Hot Spots of Phytoene Desaturase from Rhodobacter sphaeroides Inﬂuencing the Desaturation of Phytoene

: Phytoene desaturase (CrtI, E.C. 1.3.99.31) shows variable desaturation activity, thereby introducing different numbers of conjugated double bonds (CDB) into the substrate phytoene. In particular, Rhodobacter sphaeroides CrtI is known to introduce additional 6 CDBs into the phytoene with 3 CDBs, generating neurosporene with 9 CDBs. Although in-depth studies have been conducted on the function and phylogenetic evolution of CrtI, little information exists on its range of CDB-introducing capabilities. We investigated the relationship between the structure and CDB-introducing capability of CrtI. CrtI of R. sphaeroides KCTC 12085 was randomly mutage-nized to produce carotenoids of different CDBs (neurosporene for 9 CDBs, lycopene for 11 CDBs, and 3,4-didehydrolycopene for 13 CDBs). From six CrtI mutants producing different ratios of neurosporene/lycopene/3,4-didehydrolycopene, three amino acids (Leu163, Ala171, and Ile454) were identiﬁed that signiﬁcantly determined carotenoid proﬁles. While the L163P mutation was responsible for producing neurosporene as a major carotenoid, A171P and I454T produced lycopene as the major product. Finally, according to the in silico model, the mutated amino acids are gathered in the membrane-binding domain of CrtI, which could distantly inﬂuence the FAD binding region and consequently the degree of desaturation in phytoene.


Introduction
Carotenoids are a diverse group of colored isoprenoid derivatives that play distinct roles in nature [1]. More than 800 different carotenoids are synthesized in photosynthetic microorganisms, plants, and animals [2]. Naturally occurring carotenoids and their biosynthetic pathways are classified as C30, C40, and C50 based on the carbon numbers of their backbone structures [3]. Carotenoids serve several biological functions, including in coloration, photoprotective activities, and light harvesting, and are also the precursors for several plant hormones [4][5][6]. Carotenoids are widely used in the food, medical, pharmaceutical, and cosmetic industries as colorants and functional ingredients [7,8]. Despite the structural diversity and commercial importance of carotenoids, only a few simplestructured carotenoids, such as β-carotene and lycopene, are produced commercially by chemical synthesis or isolation. Their increasing industrial importance has led to renewed efforts to develop bioprocesses for the production of diverse carotenoids [8][9][10].
sphaeroides is a well-known carotenogenic bacterium similar to Pantoea agglomerans, which was previously classified as Erwinia herbicola [12], and its carotenogenic pathway has been widely used as a model system [13].

Identification of Mutation in Mutant Crtirs by Sequencing Analysis
To reveal the genetic alterations of the six mutant CrtI RS , genes encoding RIm1, RIm2, RIm3, RIm4, RIm5, and RIm6 were sequenced using the Sanger method (Table 2). A single-nucleotide point mutation was found in RIm1 and RIm2, causing an I454T amino acid change and an A171P change, respectively. To confirm the effect of the single I454T and A171P mutations on the degree of desaturation on phytoene, site-directed mutagenesis (SDM) of wild-type CrtI RS was applied to generate Ism4 (SDM_I454T) and Ism5 (SDM_A171P). HPLC analysis revealed that the two SDM mutants, Ism4 and Ism5, produced very similar carotenoid profiles to those of RIm1 and RIm2: 14.6 ± 2.7% of neurosporene, 79.4 ± 2.5% of lycopene, 6 ± 0.2% of 3,4-didehydrolycopene in the SDM mutant Ism4, and 19.2 ± 4.6% of neurosporene and 80.8 ± 4.7% of lycopene without 3,4-didehydrolycopene in the SDM mutant Ism5 (Figure 3a). The very similar carotenoid profiles observed between random mutants and SDM mutants strongly indicate that the single amino acid mutations (I454T and A171P) intrinsically altered the catalytic activity of CrtI RS toward phytoene. Similarly, an SDM mutant (Ism1, SDM_L163P) of the major neurosporene-producing RIm6 (Figure 3a) showed a carotenoid profile very similar to that of RIm6, demonstrating the intrinsically altered catalytic activity of CrtI RS .  One additional nucleotide point mutation was observed in the 2nd round mutants RIm3, RIm4, and RIm5. The 2nd single amino acid change of E186G in RIm3 led to double amino acid mutations (I454T and E186G). Similarly, the 2nd single W142R change in RIm5 generated a double amino acid mutation (A171P and W142R). Compared to RIm3 and RIm5, the single nucleotide point mutation in RIm4 caused a silent mutation. Notably, although RIm4 shared a single amino acid A171P mutation with RIm2, except for an additional single silent mutation in RIm4, the two mutants produced different carotenoid profiles (Table 2 and Figure 2a); however, the difference in carotenoid profiles between the two mutants was negligible.

