Unusual Base Pair between Two 2-Thiouridines and Its Implication for Nonenzymatic RNA Copying

2-Thiouridine (s2U) is a nucleobase modification that confers enhanced efficiency and fidelity both on modern tRNA codon translation and on nonenzymatic and ribozyme-catalyzed RNA copying. We have discovered an unusual base pair between two 2-thiouridines that stabilizes an RNA duplex to a degree that is comparable to that of a native A:U base pair. High-resolution crystal structures indicate similar base-pairing geometry and stacking interactions in duplexes containing s2U:s2U compared to those with U:U pairs. Notably, the C=O···H–N hydrogen bond in the U:U pair is replaced with a C=S···H–N hydrogen bond in the s2U:s2U base pair. The thermodynamic stability of the s2U:s2U base pair suggested that this self-pairing might lead to an increased error frequency during nonenzymatic RNA copying. However, competition experiments show that s2U:s2U base-pairing induces only a low level of misincorporation during nonenzymatic RNA template copying because the correct A:s2U base pair outcompetes the slightly weaker s2U:s2U base pair. In addition, even if an s2U is incorrectly incorporated, the addition of the next base is greatly hindered. This strong stalling effect would further increase the effective fidelity of nonenzymatic RNA copying with s2U. Our findings suggest that s2U may enhance the rate and extent of nonenzymatic copying with only a minimal cost in fidelity.

oscillation angle.The distances between detector and the crystal were set to 180-300 mm.The data were processed by HKL2000 1 or xia2 2 and DIALS.The structures were solved by molecular replacement by PHASER 3 using the structure of 3ND4 as the search model 4 .All structures were refined by Refmac5 in CCP4i 5 .After several cycles of refinement, some water molecules and metal atoms were added in Coot 6 .Data collection, phasing, and refinement statistics of the determined structures are listed in Supplementary Table S2 and S3.

Synthesis of oligonucleotides
The oligonucleotides for thermodynamic studies were synthesized on K&A H-8-SE -Oligo Synthesizer, with phosphoramidites from ChemGenes (Wilmington, MA) and reagents from Glen Research (Sterling, MA).The oligonucleotides were cleaved from the solid support and deprotected with AMA at room temperature for two hours.The mixtures were lyophilized, then the 2ʹ-TBDMS protecting group was removed by treatment with triethylamine trihydrofluoride (room temp.overnight for s 2 U-containing oligonucleotides, 65°C 2.5 hours for canonical oligonucleotide) and purified by Glen-Pak cartridge column.The purity of oligonucleotides was confirmed by Agilent 6230 TOF LC-MS; further purification by analytical HPLC was used if needed.
The oligonucleotides used for crystallography and for nonenzymatic primer extension studies were prepared on an Expedite 8909 DNA/RNA synthesizer with phosphoramidites from ChemGenes (Wilmington, MA) and reagents from Glen Research (Sterling, MA).Deprotection and purification methods were as above, except that the canonical RNAs for crystal studies were purified by PAGE.
The canonical oligonucleotides used for nonenzymatic primer extension studies were purchased from Integrated DNA Technologies (Coralville, IA).

Melting temperatures
Melting temperatures were measured using an Agilent Cary 3500 UV-Vis Spectrophotometer.For each pair of complementary oligonucleotides, samples were prepared with the desired concentration of the target oligonucleotide in 10 mM Tris-HCl (pH 8.0), 1 M NaCl and 2.5 mM EDTA.200 µL mineral oil was added to the top of the RNA solution in the cuvette to prevent the evaporation of water.Melting curves were collected by following absorbance at 260 nm as a function of temperature using a temperature ramp of 0.5-1°C/min.The readings were collected in heating-cooling cycles with respect to a control sample containing 10 mM Tris-HCl (pH 8.0), 1 M NaCl and 2.5 mM EDTA.The melting temperatures were calculated from the interpolation of sigmoidal curves.Each experiment was repeated eight times, including four sets of data from low to high temperature and four sets of data from high to low temperature.

