Structural studies of 4-aryloctahydro-pyrido[1,2-c]pyrimidine derivatives

Journal of Molecular Structure 605 (2002) 85±92

www.elsevier.com/locate/molstruc

Structural studies of 4-aryloctahydro-pyrido[1,2-c]pyrimidine derivatives Maciej Pisklak a, Franciszek Herold b, Romana Anulewicz-Ostrowska c, Iwona Wawer a,* a

Department of Physical Chemistry, Faculty of Pharmacy, The Medical University of Warsaw, Banacha 1, 02-097 Warsaw, Poland b Department of Drug Technology, Faculty of Pharmacy, The Medical University of Warsaw, Banacha 1, 02-097 Warsaw, Poland c Department of Chemistry, Warsaw University, Pasteura 1, 02-093 Warsaw, Poland Received 18 April 2001; revised 6 June 2001; accepted 6 June 2001

Abstract 13

C cross-polarisation magic angle spinning NMR data have been reported for four derivatives of 4-aryl-octahydropyrido[1,2-c]pyrimidine-1,3-dione and the X-ray diffraction data for two (with 2 0 -Me and 2 0 -OMe). The crystal structures show the presence of centrosymmetric cyclic dimers with intermolecular C1yO´ ´´H±N or C3yO´ ´ ´H±N hydrogen bonds, the con®guration at the chiral centres (C4 and C4a) was determined as RR (SS). The twisting of aromatic ring at C4 with respect to the pyrido[1,2-c]pyrimidine skeleton is about 68±1098. q 2002 Elsevier Science B.V. All rights reserved. Keywords: CPMAS NMR; X-ray diffraction; 4-Aryl-octahydro-pyrido[1,2-c]pyrimidine-1,3-diones; Cyclic dimers

1. Introduction As a part of our work on the development of novel piperazines with CNS-activity, we have prepared structural analogues of buspirone with different types of imide group. The models [1,2] suggest that steric hindrance and repulsive forces between the receptor and the imide group of the buspirone analogues are the most important determinants of ligand-binding af®nity. Therefore, a more detailed knowledge about the three-dimensional structure of novel imide-type compounds would be useful. The substrates, used to replace the imide part of the ligand, were pyrido[1,2-c]pyrimidine-1,3-dione derivatives. The structure and intermolecular interactions of 4aryl-1H,3H-hexahydro-pyrido[1,2-c]pyrimidine-1,3diones were reported recently [3] and hydrogen * Corresponding author. Tel.: 148-228236297. E-mail address: [email protected] (I. Wawer).

bonded dimers were found in the crystals. In the present study we examined the structure and interactions of 4-aryl-octahydro-pyrido[1,2-c]pyrimidine1,3-diones with the ortho and para substituted aromatic ring at C4 (Fig. 1), using a combined approach of solid state NMR, X-ray diffraction and theoretical MO methods. 2. Experimental The 4-aryl-octahydro-pyrido[1,2-c]pyrimidine-1,3dione derivatives were prepared, starting from the respective nitriles, followed by intra-molecular cyclisation of the uretanes; the synthetic pathway is described in [4,5]. The 1H and 13C NMR spectra were recorded at 500.13 and 125.76 MHz, respectively, on a Bruker DRX-500 spectrometer for CDCl3 solutions. The 2D correlations: COSY, HETCOR, HMBC and ROESY

0022-2860/02/$ - see front matter q 2002 Elsevier Science B.V. All rights reserved. PII: S 0022-286 0(01)00754-2

86

M. Pisklak et al. / Journal of Molecular Structure 605 (2002) 85±92

R2 4' 5'

3'

6'

2'

R1

1' 5 6

O

4

4a

1 2 3 4

R1 R2 CH3 H OCH3 H F H H CH3

3

N

7 8

N 2

1

H

O Fig. 1. The compounds studied with carbon numbering.