Verification of the Effect of E186g and W142r Mutations on Activity of Phytoene Desaturase by Site-Directed Mutagenesis
As RIm3 and RIm5 mutants had additional amino acid mutations (E186G in RIm3 and W142R in RIm5) in comparison with RIm1 (I454T) and RIm2 (A171P) ( Table 2). The individual effects of the W142R and E186G mutations were investigated by generating two SDM mutants, Ism2 (SDM_W142R) and Ism3 (SDM_E186G). HPLC analysis revealed One additional nucleotide point mutation was observed in the 2nd round mutants RIm3, RIm4, and RIm5. The 2nd single amino acid change of E186G in RIm3 led to double amino acid mutations (I454T and E186G). Similarly, the 2nd single W142R change in RIm5 generated a double amino acid mutation (A171P and W142R). Compared to RIm3 and RIm5, the single nucleotide point mutation in RIm4 caused a silent mutation. Notably, although RIm4 shared a single amino acid A171P mutation with RIm2, except for an additional single silent mutation in RIm4, the two mutants produced different carotenoid profiles (Table 2 and Figure 2a); however, the difference in carotenoid profiles between the two mutants was negligible.

Verification of the Effect of E186g and W142r Mutations on Activity of Phytoene Desaturase by Site-Directed Mutagenesis
As RIm3 and RIm5 mutants had additional amino acid mutations (E186G in RIm3 and W142R in RIm5) in comparison with RIm1 (I454T) and RIm2 (A171P) ( Table 2). The individual effects of the W142R and E186G mutations were investigated by generating two SDM mutants, Ism2 (SDM_W142R) and Ism3 (SDM_E186G). HPLC analysis revealed that both Ism2 and Ism3 produced slightly less lycopene in comparison with that of the wild-type CrtI RS (Figure 3a), suggesting a marginal effect of the W142R and E186G mutations on the degree of phytoene desaturation by RIm3 (E186G and I454T) and RIm5 (W142R and A171P). Since I454T and A171P mutations significantly influenced the degree of desaturation of phytoene, an SDM mutant Ism6 (SDM_I454T_A171P) was generated and its carotenoid profile was investigated. As expected, dramatical alteration of the carotenoid profile (10.1 ± 1.9% of neurosporene, 84.5 ± 2.7% of lycopene, and 5.4 ± 4.6% of 3,4-didehydrolycopene) was observed in comparison with that of the wild-type CrtI RS (Figure 3a).

Complementation of Mutant Crtirs with the P. agglomerans Crtepa and Crtbpa
Even though carotenogenic enzymes have the same catalytic function, carotenoid profiles tend to vary depending on the source of the enzymes when expressed in a heterologous host [22]. To better understand the function of the mutant CrtI RS , random and SDM mutants were coexpressed in E. coli with the heterologous P. agglomerans CrtE PA and CrtB PA . HPLC analysis revealed that the three single amino acid mutants (RIm1, RIm2, and RIm6) when coexpressed with the heterologous CrtE PA and CrtB PA produced carotenoid profiles similar to those of RIm1, RIm2, and RIm6 when coexpressed with the native CrtB RS and CrtE RS (Table 2 and Figure 2b); however, RIm3, RIm5 (two double amino acid mutants) and RIm6 (a single amino acid mutant) when coexpressed with the heterologous CrtE PA and CrtB PA produced different carotenoid profiles in comparison with those coexpressed with native CrtB RS and CrtE RS (Table 2 and Figure 2b). This suggests that the 2nd amino acid alteration of E186G in RIm3 and W142R in RIm5 might influence the conformation of the enzyme complex structure, which consequently alters the catalytic activity of mutant CrtI RS toward phytoene. Notably, unlike the single amino acid mutants (RIm1, RIm2, and RIm6), RIm4 produced different carotenoid profiles when coexpressed with heterologous and native CrtE and CrtB. Three SDM mutants Ism2 (SDM_W142R), Ism3 (SDM_E186G), and Ism6 (SDM_I454T_A171P) produced similar carotenoid profiles when coexpressed with heterologous and native CrtE and CrtB (Figure 3b).

