Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

2,6-diaminopurine promotes repair of DNA lesions under prebiotic conditions


High-yielding and selective prebiotic syntheses of RNA and DNA nucleotides involve UV irradiation to promote the key reaction steps and eradicate biologically irrelevant isomers. While these syntheses were likely enabled by UV-rich prebiotic environment, UV-induced formation of photodamages in polymeric nucleic acids, such as cyclobutane pyrimidine dimers (CPDs), remains the key unresolved issue for the origins of RNA and DNA on Earth. Here, we demonstrate that substitution of adenine with 2,6-diaminopurine enables repair of CPDs with yields reaching 92%. This substantial self-repairing activity originates from excellent electron donating properties of 2,6-diaminopurine in nucleic acid strands. We also show that the deoxyribonucleosides of 2,6-diaminopurine and adenine can be formed under the same prebiotic conditions. Considering that 2,6-diaminopurine was previously shown to increase the rate of nonenzymatic RNA replication, this nucleobase could have played critical roles in the formation of functional and photostable RNA/DNA oligomers in UV-rich prebiotic environments.


Exposure of nucleic acid strands to ultraviolet (UV) irradiation results in the formation of lesions which effectively suppress the key biological functions of DNA and may result in mutagenesis, immunosuppression, and skin cancer1,2,3. Cyclobutane pyrimidine dimers (CPDs) are among the most frequent photolesions and are formed in a [2 + 2] photocycloaddition reaction between the C=C bonds of adjacent pyrimidine bases4,5,6. While CPDs have a significant effect on the structure of the strand and its recognition, Nature developed complex machinery, i.e., photolyase enzymes, which locate the lesion and initiate the repair with photoinduced electron transfer from a flavin chromophore (see Fig. 1 and Supplementary Fig. 1 in the SI)7,8. However, photolyases evolved only after early living organisms emerged and, therefore, could not counteract photodamage formation in the very first oligonucleotides on Earth9. Therefore, more primitive molecular mechanisms of photodamage repair are likely to have protected the earliest biological systems from the ever-present threat of excessive UV-induced genetic mutation.

Fig. 1: Schematic representation of the self-repair mechanism of T=T dimers promoted by photoinduced electron transfer from 2,6-diaminopurine (D).

This mechanism is analogous to the enzymatic repair induced by the flavin chromophore acting as the electron donor (Supplementary Fig. 1 in the SI).

Recent advances in nonenzymatic template copying of nucleic acids10,11,12,13 and prebiotic formation of their components14,15,16,17,18,19 have brought us near to proposing a credible scenario for the emergence of informational polymers on early Earth. Among several different geochemical scenarios for the formation of purine and pyrimidine nucleosides14,15,16,17,18,19, involvement of UV light offers highest selectivity by driving the key chemical transformations and destroying biologically irrelevant stereoisomers20,21,22,23,24. While high prebiotic UV fluxes could pass through the anoxic Archean atmosphere and support the synthesis and selection of key DNA and RNA building blocks25, intense UV irradiation posed threat to the integrity of their oligomers. The resulting lack of compatibility between UV-assisted formation of nucleosides and the longevity of their oligomers remains one of the key unresolved issues for the origins of informational polymers on Earth. This led to the suggestion that specific DNA sequences could act as potent electron transmitters and allow photochemical charge separation leading to photoreversal of CPDs26. Nevertheless, the most photostable sequences like GATT, were demonstrated to repair only up to 25% of thymine CPDs (denoted as T=T)26. In addition, 8-oxo-purine bases, which could act as good electron donors and promote efficient photoinduced CPD repair27, were shown to nearly completely inhibit nonenzymatic template copying of nucleic acids13.

Previous theoretical works reported that charge-transfer (CT) electronic states in the GAT=T sequence are characterized by high vertical excitation energies and populated only indirectly, outside of the Franck–Condon region of the oligomer28,29. Therefore, involvement of an alternative nucleoside acting as a prebiotic flavin mimic could significantly lower the excitation energies of CT states and increase the self-repairing activity of oligonucleotides.

Here, we demonstrate that 2,6-diaminopurine (Dap) satisfies this criterion and enables highly efficient photoreversal of T=T CPDs. The subjects of our study were canonical and modified DNA trinucleotides containing a purine and two neighboring thymine (Thy) bases as model systems for photodamage repair. Consideration of trinucleotides consisting of deoxyribonucleosides is also consistent with the recent suggestion that DNA and RNA nucleosides could have emerged simultaneously on prebiotic Earth24. The self-repairing activity is manifested by higher T=T photoreversal quantum yields and electron transfer rates in the damaged trinucleotides containing Dap, when compared to canonical trimers containing adenine (Ade). It is worth noting that Dap is an excellent Watson–Crick base pairing partner for Thy and was shown to entirely replace Ade in the genome of phage S-2L of the cyanobacterium S. elongatus30,31. Furthermore, the deoxyriboside of Dap (D) was demonstrated to undergo prebiotic phosphorylation with diamidophosphate in the presence of 2-aminoimidazole, which could be followed by oligomerization with biological deoxyribosides32. Given that Dap could also enhance the rate of nonenzymatic RNA template copying33,34,35, we provide a possible bridge between the UV-assisted synthesis of nucleotides and formation of their first self-replicating oligomers.


Low energies of charge-transfer states lead to high efficiency of self-repair

Calculations of vertical and adiabatic ionization energies (VIE and AIE) provide a good first estimate of electron-donating properties of nucleobases36. Our MP2 calculations indicated that the deoxyriboside of 2,6-diaminopurine (D) has lower VIE and AIE by over 1.8 eV when compared to canonical deoxyribosides of adenine (A) and guanine (G) (see Supplementary Table 3 in the SI). Consequently, we decided to computationally investigate the photophysical properties of DNA trinucleotides containing a CPD with two dimerized thymine bases (T=T) and D or A either on the 5′ or 3′ terminus. For this purpose, we carried out 20-μs long molecular dynamics (MD) simulations of these damaged trimers to identify their most representative stacked conformers and obtain their averaged structures (see Supplementary Figs. 21 and 22 in the SI). We further performed QM/MM calculations in an aqueous environment with the algebraic diagrammatic construction to the second-order [ADC(2)]37,38 QM method, which was shown to provide accurate vertical excitation energies of the key photoactive states in DNA oligomers39, and the TZVP basis set (see the “Methods” section for more details). As expected, the repairing D•+T=T•− state has the excitation energy of 5.02 eV and exhibits configuration mixing with the bright (UV-absorbing) ππ* state on D that is lower in energy by merely 0.14 eV. Even though the repairing T=T•−D•+ state has the vertical excitation energy of 5.62 eV, the bright S1(ππ*) state on D also has a notable contribution from this CT configuration, which implies good self-repairing potential. In contrast, the analogous CT states of AT=T and T=TA have significantly higher vertical excitation energies (5.74 and 5.80 eV, respectively) and do not exhibit any clear mixing with the low-energy ππ* and nπ* states on A (see Supplementary Table 4 in the SI).

Encouraged by these predictions we performed UV-irradiations of the four damaged trinucleotides at 280 nm and analyzed the mixture at different time intervals with HPLC experiments and UV-vis spectroscopy (Fig. 2a). At this wavelength, over 95% of absorbed photons excite the purine base, whereas the extinction coefficients of the CPDs are very small40. As shown in Fig. 2a, irradiations of trimers containing D lead to rapid formation of a photoproduct with corresponding HPLC peaks visible at longer retention times than the damaged trimers. The accumulation of this photoproduct at different irradiation times is proportional to the depletion of the starting material. The differential absorption spectra of the UV-irradiated samples generated with respect to the starting materials (Fig. 2b) demonstrate a gradual increase of absorption at 264 nm, which corresponds to the absorption maximum of recovered Thy chromophores (see Fig. 1 for schematic representation of the self-repair process). In addition, we have validated by LC–MS that the photoproduct corresponds to the repaired Thy bases within the trimers (see section S2.1 in the SI).

