Hydrogen-bond interactions with an S atom as an acceptor are important in biological processes. For example, sulfur-containing nucleosides are components of the anticodon of transfer RNAs. They exhibit the same arrangement of hydrogen-donor and -acceptor groups as unmodified nucleosides, but the replacement of an O with an S atom induces changes in their properties and interactions. The thio residue can be selectively photoactivated, so that it is used as an intrinsic photolabel to probe the nucleic acid structure and to identify interactions within nucleic acids or between nucleic acids and proteins (Favre et al., 1998). Furthermore, the enhanced base-pairing specificity of thionucleosides can be utilized, for example, in the design of antisense oligonucleotides (Testa et al., 1999).
Because of its reduced electronegativity, an S atom should be a weaker hydrogen-bond acceptor than an O atom. Many theoretical and experimental investigations have been concerned with the stability of the N-HO and the N-HS hydrogen bonds, but no clear trend has emerged. Ab initio energy calculations (Sponer et al., 1997; Basilio Janke et al., 2001) and a study of the thermodynamics of RNA duplexes containing thiouridine (Testa et al., 1999) showed that a base pair connected by an N-HO hydrogen bond is more stable than that connected by an N-HS hydrogen bond. In the case of 2-thiouridine, the Watson-Crick base pair with adenine, which is connected by an N-HO hydrogen bond, is preferred over the wobble base pair with guanine, which is linked by an N-HS hydrogen bond. In contrast, 4-thiouridine increases the stability of the wobble base pair compared with the Watson-Crick base pairing. However, the IR spectroscopic red shift of the N-H stretching frequency indicated that N-HS is comparable or even stronger than the N-HO interaction (Lautié & Novak, 1980; Biswal & Wategaonkar, 2009).
In order to study the stability of the N-HS hydrogen bond in the presence of a competitive carbonyl O atom as an acceptor, we cocrystallized pyrimidin-2-amine and N-(6-acetamidopyridin-2-yl)acetamide, respectively, with the antithyroid drug 6-propyl-2-thiouracil, also known as propylthiouracil. It inhibits the synthesis of thyroid hormones and has been used for the treatment of hyperthyroidism caused by Graves' disease (Cooper, 2005). Because of its risk of serious liver injury, 6-propyl-2-thiouracil is used as a second-line drug for patients who are intolerant of other therapies (Bahn et al., 2009).
We chose pyrimidin-2-amine because of its adjacent amine and imine groups resembling the donor-acceptor site of adenine. Since it has a mirror plane bisecting the molecule along the C-NH2 bond, one pyrimidin-2-amine molecule may be hydrogen bonded to two 6-propyl-2-thiouracil molecules. Indeed, the asymmetric unit of cocrystal (I), namely pyrimidin-2-amine-6-propyl-2-thiouracil (1/2), contains two 6-propyl-2-thiouracil molecules and one pyrimidin-2-amine molecule (Fig. 1). The plane of the pyrimidin-2-amine molecule forms dihedral angles of 17.1 (1) and 10.6 (1)° with those of the thiouracil rings of molecules A and B, respectively. Different propyl side-chain conformations are observed: in molecule A, methyl C atom C9A and thiouracil ring atom C6A are synclinal, with the C8A-C9A bond almost perpendicular to the plane of the ring, while in molecule B they are antiperiplanar, with a dihedral angle of 32.7 (2)° between the plane of the thiouracil ring and the plane through the side chain (Table 1). Each 6-propyl-2-thiouracil molecule is hydrogen bonded to the pyrimidin-2-amine molecule by an R22(8) motif (Bernstein et al., 1995) characterized by one N-HS and one N-HN hydrogen bond. The O atoms participate in N-HO interactions (Table 2) connecting the 6-propyl-2-thiouracil molecules into C(6) chains running along the b axis. The packing of (I) shows layers parallel to (101) containing circular arrangements of four adjacent trimeric units with an R88(34) hydrogen-bond pattern (Fig. 2).
