Microdroplets in water-in-oil emulsions can be used as microreactors with volumes 10(3) to 10(9) times smaller than the smallest working volumes in a microtitre plate well (1-2 microL). However, many reactions and assays require multiple steps where new reagents are added at defined times, to start, modify or terminate a reaction. The most flexible way to add new reagents to pre-formed droplets is by controlled, pairwise droplet fusion. We describe a droplet-based microfluidic system capable of performing multiple operations, including pairwise droplet fusion, to analyze complex and sequential multi-step reactions. It is exemplified by performing a series of six on-chip and two off-chip operations which enable the coupled in vitro transcription and translation of cotA laccase genes in droplets and, after performing a controlled fusion with droplets containing laccase assay reagents, the end-point and kinetic analysis of the catalytic activity of the translated protein. In vitro translation and the laccase assay must be performed sequentially as the conditions for the laccase assay are not compatible with in vitro translation. Droplet fusion was performed by electrocoalescence at a rate of approximately 3000 fusion events per second and nearly 90% of droplets were fused one-to-one (one droplet containing in vitro translated laccase fused to one droplet containing the reagents for the laccase assay). The ability to uncouple the enzymatic assay from in vitro translation greatly extends the range of activities of in vitro translated proteins that can potentially be screened in droplet-based microfluidic systems. Furthermore, the system also opens up the possibility of performing a wide range of other new (bio)chemical reactions in droplets.

We describe a completely in vitro high-throughput screening system for directed evolution of enzymes based on in vitro compartmentalization (IVC). Single genes are transcribed and translated inside the aqueous droplets of a water-in-oil emulsion. Enzyme activity generates a fluorescent product and, after conversion into a water-in-oil-in-water double emulsion, fluorescent droplets are sorted using a fluorescence-activated cell sorter (FACS). Earlier in vivo studies have demonstrated that Ebg, a protein of unknown function, can evolve to allow Escherichia coli lacking the lacZ beta-galactosidase gene to grow on lactose. Here we demonstrate that we can evolve Ebg into an enzyme with significant beta-galactosidase activity in vitro. Only two specific mutations were ever seen to provide this improvement in Ebg beta-galactosidase activity in vivo. In contrast, nearly all the improved beta-galactosidases selected in vitro resulted from different mutations.