Structural basis of sequence-specific cytosine deamination by double-stranded DNA deaminase toxin DddA

Structural basis of sequence-specific cytosine deamination by double-stranded DNA deaminase toxin DddA | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Structural basis of sequence-specific cytosine deamination by double-stranded DNA deaminase toxin DddA Hideki Aihara, Lulu Yin, Ke Shi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-2031914/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 17 Jul, 2023 Read the published version in Nature Structural & Molecular Biology → Version 1 posted You are reading this latest preprint version Abstract An interbacterial deaminase toxin DddA catalyzes cytosine-to-uracil conversion in double-stranded (ds) DNA and enables CRISPR-free mitochondrial base editing, but the molecular mechanisms underlying its unique substrate selectivity remained unknown. Here we report crystal structures of DddA bound to a dsDNA substrate containing the 5'-TC target motif. The structures show that DddA binds to the minor groove of a sharply bent dsDNA and engages the target cytosine extruded from the double-helix. DddA Phe1375 intercalates in dsDNA and displaces the 5' (–1) thymine, which in turn replaces the target (0) cytosine and forms a non-canonical T-G base-pair with the juxtaposed guanine. This “domino effect” mechanism allows DddA to locate the target cytosine without flipping it into the active site. Biochemical experiments show that DNA base-mismatches enhance DddA deaminase activity and relax its sequence selectivity. Based on the structural information, we further identified DddA mutants that exhibit attenuated activity or altered substrate preference. Our studies may help design novel tools useful in genome editing or other applications. Biological sciences/Biochemistry/Enzyme mechanisms Biological sciences/Structural biology/X-ray crystallography Biological sciences/Molecular biology/DNA metabolism Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Enzymatic deamination of cytosines in DNA plays key roles in various important biological processes, including innate immune responses against viruses and transposons, antibody diversification in adaptive immunity, and the accumulation of somatic mutations in various human cancers 1 – 4 . The activity of APOBEC family single-stranded (ss) DNA cytosine deaminases has also been harnessed in base editing technologies, where an engineered Cas9-guide RNA complex directs APOBECs for site-specific C-to-T base substitutions in genomic DNA without making double-strand breaks 5 . Cytosine deamination by the APOBEC enzymes is sequence-selective; for instance, human APOBEC3A (A3A) and APOBEC3B (A3B) only deaminate cytosines in 5'-T C sequence context, which is responsible for the characteristic “APOBEC signature” mutations found widely in cancer genomes 6 , 7 . Structural studies have shown that A3A and A3B bind ssDNA substrates in a U-shaped conformation, with the thymine base 5' (–1) to the target cytosine flipped out and making specific contacts with the protein 8 . A similar mode of hairpin-shaped substrate engagement was observed for a distantly related bacterial tRNA adenosine deaminase TadA, which served as the template for an evolved DNA adenine deaminase capable of A-to-G conversion in base editing 9 , 10 . Recent studies have identified a double-stranded (ds) DNA deaminase from Burkholderia cenocepacia , DddA, an interbacterial toxin that is delivered to contacting cells by the type VI secretion system and mediates antagonism between Gram-negative bacteria 11 , 12 . Interestingly, DddA shares the strong preference for the 5'-T C target sequence with A3A, A3B, and several other APOBEC family members 12 . However, unlike APOBECs that only deaminates ssDNA, DddA selectively deaminates cytosines in dsDNA. The unique activity of DddA allowed Mok et al . to develop a CRISPR-free DddA-derived cytosine base editors (DdCBEs), which enable C-to-T base editing in mitochondrial, chloroplast, and nuclear DNA 12 – 20 . Furthermore, Cho et al. showed that a catalytically inactive DddA mutant (E1347A) fused to the TadA-derived DNA adenine deaminase mediates targeted A-to-G editing in human mitochondrial DNA, where DddA may assist in unwinding/melting of the dsDNA substrate 21 . In addition, DddA has been adapted by Gallagher et al. for genome-wide protein–DNA interaction site mapping in bacteria 22 . However, despite its useful applications, molecular mechanisms underlying the biochemical activities of DddA have remained unknown. Here we report crystal structures of DddA in complex with dsDNA and corroborating biochemical data, which together reveal a unique mechanism of substrate DNA recognition of DddA. Results Overall structure of DddA-dsDNA complex To understand how DddA interacts with dsDNA substrates, we crystallized the toxin domain (Gly1290 to Pro1422) of Burkholderia cenocepacia DddA in complex with a 14-bp dsDNA substrate containing the 5'-T C target sequence (Fig. 1 a). DddA with a substitution of the catalytically essential glutamic acid residue (E1347A) was used to capture the enzyme-substrate complex. The structure of the DddA-dsDNA complex was determined in two different crystal forms and refined to 2.6 and 2.5-Å resolution, respectively ( Supplementary Table 1 ). The crystal structures show that DddA engages the minor groove of a sharply bent dsDNA (Fig. 1 b, c). The structures obtained in the two crystal forms are very similar overall, with a root mean square deviation (r.m.s.d.) of 1.37 Å for all protein and DNA atoms and 0.45 Å for the protein backbone atoms, although they differ in the conformation of the target (0) 2'-deoxycytidine nucleotide. In the first structure (PDB ID: 8E5E), the target cytosine base is completely flipped out of the DNA double helix and captured in the active site pocket, where it interacts with the Zn ion (Fig. 2 a). In the second structure (PDB ID: 8E5D), the target cytosine is parked in the major groove via a T-shaped stacking on the edge of the adjacent (+ 1) cytosine base, and the active site pocket is occupied by a phosphate ion (Fig. 2 b). In both structures, the dsDNA substrate bound by DddA is bent away from the protein by ~ 80°, which leads to a significantly widened minor groove (groove width up to 15 Å in comparison to 6 Å in the B-form DNA; calculated using CURVES+) 23 , allowing for direct base contacts by the protein. Correspondingly, several nucleotides surrounding the 5'- −1 T C 0 motif, including G (–2) and C (+ 1) of the deaminated strand and A (–1) of the complementary strand (unpaired due to the shift of − 1 T; see below), show the A-form-like C3'-endo sugar pucker in both structures. The structure of the Zn-dependent deaminase fold of DddA in complex with DNA shows minimal changes from that in complex with the immunity protein DddI (PDB ID: 6u08) 12 , with an overall backbone r.m.s.d. of 0.50 and 0.62 Å, respectively, for the two DNA-bound structures. Besides the active site zinc ion, we observed electron density for a putative metal ion octahedrally coordinated by the backbone carbonyl oxygen of Glu1381, Thr1382, Leu1384, and Asn1417, and both the backbone and side chain oxygen atoms of Asn1415. This density was modeled as a magnesium ion, which appears to play a structural role to stabilize the DddA residues important for DNA-binding ( Supplementary Fig. 1 ). Mechanism of TC motif recognition The minor groove interaction by DddA is centered on Phe1375, which intercalates in dsDNA and displaces thymine at − 1 position (5' to the target cytosine) (Fig. 1 b). The displaced thymine in turn replaces the target (0) cytosine extruded from the double helix (Fig. 3 a). This unique arrangement is stabilized by bifurcated hydrogen bonds donated to the thymine O4 atom from the juxtaposed guanine base N1 and N2 atoms (Fig. 3 b). His1345, which is one of the Zn-coordinating residues, also donates a hydrogen-bond to the thymine O2 atom. Thus, the strong 5'-T C preference of DddA appears to reflect the favorable interaction made by − 1 T base in replacing the target cytosine in the double helix. The non-canonical T-G interaction, which is distinct from the G-T wobble pair commonly observed in RNA secondary structures, is further stabilized by van der Waals contacts made by Ala1341 and a hydrogen bond between the carbonyl oxygen of Pro1338 and the guanine base N2 atom. Met1379 complements Phe1375 and Ala1341 to form a cluster of hydrophobic side chains inserted into the minor groove, interacting with the orphan (unpaired) adenine at the − 1 position and stabilizing unstacked bases of dsDNA in the distorted conformation (Figs. 1 b, 3 a). Upstream of the 5'-T C motif, Asn1378 and Arg1403 are inserted into the DNA minor groove and interact with bases at positions − 2 to − 4, respectively, which may modestly contribute to sequence preferences ( Supplementary Fig. 2 ). Binding of DddA to the bent DNA is also supported by interaction with the backbone phosphate groups from both strands, involving residues Ser1331, Asn1339, Tyr1340, Lys1402, and Lys1420 ( Supplementary Fig. 2 ). Base-mismatches promote DddA activity Based on the highly distorted conformation of the dsDNA bound to DddA, we reasoned that base-mismatches at the target (0) or 5' (–1) position would destabilize the double helical structure of the substrate and facilitate DNA deamination by DddA. Thus, we compared DddA activity on fully base-paired, singly mis-matched (at position 0), and doubly mis-matched (at positions 0 and − 1) 14-bp dsDNA substrates (Fig. 4 a, b). DddA