Synthesis of 2-aminoimidazolium activated mononucleotides and bridged dinucleotides
The synthesis was performed following previously published procedures. 7e activated mononucleotides *A, *C, *U, *G, and *s 2 U were purified by reverse-phase flash chromatography, with a 50 g C18Aq column, over 12CVs of 0-10% acetonitrile in 2 mM TEAB buffer (pH 8.0).The fractions containing desired products were adjusted to pH 9.5-10 with NaOH, aliquoted into Eppendorf tubes, and lyophilized.
The bridged dinucleotides A*A, G*G, and s 2 U*s 2 U were purified by preparative scale HPLC using a C18 reverse phase column eluted over 20 mins in a gradient of 2-10% acetonitrile in 2 mM TEAB buffer (pH 8.0), under a flow rate of 15 mL/min.The fractions containing desired products were adjusted to pH 8 and with HCl, aliquoted into Eppendorf tubes, and lyophilized.

Synthesis of 2-aminoimidazole activated trimers *GAC & *AGG
The 5ʹ-phosphorylated trimers were prepared by solid phase synthesis on a MerMade 6 DNA/RNA synthesizer, deprotected, and purified by reverse phase flash chromatography, with a 50 g C18Aq column, over 12 CVs of 0-10% acetonitrile in 2 mM TEAB buffer (pH 8.0).After lyophilizing to dryness, the 5ʹ-phosphorylated trimers were added to 40 equivalents of 2AI and TPP, and dissolved in dry DMSO with 400 equiv. of TEA.The solution was then added to 40 equiv. of DPDS and incubated at room temperature for 6 hours.The product was then precipitated and purified with the same procedure as the activated mononucleotides.

Nonenzymatic primer extension reactions
The primer/template or the primer/template/blocker complexes were prepared in an annealing buffer with 5X final concentration: 7.5 μM primer, 12.5 μM template, 17.5 μM downstream blocker if needed, 50 mM Tris-Cl pH 8.0, 50 mM NaCl, and 1 mM EDTA.The solution was heated at 85°C for 30 s and then slowly cooled to 25°C at a rate of 0.1°C/s in a thermal cycler machine.
The annealed solution was then diluted into the primer extension reaction to concentrations of 1.5 μM primer, 2.5 μM template, 3.5 μM blocker (for the sandwich system), 200 mM Tris-Cl pH 8.0, and 100 mM MgCl2.
For primer extension reactions with activated monomer and downstream activated trimer, the monomers were freshly prepared as a 50 mM stock solution and the activated trimer was freshly prepared as a 5 mM stock solution.In a 0.2 mL PCR tube, the primer/template reaction solution were place at the bottom, while the activated monomer and trimer stock solution were placed in the lid or on the wall, followed by immediate spin down to mix the solution and initiate the reaction.
At each time point, 0.5 μL of reaction sample was added to 25 μL quench buffer containing 25 mM EDTA, 1X TBE, and 4 μM of an complementary RNA/DNA to the template in formamide.
All reactions used the primer /FAM/AGU GAG UAA CGC.The template and complementary RNA sequence were listed as below (5ʹ→ 3ʹ).

