ARTICLE pubs.acs.org/JPCA
Formamide-Based Prebiotic Synthesis of Nucleobases: A Kinetically Accessible Reaction Route Judit E. Sponer,*,†,‡ Arnost Mladek,† Jirí Sponer,†,‡ and Miguel Fuentes-Cabrera§ †
Institute of Biophysics, Academy of Sciences of the Czech Republic, Kralovopolska 135, CZ-61265 Brno, Czech Republic CEITEC-Central European Institute of Technology, Masaryk University, Campus Bohunice, Kamenice 5, CZ-62500 Brno, Czech Republic § Center for Nanophase Materials Sciences and Computer Sciences and Mathematics Division, Oak Ridge National Laboratory, Oak Ridge, P.O. Box 2008, Oak Ridge, Tennessee 37831-6494, United States ‡
bS Supporting Information ABSTRACT: Synthesis of nucleobases in nonaqueous environments is an alternative way for the emergence of terrestrial life, which could solve the fundamental problem connected to the hydrolytic instability of nucleic acid components in an aqueous environment. In this contribution, we present a plausible reaction route for the prebiotic synthesis of nucleobases in formamide, which does not require participation of the formamide trimer and aminoimidazole-carbonitrile intermediates. The computed activation energy of the proposed pathway is noticeably higher than that of the HCN-based synthetic route, but it is still feasible under the experimental conditions of the Saladino synthesis. We show that, albeit both the pyrimidine and purine ring formation utilizes the undissociated form of formamide, the dehydration product of formamide, HCN, may also play a key role in the mechanism. The rate determining step of the entire reaction path is the cyclization of the diaza-pentanimine precursor. The subsequent formation of the imidazole ring proceeds with a moderate activation energy. Our calculations thus demonstrate that the experimentally suggested reaction path without the involvement of aminoimidazole-carbonitrile intermediates is also a viable alternative for the nonaqueous synthesis of nucleobases.
1. INTRODUCTION The one-pot synthesis of nucleic acids constituents is one of the most challenging tasks related to the emergence of life on the early earth. Powner et al. have recently elaborated a convenient synthetic route leading to nucleotides in aqueous environment,1,2 which has exemplified that in the presence of phosphates and UV light nucleotides can be formed from simple inorganic molecules without a substantial kinetic barrier. Nevertheless, survival of nucleotides in aqueous environment is rather problematic, due to the well known thermodynamic instability of nucleic acid components.36 This calls for considering solvent media other than water in the prebiotic nucleic acid synthesis.7 A potential candidate for this purpose is formamide, which is one of the most abundant molecules in the universe. Recently, it has been suggested that it could have played a role in the emergence of the terrestrial life as well.7,8 Saladino et al. have elaborated a formamide-based synthetic path that could give rise to the simultaneous formation of nucleobases as well as sugars assuming that all chemistry takes place in formamide solution.9,10 Beyond providing the reaction medium, a considerable advantage of using formamide is that it is less reactive than HCN. Therefore, formamide ensures a sufficient control over the reaction speed and it may act as the precursor of a multitude of simple prebiotically relevant molecules. For example, the dehydrated form of formamide, HCN is excessively reactive and cannot be converted to pyrimidine bases efficiently. In addition, formamide is an excellent dipolar aprotic reaction medium, r 2011 American Chemical Society
which helps to overcome the hydrolytic instability of nucleobases and nucleosides. A further advantage of using formamide as a solvent is that it enables a broader temperature range for solution-phase reactions because its boiling point is about 100° higher than that of water. In addition, many reactions of formamide are catalyzed by metal oxides and other minerals, which could also be present on the primitive earth. On the basis of the results of Yamada et al.,11,12 Saladino et al.7,8 suggested that the formamide-based synthesis (for a summary of the reaction steps see Scheme 1) of pyrimidines and purines starts with the formation of formamide dimer. In the later reaction steps, the dehydration products of formamide, i.e., HCN and H2O are active as well.13 According to this, pyrimidines are afforded by the reaction of 2 HCN molecules with formamide dimer. Purines are formed from pyrimidines via the addition of formamide to the six-membered ring and a subsequent cyclization step. Note that the original synthesis by Saladino goes up to the formation of nucleosides.14 Nevertheless, in the current study, we will be restricted to the main steps of the reaction leading to pyrimidine and purine heterocycles only, which do not necessarily involve redox steps and transition metal-catalysis.15 The formamide-based synthesis also utilizes HCN. This could make the mechanism similar to the one reported by Roy et al. for Received: October 14, 2011 Revised: November 24, 2011 Published: November 30, 2011 720