deaminates cytosine in the 5'-T C motif in the fully base-paired substrate, in which the complementary strand has opposing 5'-GA (Fig. 4 b, lane 6). Using a complementary strand with single mismatch (5'-TA) led to an enhanced activity, confirming our hypothesis (Fig. 4 b, lane 4). The deamination reaction was even more efficient with a complementary strand with double mismatches (5'-TT), consistent with our structural observation that substrate engagement by DddA requires disruption of base-pairs at both positions 0 and − 1 (Fig. 4 b, lane 5). Next, we further hypothesized that base-mismatches may relax the 5'-T C requirement of DddA and examined whether DddA can deaminate cytosines preceded by different − 1 bases (5'-G C , 5'-C C , 5'-A C ) when paired with mismatched complementary strands (Fig. 4 c). For the original complementary strand with 5'-GA, which would generate mismatches at the − 1 position, we observed DddA-mediated deamination on all 3 substrates to varying extent; the activity was highest on 5'-A C and poor on 5'-G C (Fig. 4 c, lanes 9–11). For the complementary strand with opposing 5'-TT, we also observed deamination on all 3 substrates but their preferences were reversed; the activity was highest on 5'-G C and modest on 5'-A C that forms single mismatch at position 0 (Fig. 4 c, lanes 6–8). With opposing 5'-TA, the activity was high on all 3 doubly mismatched substrates (Fig. 4 c, lanes 3–5). Of note, the 5'-C C target was deaminated at both (–1 and 0) cytosines, which was confirmed by testing substrates labeled at either 5' or 3' terminus of the target strand (Fig. 4 c, lanes 4, 7, 10; Supplementary Fig. 3 ). These results show that base-mismatches at either position 0 or − 1 eliminate the 5'-T C requirement of DddA, although the sequence context matters in some cases. DddA mutants To dissect structure-function relationships, we explored amino acid substitutions for key DNA-interacting residues of DddA (Fig. 5 a, Supplementary Fig. 4 ). As mentioned above, a triad of hydrophobic residues, Ala1341, Phe1375, and Met1379 support unstacked bases of DddA-bound dsDNA in the minor groove (Figs. 1 b, 3 a). For Ala1341, which abuts against the non-canonical T-G base pair, we tested substitutions of Ser, Thr, Glu, Tyr, and Pro. Of these mutants, only DddA A1341P retained activity on the canonical substrate (5'-T C /GA), and it showed the 5'-T C preference (Fig. 5 a, b). Interestingly, although the activity of DddA A1341P on the fully base-paired substrate was weaker than that of the wild type, DddA A1341P showed higher activities than wild type on all mismatch-containing substrates (Fig. 5 c, d vs. Figure 4 b, c). The hydrophobic proline side chain inserted more deeply (than alanine) into the minor groove may interact favorably with unpaired DNA bases. For the DNA-intercalating residue Phe1375, either Ala (F1375A) or Arg (F1375R) substitution led to a complete loss of the deaminase activity while a variant with Tyr substitution (F1375Y) showed residual activity, which highlights the importance of π-stacking interaction (Fig. 5 a). DddA F1375Y also showed activities on mis-matched substrates ( Supplementary Fig. 4 ). For Met1379, either Ala (M1379A) or Arg (M1379R) substitution abolished the deaminase activity (Fig. 5 a). These results show the importance of the hydrophobic patch of DddA in DNA substrate engagement and that structural perturbation of this region affects the target preference. One of the DddA residues positioned near the DNA backbone is Glu1370, which forms a part of the rim of the deep active site pocket along with Tyr1307. In the structure with the cytosine base parked in the DNA major groove, Glu1370 side chain is pointed away from the DNA (Fig. 2 b). When the cytosine base is engaged in the active site pocket, Glu1370 appears to be oriented toward DNA with ~ 3.7 Å between the carboxyl and phosphate groups, although weak electron density suggests high flexibility of this side chain (Fig. 2 a). Substitution of either Lys (E1370K) or Arg (E1370R) for Glu1370, which installs a positive charge to interact favorably with the DNA backbone phosphate, made DddA less active than the wild type (Fig. 5 a, Supplementary Fig. 4 ). It is possible that the dynamics of this residue plays a role in flipping the target cytosine base into the active site. Lastly, replacing His1345 with Cys, an alternative Zn coordinating residue as found in some cytidine deaminases 24 , abolished the DddA activity (Fig. 5 a). Discussion Our structural studies show that DddA active site captures the target cytosine base that has completely swung out of the DNA double helix (Figs. 1 , 2 ). Similar base-flipping mechanisms have been observed for various nucleic acid repair or modifying enzymes, including DNA glycosylase, cytosine methyltransferase, dsRNA adenosine deaminase, and lesion-specific endonuclease 25 – 32 . A hallmark feature of these enzymes is the intercalation of an amino acid side chain into DNA/RNA base stacks to fill a void in the double helix 33 . Another frequently observed feature is a sharp kink in the dsDNA substrate with unstacked bases, which also facilitates base-flipping 25 , 29 , 31 , 34 . DddA uses both these strategies – the dsDNA bound by DddA is sharply bent at the base step 5' to the 5'-T C motif, and Phe1375 inserts deeply into the minor groove. However, the mechanism of base-flipping by DddA is distinct in that the intercalated phenylalanine replaces the adjacent (–1) thymine rather than the target (0) cytosine base itself (Fig. 3 a). This unique arrangement causes a “domino effect” and a shift in the register of base-pairing, with the target cytosine base extruded from the double helix. The DddA-dsDNA structure trapped with the target cytosine parked in the major groove (Fig. 2 b) may represent an intermediate conformation and suggests that DddA can locate 5'-T C motifs in double-stranded DNA without engaging the cytosine base in the active site. The mechanism of 5'-T C target recognition by DddA is distinct from that of APOBEC-family ssDNA deaminases. We showed previously that ssDNA substrates bound to A3A and A3B take a U-shaped conformation with the − 1 thymine base bound in a groove on the enzyme surface, where it forms hydrogen bonds with a key Asp side chain 8 . In contrast, the − 1 thymine in dsDNA bound to DddA remains intrahelical and is paired with a guanine base, where it makes both DNA base (guanine) and protein side chain (His1345) contacts (Fig. 3 b). The strong 5'-T C selectivity of DddA suggests that the non-canonical T-G interaction is required for the target cytosine base-flipping, which is corroborated by the dramatically relaxed target sequence selectivity of DddA on mismatch-containing dsDNA substrates. While most amino acid substitutions that affect the key DNA minor groove interaction of DddA led to a loss of the enzymatic activity, several mutant enzymes retained DNA deaminase activity (Fig. 5 , Supplementary Fig. 4 ). These attenuated DddA variants could be useful in reducing off-target mutations or alleviating cytotoxicity in base editing, as shown in recent studies 16 , 19 . In addition, the enhanced activity of DddA A1341P toward mismatch-containing substrates (Fig. 4 c vs . 5d ) suggests that it might be possible to engineer DddA to expand its targets. In this context, it is notable that recent directed evolution experiments have identified DddA11, a DddA variant containing A1341V and E1370K amino acid substitutions, which can edit non-T C targets in both mitochondrial and nuclear DNA 18 . Our studies reported here will be instrumental in further structure-based engineering of DddA for base editing or other novel applications, either as the deaminase catalytic component or a vehicle for other DNA-modifying enzymes. Methods Protein expression and purification DddA(1290–1422) with an E1347A deaminase-inactivating amino acid substitution was expressed in E. coli strain BL21(DE3) using the pET-24 a vector with a C-terminal 6xHis-tag and an additional methionine on the N-terminus. The expressed protein was purified using nickel affinity and size-exclusion chromatography. An N-terminal fragment, DddA(1290–1396), was expressed as MBP-fusion using the pMAL-c5x vector with a 8xHis-tag and a HRV 3C protease cleavage site between MBP and DddA. The wild type or various mutant derivatives of this fusion protein were expressed and purified as above, except that the MBP-His 8 tag was cleaved after the nickel affinity step by an overnight incubation with HRV 3C protease. Purified proteins were concentrated by ultrafiltration in 20 mM Tris-HCl, pH 7.4, 0.5 M NaCl, 5 mM β-mercaptoethanol, flash-frozen in liquid nitrogen, and stored at − 80°C. Protein concentrations were determined based on UV absorbance. Crystallization and structure determination DddA(1290–1422) E1347A at ~ 12 mg ml − 1 was mixed with 1.5x molar excess of a 14 bp dsDNA (5'- GCAACG T C CGGTAC/5'-GTACCG GA CGTTGC; the 5'-T C target motif is underlined) and dialyzed overnight at 4°C against 10 mM Tris-HCl, pH 7.4, 0.1 M NaCl, 0.5 mM Tris(2-carboxyethyl)phosphine. The dialyzed complex was subjected to crystallization screening in the sitting drop vapor diffusion mode. We obtained crystals in two different conditions. Crystal form #1 obtained in [0.2 M magnesium chloride, 0.1 M Tris-HCl, pH 8.5, 25% polyethylene glycol 3350] yielded the structure with the target cytosine in the active site pocket at 2.62 Å resolution. Crystal form #2 obtained in [0.2 M sodium dihydrogen phosphate, 20% polyethylene glycol 3350] yielded the structure with the target cytosine parked