Bekir Engin Eser,Yan Zhang,Li Zong,2,Zheng Guo,*

1 Department of Engineering,Faculty of Technical Sciences,Aarhus University,Denmark

2 Key Laboratory for Molecular Enzymology and Engineering,The Ministry of Education,School of Life Science,Jilin University,Changchun 130021,China

ABSTRACT Cytochrome P450s(CYPs)are ubiquitously found in all kingdoms of life,playing important role in various biosynthetic pathways as well as degradative pathways;accordingly find applications in a vast variety of areas from organic synthesis and drug metabolite production to modification of biomaterials and bioremediation.Significantly,CYPs catalyze chemically challenging C-H and C-C activation reactions using a reactive high-valent iron-oxo intermediate generated upon dioxygen activation at their heme center,while the other oxygen atom is reduced to the level of water by electrons provided through a reductase partner protein.Self-sufficient CYPs,encoding their heme domain and reductase protein in a single polypeptide,facilitate increased catalytic efficiency and render a less complicated system to work with.The self-sufficient CYP enzyme from CYP102A family (CYP102A1,BM3) is among the earliest and most-investigated model enzymes for mechanistic and structural studies as well as for biotechnological applications.An increasing number of self-sufficient CYPs from the same CYP102 family and from other families have also been reported in last decade.In this review,we introduce chemistry and biology of CYPs,followed by an overview of the characteristics of self-sufficient CYPs and representative reactions.Enzyme engineering efforts leading to novel self-sufficient CYP variants that can catalyze synthetically useful natural and non-natural (nature-mimicking) reactions are highlighted.Lastly,the strategy and efforts that aim to circumvent the challenges for improved thermostability,regio-and enantioselectivity,and total turnover number;associated with practical use of self-sufficient CYPs are reviewed.

Keywords:Biocatalysis Heme enzymes C-H activation Cytochrome P450s Self-sufficient P450s P450 BM3

1.Introduction

Enzymes are Nature’s powerful catalysts that drive biochemical transformations with sufficiently fast rates to sustain life.Enzymes have several advantages over chemical catalysts,such as;catalytic efficiency,excellent chemo-,regio-and stereoselectivity,biocompatibility and biodegradability [1–3].Moreover,they can catalyze reactions at ambient temperatures in aqueous environments,without need for harsh organics and toxic chemicals.Cytochrome P450s(CYPs) are one of the largest group of enzymes present in Nature,widely distributed among almost all forms of life from bacteria to higher organisms [4–7].These enzymes use an iron-porphyrin(heme)cofactor to carry out monooxygenation and oxidation reactions,and are part of a larger group of proteins known as hemedependent enzymes.First characterized almost 60 years ago,CYPs are the most widely investigated group of iron-dependent enzymes[8–12].

CYPs are of great interest in industry due to their ability to catalyze a wide variety of chemically challenging reactions,the most prominent being hydroxylation of unactivated C-H bonds[6,8,13].Also known as oxy-functionalization,this reaction is of great importance in organic chemistry and synthesis,since introduction of oxygen functionality opens up many possibilities for further synthetic routes for a molecule that is otherwise inert[14–16].This enables access to high-value molecules such as chemical synthons and pharmaceutical precursors.There has been extensive research efforts on C-H functionalization using free radicals,organometallic complexes and enzymes [8,14].However,an important challenge in C-H functionalization is still achieving the desired regio-and stereoselectivity [6,17,18].Unlike their chemical counterparts,CYPs catalyze these reactions with desired selectivity under mild reaction conditions,which is specifically important for areas such as pharmaceutical industry [19].For example,one of their most common and synthetically useful reactions is regio-and stereoselective hydroxylation of fatty acids for synthesis of value-added hydroxy fatty acid products with broad applications in plastics,cosmetics and pharmaceutical industry [18].Another common reaction is steroid oxidation that generates valuable drug metabolites and pharmaceutical intermediates [20,21].In some cases,CYPs reach over 99%enantiopurity for the products that they generate,with negligible amounts of by-products,making them much more advantageous compared to chemical methods [22–24].In addition,CYPs possess a versatile substrate range and catalyze various reactions including epoxidation,heteroatom oxidations,dealkylations,C-C bond scission and oxidative decarboxylations [13,18,25](Fig.1).Some CYPs,especially those from microorganisms,naturally prefer long chain hydrophobic molecules such as fatty acids or alkanes,whereas other CYPs have been shown to take a variety of substrate structures including bulky aromatics and complex organic molecules[13,18,26,27].Many of CYP catalyzed reactions are valuable transformations for biotechnology and span a wide-range of potential applications from synthesis of fine chemicals and biofuel production to bioremediation.

In addition to the versatility of native CYPs,advances in enzyme engineering methods enabled researchers to obtain enzyme variants that are well suited for biotechnology [17,26,28].Furthermore,engineered CYPs were reported to catalyze reactions that are unprecedented in nature,such as selective olefincyclopropanation and C-H amination [3,24](Fig.1).

In this review,we will briefly introduce physiological roles of CYPs as well as their chemical mechanism and classification.Then,this review will mainly focus on self-sufficient CYPs,which is the most promising subgroup for industrial applications due to their reductase partner proteins being fused to their catalytic heme domains,greatly simplifying the biocatalytic conversion process.Different groups of self-sufficient CYPs will be introduced with a focus on the biotransformations that they catalyze.Lastly,we will present research efforts directed towards expanding the application potential of CYPs in biotechnology.

2.Biology &Chemistry of Cytochrome P450s

There are over forty thousand (41514) protein sequences belonging to CYP enzymes in Cytochrome P450 Engineering Database (CYPED) as of September 2020,organized into 317 families and 1701 subfamilies (https://cyped.biocatnet.de) [29].18,851 of those CYPs belong to bacterial organisms and the rest (34196)belong to eukaryotic organisms that include animals,plants and fungi.

Fig.1.Natural and non-natural reactions catalyzed by native or engineered CYPs.

CYPs have various physiological roles in organisms ranging from bacteria to humans.There are 57 human CYP proteins that play important roles in both anabolic and catabolic processes that include steroid and lipid metabolisms as well as detoxification of xenobiotics (e.g.drug and toxins) [30].Plant CYPs are involved in the biosynthesis of molecules such as hormones,defense molecules and secondary metabolites.Fungal CYPs play roles in primary and secondary metabolite synthesis as well as denitrification[31,32].Presence of CYPs in bacteria shows great variation;whereas some bacteria have high number of various CYPs,such as streptomycetes(S.avermiltilispossess 33 CYP genes)[33],some bacteria such asE.colido not possess any CYP encoding gene[7].In bacteria,CYPs are involved in processes such as fatty acid metabolism,steroid and secondary metabolite biosynthesis,as well as in carbon assimilation required for growth[7].Bacterial CYPs are generally soluble enzymes,thus are convenient for structural and functional studies.This led to the use of bacterial CYPs as model enzymes for better understanding of human CYPs as well as exploitation of CYPs for cell-free biocatalysis [7,34].CYP101A1(P450cam) and CYP102A1 (BM3) are the most studied bacterial CYPs as model proteins [35–39].

