A rational design screening method for improving enzyme catalytic performance, protopanaxatriol synthase mutant and application thereof

Abstract

The application belongs to the field of enzyme engineering and bioinformatics, and discloses a rational design screening method for improving enzyme catalytic performance, a protopanaxatriol synthase mutant and application. The application takes protopanaxatriol synthase CYP716A53v2 (PPTS) as a transformation object, designs and constructs a mutation of a remote site through MM / PBSA guided active site mutation design and conformation bias guided remote site mutation design, and obtains advantage mutants such as T114K, I117D, I117Y, V102S, V102Q and V102N through experimental verification, and further combines to obtain a combined mutant. The application provides a rational design method with small screening amount and high hit rate for activity pocket transformation and remote conformation regulation of a terpene P450 enzyme, and has application value in biosynthesis of rare ginsenosides and other natural products.

Description

Technical Field

[0001] This invention belongs to the field of enzyme engineering and bioinformatics, specifically relating to a rational design screening method for improving enzyme catalytic performance, a protopanaxadiol synthase mutant and its application. Background Technology

[0002] Rare ginsenoside Rh1 is a type of protopanaxadiol-based ginsenoside with high pharmacological activity and application value, showing potential for development in anti-inflammatory, antioxidant, immunomodulatory, neuroprotective, and antitumor effects. However, the content of Rh1 in natural ginseng plants is low, and plant extraction is limited by factors such as raw material sources, growth cycles, and environmental conditions, making it difficult to meet the needs of large-scale preparation. Therefore, using synthetic biology methods to construct microbial cell factories to synthesize rare ginsenoside Rh1 has become an important direction for the green manufacturing of rare ginsenosides in recent years. Saccharomyces cerevisiae has a clear genetic background, high safety, and a complete endogenous mevalonic acid pathway and strong eukaryotic protein expression ability, making it a commonly used chassis for constructing terpenoid compounds. Through metabolic engineering, Saccharomyces cerevisiae can introduce carbon sources into the isoprene precursor supply system, further forming squalene, providing a precursor for the synthesis of dammarane-type ginsenoside skeletons. In the heterologous synthesis pathway of rare ginsenoside Rh1, 2,3-squalene oxide first undergoes PgDDS (dammarene diol synthase) catalysis to generate dammarene diol, followed by CYP716A47 / PgCPR1 (protopanaxadiol synthase PPDS / cytochrome reductase) catalysis to generate protopanaxadiol, then CYP716A53v2 / PgCPR2 (protopanatriol synthase PPTS / cytochrome reductase) catalysis to generate protopanatriol, and finally UGTPg100 (glycosyltransferase) to generate rare ginsenoside Rh1. Previous experimental results indicate that the step of protopanaxadiol synthase catalysis to generate protopanatriol is the key bottleneck limiting the efficiency of Rh1 synthesis. Therefore, molecular modification of protopanaxadiol synthase to improve its catalytic efficiency and substrate conversion ability is of great significance for enhancing the Rh1 synthesis flux in engineered Saccharomyces cerevisiae.

[0003] However, the engineering modification of protopanaxadiol synthase (PPTS) faces significant challenges. First, protopanaxadiol is a large tetracyclic triterpenoid substrate with a rigid structure. After entering the P450 active pocket, the substrate needs to approach the heme active site in a specific orientation to achieve selective hydroxylation at the C6 position. Second, the catalysis of P450 enzymes depends not only on substrate binding but also on the formation of a suitable pre-reaction state or near-attack conformation between the CH bond of the substrate to be hydroxylated and the Cpd I reactive oxygen species. Simply enhancing substrate binding ability does not necessarily improve catalytic efficiency; excessively strong or unreasonable hydrophobic binding may actually limit the conformational adjustment of the substrate within the pocket, making it difficult to enter a geometry conducive to the reaction. Third, the conformational flexibility, distal residue network, and redox chaperone interactions of plant P450 enzymes can all affect catalytic activity, making traditional local pocket mutations insufficient to fully address the rate-limiting problem.

[0004] Existing enzyme engineering methods mainly include directed evolution, semi-rational design, and structure-based rational design. Directed evolution typically relies on large-scale mutant libraries and high-throughput screening systems; however, the detection of complex natural products often depends on high-performance liquid chromatography (HPLC), which is costly and difficult to perform ultra-large-scale library screening. While traditional semi-rational design can reduce the screening scale by using molecular docking, conservation analysis, or binding energy scoring, static docking cannot reflect the dynamic conformational changes of the enzyme-substrate complex, and relying solely on docking scores can easily lead to false positive candidates with good binding energies but unreasonable reaction geometry. Furthermore, distal residues may affect catalytic efficiency by regulating conformational metastability, pocket opening and closing, and local flexibility, which existing methods often struggle to effectively identify.

[0005] Therefore, in the system for synthesizing rare ginsenoside Rh1 using *Saccharomyces cerevisiae* as a substrate, there is an urgent need to establish a rational design method with low screening quantity and high hit rate suitable for protopanaxadiol synthase. This method should simultaneously consider the energy contribution of residues in the active pocket to substrate binding, the effect of mutations on protein stability, whether the substrate meets the pre-reaction geometry required for P450 hydroxylation, and the regulatory role of distal residues on the conformational distribution of the enzyme-substrate complex. By combining MM / PBSA binding free energy decomposition, near-attack conformation determination, and conformational bias prediction, high-performance protopanaxadiol synthase mutants can be screened at a limited experimental validation scale, thereby removing the key rate-limiting step in the Rh1 synthesis pathway and improving the biosynthetic efficiency of rare ginsenoside Rh1 in engineered *Saccharomyces cerevisiae*. Summary of the Invention

[0006] The primary objective of this invention is to overcome the problems of low catalytic efficiency of existing protopanaxadiol synthases, large scale of traditional directed evolution screening, and insufficient accuracy of rational design based solely on static structures, and to provide a rational design screening method to improve enzyme catalytic performance. This method is based on binding free energy decomposition and conformational bias analysis. Through MM / PBSA-guided active site mutation design and conformational bias-guided distal site mutation design, it screens amino acid residues that affect substrate binding, pre-reaction conformational formation, and conformational metastable distribution, thereby obtaining protopanaxadiol synthase mutants with improved catalytic performance.

[0007] Another object of the present invention is to provide a protopanaxadiol synthase mutant obtained by the above-described rational design screening method.

[0008] Another object of the present invention is to provide a gene encoding the above-mentioned protopanaxadiol synthase mutant, a recombinant expression vector containing the gene, and a recombinant host cell.

[0009] Another object of the present invention is to provide the application of the above-mentioned protopanaxadiol synthase mutant, gene, recombinant expression vector and recombinant host cell in the preparation of protopanaxadiol and / or rare ginsenoside Rh1.

[0010] The objective of this invention is achieved through the following technical solution:

[0011] A rational design screening method to improve enzyme catalytic performance includes MM / PBSA-guided active site mutation design and conformational bias-guided distal site mutation design.

[0012] The MM / PBSA-guided active site mutation design includes the following steps:

[0013] (1) Obtain the three-dimensional structure of protopanaxadiol synthase, construct an enzyme-substrate complex model formed by protopanaxadiol synthase and protopanaxadiol, and perform molecular dynamics simulation on the enzyme-substrate complex.

[0014] (2) Select the molecular dynamics simulation trajectory after equilibrium is reached, and decompose the MM / PBSA binding free energy of amino acid residues near the original ginsenoside diol binding pocket. Based on the energy contribution of each amino acid residue to substrate binding and combined with sequence conservation analysis, screen the active sites that affect substrate binding.

[0015] (3) Virtual saturation mutations were performed on the active sites obtained by screening to construct virtual mutant models of ginsenoside triol synthase, and molecular docking was performed between each virtual mutant and ginsenoside diol.

[0016] (4) Based on the spatial distance between the hydroxylation site of the original ginsenoside diol and the active center of heme, the substrate binding conformation obtained by molecular docking was screened for pre-reaction state, and the active site candidate mutants were obtained by combining the evaluation of substrate binding ability and protein folding stability.

[0017] (5) Construct and express the candidate mutants of the active site, determine their ability to catalyze the production of protopanaxadiol to protopanatriol and further synthesize rare ginsenoside Rh1 in engineered Saccharomyces cerevisiae, and screen for active site mutants with improved catalytic performance.

[0018] The preferred amino acid sequence of the protopanaxadiol synthase PPTS is shown in SEQ ID NO.1.

[0019] The preferred nucleic acid sequence of the gene encoding protopanaxadiol synthase PPTS is shown in SEQ ID NO.2.

[0020] The molecular dynamics simulation time in step (1) is preferably 200 to 500 ns, more preferably 300 ns.

[0021] The molecular dynamics simulation trajectory after reaching equilibrium mentioned in step (2) is preferably the trajectory of the last 50 to 150 ns of the simulation trajectory, and more preferably the trajectory of the last 100 ns.

[0022] The amino acid residues near the protopanaxadiol binding pocket mentioned in step (2) are preferably amino acid residues that are no more than 5 Å away from any atom of protopanaxadiol.

[0023] The active site described in step (2) preferably includes one or more of the following: serine at position 110, threonine at position 114, isoleucine at position 117, asparagine at position 275, glutamine at position 348, proline at position 452, and phenylalanine at position 453.

