Cytochrome P450 (CYP) enzymes are a group of membrane-bound hemoproteins responsible for the synthesis of a great number of endogenous compounds including steroid hormones, bile acids, fatty acids, and eicosanoids ^[1][2][3][1,2,3]. CYPs are also major phase I metabolizing enzymes, bio-transforming xenobiotics such as drugs and carcinogens, in the body ^[4][5][4,5]. In humans, the CYP families 1, 2, and 3 contribute significantly to xenobiotic metabolism, while other CYPs are mainly involved in endogenous biotransformation [6]. Unlike prokaryotic CYPs, which are soluble, mammalian CYPs are integral membrane proteins found in the endoplasmic reticulum or mitochondria [7]. Characterisations of the structure–function relationships for CYP enzymes have been impeded by the challenges of purifying these insoluble CYPs from human tissues with sufficient quantity and activity ^[8][9][8,9]. Moreover, with the advanced development of whole-genome sequencing technologies, a large number of CYP genomic variations have been identified [10]. CYP polymorphisms, in particular, CYP2C9, CYP2C19, and CYP2D6, account for the most commonly seen variations in phase I drug metabolism clinically [11]. Nevertheless, the low frequencies of CYP variants have limited the evaluations of their impact on the pharmacokinetics of clinical drugs [12].

The heterologous expression systems provide an alternative opportunity to obtain individual CYP isoforms and their variants in evaluating the enzyme activities or in analysing protein structures under reproducible conditions [13]. Thus far, several in vitro expression systems, including mammalian cells, baculoviruses, yeast, and bacteria cells, have been documented for applications in characterising CYP enzymes [14]. Mammalian cells such as the African green monkey kidney-derived cells COS-1 and the human embryonic kidney cells HEK293 have been employed in expressing recombinant human CYP enzymes ^[15][16][15,16]. The advantages of the mammalian cell systems include no requirement for cDNA modifications, as well as adequate levels of endogenous NADPH-CYP oxidoreductase (OxR) and cytochrome b₅ to support electron transport and CYP catalytic activities [17]. However, employment of mammalian cells is often associated with high technical demand and a long duration of culture [18]. Besides, the CYP expression levels in mammalian cell cultures are usually low, which is unsuitable to study CYP variants, in particular, with low enzyme activity [14]. Baculovirus systems employ insect cells to express recombinant human CYPs, which can achieve high levels of expression [19]. Nevertheless, the technical demand and cost for insect cell cultures are high. The baculovirus systems also require the co-expression of OxR as insect cell lines are unable to express sufficient levels of OxR [17]. Yeasts such as Saccharomyces cerevisiae and Schizosaccharomyces pombe are useful in expressing human recombinant CYP ^[20][21][20,21]. The advantages of using yeast cells are low cost for maintenance, ease of culture, and a relatively high yield of CYP proteins. Moreover, the protein expression and post-translational modification processes are similar to those of higher eukaryotes, hence modifications of cDNA are usually not required [17]. Despite that yeasts contain endogenous OxR, the activity and quantity may be insufficient to fully support CYP enzyme activities, thus exogenous OxR may be essential [22]. Bacterial cells such as Escherichia coli (E. coli) demonstrate several advantages when being used as a heterologous system for human CYP expression. Culturing bacterial cells involves minimal maintenance cost as well as easier and faster cultivation. The recombinant CYP expression levels in bacteria are usually higher compared with those in yeast cells [23]. On the other hand, as human CYPs are membrane-bound, their expression in bacteria systems would require N-terminal modifications of the CYP cDNA to achieve optimal protein expression, conserve ideal folding, and maintain native biological functions ^[7][24][7,24].

2. Selections of Expression Vectors and E. coli Strains

The successful expression of CYP protein in bacteria is also influenced by the choice of plasmid vectors and E. coli strains (see Table 1). The most commonly employed CYP expression plasmid vector in E. coli is pCWori+. It was initially developed by F.W. Dahlquist and is not commercially available [23]. The overall structure of pCWori+ has been illustrated previously ^[25][65]. Essentially, it contains two tac promoters upstream of the Nde I restriction enzyme digestion site coincident with the ATG codon (start codon). Only one tac promoter (the one upstream of the polylinker site) is used, which is recognised by E. coli RNA polymerase. Upon the addition of Isopropyl β-D-1-thiogalactopyranoside (IPTG), the protein expression output is proportional to the amount of IPTG, which allows the expression of the precise level of CYP [23]. Additionally, it contains one trpA (a strong transcription terminator), the β-lactamase gene (conferring ampicillin resistance), and the lacI^q gene that encodes the Lac repressor (prevents any transcription initiated from the tac promoters without adding inducing agents) ^[25][65]. In general, the target CYP cDNA (native or modified) is introduced between the ATG start codon (contained within the Nde I site) and another restriction enzyme site, which is usually carried out by polymerase chain reaction (PCR) mutagenesis ^[26][25]. The recombinant vector was used in the transformation of various E. coli strains to produce recombinant human CYP proteins. Among them, DH5α ^{[8][9][24][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47][48]}[8,9,24,26,27,38,39,42,44,45,46,52,53,59,66,67,68,69,70,71,72,73,74,75,76] and JM109 ^{[24][26][29][49][50][51][52][53][54]}[24,25,38,64,77,78,79,80,81] strains are the most commonly used, while MV1304 ^{[7][55][56][57]}[7,43,47,49], XL-1 blue ^[58][82], and TOPP ^[59][60][83,84] have also been used. It is important to note that the E. coli strain selection can impact CYP expression levels. It was evidenced that CYP2C10 was not detectable in JM109 cells, but expressed in DH5α cells [24]. Nevertheless, no genetic markers were identified in these strains, showing a significant correlation with the capability of producing high levels of recombinant CYP proteins ^[25][65]. It is suggested to evaluate these common E. coli strains for their ability to express a particular recombinant CYP at the beginning of the study.

