Only after the addition of the ePCR step to the protocol were we able to measure fluorescent signals from beads originally bound to single molecules of DNA. Since amplification efficiency is dependent on droplet size, we formed emulsions for ePCR in the microfluidic device. The amplification efficiency of ePCR drops markedly with increasing amplicon length, and very low yields have been reported for targets the size of the CpI expression template. Still, template dilution experiments indicated that even a modest increase in the number of DNA copies displayed on the bead could allow a significant increase in protein expression. Because amplification of long templates has been shown to be more BMS-907351 efficient in larger emulsion droplet volumes, we set the oil and aqueous phase flow rates to produce droplets about 30 microns in diameter. An additional benefit of the ePCR step is improved efficiency of DNA recovery by PCR from sorted beads following FACS. Some fraction of the bead-bound genes are expected to be partially or completely degraded by nucleases during the eCFPS incubation. Following ePCR, it should be much more likely that at least one copy of each gene will remain competent for amplification. The microbead display format adopted here is a very useful option for FACS analysis of IVC-generated fluorescent signals. However, caged-biotinylated fluorogenic substrates or other convenient methods to control fluorophore/bead binding are often unavailable. It thus seems likely that an ePCR step will prove useful in many future applications of IVC. To demonstrate the ability of the screen to identify beads bound to single molecules of hydrogenase DNA, we performed a test enrichment experiment. We incubated two sets of beads with single-molecule per bead levels of biotinylated DNA templates encoding either tagged CpI or a negative control protein, chloramphenicol acetyltransferase. Both templates contained the same homoprimer annealing sites for PCR amplification. We mixed the CAT and CpI beads at a 20:1 ratio, performed the three emulsification steps of the IVC screen protocol, and sorted the recovered beads by FACS. Low- and high-fluorescence beads were sorted and collected in separate tubes. Amplification of the DNA on the sorted low-fluorescence beads gave rise to a bright CAT gene band and a very light CpI gene band, while the high-fluorescence beads yielded a bright CpI gene band and a light CAT gene band. This enrichment of CpI DNA indicates that the IVC screen was able to identify and sort beads initially bound to single molecules of DNA encoding an active hydrogenase. No enrichment was observed when the ePCR step was omitted from the protocol. The success of the enrichment experiment indicates that hydrogenase mutants can be screened in an extremely highthroughput fashion by in vitro compartmentalization using C12resazurin to link enzyme activity to the development of a florescent signal.