Traditional directed evolution operates in discrete cycles. You mutate a gene library, screen or select for improved variants, isolate winners, and repeat. Each round might take days or weeks. PACE—Phage-Assisted Continuous Evolution—collapses this into a continuous process where evolution runs autonomously, churning through hundreds of generations in a single week.

The system exploits bacteriophage biology with elegant precision. By linking the activity of an evolving protein to phage survival, PACE creates a selection pressure that operates continuously in flowing bacterial cultures. Beneficial mutations propagate; deleterious ones wash out. No human intervention required between generations. The result is evolutionary timescales compressed into laboratory schedules.

What makes PACE particularly powerful is its modularity. The core architecture—phage replication coupled to protein function—can be adapted to evolve almost any protein activity you can link to gene expression. Transcription factors, polymerases, proteases, base editors: all have been evolved through PACE to achieve functions that would be extraordinarily difficult to engineer rationally. This system represents a fundamental shift in how we approach protein engineering, moving from iterative optimization to autonomous evolutionary exploration.

The foundation of PACE rests on a single genetic manipulation: removing gene III from the M13 bacteriophage genome and placing it under control of the evolving protein's activity. Gene III encodes pIII, a coat protein absolutely required for phage infectivity. Without functional pIII, phage particles can assemble but cannot infect new host cells. This creates a life-or-death selection pressure directly tied to the protein you're evolving.

The evolving gene sits on the selection phage (SP), which propagates through host E. coli in a continuous-flow bioreactor called a lagoon. Fresh bacteria flow in constantly while phage and spent cells wash out. The dilution rate is calibrated to exceed bacterial division time but remain below phage replication rate. This means bacteria cannot evolve—they're replaced before they can divide—while phage must continuously replicate to avoid being diluted out.

Host cells contain an accessory plasmid (AP) harboring gene III under control of a promoter that responds to the activity being evolved. If the evolving protein performs its function—activating a promoter, cleaving a substrate, synthesizing a product—gene III expresses, pIII accumulates, and progeny phage become infectious. They propagate to new hosts and continue replicating. Phage carrying inactive or poorly functioning variants produce less pIII, fewer infectious progeny, and are outcompeted.

The beauty of this architecture lies in its self-sustaining nature. Every phage replication cycle represents a generation of evolution. With phage generation times around 10-20 minutes, PACE can achieve 20-50 generations per day—rates impossible with traditional screening methods. Over a week-long experiment, populations traverse evolutionary distances that would otherwise require months or years of iterative selection.

Selection stringency can be tuned by modifying the AP. Weaker promoters, lower gene III expression, or additional regulatory elements increase the activity threshold for survival. Many PACE experiments implement drift phases with relaxed selection, allowing neutral mutations to accumulate before reimposing stringent selection. This strategy helps populations escape local fitness maxima and discover novel evolutionary solutions.

Takeaway

Coupling survival to function at the genetic level transforms evolution from a process you impose into one that runs itself, limited only by how creatively you design the coupling.

Natural mutation rates are far too low for efficient directed evolution. E. coli polymerases introduce roughly one error per billion base pairs replicated—excellent for genome stability, terrible for exploring sequence space. PACE overcomes this through mutagenesis plasmids (MPs) that dramatically elevate error rates specifically during phage replication.

The original MP employs a dominant-negative variant of proofreading subunit dnaQ, which interferes with the ε subunit of DNA polymerase III. Combined with overexpression of error-prone polymerase components, this increases mutation rates approximately 100-fold. Importantly, this mutagenesis targets all DNA replication in the cell, but since bacteria wash out before dividing, only phage genomes accumulate and inherit mutations.

More sophisticated MPs have been developed to further enhance mutagenesis. Second-generation systems incorporate additional error-prone polymerases, base-modifying enzymes, and DNA damage-inducing components. Some versions achieve mutation rates exceeding 10^-3 substitutions per base pair per replication—a millionfold increase over wild-type. At these rates, every phage replication generates variants sampling nearby sequence space.

The mutagenesis must be balanced carefully. Too little, and evolution stalls as populations wait for beneficial mutations. Too much, and error catastrophe occurs—genomes accumulate so many deleterious mutations that no lineage maintains functionality. The optimal mutation rate depends on gene length, fitness landscape ruggedness, and population size. PACE experiments often begin with moderate mutagenesis and increase it as populations adapt.

Temporal control adds another dimension. Some MPs are inducible, allowing researchers to toggle mutagenesis during experiments. High mutation rates help discover novel solutions; reducing mutagenesis afterward can polish variants by purging accumulated neutral or slightly deleterious mutations. This mimics natural evolutionary dynamics where exploratory and consolidating phases alternate.

Takeaway

Effective directed evolution requires not just selection pressure but calibrated chaos—enough mutation to explore possibilities without destroying what already works.

PACE's versatility stems from its modular selection architecture. Any protein activity that can be linked to gene III expression becomes evolvable. The challenge lies in designing selection circuits where the connection between protein function and pIII production is tight, specific, and tunable. This engineering problem has driven remarkable creativity in synthetic biology.

Transcription factors represent the simplest case. Place gene III downstream of a promoter controlled by the evolving transcription factor, and DNA-binding activity directly drives selection. This approach has evolved transcription factors with altered DNA specificity, improved binding affinity, and enhanced cooperativity. The Liu laboratory used this strategy to evolve TALE proteins and subsequently base editors with dramatically improved activity.

Enzymes require more elaborate circuits. For proteases, the strategy involves fusion proteins where a protease-cleavable linker tethers a transcription factor to an inhibitory domain. Cleavage releases the transcription factor, activating gene III expression. This approach evolved TEV protease variants with novel substrate specificities—a remarkably difficult problem for rational engineering.

Polymerases presented a different challenge. The solution involved a two-hybrid-like system where the evolving polymerase must synthesize a functional gene III transcript from a template. Only polymerases capable of faithful, processive synthesis produce enough pIII for survival. This enabled evolution of DNA polymerases, RNA polymerases, and reverse transcriptases with altered substrate preferences, improved fidelity, or enhanced thermostability.

Negative selection circuits expand PACE's capabilities further. By linking undesired activities to expression of gene III-neg (which produces a dominant-negative pIII variant), researchers can select against specific functions while selecting for others. This enables evolution toward specificity rather than just activity—critical for applications requiring precise molecular recognition without off-target effects.

Takeaway

The power of an evolutionary system depends entirely on the fitness function you impose—design the selection circuit well, and evolution will find solutions you never imagined.

PACE represents a philosophical shift in protein engineering: from designing molecules to designing evolutionary systems that discover molecules. The researcher's role transforms from engineer to architect of selection pressures. You define what fitness means; evolution does the searching.

The implications extend beyond practical protein engineering. PACE provides a window into evolutionary dynamics at unprecedented resolution. Researchers can watch adaptation unfold in real time, sequence populations at multiple timepoints, and reconstruct the mutational paths that lead to novel functions. These experiments illuminate how evolution navigates fitness landscapes—knowledge applicable far beyond the laboratory.

As selection circuit designs grow more sophisticated and mutagenesis systems more powerful, PACE continues expanding into new protein families and functional challenges. The platform exemplifies synthetic biology's core promise: using biological systems to accomplish what neither rational design nor natural evolution could achieve alone.