A

ttnmeiw©

K

1

gc[oatraD(thTom

0d

Journal of Steroid Biochemistry & Molecular Biology 103 (2007) 163–169

A full conformational characterization of 13-ethylprogestogens throughtheoretical calculations and nuclear magnetic resonance spectroscopy

Diego Colombo a,∗, Patrizia Ferraboschi a, Laura Legnani b, Paolo Prestileo a, Lucio Toma b

a Dipartimento di Chimica, Biochimica e Biotecnologie per la Medicina, Universita di Milano, Via Saldini 50, 20133 Milano, Italyb Dipartimento di Chimica Organica, Universita di Pavia, Via Taramelli 10, 27100 Pavia, Italy

Received 12 April 2006; accepted 5 September 2006

bstract

The conformational preferences of a group of 13-ethylsteroids and related estranes have been determined through theoretical calculations athe B3LYP/6-31G* level in order to ascertain differences and similarities in the conformational behavior which might, in principle, influencehe activity. Attention was focussed on two geometrical features usually related to the progestational activity of this class of compounds,amely, the inversion of the A ring and the orientation of the 13-ethyl group. The present calculations show that the absence of the C10ethyl group, like in levonorgestrel, 3-ketodesogesterel, and gestodene, makes the inversion of the A ring easier than in norethisterone and

thisterone even if in any case the 1�,2�-half-chair conformation remains preferred. The preference of the ethyl group for the trans orientations maintained by all the 13-ethylprogestogens though in gestodene also the g+ gauche conformation is significantly populated. This preferenceas experimentally supported through a high field NMR spectroscopy study of the ethylsteroids.2006 Elsevier Ltd. All rights reserved.

ma

tsgaCrbceMeo

eywords: Progestins; Molecular modeling; DFT calculations; NMR

. Introduction

During the decades several synthetic steroids with pro-estational activity were discovered and introduced as contra-eptives or in postmenopausal hormonal replacement therapy1,2]. This class of compounds has been the target of vari-us structure–activity relationship studies aimed to correlatectivity to their physicochemical parameters [3–5]. However,o the best of our knowledge, a detailed and systematic explo-ation of the conformational behavior of these compounds athigh level of calculation still lacks; in fact, in papers whereFT optimized structures were considered, only the preferred

or the most probable) conformation was considered andhe easiness for compounds to access geometries of energyigher than the most stable one was not investigated [5].

he complete characterization of the conformational domainf compounds is relevant since the bioactive conformationight not to correspond to the global minimum structure but

∗ Corresponding author. Tel.: +39 0250316039; fax: +39 0250316036.E-mail address: diego.colombo@unimi.it (D. Colombo).

r

llgm

960-0760/$ – see front matter © 2006 Elsevier Ltd. All rights reserved.oi:10.1016/j.jsbmb.2006.09.036

ay be one of those accessible comprises within a range offew kcal/mol above the most stable.

The most active progestogens, e.g. 1–3, do not have thewo methyl groups at C10 and C13, distinctive of most ofteroids. The methyl at C10 is absent, so that such progesto-ens belong to the 19-norsteroid class, whereas the methylt C13 is elongated to an ethyl group. The absence of the10 methyl group might influence the geometry of the A

ing that seems strictly related to activity; in fact, it haseen observed that the A ring inversion from the 1�,2�-half-hair to the 1�,2�-half-chair gives structures which oftenxhibit a very high affinity for the progesterone receptor [6,7].oreover, also the orientation of the ethyl group has been

valuated as a feature relevant for the modulation of activityf 13-ethylsteroids through the synthesis of conformationallyestricted analogues [8].

In this paper we wish to report a modeling study of

evonorgestrel (1), 3-ketodesogestrel (2), the active metabo-ite of desogestrel, and gestodene (3), all bearing an ethylroup at position 13, in comparison with the related 13-ethylprogestogens norethisterone (4), ethisterone (5), and

1 emistry

pfsrs(cNt

2

2

mKtotcehphlbmCIatEo

2

[pptseorolmtcTkrmftmr

2

B1gCrc(a2BC(csN(p1

p1C(i

64 D. Colombo et al. / Journal of Steroid Bioch

rogesterone (6) in order to elucidate the role that the dif-erent structural features exert in the determination of theubtle differences in their geometrical preferences. Theo-etical calculations at the B3LYP/6-31G* level, a currenttandard also for medium sized molecules such as steroidssee, for example, Refs. [5,9,10]), were performed for all theompounds; in addition, for the 13-ethylsteroids high fieldMR experimental data were used to support the results of

he calculations.