CYPs are heme-dependent proteins.They contain an iron atom that has been coordinated to a porphyrin molecule through electron rich nitrogen atoms at four equatorial coordination sites (tetradentate) [9](Fig.2).In general a protein derived cysteine thiolate acts as the axial ligand (distal position).The other axial position (proximal) is occupied by water in resting CYP,but becomes vacant for oxygen binding when the catalysis is initiated.Although substrates and reaction outcomes might be different,most CYPs share a common reaction mechanism throughout their catalytic cycle [8,25,27,41].This common reaction motif includes activation of molecular oxygen by the heme group at early stages of catalysis;the activated oxygen react with the substrates at later stages.Even though reactions of molecular oxygen with organic molecules are thermodynamically favorable,these reactions exhibit sluggish kinetics due to spin mismatch,and thus require enzymatic activation [42].

The catalytic cycle of CYPs starts with binding of the substrate,which primes the heme center for reaction with molecular oxygen(Fig.3)[25,43].This priming includes reduction of ferric iron to ferrous form by the redox partner and displacement of the bound water molecule to open up a coordination site for oxygen binding.The next step after binding of molecular oxygen is the electron transfer from ferrous iron to dioxygen generating an iron(III)-superoxo intermediate.Additional electron transfer from reducing partner protein leads to formation of ferric peroxo intermediate that captures a proton to form ferric hydroperoxo (compound 0).Upon addition of a second proton,heterolytic cleavage of compound 0 takes place giving rise to a reactive Fe(IV)-oxo porphyrin Π radical cation intermediate (compound I;also shown as Fe(V)-oxo) [44],where the other oxygen is reduced to level of water.Fe(IV)-oxo,a powerful oxidant,is the reactive intermediate that abstracts a hydrogen atom from the substrate leading to ferrylhydroxo (Fe(IV)-OH) intermediate (compound II) and a substrate radical [45].A radical rebound step between compound II and the substrate radical gives the hydroxylated product.In some cases,instead of a radical rebound step,the substrate radical can lead to decarboxylation and desaturation reactions upon abstraction of a second hydrogen atom[25,46](Fig.1).For a full catalytic cycle,two electrons are required in total,which end up in the water upon reduction of one of the oxygen atoms (Fig.3).In some CYPs,also called peroxygenases,H2O2is used as both the oxygen and electron source rather than molecular oxygen[47].In this case,the catalytic cycle becomes shorter,directly starting from ferric hydroperoxo intermediate (Fig.3).This route is called peroxide shunt pathway [45].

Fig.2.Overall structure of CYP116B46 (PDB:6LAA).The upper line diagram shows the primary structure of CYP116B46 with the amino acid numbers indicated for each region.The polypeptide structure is presented as a cartoon model,with the three domains and connecting linkers in different colors.The heme domain,reductase domain,ferredoxin domain and linkers are colored in light blue,light orange,yellow orange and gray,separately.The bound ligands that Heme and FMN shown in stick models and Fe2S2shown in sphere model are labeled [40].

Fig.3.The catalytic cycle of cytochrome P450 (The grey arrows illustrate the peroxide shunt pathway and P450 uncoupling) [43].

3.Classes of Cytochrome P450s

CYPs are classified in various ways depending on their sequence identity,organization of their electron transfer systems,cellular localization,membrane association and solubility [48].The main classification,which also serves as a means for nomenclature,is the distribution of CYPs into families and subfamilies based on their sequence.Each CYP family contain proteins with over 40%amino acid sequence identity,whereas proteins with more than 55% identity are placed within a subfamily [48].CYPs are named with numbers and letters that indicate their family and subfamily.For example,CYP102A1 is the protein number 1 in the subfamily of A,which is under the CYP family of 102.CYPs are systematically compiled in dedicated databases [29,48].

Another notable classification system is according to the type and the organization of electron transfer partners that are responsible for carrying two electrons from the ultimate electron supplier NAD(P)H to the heme center,one electron at a time (Fig.2).For some CYPs,this system can be a single reductase protein with multiple domains while for some others it will be multiple separate proteins acting together.Although there are different classification systems used with varying number of classes[49–52],we prefer to present the classification with 5 groups (Fig.4).In this classification,class I represents two separate component reductase system made up of a ferredoxin reductase and a ferredoxin protein.Class I includes mainly bacterial and mitochondrial CYPs,which are soluble.Class II represents eukaryotic microsomal (membraneassociated)CYPs with a single component reductase unit(CPR protein).Class III consist of CYPs with similar reductase component as class II(CPR),however their CPR is fused to the heme domain by a short linker.Class IV enzymes have the same reductase component as class I systems,but the ferredoxin and ferredoxin reductase functionalities are organized as subunits in a single reductase domain,which is fused to heme domain forming one single polypeptide chain similar to class III CYPs.Class V includes nitric oxide reductase (P450nor),which can directly use NAD(P)H (without any reductase protein) for denitrification reactions in several fungal species [32].

4.Self-Sufficient CYP Families:Reactions Catalyzed and Their Potential for Applications

In most studied CYP systems,heme proteins and redox partners are separate polypeptides.They associate with each other during catalysis to enable the transfer of electrons.This also means that both heme protein and the redox proteins should be included together forin vitroreactions.However,“self-sufficient CYPs”(class III and IV,Fig.4)have their catalytic heme domain and redox proteins fused naturally in a single polypeptide chain.These CYPs do not require any additional protein and just the presence of NADH/NADPH,along with molecular oxygen and the substrate,is sufficient to support activity (Table 1).

Fig.4.Classification of cytochrome P450 enzymes based on redox partners.FdR,Ferredoxin reductase;Fdx,Ferredoxin;CPR,cytochrome P450 reductase;FMN,Flavin mononucleotide;FeS,iron-sulfur cluster.

