Efficient molecular evolution to generate enantioselective enzymes using a dual-channel microfluidic droplet screening platform

Abstract

Directed evolution has long been a key strategy to generate enzymes with desired properties like high selectivity, but experimental barriers and analytical costs of screening enormous mutant libraries have limited such efforts. Here, we describe an ultrahigh-throughput dual-channel microfluidic droplet screening system that can be used to screen up to ~10⁷ enzyme variants per day. As an example case, we use the system to engineer the enantioselectivity of an esterase to preferentially produce desired enantiomers of profens, an important class of anti-inflammatory drugs. Using two types of screening working modes over the course of five rounds of directed evolution, we identify (from among 5 million mutants) a variant with 700-fold improved enantioselectivity for the desired (S)-profens. We thus demonstrate that this screening platform can be used to rapidly generate enzymes with desired enzymatic properties like enantiospecificity, chemospecificity, and regiospecificity.

Fig. 1

DMDS platform for screening enzymatic enantioselectivity. a Schematic of DMDS operation. Mutant enzyme-expressing single cells are encapsulated in water-in-oil droplets with two fluorogenic substrates and lysis buffer. After the droplets are incubated for a specified time, those droplets containing the desired mutants are enriched via fluorescence-activated droplet sorting. Optical images of DMDS processes: b droplet generation; c off-chip incubation; d droplet reinjection; e fluorescence-activated droplet sorting. f To avoid crosstalk of two fluorescence signals, the droplets are excited by two spatially separated lasers, which generates two temporally separated emissions. g Sorting different populations in a mutation library with the DMDS platform is achieved via two screening modes: a cooperative mode and biased modes. h Three fluorogenic substrate designs and their enzymatic reactions yielding two different fluorescence signals. Scale bars: 100 μm

Fig. 2

Directed evolution of the enantioselectivity of AFEST. a Conceptual progression of enzymatic enantioselectivity enhancement by iterative rounds of mutagenesis and use of the DMDS process. b Cumulative improvement in the enantioselectivity of AFEST resulting from the various directed evolution steps of the present study

Fig. 3

Chemical structures of rac-1–5. Chemical names: rac-1, ibuprofen-pNP ester; rac-2, ketoprofen-pNP ester; rac-3, naproxen-pNP ester; rac-4, 2-phenylpropanoate-pNP ester; rac-5, 2-(para-methylphenyl)propanoate-pNP ester. Chemical names: rac-1, ibuprofen-pNP ester; rac-2, ketoprofen-pNP ester; rac-3, naproxen-pNP ester; rac-4, 2-phenylpropanoate-pNP ester; rac-5, 2-(para-methylphenyl)propanoate-pNP ester

Fig. 4

Structural basis for the altered enantioselectivity. AFEST, residues S160, D255, and H285 form a catalytic triad, while residues G88, G89, and A166 form an oxyanion hole by forming three hydrogen bonds with the carbonyl oxygen group of the (S)/(R)-substrate. a Docking poses of (R)/(S)-ibuprofen-pNP ester with wild-type AFEST. (R)-substrate was in a folded conformation, and the ibuprofen group was located at the substrate pocket entrance of AFEST. In contrast, (S)-substrate was in a linear conformation and the ibuprofen group was located deep inside the substrate pocket. b The introduction of the G89C mutation into IE9 resulted in significant steric hindrance against the chiral carbon-attaching methyl group of (R)-substrate, shifting the (R)-substrate outward. c The introduction of the G89P mutation into IE9 resulted in significant steric hindrance against the ibuprofen phenyl ring of (S)-substrate, forcing it to shift. The distances of the catalytic serine residue to the carbonyl group of the substrate were similar: WT-(S) 2.70 Å, WT-(R) 2.49 Å, IE9-(R) 2.89 Å, 6A8-(R) 2.62 Å, IE9-(S) 2.72 Å, and 4E12-(S) 2.87 Å