The yeast biosensor strain yJS1256 for high throughput screening was kindly provided by Prof. Dueber from the University of California at Berkeley (DeLoache et al., 2015). The A. thaliana cDNA library under the control of ADH1 promoter and ADH1 terminator in plasmid pGADT7-AD was kindly provided by Prof. Zeng Tao from Zhejiang University. To facilitate the construction of the A. thaliana cDNA overexpression library in S. cerevisiae, a helper plasmid pRS416-TEF1p was constructed by cloning TEF1 promoter, upstream and downstream homology arms of the cDNA library, as well as EcoRI and BamHI restriction sites into pRS416. Then pRS416-TEF1p was digested by EcoRI and BamHI and co-transformed with the PCR amplified cDNA library fragments with homology arms into yJS1256 for in vivo assembly (Figure 1B). The identified plasmids from the A. thaliana cDNA library were isolated using a ZymoPrep Yeast Plasmid Miniprep Kit (Zymo Research, Irvine, CA) and transformed into Escherichia coli DH5α for plasmid amplification. The Z-α-santalol producing S. cerevisiae strain was constructed by integrating four copies of the α-santalene synthase (SAS) gene (Dong et al., 2020) and one copy of CYP736A167 and CPR2 from Santalum album (XII5 locus) into the genome of BY4741 (MATa his△1 leu2△0 met15△0 ura3△0), via CRISPR-Cas9-mediated genome editing technology (Lian et al., 2017; Lian et al., 2019). In addition, the mevalonate (MVA) pathway genes were overexpressed to enhance the precursor supply for Z-α-santalol biosynthesis, with tHMG1-ERG8-ERG13-ERG20-ERG12 overexpression cassettes and EGR10-MVD1-IDI1-tHMG1 overexpression cassettes integrated into X4 and XI3 loci, respectively. Q5 polymerase, T4 DNA ligase, and all restriction enzymes used were purchased from New England Biolabs (Ipswich, MA). All chemicals were bought from Sigma (Sigma Aldrich, St. Louis, MO) unless otherwise stated.

E. coli strain DH5α for cloning and plasmid propagation was cultured at 37°C in Luria-broth (LB) medium containing 100 μg/ml ampicillin. Yeast strains were cultivated in standard yeast peptone dextrose (YPD) medium consisting of 2% glucose, 2% peptone, and 1% yeast extract. Recombinant yeast strains were grown on complete synthetic (SCD-URA) medium consisting of 0.17% yeast nitrogen base (YNB, Difco, Boom, Netherlands), 0.5% ammonium sulfate, and the appropriate amino acid drop-out mix (CSM-URA, MP Biomedicals, Solon, Ohio) supplemented with 2% glucose at 30°C.

The A. thaliana cDNA overexpression library was constructed by the in vivo DNA assembly method in S. cerevisiae. After transformation, 10⁶ independent clones were observed on SCD-URA agar plates, indicating at least a 50-fold coverage of the A. thaliana cDNA library. For biosensor-based high-throughput screening, 76 clones with the highest yellow color intensities were selected from SCD-URA agar plates and inoculated into 1 ml SCD-URA medium in a 96 deep-well plate. The yeast strains were pre-cultured for 2 days and then inoculated into fresh SCD-URA medium with an initial OD₆₀₀ of 0.1. Mid-log phase yeast cells were collected and diluted 2-fold in ddH₂O for measuring betaxanthin fluorescence intensity at 498–533 nm using a Tecan microplate reader. The fluorescence intensity (relative fluorescence units; RFU) was normalized to cell density that was determined by the same microplate reader.

Z-α-Santalol producing strains were pre-cultured in SCD-URA medium for 2 days, inoculated into 50 ml fresh medium in 250 ml shaker flasks with an initial OD₆₀₀ of 0.1, and cultured at 30°C and 250 rpm for 5 days. Then 2 ml yeast cells were collected by centrifuge at 12,000×g for 2 min and resuspended in 700 μL ethyl acetate to be disrupted by bead milling. The cell lysate was centrifuged for 10 min and the supernatant was filtered for GCMS (SHIMADZU, Japan) analysis on a DB-5MS column. 2 μL of each sample was injected with a 20:1 split mode at 280°C. The initial column temperature was 40°C and kept for 3 min. Subsequently, the temperature was increased to 130°C at a rate of 10°C/min, followed by to 180°C at a rate of 2°C/min and to 300°C at a rate of 50°C/min, and finally kept at 300°C for 10 min. The production of α-santalene and Z-α-santalol was quantified using the standard curve method. The α-santalene to Z-α-santalol conversion rate was calculated as [Z-α-santalol]/([α-santalene]+[Z-α-santalol]).

