Interplay between Charge Density Distribution ... - ACS Publications

Mar 11, 2013 - Energetic Features, and Crystal Morphology of 6‑Methyl-2-thiouracil ... methyl-2-thiouracil crystals together with comprehensive ener...
0 downloads 0 Views 4MB Size
Article pubs.acs.org/JPCC

Interplay between Charge Density Distribution, Crystal Structure Energetic Features, and Crystal Morphology of 6‑Methyl-2-thiouracil Katarzyna N. Jarzembska,*,† Radosław Kamiński,† Emmanuel Wenger,‡ Claude Lecomte,‡ and Paulina M. Dominiak† †

Department of Chemistry, University of Warsaw, Pasteura 1, 02-093 Warszawa, Poland CRM2, Jean Barriol Institute, CNRS UMR 7036, University of Lorraine, BP 70239, Boulevard des Aiguillettes, 54506 Vandoeuvre-lès-Nancy, France



S Supporting Information *

ABSTRACT: This study provides a detailed charge density distribution analysis of 6methyl-2-thiouracil crystals together with comprehensive energetic investigations. The findings constitute a link between crystal network features, electronic structure, molecular motif interaction energy, and crystal morphology. Our results revealed agreement between the topologically estimated hydrogen bond energy and the computational values. It also occurred that the methyl group and sulfur atom significantly affect crystal packing and contribute to a variety of intermolecular contacts. Nevertheless, the strongest interactions observed in the crystal structure are of electrostatic type. It is well reflected in the generated molecular electrostatic potential surfaces of the complementary character. Additionally, simple analysis of the interslab interaction energies and surface energy values led to a rough explanation of the crystal elongation direction and preferably created facets, such as the well-developed {010} crystal form. It should be also noted that complex investigations of the charge density distribution revealed significant residual peaks next to the sulfur atom, neither reliably modeled, nor justified by theory. These could be explained by the presence of a small percentage of the oxo-thiol tautomeric form in the crystal or by the S···S interactions across the crystal lattice (S−S bridge formation, etc.); however, no binding conclusions were drawn.

1. INTRODUCTION Among the pyrimidine bases, thiouracil derivatives constitute a valuable class of compounds due to their biological and physicochemical properties.1−8 A number of uracil mono- and dithio- modifications have already found medical applications.8−11 Thanks to their close relation to RNA/DNA components, they usually act as antivirial, antithyroid, or antitumor agents. For instance, 6-propyl-2-thiouracil (PTU) is a well-known antihyroid drug.12 Its methyl analogue, 6-methyl-2thiouracil (6m2tU), also exhibits some therapeutic activity; however, it is especially characteristic for its capability of forming high-quality well-shaped crystals (from aqueous solutions), suitable for high-resolution X-ray measurements and for morphological investigations. This property opens up possibilities of detailed charge density distribution studies and crystal architecture energy aspect exploration. Therefore, our analysis is devoted to meeting the interrelations between charge density distribution and interaction energy of selected molecular motifs present in the crystal network of 6m2tU and further to exploring the structural energetic features in the context of crystal morphology.

temperature. It is worth noting, as has already been reported in the literature,13 that 6m2tU tends to form nonmerohedrally twinned crystals. Therefore, crystals appropriate for the subsequent analyses were very carefully chosen. For the purpose of this study, we conducted two highresolution X-ray diffraction experiments at different temperatures, i.e., at 90 and 10 K, on two different crystals taken from the same crystallization batch. The 90 K measurement was performed on a Bruker AXS Kappa APEXII Ultra diffractometer, equipped with a Mo TXS rotating anode, multilayer optics, and an Oxford Cryosystems nitrogen gas-flow device (700 Series Cryostream). In turn, the 10 K experiment was carried out with a locally modified Nonius Kappa CCD diffractometer equipped with a molybdenum X-ray sealed tube, graphite monochromator, and orange closed-circuit cryostat.14 Full details on the two measurements together with the data processing information are given in the Supporting Information. Data processing was performed similarly in both cases. The determination of unit cell parameters and the integration of raw diffraction images were conducted with the APEX2 program package.15 Data sets were corrected for Lorentz, polarization, and oblique incidence effects. The multiscan absorption correction, frame-to-frame scaling and

2. MATERIALS AND METHODS 2.1. Crystal Synthesis and X-ray Data Collection. Charge-density-quality single crystals of 6m2tU were grown from water solution by slow evaporation of solvent at room © 2013 American Chemical Society

Received: December 10, 2012 Revised: March 8, 2013 Published: March 11, 2013 7764

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C

Article

for ith reflection wi = 1/σi2). The X−H bond lengths (X = nonhydrogen atom) were restrained to neutron-normalized distances with σ = 0.001.23 This approach provided previously results comparable with the corresponding theoretical periodic computations and neutron studies for a series of uracil derivatives.24 In the final stage of refinements, the derivation of anisotropic hydrogen atom ADPs was carried out using the SHADE server.25 During the refinement procedure the contraction−expansion parameters for all atoms, κ and κ′, were kept fixed at the UBDB2011 values. The multipole expansion was truncated at the hexadecapole (lmax = 4) and quadrupole (lmax = 2) levels for all non-hydrogen and hydrogen atoms, respectively. The general refinement strategy was as follows: (1) scale factor (which was also refined in almost all other stages), (2) atomic coordinates, (3) atomic coordinates and ADPs, (4) SHADE estimation of anisotropic hydrogen atom ADPs (which was also updated in-between other stages), (5) multipole parameters in a stepwise manner, and (6) all parameters simultaneously. According to the reasoning presented below, in the final model, the deformation terms describing the sulfur atom were fixed at the values transferred from the UBDB, whereas the sulfur atom valence population was refined. Much more details of multipole refinements can be found in the Supporting Information. 2.3. High-Resolution Data Interpretation. The interpretation of the collected high resolution X-ray data was not straightforward. A proper evaluation of the charge density model required a lot of attention. The simple multipole refinement described above (with all κ and κ′ parameters being fixed at the UBDB-transferred values) led to perhaps quite a satisfying overall fit (Δρres = −0.27/+0.35 e·Å−3); however, it did not converge to the physically reliable model. The charge density attributed to the sulfur atom valence shell was cumulated somehow above and below the aromatic ring plane, suggesting a quite peculiar electron lone pair distribution, which was not dictated by any strong intermolecular interaction in the crystal (Supporting Information). Furthermore, the theoretical charge density data, which was obtained after the multipole refinement with respect to the periodically computed structure factors, did not support the existence of such unexpectedly distributed electron density around the sulfur atom (for details see the Supporting Information). Instead, both lone electron pairs at the sulfur atom exhibited their maxima in the aromatic ring plane, forming a characteristic for a sulfur atom deformation density pattern. In order to find the origin of this effect, several additional investigations were carried out. The first attempt to refine the sulfur atom related features concerned the adjustment of the corresponding contraction−expansion parameters. These proceedings, however, neither improved the physical quality of the model nor reduced the charge density residues (Δρres = −0.31/ +0.42 e·Å−3). The other possible reason, which would explain the observed residues and unnatural charge density distribution, might be an improper description of the sulfur atom thermal motion. This supposition was partially backed up by the high-order (i.e., using only the reflections with sin θ/λ ≥ 0.8 Å−1) Fourier residual density maps (Supporting Information) for the 90 K data set. The observed residual charge density pattern is characteristic for anharmonic thermal motion of atoms, which has recently been reported and deeply analyzed in the literature.26 A departure from the harmonic model can be introduced via the Gram− Charlier series:27

merging of reflections were carried out with the SORTAV program.16 The final data collection and reduction parameters are presented in Table 1. Table 1. Selected Crystal Setting Information Together with the Data Collection Parameters for Both 90 and 10 K Measurement Temperatures formula molecular mass, Mr/a.u. crystal system space group Z F(000) a/Å b/Å c/Å α/° β/° γ/° volume, V/Å3 dcalc/g·cm−3 completeness (sin θ/λ)max/Å−1 θ range/° absorption coefficient, μ/mm−1 index ranges

no. of reflections collected/unique Rint

90 K data set

10 K data set

C5H6N2OS 142.2 monoclinic P21/c 4 296 4.3315(2) 14.4396(5) 9.5617(4) 90 90.8638(7) 90 597.97(4) 1.580 93.7% 1.19 2.56−58.00 0.445 −10 ≤ h ≤ 7 −34 ≤ k ≤ 34 −19 ≤ l ≤ 22 45681/6473 3.94%

