High-throughput directed evolution: a golden era for protein science

Abstract

Different in vivo diversification methods are emerging to enable targeted mutagenesis of both genomic and plasmid DNA.Split protein biosensors can be utilised for directed evolution of optimal biophysical properties in proteins relevant to medicine and biotechnology.Innovative screening technologies are emerging for in vivo continuous evolution to enable engineering of proteins with enhanced biophysical properties, as well as new or improved functions.Screening technologies can be utilised to enable high-throughput identification of beneficial mutants, as well as to study the relationship between sequence and function for disease- and biopharmaceutically-relevant proteins.Screens can produce rich datasets that can inform computational algorithms to better understand and predict the fundamental mechanisms governing protein behaviour and thereby facilitate protein design. Directed evolution is a robust and powerful tool for engineering new and/or improved functions in biomolecules for therapeutic and industrial applications, as well as to uncover fundamental insights into protein behaviour. It works by exploiting the principle of natural evolution and accelerating it through multiple rounds of gene diversification and selection. To evolve the desired property, an appropriate assay for the property of interest must be chosen. Here, we describe recent advances in the development of in vitro and in vivo diversification methods, as well as high-throughput assays for protein directed evolution. Using recent examples, we discuss the drawbacks and challenges of the array of diversification methods and selection assays and consider future challenges in the field. Directed evolution is a robust and powerful tool for engineering new and/or improved functions in biomolecules for therapeutic and industrial applications, as well as to uncover fundamental insights into protein behaviour. It works by exploiting the principle of natural evolution and accelerating it through multiple rounds of gene diversification and selection. To evolve the desired property, an appropriate assay for the property of interest must be chosen. Here, we describe recent advances in the development of in vitro and in vivo diversification methods, as well as high-throughput assays for protein directed evolution. Using recent examples, we discuss the drawbacks and challenges of the array of diversification methods and selection assays and consider future challenges in the field. Over the past 3.5 billion years, organisms have been adapting and evolving to increase their competitiveness. Many cellular processes are carried out by proteins for which evolution has generated functionality that is often beyond our current ability to rationally design. Consequently, protein engineers have been working to exploit and expedite Nature’s evolutionary processes to evolve and improve different protein functions since the advent of recombinant DNA technology in the 1970s (Figure 1). Directed evolution utilises the principles of Darwinian evolution, whereby genetic diversity is introduced into the test protein, which is then subjected to a selective pressure (Figure 2A ). Compared with natural evolution, directed evolution has higher mutation rates to accelerate the process. By using an appropriate genotype–phenotype screen, rare beneficial mutations are enriched and can be identified.Figure 2Directed evolution and in vitro mutagenesis.Show full caption(A) A directed evolution experiment works by creating a library of gene variants of a protein-of-interest and subjecting them to a selective pressure to identify beneficial mutations. (B) Error-prone PCR (epPCR) uses an error-prone DNA polymerase (EP DNAP) to amplify a gene of interest and introduce mutations. Alternatively, the buffer conditions can be modified to increase the mutation rate of a standard DNAP, such as by adding magnesium ions (pink) or having unbalanced dNTP concentrations (A, green; T, orange; G, yellow; C, blue). (C) DNA shuffling allows mixing of homologous sequences, such as variants of the same protein with single point mutants, to create hybrid genes combining different mutations. Libraries are created by