Improved nucleoside (2’-deoxy)ribosyltransferases maximize enzyme promiscuity while maintaining catalytic efficiency

Nucleoside analogues have been extensively used to treat viral and bacterial infections and cancer for the past 60 years. However, their chemical synthesis is complex and often requires multiple steps and a dedicated synthetic route for every new nucleoside to be produced. Wild type nucleoside 2′ -deoxyribosyltransferase enzymes are promising for biocatalysis. Guided by the structure of the enzyme from the thermophilic organism Chroococcidiopsis thermalis PCC 7203 (CtNDT) bound to the ribonucleoside analogue Immucillin-H, we designed mutants of CtNDT and the psychrotolerant Bacillus psychrosaccharolyticus (BpNDT) to improve catalytic efficiency with 3 ′ -deoxynucleosides and ribonucleosides, while maintaining nucleobase promiscuity to generate over 100 distinct nucleoside products. Enhanced catalytic efficiency towards ribonucleosides and 3 ′ -deoxyribonucleosides occurred via gains in turnover rate, rather than improved substrate binding. We determined crystal structures of two engineered variants as well as kinetic parameters with different substrates, unveiling molecular details underlying their expanded substrate scope. Our rational approach generated robust enzymes and a roadmap for reaction conditions applicable to a wide variety of substrates. Nucleoside analogues (NAs) are used to treat cancer, viral and bacterial infections. 1 They are challenging to produce synthetically, and enzymatic routes are an alternative to access novel analogues.2 Merck’s biocatalytic synthesis of the islatravir demonstrated the feasibility of an in vitro fully biocatalytic cascade for the synthesis of an anti-HIV NA. 3 Moreover, incorporating nucleoside 2 ′ - deoxyribosyltransferase (NDT) enzymes for the fully biocatalytic production of the NA anti-Sars- Cov-2 drug Molnupiravir has been proposed. 4 Novel nucleoside and nucleotide analogues chemically synthesized for targeting human cancers and bypassing resistance to common treatments have been proposed, demonstrating scope for future work on nucleoside development. 5 Importantly, recent work demonstrated the versatility of NDTs and other enzymes from nucleoside salvage pathways to generate new NAs. 6-7 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 11, 2025. ; https://doi.org/10.1101/2025.03.05.641663doi: bioRxiv preprint Figure 1: Nucleoside and nucleobase substrate scope and possibilities for novel nucleosides using CtNDT or BpNDT. Left: 1) A nucleoside donor is chosen between a ribonucleoside, a 2 ′deoxy- ribonucleoside or 3 ′deoxyribonucleoside. 2) An enzyme catalyst is chosen based on nucleoside donor identity and reaction conditions since Ct NDT operates at high temperatures, while BpNDT operates at low temperatures. 3) A base acceptor is also chosen. Right: Nucleosides produced with a combination of CtNDT or BpNDT variants. HRMS data for each product generated are on Figure S1 and Table S2. Enzymes that display broad substrate scope and the capacity to catalyse more than one type of reaction are desirable, albeit less explored. 