[0024] The virtual saturation mutation described in step (3) is preferably performed using FoldX.

[0025] The molecular docking described in step (3) is preferably performed using AutoDock Vina.

[0026] In step (3), each virtual mutant preferably generates 20 to 60 protopanaxadiol-binding conformations, more preferably 40 protopanaxadiol-binding conformations.

[0027] The pre-reaction state screening mentioned in step (4) refers to selecting a substrate binding conformation with a distance of 3.5 to 5.0 Å between the carbon atom to be hydroxylated at the C6 position of protopanaxadiol and the iron atom of heme.

[0028] The active site candidate mutants mentioned in step (4) are preferably virtual mutants that retain a substrate binding free energy change of less than 0 kcal / mol and a folding free energy change of no more than 1.0 kcal / mol.

[0029] Using the above MM / PBSA-guided active site mutation design, 20 candidate mutants were screened from 133 virtual mutants for experimental verification, resulting in active site mutants T114K, I117D, and I117Y with improved catalytic performance.

[0030] The relative yield of rare ginsenoside Rh1 in engineered brewer's yeast by T114K was 1.526 times that of wild-type protopanaxadiol synthase, representing a 52.65% increase.

[0031] The relative yield of rare ginsenoside Rh1 in engineered brewer's yeast using I117D was 1.390 times that of wild-type protopanaxadiol synthase, representing a 39.04% increase.

[0032] The relative yield of rare ginsenoside Rh1 in engineered brewer's yeast of I117Y was 1.211 times that of wild-type protopanaxadiol synthase, representing an increase of 21.06%.

[0033] The conformational bias-guided distal site mutation design includes the following steps:

[0034] (6) Construct a Compound I state complex model composed of protopanaxadiol synthase, protopanaxadiol and heme active intermediate, and perform molecular dynamics simulation on the complex model;

[0035] (7) Principal component analysis and cluster analysis were performed on the obtained molecular dynamics simulation trajectory to construct the conformational landscape of the protopanaxadiol synthase-protopanaxadiol complex, and the complex was divided into multiple conformational metastable states.

[0036] (8) Evaluate the catalytic potential of metastable states of different conformations based on the geometric criterion of near-attack conformation, and determine the catalytically favorable and catalytically unfavorable conformations;

[0037] (9) Compare the active pocket volume, nonpolarity ratio, volume fluctuation characteristics and residue contact network differences of metastable states in different conformations, and screen for distant candidate sites that may affect the distribution of metastable states in conformations.

[0038] (10) Using representative structures of catalytically favorable and catalytically unfavorable conformations as input, the conformational bias method and at least two protein sequence design models are used to predict the preference of single-point mutations for different conformational states and obtain distant candidate mutants.

[0039] (11) Construct and express the distal candidate mutant, determine its ability to synthesize rare ginsenoside Rh1 in engineered Saccharomyces cerevisiae, and screen for distal site mutants with improved catalytic performance.

[0040] The molecular dynamics simulation time for the Compound I state complex model described in step (6) is preferably 300 to 700 ns, more preferably 500 ns.

[0041] The near-attack conformation described in step (8) simultaneously satisfies the following geometric conditions:

[0042] The Fe-OH angle at the active site of heme is 100°–140°;

[0043] The CHO angle of the CH bond in protopanaxadiol to be hydroxylated is 140°–180°; and

[0044] The distance between the reactive oxygen atom of heme and the hydrogen atom to be hydroxylated from protopanaxadiol is no greater than 3.0 Å.

[0045] In step (8), the metastable conformations with a higher proportion of near-aggressive conformations are identified as catalytically favorable conformations, and the metastable conformations with a lower proportion are identified as catalytically unfavorable conformations.

[0046] The residue contact mentioned in step (9) refers to the distance between the side chain heavy atoms of non-adjacent amino acid residues not exceeding 4.5 Å.

[0047] The occupancy rate of residue contacts mentioned in step (9) is the proportion of the number of trajectory frames in which residue contacts occur to the total number of trajectory frames in the corresponding metastable conformation.

[0048] The distal candidate sites mentioned in step (9) preferably exclude highly conserved residues and residues that form a protein-protein interaction interface with cytochrome P450 reductase.

[0049] The protein sequence design model described in step (10) is preferably selected from at least two of ProteinMPNN, ESM-IF1, Frame2Seq and ThermoMPNN, and more preferably ProteinMPNN, ESM-IF1 and Frame2Seq.

[0050] The conformational bias method described in step (10) is to evaluate the mutant’s preference for different conformational states by comparing the fit scores of the same mutant sequence with catalytically favorable and catalytically unfavorable conformations.

[0051] The distal candidate mutants mentioned in step (10) preferably satisfy the following conditions: their standardized fit score relative to wild-type protopanaxadiol synthase under catalytically favorable conformation is greater than 0, and the difference between their standardized fit score under catalytically favorable conformation and their standardized fit score under catalytically unfavorable conformation is greater than 0.

[0052] Through the aforementioned conformational bias-guided distal site mutation design, protopanaxadiol synthase mutants with enhanced catalytic performance were obtained, including A125T, V102W, A125M, V102T, V102C, A125K, A125H, V102N, A125R, S232G, V102M, V102S, V102H, A125Q, V102A, V102Q, V102K, E259P, V102R, and V102G.

[0053] Preferably, the distal site mutant is selected from one of V102S, V102N, V102Q, V102H, V102R, V102T, VO2C and V102M.

[0054] More preferably, the distal site mutant is V102S, V102Q, or V102N.

[0055] Most preferably, the distal site mutant is V102S.

[0056] The V102S increases the proportion of the near-attack conformation in the protopanaxadiol synthase-protopanaxadiol complex from 1.86% in the wild type to 7.56%, an increase of approximately 4.06 times.

[0057] The relative yield of rare ginsenoside Rh1 in engineered brewer's yeast by V102S is 2.745 times that of wild-type protopanaxadiol synthase, representing an increase of 174.53%.

[0058] The relative yield of rare ginsenoside Rh1 in engineered brewer's yeast of V102N is 2.074 times that of wild-type protopanaxadiol synthase, representing an increase of 107.42%.

[0059] The relative yield of rare ginsenoside Rh1 in engineered brewer's yeast using V102Q was 1.969 times that of wild-type protopanaxadiol synthase, representing a 96.87% increase.

[0060] This invention also provides a protopanaxadiol synthase mutant, obtained through the above-described rational design screening method. The mutant uses wild-type PPTS as the parent enzyme and is any one of the following mutants:

[0061] The threonine at position 114 is mutated to lysine, i.e., T114K;

[0062] The isoleucine at position 117 is mutated to aspartic acid or tyrosine, namely I117D or I117Y.

[0063] The valine at position 102 is mutated to arginine, lysine, alanine, glycine, serine, threonine, glutamine, asparagine, methionine, histidine, cysteine, aspartic acid, or glutamic acid.

[0064] The alanine at position 125 is mutated to threonine, methionine, lysine, histidine, arginine, or glutamine.

[0065] The isoleucine at position 109 is mutated to valine and the threonine at position 116 is mutated to lysine, i.e., I109V / T116K.

[0066] The valine at position 102 is mutated to serine and the isoleucine at position 109 is mutated to valine, i.e., V102S / I109V;

[0067] The valine at position 102 is mutated to serine and the alanine at position 115 is mutated to lysine, i.e., V102S / A115K.

[0068] The 114th threonine is mutated to lysine and the 115th alanine is mutated to lysine, i.e., T114K / A115K;

[0069] The alanine at position 115 is mutated to lysine and the threonine at position 116 is mutated to lysine, namely A115K / T116K;

[0070] The valine at position 102 is mutated to serine and the threonine at position 116 is mutated to lysine, i.e., V102S / T116K.

[0071] The isoleucine at position 109 is mutated to valine and the alanine at position 115 is mutated to lysine, i.e., I109V / A115K.

[0072] Preferably, the mutant is I109V / T116K, V102S / I109V, V102S / A115K, T114K / A115K, or A115K / T116K.

[0073] The relative yield of rare ginsenoside Rh1 in engineered brewer's yeast using I109V / T116K reached 3.047 times that of wild-type PPTS.

[0074] The present invention also provides a gene that encodes the above-mentioned protopanaxadiol synthase mutant.

[0075] The nucleotide sequence of the gene is preferably optimized according to the codon preference of Saccharomyces cerevisiae.

[0076] The present invention also provides a recombinant expression vector for expressing protopanaxadiol synthase, wherein the recombinant expression vector contains the above-mentioned gene.

[0077] The present invention also provides a recombinant host cell for expressing protopanaxadiol synthase, wherein the recombinant host cell contains the above-mentioned gene or recombinant expression vector, or the gene is integrated into its genome.

[0078] The chassis cells of the recombinant host cell are preferably Saccharomyces cerevisiae.

[0079] The recombinant host cells also express dammarene diol synthase PgDDS, protopanaxadiol synthase CYP716A47, cytochrome P450 reductase PgCPR1, cytochrome P450 reductase PgCPR2, and glycosyltransferase UGTPg100.

[0080] The dammarene diol synthase PgDDS is used to catalyze the production of dammarene diol from 2,3-oxidized squalene. Its amino acid sequence is preferably shown in SEQ ID NO.3, and the nucleic acid sequence of its encoding gene is preferably shown in SEQ ID NO.4.