4.3135(7) 14.432(3) 9.5548(18) 90 91.170(6) 90 594.69(18) 1.588 89.7% 1.25 2.56−62.58 0.447 −10 ≤ h ≤ 10 −34 ≤ k ≤ 35 −19 ≤ l ≤ 21 39616/8699 4.98%

The 10 K measurement is of generally lower quality than the 90 K one. This is due to the mechanical restrictions and slight misalignments of the used helium temperature experimental setup. Thus, the 10 K data set lacks the lowest resolution reflections and contains much more problematic background signal as well as a less accurate crystal orientation matrix. Consequently, great majority of our investigations is based on the data collected at 90 K. 2.2. Structure Determination and Multipole Refinements. The crystal structure of 6m2tU was solved by a chargeflipping method with the SUPERFLIP program17 and refined on F2 in the JANA package,18 within the independent atom model (IAM). Atomic scattering factors, in their analytical form, were taken from the International Tables for Crystallography.19 All non-H atoms were refined anisotropically. Hydrogen atoms were placed geometrically within the riding model for their isotropic ADPs (dC−H = 0.96 Å, UHiso = 1.2 UCeq; dN−H = 0.87 Å, UHiso = 1.5 UNeq). The orientation of the methyl group was determined from the adequate Fourier map. These results served as a starting point for the subsequent multipole refinements. The charge density model was evaluated using the MOPRO suite20 combined with the newly prepared version of the University at Buffalo Data Bank (UBDB2011),21 based on the Hansen-Coppens multipole formalism.22 Such a proceeding was additionally facilitated by specific modifications of the LSDB code. Currently, LSDB automatically provides the input files desired for the MOPRO package. All refinements were based on F, and only the reflections fulfilling the |Fo|2 ≥ 3σ(|Fo|2) condition were taken into account, as this was found to provide reliable results. In all the cases statistical weights were used (i.e., 7765

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C ⎡ 4π 3 T (H ) = ⎢1 − i ⎢⎣ 3 +

2π 4 3

Article

3. RESULTS AND DISCUSSION 3.1. Structural Remarks. A detailed analysis of 6m2tU molecular motifs present in the solid state was already a subject of our previous publication, covering structural and energetic features of a series of uracil derivatives.24 Therefore, here, we are going to recall solely some basic aspects of the 6m2tU crystal structure. 6m2tU crystallizes in the monoclinic P21/c space group with one molecule in the asymmetric unit. Its crystal architecture consists of molecular hydrogen-bonded chains forming sheettype motifs via N−H···S contacts, parallel to the (102) crystal plane. Such molecular layers are arranged parallel to one another in space. The three main dimeric units that can be distinguished in the 6m2tU crystal structure include: (i) C, motif based on the strong N−H···O hydrogen bond; (ii) D, dimeric entity bound by N−H···S interactions; (iii) S, π-stacking motif of two adjacent parallel located molecules (Figure 1).

∑ ∑ ∑ Cjklhjhkhl j

k

l



∑ ∑ ∑ ∑ Djklmhjhkhlhm⎥⎥T0(H) j

k

l

m



where T0(H) is the harmonic temperature factor, Cjkl and Djklm are the third and fourth order tensor refineable coefficients, and hk represents the kth component of the scattering vector H. The anharmonic refinement provided a substantial improvement of the overall fit to the experimental data. The charge density residual peaks were remarkably decreased (Δρres = −0.17/+0.23 e·Å−3) (Supporting Information). To verify the legitimacy of such a model, we deeply analyzed the 10 K data set. The mentioned characteristic residual pattern, being a consequence of atomic anharmonic thermal motion, should practically vanish at such low temperature.28,29 Despite the lower quality of the 10 K data set (the high-order data is complete, though), essentially the same residual pattern as for the 90 K data set was observed (Supporting Information). Although in the latter case this effect seems to be slightly less pronounced (keeping in mind lower data quality), the results suggested that this is not anharmonicity but rather some other effect unaccounted by the model (present in both measured crystals). Such findings led us to a detailed search for fine structural effects, regarding possible twinning or some subtle disorder. A very careful analysis of the 90 K data set excluded the existence of twinning (the reciprocal space is very clear with no typical signs of twinning). The investigated residual density features might also result from some disorder present in the crystal structure of 6m2tU. Unfortunately, it was not possible to find and refine any reasonable model of disorder. Nonetheless, this statement could be confirmed by the normal probability plots, which clearly suggest that some systematic error exists in both data sets (for all models) (see the Supporting Information). On the other hand, the problems with the sulfur atom modeling may be partially caused by inappropriate atomic scattering factor attributed to this chemical element. Thus, to improve the model of the sulfur atom scattering factor, an extensive scan over different sulfur atom radial functions was performed, however, with no satisfactory results. The novel approach desired by Koritsanszky et al.30 could provide a hint and shall be tested in the future. Nevertheless, it seems that the observed effect is too significant to be related just to the model limitations. Since, for the time being, we could not prove any of the above hypotheses, we decided to proceed further with the sulfur atom described by the deformation density populations kept at the UBDB-transferred values, whereas with the atomic scattering factor modified according to the literature (ζ = 3.851 a0−1; nl = 2, 4, 6, 8).31 The valence population of the sulfur atom was refined. This choice was made on the basis of the closer resemblance of such deformation density plots to the previously mentioned theoretical data results (Supporting Information). In this model, the so far unexplained residual density peaks located in the vicinity of the sulfur atom are very significant (Δρres = −0.28/ +0.64 e·Å−3) (90 K data set, see the Supporting Information). The structural models evaluated on the basis of different charge density refinements led to slightly modified molecular geometries. A comprehensive energetic comparison of the obtained results is available in the Supporting Information.

Figure 1. Main dimeric motifs present in the 6m2tU crystal structure (C, D, and S).