. Material and methods

.1. Materials

Levonorgestrel (1) was prepared starting from the com-ercially available 18-methyl-4-estren-3,17-dione (1.0 g,eifeng, Shanghai). 3-Carbonylic group was selectively pro-

ected by reaction with 2,2-dimethylpropanediol, triethylrthoformate and p-toluensulfonic acid [11]. The 7/3 mix-ure of �5 and �5(10) 3-ketals was treated with ethynylerium chloride in tetrahydrofuran [12] to afford the 17�-thynylderivative. Removal of protecting group with 6 Mydrochloric acid (0.4 mL) in acetone (60 mL) at room tem-erature for 1 h afforded, after neutralization with sodiumydrogen carbonate and usual work-up, crude product 1. Pureevonorgestrel (1) (0.69 g, 63% overall yield) was isolatedy crystallization with ethanol. DSC (5 ◦C/min), endother-ic peak at 236 ◦C (onset at 231.87 ◦C) [�]D

25-31.1 (c 1 inHCl3). 3-Ketodesogestrel (2) was a generous gift from Poli

ndustria Chimica (Milan, Italy). Gestodene (3) was preparedccording to a previously reported method [11]. Differen-ial scanning calorimetry (DSC) was performed on a Perkinlmer DSC-7 instrument and optical rotations were measuredn a Perkin Elmer 241 polarimeter.

nmres

& Molecular Biology 103 (2007) 163–169

.2. Theoretical calculations

All calculations were carried out using the Gaussian0313] program package. The conformational space of com-ounds 1–5 was explored through optimization of all theossible starting geometries which were optimized withinhe DFT approach at the B3LYP level with the 6-31G* basiset. All the degrees of conformational freedom were consid-red, in particular the possible existence of different chairr twisted-boat conformations at the A, B, and C hexacyclicings as well as the orientation of the ethyl group at C13 andf the hydroxyl group at C17. Several conformations wereocated for each compound; Tables 1 and 2 summarize theain geometrical data together with the relative energy and

he percentage contribution to the overall population of eachonformation determined through the Boltzmann equation.he geometry of the A–C rings is described through twoind of descriptors, i.e., a significant torsion angle for eaching, τA, τB, τC, and the ring puckering coordinates deter-ined according to Cremer and Pople [14], also calculated

or the D ring. The orientation of the ethyl group is describedhrough the torsional angle τE (C14–C13–C18–C18′). The

ost representative conformations of each compound are alsoepresented as three-dimensional plots in Fig. 1.

.3. NMR spectroscopy

All NMR spectra were recorded at 298 K with aruker AVANCE-500 spectrometer operating at 500.13 and25.76 MHz for 1H and 13C, respectively, using a 5 mm sin-le pulsed field gradient (z-PFG) broadband reverse probe.hemical shifts are reported on the δ (ppm) scale and are

elative to chloroform signals (7.24 for 1H and 77.0 ppm,entral line, for 13C spectra, respectively). Compounds 1–3about 10 mg) were dissolved in CDCl3 (0.5 mL) under N2,nd their assignments were given by a combination of 1D andD COSY, HSQC and NOESY experiments, using standardruker pulse programs. Z-PFGs were used to obtain 1H–1HOSY and HSQC spectra. The pulse widths were 7.15 �s

90◦) and 13.0 �s (90◦) for 1H and 13C, respectively. Typi-ally 32,768 data points were collected for one-dimensionalpectra. Spectral widths were 11.45 ppm (5733 Hz) for 1HMR (digital resolution: 0.17 Hz per point) and 259.84 ppm

32,680 Hz) for 13C NMR (digital resolution: 1.0 Hz peroint). 2D experiments parameters were as follows. ForH–1H correlations: relaxation delay 2.0 s, 1024 × 1024 dataoint matrices (512 experiments to 1024 zero filling in F1,024 in F2), 2 or 16 transients in each experiment forOSY and NOESY, respectively, spectral width 6.0 ppm

3004.8 Hz). The NOESY spectra were generated with a mix-ng time of 1.0 s and acquired in the TPPI mode. There wereot significant differences in the results obtained at different

ixing times (0.5–1.5 s). For 13C–1H correlations (HSQC):

elaxation delay 2.5 s, 1024 × 1024 data point matrices (512xperiments to 1024 zero filling in F1, 1024 in F2), two tran-ients in each experiment, spectral width 6.0 ppm (3004.8 Hz)

