Vega Aitor, Planas Antoni, Biarnés Xevi
Laboratory of Biochemistry, Institut Químic de Sarrià, University Ramon Llull, Barcelona, Spain.
Chemistry Section, Royal Academy of Sciences and Arts of Barcelona, Spain.
FEBS J. 2025 Aug;292(16):4211-4231. doi: 10.1111/febs.70121. Epub 2025 May 5.
The high catalytic efficiency of enzymes is attained, in part, by their capacity to stabilize electrostatically the transition state of the chemical reaction. High-throughput protocols for measuring this electrostatic contribution in computer-assisted enzyme design are limited. We present here an easy-to-compute metric that captures the electrostatic complementarity of the enzyme to the charge distribution of the substrate at the transition state. We demonstrate such a complementarity for a representative dataset of glycoside hydrolases, a large family of enzymes responsible for the hydrolytic cleavage of glycosidic bonds in oligosaccharides, polysaccharides, and glycoconjugates. We have implemented this metric in BindScan, a computer-based mutational analysis protocol to assist protein engineering. We demonstrate the predictive power of BindScan with this metric for two mechanistically distinct glycoside hydrolases: Spodoptera frugiperda β-glucosidase (Sfβgly, operates via protein nucleophile catalysis) and Bifidobacterium bifidum lacto-N-biosidase (BbLnbB, operates via substrate-assisted catalysis). The metric correctly predicts sequence positions sensible to the modulation of k/K upon mutation from an experimental benchmark of 51 mutants of Sfβgly with 77% classification efficiency and identifies variants of BbLnbB with improved transglycosylation yields (up to 32%). Based on electrostatic potential and ligand affinity calculations, as implemented in BindScan, we propose a rational strategy to design glycoside hydrolase variants with improved transglycosylation efficiency for the synthesis of added-value glycoconjugates. The new reactivity metric may contribute to expanding the range of computational protocols available to assist enzyme engineering campaigns aimed at optimizing mechanistically relevant properties.
酶的高催化效率部分是通过其在静电作用下稳定化学反应过渡态的能力来实现的。在计算机辅助酶设计中，用于测量这种静电作用的高通量方案有限。我们在此提出一种易于计算的指标，该指标可捕捉酶与过渡态底物电荷分布之间的静电互补性。我们针对糖苷水解酶的代表性数据集展示了这种互补性，糖苷水解酶是一个大家族的酶，负责寡糖、多糖和糖缀合物中糖苷键的水解裂解。我们已将此指标应用于BindScan中，BindScan是一种基于计算机的突变分析方案，用于辅助蛋白质工程。我们用该指标证明了BindScan对两种机制不同的糖苷水解酶的预测能力：草地贪夜蛾β-葡萄糖苷酶（Sfβgly，通过蛋白质亲核催化作用）和两歧双歧杆菌乳糖-N-生物酶（BbLnbB，通过底物辅助催化作用）。该指标从51个Sfβgly突变体的实验基准中正确预测了对突变时k/K调节敏感的序列位置，分类效率达77%，并鉴定出转糖基化产率提高（高达32%）的BbLnbB变体。基于BindScan中实现的静电势和配体亲和力计算，我们提出了一种合理策略，以设计具有更高转糖基化效率的糖苷水解酶变体，用于合成附加值糖缀合物。这种新的反应性指标可能有助于扩大可用于辅助酶工程活动的计算方案范围，这些活动旨在优化与机制相关的特性。