2
UUA CUC ACU AGU GAG UAA CGC A GAC s 2 U *GAC GUC s 2 U GCG UUA CUC ACU AGU GAG UAA CGC A GAC A *GAC GUC A GCG UUA CUC ACU AGU GAG UAA CGC U GAC G *GAC GUC G GCG UUA CUC ACU AGU GAG UAA CGC C GAC C *AGG CCU C GCG UUA CUC ACU AGU GAG UAA CGC G AGG The Michalis-Menten experiment in Figure 4 was performed similarly, except that 2X stock solutions of bridged dinucleotides were freshly prepared and added to the primer/template buffer solution to initiate the primer extension reactions.All reactions used the primer 5ʹ -/FAM/AGU GAG UAA CGG and the blocker 5ʹ-OH-G AUG UCA GAU AU.The template and complementary DNA sequence were listed as below (5ʹ→ 3ʹ).CAU CAA CCG UUA CUC ACU AGT GAG TAA CGG TTG ATG TCA GAT AT s Us 2 U AU AUC UGA CAU Cs 2 Us 2 U CCG UUA CUC ACU AGT GAG TAA CGG AAG ATG TCA GAT AT Primer extension competition experiments The primer-template duplex was first annealed in a 16 µL solution containing 4 µM primer (/FAM/AGU GAG UAA CGC), 4.8 µM template (5ʹ-OH-GUC s 2 U GCG UUA CUC ACU), 50 mM Tris-HCl pH 8.0, 50 mM NaCl, and 1 mM EDTA.The solution was heated at 85°C for 30 s and then slowly cooled to 25°C at a rate of 0.1°C/s in a thermal cycler machine.The annealed product was then added to 16 µL of 1M Tris-HCl pH 8.0 and 8 µL 1 M MgCl2.An equimolar mixture of *s 2 U, *A, *C, and *G was freshly prepared at a 50 mM total concentration and the activated trimer *GAC was prepared at 5 mM.The reaction was initiated by adding the 32 µL of the *N stock and 8 µL of the *GAC stock to the primer-template solution to make the final reaction concentration of 5 µM primer, 6 µM template, 200 mM Tris-Cl pH 8.0, 100 mM MgCl2, 20 mM *N (5 mM *s 2 U, *A, *C & *G) and 0.5 mM *GAC.After 10 min, the reaction mixture was added to 16 µL of 0.5 M EDTA, 16 µL of 5 M ammonium acetate, and 288 µL of cold ethanol.The solution was allowed to precipitate on dry ice for 1 h before being centrifuged at 15,000 rpm for 15 min, and then washed twice with 80% ethanol.The pellet was resuspended in LC-MS grade water and desalted by ion pairing reverse phase (IP-RP) purification on multiple C18 ZipTip pipette tips (Millipore, Billerica, MA).The tip was wetted with acetonitrile and then with 2M TEAA prior to sample binding.After extensive washing with 10 mM TEAA, the sample was eluted with 50% acetonitrile, dried under vacuum and resuspended in LC-MS grade water.The eluted sample was loaded onto an Agilent 1200 HPLC coupled to an Agilent 6230 TOF equipped with a solvent degasser, column oven, autosampler, and diode array detector.The sample was separated by IP-RP-HPLC on a 100 mm × 1 mm Xbridge C18 column with 3.5 µm particle size (Water, Milford, MA) with (A) 200 mM 1,1,1,3,3,3-hexafluoro-2-propanol with 1.25 mM triethylamine, pH 7.0 and (B) methanol.The sample was eluted between 2.5% and 20% B over 28.5 min with a flow rate of 125 µL/min at 60°C.The sample was analyzed in negative mode from 239 m/z to 3200 m/z with a scan rate of 1 spectrum/s, drying gas flow of 8 L/min at 325°C, nebulizer pressure of 30 psig, capillary voltage of 3500V, fragmentor at 200 V, and skimmer at 65V.Extracted ion chromatograms were generated with the Find by Formula algorithm in Agilent's MassHunter Qualitative Analysis software using a chemical formula database of possible primer extension products.The database was generated by calculating the composition of all possible random RNA extension products.

Figure S2 .
Figure S2.Linear least-squared fits of a Van't Hoff plot of inverse melting temperature (Tm -1 )collected from optical melts at different oligonucleotide concentrations.See Table1of the main text for the sequences of duplexes.

Figure S3 .
Figure S3.Overlap areas of all base steps in duplexes (A) UU1 and SS1 or (B) UU2 and SS2 (with s 2 U modified on the red position).

Figure S5 .Figure S6 .
Figure S5.Mechanism of nonenzymatic primer extension reactions studied in this paper.(A) Primer extension through a monomer-bridged-oligonucleotide intermediate.(B) Primer extension through a bridged dinucleotide intermediate inside a primer-template-blocker system.