random fragmentation of genes, which are then joined together using primer-free PCR.View Large Image Figure ViewerDownload Hi-res image Download (PPT) (A) A directed evolution experiment works by creating a library of gene variants of a protein-of-interest and subjecting them to a selective pressure to identify beneficial mutations. (B) Error-prone PCR (epPCR) uses an error-prone DNA polymerase (EP DNAP) to amplify a gene of interest and introduce mutations. Alternatively, the buffer conditions can be modified to increase the mutation rate of a standard DNAP, such as by adding magnesium ions (pink) or having unbalanced dNTP concentrations (A, green; T, orange; G, yellow; C, blue). (C) DNA shuffling allows mixing of homologous sequences, such as variants of the same protein with single point mutants, to create hybrid genes combining different mutations. Libraries are created by random fragmentation of genes, which are then joined together using primer-free PCR. Advances in molecular biology, such as the discovery of restriction endonucleases and the invention of PCR, have facilitated the specific and rational engineering of proteins (Figure 1). These advances have permitted the study of proteins by enabling the effect of specific amino acid substitutions on a protein’s biological function and stability to be explored, as well as allowing the creation of proteins with improved biophysical properties [1.Ebo J.S. et al.An in vivo platform to select and evolve aggregation-resistant proteins.Nat. Commun. 2020; 11: 1816Crossref PubMed Scopus (11) Google Scholar, 2.Bolognesi B. et al.The mutational landscape of a prion-like domain.Nat. Commun. 2019; 10: 4162Crossref PubMed Scopus (12) Google Scholar, 3.Newberry R.W. et al.Deep mutational scanning reveals the structural basis for α-synuclein activity.Nat. Chem. Biol. 2020; 16: 653-659Crossref PubMed Scopus (22) Google Scholar, 4.Wang T. et al.Continuous directed evolution of proteins with improved soluble expression.Nat. Chem. Biol. 2018; 14: 972-980Crossref PubMed Scopus (45) Google Scholar, 5.Chen K. Arnold F.H. Tuning the activity of an enzyme for unusual environments: sequential random mutagenesis of subtilisin E for catalysis in dimethylformamide.Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 5618-5622Crossref PubMed Google Scholar]. Over the past 50 years, there have been enormous advances in molecular biology and DNA sequencing technologies (Figure 1), allowing the field to advance from experiments limited to assessing the effect of a single amino acid substitution in a single protein-of-interest (POI), to deep mutational scanning (DMS), which allows simultaneous characterisation of millions of variants. For small proteins (e.g., up to ~150 residues in length), DMS allows measurement of the functional consequence of every possible amino acid substitution at every position of the protein sequence in a single experiment [2.Bolognesi B. et al.The mutational landscape of a prion-like domain.Nat. Commun. 2019; 10: 4162Crossref PubMed Scopus (12) Google Scholar,3.Newberry R.W. et al.Deep mutational scanning reveals the structural basis for α-synuclein activity.Nat. Chem. Biol. 2020; 16: 653-659Crossref PubMed Scopus (22) Google Scholar,6.Seuma M. et al.The genetic landscape for amyloid beta fibril nucleation accurately discriminates familial Alzheimer’s disease mutations.eLife. 2021; 10e63364Crossref PubMed Scopus (11) Google Scholar]. Directed evolution has been used to engineer proteins with improved functional and biophysical properties (catalytic activity [7.Roth T.B. et al.Phage-assisted evolution of Bacillus methanolicus methanol dehydrogenase 2.ACS Synth. Biol. 2019; 8: 796-806Crossref PubMed Scopus (40) Google Scholar], binding affinity [8.Julian M.C. et al.Nature-inspired design and evolution of anti-amyloid antibodies.J. Biol. Chem. 2019; 294: 8438-8451Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar], thermal [9.Pavoor T.V. et al.An enhanced approach for engineering thermally stable proteins using yeast display.Protein Eng. Des. Sel. 2012; 25: 625-630Crossref PubMed Scopus (10) Google Scholar], and thermodynamic [10.Ren C. et al.An enzyme-based biosensor for monitoring and engineering protein stability in vivo.Proc. Natl. Acad. Sci. U. S. A. 2021; 118e2101618118Crossref Scopus (6) Google Scholar] stability); to evolve novel functions [11.Esvelt K.M. et al.A system for the continuous directed evolution of biomolecules.Nature. 