Combining temperature perturbations with X-ray crystallography to study dynamic macromolecules: A thorough discussion of experimental methods

Design of Efficient Artificial Enzymes Using Crystallographically Enhanced Conformational Sampling

“…To create efficient de novo enzymes without relying on iterative rounds of mutagenesis and high-throughput screening, we implemented an ensemble-based computational enzyme design pipeline that uses protein backbones generated by ensemble refinement of diffraction data as templates for design (Figure ). Because the diffraction data and calculated electron density maps are derived from crystals that contain billions of molecules, which are all sampling different conformational states throughout data collection, the resulting ensemble models represent a distribution of conformations that are well populated in the ensemble at equilibrium . Thus, we postulated that performing crystallographic ensemble refinement on a low activity de novo enzyme would allow sampling of conformational substates that can more accurately accommodate the TS and its interactions with catalytic residues than the single conformation modeled using traditional refinement methods, thereby facilitating the design of improved active sites.…”

“…data collection, the resulting ensemble models represent a distribution of conformations that are well populated in the ensemble at equilibrium. 21 Thus, we postulated that performing crystallographic ensemble refinement on a low activity de novo enzyme would allow sampling of conformational substates that can more accurately accommodate the TS and its interactions with catalytic residues than the single conformation modeled using traditional refinement methods, thereby facilitating the design of improved active sites. To test our hypothesis, we applied an ensemble-based design to increase catalytic efficiency of the de novo enzyme HG3 7 that catalyzes the Kemp elimination (Scheme 1), a well-established model organic transformation for benchmarking computational enzyme design methods.…”

“…To evaluate whether the catalytic efficiency of our designs was increased via the predicted molecular mechanisms (Figure b), we solved room-temperature (280 K) X-ray crystal structures of four of the most active designs (HG185, HG198, HG630, and HG649) in the presence and absence of the 6NT TS analogue. We determined structures at room temperature for direct comparison with the HG3 noncryogenic structure used to generate the ensemble and because this method can reveal conformational heterogeneity in protein structures that would not be visible at cryogenic temperatures, providing insights into the conformational ensemble that is sampled by a protein. , Crystals of HG185, HG630, and HG649 were obtained in the absence of 6NT and soaked into solutions containing this molecule prior to X-ray data collection to obtain the bound structures (Supplementary Table 5). These three enzyme variants all crystallize in space group P 2 1 with two molecules in the asymmetric unit, roughly matching the crystal form reported for HG3 .…”

Design of efficient artificial enzymes using crystallographically-enhanced conformational sampling

“…To create efficient de novo enzymes without relying on iterative rounds of mutagenesis and high-throughput screening, we implemented an ensemble-based computational enzyme design pipeline that uses protein backbones generated by ensemble refinement of diffraction data 18 as templates for design (Figure 1). Because the diffraction data and calculated electron density maps are derived from crystals that contain billions of molecules, which are all sampling different conformational states throughout data collection, the resulting ensemble models are expected to represent the entire equilibrium conformational ensemble of the target protein fold 21 . Thus, we postulated that performing crystallographic ensemble refinement on a low activity de novo enzyme would allow sampling of conformational substates that can more accurately accommodate the transition state and its interactions with catalytic residues than the single conformation modeled using traditional refinement methods, thereby facilitating the design of improved active sites.…”

“…To evaluate whether the catalytic efficiency of our designs was increased via the predicted molecular mechanisms (Figure 1b), we solved room-temperature (280 K) X-ray crystal structures of four of the most active designs (HG185, HG198, HG630 and HG649) in the presence and absence of the 6NT transition-state analogue. We used room-temperature X-ray crystallography for direct comparison with the HG3 non-cryogenic structure used to generate the ensemble and because this method can reveal conformational heterogeneity in protein structures that would not be visible at cryogenic temperatures, providing insights into the conformational ensemble that is sampled by a protein 21,24 . Crystals of HG185, HG630, and HG649 were obtained in the absence of 6NT, and soaked into solutions containing this molecule prior to X-ray data collection to obtain the analogue-bound structures (Supplementary Table 4).…”