Heterologous Expression of the Hot Pepper ABA 8’-Hydroxylase in Escherichia coli for Phaseic Acid Production

The CYP707A family genes encoding ABA 8’-hydroxylase catabolize abscisic acid (ABA), a plant stress hormone that plays an important role in stress condition, such as drought, heat, cold and salinity. Phaseic acid (PA) is a catabolic product of ABA. Recent studies have shown that PA is important for the physiological functions in plants. It is also a neuroprotective molecule that protects against ischemic brain injury in mice. To obtain enzymes for the PA production, four CaCYP707A genes (CaCYP707A1, CaCYP707A2, CaCYP707A3 and CaCYP707A4) were isolated from hot pepper. They were heterologously expressed in Escherichia coli. Among them, CaCYP707A2 showed significantly higher expression levels in both the membrane fraction and the soluble fraction. Preferred redox partners were investigated to improve the efficiency of CaCYP707A2's catalytic reaction, and NADPH-cytochrome P450 reductase (CPR) from hot pepper (CaCPR) was preferred over other redox partners (i.e., rat CPR and ferredoxin reductase/ferredoxin). The production of 8’-hydroxy ABA and PA by ABA hydroxylation activity was confirmed in CaCYP707A2 from both membrane and soluble fractions. Therefore, CaCYP707A2 is the first identified plant CYP protein that is expressed a soluble form in cytosolic fraction having stable activity. Taken together, we propose a new CYP707A protein with industrial applications for PA production without additional modifications in E. coli heterologous expression.

region allowed successful expression, increased expression through N-terminal truncation is still difficult to predict because this proline-rich region plays an essential role in correct folding [19].
In this study, we tried to establish a system that efficiently expresses plant CYP707A at cytosolic fraction in E. coli and used it to produce the PA. Four CYP707A genes were isolated from hot pepper and heterologously expressed in E. coli. Without N-terminus modification, we found only CaCYP707A2 could be functionally expressed as a soluble form in cytosolic fraction. To increase the efficiency of the catalytic activity of CaCYP707A2, a preference investigation with redox partners showed much higher interaction with plant CaCPR than with other redox partners. Finally this study confirmed the generation of products according to the reaction time in the CaCYP707A2 from the soluble fraction with high industrial applicability. This is the first report of expressed plant CYP as a soluble form in the cytosolic fraction of E. coli.

Amino Acid Sequence and Structure Analysis
In a previous study, we isolated four CaCYP707A genes (CaCYP707A1, CaCYP707A2, CaCYP707A3, and CaCYP707A4) from the cDNA of hot pepper [22]. For further investigation of the structurally important motifs and 3D structure prediction, the amino acid sequences deduced from the four CaCYP707A genes were analyzed. Accession numbers of genes used for analysis are listed in Table 1. Multiple amino acid sequence alignments were performed using Sequence Manipulation Suite: Color Align Conservatioin (https://www.bioinformatics.org/ sms2/color_align_cons.html) and Clustal W [23]. Prediction of transmembrane was performed using the TMHMM Server ver. 2.0 (http://www.cbs.dtu.dk/services/TMHMM/). The protein structural prediction was obtained from the AlphaFold Protein Structure Database [24].

Heterologous Expression in E. coli and Purification of CaCYP707A
In a previous study, four CaCYP707A genes were transformed into E. coli, and the ABA 8'-hydroxylation activity was confirmed [22]. These transformed E. coli cells were used to investigate the soluble/membrane expression of heterologously expressed CaCYP707A proteins and its redox partner preference. The E. coli strains and plasmid information used for heterologous expression are listed in Table 1. Vector constructs for protein expression are shown in Fig. 2 To confirm the protein expression in cultured E. coli, 1 ml of the culture was collected and centrifuged at 13,000 ×g for 1 min. The cell pellet was resuspended in TES buffer (100 mM Tris-HCl, pH 7.6, 500 mM sucrose and 0.5 mM EDTA) and lysed by brief sonication (1 sec on, 10 sec off) repeated five times. Soluble part proteins were obtained from the supernatant by centrifuging the suspension at 13,000 ×g for 5 min, and insoluble part proteins were obtained by second resuspending in TES buffer. The presence of the protein of interest in both parts  was confirmed using 8% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel.
To isolate the proteins, the E. coli cells containing each expression plasmid were harvested by centrifugation (5,000 g, 15 min, 4°C). The cell pellet was resuspended in TES buffer and sonicated (5 sec on, 30 sec off) 10 times to lyse. To remove cell debris, the sonicated fraction was centrifuged at 5,000 g (15 min, 4°C) and the resulting supernatant was ultracentrifuged at 100,000 g (120 min, 4°C). The soluble cytosolic fraction was collected and stored at -80°C until use. The membrane fraction was resuspended in 100 mM potassium phosphate buffer (pH 7.4) containing 20% glycerol (v/v) and subsequently homogenized with a dounce homogenizer. The homogenized membrane fraction was stored at -80°C until use. The presence of the protein of interest in both separated fractions was analyzed on 8% SDS-PAGE gel.