[0081] The protopanaxadiol synthase CYP716A47 and the cytochrome P450 reductase PgCPR1 are used to catalyze the production of protopanaxadiol from dammarenediol. The preferred amino acid sequence of the protopanaxadiol synthase CYP716A47 is shown in SEQ ID NO. 5, and the preferred nucleic acid sequence of its encoding gene is shown in SEQ ID NO. 6. The preferred amino acid sequence of the cytochrome P450 reductase PgCPR1 is shown in SEQ ID NO. 7, and the preferred nucleic acid sequence of its encoding gene is shown in SEQ ID NO. 8.

[0082] The protopanaxadiol synthase mutant and the cytochrome P450 reductase PgCPR2 are used to catalyze the production of protopanaxadiol from protopanaxadiol. The preferred amino acid sequence of the cytochrome P450 reductase PgCPR2 is shown in SEQ ID NO. 9, and the preferred nucleic acid sequence of its encoding gene is shown in SEQ ID NO. 10.

[0083] The UGTPg100 is used to catalyze the production of rare ginsenoside Rh1 from protopanaxadiol. Its amino acid sequence is preferably shown in SEQ ID NO.11, and its encoding gene nucleic acid sequence is preferably shown in SEQ ID NO.12.

[0084] The present invention also provides the application of the above-mentioned protopanaxadiol synthase mutant, gene, recombinant expression vector or recombinant host cell in the preparation of protopanaxadiol and / or rare ginsenoside Rh1.

[0085] This invention also provides a method for preparing rare ginsenoside Rh1 using brewer's yeast, comprising the following steps:

[0086] (1) The gene encoding the above-mentioned original ginsenoside synthase mutant was introduced into Saccharomyces cerevisiae with a ginsenoside synthesis pathway to obtain recombinant Saccharomyces cerevisiae;

[0087] (2) Cultivate the recombinant brewer's yeast to express the protopanaxadiol synthase mutant and synthesize rare ginsenoside Rh1; and detect and / or isolate rare ginsenoside Rh1 from the culture.

[0088] The enzymes involved in the ginsenoside synthesis pathway are preferably PgDDS, CYP716A47, PgCPR1, PgCPR2, and UGTPg100.

[0089] The recombinant brewer's yeast is preferably a brewer's yeast strain that has integrated the heterologous synthesis pathway of rare ginsenoside Rh1, and can express the above-mentioned protopanaxadiol synthase mutant.

[0090] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0091] (1) Rapid and efficient: In the past, P450 enzyme modification often required the construction of mutant libraries of tens of thousands of units and relied on expensive high-throughput screening equipment; this invention adopts a rational design strategy of "active pocket optimization and remote conformation regulation" in parallel, screening 20 mutants for verification in each direction, which greatly reduces the scale of wet experimental screening and significantly improves the hit efficiency of candidate mutants. If this dry-to-wet workflow is platformized, it will exponentially reduce the trial and error costs of enterprise R&D pipelines.

[0092] (2) This invention uses MM / PBSA-guided active site mutation design to obtain three dominant mutants, T114K, I117D, and I117Y, from 20 candidate mutants, with an experimental hit rate of 15%. Among them, the relative yields of rare ginsenoside Rh1 in engineered Saccharomyces cerevisiae by the T114K, I117D, and I117Y mutants reached 1.526 times, 1.390 times, and 1.211 times that of the wild type, respectively, representing increases of 52.65%, 39.04%, and 21.06%.

[0093] (3) Through conformational bias-guided distal site mutation design, this invention obtained 18 mutants with higher Rh1 yields than wild-type mutants among 20 distal candidate mutants, achieving an experimental hit rate of 94.74%. Among them, the relative yields of rare ginsenoside Rh1 in the V102S, V102N, and V102Q mutants reached 2.745 times, 2.074 times, and 1.970 times that of wild-type mutants, respectively, representing increases of 174.53%, 141.08%, and 107.42%. The V102S mutant also increased the proportion of the proximal attack conformation in the enzyme-substrate complex from 1.86% to 7.56%.

[0094] (4) This invention further constructed a dual-site combined mutant of protopanaxadiol synthase, verifying the potential for combined application between the active pocket mutant site and the distal regulatory mutant site. Specifically, the relative Rh1 yield of the I109V / T116K combined mutant reached 3.047 times that of the wild type; the relative Rh1 yields of the V102S / I109V, V102S / A115K, T114K / A115K, and A115K / T116K combined mutants reached 2.187 times, 2.188 times, 2.069 times, and 2.062 times that of the wild type, respectively. These results indicate that the mutant sites screened in this invention can not only be used as single-point mutations to improve the catalytic performance of PPTS, but also, through combined mutations, further obtain protopanaxadiol synthase mutants with enhanced Rh1 synthesis capacity. Attached Figure Description

[0095] Figure 1 is a flowchart of the dual-track rational design of protopanaxadiol synthase activity pocket mutation and distal conformational regulation mutation.

[0096] Figure 2 is a diagram showing the binding free energy decomposition of MM / PBSA residues in the complex of protopanaxadiol synthase and protopanaxadiol.

[0097] Figure 3 shows the distribution and sequence conservation analysis results of the substrate binding pocket residues of protopanaxadiol synthase.

[0098] Figure 4 is an evaluation diagram of substrate binding energy and folding stability of the candidate site of protopanaxadiol synthase.

[0099] Figure 5 shows the principal component analysis and metastable conformational distribution results of the protopanaxadiol synthase-protopanaxadiol complex.

[0100] Figure 6 This is a diagram showing the geometric distribution of the near-attack conformation of protopanaxadiol under different metastable conformations.

[0101] Figure 7 shows the contact occupancy of key residue pairs under different metastable conformations.

[0102] Figure 8 is a map of a low-copy plasmid containing the PPTS gene.

[0103] Figure 9 shows the yield of rare ginsenoside Rh1 generated from the screened active pocket candidate mutants.

[0104] Figure 10 shows the results of conformational bias prediction scores and the yield of rare ginsenoside Rh1 in distant candidate mutants.

[0105] Figure 11 shows the results of Rh1 synthesis performance verification of the PPTS combined mutant. Detailed Implementation

[0106] The present invention will be further described in detail below through specific embodiments.

[0107] It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0108] Experimental methods in the following examples, unless otherwise specified, are generally performed under standard conditions or as recommended by the manufacturer. Unless otherwise specified, the materials and reagents used in these examples are commercially available.

[0109] The composition of 1 mL conversion solution is: 800 μL 50% PEG3350, 200 μL 2 mol / L LiAc, 7.5 μL β-mercaptoethanol, and the remainder is deionized water.

[0110] The composition of leucine-deficient agar solid medium is as follows: 6.7 g / L amino-free yeast nitrogen source, 20 g / L glucose, 50 mg / L histidine, 50 mg / L tryptophan, 50 mg / L uracil, and 15 g / L agar powder.

[0111] The rational design method of the protopanaxadiol synthase of the present invention is as follows: Figure 1 As shown in Tables 1-4, the primers used to construct the mutants are listed in Tables 1-4.

[0112] Table 1 Primers for PPTS expression vector construction

[0113]

[0114] Table 2 Primers required for unit point mutants

[0115]

[0116] Table 3 Primers required for unit point mutants

[0117]

[0118] Table 4 Primers required for double mutants

[0119]

[0120] Note: The primers used in two-site mutations are single-point mutation plasmids, selected according to the mutation site starting with the primer name. For example, the template for V102S-I109V-F and V102S-I109V-R is pY26-PPTS. V102S .

[0121] The primers required for constructing the engineered strain S. cerevisiae P1 are shown in Table 5:

[0122] Table 5 Primers used in the construction of yeast strain P1

[0123]

[0124] Example 1: Mutation design of protopanaxadiol synthase active site based on MM / PBSA

[0125] This embodiment uses protopanaxadiol synthase CYP716A53v2 (PPTS) as the target for modification and protopanaxadiol (PPD) as the substrate. Through molecular dynamics simulations, decomposition of MM / PBSA residue binding free energy, sequence conservation analysis, virtual saturation mutagenesis, molecular docking, and pre-reaction state constraints, candidate mutants of active sites that may improve the catalytic performance of protopanaxadiol synthase were screened. The specific steps are as follows:

[0126] (1) Construction of the PPTS-PPD complex model

[0127] The three-dimensional structure of protopanaxadiol synthase PPTS (amino acid sequence shown in SEQ ID NO.1) was predicted using AlphaFold3, and the predicted protein structure was used as the initial model for subsequent molecular simulations. PPTS belongs to the cytochrome P450 monooxygenase family, and its catalytic center contains a heme prosthetic group. The heme prosthetic group was assembled into the active site of the protein, and the proximal cysteine ​​residue was coordinated with the heme iron atom.

[0128] MDLFISSQLLLLLVFCLFLFWNFKPSSQNKLPPGKTGWPIIGETLEFISCGQKGNPEKFVTQRMNKYSPDVFTTSLAGEKMVVFCGASGNKFIFSNENKLVVSWWPPAISKILTATI PSVEKSKALRSLIVEFLKPEALHKFISVMDRTTRQHFEDKWNGSTEVKAFAMSESLTFELACWLLFSINDPVQVQKLSHLFEKVKAGLLSLPLNFPGTAFNRGIKAANLIRKELSVVI KQRRSDKLQTRKDLLSHVMLSNGEGEKFFSEMDIADVVLNLLIASHDTTSSAMGSVVYFLADHPHIYAKVLTEQMEIAKSKGAEELLSWEDIKRMKYSRNVINEAMRLVPPSQGGFK VVTSKFSYANFIIPKGWKIFWSVYSTHKDPKYFKNPEEFDPSRFEGDGPMPFTFIPFGGGPRMCPGSEFARLEVLIFMHHLVTNFKWEKVFPNEKIIYTPFPFPENGLPIRLSPCTL.

