
Random mutations and natural selection are the fundamental principles of evolution and adaptation to changing environments. The ability to diversify user-defined genomic loci has recently enabled the swift introduction of genetic variations in iterative cycles. Screening of the resulting mutant libraries through selection pressure, called directed evolution (DE), has allowed the development of genetic variants with novel functions (1). While both natural selection and DE are based on similar principles, the process of DE is accelerated by utilizing mutagenesis tools in a fast-forward manner. The spontaneous mutation rate in bacteria, plant, and human cells ranges from 10−10 to 10−9 per nucleotide base in DNA replication, which requires many generations to attain a mutant with desired gain-of-function (2). On the other hand, iterative rounds of generating mutants in DE accelerate the evolutionary process. DE has been successfully applied to obtain proteins/enzymes with novel functions for basic research, to unveil biomolecules with clinical significance, and to improve plant traits (1, 3-5).
Mutant libraries can be generated through in vitro and in vivo methods (4, 6). In vitro DE approaches need iterative cycles of gene diversification, laborious screening, and selection steps. On the other hand, in vivo continuous DE can avoid several repetitive actions (like cloning and genetic transformation); consequently enabling simultaneous hypermutation and screening of mutants with favorable phenotypes. Although in vivo DE approaches can save time and are cost-effective, they require efficient means for the rapid induction of mutations. Thus, mutagenic agents capable of targeted hypermutagenesis (without altering the rest of the genome) are necessary to rapidly achieve the desired locus. Conventional mutagenic agents, such as radiation and chemicals, induce random mutations throughout the genome, which can be toxic to the host cells and may result in false screening during selection due to global mutations (7). With the improved understanding of genetics and expanded molecular biology toolbox, early targeted random mutagenesis (TRM) tools address some of these challenges. However, they rely on heterologous hosts with nonnative environments that lack natural hosts’ physiological and genetic state (8, 9).
To overcome these challenges, researchers have been exploring the potential of DNA deaminases-based base editing (BE) tools, which are developed using engineered CRISPR/Cas system and DNA deaminases (10). Cas endonuclease (e.g., Cas9) and single guide RNA (sgRNA; chimeric crRNA-tracrRNA) are the two primary components of the CRISPR/Cas system. Cas effector anchored by sgRNA recognizes and binds to the target DNA site preceded by protospacer adjacent motif (PAM) and induces double-stranded breaks (DSBs). Error-prone repair of DSBs leads to disruption of encoded DNA functions. BE tools comprising a fusion of deaminase with partial (nickase, nCas9) or fully (dead form, dCas9) nuclease deficient Cas9 can diversify targeted DNA sites without requiring donor DNA templates or introducing DSBs, making it ideal for in vivo DE studies (11). The BE tools offer the advantage of the ability to perform nucleobase substitutions unlike the homology-directed repair (HDR)-based techniques, which tend to have a higher rate of indels than precise base changes and lower efficiencies (2).
Since the development of single-BEs (12-14), there has been notable progress in expanding their editing capabilities and targeting scope, including use in diversifying nucleotides at targeted loci. Specifically, newer BE-TRM tools are developed by harnessing the base substitution ability of engineered deaminases. BEs possess several features that make them suitable for engineering novel BE-TRM suites for DE studies. For instance, the editable window (EW, DNA length accessible for nucleotide substitution) of these tools can be intentionally widened or narrowed (15). Moreover, BE-TRM tools allow simultaneous sequence diversification and selection in vivo (1, 14, 15). Furthermore, the fusion of deaminases with other components (e.g., Cas9, deaminases, DNA-repair proteins, or polymerases) has been demonstrated to be compatible in a modular fashion without compromising the functionality of either the deaminase or fused component (16).
Although there are other CRISPR/Cas-based systems with the potential for targeted mutagenesis, such as CRISPR-X (17), EvolvR (18), Target-G (19), ROS-mutator (20), DNA integrators (21, 22) and prime editing (23), some of these techniques are reviewed elsewhere (1, 4, 6, 24) and not included in this work. Instead, we focus on the recent advances in DNA deaminase-based BE-TRM tools. We summarize the emerging BE-TRM techniques for in vivo DE, their applications, and future directions for designing improved BE-TRM tools.
