Transferred multipolar atom model for 10β,17β-di­hydroxy-17α-methyl­estr-4-en-3-one dihydrate obtained from the biotransformation of methyl­oestrenolone

Biotransformation is the structural modification of compounds using enzymes as the catalysts and it plays a key role in the synthesis of pharmaceutically important compounds. 10β,17β-Di­hydroxy-17α-methyl­estr-4-en-3-one dihydrate, C₁₉H₂₈O₃·2H₂O, was obtained from the fungal biotransformation of methyl­oestrenolone. The structure was refined using the classical independent atom model (IAM) and a transferred multipolar atom model using the ELMAM2 database. The results from the two refinements have been compared. The ELMAM2 refinement has been found to be superior in terms of the refinement statistics. It has been shown that certain electron-density-derived properties can be calculated on the basis of the transferred parameters for crystals which diffract to ordinary resolution.

S1. Introduction

Biotransformation or biocatalysis is the structural modification of compounds using enzymes as the catalysts (Mahato & Garai, 1997). Biotransformation using whole-cell cultures of microorganisms is an efficient method for the production of libraries of structurally diverse molecules, starting from a single substrate (Mahato & Garai, 1997; Ahmad et al., 2014; Khan et al., 2014; Zafar et al., 2012; Leon et al., 1998; Holland, 1999). This methodology has been applied for the synthesis of bioactive molecules of varying nature (Choudhary et al., 2012; Monagas et al., 2010; Liu & Yu, 2010; Cao et al., 2015; Wang, 2015). Biotransformation has found its applications in organic synthesis as well. Due to the specific nature of enzymes, the reactions take place with high stereo- and regioselectivity (Wang et al., 2000; Wang, 2015; Palomo & Filice, 2015). Therefore, biotransformation plays a key role in the synthesis of pharmaceutically important compounds.

The chemical properties of the molecule depend on the charge density distribution (Coppens, 1998; Jayatilaka et al., 2012). A detailed knowledge of the structural parameters, stereochemistry, planarity and the mutual arrangement of the molecule is required to better understand the structure–activity relationships. In addition, knowledge of the intra- and inter­molecular inter­actions is important for the design of new drugs and modelling the inter­actions with proteins. The commonly used independent atom model (IAM) does not give all the information about the inter­molecular inter­actions and is likely to produce severe systematic errors in the refined atomic parameters (Ruysink & Vos, 1974). Experimental electron-density analysis is carried out by X-ray diffraction of monocrystals at ultra high resolution (Coppens, 1998). However, the difficult part is the separation of the anisotropic atomic mean-square displacements from the static molecular electron distribution (Hirshfeld, 1976). Proper experimental deconvolution requires diffraction data measured at ultra high resolution. However, effective thermal displacement deconvolution and meaningful electron-density distributions can be achieved even at lower resolutions by transferring the parameters from an electron-density database (Pichon-Pesme et al., 1995; Jelsch et al., 1998; Dittrich et al., 2004, 2005, 2007; Ahmed et al., 2011). Transferring electron-density parameters is comparable to constraining the atomic coordinates at their most likely values. The transferability of atomic electron densities was tested for the first time by Brock et al. (1991) who applied atomic charge-density parameters from an accurate low-temperature study of perylene to diffraction data collected at several temperatures on naphthalene and anthracene crystals.

The ELMAM database (Zarychta et al., 2007) has been extended to ELMAM2 (Domagała et al., 2012) from protein atom types to common organic molecules, and is based on optimal local coordinate systems (Domagała & Jelsch, 2008). An automatic transfer procedure of the ELMAM2 database is now available in the MoPro software (Guillot et al., 2001; Jelsch et al., 2005). Different atom types in a molecule are recognized according to the nature and number of their neighbors. For most atoms, only the first shell of neighbours is analysed, while for H and O atoms, the second and third shells are investigated, respectively (Domagała & Jelsch, 2008; Domagała et al., 2012). Using the transferability principle, a multipolar model is applied for the molecule, and only the structural parameters (scale factor, atomic coordinates and displacement parameters) are refined. The Fourier residual maps are improved, notably on the covalent bonds due to the proper electron-density modelling.

During the current study, 17α-methyl-10β,17β-di­hydroxy­estr-4-en-3-one (Scheme 1) was obtained from the fungal biotransformation of methyl­oestrenolone, a contraceptive agent (Zafar et al., 2013; Yousuf et al., 2010). The substrate belongs to the estrane class of compounds, steroids without the C-19 methyl group (Mason, 1948). Its structure has been investigated by single-crystal X-ray diffraction analysis and its molecular properties have been investigated by using the electron-density parameters from the ELMAM2 library.

S2. Experimental

S2.1. Culture preparation

Aspergillus niger (ATCC 10549) was cultured in liquid media, prepared by dissolving glucose (10.0 g), peptone (5.0 g), yeast extract (3.0 g), KH₂PO₄ (5.0 g), glycerol (10.0 ml) and NaCl (5.0 g) per litre of distilled water. The media was sterilized followed by inoculation of the fungal spores. After optimum growth, methyl­oestrenolone (1.0 g) was transferred to the culture in the form of solution in acetone (Zafar et al., 2013).

S2.2. Synthesis and crystallization

Incubation of methyl­oestrenolone was carried out for 14 d. The biomass was then separated by filtration and the filtrate was extracted with di­chloro­methane. The extract was subjected to column chromatography and size exclusion HPLC (GS-320, MeOH, retention time = 33 min) to obtain 10β,17α-di­hydroxy-17α-methyl­estr-4-en-3-one as a pure white crystalline solid. The yield of the reaction was only 0.08%.

S2.3. Structure solution and refinement

Structure refinement details for the transferred model are summarized in Table 1. The structure was solved in the orthorhombic space group P2₁2₁2₁ using SIR92 (Altomare et al., 1993) available in the WinGX package (Farrugia, 2012). An Initial IAM refinement was carried out using SHELXL2013(Sheldrick, 2015). Most of the H atoms were located in difference Fourier maps. However, H atoms on O atoms could not be ascertained in the difference Fourier maps, hence a riding model was used as was done for C atoms. After the initial cycles of refinement, the model was imported to MoPro (Jelsch et al., 2005). A I/σ cut-off of 3 was applied throughout the MoPro refinements. The bond lengths for H atoms were constrained to standard neutron distances from the Inter­national Tables of Crystallography (Allen & Bruno, 2010).

S2.3.1. IAM refinement

A full-matrix least-squares refinement was carried out against intensity data using MoPro (Jelsch et al., 2005). A SHELX-type weighting scheme was adopted with w = 1/[\s²(Fo²)+(aP)²+bP] where P = (Fo²+2Fc²)/3), a = 0.1 and b = 0.01 in order to have a goodness-of-fit close to unity. The MoPro software provides an easy way to modify the weighting scheme. The normal probability plots (Zhurov et al., 2008) are shown in the Supporting information (Fig. S1). Initially, only the scale factor was refined. Subsequently, the positions and displacement parameters were refined. The refinement was carried out until convergence. The anisotropic thermal displacement parameters for H atoms were constrained to the estimated values using the SHADE server (Madsen, 2006). The residual electron-density maps after the IAM refinement have been shown in Fig. S2(a) (see Supporting information). It can be seen from the residual maps that the electron-density peaks reside on the covalent bonds and there are no significant residual density peaks on the atoms. At the end of the IAM refinement, the crystallographic R factor R[F² > 2σ(F²)] was 6.0%, the weighted R factor wR(F²) was 14.9% while the goodness-of-fit was 1.05. The minimum and the maximum electron-density peaks after IAM refinement were -0.35 and 0.41 e Å^-3, respectively.