3.2. Charge Density Distribution and Dimer Interaction Energy Analysis. Despite some problems highlighted previously, the electron density distribution within the 6m2tU crystal was quite satisfactorily derived from the experimental data. The corresponding Fourier residual density maps together with the charge density parameters are summed up in the Supporting Information. Subsequently, the final charge density model was analyzed within the quantum theory of atoms in molecules (QTAIM), developed over the years by Bader and co-workers.32,33 QTAIM involves a discrete boundary partitioning scheme based on a topology of the electron density distribution. This theory provides a method of electron density distribution characterization by means of various descriptors, such as bond critical points (BCPs), or integrated properties (atomic charges, dipole moments, etc.). Afterward, Abramov’s investigations,34 and further Espinosa-Lecomte’s remarks,35,36 led to a method linking the charge density distribution features with the energy of a hydrogen bond. In the Espinosa-Lecomte approach, the weak interaction energy of a particular intermolecular contact is considered as being equal to approximately half of the estimated potential energy density value at the BCP (the proportionality factor is in volume atomic units). Therefore, the analysis was supplemented by the energetic considerations regarding hydrogen bond contacts in line with Espinosa-Lecomte approach. Tables 2 and 3 summarize the detected bond paths (BPs) and bond critical points (BCPs), whereas the 6m2tU molecular 7766

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C

Article

the strong hydrogen bonding contacts involving the H1 and H3 hydrogen atoms. Finally, bond ellipticity indices show that all pyrimidine ring bonds can be treated somewhat in-between single and double, among which the C5−C6 bond has the largest double bond character (ε = 0.22). In general, our results go along with the corresponding values obtained for charge density studies of related molecular crystals, such as 2-thiouracil or thymine crystals described elsewhere.37,38 In the crystalline state every molecule interacts with its surrounding. Therefore, a number of distinct intermolecular BPs and BCPs can be detected. Consequently, the three main structural dimeric motifs described previously (i.e., C, D and S, Figure 1) were supplemented by five additional ones. These were selected on the basis of the observed BPs. All the motifs are presented in Figure 3. The numerical parameters for the possible interactions, which were found significant or worth exploring, are summarized in Table 3. Three new motifs, i.e., M1, M2, and M3, involve interactions between the methyl group and the sulfur atom. On the basis of the charge density property values at BCPs, it can be assumed that M1 is the most favorable motif among the group due to the observed extra H5···S2 and N1···O4 bond paths and their features. Indeed, the M1 dimer is characterized by a quite significant interaction energy value according to the CRYSTAL single point calculation (Table 4). Surprisingly, it is twice as well stabilized as the π-stacking S motif, which remained unnoticed in our previous analysis.24 The charge density study clearly shows that the methyl group is engaged in a variety of both in-layer (M2) and off-layer (M1 and M3) interactions. This goes along with the previous remarks, when analyzing a set of uracil derivatives. Finally, the two motifs, H1 and H2, were chosen to illustrate the existing weak H···H close contacts, which are, though, beyond the range of the sum of van der Waals radii. H1 is the in-layer contact, whereas H2 is assigned to the off-layer interaction. Hydrogen-bonded chains (formed by C units) are characterized by three BPs, referring to the N−H···O, C−H···O, and S···H−C interactions. Table 3 clearly shows that the N−H···O hydrogen bond is the strongest one. The particular strength of the N−H···O contact is additionally confirmed by the deformation density and Laplacian maps, which show the polarization of the oxygen atom toward the H1 atom of the neighboring molecule (Figure 4). A significant polarization is

Table 2. Selected QTAIM Parameters of Covalent Bonds at BCPsa bond

R/Å

R1/Å

R2/Å

ρ(rBCP)/ e·Å−3

∇2ρ(rBCP)/ e·Å−5

ε

S2−C2 O4−C4 N1−C2 N1−C6 N3−C4 N3−C2 C4−C5 C5−C6 C6−C7 N1−H1 N3−H3 C5−H5 C7−H7A C7−H7B C7−H7C

1.675 1.237 1.359 1.375 1.387 1.359 1.432 1.361 1.493 1.030 1.030 1.083 1.077 1.077 1.077

0.678 0.809 0.790 0.803 0.805 0.801 0.743 0.644 0.801 0.755 0.755 0.693 0.703 0.704 0.709

0.997 0.428 0.569 0.572 0.582 0.558 0.689 0.718 0.693 0.275 0.275 0.390 0.374 0.373 0.368

1.48 2.71 2.29 2.14 2.11 2.23 1.96 2.19 1.78 2.34 2.29 1.91 1.86 1.84 1.81

−6.7 −23.7 −23.6 −21.2 −20.1 −22.2 −17.5 −19.4 −14.2 −41.3 −34.3 −21.6 −19.8 −19.6 −18.7

0.06 0.07 0.10 0.12 0.11 0.12 0.16 0.22 0.09 0.02 0.02 0.04 0.05 0.05 0.05

a

R, distance between bonded atoms; R1 and R2, distances from 1st and 2nd atom to the bond critical point, respectively; ρ, electron density; ε, bond ellipticity.

graph is presented in Figure 2. BPs and BCPs were found between all atoms in the molecule, showing that the overall topology of the electron density is consistent. In most of the cases (i.e., for C−C, C−N, C−O, C−H, and N−H bonds) the values of the electron density and Laplacian are within typical ranges for such bond types. As expected, the electron density value at the C−S bond critical point is much lower than that for the C−O bond, and it is generally the lowest one among the analyzed covalent bonds. Additionally, the related Laplacian value is less negative than those of all other bonds. Significant differences can be also observed between N−H and C−H bond types. In the case of both N−H bonds, the BCPs are shifted about 0.1 Å toward the hydrogen atoms, when referred to the analogue value of the C−H contacts. Such a trend has three origins, among which first of all (i) the electronegativity difference between nitrogen and carbon atoms (a dominating effect), (ii) the nature of a nitrogen atom which is more expanded than a carbon atom, as it owes more electrons (Table 5), and a secondary effect related to (iii)

Table 3. Selected QTAIM Parameters for Weak Intermolecular Interactions at the Positions of Respective BCPsa interaction

motif

R/Å

R1/Å

R2/Å

ρ(rBCP)/e·Å−3

∇2ρ(rBCP)/e·Å−5

G(rBCP)/kJ·mol−1·a0−3

V(rBCP)/kJ·mol−1·a0−3

O4···H1 (#1) O4···H7A (#1) H5···S2 (#1) S2···H3 (#2) S2···N3 (#3) N1···C4 (#3) H7A···C5 (#3) S2···H5 (#4) N1···O4 (#4) S2···H7B (#4) S2···H7A (#5) S2···H7C (#6) H7B···H7B (#7) H7B···H7B (#8)

C C C D S S S M1 M1 M1 M2 M3 H1 H2

1.764 2.414 3.286 2.378 3.500 3.219 2.946 3.060 3.389 3.238 3.198 2.996 2.332 2.695

1.150 1.394 1.239 1.581 1.882 1.639 1.223 1.889 1.698 2.019 1.957 1.881 1.166 1.348

0.617 1.031 2.056 0.799 1.619 1.583 1.726 1.249 1.691 1.229 1.265 1.117 1.166 1.348

0.27 0.07 0.03 0.15 0.04 0.04 0.03 0.04 0.03 0.02 0.03 0.04 0.03 0.02

1.7 1.1 0.2 0.9 0.5 0.5 0.4 0.5 0.5 0.3 0.4 0.4 0.6 0.4

66.9 23.6 4.7 29.2 11.1 10.7 8.8 11.3 9.3 6.3 7.4 9.0 11.5 7.5

−86.5 −17.0 −3.4 −35.1 −8.1 −8.0 −5.8 −7.9 −6.2 −4.1 −5.1 −6.9 −7.4 −4.5

G, kinetic energy density; V, potential energy density. Symmetry transformations: (#1) x + 1, −y + 0.5, z − 0.5; (#2) −x + 2, −y + 1, −z; (#3) x − 1, y, z; (#4) x, −y + 0.5, z + 0.5; (#5) −x + 1, y + 0.5, −z + 0.5; (#6) −x + 2, y + 0.5, −z + 0.5; (#7) −x + 2, −y, −z; (#8) −x + 1, −y, −z. a

7767

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C

Article

Figure 2. (a) Labeling of atoms and estimation of their thermal motion parameters, ADPs (50% probability level), for 6m2tU at 90 K after final multipole refinement. (b) Experimental molecular graph for 6m2tU (small red spheres, BCPs; small magenta sphere, RCP; orange solid lines, BPs).

