Figure 4. Representative amino acids transformations of variants W1-W3 and binding pose analysis in the docking model.(A) Representative amino acids transformation results of variants W1-W3. (B-G) Posture analysis of variant W1 or W2 after docking with representative substrates (L-Leu, L-Met, l-val, L-Phe, L-Arg or L-Glu).
We also evaluated the properties of variant W2 for different amino acids. As shown in Supplementary Table 3, variant W2 significantly increased catalytic efficiency toward L-Phe and L-Met, but not toward the other substrates. D1Phe decreased the most (15.4%; from 3.0 to 2.6 Å), which increased its kcat ,Km ,kcat/Km , and specific activity toward L-Phe by 107.1%, 23.5%, 64.7%, and 65.5%, respectively. D1Met decreased by 14.7% (from 2.8 to 2.4 Å), which increased kcat , Km ,kcat/Km , and specific activity by 57.7%, 26.8%, 23.8%, and 38.2%, respectively.
The mechanisms underlying improved catalytic efficiency of variant W2 were analyzed. Compared to Pmi LAAD, molecular dynamic simulations revealed an increase in root-mean-square fluctuations (RMSFs) for six residues (T105, D144, E145, E340, S412, and E417) around the substrate channel (Figure 5). The increased conformational dynamic of these mutated residues resulted in greater structural flexibility of the channel(W. Song et al. 2020; Yang et al. 2017), which changed the orientation of the substrate side-chain (phenyl or methylthio group) (Figure 4C and 4E). As a consequence, the binding posture of L-Phe and L-Met was significantly altered, decreasing θ of W2Phe and W2Met from 136.3° to 94.8° and from 128.2° to 96.8°, respectively. The reduced θ shortened D1Phe and D1Met by 0.4 Å, resulting in higher catalytic efficiency toward L-Phe and L-Met. Meanwhile, θ and D1 in W2 remained almost unchanged toward L-Leu, L-Val, L-Arg, and L-Glu (Figure 4B–4G), and so was their catalytic efficiency.