S2.3.2. ELMAM2 refinement

The electron-density multipolar parameters of Hansen and Coppen's model (Hansen & Coppens, 1978) were transferred from the ELMAM2 library (Domagała et al., 2012) using the built-in option in the MoPro software (Jelsch et al., 2005). The molecule was neutralized electrically after the transfer. The electron-density parameters were kept fixed during the ELMAM2 refinement and only the scale factor, position and displacement parameters were refined until convergence. The anisotropic thermal displacement parameters for H atoms were kept constrained to the estimated values as explained previously. The ELMAM2 refinement revealed a noticeable improvement in the refinement statistics with the crystallographic R factor R[F² > 2σ(F²)] was 4.8 %, the weighted R factor wR(F²) was 11.6%, while the goodness-of-fit was 0.95. The minimum and the maximum electron-density peaks had values of -0.26 and 0.35 e Å^-3, respectively.

S3. Results and discussion

The asymmetric unit of the title compound, (I) (Fig. 1), consists of one parent moiety along with two water molecules of crystallization, whereas the unit cell contains four molecules along with eight water molecules. The compound possesses the typical core structure of steroids having three six-membered rings and a five-membered ring. The cyclo­hexene ring (A) is not planer in shape and adopts a half-chair conformation. Atoms C2, C3, C4 and C5 lie in the same plane. However, atom C1 deviates from the mean C2—C3—C4—C5 plane by 0.713 Å, whereas atom C10 deviates by 0.218 Å. The two central cyclo­hexane rings (B and C) both adopt a typical chair conformation, whereas cyclo­pentane ring D adopts a 13β,17α-envelope conformation (Cremer & Pople, 1975). The A/B ring junction is quasi-trans, whereas the B/C and C/D ring junctions both have a trans conformation (Bartcourt, 1974). The torsion angle of the opposing methyl groups is -161.83°, as measured on C18—C13—C17—C20. All the bond lengths and angles are within the normal range for similar structures (Galdecki et al., 1990; Vasuki et al., 2002; Thamotharan et al., 2004; Shen et al., 2011).

The molecular assembly is stabilized by strong inter­molecular hydrogen bonds (Table 2) which are reinforced by a number of lateral weak H···H inter­actions. The primary aggregation motif is the O—H···O hydrogen bond between the C17 hy­droxy donor and O3 ketone acceptor (Duax & Norton, 1975), with an H···O distance of 1.77Å and a D—H···A angle of 167°. This motif is typical of steroidal assemblies and results in the P2₁ or P2₁2₁2₁ space group. The two water molecules serve to bridge the molecular chains along three dimensions through strong inter­molecular hydrogen bonding (Fig. 2). The O1W water molecule forms three strong O—H···O hydrogen bonds. Firstly, it donates its H3W atom to the O2 atom of the C17 hyr­oxy group at a distance of 1.92 Å and a D—H···A angle of 179°. Secondly, it accepts the H3 atom of the C10 hy­droxy group at a distance of 1.80 Å and a D—H···A angle of 172°. Thirdly, it donates atom H1W to the O2W atom of the other water molecule, with an H···O distance of 1.84 Å and a D—H···A angle of 176°. The O2W water molecule further links with two other neighbouring molecules by donating its H atoms to the O2 and O3 atoms of the C17 and C10 hy­droxy groups, with H···O distances of 2.07 and 1.82 Å and D—H···A angles of 161 and 171°, respectively.

S3.1. Hirshfeld surface analysis

The Hirshfeld surface emerged from an attempt to define the space occupied by a molecule in a crystal for the purpose of partitioning the crystal electron density into molecular fragments (Spackman & Byrom, 1997). First function of distance explored for mapping on surface is the distances from the Hirshfeld surface to the adjacent nucleus inside the surface (or inter­nal, d_i) and outside the surface (or external, d_e). The informative use of these qu­anti­ties one can find combination of d_i and d_e in the form of a two-dimensional fingerprint plot. Due to the difficulty of representation in two-dimensional format (printed page or computer monitor) fingerprint plot was invented. Two different but potentially useful distance measures each mapped on a three-dimensional molecular surface were invented. Surface features characteristic of different types of inter­molecular inter­actions can be identified and such features can be exposed by color coding distances from the surface to the nearest atom exterior or inter­ior to surface. In this study d_e and d_i properties are calculated separately and also proposed a contact distance, d_norm , that combines both d_e and d_i, each normalized by the van der Waals radius for the particular atoms involved in the close contact to the surface. Using a red–white– blue colouring scheme to distinguish it from the red–green–blue schemes for d_e and d_i, contacts shorter than van der Waals separations show up as red spots on a largely blue surface. And because the definition of d_norm is symmetric in both d_e and d_i, close inter­molecular contacts appear as three identical red spots (but not necessarily on the same molecule if there is more than one molecule in the asymmetric unit.

Crystal Explorer 3.1 (Wolff et al., 2012) was used to generate Hirshfeld surfaces mapped over d_norm property of the molecule. It is seen that hydrogen bond between carbonyl atom O1 and water atom H2W is associated with large red spots, while the alcohol groups present on the ring making inter­molecular contacts are shown by the regions coloured in red (Fig. 3).

The fingerprint plots (Fig. 4) of structure supports the inter­molecular inter­actions. Three types of inter­actions are found in crystal packing: H···H inter­actions account for 76.5%, O···H 21.4% and C···H 2.1%. Calculation of the enrichment ratio (ER) (Jelsch et al., 2014) shows that the H···H contacts are relatively more favoured, with an ER value of 1.13, over O···H and C···H inter­actions, whose ER values are 0.566 and 0.565, respectively.

S3.2. Electrostatic potential and the dipole moment

The presence of possible electrostatic inter­actions around the molecule can be described qualitatively by the electrostatic potential map (ESP) on the electron-density isosurface. The electrostatic potential can be calculated directly from the electron density (Su & Coppens, 1992). The electrostatic potential mapped on the isosurface of electron density had a value of 0.01 e Å^-3; Fig. 5 shows two views of the three-dimensional electron-density surface coloured according to the electrostatic potential. Fig. 5(a) shows a projection of the molecule with the hy­droxy groups projected towards the viewer, whereas in Fig. 5(b), the molecule was rotated to show it downwards. The maps show that around all the three O atoms there is an accumulation of the electron density. Atom H2 of the hy­droxy group has a strong positive region. The carbonyl O1 atom has the greatest concentration of negative charge. In fact, this is the reason behind the polar nature of the compound, as evident from relatively high dipole moment value (10.09 Debye) of the molecule (Fig. 6). The dipole moment was calculated for an electrically neutralized molecule using MoProViewer (Guillot, 2011), with the origin at the coordinate centre, and is the sum of the dipolar contributions only. The polar nature of the molecule also results in the typical head-to-tail arrangement in the crystal structures. During the process of biotransformation, the H atom on C10 is replaced by the O3 hy­droxy group, which has an accumulation of the negative charge. Since electrostatic forces are long-range forces, the molecules can recognize each other from a distance. These negative electrostatic regions can be potential binding sites for hydrogen bonds and for attack by the electrophilic species. It can be said with confidence that the binding mode of the biotransformed product will certainly be different from that of its parent molecule methyl­oestrenolone.

