Background

The modern nanobiotechnology uses the planar building elements for construction of the sophisticated synthetic DNA and RNA [1, 2]. Such elements are complementary base pairs of the adenine-thymine and the guanine-cytosine. This kind of approach allows the creation of a complex spatial structure from the separate modules called “tiles” [3–5]. The resulting structure is a bulk because of the need to bypass an immanent planarity of the complementary nucleobase pairs. In general, this approach originates in the historically acquired planar X-ray structures [6, 7], may be called “planar”. As one can see, this planar approach was determinative for choosing the base pairs self-assembly types that have been observed experimentally [8].

At the same time, the RNA shows a bigger diversity in the types of interacting base pairs [9] and employs both the canonical and non-canonical base pairs, which results in a rich variety of the secondary and tertiary structures of the RNA. It is obvious that the RNA crystal is capable of creating the non-canonical interactions as well as the non-planar forms of the nucleobase pairs by means of restriction of the degrees of freedom. This fact allows us to raise a question of the conformational capacity of the electron structures of the different nucleobase pairs. Does the free self-assembly of the essentially planar nucleic bases result in strictly planar base pairs or does it allow significantly non-planar structures as well? A positive answer to the question may lead to transition from the “planar approach” to the “stereo approach” that rely on the primary structure as the main source of an arbitrary secondary structure. Consequentially, it would allow the creation of more compact nanobiotechnological devices.

In our study [10], we prove that the model heteroassociates of the m1Cyt · m1Thy are capable of forming up to ten observable molecular complexes. The first three of them have population of ~90 % at standard conditions and are highly non-planar. The most energetically favorable structure has a T-shape, and the next two are of L-shape. Unfortunately, we failed to find any available published experimental data covering the self-assembly of the cytosine-thymine base pairs.

Presentation of the Hypothesis

As it was mentioned above, the modern nanobiotechnology constrains itself by the use of only planar nucleobase pairs. In our opinion, it results from a relative impossibility to produce stable non-planar nucleobase pairs, so creation of some complex 3D structure out the innate planar elements requires usage of the sophisticated techniques. But if one were able to create the stable non-planar elements, achievement of the complex 3D structures might be much easier.

Taking into account the great conformational variety of the nucleobase pairs in the wild [9], as well as a relative lack of the theoretical awareness of all the possible conformations, we suppose that the nucleobase pairs are capable of forming observable non-planar structures. Because of a quite good knowledge of the complementary base pairs, we believe that those structures could be found in the non-canonical non-complementary nucleobase pairs. Additionally, we needed a means to find all possible conformations of the given nucleobase pairs. We solved that problem in [11].

Testing the Hypothesis

Our hypothesis would be successful if we could find some theoretical grounds for any observable non-planar nucleobase pair that have not been revealed by the previous base-pairing experiments. And this paper testifies that we managed to find such a pair. The more important fact is that it has not been found experimentally yet.

The main goal of this paper is to obtain all possible observable m1Cyt · m1Thy heteroassociates for answering the question in the title. To fulfill this task, we used our own algorithm [11] of the input structures creating. We would like to emphasize the importance of the algorithm, which basically proceeds from an assumption that the nucleobase pairs could be stabilized by at least two intermolecular H-bonds. Aside from this assumption, the algorithm has no other restrictions and uses a modern view of the chemical nature of the H-bonds [12]. As a result, we obtained numerous optimized and stable non-planar structures of both the canonical and non-canonical nucleobase pairs. Although most of them are highly energetic complexes in the free form, we can use them for creating conditions in a crystal to achieve the exact desirable non-planar structure through restriction of the molecular degrees of freedom. That is how the wild-type RNA non-planar structures are created.

All the calculations have been carried out with the Gaussian 09 suite of the programs [13]. The relaxed geometries and their corresponding harmonic vibrational frequencies of the base pairs have been obtained using the density functional theory (DFT) with the B3LYP hybrid functional [14] for Pople’s 6-311++G(d,p) basis set in a vacuum. We performed the single-point energy calculations at the correlated MP2 level of theory [15] with the 6-311++G(2df,pd) Pople’s [16–18] basis set for B3LYP/6-311++G(d,p) geometries to consider the electronic correlation effects as accurately as possible.