Fig. 2: UV-irradiation experiments (λexc = 280 nm, 80 µW cm−2) of damaged DNA trinucleotides containing a T=T dimer and A or D.

a Analytical HPLC analysis of irradiated DT=T, T=TD, AT=T, and T=TA performed after different irradiation times. The peaks at shorter retention times correspond to damaged trinucleotides, while the peaks at longer retentions correspond to the repaired photoproducts. b Differential absorption UV spectra demonstrating changes of the spectrum of the irradiated material at different irradiation times. Both trinucleotides containing D exhibit significant increase of absorption at 264 nm, which corresponds to the absorption maximum of repaired T chromophore. Only minor increase of absorption at 264 nm can be observed for the trinucleotides containing A.

After 45 min of continuous irradiation, DT=T and T=TD underwent thymine recovery in 47.2% and 44.9% yields, respectively (Fig. 2a). In contrast, equivalent exposure of AT=T and T=TA to UV light allowed for self-repair of only 6.1% and 4.9% of the corresponding starting materials. Therefore, presence of the additional amino group in the 2 position of the purine ring significantly enhanced the self-repairing properties. This is also reflected by an order of magnitude higher self-repair quantum yields observed for damaged trimers containing D (see Supplementary Table 2 in the SI). In reality, the overall photostability of specific DNA oligomers is the resultant of rates of various interplaying photoinduced phenomena, including formation of CPDs, self-repair, and direct benign photorelaxation channels of nucleobases.

To provide a better estimate of the photostability of our trinucleotides we performed continuous irradiations of their undamaged forms in the UV-B spectral range to reach the photostationary equilibria (Fig. 3). The irradiation of DTT reached the photostationary state in <20 min with only 8% of CPDs formed, leaving over 90% of the trinucleotide intact (see Fig. 3 and Supplementary Fig. 18 in the SI). TTD also exhibits very high photostability with CPDs formed in <13% yield after the photostationary state is reached. On the contrary, continuous irradiations of ATT and TTA resulted in the formation of T=T dimers in 72% and 76% yields, respectively. This indicates that majority of the material containing the ATT and TTA sequences would be prevented from performing key prebiotically relevant functions such as self-replication.

Fig. 3: Continuous UV-irradiation experiments of undamaged DNA trinucleotides in the UV-B spectral range.

Samples were exposed to continuous UV-B flux (1.95 mW cm−2) to yield photostationary equilibria of T=T dimer formation and self-repair. Analysis of photoproducts was performed by HPLC and % of conversion of undamaged DNA trimers into the damaged trinucleotides was calculated by integration of the respective peak areas (see the SI and the Methods section for more details).

Molecular mechanism of self-repair in modified and canonical trimers

To scrutinize the above differences in photoreactivity of the damaged trimers, we performed further QM/MM calculations with ADC(2) as the QM method and the TZVP basis set. As shown in Fig. 4a, direct electron transfer from UV-excited D to the T=T dimer is energetically favorable and the corresponding S1(CT) minimum lies ~1.5 eV lower in energy than the ring-puckered S1 minimum of photoexcited D. The relative energies of the D•+T=T•− and D*T=T excited-state minima with respect to the ground-state minimum (adiabatic excitation energies) amount to 2.95 and 4.42 eV, respectively. Apart from low vertical and adiabatic excitation energies of the repairing D•+T=T•− state, electron transfer is facilitated by relatively high energy and sloped topography of the crossing between the D*T=T state and the ground state, which lies 0.4 eV above the respective S1 minimum and hinders direct photorelaxation of D. Consequently, UV-excited D is much more likely to transfer an electron to the T=T dimer than undergo direct photorelaxation in DT=T. Similarly, the somewhat lower efficiency of self-repair observed for T=TD can be explained in terms of higher vertical excitation energy of the T=T•−D•+ state and less sloped topography of the S1/S0 state crossing41 enabling smoother direct photorelaxation of D (Fig. 4b). The relatively high yields of self-repair in this trinucleotide are the consequence of a very low adiabatic excitation energy of the T=T•−D•+ state (2.52 eV). Finally, once the negative charge is transferred to the T=T dimer, it induces breaking of the C5–C5 bond in the S1(CT)/S0 state crossing. The remaining C6–C6 bond of the CPD is opened spontaneously in the vibrationally hot electronic ground state28, as schematically demonstrated in Fig. 1.

Fig. 4: Potential energy profiles calculated with the QMbases/MM approach for the key photorelaxation mechanisms of all four studied trinucleotides.

a Blue curve corresponds to self-repair of the T=T dimer triggered by the electron transfer from D. The C5–C5 bond is broken in the D•+T=T•−/S0 state crossing. Direct photorelaxation of D base for DT=T – purple curve. Self-repair is the energetically more favorable process. b Analogous photorelaxation channels in T=TD. Dashed blue line indicates mixing between the T=T•−D•+ and T=TD* states. c The T=T•−A•+ configuration is not energetically available in the S1 state and only the direct photorelaxation of A is presented. d High adiabatic excitation energy of the repairing A•+T=T•− state results in dominant role of direct photorelaxation of A in AT=T. The electronic-structure (QM) calculations were performed at the ADC(2) and MP2 levels with the TZVP basis set. The molecular structures depict the key structural changes for purine bases or the T=T in the S1/S0 minimum-energy conical intersection (MECI).

Negligible self-repairing activity of AT=T and T=TA trimers derives from high adiabatic and vertical excitation energies of the A•+T=T•− and T=T•−A•+ states. In particular, the S1 T=T•−A•+ minimum is lower in energy than the S1 T=TA* minimum corresponding to photoexcited A base by merely 0.10 eV (see Fig. 4d). We were unable to locate the S1 minimum for the A•+T=T•− state, which is the result of an even higher adiabatic excitation energy for this state and lower energies of other valence excitations outside the Frank-Condon region of T=TA (see the SI for discussion). Furthermore, A may undergo much more efficient photorelaxation than D, which is manifested by a peaked S1/S0 state crossing separated from the respective S1 minima by a modest energy barrier of merely 0.3 eV in both AT=T and T=TA trimers (Fig. 4c and d). It is worth noting that while our estimate is an upper bound to the barrier energy, this process is driven by ultrafast C2-H tilting motion, which can be facilitated by H atom tunneling.

As observed for canonical and undamaged DNA trimers and the GAT=T tetramer, the energies of various CT states are strongly affected by the structural arrangement of the oligomer28,42. In particular, Lee and Matsika42 showed that excitation energies of CT states are directly proportional to the distance between the donor–acceptor pair. Similarly, we assign the very low vertical excitation energy of the CT state in the DT=T trimer to a practically perfect spatial alignment of the internal Thy and Dap bases and the short distance between them (see section S2.2 in the SI for detailed discussion). Consequently, higher vertical excitation energy of the CT state in the T=TD trimer (by ~0.6 eV) is consistent with longer distance between the internal Thy and Dap bases. However, we do not see a straightforward correlation between the distance separating the donor and acceptor bases and the adiabatic excitation energies of the D•+T=T•− and T=T•−D•+ states (CT minima).

We provide further interpretation of our experimental observations based on calculations of approximate charge-transfer rates within the Marcus theory of electron transfer and considering couplings between locally excited states on purine bases and CT states, derived from the Boys localization approach43 (see Supplementary Table 5 and the associated discussion in the SI for more details). The estimated rate constants calculated using the geometries and energies from Fig. 4 indicate that charge separation in photoexcited D-containing trimers occurs on a timescale of hundreds of femtoseconds and could indeed outcompete the direct photorelaxation of D (see Table 1). In contrast, the estimated timescale of electron transfer from excited A to T=T is over an order of magnitude longer than in the case of trimers containing D (~5.82 ps for T=TA). Considering that UV-excited A undergoes direct photodeactivation to the electronic ground state within <2.0 ps, its excited-state lifetime is too short to permit photoreversal of the T=T dimer44.