D. Colombo et al. / Journal of Steroid Biochemistry & Molecular Biology 103 (2007) 163–169 165

Table 1Geometrical features, relative energies, and equilibrium percentages of the selected conformations of compounds 1–6 representative of the various geometriesof the tricyclic skeleton (the bold data highlight the geometrical differences with respect to the corresponding global minimum conformation)

Erel

(kcal/mol)

% τA (◦) τB (◦) τC (◦) τE (◦) τOH (◦) A ringa B ringa C ringa D ringa

Q φ2 θ Q φ2 θ Q φ2 θ q2 φ2

1A 0.00 88.0 −55 53 −56 −165 165 0.45 23 56 0.54 173 11 0.57 249 4 0.47 1881B 1.18 12.0 54 53 −56 −165 165 0.44 196 127 0.59 290 4 0.58 250 3 0.47 1881C 5.52 0.0 −53 −46 −57 −164 165 0.48 355 53 0.72 249 82 0.58 242 2 0.48 1901D 11.83 0.0 −56 51 49 −172 167 0.46 16 56 0.56 185 17 0.75 325 80 0.50 1932A 0.00 93.0 −56 57 −59 −170 165 0.46 12 55 0.51 132 15 0.59 302 3 0.43 1902B 1.53 7.0 54 57 −61 −170 166 0.44 203 125 0.57 44 7 0.61 68 2 0.48 1882C 6.67 0.0 −54 −54 −61 −169 166 0.51 349 55 0.70 256 88 0.62 53 6 0.49 1892D 9.10 0.0 −57 48 50 −172 165 0.47 16 56 0.58 196 21 0.74 321 81 0.50 1933A 0.00 88.7 −55 53 −55 −166 161 0.45 23 56 0.54 174 12 0.59 263 8 0.41 1833B 1.22 11.3 54 54 −55 −165 162 0.44 196 127 0.59 280 4 0.59 267 7 0.41 1833C 5.56 0.0 −53 −45 −56 −166 162 0.48 355 53 0.72 248 82 0.59 268 5 0.41 1833D 13.65 0.0 −56 50 43 −177 167 0.46 18 56 0.57 191 18 0.73 320 78 0.44 1854A 0.00 86.5 −55 53 −53 – 167 0.45 23 56 0.54 170 11 0.57 282 5 0.46 1864B 1.10 13.5 54 53 −54 – 166 0.44 196 127 0.59 296 4 0.57 291 4 0.47 1864C 5.45 0.0 −54 −50 −54 – 166 0.47 357 53 0.73 254 84 0.58 328 4 0.47 1864D 11.39 0.0 −56 51 46 – 168 0.46 17 56 0.56 187 18 0.74 323 79 0.49 1945A 0.00 95.3 −54 54 −54 – 167 0.44 17 54 0.54 169 6 0.57 271 5 0.46 1865B 1.78 4.7 56 55 −54 – 167 0.45 203 125 0.58 350 5 0.58 300 4 0.47 1865C 6.06 0.0 −48 −52 −54 – 166 0.45 348 54 0.72 261 84 0.59 347 5 0.47 1865D 10.39 0.0 −54 53 44 – 168 0.44 11 55 0.56 182 12 0.73 324 78 0.49 1946Ab 0.00 95.9 −54 53 −54 – – 0.44 17 54 0.54 169 6 0.57 267 5 0.47 1906Bb 1.86 4.1 55 54 −55 – – 0.44 203 125 0.58 348 5 0.58 291 4 0.47 1916Cb 6.25 0.0 −48 −51 −55 – – 0.45 348 53 0.72 261 84 0.59 336 5 0.48 1916Db 11.48 0.0 −54 53 46 – – 0.44 12 54 0.56 183 12 0.74 325 79 0.50 195

τA: C10–C1–C2–C3; τB: C5–C6–C7–C8; τC: C9–C11–C12–C13; τE: C14–C13–C18–C18′; τOH: C16–C17–O–H.a Ring puckering coordinates.b From ref. [10].

