A rhythmically pulsing leaf-spring DNA-origami nanoengine that drives a passive follower

The study addresses the challenge of creating chemically fueled nanoscale engines capable of generating directed, rhythmic motion and transmitting force to downstream passive components. While many synthetic molecular machines rely on Brownian motion with limited directed force transmission, the goal here is to develop a biohybrid DNA-origami nanoengine powered by T7 RNA polymerase and NTPs to achieve periodic pulsing motion and to demonstrate mechanical coupling to a passive follower. The purpose is to enable modular, bottom-up construction of nanodevices that perform tasks such as pumping, transport, and sensing with active, controllable motion, advancing beyond prior Brownian ratchets or externally driven devices.

Prior work has demonstrated synthetic nanoscale pumps, rotors, walkers, and field-driven ratchets, as well as DNA-origami mechanisms with programmable mechanical properties. A previous biohybrid DNA rotor-stator driven by non-covalently bound T7 RNAP rotated a catenated wheel, but direct transfer of generated motion to a passive element remained unrealized. The present work builds on advances in DNA origami compliant mechanisms, light-driven nanomachines, and coarse-grained DNA simulations (oxDNA), addressing the gap in chemically fueled, rhythmic engines that can actuate passive followers.

• Design and construction: A DNA-origami nanoengine with two rigid 18-helix-bundle arms connected by a compliant leaf-spring flexure. A dsDNA transcription template (dsDNA-t) bridges the inner faces of the arms. A HaloTag–T7 RNA polymerase (HT-T7RNAP) is covalently tethered via a chloroalkane-modified staple near the promoter, enabling localized, fuel-driven transcription to pull the arms closed and subsequent active reopening upon termination.
• Variants: A nicked-nanoengine (nNE) with two nicks near dsDNA-t arm attachments to relieve torsional stress; a soft-hinge variant removing two staples in the flexure; attachment variations with dsDNA-t singly attached to one arm; D (driver) and F (passive follower) constructs assembled via complementary ssDNA overhangs (including F-soft-hinge and F-ss-hinge versions).
• Bulk transcription assay: Molecular beacon (MB) fluorescence reports RNA output. Conditions typically 10 nM origami, 2 equiv HT-T7RNAP, 2 mM NTPs, 37 °C; fluorescence monitored over ≥3.5 h. Relative transcription rates compared across constructs; PAGE verification of RNA length.
• TEM and AFM: Negative-stain TEM for structural verification and extensive angle distribution measurements under transcription and no-transcription conditions. RELION used for 2D class averaging; ImageJ/Fiji angle measurement; AFM for assembly verification and feature spacing.
• smFRET: Single-molecule TIRF-based FRET with Cy3/Cy5 labels on opposing arms to monitor inter-arm distance. Conditions at 25 °C with oxygen scavenging. FRET trajectories analyzed for dwell times (open/closed) and transition times (open→closed, closed→open). NTP-toggling (“switch”) experiments to start/stop engine by buffer exchange.
• Coarse-grained MD (oxDNA): Equilibrium and non-equilibrium pulling/relaxation simulations across designs (nNE, NE, NTS, with/without secondary structures in flexure) at 23 °C and 37 °C. Estimation of equilibrium angle distributions, torsional spring constants (equipartition), and relative closing/opening rates under applied 16 pN force and upon release.
• Driver–Follower assembly and analysis: Equimolar D and F annealed; AFM/TEM verification; quantification of D–F formation yield; TEM angle distributions for D–F under transcription and no-transcription; bulk transcription rates for D alone vs D–F variants.

