Contingency and chance erase necessity in the experimental evolution of ancestral proteins

High-throughput protein binder discovery by rapid in vivo selection

Proteins that selectively bind to a target of interest are foundational components of research pipelines1,2, diagnostics3, and therapeutics4. Current immunization-based5,6, display-based7-14, and computational approaches15-1718 for discovering binders are laborious and time-consuming - taking months or more, suffer from high false positives - necessitating extensive secondary screening, and have a high failure rate, especially for disordered proteins and other challenging target classes. Here we establish Phage-Assisted Non-Continuous Selection of Protein Binders (PANCS-binders), an in vivo selection platform that links the life cycle of M13 phage to target protein binding though customized proximity-dependent split RNA polymerase biosensors, allowing for complete and comprehensive high-throughput screening of billion-plus member protein variant libraries with high signal-to-noise. We showcase the utility of PANCS-Binders by screening multiple protein libraries each against a panel of 95 separate therapeutically relevant targets, thereby individually assessing over 1011 protein-protein interaction pairs, completed in two days. These selections yielded large, high-quality datasets and hundreds of novel binders, which we showed can be affinity matured or directly used in mammalian cells to inhibit or degrade targets. PANCS-Binders dramatically accelerates and simplifies the binder discovery process, the democratization of which will help unlock new creative potential in proteome-targeting with engineered binder-based biotechnologies.

Functional differences between rodent and human PD-1 linked to evolutionary divergence

Mechanistic understanding of the inhibitory immunoreceptor PD-1 is largely based on mouse models, but human and mouse PD-1 share only 59.6% amino acid identity. Here, we found that human PD-1 is more inhibitory than mouse PD-1, owing to stronger interactions with the ligands PD-L1 and PD-L2 and more efficient recruitment of the effector phosphatase Shp2. In a mouse melanoma model with adoptively transferred T cells, humanization of a PD-1 intracellular domain disrupted the antitumor activity of CD8+ T cells and increased the magnitude of anti-PD-1 response. We identified a motif highly conserved across vertebrate PD-1 orthologs, absent in rodents, as a key determinant for differential Shp2 recruitment. Evolutionary analysis suggested that PD-1 underwent a rodent lineage-specific functional attenuation during evolution. Together, our study uncovers species-specific features of the PD-1 pathway, with implications for PD-1 evolution and differential anti-PD-(L)1 responses in mouse models and human patients.

Specialization Restricts the Evolutionary Paths Available to Yeast Sugar Transporters

Functional innovation at the protein level is a key source of evolutionary novelties. The constraints on functional innovations are likely to be highly specific in different proteins, which are shaped by their unique histories and the extent of global epistasis that arises from their structures and biochemistries. These contextual nuances in the sequence-function relationship have implications both for a basic understanding of the evolutionary process and for engineering proteins with desirable properties. Here, we have investigated the molecular basis of novel function in a model member of an ancient, conserved, and biotechnologically relevant protein family. These Major Facilitator Superfamily sugar porters are a functionally diverse group of proteins that are thought to be highly plastic and evolvable. By dissecting a recent evolutionary innovation in an α-glucoside transporter from the yeast Saccharomyces eubayanus, we show that the ability to transport a novel substrate requires high-order interactions between many protein regions and numerous specific residues proximal to the transport channel. To reconcile the functional diversity of this family with the constrained evolution of this model protein, we generated new, state-of-the-art genome annotations for 332 Saccharomycotina yeast species spanning ∼400 My of evolution. By integrating phylogenetic and phenotypic analyses across these species, we show that the model yeast α-glucoside transporters likely evolved from a multifunctional ancestor and became subfunctionalized. The accumulation of additive and epistatic substitutions likely entrenched this subfunction, which made the simultaneous acquisition of multiple interacting substitutions the only reasonably accessible path to novelty.