BE tools are analogous to CRISPR/Cas system, enabling the diversification of targeted DNA sequences. Typical single-BEs allow modification of one type of nucleotide at a time. Researchers have engineered dual BEs (DuBEs), that allow concurrent conversion of more than one type of nucleotide, by fusing multiple deaminases. The fusion of mobile elements, such as polymerase enzymes, to BE components has led to the development of a new class of BE-TRM tools, which adds more versatility to the possible genetic diversification of targeted sites. In this section, we summarize the design and development of these tools (Table 1) focusing on their use for DE.
The general architecture of the BE tool consists of a fusion of a Cas9 variant, DNA deaminase, and in some cases, DNA repair factors (e.g., uracil N-glycosylase inhibitor, UGI) to achieve efficient editing (Fig. 1A) (25). Single-BEs can be categorized into three major classes, cytosine BEs (CBEs), adenine BEs (ABEs), and glycosylase BEs (CGBEs), based on their targeted nucleotides and subsequent mutations (Fig. 1B-D).
The first CRISPR/Cas9-targeted deamination was demonstrated in 2016. Two independent research groups developed CBEs that perform C-to-U conversion as intermediate state and thereby C-to-T editing using different cytidine deaminases, rat deaminase APOBEC1 (rAPOBEC1) and PmCDA1 from sea lamprey, named BE3 (13) and Target-AID (12), respectively. Both tools can access slightly different EW lengths (Table 1). The CBEs increase the frequency of C-to-T substitution, especially after the suppressing uracil DNA N-glycosylase (UNG) activity by gene deletion or by specific inhibitors like UGI. If the U-containing intermediate is exposed to the base excision, the outcome is often a C-to-G conversion rather than a C-to-T. This aspect was exploited to develop CGBEs by modifying the DNA repair mechanisms (26-32).
ABE is another primary BE tool first engineered using laboratory-evolved TadA mutant (TadA*) to perform A-to-Inosine (I) conversion, which DNA repair machinery read as G, ultimately producing A-to-G editing (14). As no a known naturally occurring enzyme catalyzes the deoxyadenosine deamination in DNA, all the available ABEs use TadA-evolved adenine deaminases (14, 33-35). Recently, A-to-Y (AYBE) tool was developed by fusing TadA-8e with N-methylpurine DNA glycosylase (MPG) (36), which expands the scope of possible nucleotide diversification.
BEs can diversify the targeted loci but to a limited scope due to shorter EW length and their inability to edit more than one type of nucleotide. However, in some instances, the features of BEs provide multiple avenues to exploit their use in DE, which includes adjustable EW length, simultaneous conversion of multiple nucleotides, and randomized outcomes from a single nucleobase (e.g., A-to-Y). For instance, EW length can be altered when a different Cas domain is used (e.g., circular permutant Cas9, Cas9 variants/orthologs), when the deaminase domain changes (variants/orthologs of deaminases), or when modifying the linker length between Cas-deaminase (34). Natural or engineered deaminases with variable substrate preferences and kinetic parameters exhibit different EW widths. In addition, recruiting higher doses of deaminase at a target site (e.g., MS2 aptamer system in CRISPR-X or Suntag in BE-PLUS) broadens the EW length (15). Thus, combining the rapidly expanding suite of deaminase and Cas enzymes might enhance the capabilities of BEs for DE studies.
Any newer deaminase would be an excellent starting material for designing DuBEs or related BE-TRM tools (Fig. 1E). The first-generation DuBEs, also called ACBEs, consisted of ABE (ABE7.10 version) and CBE components fused to install both C-to-T and A-to-G edits (37-41). Some ACBEs consisted of a fusion of different deaminases at the opposite end of nCas9 (38, 39, 41) and some contained the N-terminal fusion of both deaminases (37, 40). In all the DuBEs, TadA fusion to the N-terminus was preferred since C-terminal fusion was reported to lack activity (40). The ABE7.10 version converts A-to-G with relatively low efficiency, reducing the synchronized BE events’ frequency.