dx.doi.org/10.1021/jp209886b | J. Phys. Chem. A 2012, 116, 720–726
The Journal of Physical Chemistry A
ARTICLE
Scheme 1
the prebiotic adenine formation.16 Nevertheless, the formamidebased route is fundamentally different since in the initial reaction step it involves formation of a formamide dimer species. In contrast, the HCN-based route1719 utilizes HCN from the earliest stages of the reaction, involving the dimeric, trimeric (aminomalononitrile), and tetrameric (2,3-diaminomaleonitrile) forms of HCN. A further, concurrent route of the HCN-based adenine formation is described by Glaser et al.20 The linkage between the formamide and HCN-based reaction routes is illustrated by the recent work of Barks et al.,21 which demonstrates that nucleobases can form from UV irradiated formamide according to the HCN chemistry. In the current work, we follow a similar computational strategy as the one applied by Roy et al. for the computational description of the HCN-based reaction route leading to adenine.16 On the basis of the intermediates suggested by Saladino et al. experimentally,7,8 we construct a reaction pathway interconnecting these species using ab initio electronic structure calculations. Quantum chemistry is a viable complement to the experiments since it enables the study of species with very short lifetimes and thus to localize transition states. In addition, it provides an atomic level insight into the transformations of the reacting single molecules along the pathway, which is often not apparent in experimental studies when dealing with complex mixtures of various species. We evaluate the activation and reaction free energies of all suggested reaction steps in order to get an insight into the kinetics and thermodynamics of one of the suggested scenarios of the formamide-based purine and pyrimidine synthesis. In particular, we disclose the factors that make the formamide-based reaction route energetically more demanding than the traditional HCN-based chemistry. Moreover, we elucidate the catalytic role of water on the mechanism, suggested by Saladino et al.7
atomic radii. As a consequence, an individual sphere was assigned to each hydrogen atom without regard to its connectivity, and thus, hydrogens were treated explicitly. Free energies of the studied compounds in gas phase (Ggas) were calculated from the total electronic energy computed in gas phase (Etot,gas) and from the thermal and entropic correction terms to the Gibbs free energy computed in gas phase (δGgas) via harmonic approximation from frequency calculations Ggas ¼ Etot, gas þ δGgas
ð1Þ
Free energies of the studied compounds in solution (Gsol) were obtained using the total electronic energies computed in solution (Etot,sol) supplemented with the thermal and entropic correction terms to the Gibbs free energy computed in gas phase (δGgas) as well as the solvent correction to the free energy (δGsol) Gsol ¼ Etot, sol þ δGgas þ δGsol
ð2Þ
In the case of water, δGsol accounts for the solute cavitation energy correction calculated using the model of Pierotti.27 For formamide, δGsol was assumed to be 0, due to the lack of empiricial parameters for calculating nonelectrostatic components of the solvation free energy. We have made a set of test calculations to evaluate the impact of neglecting the nonelectrostatic components of the solvent correction to the free energy in an aqueous environment for the elementary steps of pyrimidine ring formation (summary of the computed results obtained for the HCN addition, CN-addition, and cyclization reaction steps is presented in Table S1 in the Supporting Information). These calculations have conclusively shown that the free energy profiles with and without including the nonelectrostatic components are essentially the same with a maximum deviation of the estimated activation and reaction free energies of 4 and 2 kcal/mol, respectively.