Spectroscopic Analysis
The CaCYP707A1, CaCYP707A2, CaCYP707A3, and CaCYP707A4 concentrations in both the membrane fraction and the soluble fraction were calculated using the CO difference spectra as previously described [25]. Sodium dithionite was added to reduce the CYPs, and CO gas was injected to create a CO saturated condition. UVvisible spectra were recorded for absorbance from 400 to 500 nm using a Shimadzu UV-1601 spectrophotometer.

Substrate Binding Titrations
Spectral determination of the K D values for the binding of (+)-ABA substrates to the CaCYP707A2 was performed as previously described [26]. CaCYP707A2 was diluted to 0.5 μM in 100 mM potassium phosphate buffer (pH 7.4) and divided into two 1.0 ml glass cuvettes and a baseline was set. The (+)-ABA were dissolved in dimethylsulfoxide, and the final dimethylsulfoxide concentrations were < 1% (V/V). For spectrophotometric titration, ABA was added to CaCYP707A2 at a concentration ranging from 0 to 80 μM, a difference in absorbance between 350 to 500 nm was recorded using a Shimadzu UV-1601 spectrophotometer. The absorption difference between 389 and 420 nm was plotted against the substrate concentration. The binding affinities were estimated using the Graph-Pad Prism software as previously described (Graph-Pad, USA) [27].

Analysis of Enzymatic Activities of CaCYP707A2
Reaction mixtures were constructed to confirm the ABA hydroxylation activity of CaCYP707A protein. The reaction mixtures had a total volume of 250 μl and contained 50 pmol CaCYP707A2 (both the membrane fraction and the soluble fraction), 100 μM ABA, 100mM potassium phosphate buffer (pH 7.4), and an NADPH-generating system (final concentration of 10 mM glucose-6-phosphate, 0.5 mM NADP + , and 1.0 IU of yeast glucose-6phosphate dehydrogenase/ml). Three redox partners were used to compare the CaCYP707A2 catalyzed reactions: rat CPR (100 pmol), CaCPR (100 pmol), FdR/Fdx (0.1 units of FdR/ 20 μg of Fdx). For the membrane fraction protein, the reactions were performed at 28°C for 4 h. For soluble fraction proteins, the reactions were performed at 28°C for 15, 30, 45, 60 and 120 min. After incubation, the reactions were stopped by adding 25 μl of 1 N HCl. The reaction mixtures were extracted with 500 μl of ice-cold ethyl acetate. The products were extracted with centrifugation at 3,000 rpm for 5 min. The upper layer was transferred to a clean test tube, this process was repeated two more times. The solvent was evaporated under a gentle stream of nitrogen. The dried products were redissolved in 180 μl of mobile phase solution and transferred to a glass vial for HPLC analysis.
These dissolved products were injected into a Gemini C18 column (4.6 mm × 150 mm, 5 μm, Phenomenex) equipped with a HPLC system (Shimadzu Nexera LC-40D XS, Japan) and UV detector. The mobile phase solution consisted of water that contained 0.05% acetic acid and 37% methanol (v/v). The flow rate was 1.0 ml/min and the absorbance was measured at 254 nm.

Statistical Analysis
All experiments were performed three times. The data were presented as means with ± standard deviation of three determinations.

Analysis of Hot Pepper CYP707A Amino Acid Sequence
ABA 8'-hydroxylase is an enzyme that catabolize ABA and plays a key role in PA production [28]. In a previous study, we isolated four ABA 8'-hydroxylase genes from hot pepper (CaCYP707A1, CaCYP707A2, CaCYP707A3, and CaCYP707A4) [22]. By comparing the amino acid sequences deduced based on the DNA sequence information of the isolated genes, three highly conserved regions (oxygen binding site, E-R-R triad site, and hemebinding loop) essential for CYP catalytic activity were identified (Fig. 3). The transmembrane domain located at the N-terminus was predicted through the TMHMM Server, and their sequence similarity was revealed to be very low.