were measured using standard pulse sequences and parameters from Bruker library. Cross polarisation magic angle spinning solid-state 13C NMR spectra were recorded at 75.5 MHz on a Bruker MSL-300 instrument. Powder samples were spun at 10.0± 10.5 kHz in 4 mm ZrO2 rotor, contact time of 4± 5 ms, repetition time of 6 s and spectral width of 20 kHz were used for accumulation of 700±1200 scans. Dipolar dephasing pulse sequence (with 50 ms delay time inserted before acquisition) was used to selectively observe the nonprotonated carbons. Chemical shifts were calibrated indirectly through the glycine CO signal recorded at 176.0 ppm, relative to TMS. The single crystals of 1 and 2, suitable for X-ray analysis, were grown from ethanol, by slow evaporation. The data for crystal 1 were collected on a KM4 KUMA-diffractometer, with graphite monochromated CuKa radiation. The u±2u scan technique and a variable scan speed ranging from 1.2 to 18.08/min, depending on re¯ection intensity, were applied. All measurements of crystal 2 were performed on a Kuma KM4CCD k -axis diffractometer with graphitemonochromated Mo Ka radiation. The crystal was positioned at 65 mm from the KM4CCD camera, 256 frames were measured at 1.88 intervals with a counting time of 15 s. Intensity data were corrected for Lorenz and polarisation effects. The structures were solved by direct methods [6] and re®ned using shelxl [7]. The re®nement was based on F 2 for all re¯ections except those, with highly negative F 2. Weighted R factors wR and all goodness-of-®t S values are based on F 2. Conventional R factors are

based on F with F set to zero for negative F 2. The F02 . 2s…F02 † criterion was used only for calculating R factors and is not relevant to the choice of re¯ections for the re®nement. The R factors based on F 2, are about twice as large as those based on F. All hydrogen atoms were located from a differential map and re®ned isotropically. Scattering factors were taken from Tables 6.1.1.4 and 4.2.4.2 of Ref. [8]. Crystal data together with the data collection and structure re®nement details are listed in Table 1. All positional, geometric and thermal parameters are deposited as supplementary material at the Cambridge Crystallographic Data Centre (CCDC Identi®cation Number 154877 (2 0 -Me) and 154876 (2 0 -OMe)). For the calculation of NMR shielding constants the GIAO-CHF approach was used (from gaussian 98 package [9]). The calculations were performed on SG Origin200 workstation. Semi-empirical calculations were performed with PM3 method implemented in hyperchem 5.01 package [10].

3. Results and discussion 3.1. Solid state NMR The 13C CP MAS NMR spectra were recorded for solid compounds 1±4, and the spectrum of 2 (2 0 OMe) is illustrated in Fig. 2. The majority of the 13 C resonances could be assigned directly by comparison with the solution data, and the chemical shifts have been listed in Table 2. The signals of carbons linked to nitrogen atom (C8, C4a, C1yO, C3yO) are broader and/or split into asymmetric doublets, due to the residual dipolar 13C± 14N coupling. The signal of C1yO carbon, which is bound to two nitrogens, is broader than that of C3yO, and this made it possible to distinguish between their resonances (Fig. 2). The number of resonances does not exceed the number of carbons in the spectra of 1±4, indicating that there is no polymorphism in the solid phase. In the solution, molecules can freely undergo intra-molecular rotation; in solids, conformational equilibration is usually inhibited, revealing speci®c environments with different chemical shifts. Thus, the differences: D ˆ dsolution 2 dsolid . 1 ppm; provide new information

M. Pisklak et al. / Journal of Molecular Structure 605 (2002) 85±92

87

Table 1 Crystal data and structure re®nement for 1 and 2

Molecular formula Molecular weight Crystal system Space group Z Ê) A (A Ê) B (A Ê) C (A B (8) Ê 3) Volume (A Density(calcd) (g cm 23) F(000) (e) Ê) Wavelength (A Absorption coef®cient (mm 21) Crystal size (mm 3) Index range Re¯ections collected/unique No. of parameters Final R indices ‰I . 2s…I†Š R indices (all data) Ê 23) Max./min. Dr (e A

