Use of Protein Engineering to Elucidate Electron Transfer Pathways between Proteins and Electrodes

Abstract

Herein, we review protein engineering tools for electron transfer enhancement and investigation in bioelectrochemical systems. We present recent studies in the field while focusing on how electron transfer investigation and measurements were performed and discuss the use of protein engineering to interpret electron transfer mechanisms.

1. Introduction

Electrochemical biosensors are analytical devices that allow the detection of a particular molecule of interest (analyte) by combining a biological element with an electrode that can communicate with each other to produce a measurable electronic signal.¹ One example of a well-known commercially available electrochemical biosensor is the glucometer used by diabetes patients for real-time measurements of blood glucose concentrations.² These biosensors use glucose-oxidizing enzymes to harvest electrons from glucose molecules and transfer them to an electrode.³ In most cases, redox enzymes used in biosensors have their electroactive cofactors buried within the insulating protein matrix, interfering with efficient electron transfer (ET) to and from the electrode.⁴ As electrochemical biosensors are rapidly developing, there is a constant need to improve these systems.

ET between proteins and electrodes was the focus of many studies since the 1980s, where the Marcus theory for ET in proteins was developed.⁵ Winkler and Gray used chemically modified proteins to study the effect of distance on ET in proteins.^6,7 At the same time, the Armstrong team focused their research on the direct electrochemistry of redox proteins, mainly cytochromes, paving the road toward the study and development of direct electron transfer systems.⁸

Direct electron transfer (DET) bioelectrochemical systems are systems where the electron transfer is directly transferred between an enzyme and an electrode, without the need for mediation or encapsulation of the enzyme using complex polymers.⁹ In those systems, diffusion of an exogenous electroactive molecule does not play a role. Furthermore, in most cases, the observed reduction potential is that of the redox-active site of the protein or very close to it.⁹ To achieve DET, several factors need to be considered.

One factor is the availability of the redox-active site and the extent of its surface exposure. A surface-exposed ET domain will allow DET,¹⁰ while an ET domain buried within the insulating protein matrix will interfere with efficient ET. One way of achieving an exposed ET domain is by protein truncation, where some of the protein sequence is deleted at the DNA level or by using a specific protease. The truncation will usually be performed from the C-termini of the enzyme in a way that does not interfere with the protein activity or interactions while exposing the relevant electroactive molecule and make it more approachable. Fusing an electron-transferring domain with a surface-exposed electroactive molecule to a non-DET enzyme is another approach that has been shown to allow DET.¹¹ It is essential that the electron-transferring domain be linked via a flexible peptide linker to enable accessibility to the enzyme’s active site and transfer the electrons to the electrode.^10,11

Another factor is the distance between the redox-active site and the electrode. According to Marcus’s theory, the theoretical maximal distance for ET in proteins via quantum tunneling was calculated to be 20 Å.⁵ Longer-range electron transfer is possible in some enzymes, such as in the case of multiple sulfur–iron cluster hydrogenases, where a chain of cofactors enables ET from the buried redox-active site through other cofactors and to the outer surface of the enzyme.¹² Yet, close proximity between the distal cofactor in the chain and the electrode should be achieved for an efficient DET.¹³ One way of achieving proximity without the need for encapsulation of the protein in a polymer or hydrogel is the use of site-specific immobilization where the protein is directly linked to the electrode from a designated site that is rationally designed to achieve this proximity. The rational design of a protein requires information about the protein structure. Such data can be deduced from a three-dimensional protein structure that was either experimentally solved or modeled based on homology with similar proteins. It can also rely on previous mutagenesis studies of the enzyme to shed light on the relevant sites for its genetic manipulation. Examples are site-directed mutagenesis or the fusion of an affinity tag for protein tethering to the electrode.

Using site-specific immobilization allows the control of the enzyme orientation toward the electrode, which is another essential factor for efficient DET.¹⁴ Orienting the enzyme in a way that will position the active site of the enzyme toward the electrode can also result in increased stability and sensitivity due to the uniformity of the enzyme-fabricated electrode.^15,16 In such an electrode, the whole enzyme population will participate in the ET process, as opposed to mixed populations with different orientations, where some of the enzyme population will not be optimally oriented for an efficient ET.^3,14,17,18 There are several physical limitations to achieving uniform population of optimally oriented enzymes on an electrode such as electrode surface morphology and protein–protein and protein–electrode interactions that can lead to partial enzyme misfolding.¹⁹ As a result, one should take into account that the calculated ET properties will be an average of the overall orientations on the electrode, despite a dominant orientation that is more prevalent.²⁰

Enzyme immobilization to electrodes can be through a covalent attachment, such as the thiol side chain of cysteines for the formation of a gold–sulfur bond to a gold electrode.²¹ Another immobilization approach is through a coordinative attachment like in the case of metal ions binding to a polyhistidine tag.²² The π–π stacking interactions by noncovalent interactions between two aromatic surfaces are also broadly used for immobilization.²³ An additional way of enzyme immobilization is electroactive wiring, where a redox-active polymer, such as an osmium polymer, can be deposited on an electrode and used to entrap the enzyme of interest.²⁴

In many cases, protein engineering for DET can help decipher internal ET mechanisms and ET to the electrode of an enzyme of interest. Measurements of engineered proteins (rather than native ones) may serve as an additional tool for ET mechanism investigation, as we present in this review. Understanding ET processes within the protein and from the protein to an electrode using DET is essential for making biosensing systems more efficient. Herein, we will review the different approaches for protein engineering for DET studies (Figure 1) while focusing on how DET was measured and whether it had contributed to further understanding of ET processes.

Figure 1.

Schematic illustration of protein engineering tools described in this review for ET investigation and DET enhancement. Catalytic subunit (purple) with the electroactive cofactor (green hexagon) and electron-transferring domain (red) with two electroactive moieties (green/white rhombus). Fusion enzymes can have an affinity tag or an electron-transferring domain (blue oval represents both). Mutation site is represented by a black star, whereas a star denotes a natural amino acid, and an asterisk denotes an unnatural amino acid.