The topography of the electrostatic potential in the molecule under study is significantly different from that of the estrone molecule (Zhurova et al., 2006) due to opposite locations of the carbonyl O atom and the hy­droxy group in two molecules. The addition of the O3 hy­droxy group in the current study renders the syn sides of rings A, B and C significantly negative or neutral, whereas the anti sides of the rings are electrostatically more positive. Ring D has the greatest concentration of positive potential. The same has been indicated by the direction of the dipole moment vector.

S3.3. Topology of inter­molecular inter­actions

A qu­anti­tative analysis of the inter­molecular inter­actions of the title molecule was carried out on the basis of topological analysis using the Bader quantum theory of atoms in molecules (Bader, 1990). The literature reports the successful use of the transferability principle to precisely qu­antify the electron-density-derived properties for the crystals which diffract to ordinary resolutions (Domagała et al. 2011; Bibila Mayaya Bisseyou et al. 2012; Dadda et al., 2012). All the three O atoms of the parent moiety are involved in the formation of some O—H···O type hydrogen-bond inter­actions at very short distances of less than 2 Å. The electron-density values at the critical points are significantly higher than other types of inter­actions. Moreover, the presence of 3,-1-type bond-critical points and a positive Laplacian values establish that they are strong hydrogen bonds. Table S1 (see Supporting information) gives the topological parameters of the inter­molecular inter­actions.

S4. Conclusion

This paper reports the crystal structure and charge–density properties of a molecule which was obtained as a biotransformation product of the steroid drug methyl­estrenolone. The structure is stabilized by a number of O—H···O and C—H···O hydrogen bonds which are assisted by a number H···H inter­actions. Even though the diffraction data were limited to ordinary resolution, the charge–density properties, electrostatic potential, dipole moment and topology of inter­molecular inter­actions could be estimated using the transferred parameters from the ELMAM2 database. The results show that the ELMAM2 model is better. The procedure is very rapid and simplified in the MoPro software although it may require some manual inter­vention. The method is very suitable for high throughput screening of diverse targets, especially biologically active molecules, if diffraction data is available.

Structure description

Biotransformation or biocatalysis is the structural modification of compounds using enzymes as the catalysts (Mahato & Garai, 1997). Biotransformation using whole-cell cultures of microorganisms is an efficient method for the production of libraries of structurally diverse molecules, starting from a single substrate (Mahato & Garai, 1997; Ahmad et al., 2014; Khan et al., 2014; Zafar et al., 2012; Leon et al., 1998; Holland, 1999). This methodology has been applied for the synthesis of bioactive molecules of varying nature (Choudhary et al., 2012; Monagas et al., 2010; Liu & Yu, 2010; Cao et al., 2015; Wang, 2015). Biotransformation has found its applications in organic synthesis as well. Due to the specific nature of enzymes, the reactions take place with high stereo- and regioselectivity (Wang et al., 2000; Wang, 2015; Palomo & Filice, 2015). Therefore, biotransformation plays a key role in the synthesis of pharmaceutically important compounds.

The chemical properties of the molecule depend on the charge density distribution (Coppens, 1998; Jayatilaka et al., 2012). A detailed knowledge of the structural parameters, stereochemistry, planarity and the mutual arrangement of the molecule is required to better understand the structure–activity relationships. In addition, knowledge of the intra- and inter­molecular inter­actions is important for the design of new drugs and modelling the inter­actions with proteins. The commonly used independent atom model (IAM) does not give all the information about the inter­molecular inter­actions and is likely to produce severe systematic errors in the refined atomic parameters (Ruysink & Vos, 1974). Experimental electron-density analysis is carried out by X-ray diffraction of monocrystals at ultra high resolution (Coppens, 1998). However, the difficult part is the separation of the anisotropic atomic mean-square displacements from the static molecular electron distribution (Hirshfeld, 1976). Proper experimental deconvolution requires diffraction data measured at ultra high resolution. However, effective thermal displacement deconvolution and meaningful electron-density distributions can be achieved even at lower resolutions by transferring the parameters from an electron-density database (Pichon-Pesme et al., 1995; Jelsch et al., 1998; Dittrich et al., 2004, 2005, 2007; Ahmed et al., 2011). Transferring electron-density parameters is comparable to constraining the atomic coordinates at their most likely values. The transferability of atomic electron densities was tested for the first time by Brock et al. (1991) who applied atomic charge-density parameters from an accurate low-temperature study of perylene to diffraction data collected at several temperatures on naphthalene and anthracene crystals.

The ELMAM database (Zarychta et al., 2007) has been extended to ELMAM2 (Domagała et al., 2012) from protein atom types to common organic molecules, and is based on optimal local coordinate systems (Domagała & Jelsch, 2008). An automatic transfer procedure of the ELMAM2 database is now available in the MoPro software (Guillot et al., 2001; Jelsch et al., 2005). Different atom types in a molecule are recognized according to the nature and number of their neighbors. For most atoms, only the first shell of neighbours is analysed, while for H and O atoms, the second and third shells are investigated, respectively (Domagała & Jelsch, 2008; Domagała et al., 2012). Using the transferability principle, a multipolar model is applied for the molecule, and only the structural parameters (scale factor, atomic coordinates and displacement parameters) are refined. The Fourier residual maps are improved, notably on the covalent bonds due to the proper electron-density modelling.

During the current study, 17α-methyl-10β,17β-di­hydroxy­estr-4-en-3-one (Scheme 1) was obtained from the fungal biotransformation of methyl­oestrenolone, a contraceptive agent (Zafar et al., 2013; Yousuf et al., 2010). The substrate belongs to the estrane class of compounds, steroids without the C-19 methyl group (Mason, 1948). Its structure has been investigated by single-crystal X-ray diffraction analysis and its molecular properties have been investigated by using the electron-density parameters from the ELMAM2 library.

Aspergillus niger (ATCC 10549) was cultured in liquid media, prepared by dissolving glucose (10.0 g), peptone (5.0 g), yeast extract (3.0 g), KH₂PO₄ (5.0 g), glycerol (10.0 ml) and NaCl (5.0 g) per litre of distilled water. The media was sterilized followed by inoculation of the fungal spores. After optimum growth, methyl­oestrenolone (1.0 g) was transferred to the culture in the form of solution in acetone (Zafar et al., 2013).

The asymmetric unit of the title compound, (I) (Fig. 1), consists of one parent moiety along with two water molecules of crystallization, whereas the unit cell contains four molecules along with eight water molecules. The compound possesses the typical core structure of steroids having three six-membered rings and a five-membered ring. The cyclo­hexene ring (A) is not planer in shape and adopts a half-chair conformation. Atoms C2, C3, C4 and C5 lie in the same plane. However, atom C1 deviates from the mean C2—C3—C4—C5 plane by 0.713 Å, whereas atom C10 deviates by 0.218 Å. The two central cyclo­hexane rings (B and C) both adopt a typical chair conformation, whereas cyclo­pentane ring D adopts a 13β,17α-envelope conformation (Cremer & Pople, 1975). The A/B ring junction is quasi-trans, whereas the B/C and C/D ring junctions both have a trans conformation (Bartcourt, 1974). The torsion angle of the opposing methyl groups is -161.83°, as measured on C18—C13—C17—C20. All the bond lengths and angles are within the normal range for similar structures (Galdecki et al., 1990; Vasuki et al., 2002; Thamotharan et al., 2004; Shen et al., 2011).