Structural Evaluation of Mutant Crtirs Using Computational Model Analysis
To understand the correlation between the structural changes and the observed activity of mutant CrtI RS , an in silico model of CrtI RS was created using the I-TASSER program [29] with the Protein Data Bank (PDB) templates of Nonlabens dokdonensis DSW-6 γ-carotenoid desaturase (4REP, [30]) and Pantoea ananatis phytoene desaturase (4DGK, [18]) ( Figure 4a). As FAD, a redox-active cofactor, is present in the active site region of CrtI RS , in silico ligand docking was simulated using COACH-D [31] with the PDB of FAD binding residues of Pseudomonas savastanoi pv. phaseolicola oxidoreductase. Predicted FAD binding sites of CrtI RS were the residues 16-17, 19-21, Figures 4b and 5a). Interestingly, the five mutated amino acids of CrtI RS (Trp142, Leu163, Ala171, Glu186, and Ile454) were present in the putative membrane-binding domain (red in Figure 5a and blue in Figure 5b). The membrane-binding domain was previously predicted to influence the hydrophobic residues (cyan in Figure 5a), which are involved in the FAD-associated tunnel of CrtI from P. ananatis [18]; therefore, different degrees of phytoene desaturation (the observed differences in carotenoid profiles) in mutant CrtI RS could be attributed to the alteration of the FAD binding environment in the active site region of CrtI RS . the five mutated amino acids of CrtIRS (Trp142, Leu163, Ala171, Glu186, and Ile454) were present in the putative membrane-binding domain (red in Figure 5a and blue in Figure  5b). The membrane-binding domain was previously predicted to influence the hydrophobic residues (cyan in Figure 5a), which are involved in the FAD-associated tunnel of CrtI from P. ananatis [18]; therefore, different degrees of phytoene desaturation (the observed differences in carotenoid profiles) in mutant CrtIRS could be attributed to the alteration of the FAD binding environment in the active site region of CrtIRS.

Error-Prone PCR Mutagenesis
Random mutagenesis of the R. sphaeroides crtI gene on the plasmid pUCM was performed using a previously reported error-prone PCR method [28]. Briefly, the mutagenic PCR condition of 1.5 mM MgCl 2 , unbalanced dNTP ratio (ATP:TTP:CTP:GTP, 1:1:1:4), and Taq polymerase were utilized to incorporate mismatched bases into the crtI gene with primers (5 -GCTCTAGAAGGATTACAAAATGCCCTCGATCTCGCCC-3 and 5 -CGGAATTCTCATTCCGCGGCAAGCCT-3 ) flanking the crtI gene on the pUCM vector. The PCR products were purified using a gel DNA extraction kit (Macrogen, Inc., Seoul, Korea), followed by digestion with the restriction enzymes EcoRI and XbaI. The fragments were cloned into the corresponding site of vector pUCM and then transformed into E. coli harboring pACM-E RS -B RS . The cells grown on M9 agar plates were supplemented with ampicillin (100 µg/mL) and chloramphenicol (50 µg/mL) at 37 • C for 24 h and then incubated at 20 • C until colonies developed. Colonies on M9 agar plates were visually screened, and those with color changes were selected and restreaked on LB agar plates to isolate pure colonies. The isolated cells were cultured in LB medium supplemented with ampicillin (100 µg/mL) and chloramphenicol (50 µg/mL), and pUCM plasmids containing the mutagenic crtI gene were isolated from 1% (v/v) agarose gel with a gel DNA extraction kit (Macrogen). Mutagenic crtI gene sequences were verified by Sanger sequencing (Macrogen). Fresh transformed E. coli strains harboring pACM-E RS -B RS and pUCM_mutant_crtI were prepared and cultured in Terrific Broth (12 g/L tryptone, 24 g/L yeast extract, 0.17 M KH 2 PO 4 , 0.72 M K 2 HPO 4 , and 10 g/L glycerol) supplemented with ampicillin (100 µg/mL) and chloramphenicol (50 µg/mL) to investigate the carotenoid profile.