Table 1 Self-sufficient CYPs (natural heme/reductase fusion) characterized to date.References for each enzyme is given in the text

4.1.CYP102 family

CYP102 family is constituted of bacterial self-sufficient CYPs possessing a heme domain fused to a reductase domain that consists of eukaryotic-like FAD and FMN subdomains (Fig.4,Class III).There are 29 subfamilies assigned to CYP102 family with a total of 694 protein sequences (https://cyped.biocatnet.de).CYP102A1 from the soil bacteriumBacillus Megaterium,widely known as BM3,was the first enzyme of the family to be characterized in 1970s by Fulco and coworkers.Since then it has been the model enzyme of CYPs for structural and mechanistic studies as well as for biotechnological applications [38,53,54].BM3 is composed of a 55000 catalytic heme domain and a 65000 reductase domain.Although the main reaction of BM3 is hydroxylation of the subterminal(ω-1,ω-2 and ω-3)positions of medium to long chain saturated and unsaturated fatty acids,it can also catalyze hydroxylation of various other compounds including fatty amides and fatty alcohols as well as epoxidation of unsaturated fatty acids [38,55–57](Fig.5A).The self-sufficiency,substrate promiscuity and very high-turnover numbers of BM3(e.g.17.000 s-1for arachidonic acid hydroxylation) attracted great attention from biotechnologists[49,58–60].Various studies on BM3 aimed enzymatic production of drug metabolites or drug precursors for pharmaceutical purposes and synthesis of functionalized fatty acids(e.g.hydroxy fatty acids) or alkanes (e.g.fatty alcohols) as high value oleochemicals for applications in materials,flavor and cosmetics industry.Many engineering studies were applied to improve and diversify catalytic utility of BM3,as described in the following sections of this review.

In addition to BM3,various other CYP102 members were characterized (Table 1).These include CYP102A2 and CYP102A3 fromBacillus subtilis[61],CYP102A5 fromBacillus cereus[62],CYP102A7 fromBacillus licheniformis[63]and CYP102D1 fromStreptomyces avermitilis[64].Recently,two new CYP102 family enzymes isolated fromKtedonobacter racemifer,named as Krac9955 and Krac0936,were reported to hydroxylate fatty acids [65].Although Krac0936 shows similar regioselectivity compared to BM3 and many other CYP102 members by hydroxylating ω-1 to ω-3 positions,Krac9955 has a unique regioselectivity profile where hydroxylation of ω-4,ω-5 and ω-6 positions take place to a significant extent with C13-C15 chain length saturated fatty acids.Moreover,Krac9955 prefers alkybenzoic and alkyloxybenzoic acids over fatty acid substrates [66].In another study,a genome mining approach aimed for identification of homologs with moderate to low sequence identity to CYP102A and other self-sufficient families in Class III[67].This led to characterization of two new CYP102 members;CYP102A25 from the alkaliphileB.marmarensisand CYP102A26 from the halophilic bacteriumP.halophilus,with different substrate and regioselectivity profiles.CYP102A25 exhibited substantially higher activity towards myristoleic acid hydroxylation in comparison to other medium to long chain fatty acids,whereas CYP102A26 showed diverse substrate scope with similar activity levels towards C14-C18 saturated and unsaturated fatty acids.Both enzymes hydroxylated ω-1 to ω-3 positions [67].Another CYP102 member,CYP102G4 fromStreptomyces cattleya,exhibits a unique hydroxylation activity towards aromatic compounds,such as flavone,benzophenone and indole,in addition to canonical fatty acid substrates [68].E.colicells expressing CYP102G4 can efficiently produce the eco-friendly indigo dye from indole,which makes it the only bacterial wild-type CYP that is able to catalyze this reaction[69].Structural analysis based on a homology model indicated that the enzyme has a larger active-site cavity than typical CYP102s,enabling acceptance of bulky aromatic substrates [68].In a recent study,CYP102D2 fromDeinococcus apachensis,has been characterized [70].This enzyme has been shown to carry out benzylic hydroxylation of a variety of compounds including methyl 2-phenylacetate,allylbenzene,and ethylbenzene to produce the corresponding chiral alcohols with high to excellent enantioselectivities (86%–99% ee).Unlike other CYP102 family enzymes that exhibit strict cofactor specificity towards NADPH,CYP102G4 and CYP102D2 can efficiently use both NADH and NADPH as their electron donating cofactors [68,70].More recently,characterization of two CYP102 members,both fromBacillus amyloliquefaciensDSM 7,has been performed [71].A detailed investigation of product scope with varying chain length fatty acids indicated distinct regioselectivity profiles for hydroxylation.While one of the enzymes (BAMF0695) perform hydroxylation mainly at ω-1,ω-2 and ω-3 positions of C11 to C18 chainlength fatty acids,the other homolog (BAMF2522) exhibited hydroxylation for all positions from ω-1 to ω-7 for palmitic acid,with ω-7 product making up over 20% of total product content.This makes BAMF2522 the only wild-type CYP102 enzyme that is able to hydroxylate ω-6 and ω-7 positions of palmitic acid.Moreover,BAMF0695 can generate ω-1 hydroxy product of arachidonic acid(19-HETE),which has not been reported earlier with any nonmammalian CYP [71].(S)-stereoisomer of 19-HETE is known to play a role in the regulation of vascular function [72].

Fig.5.Selected reactions of wild-type and engineered self-sufficient CYPs mentioned in the text.

4.2.CYP116 family

CYP116 is another bacterial family of CYPs that have their reductase partners fused to their catalytic domain.The type and organization of their redox proteins is different from those of CYP102 family (Fig.4,Class IV).CYP116 enzymes use two reductase protein domains attached to their catalytic heme domain;a 2Fe-2S cluster containing ferredoxin and a FMN containing flavin domains [42].Electrons from NADPH are initially transferred to FMN before being transferred to the heme domain through ferredoxin.In the recent decade,various CYP116 members that exhibit diverse substrate scope and reaction profiles were characterized(Table 1).