The molecular assembly is stabilized by strong inter­molecular hydrogen bonds (Table 2) which are reinforced by a number of lateral weak H···H inter­actions. The primary aggregation motif is the O—H···O hydrogen bond between the C17 hy­droxy donor and O3 ketone acceptor (Duax & Norton, 1975), with an H···O distance of 1.77Å and a D—H···A angle of 167°. This motif is typical of steroidal assemblies and results in the P2₁ or P2₁2₁2₁ space group. The two water molecules serve to bridge the molecular chains along three dimensions through strong inter­molecular hydrogen bonding (Fig. 2). The O1W water molecule forms three strong O—H···O hydrogen bonds. Firstly, it donates its H3W atom to the O2 atom of the C17 hyr­oxy group at a distance of 1.92 Å and a D—H···A angle of 179°. Secondly, it accepts the H3 atom of the C10 hy­droxy group at a distance of 1.80 Å and a D—H···A angle of 172°. Thirdly, it donates atom H1W to the O2W atom of the other water molecule, with an H···O distance of 1.84 Å and a D—H···A angle of 176°. The O2W water molecule further links with two other neighbouring molecules by donating its H atoms to the O2 and O3 atoms of the C17 and C10 hy­droxy groups, with H···O distances of 2.07 and 1.82 Å and D—H···A angles of 161 and 171°, respectively.

The Hirshfeld surface emerged from an attempt to define the space occupied by a molecule in a crystal for the purpose of partitioning the crystal electron density into molecular fragments (Spackman & Byrom, 1997). First function of distance explored for mapping on surface is the distances from the Hirshfeld surface to the adjacent nucleus inside the surface (or inter­nal, d_i) and outside the surface (or external, d_e). The informative use of these qu­anti­ties one can find combination of d_i and d_e in the form of a two-dimensional fingerprint plot. Due to the difficulty of representation in two-dimensional format (printed page or computer monitor) fingerprint plot was invented. Two different but potentially useful distance measures each mapped on a three-dimensional molecular surface were invented. Surface features characteristic of different types of inter­molecular inter­actions can be identified and such features can be exposed by color coding distances from the surface to the nearest atom exterior or inter­ior to surface. In this study d_e and d_i properties are calculated separately and also proposed a contact distance, d_norm , that combines both d_e and d_i, each normalized by the van der Waals radius for the particular atoms involved in the close contact to the surface. Using a red–white– blue colouring scheme to distinguish it from the red–green–blue schemes for d_e and d_i, contacts shorter than van der Waals separations show up as red spots on a largely blue surface. And because the definition of d_norm is symmetric in both d_e and d_i, close inter­molecular contacts appear as three identical red spots (but not necessarily on the same molecule if there is more than one molecule in the asymmetric unit.

Crystal Explorer 3.1 (Wolff et al., 2012) was used to generate Hirshfeld surfaces mapped over d_norm property of the molecule. It is seen that hydrogen bond between carbonyl atom O1 and water atom H2W is associated with large red spots, while the alcohol groups present on the ring making inter­molecular contacts are shown by the regions coloured in red (Fig. 3).

The fingerprint plots (Fig. 4) of structure supports the inter­molecular inter­actions. Three types of inter­actions are found in crystal packing: H···H inter­actions account for 76.5%, O···H 21.4% and C···H 2.1%. Calculation of the enrichment ratio (ER) (Jelsch et al., 2014) shows that the H···H contacts are relatively more favoured, with an ER value of 1.13, over O···H and C···H inter­actions, whose ER values are 0.566 and 0.565, respectively.

The presence of possible electrostatic inter­actions around the molecule can be described qualitatively by the electrostatic potential map (ESP) on the electron-density isosurface. The electrostatic potential can be calculated directly from the electron density (Su & Coppens, 1992). The electrostatic potential mapped on the isosurface of electron density had a value of 0.01 e Å^-3; Fig. 5 shows two views of the three-dimensional electron-density surface coloured according to the electrostatic potential. Fig. 5(a) shows a projection of the molecule with the hy­droxy groups projected towards the viewer, whereas in Fig. 5(b), the molecule was rotated to show it downwards. The maps show that around all the three O atoms there is an accumulation of the electron density. Atom H2 of the hy­droxy group has a strong positive region. The carbonyl O1 atom has the greatest concentration of negative charge. In fact, this is the reason behind the polar nature of the compound, as evident from relatively high dipole moment value (10.09 Debye) of the molecule (Fig. 6). The dipole moment was calculated for an electrically neutralized molecule using MoProViewer (Guillot, 2011), with the origin at the coordinate centre, and is the sum of the dipolar contributions only. The polar nature of the molecule also results in the typical head-to-tail arrangement in the crystal structures. During the process of biotransformation, the H atom on C10 is replaced by the O3 hy­droxy group, which has an accumulation of the negative charge. Since electrostatic forces are long-range forces, the molecules can recognize each other from a distance. These negative electrostatic regions can be potential binding sites for hydrogen bonds and for attack by the electrophilic species. It can be said with confidence that the binding mode of the biotransformed product will certainly be different from that of its parent molecule methyl­oestrenolone.

The topography of the electrostatic potential in the molecule under study is significantly different from that of the estrone molecule (Zhurova et al., 2006) due to opposite locations of the carbonyl O atom and the hy­droxy group in two molecules. The addition of the O3 hy­droxy group in the current study renders the syn sides of rings A, B and C significantly negative or neutral, whereas the anti sides of the rings are electrostatically more positive. Ring D has the greatest concentration of positive potential. The same has been indicated by the direction of the dipole moment vector.

A qu­anti­tative analysis of the inter­molecular inter­actions of the title molecule was carried out on the basis of topological analysis using the Bader quantum theory of atoms in molecules (Bader, 1990). The literature reports the successful use of the transferability principle to precisely qu­antify the electron-density-derived properties for the crystals which diffract to ordinary resolutions (Domagała et al. 2011; Bibila Mayaya Bisseyou et al. 2012; Dadda et al., 2012). All the three O atoms of the parent moiety are involved in the formation of some O—H···O type hydrogen-bond inter­actions at very short distances of less than 2 Å. The electron-density values at the critical points are significantly higher than other types of inter­actions. Moreover, the presence of 3,-1-type bond-critical points and a positive Laplacian values establish that they are strong hydrogen bonds. Table S1 (see Supporting information) gives the topological parameters of the inter­molecular inter­actions.

This paper reports the crystal structure and charge–density properties of a molecule which was obtained as a biotransformation product of the steroid drug methyl­estrenolone. The structure is stabilized by a number of O—H···O and C—H···O hydrogen bonds which are assisted by a number H···H inter­actions. Even though the diffraction data were limited to ordinary resolution, the charge–density properties, electrostatic potential, dipole moment and topology of inter­molecular inter­actions could be estimated using the transferred parameters from the ELMAM2 database. The results show that the ELMAM2 model is better. The procedure is very rapid and simplified in the MoPro software although it may require some manual inter­vention. The method is very suitable for high throughput screening of diverse targets, especially biologically active molecules, if diffraction data is available.

Synthesis and crystallization

Incubation of methyl­oestrenolone was carried out for 14 d. The biomass was then separated by filtration and the filtrate was extracted with di­chloro­methane. The extract was subjected to column chromatography and size exclusion HPLC (GS-320, MeOH, retention time = 33 min) to obtain 10β,17α-di­hydroxy-17α-methyl­estr-4-en-3-one as a pure white crystalline solid. The yield of the reaction was only 0.08%.