2. Protein Engineering for the Investigation of DET and ET Mechanisms

2.1. Protein Truncation and Partial Deletion

The approach of protein engineering by truncation includes the use of genetic tools for the deletion of domains of the protein-encoding sequence²⁵ or the use of specific proteases for the digestion of protein subunits.^26,27 The rationale behind the removal of a peptide or a protein domain is to expose the redox-active site of the protein or to enable a close proximity of the active site to the electrode surface.²⁸ When deletion is performed in a protein with multiple redox-active sites, such as an ET subunit with a multiheme cofactor, it enables the electrochemical characterization of different cofactors within the ET subunit.

Engineering of an enzyme that has a multiheme subunit was performed in several studies to decipher ET pathways. Partial or full deletion of multiheme subunits was performed for the investigation and characterization of ET mechanism. The DNA sequence encoding for the ET subunit (β subunit), which is a three heme-c containing subunit, of flavin-adenine dinucleotide-dependent glucose dehydrogenase (FAD-GDH) from Burkholderia cepacia was deleted for the removal of two out of three heme-binding sites (Figure 2).²⁹ Deletion enabled the investigation of the role of the remaining heme-c domain in the ET process. The resulting ET subunit showed an ability to accept electrons from glucose oxidation by the catalytic subunit using absorbance spectroscopy. It has also demonstrated DET abilities to an electrode, while the investigation of the engineered protein has shed light on the ET mechanism of the full-length protein. It was concluded that a single heme molecule in the ET subunit enabled DET to an electrode. Spectroelectrochemical measurements were used in this study to determine the formal potential of the ET subunit containing the single heme-c site. From these measurements, it was found that the heme domain that was investigated had the lowest formal redox potential of −284 mV vs Ag/AgCl compared to the other two heme domains in the full-length ET subunit with redox potentials of −85 and −155 mV vs Ag/AgCl. Such low ET potential can be utilized for glucose biosensing with reduced interference from other metabolites in a physiological context.

Figure 2.

ET domain partial deletion. Hypothesized ET pathway of FAD-GDH from Burkholderia cepacia. (A) DET pathway of the full-length native protein complex to an electrode. (B) DET pathway of the deletion mutant through the third heme only. Reprinted with permission from ref (29). Copyright 2018 Elsevier.

Besides a partial deletion, as performed in the above-mentioned study, a complete deletion of the ET subunit has also enabled the discovery of novel roles of an ET subunit in an enzyme of interest. An ET subunit containing a three heme-c domain was deleted from a d-fructose dehydrogenase (FDH) from Gluconobacter japonicas.³⁰ In this study, the entire DNA sequence encoding for the heme-c subunit was deleted. The use of cyclic voltammetry (CV) and spectroelectrochemistry measurements revealed that the heme-c subunit of FDH is essential for DET to an electrode. CV measurements with and without electrochemical mediator molecules have shown that the deletion did not compromise fructose oxidation, but only DET was affected. That study was followed by another study by the same team, where the ET subunit of the same protein was only partially deleted to remove one heme moiety out of three.³¹ A truncated form of the enzyme was used to investigate the ET mechanism of the DET-type enzyme. Deletion of the third heme moiety in the ET subunit resulted in increased DET efficiency. Comparison of the truncated enzyme to the native enzyme revealed that the deleted heme domain is not participating in DET to the electrode. By analyzing the results from CV measurements, using the limiting current density equation, the authors have deduced that the enzyme surface coverage on the electrode was increased by downsizing the enzyme footprint, and that enzyme orientation toward the electrode in the truncated form is more suitable for DET to the electrode. Further understanding the role of the third heme in the ET subunit was achieved by an in vivo oxygen consumption measurement, which has shown that ET to the respiratory system was ten times lower when using the truncated enzyme. This result implies that the third heme domain is participating in bacterial intramolecular processes; however, it is not essential in vitro. In this case, measurements of in vivo oxygen consumption for the determination of the ET pathway were performed on the truncated and the full-length enzyme, allowing new insights regarding the role of the different heme domains.

In some cases, protein truncation was performed using specific proteases to separate domains from the enzymatic complex. When using a protease, it is very important to verify that the digestion site is unique and is located within a nonstructured domain to prevent loss in enzyme catalytic activity. Such an approach was taken with cellobiose dehydrogenase (CDH) from Myriococcum thermophilum, where its catalytic subunit was separated from its ET subunit by a protease separating it through its polypeptide linker region.²⁷ CDH is a DET-type enzyme with a catalytic subunit and a b-type cytochrome subunit connected by a flexible polypeptide linker. The electrons from substrate oxidation by the catalytic domain are transferred to an electrode through the cytochrome domain. The two subunits of CDH were separated using protease digestion, and then an investigation of the interdomain ET mechanism and interactions in different pH values was performed. Previous studies suggested that the interaction between the two domains of CDH is affected by the charge repulsion of negatively charged patches on the surface of the two domains. Using hydrogen/deuterium exchange coupled with mass spectrometry and native ion mobility, mass spectrometry measurements made the experimental investigation of the charge repulsion model of the pH-dependent interdomain ET possible. It was also shown that the interdomain ET is independent of the physical contact between the two subunits of CDH but is controlled by the environmental pH that affects the charge at the interdomain interface. In this study, the combination of simulations of the protein surface electrostatics with the experimental data resulted in a better understanding of the subunit interactions and ET mechanism.

Different aspects of using protein truncation by deletions are in use to improve bioelectrochemical–biotechnological tools. Geobacter sulfurreducens is an obligate anaerobic bacterium, which requires a specific anaerobic atmosphere for its proper growth. This bacterial strain has been studied in the past two decades for the production of a highly conductive bacterial pili that can be used as a conductive nanowire. The gene for a PilA subunit that assembles to conductive pili nanowires in Geobacter sulfurreducens was partially deleted for the bottom-up fabrication of these nanowires.³² N-Terminus deletions of the pili nanowire constructive subunit (PilA) were performed, and the deletion mutants were expressed in Escherichia coli (E. coli). It was found that the deletion of 19 amino acids from the N-terminus of the protein enables the proper self-assembly of the protein nanowire. Conductive atomic force microscopy and scanning tunneling microscopy measurements enabled the characterization of the bottom-up fabrication of conductive bacterial pili, which was shown to have similar conductivity as native pili.