CYP116B2 (also known as P450Rhf) fromRhodococcussp.and CYP116B3 fromRhodococcus ruberDSM 44319 are the earliest characterized members of CYP116 family [73–75].CYP116B2 can catalyzeO-dealkylation of 7-ethoxycoumarin and various substituted alkyl aryl ethers [76].In addition to dealkylation of 7-ethoxycoumarin,CYP116B3 is capable of hydroxylating many aromatic compounds including napthalene,toluene,ethyl benzene and fluorine.Hydroxylation takes place at the ring position in the absence of a side chain on the substrate,but at benzylic position if an alkyl side chain is present[75].Recently,CYP116B2 has been shown to catalyze biotransformation of the anti-inflammatory drug diclofenac to 5-hydroxydiclofenac,an important human metabolite required as a standard for environmental and pharmaceutical studies [77](Fig.5B).In this study,researchers have demonstrated use ofE.coliwhole cells expressing CYP116B2 for the preparative scale synthesis of 5-hydroxydiclofenac with high regioselectivityandgoodyield.CYP116B1fromCupriavidus metalliduranshas been demonstrated to hydroxylate the herbicidesS-ethyl dipropylthiocarbamate (EPTC) andS-propyl dipropylthiocarbamate (vernolate) and toN-dealkylate vernolate[78].CYP116B4 fromLabrenzia aggregatais another member of the CYP116 family that exhibits broad substrate spectrum for hydroxylation reactions ranging from alkylbenzenes and aromatic bicyclic molecules to terpenoids [79].Moreover,other reaction types including epoxidation,O-dealkylation and sulfoxidation were observed.CYP116B5 is a more recently characterized enzyme fromAcinetobacter radioresistensand catalyzes hydroxylation of long and medium chain alkanes[80].Growth studies(alkanes were supplied as sole energy source to the bacteria expressing the enzyme)indicated that long chain alkanes (C24 and C36) are subjected to subterminal hydroxylation whereas medium chain (C14 and C16)alkanes were hydroxylated at terminal positions.Partially purified CYP116B5 converted tetradecane into 1-tetradecanol.The structure of the heme domain of CYP116B5 has been recently solved[81].In a more recent study,researchers characterized a panel of thermostable CYP116B enzymes that exhibit broad substrate scope and diverse reaction range that includes hydroxylation,sulfoxidation,demethylation and epoxidation of various aromatic compounds [82].These enzymes include CYP116B65 fromA.thermoflava,CYP116B64 fromA.xiamenense,CYP116B63 fromJ.thermophila,CYP116B29 fromT.bisporaand CYP116B46 fromT.thermophilus.The enzymes also showed enhanced stability and expression levels compared to their mesophilic counterparts.Further screening of these newly discovered CYP116 members towards C8-C12 fatty acids led to demonstration of CYP116B46 as being capable of efficient hydroxylation of decanoic acid at C5(γ) carbon with high enantio-and regioselectivity.The resulting(S)-5-hydroxydecanoic acid can be further lactonized under acidic conditions to yield (S)-δ-decalactone,a valuable fragrance compound,with 90% C5 selectivity and over 90% enantiomeric excess(ee)in preparative scale [22].This represents the first example of a selective enzymatic C5 hydroxylation of a saturated fatty acid and a promising approach towards fatty acid valorization.Recently solved crystal structures of the CYP116B46 heme domain[83]and the full-length protein with its reductase partners [42]are expected to guide enzyme engineering efforts towards further optimization of CYP116 family for biotechnological applications.CYP116B62 is another recently characterized enzyme identified by genome mining from the halophilic bacteriumHalomonas sp.,a salt-tolerant microorganism [84].This enzyme exhibits high expression levels and diverse substrate scope making it advantageous for synthetic purposes.CYP116B62 mainly catalyzes demethylation of various aromatic substrates,but also can carry out hydroxylation and sulfoxidation reactions.

4.3.CYP505 family

CYP505 family of enzymes are eukaryotic self-sufficient CYPs from fungi [31,32].They are similar to bacterial CYP102 family in terms of function and organization of their reductase partner(Fig.4,Class III),thus are considered their fungal orthologues.Similar to their bacterial counterparts,many of the CYP505 family members catalyze hydroxylation of fatty acids at sub-terminal positions (ω-1 to ω-3).CYP505A1 (also known as P450foxy) fromFusarium oxysporumwas the first family member to be characterized in mid-1990s [85],and since then was studied in detail[32,86,87].The enzyme has very similar catalytic properties to BM3.Although CYP505A1 was shown to be a membrane associated protein in its native organism,its recombinant expression in yeast yielded active protein in soluble form [32,86].CYP505B1 fromFusarium verticillioideshas been proposed to play a role as a hydroxylase in the biosynthesis of the mycotoxin fumanisin [88].Another CYP505 has been discovered through screening efforts of a set of bacterial and fungal CYPs with high sequence resemblance to BM3 [87].Among these,CYP505X fromAspergillus fumigatusdemonstrated notable stability and organic solvent tolerance[87].RestingE.coliwhole cell preparations of this enzyme have been demonstrated to catalyze efficient oxidation of capsaicin,an active component of chili peppers and an active pharmaceutical ingredient of topical analgesics [89](Fig.5C).Both hydroxylated and an epoxylated products were generated at a rate of 4.4 μmol·min-1with 85% total conversion.Biotransformations for the hydroxylation of the drugs chlorzoxasone and ibuprofen were also shown byP.pastoriswhole cells expressing CYP505X,with over 25% conversion levels [90].

In recent years,various other members of CYP505 family have been reported (Table 1).CYP505A30 fromMyceliophthora thermophileis a fatty acid hydroxylase with similar characteristics as CYP505A1,but with higher preference for ω-1 position [91].The enzyme can also hydroxylate pharmaceutical agents including ibuprofen and capsaicin [92].Although the enzyme is from thermophilic fungus,it exhibited only moderate thermostability(Tm=58 °C),but a high tolerance to cosolvents and a broad pH range were observed [91,92].CYP505D6 from white-rot fungusPhanerochaete chrysosporiumis another recently identified enzyme that exhibited hydroxylation activity on fatty alcohols as well as fatty acids with diverse regiospecificity [93].Hydroxylation from ω-1 to ω-6 positions have been observed with C9-C15 fatty alcohols.Hydroxylation of ω-7 position was also noted with 1-dodecanol.Moreover,CYP505D6 is also capable of hydroxylating polycyclic aromatics such as naphthalene,consistent with the lignin-degrading activity of its native organismPhanerochaete chrysosporium[93].

CYP505D13 is another family member characterized recently.This enzyme fromGanoderma lucidumenabled biosynthesis of squalene-type triterpenoids (linearized triterpenoids with important bioactivities) inS.cerevisiaewhen recombinantly expressed[94].According toin vitroassays,CYP505D13 has been proposed to catalyze epoxidation of 2,3-oxidosqualene,although no direct evidence has been provided so far [94].

Very recently,characterization of CYP505E3 identified inAspergillus terreushas been reported[95].The enzyme has novel regioselectivity and exhibits hydroxylation of the in-chain positions,including ω-7 and ω-8,of its fatty substrates.ω-7 hydroxylated species is the major product for all alkanes from C10 to C16 chain length (over 70% selectivity) and for some of the fatty acid and fatty alcohol substrates.Hydroxylation of dodecanoic acid or 1-dodecanol at ω-7 position led to demonstration of promising routes for production of δ-dodecalactone,a high value flavor compound,through lactonization reactions following hydroxylation by CYP505E3 [95](Fig.5D).