8 Here we employed two nucleoside 2 ′ - deoxyribosyltransferases, the enzyme from the thermophilic organism Chroococcidiopsis thermalis PCC 7203 ( CtNDT) and the enzyme from the psychrotolerant organism Bacillus psychrosaccharolyticus ( BpNDT) to produce a series of over 50 NAs, several of which unprecedented (Figure 1). Furthermore, we explored the effects of reaction conditions, including pH, temperature, substrate concentrations and usage of a couple enzyme in efforts to drive the reaction equilibrium towards desired products and decrease nucleoside hydrolysis uncoupled from nucleoside transfer. We established a workflow to optimize reaction conditions and determine factors affecting reaction yield, with implications for others working in nucleoside biocatalysis. We also determined the structure of a double mutant with improved catalytic efficiency towards 3 ′ - deoxynucleosides and ribonucleosides, establishing a path to guide future NDT engineering which relies on modulating the sterics and electrostatics surrounding a conserved and essential catalytic glutamate residue. Inspired by the structures of Trypanossoma brucei (TbNDT) 9, BpNDT10 and our recent work on CtNDT11, we performed structure-based enzyme engineering to produce modified CtNDT enzyme variants with expanded substrate scope. According to reports with other NDT enzymes 7, 12, we first explored the natural promiscuity of CtNDT and BpNDT. CtNDT has a preference for .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 11, 2025. ; https://doi.org/10.1101/2025.03.05.641663doi: bioRxiv preprint purine-based 2′ -deoxynucleosides, while Bp NDT prefers pyrimidine-based 2 ′ -deoxynucleosides (Figure S1). Both enzymes can utilize a vast array of non-canonical nucleobases as substrates, with CtNDT possessing a broader substrate scope. Figure 1 depicts the NAs produced here, employing wild type and enzyme variants discussed below. Figure S1 and Table S2 show more details on substrates tested and products obtained. Importantly, wild type CtNDT can utilize ribonucleoside and 3 ′ -deoxynucleoside substrates, albeit with lower efficiency, demonstrating less strict selection of 2 ′ - deoxynucleosides, and a possible strategy to produce 3 ′ -deoxynucleosides.7, 13 As proof of concept, we determined cordycepin (3 ′ - deoxyadenosine), a natural product currently employed to treat various types of cancers, can be used as a substrate by CtNDT. 14 This enabled the generation of other 3 ′ -deoxynucleoside derivatives using CtNDT variants. Importantly, bulky nucleobases such as 6-(benzyloxy)-9H- purine were accepted as substrates, as well as 3- aminopyridin-2(1H)-one, which is not typically considered a nucleobase. CtNDT has an average sized substrate binding pocket (Figure S3), with a solvent accessible volume of ~ 150 /i1 3, smaller than Lactobacillus leichmannii (LlNDT, ~ 170 /i2 3), which was also shown to accept many different nucleobases and 2 ′ -deoxynucleoside as substrates.7, 15 A flexible loop that acts as a “lid” and could potentially allow bulkier substrates to be used was observed in an “open” or “closed” conformation (Figure 2a and Figure S2). The synthesis of purine nucleosides with bulkier, expanded bases using traditional chemical synthesis has been challenging, despite some showing promise as anticancer and antiviral compounds 16. An asparagine close to the 3 ′ -OH group of ribonucleosides has been proposed as a “gatekeeping residue” controlling acceptance of ribonucleoside substrates, and it is replaced by an aspartate on other enzymes that do not utilize ribonucleosides as substrates. CtNDT lacks this “gatekeeping” asparagine (N53 on TbNDT and D62 on Ct NDT). Recent work has exploited a Figure 2: Structural features of en- zyme variants characterized here. a ) O v e rl ay o f s tr uc tur e s o f w il d ty pe CtNDT (pdb 8PQP, purple/pink) and CtNDT Y7F bound to ImmucillinH (teal/green). Surface colored to depict electrostatic potential. Loop covering active site and Gln46 occupy different positions in the open and closed con- formation shown. b) Same overlay as depicted in a, rotated to display the nucleobase and sugar conformations. c) Overlay of structures of CtNDT Y7F (teal/green) and CtNDTY7F_A9S (blue) bound to ImmH and cordycepin, re- spectively). d) Details of interactions between Ct NDT Y7F_A9S and cordy- cepin. Key distances are shown, and residues mutated (Tyr7, Ser9) are shown in dark blue. Loop covering the active site depicted for reference. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 11, 2025. ; https://doi.org/10.1101/2025.03.05.641663doi: bioRxiv preprint Figure 3: Kinetic characterization of enzyme variants with Ino and Ade as substrates: a) Le raw HPLC data used to obtain initial rates at different substrate concentrations (inset) and genera Michaelis-Menten plots (right). b) Summary of kinetic parameters obtained for each enzyme varia Experiments were conducted in trip licate and data are shown as average and spread error deriv from standard error of fits. c) Residues mutated in variants under study. On the top the over tetramer is shown, and in the bottom a zoomed- in image of the substrate (cordycepin, dark blue) a mutated residues (marine blue), depicted here from CtNDTY7F-A9S. double mutant (residues equivalent to Y7F/D62N in CtNDT) in the enzyme from L. leichmannii 15 to improve acceptance of ribonucleoside substrates, but an improvement in product conversion was not observed for this mutant. A comparison between the structures of CtNDT mutants and LhNDT and Ll NDT is shown on Figure S3. The mutant Ct NDT D62N does not display improved kinetic parameters with ribonucleoside substrates (Figure 3), hence a different nucleoside selection strategy is likely taking place. Prior work carried out the mutation of a tyrosine to phenylalanine in the vicinity of the sugar binding pocket of NDT enzymes to improve ribonucleoside substrate utilization. 17 We have previously shown that this tyrosine residue in CtNDT (Y7) does not directly interact with the 2′ -OH group, and instead positions the catalytic E88 for reaction. 11 This tyrosine residue is conserved in NDT enzymes, and further explored below. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 11, 2025. ; https://doi.org/10.1101/2025.03.05.641663doi: bioRxiv preprint Figure 4: Steps to optimize reaction conditions and determine kinetic parameters, while also im- proving reaction yield. After choosing a CtNDT variant to test (depending on the nucleoside sub- strate as wild type, CtNDT Y7F and CtNDTY7F-A9S were shown to possess different preferences), step 1 is to determine the best pH and tempera- ture to monitor the reaction, step 2 is to investi- gate whether xanthine oxidase can improve reac- tion yield, step 3 is to determine optimal substrate ratio, and step 4 is to determine the optimal reac- tion time, if the objective is to improve yield. When time is evaluated, one must also consider enzyme stability in the pH/temperature range. In the bottom right, optimal conditions were estab- lished. Crucial to our engineering efforts, as well as future applications of NDTs as biocatalysts, we obtained crystal structures of engineered mutants bound to ribonucleoside analogues, providing atomic-level detail into how changes in the ribosyl-binding pocket accommodate these modified substrates (Figure 2 and Tables S4 and S5). To understand the interaction between CtNDT and ribonucleosides, we co-crystallized the variant CtNDT Y7F, which possesses higher catalytic efficiency with ribonucleoside substrates than CtNDT WT with the nucleoside analogue Immucillin-H (ImmH). This analogue is a purine nucleoside phosphorylase inhibitor, rationally