Table 4. Selected Dimer Motif Intermolecular Interaction Energies Calculated at the DFT(B3LYP)/pVTZ Level of Theory, Accounted for BSSE and Dispersiona motif

Eint,I/kJ·mol−1

−1 Eopt int,I/kJ·mol

Etop,I/kJ·mol−1

C D S M1 M2 M3 H1 H2

−52.2 −32.5 −12.5 −24.7 −8.5 −8.3 0.9 4.1

−52.4 −34.0 −12.1 −25.4 −8.5 −8.6

−53.5 −35.1b −10.9 −9.1 −2.6 −3.5 −3.7 −2.3

a

Eint,I and Eopt int,I are the interaction energies of dimers extracted from the X-ray determined crystal structure and the optimized structure of 6m2tU, respectively. The last column includes the corresponding hydrogen bond energy values (and other short contacts; Etop,I) derived from the experimental potential energy density according EspinosaLecomte’s approach. bAbout −35 kJ·mol−1 for all tested charge density models of sulfur atom.

is additive (as is dispersion energy to some extent) and that is also the main component of hydrogen-bonding interaction, the partial contributions coming from all three contact types were simply added leading to a final value of −53.5 kJ·mol−1. The energy density results indicate a quite significant influence of the C−H···O contact on the C motif stability. These remarks together with the observations made on the M-type dimers suggest that the methyl group interactions with the adjacent molecules are crucial, both for the molecular layer stability and for the strength of interlayer contacts. This is naturally a result of the overall 6m2tU crystal packing, where the relatively bulky methyl substituent substantially influences the preferred mutual orientation adopted by 6m2tU molecules. The second most stabilized motif is the D dimer. It is formed by two equal S···H close contacts. Once again, the dimer interaction energy calculated via the CRYSTAL package is very well reflected by the charge density derived hydrogen bond energy. Furthermore, this energy remains very much the same, no matter what sulfur atom model is applied. As can be expected on the basis of Table 4, EspinosaLecomte’s approach is indeed much more appropriate in the case of stronger intermolecular interactions than weaker and less electrostatic ones, such as M and H-type dimeric interactions. This observation is also illustrated in Figure 5. A deeper insight into electronic charge distribution can also be achieved by the analysis of atomic charges. These are presented in Table 5 along with the respective atomic basin volumes.

Figure 3. Bond paths (BPs) and respective bond critical points (BCPs) for all structural motifs present in the 6m2tU crystal structure (C, N− H···O chain; D, N−H···S dimer; S, π-stacking; M1, M2, M3, interaction with methyl groups; H1, H2, H···H contacts). Table 3 contains the dimer symmetry transformation information.

also visible for the H1 hydrogen atom in the Laplacian map, when the hydrogen quadrupolar parameters are refined. The two remaining interactions, i.e. C−H···O and S···H−C, also contribute to a particular stabilization of the C motif. This all makes the supramolecular chains very favorable while forming the crystal. The total interaction energy of the 6m2tU molecules building the C unit amounts to about −52 kJ·mol−1 (Table 4), which corresponds to the energy of two medium-strength hydrogen bond contacts. The C motif energy value obtained from the single-point CRYSTAL calculation is well reproduced by Espinosa-Lecomte’s estimation grounded on the potential energy density at the hydrogen bond critical point. As electrostatic interaction energy 7768

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C

Article

Figure 4. (a) Deformation density map (N1−H1−O4 plane, contours at 0.05 e·Å−3, blue solid lines, positive contours; red dashed lines, negative contours). (b) Laplacian of electron density map (contours at e·Å−5 shown in logarithmic scale).

Table 5. AIM Charges (QAIM) and Volumes (VAIM) for the 6m2tU Molecule Present in the Crystal Structure atom

QAIM/e

VAIM/Å3

S2 O4 N1 N3 C4 C2 C5 C6 C7 H1 H3 H5 H7A H7B H7C

−0.273 −1.161 −1.001 −1.063 +1.330 +0.558 +0.027 +0.278 −0.100 +0.510 +0.415 +0.116 +0.112 +0.140 +0.118

32.815 16.622 12.154 13.318 5.546 7.756 11.019 8.677 10.935 2.234 3.153 6.497 5.761 6.109 6.605

Two of them are engaged in the D and S motifs. Nevertheless, a detailed analysis of the sulfur atom is a bit hampered by the unrefined multipolar populations. The transferred UBDB parameters impose a certain fixed distribution of the lone electron pairs (and consequently valence shell charge concentrations (VSCCs)) on the sulfur atom. One of VSCCs is involved in the interaction with the H3 hydrogen atom, which stabilizes the D motif. The second one points toward the H5 atom. The remaining atoms interacting with the sulfur atom are engaged in less significant intermolecular contacts. On the whole, the sulfur atom seems to substantially affect the crystal stability and so the formed crystal architecture. At the end, it is worth noting that the presented topological features correspond very well with the respective values obtained for the theoretical charge density model (i.e., the one derived on the basis of the computed structure factors; for details see the Supporting Information). 3.3. Electrostatic Potential Analysis. The multipole model allows for reconstructing the electron density distribution of a crystal structure, and on that basis, for deriving various electrostatic properties, e.g., electrostatic potential (ESP). In turn, such an electrostatic potential mapped onto the molecular surface may provide some additional insight into the nature of intermolecular contacts inside the crystalline state. As suggested

Figure 5. Correlation plot of Etop,I (sum of interaction energies obtained via the Espinosa-Lecomte approach) vs Eint,I (CRYSTAL-computed total interaction energy). Dashed line, Etop,I = Eint,I, serves as a reference line.

Oxygen and nitrogen atoms are, as expected, negatively charged to a high degree. There is just a small charge difference between the N1 and N3 atoms (ca. 0.06 e), which can be disregarded. Carbon atom ring members reveal a rather predictable pattern, i.e. the atoms having more electronegative neighbors are more positively charged. The C7 carbon atom (from the methyl group) is the only one bearing a negative charge, which is anyway rather small (−0.10 e). All hydrogen atoms are positively charged. Among them, H1 and H3, primarily connected to electronegative nitrogen atoms and, second, involved in the N− H···O and N−H···S hydrogen bonds, are characterized by the most positive charges (about +0.5 e). Even though the sulfur atom exhibits a rather small negative charge (ca. −0.27 e), it forms many different interatomic contacts in the solid state. This is mainly due to its size (about 33 Å3) and high polarizability. As a result, seven diverse intermolecular BCPs and the respective BPs, leading to the sulfur atom, were found. 7769

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C

Article

Figure 6. (a) Molecular electrostatic potential mapped on the Hirshfeld surface (semitransparent representation) together with two neighboring molecules forming C and D motifs (units are e·Å−1; small red spheres, BCPs; yellow lines, BPs). (b) Layer fragment with Hirshfeld surface colored with molecular ESP. (c) Layer fragment with Hirshfeld surface colored with ESP coming from a central molecule together with its surrounding in a crystal structure.

by Spackman et al.,39 the problem of overlapping molecular domains can be overcome when Hirshfeld surfaces are employed. Therefore, we have computed the electrostatic potential solely for the central molecule and mapped it onto the Hirshfeld surface. Figure 6a illustrates the Hirshfeld surface along the Cand D-type dimeric motif bond paths. The nature of these interactions is well visible as the complementary ESP region contacts. In the case of D, the electropositive region of H3 hydrogen atoms interacts with the electronegative region of the S2 sulfur atom. Essentially the same pattern is observed for the C motif. ESP distribution clearly shows why 6m2tU forms layers, where molecules are bound by strong, electrostatic in nature, hydrogen bonds. Secondary organization is obtained by the πstacking interactions, being more dispersive in nature. Naturally, in the solid state the molecular electron density (and thus ESP) is influenced by the surrounding molecules. The ESP computed by taking into account contributions of the neighboring moieties is presented in Figure 6c. Interactions with other molecules change the ESP dramatically. The influence of the neighboring molecules counterbalances the electrostatic potential of a single 6m2tU moiety, leading to a more uniform electrostatic potential distribution. 3.4. Crystal Morphology and Structural Motif Stabilization Energy. In general, 6m2tU forms two types of crystals from water solution. These are crystal prisms with well-defined facets (Figure 7) and usually less numerous thin elongated plates not suitable for high quality X-ray experiments. What is more, as