1

2

C15H18N2O2 258.31 Monoclinic P2…1†=n 4 7.2480 (10) 27.506 (6) 7.2750 (10) 110.63 (3) 1357.4 (4) 1.264 552 1.54178 0.682 0:35 £ 0:30 £ 0:30 28 # h # 5 0 # k # 27 24 # l # 8 1861/1763 ‰R…int† ˆ 0:0635Š 194 R1 ˆ 0:0928, wR2 ˆ 0:2258 R1 ˆ 0:1308, wR2 ˆ 0:2781 0.324/20.517

C15H18N2O3 274.31 Monoclinic P2…1†=n 4 7.519 (2) 17.652 (4) 10.695 (2) 103.92 (3) 1377.8 (5) 1.322 584 0.71073 0.093 0:4 £ 0:3 £ 0:2 25 # h # 10 223 # k # 23 214 # l # 13 6317/2157 ‰R…int† ˆ 0:0749Š 200 R1 ˆ 0:0771, wR2 ˆ 0:1955 R1 ˆ 0:1023, wR2 ˆ 0:2220 0.264/20.239

on the dynamics and molecular interactions in the two phases. The 4-aryl-hexahydro-1H,3H-pyrido[1,2-c]pyrimidine-1,3-diones, studied previously [3], formed hydrogen bonded dimers. Similar dimers can also be expected in solid 4-aryl-octahydro-pyrido[1,2-c]pyrimidine-1,3-dione derivatives, containing CyO and NH functional groups. In solution, these groups are

involved in a weak interaction with solvent molecules (CDCl3). Remarkable deshielding of carbonyl carbons C1yO or C3yO can be indicative of the formation of intermolecular CyO´ ´ ´H±N hydrogen bonds, in the solid phase. The inspection of the differences D suggests that the dimers should be connected by the C3yO´ ´ ´H±N bonds …D ˆ 24:7 ppm† in the crystals of 1 and by the C1yO´ ´ ´H±N bonds in 3 and 4

Fig. 2. 13C CP MAS spectrum of solid 2.

88

M. Pisklak et al. / Journal of Molecular Structure 605 (2002) 85±92

Table 2 13 C NMR chemical shifts (d , ppm) and calculated shielding constants (s , ppm) for solid 4-aryl-octahydro-pyrido[1,2-c]pyrimidine-1,3-diones (the values for CDCl3 solutions are given in parentheses) Carbon atom

C1 C3 C4 C4a C5 C6 C7 C8 C1 0 C2 0 C3 0 C4 0 C5 0 C6 0 R a

1 (2 0 -Me)

2 (2 0 -OMe)

3 (2 0 -F)

4 (4 0 -Me)

d (ppm)

s (ppm)

d (ppm)

s (ppm)

d (ppm)

d (ppm)

154.44 (152.70) 174.31 (169.61) 48.65 (50.18) 57.97 (58.19) 30.06 (31.88) 23.58 (23.66) 24.08 (24.31) 43.31 (44.44) 134.85 (134.10) 138.60 (136.82) 129.95 (131.09) 129.50 (128.03) 124.14 (126.88) 128.88 (128.31) 20.22 (20.18)

35.78 18.81 149.95 153.74 172.32 181.68 177.98 163.58 68.49 64.82 69.61 71.14 75.56 65.26 177.79

154.11 (153.49) 169.10 (169.80) 53.94 (50.22) 54.62 (56.13) 33.45 (32.03) 23.96 (23.48) 23.96 (24.16) 43.90 (44.04) 121.81 (123.98) 156.72 (157.03) 113.16 (111.55) 131.99 (129.45) 119.79 (120.90) 133.96 (131.24) 55.84 (55.60)

35.29 23.08 150.85 148.84 170.08 178.27 176.42 157.75 75.61 45.70 88.10 71.08 81.11 63.66 152.14

156.50 (152.89) 169.70 (168.43) 51.10 (48.95) 55.66 (56.40) 31.76 (31.78) 23.50 (23.29) 23.50 (24.20) 42.93 (44.15) 122.01 (122.21) 161.30 (161.77/159.80) a 115.34 (116.04) 132.30 (130.14) 126.8 (124.63) 134.82 (131.13)