Determinants of Unfolding Cooperativity and Binding Are Decoupled in a DNA Binding Domain

The relative magnitudes of noncovalent stabilization energies or the coupling free energies in folded proteins are anisotropically distributed, uniquely influencing folding and functional behaviors. In this regard, the fructose repressor (FruR) DBD belonging to the LacR repressor family harbors a three-residue insertion─KQY─between the canonical second and third helices. This sequence insertion promotes a strong Tyr-Tyr stacking interaction that is not observed in related homologues. Combining experiments with simulations, we show that the Tyr-Tyr stacking contributes to a decoupled unfolding due to the localization of a large part of the stabilization energy in this specific structural region. This leads to melting temperatures from different probes spanning nearly 10 K, while concomitantly stabilizing a partially structured intermediate state. Disruption of the aromatic stacking interaction via an alanine mutation promotes a molten-globular state whose native ensemble is replete with non-native interactions while displaying enhanced thermodynamic fluctuations and minimal calorimetric cooperativity. Surprisingly, the molten-globular variant of FruR DBD binds to the operator site on DNA with an affinity similar to that of the wild-type but with altered secondary-structure characteristics in the bound state, underscoring the chaperone-like role of DNA through its large negative electrostatic potential. FruR DBD thus appears to be at the verge of disorder as expected of an entropically destabilizing three-residue insertion but is rescued by the aromatic stacking interaction that distinctly dictates the finer details of stability, cooperativity, and binding.

Understanding epistatic networks in the B1 β-lactamases through coevolutionary statistical modeling and deep mutational scanning

Throughout evolution, protein families undergo substantial sequence divergence while preserving structure and function. Although most mutations are deleterious, evolution can explore sequence space via epistatic networks of intramolecular interactions that alleviate the harmful mutations. However, comprehensive analysis of such epistatic networks across protein families remains limited. Thus, we conduct a family wide analysis of the B1 metallo-β-lactamases, combining experiments (deep mutational scanning, DMS) on two distant homologs (NDM-1 and VIM-2) and computational analyses (in silico DMS based on Direct Coupling Analysis, DCA) of 100 homologs. The methods jointly reveal and quantify prevalent epistasis, as ~1/3rd of equivalent mutations are epistatic in DMS. From DCA, half of the positions have a >6.5 fold difference in effective number of tolerated mutations across the entire family. Notably, both methods locate residues with the strongest epistasis in regions of intermediate residue burial, suggesting a balance of residue packing and mutational freedom in forming epistatic networks. We identify entrenched WT residues between NDM-1 and VIM-2 in DMS, which display statistically distinct behaviors in DCA from non-entrenched residues. Entrenched residues are not easily compensated by changes in single nearby interactions, reinforcing existing findings where a complex epistatic network compounds smaller effects from many interacting residues.

Evolutionary fingerprint in rodent PD1 confers weakened activity and enhanced tumor immunity compared to human PD1

Mechanistic understanding of the immune checkpoint receptor PD1 is largely based on mouse models, but human and mouse PD1 orthologs exhibit only 59.6% identity in amino acid sequences. Here we show that human PD1 is more inhibitory than mouse PD1 due to stronger interactions with the ligands PDL1 and PDL2 and with the effector phosphatase Shp2. A novel motif highly conserved among PD1 orthologs in vertebrates except in rodents is primarily responsible for the differential Shp2 recruitment. Evolutionary analysis suggested that rodent PD1 orthologs uniquely underwent functional relaxation, particularly during the K-Pg boundary. Humanization of the PD1 intracellular domain disrupted the anti-tumor activity of mouse T cells while increasing the magnitude of anti-PD1 response. Together, our study uncovers species-specific features of the PD1 pathway, with implications to PD1 evolution and differential anti-PD(L)1 responses in mouse models and human patients.

Specialization restricts the evolutionary paths available to yeast sugar transporters

Functional innovation at the protein level is a key source of evolutionary novelties. The constraints on functional innovations are likely to be highly specific in different proteins, which are shaped by their unique histories and the extent of global epistasis that arises from their structures and biochemistries. These contextual nuances in the sequence-function relationship have implications both for a basic understanding of the evolutionary process and for engineering proteins with desirable properties. Here, we have investigated the molecular basis of novel function in a model member of an ancient, conserved, and biotechnologically relevant protein family. These Major Facilitator Superfamily sugar porters are a functionally diverse group of proteins that are thought to be highly plastic and evolvable. By dissecting a recent evolutionary innovation in an α-glucoside transporter from the yeast Saccharomyces eubayanus, we show that the ability to transport a novel substrate requires high-order interactions between many protein regions and numerous specific residues proximal to the transport channel. To reconcile the functional diversity of this family with the constrained evolution of this model protein, we generated new, state-of-the-art genome annotations for 332 Saccharomycotina yeast species spanning approximately 400 million years of evolution. By integrating phylogenetic and phenotypic analyses across these species, we show that the model yeast α-glucoside transporters likely evolved from a multifunctional ancestor and became subfunctionalized. The accumulation of additive and epistatic substitutions likely entrenched this subfunction, which made the simultaneous acquisition of multiple interacting substitutions the only reasonably accessible path to novelty.