[0129] The three-dimensional structure of the substrate PPD was obtained, and hydrogenation, energy minimization, and charge calculations were performed. AutoDockVina was used to dock PPD to the vicinity of the heme active site of PPTS, establishing a docking region containing the heme prosthetic group and its surrounding substrate binding pocket, and thoroughly sampling the substrate conformation. Based on the molecular docking score, the spatial matching degree between the substrate and the active pocket, and the relative position of the C6 position of PPD to be hydroxylated and the heme iron atom, a reasonable binding conformation was selected to obtain an initial model of the PPTS-PPD-heme complex.

[0130] (2) Molecular dynamics simulation of the complex

[0131] Molecular dynamics simulations of the complex were performed using Gromacs 2025.4 software. The protein was simulated using the Amber14SB force field, the PPD using the GAFF2 force field, and the water molecule using the TIP3P water model. The active site of the PPTS heme was set to the resting state of Fe(III) five-coordinate high-spin sixtt trivalent iron (IC6). The heme prosthetic group and its coordinating cysteine ​​residues were assigned to AMBER-compatible force field parameters for the P450 IC6 state as reported in the literature to maintain the coordination relationship between the heme iron atom and the proximal cysteine ​​residue.

[0132] The complex was placed in a cubic water box, with the minimum distance between the protein surface and the water box boundary set to 1.0 nm, and an appropriate amount of Na was added. + and Cl - The total charge of the simulated system was neutralized. The cutoff radii for both electrostatic and van der Waals interactions were set to 1.0 nm, and the integration time step was set to 2 fs.

[0133] First, the steepest descent method was used to minimize the energy of the simulated system. Then, with positional constraints imposed on the protein backbone atoms and PPD heavy atoms, 5 ns of NVT equilibrium simulation and 5 ns of NPT equilibrium simulation were performed sequentially. The temperature for the NVT equilibrium simulation was set to 300 K, and the pressure for the NPT equilibrium simulation was set to 1 bar. After the system reached equilibrium, 300 ns of unconstrained production-period molecular dynamics simulations were performed at 300 K and 1 bar.

[0134] (3) Stability analysis of the molecular dynamics simulation system

[0135] Root mean square bias (RMSB) analysis was used to analyze the changes in the PPTS backbone structure over simulation time to evaluate the overall conformational stability of the protein. Root mean square fluctuation (RMSF) analysis was used to analyze the flexibility changes of individual amino acid residues during the simulation. Simultaneously, the distance between the C6 position of the PPD to be hydroxylated and the heme iron atom was calculated over simulation time to evaluate the binding stability of the substrate in the active pocket. After initial equilibration, the overall RMSB of the PPTS-PPD complex remained stable, and the RMSF of each residue did not show an abnormal overall increase. The distance between the C6 position of the PPD to be hydroxylated and the heme iron atom remained relatively stable during the simulation, indicating that the constructed enzyme-substrate complex can be used for subsequent binding free energy and active pocket analysis.

[0136] (4) Decomposition of MM / PBSA binding free energy of active pocket residues

[0137] The last 100 ns of the molecular dynamics trajectory during the 300 ns production phase was selected as the equilibrium analysis window. Using the PPD as the center, amino acid residues with a distance of no more than 5 Å from any substrate atom were screened, resulting in 24 substrate-binding pocket residues.

[0138] The binding free energy of PPTS and PPD was calculated using the MM / PBSA method, and the binding free energy was further decomposed to individual amino acid residues to evaluate the energy contribution of different residues to substrate binding. The more negative the energy contribution of a residue, the more favorable its contribution to substrate binding; a positive or near-zero energy contribution indicates that the residue may have an unfavorable interaction with the substrate or a weak contribution to substrate binding.

[0139] like Figure 2As shown, P106, S110, P346, Q348, G349, F351, W373, R414, and P452 exhibit unfavorable energy contributions to PPD binding; T114, I117, N275, T283, and F453 have weaker energy contributions. Using a binding free energy contribution greater than -0.5 kcal / mol as a reference condition for weak or unfavorable contributions, a total of 14 initial candidate residues that may affect substrate binding were screened.

[0140] (5) Sequence conservation analysis of candidate residues

[0141] Multiple sequence alignment and conservation analysis were performed on the amino acid sequences of PPTS and its homologous P450 protein. To minimize the adverse effects of mutations on the overall folding, heme binding, and key catalytic functions of P450, residues with lower conservation and located near the substrate-binding pocket were preferentially retained, while highly conserved residues or those potentially involved in maintaining the core structure of P450 were excluded. Figure 3 By combining the MM / PBSA residue energy decomposition results with sequence conservation analysis results, seven candidate active sites—S110, T114, I117, N275, Q348, P452, and F453—were ultimately identified for subsequent virtual saturation mutagenesis design. Figure 4 ).

[0142] (6) Virtual saturation mutations at candidate sites

[0143] Virtual saturation mutations were performed on S110, T114, I117, N275, Q348, P452, and F453 using FoldX software. For each candidate site, wild-type amino acids were replaced with the remaining natural amino acids, and the local side chain conformation of the mutants was optimized, ultimately constructing 133 virtual single-point mutant models of protopanaxadiol synthase. The change in folding free energy of each mutant relative to wild-type PPTS was calculated. The calculation formula is as follows: .in, The lower the value, the smaller the impact of the mutation on the stability of the protein structure; A significant increase indicates that the mutation may disrupt protein folding stability.

[0144] (7) High-throughput molecular docking of virtual mutants

[0145] AutoDock Vina was used to perform molecular docking between 133 virtual mutants and PPD. To fully sample the possible binding modes of the substrate in the active pocket, 40 candidate binding conformations were generated for each mutant, resulting in a total of 5320 mutant-substrate binding conformations.

[0146] Since molecular docking scores only reflect the relative binding tendency between proteins and substrates and cannot directly determine whether the substrate has a reasonable catalytic orientation, a geometric constraint of the pre-reaction state is further introduced on the basis of docking scores.

[0147] (8) Screening of candidate mutants based on pre-reaction state

[0148] When protopanaxadiol synthase catalyzes the C6 hydroxylation of PPD, the C6 position of PPD needs to be close to the active site of heme. A binding conformation in which the distance between the C6 carbon atom of PPD to be hydroxylated and the iron atom of heme is in the range of 3.5–5.0 Å is defined as a candidate conformation that meets the distance requirement of the pre-reaction state.

[0149] For each virtual mutant, firstly, conformations that meet the above distance conditions are screened from 40 docking conformations. Then, the conformation with the lowest molecular docking score is selected as the representative binding conformation for that mutant. This method eliminates false positive binding modes with low docking scores but unreasonable substrate catalytic orientation.

[0150] Further calculations were performed on the substrate binding free energy changes of each mutant relative to the wild type. The calculation formula is as follows: Preferred retention Mutants with a substrate binding capacity of <0 kcal / mol, i.e., mutants with superior substrate binding capacity compared to wild-type mutants; excluding... Mutants with a concentration >1.0 kcal / mol were selected to avoid significant disruption to protein structural stability caused by candidate mutations.

[0151] Following a stepwise screening process based on pre-reaction state distance, substrate binding ability, and protein folding stability, 20 candidate mutants for active site validation were ultimately obtained: T114W, N275I, S110F, F453L, S110W, S110Y, S110R, S110K, S110Q, T114K, T114H, I117D, I117V, S110E, T114M, T114L, T114E, S110P, I117Y, and S110C. These 20 candidate mutants were used for subsequent construction of the protopanaxadiol synthase mutant gene and experimental validation of the synthesis performance of rare ginsenoside Rh1 in engineered Saccharomyces cerevisiae.

[0152] Example 2: Design of distal site mutation of protopanaxadiol synthase based on conformational bias

[0153] This embodiment focuses on the enzyme-substrate complex formed by protopanaxadiol synthase (PPTS) and protopanaxadiol (PPD). Through the construction of a Compound I state model, molecular dynamics simulation, principal component analysis, conformational metastable state identification, near-attack conformation evaluation, active pocket kinetic analysis, residue contact network analysis, and conformational bias prediction, distal candidate mutants that may improve catalytic performance by regulating protein conformational distribution are screened. The specific steps are as follows:

[0154] (1) Construction of the PPTS-PPD Compound I-state complex model

[0155] Based on the enzyme-substrate complex formed by protopanaxadiol synthase (PPTS) and PPD, the heme active site was set as the Compound I state in the cytochrome P450 catalytic cycle. Compound I is a high-valent ferro-oxygen reactive intermediate with direct catalytic significance in the P450-catalyzed hydroxylation reaction, and can be used to characterize the relative geometric relationship between the CH bond of the substrate to be hydroxylated and reactive oxygen species. The heme force field parameters of the Compound I state adopted the publicly reported parameters of P450 high-valent ferro-oxygen intermediates. The heme prosthetic group was placed in the PPTS active site, and the proximal cysteine ​​residue was deprotonated to coordinate with the heme iron atom to maintain the structural stability of the heme active site during the simulation process. PPD was placed in the PPTS active pocket to obtain the initial model of the PPTS-PPD Compound I state complex.