designed to mimic the transition state of the reaction catalysed by that enzyme. Importantly, it is a non-hydrolysable analogue containing hydroxyl groups on positions 2 ′ and 3′ . Figure 2a compares the structures of CtNDTWT and CtNDTY7F, depicting key residues participating in the reaction and substrate selection. No major differences are present in the protein backbone, but the wild-type enzyme binds ImmH in a distorted conformation in relation to the ribose ring, bringing the 5 ′ -OH closer to D82, likely to avoid clashes with the 2 ′ - OH of ImmH (Figure 2a). A Dali search indicates the closest NDT homologues in the PDB are the proteins from Trypanosoma cruzi (pdb 2f67 9, rmsd 0.85 Å), Leishmania mexicana (6qai 18, rmsd 0.87 Å), and Lactobacillus helveticus (1s2g19, rmsd 0.89 Å – henceforth referred to as LhNDT). LhNDT complex with 2 ′ - deoxyadenosine allows a comparison between residues interacting with base and nucleoside moieties (Figure S3a). Following our interaction map with ImmH, a ribonucleoside analogue, we designed mutants targeting residues surrounding the 2 ′ -OH group, aimed at further tailoring substrate selection towards modified sugars. Our rationale was centred in ( i) opening up the substrate binding pocket by mutating P40 and Y7 and ( ii) and exploring additional hydrogen bonding interactions to better position a 2 ′ -OH group by mutating A9 and V85. Single mutants CtNDT A9S, CtNDTV85S, CtNDTY7A, CtNDTP40A, double mutant CtNDTY7F-A9S and triple mutants CtNDTY7F-A9S-V85 and CtNDTY7F-A9S-P40A were evaluated with inosine as sugar donor and adenine as sugar acceptor. These mutants had modest changes on K M-inosine, and the most pronounced effects were driven by kcat (Figure 3 and Figure S4) Figure 2b shows the complex structure of CtNDT Y7F-A9S and Cordycepin. A potential additional interaction between S9 and the catalytic E88 (distance 2.5 Å) is formed. This could be important to compensate for the loss in positioning previously conferred by Y7, mutated to a phenylalanine in this variant. In agreement with this, CtNDT Y7F-A9S showed an improvement on kcat/KM-inosine of eight-fold in comparison to wild type, and a two-fold improvement towards CtNDT Y7F. Furthermore, as depicted in Figure 2, CtNDTY7F and CtNDTY7F-A9S ligands harbouring a 2 ′ -OH adopt a substrate-like conformation, unlike what is observed in CtNDTWT bound to ImmH, which binds the ribosyl moiety in a distorted conformation. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 11, 2025. ; https://doi.org/10.1101/2025.03.05.641663doi: bioRxiv preprint Since pyrimidines were not efficient substrates for CtNDT, we turned to BpNDT as it was previously shown to use 2'-deoxynucleoside pyrimidine substrates. 10, 20 In our enzymatic assays, analogues of 2-deoxyuridine were produced. Given 5-fluoro-2 ′ -deoxyuridine (Floxuridin) and capecitabine are FDA approved to treat different cancers, new analogues are desirable. We generated the mutant BpNDTY5F, which is equivalent to CtNDTY7F since BpNDTWT cannot utilize ribonucleosides as substrates. BpNDTY5F uses guanosine and uracil as substrates to produce 5-fluorouridine as a product (Figure S1, Table S2). Cordycepin and clofarabine were not substrates for Bp NDTWT or BpNDTY5F, demonstrating this enzyme has a narrower substrate scope than CtNDT. Both BpNDT21 and CtNDT22 have been previously immobilized, further illustrating the potential of these enzymes in nucleoside production. To demonstrate the applicability of employing CtNDT, CtNDTY7F-A9S and BpNDT to produce novel compounds, we developed a substrate scope and reaction condition