Figure 7. 6m2tU monocrystal with the indexed crystal faces.

already noted, 6m2tU crystals exhibit some tendency to twin. The observed crystal morphology, growth, or twinning, are all complex problems involving kinetic and thermodynamic processes, and obviously solvation effects. Nevertheless, a detailed analysis of the crystal architecture features may shed some light on the above-mentioned aspects. Therefore, taking into account the layered crystal network of 6m2tU, the interlayer interactions become worth exploring. As clearly seen in Figure 8, the previously analyzed slabs parallel to the (102) crystal plane, are not the only easily distinguished ones.24 The regularity of 6m2tU crystal architecture and the simplicity of molecules allow for detecting differently oriented 7770

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C

Article

Figure 8. (a) Projection of the 6m2tU crystal packing along the Y axis. (b) Selected molecular slabs parallel to different crystal planes illustrated along the X axis.

usually significantly change the properties of pure substance crystal faces. Having carefully looked at the 6m2tU crystal morphology and its face indices, one can notice that the crystal is elongated in the [100] direction, exhibiting a well-developed {010}-type crystal faces, medium-size {001}-type ones, and the smallest facets assigned to the {102} form. There are also other less significant crystal faces visible, e.g. the {021̅}-type ones. Among the studied crystal faces, {001} are characterized by the lowest interlayer interaction energy of the corresponding molecular slabs and the highest surface energy value (Figure 8a, Table 6). The {100} set of facets seems to be very similar in nature, however, less energetically extreme. The related molecular slab types, SB3 and SB4, respectively, interact, among others, via the N−H···O hydrogen bonds (motif C), what leads to the high interlayer interaction energy magnitudes. This may suggest that the crystal growth is faster along the [001] or [100] direction, and the formed surfaces of the adequate facets are limited in size. Indeed, depending on the solvent, either [100] or [001] is parallel to the elongation direction of the crystal. In water solution a 6m2tU crystal is elongated along the X axis. Nevertheless, it is exposed with the {102} facet types in that direction, instead of {100}, due to their better stability. According to the literature reports,13 in the 1:1 ratio mixture of glacial acid and water, the obtained crystals are extended along the [001] direction. Therefore, this is a complementary situation to the one observed in water solution. Molecules of 6m2tU are easily attached to such molecular surfaces contributing to the crystal elongation process. However, when such a surface is formed, it might attract also water molecules. Water molecules may in turn improve the surface stability and slow down the crystal growing in the corresponding direction. Such solvent influence may also cause some defects in the resulting crystal structure or stimulate twinning. On the other hand, in line with our expectations, the most developed {010}-type crystal facet corresponds to the molecular slabs (SB2) of the greatest interlayer interaction energy value (i.e., the least interlayer interaction strength) and the smallest surface energy. The molecular layers of the SB2 type interact via the N−H···S hydrogen bonds (motif D) and other less significant contacts (motifs M2, M3, H1, and H2). The remaining crystal faces listed in Table 6, {110} and {021̅}, constitute examples of more rough molecular surfaces, which are less likely formed. It might also result from the specificity of

molecular layers. These various molecular layers also interact with one another in different manners, involving either mostly dispersive contacts, or some stronger hydrogen bond interactions. For the purpose of the analysis, we have chosen molecular slabs related to the measured crystal face indices and to the ones previously mentioned in the literature.13 To characterize each of the slab types, we considered two energetic parameters, i.e., interlayer interaction energy and surface energy. The interlayer interaction energy was calculated at the DFT(B3LYP)/pVTZ level of theory following the procedure described in the literature,24,40,41 whereas the evaluated surface energy constitutes a rough estimation provided on the basis of the classical approach applied to metal or metal oxide surfaces. The obtained energetic values for the selected molecular slabs are put together in Table 6, and all of the computational details are available from the Supporting Information. Table 6. Interlayer Interaction Energy (Eintl) and Estimated Surface Energy (Esurf) of the Selected Molecular Slabsa slab

Miller index

Eintl/kJ·mol−1

Esurf/J·m−2

SB1 SB2 SB3 SB4 SB5 SB6

(102) (010) (001) (100) (110) (021̅)

−48 −33 −102 −83 −93 −74

0.104 0.145 0.296 0.220 0.230 0.255

a

Miller indices indicate the crystal plane family being parallel to a chosen slab (see Figure 8).

In general, the calculated interlayer energies, together with the surface energy magnitudes, should explain to some extent the crystal shape and its growth tendencies. The strength of an interslab contact provides some information about the growth rate in the related crystallographic direction, which should be more stimulated when the adjacent molecular layers interact stronger. In turn, the energy estimated per molecular area somewhat reflects the relative stability of the molecular surfaces being compared. The greater the surface energy is, the less stable the molecular surface and, thus, the corresponding crystal faces are less likely formed. Of course, one should keep in mind the other factors governing crystal formation. Solvent molecules 7771

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C

Article

Scheme 1. Two Possible Tautomeric Forms of 6m2tU Related by the Proton Transfer between Nitrogen and Sulfur Atoms, (a) Tautomer II (Oxo-Thiol Form) and (b) Tautomer I (Oxo-Thione Form), and Hypothetical Resonance Structures (c and d) Suggesting the Charge Redistribution Across the S−C−N Structural Fragments

crystallization process, its kinetics, and from possible unfavorable consequences for other, more stable, crystal faces. Despite many different effects affecting the crystal growth, our study shows that the simple analysis of energetic features of crystal network motifs may also lead to meaningful conclusions. After all, crystal thermodynamic properties, solvation effects, and crystallization kinetics in general, are all factors responsible for the final product, i.e., crystal and its features. Much of the information is also reflected in the resulting crystal architecture. Generally, balanced and advantageous stabilization energy contributes to the overall thermodynamic and mechanical crystal stability. 3.5. Sulfur Atom Residual Density Interpretation. Difficulties with the interpretation of the residual density around the sulfur atom (for data sets collected at both temperatures) and its characteristics led us to investigate the phenomenon of tautomerism in the solid state. In general, nucleic acid bases may exist, with different doses of probability, in several tautomeric forms. Numerous computational and spectroscopic studies have already been devoted to this problem, as it is believed to be tightly related to the base-pairing mutations.5,6,42,43 In the case of uracil derivatives the energy difference between the normal and rare tautomers is greater than in the case of guanine or cytosine. For monosubstituted uracil modifications the oxo-thione form is the most favorable one, especially in the solid state (Scheme 1b).44 On the other hand, according to the recent computational and spectroscopic studies of thio-uracil derivatives, single molecules might be quite easily photoinduced and transformed from the thione to thiol form (Scheme 1a,b).5,6 Depending on the chemical environment, the tautomeric transformation process might be either an intramolecular or intermolecular proton transfer. Nowadays a lot of research deals with the proton transfer phenomenon taking place in the crystal lattice.45−47 Is it then possible that 6m2tU exists in more than one tautomeric form in its crystal, e.g., the ones shown in Scheme 1a,b? To approach this problem, we performed a set of computational investigations together with a careful analysis of the unexplained residual density features. The two high residual density peaks (Δρmax = +0.64 e·Å−3) observed in the vicinity of sulfur atom are presented in Figure 9a. Their location out of the plane of the 6m2tU aromatic ring is not supported by the UBDB-transferred sulfur atom deformation parameters for the CS fragment (Figure 9b). Nevertheless, if the second tautomer (tautomer II) is taken into consideration, one can find that the arrangement of the unexplained charge density resembles the electron lone pair distribution of the C− S−H group located in the molecular plane (Figure 9c,d). Therefore, to supply this assumption, we modeled a hypothetical geometry of tautomer II in the crystal lattice. For that purpose, an artificial crystal network, analogous to the one of tautomer I (the oxo-thione form) and based on tautomer II, was created and

