Thermodynamic Insights of Base Flipping in TNA Duplex: Force Fields, Salt Concentrations, and Free-Energy Simulation Methods

Open AccessCCS ChemistryRESEARCH ARTICLE1 Feb 2021Thermodynamic Insights of Base Flipping in TNA Duplex: Force Fields, Salt Concentrations, and Free-Energy Simulation Methods Zhaoxi Sun John Z.H. Zhang *Corresponding author: E-mail Address: [email protected] State Key Laboratory Precision Spectroscopy, School Chemistry Molecular Engineering, East China Normal University, Shanghai 200062 Google Scholar More articles by this author NYU-ECNU Center for Computational at NYU Shanghai, Department Chemistry, New York NY, NY 10003 https://doi.org/10.31635/ccschem.020.202000202 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Threofuranosyl nucleic acid (TNA) is an analogue DNA with a shift the internucleotide linkages from wild-type 5’-to-3’ direction 3’-to-2.’ This alteration leads higher chemical stability, less reactive groups, lower conformational flexibility. Experimental observations indicate that these characteristic changes are attributable minimal perturbation interaction network, but thermodynamic stability duplex remains unaltered mutation. We applied equilibrium nonequilibrium free-energy simulations employing three popular assisted model building energy refinement (AMBER) force fields nucleotides investigate mutation-dependent behavior base flipping T (DNA) residue T-to-TFT mutation computationally. The were performed similarly, as described base-paired state. However, after exploring high-energy regions simulations, we observed behaved differently. Previous reports conclude net-neutral excess-salt provided similar results. Nonetheless, our simulation indicated presence excess salt affected stability. barrier along base-flipping pathway was generally elevated upon addition salts, relative height barriers duplexes did not change significantly. phenomenon emphasizes importance adding sufficient salts scheme reproduce experimental condition. Download figure PowerPoint Introduction capacity Watson–Crick (WC) pair various alternative sugars (e.g., hexopyranoses, pentopyranoses, tetrofuranoses).1–8 One simplest analogues threofuranosyl (TNA), which shifted 3’-to-2’ one.9,10 generated has been shown have following benefits: (1) It able form stable DNA, RNA, TNA,4,7,11–16 enables it transfer information or RNA. (2) enzymatic nonenzymatic polymerization nucleotide possible due cross-pairing other polynucleotides.17–23 (3) often assumed be good predecessor RNA,24,25 higher, number groups smaller, fewer side reactions, thus, more faithful copying. (4) Regioselectivity no longer problem only two hydroxyl each sugar moiety polymer.24 (6) A direct outcome linkage alternation covalent bonds between consecutive phosphates, is, 6 5 TNA, reduces flexibility, possibly making suitable storage than DNA. Favorably, network mutation.2 For instance, stacking interactions virtually unchanged TNA-modified Dickerson–Drew dodecamer (DDD). O4’ atoms tetrose also share remarkably those deoxyribose DDD duplex.26 unique functional role makes them one key targets understanding biological processes. genetic code deposited molecule hides inside their triplex structures, inaccessible biomolecules. When activated surrounding conditions intrinsic fluctuation flips outward exposed solvent, enabling bases interact enzymes triggered processes such modifications sequence.27,28 Computer powerful tool study dynamics DNA-related molecules atomic details.29–36 Technically, molecular (MD) usually unable vast difference timescale event time step integrating equations motion. Specifically, penalty about 10 kcal/mol.37–42 As result, millisecond (ms).43 By contrast, ensure numerical motion, MD could exceed several femtoseconds (fs). Such huge gap accessible hard acquire converged statistics computer simulations. Thus, obtain statistically meaningful data, equilibrium-enhanced sampling techniques umbrella sampling44–48 replica-exchange methods49–55 employed. protonation-dependent (P-DBBF) RNA systems, P-DBBF found couple syn-to-anti transformation guanine (G) group G-adenine (A) mismatch.39 Although technique steered (SMD)56–63 frequently employed thermodynamics flipping, they do significant potentials showed performance construction landscape compared techniques.38,64 Successful applications reported recent years include, variation landscapes sulfur (S) substitution G G-cytosine (C) pairs G-thymine (T) mismatches.64 existing publications focus on behaviors systems synthetic nucleobase naturally occurring mutations.39,64–73 influenced moiety, studied. Therefore, selected polynucleotide effect duplex. profiles mutant constructed Comparisons performances different methods provided. All-atom descriptions computational community increase computation power. mostly used derivatives bsc174 OL15,75–77 both based earlier AMBER field named parm9978 its modification bsc0 counterpart.79 dihedral parameter most challenging part development fields. Recent established term angle torsion potentials. physical meaning well-defined; however, pieces evidence suggest closely related contributions electronic structures final tuning field. benchmark studies last-generation similarly describing structural ensemble canonical long brute-force comparable nuclear magnetic resonance (NMR)-derived ones.40 work AT tracts, significantly under fields.38 Additionally, mutated exhibited behavior,64 provide accurate quantitative results substitution-induced pathway. Accordingly, current work, all fields, including oldest newer OL15 bsc1 assess abilities (i.e., duplexes). envisaged ion concentration solvated system might influence kinetic regarding Since polyanions negative charges outer phosphate neutralization occurs Na+ cations. achieve