Pangenomes at the limits of evolution

Evolutionary pathways can be random or deterministic. In a recent article, Beavan et al. investigate this balance by applying machine learning models to microbial pangenomes. The presence of almost one-third of genes can be reliably inferred, indicating a surprising amount of predictable evolution.

Folded Alpha Helical Putative New Proteins from Apilactobacillus kunkeei

The emergence of new proteins is a central question in biology. Most tertiary protein folds known to date appear to have an ancient origin, but it is clear from bioinformatic analyses that new proteins continuously emerge in all organismal groups. However, there is a paucity of experimental data on new proteins regarding their structure and biophysical properties. We performed a detailed phylogenetic analysis and identified 48 putative open reading frames in the honeybee-associated bacterium Apilactobacillus kunkeei for which no or few homologs could be identified in closely-related species, suggesting that they could be relatively new on an evolutionary time scale and represent recently evolved proteins. Using circular dichroism-, fluorescence- and nuclear magnetic resonance (NMR) spectroscopy we investigated six of these proteins and show that they are not intrinsically disordered, but populate alpha-helical dominated folded states with relatively low thermodynamic stability (0-3 kcal/mol). The NMR and biophysical data demonstrate that small new proteins readily adopt simple folded conformations suggesting that more complex tertiary structures can be continuously re-invented during evolution by fusion of such simple secondary structure elements. These findings have implications for the general view on protein evolution, where de novo emergence of folded proteins may be a common event.

A morphological basis for path-dependent evolution of visual systems

Path dependence influences macroevolutionary predictability by constraining potential outcomes after critical evolutionary junctions. Although it has been demonstrated in laboratory experiments, path dependence is difficult to demonstrate in natural systems because of a lack of independent replicates. Here, we show that two types of distributed visual systems recently evolved twice within chitons, demonstrating rapid and path-dependent evolution of a complex trait. The type of visual system that a chiton lineage can evolve is constrained by the number of openings for sensory nerves in its shell plates. Lineages with more openings evolve visual systems with thousands of eyespots, whereas those with fewer openings evolve visual systems with hundreds of shell eyes. These macroevolutionary outcomes shaped by path dependence are both deterministic and stochastic because possibilities are restricted yet not entirely predictable.

Evolution of a plant growth-regulatory protein interaction specificity

Specific protein-protein interactions (PPIs) enable biological regulation. However, the evolution of PPI specificity is little understood. Here we trace the evolution of the land-plant growth-regulatory DELLA-SLY1/GID2 PPI, revealing progressive increase in specificity of affinity of SLY1/GID2 for a particular DELLA form. While early-diverging SLY1s display relatively broad-range DELLA affinity, later-diverging SLY1s tend towards increasingly stringent affinity for a specific DELLA A' form generated by the growth-promoting phytohormone gibberellin (GA). Our novel mutational strategy reveals amino acid substitutions contributing to the evolution of Arabidopsis thaliana SLY1 A' specificity, also showing that routes permitting reversion to broader affinity became increasingly constrained over evolutionary time. We suggest that progressive affinity narrowing may be an important evolutionary driver of PPI specificity and that increase in SLY1/GID2-DELLA specificity enabled the enhanced flexibility of plant physiological environmental adaptation conferred by the GA-DELLA growth-regulatory mechanism.

Multifractal foundations of biomarker discovery for heart disease and stroke

Any reliable biomarker has to be specific, generalizable, and reproducible across individuals and contexts. The exact values of such a biomarker must represent similar health states in different individuals and at different times within the same individual to result in the minimum possible false-positive and false-negative rates. The application of standard cut-off points and risk scores across populations hinges upon the assumption of such generalizability. Such generalizability, in turn, hinges upon this condition that the phenomenon investigated by current statistical methods is ergodic, i.e., its statistical measures converge over individuals and time within the finite limit of observations. However, emerging evidence indicates that biological processes abound with nonergodicity, threatening this generalizability. Here, we present a solution for how to make generalizable inferences by deriving ergodic descriptions of nonergodic phenomena. For this aim, we proposed capturing the origin of ergodicity-breaking in many biological processes: cascade dynamics. To assess our hypotheses, we embraced the challenge of identifying reliable biomarkers for heart disease and stroke, which, despite being the leading cause of death worldwide and decades of research, lacks reliable biomarkers and risk stratification tools. We showed that raw R-R interval data and its common descriptors based on mean and variance are nonergodic and non-specific. On the other hand, the cascade-dynamical descriptors, the Hurst exponent encoding linear temporal correlations, and multifractal nonlinearity encoding nonlinear interactions across scales described the nonergodic heart rate variability more ergodically and were specific. This study inaugurates applying the critical concept of ergodicity in discovering and applying digital biomarkers of health and disease.