Its participation in the `base pairing' of (I) suggests that the S atom competes as an acceptor with the O atom. Hence, we were also interested in whether both S and O atoms can be hydrogen bonded simultaneously to a complementary molecule. Since 6-propyl-2-thiouracil possesses an acceptor-donor-acceptor site, we cocrystallized it with N-(6-acetamidopyridin-2-yl)acetamide, which exhibits a donor-acceptor-donor site.
Cocrystal (II), namely N-(6-acetamidopyridin-2-yl)acetamide-6-propyl-2-thiouracil (1/2), contains three symmetry-independent complexes, each consisting of two 6-propyl-2-thiouracil molecules and one N-(6-acetamidopyridin-2-yl)acetamide molecule (Fig. 3). The molecular structures of the six propylthiouracil and the three N-(6-acetamidopyridin-2-yl)acetamide molecules are very similar. The r.m.s. deviation from the mean plane through the non-H atoms of each 6-propyl-2-thiouracil molecule varies from 0.012 to 0.028 Å, confirming their planarity. All side chains show an extended conformation, with C8 antiperiplanar to N1 and C9 antiperiplanar to C6 (Table 3). Both N-H bonds of the N-(6-acetamidopyridin-2-yl)acetamide molecules are directed to the same side of the side chains as the pyridine N atom, while the methyl groups are antiperiplanar to ring atoms C2 and C6 (Table 4). Thus, dihedral angles ranging from 12.6 (1) to 15.6 (1)° are formed between the planes through one of the amide groups and the pyridine ring (Table 5). The hydrogen-bond patterns within the three complexes are also identical. The 6-propyl-2-thiouracil molecules are linked into dimers by an R22(8) motif involving two N-HS hydrogen bonds. In addition, an N-HO interaction connects one 6-propyl-2-thiouracil molecule to an N-(6-acetamidopyridin-2-yl)acetamide molecule. However, the geometric arrangements of the complexes show some flexibility. The planes through the two 6-propyl-2-thiouracil molecules of a dimer enclose a dihedral angle ranging from 4.4 (1) to 11.9 (1)°, while dihedral angles ranging from 2.0 (1) to 9.7 (1)° are observed between the planes through the pyridine ring and the neighbouring 6-propyl-2-thiouracil molecule (Table 6). In the packing, all three complexes are twisted by 17° with respect to each other and are connected by N-HO hydrogen bonds into chains running along  (Fig. 4). Furthermore, a second chain is formed consisting of N-HO-bonded symmetry-related complexes aligned along the b axis. Altogether, an extended three-dimensional network of hydrogen bonds is observed (Table 7).
In spite of the appropriate arrangement of donor and acceptor groups, 6-propyl-2-thiouracil does not form three hydrogen bonds to N-(6-acetamidopyridin-2-yl)acetamide in (II), but undergoes homodimerization without participation of the carbonyl O atom. In order to further investigate its preferred hydrogen-bonding interactions, we analysed related crystal structures. Two structures containing 6-propyl-2-thiouracil are present in the Cambridge Structural Database (CSD, Version 5.32 of November 2010, plus two updates; Allen, 2002), namely a 1,4-dioxane solvate (refcode BUWYOH; Okabe et al., 1983) and a charge-transfer complex with diiodine (refcode HAFLAC; Antoniadis et al., 2003). The latter structure is not further considered, since the S atom is connected to the diiodine molecule and hence can hardly participate as a hydrogen-bond acceptor. In the 1,4-dioxane solvate, only the carbonyl O atom takes part in the hydrogen bonding and connects the 6-propyl-2-thiouracil molecules into chains, while no N-HS interactions are observed. In the solvent-free structure of the selenium analogue of 6-propyl-2-thiouracil (refcode PELHEU; Antoniadis et al., 2006), the molecules are hydrogen bonded into chains by R22(8) interactions involving either N-HSe or N-HO hydrogen bonds. We therefore undertook crystallization experiments with 6-propyl-2-thiouracil alone to study whether similar interactions can be observed.