In the second-generation DuBEs, ABE7.10 was replaced with high-performing monomeric ABE variants to increase the likelihood of greater simultaneous A-to-G and C-to-T edits. Notably, TadA-8e (ABE8e) adoption in DuBEs showed higher concurrent BE activities in distinctive DuBE architectures developed by various research teams, such as iACBEs (42), ACBE (43), CABE-RY (44), ACEs (5), Dual BEs (45), and hyA&C-BEmax (46) (Table 1). Recently, ACBE variants without the UGI domain or with the UGI domain replaced by uracil-DNA glycosylase (UNG) have shown increased randomization of deaminated Cs into all possible outcomes (C-to-G/T/A) along with A-to-G and indels, termed as AGBEs (47). Overall, expanding the DuBEs toolset provides multiple means for genetic diversification and identification of functional mutants through screening in DE campaigns.
The fusion of different deaminases to Cas protein provides a mode for TRM of desired loci. However, its lack of processivity and limited EW length restricts its broader applicability for DE applications. An alternative, EvolvR, was demonstrated in 2018 using nCas9 fused to error-prone DNA polymerase (DNAP) (18). Although the EvolvR system was shown to introduce mutations up to 350 bp in the targeted DNA region in Escherichia coli, it is not yet reported for targeted hypermutation in eukaryotic cells. In the same year, MutaT7, an in vivo TRM system based on T7 RNA polymerase (T7RNAP), was first designed by tethering rAPOBEC fused to T7RNAP (48). T7RNAP does not need ancillary transcription factors to transcribe DNA. The MutaT7 tool was first reported in E. coli and further employed in yeast (TRIDENT) (49), mammalian cells (TRACE) (8), and plants (50).
Besides, the dCas9-based approach effectively regulated the EW length in the MutaT7 system (7). Also, a more active CBE variant, PmCDA1, improved from 7 to 20-fold mutation frequency (7, 51). Earlier MutaT7 versions comprised only CBE fusions with T7RNAP, thus curtailing only C:G-to-T:A transition mutations (48, 52). To expand the mutational spectra accessible by Muta7-based tools, TadA-based ABE7.10 (7, 53), yeTadA1.0 (49), and TadA-8e (53, 54) were fused to T7RNAP, demonstrating the utility of chimera for A:T-to-G:C substitutions. Recently, Seo and coworkers reported the combined use of PmCDA1 and TadA-8e in the eMutaT7transition approach (54). They fused these two deaminases separately to T7RNAP to install C:G-to-T:A and A:T-to-G:C mutations. Overall, adenosine and cytidine deaminase linked to T7RNAP can facilitate all transition mutations simultaneously, an approach that broadens the possible mutation spectra for DE studies (Fig. 1F). In the future, continued expansion and design of novel polymerase-based BE-TRM tools may enable the induction of all potential transition and transversion mutations, and further allow the use of these tools across all types of cells.
The various editing outcomes generated by BE-TRM tools enable precision editing and sequence diversification. Although each arsenal of the BE-TRM toolbox is capable of having diverse applications in different organisms (Fig. 2), herein, we categorize them into three broad categories: basic research for medical uses, agriculture, and synthetic biology. In this section, we highlight some of these aspects that may also intersect among these categories.
The use of BE-TRM tools that can directly modify targeted DNA sequences is highly desirable than random mutagenic agents. Single-BE tools are valuable for installing precise base mutations. Conversely, establishing a saturated mutagenic population for a specific gene through DE is an efficient approach for genotype-phenotype mapping, which can help identify the relationships between mutations and their corresponding phenotypes. In vivo continuous DE using cellular systems allows researchers to identify potential mutational pathways and test the effects of mutations on resistance emergence in a short time frame, accelerating the design of new drugs. Identifying drug-resistant mutations in clinically relevant targets is crucial for drug design or gene therapy and also for providing insights into evolutionary principles. Table 1 lists several examples, such as rpoB evolution by iACBEs (42), T7-DIVA-mediated evolution of TEM-1 encoding TEM-1 β-lactamase that confers resistance to penicillin, cephalosporin, and related β-lactams (7), and PSMB5 evolution by CRISPR-X, providing novel mutation set responsible for bortezomib resistance (17).