2011; 472: 499-503Crossref PubMed Scopus (357) Google Scholar]; to enable characterisation and understanding of the function and stability of natural proteins [2.Bolognesi B. et al.The mutational landscape of a prion-like domain.Nat. Commun. 2019; 10: 4162Crossref PubMed Scopus (12) Google Scholar,3.Newberry R.W. et al.Deep mutational scanning reveals the structural basis for α-synuclein activity.Nat. Chem. Biol. 2020; 16: 653-659Crossref PubMed Scopus (22) Google Scholar]; as well as to engineer entire organisms [12.Mans R. et al.Under pressure: evolutionary engineering of yeast strains for improved performance in fuels and chemicals production.Curr. Opin. Biotechnol. 2018; 50: 47-56Crossref PubMed Scopus (9) Google Scholar]. These techniques have subsequently allowed research to extend beyond the confines of Nature; the directed evolution of novel tRNAs has enabled engineering of proteins beyond the 20 canonical amino acids [13.Chin J.W. et al.Addition of a photocrosslinking amino acid to the genetic code of Escherichia coli.Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 11020-11024Crossref PubMed Scopus (501) Google Scholar, 14.DeBenedictis E.A. et al.Systematic molecular evolution enables robust biomolecule discovery.Nat. Methods. 2022; 19: 55-64Crossref PubMed Scopus (2) Google Scholar, 15.Wang L. et al.Expanding the genetic code of Escherichia coli.Science. 2001; 292: 498-500Crossref PubMed Google Scholar, 16.Park H.-S. et al.Expanding the genetic code of Escherichia coli with phosphoserine.Science. 2011; 333: 1151-1154Crossref PubMed Scopus (246) Google Scholar, 17.Reddington S.C. et al.Directed evolution of GFP with non-natural amino acids identifies residues for augmenting and photoswitching fluorescence.Chem. Sci. 2015; 6: 1159-1166Crossref PubMed Google Scholar, 18.Dumas A. et al.Designing logical codon reassignment – expanding the chemistry in biology.Chem. Sci. 2015; 6: 50-69Crossref PubMed Google Scholar], and the evolution of quadruplet tRNAs has the potential to further expand the genetic code with up to 255 unique amino acids possible (in principle) [14.DeBenedictis E.A. et al.Systematic molecular evolution enables robust biomolecule discovery.Nat. Methods. 2022; 19: 55-64Crossref PubMed Scopus (2) Google Scholar,19.DeBenedictis E.A. et al.Multiplex suppression of four quadruplet codons via tRNA directed evolution.Nat. Commun. 2021; 12: 5706Crossref PubMed Scopus (4) Google Scholar]. Furthermore, directed evolution has been invaluable in aiding in the optimisation of computationally designed proteins [20.Bunzel H.A. et al.Emergence of a negative activation heat capacity during evolution of a designed enzyme.J. Am. Chem. Soc. 2019; 141: 11745-11748Crossref PubMed Scopus (23) Google Scholar, 21.Tinberg C.E. et al.Computational design of ligand-binding proteins with high affinity and selectivity.Nature. 2013; 501: 212-216Crossref PubMed Scopus (279) Google Scholar, 22.Khersonsky O. et al.Optimization of the in-silico-designed kemp eliminase KE70 by computational design and directed evolution.J. Mol. Biol. 2011; 407: 391-412Crossref PubMed Scopus (125) Google Scholar], as well as facilitating de novo protein design pipelines [23.Cao L. et al.De novo design of picomolar SARS-CoV-2 miniprotein inhibitors.Science. 2020; 370: 426-431Crossref PubMed Scopus (178) Google Scholar,24.Chevalier A. et al.Massively parallel de novo protein design for targeted therapeutics.Nature. 2017; 550: 74-79Crossref PubMed Scopus (193) Google Scholar]. In this review, we outline different diversification methods, including in vitro and in vivo approaches, as well as novel selection strategies for evolving proteins with optimised functional and biophysical properties. Throughout we draw on recent examples that exemplify the power of these approaches to answer fundamental biological questions, as well as to address problems in medicine and biotechnology. The first step in any directed evolution experiment is the creation of genetic diversity upon which selection pressures can be applied. Early work developing directed evolution techniques for engineering enzymes in the 1990s used random mutagenesis technologies to create genetic diversity [5.Chen K. Arnold F.H. Tuning the activity of an enzyme for unusual environments: sequential random mutagenesis of subtilisin E for catalysis in dimethylformamide.Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 5618-5622Crossref PubMed Google Scholar,25.Chen K. Arnold F.H. Enzyme engineering for nonaqueous solvents: random mutagenesis to enhance activity of subtilisin E in polar organic media.Bio/Technology. 1991; 9: 1073-1077Crossref PubMed Scopus (0) Google Scholar,26.Currin A. et al.The evolving art of creating genetic diversity: from directed evolution to synthetic biology.Biotechnol. Adv. 2021; 50107762Crossref PubMed Scopus (5) Google Scholar]. Error-prone PCR (epPCR; Box 1) is by far the most popular of these techniques owing to its ease of use (Figure 2B). epPCR is still widely employed and has been used successfully to engineer the properties of proteins, such as to increase aggregation-resistance [1.Ebo J.S. et al.An in vivo platform to select and evolve aggregation-resistant proteins.Nat. Commun. 2020; 11: 1816Crossref PubMed Scopus (11) Google Scholar], to improve enzyme activity [27.Nearmnala P. et al.An in vivo selection system with tightly regulated gene expression enables directed evolution of highly efficient enzymes.Sci. Rep. 2021; 11: 11669Crossref PubMed Scopus (1) Google Scholar], to evolve complex proteins useful for biotechnology [28.Tetter S. et al.Evolution of a virus-like architecture and packaging mechanism in a repurposed bacterial protein.Science. 2021; 372: 1220-1224Crossref PubMed Scopus (13) Google Scholar], and to determine protein fitness landscapes using DMS [6.Seuma M. et al.The genetic landscape for amyloid beta fibril nucleation accurately discriminates familial Alzheimer’s disease mutations.eLife. 2021; 10e63364Crossref PubMed Scopus (11) Google Scholar,10.Ren C. et al.An enzyme-based biosensor for monitoring and engineering protein stability in vivo.Proc. Natl. Acad. Sci. U. S. A. 2021; 118e2101618118Crossref Scopus (6) Google Scholar]. Another often-used method for in vitro gene diversification is DNA shuffling (Box 1), wherein libraries are created by random fragmentation and recombination of homologous DNA sequences (Figure 2C) [29.Stemmer W.P. Rapid evolution of a protein in vitro by DNA shuffling.Nature. 1994; 370: 389-391Crossref PubMed Scopus (1582) Google Scholar]. Since its invention, DNA shuffling has been widely used and adapted to engineer a range of properties, including improved thermostability [30.Hao J. Berry A. A thermostable variant of fructose bisphosphate aldolase constructed by directed evolution also shows increased stability in organic solvents.Protein Eng. Des. Sel. 2004; 17: 689-697Crossref PubMed Scopus (80) Google Scholar] or catalytic activity [27.Nearmnala P. et al.An in vivo selection system with tightly regulated gene expression enables directed evolution of highly efficient enzymes.Sci. Rep. 2021; 11: 11669Crossref PubMed Scopus (1) Google Scholar], and, most recently, to develop chemogenetic fluorescent reporters with tuneable fluorescent properties [31.Benaissa H. et al.Engineering of a fluorescent chemogenetic reporter with tunable color for advanced live-cell imaging.Nat. Commun. 2021; 12: 6989Crossref PubMed Scopus (4) Google Scholar].Box 1Traditional approaches to create librariesError-prone PCR (epPCR) works by using an error-prone DNA polymerase (DNAP) to randomly generate mutations during PCR amplification, or by modifying the buffer components to decrease the fidelity of a standard DNAP (see Figure 2B in main text) [97.Leung D. et al.A method for random mutagenesis of a defined DNA segment using a modified polymerase chain reaction.Technique. 