Heterogeneous Expression of CaCYP707A in E. coli
CaCYP707A proteins from hot pepper were targeted to express CYP proteins with stable activity in a heterologous expression system using E. coli. Various factors such as temperature, incubation time, aeration, and medium supplementation affect the expression of CYP proteins, and low-temperature incubation has been reported to produce stable protein folding without aggregation [29]. To estimate the expressed a CYP protein, the CO difference spectrum was measured for each incubation temperature and incubation time in E. coli, which is transformed with the vector expressing the CaCYP707A proteins. The reduced-CO difference spectrum of CYP proteins is a feature of CYP proteins that is widely used for qualitative and quantitative estimation [30]. The protein expression at different induction times (20,24,30,36,42, and 48 h) in E. coli cells were performed at several temperatures (25,28,32, and 37°C). The optimal expression levels of CaCYP707As were observed at 28°C with about 32 h culture (data not shown). Base on this, the expression of CaCYP707A proteins were investigated in the membrane fraction and soluble fraction isolated from E. coli (Fig. 4). In case of CO difference spectrum of the CaCYP707A1 membrane fraction, the Soret absorption peak was observed at 420 nm but not at 450 nm. Similar results were confirmed in the membrane fractions of CaCYP707A3 and CaCYP707A4, indicating that the isolated CaCYP707A1, CaCYP707A3 and CaCYP707A4 proteins can easily be deactivated. Meanwhile, the membrane fraction of CaCYP707A2 showed a clear absorption peak at 450 nm. The estimated expression of the enzyme using the CO difference spectrum was 29.4 nM for CaCYP707A2, but 0.3 nM and 2.4 nM for CaCYP707A1 and CaCYP707A4, respectively. CaCYP707A3 was not identified. In addition, CaCYP707A2 showed an absorption peak of 450 nm in the soluble fraction. However, CO difference of spectrum of other three CaCYP707As did not show. Expression in the soluble fraction was the highest at 3.2 nM for CaCYP707A2, but at 0.3 nM for CaCYP707A3. CaCYP707A1 and CaCYP707A4 from the soluble fraction were not identified. Therefore, CaCYP707A2 was selected for further experiments.
From the CO difference spectrum results, the expression of CaCYP707A2 was estimated in the cytosolic and membrane fractions of E. coli. To confirm this, the cultured E. coli was prepared under non-denaturing conditions and the presence of protein was evaluated using SDS-PAGE (Fig. 5). The CaCYP707A2 protein was found in both soluble and insoluble parts (Fig. 5A). Additionally, the presence of CaCYP707A2 protein was found in the same position as E. coli cell lysate in the separated soluble cytosolic fraction and membrane fraction (Fig. 5B). This indicates that it is consistent with the estimated expression result using the CO difference spectrum.
To confirm whether (+)-ABA can bind to CaCYP707A2, a spectral binding titration assay was performed (Fig. 6). By measuring the spectrum for the gradual increase in ABA concentration using CaCYP707A2, a type I difference spectrum with a peak at 386 nm and a trough at 419 nm was observed. The spectrally determined dissociation constants (K D ) of CaCYP707A2 to ABA was 5.8 ± 0.93 μM.

Dependence of CaCYP707A2 Activity on the Redox Partner
The CYP catalysis requires electron-transferring redox proteins, which is one of the factors determining the rate of CYP catalysis [31]. Therefore, studies have attempted to identify the preferred redox partner according to CYP as a strategy for optimizing catalysis [32,33]. The catalytic activity of CaCYP707A2 was investigated using redox partners including CaCPR, rat CPR, and FdR/Fdx, to determine the best redox partner to achieve the optimal electron transfer pathway in catalysis (Fig. 7). When CaCYP707A2 was incubated with CaCPR, 8'-hydroxy ABA and PA were produced due to the ABA hydroxylation activity. Although 8'-hydroxy ABA and PA were detected in  the catalytic reaction using rat CPR and FdR/Fdx, the amounts were much smaller than those observed after using CaCPR. These results indicate that interaction with redox partners affects the catalysis of CaCYP707A2, and CaCPR is the most preferred partner.

ABA Hydroxylation Activity of CaCYP707A2 Protein
In the CO difference spectrum results, the CYP-specific absorption peak of 450 nm was observed not only in the membrane fraction, but also in the soluble fraction of E. coli where CaCYP707A2 was heterologously expressed. Additional experiments were performed with a time course based on reaction time to investigate whether CaCYP707A2 protein from the soluble fraction has ABA hydroxylation activity and whether incubation time has an effect on the production of 8'-hydroxy ABA and PA (Fig. 8). The redox partner for electron transfer in the catalytic reaction was used CaCPR, which showed the highest interaction with CaCYP707A2. In the case of 8'hydroxy ABA, the amount of 8'-hydroxy ABA gradually increased as the reaction time increased. PA also showed a gradually increase as the reaction time increased until 2 h. The PA generation ratio of the reaction products was 0.2 to 0.3 when incubated for 15 to 60 min. Interestingly, after 2 h of incubation, it showed an increase over 0.7. These results indicate that the CaCYP707A2 protein from the soluble fraction has the same ABA hydroxylation activity as the CaCYP707A2 protein from the membrane fraction, and that the conversion ratio of 8'-hydroxy ABA to PA increases with time. Taken together, these findings suggest that the ABA hydroxylation activity of CaCYP707A2 is applicable to catalytic reaction systems for PA production.