Molecular determinants of protein evolvability

The plethora of biological functions that sustain life is rooted in the remarkable evolvability of proteins. An emerging view highlights the importance of a protein's initial state in dictating evolutionary success. A deeper comprehension of the mechanisms that govern the evolvability of these initial states can provide invaluable insights into protein evolution. In this review, we describe several molecular determinants of protein evolvability, unveiled by experimental evolution and ancestral sequence reconstruction studies. We further discuss how genetic variation and epistasis can promote or constrain functional innovation and suggest putative underlying mechanisms. By establishing a clear framework for these determinants, we provide potential indicators enabling the forecast of suitable evolutionary starting points and delineate molecular mechanisms in need of deeper exploration.

Innate Conformational Dynamics Drive Binding Specificity in Anti-Apoptotic Proteins Mcl-1 and Bcl-2

The structurally conserved B-cell lymphoma 2 (Bcl-2) family of protein function to promote or inhibit apoptosis through an exceedingly complex web of specific, intrafamilial protein-protein interactions. The critical role of these proteins in lymphomas and other cancers has motivated a widespread interest in understanding the molecular mechanisms that drive specificity in Bcl-2 family interactions. However, the high degree of structural similarity among Bcl-2 homologues has made it difficult to rationalize the highly specific (and often divergent) binding behavior exhibited by these proteins using conventional structural arguments. In this work, we use time-resolved hydrogen deuterium exchange mass spectrometry to explore shifts in conformational dynamics associated with binding partner engagement in the Bcl-2 family proteins Bcl-2 and Mcl-1. Using this approach combined with homology modeling, we reveal that Mcl-1 binding is driven by a large-scale shift in conformational dynamics, while Bcl-2 complexation occurs primarily through a classical charge compensation mechanism. This work has implications for understanding the evolution of internally regulated biological systems composed of structurally similar proteins and for the development of drugs targeting Bcl-2 family proteins for promotion of apoptosis in cancer.

Evolutionary repeatability of emergent properties of ecological communities

Most species belong to ecological communities where their interactions give rise to emergent community-level properties, such as diversity and productivity. Understanding and predicting how these properties change over time has been a major goal in ecology, with important practical implications for sustainability and human health. Less attention has been paid to the fact that community-level properties can also change because member species evolve. Yet, our ability to predict long-term eco-evolutionary dynamics hinges on how repeatably community-level properties change as a result of species evolution. Here, we review studies of evolution of both natural and experimental communities and make the case that community-level properties at least sometimes evolve repeatably. We discuss challenges faced in investigations of evolutionary repeatability. In particular, only a handful of studies enable us to quantify repeatability. We argue that quantifying repeatability at the community level is critical for approaching what we see as three major open questions in the field: (i) Is the observed degree of repeatability surprising? (ii) How is evolutionary repeatability at the community level related to repeatability at the level of traits of member species? (iii) What factors affect repeatability? We outline some theoretical and empirical approaches to addressing these questions. Advances in these directions will not only enrich our basic understanding of evolution and ecology but will also help us predict eco-evolutionary dynamics. This article is part of the theme issue 'Interdisciplinary approaches to predicting evolutionary biology'.

BCL-G: 20 years of research on a non-typical protein from the BCL-2 family

Proteins from the BCL-2 family control cell survival and apoptosis in health and disease, and regulate apoptosis-unrelated cellular processes. BCL-Gonad (BCL-G, also known as BCL2-like 14) is a non-typical protein of the family as its long isoform (BCL-G_L) consists of BH2 and BH3 domains without the BH1 motif. BCL-G is predominantly expressed in normal testes and different organs of the gastrointestinal tract. The complexity of regulatory mechanisms of BCL-G expression and post-translational modifications suggests that BCL-G may play distinct roles in different types of cells and disorders. While several genetic alterations of BCL2L14 have been reported, gene deletions and amplifications prevail, which is also confirmed by the analysis of sequencing data for different types of cancer. Although the studies validating the phenotypic consequences of genetic manipulations of BCL-G are limited, the role of BCL-G in apoptosis has been undermined. Recent studies using gene-perturbation approaches have revealed apoptosis-unrelated functions of BCL-G in intracellular trafficking, immunomodulation, and regulation of the mucin scaffolding network. These studies were, however, limited mainly to the role of BCL-G in the gastrointestinal tract. Therefore, further efforts using state-of-the-art methods and various types of cells are required to find out more about BCL-G activities. Deciphering the isoform-specific functions of BCL-G and the BCL-G interactome may result in the designing of novel therapeutic approaches, in which BCL-G activity will be either imitated using small-molecule BH3 mimetics or inhibited to counteract BCL-G upregulation. This review summarizes two decades of research on BCL-G.