[0156] (2) Molecular dynamics simulation of the Compound I state complex

[0157] Unbiased molecular dynamics simulations of the PPTS-PPD Compound I-state complex were performed using Gromacs software. The protein was subjected to an Amber14SB force field, the PPD to a GAFF2 force field, and the solvent to a TIP3P water model. The complex was placed in a cubic water tank with a minimum distance of 1.0 nm between the protein surface and the tank boundary, and an appropriate amount of Na was added. + and Cl - The total charge of the system was neutralized. The cutoff radii for both electrostatic and van der Waals interactions were set to 1.0 nm, and the integration time step was set to 2 fs. The system was first simulated using the steepest descent method for energy minimization, followed by NVT and NPT equilibrium simulations. After system equilibrium was achieved, unconstrained production-period molecular dynamics simulations were performed at 300 K and 1 bar for 500 ns.

[0158] The root mean square deviations (RMS) of the PPTS protein backbone and the PPD heavy atoms were calculated to evaluate the stability of the enzyme-substrate complex during the simulation. The results showed that neither the PPTS protein backbone nor the PPD exhibited sustained large shifts during the simulation, indicating that the obtained trajectories can be used for subsequent conformational landscape and near-attack conformational analysis.

[0159] (3) Principal component analysis and metastable state classification of PPTS-PPD complex

[0160] Principal component analysis was performed on the molecular dynamics simulation trajectory during the 500 ns production period. First, the protein backbone atoms in the trajectory were structurally superimposed to remove overall translational and rotational motions. Then, the atomic coordinate covariance matrix was calculated, and eigenvalue decomposition was performed on the covariance matrix to obtain the principal components describing the main conformational changes of the complex. The molecular dynamics simulation trajectory was projected onto a two-dimensional space composed of the first principal component PC1 and the second principal component PC2 to construct a conformational distribution map of the PPTS-PPD complex. Further, a clustering method was used to classify the trajectory frames in the principal component space, dividing the complex into four main metastable conformations: S1, S2, S3, and S4. Figure 5 The population proportions of metastable states in each conformation are as follows: S1 state accounts for 54.09% of the total trajectory frames, S2 state accounts for 19.26%, S3 state accounts for 17.18%, and S4 state accounts for 9.47%. Among them, S1 is the dominant conformation state with the highest proportion in the trajectory. The conformation corresponding to the region with the highest metastable state density is selected as the representative conformation of that state for subsequent active pocket analysis, residue contact analysis, and conformational bias prediction.

[0161] (4) Proximal attack conformation analysis of PPTS catalytic selectivity

[0162] To evaluate the catalytic selectivity of PPTS for different potential hydroxylation sites of PPD, the spatial geometric relationships between the C6α, C6β, C7α, C7β, C15α, and C15β sites of PPD and the Compound I reactive oxygen species were analyzed. The near-attack conformation refers to a conformation in which the reactant atom of the substrate and the catalytically active center are close in distance and angle to the geometric state required for the chemical reaction. In this embodiment, a trajectory frame that simultaneously satisfies the following conditions is defined as a near-attack conformation: the Fe-OH angle is 100°–140°, the CHO angle is 140°–180°, and the distance between the Compound I reactive oxygen atom and the hydrogen atom of the substrate to be hydroxylated is no greater than 3.0 Å. Here, C and H represent the carbon atom of PPD to be hydroxylated and the hydrogen atom connected to it, respectively, and O represents the reactive oxygen atom in the Compound I state.

[0163] Statistical analysis of the entire 500 ns trajectory revealed that the C6α site of PPD was more likely to satisfy the near-attack conformation geometry conditions than the C6β, C7α, C7β, C15α, and C15β sites. This indicates that PPTS can preferentially locate the C6α site of PPD to a spatial position conducive to hydroxylation during the dynamic process, which is consistent with the selectivity of PPTS in catalyzing the formation of protopanaxadiol from PPD.

[0164] (5) Evaluation of near-attack conformations in metastable states of different conformations

[0165] Statistically analyze the proportion of trajectory frames that satisfy the geometric conditions of the near-attack conformation in the four metastable states of conformation S1 to S4. Figure 6 The results are as follows: In state S1, the proportions of conformations satisfying the Fe-OH angle, CHO angle, OH distance, and all near-attack conformation conditions were 98.24%, 11.68%, 2.13%, and 0.83%, respectively; in state S2, the proportions were 95.48%, 13.91%, 8.85%, and 4.67%; in state S3, the proportions were 98.31%, 11.99%, 6.29%, and 2.15%; and in state S4, the proportions were 87.54%, 3.17%, 45.09%, and 1.48%. Throughout the entire 500 ns simulated trajectory, the proportions of conformations satisfying the Fe-OH angle, CHO angle, OH distance, and all near-attack conformation conditions were 96.70%, 11.36%, 8.21%, and 1.86%, respectively.

[0166] The results showed that although S1 was the dominant conformation state with the highest population proportion, its near-aggressive conformation proportion was only 0.83%, lower than that of S2, S3, and S4 states. Among them, the near-aggressive conformation proportion was the highest in state S2, reaching 4.67%, indicating that S2 was more conducive to the formation of catalytically ready conformations than S1. Therefore, S1 was defined as a relatively catalytically unfavorable conformation, and S2 was defined as a relatively catalytically favorable conformation. Representative conformations of S1 and S2 were selected for subsequent conformational bias prediction.

[0167] (6) Kinetic analysis of active pockets in metastable states of different conformations

[0168] To elucidate the structural basis for the differences in metastable catalytic potential across different conformations, the active pocket volume, nonpolar ratio, and volume fluctuation characteristics of the S1–S4 states were analyzed. The pocket parameters for each metastable conformation are as follows: the median active pocket volume of the S1 state is 259.0856 Å. 3 The average nonpolar proportion was 77.15%, and the volume fluctuation index was 0.3795; the median active pocket volume of the S2 state was 291.7469 Å. 3The average nonpolar proportion was 60.43%, and the volume fluctuation index was 0.3716; the median active pocket volume of the S3 state was 272.2901 Å. 3 The average nonpolar proportion was 68.75%, and the volume fluctuation index was 0.3975; the median active pocket volume of the S4 state was 252.6548 Å. 3 The average nonpolarity ratio is 58.45%, and the volume fluctuation index is 0.3204.

[0169] The results show that the active pocket in the S1 state has a high proportion of nonpolar molecules and a relatively compact spatial configuration. Although this state can stabilize the binding of PPD through hydrophobic interactions, it may limit the conformational adjustment of PPD within the active pocket, making it difficult for PPD to enter the ideal near-attack conformation. In contrast, the active pockets in the S2 and S3 states have larger volumes and relatively open microenvironments, which are more conducive to PPD adjusting its reaction posture and forming catalytically relevant conformations.

[0170] (7) Analysis of metastable residue contact networks in different conformations

[0171] To further identify distal residues that may be involved in conformational changes, the residue contact network in each metastable state (S1–S4) was analyzed. A residue contact was defined as a distance of no more than 4.5 Å between the side chain heavy atoms of two non-adjacent amino acid residues. The contact occupancy for any residue pair was calculated using the following formula:

[0172] .

[0173] During the calculation, directly adjacent residue pairs on the primary sequence were excluded to avoid interference from covalent linkages in the contact network analysis. Residue contact occupancy matrices for states S1, S2, S3, and S4 were constructed, and the differences between state S1 and states S2–S4 were compared. The results showed that the overall protein backbone contact patterns were relatively consistent across different metastable conformations, but significant differences existed in the contact occupancy of several local residue pairs. Figure 7 ).

[0174] For example, some residue pairs have a high contact occupancy rate in the S1 state, but the contact is significantly weakened in the S2, S3, or S4 states; other residue pairs have weak contact in the S1 state, but the contact is significantly enhanced in other states. These changes indicate that the transition between different metastable conformations is accompanied by a rearrangement of the local residue interaction network. These residues may regulate the morphology of active pockets, substrate localization, and the formation of proximity-attack conformations by altering the local stability and motion coupling relationships of proteins. Based on residue pairs with significant changes in contact occupancy rates between S1 and other metastable states, 31 candidate residues that may be involved in conformational regulation were initially identified.

[0175] (8) Further screening of distant candidate sites

[0176] Sequence conservation analysis was performed on the 31 candidate residues, and highly conserved residues that might maintain the core structure and function of P450 were excluded. A protein complex model of PPTS and cytochrome P450 reductase PgCPR1 was further constructed to identify the interaction interface residues between the two, and interface residues that might affect electron transfer and protein-protein interactions between P450 and the reductase were excluded. After conservation screening and protein-protein interaction interface screening, 23 distal candidate sites for conformational bias prediction were finally obtained. These candidate sites do not have direct contact with PPD and the heme active site; their potential role is mainly to indirectly affect the catalytic process by regulating the local protein contact network, flexibility, and conformational metastable distribution.