156.46 (152.70) 170.23 (169.82) 53.51 (53.40) 58.69 (58.36) 31.92 (32.02) 23.34 (23.73) 23.33 (24.32) 43.82 (44.61) 132.74 (132.46) 128.14 (128.58) 133.66 (129.72) 136.81 (137.93) 129.49 (129.72) 128.14 (128.58) 20.74 (21.11)

Coupling constant 1 JC±F ˆ 247 Hz:

(D ˆ 23:6 and 23.8 ppm, respectively). No clear prediction, on the basis of chemical shifts, could be made for 2, since the difference D ˆ 20:6 ppm for C1yO, is too small (for C3yOD ˆ 10:6 ppm). Solid-state NMR spectra were measured prior to X-ray diffraction studies, and therefore, we decided to determine the crystal structures of 1, 2 and 3 (see further). The locked conformation of the phenyl ring at C4 (which is twisted with respect to the pyrido-[1,2c]pyrimidine system) results in non-equivalence of all aromatic carbons. Separate signals of C2 0 and C6 0 as well as of C3 0 and C5 0 also appear for 4 with a symmetric substituent (4 0 -Me). Dipolar dephasing pulse sequence (eliminating the signals of protonated carbons), was helpful in the detection of quaternary carbons. However, the proposed assignments for carbons bearing hydrogen (C2 0 /C6 0 or C3 0 /C5 0 pairs) could be reversed for 4, and with any other sequence, the values of D are unreasonably larger. The differences between the solution and solid-state chemical shifts for aromatic carbons range from 23.9 to 13.4 ppm. 3.2. X-ray diffraction studies X-Ray crystal structures were determined for 1, 2

and 3 [5]. Crystal data and structure re®nement parameters for 1and 2 are listed in Table 1. Table 3 lists selected bond lengths, angles and distances. In spite of the presence of two chiral atoms C4 and C4a, absolute chirality cannot be de®ned, because, 1 and 2 belong to the space group P2…1†n and 3 belongs to P2…1†c. The study established 1, 2 and 3 as the RR/SS pair. All three compounds form centrosymmetric cyclic dimers linked by two C1(or C3)yO´ ´´H±N hydrogen bonds. The pyrimidine ring adopts a ¯attened conformation, due to the presence of two sp 2 hybridised carbonyl carbons. The fragment O1±C1±N2±C3 is almost planar (torsional angles are 173.5, 179.8 and 2175.6 for 1, 2 and 3, respectively, whereas torsional angles which include the C4a and C4 atoms are larger (Table 3)). The hydrogen bond interaction C1yO´ ´ ´H±N, within the dimer, is only intermolecular in 3. In the crystals of 1, the dimers are created by the C3yO´ ´ ´H±N interaction (Fig. 3), and additionally, C1yO (being outside the cycle) is involved in an intermolecular contact with hydrogen atom H4a (the H4a´ ´´O distance is 0.241 nm). Carbonyl group C1yO forms hydrogen bonds within cyclic dimer of 2, whereas, the second carbonyl group interacts with C4 0 ±H of the aromatic ring of the parent molecule

M. Pisklak et al. / Journal of Molecular Structure 605 (2002) 85±92 Table 3 Selected XRD data for crystals of 4-aryl-octahydro-pyrido[1,2-c]pyrimidine-1,3-diones 1 (2 0 -Me) Ê) Bond lengths (A C1yO C3yO N9±C1 C1±N2 N2±C3 C1 0 ±C4

1.220 (5) 1.240 (5) 1.340 (5) 1.400 (5) 1.350 (5) 1.544 (5)

2 (2 0 -OMe) 1.219 (4) 1.205 (4) 1.339 (5) 1.387 (4) 1.362 (4) 1.513 (5)

3 (2 0 -F), Ref. [5] 1.210 (4) 1.190 (4) 1.323 (4) 1.357 (4) 1.357 (4) 1.482 (4)

Torsion angles (deg) C3±C4±C1 0 ±C2 0 C4a ±C4±C1 0 ±C2 0 N2±C1±N9±C4a C1±N2±C3±C4 N2±C3±C4±C4a N9±C1±N2±C3 O1±C1±N2±C3 C3±C4±C4a ±N9