2. COMPUTATIONAL DETAILS Quantum chemical calculations were carried out at the B3LYP level of theory22,23 using tight convergence criteria, the 6-311++G(2d, 2p) set of atomic orbitals, and the Gaussian09 suite of programs.24 The same approach was successfully used in our recent theoretical study to describe some of the key steps of the Sutherland synthesis in ref 6. Gradient optimizations were used to obtain both the gas phase optimized geometries and vibrational frequencies. Local energy minima and transition states were verified with frequency calculations. The presence of a polar solvent (ε = 78.4 for water and ε = 108.9 for formamide) was mimicked with single-point calculations using the polarizable conductor continuum solvent model (C-PCM)25,26 in the standard parametrization supported by the Gaussian09 program: the average surface of a tesserae was 0.4 Å2, and the minimum radius of the added spheres used to create the solvent excluded surface was 0.2 Å; the United Force Field (UFF) model and a scaling factor of 1.1 were used to define the
3. RESULTS AND DISCUSSION 3.1. Formation of Pyrimidines. Under the default reaction conditions (ca. 430 K) of the Saladino synthesis,7 HCN is abundantly present in the reaction mixture, being the dehydration product of formamide.13 According to Yamada et al.,11,12 a plausible reaction route leading to pyrimidines involves the addition of two HCN molecules to a formamide dimer (hereafter denoted as 1, see Scheme 2). The reaction steps (which may proceed also in a reverse order compared to the one described below) lead to the formation of 3,5-dihydroxy-5-cyano-2,4-diazapentan-1-imine, whose cyclization product is the reduced form of pyrimidine bases. Figure 1 depicts the computed free energy profile for the above-mentioned reaction steps. 3.1.1. Reaction of Formamide Dimer with HCN. The initial step of the synthetic pathway leading to pyrimidine bases is the reaction of 1 with HCN. At first, we assumed that HCN attacks at the amino group of 1, similar to the initial step of the HCN-based 721
dx.doi.org/10.1021/jp209886b |J. Phys. Chem. A 2012, 116, 720–726
The Journal of Physical Chemistry A
ARTICLE
Scheme 2
water molecule into the computational model (Cartesian coordinates of the optimized initial state, transition state, and product geometries along with a short description of the computational models is available in the Supporting Information) may influence the kinetics of this reaction step. Indeed, we have observed a rather substantial decrease of the activation energy (to 43.8, 37.7, and 35.6 kcal/mol in gas phase, bulk formamide, and bulk water, respectively) in the presence of catalytic water molecules. In the water-mediated case, the amino proton is transferred to the cyano nitrogen with the assistance of the water molecule, whereas a high-energy three-centered bond is formed in the noncatalyzed case leading to a significant increase of the activation energy. A similar situation is described by Roy et al.16 for the water-assisted addition of HCN to the amino group of 4-aminoimidazole-5-carbonitrile. As mentioned in ref 16, in the HCN-based route, the catalytic water molecules can be replaced with N-bases as well. We have evaluated whether the same applies also for the current case. Replacing the water catalyst with ammonia, our computed activation energies increase only ca. 6 kcal/mol (i.e., 49.3, 44.2, and 41.7 kcal/mol in gas phase, formamide, and water, respectively) as compared to the values obtained with water. Thus, in this reaction step, N-bases may act as catalysts too; nonetheless, they are noticeably less efficient than water. 3.1.2. CN Attack on 3-Hydroxy-2,4,6-triaza-hex-5-en-1-one. According to the reaction pathway proposed by Saladino et al., the product of the previous reaction step, i.e., 3-hydroxy-2,4,6-triaza-hex-5-en-1-one (2, see Scheme 2), reacts with a further HCN molecule to form pyrimidine heterocycles. We tried to describe this reaction step assuming that a neutral HCN
Figure 1. Free energy profile of the reaction route leading to the formation of the 6-membered heterocyclic ring. Blue, gas phase; black, formamide; red, aqueous environment. The energies were computed at the B3LYP/6-311++G(2d,2p) level. Bulk solvent effects were treated using the C-PCM approximation. Note, that the first two steps, i.e., HCN addition and CN addition, may proceed in reverse order.