[0177] (9) Selection of representative conformations of S1 and S2

[0178] The trajectory frames corresponding to the positions with the highest principal component spatial density in the metastable states S1 and S2 were selected as representative conformations of S1 and S2, respectively. The root mean square deviation of the main chain between the two representative conformations was 1.171 Å, indicating that their overall structures are similar, but there are differences in local conformations and interaction networks that can be identified by the conformation-sensitive model. S1 and S2 were selected as the inputs for conformation bias prediction, where S1 represents a conformational state with a high population proportion but low catalytic potential, and S2 represents a catalytically favorable state with a low population proportion but a high proportion of near-aggressive conformations.

[0179] (10) Prediction of distant mutations and multi-model cross-screening and identification of candidate mutants based on conformational bias method

[0180] Using representative S1 and S2 conformations as inputs, four protein backfolding models—ProteinMPNN, ESM-IF1, Frame2Seq, and ThermoMPNN—were employed to perform single-point mutation scanning at 23 distal candidate sites based on conformational bias. For each candidate mutant, its structural preference score relative to the wild-type sequence was calculated within both the S1 and S2 structural contexts. To ensure comparability of scores across different structures and models, the scores obtained within each conformation were normalized to 0-1. For any mutant, the S2 conformational bias score was calculated using the following formula: .in, This indicates the normalized fit score of the mutant relative to the wild type in the S2 conformation. This represents the normalized fit score of the mutant relative to the wild type in the S1 conformation. A mutant that simultaneously meets the following criteria is defined as an S2-preferred mutant: >0 and >0. The above screening criteria can avoid selecting mutants simply because they have a high absolute score in the S2 conformation, but require that the mutant has a more obvious relative preference for the S2 conformation compared to the S1 conformation.

[0181] The conformational bias prediction results of four models—ProteinMPNN, ESM-IF1, Frame2Seq, and ThermoMPNN—were compared. The results showed that the prediction trend of ThermoMPNN differed significantly from the other three models. Therefore, the prediction results of ProteinMPNN, ESM-IF1, and Frame2Seq were used as the primary basis for ranking candidate mutants. For mutants predicted as S2-biased by multiple models, the average S2 conformational bias score in the ProteinMPNN, ESM-IF1, and Frame2Seq models was calculated. These mutants were then ranked from highest to lowest based on their average conformational bias score, and the top 20 candidate mutants were selected for further experimental validation. The candidate mutants include A125T, V102W, A125M, V102T, V102C, A125K, A125H, V102N, A125R, S232G, V102M, V102S, V102H, A125Q, V102A, V102Q, V102K, E259P, V102R, and V102G.

[0182] Example 3: Construction of protopanaxadiol synthase mutant and verification of the synthetic performance of rare ginsenoside Rh1

[0183] In this embodiment, Saccharomyces cerevisiae Sq-0 was used as the starting strain (disclosed in Chinese Invention Patent CN120574697B, which can be prepared according to Example 1 of Chinese Invention Patent CN120574697B). The expression cassettes of dammarene diol synthase PgDDS (GenBank accession number: AB265170.1), protopanaxadiol synthase CYP716A47 (GenBank accession number: JN604537.1), cytochrome P450 reductase PgCPR1 (GenBank accession number: AIC73829.1), cytochrome P450 reductase PgCPR2 (GenBank accession number: QRG28713.1), and glycosyltransferase UGTPg100 (GenBank accession number: A0A0K0PVW1.1) were integrated to obtain the recombinant strain S. cerevisiae P1. The specific operations are as follows: The expression module PgDDS is integrated at the X-3 site (chromosome coordinates Chr X: 223616...224744), the expression module genes CYP716A47 and PgCPR1 are integrated at the XI-3 site (chromosome coordinates Chr XI: 93378...94567), and the expression module genes PgCPR2 and UGTPg100 are integrated at the LPP1 site (SGD number: S000002911). In this embodiment, the specific integration method uses the CRISPR-Cas9 system to edit the selected genomic sites of *Saccharomyces cerevisiae*, and designs the 20 nt sequence of sgRNA on the website CHOPCHOP. The 20 nt sequences of the X-3, XI-3, and LPP1 integration sites are: GACACATTAGTCTCGTATGT, GTAGAAATCAGACGCACGCT, and ATGAAACTTGAATGTCCGCT, respectively. When constructing the Cas9-sgRNA plasmid, a designed 20 nt sequence was used as the homologous arm. Primers X-3-sgRNA-F and X-3-sgRNA-R, XI-3-sgRNA-F and XI-3-sgRNA-R, and LPP1-sgRNA-F and LPP1-sgRNA-R were used to target p426-P TEF1 -SpCas9-T CYC1 -P SNR52 -sgRNA-T SUP4(The plasmid constructed according to Example 1 of Chinese Invention Patent CN120574697B, as disclosed in Chinese Invention Patent CN120574697B, was linearized by PCR. After recovering the fragments at the three sites, 100 ng of the linearized plasmid was transformed into E. coli DH5α competent cells. To ensure the correctness of the plasmid construction, three positive transformants were randomly selected for culture. After extracting the plasmid, it was sent to Shanghai Sangon Biotech Co., Ltd. for sequencing verification. At this time, Cas9-sgRNA plasmids with sgRNA sequences targeting the X-3, XI-3, and LPP1 sites were constructed, namely Cas9-sgRNA-A, Cas9-sgRNA-B, and Cas9-sgRNA-C.

[0184] PgDDS gene editing involves introducing upstream and downstream homologous arms at the X-3 site using primers. After PCR amplification and fusion PCR assembly, the resulting product is purified into a high concentration of T using a kit. ADH1 -PgDDS-Pg al1,10 Linear donor DNA. The specific procedure is as follows: Using the PUC57-PgDDS plasmid synthesized by Sangon Biotech (the nucleotide sequence of PgDDS is shown in SEQ ID NO.4) as a template, and using PgDDS-F and PgDDS-R as primers, the PgDDS fragment was obtained; using the pGAL1,10-MCS-His-MCS-Flag-URAp426Gal plasmid (Beyotime Biotech) as a template, and using X-3-T... ADH1 -F and X-3-T ADH1 -R is a primer, yielding the terminator T. ADH1 ; with X-3-P gal1,10 -F and X-3-P gal1,10 -R is a primer, resulting in the bidirectional promoter X-3-P. gal1,10 The bidirectional promoter X-3-P gal1,10 PgDDS gene fragment and terminator T ADH1 Mix the ingredients at a molar ratio of 1:1:1, add Gibson assembly reaction solution, and ligate at 50℃ for 45 min. Transform the vector into *E. coli* DH5α strain. Positive single clones are screened using the ampicillin resistance of the plasmid. Successful ligation is verified by colony PCR and sequencing. The correct vector is then used with primer X-3-T. ADH1 -F and X-3-P gal1,10 -R was used for PCR amplification, and the gene fragment was purified and recovered using the SanPrep column PCR product purification kit to obtain donor DNA-A with PgDDS inserted at the X-3 site.

[0185] Gene editing of CYP716A47 and PgCPR1 involved introducing upstream and downstream homologous arms at the XI-3 site using primers. After PCR amplification and fusion PCR assembly, the resulting product was purified into a high concentration of P using a kit. gal1,10 -CYP716A47-T ALT1 -P gal7 -PgCPR1-T CYC1 Linear donor DNA. The specific procedures are as follows: Using the PUC57-CYP716A47 plasmid synthesized by Sangon Biotech (the nucleotide sequence of CYP716A47 is shown in SEQ ID NO. 6) as a template, and using CYP716A47-F and CYP716A47-R primers, the CYP716A47 fragment was obtained. Using the PUC57-PgCPR1 plasmid synthesized by Sangon Biotech (the nucleotide sequence of PgCPR1 is shown in SEQ ID NO. 8) as a template, and using PgCPR1-F and PgCPR1-R primers, the PgCPR1 fragment was obtained. Using the pGAL1,10-MCS-His-MCS-Flag-URAp426 Gal plasmid as a template, and using XI-3-P... gal1,10 -F and XI-3-P gal1,10 -R is a primer, resulting in the bidirectional promoter XI-3-P. gal1,10 Using the Saccharomyces cerevisiae CEN PK2-1D genome as a template, and with T... ALT1 -F and T ALT1 -R is a primer, yielding the ALT1 GAA1 Terminator, abbreviated as T terminator. ALT1 Using Pgal7-F and Pgal7-R as primers, the promoter Pgal7 was obtained; using XI-3-T CYC1 -F and XI-3-T CYC1 -R is a primer, yielding the terminator XI-3-T CYC1 The CYP716A47 gene fragment and the bidirectional promoter XI-3-P were used. gal1,10 PgCPR1 gene fragment, terminator T ALT1 and terminator XI-3-T CYC1 Mix the ingredients in a molar ratio of 1:1:1:1:1, add Gibson assembly reaction solution, and ligate at 50℃ for 45 min. Transform the vector into *E. coli* DH5α strain. Positive single clones were screened using the ampicillin resistance of the plasmid. Successful ligation was verified by colony PCR and sequencing. The correct vector was then used with primer XI-3-P. gal1,10 -F and XI-3-T CYC1-R was used for PCR amplification, and the gene fragment was purified and recovered using the SanPrep column PCR product purification kit to obtain donor DNA-B with CYP716A47 and PgCPR1 inserted at the XI-3 site.

