Highly L and D enantioselective variants of horseradish peroxidase discovered by an ultrahigh-throughput selection method

Abstract

A highly efficient selection method for enhanced enzyme enantioselectivity based on yeast surface display and fluorescence-activated cell sorting (FACS) is developed and validated. Its application to horseradish peroxidase has resulted in enzyme variants up to 2 orders of magnitude selective toward either substrate enantiomer at will. These marked improvements in enantioselectivity are demonstrated for the surface-bound and soluble enzymes and rationalized by computational docking studies.

Fig. 1.

Chemical structures of the reducing substrates used in the present study to assess HRP's enantioselectivity (asterisks designate stereogenic centers); the tyrosinol (chiral) portion is shown in green.

Fig. 2.

Schematic representation of the ultrahigh-throughput selection method for yeast-surface-bound HRP variants with enhanced l or d enantioselectivity. HRP, expressed as a fusion protein to the c-Myc tag and Aga2p mating agglutinin protein, is displayed on the yeast surface via disulfide bridges between the Aga2p and Aga1p proteins. Enzymatic oxidation of the l and d enantiomers of tyrosinol (shown in green) conjugated to fluorescent dyes (Dye 1 and Dye 2) yields phenoxyl radicals that then nonenzymatically react with Tyr residues of membrane-bound proteins; this reaction leads to labeled cells with fluorescence intensity that depends on the enantioselectivity of HRP. The enzymatic activity is normalized via fluorescently labeled antibodies against the c-Myc tag (magenta star). Multiparameter FACS is used to isolate cells with the highest ratio of fluorescence intensities (Dye 1/Dye 2 or Dye 2/Dye 1) encoding l or d selective HRP variants.

Fig. 3.

Multiparameter FACS analysis of surface-bound, wild-type HRP incubated with l-1 + l-2, d-tyrosinol-A647, and H₂O₂. The regions outlined by trapezoids schematically represent library sort gates, used to isolate l selective (cells with high A488 and low A647 fluorescence) and d selective (cells with low A488 and high A647 fluorescence) HRP variants.

Fig. 4.

Enantioselectivities of l selective variants toward 1 and 3 (A) and 2 and 4 (B) as well as d selective variants toward 1 and 3 (C) and 2 and 4 (D), discovered in each round of evolution. Red, yellow, blue, and green bar colors represent substrates 1, 2, 3, and 4, respectively; hatched and solid bars designate surface-bound and soluble HRP, respectively. The l and d letters designate the direction of enantiopreference; the Roman numerals after the letters define the round of directed evolution; the r and/or s letters after the Roman numerals indicate whether these variants were isolated from the random or saturated mutagenesis libraries, respectively. Mutations: LIr (Arg178Gln), LIs (Phe68Leu, Gly69Ala, Asn72Glu, Ser73Leu, Ala74Tyr) (14), LIrs (LIr + LIs), LIIr (LIrs + Gln147Arg), LIII (LIIr + Asn158Asp), DIs (Phe68Glu, Gly69Pro, Asn72Lys) (14), DIIs (DIs + Asn137Arg, Ala140His, Phe142Lys, Phe143Met), and DIII (DIIs + Ser167Ile).

Fig. 5.

Modeled complexes of wild-type HRP with d-3 (A), l-3 (B), d-4 (C), and l-4 (D). For clarity, only the active site of the enzyme is shown with the heme moiety in orange, substrate in blue, and some mutated residues in green. Distances indicated are in angstroms. The Arg-178 residue is shown in a double rotamer configuration as it appears in the crystal structure (29); only 1 rotamer configuration was used in docking experiments. See Methods for details of how these models were built.