in the DNA major groove at 2.49 Å resolution. The DddA-dsDNA crystals were cryo-protected by brief soaking in the respective well solution supplemented with 20% ethylene glycol and flash cooled by plunging in liquid nitrogen. X-ray diffraction data were collected at the NE-CAT beamline 24-ID-C of the Advanced Photon Source (Lemont, IL) and processed using XDS 35 . The structures were determined by molecular replacement with PHASER 36 using the previously reported inhibitor-bound DddA structure 12 (PDB ID: 6u08) as the search model. Iterative model building and refinement were conducted using COOT 37 and PHENIX 38 . A summary of crystallographic data statistics is shown in Supplementary Table 1 . Figures were generated using PyMOL ( https://pymol.org/2/ ). DddA activity assay To reconstitute the active enzyme, DddA(1290–1396) was mixed with 10x molar excess of a chemically synthesized C-terminal peptide corresponding to residues 1397 to 1422 (GAIPVKRGATGETKVFTGNSNSPKSP). The deaminase assay was conducted with a 5'-fluorescein-labeled 14-mer DNA oligo (5'-GCAACG T C CGGTAC-3') or its variants with different − 1 bases (5'-G C , 5'-C C , 5'-A C ), annealed to an unlabeled 14-mer complementary DNA strand (5'-GTACCG GA CGTTGC) or its variants with 5'-TT, 5'-TA, 5'-GC, 5'-GG, or 5'-GT in place of the underlined 5'-GA. The reactions contained 200 nM dsDNA substrate, 10 µM DddA(1290–1396), 100 µM DddA(1397–1422), 40 mM Tris-HCl, pH 7.4, 50 mM KCl, 1.0 mM MgCl 2 , 1.0 mM dithiothreitol. Following an incubation at 37°C for 50 min, pfuEndoQ was added to the final concentration of 1.0 µM and the samples were further incubated at 60°C for 30 min to cleave deaminated products 29 . The reactions were stopped by the addition of formamide to65 % and heating to 95°C for 10 min. The products were separated by gel electrophoresis on a15 % polyacrylamide TBE-Urea denaturing gel and visualized by scanning on a Typhoon FLA 9500 imager. For every experiment, the activity of pfuEndoQ was verified on a control DNA oligo with dU (2'-deoxyuridine) in place of the target C. Specifically in the experiment shown in Supplementary Fig. 3, 3'-fluorescein-labeled DNA substrates were used. Declarations Acknowledgement This work was supported by grants from the US National Institutes of Health (NIGMS R35-GM118047 and NCI P01-CA234228 to H.A.). X-ray diffraction data were collected at the Northeastern Collaborative Access Team beamlines, which are funded by the US National Institutes of Health (NIGMS P30 GM124165). The Pilatus 6M detector on 24-ID-C beamline is funded by a NIH-ORIP HEI grant (S10 RR029205). This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357. We thank Reuben Harris for helpful suggestions. Data Availability Atomic coordinates and structure factors have been deposited in the Protein Data Bank (PDB) under accession codes 8E5E and 8E5D. All other data are available from the authors upon request. Author Contributions L.Y. performed protein purification, crystallization, and biochemical analyses. K.S. performed crystallization, X-ray data collection, and structure determination. H.A. managed the project and wrote the paper with inputs from all authors. Competing Interests The authors have no competing interests to declare. References Feng, Y., Seija, N., Di Noia, J.M. & Martin, A. AID in Antibody Diversification: There and Back Again. Trends Immunol 41 , 586–600 (2020). Green, A.M. & Weitzman, M.D. The spectrum of APOBEC3 activity: From anti-viral agents to anti-cancer opportunities. DNA Repair (Amst) 83 , 102700 (2019). Muramatsu, M. et al. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 102 , 553–63 (2000). Swanton, C., McGranahan, N., Starrett, G.J. & Harris, R.S. APOBEC Enzymes: Mutagenic Fuel for Cancer Evolution and Heterogeneity. Cancer Discov 5 , 704–12 (2015). Komor, A.C., Kim, Y.B., Packer, M.S., Zuris, J.A. & Liu, D.R. Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage. Nature 533 , 420–4 (2016). Alexandrov, L.B. et al. Signatures of mutational processes in human cancer. Nature 500 , 415 – 21 (2013). Jarvis, M.C., Ebrahimi, D., Temiz, N.A. & Harris, R.S. Mutation Signatures Including APOBEC in Cancer Cell Lines. JNCI Cancer Spectr 2 (2018). Shi, K. et al. Structural basis for targeted DNA cytosine deamination and mutagenesis by APOBEC3A and APOBEC3B. Nat Struct Mol Biol 24 , 131–139 (2017). Gaudelli, N.M. et al. Programmable base editing of A*T to G*C in genomic DNA without DNA cleavage. Nature 551 , 464–471 (2017). Losey, H.C., Ruthenburg, A.J. & Verdine, G.L. Crystal structure of Staphylococcus aureus tRNA adenosine deaminase TadA in complex with RNA. Nat Struct Mol Biol 13 , 153–9 (2006). de Moraes, M.H. et al. An interbacterial DNA deaminase toxin directly mutagenizes surviving target populations. Elife 10 (2021). Mok, B.Y. et al. A bacterial cytidine deaminase toxin enables CRISPR-free mitochondrial base editing. Nature 583 , 631–637 (2020). Guo, J. et al. DdCBE mediates efficient and inheritable modifications in mouse mitochondrial genome. Mol Ther Nucleic Acids 27 , 73–80 (2022). Kang, B.C. et al. Chloroplast and mitochondrial DNA editing in plants. Nat Plants 7 , 899–905 (2021). Lee, H. et al. Mitochondrial DNA editing in mice with DddA-TALE fusion deaminases. Nat Commun 12 , 1190 (2021). Lei, Z. et al. Mitochondrial base editor induces substantial nuclear off-target mutations. Nature 606 , 804–811 (2022). Lim, K., Cho, S.I. & Kim, J.S. Nuclear and mitochondrial DNA editing in human cells with zinc finger deaminases. Nat Commun 13 , 366 (2022). Mok, B.Y. et al. CRISPR-free base editors with enhanced activity and expanded targeting scope in mitochondrial and nuclear DNA. Nat Biotechnol (2022). Mok, Y.G. et al. Base editing in human cells with monomeric DddA-TALE fusion deaminases. Nat Commun 13 , 4038 (2022). Silva-Pinheiro, P. et al. In vivo mitochondrial base editing via adeno-associated viral delivery to mouse post-mitotic tissue. Nat Commun 13 , 750 (2022). Cho, S.I. et al. Targeted A-to-G base editing in human mitochondrial DNA with programmable deaminases. Cell 185 , 1764–1776 e12 (2022). Gallagher, L.A. et al. Genome-wide protein-DNA interaction site mapping in bacteria using a double-stranded DNA-specific cytosine deaminase. Nat Microbiol 7 , 844–855 (2022). Blanchet, C., Pasi, M., Zakrzewska, K. & Lavery, R. CURVES + web server for analyzing and visualizing the helical, backbone and groove parameters of nucleic acid structures. Nucleic Acids Res 39 , W68-73 (2011). Teh, A.H. et al. The 1.48 A resolution crystal structure of the homotetrameric cytidine deaminase from mouse. Biochemistry 45 , 7825–33 (2006). Hollis, T., Ichikawa, Y. & Ellenberger, T. DNA bending and a flip-out mechanism for base excision by the helix-hairpin-helix DNA glycosylase, Escherichia coli AlkA. EMBO J 19 , 758–66 (2000). Hong, S. & Cheng, X. DNA Base Flipping: A General Mechanism for Writing, Reading, and Erasing DNA Modifications. Adv Exp Med Biol 945 , 321–341 (2016). Klimasauskas, S., Kumar, S., Roberts, R.J. & Cheng, X. HhaI methyltransferase flips its target base out of the DNA helix. Cell 76 , 357–69 (1994). Matthews, M.M. et al. Structures of human ADAR2 bound to dsRNA reveal base-flipping mechanism and basis for site selectivity. Nat Struct Mol Biol 23 , 426–33 (2016). Shi, K. et al. Structural basis for recognition of distinct deaminated DNA lesions by endonuclease Q. Proc Natl Acad Sci U S A 118 (2021). Slupphaug, G. et al. A nucleotide-flipping mechanism from the structure of human uracil-DNA glycosylase bound to DNA. Nature 384 , 87–92 (1996). Vassylyev, D.G. et al. Atomic model of a pyrimidine dimer excision repair enzyme complexed with a DNA substrate: structural basis for damaged DNA recognition. Cell 83 , 773–82 (1995). Yang, C.G. et al. Crystal structures of DNA/RNA repair enzymes AlkB and ABH2 bound to dsDNA. Nature 452 , 961–5 (2008). Hendershot, J.M. & O'Brien, P.J. Critical role of DNA intercalation in enzyme-catalyzed nucleotide flipping. Nucleic Acids Res 42 , 12681–90 (2014). Olmon, E.D. & Delaney, S. Differential Ability of Five DNA Glycosylases to Recognize and Repair Damage on Nucleosomal DNA. ACS Chem Biol 12 , 692–701 (2017). Kabsch, W. Xds. Acta Crystallogr D Biol Crystallogr 66 , 125–32 (2010). McCoy, A.J. et al. Phaser crystallographic software. J Appl Crystallogr 40 , 658–674 (2007). Emsley, P., Lohkamp, B., Scott, W.G. & Cowtan, K. Features and development of Coot. Acta Crystallogr D Biol Crystallogr 66 , 486–501 (2010). Adams, P.D. et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr 66 , 213–21 (2010). Jurrus, E. et al. Improvements to the APBS biomolecular solvation software suite. Protein Sci 27 , 112–128 (2018). Additional Declarations There is NO Competing Interest. Supplementary Files DddAsupplementv4.pdf Cite Share Download PDF Status: Published Journal Publication published 17 Jul, 2023 Read the published version in Nature Structural & Molecular Biology → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-2031914","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":153319023,"identity":"5b2187de-04a2-4184-a154-5152543c598b","order_by":0,"name":"Hideki Aihara","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuUlEQVRIiWNgGAWjYDACZh4gYSDBww/lEqulwkZGsoFoLQwgLWfSbAwOEKuFv5334OfKtsM8xjey0x4wVFgnNhDSInGYL1nyLFCL2Y3c7QYMZ9IJa2E4zGMg2QjRsk2Cse0wYS3yQCf9BGkxngHS8o8ILQZA8yUbzqTxGEiAtDQQocUQqMWyocKGR+LM220SCcfSjQlqkTt/xvhmg4GEPX870JYPNdayBLWgggTSlI+CUTAKRsEowAUADDY7TkdPdokAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0001-7508-6230","institution":"University of