Table 2Geometrical features and relative energies of the selected conformations of compounds 1–3 representative of the various orientations of the ethyl group at C13and the hydroxyl group at C17

Erel

(kcal/mol)

τA (◦) τB (◦) τC (◦) τE (◦) τOH (◦) A ringa B ringa C ringa D ringa

Q φ2 θ Q φ2 θ Q φ2 θ q2 φ2

1A 0.00 −55 53 −56 −165 165 0.45 23 56 0.54 173 11 0.57 249 4 0.47 1881Ab 0.25 −55 53 −56 −164 74 0.45 23 56 0.54 173 11 0.57 249 4 0.47 1881Ac 2.87 −55 53 −51 56 165 0.45 22 55 0.54 171 11 0.58 284 9 0.43 1891Ad 2.91 −55 53 −51 58 76 0.45 23 56 0.54 172 11 0.58 282 9 0.43 1881Ae 4.01 −55 53 −47 −90 165 0.45 22 55 0.53 168 11 0.55 319 10 0.49 1831Af 4.50 −55 53 −47 −88 77 0.45 22 55 0.53 168 11 0.55 320 10 0.49 1832A 0.00 −56 57 −59 −170 165 0.46 12 55 0.51 132 15 0.59 302 3 0.43 1902Ab 0.30 −56 57 −59 −168 74 0.46 12 55 0.51 132 15 0.59 41 4 0.48 1892Ac 2.27 −56 58 −58 54 166 0.46 11 54 0.51 130 16 0.60 302 3 0.43 1902Ad 2.32 −56 58 −57 55 76 0.46 11 54 0.51 130 17 0.60 302 3 0.42 1902Ae 3.78 −56 57 −56 −51 165 0.46 11 54 0.51 131 16 0.57 39 5 0.50 1922Af 4.09 −56 57 −56 −51 78 0.46 12 54 0.51 130 16 0.57 38 5 0.50 1933A 0.00 −55 53 −55 −166 161 0.45 23 56 0.54 174 12 0.59 263 8 0.41 1833Ab 0.40 −55 53 −55 −165 81 0.45 23 56 0.54 175 12 0.58 263 8 0.40 1843Ac 0.55 −55 53 −49 53 162 0.45 22 55 0.54 172 12 0.59 288 12 0.38 1833Ad 0.76 −55 53 −49 54 82 0.45 22 56 0.54 172 12 0.59 288 12 0.38 1843Ae 4.63 −55 53 −45 −92 166 0.45 22 55 0.54 168 12 0.57 308 6 0.44 1833Af 5.26 −55 54 −45 −90 86 0.45 22 55 0.54 169 11 0.57 308 13 0.43 1833Ag 5.55 −55 53 −52 −45 166 0.45 22 56 0.54 175 12 0.56 293 6 0.44 1833Ah 6.05 −55 53 −52 −46 85 0.45 22 56 0.54 174 12 0.56 294 6 0.44 183

τA: C10–C1–C2–C3; τB: C5–C6–C7–C8; τC: C9–C11–C12–C13; τE: C14–C13–C18–C18′; τOH: C16–C17–O–H.a Ring puckering coordinates.

166 D. Colombo et al. / Journal of Steroid Biochemistry & Molecular Biology 103 (2007) 163–169

st signi

ibs

3

3

icoartTpgft2actrct3

ptIgccearCcaomai3mo

dtitD

Fig. 1. Three-dimensional plot of the mo

n the proton domain and 180.0 ppm (22,638.6 Hz) in the car-on domain. All 2D spectra were processed with the Brukeroftware package.

. Results

.1. Conformational properties

The conformers located for each compound were groupedn families characterized by the same geometry of the tetra-yclic skeleton; inside each family the members differ in therientation of the ethyl group at C13 and the hydroxyl groupt C17. In Table 1 the most stable members of each family areeported while Table 2 shows for the 13-ethylsteroids 1–3 allhe members of the family to which the global minima belong.he preferred conformation of compounds 1–5, located in theresent study, corresponds, in the tetracyclic skeleton, to theeometry already determined at the same level of calculationsor progesterone 6 [10]. In fact, it can be seen from Table 1hat the most stable conformation of each compound (1A,A, 3A, 4A, 5A, and 6A) show very close values of torsionalngles and puckering coordinates. A careful analysis of theonformational freedom of the tetracyclic skeleton allowedo determine the facility of inversion of the three hexacyclic

ings. As already reported for progesterone 6, also in theases of 1–5 the A ring inversion from the 1�,2�-half-chairo the 1�,2�-half-chair conformation (1A → 1B, 2A → 2B,A → 3B, 4A → 4B, 5A → 5B) is the easiest among all the

Ttip

ficant conformations of compounds 1–5.