The DuBE (iACBE) toolset was employed to evolve the RNA polymerase rpoB gene implicated in rifampicin resistance (RifR) in E. coli (42). Through mutant library screening, novel rpoB mutants conferring RifR were identified, which may offer potential insights into antibiotic resistance mechanisms. In another study, the AGBE was used to generate a mutant population of the human diphtheria toxin (DT) receptor (hDTR) gene in human cells, enabling the identification of mutants responsible for DT insensitivity (47). These findings hold promise for the design of more precise treatments. Furthermore, MutaT7 (TRIDENT) was employed to evolve Plasmodium falciparum dihydrofolate reductase (PfDHFR) in yeast, which may help to understand the structural aspects of drug resistance (49). Recently, the DuBE (A&C-BEmax) tool was successfully used to introduce desired mutations in the promoter regions of γ-globin genes (HBG1 and HBG2) (40), which may pave the way for therapeutic gene editing treatments for β-thalassemia.
The introduction of valuable traits in agricultural crops has become a necessity in the 21st century to meet the increasing global food demand, especially in the face of changing climates (55). Traits such as resistance to multiple stresses, herbicide tolerance, enhanced yield and quality, and efficient utilization of resources like water and fertilizers, have drawn significant interest in the field of agriculture. However, during the process of domestication, crops have undergone extensive selection, resulting in a reduction in genetic diversity within their gene pool (16). Moreover, naturally, certain crops have limited wild relatives and exhibit fewer genetic variations, which presents challenges for conventional breeding methods aimed at improving crop traits. Therefore, instead of solely relying on natural genetic diversity, targeted random mutagenesis can serve as an initial step in generating a mutant population and can be used in the development of designer crops with superior performance (2).
Herbicide tolerance-related genes have already been targeted in rice using different DE approaches (5, 37, 56). For example, DuBE (ACE5)-mediated evolution strategy was used to gain novel herbicide-resistant allelic variants in acetyl-coenzyme A carboxylase (OsACC) (5). As DE in a plant platform is far less potent than one in a microbial platform, a preferable strategy is to run DE on a plant gene in microbial cells, and then return the improved gene to the plant by genome editing. This strategy was recently demonstrated for the evolution of Arabidopsis arogenate dehydratase (AtADT2) for resistance to feedback inhibition by phenylalanine (57). Evolving plants or animal genes in microbes have several advantages, including short generation times, availability of several efficient BE-TRM tools in microbes, and more powerful evolutionary campaigns (6). In the future, targeting cis-regulatory elements of promoters that regulate gene expression may be an exciting avenue for crop improvement and food security.
In synthetic biology, DE is extensively applied to create or improve biological molecules for various applications, including drug discovery, protein engineering, and metabolic engineering. In particular, the applications of DE tools in synthetic biology range from developing enzymes with improved functions to designing novel proteins, biosensors, and genetic circuits (51, 58). BE-TRM tools provide efficient means to diversify biomolecules of interest through random mutagenesis, followed by iterative rounds of screening or selection to identify variants with the desired properties. By mimicking natural evolution in the laboratory, DE enables the creation of biomolecules with improved function, stability, specificity, and activity (1). In the future, BE-TRM tools would be valuable for exploring the vast sequence space of biomolecules, allowing researchers to harness the power of evolution to solve complex biological problems and create new biological systems with unique properties.
The ideal TRM tools should possess high specificity to reduce off-targets and generate a greater sequence diversity within user-defined DNA segments. Although single-BE tools can generate base substitutions, their narrow EW length demands multiple gRNAs to evolve a targeted locus; consequently, limiting their applicability. In addition, single-component mutation tools exhibit nucleotide or substitution bias, low mutation rates, and generate limited sequence diversity, thereby restricting the fine-tuning of user-intended mutation diversity (18). Notably, DuBE tools provide higher mutation rates and better localized sequence diversity than single-BEs. However, both single and DuBE platforms require Cas effector activity with PAM sequence restrictions. This has been addressed by evolving several Cas9 variants with relaxed PAM recognition, improved efficiency, and reduced off-target activities (34).