Minnesota","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Hideki","middleName":"","lastName":"Aihara","suffix":""},{"id":153319024,"identity":"8f3f4b1e-f660-4172-aa79-b3a1af175816","order_by":1,"name":"Lulu Yin","email":"","orcid":"","institution":"University of Minnesota","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Lulu","middleName":"","lastName":"Yin","suffix":""},{"id":153319025,"identity":"7b70c77d-9fa4-4e96-91a5-1c5227ceec11","order_by":2,"name":"Ke Shi","email":"","orcid":"","institution":"University of Minnesota","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Ke","middleName":"","lastName":"Shi","suffix":""}],"badges":[],"createdAt":"2022-09-05 03:16:03","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-2031914/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-2031914/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41594-023-01034-3","type":"published","date":"2023-07-17T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":29359181,"identity":"18346790-c1fa-493e-a0db-d7139bb97739","added_by":"auto","created_at":"2022-11-21 22:12:39","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":721668,"visible":true,"origin":"","legend":"\u003cp\u003eCrystal structure of the DddA-dsDNA complex\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e Sequence of the 14 bp DNA substrate used in our crystallographic studies, designed based on the sequence preference of DddA in \u003cem\u003eE. coli\u003c/em\u003e reported previously\u003csup\u003e12\u003c/sup\u003e. \u003cstrong\u003eb\u003c/strong\u003e Overall view of the DddA-dsDNA complex, with the target cytosine flipped out of the double helix and engaged in the enzyme active site. The color scheme for nucleotides at the –1 and 0\u003csup\u003eth\u003c/sup\u003e positions follows that in \u003cstrong\u003ea\u003c/strong\u003e. \u003cstrong\u003ec\u003c/strong\u003e Alternative view of the DddA-dsDNA complex, with the DddA molecular surface colored according to electrostatic potential (–2.5 kT/e in red to +2.5 kT/e in blue) as calculated by APBS\u003csup\u003e39\u003c/sup\u003e.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-2031914/v1/00460cd680a884c284b8f93a.png"},{"id":29359182,"identity":"5ae935ed-aa19-4641-93b5-e077b560ab5e","added_by":"auto","created_at":"2022-11-21 22:12:40","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":412499,"visible":true,"origin":"","legend":"\u003cp\u003eAlternative conformations of the target (0) cytosine\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e Structure of the DddA-dsDNA complex as in Fig. 1b, with the target cytosine engaged in the enzyme active site pocket. \u003cstrong\u003eb\u003c/strong\u003eStructure of the DddA-dsDNA complex in an alternative conformation, in which the extrahelical target cytosine is stacked on the +1 cytosine base.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-2031914/v1/3013b63a3970b000ea63980f.png"},{"id":29358442,"identity":"ef4aad41-a5b6-4829-8dd9-2a29ea827652","added_by":"auto","created_at":"2022-11-21 22:04:39","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":724910,"visible":true,"origin":"","legend":"\u003cp\u003eStructural basis of 5'-T\u003cstrong\u003eC\u003c/strong\u003e target recognition\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e DNA minor groove interaction by DddA centered on a cluster of hydrophobic residues (Ala1341, Phe1375, Met1379). The –1 thymine base in green is displaced by Phe1375, which in turn replaces the extrahelical target (0) cytosine. \u003cstrong\u003eb\u003c/strong\u003e The –1 thymine base in the shifted register, stabilized by hydrogen bonds (yellow dashed lines with distances shown in angstroms) to the juxtaposed guanine base and a zinc-coordinating residue His1345. Van der Waals radii for Ala1341, Phe1375, and Met1379 are indicated by dots.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-2031914/v1/edc65c52e6b9e6e48659b172.png"},{"id":29358444,"identity":"2083c75b-6756-44ab-9dbe-65ded85d1dc2","added_by":"auto","created_at":"2022-11-21 22:04:39","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":307417,"visible":true,"origin":"","legend":"\u003cp\u003eBase-mismatches enhance DddA activity and eliminate its 5'-T\u003cstrong\u003eC\u003c/strong\u003e requirement\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e Sequence variation of the top and bottom strands used to generate mismatch-containing DNA substrates. FAM denotes fluorescein. \u003cstrong\u003eb\u003c/strong\u003e Deamination by wild type DddA of the T\u003cstrong\u003eC\u003c/strong\u003e-containing target strand, annealed to the fully base-paired (lane 6) or mismatched bottom strands (lanes 4, 5). The top bands are uncleaved 14-mer substrate DNAs, whereas the bottom bands are deamination products subsequently cleaved by the lesion-specific endonuclease pfuEndoQ\u003csup\u003e29\u003c/sup\u003e. The control ‘dU DNA’ contains 2'-deoxyuridine in place of the target C. \u003cstrong\u003ec\u003c/strong\u003e Deamination by wild type DddA of the non-T\u003cstrong\u003eC\u003c/strong\u003e target strands, annealed to mismatched bottom strands. The top and bottom strand sequences for the –1 and 0\u003csup\u003eth\u003c/sup\u003e positions are shown above each lane. Lane 12 shows a reaction on the canonical substrate (fully base-paired T\u003cstrong\u003eC\u003c/strong\u003e target) for reference. 5' fluorescein-labeled substrates were used in all experiments in Fig. 4.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-2031914/v1/fb56505e9292b10242449d80.png"},{"id":29358446,"identity":"b7332d2c-f034-4d09-a93f-220eb8a23c4c","added_by":"auto","created_at":"2022-11-21 22:04:40","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":442605,"visible":true,"origin":"","legend":"\u003cp\u003eDddA mutants\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e Activities of DddA point mutants on the canonical (fully base-paired T\u003cstrong\u003eC\u003c/strong\u003e target) substrate. \u003cstrong\u003eb\u003c/strong\u003e A comparison between wild type and A1341P on the fully base-paired DNA substrates with various –1 bases. \u003cstrong\u003ec\u003c/strong\u003e Activities of DddA A1341P on the T\u003cstrong\u003eC\u003c/strong\u003e-containing target strand annealed to the fully base-paired or mismatched bottom strands. \u003cstrong\u003ed\u003c/strong\u003eDeamination by DddA A1341P of the non-T\u003cstrong\u003eC\u003c/strong\u003e target strands annealed to mismatched bottom strands (the same set of substrates as in Fig. 4c). 5' fluorescein-labeled substrates were used in all experiments in Fig. 5.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-2031914/v1/09f4e2f50fad42789a08da6f.png"},{"id":40356609,"identity":"fec60c76-d660-446a-ab14-aa66b67e2ebf","added_by":"auto","created_at":"2023-07-21 07:09:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2435963,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2031914/v1/ef2e1d8d-8313-4eb6-bcf0-35f18b1aee82.pdf"},{"id":29358447,"identity":"870ef1f2-5f0a-42af-a1bc-3705c1d38aaf","added_by":"auto","created_at":"2022-11-21 22:04:40","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":5014887,"visible":true,"origin":"","legend":"","description":"","filename":"DddAsupplementv4.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2031914/v1/83f8c03bc7687853ad630c8a.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Structural basis of sequence-specific cytosine deamination by double-stranded DNA deaminase toxin DddA","fulltext":[{"header":"Introduction","content":"\u003cp\u003eEnzymatic deamination of cytosines in DNA plays key roles in various important biological processes, including innate immune responses against viruses and transposons, antibody diversification in adaptive immunity, and the accumulation of somatic mutations in various human cancers\u003csup\u003e\u003cspan additionalcitationids=\"CR2 CR3\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. The activity of APOBEC family single-stranded (ss) DNA cytosine deaminases has also been harnessed in base editing technologies, where an engineered Cas9-guide RNA complex directs APOBECs for site-specific C-to-T base substitutions in genomic DNA without making double-strand breaks\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. Cytosine deamination by the APOBEC enzymes is sequence-selective; for instance, human APOBEC3A (A3A) and APOBEC3B (A3B) only deaminate cytosines in 5'-T\u003cb\u003eC\u003c/b\u003e sequence context, which is responsible for the characteristic \u0026ldquo;APOBEC signature\u0026rdquo; mutations found widely in cancer genomes\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Structural studies have shown that A3A and A3B bind ssDNA substrates in a U-shaped conformation, with the thymine base 5' (\u0026ndash;1) to the target cytosine flipped out and making specific contacts with the protein\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. A similar mode of hairpin-shaped substrate engagement was observed for a distantly related bacterial tRNA adenosine deaminase TadA, which served as the template for an evolved DNA adenine deaminase capable of A-to-G conversion in base editing\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eRecent studies have identified a double-stranded (ds) DNA deaminase from \u003cem\u003eBurkholderia cenocepacia\u003c/em\u003e, DddA, an interbacterial toxin that is delivered to contacting cells by the type VI secretion system and mediates antagonism between Gram-negative bacteria\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Interestingly, DddA shares the strong preference for the 5'-T\u003cb\u003eC\u003c/b\u003e target sequence with A3A, A3B, and several other APOBEC family members\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. However, unlike APOBECs that only deaminates ssDNA, DddA selectively deaminates cytosines in dsDNA. The unique activity of DddA allowed Mok \u003cem\u003eet al\u003c/em\u003e. to develop a CRISPR-free DddA-derived cytosine base editors (DdCBEs), which enable C-to-T base editing in mitochondrial, chloroplast, and nuclear DNA\u003csup\u003e\u003cspan additionalcitationids=\"CR13 CR14 CR15 CR16 CR17 CR18 CR19\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Furthermore, Cho \u003cem\u003eet al.