.

## **3. Materials and Methods**

## *3.1. Bacterium and Growth Conditions*

*S. ovata* DSM-2662 [51] was obtained from the Deutsche Sammlung Mikroorganismen und Zellkulturen (DSMZ, Braunschweig, Germany). *S. ovata* strains were routinely maintained in the DSMZ-recommended 311 medium at 30 ◦C and at pH 6.8 with a H2:CO2 (80:20) atmosphere (1.7 atm). Casitone, sodium sulfide, yeast extract, and resazurin were omitted from the 311 medium. For MES experiments, cysteine was also omitted from the 311 medium.

## *3.2. Preparation of Cell Suspension*

Triplicate of 300 mL of anoxic cultures of *S. ovata* were harvested by centrifugation and washed two times with 311 sterile medium before being resuspended in a final volume of 3 mL. H2:CO2-grown *S. ovata* cells were harvested when the optical density was ca. 0.3 (545 nm). Cell suspensions were then used to inoculate the cathodic chamber of MES reactors containing 250 mL of sterile 311 medium.

## *3.3. Cell-Free Spent Medium of S. ovata*

*S. ovata* cultures at an OD545 of 0.2 catalyzing the conversion of CO2 to acetate for 4 weeks in a MES reactor with a cathode potential of −900 mV vs. Ag/AgCl/3M KCl were filtered two times with 0.45 μm pore size filters. 250 mL of cell-free spent medium was then injected in the cathodic chamber of a MES reactor.

## *3.4. MES Reactor and H2 Evolution*

Three-electrode H-cell glass bioreactor (Adams and Chittenden Scientific Glass, Berkeley, CA, USA) systems were used for H2 evolution experiments during MES. The anode chamber was filled with 250 mL of sterile 311 medium, whereas the cathode chamber contained S. ovata cell suspension, *S. ovata* cell-free spent medium or sterile 311 medium. The anode and cathode chambers were separated by a Nafion 115 membrane (DuPont Inc., Wilmington, DE, USA). Graphite plates (type HLM, SGL carbon, Wiesbaden, Germany) with a normalized surface area of 35.5 cm2 were used as both anode and cathode. The anode and cathode chamber was stirred at 300 rpm and bubbled with N2:CO2 (80:20) gas mixture at a flow rate of 18.5 ± 1.0 mL/min (ADM 2000 Flowmeter, Agilent, Santa Clara, CA, USA). For the data presented here, the stirring and bubbling in cathode chamber was paused for 25 min during H2 measurement to avoid interference of accuracy measurement. Fixed potentials were applied to the cathode from −900 to −400 mV vs. Ag/AgCl using a multi-potentiostat (CHI 1000C, CH Instrument, Inc., Austin, TX, USA).

H2 evolution was measured in close proximity to the surface of the graphite cathode of a MES reactor maintained at 25 ◦C with a hydrogen microsensor (H2-500, Unisense, Aarhus, Denmark). The microsensor with a tip surface area of 0.2 mm2 was installed in the MES reactor and the distance between the tip of the H2 microsensor and the cathode surface was ca. 2 mm (Figure S1). Before each experiment, the microsensor was calibrated with a gas mixture containing 7% H2. The sensor data was logged continuously every second using a data logger (Microsensor Multimeter, Unisense, Aarhus, Denmark). H2 evolution was measured for a maximum of 25 min until H2 concentration curves reached an equilibrium plateau. The rates of H2 evolution presented in this study are the slopes of H2 concentration curves before this plateau.

## *3.5. High-Performance Liquid Chromatography (HPLC)*

Acetate concentration was measured with an HPLC apparatus equipped with a HPX-87H anion exchange column (Bio-Rad Laboratories Inc., Hercules, CA, USA) at a temperature of 30 ◦C, with 5 mM H2SO4 as the mobile phase, and a flow rate of 0.6 mL/min.

## *3.6. SDS-PAGE and Mass Spectrometry*

Proteins from 220 mL of *S. ovata* cell-free spent medium were concentrated 375 times with Amicon Ultra-15 centrifugal filter devices with a nominal molecular weight limit of 3 kDa (Merck Millipore, Hellerup, Denmark). Protein concentration was measured with a Coomassie Plus Assay kit (ThermoScientific, Hvidovre, Denmark) and 1.5 μg of protein from two cell-free spent medium samples were loaded on a Sodium Dodecyl Sulfate Poly-Acrylamide Gel Electrophoresis (SDS-PAGE, 12.5%). The PageRuler prestained protein ladder (ThermoScientific) was loaded on the same gel to evaluate the molecular weight of protein bands. After protein separation, the SDS-PAGE was stained with the GelCode Blue stain reagent (Life Technologies, Carlsbad, CA, USA). Revealed protein bands were excised and sent to Alphalyse (Odense, Denmark) for protein identification by matrix-assisted laser desorption/ionization-tandem mass spectrometry (MALDI-MS/MS).