Table 1 Electron transfer rate constants and the corresponding timescales of electron transfer from the purine base to the T=T CPD estimated for three of the studied trinucleotides.

Self-repair of internal CPDs in longer nucleic acid oligomers

The above results demonstrate that the photochemistry of damaged DNA trimers is characterized by the competition between purine photorelaxation and electron transfer processes and that the presence of D could induce efficient self-repair of CPDs on the termini of oligomers. To explore whether D could induce self-repair of internal CPDs, we performed analogous QM/MM simulations for the ADT=TA pentamer having an adenosine residue on each terminus. To examine the excited-state charge-transfer properties of ADT=TA we performed MD simulations and selected a conformer with all five nucleobases stacked, so that it mimics the structural arrangement of even longer single strands. We next performed calculations of vertical excitation energies and optimizations of the key minima of the S1 excited state at the ADC(2)/def2-SVP level (note that a smaller basis set was applied because of increased computational costs). The lowest-lying excited state calculated for the optimized ground-state geometry of ADT=TA has the energy of 4.90 eV and can be characterized as the bright (UV-absorbing) ππ* excitation on D mixed with the AD•+T=T•−A electronic configuration responsible for self-repair (see Supplementary Fig. 26 and the associated discussion in the SI for more details). In addition, the S1(CT) minimum of the AD•+T=T•−A state is lower in energy than the S1 minimum of the locally excited state on D by 0.72 eV. This value, referred to as the driving force of electron transfer, is approximately two times lower in magnitude than in the case of the DT=T and T=TD trimers. However, the CT minimum is still considerably lower in energy than the AD*T=TA (ππ*) minimum, which together with the low vertical excitation energy of the AD•+T=T•−A (πDπTT*) excitation shows that electron transfer from D to the T=T is also an energetically favorable process in the studied pentamer. Therefore, we expect the stacked conformer of ADT=TA to be involved in highly efficient self-repair of its internal CPD.

Our QM/MM simulations indicate that excited-state electron transfer between the D and T=T fragments is comparably as efficient inside a well-stacked DNA single strand as in the investigated DNA trimers. However, according to the classical MD simulations and conformational analysis, the ADT=TA oligomer spent <5% of time in the fully stacked arrangement and it sampled various partly or fully unstacked conformations for the remaining part of the simulation. In contrast, the stacked conformer of AT=T was populated for more than 50% of time during analogous simulations (see the SI for more details). Therefore, any potential irradiation experiments on the ADT=TA damaged pentamer would largely reflect the averaged photoreactivity of random coil structural arrangements. While such random coils would be present on the early Earth, the results would be less conclusive toward internal CPDs. In contrast, longer and more stable DNA double strands exhibit very complex photochemistry with multiple possible photoproducts26, owing to a large number of stacked chromophores. Consequently, we focused here only on short single-stranded oligomers, to clearly disentangle CPD self-repair from other competing photorelaxation channels. Self-repair in a flexible single-stranded DNA oligomer could also operate in partly unstacked conformations. We expect that conformers having a sufficiently close distance between the Dap base and the T=T dimer, and unstacked flanking bases, exhibit very similar photochemistry to the studied DNA trimers. Further experimental work is underway to test these predictions.

Prebiotically plausible synthesis of β-2,6-diaminopurine 2′-deoxyriboside

Considering its photostabilizing activity in DNA trimers, we sought to determine whether β-2,6-diaminopurine 2′-deoxyriboside (β-D) might be available on a prebiotic Earth for incorporation into nucleic acids. Recently, Xu et al.22 reported a proof of principle synthesis of deoxyribosides, including deoxyadenosine, by transglycosylation of 2′-deoxy-2-thiouridine (2tU) with adenine, to provide 4% and 6% of β- and α-A, respectively. The synthesis, although low yielding, demonstrates that a prebiotic world that contained RNA may also have contained small amounts of deoxyadenosine. Since Dap has previously been synthesized in a prebiotic fashion45, and has even been discovered on meteorites46, we reasoned that it may also have accumulated and participated in this transglycosylation reaction. Indeed, after evaporating an aqueous mixture of 2tU and Dap (15 equiv.) and heating to 100 °C for 40 h, β-D was formed (2%) along with at least one isomer (probably the α-diastereomer, 4%) and other minor compounds (see Fig. 5). Thus, the reaction proceeds similarly to that of the reported synthesis of deoxyadenosine, providing similar proof of principle that DNA containing Dap may have indeed been present in a prebiotic world.

Fig. 5: Transglycosylation reaction of 2′-deoxy-2-thiouridine (2tU) furnishing 2,6-diaminopurine 2′-deoxyriboside (D).

Transglycosylation of 2tU in the dry state with D gives a mixture of α and β stereoisomers of the D deoxyriboside alongside loss of the 2tUra base.

Photostability of D vs A

UV irradiation was suggested as the key environmental factor, which selected adenosine, guanosine, cytidine, uridine, and thymidine as the most photostable canonical nucleosides47. Therefore, we performed comparative continuous irradiations of D and A within the UV-C spectral domain, to establish whether D could survive in the prebiotic environment as a potential building block. Our HPLC analysis (Supplementary Fig. 19 in the SI) performed for different irradiation times demonstrates that D undergoes photodegradation approximately two times more rapidly than A when exposed to UV-C radiation (254 nm). Approximately 11.6% and 5.9% of D and A were decomposed, respectively, after 120 min of continuous irradiation (Supplementary Fig. 19 in the SI). Even though D undergoes photodegradation more rapidly than A, the difference in photostability in the UV-C range is not overwhelming. Analogous irradiation experiments performed in the UV-B spectral range resulted in the photodegradation of D in merely 16.4% after 120 min, whereas A remained practically photostable under these conditions and its depletion was below the accuracy threshold of our analyses (Supplementary Fig. 20 in the SI). It worth noting that the measured photon flux in the UV-B region was an order of magnitude higher than in the UV-C range and several times higher than the overall UV flux estimated for the early Earth (see the following section for more details). Therefore, we conclude that D is sufficiently photostable to have been involved in the formation of the first informational polymer.

Based on our ADC(2) calculations, we ascribe the lower photostability of D to the fact that its key photorelaxation mechanisms are associated with out-of-plane rotation of amino groups that require longer timescales than the ultrafast C2-H bond rotation in A48. Thus, despite a modest energy barrier, rotation of the C2-H bond occurs more rapidly in aqueous environment than rotation of the C2-NH2 group, which forms hydrogen bonds with nearby water molecules. Furthermore, the C2-NH2 rotation in D leads to a sloped S1/S0 state crossing that lies 0.3 eV above the C2-puckered S1 minimum (see Fig. 6). This additionally hinders damageless photodeactivation of D. Consequently, the anticipated longer lifetimes of the electronically excited bi-radical state in D are likely to entail destructive bimolecular photoreactions.

Fig. 6: Potential energy profiles for the key photorelaxation mechanisms of 2,6-diaminopurine and adenine 2′-deoxyribosides (D and A).

The electronic-structure calculations for ground and excited states were performed at the MP2 and ADC(2) levels, respectively, using the TZVP basis set. MECP minimum-energy crossing point.

Comparison of irradiation conditions and prebiotic UV fluxes

The integrated surface flux between 200 and 300 nm delivered to the surface of early Earth was estimated to be ~2700 erg s−1 cm−223. Therefore, we aimed to irradiate our samples with UV fluxes that would be consistent with this value within an order of magnitude in terms of intensity. Our tunable UV lamp used for the irradiation of damaged DNA trimers at 280 nm delivered flux of 800 erg s−1 cm−2, which is in good agreement with the UV environment expected for prebiotic Earth. In this experiment, we focused on irradiating our samples with the same wavelength applied previously by Pan et al.40 in analogous irradiation experiments of canonical DNA trimers. Irradiation experiments aiming to determine the photostationary equilibria of undamaged DNA trimers and the photostability of D and A nucleosides delivered higher fluxes of 17,000 erg s−1 cm−2 in the UV-B spectral range. It is important to emphasize that even though we did not perform broadband irradiation experiments that would fully cover the spectrum of the young Sun, our work provides a good comparative assessment of the self-repair properties and photostability of trinucleotides and nucleosides comprising Ade and Dap bases. Consequently, data presented in Fig. 3 and Supplementary Fig. 19 indicate that the D nucleosides and the studied trimers containing D could withstand days of continuous exposure to UV light on early Earth. Broadband irradiation experiments aiming to estimate the half-life of short D-containing DNA oligomers in UV-rich prebiotic environments are currently underway.