In the RNAP-based TRM approach, polymerases with higher processivity are desirable for wider EWs, but it is challenging to restrict the DE of only desired DNA segments or regulate the EW length. Recently, the DNA-bound sgRNA/dCas9 complex was shown to impede elongation by T7 Pol-BE fusions, protecting the downstream DNA region (7). This implies that it is an effective way to control the EW length and concentrate mutagenesis in defined DNA segments. Also, an inducible expression of the TRM molecules reduces potential toxicities associated with the constitutive expression of BE deaminases, T7RNAP, or other accessory enzymes such as UGI. This practical approach is effectively applied to control the TRM process in bacteria and has an excellent potential to be used in eukaryotes whenever possible.
Insertion of the T7 promoter at the targeted locus is a prerequisite for using T7RNAP-based TRM tools. The EvolvR method based on employing error-prone DNA polymerase can be exploited as an alternative method for diversifying target DNA sequences. However, it still lacks temporal regulation over mutation spectra and suffers from low mutation rates (18). In that scenario, designing DNAP or RNAP-based BE-TRM tools with higher mutation rates would serve the purpose of evolving targeted loci under in vivo conditions.
The intracellular DNA repair systems are central to the editing outcome of CRISPR-based tools including BEs. Although most DNA repair pathways are highly conserved across different species, the outcome of DNA repair may vary according to the choice of repair systems utilized by a particular organism. In addition, in vivo continuous DE methods can suffer from off-target mutations throughout the genome, leading to cellular toxicity or false interpretation of results. Alternatively, off-target mutations may allow the cells to circumvent the screening and selection criteria. The major challenge in the continuous evolution field is linking genotype to phenotype. To tackle this issue, the combined use of automation and machine learning at different stages in DE will be instrumental in reducing time and cost unlike manual operations (59), as validated in the PEACE system (58). Indeed, the “design-build-test-learn” cycle operating under in vivo continuous DE conditions automatically manner would enable the evolution of any chosen genetic element.
Most DE studies conducted so far have primarily focused on bacteria and yeasts. However, there is a significant untapped potential to induce targeted genetic diversification in higher organisms, offering a promising opportunity to accelerate DE applications using TRM tools. Nonetheless, it is crucial to develop TRM tools that can precisely and efficiently induce genetic diversity to achieve this objective because genes evolved within a (heterologous) prokaryotic genetic context may not necessarily exhibit their intended functions in the native environment of eukaryotic cells. Several factors, including misfolding and protein aggregation, unexpected intermolecular interactions, or unforeseen modifications, can contribute to this outcome. Therefore, it is imperative to focus on developing efficient TRM tools that can operate within the native cellular environment of multicellular eukaryotes, including humans, animals, and plants. The application of TRM-mediated DE holds immense potential at the intersection of plant biology, synthetic biology, and medicine. For instance, the utilization of BE-TRM tools, alongside other genetic engineering methods, could significantly facilitate the use of plants as hosts in molecular farming, enabling the production of pharmaceutical compounds, enzymes, and other biomolecules.
This work was supported by the National Research Foundation of Korea (grants NRF 2021R1I1A3057067, 2021R1A5A8029490, 2022R1A2C3010331) and the Program for New Plant Breeding Techniques (NBT, PJ01686702), Rural Development Administration (RDA), Korea.
The authors have no conflicting interests.