Refinement details

Structure refinement details for the transferred model are summarized in Table 1. The structure was solved in the orthorhombic space group P2₁2₁2₁ using SIR92 (Altomare et al., 1993) available in the WinGX package (Farrugia, 2012). An Initial IAM refinement was carried out using SHELXL2013(Sheldrick, 2015). Most of the H atoms were located in difference Fourier maps. However, H atoms on O atoms could not be ascertained in the difference Fourier maps, hence a riding model was used as was done for C atoms. After the initial cycles of refinement, the model was imported to MoPro (Jelsch et al., 2005). A I/σ cut-off of 3 was applied throughout the MoPro refinements. The bond lengths for H atoms were constrained to standard neutron distances from the Inter­national Tables of Crystallography (Allen & Bruno, 2010).

A full-matrix least-squares refinement was carried out against intensity data using MoPro (Jelsch et al., 2005). A SHELX-type weighting scheme was adopted with w = 1/[\s²(Fo²)+(aP)²+bP] where P = (Fo²+2Fc²)/3), a = 0.1 and b = 0.01 in order to have a goodness-of-fit close to unity. The MoPro software provides an easy way to modify the weighting scheme. The normal probability plots (Zhurov et al., 2008) are shown in the Supporting information (Fig. S1). Initially, only the scale factor was refined. Subsequently, the positions and displacement parameters were refined. The refinement was carried out until convergence. The anisotropic thermal displacement parameters for H atoms were constrained to the estimated values using the SHADE server (Madsen, 2006). The residual electron-density maps after the IAM refinement have been shown in Fig. S2(a) (see Supporting information). It can be seen from the residual maps that the electron-density peaks reside on the covalent bonds and there are no significant residual density peaks on the atoms. At the end of the IAM refinement, the crystallographic R factor R[F² > 2σ(F²)] was 6.0%, the weighted R factor wR(F²) was 14.9% while the goodness-of-fit was 1.05. The minimum and the maximum electron-density peaks after IAM refinement were -0.35 and 0.41 e Å^-3, respectively.

The electron-density multipolar parameters of Hansen and Coppen's model (Hansen & Coppens, 1978) were transferred from the ELMAM2 library (Domagała et al., 2012) using the built-in option in the MoPro software (Jelsch et al., 2005). The molecule was neutralized electrically after the transfer. The electron-density parameters were kept fixed during the ELMAM2 refinement and only the scale factor, position and displacement parameters were refined until convergence. The anisotropic thermal displacement parameters for H atoms were kept constrained to the estimated values as explained previously. The ELMAM2 refinement revealed a noticeable improvement in the refinement statistics with the crystallographic R factor R[F² > 2σ(F²)] was 4.8 %, the weighted R factor wR(F²) was 11.6%, while the goodness-of-fit was 0.95. The minimum and the maximum electron-density peaks had values of -0.26 and 0.35 e Å^-3, respectively.

Computing details

Data collection: SMART (Bruker, 2000); cell refinement: SAINT (Bruker, 2000); data reduction: SAINT (Bruker, 2000); program(s) used to solve structure: SIR92 (Altomare et al., 1993); program(s) used to refine structure: MoPro (Jelsch et al., 2005); molecular graphics: Mercury (Macrae et al., 2006) and MoProViewer (Guillot, 2011); software used to prepare material for publication: pubCIF (Westrip, 2010).

Figures

	Fig. 1. The molecular structure of the title compound, showing the atom-numbering scheme for the non-H atoms. Displacement ellipsoids are drawn at the 50% probability level.
	Fig. 2. A view of molecular packing showing the role of the water molecules in bridging the molecular chains.
	Fig. 3. A Hirshfeld surface of the parent molecule showing the interacting molecules.
	Fig. 4. Fingerprint plots of the title compound showing the percentage of various interactions.
	Fig. 5. A three-dimensional electron-density isosurface valued 0.01 (e Å^-13) coloured according to the electrostatic potential.
	Fig. 6. The total dipole moment of the molecule computed on the basis of transferred parameters.

17α-Methyl-10β,17β-dihydroxyestr-4-en-3-one dihydrate

Crystal data

C19H28O3·2H2O	F(000) = 744
Mr = 340.44	Dx = 1.215 Mg m^−3
Orthorhombic, P212121	Mo Kα radiation, λ = 0.71073 Å
Hall symbol: P 2ac 2ab	Cell parameters from 507 reflections
a = 6.9028 (6) Å	θ = 2.0–28.3°
b = 11.0334 (10) Å	µ = 0.09 mm^−1
c = 24.451 (2) Å	T = 100 K
V = 1862.2 (3) Å^3	Block, colorless
Z = 4	0.27 × 0.10 × 0.07 mm

Data collection

Bruker SMART CCD detector diffractometer	2792 independent reflections
Radiation source: fine-focus sealed tube	2792 reflections with > 2.0σ(I)
Graphite monochromator	Rint = 0.052
ω and phi scan	θmax = 28.3°, θmin = 2.0°
Absorption correction: multi-scan (SADABS; Bruker, 2000)	h = 0→9
Tmin = 0.977, Tmax = 0.994	k = 0→14
13666 measured reflections	l = −31→32

Refinement

Refinement on F^2	Primary atom site location: structure-invariant direct methods
Least-squares matrix: full	Secondary atom site location: difference Fourier map
R[F^2 > 2σ(F^2)] = 0.048	Hydrogen site location: difference Fourier map
wR(F^2) = 0.116	H-atom parameters constrained
S = 0.95	w = 1/[σ^2(Fo^2) + (0.08P)^2 + 0.01P] where P = (Fo^2 + 2Fc^2)/3
2792 reflections	(Δ/σ)max = 0.029
217 parameters	Δρmax = 0.35 e Å^−3
0 restraints	Δρmin = −0.26 e Å^−3

Special details

Refinement. Refinement of F² against reflections. The threshold expression of F² > 2sigma(F²) is used for calculating R-factors(gt) and is not relevant to the choice of reflections for refinement. R-factors based on F² are statistically about twice as large as those based on F, and R-factors based on ALL data will be even larger.

