A partially self-regenerating synthetic cell

The study addresses the challenge of constructing a bottom-up, self-replicating synthetic cell by demonstrating partial self-regeneration of a transcription–translation system. Building on the concept of von Neumann’s universal constructor, a biochemical analog would replicate itself using DNA-encoded instructions, supplied building blocks, and energy. Prior work has shown key subsystems of a synthetic cell such as compartmentalization, metabolism, communication, and aspects of DNA and RNA replication. However, a major hurdle remains: achieving a transcription–translation system that can regenerate its own essential components while operating continuously. The authors propose and test a minimal recombinant system capable of regenerating essential proteins for transcription and translation and sustaining function in a controlled microfluidic environment. They hypothesize that careful management of resource competition and optimal resource allocation among expressed genes are critical to robust self-regeneration.

The paper situates its work within efforts toward bottom-up synthetic cells, referencing advances in compartmentalization, mobility and morphology, metabolism, communication, and nucleic acid replication. The universal constructor concept by von Neumann is cited as a theoretical underpinning, with prior implementations in cellular automata. Experimental analogs include autocatalytic chemical systems and self-replicating ribozymes. Cell-free systems, particularly the PURE system, enable defined transcription–translation but face reduced capacity over time and sensitivity to resource limitations. Recent progress includes in vitro DNA replication cycles and steps toward ribosome biogenesis and RNA synthesis, but integrating these into a self-regenerating system remains challenging. Resource loading and competition effects in cell-free systems and cells are recognized as key determinants of performance, motivating models of resource allocation to guide design.

• Platform: Continuous cell-free reactions were run in a custom microfluidic chemostat device with eight independent ~200 µL reactors. Each reactor experienced periodic dilution: 20% of volume replaced every 15 minutes by sequentially introducing energy solution, protein/ribosome solution, and DNA solution to achieve target fractions (approximately 8%, 8%, and 4% per cycle). Temperature was controlled at 34 °C.
• Cell-free system: A customized PURE system was prepared based on published formulations, enabling omission of selected components (e.g., specific aminoacyl-tRNA synthetases, aaRSs, or T7 RNA polymerase) to test essentiality and self-regeneration. TCEP replaced DTT to improve stability. Concentrations of certain aaRSs were reduced to sensitize washout dynamics and allow clearer detection of self-regeneration.
• Experimental phases: A three-phase protocol was implemented: (1) Kick-start (first 4 h): complete PURE with linear DNA templates for eGFP (reporter) and the target protein(s) to be regenerated to bootstrap expression; (2) Self-regeneration: switch to a modified PURE mixture lacking the target protein(s), while continuing to supply DNA templates, to test whether the system can sustain function by regenerating the missing components; (3) Washout: omit the target DNA templates to verify loss of function if self-regeneration ceases.
• Readouts: eGFP fluorescence served as a quantitative proxy for protein synthesis capacity and system function, monitored by automated fluorescence microscopy every 5 minutes. mScarlet was used in some controls. Data were normalized to positive controls.
• Design of regeneration tests: Experiments targeted (a) regeneration of individual essential translation proteins (e.g., Asparaginyl-tRNA synthetase, AsRS; Leucyl-tRNA synthetase, LeuRS), (b) regeneration of an essential transcription protein (T7 RNA polymerase), and (c) simultaneous regeneration of multiple proteins (combinations of T7 RNAP with two aaRSs, and sets of 4–7 aaRSs). For each, DNA template concentrations were titrated to map performance versus loading.
• Modeling: A minimal resource-competition model described coupled transcription–translation of eGFP and T7 RNAP with a single lumped limiting resource (R) consumed by transcription and translation. The model comprised seven ODEs with three parameters, enabling qualitative predictions about resource loading (competition between genes) and resource allocation (distribution between TX/TL), and the existence of optimal DNA template ratios. Simulations mimicked phase switches at 4 h and 16 h and explored low, medium, and high template regimes.
• Controls and validation: Batch PURE reactions validated expression and essentiality of targeted components. Negative controls omitted target DNA in ΔPURE mixes; positive controls used full PURE with eGFP DNA. Device operation details, calibration curves for eGFP, and component compositions are provided in supplementary materials.