The Gibbs free energy G values for all the structures were obtained at a room temperature (T = 298.15 K) in the following way:

$$ G = {E}_{\mathrm{el}} + {E}_{\mathrm{corr}}, $$
(1)

where E el is the electronic energy and E corr is the thermal correction.

Bader’s quantum theory “atoms in molecules” (QTAIM) was applied to analyze the electron density [19]. The topology of the electron density was examined using the program package AIMAll [20] with all the default options. The wave functions were obtained at the B3LYP/6-311++G(d,p) level of the theory. The presence of a bond critical point (BCP) of (3,-1) type [19] and a bond path between hydrogen donor and acceptor, as well as the positive value of the Laplacian at this BCP (Δρ ≥ 0), was considered as three criteria for the H-bond formation [19, 21].

Energies of the classical intermolecular H-bonds in base pairs were evaluated by the empirical Iogansen’s formula [22]:

$$ {E}_{\mathrm{HB}}=0.33\cdot \sqrt{\Delta v-40}, $$
(2)

where Δν is the magnitude of the redshift (relative to the free molecule) of the stretching mode of the H-bonded groups involved in the H-bonding. The partial deuteration, namely the partial deuteration of the amino group, was applied to eliminate the effect of the vibrational resonances [23].

Energies of the non-canonical so-called weak intermolecular CH..O/N H-bonds were evaluated by the empirical Espinosa-Molins-Lecomte (EML) formula [24, 25] based on the electron density distribution at the (3,-1) BCPs of the H-bonds:

$$ {E}_{\mathrm{HB}} = 0.5\times V(r), $$
(3)

where V(r) is the value of a local potential energy at the (3,-1) BCPs.

Relative strengths of van der Waals contacts were estimated by means of Grunenberg’s compliance constant formalism [26–28], calculated by the Compliance 3.0.2 program.

The main advantage of this approach is invariance of the compliance constants. The physical meaning of the compliance constants is deduced from their definition as a partial second derivative of the potential energy due to an external force:

$$ {C}_{ij}=\frac{\partial^2E}{\partial {f}_i\partial {f}_j}, $$
(4)

In other words, the compliance constants measure the displacement of an internal coordinate, resulting from a unit force acting on it. As follows from this definition, a lower numerical value of compliance constant corresponds to a stronger bond.

To study the charge transfer property in the interacting orbitals of the non-canonical intermolecular CH..O/N contacts, we used a natural bond orbital (NBO) analysis [29], which interprets the electronic wave function in terms of a set of occupied Lewis and a set of unoccupied non-Lewis localized orbitals. A second-order Fock matrix analysis was carried out to evaluate interactions between donor (i) and acceptor (j) bonds. The result of such interaction is a migration of the electron density from the idealized Lewis structure into an empty non-Lewis orbital σ *. For each donor (i) and acceptor (j) bond, the stabilization energy is

$$ {E}^{(2)}=\varDelta {E}_{ij}={q}_{ij}\frac{F{\left(i,j\right)}^2}{\varepsilon_j-{\varepsilon}_i}, $$
(5)

where q i is the donor orbital occupancy, ε j and ε i are the diagonal elements, and F(i,j) is the off diagonal element of the NBO Fock matrix.

The atomic numbering scheme for nucleobases is conventional [30].

Implications of the Hypothesis

We obtained the set of 51 complexes of the heteroassociates of m1Cyt · m1Thy. This set contains both the common and the rare tautomeric forms.

Total occupancy of the obtained complexes 1 (30.50 %) + 2 (26.66 %) + 3 (22.90 %) + 4 (7.64 %) + 5 (3.80 %) + 6 (3.09 %) + 7 (1.93 %) + 8 (1.14 %) + 9 (0.72 %) + 10 (0.67 %) constitute 99.04 % at normal conditions and defines all the experimentally observable complexes (see Table 1). For the purpose of achieving the declared goal, we will focus on these ten complexes in the rest of this paper.

Table 1 Common physico-chemical properties of observable heteroassociates of the m1Cyt · m1Thy
Full size table

The most energetically favorable complex 1 has a T-shape structure and is stabilized by means of two intermolecular H-bonds and two attractive van der Waals contacts (see Fig. 1).

Fig. 1
figure 1

Stereo image of the most energetically favorable heteroassociate of m1Cyt · m1Thy. Its T-shaped and stabilized by means of two H-bonds and two van der Waals contacts (represented by dotted line; their corresponding bond lengths H…B/A…B are given in Å). Note that the 1-methyl group of m1Thy actively interacts with the m1Cyt

Full size image

We have found out that the obtained heteroassociates of the m1Cyt · m1Thy are stabilized by means of the intermolecular H-bonds of NH..N/O, CH..N/O types and attractive van der Waals contacts of the N..N/O/C, O..O/C. Energies of the classical H-bonds lie in diapason of 3.70 ÷ 7.86 kcal/mol. Energies of the non-canonical H-bonds lie in diapason of 0.66 ÷ 2.65 kcal/mol.

The observable heteroassociates of the m1Cyt · m1Thy have several different shapes (see Fig. 2): the T-shape is represented exclusively by the most energetically favorable complex 1, an L-shape (complexes 2, 3), a spiral shape (complexes 4, 6), the planar structures (complexes 5, 7, 9, 10), and a severely non-planar structure of the complex 8 which cannot be classified as one of above. As one can see, the most expected shapes are the L-shape (49.56 %), the T-shape (30.5 %), the spiral shape (10.73 %), the planar shape (7.3 %), and finally the complex 8 shape (1.14 %). This fact gives us the positive answer to the question in the title.

Fig. 2
figure 2

Geometrical structure of observable at normal conditions heteroassociates of the m1Cyt · m1Thy. The non-planar heteroassociates are represented in two projections while the planar ones are represented by one projection. Numbers under heteroassociates correspond to the row ‘complex’ of Table 1. H-bonds and van der Waals contacts (represented by dotted lines and their corresponding bond lengths H…B/A…B are given in Å)

Full size image

We believe that such unusual geometries are the result of a strong involvement of a methyl group in stabilization of the heteroassociates. As one can see from Table 1, the methyl groups (C1H, C5H) are capable of forming the strong enough CH..O/N H-bonds with energies in diapason of 0.66 ÷ 2.28 kcal/mol. At the same time, one can observe the strong stabilizing van der Waals contacts (see Table 2), which are present in five out of ten observable heteroassociates, and have energies in the diapason of 0.29 ÷ 1.19 kcal/mol. The especially strong van der Waals contacts are surveyed in the most energetically favorable complex 1 with the energies of 1.05 and 1.19 kcal/mol respectively.

Table 2 Intermolecular CH..O/N H-bonds and van der Waals contacts in the observable heteroassociates of the m1Cyt · m1Thy
Full size table

In our opinion, the experiment will show plenty of the mixed L-shapes and T-shapes, with quite rare inclusions of the spiral and planar shapes. This in turn presents a possibility of a self-assembly of a layered structure upon the self-assembled non-planar structures. This hypothesis will be tested in our future study.

For the first time, from the first principles, we show that the conformational capacity of the nucleic base pairs is much richer from the perspective of the self-assembly than it is used to consider in the modern nanobiotechnology. The capability of the modified cytosine and the modified thymine to create the significantly non-planar structures opens the way for the novel “stereo approach” to construction of the nanobiotechnological devices. We believe that the modern nanobiotechnological basis can and should be enriched by the new nucleic base pairs with an innate ability for the non-planar structures.

We think there might be at least three possible ways of a future development of the “stereo approach”.

The first way is to find the other non-planar nucleobase pairs that would be much more convenient for usage in a real industry than those that we have found and presented in this paper. The next way might be searching for an appropriate nucleobase modification in order to achieve the same goal. And the third way is to try to synthesize similar to the non-planar nucleobase pair structures which would behave themselves both as non-planar and energetically favorable at the same time.