DOI:10.1002/cbic.201600338 Communications Reverse Transcription of Threose Nucleic Acid by aNaturally Occurring DNA Polymerase Matthew R. Dunnand John C. Chaput*[a] Recent advances in polymerase engineering have enabled the virus, which is sold commercially under the trade name Super- replication of xenonucleic acid (XNA) polymers with backbone script II (SSII), can functionasareasonablyefficient threose nu- structures distinct from those found in nature. By introducing cleic acid(TNA)-dependentDNA polymerase.[12,13] TNA is an ar- aselectiveamplification step into the replication cycle, func- tificial geneticpolymer in which the natural ribose sugar found tional XNA molecules have been isolated by in vitro selection in RNA has been replaced with an unnatural four-carbon with binding and catalytic activity.Despite these successes, threose sugar (Scheme 1). The observation that SSII could re- coding and decoding geneticinformationinXNA polymers verse transcribe TNA into DNA was attributed to the A-like hel- remains limited by the fidelity and catalytic efficiency of engi- ical geometry of the TNA/DNA duplex,[15] which closely approx- neered XNA polymerases. In particular,the process of reverse imates the naturalhelical geometry of DNA/RNA hybrids. transcribing XNA back into DNA for amplification by PCR has When used in combinationwith Kod-RI, an engineered TNA been problematic.Here, we show that Geobacillus stearother- polymerase created by combinatorial libraryscreening and mophilus (Bst) DNA polymerase Ifunctions as an efficient and scaffold sampling, the aggregate fidelity for acomplete replica- 2 faithful threose nucleic acid(TNA)-dependent DNA polymerase. tion cycle of DNA TNA DNAis~1.510À ,which is slightly ! ! Bst DNA polymerase generates ~twofold more cDNA with threefold fewer mutationsthan Superscript II (SSII), which was previously the best TNA reverse transcriptase. Notably,Bst also functions under standard magnesium-dependentconditions, whereas SSII requires manganese ions to relax the enzyme’s substrate specificity.Wefurther demonstrate that Bst DNA polymerase cansupport the in vitro selection of TNA aptamers by evolving aTNA aptamertohuman a-thrombin. Recent advances in polymerase engineering have enabled the replication of awide range of xenonucleic acid (XNA) polymers with backbonestructures that are unique relative to those found in natural DNA and RNA.[1–3] Although such systems Scheme1.Constitutional structure for the linearized backboneofthreose have been used to evolve XNA molecules with ligand binding nucleic acid (TNA;left) and DNA (right). TNA is an unnatural genetic polymer composed of repeating a-l-threose sugars vicinallyconnected by 2 ,3 -phos- and catalytic activities,[4–6] the current generation of engineered ’ ’ phodiester bonds.Differences in backbonelinkages are highlighted in red. polymerases function with reduced activity relative to natural polymerases.[7,8] The development of XNA polymeraseswith strong reverse transcriptase activity,inparticular, has posed a higher than one misincorporation per 100 nucleotides.[16] significant bottleneck in the in vitro replicationofXNA poly- In an effort to improvethe fidelity of TNA replication, we re- mers. Although considerable progress has been made in the examined apreviousscreen of commercial polymerasesthat development of DNA-dependentXNA polymerases (also were tested for the ability to copy TNAinto DNA.[12] In that known as XNA synthases),[9] methodologies developed to engi- study,the Geobacillus stearothermophilus DNA polymerase I neer these enzymes are often restricted to DNA-templated re- (Bst) large fragment was identified as one of asmall number of actions.[6,10,11] Consequently,most XNA reversetranscriptases DNA polymerasesthat were capable of extending aDNA have been identified by screeningpolymerases for DNA syn- primerannealed to ashort synthetic TNAtemplate with dNTP thesis activity on an XNA template rather than by directed evo- substrates. More recently,Bst was shown to transcribe and re- lution.[6,12–14] verse transcribe limited stretches of glycerolnucleic acid (GNA) Previously,weshowed that an engineered version of are- polymers,[17,18] as well as copy RNA templatesinto DNA,[19] sug- verse transcriptase isolated from the Moloney murineleukemia gesting that this enzyme could have broader template specific- ity than is typical of most naturallyoccurring DNA poly- [a] Dr.M.R.Dunn, Prof. J. C. Chaput merases. Department of PharmaceuticalSciences, University of California Irvine We began by examining the ability of Bst DNA polymerase Irvine, CA 92697 (USA) E-mail:[email protected] to copy alibrary of TNA templates into DNA. We chose to Supportinginformation for this article can be found under http:// monitorthe polymerase extension