Yeast cells in biological triplicates were grown at 30°C overnight in 5 ml of SCD medium for 2 days and inoculated into 50 ml SCD medium in 250 ml shaker flasks with an initial OD₆₀₀ of 0.1. Then cells were harvested in the early stationary phase by centrifugation at 4,000 g for 15 min at 4°C, with the total RNA extracted and sequenced (RNA-Seq) by Shanghai Majorbio Bio-pharm Technology Co., Ltd. RNA-Seq data manipulation and differential gene expression profiling were performed on the free online Majorbio Cloud Platform (www.majorbio.com) with default settings. The RNA-Seq data are available from the NCBI Sequence Read Archive, with an accession number PRJNA760804.

A major challenge for P450 engineering is the lack of a high throughput screening method. The betaxanthin-producing yeast strain yJS1256 developed by the Dueber group (DeLoache et al., 2015) was employed as a biosensor for high-throughput screening of yeast strains with improved functional expression of P450s. The biosensor is composed of a P450 enzyme mutant (CYP76AD1^W13L−F309L) from Beta vulgaris and an L-DOPA (dihydroxyphenylalanine) dioxygenase (DOD) from Mirabilis jalapa, which can catalyze the conversion of l-tyrosine to produce betaxanthin (Figure 1A). In addition, two transporter genes, encoding Qdr2p and Yor1p, were deleted to accumulate betaxanthin intracellularly (Savitskaya et al., 2019). With the introduction of an additional copy of CYP76AD1^W13L−F309L, the production of betaxanthin was significantly improved, indicating that betaxanthin biosynthesis in yeast was limited by CYP76AD1 (Supplementary Figure S1), which was consistent with the previous report (DeLoache et al., 2015). Considering the color and fluorescence characteristics of betaxanthin and rate-limiting of CYP76AD1 in betaxanthin biosynthesis, a biosensor-based high throughput screening method was established to isolate yeast mutants with improved production of betaxanthin (Figure 1B). Then the A. thaliana cDNA library was cloned into the single-copy plasmid pRS416 and overexpressed in S. cerevisiae. After yeast transformation, 20 clones were randomly selected for diagnostic PCR verification to investigate the diversity of the A. thaliana cDNA library in S. cerevisiae (Supplementary Figure S2A). Clones with obvious color changes were selected and inoculated into SCD-URA medium to measure the change in fluorescence intensities (Supplementary Figure S2B). Compared with the control strain containing empty plasmid, the fluorescence intensities of 25 clones were increased to varying degrees. The plasmids from these clones were extracted and sequenced to identify the candidate genes which could improve the production of betaxanthin in yeast.

After the extraction of plasmids from the isolated clones, they were re-transformed into the biosensor strain yJS1256 to verify the ability of these genes to improve the functional expression of CYP76AD1 and accordingly the production of betaxanthin in yeast. Three plasmids, A3, A6, and E9 harboring the genes of A. thaliana glycine-rich RNA-binding protein (AtGRP7), membrane steroid binding protein 1 (AtMSBP1), and A. thaliana CO-like four protein (AtCOL4), showed 1.32-, 1.86-, and 1.10-fold improvement in betaxanthin production, respectively (Figures 2A,B). AtGRP7 (A3) participates in the negative feedback loop of circadian rhythm regulation and pre-mRNA splicing and plays an important role in a complex network of transcripts in A. thaliana (Koster et al., 2014; Streitner et al., 2010). AtMSBP1 (A6), an ER-located protein, demonstrates steroid-binding activity in vitro and is involved in the inhibition of elongation and brassinosteroid signaling (Shi et al., 2011). More importantly, MSBP1 and its homologue MSBP2 can form homomers and heteromers on the ER membrane, which interact with three monolignin P450 enzymes to form MSBP-P450 protein complexes, thereby improving the stability and activity of P450s as well as the production of lignin (Gou et al., 2018). AtCOL4 (E9), a putative novel transcription factor, has transcriptional activation activity and is an important regulator of plant tolerance to abiotic stress (Min et al., 2015).

Afterwards, the synergistic or additive interactions among AtGRP7, AtMSBP1, and AtCOL4 were further investigated. Considering the stability of heterologous genes, these gene expression cassettes were integrated into XII5, X4, and XI3 loci of the yeast genome for combinatorial optimization (Figure 2B). When AtGPR7 (A3) or AtCOL4 (E9) was expressed together with MSBP1 (A6), synergistic and/or additive effects on the production of betaxanthin were observed, representing 2.21- and 2.18-fold higher than the control strain, respectively. In contrast, the effect of simultaneous expression of AtGPR7 (A3) and AtCOL4 (E9) was not obvious. The highest production of betaxanthin was achieved in the strain with these three genes (A3-A6-E9) being expressed simultaneously, representing a 2.36-fold improvement over the control strain. These results indicated that there might be synergistic interactions between AtMSBP1 and AtGRP7, as well as AtMSBP1 and AtCOL4, but not AtGRP7 and AtCOL4.