3. Bacteria Culture and Protein Expression Conditions

The typical bacteria culture and protein expression start with the initial culture of transformed E. coli strain in LB media supplemented with ampicillin (50–100 µg/mL) overnight at 37 °C (the optimal growth temperature for E. coli), followed with growing in Terrific Broth (TB) media containing ampicillin for an extended number of hours. The protein expression is subsequently induced by adding an inducing agent such as IPTG ^[31][42]. Factors involved in this process that may affect the yield of CYP protein expression include the ratio of LB to TB, OD₆₀₀ readings upon initiation of protein expression, temperature, shaking speed, expression duration, concentrations of IPTG, with or without δ-aminolevulinic acid (δ-ALA), and other more specific conditions for a particular CYP isoform (see Table 1). TB is a type of phosphate-buffered media that maintains a neutral pH level and comprises readily utilisable carbon sources ^[25][65]. The LB culture-to-TB culture ratio is usually maintained at 1:100 (e.g., 10 mL of LB culture to 1 L of TB) ^[31][38][42,66]. The TB media is often supplemented with trace elements to maintain CYP enzyme stability. Different studies applied different trace element compositions. As reported by Ahn and colleagues, trace elements expressing CYP1A2 in E. coli included 50 µM FeCl₃, 1 mM MgCl₂, and 2.5 mM (NH₄)₂SO₄ ^[27][26]. It is common for 1 mM thiamine (also known as vitamin B1) to be added to the TB culture media to ensure rapid E. coli growth ^[64][87]. The typical OD₆₀₀ values of 0.4 to 0.8 representing the mid-exponential bacterial growth phase were mostly used prior to induction ^[7][28][7,27]. Arabinose was required to induce the chaperon GroES-GroEL ^[9][35][62][9,48,52]. IPTG is a compound that mimics the molecular structure of allolactose that triggers the transcription of lac operon in E. coli. Hence, IPTG is used for protein expression induction where the gene expression is controlled by the lac operator, including pCWori+, the most commonly used vector for heterologous CYP protein expression in E. coli ^[25][65]. The majority of the studies employed 1 mM IPTG to induce CYP expression in E. coli cells, while exceptions were found in the expressions of CYP2D6 (1.5 mM IPTG) ^[30][39], CYP3A5 (0.1 mM IPTG) ^[44][72], CYP2S1, and CYP39A1 (0.5 mM) ^[62][63][48,86]. Δ-ALA, a well-known heme precursor, is involved in the pathway of protoporphyrin IX synthesis, and thus heme synthesis ^[65][88]. E. coli cells are able to produce heme-containing proteins with their endogenous heme biosynthesis system. The current results show that, although not an exclusive requirement for maximal production of all human CYP proteins in E. coli, the supplementation of δ-ALA could enhance the expression dramatically ^[25][65]. δ-ALA is readily taken up by E. coli cells, followed by heme synthesis catalysed by bacterial enzymes, which is subsequently inserted into the recombinant CYP polypeptide to form an enzymatically active protein ^[66][89]. The most commonly used final concentration of δ-ALA added before induction is 0.5 mM, with exceptions such as 1 mM for CYP3A5 ^[44][72] and 1.5 mM for CYP1A2 ^[27][26]. The addition of other chemicals to expression media was more specific to one or a group of CYP proteins. 4-methyl pyrazole, an inhibitor of CYP2E1 with high affinity, was added to the expression culture to stabilise the protein ^[7][58][67][7,82,90]. Bactopeptone was seeded in a TB medium to enhance cell growth in several studies ^{[24][27][43][56]}[24,26,47,71]. The employment of 37 °C for protein expression usually results in recombinant CYP accumulating as inclusion bodies. A lower expression temperature has been shown to produce more stable proteins without aggregation ^[68][91]. Nevertheless, expression temperatures below 25 °C lead to a dramatic drop in the expression level [23]. The optimal expression temperature during protein induction is often within a rather narrow range, and thus sensitive to drastic fluctuations in the temperature of the incubator. The typical induction temperature is not higher than 30 °C (mostly 28–30 °C). Certain human CYP proteins can be expressed with higher yields under higher temperatures, such as CYP2A6, CYP2E1, and CYP1A2, which were expressed at a comparable level and activities at 37 °C ^[27][69][26,92]. Moreover, the shaking speed and length of incubation during induction may also influence the optimal expression levels. The culture media in flasks shaken vigorously at 100–200 rpm were routinely performed to obtain optimal yields ^[48][56][47,76]. During the induction phase, the incubation usually lasts for 24–72 h. For instance, Bui and Hankinson reported that the growth of E. coli at 30 °C for 24 h provided the best expression conditions for a recombinant CYP2S1 ^[62][48].