testing matrix as summarized on Figure 1 and detailed on Figure 4. Depending on the enzyme variant, an optimal nucleoside is chosen to act as “sugar donor”, while nucleobase is varied. When sufficient quantities of nucleobase are available, increasing the ratio nucleobase/nucleoside can increase reaction yields to up to 99% when considering the limiting substrate. Other factors influence reaction yields, including pH, temperature and reaction times, and we hypothesize this is due to reaction kinetics x reaction equilibrium when different substrates are employed. In cases where 2′-deoxyinosine acts as nucleoside sugar donor, adding xanthine oxidase increases product formation specially in earlier time points, but since uric acid is a substrate for Ct NDT with a k cat/KM-uric_acid = 0.22 mM-1s-1 in the same range as observed for guanine for example (0.33 mM-1s-1), over time it is consumed as a nucleobase substrate, shifting the equilibrium towards initial conditions. Applying our condition matrix, we determined yield and developed a purification protocol for 2- fluoro-3 ′ -deoxyadenosine (or 2F-3 ′ d-Ado or 2F- cordycepin, compound 3'deoxy-3), and 2- fluoroadenosine (compound ribo-3) using both substrates at a 1:1 ratio, and producing N-(9H- purin-6-yl)benzamide (compound ribo-6) and 2 ′ - deoxy-2-amino-adenosine (compound 2'deoxy- 54) using a 10:1 nucleobase to nucleoside substrate ratio. Yields were 13%, 80%, 78%, and 73% respectively. 2F-cordycepin was previously synthesized with 2% overall final yield starting from adenosine and shown to be a potent anti- trypanosomal compound 23. Furthermore, NDTs can also act as nucleoside hydrolases in the absence or under limiting concentrations of the nucleobase substrate in the second half reaction. We evaluated nucleoside substrate hydrolysis to have a variable effect on reaction yield (from no hydrolysis to up to 23% hydrolysis of 3'deoxy- adenosine when CtNDT Y7F-A9S was employed (Figure S5). In summary, we engineered NDT enzymes to broaden substrate scope and generate over 50 nucleoside analogues, several of which currently have no synthetic route proposed. We established a workflow to determine optimal reaction conditions and generated a mutant ( CtNDT Y7F- A9S) with altered nucleoside substrate specificity towards 3'-deoxynucleosides and ribonucleoside substrates. ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website. Materials and methods, raw data for LC-MS and HPLC traces, intact mass for proteins employed here, saturation curves and additional figures and tables are available. AUTHOR INFORMATION Corresponding Author *Clarissa Melo Czekster – School of Biology, Biomedical Sciences Research Complex, University of St Andrews, St Andrews, Fife KY16 9ST, United Kingdom; orcid https://orcid.org/0000-0002-7163-4057; Email: cmc27@st- andrews.ac.uk Author Contributions C.M. Czekster, R.G. da Silva, and D.J. Harrison conceptualized the project and discussed data, discussed and proofread the article; T. Lebl provided support with compound characterisation; P. Tang, A. Dickson, C.J. Harding, G.M. Zickuhr, and S. Devi performed experiments, wrote and proofread the manuscript. Funding Sources PT was funded by IBioIC (IBioIC 2020-2-1) and by a University of St Andrews Impact grant, CMC was funded by the Wellcome .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 11, 2025. ; https://doi.org/10.1101/2025.03.05.641663doi: bioRxiv preprint trust (217078/Z/19/Z). CMC and D.H. were funded by research grants from NuCana plc. Notes D.J.H. is part-time employed by NuCana plc. A.D. is employed by Compass Pathways. ACKNOWLEDGMENT We thank the mass spectrometry facility in St Andrews for support in mass spectrometry. ABBREVIATIONS 2′-deoxyribosyltransferase from Chroococcidiopsis thermalis PCC 7203 ( CtNDT, Uniprot K9TVX3) and Bacillus psychrosaccharolyticus ( BpNDT, Uniprot A0A3G5BRZ6), Immucillin-H (ImmH). PDB Accession codes 9EMX - CtNDTY7F-A9S bound to Cordycepin, and 9EMW - CtNDTY7F bound to ImmH.