The crystal structure of 6-propyl-2-thiouracil, (III), is isostructural with PELHEU (Fig. 5). The thiouracil rings of the two independent molecules are planar [r.m.s. deviations = 0.006 (A) and 0.016 Å (B) for all non-H atoms] and the propyl side chains are again extended but slightly twisted, with the planes through the ring and the side chain enclosing dihedral angles of 26.0 (2)° in A and 29.8 (2)° in B (Table 8). The 6-propyl-2-thiouracil molecules are connected into vaulted chains running along the b axis by two kinds of hydrogen-bond interactions (Table 9). Although both show the same R22(8) graph set, the hydrogen-bond pattern consists of either two N-HO or two N-HS interactions (Fig. 6). In the crystal packing, two adjacent chains form a tubular arrangement stabilized by van der Waals interactions (Fig. 7).
In (I)-(III), 6-propyl-2-thiouracil exhibits different side-chain conformations. The dihedral angle between the planar thiouracil ring and the plane through the side chain varies from 2.3 (1) to 89.4 (1)°, although an extended arrangement is preferred. Since sufficient donor groups are available both O and S atoms participate in the hydrogen bonding. The N-HO hydrogen bonds have different functions: they connect hydrogen-bonded 6-propyl-2-thiouracil molecules either with themselves [in (I)] or with the other cocrystal component [in (II)], thus forming chains, or they stabilize homodimers of 6-propyl-2-thiouracil with an R22(8) pattern [in (III)]. In contrast, the S atoms are only involved in R22(8) hydrogen-bond formation linking the 6-propyl-2-thiouracil molecules into a heterodimer [in (I)] or into a homodimer [in (II) and (III)].
From the hydrogen-bond interactions in the three structures [(I)-(III)], it is not evident whether an N-HO or an N-HS hydrogen bond is stronger. A CSD search of six-membered ring compounds with hydrogen-bonding sites similar to 2-thiouracil yielded four different types of R22(8) patterns. 19 entries showed R22(8) motifs characterized by two N-HO hydrogen bonds; the S atoms take part as acceptors only in five of them [refcodes LACJIJ (Tashkhodzhaev et al., 2002), XUHJIY (Pawlowski et al., 2009), XEXWAZ, XEXWED and XEXWIH (Balalaie et al., 2006)], whereby chains stabilized by N-HS hydrogen bonds are observed only in XUHJIY. Nine structures contain two different R22(8) patterns with either two N-HO or two N-HS hydrogen bonds [refcodes CASPUI (Hu et al., 2005), CUKBOA (Hori et al., 2009), GEMCAC (Read et al., 1988), PABPAL (Chierotti et al., 2010), RAPNAY (Long et al., 2005), TURCIL01 (Tiekink, 1989), TURCIL02 (Munshi & Guru Row, 2006), WIVJAM (Coxall et al., 2000) and ZEWDOU (Ferrari et al., 1995)]. Furthermore, six entries showed R22(8) interactions consisting of two N-HS hydrogen bonds [refcodes FALWOF (Orzeszko et al., 2004), JESWEK (Xue et al., 2006), MTURAC (Hawkinson, 1975), PABNIR (Chierotti et al., 2010), VOKBUT (Luo et al., 2008) and ZUWMUZ (Branch et al., 1996)]. In three of these structures, the O atoms do not participate in hydrogen bonds. Finally, R22(8) motifs with one N-HO and one N-HS hydrogen bond are only observed in EAZTHY (Voutsas et al., 1978).
The CSD study might suggest that an N-HO is more stable than an N-HS interaction, but some structures revealed hydrogen-bond interactions only with C=S as a supposedly weaker acceptor group. Although the R22(8) motif with two N-HO hydrogen bonds is more abundant in the CSD, it is not formed in two of our three structures. A closer examination of the hydrogen-bonding interactions between 6-propyl-2-thiouracil and pyrimidin-2-amine in (I) revealed unusually large NS distances [N21S2A = 3.6234 (18) Å and N21S2B = 3.5145 (17) Å]. Presumably the complex is further stabilized by a weak C-HO interaction, which leads to a slightly twisted arrangement of the molecules. The hydrogen-bond pattern with the O atom as an acceptor appears to be essential for the packing in (I). If the N-HO hydrogen bond was instead present in the R22(8) motif (an interaction similar to the 2-thiouracil-adenine Watson-Crick base pair), the heterodimer between 6-propyl-2-thiouracil and pyrimidin-2-amine would be further stabilized by a C-HS instead of a C-HO interaction and the 6-propyl-2-thiouracil chains linked by N-HS instead of N-HO hydrogen bonds. The C-HS interaction and chains connected by N-HS hydrogen bonds seem to be less stable, since they are rarely observed in crystal structures (Domagala et al., 2003; Pawlowski et al., 2009).