PROT-ON: A structure-based detection of designer PROTein interface MutatiONs

The mutation-induced changes across protein-protein interfaces have often been observed to lead to severe diseases. Therefore, several computational tools have been developed to predict the impact of such mutations. Among these tools, FoldX and EvoEF1 stand out as fast and accurate alternatives. Expanding on the capabilities of these tools, we have developed the PROT-ON (PROTein-protein interface mutatiONs) framework, which aims at delivering the most critical protein interface mutations that can be used to design new protein binders. To realize this aim, PROT-ON takes the 3D coordinates of a protein dimer as an input. Then, it probes all possible interface mutations on the selected protein partner with EvoEF1 or FoldX. The calculated mutational energy landscape is statistically analyzed to find the most enriching and depleting mutations. Afterward, these extreme mutations are filtered out according to stability and optionally according to evolutionary criteria. The final remaining mutation list is presented to the user as the designer mutation set. Together with this set, PROT-ON provides several residue- and energy-based plots, portraying the synthetic energy landscape of the probed mutations. The stand-alone version of PROT-ON is deposited at https://github.com/CSB-KaracaLab/prot-on. The users can also use PROT-ON through our user-friendly web service http://proton.tools.ibg.edu.tr:8001/ (runs with EvoEF1 only). Considering its speed and the range of analysis provided, we believe that PROT-ON presents a promising means to estimate designer mutations.

Prediction of designer-recombinases for DNA editing with generative deep learning

Site-specific tyrosine-type recombinases are effective tools for genome engineering, with the first engineered variants having demonstrated therapeutic potential. So far, adaptation to new DNA target site selectivity of designer-recombinases has been achieved mostly through iterative cycles of directed molecular evolution. While effective, directed molecular evolution methods are laborious and time consuming. Here we present RecGen (Recombinase Generator), an algorithm for the intelligent generation of designer-recombinases. We gather the sequence information of over one million Cre-like recombinase sequences evolved for 89 different target sites with which we train Conditional Variational Autoencoders for recombinase generation. Experimental validation demonstrates that the algorithm can predict recombinase sequences with activity on novel target-sites, indicating that RecGen is useful to accelerate the development of future designer-recombinases.

Simple mechanisms for the evolution of protein complexity

Proteins are tiny models of biological complexity: specific interactions among their many amino acids cause proteins to fold into elaborate structures, assemble with other proteins into higher-order complexes, and change their functions and structures upon binding other molecules. These complex features are classically thought to evolve via long and gradual trajectories driven by persistent natural selection. But a growing body of evidence from biochemistry, protein engineering, and molecular evolution shows that naturally occurring proteins often exist at or near the genetic edge of multimerization, allostery, and even new folds, so just one or a few mutations can trigger acquisition of these properties. These sudden transitions can occur because many of the physical properties that underlie these features are present in simpler proteins as fortuitous by-products of their architecture. Moreover, complex features of proteins can be encoded by huge arrays of sequences, so they are accessible from many different starting points via many possible paths. Because the bridges to these features are both short and numerous, random chance can join selection as a key factor in explaining the evolution of molecular complexity.

Epistatic drift causes gradual decay of predictability in protein evolution

Epistatic interactions can make the outcomes of evolution unpredictable, but no comprehensive data are available on the extent and temporal dynamics of changes in the effects of mutations as protein sequences evolve. Here, we use phylogenetic deep mutational scanning to measure the functional effect of every possible amino acid mutation in a series of ancestral and extant steroid receptor DNA binding domains. Across 700 million years of evolution, epistatic interactions caused the effects of most mutations to become decorrelated from their initial effects and their windows of evolutionary accessibility to open and close transiently. Most effects changed gradually and without bias at rates that were largely constant across time, indicating a neutral process caused by many weak epistatic interactions. Our findings show that protein sequences drift inexorably into contingency and unpredictability, but that the process is statistically predictable, given sufficient phylogenetic and experimental data.