ossible ring inversions though a subtle effect is caused byhe methyl group at C10 or the exomethylene group at C11.n the 19-norsteroids (1–4) the absence of the C10 methylroup makes the difference in energy between the 1�,2�-half-hair and the 1�,2�-half-chair conformations smaller than inompounds 5 and 6 which bear a methyl group at C10. Forxample, in compound 1 the energy difference between 1And 1B is 1.18 kcal/mol, while in progesterone 6, the cor-esponding difference between 6A and 6B is 1.86 kcal/mol.onsequently, the percentage contribution of the 1�,2�-half-hair conformation to the overall population is enhanced frombout 4% in 6 to 12% in 1. The exomethylene group presentnly in 2, instead, disadvantages the 1�,2�-half-chair confor-ation (�E = 1.53 kcal/mol) lowering its contribution to 7%

nd making it closer to that found in progesterone. The B ringnversion is much more difficult, in particular in the case of-ketodesogestrel (2A → 2C, �E = 6.67 kcal/mol) and evenore difficult is the C ring inversion, especially in the case

f gestodene (3A → 3D, �E = 13.65 kcal/mol).In levonorgestrel (1), 3-ketodesogestrel (2), and gesto-

ene (3) the ethyl group at C13 might in principle assumehree different orientations: the so called trans conformationn which the C18′ methyl is trans with respect to C14 andhe two gauche conformations having this methyl above the

ring or the C ring (g+ and g−, respectively). The data in

able 2 as well as the plots in Fig. 2 show that, while in all the

hree 13-ethylsteroids the trans orientation (τE about −170◦)s preferred, significant differences are found in the energyrofiles for the rotation of the ethyl group. The presence of

D. Colombo et al. / Journal of Steroid Biochemistry & Molecular Biology 103 (2007) 163–169 167

F

tot3tvbsmm

giacro4

wpth

3

(aapapctTsbaa

Table 31H NMR chemical shifts (ppm) of compounds 1–31H 1 2 3

1� 1.55 1.52 1.471� 2.26 2.50 2.212� 2.39 2.39 2.362� 2.24 2.30 2.174 5.81 5.86 5.796� 2.46 2.46 2.466� 2.25 2.20 2.287� 1.07 1.09 1.107� 1.80 1.80 1.968 1.43 1.45 1.559 0.87 1.50 0.8810 2.08 2.56 2.0911� 1.88 – 1.8611� 1.06 – 1.1411′a – 4.81 –11′b – 5.04 –12� 1.55 2.27 1.7912� 2.00 2.65 1.8714 1.53 1.78 2.1715 – – 5.8615� 1.65 1.64 –15� 1.33 1.38 –16 – – 5.6516� 2.31 2.36 –16� 2.09 2.11 –18(proS) 1.51 1.46 1.7018(proR) 1.51 1.40 1.4712O

3H7�, H7� and H15�, H15� and H16�, H15� and H18(proS),H16� and H18(proS), H9 and H14, H9 and H7�, H10 andH2�, H10 and H11′A, H1� and H11′A, CH318′ and H12�,

Table 413C NMR chemical shifts (ppm) of compounds 1–313C 1 2 3

1 26.57 28.28 26.282 36.53 37.00 36.303 199.89 200.05 199.944 124.66 125.66 124.575 166.51 166.52 166.296 35.52 35.36 35.117 30.69 30.16 30.618 40.91 41.93 37.929 48.91 53.69 49.29

10 42.49 37.62 42.2311 26.21 146.36 25.7611′ – 108.87 –12 28.45 40.53 29.1713 47.97 50.26 55.4814 50.76 52.02 56.0215 22.45 21.89 131.1716 39.55 39.74 135.9117 81.43 80.93 83.30

ig. 2. Energy profile for rotation of the 13-ethyl group of compounds 1–3.

he C15–C16 double bond in gestodene 3 makes the regionver the D ring more accessible than in 1 and 2 loweringhe energy of the corresponding g+ gauche conformationAc which remains, however, about 0.5 kcal/mol less stablehan the trans conformation 3A. This result confirms pre-ious molecular mechanics calculations [15] which foresawoth conformations populated; however, while in the presenttudy the trans conformation was predicted as the most stable,olecular mechanics predicted the g+ gauche conformationore stable than the trans one by more than 1 kcal/mol.In levonorgestrel (1) and 3-ketodesogestrel (2) the g+

auche conformations 1Ac and 2Ac, in which the ethyl groups over the D ring, are 2.87–2.27 kcal/mol less stable than 1And 2A, respectively. Energy minima were also located inorrespondence to the passage of the ethyl group over the Cing, the g− gauche conformations 1Ae–f, 2Ae–f, 3Ae–h; allf them are less stable than the global minima by more thankcal/mol.