1989; 1: 11-15Google Scholar,98.Wang Y. et al.Directed evolution: methodologies and applications.Chem. Rev. 2021; 121: 12384-12444Crossref PubMed Scopus (20) Google Scholar]. Reaction components can be modified to increase the mutation rate, such as using unbalanced dNTP concentrations, increasing the concentration of magnesium ions, increasing the number of PCR cycles, or adding manganese ions [98.Wang Y. et al.Directed evolution: methodologies and applications.Chem. Rev. 2021; 121: 12384-12444Crossref PubMed Scopus (20) Google Scholar]. However, epPCR has limitations: often the DNAP has a bias for certain nucleotide substitutions over others, which can affect the amino acids available for a particular codon and, as a result, there is high chance for synonymous substitutions, therefore reducing library diversity. Mutations acquired in early PCR cycles can become dominant in the library compared with those acquired in later cycles, thereby biasing the library towards mutants acquired in early cycles. Furthermore, in epPCR, consecutive mutation of two bases is rare, which can further reduce the possible amino acids available; it requires large amounts of screening in order to sample the entire library; and can result in insertions and deletions (although at low frequencies), as well as the introduction of stop codons [26.Currin A. et al.The evolving art of creating genetic diversity: from directed evolution to synthetic biology.Biotechnol. Adv. 2021; 50107762Crossref PubMed Scopus (5) Google Scholar,98.Wang Y. et al.Directed evolution: methodologies and applications.Chem. Rev. 2021; 121: 12384-12444Crossref PubMed Scopus (20) Google Scholar].DNA shuffling makes use of fragmentation and recombination of homologous genes (see Figure 2C in main text) [29.Stemmer W.P. Rapid evolution of a protein in vitro by DNA shuffling.Nature. 1994; 370: 389-391Crossref PubMed Scopus (1582) Google Scholar]. Genes are fragmented using DNase I and recombined using primer-free PCR where fragments with sufficient overlap will anneal to each other and be amplified. This approach is especially useful for mixing and combining a library of mutants that have already been evolved and selected as beneficial, in order to combine advantageous characteristics and improve them further. Error-prone PCR (epPCR) works by using an error-prone DNA polymerase (DNAP) to randomly generate mutations during PCR amplification, or by modifying the buffer components to decrease the fidelity of a standard DNAP (see Figure 2B in main text) [97.Leung D. et al.A method for random mutagenesis of a defined DNA segment using a modified polymerase chain reaction.Technique. 1989; 1: 11-15Google Scholar,98.Wang Y. et al.Directed evolution: methodologies and applications.Chem. Rev. 2021; 121: 12384-12444Crossref PubMed Scopus (20) Google Scholar]. Reaction components can be modified to increase the mutation rate, such as using unbalanced dNTP concentrations, increasing the concentration of magnesium ions, increasing the number of PCR cycles, or adding manganese ions [98.Wang Y. et al.Directed evolution: methodologies and applications.Chem. Rev. 2021; 121: 12384-12444Crossref PubMed Scopus (20) Google Scholar]. However, epPCR has limitations: often the DNAP has a bias for certain nucleotide substitutions over others, which can affect the amino acids available for a particular codon and, as a result, there is high chance for synonymous substitutions, therefore reducing library diversity. Mutations acquired in early PCR cycles can become dominant in the library compared with those acquired in later cycles, thereby biasing the library towards mutants acquired in early cycles. Furthermore, in epPCR, consecutive mutation of two bases is rare, which can further reduce the possible amino acids available; it requires large amounts of screening in order to sample the entire library; and can result in insertions and deletions (although at low frequencies), as well as the introduction of stop codons [26.Currin A. et al.The evolving art of creating genetic diversity: from directed evolution to synthetic biology.Biotechnol. Adv. 2021; 50107762Crossref PubMed Scopus (5) Google Scholar,98.Wang Y. et al.Directed evolution: methodologies and applications.Chem. Rev. 2021; 121: 12384-12444Crossref PubMed Scopus (20) Google Scholar]. DNA shuffling makes use of fragmentation and recombination of homologous genes (see Figure 2C in main text) [29.Stemmer W.P. Rapid evolution of a protein in vitro by DNA shuffling.Nature. 