Methods for the directed evolution of biomolecular interactions

Noncovalent interactions between biomolecules such as proteins and nucleic acids coordinate all cellular processes through changes in proximity. Tools that perturb these interactions are and will continue to be highly valuable for basic and translational scientific endeavors. By taking cues from natural systems, such as the adaptive immune system, we can design directed evolution platforms that can generate proteins that bind to biomolecules of interest. In recent years, the platforms used to direct the evolution of biomolecular binders have greatly expanded the range of types of interactions one can evolve. Herein, we review recent advances in methods to evolve protein-protein, protein-RNA, and protein-DNA interactions.

The mutation effect reaction norm (mu-rn) highlights environmentally dependent mutation effects and epistatic interactions

Since the modern synthesis, the fitness effects of mutations and epistasis have been central yet provocative concepts in evolutionary and population genetics. Studies of how the interactions between parcels of genetic information can change as a function of environmental context have added a layer of complexity to these discussions. Here I introduce the "mutation effect reaction norm" (Mu-RN), a new instrument through which one can analyze the phenotypic consequences of mutations and interactions across environmental contexts. It embodies the fusion of measurements of genetic interactions with the reaction norm, a classic depiction of the performance of genotypes across environments. I demonstrate the utility of the Mu-RN through the signature of a "compensatory ratchet" mutation that undermines reverse evolution of antimicrobial resistance. More broadly, I argue that the mutation effect reaction norm may help us resolve the dynamism and unpredictability of evolution, with implications for theoretical biology, genetic modification technology, and public health. This article is protected by copyright. All rights reserved.

Co-option of the same ancestral gene family gave rise to mammalian and reptilian toxins

Background

Evolution can occur with surprising predictability when organisms face similar ecological challenges. For most traits, it is difficult to ascertain whether this occurs due to constraints imposed by the number of possible phenotypic solutions or because of parallel responses by shared genetic and regulatory architecture. Exceptionally, oral venoms are a tractable model of trait evolution, being largely composed of proteinaceous toxins that have evolved in many tetrapods, ranging from reptiles to mammals. Given the diversity of venomous lineages, they are believed to have evolved convergently, even though biochemically similar toxins occur in all taxa.

Results

Here, we investigate whether ancestral genes harbouring similar biochemical activity may have primed venom evolution, focusing on the origins of kallikrein-like serine proteases that form the core of most vertebrate oral venoms. Using syntenic relationships between genes flanking known toxins, we traced the origin of kallikreins to a single locus containing one or more nearby paralogous kallikrein-like clusters. Additionally, phylogenetic analysis of vertebrate serine proteases revealed that kallikrein-like toxins in mammals and reptiles are genetically distinct from non-toxin ones.

Conclusions

Given the shared regulatory and genetic machinery, these findings suggest that tetrapod venoms evolved by co-option of proteins that were likely already present in saliva. We term such genes ‘toxipotent’—in the case of salivary kallikreins they already had potent vasodilatory activity that was weaponized by venomous lineages. Furthermore, the ubiquitous distribution of kallikreins across vertebrates suggests that the evolution of envenomation may be more common than previously recognized, blurring the line between venomous and non-venomous animals.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12915-021-01191-1.

A System for the Evolution of Protein-Protein Interaction Inducers

Molecules that induce interactions between proteins, often referred to as "molecular glues", are increasingly recognized as important therapeutic modalities and as entry points for rewiring cellular signaling networks. Here, we report a new PACE-based method to rapidly select and evolve molecules that mediate interactions between otherwise noninteracting proteins: rapid evolution of protein-protein interaction glues (rePPI-G). By leveraging proximity-dependent split RNA polymerase-based biosensors, we developed E. coli-based detection and selection systems that drive gene expression outputs only when interactions between target proteins are induced. We then validated the system using engineered bivalent molecular glues, showing that rePPI-G robustly selects for molecules that induce the target interaction. Proof-of-concept evolutions demonstrated that rePPI-G reduces the "hook effect" of the engineered molecular glues, due at least in part to tuning the interaction affinities of each individual component of the bifunctional molecule. Altogether, this work validates rePPI-G as a continuous, phage-based evolutionary technology for optimizing molecular glues, providing a strategy for developing molecules that reprogram protein-protein interactions.