2 127.3 108.8 2 13.5 2 5.7 35.5 2 7.5 173.5 2 51.8

2 64.3 68.1 2 10.2 2 11.5 28.7 1.4 179.8 2 35.7

56.4 2 71.0 14.5 3.6 2 27.8 4.3 2 175.6 43.1

Hydrogen bonding Ê) Distance N2´´´O11 (A H´´´O11

C3yO´´ ´H±N 2.936 2.101

C1y0´´ ´N±H 2.885 2.037

C1yO´´´H±N 2.812 1.962

Fig. 3. Hydrogen bonded dimer of 1 with 50% probability thermal ellipsoids.

89

90

M. Pisklak et al. / Journal of Molecular Structure 605 (2002) 85±92

Fig. 4. Hydrogen bonding pattern in 2.

(the H4 0 ´ ´´O distance is 0.2513 nm). The crystal packing scheme for 2 is illustrated in Fig. 4. In all the three structures H4 and H4a are axial with trans orientation, and, consequently, the phenyl ring is always in an equatorial position. The twist angle of the aromatic ring, with respect to the pyrido-[1,2-c]pyrimidine system is 68±1088. The steric hindrance, which is higher in 1 and 2 with voluminous substituents at C2 0 (Me, OMe), can de diminished not only by twisting the ring but also by elongating the C4±C1 0 bond. This bond is shorter (0.148 nm) in 3 with a small 2 0 -F substituent and longer (0.151, 0.154 nm) in 1 and 2. 3.3. Solution conformation and molecular modelling Structural characterisation of 1±4 in solution is of much interest, since stereochemistry and conformational equilibrium could be important in binding

with the receptor. The assignment of 1H and 13C chemical shifts of 1±4 in solution is given in [5]. The combined use of 2D spectra, especially COSY and ROESY helped in the conformational analysis. At the ®rst sight, the 1H NMR results suggest that the saturated hydrocarbon fragment is relatively rigid, without fast intra-molecular dynamics (i.e. inversion of the saturated ring by pseudo-rotation around C±C bonds), since separate 1H resonances are observed for axial and equatorial methylene protons at C5, C6, C7 and C8 (Table 4). The separation of signals is: 0.34 and 0.24 ppm for H5a,e and H7a,e, respectively, i.e. less than 0.5 ppm typically found for cyclohexane or cyclohexene derivatives. This might indicate that conformational averaging still takes place. The most striking feature is a very big difference (of 1.7 ppm) in the chemical shift of hydrogens H8a and H8e, the probable reason for which is the proximity of carbonyl C1yO group.

M. Pisklak et al. / Journal of Molecular Structure 605 (2002) 85±92

91

Table 4 1 H NMR chemical shifts d (ppm) and shielding constants s (ppm) calculated with GIAO-CHF method for methylene groups of 1 and 2 (D 0 ˆ d ax 2 deq ; D 00 ˆ s ax 2 s eq ) 1

H5 H6 H7 H8

Axial Equatorial Axial Equatorial Axial Equatorial Axial Equatorial

2 0

d (ppm)

D (ppm)

s (ppm)

D (ppm)

d (ppm)

D 0 (ppm)

s (ppm)

D 00 (ppm)