In summary, we have demonstrated that D might have served as an important component of the primordial genetic alphabet by significantly enhancing the photostability of DNA oligomers and increasing the self-repair rates of cyclobutane pyrimidine dimers (CPDs). These effects result from a decreased energy threshold for photoinduced charge separation and enhanced efficient intra-strand electron transfer. In particular, D readily inserts an electron into a CPD which is followed by self-repair of the damaged site. Thus, D can be considered as a prebiotic flavin mimic enabling DNA self-repair with ~90% yield. To further reinforce its prebiotic relevance, we have shown that D could have been formed on early Earth alongside A through a transglycosylation reaction. Alternative syntheses of purine (deoxy)nucleosides raise the prospect that D might also be a product of these alternative chemistries, perhaps in more significant yields15,17,24,49,50. Given its only moderately lower photostability relative to A, D could survive long enough in a UV-rich environment to be incorporated into oligomers and play a further role in enzyme-free DNA self-replication. Ongoing questions remain as to the reason and mechanism by which D was eventually removed from the genetic alphabet. It is worth noting that the additional amino group in the 2 position of the purine ring could affect functional tertiary structures of RNA or hinder separation of duplexes. Consequently, if D had provided limited functionality compared to, for example, A, it might have not persisted after the emergence of the first CPD-repairing enzyme. The overarching and unintuitive result of this study is that the formation of highly photostable molecular assemblies could be achieved by joining moderately photostable or even photounstable building blocks, here represented by D and pyrimidine-rich DNA fragments. We anticipate that these findings will shed further light on UV-induced prebiotic selection, as it has been widely postulated that Nature selected only the most photostable building blocks to protect their biomolecular polymers from the photodamage caused by continuous and intense UV irradiation47. The result of this work could be also used to support the design of self-healing macromolecules51 or to enhance the electron transfer properties of DNA52 through the design of photostable synthetic XNAs.


Irradiation experiments


All oligonucleotides (ATT, TTA, DTT, TTD) were purchased from Future Synthesis, Poland. The samples were HPLC-purified by the Future Synthesis. 2′-deoxyadenosine and 2,6-diaminopurine 2′-deoxyriboside were purchased from Merck and Alfa Aesar, respectively. All solutions were prepared using ultra-pure water produced by a Millipore purification system.

Generation of AT=T, T=TA, DT=T, T=TD

Illumination conditions

The thymidine photodimerization was conducted in the pH 7.0 buffer solution of 10 mM phosphate containing 100 mM sodium chloride. All samples were illuminated with UV-B light in quartz tubes with the use of Irradiation Chamber BS-04 (Opsytec Dr. Grӧbel, Ettilngen) at an average power of 1.95 mW measured by a Maestro11 power/energy meter, equipped with an 11PD photodetector (Standa, Lithuania). The accumulation of photoproduct was performed using HPLC (see Supplementary Fig. 4) followed by characterization with the use of mass spectrometry (LC–MS and LC–MS/MS, see Supplementary Figs. 516).

HPLC analysis

Damaged trinucleotides were isolated using reversed-phase HPLC method (Dionex Ultimate 3000 System with a Diode Array Detector, which was set at 260 nm). The mobile phases used in this study were 0.1% HCOOH and 80% ACN in water (gradient elution from 0 to 25% ACN in 15 min) with a flow rate of 1 mL/min. The separations were conducted on the Wakopak Handy ODS column (4.6 × 150 mm, 5 μm in particle size, and 100 Å in pore size).

Mass spectrometry

Ultra-High Performance Liquid Chromatography system, Nexera X2 (Shimadzu) was coupled to ESI-TripleTOF 5600+ (AB SCIEX) mass spectrometer operated in negative ion mode. The separations were performed under the following conditions: Kinetex column (Phenomenex C18, 100 Å, 2.1 × 150 mm); flow rate 0.3 mL/min; a gradient elution with 80% ACN and 0.1% HCOOH in water as mobile phases (from 0 to 25% ACN in 15 min). The MS experiments included the following parameters: −4.5 kV ion spray voltage (ISV); −80 V declustering potential (DP); −10 V collision energy (CE); 200 °C source temperature.

Self-repair of T=T lesions

Illumination conditions

All samples were illuminated at 280 nm in quartz capillaries with the use of a xenon lamp. The 280-nm wavelength of incident light was selected using a monochromator (M250 Optel, Poland). The average incident light intensity for the above irradiation setup was 80 µW. This value was determined with a Maestro11 laser power/energy meter, equipped with an 11PD photodetector (Standa, Lithuania). To provide stable measuring conditions at pH 7, the oligonucleotides were dissolved in water and buffered by phosphate buffer at a concentration of 10 mM. For each illumination experiment, the concentration of trinucleotides was equal to 20 µM. Illumination was performed in different time intervals up to 45 min. Between the illumination steps absorbance spectra from 230 to 340 nm were recorded. All experiments were performed at least in triplicate and followed by HPLC analysis and characterization of repair product by mass spectrometry (LC–MS and LC–MS/MS).

HPLC analysis

Irradiated and non-irradiated samples of trinucleotides were analyzed using the method described in the section “Generation of AT=T, T=TA, DT=T, T=TD”.

Mass spectrometry

All analyses were performed using the method described in the section “Generation of AT=T, T=TA, DT=T, T=TD”.

Photodegradation of D and A

Illumination conditions

The water solution of A and D (10−4 M) containing phosphate buffer (10 mM, pH = 7.0) was illuminated at UV-C radiation in quartz capillaries for 30, 60, 90, and 120 min. All samples were deoxygenated before irradiation by purging with argon and illuminated in Irradiation Chamber BS-04 (Opsytec Dr. Grӧbel, Ettilngen) used at an average power of 170 µW (UV-C) or 1.7 mW (UV-B) measured by a Maestro11 power/energy meter, equipped with an 11PD photodetector (Standa, Lithuania). All experiments were performed at least in duplicate and followed by HPLC analysis. The percentage of photodegradation was calculated by integration of the respective peak areas.

HPLC analysis

Irradited and non-irradiated samples of A and D were analyzed with the use of reversed-phase HPLC method on Dionex UltiMate 3000 System with Diode Array Detector. The detection of the effluents was carried out at 260 nm. Separations were performed using a Wakopak Handy ODS column (C18, 4.6 × 150 mm, 5 μm; 100 Å), 0.1% HCOOH and 80% ACN in water as mobile phases (gradient elution from 0 to 10% ACN in 10 min) and a flow rate of 1 mL/min.