The hydroxyl group at C17 has two accessible orientationsith a very small difference in energy. In all the cases thereferred one is the same (τOH about 165◦), in agreement withhe orientation found in the X-ray structure of 17-ethynyl-17-ydroxysteroids [15–18].

.2. NMR data

Complete 1H and 13C NMR signal assignmentsTables 3 and 4) of the spectra of compounds 1–3 werechieved using a combination of 1D and 2D (COSY, HSQCnd NOESY) experiments recorded in CDCl3 at 298 K. Inarticular starting from the cross peak between the char-cteristic vinylic H4 proton (5.79–5.86 ppm) and the H6�roton (2.46 ppm) in the NOESY spectra of all the studiedompounds it was possible to assign the resonances of allhe other protons of 1–3 on the basis of their 2D spectra.he following nOe contacts were observed in the NOESY

pectra of the studied compounds. Levonorgestrel (1): nOeetween H4 and H6�, H6� and H7�, H7� and H15�, H15�nd H16�, H16� and H18, H9 and H14, H9 and H12�, H10nd H11�, H11� and H1�, H1� and H2�, CH318′ and H12�.

1122

8′ 0.99 1.03 0.841 2.58 2.60 2.56H 1.83 1.92 2.63

-Ketodesogestrel (2): nOe between H4 and H6�, H6� and

8 18.92 19.86 20.218′ 9.58 9.07 10.970 87.74 87.61 82.871 74.20 74.36 74.76

168 D. Colombo et al. / Journal of Steroid Biochemistry & Molecular Biology 103 (2007) 163–169

Table 5Experimental 1H NMR coupling constants (Hz) of compounds 1–3 in comparison with the values calculated with the electronegativity-modified Karplusrelationship

J 1 2 3

Calculated Experimental Calculated Experimental Calculated Experimental

1�,2� 3.6 nd 3.3 4.0 3.6 nd1�,2� 12.3 nd 12.8 13.0 12.3 nd1�,10 10.4 11.0 11.4 9.7 10.5 10.81�,10 5.2 4.0 4.0 4.3 5.2 4.01�,2� 4.3 nd 4.0 4.3 4.2 nd1�,2� 3.3 nd 3.0 4.3 3.3 nd4,6� – 2.1 – 2.2 – 2.04,10 – 2.1 – 2.2 – 2.06�,7� 3.7 4.2 3.2 3.3 3.7 3.86�,7� 2.8 2.8 3.3 3.1 2.8 2.36�,7� 13.2 nd 13.3 14.0 13.2 13.66�,7� 3.8 2.8 3.2 5.0 3.8 5.57�,8 12.4 11.0 12.4 11.3 12.4 11.77�,8 2.9 2.8 3.1 nd 3.0 3.28,9 12.0 11.0 12.1 nd 11.9 11.99,10 12.0 11.0 11.7 10.5 12.0 10.99,11� 3.6 4.0 – – 3.8 4.59,11� 12.2 10.5 – – 12.2 11.511�,12� 3.4 3.5 – – 3.7 4.511�,12� 3.3 3.3 – – 3.0 nd11�,12� 13.3 nd – – 13.2 12.111�,12� 3.4 3.3 – – 3.8 5.38,14 12.1 11.0 12.1 13.0 12.1 10.214,15� 5.8 7.0 5.8 7.5 – –14,15� 11.5 12.0 11.6 12.0 – –15�,16� 12.1 10.2 12.1 10.0 – –15�,16� 2.7 3.5 2.8 3.6 – –15�,16� 5.1 6.3 5.1 6.0 – –15�,16� 12.1 12.0 12.1 12.0 – –14,15 – – – – – 1.911