reaction by denaturing dx.doi.org/10.1002/cbic.201600338. polyacrylamide gel electrophoresis (PAGE), which makes it pos- ChemBioChem 2016, 17,1804 –1808 1804 2016 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Communications sible to measurethe fractionofprimer that is extendedtofull- full-length DNA product, regardless of whether MnCl2 was length product and the level of truncated products that accu- present or absent from the reactionmixture (Figure 1B). In mulate during the course of the reaction. Alibrary of TNAmol- total, Bst extended ~60 %ofthe DNA primer into full-length ecules was constructed by copying aDNA library into TNA cDNA product, whereas SSII supplemented with manganese using Kod-RI and asynthetic DNA primer carrying an IR800 ions extended only ~25 %ofthe primer to full-length product. dye at the 5’-end. This reaction produced alibrary of IR-labeled This striking difference demonstrated that, within the context TNA templates that were uniform in length (90 nt) but carried of TNA reversetranscription,Bst DNA polymerase functions an internal random region of 50 unbiasednucleotide positions. with higher primer extensionefficiency and lower sequence Bst and SSII werethen challenged to copy the TNA library bias than SSII. To improvethe efficiency of cDNA synthesis, we back into DNA by extending a5’-IR700 labeled-DNAprimer determined the magnesium ion dependency and polymerase with dNTP substrates (Figure 1A). Both reactions wereper- concentration requiredfor optimal TNA-dependent DNA syn- formed at their optimal reaction temperatures (428Cfor SSII thesis by Bst DNA polymerase.Bygradually supplementing the and 558Cfor Bst) in their preferred commercial buffers (First reactionbuffer with MgCl2 or increasing the enzymeconcen- Strand buffer [50 mm Tris·HCl, 75 mm KCl, 3mm MgCl2,pH8.3] tration,wefound that Bst functioned with optimal activity at 1 and Thermpol buffer [20 mm Tris·HCl, 10 mm (NH4)2SO4,10mm 6mm MgCl2 and an enzyme concentration of 1.6 UmLÀ (Fig- KCl, 2mm MgSO4,0.1%Triton X-100, pH 8.8],respectively). Re- ure 2A,B). Under each polymerase’s optimal conditions, a actions were performed in the presence and absence of MnCl2, time-course analysis revealed that Bst DNA polymerase con- which is knowntoimprove polymerase incorporation of non- verted ~80%ofthe startingprimerinto full-lengthcDNA prod- cognatesubstrates by lowering substrate specificity.[20,21] uct, whereas SSII supplemented with manganese ions achieved Consistent with previous reports,SSII was found to generate only ~40 %product conversion(Figure 2C). In both cases, after modest amountsoffull-length DNA product in the presence of an initial burst, the reaction profile reaches aplateau after 2h manganese ions butfailed to extend the primerinthe absence of incubation;this suggests that longerincubation times are of MnCl2 (Figure 1B,Figure S1 in the Supporting Informa- unlikely to yield higher amounts of elongated product. Analy- tion).[13] In contrast, Bst produced nearly equivalent amounts of sis of the gel revealed thatSSII generated noticeable amounts of truncated product, as indicated by agrowingsmear be- neath the full-length product band (Figure S1). The time- course reactions performed under optimal conditions mirrored our initial data and furthersupported the observation that Bst DNA polymerase functions with higherprimer-extensioneffi- ciency andlower sequence bias than SSII. To ensure that the improved activity associated with Bst DNA polymerase did not come at anegative cost to poly- merase fidelity,wemeasured the aggregate fidelity for TNA replication. This term refers to the fidelity observed when aDNA template is taken through acomplete replication cycle (DNA TNA DNA),which is operationally different than the ! ! more restricted view of fidelity as asingle-nucleotide incorpo- Figure 1. SSII and Bst-mediated synthesis of cDNA on alibraryofTNA tem- plates. A) Schematic representation of aTNA reverse transcription reaction. ration event. Aggregate fidelity is an importantparameter of An IR-labeled DNA primer (gray) was annealed to anaı¨ve TNA library (blue) polymerase function, asitreflects the combined effects of nu- and extended with dNTPs.Ndenotesarandomregion containingall four cleotide misincorporations, insertions, anddeletions, as well as nucleobases. B) The amount of full-length cDNA product generated by Bst the sequence context in whichthese mistakes occur.[6,8,9,13, 16] and SSII DNA polymerases. Primer-extensionreactions were performedfor Accordingly,aDNA template of known sequence is transcribed 1hunder standard reaction conditionsinthe presence or absence of MnCl2. Full-length data derived from the gel are showninFigureS1. into TNA by using Kod-RI,PAGE purified, reverse transcribed Figure 2. Bst-mediated TNA reverse