Discussion
Plant CYPs are involved in the biosynthetic pathway of various anticancer and therapeutic agents derived from plants, such as Taxol, terpenoid indole alkaloid family substances, and a precursor of artemisinin (a powerful antimalarial drug) [34,35]. The application of plant CYPs to produce these high-value metabolites is an important goal in metabolic engineering.
However, two characteristics of plant CYPs often limit widespread application in metabolic engineering. First, CYP catalysis requires additional redox partners for sustained electron transport [36]. When constructing a metabolic engineering system through heterologous expression, the lack of information on redox partners and low accessibility to CYP hinder the catalytic reaction [4]. In the field of metabolic engineering, studying the interaction between CYP and redox partners to increase the efficiency of catalysis is important. Bacterial FdR/Fdx was used to find the optimal redox partner that interacts in CYP catalysis, and the results showed that depending on the CYP, using of a non-homologous surrogate redox partner can be helpful for optimizing the catalysis reaction [32,37]. To increase the interaction efficiency of CYP and its redox partner, rat CPR was fused with CYP and the expression site of rat CPR in bacteria was adjusted for efficient substrate delivery [38,39]. In addition, the characteristics of plant CPR were also investigated to understand the activity and role of CPR in plants [21]. In this study, the ABA hydroxylation activity of CaCYP707A2 produced more products with CaCPR than with rat CPR and FdR/Fdx (Fig. 7). These results indicate that selection of optimal redox partners is important to increase the efficiency of CYP catalysis, suggesting that the application of CaCPR to metabolic engineering using plant CYP is more effective.
Second, plant CYPs tend to exhibit low stability, low activity, and low expression levels in non-native hosts. Statistical data showed that out of over 250 plant CYPs, about 40% were difficult to express in yeast [8]. In this study, four CaCYP707A proteins were heterologously expressed in E. coli. In the CO difference spectra results of the CaCYP707A1, CaCYP707A3, and CaCYP707A4 proteins, the CYP-specific absorption peak of 450 nm was poorly observed (Fig. 4). These results suggest that heterologous expressions of CaCYP707A1, CaCYP707A3, and CaCYP707A4 proteins in E. coli have low stability and low activity. In contrast, CaCYP707A2 showed a CYPspecific absorption peak 0f 450 nm (Fig. 4). Notably, this result was observed not only in the membrane fraction but also in the soluble fraction. In addition, the CaCYP707A2 protein that was isolated from this soluble fraction was confirmed to have ABA hydroxylation activity related to 8'-hydroxy ABA and PA production (Fig. 8). This suggests that there is a difference between CaCYP707A2 and other CaCYP707A proteins that plays a crucial role in the stability of expression and maintenance of enzyme activity.
In the amino acid alignment results among CaCYP707A1, CaCYP707A2, CaCYP707A3, and CaCYP707A4, many regions were highly similar, but the N-terminal region was not similar to each other (Fig. 3). In particular, it is necessary to note to the amino acid sequence between the predicted transmembrane region of each protein (green box) and the proline-rich motif (blue box) (Fig. 3). The sequences for this part of CaCYP707A1, CaCYP707A3, and CaCYP707A4 contain hydrophobic amino acid residues, whereas the sequence of CaCYP707A2 consists mostly of hydrophilic amino acid residues (underlined in red) (Fig. 3). Studies conducted to obtain a catalytically active protein expressed in the cytoplasm of E. coli showed that introducing positive charges into the N-terminal linker region can lower the proportion of the protein localized in the inner membrane of E. coli [40].
The hydrophobic sequence length of the transmembrane domain and the distribution of amino acid residues may adopt different orientations or tilt angles when binding to the membrane [41]. The angle of the transmembrane domain affects the membrane binding of CYP proteins [42]. To further understand the protein structure, AlphaFold was used to predict the 3D structures of the CaCYP707A proteins ( Fig. 9) [24]. In general, eukaryotic membrane-bound CYPs have a transmembrane-helix domain that is located at the N-terminus, a linker region, and a globular domain structure that performs a catalytic action [42]. The structures of the predicted CaCYP707A proteins were also observed similarly to the reported structure of the transmembranehelix domain, linker region, and globular domain. In the 3D structure predicted through the AlphaFold, however, the transmembrane domain of the CaCYP707A2 protein shows a different tilt angle than other CaCYP707A proteins (Fig. 9). This result suggests that the distribution of hydrophilic and hydrophobic amino acid residues in the linker region of CaCYP707A2 protein may affect the membrane binding interaction and solubility of the CYP protein by changing the tilt angle of the transmembrane domain. Further experiments are needed to determine whether the linker region affects CYP solubilization and expression stability enhancement. The replacement of the linker region of membrane anchor proteins and additional sequence modifications will increase the availability for stable activity and expression of a variety of membrane-bound CYPs, including plant CYPs.