From these examples, it can be reasoned that protein engineering by truncation or deletion is a powerful tool for investigating ET mechanisms, enhancing DET, and determining ET properties. By combining various molecular tools for the measurement of truncated proteins and their comparison with native proteins, researchers can gain new insights as of how the electrons are being transferred within those complex biological molecules and from the enzymes to electrodes.

2.2. Fusion Proteins

Fusion proteins result from the genetically encoded attachment of a protein domain or a polypeptide, to the N- or C-termini of a protein of interest, in order to add a new property not present in the native one.³³ Fusion typically takes place at the DNA level, where the relevant gene is edited to have the additional fragment or domain inserted. In some cases, the protein is engineered to have a different ET domain, and in other cases, an affinity tag is added for enzyme attachment to an electrode.

Addition of an ET domain to an enzyme can be achieved by the fusion of a gene encoding for an ET domain such as a cytochrome. When such a fusion is performed, it is recommended to use a flexible polypeptide linker that will allow the movement of the ET subunit in a manner that will yield an efficient ET to an electrode. A minimal cytochrome c domain from magnetotactic bacteria magneto-ovoid bacterium MO-1 was fused to the C-termini of FAD-GDH derived from Burkholderia cepacia.¹¹ In this study, the smallest known natural cytochrome c (23 amino acids long) was used, with its heme cofactor exposed to the solution, and fused to the C-termini of FAD-GDH (FGM) via a flexible polypeptide linker. The fusion protein was then characterized using in-gel heme staining and spectrophotometry to verify the presence of a mature cytochrome c. For the electrochemical investigation, square-wave voltammetry (SWV) was used to enable high-resolution measurement of the Faradaic currents, which originate from the electroactive sites of the enzyme. Using SWV enabled the estimation of the redox potentials of the FAD and heme domains by comparison to wild-type GDH (that lacks the cytochrome domain), which allowed the assignment of the two peaks observed from the SWV to the FAD or heme domains. Enhanced DET was demonstrated by comparing the catalytic currents observed in the presence of glucose for the native and fusion enzymes. Chronoamperometric measurements were used with the supplementation of other sugars and interfering molecules to verify a correctly folded fusion enzyme, assuming that the correctly folded enzyme will retain its substrate specificity. As described herein, during the characterization of a fusion enzyme, it is essential to compare it to the native enzyme.

Different cytochromes can be used as an ET domain for achieving DET. A flavocytochrome b2 was fused to a “virtual lactate dehydrogenase” (LDH) originating l-lactate oxidase from Aerococcus viridans via a flexible hinge peptide.³⁴ In that case, the characterization began with the measurement of absorbance spectra to investigate intra-ET from the FMN cofactor of the enzyme to the heme-b molecule of the cytochrome. Then, using chronoamperometry, DET of the fusion enzyme was demonstrated by showing its ability to oxidize lactate in concentrations 4 times higher than those of the enzyme lacking the cytochrome. One of the important factors for successful DET is cytochrome maturation. Using an endogenous cytochrome from E. coli in case that E. coli is the host organism enables high expression and maturation levels of the engineered enzyme. For that reason, cytochrome b₅₆₂ from E. coli was fused to the N- or C-termini of FAD-GDH from Botryotinia fuckeliana via a flexible polypeptide linker from CDH.³⁵ Estimation of the intra-ET was achieved using absorbance spectra measurements, where the performance of C- and N-termini fusion proteins were compared. The comparison of the two fusion variants revealed that fusing the cytochrome domain to the N-terminus of the enzyme increased the intra-ET rates, which resulted in higher current density as was measured by chronoamperometry. This work underscores the importance of the direction and distance of the added ET domain and demonstrates that the overall ET rate can be enhanced when taking into account intra-ET rates.

Using cytochromes as an ET domain is one example of fusion enzymes that can directly communicate with an electrode. Another valuable way for achieving DET is enzyme immobilization unto electrodes using an affinity tag. The enzyme is oriented to allow DET, and its electroactive site is positioned to be in close proximity to the electrode surface. When using an affinity tag for controlled enzyme orientation, analysis of the electrode surface should be performed by different techniques for the enzyme–electrode interface characterization. Some of the methods that are available for this purpose are electrochemical impedance spectroscopy (EIS),³⁶⁻³⁹ atomic force microscopy (AFM),⁴⁰⁻⁴² surface plasmon resonance (SPR) spectroscopy,⁴³⁻⁴⁵ and quartz crystal microbalance (QCM).⁴⁶⁻⁴⁸

Recently, a fusion of FAD-GDH derived from Burkholderia cepacia and a gold-binding peptide (GBP) in its N- or C-termini was engineered.¹⁷ GBP is a 12 amino acid peptide that enables the enzyme’s direct attachment to an Au electrode surface. The attachment of the fusion enzyme to the gold surface was first characterized using SPR spectroscopy in a flow cell and was compared to the native enzyme. From that comparison, it could be deduced that the affinity of the fusion enzyme toward gold has increased. Next, to investigate the enzyme orientation on the electrode, AFM measurements were taken. The native enzyme showed 3 times higher film thickness variation on the surface, which means a mixed population of enzyme orientations. To find the optimal working potential of the fusion enzyme, polarization measurements were performed. Another verification for the enzyme orientation was enabled by measuring the open-circuit potential (OCP) of the enzyme–electrode in the presence or absence of a substrate. It was shown that the OCP of the fusion enzyme has significantly decreased in the presence of a substrate, which indicates a more significant fraction of “correctly oriented” enzymes on the surface when compared to the native enzyme. EIS was used to determine the best fusion option between the N- or C-termini GBP. EIS was used to analyze the charge transfer resistance of the different variants. It was lower for the C-terminus fused variant, meaning that electron tunneling is more efficient using this variant. The fusion enzyme was characterized as a glucose biosensor on a gold screen-printed electrode in a follow-up study.⁴⁹ In this study, analysis of binding kinetics was performed using QCM measurements by monitoring the change in resonance frequency of the piezoelectric quartz crystal, and DET on the screen-printed electrode was demonstrated using CV and chronoamperometry.