4.4.Other self-sufficient members

In addition to the self-sufficient families described above,there are two other CYP families that are generally classified as selfsufficient,although their redox properties are quite different.One such example is CYP74 family,which are plant enzymes with unique electron transfer properties that require neither oxygen nor reducing equivalents from a reductase partner and/or NADPH[96].Instead,they use the acyl hydroperoxide part of their substrate as an oxygen donor.CYP74A,allene oxide synthase;CYP74B and CYP74C,fatty acid hydroperoxide lyase;and CYP74D,divinyl ether synthase are the family members studied until this date[96].

A second unique group is the CYP55 family,which are nitric oxide (NO) reductases from fungi that play an important role in denitrification process in Nature [97].They are considered selfsufficient enzymes since they are not associated with any reductase protein and only require NAD(P)H for their activity (Fig.4,Class V).Their primary reaction is the reduction of NO to N2O using electrons transferred to the heme center directly from NAD(P)H.Although CYP55 enzymes have not been yet explored for synthetic purposes,their application potential for other purposes,such as NO quantification and NO scavenging,have been demonstrated[98,99].

5.Efforts for Expanding Use and Application Potential of Selfsufficient CYPs in Biotechnology

5.1.Enzyme engineering

An important approach to reach excellent levels of desired enzymatic performance,such as activity levels,selectivity and stability is engineering of enzymes through modification of protein sequence using molecular biology tools [125,126].Targeted (rational design),random and semi-rational engineering methods can be applied depending on the desired purpose and available information on the target enzymes.Almost 60 years of research on CYPs led to accumulation of vast amount of information on the catalytic mechanisms,homologous sequences,assay methods and sequence-structure–function relationships in various CYPs[10,13].Protein engineering,in the light of this information,enables one to alter various properties of CYPs.Even no information is available on a CYP enzyme,protein sequence itself is sufficient to carry out efficient random engineering in most cases [3].

Due to their distinct advantages for biotechnology,selfsufficient CYPs have been the focus of intense protein engineering efforts among all CYP families.In general,engineering studies have been carried out with the purpose to increase activity,to increase stability(solvent,pH,process conditions),to tune the enzymes for a desired substrate and product outcome and,in some cases,to achieve non-natural reactions.Different engineering strategies have been applied including rational design,directed evolution and reductase partner engineering.BM3 has been at the center of most engineering studies due to its high catalytic efficiency,easy expression as well as availability of structural data[38,127].Efforts on this enzyme led to high-value products such as ω-hydroxy fatty acids,chiral alcohols and drug metabolites.We compiled below some of the interesting engineering studies with application potential.

Various engineering efforts aimed to alter the regioselectivity of BM3 for the hydroxylation of fatty acids.In one study,Urlacher and coworkers targeted substrate binding residues by site-directed mutagenesis to affect the size of the binding pocket as well as the interactions with carboxylate group of the substrate in order to shift the regioselectivity towards the carboxylate end of the molecule[100].Two resulting variants with triple mutations were able hydroxylate lauric acid at ω-4 to ω-9 positions,which are not accessible by wild-type BM3,albeit at the expense of significant reduction in activity levels and coupling efficiency (coupling of product formation to NAD(P)H oxidation).ω-7 and ω-8 hydroxylated products are biotechnologically important since they can be used for production of γ-and δ-lactones as flavor agents [101].In another study aiming regioselectivity shift in BM3,a combination of random and targeted mutagenesis (through iterative cycles)was used to evolve mutant variants that are capable of terminaloxidation of palmitic acid.The resulting ω-hydroxy fatty acid is a high-value product,which can be used as a monomer in bioplastic production.A BM3 mutant with 11 mutations was the best variant,demonstrating ω-hydroxy product formation with 74%regioselectivity [102](Fig.5E).

Table 2 Summary of common approaches for expanding use of CYPs in biotechnology

Engineering studies were also conducted to obtain variants that can catalyze alcohol formation from various substrates in a regioand enantioselective manner.Such functionalized molecules are useful materials and building blocks in manufacturing industry,for example in the production of detergents,pharmaceuticals and oleochemicals.In one study,researchers identified A74G/L188Q double mutant of BM3 from a library of 65 mutants generated by mutation of five active-site residues.This mutant can efficiently carry out asymmetric allylic hydroxylation of a broad range of ωalkenoic acids,such as 7-octenoic and 10-undecenoic acids and their esters in a selective manner,generatingS-configured secondary allylic alcohols,which otherwise requires multiple steps or leads to poor stereoselectivity by chemical methods [23].The practical potential of this double mutant was demonstrated with preparative scale reactions where yields up to 80% and excellent enantioselectivities (ee over 99%) were observed for production of chiral (ω-2)-hydroxy-ω-alkenoic acids,which are important building blocks for the synthesis of biologically active compounds including antibiotics and oxylipins.BM3 was also subjected to engineering for oxidation of alkanes to desired alcohols in a selective manner [103].A combination of rational design and directed evolution approaches resulted in BM3 variants that can efficiently produceS-2-octanol andR-2 alcohols (with 7 or more carbons)with over 40% ee and with high turnover numbers [104].Multiple evolutionary cycles were further applied to the promising variants in the same library,which finally generated a variant with greatly enhanced hydroxylation activity towards propane (9000-fold increase inkcat/Km) [104,105].Similar engineering studies also led to a BM3 variant capable of converting ethane into ethanol[106].In a different study,screening of a BM3 library constructed by semi-random engineering led to F87A/A328V mutant that can efficiently hydroxylate cyclooctane,cyclodecane and cyclododecane [107](Fig.5F).Moreover,saturation mutagenesis at selected active-site amino acids of BM3 and combination of resulting beneficial mutations led to a variant library,where some mutants performed ω-hydroxylation of octane with good selectivities (up to 52%) [108].

In a recent study on BM3 by the groups of Reetz and Munro,engineering attempts through semi-rational mutagenesis led to BM3 variants that performed excellent regio-and diastereoselective hydroxylation of various steroid compounds at C16 position with unusually high activity levels (over 80% conversion) [20](Fig.5G).Authors of this study used a novel engineering strategy where they combined structural and mechanistic information as well as insights from molecular dynamics simulations and from mutability landscape (maps of beneficial,neutral,and detrimental residues) [109]in order to limit amino acid exchanges for saturation mutagenesis of the residues lining the binding site for multiple rounds of iterative saturation mutagenesis.The novelty of this study lies in the fact that both selectivity and activity have been increased simultaneously without any trade-offs.