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
O2	0.8602 (3)	0.00169 (19)	0.10105 (8)	0.0423 (3)
O3	0.5688 (3)	0.60762 (16)	0.06832 (8)	0.0372 (3)
C9	0.5046 (4)	0.4361 (2)	0.12707 (11)	0.0256 (4)
H9A	0.41191	0.40931	0.16156	0.04780
C14	0.7637 (4)	0.2990 (2)	0.16183 (11)	0.0294 (4)
H14A	0.66047	0.27174	0.19375	0.05083
C5	0.4790 (4)	0.6495 (3)	0.16041 (11)	0.0307 (4)
C8	0.7149 (4)	0.4297 (3)	0.14791 (11)	0.0284 (4)
H8A	0.81266	0.46121	0.11542	0.04817
C11	0.4658 (5)	0.3443 (3)	0.08087 (12)	0.0331 (4)
H11A	0.31178	0.34602	0.07072	0.05949
H11B	0.54685	0.37136	0.04450	0.05485
C10	0.4443 (4)	0.5670 (2)	0.11168 (11)	0.0281 (4)
C7	0.7392 (5)	0.5118 (3)	0.19804 (12)	0.0416 (5)
H7A	0.65417	0.47503	0.23184	0.06419
H7B	0.89122	0.51245	0.21040	0.06902
C13	0.7357 (4)	0.2086 (2)	0.11473 (10)	0.0264 (4)
C18	0.8680 (5)	0.2376 (3)	0.06591 (12)	0.0406 (5)
H18A	1.01710	0.23352	0.07871	0.06905
H18B	0.84316	0.17249	0.03384	0.07116
H18C	0.83579	0.32719	0.05098	0.06451
C12	0.5237 (4)	0.2144 (3)	0.09678 (13)	0.0342 (4)
H12A	0.43130	0.18307	0.13014	0.05931
H12B	0.50184	0.15422	0.06188	0.06311
O1	0.0599 (4)	0.8371 (2)	0.16287 (11)	0.0586 (4)
C17	0.8069 (5)	0.0905 (3)	0.14221 (12)	0.0335 (4)
C1	0.2322 (5)	0.5695 (3)	0.09338 (13)	0.0383 (4)
H1A	0.22059	0.53019	0.05251	0.06126
H1B	0.14759	0.51247	0.12100	0.06175
C15	0.9659 (5)	0.2692 (3)	0.18443 (13)	0.0415 (4)
H15A	0.97576	0.29139	0.22786	0.07100
H15B	1.07860	0.31790	0.16212	0.06733
C4	0.3475 (5)	0.7313 (3)	0.17698 (12)	0.0373 (5)
H4A	0.38181	0.78734	0.21204	0.06920
C6	0.6738 (5)	0.6412 (3)	0.18718 (13)	0.0398 (5)
H6A	0.66908	0.68972	0.22609	0.07267
H6B	0.78006	0.68447	0.16047	0.06481
C3	0.1726 (5)	0.7590 (3)	0.14602 (13)	0.0415 (5)
C2	0.1439 (5)	0.6959 (3)	0.09252 (13)	0.0422 (5)
H2B	0.21031	0.74961	0.06003	0.06146
H2C	−0.01102	0.68996	0.08388	0.07212
C20	0.6574 (6)	0.0294 (3)	0.17962 (15)	0.0538 (6)
H20A	0.61344	0.09206	0.21101	0.07347
H20B	0.53306	0.00390	0.15564	0.07587
H20C	0.72042	−0.05018	0.19795	0.07833
C16	0.9857 (5)	0.1317 (3)	0.17549 (14)	0.0492 (6)
H16A	0.99000	0.08430	0.21468	0.06984
H16B	1.11854	0.11132	0.15304	0.06804
O2W	0.2174 (4)	0.9836 (2)	0.02964 (10)	0.0615 (4)
O1W	0.5632 (4)	0.8548 (2)	0.04914 (11)	0.0612 (4)
H3W	0.65624	0.91064	0.06619	0.07568
H2W	0.16287	0.96002	−0.00594	0.08736
H1W	0.45287	0.90897	0.04218	0.07005
H4W	0.10077	0.99330	0.05231	0.07847
H2	0.91659	−0.06430	0.12264	0.06365
H3	0.55537	0.69538	0.06344	0.05217

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
O2	0.0494 (14)	0.0297 (10)	0.0477 (12)	0.0106 (10)	−0.0023 (11)	−0.0037 (10)
O3	0.0495 (13)	0.0286 (10)	0.0335 (10)	0.0054 (9)	0.0103 (10)	0.0027 (9)
C9	0.0251 (14)	0.0208 (13)	0.0310 (13)	0.0023 (11)	−0.0025 (12)	0.0023 (11)
H9A	0.04670	0.04902	0.04769	0.00082	0.01143	0.00489
C14	0.0318 (15)	0.0293 (15)	0.0271 (13)	0.0026 (12)	−0.0014 (13)	−0.0015 (12)
H14A	0.05833	0.05167	0.04248	0.00608	0.00709	0.00416
C5	0.0324 (15)	0.0295 (14)	0.0302 (14)	0.0019 (13)	0.0011 (13)	−0.0017 (12)
C8	0.0258 (15)	0.0297 (14)	0.0298 (13)	0.0016 (12)	−0.0043 (13)	−0.0050 (12)
H8A	0.04245	0.04858	0.05347	−0.00291	0.00583	0.00180
C11	0.0363 (17)	0.0274 (14)	0.0354 (15)	0.0027 (13)	−0.0138 (14)	−0.0047 (12)
H11A	0.04853	0.05005	0.07988	0.00065	−0.01864	−0.00748
H11B	0.07035	0.05036	0.04386	0.00319	0.00206	−0.00255
C10	0.0280 (15)	0.0282 (14)	0.0280 (13)	0.0042 (13)	−0.0018 (12)	−0.0013 (12)
C7	0.047 (2)	0.0402 (17)	0.0380 (16)	0.0112 (16)	−0.0182 (16)	−0.0155 (14)
H7A	0.08819	0.05911	0.04526	0.01453	−0.00217	−0.00440
H7B	0.06547	0.05982	0.08177	0.01011	−0.03232	−0.01591
C13	0.0279 (15)	0.0260 (14)	0.0254 (13)	0.0023 (12)	−0.0006 (12)	−0.0006 (11)
C18	0.0470 (19)	0.0396 (17)	0.0351 (15)	0.0078 (15)	0.0107 (15)	0.0018 (14)
H18A	0.05183	0.08156	0.07376	0.01050	0.01130	0.00585
H18B	0.09546	0.06445	0.05357	0.00781	0.01251	−0.01428
H18C	0.07743	0.05273	0.06337	0.01619	0.02049	0.00983
C12	0.0330 (16)	0.0255 (14)	0.0441 (17)	−0.0008 (13)	−0.0098 (14)	−0.0035 (13)
H12A	0.05198	0.05114	0.07481	−0.00202	0.00570	0.00851
H12B	0.07347	0.04768	0.06819	0.00511	−0.01940	−0.01524
O1	0.0635 (16)	0.0445 (13)	0.0679 (15)	0.0312 (12)	0.0027 (14)	−0.0031 (12)
C17	0.0392 (17)	0.0268 (14)	0.0345 (14)	0.0059 (13)	0.0000 (14)	0.0005 (12)
C1	0.0348 (17)	0.0327 (16)	0.0475 (17)	0.0057 (15)	−0.0096 (16)	−0.0017 (14)
H1A	0.06418	0.05552	0.06409	0.00766	−0.02065	−0.01060
H1B	0.04541	0.05408	0.08575	0.00055	0.00393	0.00879
C15	0.0399 (17)	0.0381 (16)	0.0465 (17)	0.0102 (15)	−0.0190 (16)	−0.0064 (15)
H15A	0.08518	0.06620	0.06163	0.02104	−0.02708	−0.01164
H15B	0.04933	0.05954	0.09310	0.00426	−0.00746	0.00014
C4	0.0448 (19)	0.0311 (15)	0.0359 (16)	0.0092 (15)	−0.0003 (15)	−0.0048 (13)
H4A	0.09336	0.06189	0.05236	0.02363	−0.00845	−0.01741
C6	0.0430 (18)	0.0276 (14)	0.0488 (17)	0.0035 (14)	−0.0079 (16)	−0.0134 (14)
H6A	0.09630	0.05963	0.06208	0.01900	−0.02904	−0.02275
H6B	0.05807	0.05270	0.08365	−0.00254	−0.01302	−0.00184
C3	0.0444 (18)	0.0332 (16)	0.0468 (18)	0.0137 (15)	0.0036 (16)	0.0044 (15)
C2	0.0424 (19)	0.0380 (17)	0.0463 (19)	0.0111 (15)	−0.0049 (16)	0.0007 (15)
H2B	0.07258	0.04982	0.06200	0.00514	−0.00692	0.00162
H2C	0.04959	0.06202	0.10476	0.00974	−0.01822	−0.00654
C20	0.069 (2)	0.0359 (18)	0.057 (2)	0.0068 (17)	0.016 (2)	0.0145 (16)
H20A	0.09266	0.05593	0.07182	0.00811	0.02421	0.00423
H20B	0.06954	0.06561	0.09246	−0.00867	−0.00135	0.01070
H20C	0.08983	0.05419	0.09097	0.01843	0.01026	0.02473
C16	0.056 (2)	0.0384 (17)	0.0535 (19)	0.0143 (16)	−0.0180 (18)	0.0025 (16)
H16A	0.08864	0.06098	0.05989	0.01685	−0.01541	0.00370
H16B	0.05549	0.06438	0.08425	0.01446	−0.00074	−0.00796
O2W	0.0669 (18)	0.0622 (16)	0.0553 (14)	0.0038 (14)	−0.0199 (13)	−0.0031 (13)
O1W	0.0709 (17)	0.0379 (12)	0.0749 (16)	−0.0040 (13)	−0.0093 (15)	0.0007 (12)
H3W	0.07694	0.04788	0.10221	−0.00752	0.00251	0.00962
H2W	0.12730	0.06475	0.07003	0.02179	−0.02287	0.00288
H1W	0.08313	0.04872	0.07831	0.00997	−0.00116	0.01020
H4W	0.09507	0.05961	0.08073	0.01899	−0.01069	−0.00002
H2	0.07789	0.04661	0.06645	0.01823	−0.00426	−0.00233
H3	0.06254	0.04085	0.05312	0.00376	0.00532	0.00679