Enzyme orientation on electrodes can be performed on a variety of surfaces and materials. As GBP is used for the attachment to gold surfaces, other peptides can be used for the attachment to other electrode materials. A methionine-rich (Met-rich) peptide present in the native amino acids sequence of copper efflux oxidase (CueO) from E. coli, close to its C-termini, was examined to enable its rapid oriented immobilization to carbon nanotubes (CNTs).⁵⁰ Several protein-engineering approaches were used for that purpose. Deletion of the Met-rich region was performed to test its role in protein adsorption on CNTs. The deletion variant of the enzyme showed high laccase activity compared to that of the native enzyme using colorimetric assay while at the same time showing lower DET abilities when tested on an electrode. The Met-rich peptide was also mutated to a serine-rich peptide by replacing all methionine residues with the hydrophilic amino acid serine, which resulted in poor adhesion of the enzyme on CNTs. In the next step, two fusion proteins were prepared for two different applications. An enhanced green fluorescent protein (eGFP) was fused to the Met-rich sequence in its C-termini to visualize the adhesion of the Met-rich-containing protein on carbon surfaces using fluorescent imaging. Next, another laccase that does not have a Met-rich peptide or a similar peptide was fused to a spore coat protein A (CotA) from Bacillus licheniformics through its N- and C-termini. Using CV measurements, both fusions showed ET capabilities superior to those of the native enzyme. Between the two, the C-terminus fusion had the best ET efficiency, which was expected according to the modeled enzyme orientation. In this study, molecular dynamics (MD) simulations were used to understand better the process of enzyme-oriented wiring on the electrode (Figure 3). A combination with CV measurements of a variety of mutants showed the effect of orientation on ET.

Figure 3.

Fusion of an oriented adsorption motif. The adsorption process of CueO on CNT (A) in the presence and (B) in the absence of a Met-rich motif, as snapshots from MD simulations. Reprinted from ref (50). Copyright 2021 American Chemical Society.

Not only Met-rich but also other peptides can be used for the attachment to CNT, like the one used for the attachment of P450 mono-oxygenase from Bacillus megaterium (BM3) to CNT through a C-terminally fused peptide: IFRLSWGTYFS.⁵¹ For the estimation of direct charge transfer efficiency from the electrode to the enzyme, SWV measurements were performed, showing peak currents for the fusion enzyme higher than those of the native one. Then, a direct electrochemical detection (DED) system was designed using a glassy carbon electrode, placed near the surface of the enzyme-electrode, and poised at a positive potential for the detection of the product produced by a DET-catalyzed reaction of the wired enzyme. The DED system was utilized for the characterization of the fusion enzyme DET abilities. It was shown that wiring through the CNT binding peptide results in an orientation that allows ET efficiency higher than that of the native enzyme. Such a DED system can aid in the improvement and investigation of other enzymatic systems where the DET-catalyzed reaction may yield different electroactive products. In a follow-up study, the electron transport pathways of the enzyme were investigated by mutating amino acids known to participate in the three possible ET pathways to alanine (an alanine screen approach).⁵² Several mutant fusion enzymes were prepared, while in each one, only one ET pathway out of the three was available, while each had the CNT-binding peptide fused at the C-terminus. Comparison of the DET abilities of the mutant enzymes was performed using CV and chronoamperometric detection, and the k_ET values of the different ET pathways were calculated using Laviron’s equation for a surface-confined protein. CV measurements.⁵³

A broadly used affinity tag in biochemistry for the purification of proteins using an immobilized metal affinity column is the poly histidine tag. Having a tag of six consecutive histidines with several imidazole residues allows forming a coordinative bond of the imidazole rings with divalent metal cations such as Ni²⁺, Co²⁺, Zn²⁺, and Cu²⁺. The histidine tag was utilized to control the orientation of a multicopper oxidase from Pyrobaculum aerophilum on a multiwalled carbon nanotube (MWCNT) electrode, modified with pyrene-nitrilotriacetic acid-Ni²⁺ (Ni-NTA), by fusing it to the N- or C-termini of the enzyme.²² CV measurements indirectly analyzed the orientation of the fusion enzyme by comparing the cathodic currents resulting from O₂ reduction. Higher currents were observed from the C-terminus fusion enzyme compared to the N-terminus fused version. Another way for the orientation determination was by measuring the hydrogen peroxide (H₂O₂) generated in the electrolyte solution. Multicopper oxidases have two electroactive sites referred to as T1Cu and T2/T3Cu sites. Accumulation of hydrogen peroxide is increased when the T2/T3Cu site of the enzyme is closer than the T1Cu site to the electrode. In this case, the T1Cu site is bypassed. Thus, O₂ undergoes only a 2e^– reduction to H₂O₂ rather than a 4e^– reduction process to H₂O. H₂O₂ concentrations were measured after the electrocatalytic reduction by a simple colorimetric assay that showed that the T1Cu site is closer to the electrode when the enzyme attachment to the electrode is through its C-terminus.

Using the N-terminus histidine tag fusion enzyme with the same wiring approach has enabled the investigation of an immobilization property that was coined “swing”.⁵⁴ The swing property is the actual mobility of the wired enzyme, its ability to move on its anchoring point, whereas ET occurs when the redox-active site approaches the electrode surface. A DET system was used and enabled, in this case, the investigation of a property that could not be measured otherwise. QCM with dissipation (QCMD) measurements were used to estimate the mobility of the wired enzyme, which correlates with the rate of oscillation decay, as evaluated by the dissipation value (ΔD). It was shown that lower enzyme densities on the electrode surface led to higher swing abilities, which resulted in higher DET efficiency.