\u003c/em\u003e showed that a catalytically inactive DddA mutant (E1347A) fused to the TadA-derived DNA adenine deaminase mediates targeted A-to-G editing in human mitochondrial DNA, where DddA may assist in unwinding/melting of the dsDNA substrate\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. In addition, DddA has been adapted by Gallagher \u003cem\u003eet al.\u003c/em\u003e for genome-wide protein\u0026ndash;DNA interaction site mapping in bacteria\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. However, despite its useful applications, molecular mechanisms underlying the biochemical activities of DddA have remained unknown. Here we report crystal structures of DddA in complex with dsDNA and corroborating biochemical data, which together reveal a unique mechanism of substrate DNA recognition of DddA.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv class=\"Section2\" id=\"Sec3\"\u003e\n \u003ch2\u003eOverall structure of DddA-dsDNA complex\u003c/h2\u003e\n \u003cp\u003eTo understand how DddA interacts with dsDNA substrates, we crystallized the toxin domain (Gly1290 to Pro1422) of \u003cem\u003eBurkholderia cenocepacia\u003c/em\u003e DddA in complex with a 14-bp dsDNA substrate containing the 5\u0026apos;-T\u003cstrong\u003eC\u003c/strong\u003e target sequence (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003ea). DddA with a substitution of the catalytically essential glutamic acid residue (E1347A) was used to capture the enzyme-substrate complex. The structure of the DddA-dsDNA complex was determined in two different crystal forms and refined to 2.6 and 2.5-\u0026Aring; resolution, respectively (\u003cstrong\u003eSupplementary Table\u0026nbsp;1\u003c/strong\u003e). The crystal structures show that DddA engages the minor groove of a sharply bent dsDNA (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eb, c). The structures obtained in the two crystal forms are very similar overall, with a root mean square deviation (r.m.s.d.) of 1.37 \u0026Aring; for all protein and DNA atoms and 0.45 \u0026Aring; for the protein backbone atoms, although they differ in the conformation of the target (0) 2\u0026apos;-deoxycytidine nucleotide. In the first structure (PDB ID: 8E5E), the target cytosine base is completely flipped out of the DNA double helix and captured in the active site pocket, where it interacts with the Zn ion (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ea). In the second structure (PDB ID: 8E5D), the target cytosine is parked in the major groove via a T-shaped stacking on the edge of the adjacent (+\u0026thinsp;1) cytosine base, and the active site pocket is occupied by a phosphate ion (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eb). In both structures, the dsDNA substrate bound by DddA is bent away from the protein by ~\u0026thinsp;80\u0026deg;, which leads to a significantly widened minor groove (groove width up to 15 \u0026Aring; in comparison to 6 \u0026Aring; in the B-form DNA; calculated using CURVES+)\u003csup\u003e23\u003c/sup\u003e, allowing for direct base contacts by the protein. Correspondingly, several nucleotides surrounding the 5\u0026apos;-\u003csup\u003e\u0026minus;1\u003c/sup\u003eT\u003cstrong\u003eC\u003c/strong\u003e\u003csup\u003e0\u003c/sup\u003e motif, including G (\u0026ndash;2) and C (+\u0026thinsp;1) of the deaminated strand and A (\u0026ndash;1) of the complementary strand (unpaired due to the shift of \u0026minus;\u0026thinsp;1 T; see below), show the A-form-like C3\u0026apos;-endo sugar pucker in both structures.\u003c/p\u003e\n \u003cp\u003eThe structure of the Zn-dependent deaminase fold of DddA in complex with DNA shows minimal changes from that in complex with the immunity protein DddI (PDB ID: 6u08)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e, with an overall backbone r.m.s.d. of 0.50 and 0.62 \u0026Aring;, respectively, for the two DNA-bound structures. Besides the active site zinc ion, we observed electron density for a putative metal ion octahedrally coordinated by the backbone carbonyl oxygen of Glu1381, Thr1382, Leu1384, and Asn1417, and both the backbone and side chain oxygen atoms of Asn1415. This density was modeled as a magnesium ion, which appears to play a structural role to stabilize the DddA residues important for DNA-binding (\u003cstrong\u003eSupplementary Fig.\u0026nbsp;1\u003c/strong\u003e).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\" id=\"Sec4\"\u003e\n \u003ch2\u003eMechanism of TC motif recognition\u003c/h2\u003e\n \u003cp\u003eThe minor groove interaction by DddA is centered on Phe1375, which intercalates in dsDNA and displaces thymine at \u0026minus;\u0026thinsp;1 position (5\u0026apos; to the target cytosine) (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eb). The displaced thymine in turn replaces the target (0) cytosine extruded from the double helix (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ea). This unique arrangement is stabilized by bifurcated hydrogen bonds donated to the thymine O4 atom from the juxtaposed guanine base N1 and N2 atoms (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eb). His1345, which is one of the Zn-coordinating residues, also donates a hydrogen-bond to the thymine O2 atom. Thus, the strong 5\u0026apos;-T\u003cstrong\u003eC\u003c/strong\u003e preference of DddA appears to reflect the favorable interaction made by \u0026minus;\u0026thinsp;1 T base in replacing the target cytosine in the double helix. The non-canonical T-G interaction, which is distinct from the G-T wobble pair commonly observed in RNA secondary structures, is further stabilized by van der Waals contacts made by Ala1341 and a hydrogen bond between the carbonyl oxygen of Pro1338 and the guanine base N2 atom. Met1379 complements Phe1375 and Ala1341 to form a cluster of hydrophobic side chains inserted into the minor groove, interacting with the orphan (unpaired) adenine at the \u0026minus;\u0026thinsp;1 position and stabilizing unstacked bases of dsDNA in the distorted conformation (Figs. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eb, \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ea). Upstream of the 5\u0026apos;-T\u003cstrong\u003eC\u003c/strong\u003e motif, Asn1378 and Arg1403 are inserted into the DNA minor groove and interact with bases at positions \u0026minus;\u0026thinsp;2 to \u0026minus;\u0026thinsp;4, respectively, which may modestly contribute to sequence preferences (\u003cstrong\u003eSupplementary Fig.\u0026nbsp;2\u003c/strong\u003e). Binding of DddA to the bent DNA is also supported by interaction with the backbone phosphate groups from both strands, involving residues Ser1331, Asn1339, Tyr1340, Lys1402, and Lys1420 (\u003cstrong\u003eSupplementary Fig.\u0026nbsp;2\u003c/strong\u003e).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\" id=\"Sec5\"\u003e\n \u003ch2\u003eBase-mismatches promote DddA activity\u003c/h2\u003e\n \u003cp\u003eBased on the highly distorted conformation of the dsDNA bound to DddA, we reasoned that base-mismatches at the target (0) or 5\u0026apos; (\u0026ndash;1) position would destabilize the double helical structure of the substrate and facilitate DNA deamination by DddA. Thus, we compared DddA activity on fully base-paired, singly mis-matched (at position 0), and doubly mis-matched (at positions 0 and \u0026minus;\u0026thinsp;1) 14-bp dsDNA substrates (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ea, b). DddA deaminates cytosine in the 5\u0026apos;-T\u003cstrong\u003eC\u003c/strong\u003e motif in the fully base-paired substrate, in which the complementary strand has opposing 5\u0026apos;-GA (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eb, lane 6). Using a complementary strand with single mismatch (5\u0026apos;-TA) led to an enhanced activity, confirming our hypothesis (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eb, lane 4). The deamination reaction was even more efficient with a complementary strand with double mismatches (5\u0026apos;-TT), consistent with our structural observation that substrate engagement by DddA requires disruption of base-pairs at both positions 0 and \u0026minus;\u0026thinsp;1 (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eb, lane 5).