Engineering studies were also carried out to obtain BM3 variants that are stable in organic solvents.Enzyme engineering by directed evolution led to variants that exhibit 3-to 6-fold higher specific activities compared to wild-type BM3 in 25% (v/v) DMSO.Some variants also exhibited higher resistance to some other organic cosolvents including THF,DMF and ethanol [110].BM3 heme domain was also evolved with the aim of generating peroxygenases that can use H2O2instead of oxygen for hydroxylation reactions.Such variants circumvented the need for the reductase partner and for NADPH [111].Moreover,a thermostable variant of the “evolved peroxygenase”BM3 heme domain was obtained in a follow-up study where the protein was subjected to further directed evolution that consisted of random mutagenesis and DNA shuffling,leading to significantly higher (50 to 250 times)half-life at 57.5 °C [112].

Furthermore,elegant engineering studies demonstrated ability of BM3 mutants to perform new-to-Nature reactions (reactions not present in nature,Fig.1B),including cyclopropanation by carbene transfer and amination by nitrene transfer reactions,which are valuable transformations for the synthesis of pharmaceuticals.Using directed evolution,Arnold and coworkers obtained variants of BM3 that use ethyldiazoacetate to generate an iron-carbenoid(analogous to iron-oxo intermediate) as the reactive intermediate to carry out stereoselective cyclopropanation of olefins [113,114].In fact,different variants,each of which can generate a specific stereoisomer out of the four possible isomers,were successfully evolved.A key mutation that converted axial cysteine ligand of the heme to a serine residue further rendered a variant that can efficiently catalyze styrene cyclopropanation using whole cells with total turnover numbers of up to 70.000 [115].This variant was named as P411 BM3,since the Soret band shifted from 450 nm to 411 nm due to the axial ligand change.Such a mutant activity demonstrates the power of enzyme engineering,specifically directed evolution,in obtaining novel variants that can carry out synthetically valuable transformations under mild conditions.Cyclopropanation capable BM3 mutants were also demonstrated as catalysts in the synthesis of the cyclopropane core of the antidepressant drug levomilnacipran in preparative scale[116](Fig.5H).Arnold and coworkers also evolved BM3 variants that can catalyze nitrene transfer for intra-and intermolecular C-H aminations[24,117].Highly enantioselective variants that can catalyze benzylic C-H amination were obtained.In addition,P450 enzymes(P411 variants)were engineered by Arnold group to catalyze enantiodivergent synthesis of fluoroalkyl-containing molecules with>99%ee value and total turnovers of up to 4070[118].Using directed evolution,BM3 was also evolved for selective silane oxidation to generate silanols,which further expands the tool box of nonnative reactions of CYPs [119].

A different enzyme engineering approach,named as chimeragenesis,was used successfully to alter the regiospecificity and substrate selectivity of BM3 in a number of studies carried out by Walker lab.In chimeragenesis,enzyme variants were generated by the exchange of short peptide sequences at the substrate binding site with the corresponding residues from other CYPs.This approach led to chimeras that can hydroxylate palmitic acid at positions from ω-4 to ω-6[120,121].In one such study,a continuous peptide sequence of 10 residues close to the bound substrate of BM3 (residues 73–82) were replaced by the corresponding residues from the insect enzyme CYP4C7,for which farnesol is a native substrate.An additional mutation at F87 was also introduced.Resulting variants exhibited 5.7-fold increased activity and altered regioselectivity towards farnesol oxidation [122].

You re so darn negative and boring! I retorted. I happen to like crowds. They make me feel like I m part of something. I promise I won t buy anything but I m going to look around for fun anyways. You go for coffee and I ll meet you back here in half an hour.

In addition to BM3,a few engineering studies aimed other selfsufficient CYP families.For example in a recent study,CYP505A30 has been subjected to targeted mutagenesis based on a homology model and on the corresponding BM3 mutations explored previously in the literature[92].Five amino acids were targeted in generating a triple and a quintuple variant.However,mutants did not exhibit significant change compared to the wild-type enzyme in terms of substrate scope,albeit a more pronounced difference was observed with respect to product distribution,especially in the hydroxylation of ibuprofen [92].In another study,a single active-site mutant (V51Y) of CYP505D6 exhibited increased ratio of ω-1 to ω-3 hydroxy fatty acids at the expense of decreasing ω-4 to ω-6 hydroxy products from lauric acid,whereas the opposite mutation in BM3 (Y51V) enabled hydroxylation of ω-5 to ω-7 positions,which is not observed with wild-type BM3 [93].CYP505X fromAspergillus fumigatushas also been subjected to rational protein engineering and expressed inP.pastoriscells,which served as whole cell biocatalyst [90].In comparison to the wild-type enzyme,a five-residue mutant of CYP505X exhibited higher conversion levels towards oxidation of five of the xenobiotic compounds tested.Specifically,monooxygenation of ibuprofen was 2-fold higher than the wild-type enzyme and resulted in a different oxidized metabolite profile.A preparative scale reaction with the mutant enzyme yielded over 80% conversion in 22 hours with 500 mg of substrate and generated three monooxygenated metabolites,two of which have been hydroxylated at the isobutyl moiety of ibuprofen[90].In a recent study,mutation of the activesite residue F89 (corresponding to F87 in BM3) to isoleucine in CYP102 fromB.amyloliquefaciens(BAMF2522) shifted hydroxylation regioselectivity significantly towards in-chain position for palmitic acid substrate[71].Whereas wild-type enzyme hydroxylates positions from ω-1 to ω-7,F89I mutant was able to generate ω-1 to ω-9 hydroxy products of palmitic acid,with ω-7,ω-8 and ω-9 hydroxy fatty acids constituting over 70% of the total product.Inchain hydroxy fatty acids are potential precursors for the synthesis of fatty acid esters of hydroxy fatty acids (FAHFAs),a group of recently discovered natural lipids with anti-diabetic and antiinflammatory properties [123].A double active-site mutant(A331V/F89I) of the other CYP102 homolog (BAMF0695) fromB.amyloliquefaciensexhibited increased ratio of ω-1 hydroxylation over ω-2 and ω-3 positions compared to wild-type enzyme [71].

Native self-sufficient enzymes gave researchers inspirations to generate artificial fusion enzymes where heme domains of CYPs from non-self-sufficient families were artificially fused to reductase partners to give them the ability of self-sufficiency [50].The benefit of such an approach is that it becomes possible to take advantage of the redox self-sufficiency for the non-self-sufficient CYPs that may catalyze unique reactions with high potential for applications.One good example for such a case is CYP105AS1,which is used in pravastatin synthesis,a cholesterol reducing blockbuster drug (Fig.5I).In addition to being engineered to catalyze the hydroxylation of compactin to pravastatin with the desired stereoselectivity,it has been fused to the reductase partner of CYP116B2 (FMN and Fe-S) and exhibited efficient conversion levels under industrial production conditions [124].

