Affinity-tag-based microfluidic protein isolation enables high-resolution Cryo-EM from minimal starting material

Abstract

Cryo-EM is central to high-resolution structure determination, but conventional sample preparation consumes substantial amounts of protein and can destabilize sensitive complexes. Microfluidic approaches can miniaturize and accelerate these workflows while retaining the particle numbers required for single-particle analysis. Here, we present a generalized microfluidic isolation strategy that captures proteins directly from cell lysates or in vitro translation reactions using genetically encoded tags. The platform supports both affinitybased (ALFA-nanobody) and covalent (SpyTag3/SpyCatcher3) capture. Specificity derives from two sequential, orthogonal steps: (i) tag-mediated capture and concentration of the target protein on beads, and (ii) photoelution from the bead surface. By confining the sample to nanoliter volumes and minimizing transport, the approach suppresses non-specific carryover. Importantly, the isolation workflow is coupled to grid preparation, enabling preparation of three cryo-EM or negative-stain grids from a 25 nL eluate. Using this workflow, we isolated E. coli ferritin A, {beta}-galactosidase, and P. aeruginosa VgrG1 from cell lysate, obtaining reconstructions at 1.9-2.6 [A] resolution with B-factors comparable to optimized conventional workflows. We further isolated ferritin A from an in vitro translation reaction, yielding a 2.04 [A] reconstruction. A standardized 50 {micro}L starting volume enabled more than 16 independent microfluidic isolations. Tag-based microfluidic isolation thus provides a broadly applicable route to cryo-EM sample preparation, reducing sample volumes more than 1000-fold, reducing preparation times more than 3-fold for lysate-based workflows and up to 10-fold when combined with cell-free expression, and enabling high-throughput structural screening directly from in vitro translation reactions. Significance StatementCryo-electron microscopy (cryo-EM) enables protein structure determination at atomic resolution, but conventional sample preparation requires large amounts of purified protein and laborious workflows that can damage fragile complexes. We present a microfluidic platform that captures tagged proteins directly from minute amounts of crude lysates or cell-free synthesis reactions and deposits them onto cryo-EM grids in a single integrated workflow. Using this approach, we determine three protein structures at high resolution. By reducing sample volumes by roughly three orders of magnitude, the platform expands cryo-EM to proteins inaccessible to conventional workflows and provides a foundation for automated, high-throughput structural analysis.