Using such an affinity tag can allow the investigation of the effect of enzyme orientation on DET properties of an enzyme. That effect was investigated using a DET-type enzyme by fusing a histidine tag to several sites using different variants of the same enzyme. Pyrroloquinoline quinone-dependent aldehyde dehydrogenase (PQQ-AlDH) from Gluconobacter sp. was fused to a six-histidine tag at the N- or C-termini of either one of its three subunits, resulting in six different fusion mutants.⁵⁵ PQQ-AlDH has three different subunits as follows: a PQQ cofactor and a heme-c in one subunit and three heme-c molecules in the second. The third subunit is a small subunit that links the two together for a full-length correctly folded protein and does not have an electroactive role. A gold electrode functionalized with Ni-NTA has been used for attaching the different variants through the histidine tag. A combination of EIS and CV was used to verify the self-assembled monolayer on the gold electrode and its ability to bind metal ions. EIS showed an increase in ET resistance with the electrode modification with the Ni-NTA linker; then, for the verification of metal binding ability, CV was used after the attachment of Cu ions to the modified surface to detect Cu¹⁺ oxidation according to its known oxidation peak. Comparing the fusion enzyme variants using CV measurements and subsequent plotting of Lineweaver–Burk plots revealed that enzyme orientation is crucial for an efficient DET. Some of the variants did not display DET while wired through the N-terminus, whereas the C-terminus wiring of the same subunit resulted in DET.

As exemplified in this section, protein engineering by the fusion of different tags or domains can provide new tools for DET enhancement and ET mechanism investigation. The use of cytochrome domains can introduce DET abilities to enzymes lacking this ability or enhance DET in enzymes that already have this property. Attachment of enzymes through a specific tag enables the determination of enzyme orientation and the use of new techniques to investigate ET pathways.

2.3. Point Mutations

In recent years, the 3D structure of many proteins has been made available thanks to recent advances in protein structure determination technologies such as cryo-electron microscopy and the use of artificial intelligence software for protein structure predictions such as “alpha-fold”. The knowledge of protein structure is essential when point mutations are introduced into the amino acid sequence of a protein. Point mutations can be utilized to investigate ET by the replacement of aromatic amino acids in strategic sites that participate in ET through a hopping mechanism,^56,57 alter the sites participating in branched ET pathways in proteins like photosystems,^58,59 and modify the coordination sphere of a reactive metal ion that participates in ET.⁶⁰ Since our aim in this review is DET systems, we focus this part on mutations that can be rationally designed to allow site-specific immobilization to an electrode, which is less limited than adding a tag at the protein’s N- or C-termini while retaining enzyme activity. Considering the known structure of a protein, modification of residues from unstructured regions are, in most cases, permissive mutation sites that enable lowering the chances for enzyme inactivation by misfolding. In many cases, the target site will be mutated to a cysteine, an amino acid with a thiol side chain, to form a self-assembled monolayer on gold surfaces. The gold–sulfur bond is a spontaneous forming semicovalent bond that occurs at room temperature, creating a linkage between gold atoms to a thiol.⁶¹ When using cysteine or any other amino acid for surface immobilization, there is a need that no other residues of the same amino acid will be surface exposed to avoid nonspecific immobilization and undesired protein orientation.

Glucose oxidase from Aspergillus niger was mutated to have a single cysteine instead of each one of five different residues on the protein surface.²¹ Mutation sites were selected in strategic locations that are proximal to the FAD cofactor, at varying distances from the cofactor itself. The mutants were then conjugated through a single cysteine residue to maleimide-modified Au nanoparticles (Au-NPs) which were subsequently dropped on a gold electrode for electrochemical measurements. DET of the different mutants was tested using CV measurements in the presence of glucose, where only the mutant that had the shortest distance from the FAD (13.8 Å) exhibited DET. A control experiment in the absence of Au-NPs revealed that DET from the FAD to the electrode was enabled through the conjugated Au-NP. To verify that the FAD cofactor is responsible for the ET signal, OCP was measured in response to glucose additions, resulting in a steady-state potential close to the theoretical FAD redox potential. The shift in the potential indicated that another reaction originating from H₂O₂ production is occurring on the electrode surface. OCP, which was close to its theoretical value, could also infer that most of the enzymes in the sample were site-specifically conjugated to the Au-NPs. This study emphasizes the importance of distance for achieving DET and the use of potentiometric measurements for the ET pathway investigation.

Single cysteine mutants can be utilized for the investigation of the effect of enzyme orientation on DET when the 3D structure of the enzyme is known. Single cysteine mutants were introduced in a CDH from Myriococcum thermophilum, where seven different positions for a single cysteine variant were prepared, located on different regions of the catalytic domain, for the investigation of orientation-dependent DET (Figure 4).⁶² First, attachment efficiency to a maleimide-modified surface was tested using SPR measurements on a maleimide-modified gold-sputtered disc. All mutants showed highly efficient binding abilities, and a control experiment using the native enzyme showed no binding without having the surface-exposed cysteine residue. The mutants were then site-specifically attached to a glassy carbon electrode (GCE) through a maleimide functional group for further electrochemical investigation. CV measurements showed that all mutants had DET with different efficiencies, which was surprising and implied that the cytochrome domain mobility enables the rescue of DET ability even through long distances. The mobility factor of the cytochrome was then investigated using a combination of SPR measurements with molecular dynamic simulations and revealed the importance of cytochrome mobility for the interactions with the FAD cofactor at the catalytic domain and with the electrode surface. In this study, the point mutations on a DET-type enzyme enabled the investigation of a new factor, cytochrome mobility, using standard measurements and modeling tools.

Figure 4.

Use of point mutations for the orientation of CDH from Myriococcum thermophilum. Spheres indicate mutations to cysteine in seven distinct single cysteine mutants for maleimide coupling to an electrode. (A) Back, front, bottom, and top oriented mutants. (B) Left and right oriented mutants. Reprinted from ref (62). Copyright 2019 American Chemical Society.

Point mutations can also be used to enhance intramolecular ET (IET), which can result in increased DET efficiency. A DET-type enzyme FAD-GDH from Aspergillus flavus was fused to a b-type cytochrome domain of CDH from Phanerochaete chrysosporium, which was mutated to have a positively charged amino acid (lysine) for altered electronic potential distribution on the protein surface.⁶³ A docking simulation and surface electrostatic potential estimation were used for locating the relevant residues for point mutations. IET efficiency was then measured using absorbance spectrum analysis, tracking the change in the spectrum of the b-type cytochrome in response to glucose oxidation by the catalytic subunit, indicating heme reduction. After verifying the increased IET efficiency, chronoamperometric measurements were used to analyze DET abilities of the mutants compared to the native enzyme. This example highlights the importance and robustness of using computational models and simulations for detecting relevant sites for point mutations.