path described by Roy et al.16 If no catalysts are used, the reaction would require a rather substantial activation energy (61.9, 60.3, and 57.9 kcal/mol in gas phase, bulk formamide, and bulk water, respectively). Saladino et al. reported that trace amounts of water exert a robust catalytic effect.7 We investigated whether inclusion of a 722
dx.doi.org/10.1021/jp209886b |J. Phys. Chem. A 2012, 116, 720–726
The Journal of Physical Chemistry A
ARTICLE
water warrant that the cyclization step requires pretty harsh conditions, which might explain the experimentally observed low reaction yield in the temperature range where water is present in liquid phase. This underlines the importance of formamide as a reaction medium in the Saladino synthesis. 3.1.4. Dehydration Reaction Leading to Double Bond Formation. The direct product of the cyclization reaction, beyond HCN and CN, is 2,6-dihydroxy-4,5-diimino-hexahydropyrimidine (4, Scheme 2). Saladino et al. (ref 7) and Yamada et al. (ref 11) proposed that the ultimate product of the HCN-addition to a formamide dimer is 6-hydroxy-4,5-diimino-3,4,5,6-tetrahydropyrimidine (5, Scheme 2), which can be deduced from 4 by a dehydration step involving the N1(H) and C2(OH) positions. We have found a kinetically feasible route for this dehydration reaction, which comprises two steps (its free energy profile is depicted in Figure 2). In the first step, a HCN molecule protonates the hydroxyl group at C2, which leads to the elimination of a water molecule from the reaction complex and to the formation of a carbocationic center at C2. This reaction step requires a relatively low activation energy in all three studied environments (21.6, 20.9, and 19.2 kcal/mol in gas phase, bulk formamide, and water, respectively). The carbocation stabilizes in spontaneous proton loss from the neighboring N1, initiated by the CN ion formed in the previous reaction step. This leads to the formation of 5, which can be considered as the direct precursor of pyrimidine bases. This reaction is basically an acid catalyzed conversion of hemiaminals to imines, which is one of the elementary steps of the alkylimino-deoxo-bisubstitution. 3.2. Route to Purines. According to Saladino7 and Yamada et al.,11,12 formation of the purine ring is preceded by the addition of ammonia to the ketimino functional groups at C4 and C5 of 5, resulting in 6-hydroxy-4,4,5,5-tetramino-3,4,5,6-tetrahydropyrimidine (6, Scheme 2) in an equilibrium reaction. Our computations on model compounds (for details, see the Supporting Information) show that this reaction is analogous to the reaction of ammonia with carbonyl compounds. For the latter reaction, high-level quantum chemical calculations have shown that it proceeds with a very low activation energy with the assistance of catalytic water molecules.2830 The free energy profile of the reaction leading from 6 to the bicyclic product is summarized in Figure 3. We have found that the addition of formamide to 6 requires proton catalysis, in order to enhance the electrophilicity of the carbonyl group of formamide. Albeit the direct protonation of formamide by HCN is thermodynamically not favored, clay minerals may play the role of proton donors, which are known to catalyze the Saladino synthesis.31 Attack of the resulting NH2CHOH+ carbocation to one of the amino groups connected to C5 of 6 requires an activation energy of about 12.7, 14.1, and 12.7 kcal/mol in gas phase, bulk formamide, and water, respectively. (Likewise, we have found that the activation energy for the analogous attack to an amino group connected to C4 is roughly the same, i.e., 13.7, 13.6, and 12.9 kcal/mol in gas phase, formamide, and water, respectively. This was the reason why we have described only the reaction route involving the amino group bound to C5.) The product of the formamide addition (7, Scheme 2) loses water in a subsequent reaction step, which proceeds with an activation energy of 32.0, 29.7, and 29.9 kcal/mol in gas phase, bulk formamide, and bulk water environments, respectively. (Prior to this reaction step the CHOHNH2 moiety connected to the amino group at C5 of the pyrimidine ring must undergo a small conformational change. Nevertheless, according to our
Figure 2. Free energy profile for the dehydration step of the hexahydropyrimidine intermediate (4). Blue, gas phase; black, formamide; red, aqueous environment. The energies were computed at the B3LYP/ 6-311++G(2d,2p) level. Bulk solvent effects were treated using the C-PCM approximation. Numbers in parentheses refer to the free energy changes calculated relative to the initial state complex formed from the formamide dimer, HCN, and water.
molecule attacks the carbonyl functionality of 2. Since all our attempts failed to find a plausible transition state, we assumed that ammonia (that was reported to be present in the reaction mixture as well;7,11,12 pKa = 9.2 and 10.5 in water and DMSO, respectively) may deprotonate HCN (pKa = 9.3 and 12.9 in water and DMSO, respectively) to give CN, which may be a more active nucleophile. For this scenario, we indeed managed to allocate a transition state. Due to the anionic mechanism, there is a substantial difference in the activation energies computed in gas phase and polar solvents. Whereas the gas phase activation energy for this reaction step is only 4.3 kcal/mol, in bulk formamide and bulk water it is pretty remarkable, amounting to 28.3 and 24.1 kcal/mol, respectively. 3.1.3. Formation of the Six-Membered Ring. Our calculations show that the product of the cyanide addition, the O(5)deprotonated form of 3,5-dihydroxy-5-cyano-2,4-diaza-pentan1-imine (3, Scheme 2), is able to pick up protons from HCN and that the protonation proceeds with a negligible activation energy (