1. Egli, M.; Flavell, A.; Pyle, A. M.; Wilson, W. D.; Haq, S. I.; Luisi, B.; Fisher, J.; Laughton, C.; Allen, S.; Engels, J.; Grasby, J. A.; Neidle, S., Introduction and Overview. In Nucleic Acids in Chemistry and Biology , Blackburn, G. M.; Gait, M. J.; Loakes, D.; Williams, D. M., Eds. The Royal Society of Chemistry: 2006; p 0. 2. Lapponi, M. J.; Rivero, C. W.; Zinni, M. A.; Britos, C. N.; Trelles, J. A., New developments in nucleoside analogues biosynthesis: A review. Journal of Molecular Catalysis B: Enzymatic 2016, 133, 218-233. 3. Huffman, M. A.; Fryszkowska, A.; Alvizo, O.; Borra- Garske, M.; Campos, K. R.; Canada, K. A.; Devine, P. N.; Duan, D.; Forstater, J. H.; Grosser, S. T.; Halsey, H. M.; Hughes, G. J.; Jo, J.; Joyce, L. A.; Kolev, J. N.; Liang, J.; Maloney, K. M.; Mann, B. F.; Marshall, N. M.; McLaughlin, M.; Moore, J. C.; Murphy, G. S.; Nawrat, C. C.; Nazor, J.; Novick, S.; Patel, N. R.; Rodriguez-Granillo, A.; Robaire, S. A.; Sherer, E. C.; Truppo, M. D.; Whittaker, A. M.; Verma, D.; Xiao, L.; Xu, Y.; Yang, H., Design of an in vitro biocatalytic cascade for the manufacture of islatravir. Science 2019, 366 (6470), 1255-1259. 4. Del Arco, J.; Acosta, J.; Fernandez-Lucas, J., New trends in the biocatalytic production of nucleosidic active pharmaceutical ingredients using 2'-deoxyribosyltransferases. Biotechnol Adv 2021, 51, 107701. 5. Wagner, A. G.; Lang, T. B. D.; Ledingham, E. T.; Ghosh, A.; Brooks, D.; Eskandari, R.; Suthagar, K.; Almo, S. C.; Lamiable-Oulaidi, F.; Tyler, P. C.; Schramm, V. L., Transition State Analogs of Human DNPH1 Reveal Two Electrophile Migration Mechanisms. J Med Chem 2025. 6. Turner N, W. M., Finnigan W, Birmingham W, Heath R, Derrington S, Schnepel C, Hayes M, Smith P, Falcioni F A biocatalytic platform for the synthesis of 2 ′-functionalized nucleoside analogues. . This content is a preprint and has not been peer-reviewed. ed.; ChemRxiv, 2023. 7. Salihovic, A.; Ascham, A.; Taladriz-Sender, A.; Bryson, S.; Withers, J. M.; McKean, I. J. W.; Hoskisson, P. A.; Grogan, G.; Burley, G. A., Gram-scale enzymatic synthesis of 2'- deoxyribonucleoside analogues using nucleoside transglycosylase-2. Chem Sci 2024, 15 (37), 15399-407. 8. Thorpe, T. W.; Marshall, J. R.; Turner, N. J., Multifunctional Biocatalysts for Organic Synthesis. J Am Chem Soc 2024, 146 (12), 7876-7884. 9. Bosch, J.; Robien, M. A.; Mehlin, C.; Boni, E.; Riechers, A.; Buckner, F. S.; Van Voorhis, W. C.; Myler, P. J.; Worthey, E. A.; DeTitta, G.; Luft, J. R.; Lauricella, A.; Gulde, S.; Anderson, L. A.; Kalyuzhniy, O.; Neely, H. M.; Ross, J.; Earnest, T. N.; Soltis, M.; Schoenfeld, L.; Zucker, F.; Merritt, E. A.; Fan, E.; Verlinde, C. L.; Hol, W. G., Using fragment cocktail crystallography to assist inhibitor design of Trypanosoma brucei nucleoside 2-deoxyribosyltransferase. J Med Chem 2006, 49 (20), 5939-46. 10. Fresco-Taboada, A.; Fernández-Lucas, J.; Acebal, C.; Arroyo, M.; Ramón, F.; De la Mata, I.; Mancheño, J. M., 2 ′- Deoxyribosyltransferase from Bacillus psychrosaccharolyticus: A Mesophilic-Like Biocatalyst for the Synthesis of Modified Nucleosides from a Psychrotolerant Bacterium. Catalysts 2018, 8 (1), 8. 