To further demonstrate the general applicability of AtGRP7, AtMSBP1, and AtCOL4 from A. thaliana in improving the functional expression of P450s, these genes were overexpressed in a Z-α-santalol producing yeast strain and their effects on the functional expression of CYP736A167, whose gene product catalyzes the hydroxylation of α-santalene to produce Z-α-santalol, was investigated (Figure 3A). Based on the previously constructed α-santalene producing strain (Dong et al., 2020), CYP736A167 and CPR2 were further integrated into the yeast genome to produce Z-α-santalol. To increase the accumulation of α-santalene, the substrate of CYP736A167 and the precursor of Z-α-santalol biosynthesis, genes involved in MVA pathway were overexpressed, including tHMG1, ERG8, ERG10, ERG12, ERG13, ERG20, IDI1, and MVD1. As shown in Supplementary Figures S3, S4, α-santalene was accumulated to a relatively high level, indicating that the production of Z-α-santalol was limited by the low activity of CYP736A167. The introduction of AtGRP7 (A3), AtMSBP1 (A6), and AtCOL4 (E9) all demonstrated positive effects on the hydroxylation of α-santalene to Z-α-santalol, with the conversion rate increased for 1.89-, 1.71-, and 1.73-fold, respectively. What’s more, when these three genes (A3-A6-E9) were overexpressed simultaneously (integrated into XI2, XII2, and XI3 loci of the yeast genome), the conversion rate was the highest, representing a 2.97-fold improvement when compared with the control strain (Figure 3B). Surprisingly, the titer of Z-α-santalol was not increased as significantly as the α-santalene to Z-α-santalol conversion rate (Supplementary Figure S5). To figure out the possible reasons, the expression level of MVA pathway genes were profiled. As shown in Supplementary Figure S6, all the MVA pathway genes except for ERG10 were down-regulated to different degrees. In other words, the advantage of overexpressing AtGRP7 (A3), AtMSBP1 (A6), and AtCOL4 (E9) in improving the functional expression of CYP736A167 should be combined with other metabolic engineering strategies to enhance the MVA pathway fluxes, which has been well established in yeast.

Finally, the molecular mechanisms of AtGRP7, AtMSBP1, and AtCOL4 in improving P450 expression were explored using transcriptomic analysis. mRNAs were extracted from yJS1256-AtMSBP1 (A6), yJS1256-AtGRP7-AtMSBP1 (A3-A6), and yJS1256-AtGRP7-AtMSBP1-AtCOL4 (A3-A6-E9) and sent for next-generation sequencing (RNA-Seq). Compared with the control strain yJS1256, 163 (105 up-regulated and 58 down-regulated), 405 (317 up-regulated and 88 down-regulated), and 76 (56 up-regulated and 20 down-regulated) genes had expression level changes by more than 2-fold with a threshold of <0.05 in yJS1256-AtMSBP1 (A6), yJS1256-AtGRP7-AtMSBP1 (A3-A6), and yJS1256-AtGRP7-AtMSBP1-AtCOL4 (A3-A6-E9) (Figure 4A and Supplementary Figure S7). Gene ontology enrichment analysis revealed that the expression level of genes associated with cytoplasmic translation, ribosomal small subunit assembly, peptide biosynthetic process, amide biosynthetic process, and peptide metabolic process were significantly changed (Figure 4B and Supplementary Figure S8). Specifically, some genes involved in integral component of membrane, such as COX3, COS12, COB, and NCW1 were up-regulated. Genes related to response to stress conditions, including GPX2 and GRX5, were also up-regulated. HAC1, a transcription factor, which was reported to induce the expression of genes related to the folding capacity of endoplasmic reticulum (ER) (Schuck et al., 2009), was apparently up-regulated. Furthermore, MUP3 was found to be significantly up-regulated in all strains, yJS1256-AtMSBP1 (A6), yJS1256-AtGRP7-AtMSBP1 (A3-A6), and yJS1256-AtGRP7-AtMSBP1-AtCOL4 (A3-A6-E9). MUP3 encodes a methionine permease, whose overexpression increases methionine availability and is important to maintain the stability and abundance of membrane proteins (Lee et al., 2019). HMX1, an ER-localized heme oxygenase, was down-regulated in all three strains. The deletion of HMX1 has been proved to increase heme concentration and improve the activity of CYP76AD1 and CYP102A1 in yeast (Savitskaya et al., 2019). In short, a variety of genes, including some genes with unknown functions, demonstrated significant changes at the transcriptional level, indicating complex regulatory mechanisms of the expression and folding of P450s.