physiological concentrations 0.1 M 0.15 M, obtained altered concentrations.80–83 Meanwhile, absence noted unbiased where phase space explored close structure. Moreover, enhanced some high-free-energy visited, make conclusions inapplicable. aimed performing reaction effect. Preparation original 2’-deoxy-T (l)-alpha-threofuranosyl-T (TFT) previously described.26 crystal structure data derived pdbid 1N1O.26 illustrations depicted Figure 1a, middle compound showing TFT revealed 5’-3’ backbone 3’-2’ backbone. seventh center duplex, likely level terminal fraying effect, led us residue. high-quality charge scheme, AM1-BCC,84,85 reliable amber program, AMBER14SB,86 general (GAFF)87 comparative parameterize residue, work. studied investigating variations DNA<->TNA Three include combination field,75–77 another bsc1,74 field.79 built bsc0, perspective, present represents assessment improvement made incorporated into whole water model, TIP3P.88,89 box (i.e. DNA/TNA) replicated periodic boundary conditions. Then added nonpolarizable spherical counter ions neutralization, well (Na–Cl pairs) profiles. maximum 90 mM consistent study.26 Details preparation included Supporting Information. 1 | (a) An illustration 7th T/TFT TNA/DNA comparison TNA. (b) definition collective variable (CV) describe event. masses (COMs) four (heavy only) circle, T7/TFT7, base, T8 (the 3’ base), T8-A17, define dihedral. Free-energy method stratification bias sampling.38 coordinate had slow degree freedom, important process interest. In case, dihedral, defined centers (COMs), CV. Notably, pseudo-dihedral T8-A17. Only heavy method. illustrated 1b. Similar CVs widely systems.37,39,64 reweighted variational profile (vFEP)90 recover expectations observables original, ensemble. convergence check block averaging method, simulation. production run divided blocks monitor landscape. identical PMF estimates blocks, reached. Hence, nonequilibrated identified omitted, well-equilibrated later comparisons. method38,64 time-dependent biasing potential drive state another. pulling process, ranging 0° 360 °, series smaller segments. bidirectional initiated equilibrated accumulate microscopic works. Crooks’ Equation (CE)91 Bennett Acceptance Ratio (BAR)92 reweight statistics. Further, tested new criterion reweighting proposed originally previous SHAKE93 algorithm, motion degrees freedom coordinates, examine bond-length constraints involving hydrogen minimize bonds.94 Langevin dynamics95 collision frequency 4 ps−1 temperature regulation 300 K. Isotropic position scaling Berendsen barostat regulate pressure. fs integrate cutoff point Å nonbonded real applied, long-range electrostatic forces determined particle mesh Ewald (PME) method.96 AMBER97 16 suite, analyses homemade codes. full details Results Discussion Equilibrium first thing convergence, it. ns, 2 ns estimate profiles, estimated minimum window attained results.38,64 To possible, Information Figures S1 (OL15), S2 (bsc1), S3 (bsc0). Note study, statistical error bootstrap analysis much systematic error.38 hindered mainly elimination rather variance minimization. apply profile. gradual equilibration window. last fluctuations observed. Based outcomes, use overall calculate mean (PMFs). summary Table S2. require lengths equilibration, total times varying differed, resulting summarized S1. empirical inference µs required system, simpler existence fluctuations, even shorter µs).38,64 any result sub-µs should treated care. Nonequilibrium coupled optimal estimator, very robust constructing flipping.38,64 checked applicability cases. illustrative discussion since similar. critical autocorrelation mechanical bias, viz, CV, acquired inefficiency calculated during initial configurational procedure, S4a. bound ps, interval ps. minor value inefficiency, although typical utilized calculations Previously, correlation size dimensionless standard deviation (SD), overlap scalar. if SD scalar, reached, agreed time-invariant monotonically decreasing further sampling. before checking profile, initially glance sample SD-smaller-than-overlap criterion. S4b shows overlapping noticed 5-sample intersected while larger sizes 50 100 samples, seemed satisfied, reached; accordingly, deduced ∼ samples. sought determination plotting iteration dependence S4c. iteration, five samples set considered when converged, remain fluctuate eight iterations, lower, After tenth sampling, reached evolution provides hints simulations.57,98 If converges, SDs states decrease sampling.57,98 Another worth noting biased free itself.98 envisioned useful otherwise PMF). S4d, apparent iterations displayed values, became starting until up attain stratification. convergence-check mentioned above become apparent, explained why calculation efficiency technique. S5 S6. bit better OL15. avoid exaggeration still 50. versus turned consistency methods. S7, presented comparisons using excellent agreement duplexes, validity systems. Under field, (vFEP) reweighting. check, S3. needed per 72 windows reach yielded 288 ns. earlier, concluded iterations) occur 180 segments leading 27 These findings faster flipping. Also, sometimes investigation technique.38,64 necessarily preferred, there nonrigorous parts processing timing data. neither spacing nor optimal, influencing factor comparison. calculation, modest conclusion accuracy Sugar-moiety-mutation-induced main lies moiety. directional, 5’ 3’, position, quasi-trans--diaxial 3’-’2’ phosphodiester linkages. connecting phosphates decreased TNA.4,7,11–12 stability.2,26 investigated whether modeling. 3, From know positions global minima almost same, fashion. simulation, simulating region pretty developed many benchmarks simulations,40 energy, great significance focused shape 3 Comparison same differen