Geometric parameters (Å, º)

O2—C17	1.4520 (12)	C18—H18B	1.0770
O2—H2	0.9800	C12—H12B	1.0920
O3—C10	1.4364 (12)	C12—H12A	1.0920
O3—H3	0.9800	O1—C3	1.2325 (15)
C9—C10	1.5493 (12)	C17—C16	1.5462 (15)
C9—C8	1.5400 (12)	C17—C20	1.5354 (15)
C9—C11	1.5407 (12)	C1—C2	1.5226 (14)
C9—H9A	1.0990	C1—H1A	1.0920
C14—C13	1.5360 (12)	C1—H1B	1.0920
C14—C8	1.5192 (13)	C15—C16	1.5387 (15)
C14—C15	1.5364 (14)	C15—H15A	1.0920
C14—H14A	1.0990	C15—H15B	1.0920
C5—C10	1.5181 (13)	C4—C3	1.4572 (16)
C5—C6	1.4979 (14)	C4—H4A	1.0830
C5—C4	1.3433 (15)	C6—H6A	1.0920
C8—C7	1.5335 (13)	C6—H6B	1.0920
C8—H8A	1.0990	C3—C2	1.4950 (15)
C11—C12	1.5380 (13)	C2—H2B	1.0920
C11—H11A	1.0920	C2—H2C	1.0920
C11—H11B	1.0920	C20—H20A	1.0770
C10—C1	1.5314 (14)	C20—H20C	1.0770
C7—C6	1.5210 (14)	C20—H20B	1.0770
C7—H7A	1.0920	C16—H16A	1.0920
C7—H7B	1.0920	C16—H16B	1.0920
C13—C17	1.5459 (12)	O2W—H4W	0.9830
C13—C12	1.5289 (14)	O2W—H2W	0.9830
C13—C18	1.5366 (13)	O1W—H1W	0.9830
C18—H18A	1.0770	O1W—H3W	0.9830
C18—H18C	1.0770
C17—O2—H2	103.235	C17—C13—C12	117.7 (10)
C10—O3—H3	109.961	C17—C13—C18	108.9 (9)
C10—C9—C8	112.1 (9)	C12—C13—C18	109.7 (10)
C10—C9—C11	112.8 (9)	H18A—C18—H18C	109.504
C10—C9—H9A	106.331	H18A—C18—H18B	109.624
C8—C9—C11	112.1 (9)	H18C—C18—H18B	109.438
C8—C9—H9A	106.400	H12B—C12—H12A	108.064
C11—C9—H9A	106.529	C16—C17—C20	111 (1)
C13—C14—C8	114.9 (9)	C2—C1—H1A	108.756
C13—C14—C15	104.2 (9)	C2—C1—H1B	108.792
C13—C14—H14A	105.832	H1A—C1—H1B	107.362
C8—C14—C15	119.1 (10)	C16—C15—H15A	110.737
C8—C14—H14A	105.953	C16—C15—H15B	110.551
C15—C14—H14A	105.958	H15A—C15—H15B	109.310
C10—C5—C6	116.6 (10)	C3—C4—C5	123 (1)
C10—C5—C4	122 (1)	C3—C4—H4A	118.236
C6—C5—C4	121 (1)	H6A—C6—H6B	109.074
C7—C8—H8A	108.869	C2—C3—C4	118 (1)
C12—C11—H11A	109.041	H2B—C2—H2C	107.637
C12—C11—H11B	109.094	H20A—C20—H20C	109.926
H11A—C11—H11B	108.011	H20A—C20—H20B	109.307
C6—C7—H7A	108.739	H20C—C20—H20B	109.621
C6—C7—H7B	109.115	H16A—C16—H16B	108.643
H7A—C7—H7B	108.026	H4W—O2W—H2W	102.395
C17—C13—C14	100.4 (8)	H1W—O1W—H3W	101.439
O2—C17—C13—C14	−159.4 (5)	H8A—C8—C7—H7B	−56.03
O2—C17—C13—C12	83.8 (2)	C11—C9—C10—C1	54.5 (2)
O2—C17—C13—C18	−41.8 (2)	C11—C9—C8—C7	−173.4 (2)
O2—C17—C16—C15	139.8 (4)	C11—C12—C13—C17	168.8 (2)
O2—C17—C16—H16A	−100.15	C11—C12—C13—C18	−66.0 (2)
O2—C17—C16—H16B	19.66	H11A—C11—C9—C10	−58.29
O2—C17—C20—H20A	179.95	H11A—C11—C12—C13	−176.56
O2—C17—C20—H20C	59.39	H11A—C11—C12—H12B	62.55
O2—C17—C20—H20B	−60.58	H11A—C11—C12—H12A	−55.70
O3—C10—C9—C8	61.4 (2)	H11B—C11—C9—C10	59.47
O3—C10—C9—C11	−66.3 (2)	H11B—C11—C12—C13	65.69
O3—C10—C9—H9A	177.29	H11B—C11—C12—H12B	−55.20
O3—C10—C5—C6	−67.3 (2)	H11B—C11—C12—H12A	−173.45
O3—C10—C5—C4	108.7 (2)	C10—C9—C8—C7	58.5 (2)
O3—C10—C1—C2	−76.7 (2)	C10—C9—C11—C12	−179.3 (2)
O3—C10—C1—H1A	45.09	C10—C5—C6—C7	−47.1 (2)
O3—C10—C1—H1B	161.98	C10—C5—C6—H6A	−168.08
C9—C10—O3—H3	−166.84	C10—C5—C6—H6B	73.38
C9—C10—C5—C6	49.4 (2)	C10—C5—C4—C3	−8.00 (19)
C9—C10—C5—C4	−134.6 (3)	C10—C5—C4—H4A	−179.92
C9—C10—C1—C2	164.1 (2)	C10—C1—C2—C3	−52.5 (2)
C9—C10—C1—H1A	−74.19	C10—C1—C2—H2B	68.47
C9—C10—C1—H1B	42.70	C10—C1—C2—H2C	−173.32
C9—C8—C14—C13	56.4 (2)	C7—C8—C14—C13	177.6 (2)
C9—C8—C14—C15	−179.1 (2)	C7—C8—C14—C15	−57.9 (2)
C9—C8—C14—H14A	−60.03	C7—C6—C5—C4	136.9 (4)
C9—C8—C7—C6	−53.9 (2)	H7A—C7—C6—H6A	48.50
C9—C8—C7—H7A	66.79	H7A—C7—C6—H6B	166.79
C9—C8—C7—H7B	−175.31	H7B—C7—C6—H6A	−69.11
C9—C11—C12—C13	−55.6 (2)	H7B—C7—C6—H6B	49.18
C9—C11—C12—H12B	−176.51	C13—C17—O2—H2	173.56
