Filippo Caschera

Automated synthetic cell-based screening for designed proteins with emergent functions @NatureComms 1. The paper introduces PUREdrop, an automated microfluidic screening platform that expresses protein libraries inside thousands of picoliter “synthetic cells” (water-in-oil droplets) per construct, then deposits each droplet population into predefined wells of a 96-well plate for time-lapse microscopy. 2. Key idea: emergent functions (e.g., cytoskeletal self-organization) require both confinement and spatiotemporal readouts. PUREdrop is positioned between bulk plate assays (good genotype-to-well mapping, poor spatial resolution) and ultrahigh-throughput droplet sorting (high throughput, but typically single-timepoint scalar readouts). 3. Hardware/workflow: a 5-way hydrodynamic focusing junction co-encapsulates DNA (IA1) with PURE TXTL reagents (IA2) into monodisperse ~20 µm droplets at ~3 kHz. An autosampler sequentially swaps DNA templates from a 96-well input plate; a fraction collector routes each droplet population into the correct destination well. 4. Automation details that enable reliable library screening: pneumatic valves switch between priming/production modes; ethanol+water cleaning cycles reduce DNA carryover; real-time pressure/flow monitoring stabilizes droplet generation and improves run-to-run reproducibility during multi-well operation. 5. Practical impact: reagent efficiency. By decoupling DNA from the TXTL stream and producing droplets on demand, the platform generates ~3×10^5 droplets using ~0.87 µL of key PUREfrex2.0 components per reaction, reported as ~7.4-fold lower PURE consumption versus manual emulsification workflows. 6. Validation experiment: alternating sfGFP and mCherry DNA across 48 wells showed clear per-well channel segregation with minimal cross-contamination. Time-lapse traces enabled quantitative expression-kinetics metrics (e.g., T0.5 and slope at T0.5), revealing reproducible construct-specific maturation/production dynamics. 7. End-to-end “design-to-screen” pipeline: Golden Gate assembly in plate format (Echo acoustic handling compatible), bacterial growth, PCR from lysate to generate PURE-ready linear templates, normalization, then a ~5-hour PUREdrop run for 48 constructs followed by overnight imaging and droplet-level segmentation/feature extraction. 8. Showcase 1 (designed proteins): screening computationally redesigned FtsZ variants. The authors used conservation/functional-site masking + ProteinMPNN redesign, AlphaFold2-based selection, and then screened 24 designs (plus WT and blinded replicates) for higher-order bundling dynamics inside droplets. 9. Phenotype quantification goes beyond “does it polymerize”: they track time-resolved bundle area trajectories (including onset timing and FWHM as a persistence descriptor) and morphology proxies (skeleton length; intensity coefficient of variation as a bundling compactness proxy). Hits included variants with earlier onset and more compact bundling (mpnnFtsZv18) and variants with delayed onset and reduced bundling (mpnnFtsZv14). 10. Showcase 2 (multi-component emergent function): screening FtsZ modulators by co-expression and reading out changes in FtsZ organization. Distinct modulators shifted lifetime and morphology; ZipA stood out by anchoring FtsZ assemblies to droplet–droplet interfaces and yielding ring-like structures, despite the compartments being surfactant-stabilized rather than membrane-bound. 💻Code: github.com/KANahas/PUREdrop.… 📜Paper: doi.org/10.1038/s41467-026-7… #SyntheticBiology #BottomUpBiology #SyntheticCells #Microfluidics #CellFree #ProteinDesign #ProteinEngineering #TXTL #FtsZ #SelfOrganization #Automation #HighThroughputScreening