[0186] Gene editing of PgCPR2 and UGTPg100 involved introducing primers into the upstream and downstream homologous arms of the LPP1 site. After PCR amplification and fusion PCR assembly, the resulting product was purified into a high concentration of T using a kit. ADH1 -PgCPR2-P gal1,10 -UGTPg100-T ALT1 Linear donor DNA. The specific procedures are as follows: Using the PUC57-PgCPR2 plasmid synthesized by Sangon Biotech (the nucleotide sequence of PgCPR2 is shown in SEQ ID NO.10) as a template, and using PgCPR2-F and PgCPR2-R primers, the PgCPR2 fragment was obtained. Using the PUC57-UGTPg100 plasmid synthesized by Sangon Biotech (the nucleotide sequence of UGTPg100 is shown in SEQ ID NO.12) as a template, and using UGTPg100-F and UGTPg100-R primers, the UGTPg100 fragment was obtained. Using the pGAL1,10-MCS-His-MCS-Flag-URAp426 Gal plasmid as a template, and using LPP1-T... ADH1 -F and LPP1-T ADH1 -R is a primer, yielding the terminator LPP1-T ADH1 ; with LPP1-P gal1,10 -F and LPP1-P gal1,10 -R is a primer, resulting in the bidirectional promoter LPP1-P gal1,10 Using the CENPK2-1D genome of *Saccharomyces cerevisiae* as a template, and LPP1-T... ALT1 -F and LPP1-T ALT1 -R is a primer, yielding the terminator LPP1-T ALT1 The PgCPR2 gene fragment and the bidirectional promoter LPP1-P were used. gal1,10 UGTPg100 gene fragment, terminator LPP1-T ADH1 and terminator LPP1-T ALT1 Mix the ingredients in a molar ratio of 1:1:1:1:1, add Gibson assembly reaction solution, and ligate at 50℃ for 45 min. Transform the vector into *E. coli* DH5α strain. Positive single clones were screened using the ampicillin resistance of the plasmid. Successful ligation was verified by colony PCR and sequencing. The correct vector was then used with primers LPP1-T. ADH1 -F and LPP1-T ALT1-R was used for PCR amplification, and the gene fragment was purified and recovered using the SanPrep column PCR product purification kit to obtain donor DNA-C with PgCPR2 and UGTPg100 inserted at the LPP1 site.

[0187] Competent cells of wild-type S. cerevisiae Sq-0 were prepared and edited according to the instructions of the ZYMO Frozen-EZ Yeast Transformation II Kit. During the experiment, approximately 500 ng of Cas9-sgRNA plasmid and 1 μg of linear donor DNA were added. The transformation system was directly plated on uracil-YNB auxotrophic selection plates and incubated at 30°C for 4–5 days to form single colonies. Initial screening of yeast transformants was performed using colony PCR: single colonies from the uracil-YNB auxotrophic plates were picked and placed in tubes containing the PCR system. Primer pairs were used at X-3, XI-3, and LPP1 sites to verify whether the target gene was integrated into the yeast genome. To ensure the continued use of the uracil URA3 selection tag in subsequent experiments, the tag integrated into the gene was discarded. The specific procedure is as follows: Transformants with correct initial screening bands are inoculated into YPD liquid medium and cultured in a shaker at 220 rpm and 30℃ for 16 h. The culture is then spread onto solid YPD plates containing 5-fluoroorotic acid for negative screening. Colonies growing on the plates are then verified by PCR again. Correctly verified strains indicate that the labels have been discarded. Transformations are performed sequentially using Cas9-sgRNA-A / donor DNA-A, Cas9-sgRNA-B / donor DNA-B, and Cas9-sgRNA-C / donor DNA-C, respectively. This results in the selection of engineered strains *S. cerevisiae* P1 with PgDDS inserted at site X-3, CYP716A47 and PgCPR1 inserted at site XI-3, and PgCPR2 and UGTPg100 inserted at site LPP1.

[0188] Codon optimization of the PPTS coding gene was performed using the online tool GenSmart Optimization (GenScript Biotechnology Co., Ltd.). The nucleotide sequence was synthesized by Sangon Biotech (Shanghai) Co., Ltd. as shown below or as SEQ ID NO.2. The low-copy plasmid pY26-P was used. TEFUsing the plasmid (obtained from the BioVector NTCC plasmid vector strain cell gene deposit center) as a template, the plasmid framework was amplified using primers pY26-1F and pY26-1R. Using the PPTS coding gene as a template, the PPTS gene fragment was amplified using primers PPTS-F and PPTS-R. Using the Saccharomyces cerevisiae CEN PK2-1D genome as a template, the promoter Gal1 gene fragment was amplified using primers Gal1-F and Gal1-R. After recovering the PCR products, the three gene fragments were assembled using Gibson assembly technology at a 1:1:1 ratio to construct the pY26-PPTS plasmid. Figure 8 ).

[0189]

[0190] To compare the catalytic performance of different protopanaxadiol synthase mutants, the single-point mutants obtained from Examples 1 and 2 were cloned into the same low-copy free expression plasmid. The specific procedures were as follows: using primer sequences synthesized by the company, the plasmid pY26-PPTS was amplified with primers V102S-F / V102S-R to obtain pY26-PPTS. V102S After verification by nucleic acid gel electrophoresis, the circular p fragment was recovered and transformed into *E. coli* DH5α and cultured for 12 h. One single colony was randomly selected and cultured for 12 h, from which the plasmid was extracted and sent for sequencing verification. The remaining mutants were the same as those for pY26-PPTS. V102S Construction Method. Wild-type and candidate mutant plasmids were transformed into recombinant strain *S. cerevisiae* P1, respectively. The transformation system consisted of 1 μg plasmid, 100 μL transformation buffer, and 3 μL salmon sperm DNA at a concentration of 10 mg / mL, thoroughly mixed. The mixture was incubated at 37°C for 30 min, followed by centrifugation at 8000 g for 3 min, and the supernatant was discarded. The resulting cells were resuspended in sterile deionized water to obtain the transformation culture, which was then plated onto leucine-deficient agar plates and incubated at 30°C for 3 days. Colonies were then selected and cultured in YPD medium supplemented with 20 g / L glucose for 144 h at 220 rpm and 30°C. The yield of rare ginsenoside Rh1 and the accumulation of protopanaxadiol were then measured (referring to Example 3 of Chinese Invention Patent CN120574697B for detection).

[0191] The experimental results of the 20 active site candidate mutants obtained in Example 1 are as follows: Figure 9 As shown, the relative Rh1 yields of the T114K, I117D, and I117Y mutants were 1.526 times, 1.390 times, and 1.211 times that of the wild type (47.12 mg / L), respectively, representing increases of 52.65%, 39.04%, and 21.06%. Among the 20 candidate mutants with active sites, three mutants showed higher Rh1 yields than the wild type, resulting in an experimental hit rate of 15%.

[0192] The experimental results of the distant candidate mutants obtained in Example 2 are as follows: Figure 10 As shown, excluding the control mutant I117V, a total of 20 distal candidate mutants were validated, of which 18 mutants had higher Rh1 yields than the wild type, with an experimental hit rate of 94.74%. Among them, the relative Rh1 yields of the V102S, V102N, and V102Q mutants were 2.745 times (129.34 mg / L), 2.074 times, and 1.969 times that of the wild type, respectively, representing increases of 174.53%, 107.42%, and 96.87%.

[0193] Example 4: Construction of combinatorial mutants and verification of the synthetic performance of rare ginsenoside Rh1

[0194] To further verify the potential of combined applications between dominant mutation sites, this embodiment constructs a PPTS dual-site combined mutant based on a single-site mutant. According to the effective mutation sites screened in Examples 1 and 2, the active pocket mutation sites T114K, I117D, and I117Y, and the distal regulatory mutation sites V102S, A115K, V102N, and V102Q were selected as the basis for combined mutations. Furthermore, residues such as I109, I112, L113, and T116, located in the vicinity of the aforementioned dominant sites and potentially involved in local conformational regulation and substrate-binding pocket microenvironment remodeling, were selected to construct dual-site combined mutants to verify the combined application effect of active pocket mutations and distal conformational regulatory mutations. The following models were constructed: V102S / I109V, V102S / I112Y, V102S / L113A, V102S / T114K, V102S / A115K, V102S / T116K, I109V / I112Y, I109V / L113A, I109V / T114K, I109V / A115K, and I109V / T11. Two-site combined mutants, including 6K, I112Y / L113A, I112Y / T114K, I112Y / A115K, I112Y / T116K, L113A / T114K, L113A / A115K, L113A / T116K, T114K / A115K, T114K / T116K, and A115K / T116K, were obtained using site-directed mutagenesis. These combined mutant PPTS encoding genes were then cloned into the same low-copy-count free expression plasmids as in Example 3, and fermented under the transformation and fermentation conditions described in Example 3. After 144 h, the yield of rare ginsenoside Rh1 and the accumulation of protopanaxadiol were measured.

[0195] The results are as follows Figure 11The results showed that the I109V / T116K hybrid mutant had the highest relative Rh1 yield, reaching 3.047 times (143.57 mg / L) that of the wild type. The relative Rh1 yields of the V102S / I109V, V102S / A115K, T114K / A115K, and A115K / T116K hybrid mutants were 2.187 times, 2.188 times, 2.069 times, and 2.062 times that of the wild type, respectively. The relative Rh1 yields of the V102S / T116K and I109V / A115K hybrid mutants were 1.976 times and 1.876 times that of the wild type, respectively, which were also significantly higher than those of the wild type. In addition, the relative Rh1 yields of the V102S / L113A, L113A / A115K, I112Y / A115K, I112Y / T116K, and V102S / I112Y combined mutants were 1.652 times, 1.309 times, 1.294 times, 1.179 times, and 1.155 times that of the wild type, respectively, showing varying degrees of improvement.