The hydrogen-bond interactions in (II) can be rationalized by similar arguments. If the homodimer of 6-propyl-2-thiouracil was linked by an R22(8) motif with two N-HO hydrogen bonds, the N-HS hydrogen bonds would connect the 6-propyl-2-thiouracil and N-(6-acetamidopyridin-2-yl)acetamide molecules into chains. The desired heterodimer with three hydrogen bonds is not observed. This is probably due to the fact that the intramolecular distances between the hydrogen donor and acceptor groups of 6-propyl-2-thiouracil do not match with those of N-(6-acetamidopyridin-2-yl)acetamide (pyrimidine-thio NS ca 2.7 Å and pyridine-amide NN ca 2.3 Å). Therefore, formation of the desired complex may result in a strained arrangement; no such cocrystal has yet been reported in the CSD. The hydrogen-bonding interactions in (III) are similar to those of its selenium analogue and to the nine entries from the CSD study (see above). In none of the three structures, (I)-(III), is an R22(8) motif with one N-HO and one N-HS hydrogen bond observed.
Obviously, the relative strength of the N-HO and N-HS hydrogen bonds cannot be clearly judged, since there are many factors affecting hydrogen-bond formation in the crystal. All donor groups will strive to form hydrogen bonds with available acceptor groups within a favourable crystal packing. This complex situation might explain why previous theoretical and experimental studies revealed different relative stabilities for the N-HO and N-HS hydrogen bonds. As a result of our investigation, C=O and C=S are indeed competitive acceptor groups.
The authors thank Dr Michael Bolte for helpful discussions.
Allen, F. H. (2002). Acta Cryst. B58, 380-388.
Antoniadis, C. D., Blake, A. J., Hadjikakou, S. K., Hadjiliadis, N., Hubberstey, P., Schröder, M. & Wilson, C. (2006). Acta Cryst. B62, 580-591.
Antoniadis, C. D., Corban, G. J., Hadjikakou, S. K., Hadjiliadis, N., Kubicki, M., Warner, S. & Butler, I. S. (2003). Eur. J. Inorg. Chem. pp. 1635-1640.
Bahn, R. S., Burch, H. S., Cooper, D. S., Garber, J. R., Greenlee, C. M., Klein, I. L., Laurberg, P., McDougall, I. R., Rivkees, S. A., Ross, D., Sosa, J. A. & Stan, M. N. (2009). Thyroid, 19, 673-674.
Balalaie, S., Bararjanian, M. & Rominger, F. (2006). J. Heterocycl. Chem. 43, 821-826.
Basilio Janke, E. M., Dunger, A., Limbach, H.-H. & Weisz, K. (2001). Magn. Reson. Chem. 39, 177-182.
Bernstein, J., Davis, R. E., Shimoni, L. & Chang, N.-L. (1995). Angew. Chem. Int. Ed. Engl. 34, 1555-1573.
Biswal, H. S. & Wategaonkar, S. (2009). J. Phys. Chem. A, 113, 12763-12773.
Blessing, R. H. (1995). Acta Cryst. A51, 33-38.
Branch, C. L., Eggleston, D. S., Haltiwanger, R. C., Kaura, A. C. & Tyler, J. W. (1996). Synth. Commun. 26, 2075-2084.
Chierotti, M. R., Ferrero, L., Garino, N., Gobetto, R., Pellegrino, L., Braga, D., Grepioni, F. & Maini, L. (2010). Chem. Eur. J. 16, 4347-4358.
Cooper, D. S. (2005). N. Engl. J. Med. 352, 905-917.
Coxall, R. A., Harris, S. G., Henderson, D. K., Parsons, S., Tasker, P. A. & Winpenny, R. E. P. (2000). J. Chem. Soc. Dalton Trans. pp. 2349-2356.