In addition to enzyme engineering,there are metabolic engineering approaches involving CYPs,which recently gained increased attention.These synthetic biology approaches involves multiple enzymes for generation of high value productsin vivo,where CYPs serve as important enzymes catalyzing key tailoring steps.Moreover,CYPs have been also included as key components inin vitrocascades.For more details and literature examples,we refer the reader to extensive reviews and research articles on metabolic engineering [26,128–130]and on enzymatic cascades[131–133].

5.2.Discovery of new self-sufficient CYPs

Genome mining and discovery of new CYPs,especially selfsufficient enzymes,is a promising avenue to uncover CYPs with desired properties for biocatalysis.Especially,recent metagenomics and proteomics approaches render great possibilities for the identification of genes and proteins from cell free mixtures of genes and proteins [134,135].Although there are various selfsufficient enzymes studied to date,they represent only a small portion from the available genomic information present in databases.Through bioinformatics analysis,new orthologues from selfsufficient families can be identified.With the current ease of gene synthesis,these homologous CYPs can be expressed from a synthetic gene encoding the protein,followed by activity screening towards various target substrates under different conditions.Gene mining from thermostable organisms or organisms living in extreme conditions are especially promising,since their enzymes are more likely to be stable under high temperature and other harsh process conditions,thus are more suited for industrial applications.For example,such efforts led to discovery of various thermostable CYPs from different families,including members of selfsufficient CYP116B family as described above [82,84].

5.3.Use of whole cell catalysts

Most CYPs require stoichiometric amounts of NAD(P)H for each of their catalytic cycle as electron donors [59].Use of whole cell catalysts is one important approach to avoid external supply of the expensive NAD(P)H cofactor [136].Cells have intrinsic NAD(P)H at decent concentrations that can support activity of CYPs for many turnovers without additional supply.Another main advantage of using whole cell systems is that there is no requirement for protein isolation.This is an important benefit especially for industrial scale bioconversions as well as for screening of large number of mutant libraries.Moreover,the intrinsic scavengers of the cell can neutralize any reactive oxygen species that can be formed due to uncoupled turnovers.However,when using whole cells,issues such as limited substrate uptake or product release,product degradation,side-reactions from other cell components and the need for more complex downstream processing should be considered[6].Many examples of CYP-catalyzed biotransformations given throughout the text have been performed using whole cell catalysts [71,77,89,90,115].

5.4.Cofactor recycling

Another important area of CYP research that aims minimizing dependence on NAD(P)H is cofactor recycling.Regeneration of the expensive cofactor molecule is a promising approach to reduce the costs associated with otherwise continuous stoichiometric supply of the cofactor.Once a suitable recycling system is established,only catalytic amounts of the cofactor is required.There are common enzymatic regeneration systems,often constructed with various redox enzymes.Glucose dehydrogenase,formate dehydrogenase,phosphite dehydrogenase and alcohol dehydrogenease systems are the most widely used ones and are added to the CYP reaction mixture along with their substrates[137–139].These enzymes then use NAD(P)as an electron acceptor to catalyze their respective oxidation reaction,where NAD(P)H is generated and becomes available for CYP catalytic cycle.Successful cofactor regeneration examples were reported for fatty acid oxidation by BM3 [140–142].We refer readers to detailed reviews on various cofactorregenerationsystemsandtheirapplications[137,139,143,144].

5.5.Replacing reductase system with H2O2

One other way to circumvent the need for expensive cofactors is taking advantage of the peroxide shunt pathway,in which only H2O2is required as the supply of both electrons and oxygen,without the need of any intermediate protein.This approach significantly reduces the complexity and cost of biotransformations catalyzed by CYPs [145].Some CYPs (CYP152 family) already uses H2O2as their native electron and oxygen donor and named as P450 peroxygenases [146,147].These enzymes bind H2O2with high affinity and exhibit efficient turnover numbers.Many other CYPs can go through the peroxide shunt pathway if H2O2is included in the reaction.However,use of H2O2has its own challenges[26].One main issue is the toxicity of H2O2for the protein [148].Enzymes can easily become deactivated at high concentrations of H2O2that is needed for efficient catalysis.Thus,in situcontrolled generation of H2O2at low concentrations has been explored to overcome the problems associated with H2O2usage in CYP152 family [149].There are a few examples,where a self-sufficient CYP was engineered to utilize the peroxide shunt pathway for catalysis [111,150].BM3 heme domain variants generated by directed evolution were able to act as efficient peroxygenases,utilizing H2O2to replace reductase domain,oxygen and NADPH.These variants were demonstrated to catalyze the production of human metabolites of the drug propranolol[145].Another recently demonstrated example can be considered as a unique substrate engineering strategy,where exogenous small molecules were used to convert wild-type BM3 monooxygenase into a peroxygenase enzyme [151].These small molecules had dual-functionality comprised of a binding group to the enzyme and a basic group to activate H2O2.N-(ω-imidazolyl fatty acyl)-l-amino acids successfully served this purpose as general acid-base co-catalysts for epoxidation of styrene with significantly increased activity and enantioselectivity,and for hydroxylation of short-chain alkanes with better or comparable activities to other P450 peroxygenases [151,152].Similar approach has also been used very recently by the same research group to demonstrate highly regioselectiveOdemethylation of various aromatic ethers,an important step in lignin bioconversion,using peroxide-driven BM3 variants from a semi-rational mutant library[153].Moreover CYP116B5,a recently characterized self-sufficient enzyme,has been shown to efficiently utilize H2O2for hydroxylation of aromatic compounds including pnitrophenol and the drug compound diclofenac,withkcatvalues ranging from 0.06 min-1to 2.7 min-1andKmvalues in the range of 10–100 μmol·L-1[154].The enzyme can also catalyze peroxide-drivenN-desmethylation of tamoxifen,another drug compound.With its high reduction potential and stability towards presence of H2O2at high concentrations (up to 5 mmol·L-1),this enzyme is a good candidate for bioremediation and drug metabolite synthesis.

5.6.Use of artificial cofactor mimics

Another promising area of research in circumventing the problems associated with the high cost and instability (under process conditions)of NADH/NADPH cofactors is the use of artificial cofactor mimics[155,156].Although cofactor mimics have been known for quite some time due to their common use in mechanistic enzymology studies,their use for biocatalysis is quite new,with promising examples appearing only in the last decade [155–157].While this strategy does not work with all cofactor utilizing enzymes and seems to be limited to certain enzyme groups,a successful example has been shown for CYPs;N-Benzyl-1,4-dihydronicotinamide was used as an NADH biomimic to support catalysis by a BM3 variant at levels comparable to NADPH,whereas the wild-type BM3 was inactive [158].