• Individual aaRS regeneration: Both AsRS and LeuRS were successfully self-regenerated. Optimal DNA template concentrations produced eGFP levels comparable to positive controls during self-regeneration and complete loss upon washout. For AsRS, ~0.1 nM DNA supported robust self-regeneration; 0.05 nM failed (eGFP declined similarly to negative control). For LeuRS, ~0.5 nM was optimal; 0.1 nM gave slightly lower expression. Higher-than-optimal DNA still enabled regeneration but reduced eGFP due to resource loading.
• Resource competition: eGFP DNA was fixed at 2 nM, approximately saturating system capacity (~1 nM loading threshold in their PURE system). Adding target protein DNA above minimal levels induced resource competition, lowering eGFP output. The model predicted and experiments confirmed that eGFP levels during regeneration decrease with increasing target DNA due to loading and can exhibit peaks depending on allocation dynamics.
• T7 RNAP regeneration: T7 RNAP was regenerated across DNA inputs spanning three orders of magnitude, with an optimal regime around 0.5 nM supporting steady-state regeneration and sustained eGFP synthesis for over 85 hours when the washout phase was omitted (long-term operation up to at least 26 hours of regeneration window explicitly shown; steady-state projected/observed beyond). Very low DNA (<~0.05 nM) yielded little eGFP due to insufficient RNAP synthesis; high DNA led to strong loading effects and reduced eGFP yields.
• Multi-component regeneration (T7 RNAP + aaRSs): Initial attempts using DNA levels based on single-protein minima failed to sustain multi-component regeneration. Increasing DNA concentrations while keeping ratios constant enabled regeneration for 20–25 hours, but activity eventually ceased. Adjusting DNA ratios—retaining relatively high T7 RNAP DNA (e.g., ~0.2 nM) with matched aaRS DNA—improved robustness and sustained regeneration.
• Regeneration of multiple aaRSs (4–7 components): With optimized DNA inputs (≥0.2 nM per component), the system achieved successful self-regeneration for up to ~22 hours. At 0.1 nM, activity eroded within ~10 hours. Accounting for load from expressing 4–7 additional proteins, eGFP yield during self-regeneration reached roughly 50% of the theoretical maximum (positive control), indicating substantial remaining synthesis capacity under load.
• Yield–robustness trade-offs: The authors define yield (nonessential output, e.g., eGFP level during regeneration) and robustness (ability to sustain regeneration ≥24 h), showing Pareto-like behavior. Increasing expression of an essential protein moves the system toward a robustness of 1 past a critical concentration but can decrease yield due to loading if overexpressed. Multi-component regeneration exhibits narrower optimal DNA ranges and stronger trade-offs; nonetheless, high robustness is achievable with careful allocation.
• Overall: Demonstrated partial self-regeneration of essential transcription and translation proteins in a continuous microfluidic environment, mapped DNA concentration windows for robust operation, and validated a minimal model capturing resource loading and allocation effects critical for system design.

The findings directly address the feasibility of a biochemical constructor capable of partial self-regeneration, a key step toward a self-replicating synthetic cell. By showing that essential components for transcription (T7 RNAP) and translation (aaRSs) can be continuously regenerated from DNA templates while maintaining system function, the work validates that encoded instructions plus supplied energy and substrates can maintain the core expression machinery. The experiments expose the central role of resource competition and allocation: beyond minimal expression thresholds for essential proteins, additional DNA increases system load and diminishes yields of other outputs (e.g., eGFP). The minimal ODE model explains qualitative trends, guiding selection of DNA template concentrations and ratios to balance yield and robustness. Importantly, simultaneous regeneration of multiple components is possible but requires precise tuning to avoid collapse due to overloading. Conceptually, this aligns with von Neumann’s universal constructor framework: a core constructor (RNAs/proteins) reads DNA instructions to produce copies of its own parts. The study frames system performance in terms of yield and robustness, akin to fitness landscapes, suggesting Pareto-optimal fronts for design. This provides actionable principles for integrating additional subsystems such as DNA replication and ribosome biogenesis.

This work demonstrates a partially self-regenerating synthetic cell-like system: a defined, continuous transcription–translation platform that regenerates essential proteins (aaRSs and T7 RNAP) from DNA, sustains activity for many hours to days, and can regenerate multiple components simultaneously with appropriate DNA ratios. The authors introduce an experimental framework (kick-start/self-regeneration/washout in microfluidic chemostats) and a minimal resource-based model that together enable rational tuning of resource loading and allocation. These advances provide design principles—balancing yield and robustness via DNA concentration and ratio optimization—toward constructing a biochemical universal constructor. Future research should: (i) increase synthesis rates and efficiencies to meet the demands of regenerating all PURE components; (ii) achieve functional ribosome and tRNA biogenesis; (iii) demonstrate steady-state DNA self-replication integrated with expression; and (iv) implement feedback control and genome-level tuning (promoters, RBSs, translation factors) to maintain optimal allocation as system complexity scales.

• Insufficient synthesis throughput: The current PURE-based system operates below the synthesis rates required to regenerate all components; the authors estimate roughly a 25-fold increase in synthesis rate would be needed and that only ~50% of PURE proteins could currently be regenerated.
• Scope of regeneration: Ribosome and tRNA synthesis were not demonstrated; regeneration was limited to selected essential proteins (aaRSs and T7 RNAP).
• Sensitivity to loading: Robust operation depends on narrow DNA concentration windows; higher-than-optimal inputs induce resource competition that reduces yield and can compromise robustness, particularly in multi-component settings.
• Long-term stability and generalizability: While T7 RNAP could be sustained for extended periods, multi-component regeneration eventually ceased under some conditions, indicating fragility. Broader applicability across diverse proteins with varying activities/folding requirements remains to be established.
• Modeling granularity: The minimal lumped-resource model captures qualitative behavior but may not account for all observed phenomena or specific resource species dynamics, limiting quantitative predictive power.