Domagala, M., Grabowski, S. J., Urbaniak, K. & Mlostón, G. (2003). J. Phys. Chem. A, 107, 2730-2736.
Favre, A., Saintomé, C., Fourrey, J.-L., Clivio, P. & Laugâa, P. (1998). J. Photochem. Photobiol. B, 42, 109-124.
Ferrari, M. B., Fava, G. G., Pelosi, G., Rodriguez-Argüelles, M. C. & Tarasconi, P. (1995). J. Chem. Soc. Dalton Trans. pp. 3035-3040.
Flack, H. D. (1983). Acta Cryst. A39, 876-881.
Hawkinson, S. W. (1975). Acta Cryst. B31, 2153-2156.
Hori, A., Ishida, Y., Kikuchi, T., Miyamoto, K. & Sakaguchi, H. (2009). Acta Cryst. C65, o593-o597.
Hu, S.-L., Yin, G.-D. & Wu, A.-X. (2005). Acta Cryst. E61, o2408-o2409.
Lautié, A. & Novak, A. (1980). Chem. Phys. Lett. 71, 290-293.
Le Page, Y. (1987). J. Appl. Cryst. 20, 264-269.
Le Page, Y. (1988). J. Appl. Cryst. 21, 983-984.
Long, T., Zhou, H.-B. & Wu, A.-X. (2005). Acta Cryst. E61, o2169-o2171.
Luo, W., Yu, Q.-S., Tweedie, D., Deschamps, J., Parrish, D., Holloway, H. W., Li, Y., Brossi, A. & Greig, N. H. (2008). Synthesis, 21, 3415-3422.
Macrae, C. F., Bruno, I. J., Chisholm, J. A., Edgington, P. R., McCabe, P., Pidcock, E., Rodriguez-Monge, L., Taylor, R., van de Streek, J. & Wood, P. A. (2008). J. Appl. Cryst. 41, 466-470.
Munshi, P. & Guru Row, T. N. (2006). Acta Cryst. B62, 612-626.
Okabe, N., Fujiwara, T., Yamagata, Y. & Tomita, K. (1983). Bull. Chem. Soc. Jpn, 56, 1543-1544.
Orzeszko, B., Kazimierczuk, Z., Maurin, J. K., Laudy, A. E., Starosciak, B. J., Vilpo, J., Vilpo, L., Balzarini, J. & Orzeszko, A. (2004). Il Farmaco, 59, 929-937.
Pawlowski, M., Lendzion, A., Szawkalo, J., Leniewski, A., Maurin, J. K. & Czarnocki, Z. (2009). Phosphorus Sulfur Silicon Relat. Elem. 184, 1307-1313.
Read, G., Randal, R., Hursthouse, M. B. & Short, R. (1988). J. Chem. Soc. Perkin Trans. 2, pp. 1103-1105.
Sheldrick, G. M. (2008). Acta Cryst. A64, 112-122.
Spek, A. L. (2009). Acta Cryst. D65, 148-155.
Sponer, J., Leszczynski, J. & Hobza, P. (1997). J. Phys. Chem. A, 101, 9489-9495.
Stoe & Cie (2001). X-AREA. Stoe & Cie, Darmstadt, Germany.
Tashkhodzhaev, B., Turgunov, K. K., Usmanova, B., Averkiev, B. B., Antipin, M. Yu. & Shakhidoyatov, Kh. M. (2002). Zh. Strukt. Khim. 43, 944-948.
Testa, S. M., Disney, M. D., Turner, D. H. & Kierzek, R. (1999). Biochemistry, 38, 16655-16662.
Tiekink, E. R. T. (1989). Z. Kristallogr. 187, 79-84.
Voutsas, G. P., Venetopoulos, C. C., Kálmán, A., Párkányi, L., Hornyák, G. & Lempert, K. (1978). Tetrahedron Lett. 19, 4431-4434.
Westrip, S. P. (2010). J. Appl. Cryst. 43, 920-925.
Xue, S.-J., Wang, Q.-D. & Li, J.-Z. (2006). Acta Cryst. C62, o666-o668.