11. Tang, P.; Harding, C. J.; Dickson, A. L.; da Silva, R. G.; Harrison, D. J.; Czekster, C. M., Snapshots of the Reaction Coo rdinate of a Thermophilic 2 ′- Deoxyribonucleoside/ribonucleoside Transferase. ACS Catalysis 2024, 14 (5), 3090-3102. 12. Ye, W.; Paul, D.; Gao, L.; Seckute, J.; Sangaiah, R.; Jayaraj, K.; Zhang, Z.; Kaminski, P. A.; Ealick, S. E.; Gold, A.; Ball, L. M., Ethenoguanines undergo glycosylation by nucleoside 2'-deoxyribosyltransferases at non-natural sites. PLoS One 2014, 9 (12), e115082. 13. Tang, P.; Harding, C. J.; Dickson, A. L.; da Silva, R. G.; Harrison, D. J.; Czekster, C. M., Snapshots of the Reaction Coordinate of a Thermophilic 2'- Deoxyribonucleoside/ribonucleoside Transferase. ACS Catal 2024, 14 (5), 3090-3102. 14. Yoon, S. Y.; Park, S. J.; Park, Y. J., The Anticancer Properties of Cordycepin and Their Underlying Mechanisms. Int J Mol Sci 2018, 19 (10). 15. Salihovic, A.; Ascham, A.; Rosenqvist, P. S.; Taladriz- Sender, A.; Hoskisson, P. A.; Hodgson, D. R. W.; Grogan, G.; Burley, G. A., Biocatalytic synthesis of ribonucleoside analogues using nucleoside transglycosylase-2. Chem Sci 2025, 16 (3), 1302-1307. 16. Seley-Radtke, K. L.; Yates, M. K., The evolution of nucleoside analogue antivirals: A review for chemists and non- chemists. Part 1: Early structural modifications to the nucleoside scaffold. Antiviral Res 2018, 154, 66-86. 17. Del Arco, J.; Mills, A.; Gago, F.; Fernandez-Lucas, J., Structure-Guided Tuning of a Selectivity Switch towards Ribonucleosides in Trypanosoma brucei Purine Nucleoside 2'- Deoxyribosyltransferase. Chembiochem 2019, 20 (24), 2996- 3000. 18. Crespo, N.; Sanchez-Murcia, P. A.; Gago, F.; Cejudo- Sanches, J.; Galmes, M. A.; Fernandez-Lucas, J.; Mancheno, J. M., 2'-Deoxyribosyltransferase from Leishmania mexicana, an efficient biocatalyst for one-pot, one-step synthesis of nucleosides from poorly soluble purine bases. Appl Microbiol Biotechnol 2017, 101 (19), 7187-7200. 19. Anand, R.; Kaminski, P. A.; Ealick, S. E., Structures of purine 2'-deoxyribosyltransferase, substrate complexes, and the ribosylated enzyme intermediate at 2.0 A resolution. Biochemistry 2004, 43 (9), 2384-93. 20. Carberry, A. E.; Devi, S.; Harrison, D. J.; da Silva, R. G., Human 2'-Deoxynucleoside 5'-Phosphate N-Hydrolase 1: The Catalytic Roles of Tyr24 and Asp80. Chembiochem 2024, e202400047. 21. Fresco-Taboada, A.; Serra, I.; Fernandez-Lucas, J.; Acebal, C.; Arroyo, M.; Terreni, M.; de la Mata, I., Nucleoside 2'- deoxyribosyltransferase from psychrophilic bacterium Bacillus psychrosaccharolyticus--preparation of an immobilized biocatalyst for the enzymatic synthesis of therapeutic nucleosides. Molecules 2014, 19 (8), 11231-49. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 11, 2025. ; https://doi.org/10.1101/2025.03.05.641663doi: bioRxiv preprint 22. Martínez, S. A. H.; Tang, P.; Parra-Saldívar, R.; Melchor-Martínez, E. M.; Czekster, C. M., Immobilized Nucleoside 2 ′-Deoxyribosyltransferases from Extremophiles for Nucleoside Biocatalysis. ACS Omega 2024. 23. Vodnala, S. K.; Lundback, T.; Yeheskieli, E.; Sjoberg, B.; Gustavsson, A. L.; Svensson, R.; Olivera, G. C.; Eze, A. A.; de Koning, H. P.; Hammarstrom, L. G.; Rottenberg, M. E., Structure-activity relationships of synthetic cordycepin analogues as experimental therapeutics for African trypanosomiasis. J Med Chem 2013, 56 (24), 9861-73. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 11, 2025. ; https://doi.org/10.1101/2025.03.05.641663doi: bioRxiv preprint 9 Insert Table of Contents artwork here .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 11, 2025. ; https://doi.org/10.1101/2025.03.05.641663doi: bioRxiv preprint