\u003c/p\u003e\n \u003cp\u003eNext, we further hypothesized that base-mismatches may relax the 5\u0026apos;-T\u003cstrong\u003eC\u003c/strong\u003e requirement of DddA and examined whether DddA can deaminate cytosines preceded by different \u0026minus;\u0026thinsp;1 bases (5\u0026apos;-G\u003cstrong\u003eC\u003c/strong\u003e, 5\u0026apos;-C\u003cstrong\u003eC\u003c/strong\u003e, 5\u0026apos;-A\u003cstrong\u003eC\u003c/strong\u003e) when paired with mismatched complementary strands (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ec). For the original complementary strand with 5\u0026apos;-GA, which would generate mismatches at the \u0026minus;\u0026thinsp;1 position, we observed DddA-mediated deamination on all 3 substrates to varying extent; the activity was highest on 5\u0026apos;-A\u003cstrong\u003eC\u003c/strong\u003e and poor on 5\u0026apos;-G\u003cstrong\u003eC\u003c/strong\u003e (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ec, lanes 9\u0026ndash;11). For the complementary strand with opposing 5\u0026apos;-TT, we also observed deamination on all 3 substrates but their preferences were reversed; the activity was highest on 5\u0026apos;-G\u003cstrong\u003eC\u003c/strong\u003e and modest on 5\u0026apos;-A\u003cstrong\u003eC\u003c/strong\u003e that forms single mismatch at position 0 (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ec, lanes 6\u0026ndash;8). With opposing 5\u0026apos;-TA, the activity was high on all 3 doubly mismatched substrates (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ec, lanes 3\u0026ndash;5). Of note, the 5\u0026apos;-C\u003cstrong\u003eC\u003c/strong\u003e target was deaminated at both (\u0026ndash;1 and 0) cytosines, which was confirmed by testing substrates labeled at either 5\u0026apos; or 3\u0026apos; terminus of the target strand (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ec, lanes 4, 7, 10; \u003cstrong\u003eSupplementary Fig.\u0026nbsp;3\u003c/strong\u003e). These results show that base-mismatches at either position 0 or \u0026minus;\u0026thinsp;1 eliminate the 5\u0026apos;-T\u003cstrong\u003eC\u003c/strong\u003e requirement of DddA, although the sequence context matters in some cases.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\" id=\"Sec6\"\u003e\n \u003ch2\u003eDddA mutants\u003c/h2\u003e\n \u003cp\u003eTo dissect structure-function relationships, we explored amino acid substitutions for key DNA-interacting residues of DddA (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea, \u003cstrong\u003eSupplementary Fig.\u0026nbsp;4\u003c/strong\u003e). As mentioned above, a triad of hydrophobic residues, Ala1341, Phe1375, and Met1379 support unstacked bases of DddA-bound dsDNA in the minor groove (Figs.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eb, \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ea). For Ala1341, which abuts against the non-canonical T-G base pair, we tested substitutions of Ser, Thr, Glu, Tyr, and Pro. Of these mutants, only DddA A1341P retained activity on the canonical substrate (5\u0026apos;-T\u003cstrong\u003eC\u003c/strong\u003e/GA), and it showed the 5\u0026apos;-T\u003cstrong\u003eC\u003c/strong\u003e preference (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea, b). Interestingly, although the activity of DddA A1341P on the fully base-paired substrate was weaker than that of the wild type, DddA A1341P showed higher activities than wild type on all mismatch-containing substrates (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ec, d vs. Figure\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eb, c). The hydrophobic proline side chain inserted more deeply (than alanine) into the minor groove may interact favorably with unpaired DNA bases. For the DNA-intercalating residue Phe1375, either Ala (F1375A) or Arg (F1375R) substitution led to a complete loss of the deaminase activity while a variant with Tyr substitution (F1375Y) showed residual activity, which highlights the importance of \u0026pi;-stacking interaction (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea). DddA F1375Y also showed activities on mis-matched substrates (\u003cstrong\u003eSupplementary Fig.\u0026nbsp;4\u003c/strong\u003e). For Met1379, either Ala (M1379A) or Arg (M1379R) substitution abolished the deaminase activity (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea). These results show the importance of the hydrophobic patch of DddA in DNA substrate engagement and that structural perturbation of this region affects the target preference.\u003c/p\u003e\n \u003cp\u003eOne of the DddA residues positioned near the DNA backbone is Glu1370, which forms a part of the rim of the deep active site pocket along with Tyr1307. In the structure with the cytosine base parked in the DNA major groove, Glu1370 side chain is pointed away from the DNA (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eb). When the cytosine base is engaged in the active site pocket, Glu1370 appears to be oriented toward DNA with ~\u0026thinsp;3.7 \u0026Aring; between the carboxyl and phosphate groups, although weak electron density suggests high flexibility of this side chain (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ea). Substitution of either Lys (E1370K) or Arg (E1370R) for Glu1370, which installs a positive charge to interact favorably with the DNA backbone phosphate, made DddA less active than the wild type (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea, \u003cstrong\u003eSupplementary Fig.\u0026nbsp;4\u003c/strong\u003e). It is possible that the dynamics of this residue plays a role in flipping the target cytosine base into the active site. Lastly, replacing His1345 with Cys, an alternative Zn coordinating residue as found in some cytidine deaminases\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e, abolished the DddA activity (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea).\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur structural studies show that DddA active site captures the target cytosine base that has completely swung out of the DNA double helix (Figs.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Similar base-flipping mechanisms have been observed for various nucleic acid repair or modifying enzymes, including DNA glycosylase, cytosine methyltransferase, dsRNA adenosine deaminase, and lesion-specific endonuclease\u003csup\u003e\u003cspan additionalcitationids=\"CR26 CR27 CR28 CR29 CR30 CR31\" citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. A hallmark feature of these enzymes is the intercalation of an amino acid side chain into DNA/RNA base stacks to fill a void in the double helix\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. Another frequently observed feature is a sharp kink in the dsDNA substrate with unstacked bases, which also facilitates base-flipping\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e,\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e,\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e,\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. DddA uses both these strategies \u0026ndash; the dsDNA bound by DddA is sharply bent at the base step 5' to the 5'-T\u003cb\u003eC\u003c/b\u003e motif, and Phe1375 inserts deeply into the minor groove. However, the mechanism of base-flipping by DddA is distinct in that the intercalated phenylalanine replaces the adjacent (\u0026ndash;1) thymine rather than the target (0) cytosine base itself (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). This unique arrangement causes a \u0026ldquo;domino effect\u0026rdquo; and a shift in the register of base-pairing, with the target cytosine base extruded from the double helix. The DddA-dsDNA structure trapped with the target cytosine parked in the major groove (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb) may represent an intermediate conformation and suggests that DddA can locate 5'-T\u003cb\u003eC\u003c/b\u003e motifs in double-stranded DNA without engaging the cytosine base in the active site.\u003c/p\u003e \u003cp\u003eThe mechanism of 5'-T\u003cb\u003eC\u003c/b\u003e target recognition by DddA is distinct from that of APOBEC-family ssDNA deaminases. We showed previously that ssDNA substrates bound to A3A and A3B take a U-shaped conformation with the \u0026minus;\u0026thinsp;1 thymine base bound in a groove on the enzyme surface, where it forms hydrogen bonds with a key Asp side chain\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. In contrast, the \u0026minus;\u0026thinsp;1 thymine in dsDNA bound to DddA remains intrahelical and is paired with a guanine base, where it makes both DNA base (guanine) and protein side chain (His1345) contacts (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb). The strong 5'-T\u003cb\u003eC\u003c/b\u003e selectivity of DddA suggests that the non-canonical T-G interaction is required for the target cytosine base-flipping, which is corroborated by the dramatically relaxed target sequence selectivity of DddA on mismatch-containing dsDNA substrates.\u003c/p\u003e \u003cp\u003eWhile most amino acid substitutions that affect the key DNA minor groove interaction of DddA led to a loss of the enzymatic activity, several mutant enzymes retained DNA deaminase activity (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, \u003cb\u003eSupplementary Fig.\u0026nbsp;4\u003c/b\u003e). These attenuated DddA variants could be useful in reducing off-target mutations or alleviating cytotoxicity in base editing, as shown in recent studies\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. In addition, the enhanced activity of DddA A1341P toward mismatch-containing substrates (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec \u003cem\u003evs\u003c/em\u003e. \u003cb\u003e5d\u003c/b\u003e) suggests that it might be possible to engineer DddA to expand its targets. In this context, it is notable that recent directed evolution experiments have identified DddA11, a DddA variant containing A1341V and E1370K amino acid substitutions, which can edit non-T\u003cb\u003eC\u003c/b\u003e targets in both mitochondrial and nuclear DNA\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Our studies reported here will be instrumental in further structure-based engineering of DddA for base editing or other novel applications, either as the deaminase catalytic component or a vehicle for other DNA-modifying enzymes.