C9—C11—C12—H12A	65.24	C13—C17—C16—C15	21.2 (4)
H9A—C9—C10—C5	60.97	C13—C17—C16—H16A	141.31
H9A—C9—C10—C1	−61.91	C13—C17—C16—H16B	−98.88
H9A—C9—C8—C14	64.47	C13—C17—C20—H20A	−57.64
H9A—C9—C8—C7	−57.37	C13—C17—C20—H20C	−178.20
H9A—C9—C8—H8A	−176.59	C13—C17—C20—H20B	61.83
H9A—C9—C11—C12	−62.98	C13—C14—C15—C16	−32.4 (3)
H9A—C9—C11—H11A	58.02	C13—C14—C15—H15A	−151.46
H9A—C9—C11—H11B	175.77	C13—C14—C15—H15B	86.60
C14—C13—C17—C16	−40.3 (3)	C18—C13—C17—C16	77.3 (2)
C14—C13—C17—C20	80.2 (2)	C18—C13—C17—C20	−162.2 (2)
C14—C13—C12—C11	56.1 (2)	C18—C13—C14—C15	−70.1 (2)
C14—C13—C12—H12B	177.12	C18—C13—C12—H12B	55.00
C14—C13—C12—H12A	−64.75	C18—C13—C12—H12A	173.13
C14—C13—C18—H18A	60.14	H18A—C18—C13—C17	−49.98
C14—C13—C18—H18C	−59.84	H18A—C18—C13—C12	179.91
C14—C13—C18—H18B	−179.71	H18B—C18—C13—C17	70.17
C14—C8—C9—C10	−179.7 (2)	H18B—C18—C13—C12	−59.94
C14—C8—C9—C11	−51.6 (2)	H18C—C18—C13—C17	−169.96
C14—C8—C7—C6	−174.7 (2)	H18C—C18—C13—C12	59.92
C14—C8—C7—H7A	−53.95	C12—C13—C17—C16	−157.2 (5)
C14—C8—C7—H7B	63.95	C12—C13—C17—C20	−36.6 (3)
C14—C15—C16—C17	6.6 (2)	C12—C13—C14—C15	169.2 (2)
C14—C15—C16—H16A	−113.24	H12A—C12—C13—C17	47.93
C14—C15—C16—H16B	126.57	H12B—C12—C13—C17	−70.20
H14A—C14—C13—C17	−66.16	O1—C3—C2—C1	−152.1 (5)
H14A—C14—C13—C12	57.67	O1—C3—C2—H2B	86.61
H14A—C14—C13—C18	178.44	O1—C3—C2—H2C	−30.76
H14A—C14—C8—C7	61.18	O1—C3—C4—H4A	−6.40
H14A—C14—C8—H8A	−178.94	C17—C13—C14—C15	45.3 (3)
H14A—C14—C15—C16	79.03	C17—C16—C15—H15A	125.86
H14A—C14—C15—H15A	−40.05	C17—C16—C15—H15B	−112.84
H14A—C14—C15—H15B	−161.99	C1—C10—O3—H3	72.62
C5—C10—O3—H3	−49.44	C1—C10—C5—C6	171.3 (2)
C5—C10—C9—C8	−54.9 (2)	C1—C10—C5—C4	−12.7 (2)
C5—C10—C9—C11	177.4 (2)	C1—C2—C3—C4	32.0 (3)
C5—C10—C1—C2	42.8 (3)	H1A—C1—C2—C3	−174.53
C5—C10—C1—H1A	164.55	H1A—C1—C2—H2B	−53.58
C5—C10—C1—H1B	−78.56	H1A—C1—C2—H2C	64.62
C5—C6—C7—C8	48.2 (2)	H1B—C1—C2—C3	68.83
C5—C6—C7—H7A	−72.75	H1B—C1—C2—H2B	−170.23
C5—C6—C7—H7B	169.65	H1B—C1—C2—H2C	−52.02
C5—C4—C3—O1	−178.33 (16)	C15—C16—C17—C20	−102.0 (3)
C5—C4—C3—C2	−2.3 (2)	H15A—C15—C16—H16A	6.03
C8—C14—C13—C17	177.3 (2)	H15A—C15—C16—H16B	−114.16
C8—C14—C13—C12	−58.8 (2)	H15B—C15—C16—H16A	127.33
C8—C14—C13—C18	61.9 (2)	H15B—C15—C16—H16B	7.14
C8—C14—C15—C16	−161.9 (2)	C4—C3—C2—H2B	−89.33
C8—C14—C15—H15A	79.02	C4—C3—C2—H2C	153.30
C8—C14—C15—H15B	−42.92	C4—C5—C6—H6A	15.88
C8—C9—C10—C1	−177.8 (2)	C4—C5—C6—H6B	−102.65
C8—C9—C11—C12	53.0 (2)	H4A—C4—C3—C2	169.65
C8—C9—C11—H11A	174.00	H4A—C4—C5—C6	−4.11
C8—C9—C11—H11B	−68.24	C6—C5—C4—C3	167.8 (2)
C8—C7—C6—H6A	169.42	C20—C17—O2—H2	−61.44
C8—C7—C6—H6B	−72.29	C20—C17—C16—H16A	18.10
H8A—C8—C14—C13	−62.51	C20—C17—C16—H16B	137.91
H8A—C8—C14—C15	61.98	H20A—C20—C17—C16	58.97
H8A—C8—C9—C10	−60.73	H20B—C20—C17—C16	178.44
H8A—C8—C9—C11	67.35	H20C—C20—C17—C16	−61.59
H8A—C8—C7—C6	65.34	C16—C17—O2—H2	59.18
H8A—C8—C7—H7A	−173.92

Hydrogen-bond geometry (Å, º)

D—H···A	D—H	H···A	D···A	D—H···A
O2W—H4W···O2^i	0.98	2.05	3.0279 (7)	176
O2W—H2W···O3^ii	0.98	1.82	2.7933 (6)	171
C14—H14A···C20	1.10	2.70	3.0951 (7)	101
C11—H11A···C1	1.09	2.59	2.9771 (8)	100
C18—H18A···C16	1.08	2.63	3.0340 (8)	102
C12—H12A···C20	1.09	2.60	3.0209 (8)	102
C2—H2B···O2W	1.09	2.69	3.5634 (7)	137
O1W—H1W···O2W	0.98	1.85	2.8189 (7)	169
O3—H3···O1W	0.98	1.79	2.7672 (6)	172
O1W—H3W···O2^iii	0.98	1.93	2.9053 (6)	172
O2—H2···O1^iv	0.98	1.77	2.7352 (6)	168
C6—H6B···O1^v	1.09	2.56	3.4828 (7)	141

Symmetry codes: (i) x−1, y+1, z; (ii) x−1/2, −y+3/2, −z; (iii) x, y+1, z; (iv) x+1, y−1, z; (v) x+1, y, z.