Site-Directed Mutagenesis
Phusion High-Fidelity DNA Polymerase (New England BioLabs, Inc., Ipswich, MA, USA) was used to perform SDM of the crtI gene. Mutagenesis primers were designed according to the desired crtI gene mutations. After PCR amplification, the PCR product was digested with DpnI for 5 h and transformed into E. coli. The sequence changes in SDM were verified by Sanger sequencing (Macrogen).

Analysis of Carotenoid Production
Carotenoid extraction was performed using a previously described extraction method [22]. Briefly, 50 mL of culture was harvested and separated into a cell pellet and culture medium. Carotenoids were repeatedly extracted with a total of 20 mL of acetone until all visible color disappeared from the cell pellet. Equal volumes of water and hexane were added to the acetone extract and vortex-mixed. The upper carotenoid-containing solvent layer was carefully collected and dehydrated with 0.1 g anhydrous sodium sulfate (Sigma-Aldrich, St-Louis, MO, USA) for 20 min. After centrifugation (4 • C and 13,000 rpm), the supernatant was collected and completely dried using a Genevac EZ2 centrifugal evaporator (Genevac, Inc., Vally Center, NY, USA). The dried residue was resuspended in 500 µL acetone and 20 µL of aliquot was subjected to an Agilent 1260 series HPLC (Agilent technologies, Palo Alto, CA, USA) system equipped with an Agilent photodiode array detector and Zorbax eclipse XDB-C18 column (4.6 × 150 mm, silica particle, 80 Å, 5 µm; Agilent Technologies). The column temperature was maintained at 35 • C and the flow rate was 1 mL/min. Acetonitrile, methanol, and isopropanol (80:15:5, v/v/v) were used for isocratic elution. UV/Vis analysis of neurosporene, lycopene, and 3,4-didehydrolycopene was carried out at wavelengths of 440 nm, 470 nm, and 490 nm, respectively. The relative ratio of each carotenoid profile was calculated by comparing the peak area of each carotenoid in the LC chromatogram generated by OpenLab ChemStation ® software (Agilent Technologies). The results are expressed as means ± standard deviations of three replicates.

Computational Modeling of Phytoene Desaturase
To compare the functional differences of mutant CrtI enzymes, protein structures of CrtI mutants were computationally predicted using I-TASSER [29]. Two protein templates were used to construct CrtI protein models: γ-carotenoid desaturase (PDB ID: 4repA) from N. dokdonensis DSW-6 [30] and phytoene desaturase (PDB ID: 4dgkA) from P. ananatis [18]. Starting with the protein structures, FAD was docked into the crystal structures using COACH-D [31]. The structure of the FAD ligand was prepared using Chemsketch [32] prior to performing docking simulations. The model structures were visualized using the PyMol Molecular Graphics System (ver 2.0.4, Schrödinger, LLC, New York, NY, USA).

Conclusions
In this study, phytoene desaturase of R. sphaeroides (CrtI RS ), a catalyst in the 3-step desaturation of phytoene, was randomly mutated to alter its catalytic activity towards phytoene. CrtI RS mutants produced different ratios of neurosporene (9 CDBs)/lycopene (11 CDBs)/3,4-didehydrolycopene (13 CDBs). Leu163, Ala171, and Ile454 were particularly important residues in determining product alteration between neurosporene, lycopene, and 3,4-didehydrolycopene ( Table 2). The evaluation of an in silico model of CrtI RS concluded that the mutated amino acids were gathered in the membrane-binding domain, which could distantly influence the FAD binding region [18]. As CrtI RS is a bacterial carotene desaturase, the microenvironment of the cofactor FAD-binding region is important for the desaturation of phytoene catalyzed by a sole CrtI [33]. Although the distant influence of mutated residues on the FAD binding region of CrtI RS was demonstrated through the altered activity of mutant CrtI RS , the mechanism of FAD reoxidation in successive phytoene desaturation requires further investigation. Notably, phytoene desaturation and FAD reoxidation by quinones are separate events in plant-type phytoene desaturase [34]. This can provide insights into how bacterial CrtI simultaneously modulate phytoene desaturation and FAD reoxidation.