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eProtein expression and purification\u003c/h2\u003e \u003cp\u003eDddA(1290\u0026ndash;1422) with an E1347A deaminase-inactivating amino acid substitution was expressed in \u003cem\u003eE. coli\u003c/em\u003e strain BL21(DE3) using the pET-24 a vector with a C-terminal 6xHis-tag and an additional methionine on the N-terminus. The expressed protein was purified using nickel affinity and size-exclusion chromatography. An N-terminal fragment, DddA(1290\u0026ndash;1396), was expressed as MBP-fusion using the pMAL-c5x vector with a 8xHis-tag and a HRV 3C protease cleavage site between MBP and DddA. The wild type or various mutant derivatives of this fusion protein were expressed and purified as above, except that the MBP-His\u003csub\u003e8\u003c/sub\u003e tag was cleaved after the nickel affinity step by an overnight incubation with HRV 3C protease. Purified proteins were concentrated by ultrafiltration in 20 mM Tris-HCl, pH 7.4, 0.5 M NaCl, 5 mM β-mercaptoethanol, flash-frozen in liquid nitrogen, and stored at \u0026minus;\u0026thinsp;80\u0026deg;C. Protein concentrations were determined based on UV absorbance.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eCrystallization and structure determination\u003c/h2\u003e \u003cp\u003eDddA(1290\u0026ndash;1422) E1347A at ~\u0026thinsp;12 mg ml\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e was mixed with 1.5x molar excess of a 14 bp dsDNA (5'- GCAACG\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eT\u003c/span\u003e\u003cspan type=\"BoldUnderline\" class=\"BoldUnderline\" name=\"Emphasis\"\u003eC\u003c/span\u003eCGGTAC/5'-GTACCG\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eGA\u003c/span\u003eCGTTGC; the 5'-T\u003cb\u003eC\u003c/b\u003e target motif is underlined) and dialyzed overnight at 4\u0026deg;C against 10 mM Tris-HCl, pH 7.4, 0.1 M NaCl, 0.5 mM Tris(2-carboxyethyl)phosphine. The dialyzed complex was subjected to crystallization screening in the sitting drop vapor diffusion mode. We obtained crystals in two different conditions. Crystal form #1 obtained in [0.2 M magnesium chloride, 0.1 M Tris-HCl, pH 8.5, 25% polyethylene glycol 3350] yielded the structure with the target cytosine in the active site pocket at 2.62 \u0026Aring; resolution. Crystal form #2 obtained in [0.2 M sodium dihydrogen phosphate, 20% polyethylene glycol 3350] yielded the structure with the target cytosine parked in the DNA major groove at 2.49 \u0026Aring; resolution. The DddA-dsDNA crystals were cryo-protected by brief soaking in the respective well solution supplemented with 20% ethylene glycol and flash cooled by plunging in liquid nitrogen. X-ray diffraction data were collected at the NE-CAT beamline 24-ID-C of the Advanced Photon Source (Lemont, IL) and processed using XDS\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. The structures were determined by molecular replacement with PHASER\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e using the previously reported inhibitor-bound DddA structure\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e (PDB ID: 6u08) as the search model. Iterative model building and refinement were conducted using COOT\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e and PHENIX\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. A summary of crystallographic data statistics is shown in \u003cb\u003eSupplementary Table\u0026nbsp;1\u003c/b\u003e. Figures were generated using PyMOL (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pymol.org/2/\u003c/span\u003e\u003cspan address=\"https://pymol.org/2/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eDddA activity assay\u003c/h2\u003e \u003cp\u003eTo reconstitute the active enzyme, DddA(1290\u0026ndash;1396) was mixed with 10x molar excess of a chemically synthesized C-terminal peptide corresponding to residues 1397 to 1422 (GAIPVKRGATGETKVFTGNSNSPKSP). The deaminase assay was conducted with a 5'-fluorescein-labeled 14-mer DNA oligo (5'-GCAACG\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eT\u003c/span\u003e\u003cspan type=\"BoldUnderline\" class=\"BoldUnderline\" name=\"Emphasis\"\u003eC\u003c/span\u003eCGGTAC-3') or its variants with different \u0026minus;\u0026thinsp;1 bases (5'-G\u003cb\u003eC\u003c/b\u003e, 5'-C\u003cb\u003eC\u003c/b\u003e, 5'-A\u003cb\u003eC\u003c/b\u003e), annealed to an unlabeled 14-mer complementary DNA strand (5'-GTACCG\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eGA\u003c/span\u003eCGTTGC) or its variants with 5'-TT, 5'-TA, 5'-GC, 5'-GG, or 5'-GT in place of the underlined 5'-GA. The reactions contained 200 nM dsDNA substrate, 10 \u0026micro;M DddA(1290\u0026ndash;1396), 100 \u0026micro;M DddA(1397\u0026ndash;1422), 40 mM Tris-HCl, pH 7.4, 50 mM KCl, 1.0 mM MgCl\u003csub\u003e2\u003c/sub\u003e, 1.0 mM dithiothreitol. Following an incubation at 37\u0026deg;C for 50 min, pfuEndoQ was added to the final concentration of 1.0 \u0026micro;M and the samples were further incubated at 60\u0026deg;C for 30 min to cleave deaminated products\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. The reactions were stopped by the addition of formamide to65 % and heating to 95\u0026deg;C for 10 min. The products were separated by gel electrophoresis on a15 % polyacrylamide TBE-Urea denaturing gel and visualized by scanning on a Typhoon FLA 9500 imager. For every experiment, the activity of pfuEndoQ was verified on a control DNA oligo with dU (2'-deoxyuridine) in place of the target C. Specifically in the experiment shown in Supplementary Fig.\u0026nbsp;3, 3'-fluorescein-labeled DNA substrates were used.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by grants from the US National Institutes of Health (NIGMS R35-GM118047 and NCI P01-CA234228 to H.A.). X-ray diffraction data were collected at the Northeastern Collaborative Access Team beamlines, which are funded by the US National Institutes of Health (NIGMS P30 GM124165). The Pilatus 6M detector on 24-ID-C beamline is funded by a NIH-ORIP HEI grant (S10 RR029205). This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357. We thank Reuben Harris for helpful suggestions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAtomic coordinates and structure factors have been deposited in the Protein Data Bank (PDB) under accession codes 8E5E and 8E5D. All other data are available from the authors upon request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eL.Y. performed protein purification, crystallization, and biochemical analyses. K.S. performed crystallization, X-ray data collection, and structure determination. H.A. managed the project and wrote the paper with inputs from all authors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no competing interests to declare.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003e\u003cspan\u003eFeng, Y., Seija, N., Di Noia, J.M. \u0026amp; Martin, A. AID in Antibody Diversification: There and Back Again. Trends Immunol \u003cstrong\u003e41\u003c/strong\u003e, 586\u0026ndash;600 (2020).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGreen, A.M. \u0026amp; Weitzman, M.D. The spectrum of APOBEC3 activity: From anti-viral agents to anti-cancer opportunities. DNA Repair (Amst) \u003cstrong\u003e83\u003c/strong\u003e, 102700 (2019).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMuramatsu, M. et al. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell \u003cstrong\u003e102\u003c/strong\u003e, 553\u0026ndash;63 (2000).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSwanton, C., McGranahan, N., Starrett, G.J. \u0026amp; Harris, R.S. APOBEC Enzymes: Mutagenic Fuel for Cancer Evolution and Heterogeneity. Cancer Discov \u003cstrong\u003e5\u003c/strong\u003e, 704\u0026ndash;12 (2015).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eKomor, A.C., Kim, Y.B., Packer, M.S., Zuris, J.A. \u0026amp; Liu, D.R. Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage. Nature \u003cstrong\u003e533\u003c/strong\u003e, 420\u0026ndash;4 (2016).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eAlexandrov, L.B. et al. Signatures of mutational processes in human cancer. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e500\u003c/strong\u003e, 415 \u0026ndash; 21 (2013).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eJarvis, M.C., Ebrahimi, D., Temiz, N.A. \u0026amp; Harris, R.S. Mutation Signatures Including APOBEC in Cancer Cell Lines. JNCI Cancer Spectr \u003cstrong\u003e2\u003c/strong\u003e(2018).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eShi, K. et al. Structural basis for targeted DNA cytosine deamination and mutagenesis by APOBEC3A and APOBEC3B. Nat Struct Mol Biol \u003cstrong\u003e24\u003c/strong\u003e, 131\u0026ndash;139 (2017).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGaudelli, N.M. et al. Programmable base editing of A*T to G*C in genomic DNA without DNA cleavage. Nature \u003cstrong\u003e551\u003c/strong\u003e, 464\u0026ndash;471 (2017).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eLosey, H.C., Ruthenburg, A.J. \u0026amp; Verdine, G.L. Crystal structure of Staphylococcus aureus tRNA adenosine deaminase TadA in complex with RNA. Nat Struct Mol Biol \u003cstrong\u003e13\u003c/strong\u003e, 153\u0026ndash;9 (2006).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003ede Moraes, M.H. et al. An interbacterial DNA deaminase toxin directly mutagenizes surviving target populations. Elife \u003cstrong\u003e10\u003c/strong\u003e(2021).