Synthetic procedures

Solid-state glycosylation of 2,6-diamonopurine (Dap) with 2′-deoxy-2-thiouridine (2tU) to form β-2,6-diaminopurine 2′-deoxyriboside (β-D)

0.20 mL of a standard solution (22 mM in MeOH) of 2′-deoxy-2-thiouridine 2tU (1.0 mg, 4.4 μmol, 1.0 equiv.) was evaporated in a 10 mL round-bottom flask, leaving 2tU as a white solid distributed on the flask. 2,6-Diaminopurine Dap (9.3 mg, 66 μmol, 15.0 equiv.), and milliQ-filtered water (pH 8 by addition of traces of NaOH) (0.9 mL) were added and the suspension was sonicated until all the 2′-deoxy-2-thiouridine had been dissolved (the diaminopurine remains insoluble). The pH at this stage was measured at 8. This mixture was then concentrated on a rotary evaporator at 45 °C and ~5 mbar until the solvent had been evaporated, leaving a white deposit on the inside of the flask. The flask was then immersed in an oil bath at 100 °C. After 40 h, the material was dissolved in D2O and 1H NMR spectra recorded. The supernatant of this mixture was used for HPLC and LC–MS analysis, showing the presence of mainly diaminopurine Dap, 2-thiouracil 2tUra, and small quantities of β-2,6-diaminopurine 2′-deoxyriboside (β-D) and an isomer, presumably the α-configured diastereomer (a standard of this material is not available, however, the reaction appears to have a similar outcome, and similar retention times, to the reported reaction between 2′-deoxy-2-thiouridine 2tU and adenine which forms α- and β-configured diastereomers. The isomer has the same molecular mass and an almost identical UV-Vis spectrum to β-D). The identity of β-2,6-diaminopurine 2′-deoxyriboside β-D was confirmed by spiking experiments and measuring by both NMR and HPLC (Figs. S2, S3). The yield could not be determined accurately by NMR because of overlapping signals. Development of an analytical HPLC method (see below) gave conversions after 40 h of 2% and 4% for β-D and the presumed α-diastereomer, respectively. Analytical HPLC was used to monitor the reaction, using a Thermofisher Ultimate 3000 UPLC and Waters Atlantis T3 C18 column (5 μm, 4.6 mm × 150 mm). UV absorption calibration curves for standards of 2′-deoxy-2-thiouridine 2tU, thiouracil 2tUra, β-2,6-diaminopurine 2′-deoxyriboside β-D were collected. UV absorption for known peaks (verified by spiking) was recorded and calibration curves used to determine relative amounts. Recovery was calibrated to aliquots from time zero at which the yield of starting material 2′-deoxy-2-thiouridine 2tU was set to 100%.

Molecular dynamics simulations

Force-field parameters for the T=T dimer were taken from our previous study28. Both the AT=T and T=TA trinucleotides were solvated in a truncated octahedral water box with the distance between solute and box border being at least 10 Å. For each trinucleotide we used two different water models, SPC/E53 and OPC54. The box was first neutralized by 2 K+ ions and then 0.15 M KCl was added, using Joung and Cheatham parameters for SPC/E and TIP4Pew, respectively55. The simulations were carried out with the parmOL15 force field56, which contains several dihedral reparametrizations57,58,59 of the DNA force field developed by Cornell et al.60. Standard equilibration protocol was employed first. Five hundred steps of steepest descent minimization were followed by 500 steps of conjugate gradient minimization with 25 kcal mol−1 A−2 position restraints on DNA atoms. We then imposed position restraints of 25 kcal mol−1 A−2 on the DNA fragment and heated the system from 0 to 300 K over 100 ps. Heating was performed at constant volume. The system was then minimized with 5 kcal mol−1 A−2 restraints on the DNA oligonucleotide using 500 steps of steepest descent method followed by 500 steps of the conjugate gradient optimization approach. The whole system was then equilibrated for 50 ps at a constant pressure of 1 bar and constant temperature of 300 K. We further performed an analogous series of consecutive minimizations and equilibrations applying decreasing restraints on the atomic position of the DNA fragment of 4, 3, 2, and 1 kcal mol−1 A−2. The final equilibration was carried out using position restraints of 0.5 kcal mol−1 A−2 and starting velocities from the previous equilibration step and was followed by free (unrestrained) molecular dynamics simulation for 50 ps. Temperature and pressure coupling constants were set to 0.2 ps during the equilibration, and to 5 ps during the final phase of molecular dynamics. Production phase was started afterward and each system was simulated for 10 μs. The time step was set to 4 fs with application of the hydrogen mass repartitioning61 and the SHAKE62 and SETTLE63 algorithms. The temperature and pressure were set at 300 K and 1 bar, respectively, using Berendsen weak-coupling thermostat and barostat64. Electrostatic interactions were treated using the Particle Mesh Ewald method (PME). The non-bonded cutoff was set to 9 Å. The calculations were performed with the CUDA accelerated pmemd module of AMBER 1665.

Each trajectory was clustered to find the important conformers. We applied a modified15 algorithm of Rodriguez and Laio66, and used eRMSD as a metric67. It is a nucleic acid-specific measure that determines the similarity of two structures based on relative orientation of its nucleobases and is a suitable tool for distinguishing stacked and unstacked conformations or syn- and anti-oriented nucleotides. Only clusters with a population exceeding 1% were selected for further inspection.

QM/MM simulations

Preparation of starting solvated oligonucleotide structures

After analyzing and clustering the trajectories simulated for the AT=T and T=TA trimers we selected the most representative stacked conformers for the QM/MM simulations (see Supplementary Figs. 21 and 22 and the accompanying discussion). These conformers have the structural features of longer oligonucleotides and therefore our qualitative observations and results of the excited-state calculations are also relevant for longer DNA fragments. The stacked structural arrangement of AT=T is also the highest populated conformer in both MD simulations using the SPC/E and OPC solvent models. Even though the most populated conformer of T=TA is an unstacked structure, the stacked conformer is among the most populated structures of this trinucleotide (see Supplementary Figs. 21 and 22). In order to perform QM/MM calculations on equivalent structures of native trimers and those modified with 2,6-diaminopurine (Dap) we substituted the C2-H position of adenine with an amino group and generated the corresponding DT=T and T=TD stacked structures. Each trinucleotide was then solvated in a sphere of explicit water molecules with the radius of 24 Å. The solvent was then equilibrated for 10 ps with position constraints imposed on the solute molecule (trinucleotide). The resulting geometries were used in subsequent ground-state geometry optimizations.

QM/MM setups

The QM/MM calculations were performed in two different partitioning setups involving either the whole trinucleotide in the QM region (QMDNA/MM setup) or just the nucleobases treated at the QM level (QMbases/MM setup), as presented in Fig. 7. The remaining part of the system was described with the same force-field suite as in the MD simulations, including the SPC/E53 force field for water. We applied the link hydrogen atom scheme in the QMbases/MM setup with the MM charge of the link H atom being set to 0. The boundary between the QMbases and MM regions bisected the N-glycosidic bonds. We employed the electrostatic embedding framework to all of the QM/MM calculations. We performed these simulations using our in-house modification of the QM/MM interface68 of the AMBER suite of programs65 to enable QM calculations employing the TURBOMOLE 7.3 program69.

Fig. 7: QM/MM setups used in the geometry optimizations and energy calculations.

Solid balls and sticks models show the two considered QM regions, while transparent residues (water molecules in both setups and the sugar-phosphate backbone on the right) correspond to the MM region described at the force-field level. The MM region interacts with the QM part via point charges (electrostatic embedding). ADC(2)/TZ denotes the ADC(2) method with the TZVP basis set.

Geometry optimizations

We optimized the minimum-energy geometries of the four selected structures of damaged DNA trimers using the cost-effective composite PBEh-3c70 hybrid DFT scheme for the electronic-structure calculations and the QMDNA/MM setup. The PBEh-3c method was recently demonstrated to provide accurate geometries and energies of short nucleic acid fragments71. Initial optimizations of the ground-state geometries were performed for the whole solvated system using the AMBER limited-memory BFGS algorithm. We performed tighter optimization of the internal region of the sphere (within the radius of 12.5 Å from the most central atom of each conformer) using the open-source external geometry optimizer XOPT, which employs the rational function and approximate normal coordinates schemes72,73. The outer region of the sphere was kept frozen during this optimization procedure. This approach was used for all ground-state and excited-state geometry optimizations.