Protein engineering by point mutations, when combined with electrochemical measurements, can also be used for the investigation of enzyme states and conformations. Three mutants of bilirubin oxidase from Magnaporthe oryzae were designed to have a single cysteine on the surface with different distances from the active site.⁶⁴ The cysteine residue was used for the site-specific binding to a maleimide-modified MWCNT on GCE. Using CV measurements, the different mutants were analyzed, and a comparison between them demonstrated the effect of orientation on DET. In this study, the DET system was used to investigate how chloride ions affect the catalytic activity of the enzyme. Another use of point mutations on the DET system in this research was studying the different states (alternative resting or resting oxidized forms) of the immobilized enzyme in the presence of chloride ions, measured by the analysis of CVs from various experiments.

Overpotentials of an enzyme of interest can be reduced by point mutations targeting metal-coordinating residues in the enzyme active site. Point mutations were introduced into two different sites of CueO from E. coli, one was targeting the axial ligand of the T1Cu site and the other targeting a site that forms a hydrogen bond with the T1Cu coordinating site.⁶⁵ Both mutation sites were chosen according to previous studies on other laccases that showed their effect on overpotentials. The enzyme was entrapped on GCE covered with carbon aerogel and showed DET by CV measurements. A peak observed in the CV measurements was identified as the T1Cu site using spectroelectrochemical titration measurements. To demonstrate the effect of point mutations on the redox potential of the enzyme, voltammetry measurements were performed using a rotating disc electrode to eliminate mass transfer limitations. Point mutations have been shown to enable the tuning of the potential of T1Cu site. Two different mutants have been used to alter the midpoint potential of the native enzyme, which was measured to be 300 mV vs Ag/AgCl reference electrode. Changing the axial T1Cu ligand to leucine shifted the potential to 400 mV, and the mutation to glutamine decreased the overpotentials for oxygen reduction to 200 mV vs Ag/AgCl.

The use of MD simulations for protein engineering can help find a relevant site of mutagenesis. Point mutations on the axial ligand of T1Cu site of laccase from Trametes versicolor were investigated using MD for enhanced DET abilities.²³ At first, DET properties of the native enzyme on CNTs and graphene was estimated. The laccase–electrode interface was imaged using scanning electron microscopy (SEM) to visualize the enzyme adsorbed on the graphene and CNT electrodes. It was shown that the enzyme was immobilized uniformly on the graphene surface instead of the CNT surface, which resulted in agglomerated immobilization. Then CV and EIS were used for the estimation of DET and electrode enzyme coverage, respectively. Immobilization on the graphene electrode showed higher DET efficiency. Further understanding of the forces and factors enabling improved DET was achieved using MD simulations to investigate the enzyme immobilization process on the graphene surface. Then, detection of possible mutation sites that can increase DET abilities of that enzyme–electrode system was enabled by understanding the immobilization process from the simulation. In this particular study, no mutations were performed; MD simulations allowed the thorough understanding of the enzyme–electrode interface and immobilization process in a way that enabled planning of new mutants for optimal DET properties.

As can be seen from these examples, point mutations can be used for a variety of purposes. It can yield a specific anchoring point for enzyme immobilization in an oriented way, can be used for the tuning of enzymes’ redox potential, and can alter the total charge on a protein surface.

2.4. Incorporation of Unnatural Amino Acids

Unnatural amino acids (UAAs) are synthetic molecules with amino acids backbone, while their side chain is a functional group that is not one of the 20 natural building blocks of proteins.⁶⁶ The site-specific incorporation of UAAs into proteins requires an orthogonal translation system, which is a combination of an aminoacyl tRNA synthetase and a tRNA that enable the amino acylation of the tRNA with a UAA, by its incorporation into the elongating polypeptide during protein translation on the ribosome.⁶⁷ Site-specific incorporation can be done by suppressing one out of three possible stop codons (usually the amber stop codon—TAG) using a suppressor tRNA.⁶⁸ UAAs can be used in bioelectrochemistry as an orthogonal “chemical handle” by having a synthetic reactive group used for enzyme site-specific immobilization on an electrode.⁶⁹⁻⁷² The use of UAAs requires mutating the enzyme of interest. It is also recommended to have a 3D structure or structure prediction of the target enzyme to rationally identify relevant UAA incorporation sites.

Our team was the first to introduce UAAs for site-specific immobilization of enzymes to electrodes. An alcohol dehydrogenase II (ADHII) from Zymomonas mobilis was expressed on the surface of E. coli, with a site-specifically incorporated para-azido-l-phenylalanine (Az-Phe).⁷⁰ Three different mutants were prepared where the azide residue of the UAA was used for site-specific and bio-orthogonal reaction with an alkyne-containing redox-active linker, through a copper(I)-catalyzed azide–alkyne cycloaddition (“click”) reaction, while having a thiol on its other side, for a thiol–gold immobilization. AFM measurements were used to verify the attachment of the surface-displaying-enzyme bacteria to the gold surface after reacting them with the specific linker. TEM imaging was used for attachment verification by conjugating Au-NPs that can be bound to the free thiol residue on the surface-displayed enzymes. CV measurements were used for the characterization of the different mutants’ ET properties. The measurements showed that the site-specifically immobilized enzymes enabled ET to an electrode in response to ethanol addition in contrast with the nonimmobilized enzyme or an enzyme that was expressed in the absence of Az-Phe.

