Paeonia lactiflora subsp. pinene

2.1. Homology modeling of P. lactiflora pinene synthase The crystal structures of pinene synthases have not been reported. Therefore, we constructed a homolog model of P. lactiflora pinene synthase using the crystal structure of S-limonene synthase as the template. In order to investigate the catalytic mechanism, we docked three carbocation intermediates (terpinyl, pinyl and thujyl cations) into its active site (Fig. 1B–D). The docking results suggest that carbocations bind to a similar location in the active pocket. Nearby residues include R305, S311, W314, I335, I338, N339, D342, Y417, I443, A444, I448, F482, D486, S491 and H571. Some of these residues could make hydrophobic interactions with the carbon skeleton of carbocations. Certain polar or aromatic residues may stabilize the positive charge and influence the product profile of the enzyme. 2.2. Alanine scanning mutagenesis identifies mutations that cause an increase in sabinene production We first identified residues within 5 Å from the docked carbocations and performed alanine-scanning mutagenesis on them. Most of these mutations do not show dramatic changes in their product profiles. However, two mutations show interesting alteration in the product profile. One of them, F 482 A, essentially converts the enzyme to sabinene synthase. About 90.8% of its products is sabinene, while 97.7% of the products of the WT enzyme is α- pinene (Table 1, Fig. 2A–B). Another mutation, I 335 A, also causes an increase in sabinene production, though not as dramatic as F 482 A (Table 1). Importantly, although both mutants show an altered product profile, their overall activities are only moderately decreased (Table 1). 2.3. Phenylalanine at the 482 position is required for pinene production We hypothesized that the aromatic ring of F 482 might play a role in stabilizing the pinyl cation. To investigate this possibility, we mutated F 482 to Val and Ile, two nonpolar residues with different side chain sizes. We also mutated it to Thr, Arg and Trp, representing polar, charged or non-Phe aromatic residues. The results suggest that any mutations that changes the phenylalanine at this position leads to an increase in sabinene production and almost abolishes the α- pinene production (Table 1). Replacing phenylalanine with nonpolar residues (F 482 I, F 482V) all effectively convert the enzyme to sabinene synthase (Fig. 2C–D). More than 90% of products made by these mutants are sabinene (Table 1). Interestingly, F 482T also predominantly produces sabinene, which seems to suggest that a polar residue here is not sufficient to stabilize the pinyl cation (Table 1, Fig. 2E). Other mutations (F 482 Y, F 482 R and F 482 W) retain very little activities (Table 1). In order to verify that the mutant enzymes indeed produce sabinene, we analyzed the reaction product of F 482L using the high resolution hybrid quadrupole-orbitrap GC-MS/MS system (Fig. 3A–C). By comparing with the sabinene standard, the peak eluted at 9.29 min was identified as sabinene. Its mass spectrum was measured and could be readily matched with that of sabinene standard (Fig. 3B–C). It is interesting to note that replacement of F 482 to other aromatic residues (Tyr or Trp) causes a drastic decrease in enzyme activities. To investigate this issue, we first docked the pinyl cation into the active pockets of WT and F 482 Y. It could be seen that in these two structures, the conformation of the docked pinyl cation is quite different (Fig. 4A). We then performed a 2 ns molecular dynamics simulation. The results suggest more conformational changes (Fig. 4B–C). Therefore, the extra hydroxyl group in Tyr could significantly change the conformation of the carbocation, which may explain the decreased activity and altered product profile. 2.4. S 491 A mutation enhances the overall activity of pinene synthase without changing its product profile During our Ala scanning experiments, we identified a mutation S 491 A that enhances overall activities (Fig. 5A, D, Table 1). The overall activity of S 491 A mutant is about 29% higher than that of the WT enzyme (Table 1). Furthermore, S 491 A does not change the product selectivity, as more than 90% of their products remains α- pinene (Fig. 5A, Table 1). What is the mechanism of S 491 A to enhance the enzymatic activity? Since S 491 A still predominantly produces pinene, we docked the pinyl cation into the active pocket of the mutant enzyme and performed a 2 ns of molecular dynamics simulation. We also performed the same dynamic simulation on the WT enzyme in complex with the pinyl cation. Our data suggests that the active site residues in S 491 A have a much smaller RMSD than those in the WT enzyme, indicating the mutation S 491 A could cause the active pocket to be more rigid (Fig. 6A, STable 2). Interestingly, simulation on the WT or S 491 A enzymes in complex with the terpinyl or thujyl cation does not show the same differences in RMSD values of active site residues (Fig. 6B–C, STable 2). This probably indicates that the S 491 A mutation increases the rigidity of the active pocket only when it binds to the pinyl cation. To better quantify the mobility of protein structures, we reanalyzed the simulation data using MDLovoFit (Martínez, 2015). This way we could subdivide the protein into regions with high and low mobility (Fig. 7A–C, STable 3). Interestingly, in S 491 A, fewer residues are classified as of high mobility than residues in WT enzymes (59% vs 63%). Furthermore, for residues classified with the highly mobile subset, those in S 491 A have lower average RMSD than those in WT proteins (0.253 nm vs. 0.278 nm). Therefore, there is a slight but significant increase in rigidity in S 491 A mutant protein. 2.5. Combination of S 491 A with F 482 mutants has mixed results in the overall activity Since all four sabinene synthase mutants show some decreases in the catalytic activity, we wondered whether we may restore their activities to that of the WT pinene synthase by combing with S 491 A. We produced double mutants F 482 A / S 491 A and F 482 I / S 491 A. Both seem to retain the same product selectivity of the original single mutants, though F 482 A / S 491 A shows an insignificant decrease in sabinene production (P =0.39; Fig. 5B–C, Table 1). In overall activities, F 482 I / S 491 A shows an insignificant increase in activity (P =0.27), while F 482 A / S 491 A has a significant decrease in activity (P < 0.05; Fig. 5D, Table 1). Therefore, it seems that the effect of combining two mutants is more complicated than we expected. We asked whether the combination of F 482 A and S 491 A mutations could affect the rigidity of the protein. After a 2 ns of molecular dynamics simulation and data analysis using MDLovoFit, we found the rigidity of F 482 A / S 491 A is similar to that of S 491 A (STable 3). Apparently, the decrease of activity by the combination of F 482 A with S 491 A cannot be explained by the overall RMSD change. To further investigate this issue, we analyze the RMSF values based on the simulation data (SFig. 1 A –C). Compared with S 491 A and WT, F 482 A / S 491 A shows an RMSF spike at a region between residues 219 and 223. In the structural model, this region forms a loop in the protein structural model (SFig. 2). However, this loop is quite far from the active site (about 40 Å). Therefore, its mobility could only affect the enzyme activity indirectly. For example, it could influence the closure of the active pocket. We also superimposed the structural model of WT, S 491 A and F 482 A / S 491 A in complex with the docked pinyl cation (SFig. 3). In the structural model, the pinyl cation is docked between the residue 491 and 482 (SFig. 3). Interestingly, in WT and S 491 A, the position of the pinyl cation is similar, while in F 482 A / S 491 A, the pinyl cation is shifted towards the A 482. This change in the carbocation position may impair the enzyme activity (SFig. 3).Published as part of Xu, Jingwei, Peng, Guanzu, Xu, Jinkun, Li, Yi, Tong, Li & Yang, Dong, 2021, Probing of the plasticity of the active site in pinene synthase elucidates its potential evolutionary mechanism, pp. 1-10 in Phytochemistry (112573) (112573) 181 on pages 2-4, DOI: 10.1016/j.phytochem.2020.112573, http://zenodo.org/record/829106