CbHHCHaapalNwaiscfw

o1

H(gaulftwcct

4

e

4,16 – – –5,16 – – –

H318′ and OH, CH318′ and H11′b. Gestodene (3): nOeetween H4 and H6�, H9 and H14, H9 and H7�, H9 and1�, H9 and H12�, H7� and H15, H16 and OH, H1� and2�, H1� and H10, H1� and H11�, CH318′ and H11�,H318′ and H12�, CH318′ and OH, H8 and H18(pro R),11� and H18(pro R). These nOe data were useful for the

ssignment of the � or � configuration of geminal protonsnd also supported the calculated preferred conformations. Inarticular all these contacts correspond to distances of <3 As measured on the computed 1A–3A (Fig. 1) most popu-ated conformations of compounds 1–3. In addition, in theOESY spectrum of gestodene 3 these other nOe contactsere observed: CH318′ and H8, H11� and H18(proS), H8

nd H18(proS) (weak); they correspond to a distance of <3 An the g+ gauche conformation 3Ac and confirm that it is aignificantly populated conformation. The intensities of theseontacts, in comparison with those referable to the trans con-ormation, also confirm the results of the DFT calculations

hich predict the latter conformation as the most populated.Even if some protons in the 1H NMR spectra res-

nated as complex multiplets (H1, H12, H14, and H18 at.48–1.60 ppm in the case of levonorgestrel (1), H1, H9, H18,

hswg

– – 3.4– – 5.8

8, and H15 at 1.35-1.55 ppm in the case of 3-ketodesogestrel2) and H1, H2, and H14 at 2.20–2.30 ppm in the case ofestodene (3)) many of these signals resulted well resolvednd their coupling could be measured. The obtained val-es are reported in Table 5 in comparison with the calcu-ated constants of compounds 1–3. For each populated con-ormer the 1H vicinal coupling constants were calculated withhe electronegativity-modified Karplus relationship [19] andere weighted averaged on the basis of the population per-

entages. The close agreement of the experimental with thealculated values supports the conformational preferences ofhe three steroids determined through the calculations.

. Discussion

The conformational preferences of a group of 13-thylsteroids and related estranes have been determined at a

igh level of calculations in order to ascertain differences andimilarities which might in principle affect activity. Attentionas focussed on two geometrical features related to the pro-estational activity of this class of compounds, namely, the

emistry

ig

ms1oc

tbtHthplftudadg

sprf

A

MWsf

R

[

[

[

[

[

[

[

[

[

D. Colombo et al. / Journal of Steroid Bioch

nversion of the A ring and the orientation of the 13-ethylroup.

The present calculations show that the absence of the C10ethyl group in all the 19-norsteroids (1–4) makes the inver-

ion of the A ring easier. However, the stabilization of the�,2�-half-chair conformation is not sufficient to revert therder of stability present in 5–6 so that the 1�,2�-half-chaironformation remains largely preferred.

The preference of gestodene (3) for a g+ gauche orien-ation of the 13-ethyl group has been hypothesized on theasis of molecular mechanics calculations and by the detec-ion of this molecular geometry in the crystalline state [15].owever, conformationally restricted analogues of 1, syn-

hesized and tested for their oral contraceptive activity [8],ave excluded this gauche as the active conformation. Theresent theoretical study, performed at the B3LYP/6-31G*

evel of calculations, and the high field NMR data suggestor gestodene (3) a preference for the trans conformation,hough also the gauche conformation is significantly pop-lated. This strengthens the hypothesis that the C15–C16ouble bond in 3, like the C11–C11′ double bond in 2, makesn autonomous contribution to the activity which does notepend from a particular conformation adopted by the ethylroup.

It can be concluded that the modulation of activity in theeries of compounds of the present study relies on specificroperties or interactions of their proper functional groupsather than on their effect on the overall shape or on the con-ormational preferences of the molecules.

cknowledgements

This work was financially supported by Universita diilano (FIRST funds) and Universita di Pavia (FAR funds).e thank Professor Fiamma Ronchetti for helpful discus-

ions and Dr. Paride Grisenti (Poli Industria Chimica, Milan)or providing us with a sample of 3-ketodesogestrel.

eferences

[1] J.R. Pasqualini, Progestins: present and future, J. Steroid Biochem.Mol. Biol. 59 (5–6) (1996) 357–365.

[2] A.E. Schindler, C. Campagnoli, R. Druckmann, J. Huber, J.R.Pasqualini, K.W. Schweppe, J.H.H. Thijssen, Classification and phar-macology of progestins, Maturitas 46S1 (2003) S7–S16.

[3] R. Bursi, M.B. Groen, Application of (quantitative) structure–activityrelationships to progestagens: from serendipidity to structure-based

design, Eur. J. Med. Chem. 35 (9) (2000) 787–796.