[0196] Among them, the T114K / A115K combined mutant contains both active pocket mutation and distal regulatory mutation, and its relative Rh1 yield is 2.069 times that of the wild type, indicating that active pocket mutation and distal regulatory mutation can be used together to improve the catalytic performance of PPTS; the relative Rh1 yields of the V102S / A115K and A115K / T116K combined mutants are 2.188 times and 2.062 times that of the wild type, respectively, indicating that there is also potential for combined modification of distal regulatory sites.

[0197] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A rational design screening method for improving enzyme catalytic performance, characterized in that: This includes MM / PBSA-guided active site mutation design and conformational bias-guided distal site mutation design; The MM / PBSA-guided active site mutation design includes the following steps: (1) Obtain the three-dimensional structure of protopanaxadiol synthase, construct an enzyme-substrate complex model formed by protopanaxadiol synthase and protopanaxadiol, and perform molecular dynamics simulation on the enzyme-substrate complex. (2) Select the molecular dynamics simulation trajectory after equilibrium is reached, and decompose the MM / PBSA binding free energy of amino acid residues near the original ginsenoside diol binding pocket. Based on the energy contribution of each amino acid residue to substrate binding and combined with sequence conservation analysis, screen the active sites that affect substrate binding. (3) Virtual saturation mutations were performed on the active sites obtained by screening to construct virtual mutant models of ginsenoside triol synthase, and molecular docking was performed between each virtual mutant and ginsenoside diol. (4) Based on the spatial distance between the hydroxylation site of the original ginsenoside diol and the active center of heme, the substrate binding conformation obtained by molecular docking was screened for pre-reaction state, and the active site candidate mutants were obtained by combining the evaluation of substrate binding ability and protein folding stability. (5) Construct and express the candidate mutants of the active site, determine their ability to catalyze the production of protopanaxadiol to protopanatriol and further synthesis of rare ginsenoside Rh1 in engineered brewer's yeast, and screen for active site mutants with improved catalytic performance; The conformational bias-guided distal site mutation design includes the following steps: (6) Construct a Compound I state complex model composed of protopanaxadiol synthase, protopanaxadiol and heme active intermediate, and perform molecular dynamics simulation on the complex model; (7) Principal component analysis and cluster analysis were performed on the obtained molecular dynamics simulation trajectory to construct the conformational landscape of the protopanaxadiol synthase-protopanaxadiol complex, and the complex was divided into multiple conformational metastable states. (8) Evaluate the catalytic potential of metastable states of different conformations based on the geometric criterion of near-attack conformation, and determine the catalytically favorable and catalytically unfavorable conformations; (9) Compare the active pocket volume, nonpolarity ratio, volume fluctuation characteristics and residue contact network differences of metastable states in different conformations, and screen for distant candidate sites that may affect the distribution of metastable states in conformations. (10) Using representative structures of catalytically favorable and catalytically unfavorable conformations as input, the conformational bias method and at least two protein sequence design models are used to predict the preference of single-point mutations for different conformational states and obtain distant candidate mutants. (11) Construct and express the distal candidate mutant, determine its ability to synthesize rare ginsenoside Rh1 in engineered Saccharomyces cerevisiae, and screen for distal site mutants with improved catalytic performance.

2. The rational design screening method for improving enzyme catalytic performance according to claim 1, characterized in that: The amino acid sequence of the protopanaxadiol synthase PPTS is shown in SEQ ID NO.

1.

3. The rational design screening method for improving enzyme catalytic performance according to claim 1, characterized in that: The molecular dynamics simulation time described in step (1) is 200–500 ns; The molecular dynamics simulation trajectory after reaching equilibrium mentioned in step (2) is the trajectory of the last 50 to 150 ns of the simulation trajectory; The amino acid residues near the protopanaxadiol binding pocket mentioned in step (2) are amino acid residues that are no more than 5 Å away from any atom of protopanaxadiol; The virtual saturation mutation described in step (3) is performed using FoldX; The molecular docking described in step (3) was performed using AutoDock Vina; Each of the virtual mutants described in step (3) generates 20 to 60 protopanaxadiol-binding conformations; The pre-reaction state screening mentioned in step (4) refers to selecting a substrate binding conformation with a distance of 3.5 to 5.0 Å between the carbon atom to be hydroxylated at the C6 position of protopanaxadiol and the iron atom of heme. The active site candidate mutant mentioned in step (4) is a virtual mutant that retains a substrate binding free energy change of less than 0 kcal / mol and a folding free energy change of no more than 1.0 kcal / mol; The molecular dynamics simulation time for the Compound I state complex model described in step (6) is 300–700 ns; The near-attack conformation described in step (8) simultaneously satisfies the following geometric conditions: the Fe-OH angle of the heme active center is 100° to 140°, the CHO angle of the CH bond to be hydroxylated in protopanaxadiol is 140° to 180°, and the distance between the active oxygen atom of heme and the hydrogen atom to be hydroxylated in protopanaxadiol is no greater than 3.0 Å; In step (8), the metastable conformations with a higher proportion of near-aggressive conformations are identified as catalytically favorable conformations, and the metastable conformations with a lower proportion are identified as catalytically unfavorable conformations. The residue contact mentioned in step (9) refers to the distance between the side chain heavy atoms of non-adjacent amino acid residues not exceeding 4.5 Å; The occupancy rate of residue contacts mentioned in step (9) is the proportion of the number of trajectory frames in which residue contacts occur to the total number of trajectory frames in the corresponding metastable conformation. The distal candidate sites described in step (9) exclude highly conserved residues and residues that form a protein-protein interaction interface with cytochrome P450 reductase. The protein sequence design model described in step (10) is selected from at least two of ProteinMPNN, ESM-IF1, Frame2Seq and ThermoMPNN; The conformational bias method described in step (10) is to evaluate the mutant’s preference for different conformational states by comparing the fit scores of the same mutant sequence with catalytically favorable and catalytically unfavorable conformations. The distal candidate mutants described in step (10) satisfy the following: the standardized fit score relative to wild-type protopanaxadiol synthase is greater than 0 in the catalytically favorable conformation, and the difference between the standardized fit score in the catalytically favorable conformation and the standardized fit score in the catalytically unfavorable conformation is greater than 0.

4. A protopanaxadiol synthase mutant, characterized in that: The mutant obtained by the rational design screening method according to any one of claims 1 to 3, wherein the mutant uses wild-type protopanaxadiol synthase as the parent enzyme, and is any one of the following mutants: The threonine at position 114 is mutated to lysine, i.e., T114K; The isoleucine at position 117 is mutated to aspartic acid or tyrosine, namely I117D or I117Y. The valine at position 102 is mutated to arginine, lysine, alanine, glycine, serine, threonine, glutamine, asparagine, methionine, histidine, cysteine, aspartic acid, or glutamic acid. The alanine at position 125 is mutated to threonine, methionine, lysine, histidine, arginine, or glutamine. The isoleucine at position 109 is mutated to valine and the threonine at position 116 is mutated to lysine, i.e., I109V / T116K. The valine at position 102 is mutated to serine and the isoleucine at position 109 is mutated to valine, i.e., V102S / I109V; The valine at position 102 is mutated to serine and the alanine at position 115 is mutated to lysine, i.e., V102S / A115K. The 114th threonine is mutated to lysine and the 115th alanine is mutated to lysine, i.e., T114K / A115K; The alanine at position 115 is mutated to lysine and the threonine at position 116 is mutated to lysine, namely A115K / T116K; The valine at position 102 is mutated to serine and the threonine at position 116 is mutated to lysine, i.e., V102S / T116K. The isoleucine at position 109 is mutated to valine and the alanine at position 115 is mutated to lysine, i.e., I109V / A115K.

5. A nucleic acid, characterized in that: The nucleic acid encodes the protopanaxadiol synthase mutant of claim 4.

6. A recombinant expression vector for expressing protopanaxadiol synthase, characterized in that: It contains the nucleic acid as described in claim 5.

7. A recombinant host cell for expressing protopanaxadiol synthase, characterized in that: It contains the nucleic acid as described in claim 5 or the recombinant expression vector as described in claim 6.

8. The recombinant host cell according to claim 7, characterized in that: The recombinant host cells also express dammarene diol synthase PgDDS, protopanaxadiol synthase CYP716A47, cytochrome P450 reductase PgCPR1, cytochrome P450 reductase PgCPR2, and glycosyltransferase UGTPg100.

9. A method for preparing rare ginsenoside Rh1 using brewer's yeast, characterized in that... Includes the following steps: (1) The nucleic acid encoding the protopanaxadiol synthase mutant of claim 4 is introduced into Saccharomyces cerevisiae with a ginsenoside synthesis pathway to obtain recombinant Saccharomyces cerevisiae; (2) Cultivate the recombinant brewer's yeast to express the protopanaxadiol synthase mutant and synthesize rare ginsenoside Rh1; and detect and / or isolate rare ginsenoside Rh1 from the culture.

Images