1.28 1.67 1.42 1.81 1.52 1.74 2.73 4.43

2 0.39

29.99 29.45 30.02 29.83 29.74 29.87 28.50 26.91

0.54

1.20 1.64 1.35 1.72 1.53 1.82 2.68 4.41

2 0.44

30.76 30.31 30.83 30.37 30.25 30.32 29.37 27.67

0.45

2 0.39 2 0.22 2 1.7

In the ROESY spectra, the interactions were observed for the geminal axial and equatorial hydrogens and also for the axial and equatorial neighbours. The most interesting were the cross-peaks between each of the axially oriented hydrogens, H8a and H4a, H6a and also between H7a and H4a, H5a, which made the identi®cation of upward and downward side of the bicyclic system, possible. The observed contacts are in agreement with those expected from the X-ray diffraction data. This indicates that, there is no signi®cant change between the solution and solid-state conformation of compounds 1±4. Weak H4/H4a cross-peaks in the ROESY spectrum of 4 (4 0 -Me) con®rm that, these protons are on the opposite side of the bicyclic system, similar to that in 1±3. The coupling constants 3J(H4/H4a) are 9.5±10.5 Hz, in agreement with the trans arrangement of these two hydrogens, found in the crystal. Optimisation of the geometry for 1 and 2 was performed by means of the PM3 MO method. Molecular modelling clearly showed that there is no suf®cient space for free rotation of the aryl substituent, around C4±C1 0 bond, since H2 0 and H6 0 are too close to H5a, H5e, H4a and C3yO. The steric hindrance is even higher when H2 0 is replaced by a voluminous substituent (2 0 -Me or 2 0 -OMe). Then, the twisting of the phenyl ring helps to release a part of steric energy. The NMR shielding constants were calculated for 1 and 2 (using the PM3 geometry for one isolated molecule). The linear correlations were obtained between experimental 13C CPMAS chemical shifts and the calculated carbon shieldings (for compound 1 s ˆ 21:0518d 1 205:3; R ˆ 0:997 and for compound 2,

00

0.19 0.13 1.59

2 0.37 2 0.29 2 1.73

0.46 0.07 1.7

s ˆ 21:0564d 1 205:8; R ˆ 20:998). Larger discrepancies were noticed for carbonyl carbons, because, in the solid phase, C1yO or C3yO are involved in hydrogen bonding. The comparison of 1H chemical shifts and calculated shielding constants showed (Table 4) that, the big difference between H8a and H8e is reproduced by MO calculations. Shielding constants for methylene hydrogens H5±H7 are within the range 29.45±30.02 ppm. The signi®cantly smaller s value for H8e (26.91 ppm) can be explained by its localisation: it is coplanar with the carbonyl group and thus in the deshielding zone, produced by the CyO bond. Also, a small difference in the chemical shift between H7a and H7e is a result of electron density distribution and not the effect of conformational averaging at C7 (pseudorotation around the C6±C7 bond).

Acknowledgements The X-ray measurements were undertaken in the Crystallographic Unit of the Physical Chemistry Lab. the Chemistry Department of the University of Warsaw. This research was supported by grant FW-28/N/ 2000 from the Medical University of Warsaw.

References [1] I. Sylte, Z. Chilmonczyk, S.V. Dahl, J. Cybulski, O. Edvardsen, J. Pharm. Pharmacol. 49 (1997) 698. [2] Z. Chilmonczyk, A. Szelejewska-Wozniakowska, J. Cybulski,

92

[3] [4] [5] [6] [7] [8] [9]

M. Pisklak et al. / Journal of Molecular Structure 605 (2002) 85±92 A.E. Kozioø, M. Gdaniec, Arch. Pharm. Pharm. Med. Chem. 330 (1997) 146. F. Herold, D. Maciejewska, I. Wolska, J. Phys. Org. Chem. 13 (2000) 213. F. Herold, I. Wolska, E. Helbin, M. KroÂl, J. Kleps, J. Heterocycl. Chem. 36 (1999) 389. F. Herold, J. Kleps, R. Anulewicz-Ostrowska, B. SzczeÎsna, J. Heterocycl. Chem., submitted for publication. G.M. Sheldrick, Acta Crystallogr. A46 (1990) 467. G.M. Sheldrick, shelxl 93. Program for the Re®nement of Crystal Structures, University of GoÈttingen, Germany. A.J.C. Wilson (Ed.), International Tables for Crystallography, vol. C, Kluwer Academic Publishers, Dordrecht, 1992. M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, V.G. Zakrzewski, J.A. Montgomery

Jr., R.E. Stratmann, J.C. Burant, S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A. Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V. Ortiz, A.G. Baboul, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, J.L. Andres, C. Gonzalez, M. Head-Gordon, E.S. Replogle, J.A. Pople, gaussian 98, Revision A.7, Gaussian, Inc., Pittsburgh, PA, 1998. [10] hyperchem 5.02 Package, Hypercube: Waterloo, Ontario, Canada, 1997.