5.7.Light-driven CYP catalysis

One other novel approach to increase efficiency of CYPs is to take advantage of light activation.Use of Ru (II)-diimine functionalized metalloproteins for such purpose has been applied to many enzymes including CYPs [159–163].An efficient light-driven BM3 biocatalyst was demonstrated by Cheruzel and coworkers in a study where they generated a hybrid structure containing an engineered BM3 heme domain and Ru(II) photosensitizer,which is covalently attached to the heme domain at the proximal heme site through a non-natural cysteine.This approach enabled visible light-driven hydroxylation of dodecanoic acid with a total turnover number of over 900 and with a high initial rate,avoiding the need for NADPH and reductase partner.Ru(I) species,photogenerated from Ru(II) under flash quench reducing condition,is able to provide electrons to the heme center for the hydroxylation reaction[163].A detailed analysis of this system was also performed by X-ray crystallography,site-directed mutagenesis and transient absorption measurements [164].Light-driven catalysis by BM3 variants were also shown insideE.colicells with eosin Y as the photosensitizer.Eosin Y can enter the cell and bind to the heme domain,transferring visible light induced electrons directly to the heme center.Using this approach,bioconversion of drug molecules such as simvastatin and omeprazole has been demonstrated using light illumination,without the need for NADPH and a reductase protein [165].

5.8.Optimization of reaction conditions

Optimization of reaction conditions is another important consideration for feasible application of CYPs under industrial settings.For example,reaction engineering has been carried out in a recent study for the hydroxylation of dodecanoic acid by BM3[166].This study shows that oxygen supply and substrate solubilization(availability) are the main process issues that prevents high reaction output;optimization of these parameters led to increased space–time yields and overall conversion levels,while maintaining selectivity.Uncoupled reaction(utilization of NADPH and O2without product formation) with the hydroxylated product has been identified as another bottleneck.Further reaction engineering studies will lead to more optimized processes for large-scale production by CYP enzymes.

5.9.Machine learning approaches

Recent years witnessed other novel approaches for increasing enzyme efficiency in general,which can also be applied to improve self-sufficient CYPs.One of these modern approaches include use of machine learning models to make predictions of catalytic efficiency,substrate specificity,selectivity,stability and solubility of natural or engineered enzymes [167–169].The basic idea of machine learning approach is to train mathematical models on available input (feature) and output data (label),such as protein sequence,protein structure,reaction conditions or substrate structure as the input and catalytic efficiency,selectivity and stability as the output data.Then,high quality predictions for sequence variations(mutation sites)or reaction conditions for the desired output can be made based on the trained models.Thus,machine learning approaches emerge as an efficient tool to guide enzyme engineering [170,171].With the help of such anin silicoapproach,experimental effort for generating and screening mutant libraries can be reduced to a great extent.Various studies recently appeared in literature presenting machine learning applied to biocatalyst design and improvement,including CYP102 family [172–175].For example,Arnold group demonstrated a successful example of machine learning through use of Gaussian processes to guide enzyme engineering in order to generate thermostable chimeric CYP102 enzymes [174].It is very likely that machine learning approaches for enzyme improvement will increase rapidly in the near future,which will also be beneficial for the applicability of self-sufficient CYPs.

6.Conclusions &Outlook

There has been extensive investigation of CYPs using structural,spectroscopic and computational tools over the past almost 60 years,revealing detailed mechanistic understanding of the reaction pathways,identity of reactive intermediates,enzyme kinetics,substrate specificity,chemo-,regio-and enantioselectivity of catalysis.Among many CYP groups identified and studied to date,especially self-sufficient CYPs present promising potential for application purposes due to their significant advantages discussed in this review.Various native and engineered self-sufficient CYPs have been demonstrated to catalyze a diverse array of synthetically useful reactions.Moreover,BM3 variants have been shown to catalyze non-natural reactions of synthetic significance under mild conditions.However,there are still various challenges against practical use of CYPs for synthesis and production purposes.Particularly,low turnover numbers and poor regioselectivity towards non-natural substrates as well as poor stability and NADPH requirement are the main barriers against their industrial applicability.Enzymes with slow catalytic rates and low turnover numbers increase the production cost and thus are not feasible.Poor stability at high temperatures often limits efficient usage of enzymes in bioreactors.Moreover,stability towards high substrate/product concentrations as well as tolerance to organic solvents are other important concerns that need to be addressed for construction of efficient industrial bioprocesses.In the case of poor regioselectivity,a mixture of products are obtained,which decreases the overall yield and requires costly separation process of the target products.Uncoupling of product formation from NAD(P)H oxidation is another challenge with CYPs,specifically when non-natural substrates are used.

There is a great effort from the research community to address these bottlenecks,mainly due to highly valuable transformations catalyzed by CYPs under eco-friendly conditions.As discussed in this review,efforts through various approaches that span fields from enzyme engineering and cofactor regeneration to reaction engineering and machine learning are all promising avenues to render CYPs as feasible biocatalysts for use in biotechnology applications.Such efforts will help to convert current chemical processes to bioprocesses compatible with green chemistry principles [1].Moreover,new processes that are not possible with chemical methods can be developed,taking advantage of the great diversity and evolvability of CYPs,especially of the self-sufficient family [3].Current proof-of-concept examples from the literature highlight such potential.

On another note,studies on self-sufficient CYPs mainly targeted BM3 as the model enzyme.Only a few studies have been conducted on other self-sufficient CYPs so far.In fact,research on other self-sufficient families such as CYP505 and CYP116 intensified only in the last decade.These recent efforts led to discovery of novel self-sufficient homologs that can catalyze reactions not observed before with BM3 or its variants,such as hydroxylation of ω-7 position of medium to long chain fatty acids[95].However,the studies on the new CYPs are currently limited to identification and characterization efforts.More engineering studies on these newly discovered groups in the future,similar to those applied on BM3,can render novel enzymes for biotechnology.Besides,continuous efforts to discover and characterize new enzymes,assisted by metagenomics and proteomics studies,will greatly contribute to the exploitation of the great potential of self-sufficient CYPs.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

FinancialsupportsfromNovoNordiskFoundation(NNF16OC0021740),Aarhus Universitets Forskningsfond AUFFNOVA(AUFF-E-2015-FLS-9-12)and Danmarks Frie Forskningsfond(DFF Technology and Production,0136-00206B) are greatly acknowledged.Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2020.12.002.