Major BE and BE-TRM tools and their key features are enlisted. Major versions are summarized in case of multiple combinations tested in a report
Tool | Effectors | Outcome | Organism | Mutation window | Target | Ref. |
---|---|---|---|---|---|---|
Major single-base editors | ||||||
CBE | rAPOBEC1-nCas9-2xUGI | C-to-T | Human cells | 4 to 8 | APOE4 (Alzheimer’s disease) | (13) |
nCas9-PmCDA1-1xUGI | C-to-T | Human cells | 1 to 5 | HPRT (6-thioguanineR) | (12) | |
ABE7.10 | ABE7.10-nCas9 | A-to-G | Human cells | 4 to 8 | HBG1/HBG2 promoters (fetal hemoglobin) | (14) |
ABE8e | (TadA-8e)-nCas9 | A-to-G | Human cells | 4 to 8 | BCL11A and HBG1/HBG2 promoters (fetal hemoglobin) | (33) |
ABE8 versions | ABE8.20m-nCas9 | A-to-G | Human cells | 4 to 8 | HBG1/HBG2 promoters (fetal hemoglobin) | (35) |
CGBEs | eUNG-BE4max(R33A)-nCas9 (CGBE1) | C-to-G** | Human cells | 4 to 8 | - | (26) |
BE4max(R33A)-nCas9 (miniCGBE1) | C-to-G** | Human cells | 4 to 8 | - | ||
GBEs | eUNG-nCas9-AID | C-to-A** | Escherichia coli | - | - | (27) |
APOBEC-nCas9-hUNG | C-to-G** | Human cells | 3 to 7 | - | ||
CGBEs | rAPOBEC-nCas9-rXRCC1 | C-to-G** | Human cells | 2 to 8 | - | (31) |
CGBEs | 10 CGBEs fusing various DNA repair proteins | Mammalian cells | 3 to 11a | - | (30) | |
CGBEs | eUNG-YE1-nCas9 (eOPTI-CGBE), cUNG-YE1-nCas9 (cOPTI-CGBE) | C-to-G** | Mammalian cells, mouse embryos | 4 to 7 | - | (32) |
GBE2.0 | APOBEC(R33A)-nCas9-Rad51-scUNG1 | C-to-G** | Human cells | 3 to 9 | STK11, MEN1, PRNP, TP53, FBLN5 (various disease-related genes) | (29) |
AYBEs | (TadA-8e)-nCas9-hMPG | A-to-Y(C/T) | Human cells | 5 to 9 | - | (36) |
Dual base editors | ||||||
STEMEs | hAPOBEC3A-ABE7.10-nCas9-1xUGI | C-to-T and A-to-G | Rice | C: 1 to 17 A: 4 to 8 |
OsACC (HerbicideR) | (37) |
A&C-BEmax | hAID-ABE7.10-nCas9-2xUGI | C-to-T and A-to-G | Human cells | C: 2 to 15 A: 4 to 8 |
HBG1/HBG2 promoter (β-thalassemia) | (40) |
SPACE | miniABEmax(V82G)- nCas9-PmCDA1-2xUGI | C-to-T and A-to-G | Human cells | C: 2 to 7 A: 4 to 7 |
- | (41) |
Target-ACEmax | ABE7.10-nCas9-PmCDA1-1xUGI | C-to-T and A-to-G | Human cells | C: 1 to 11 A: 4 to 8 |
- | (39) |
ACBE-16N | ABE7.10-(16aa Link)-nCas9-PmCDA1-1xUGI | C-to-T and A-to-G | Human cells | C: 1 to 10 A: 4 to 8 |
- | (38) |
pDuBE1 | (TadA-8e)-nCas9-(LjCDA1L-4)- 3xT2A-1xUGI | C-to-T and A-to-G | Rice | C: 1 to 10 A: 4 to 11 |
OsACC, OsALS1 (HerbicideR) | (56) |
iACBEs | ABE9e-evoCDA1-nCas9-2xUGI (iACBE4) | C-to-T and A-to-G | Escherichia coli | C: -6 to 15 A: 3 to 8 |
rpoB (RifR) | (42) |
ACBE | ABE9e-PmCDA1-nCas9 | C-to-T and A-to-G | Bacillus subtilis | C: 1 to 9 A: 4 to 8 |
PsdB (NisinR) | (43) |
CABE-RY | hA3A(Y130F)-TadA-8e(V106W)- SpRY-2xUGI (CABE4) | C-to-T and A-to-G | Human cells | C: 4 to 14 A: 3 to 8 |