1994; 370: 389-391Crossref PubMed Scopus (1582) Google Scholar]. Genes are fragmented using DNase I and recombined using primer-free PCR where fragments with sufficient overlap will anneal to each other and be amplified. This approach is especially useful for mixing and combining a library of mutants that have already been evolved and selected as beneficial, in order to combine advantageous characteristics and improve them further. Targeted gene mutagenesis methods have been developed to overcome the limitations of classic random mutagenesis methods and have been reviewed at length elsewhere [26.Currin A. et al.The evolving art of creating genetic diversity: from directed evolution to synthetic biology.Biotechnol. Adv. 2021; 50107762Crossref PubMed Scopus (5) Google Scholar]. In short, recent advances in solid-phase DNA synthesis methods allow tight control over designed libraries to eliminate amino acid bias, allowing the effect of defined sets of amino acid substitutions in focussed regions of interest [e.g., the complementarity determining regions (CDRs) of antibodies] or the entire primary sequence to be determined [26.Currin A. et al.The evolving art of creating genetic diversity: from directed evolution to synthetic biology.Biotechnol. Adv. 2021; 50107762Crossref PubMed Scopus (5) Google Scholar]. Such libraries are particularly useful for DMS experiments as they allow understanding of protein functional landscapes and can be used to uncover the contribution of the identity (e.g., amino acid side chain chemistry) and individual residues to protein function, stability, and/or aggregation [32.Fowler D.M. Fields S. Deep mutational scanning: a new style of protein science.Nat. Methods. 2014; 11: 801-807Crossref PubMed Scopus (440) Google Scholar]. In vivo mutagenesis approaches involve altering the genome sequence of an organism via the addition of mutagens (such as chemicals or UV light), or the use of hypermutator strains that contain deletions or modifications in genes for enzymes involved in proofreading, mismatch-repair, and base-excision (such as XL1-Red) [33.Badran A.H. Liu D.R. Development of potent in vivo mutagenesis plasmids with broad mutational spectra.Nat. Commun. 2015; 6: 8425Crossref PubMed Scopus (0) Google Scholar,34.Greener A. et al.An efficient random mutagenesis technique using an E. coli mutator strain.Mol. Biotechnol. 1997; 7: 189-195Crossref PubMed Google Scholar]. Alternatively, various examples of mutagenic plasmids expressing different mutagenic enzymes involved in mismatch repair, translesion synthesis, and proof-reading have been developed with a wide range of induced mutagenic potency to globally increase the mutation rate in Escherichia coli [33.Badran A.H. Liu D.R. Development of potent in vivo mutagenesis plasmids with broad mutational spectra.Nat. Commun. 2015; 6: 8425Crossref PubMed Scopus (0) Google Scholar]. These strategies have the potential to yield high mutation rates (up to 322 000-fold over wild type E. coli). Such methods can be problematic as the accumulation of mutations throughout the E. coli genome can result in toxic mutations if they occur within essential regions of the genome. Alternatively, these mutations accumulating outside of the gene-of-interest (GOI) could allow the bacteria to circumvent the selection pressure. To overcome these limitations, targeted in vivo mutagenesis strategies have been developed. An early example of this strategy is the use of a mutated E. coli polymerase I (pol I) that selectively mutates genes on a ColE1 plasmid (although mutations are limited to within a few kb of the ColE1 origin) [35.Allen J.M. et al.Roles of DNA polymerase I in leading and lagging-strand replication defined by a high-resolution mutation footprint of ColE1 plasmid replication.Nucleic Acids Res. 2011; 39: 7020-7033Crossref PubMed Scopus (19) Google Scholar,36.Camps M. et al.Targeted gene evolution in Escherichia coli using a highly error-prone DNA polymerase I.Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 9727-9732Crossref PubMed Scopus (112) Google Scholar]. Furthermore, pol I still replicates parts of the genome, which can result in off-target mutations [35.Allen J.M. et al.Roles of DNA polymerase I in leading and lagging-strand replication defined by a high-resolution mutation footprint of ColE1 plasmid replication.Nucleic Acids Res. 2011; 39: 7020-7033Crossref PubMed Scopus (19) Google Scholar]. A popular method of in vivo mutagenesis is fusing specific DNA binding proteins to DNA-mutating enzymes. An example of this is MutaT7, wherein a cytidine deaminase is fused to T7 RNA polymerase (RNAP) to continuously direct mutations to specific, well-defined, DNA regions of any length in E. coli [37.Moore C.L. et al.A processive protein chimera introduces mutations across defined DNA regions in vivo.J. Am. Chem. Soc. 2018; 140: 11560-11564Crossref PubMed Scopus (34) Google Scholar]. This allows targeted mutagenesis of genes under the control of the T7 promoter (Figure 3A ). However, this approach has the potential to accumulate off-target effects, which can be problematic, particularly in the promoter regions. For example, they can potentially inhibit expression of the GOI, or lead to escape mutations, which allow the cells to evade the selection pressure applied without evolving the GOI. Furthermore, as this method utilises cytidine deaminases, their specific activity is limited to C>T and G>A mutations. Alternative cytidine deaminases have been employed to increase the mutation rate and expand the applicability of this method [38.Park H. Kim S. Gene-specific mutagenesis enables rapid continuous evolution of enzymes in vivo.Nucleic Acids Res. 2021; 49e32Crossref Scopus (8) Google Scholar], and MutaT7 has also been adapted for use in eukaryotic cells (TRACE; T7 polymerase-driven continuous editing) [39.Chen H. et al.Efficient, continuous mutagenesis in human cells using a pseudo-random DNA editor.Nat. Biotechnol. 2020; 38: 165-168Crossref PubMed Scopus (27) Google Scholar]. A similar method (EvolvR), developed for use in both yeast and bacteria, utilises a fusion between an error-prone DNA polymerase (DNAP) and a nickase-Cas9 (nCas9), which allows mutations within a region adjacent to the Cas9 nick site (Figure 3B) [40.Halperin S.O. et al.CRISPR-guided DNA polymerases enable diversification of all nucleotides in a tunable window.Nature. 2018; 560: 248-252Crossref PubMed Scopus (133) Google Scholar,41.Tou C.J. et al.Targeted diversification in the S. cerevisiae genome with CRISPR-guided DNA polymerase I.ACS Synth. Biol. 2020; 9: 1911-1916Crossref PubMed Scopus (0) Google Scholar]. The mutation rate can be tuned by using polymerases with different fidelities (~10–7–10–3 per base) and this method enables all possible nucleotide substitutions, unlike those utilising cytidine deaminases. The approach is limited due to elevated (~1011–108 per base) off-target mutation rates and the narrow mutation window within the sequence (most mutations occur within 50 bp of the nick site). T7-targeted dCas9-limited in vivo mutagenesis (T7-DIVA) utilises a similar method whereby T7 RNAP fused to a cytidine deaminase is used to introduce mutations (Figure 3C) [42.Álvarez B. et al.In vivo diversification of target genomic sites using processive base deaminase fusions blocked by dCas9.Nat. Commun. 2020; 11: 6436Crossref PubMed Scopus (12) Google Scholar]. The GOI can remain under the control of its genomic promoter and a T7 promoter is inserted downstream of the GOI on the antisense strand. This allows the T7 RNAP to translocate along the GOI and to introduce mutations without altering the endogenous 5′ promoter. A catalytically dead Cas9 (dCas9) is used as a ‘roadblock’ demarcating the boundaries of the mutagenesis, enabling targeted in vivo mutagenesis of specific genes. However, as this method requires introduction of a downstream T7 promoter, it is unable to mutate specific regions of a GOI. Error-prone DNA replication utilising the native yeast retrotransposon (see Glossary) Ty1 has been developed for selective mutation of genes inserted between long terminal repeats (Figure 3D) [43.Crook N. et al.In vivo continuous evolution of genes and pathways in yeast.Nat. Commun. 2016; 7: 13051Crossref PubMed Scopus (66) Google Scholar]. The replication cycle of Ty1 occurs via an RNA intermediate that is converted into complementary DNA through an encoded reverse transcriptase and re-integrated back into the genome. Heterologous gene expression from Ty1 has previously been demonstrated and the replication cycle has been shown to be error-prone [43.Crook N. et al.In vivo continuous evolution of genes and pathways in yeast.Nat. Commun. 2016; 7: 13051Crossref PubMed