Excited-state QM/MM calculations

Calculations of excited-state energies and gradients within the QM/MM framework were performed using the algebraic diagrammatic construction to the second-order [ADC(2)]37,38,74 method for electronic-structure calculations, whereas corresponding energies and gradients for the electronic ground-state were obtained at MP2 level. All of the ADC(2)/MP2 calculations were performed using the TZVP basis set. Vertical excitation energies contained in the Supplementary Table 4 were calculated using the QMDNA/MM setup on top of the geometries optimized at the PBEh-3c level. Optimizations of different minima on the S1 potential energy surface and S1/S0 minimum-energy crossing points (MECPs) were performed using the QMbases/MM setup. The orbital characters of excited electronic states were determined based on analysis of molecular orbitals and key contributing configurations. In addition, we established the charge-transfer numbers and performed electron-hole population analysis using the TheoDORE 1.5.1 package75,76. MECP optimizations were performed using the algorithm of Levine, Coe, and Martínez77, which allows to locate highly representative geometries of S1(ππ*)/S0 conical intersections. Even though the single-reference ADC(2)/MP2 approach does not provide a correct description of the entire topography of S1/S0 conical intersections, it was demonstrated that it can accurately describe geometries of S1(ππ*)/S0 MECPs in biomolecular systems with conjugated π bonds including qualitatively correct shapes of PE surfaces from the Frank-Condon region up to the state crossing78,79. While much less accurate results were obtained for S1(nπ*)/S0 conical intersections79, in this work, we only considered S1(ππ*)/S0 MECPs with only one dominant electronic configuration contributing to the excited-state character. Therefore, we expect that our approach provides at least qualitatively correct description of the mechanism. The energies of the key stationary points (S0 and S1 minima, MECPs, and vertical excitation energies) were used to construct illustrative PE surfaces in Fig. 3. The barrierless profiles between the Franck–Condon region of each trinucleotide and different S1 minima were confirmed by excited-state geometry optimizations. The barrier heights between the charge-transfer S1 minima and corresponding MECPs responsible for C5–C5 bond cleavage were estimated based on a relaxed potential energy scan along the C5–C5 distance performed for the whole trimer (see Supplementary Fig. 25 for more details). The barrier heights (presented in Fig. 3) between the S1 minima of the purine bases and the corresponding S1/S0 MECPs were taken from the calculations performed for single nucleosides.

Simulations of the ADT=TA pentamer

Classical molecular dynamics simulations and subsequent ground-state geometry optimizations within the QM/MM framework for the ADT=TA pentamer were performed in an analogous way to the AT=T and T=TA trimers. The excited-state QM/MM calculations for this pentamer were also performed in an analogous way to the studied DNA trimers (using the ADC(2) method), except that we used the smaller def2-SVP basis set. See also the final section of the Supplementary Information and Supplementary Fig. 26, for more details.

Excited-state QM simulations of A and D nucleosides

The excited-state energies and geometries for adenosine and 2,6-diaminopurine nucleosides were calculated in the gas phase using the ADC(2) method for electronic excitations. The ground-state geometries and energies were obtained at the MP2 level. Similarly as in the case of the QM/MM calculations, for trinucleotides we used the TZVP basis set to keep the methodology consistent. MECPs for isolated nucleosides were optimized using the same algorithm as described above. The potential energy profiles presented in Fig. 5 in the main article were constructed by means of linear interpolation in internal coordinates between the respective S0 geometries (Frank-Condon region), S1 geometries, and S1/S0 MECPs. All of the gas phase electronic-structure calculations were performed with the TURBOMOLE 7.3 program69.

Data availability

The authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Information files.

Code availability

The external optimizer XOPT is available under the link The modified interface between the TURBOMOLE 7.3 and AMBER 16 programs, which enabled performing QM/MM simulations with the ADC(2) electronic-structure method was deposited on the Figshare platform under the link with Most recent revision of this code can be obtained upon reasonable request.


  1. 1.

    Taylor, J. S. Unraveling the molecular pathway from sunlight to skin cancer. Acc. Chem. Res. 27, 76–82 (1994).

    CAS  Article  Google Scholar 

  2. 2.

    Pagès, V. & Fuchs, R. P. How DNA lesions are turned into mutations within cells? Oncogene 21, 8957–8966 (2002).

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  3. 3.

    Kripke, M. L., Cox, P. A., Alas, L. G. & Yarosh, D. B. Pyrimidine dimers in DNA initiate systemic immunosuppression in UV-irradiated mice. Proc. Natl Acad. Sci. USA 89, 7516–7520 (1992).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  4. 4.

    Schreier, W. J. et al. Thymine dimerization in DNA is an ultrafast photoreaction. Science 315, 625–629 (2007).

    ADS  CAS  PubMed  PubMed Central  Article  Google Scholar 

  5. 5.

    Mouret, S. et al. Cyclobutane pyrimidine dimers are predominant DNA lesions in whole human skin exposed to UVA radiation. Proc. Natl Acad. Sci. USA 103, 13765–13770 (2006).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  6. 6.

    Rauer, C., Nogueira, J. J., Marquetand, P. & González, L. Cyclobutane thymine photodimerization mechanism revealed by nonadiabatic molecular dynamics. J. Am. Chem. Soc. 138, 15911–15916 (2016).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  7. 7.

    Mees, A. et al. Crystal structure of a photolyase bound to a CPD-like DNA lesion after in situ repair. Science 306, 1789–1793 (2004).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  8. 8.

    Tan, C. et al. The molecular origin of high DNA-repair efficiency by photolyase. Nat. Commun. 6, 7302 (2015).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  9. 9.

    Gehring, W. & Rosbash, M. The coevolution of blue-light photoreception and circadian rhythms. J. Mol. Evol. 57, S286–S289 (2003).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  10. 10.

    Deck, C., Jauker, M. & Richert, C. Efficient enzyme-free copying of all four nucleobases templated by immobilized RNA. Nat. Chem. 3, 603–608 (2011).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  11. 11.

    Mansy, S. S. et al. Template-directed synthesis of a genetic polymer in a model protocell. Nature 454, 122–125 (2008).

    ADS  CAS  PubMed  PubMed Central  Article  Google Scholar 

  12. 12.

    Adamala, K. & Szostak, J. W. Nonenzymatic template-directed RNA synthesis inside model protocells. Science 342, 1098–1100 (2013).

    ADS  CAS  PubMed  PubMed Central  Article  Google Scholar 

  13. 13.

    Kim, S. C., O’Flaherty, D. K., Zhou, L., Lelyveld, V. S. & Szostak, J. W. Inosine, but none of the 8-oxo-purines, is a plausible component of a primordial version of RNA. Proc. Natl Acad. Sci. USA 115, 13318–13323 (2018).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  14. 14.

    Saladino, R. et al. Meteorite-catalyzed syntheses of nucleosides and of other prebiotic compounds from formamide under proton irradiation. Proc. Natl Acad. Sci. USA 112, E2746–E2755 (2015).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  15. 15.

    Becker, S. et al. A high-yielding, strictly regioselective prebiotic purine nucleoside formation pathway. Science 352, 833–836 (2016).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  16. 16.

    Saladino, R. et al. Proton irradiation: a key to the challenge of N-glycosidic bond formation in a prebiotic context. Sci. Rep. 7, 14709 (2017).

    ADS  PubMed  PubMed Central  Article  CAS  Google Scholar 

  17. 17.

    Becker, S. et al. Unified prebiotically plausible synthesis of pyrimidine and purine RNA ribonucleotides. Science 366, 76–82 (2019).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  18. 18.

    Teichert, J. S., Kruse, F. M. & Trapp, O. Direct prebiotic pathway to DNA nucleosides. Angew. Chem. Int. Ed. 58, 9944–9947 (2019).

    CAS  Article  Google Scholar 

  19. 19.

    Yadav, M., Kumar, R. & Krishnamurthy, R. Chemistry of abiotic nucleotide synthesis. Chem. Rev. 120, 4766–4805 (2020).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  20. 20.

    Powner, M. W., Gerland, B. & Sutherland, J. D. Synthesis of activated pyrimidine ribonucleotides in prebiotically plausible conditions. Nature 459, 239–242 (2009).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  21. 21.