Apart from improvements in DET abilities, UAA incorporation for site-specific enzyme immobilization can be used for investigation of the ET mechanism of a protein of interest. UAAs were incorporated into the sequence of FAD-GDH derived from Burkholderia cepacia fused to a minimal cytochrome c domain (FGM) for its site-specific immobilization to an electrode to allow this investigation.¹⁵ The UAA incorporation sites were located in a nonstructured region of the enzyme, with distances of less than 20 Å from the FAD cofactor or the heme domain, to allow efficient DET in two different mutants (Figure 5). Another mutant, lacking the cytochrome domain, had a UAA incorporated at the same site close to the FAD cofactor for a better understanding of the ET pathways of the wired enzyme. The UAA used was propargyl-l-lysine (PrK), which has an alkyne side chain that can be reacted with an azide through a “click” reaction. Using a pyrene-azide linker enabled the site-specific attachment of the enzyme in an oriented way to a graphene-based electrode. First, the site-specific binding on the surface was investigated using AFM. It was shown that using PrK for site-specific wiring resulted in scattered proteins with the expected height of the enzyme on the surface, while using a nonspecific immobilization method resulted in protein aggregation. Then, differential pulse voltammetry (DPV) measurements were used as another verification for the wiring of the enzyme and the analysis of the electroactive sites participating in the ET processes. The results showed a significant difference in the redox peaks observed from the site-specifically wired enzyme compared to the nonspecific wired one. CV and chronoamperometry measurements were used for the determination of enzymes’ electrochemical kinetic constants. Since no significant peaks were observed in the CV measurements, Laviron’s method for calculating k_ET could not be used. To do so, we have used multistep amperometry (MSA) measurements, where the k_ET values could be extracted from the linearization of a monoexponential nonlinear regression of the observed current decay. Different orientations of the same enzyme showed different DET kinetics and allowed the investigation of the ET mechanism of the enzyme by the comparison of the various mutants. In another study, CueO from E. coli was site-specifically wired, using three mutants with PrK incorporated close to the Cu1 site or to the trinuclear copper cluster (TNC) site, and the third mutant was far from both.⁷² Enzyme orientation was analyzed using CV in an anaerobic atmosphere to measure the Faradaic currents originating from the electroactive sites while avoiding catalytic currents from oxygen reduction. MSA measurements were used to calculate k_ET values, and LSV measurements were used for DET comparison of the differently oriented mutants. In both cases, site-specific wiring using UAA enabled the orientation of the enzymes in a manner that resulted in ET mainly from the electrode-proximal electroactive site, which allowed the calculation of ET rate constants and the investigation of the ET pathway in an exact fashion.

Figure 5.

Use of UAAs for the site-specific immobilization of FGM on an electrode. (A) Modeled structure of FGM. PrK incorporation sites are indicated in cyan. The double-headed arrows indicate protein dimensions. (B) Distance measurements between the PrK residue and the relevant active site–FAD cofactor (i) and heme moiety (ii). (C) Expected orientation after site-specific wiring through PrK sites with pyrene-azide (PDAz) linker. Nonspecific immobilization using pyrene carboxylic acid (PCA) :FGM-S247PCA. Reprinted with permission from ref (15). Copyright 2021 Elsevier.

UAA incorporation can be used to lower the working potential of an enzyme by incorporating it in a close proximity to the cofactor located in the enzymes’ active site. Glycine oxidase from Bacillus subtilis and l-tryptophan oxidase from Chromobacterium violaceum were mutated to have site-specifically incorporated 2-amino-3-(4-mercaptophenyl)propanoic acid (or p-thiolphenylalanine, TF) with high proximity to the FAD binding site of the enzymes.⁷³ TF can be specifically reacted with boron-dipyrromethene (Bodipy373), which can be attached to the CNT surface through π–π stacking interactions (as shown by molecular dynamics simulations in this study), through a thiol–chlorine nucleophilic substitution (S-click) reaction. Enzyme attachment to the CNT surface was characterized using AFM measurements of the Bodipy373-conjugated enzyme compared to a nonconjugated one. The conjugated enzyme has shown DET characteristics when site-specifically wired and oriented, while the nonconjugated one did not, as measured by CV in the presence of a substrate. Investigation of the origin of oxidation current was performed by absorbance spectrum analysis and low-temperature electron paramagnetic resonance (EPR) measurements, where the results indicated it is the FAD cofactor. Using the site-specific wiring resulted in high sensitivity toward the substrate and low working potential that can decrease the effect of interfering molecules when used as a biosensor.

It was shown that if carefully chosen, the use of site-specific wiring through UAA can increase the stability of an enzyme–electrode system where 4-azido-l-phenylalanine (AzF) was site-specifically incorporated into small laccase from Streptomyces coelicolor with proximity to one of the active sites or far from both.⁷⁴ The enzyme was conjugated through the azido moiety of AzF to MWCNT functionalized with cyclo-octynyloxyethyl 1-pyrenebutyrate through copper-free cyclo-octyne–azide cycloaddition reaction. As a control, a 1-pyrenebutanoic acid succinimidyl ester linker was used, which can bind enzymes in a non-oriented nonspecific way. It was shown that the site-specific immobilization had superior DET kinetics compared to the nonspecific attachment, as measured by chronoamperometric measurements. In that case, one of the UAA positions, which had the highest ET efficiency, was also tested for enzyme–electrode stability over time. A site-specific wired enzyme–electrode was compared to the nonspecifically wired enzyme–electrode, and the chronoamperometric measurement was performed on day 0 and day 8. It was shown that the site-specifically wired enzyme preserved much more of its activity compared to the nonspecifically wired one.

The UAA 3-amino-l-tyrosine (NH2Tyr) was incorporated into myoglobin, an electron-transferring protein, from Physeter catodon to test the effect of pH on its ET properties.⁷⁵ The NH2Tyr site was used for the specific conjugation reaction to acryloyl functional group, which was attached to a gold electrode, through a Diels–Alder reaction, forming a benzoxazine ring. In this study, the surface was analyzed by AFM measurements, comparing site-specific and nonspecific immobilization, allowing the calculation of protein concentrations on the surface. To test the influence of pH on ET, CV measurements were conducted under different pH values. An electrocatalysis experiment, for the conversion of thioanisole to sulfoxide products, was performed to show the importance of having a protein monolayer for enhanced enzyme activity. Gas chromatography measurements were used for the analysis of sulfoxide product formation.