Figure 9. (a) Residual electron density near the sulfur atom (isosurface drawn at +0.3 e·Å−3). (b) UBDB-reconstructed static deformation density for the sulfur atom (+0.2 e·Å−3) for computed tautomer I. (c) Static deformation density taken from the ELMAM library (+0.2 e·Å−3) for computed tautomer II. (d) Deformation density taken from the UBDB databank (+0.2 e·Å−3) for computed tautomer II.

optimized in the CRYSTAL package (for details see the Supporting Information). Having the potential molecular geometry, we could generate the deformation density maps of tautomer II with the aid of the UBDB and ELMAM databases (Figures 9c,d).21,48 Indeed, the shape and positions of the C−S−H sulfur atom valence charge density fitted quite well the earlier unexplained residual density features. This seems to be an evidence of the presence of both tautomers in the crystal structure. Unfortunately, it was not possible to refine those fine features of the electron density distribution (the sulfur atom has 10 core electrons which presumably strongly dominate the description of this atom). Thus, we could only estimate the percentage of the possible second tautomer content at approximately 10% (calculated as Δρmax × 100%/Pv,S). The computational results revealed about 90 kJ·mol−1 energy difference between the molecule of the oxo-thione form and that of the hypothetically formed tautomer II (Supporting Information). Despite the substantially less beneficial molecular energy of tautomer II with respect to tautomer I, the cohesive energy of the hypothetical crystal structure constructed of pure tautomer II seems to be more advantageous, which is of about 30 kJ·mol−1 (tautomer I, Ecoh = −132 kJ·mol−1; tautomer II, Ecoh = −162 kJ·mol−1). Nevertheless, considering the balance of the 7772

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C

Article

However, the final judgment cannot be based solely on such premises. One could also think of the oxidation process taking place in a crystal, which could lead to the S−S bond formation. This is a very common phenomenon encountered in biological systems containing thiouracil fragments.7,50−53 These moieties relatively easily undergo the oxidation reaction when exposed to iodine or affected by X-ray radiation in air-saturated aqueous solution. In the case of 6m2tU crystal, the shortest sulfur···sulfur contacts are characterized by the interatomic distance equal to 3.9424(2) Å, which is close to the sum of the van der Waals radii (Figure 11).

molecular and cohesive energy, the oxo-thione form still remains the more beneficial one. The energy results obtained for the selected dimer motifs show much more negative values for the hydrogen-bonded oxo-thiol species. This surely contributes to the overall crystal lattice stability value. The corresponding dimer interaction energy values are grouped in Table 7. These results Table 7. Selected Dimer Motif Intermolecular Interaction Energies Calculated for the Optimized Hypothetical Tautomer II Crystal Structure (Analogous to That of Tautomer I) at the DFT(B3LYP)/pVTZ Level of Theory, Accounted for BSSE and Dispersion motif

−1 Eopt int,II/kJ·mol

C D S M1 M2 M3 H1 H2

−61.1 −82.4 −1.8 −20.4 +1.0 −3.0 − −

reflect stronger electrostatic interactions, and thus an increased stability of C and D units for tautomer II, when compared to the oxo-thione form of 6m2tU (Table 4). This is caused by a higher charge separation. As a consequence, the potential of the N3 atom is more negative, whereas for the S−H fragment, it is more positive, as illustrated in Figure 10. On the other hand, a more pronounced charge partition causes lower values of interlayer contacts.

Figure 11. Closest contacts between sulfur atoms, S···S, present in the 6m2tU crystal structure illustrated together with the largest residual electron density peaks (isosurfaces drawn at +0.25 e·Å−3).

Nevertheless, the residual peaks are neither located at the line connecting the neighboring sulfur atoms, nor perpendicular to the C−S···S (−x+1, −y+1, −z) plane. Interestingly, the residual density peaks match the direction indicated by the S···S (x−1, y, z) contacts shown in Figure 11. In this case, uracil rings are mutually arranged similarly to the S−S connected 2-thiouracil fragments described in Wenska et al.54 Naturally, the intersulfur distance is over two times longer in the 6m2tU crystal than for the adequate S−S bond. It may suggest that in the crystal small part of molecules is oxidized due to their interaction with the Xray beam for example. However, the effect of a fraction of the S− S bonded 6m2tU molecules would presumably have a more significant effect on the whole collected charge density data. All in all, in order to fully investigate the problem, solid state NMR and/or some photocrystallographic studies could be an option and may lead to the final answer, as it is difficult to grow 6m2tU crystals to the size suitable for the neutron diffraction studies. Such investigations are scheduled for the near future.

Figure 10. Electrostatic potential (e·a0−1) mapped on the electron density isosurface of tautomers I and II, computed at the DFT(B3LYP)/aug-cc-pVDZ level of theory in the Gaussian package.49

Although the above observations and results seem to be quite convincing, the explanation of the charge density features requires some additional research. Generally, tautomerism is very rarely observed in the solid state, however, it could be stimulated by light pulses or other form of irradiation (also X-rays). Other observations and tests, e.g., Hirshfeld rigid-bond test, bond lengths, charge density features in the vicinity of nitrogen atoms, etc., do not support this hypothesis. An alternative explanation employs the contribution from the hypothetical resonance structures (Scheme 1c,d) to the charge density distribution. In such a case, the electron density should be slightly moved from the nitrogen atoms toward the sulfur atom. This can, in turn, account for the “excess” of the electron density at the sulfur atom and for more single bond character of the C−S bond, influencing the charge density distribution at the S atom. Additionally, the N−C bonds should gain some more double bond character.

4. SUMMARY AND CONCLUSIONS The study provides a detailed charge density distribution analysis together with comprehensive energetic investigations. Furthermore, the findings constitute a link between crystal network features, electronic structure, molecular motif interaction energy, and crystal morphology. Results show the importance of careful analysis of obtained charge density data and attentive charge density distribution model evaluation. A proper interpretation of anharmonic motion parameters resulting from the multipole refinement might be especially problematic. Therefore, we emphasize the need of a very low temperature (down to 10 K) X7773

