A general strategy for expanding polymerase function by droplet microfluidics

Abstract

Polymerases that synthesize artificial genetic polymers hold great promise for advancing future applications in synthetic biology. However, engineering natural polymerases to replicate unnatural genetic polymers is a challenging problem. Here we present droplet-based optical polymerase sorting (DrOPS) as a general strategy for expanding polymerase function that employs an optical sensor to monitor polymerase activity inside the microenvironment of a uniform synthetic compartment generated by microfluidics. We validated this approach by performing a complete cycle of encapsulation, sorting and recovery on a doped library and observed an enrichment of ∼1,200-fold for a model engineered polymerase. We then applied our method to evolve a manganese-independent α-L-threofuranosyl nucleic acid (TNA) polymerase that functions with >99% template-copying fidelity. Based on our findings, we suggest that DrOPS is a versatile tool that could be used to evolve any polymerase function, where optical detection can be achieved by Watson-Crick base pairing.

Figure 1. Droplet-based optical polymerase sorting.

(a) We have developed a fluorescent reporter system that produces an optical signal when a primer–template complex is extended to full-length product. The reporter consists of a primer–template complex (pink and green) containing a downstream fluorophore that is quenched when a DNA-quencher (black) anneals to the unextended region. (b) The assay was designed with a metastable probe to allow dissociation at elevated temperatures, where thermophilic polymerases function with optimal activity. Red arrow marks the maximium fluorescence observed in the absence of the quencher probe. (c) Flourophore (F)/quencher (Q) pairs were screened to identify a dye pair with the maximum signal-to-noise ratio. (d) Primer-extension analysis by denaturing PAGE (top) and fluorescence (bottom) for 9n and 9n-GLK polymerases using dNTP and NTP substrates. Negative control: no NTPs. Positive control: dNTPs or no DNA-quencher probe. (e) Single-emulsion droplets containing a functional 9n-GLK polymerase that extends a primer–template complex with RNA (top) and non-functional (bottom) wild-type 9n polymerase. The panel shows a cartoon depiction of the droplet, a bright-field micrograph of encapsulated E. coli (arrow), a fluorescence micrograph of the same field of view and an overlay of the two images. Scale bars, 10 μm. (f) Flow cytometry analysis of 9n and 9n-GLK polymerases following NTP extension in water-in-oil-in-water (w/o/w) droplets.

Figure 2. Model selection of an engineered polymerase with RNA synthesis activity.

(a) Overview of the microfluidic polymerase enrichment strategy. A pool of polymerase genes containing functional (green) and non-functional (blue) members are expressed in E. coli and encapsulated in w/o droplets generated in a microfluidics device. Polymerases are liberated from their bacteria by heat lysis and incubated at 55 °C to allow for primer extension. Using a second microfluidics device, droplets are emulsified into a bulk aqueous phase to generate water-in-oil-in-water compartments (w/o/w). Fluorescent w/o/ws are FACS sorted and the vectors encoding functional polymerases are recovered. (b) Vector design. The 9n-GLK vector was engineered to contain a unique NotI restriction site. Control digestion showing that NotI only cuts PCR-amplified DNA from the 9n-GLK vector. (c) Following a complete cycle of selection and amplification (see Supplementary Fig. 1) PCR-amplified DNA was digested with NotI to measure the enrichment of 9n-GLK from libraries that were doped at levels of 1:100, 1:1,000 and 1:10,000 (9n-GLK to 9n). NotI digestion of the PCR-amplified DNA reveals an enrichment of ∼1,200-fold per round of microfluidics selection.

Figure 3. Selection of a Mn²⁺-independent TNA polymerase from a focused library.

(a) Constitutional structure for the linearized backbone of threose nucleic acid (TNA). (b) Positions 409, 485 and 664 mapped onto the structure of 9n DNA polymerase (PDB: 4K8X). Polymerases isolated after one round of selection were analysed for TNA synthesis activity in the absence of Mn²⁺. Activity is defined as the amount of full-length product generated in 18 h. Basal activity of wild-type 9n polymerase (dashed grey line). (c) Time course of TNA synthesis for 9n-YRI and 9n-NVA polymerases compared with wild-type 9n. (d) Fidelity analysis of 9n-YRI polymerase in the presence and absence of manganese ions yields a mutational profile of 8 errors per 100 bases and 2 errors per 1,000 bases, respectively.