• Bulk transcription confirms functionality and importance of covalent tethering: • Fully assembled nanoengine shows transcription rate ~5× higher than intermolecular controls lacking covalent tethering; estimated 2.3 ± 0.8 transcripts per minute at 37 °C (2 mM NTP). • Introducing two nicks in dsDNA-t (nNE) yields ~2× higher transcription rate versus non‑nicked (Fig. 2b), consistent with torsional stress relief. • Controls lacking promoter or tether largely inactive; preblocking HaloTag reduces activity as expected; adding free chloroalkane post-attachment has minimal effect.
• Angle distributions (TEM) evidence pulsing/closing under transcription: • No-transcription median angle ~67° (mean 64.13° ± 20.04°, n = 5,135); transcription shifts distribution to more acute angles (median ~58°, mean 56.51° ± 21.79°, n = 3,266), highly significant (P < 3 × 10^-5). • 2D classes under transcription include angles as small as ~17°, rarely observed without transcription.
• smFRET reveals rhythmic cycles and kinetics: • Two dominant FRET states: low EFRET ≈ 0.2 (open, inter-dye ~6.8 nm) and high EFRET ≈ 0.7 (closed, ~4.3 nm), matching TEM-derived distances. • Cycle times: 18.4 ± 2 s at 1 mM each NTP (N = 112, n = 178) vs 11.9 ± 2 s at 5 mM each NTP (N = 164, n = 311), consistent with bulk estimates (~one new transcript per 12 ± 5 s at 2 mM NTP, 37 °C). • Component times: T_o (open dwell) decreases with NTP (9.5 ± 0.2 s → 5.1 ± 0.1 s); T_t‑C (open→closed transition) decreases (1.3 ± 0.2 s → 0.7 ± 0.2 s); T_c (closed dwell) ~5.3–5.4 s; T_t‑O (closed→open transition) ~0.45–0.53 s (largely NTP‑independent). Estimated elongation ~68 nt/s at 5 mM each NTP, matching literature. • Abortive events observed at low NTP (0.1 mM) with frequency 0.13 ± 0.07 s^-1; reduced at 5 mM (0.06 ± 0.03 s^-1). • NTP-switching: Engines static without NTP, become dynamic upon NTP addition, and stall in current state upon NTP removal, demonstrating reversible control.
• Simulations support mechanical interpretation: • Equilibrium angle distributions and torsional spring constants quantified; prevention of secondary structure in flexure (NS) alters angle and opening rates; trends in opening/closing rates under applied/released force align with hinge mechanics. • Proximity between Halo-tagged site and promoter in nNE favors productive spacing relative to loose attachments (distance histogram).
• Driver–Follower coupling demonstrates motion transmission: • D–F formation yield 60% ± 20% (199 TEM micrographs, n = 2,718). • D–F angle distributions shift to more acute under transcription: no-transcription mean 64.09° ± 17.38° (n = 1,074) vs transcription 51.37° ± 17.38° (n = 1,190), highly significant (P = 8 × 10^-8). • F-ss-hinge alone has broad, obtuse angles (median 113°); when coupled to D, median shifts to 71° (no transcription) and 56° (with transcription), with markedly narrower distributions. • Transcription speeds relative to D alone (set to 1.00 ± 0.12): D–F‑ss‑hinge 1.15 ± 0.20 (P = 0.001); D–F‑soft‑hinge 1.16 ± 0.24 (P = 0.01); D–F 1.10 ± 0.24 (P = 0.06).

The results demonstrate a chemically fueled, rhythmic nanoengine whose motion is governed by the interplay between transcription-driven pulling by a covalently tethered T7 RNAP and the restoring force of a DNA-origami leaf-spring. Bulk assays and smFRET confirm cyclic open–closed pulsation, with kinetics modulated by NTP concentration and in agreement with known T7 RNAP behavior. TEM angle distributions reveal a statistically significant shift toward more acute angles under transcription, consistent with active closing driven by polymerase progression and relaxation upon termination. The NTP-switch experiments establish reversible on/off control, stalling in defined conformations upon fuel removal. Coarse-grained MD simulations quantify the hinge mechanics, spring constants, and design-dependent effects (e.g., nicks relieving torsional stress; secondary structure in flexure), supporting the experimental observations. Crucially, coupling the active driver to a passive follower transmits motion non-stochastically, narrowing angle distributions and shifting medians under transcription, thereby establishing a functional driver–follower module and suggesting applicability to powering more complex nanodevices.

This work presents a bottom-up, DNA-origami-based nanoengine that autonomously executes rhythmic, fuel-driven pulsing and actively transmits motion to a passive follower, forming a functional driver–follower pair. The system integrates a covalently tethered T7 RNAP for NTP-driven actuation, a compliant leaf-spring hinge for energy storage and release, and modular interfaces enabling coupling to passive units. Quantitative smFRET kinetics, TEM angle statistics, bulk transcription measurements, and oxDNA simulations collectively validate the mechanism and performance. The modularity of DNA origami suggests broad adaptability of the driver to other nanostructures to achieve larger or more complex rearrangements. Future directions include implementing clutch-like, light-controlled connectors for dynamic coupling/decoupling and photoresponsive promoter elements for optical control of motion even in the presence of fuel.

• Angle distributions are broad due to the transient nature of opening–closing cycles in bulk and surface-deposited samples; TEM snapshots capture heterogeneous states and can be influenced by out-of-plane bending/distortion upon surface adsorption.
• Differences in assay temperatures (smFRET at 25 °C vs bulk at 37 °C) may affect rates; a subset of devices may be inactive in bulk.
• D–F assembly yield is partial (60% ± 20%), and some constructs require specific hinge designs (e.g., F-ss-hinge) to optimize transmission characteristics.
• Engine operation depends on biochemical fuel (NTPs) and polymerase properties, including abortive initiation at low NTP and stalling upon fuel removal.