Hidden Specificities in Enzyme Catalysis: Structural Basis of Substrate Structure‐Selectivity Relationship of a Ketoreductase

Abstract

Enzymes often convert both physiological and non‐physiological substrates with high stereoselectivity; yet, for some enzymes, opposite product chirality is observed. A possible explanation is the existence of hidden specificities becoming apparent when non‐physiological substrates confer different substrate–enzyme interactions than the physiological substrate. To test this hypothesis, a series of α‐methylated β‐keto esters were converted with Tyl‐KR1, a ketoreductase from polyketide synthesis in Streptomyces fradiae. The conversions of six substrates with different physicochemical properties exhibited enantioselectivities ranging from 84 % ee for R,R to 84 % ee for S,S, yet high and uniform diastereoselectivity (anti, d.r.>9:1). The exchange of a single atom, namely an oxygen ester instead of a thioester, led to almost complete loss of enantioselectivity (<5 % ee). An additional S,S‐selective binding mode as a hidden specificity in Tyl‐KR1 has been identified through molecular modeling and site‐directed mutagenesis.

SeMPI 2.0—A Web Server for PKS and NRPS Predictions Combined with Metabolite Screening in Natural Product Databases

“…It can be hypothesized that various algorithm-independent parameters are responsible for the difficulty of substrate specificity classification. First, substrate variability hinders the predictability of one outcome class [38]. Second, substrate classes with few known instances are difficult to predict.…”

Site‐Directed Mutagenesis of Modular Polyketide Synthase Ketoreductase Domains for Altered Stereochemical Control

“…values and syn : anti d.r . values rather than the ratio of each of the four stereoisomers) [10,15,64] . With an ethanethiol‐linked thioester, a mixed stereochemical outcome between the native B1 isomer (2 R ,3 R ) and minor product of the enantiomeric (2 S ,3 S ) product formed [8,15,64] .…”

“…The activity of TylKR1 has been tested on a more extensive range of ester and thioester substrates than any other KR to date. Natural stereoselectivity was observed with 2‐methyl‐3‐oxopentanoyl‐SNAC, creating a large excess of (2 R ,3 R ) product along with 3 R product of 3‐oxopentanoyl‐SNAC as well as the (2 R ,3 R ) products of 2‐methyl‐3‐oxohexanoyl‐SNAC, 2‐ethyl‐3‐oxopentanoyl‐SNAC, and 2‐acetamidoethyl 2‐methyl‐3‐oxopentanoate [10,15,64] . However, truncating the substrate by one carbon to the corresponding 2‐methyl‐3‐oxobutanoyl‐SNAC entirely altered the stereoselectivity, a phenomenon observed with the corresponding oxo‐ester as well (note: we have not included substrates from references 15 and 64 in Table 4 due to the study only reporting e.e .…”

“…values rather than the ratio of each of the four stereoisomers) [10,15,64] . With an ethanethiol‐linked thioester, a mixed stereochemical outcome between the native B1 isomer (2 R ,3 R ) and minor product of the enantiomeric (2 S ,3 S ) product formed [8,15,64] . Such a switch in selectivity can only be explained by the substrates binding differently in the active site such that there is a switch between the re face and si face of the ketone reduced by TylKR1 [15] …”

“…Later, it was found that standalone KR domains are extremely stable and active, and thus, it was possible to measure stereoselectivity, stereospecificity, and kinetics [14] . Consequently, numerous studies have been performed on excised KR domains using model diffusible substrates such as β‐keto SNAC thioesters [15–18] …”