4. Membrane Isolation

At the end of protein expression, bacterial cells are harvested by centrifugation, followed by membrane isolation prior to purification. The general steps of membrane isolation include suspension of harvested cells, lysis of cells, removal of cell debris, and membrane fraction sedimentation by ultra-centrifugation. Different studies applied different protocols in terms of suspension buffer, lysis of cell methods (by a high-pressure homogenizer, lysozyme, and ultrasonic energy), choice of a protease inhibitor, and collection of membrane fraction sedimentation. The harvested cells were usually suspended in phosphate buffers ^[57][63][49,86] or tris acetate buffers ^[29][31][38,42] with a pH range of 7.4–7.8 containing additional common compositions such as ethylenediamine tetraacetic acid (EDTA), sucrose, dithiothreitol (DTT), and glycerol. All of the steps were carried out at 4 °C. Both buffers functioned equally well in suspending bacterial cells expressing various recombinant human CYP proteins. Bacteria cells were suspended in a concentrated sucrose solution supplemented with EDTA, which were subsequently re-suspended in cold water. Under this condition, the bacteria cells shrink as a result of the high osmotic strength of the sucrose solution. EDTA plays a role in releasing lipopolysaccharide (LPS) from the cell envelope of bacterial cells, hence increasing the permeability of the outer membrane. Cold water leads to the rapid enlargement of cell size, resulting in the release of periplasmic proteins. This technique for the recovery of recombinant protein from E. coli is known as an osmotic shock ^[70][93]. Serious challenges have occurred in preserving protein stability and activity in biological applications as they are just marginally stable ^[71][94]. DTT is one of the protein reductants responsible for breaking down protein disulfide bridges and stabilizing enzymes ^[72][95]. Moreover, the most widely employed co-solvents for protein stabilization are polyols and, among polyols, glycerol is one of the most commonly used to stabilize and avoid aggregation of the protein ^[73][74][96,97]. Cell lysis can be defined as the destruction of the outer boundary or cell membrane to release inter-cellular materials. Cell lysis methods can be classified into mechanical (such as high-pressure homogenizer and bead mill) and non-mechanical approaches (including physical and chemical disruption) ^[75][98]. For the lysis of E. coli cells to obtain expressed human CYP proteins, mechanical approaches that use high-pressure homogenizer and non-mechanical techniques employing ultrasonic cavitation and enzymatic cell lysis were often recorded. A high-pressure homogenizer disrupts the membrane of cells by forcing them through an orifice valve ^[7][63][7,86]. Additionally, lysozyme is usually added to the suspended cell solution and incubated on ice or at 4 °C with stirring or shaking for 30 min ^[8][36][8,53]. Lysozyme is specific towards bacterial cells and reacts with the peptidoglycan layer, leading to the breaking of the glycosidic bond in the bacterial cell wall ^[76][99]. Ultrasonic cavitation is routinely applied in laboratories to disrupt cells. Ultrasound waves generate ultrasonic energy, which is transferred into the liquid solution and results in negative pressure. Once the negative pressure is lower than the vapour pressure of the liquid, vapour-filled bubbles are formed in the liquid solution. Then, when the bubbles grow to the size at which the ultrasonic energy is insufficient to maintain the vapour inside, they collapse and release a large amount of mechanical energy in the form of a shock wave, leading to cell rupture ^[77][100]. One of the disadvantages of ultrasonic cavitation is the generation of a large amount of heat, which may degrade enzymes ^[75][98]. During the lysis of E. coli, cells to isolate recombinant CYP proteins, a few rounds of ultrasonic treatment along with intervals on the ice were carried out in an ice bath to maintain cold conditions ^[24][26][24,25]. Upon lysis of cells, proteases are also released and their digestive functions are triggered, which can degrade isolated CYP enzyme proteins. Hence, the addition of protease inhibitors is required to preserve protein from imminent natural degradation. The majority of the proteases found in E. coli cells belong to the class of the serine protease group. Among the many classes of protease inhibitors, phenylmethylsulfonyl fluoride (PMSF) that inhibits serine protease irreversibly by deactivating the serine hydroxyl group is the most commonly used ^[78][101]. More recently, protease inhibitor cocktails comprising a mixture of several inhibitor compounds are more preferred in targeting a wide range of proteases that degrade enzymes via different mechanisms ^[28][54][27,81].