- | (44) |
ACEs | (TadA-8e)-evoCDA1-nCas9-1xUGI (ACE5) | C-to-T and A-to-G | Rice | C: 1 to 14 A: 4 to 8 |
OsACC (HerbicideR) | (5) |
Dual BEs | TadA-(TadA-8e)-nCas9-AID10-UGI | C-to-T and A-to-G | Arabidopsis protoplasts | C: 11 to 14 A: 4 to 8 |
- | (45) |
AID10-TadA-(TadA-8e)-nCas9-UGI | C-to-T and A-to-G | Arabidopsis protoplasts | C: 1 to 11 A: 4 to 8 |
|||
hyA&C-BEmax | hAID-(TadA-8e)-Rad51DBD-nCas9-2xUGI | C-to-T and A-to-G | Mammalian cells, embryos | C: 2 to 13 A: 3 to 9 |
HBG1/HBG2 promoter (β-thalassemia) | (46) |
AGBEs | hAPOBEC3Ai- TadA-8e (V106W)-nCas9 (miniAGBE-4) | C-to-G/T/A, A-to-G and indels | Mammalian cells, embryos | C: 3 to 13 A: 4 to 8 |
hDTR (Diphtheria toxinR) | (47) |
Pol-BE fusions | ||||||
MutaT7b | rAPOBEC1-T7RNAP | C-to-T | E. coli (Δung) | Multi-kbc | - | (48) |
TRACEb | hAID-T7RNAP (native or mutant)-1xUGI | C-to-T | Mammalian cells | Multi-kbc | MEK1 (Inhibitor assay) | (8) |
hAID-T7RNAP (mutant)-1xUGI | C-to-T | Tobacco, rice | Multi-kbc | OsALS1 (HerbicideR) | (50) | |
T7-DIVAb | hAID-T7RNAP, rAPOBEC-T7RNAP, PmCDA1-T7RNAP | C-to-T | E. coli (Δung) | Multi-kbc | TEM-1 (CeftazidimeR) | (7) |
TadA*-T7RNAP | A-to-G | E. coli | Multi-kbc | |||
TRIDENTb | PmCDA1-T7RNAP | C-to-T** | Yeastd | Multi-kbc | TadA*, PfDHFR (PyrimethamineR) | (49) |
ABE7.10-T7RNAP | A-to-G** | Yeastd | ||||
yeTadA1.0- T7RNAP | A-to-G** | Yeastd | ||||
eMutaT7b | PmCDA1-T7RNAP and free UGI | C-to-T | E. coli | Multi-kbc | TEM-1 (CeftazidimeR), DegP (heat stress) | (51) |
PmCDA1-T7RNAP and free UGI | C-to-T | E. coli (plantized) | Multi-kbc | AtADT (PhenylalanineR) | (57) | |
MutaEco | αEcoRNAP-PmCDA1 | C-to-T | E. coli (Δung) | Multi-kbc | - | (52) |
eMutaT7transition | PmCDA1-T7RNAP + free UGI + (TadA-8e)- T7RNAP | C-to-T and A-to-G | E. coli | Multi-kbc | TEM-1 (CeftazidimeR) | (54) |
MutaT7 toolkit | ABE7.10-T7RNAP, (TadA-8e)-T7RNAP, rAPOBEC1-T7RNAP | C-to-T and A-to-G | E. coli (Δung) | Multi-kbc | rpsL (AmpicillinR) | (53) |
PEACE | AID-T7RNAP(T3) + toxin-antitoxin system | C-to-T | E. coli | Multi-kbc | T7RNAP (Pro specificity) | (58) |
Deaminase-based other TRM systems | ||||||
TAM | hAIDx-dCas9 + free UGI (+ Pool of sgRNAs) | C-to-N | Mammalian cells | 5 to 9 | ABL (ImatinibR) | (25) |
CRISPR-X | dCas9 + (Pool of sgRNAs) | C-to-N | Mammalian cells | −50 to +50 | PSMB5 (Bortezomib inhibitor assay) | (17) |
**Additional substitutions were (C-to-T/A) observed; aVariable editing window length according to fused protein; bEvolves DNA segment downstream of T7 promoter up to T7 terminator; cVariable editing window length in kilobase (kb) pairs according to the position of T7 terminator; dYeast strains were engineered by mutating or overexpressing the different DNA-repair factors.
![]() |
![]() |