In all cases mentioned above, it could be seen that using UAAs can be efficiently utilized for site-specific wiring of enzymes to electrodes to promote DET and for the investigation of DET systems. The functional side group of UAAs decreases the nonspecific binding of contaminant proteins that can react with natural moieties but not with synthetic ones, which enables protein purification at the same time. However, incorporating UAAs into a protein results in decreased expression efficiency. In many cases, it is balanced with the fact that there is no need for additional protein purification steps that lead to protein losses from cell lysates due to a direct electrode attachment of the target protein from the lysate. Low protein expression levels could also be resolved by tuning the expression conditions toward high protein quantities.⁷⁶ Another advantage of UAAs is that they genetically encoded to be incorporated theoretically at any site, allowing increased flexibility when planning incorporation, site of immobilization or site of enzyme modification.

3. Conclusion

Herein, we have reviewed the use of protein engineering to enhance DET and the investigation of ET pathways. We have focused on how the available techniques for protein engineering enable the measurements of electrochemical parameters of various enzymes and DET characterization process of engineered proteins.

Protein truncations are a valuable tool for investigating ET pathways, as it allows the isolation of one electroactive domain and its investigation as an isolated ET domain. Another advantage of using truncations is that it exposes an ET domain, which can sometimes be buried within an insulating protein matrix, and enable DET to an electrode. Deleting some part of a protein can also be used to characterize the role of the deleted moiety by comparing its communication with an electrode to that of the native enzyme. Protein truncations also have some disadvantages—it is expected that deletion of some part of a protein will affect its complete 3D structure and may result in misfolded inactivated protein. Another point is that exposure of the active site of an enzyme can result in decreased substrate specificity, which leads to an unwanted reduction or oxidation of interfering molecules.

Protein fusions can enhance ET properties by adding an ET subunit or an affinity tag that affords specific wiring to the electrode, but is dependent on the linking peptide length and amino acids content. Fusion enzymes characterization is usually done by voltammetric techniques such as CV, DPV, and chronoamperometry. In some cases, absorbance spectrum measurements allow the analysis of IET, as in the fusion of a cytochrome domain. Adding an affinity tag to an enzyme enables its oriented attachment to an electrode with high proximity, which in some cases leads to increased ET rates. The characterization of the enzyme–electrode interface is characterised by techniques such as SPR, AFM, and SEM. The height and footprint of the enzyme particles on the surface is estimated and compared to the estimated size from the 3D structure. However, since these methods are not a direct determination of protein structure, these results should be taken with caution. The addition of an affinity tag or an ET domain is often restricted to the N- or C-termini of the enzyme, limiting the degrees of freedom for protein orientation relative to the electrode, as opposed to point mutations, where the site-specific wiring of enzymes is enabled.

Point mutations are usually used to result in a single amino acid with a reactive side chain on an exposed face of the protein, such as cysteine. Also, in that case, characterization of the enzyme–electrode interface is needed and performed using AFM or SPR. Preparing several mutants and their comparison is usually made to find the site that will afford the best ET properties using CV and chronoamperometry. Several mutants for different orientations on the electrodes can help decipher ET pathways of the enzyme and characterize the various active sites of the same protein. Being able to site-specifically attach an enzyme through a naturally occurring amino acid requires a highly purified enzyme sample with a single point of attachment on its surface to prevent protein contaminations or unwanted orientations. Using UAAs can increase the specificity of wiring using bio-orthogonal chemistry that will react with the unique synthetic residue that can be found only in the protein of interest.

UAA incorporation is usually used for the site-specific wiring of an enzyme to an electrode for ET properties investigation or to achieve DET. The UAA can be incorporated in almost any site of the protein sequence, which allows high flexibility in the enzyme engineering process. This advantage enables the broad investigation of ET processes in a native-like protein due to the possibility of incorporating the UAA in almost any site without changing other amino acids in the protein sequence. UAA incorporation also has some disadvantages—the UAA itself is sometimes costly, protein expression can be less efficient, resulting in lower protein yields, and having a 3D structure or prediction of the protein structure is mandatory for planning the desired incorporation site.

Investigation of protein–electrode interface is still not perfect, suffering from yet difficult to measure factors. One example is that the resolution of immobilization orientation determination is insufficient. Imaging or demonstrating enzyme orientation in high resolution is vital for measuring the distance between an enzyme and an electrode or an enzyme active site and the electrode. Measuring such factors will allow the design of new linkers and enzymes in a rational way for improved wiring. Overcoming these limitations is currently done by MD simulations, but still no direct measurement is available. Another factor is the ability to test the enzyme folding on the electrode surface; is the enzyme correctly folded on the electrode surface after immobilization? Is the protein misfolded due to interactions between the electrode material and the protein surface, or is its core surface-exposed? Those questions can be answered only in an indirect manner, such as testing enzyme–substrate specificity and its thermodynamic and kinetic properties, or by the combination of modeled structure and AFM measurements.

Looking toward the near future, MD simulations and the high availability of protein structure and structure predictions will accelerate protein engineering efforts. Tools of structure predictions are becoming more accurate. They can be used in MD simulations, which can then direct the research to the relevant amino acids that can be replaced with natural or unnatural amino acids. Understanding the enzyme–surface interface can contribute to the enhancement of ET and, in some cases, achieve DET. UAAs and their incorporation into proteins are becoming more common and are in a constant improvement in the direction of more efficient incorporation and expression levels. Using these advanced tools will allow the understanding of redox systems and improve bioelectrochemical systems.

Glossary

Vocabulary

direct electron transfer (DET)

an electron transfer path where electrons are transferred directly from an enzyme to an electrode without additional mediating molecules

square-wave voltammetry (SWV)

an advanced measurement technique where the electrode potential is biased between high and low potentials with a controlled frequency factor; the current is measured after a constant time interval from the potential change, allowing more accurate measurements of Faradaic currents without regeneration of the electrode’s diffusion layer

fusion protein

an engineered protein where an additional electron transferring domain or affinity tag was added to its genetic sequence

deletion/truncation mutant

a protein that some of its encoding DNA sequence was erased for the partial or complete removal of a domain; truncation is a specific case of deletion where the sequence is erased from the end of the fully elongated protein

unnatural amino acids (UAAs)

amino acids that are not encoded in natural proteins, constantly being used as a tool in synthetic biology such as bioorthogonal chemical handles, biophysical probes, metal chelators, fluorophores, catalysts, etc.

The authors declare no competing financial interest.