[4] R. Vendrame, M.M.C. Ferreira, C.H. Collins, Y. Takahata,Structure–activity relationships (SAR) of contraceptive progestogensstudied with four different methods using calculated physicochemicalparameters, J. Mol. Graph. Model. 20 (4) (2002) 345–358.

[

& Molecular Biology 103 (2007) 163–169 169

[5] R. Kiralj, Y. Takahata, M.M.C. Ferreira, QSAR of progestogens: use ofa priori and computed molecular descriptors and molecular graphics,QSAR Comb. Sci. 22 (4) (2003) 430–448.

[6] Z. Galdecki, P. Grochulski, Z. Wawrzak, E. Galdecka, W.L. Duax,P.D. Strong, Crystal structure of 4,11-pregnadiene-3,20-dione and 21-methyl-20-oxa-4-pregnene-3,20-dione, J. Chem. Crystallogr. 26 (7)(1996) 497–502.

[7] J.P. Raynaud, T. Ojasoo, M.M. Bouton, D. Philibert, in: E.J. Ariens(Ed.), Drug Design, Academic Press, New York, 1979, pp. 169–214.

[8] A.I.A. Broess, M.B. Groen, H. Hamersma, Synthesis of conforma-tionally restricted progestational 13-ethylsteroids, Bioorg. Med. Chem.Lett. 7 (23) (1997) 2929–2934.

[9] A. Szyczewsky, K. Holderna-Natkaniec, I. Natkaniec, Progesterone andtestosterone studies by neutron scattering and nuclear magnetic reso-nance methods and quantum chemistry calculations, J. Mol. Struct. 693(1–3) (2004) 49–71.

10] D. Colombo, P. Ferraboschi, P. Prestileo, L. Toma, A comparativemolecular modeling study of dydrogesterone with other progestationalagents through theoretical calculations and nuclear magnetic resonancespectroscopy, J. Steroid Biochem. Mol. Biol. 98 (1) (2006) 56–62.

11] P. Ferraboschi, P. Grisenti, A. Onofri, P. Prestileo, A simple synthesisof gestodene from 18-methyl-4-estren-3,17-dione, Synlett (10) (2004)1838–1840.

12] S. Ferrino, F. Lopez-Tapia, H. Salgano-Zamora, An improved alter-native norgestrel preparation, Synthetic Commun. 26 (7) (1996)1461–1465.

13] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,J.R. Cheeseman, J.A. Montgomery Jr., T. Vreven, K.N. Kudin, J.C.Burant, J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B. Mennucci,M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H. Nakatsuji, M.Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox,H.P. Hratchian, J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R.E.Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochter-ski, P.Y. Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg,V.G. Zakrzewski, S. Dapprich, A.D. Daniels, M.C. Strain, O. Farkas,D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J.V. Ortiz,Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu,A. Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith,M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W.Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A. Pople, Gaus-sian 03, Revision B. 02, Gaussian Inc., Pittsburgh, PA, 2003.

14] D. Cremer, J.A. Pople, A general definition of ring puckering coordi-nates, J. Am. Chem. Soc. 97 (6) (1975) 1354–1358.

15] V.J. van Geerestein, A.J.M. Duisenberg, M.H.G. Duits, J.A. Kanters,J. Kroon, Structure of modifications (I) and (II) of 13-ethyl-17�-hydroxy-18,19-dinor-17�-pregna-4,15-dien-20-yn-3-one (gestodene),Acta Cryst. C 43 (12) (1987) 2402–2405.

16] N.J. DeAngelis, T.H. Doyne, L. Grob, The crystal and molecular struc-ture of d-norgestrel, a progestational steroid, Acta Cryst. B 31 (8) (1975)2040–2043.

17] V.J. van Geerestein, J.A. Kanters, J. Kroon, Structure of 13-ethyl-17�-hydroxy-11-methylene-18,19-dinor-17�-pregn-4-en-20-yn-3-one (3-ketodesogestrel), Acta Cryst. C 43 (12) (1987) 2398–2401.

18] J.P. Mornon, G. Lepicard, J. Delettre, Crystal structure of norethin-drone (19-nor-17�-ethynyltestosterone), C.R. Acad. Sci. Paris C 282

(8) (1976) 387–390.

19] C.A.G. Haasnoot, F.A.A.M. de Leeuw, C. Altona, The relationshipbetween proton–proton NMR coupling constants and substituent elec-tronegativities. Part I. An empirical generalization of the Karplus equa-tion, Tetrahedron 36 (19) (1980) 2783–2792.