    Xu, J. et al. A prebiotically plausible synthesis of pyrimidine β-ribonucleosides and their phosphate derivatives involving photoanomerization. Nat. Chem. 9, 303–309 (2017).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  22. 22.

    Xu, J., Green, N. J., Gibard, C., Krishnamurthy, R. & Sutherland, J. D. Prebiotic phosphorylation of 2-thiouridine provides either nucleotides or DNA building blocks via photoreduction. Nat. Chem. 11, 457–462 (2019).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  23. 23.

    Roberts, S. J. et al. Selective prebiotic conversion of pyrimidine and purine anhydronucleosides into Watson-Crick base-pairing arabino-furanosyl nucleosides in water. Nat. Commun. 9, 4073 (2018).

    ADS  PubMed  PubMed Central  Article  CAS  Google Scholar 

  24. 24.

    Xu, J. et al. Selective prebiotic formation of RNA pyrimidine and DNA purine nucleosides. Nature 582, 60–66 (2020).

    ADS  CAS  PubMed  PubMed Central  Article  Google Scholar 

  25. 25.

    Ranjan, S. & Sasselov, D. D. Constraints on the early terrestrial surface UV environment relevant to prebiotic chemistry. Astrobiology 17, 169–204 (2017).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  26. 26.

    Bucher, D. B., Kufner, C. L., Schlueter, A., Carell, T. & Zinth, W. UV-induced charge transfer states in DNA promote sequence selective self-repair. J. Am. Chem. Soc. 138, 186–190 (2016).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  27. 27.

    Nguyen, K. V. & Burrows, C. J. A prebiotic role for 8-oxoguanosine as a flavin mimic in pyrimidine dimer photorepair. J. Am. Chem. Soc. 133, 14586–14589 (2011).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  28. 28.

    Szabla, R., Kruse, H., Stadlbauer, P., Šponer, J. & Sobolewski, A. L. Sequential electron transfer governs the UV-induced self-repair of DNA photolesions. Chem. Sci. 9, 3131–3140 (2018).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  29. 29.

    Piccinni, V., Reiter, S., Keefer, D. & de Vivie-Riedle, R. Multiscale conformational sampling reveals excited-state locality in DNA self-repair mechanism. J. Phys. Chem. A 124, 9133–9140 (2020).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  30. 30.

    Khudyakov, I. Y. A., Kirnos, M. D., Alexandrushkina, N. I. & Vanyushin, B. F. Cyanophage S-2L contains DNA with 2,6-diaminopurine substituted for adenine. Virology 88, 8–18 (1978).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  31. 31.

    Kirnos, M. D., Khudyakov, I. Y., Alexandrushkina, N. I. & Vanyushin, B. F. 2-Aminoadenine is an adenine substituting for a base in S-2L cyanophage DNA. Nature 270, 369–370 (1977).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  32. 32.

    Krishnamurthy, R., Jiménez, E. I. & Gibard, C. Prebiotic phosphorylation and concomitant oligomerization of deoxynucleosides to form DNA. Angew. Chem. Int. Ed. 60, 10775–10783 (2021).

  33. 33.

    Schrum, J. P., Ricardo, A., Krishnamurthy, M., Blain, J. C. & Szostak, J. W. Efficient and rapid template-directed nucleic acid copying using 2′-amino-2′,3′-dideoxyribonucleoside−5′-phosphorimidazolide monomers. J. Am. Chem. Soc. 131, 14560–14570 (2009).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  34. 34.

    Hartel, C. & Göbel, M. W. Substitution of adenine by purine-2,6-diamine improves the nonenzymatic oligomerization of ribonucleotides on templates containing thymidine. Helv. Chim. Acta 83, 2541–2549 (2000).

    CAS  Article  Google Scholar 

  35. 35.

    Kozlov, I. A. & Orgel, L. E. Nonenzymatic oligomerization reactions on templates containing inosinic acid or diaminopurine nucleotide residues. Helv. Chim. Acta 82, 1799–1805 (1999).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  36. 36.

    Bucher, D. B., Pilles, B. M., Carell, T. & Zinth, W. Charge separation and charge delocalization identified in long-living states of photoexcited DNA. PNAS 111, 4369–4374 (2014).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  37. 37.

    Hättig, C. Structure optimizations for excited states with correlated second-order methods: CC2 and ADC(2). Adv. Quantum Chem. 50, 37–60 (2005).

    ADS  Article  CAS  Google Scholar 

  38. 38.

    Dreuw, A. & Wormit, M. The algebraic diagrammatic construction scheme for the polarization propagator for the calculation of excited states. Wiley Interdiscip. Rev. Comput. Mol. Sci. 5, 82–95 (2015).

    CAS  Article  Google Scholar 

  39. 39.

    Spata, V. A., Lee, W. & Matsika, S. Excimers and exciplexes in photoinitiated processes of oligonucleotides. J. Phys. Chem. Lett. 7, 976–984 (2016).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  40. 40.

    Pan, Z., Chen, J., Schreier, W. J., Kohler, B. & Lewis, F. D. Thymine dimer photoreversal in purine-containing trinucleotides. J. Phys. Chem. B 116, 698–704 (2012).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  41. 41.

    Martinez, T. J. Seaming is believing. Nature 467, 412–413 (2010).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  42. 42.

    Lee, W. & Matsika, S. Role of charge transfer states into the formation of cyclobutane pyrimidine dimers in DNA. Faraday Discuss. 216, 507–519 (2019).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  43. 43.

    Subotnik, J. E., Yeganeh, S., Cave, R. J. & Ratner, M. A. Constructing diabatic states from adiabatic states: extending generalized Mulliken–Hush to multiple charge centers with Boys localization. J. Chem. Phys. 129, 244101 (2008).

    ADS  PubMed  Article  CAS  PubMed Central  Google Scholar 

  44. 44.

    Pecourt, J.-M. L., Peon, J. & Kohler, B. DNA excited-state dynamics: ultrafast internal conversion and vibrational cooling in a series of nucleosides. J. Am. Chem. Soc. 123, 10370–10378 (2001).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  45. 45.

    Sanchez, R. A. & Orgel, L. E. Studies in prebiotic synthesis: V. Synthesis and photoanomerization of pyrimidine nucleosides. J. Mol. Biol. 47, 531–543 (1970).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  46. 46.

    Callahan, M. P. et al. Carbonaceous meteorites contain a wide range of extraterrestrial nucleobases. Proc. Natl Acad. Sci. USA 108, 13995–13998 (2011).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  47. 47.

    Beckstead, A. A., Zhang, Y., Vries, M. Sde & Kohler, B. Life in the light: nucleic acid photoproperties as a legacy of chemical evolution. Phys. Chem. Chem. Phys. 18, 24228–24238 (2016).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  48. 48.

    Plasser, F. et al. Surface hopping dynamics with correlated single-reference methods: 9H-adenine as a case study. J. Chem. Theory Comput. 10, 1395–1405 (2014).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  49. 49.

    Fialho, D. M., Roche, T. P. & Hud, N. V. Prebiotic syntheses of noncanonical nucleosides and nucleotides. Chem. Rev. 120, 4806–4830 (2020).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  50. 50.

    Kim, H.-J. & Benner, S. A. Prebiotic stereoselective synthesis of purine and noncanonical pyrimidine nucleotide from nucleobases and phosphorylated carbohydrates. Proc. Natl Acad. Sci. USA 114, 11315–11320 (2017).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  51. 51.

    Urban, M. W. et al. Key-and-lock commodity self-healing copolymers. Science 362, 220–225 (2018).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  52. 52.

    Crespo-Hernández, C. E., Cohen, B. & Kohler, B. Base stacking controls excited-state dynamics in A·T DNA. Nature 436, 1141–1144 (2005).

    ADS  PubMed  Article  CAS  PubMed Central  Google Scholar 

  53. 53.