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C ray diffraction experiment, which in our case allowed for excluding the supposition of anharmonic behavior of the sulfur atom. On the other hand, once the model was properly derived, a number of interesting relations could have been found. Our results revealed accordance between the topologically estimated hydrogen bond energy and the computational values, when the electrostatic component is dominant. It also occurred that the methyl group and sulfur atom significantly affect crystal packing and contribute to a variety of intermolecular contacts. Nevertheless, the strongest interactions observed in the crystal structure are of electrostatic type. It is well reflected in the generated molecular electrostatic potential surfaces of the complementary character. Additionally, 6m2tU exhibits a layered architecture which can be quite easily associated with its crystal morphology. Simple analysis of the interslab interaction energies and surface energy values led to the rough explanation of the crystal elongation direction and preferably created facets, such as the well-developed {010} crystal form. A lot of effort was also put worth so as to explain the residual charge density features, especially in the vicinity of the sulfur atom. The residual density peaks seem to be too significant to result solely from the limitations of the Hansen−Coppens formalism. Complex analysis of the charge density distribution resulted in a hypothesis concerning the presence of a second tautomer in the crystal structure, the oxo-thiol form. However, no other observations and tests supported this idea. Moreover, it was impossible to reliably refine a disordered model composed of two tautomers. Nonetheless, the percentage contribution of each tautomer might be estimated as 1:9 for the benefit of the 90 kJ·mol−1 more stable oxo-thione form. Interestingly, despite the higher molecular energy of the oxo-thiol tautomer, it may form more advantageous crystal architecture due to much more pronounced electrostatic interactions, as indicated by the electrostatic potential and dimer interaction energy values. To finally confirm or reject the hypothesis, a set of further investigations need to be conducted. Another potential explanation employs the contribution of resonance forms, which affect the charge density distribution most significantly at the sulfur atom. A possible oxidation of the 6m2tU moiety leading to the S−S bound thiouracil species in the crystal lattice was also considered and seems interesting, but cannot be fully approved either. At the end, it should be noted that the new version of the LSDB program provides the MOPRO-type input files with the transferred UBDB atom type parameters and template files containing suggested atomic restraints and constraints. The current LSDB version is available at crystal.chem.uw.edu.pl/ software.



ACKNOWLEDGMENTS



REFERENCES

P.M.D. and K.N.J. would like to thank the National Science Centre in Poland (grant no. NN204 129138) for financial support, whereas R.K. thanks the Foundation for Polish Science for funding within the International Ph.D. Projects’ programmes. K.N.J. and R.K. also acknowledge the Iuventus Plus Ministry of Science and Higher Education Grant No. 120000-501/59 PM121 for financial support. The authors thank the Wrocław Centre for Networking and Supercomputing for providing computer facilities. The authors thank Christian Jelsch and Benoı̂t Guillot (Nancy, France) for substantial help with the MOPRO package.

(1) Ajitkumar, P.; Cherayil, J. D. Thionucleosides in Transfer Ribonucleic-Acid − Diversity, Structure, Biosynthesis, and Function. Microbiol. Rev 1988, 52, 103−113. (2) Bergstrom, D. E.; Leonard, N. J. Photoreaction of 4-Thiouracil With Cytosine − Relation To Photoreactions in Escherichia-Coli Transfer Ribonucleic-Acids. Biochemistry 1972, 11, 1−9. (3) Lipsett, M. N. Behavior of 4-Thiouridine in E. Coli s-RNA Molecule. Biochem. Biophys. Res. Commun. 1965, 20, 224−229. (4) Lipsett, M. N.; Peterkof, A. Enzymatic Thiolation of E. Coli sRNA. Proc. Natl. Acad. Sci. U.S.A. 1966, 55, 1169−1174. (5) Khvorostov, A.; Lapinski, L.; Rostkowska, H.; Nowak, M. J. Unimolecular Photochemistry of 4-Thiouracils. Photochem. Photobiol. 2005, 81, 1205−1211. (6) Khvorostov, A.; Lapinski, L.; Rostkowska, H.; Nowak, M. J. UVinduced Generation of Rare Tautomers of 2-Thiouracils: A Matrix Isolation Study. J. Phys. Chem. A 2005, 109, 7700−7707. (7) Miller, W. H.; Roblin, R. O.; Astwood, E. B. Studies in Chemotherapy. XI. Oxidation of 2-Thiouracil and Related Compounds by Iodine. J. Am. Chem. Soc. 1945, 67, 2201−2204. (8) Clercq, E. D.; Balzarini, J. Knocking Out Human Immunodeficiency Virus Through Non-nucleoside Reverse Transcriptase Inhibitors Used as Single Agents or in Combinations: A Paradigm for the Cure of AIDS? Farmaco 1995, 50, 737−747. (9) Tanaka, H.; Takashima, H.; Ubasawa, M. Synthesis and Antiviral Activity of 6-Benzyl Analogs of 1-[(2-Hydroxyethoxy)methyl]-6(phenylthio)thymine (HEPT) as Potent and Selective Anti-HIV-1 agents. J. Med. Chem. 1995, 38, 2860−2865. (10) Crooks, J. Thyroid and antythyroid drugs; Elsevier: Amsterdam, 1972. (11) Nugent, R. A.; Schlachter, S. T.; Murphy, M. J. Pyrimidine Thioethers: A Novel Class of HIV-1 Reverse Transcriptase Inhibitors with Activity Against BHAP-resistant HIV. J. Med. Chem. 1998, 41, 3739−3803. (12) Nakamura, H.; Noh, J. Y.; Itoh, K.; Fukata, S.; Miyauchi, A.; Hamada, N. Comparison of Methimazole and Propylthiouracil in Patients with Hyperthyroidism Caused by Graves’ Disease. J. Clin. Endocrinol. Metab. 2007, 92, 2157−2162. (13) Parry, G. S.; Strachan, F. Preliminary Investigation of Some Heterocyclic Compounds. Acta Crystallogr. 1958, 11, 303−304. (14) Fertey, P.; Argoud, R.; Bordet, P.; Reymann, J.; Palin, C.; Bouchard, C.; Bruyere, R.; Wenger, E.; Lecomte, C. A Mini-goniometer for X-ray Diffraction Studies Down to 4 K on Four-circle Diffractometers Equipped with Two-dimensional Detectors. J. Appl. Crystallogr. 2007, 40, 526−531. (15) APEX2; Bruker AXS Inc.: Madison, WI, 2010. (16) Blessing, R. H. An Empirical Correction for Absorption Anisotropy. Acta Crystallogr. 1995, A51, 33−38. (17) Palatinus, L.; Chapuis, G. Superflip − A Computer Program for the Solution of Crystal Structures by Charge Flipping in Arbitrary Dimensions. J. Appl. Crystallogr. 2007, 40, 786−790. (18) Petricek, V.; Dusek, M.; Palatinus, L. JANA2006. The crystallographic computing system; Institute of Physics: Praha, Czech Republic, 2006.

ASSOCIATED CONTENT

S Supporting Information *

Details of X-ray measurements, collected data, structural and multipolar refinement schemes, full QTAIM results, energetic comparison of different multipolar, TAAM and IAM models, and also cif data. This material is available free of charge via the Internet at http://pubs.acs.org.





Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest. 7774

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775

The Journal of Physical Chemistry C

Article

(19) Fuess, H., Ed., International Tables for Crystallography, Vol. C, Mathematical, physical and chemical tables; International Union of Crystallography: Chester, 2006. (20) Jelsch, C.; Guillot, B.; Lagoutte, A.; Lecomte, C. Advances in Protein and Small-Molecule Charge-Density Refinement Methods using MoPro. J. Appl. Crystallogr. 2005, 38, 38−54. (21) Jarzembska, K. N.; Dominiak, P. M. New Version of the Theoretical Databank of Transferable Aspherical Pseudoatoms, UBDB2011 − Towards Nucleic Acid Modelling. Acta Crystallogr. 2012, A68, 139−147. (22) Hansen, N. K.; Coppens, P. Testing Aspherical Atom Refinements on Small-Molecule Data Sets. Acta Crystallogr. 1978, A34, 909− 921. (23) Allen, F. H.; Bruno, I. J. Bond Lengths in Organic and MetalOrganic Compounds Revisited: X−H Bond Lengths From Neutron Diffraction Data. Acta Crystallogr. 2010, B66, 380−386. (24) Jarzembska, K. N.; Kubsik, M.; Kaminski, R.; Wozniak, K.; Dominiak, P. M. From a Single Molecule to Molecular Crystal Architectures: Structural and Energetic Studies of Selected Uracil Derivatives. Cryst. Growth Des. 2012, 12, 2508−2524. (25) Madsen, A. Ø. SHADE Web Server for Estimation of Hydrogen Anisotropic Displacement Parameters. J. Appl. Crystallogr. 2006, 39, 757−758. (26) Zhurov, V. V.; Zhurova, E. A.; Pinkerton, A. A. Optimization and Evaluation of Data Quality for Charge Density Studies. J. Appl. Crystallogr. 2008, 41, 340−349. (27) Sorensen, H. O.; Stewart, R. F.; McIntyre, G. J.; Larsen, S. Simultaneous Variation of Multipole Parameters and Gram-Charlier Coefficients in a Charge-Density Study of Tetrafluoroterephthalonitrile Based on X-ray and Neutron Data. Acta Crystallogr. 2003, A59, 540− 550. (28) Poulin, A. Ph.D. Thesis, Poznań, Poland, 2013. (29) Zhurov, V. V.; Zhurova, E. A.; Stash, A. I.; Pinkerton, A. A. Importance of the Consideration of Anharmonic Motion in ChargeDensity Studies: A Comparison of Variable-Temperature Studies on Two Explosives, RDX and HMX. Acta Crystallogr. 2011, A67, 160−173. (30) Koritsanszky, T.; Volkov, A.; Chodkiewicz, M. Struct. Bonding (Berlin); Springer-Verlag: Berlin, Germany, 2010; DOI: 10.1007/ 430_2010_32. (31) Dominiak, P. M.; Coppens, P. Finding Optimal Radial-Function Parameters for S Atoms in the Hansen-Coppens Multipole Model Through Refinement of Theoretical Densities. Acta Crystallogr. 2006, A62, 224−227. (32) Bader, R. F. W. Atoms in Molecules. A Quantum Theory; Oxford University Press: New York, 1994. (33) Bader, R. F. W.; Bond Path, A. A Universal Indicator of Bonded Interactions. J. Phys. Chem. A 1998, 102, 7314−7323. (34) Abramov, Y. A. On the Possibility of Kinetic Energy Density Evaluation from the Experimental Electron-Density Distribution. Acta Crystallogr. 1997, A53, 264−272. (35) Espinosa, E.; Lecomte, C.; Molins, E. Experimental Electron Density Overlapping in Hydrogen Bonds: Topology vs. Energetics. Chem. Phys. Lett. 1999, 300, 745−748. (36) Espinosa, E.; Molins, E.; Lecomte, C. Hydrogen Bond Strengths Revealed by Topological Analyses of Experimentally Observed Electron Densities. Chem. Phys. Lett. 1998, 285, 170−173. (37) Munshi, P.; Row, T. N. G. Intra- and Intermolecular Interactions in Small Bioactive Molecules: Cooperative Features From Experimental and Theoretical Charge-Density Analysis. Acta Crystallogr. 2006, B62, 612−626. (38) Jarzembska, K. N.; Goral, A. M.; Gajda, R.; Dominiak, P. M. Hoogsteen−Watson−Crick 9-Methyladenine:1-Methylthymine Complex: Charge Density Study in the Context of Crystal Engineering and Nucleic Acid Base Pairing. Cryst. Growth Des. 2012. (39) Spackman, M. A.; McKinnon, J. J.; Jayatilaka, D. Electrostatic Potentials Mapped on Hirshfeld Surfaces Provide Direct Insight Into Intermolecular Interactions in Crystals. CrystEngComm 2008, 10, 377− 388.

(40) Durka, K.; Jarzembska, K. N.; Kaminski, R.; Lulinski, S.; Serwatowski, J.; Wozniak, K. Structural and Energetic Landscape of Fluorinated 1,4-Phenylenediboronic Acids. Cryst. Growth Des. 2012, 12, 3720−3734. (41) Hoser, A. A.; Jarzembska, K. N.; Dobrzycki, L.; Gutmann, M. J.; Wozniak, K. Differences in Charge Density Distribution and Stability of Two Polymorphs of Benzidine Dihydrochloride. Cryst. Growth Des. 2012, 12, 3526−3539. (42) Les, A.; Adamowicz, L. Tautomerism of 2-Thiouracil and 4Thiouracil − Abinitio Theoretical-Study. J. Am. Chem. Soc. 1990, 112, 1504−1509. (43) Les, A.; Adamowicz, L. Theoretical Abinitio Study of The Protomeric Tautomerism Of 2-Hydroxypyrimidine, 4-Hydroxypyrimidine, and Their Derivatives. J. Phys. Chem. 1990, 94, 7021−7032. (44) Rostkowska, H.; Szczepaniak, K.; Nowak, M. J.; Leszczynski, J.; Kubulat, K.; Person, W. B. Studies of Thiouracils. 2. Tautomerism and Infrared-Spectra Of Thiouracils − Matrix-Isolation and Abinitio Studies. J. Am. Chem. Soc. 1990, 112, 2147−2160. (45) Thomas, L. H.; Blagden, N.; Gutmann, M. J.; Kallay, A. A.; Parkin, A.; Seaton, C. C.; Wilson, C. C. Tuning Proton Behavior in a Ternary Molecular Complex. Cryst. Growth Des. 2010, 10, 2770−2774. (46) Parkin, A.; Seaton, C. C.; Blagden, N.; Wilson, C. C. Designing Hydrogen Bonds with Temperature-Dependent Proton Disorder:? The Effect of Crystal Environment. Cryst. Growth Des. 2007, 7, 531−534. (47) Parkin, A.; Woźniak, K.; Wilson, C. C. From Proton Disorder to Proton Migration: A Continuum in the Hydrogen Bond of a Proton Sponge in the Solid State. Cryst. Growth Des. 2007, 7, 1393−1398. (48) Domagala, S.; Fournier, B.; Liebschner, D.; Guillot, B.; Jelsch, C. An Improved Experimental Databank of Transferable Multipolar Atom Models − ELMAM2. Construction Details and Applications. Acta Crystallogr. 2012, A68, 337−351. (49) Frisch, M. J.; et al. Gaussian 09; Gaussian, Inc.: Wallingford, CT, 2004. (50) Sorbo, B.; Lundberg, B. Radiolysis of 4-Thioruracil and Of 4Thiouracil Moieties of 4-Thiouridine and Transfer Ribonucleic Acid From Escherichia Coli in Aqueous Solutions. Biochim. Biophys. Acta 1968, 157, 445−454. (51) Lipsett, M. N. Isolation of 4-Thiouridylate Disulfide From Oxidized Transfer Ribonucleic Acid of Escherichia Coli. J. Biol. Chem. 1967, 242, 4067−4071. (52) Lipsett, M. N.; Doctor, B. P. Studies on Tyrosine Transfer Ribonucleic Acid: A Sulfur-Rich Species From Escherichia Coli. J. Biol. Chem. 1967, 242, 4072−4077. (53) Coleman, R. S.; McCary, J. L.; Perez, R. J. Thionucleoside Disulfides as Covalent Constraints of DNA Conformation. Tetrahedron 1999, 55, 12009−12022. (54) Wenska, G.; Filipiak, P.; Asmus, K.-D.; Bobrowski, K.; Koput, J.; Marciniak, B. Formation of a Sandwich-Structure Assisted, Relatively Long-Lived Sulfur-Centered Three-Electron Bonded Radical Anion in the Reduction of a Bis(1-substituted-uracilyl) Disulfide in Aqueous Solution. J. Phys. Chem. B 2008, 112, 10045−10053.

7775

dx.doi.org/10.1021/jp312158m | J. Phys. Chem. C 2013, 117, 7764−7775