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMok, B.Y. et al. A bacterial cytidine deaminase toxin enables CRISPR-free mitochondrial base editing. Nature \u003cstrong\u003e583\u003c/strong\u003e, 631\u0026ndash;637 (2020).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGuo, J. et al. DdCBE mediates efficient and inheritable modifications in mouse mitochondrial genome. Mol Ther Nucleic Acids \u003cstrong\u003e27\u003c/strong\u003e, 73\u0026ndash;80 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eKang, B.C. et al. Chloroplast and mitochondrial DNA editing in plants. Nat Plants \u003cstrong\u003e7\u003c/strong\u003e, 899\u0026ndash;905 (2021).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eLee, H. et al. Mitochondrial DNA editing in mice with DddA-TALE fusion deaminases. Nat Commun \u003cstrong\u003e12\u003c/strong\u003e, 1190 (2021).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eLei, Z. et al. Mitochondrial base editor induces substantial nuclear off-target mutations. Nature \u003cstrong\u003e606\u003c/strong\u003e, 804\u0026ndash;811 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eLim, K., Cho, S.I. \u0026amp; Kim, J.S. Nuclear and mitochondrial DNA editing in human cells with zinc finger deaminases. Nat Commun \u003cstrong\u003e13\u003c/strong\u003e, 366 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMok, B.Y. et al. CRISPR-free base editors with enhanced activity and expanded targeting scope in mitochondrial and nuclear DNA. Nat Biotechnol (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMok, Y.G. et al. Base editing in human cells with monomeric DddA-TALE fusion deaminases. Nat Commun \u003cstrong\u003e13\u003c/strong\u003e, 4038 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSilva-Pinheiro, P. et al. In vivo mitochondrial base editing via adeno-associated viral delivery to mouse post-mitotic tissue. Nat Commun \u003cstrong\u003e13\u003c/strong\u003e, 750 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eCho, S.I. et al. Targeted A-to-G base editing in human mitochondrial DNA with programmable deaminases. Cell \u003cstrong\u003e185\u003c/strong\u003e, 1764\u0026ndash;1776 e12 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGallagher, L.A. et al. Genome-wide protein-DNA interaction site mapping in bacteria using a double-stranded DNA-specific cytosine deaminase. Nat Microbiol \u003cstrong\u003e7\u003c/strong\u003e, 844\u0026ndash;855 (2022).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eBlanchet, C., Pasi, M., Zakrzewska, K. \u0026amp; Lavery, R. CURVES + web server for analyzing and visualizing the helical, backbone and groove parameters of nucleic acid structures. Nucleic Acids Res \u003cstrong\u003e39\u003c/strong\u003e, W68-73 (2011).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eTeh, A.H. et al. The 1.48 A resolution crystal structure of the homotetrameric cytidine deaminase from mouse. Biochemistry \u003cstrong\u003e45\u003c/strong\u003e, 7825\u0026ndash;33 (2006).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHollis, T., Ichikawa, Y. \u0026amp; Ellenberger, T. DNA bending and a flip-out mechanism for base excision by the helix-hairpin-helix DNA glycosylase, Escherichia coli AlkA. EMBO J \u003cstrong\u003e19\u003c/strong\u003e, 758\u0026ndash;66 (2000).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHong, S. \u0026amp; Cheng, X. DNA Base Flipping: A General Mechanism for Writing, Reading, and Erasing DNA Modifications. Adv Exp Med Biol \u003cstrong\u003e945\u003c/strong\u003e, 321\u0026ndash;341 (2016).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eKlimasauskas, S., Kumar, S., Roberts, R.J. \u0026amp; Cheng, X. HhaI methyltransferase flips its target base out of the DNA helix. Cell \u003cstrong\u003e76\u003c/strong\u003e, 357\u0026ndash;69 (1994).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMatthews, M.M. et al. Structures of human ADAR2 bound to dsRNA reveal base-flipping mechanism and basis for site selectivity. Nat Struct Mol Biol \u003cstrong\u003e23\u003c/strong\u003e, 426\u0026ndash;33 (2016).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eShi, K. et al. Structural basis for recognition of distinct deaminated DNA lesions by endonuclease Q. Proc Natl Acad Sci U S A \u003cstrong\u003e118\u003c/strong\u003e(2021).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSlupphaug, G. et al. A nucleotide-flipping mechanism from the structure of human uracil-DNA glycosylase bound to DNA. Nature \u003cstrong\u003e384\u003c/strong\u003e, 87\u0026ndash;92 (1996).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eVassylyev, D.G. et al. Atomic model of a pyrimidine dimer excision repair enzyme complexed with a DNA substrate: structural basis for damaged DNA recognition. Cell \u003cstrong\u003e83\u003c/strong\u003e, 773\u0026ndash;82 (1995).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eYang, C.G. et al. Crystal structures of DNA/RNA repair enzymes AlkB and ABH2 bound to dsDNA. Nature \u003cstrong\u003e452\u003c/strong\u003e, 961\u0026ndash;5 (2008).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHendershot, J.M. \u0026amp; O\u0026apos;Brien, P.J. Critical role of DNA intercalation in enzyme-catalyzed nucleotide flipping. Nucleic Acids Res \u003cstrong\u003e42\u003c/strong\u003e, 12681\u0026ndash;90 (2014).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eOlmon, E.D. \u0026amp; Delaney, S. Differential Ability of Five DNA Glycosylases to Recognize and Repair Damage on Nucleosomal DNA. ACS Chem Biol \u003cstrong\u003e12\u003c/strong\u003e, 692\u0026ndash;701 (2017).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eKabsch, W. Xds. Acta Crystallogr D Biol Crystallogr \u003cstrong\u003e66\u003c/strong\u003e, 125\u0026ndash;32 (2010).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eMcCoy, A.J. et al. Phaser crystallographic software. J Appl Crystallogr \u003cstrong\u003e40\u003c/strong\u003e, 658\u0026ndash;674 (2007).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eEmsley, P., Lohkamp, B., Scott, W.G. \u0026amp; Cowtan, K. Features and development of Coot. Acta Crystallogr D Biol Crystallogr \u003cstrong\u003e66\u003c/strong\u003e, 486\u0026ndash;501 (2010).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eAdams, P.D. et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr \u003cstrong\u003e66\u003c/strong\u003e, 213\u0026ndash;21 (2010).\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eJurrus, E. et al. Improvements to the APBS biomolecular solvation software suite. Protein Sci \u003cstrong\u003e27\u003c/strong\u003e, 112\u0026ndash;128 (2018).\u003c/span\u003e\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-2031914/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-2031914/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"An interbacterial deaminase toxin DddA catalyzes cytosine-to-uracil conversion in double-stranded (ds) DNA and enables CRISPR-free mitochondrial base editing, but the molecular mechanisms underlying its unique substrate selectivity remained unknown. Here we report crystal structures of DddA bound to a dsDNA substrate containing the 5'-TC target motif. The structures show that DddA binds to the minor groove of a sharply bent dsDNA and engages the target cytosine extruded from the double-helix. DddA Phe1375 intercalates in dsDNA and displaces the 5' (–1) thymine, which in turn replaces the target (0) cytosine and forms a non-canonical T-G base-pair with the juxtaposed guanine. This “domino effect” mechanism allows DddA to locate the target cytosine without flipping it into the active site. Biochemical experiments show that DNA base-mismatches enhance DddA deaminase activity and relax its sequence selectivity. Based on the structural information, we further identified DddA mutants that exhibit attenuated activity or altered substrate preference. Our studies may help design novel tools useful in genome editing or other applications.","manuscriptTitle":"Structural basis of sequence-specific cytosine deamination by double-stranded DNA deaminase toxin DddA","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2022-11-21 22:04:34","doi":"10.21203/rs.3.rs-2031914/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-structural-and-molecular-biology","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"nsmb","sideBox":"Learn more about [Nature Structural \u0026 Molecular Biology](http://www.nature.com/nsmb/)","snPcode":"","submissionUrl":"","title":"Nature Structural \u0026 Molecular Biology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"9af0e0e7-26f8-4e81-9162-a11699a7fb84","owner":[],"postedDate":"November 21st, 2022","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":17082740,"name":"Biological sciences/Biochemistry/Enzyme mechanisms"},{"id":17082741,"name":"Biological sciences/Structural biology/X-ray crystallography"},{"id":17082742,"name":"Biological sciences/Molecular biology/DNA metabolism"}],"tags":[],"updatedAt":"2023-07-21T07:09:52+00:00","versionOfRecord":{"articleIdentity":"rs-2031914","link":"https://doi.org/10.1038/s41594-023-01034-3","journal":{"identity":"nature-structural-and-molecular-biology","isVorOnly":false,"title":"Nature Structural \u0026 Molecular Biology"},"publishedOn":"2023-07-17 04:00:00","publishedOnDateReadable":"July 17th, 2023"},"versionCreatedAt":"2022-11-21 22:04:34","video":"","vorDoi":"10.1038/s41594-023-01034-3","vorDoiUrl":"https://doi.org/10.1038/s41594-023-01034-3","workflowStages":[]},"version":"v1","identity":"rs-2031914","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-2031914","identity":"rs-2031914","version":["v1"]},"buildId":"-HB7Z8yhvgn0wM9Nzuekk","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}