#### *3.7. Hydrogenase Activity Assay*

Hydrogen-evolving hydrogenase activity in the cell-free spent medium from a S. ovata-driven MES reactor and in 311 sterile fresh medium was measured in triplicate by monitoring the decrease in the absorbance of dithionite-reduced methyl viologen as described previously [52,53]. Briefly, cell-free spent medium or sterile medium were combined with 0.1M HEPES buffer (pH 8.0) and 100 μM reduced methyl viologen in an anoxic rubber-stoppered cuvette. Change in the absorbance at 604 nm was monitored over time at room temperature. The extinction coefficient of methyl viologen at 604 nm is 13.9 mM−<sup>1</sup> cm<sup>−</sup>1.

## *3.8. Energy-Dispersive X-ray Spectroscopy (EDS)*

Duplicate electrode samples were air-dried and examined with a Quanta 200 FEG scanning electron microscope (FEI, Hillsboro, OR, USA). EDS data were collected at an accelerating voltage of 20 kV under high vacuum conditions, with 10 mm working distance and 4.0 spot size.

## *3.9. Equations*

The Nernst Equation (1) was used to calculate theoretical cathode potential (E) at which H2 evolution starts as described in Vincent et al., 2007 [30].

$$\mathbf{E} = \mathbf{E}^0 + 2.3 \mathbf{R} \mathbf{T} / n\mathbf{F} \log \| (\mathbf{a} \mathbf{H}^+)^2 / \mathbf{p} (\mathbf{H}\_2) \| \tag{1}$$

where E<sup>0</sup> is the standard reduction potential for H2, R is the gas constant, T is the absolute temperature, n is the number of electrons involved (2 e<sup>−</sup> for H2 evolution), F is the Faraday constant, aH<sup>+</sup> is the activity of H<sup>+</sup> and p(H2) is the H2 partial pressure.

Henry's law 2 was used to calculate H2 partial pressure in the MES reactor gas phase.

$$\text{pH}\_2 = \text{k}\_\text{H}\text{C} \tag{2}$$

where kH is the Henry's law constant for H2 and C is the concentration of H2 in solution.

## **4. Conclusions**

This study showed that in a *S. ovata*-driven MES reactor, faster H2 evolution at higher cathode potential was enabled because the presence of microbial catalyst modified the cathodic solution chemistry. A higher accumulation of H2 means more reducing power, which should lead to lower requirements for electrical energy input for microbial CO2 reduction. The observed acceleration of H2 evolution could be due to the deposition of cobalt and nickel at the surface of the cathode caused by *S. ovata* catalytic activity. However, given the complexity of chemical species in *S. ovata* cell-free filtrate, it is possible that a more complicated synergistic effect is involved in H2 evolution during MES with *S. ovata*.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4344/9/2/166/s1, Figure S1: Experimental setup for H2 measurement in close proximity to the cathode surface of a MES reactor, Figure S2: H2 evolution profile over a period of 25 min with fresh sterile medium in the cathode chamber of a MES reactor, Figure S3: H2 evolution profile over a period of 25 min with *S. ovata* cell suspension in the cathode chamber of a MES reactor, Figure S4: H2 evolution profile over a period of 25 min with *S. ovata* cell-free spent medium in the cathode chamber of a MES reactor, Figure S5: Evolution of acetate concentration over time in a MES reactor filled with sterile medium or with *S. ovata* cell-free spent medium with a cathode set at a potential of either −600 mV or −900 mV vs. Ag/AgCl, Figure S6: Proteins found in the cell-free spent medium of electrosynthetic *S. ovata*.

**Author Contributions:** P.-L.T. and T.Z. conceived the project and designed the experiments. N.F. assembled and operated MES reactors. N.F. prepared *S. ovata* cell suspension, cell-free spent medium and measured H2 evolution with a hydrogen microsensor. Energy-dispersive X-ray spectroscopy, protein identification via SDS-PAGE and hydrogenase activity assay were done by P.-L.T. and T.Z. P.-L.T. and T.Z. wrote the manuscript with feedback from N.F.

**Funding:** This work was funded by the Novo Nordisk Foundation, the Chinese Thousand Talents Plan Program and Wuhan University of Technology.

**Acknowledgments:** We thank Daniel Höglund and Dawid Mariusz Lizak for their assistance with sampling and data processing.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Catalysts* Editorial Office E-mail: catalysts@mdpi.com www.mdpi.com/journal/catalysts

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18