    Berendsen, H. J. C., Grigera, J. R. & Straatsma, T. P. The missing term in effective pair potentials. J. Phys. Chem. 91, 6269–6271 (1987).

    CAS  Article  Google Scholar 

  54. 54.

    Izadi, S., Anandakrishnan, R. & Onufriev, A. V. Building water models: a different approach. J. Phys. Chem. Lett. 5, 3863–3871 (2014).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  55. 55.

    Joung, I. S. & Cheatham, T. E. Determination of alkali and halide monovalent ion parameters for use in explicitly solvated biomolecular simulations. J. Phys. Chem. B 112, 9020–9041 (2008).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  56. 56.

    Zgarbová, M. et al. Refinement of the sugar–phosphate backbone torsion beta for AMBER force fields improves the description of Z- and B-DNA. J. Chem. Theory Comput. 11, 5723–5736 (2015).

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  57. 57.

    Pérez, A. et al. Refinement of the AMBER force field for nucleic acids: improving the description of α/γ conformers. Biophys. J. 92, 3817–3829 (2007).

    ADS  PubMed  PubMed Central  Article  CAS  Google Scholar 

  58. 58.

    Krepl, M. et al. Reference simulations of noncanonical nucleic acids with different χ variants of the AMBER force field: quadruplex DNA, quadruplex RNA, and Z-DNA. J. Chem. Theory Comput. 8, 2506–2520 (2012).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  59. 59.

    Zgarbová, M. et al. Toward improved description of DNA backbone: revisiting epsilon and zeta torsion force field parameters. J. Chem. Theory Comput. 9, 2339–2354 (2013).

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  60. 60.

    Cornell, W. D. et al. A second generation force field for the simulation of proteins, nucleic acids, and organic molecules. J. Am. Chem. Soc. 117, 5179–5197 (1995).

    CAS  Article  Google Scholar 

  61. 61.

    Hopkins, C. W., Le Grand, S., Walker, R. C. & Roitberg, A. E. Long-time-step molecular dynamics through hydrogen mass repartitioning. J. Chem. Theory Comput. 11, 1864–1874 (2015).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  62. 62.

    Ryckaert, J.-P., Ciccotti, G. & Berendsen, H. J. C. Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. Comput. Phys. 23, 327–341 (1977).

    ADS  CAS  Article  Google Scholar 

  63. 63.

    Miyamoto, S. & Kollman, P. A. Settle: an analytical version of the SHAKE and RATTLE algorithm for rigid water models. J. Computational Chem. 13, 952–962 (1992).

    CAS  Article  Google Scholar 

  64. 64.

    Berendsen, H. J. C., Postma, J. P. M., van Gunsteren, W. F., DiNola, A. & Haak, J. R. Molecular dynamics with coupling to an external bath. J. Chem. Phys. 81, 3684–3690 (1984).

    ADS  CAS  Article  Google Scholar 

  65. 65.

    Case, D. A. et al. AMBER 2016 (2016).

  66. 66.

    Rodriguez, A. & Laio, A. Clustering by fast search and find of density peaks. Science 344, 1492–1496 (2014).

    ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

  67. 67.

    Bottaro, S., Di Palma, F. & Bussi, G. The role of nucleobase interactions in RNA structure and dynamics. Nucleic Acids Res. 42, 13306–13314 (2014).

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  68. 68.

    Götz, A. W., Clark, M. A. & Walker, R. C. An extensible interface for QM/MM molecular dynamics simulations with AMBER. J. Comput. Chem. 35, 95–108 (2014).

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  69. 69.

    Ahlrichs, R., Bär, M., Häser, M., Horn, H. & Kölmel, C. Electronic structure calculations on workstation computers: The program system turbomole. Chem. Phys. Lett. 162, 165–169 (1989).

    ADS  CAS  Article  Google Scholar 

  70. 70.

    Grimme, S., Brandenburg, J. G., Bannwarth, C. & Hansen, A. Consistent structures and interactions by density functional theory with small atomic orbital basis sets. J. Chem. Phys. 143, 054107 (2015).

    ADS  PubMed  Article  CAS  PubMed Central  Google Scholar 

  71. 71.

    Szabla, R., Havrila, M., Kruse, H. & Šponer, J. Comparative assessment of different RNA tetranucleotides from the DFT-D3 and force field perspective. J. Phys. Chem. B 120, 10635–10648 (2016).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  72. 72.

    Kruse, H. & Šponer, J. Towards biochemically relevant QM computations on nucleic acids: controlled electronic structure geometry optimization of nucleic acid structural motifs using penalty restraint functions. Phys. Chem. Chem. Phys. 17, 1399–1410 (2014).

    PubMed  Article  CAS  PubMed Central  Google Scholar 

  73. 73.

    Eckert, F., Pulay, P. & Werner, H.-J. Ab initio geometry optimization for large molecules. J. Comput. Chem. 18, 1473–1483 (1997).

    CAS  Article  Google Scholar 

  74. 74.

    Trofimov, A. B. & Schirmer, J. An efficient polarization propagator approach to valence electron excitation spectra. J. Phys. B: Mol. Opt. Phys. 28, 2299–2324 (1995).

    ADS  CAS  Article  Google Scholar 

  75. 75.

    Plasser, F., Bäppler, S. A., Wormit, M. & Dreuw, A. New tools for the systematic analysis and visualization of electronic excitations. II. Applications. J. Chem. Phys. 141, 024107 (2014).

    ADS  PubMed  Article  CAS  PubMed Central  Google Scholar 

  76. 76.

    Plasser, F., Wormit, M. & Dreuw, A. New tools for the systematic analysis and visualization of electronic excitations. I. Formalism. J. Chem. Phys. 141, 024106 (2014).

    ADS  PubMed  Article  CAS  PubMed Central  Google Scholar 

  77. 77.

    Levine, B. G., Coe, J. D. & Martínez, T. J. Optimizing conical intersections without derivative coupling vectors: application to multistate multireference second-order perturbation theory (MS-CASPT2). J. Phys. Chem. B 112, 405–413 (2008).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  78. 78.

    Tuna, D. et al. Assessment of approximate coupled-cluster and algebraic-diagrammatic-construction methods for ground- and excited-state reaction paths and the conical-intersection seam of a retinal-chromophore model. J. Chem. Theory Comput. 11, 5758–5781 (2015).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  79. 79.

    Szabla, R., W. Góra, R. & Šponer, J. Ultrafast excited-state dynamics of isocytosine. Phys. Chem. Chem. Phys. 18, 20208–20218 (2016).

    CAS  PubMed  Article  PubMed Central  Google Scholar 

Download references


The authors thank Prof. John D. Sutherland, Prof. Dimitar D. Sasselov, and Dr. Corinna Kufner for helpful discussions. This research was supported by the Simons Foundation (494188 to R.S.), the Foundation for Polish Science (START Fellowship to R.S.), the Czech Science Foundation (project 21-23718S to J.S. and P.St.), the project SYMBIT (reg. number: CZ.02.1.01/0.0/0.0/15_003/0000477 to J.S. and H.K.) financed by the ERDF and a grant from the Ministry of Science and Higher Education in Poland (DS/531-T080-D494-20 to J.R.).

Author information




M.Z., P.S., and J.R. irradiation experiments; N.J.G. nucleoside synthesis; P.St. and J.S. molecular dynamics simulations; H.K. code development; R.S. idea, quantum chemical simulations, and coordination of the project. R.S. wrote the main article with input from all coauthors. M.Z., N.J.G., P.St., and R.S. assembled the SI and “Methods” section.

Corresponding authors

Correspondence to Rafał Szabla or Magdalena Zdrowowicz.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature Communications thanks Hannes Mutschler and the other anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Szabla, R., Zdrowowicz, M., Spisz, P. et al. 2,6-diaminopurine promotes repair of DNA lesions under prebiotic conditions. Nat Commun 12, 3018 (2021).

Download citation


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing