Synthesis , X-ray Crystal Structure and Theoretical Calculations of Antileishmanial Neolignan Analogues

A síntese e a estrutura cristalina por difração de raios-X de dois análogos de neolignanas, 2-(4-clorofenil)-1-feniletanona (20) e 2-[tio(4-clorofenil)]-1-(3,4-dimetoxifenil)propan-1-ona (12) são descritas. O composto 12 apresenta atividade intracelular contra Leishmania donovani e Leishmania amazonensis de amastigotas que causam a leishmaniose tegumentar e visceral. Além disso, a teoria do funcional de densidade (DFT) com o funcional híbrido B3LYP foi empregado para calcular um conjunto de descritores moleculares para dezenove análogos sintéticos de neolignanas com atividades antileishmaniose. Posteriormente, a análise discriminante stepwise foi realizada para investigar possíveis relações entre a estrutura molecular e atividades biológicas. Por meio dessa análise os compostos foram classificados em dois grupos ativos e inativos de acordo com seu grau de atividade biológica, e as propriedades mais importantes foram as cargas de alguns átomos, a afinidade eletrônica e o ClogP.


Introduction
The leishmaniases are parasitic diseases caused by protozoa of the genus Leishmania and remain a severe public health problem, particularly in many tropical and subtropical regions.The infection is transmitted by bite from female sandflies, which are of the genus Phlebotomus or Lutzomyin.6][7][8][9] Unfortunately non-availability of satisfactory chemotherapeutic agents and failure to develop an effective vaccine are considered to be two stumbling blocks in the combat of this disease. 10reatment approaches and responses to chemotherapy vary by regions.2][13] Therefore, there remains an urgent need for development of less toxic drugs that are effective against all forms of leishmaniases.][16] Neolignans are groups of compounds that show a wide range of biological effects including antifungal, [17][18][19] anti-schistosomal, [20][21][22][23] antiplasmodial, 24 trypanocidal, [25][26][27] antibacterial, 28 anti-PAF, 29,30 antipsychotic, 31 antioxidant, 32,33 activities, and biological activity against Escherichia coli, 34,35 Paracoccidioides brasiliensis. 36Usually, neolignans are organic dimmers derived from oxidative coupling of allyl and propenyl phenols. 16,22Previous studies have evaluated the antileishmanials activities of twenty-two sulfur and oxygen synthetic analogues of neolignans against parasite species that cause cutaneous and visceral Leishmaniasis. 14,15hese compounds were synthesized and their activities against both intracellular amastigotes of L. donovani and L. amazonensis were compared.

Experimental
Instruments 1 H NMR (100 MHz) spectrum was recorded on a Varian XL-100 spectrometer.Chemical shifts were reported in ppm from tetramethylsilane on the d scale and coupling constants J are expressed in Hz.IR spectra was recorded in KBr film and measured with a Bomen model MB series II spectrophotometer.Electrothermal melting point apparatus are uncorrected.

General
Within the great structural variety of neolignans the β-ketoether, β-ketosulfide, β-ketosulfoxide and  β-ketosulfone derivatives are closely related to natural 8,4'-oxyneolignans, which are of interest because of their moderate antileishmaniasis activity against both intracellular amastigotes of L. donovani and L. amazonensis. 14Insight into the biological and physicochemical functions of complex neolignans at the molecular level requires a precise understanding of their three-dimensional structures.Figure 3 shows the basic skeletons of the twenty compounds studied and Table 1 lists their classes and biological response.Two of these neolignan derivatives, β-ketosulfide 12 and β-ketoether 20 were synthesized from condensation reactions among α-bromoketone and thiophenol or phenol derivatives, respectively, with their structures elucidated by X-ray crystal diffraction.The synthesis of 12 was mentioned previously, 15 but the crystal data and structural features were not published.Here, we report the synthesis and the X-ray crystallographic studies of 12 and 20 with a comparison of their three-dimensional structures.

Crystallographic data
Suitable colorless single crystals of the neolignan derivatives 12 and 20, with approximate dimensions 0.05×0.05×0.10 and 0.05×0.10×0.10mm, respectively, were selected and the diffraction data were collected using an Enraf-Nonius CAD4 diffractometer with graphite monochromated K α Mo radiation (l = 0.71073 Å), in the w-2q scan mode, at room temperature.The unit cell parameters were determined using 25 automatically centered reflections.Intensities of the reflections were corrected by absorption factors [m (MoK α ) = 0.308 and 0.367, for compounds 12 and 20, respectively] using the PSISCAN method. 46Information concerning to the crystallographic data collection and refinement of the structures are given in Table 2.The structures were solved by SIR-92 47 and refined by full matrix least squares and difference Fourier synthesis by SHELXL-97, 48 using the WinGX software package. 49All non-hydrogen atoms were refined anisotropically.The hydrogen atoms were located in their ideal positions and not refined.The structural analysis was performed by PLATON system. 50The graphic representations of the molecules were made using ORTEP3 for Windows. 51The crystal data are deposited at Cambridge Crystallographic Data Centre, CCDC 703864 and 703867, for compound 12 and 20, respectively.

Computational methods
The molecular geometries obtained from X-ray diffraction data of 12 and 20 were used here as a starting point for our calculations of the remaining molecules, just changing the substituent.All compounds were optimized using the B3LYP hybrid functional, 52,53 together with the 6-311++G(d,p) basis sets in the Gaussian 03 molecular package. 54Vibrational analysis was carried out for the complete equilibrium geometry obtained by the procedure in the Gaussian 03 package at the DFT level with the B3LYP/6-311++G(d,p) level in the gas phase, ensuring that each gradient optimization located was indeed a true minimum energy structure (no imaginary frequencies).In addition, the conformational analysis were carried out to confirm the minimum energy structure to molecules 12 and 20, by carrying out a series of partial optimizations constraining the concerned dihedral angle step by step within the appropriate range, with a step size of 5°, these calculations were carried out using the B3LYP/6-31G* basis set, the dihedral angles analyzed were C1-C7-C8-O and C4'-O-C8-C7 for molecule 12 and C1-C7-C8-S and C4'-S-C8-C7 for molecule 20.Previous studies about conformational preferences in solid state (crystal) and in solution has been published [55][56][57] for a large number of compounds analogues of the compounds 12 and 20, studied in the present paper.The geometrical structures of the radicals studied were optimized independently from the neutral molecules prior to the calculations of energies, treated as open shell systems DFT/UB3LYP.Thus, the more relevance electronic descriptors to antileishmaniasis activities were calculated such as: eletrostatic potential atomic charges (Q N -net atomic charge on atom N), occupied and unoccupied molecular orbital energies (ε HOMO and ε LUMO ), electronegativity (c), hardness (η), softness (1/η), chemist potential (µ), electrophilic index (w) and electronic affinity (EA) were calculated with the DFT/ B3LYP level, together with the 6-311++G(d,p) basis sets.
The c was calculated as the sum mean between the energies of HOMO and LUMO (c = -(ε HOMO + ε LUMO )/2). 58he η is simply the energy difference between LUMO and HOMO energies, while the 1/η is the inverse of the hardness. 59The µ is simply Koopmans' approximation. 60rom equation w = µ 2 /2η we obtained the electrophilic index. 58The IP was calculated as the energy differences between a radical cation (Ec) and the respective neutral molecule (En); IP = Ec -En. 61The volume (Vol), molecular refractivity (MR), polarizability (Pol), partition coefficient (ClogP) and hydration energy (HE) were obtained by using the Hyperchem 7.5 molecular package. 62The transport of a compound through membranes can be modeled by molecular hydrophobicity, which can be described by octanol/water partition coefficients (ClogP).The value of this property was obtained by using the Chem-3D molecular package. 63Table 3 shows almost molecular descriptors.
The atomic charges were obtained by employing the electrostatic potential method, which was used because the charges derived from the electrostatic potential method are physically more satisfactory than Mulliken's charges, especially when related to biological activity.The choice of the best descriptors to correlate with the biological activities was performed using stepwise discriminant analysis (SDA) built in the Minitab 14 statistical software. 64The molecular electrostatic potential (MEP) surface was generated using the geometry optimized in B3LYP/6-311++G(d,p) and an isodensity surface of 0.002 a.u.On the MEP surface, regions indicating the excess of negative potential correspond to excess negative charges, i.e., attraction of the positively charged probe.The MEP surface was calculated and analyzed visually using the PC Spartan PRO molecular package. 65

Synthesis
The molecules 12 and 20 (Figure 4) were synthesized through procedures described in the literature 14,15 using the condensation reaction of α-bromoketones with a phenol or thiophenol derivatives, in basic medium.Compound 12 was synthesized from the condensation reaction of α-bromo-3,4-dimethoxypropiophenone and 4-chlorothiophenol in anhydrous ethyl methyl ketone, in 81.7% yield.The spectroscopic data of compound 12 is in concordance with the literature. 14,15Compound 20 was synthesized in the same conditions, from reaction of α-bromoacetophenone and 4-chlorophenol, in 95% yield.Spectroscopic data and physical properties of compound 20 are described for the first time in this paper.
As shown in the crystal representation (see Figures 1 and  2), both compounds have a Cl atom in position 4′ (ring B), whereas compound 12 has also an -OCH 3 group in positions 3 and 4 (ring A).The bond distances for the single bonds C-O are 1.418(4) and 1.369(4) Å for compound 20, and the C-S, as expected, are 1.838(3) and 1.776(3) Å for 12.No significant differences were found in the bond distances and angles of the two molecules, the others bond lengths C=O, C-Cl and C-C are in the expected ranges.The distances and angles parameters are listed in Table 4 and 5, respectively.The most relevant structural difference is that the structure of 20 is almost planar, with the dihedral angle between the two benzene planes of 2.8(1)°, while the structure of 12 is itself twisted out of the plane of the aromatic rings, as can be seen from the C1-C2-S-C9 torsion angle of 123.3(5)°.The dihedral angle between the rings A and B is 34.4(1)°, which presumably reflects some flexibility of the molecule, enabling it to rotate and deform in order to minimize any unfavorable intramolecular interactions due to the presence of the methyl group.Also, in compound 12 the keto atom is more coplanar in relation to the ring A. The keto atom O1 is 0.254(3) Å out of the least-squares plane of the ring A. In compound 20 the same keto atom is 0.464(2) Å out of the plane A. The three-dimensional structure of the compound 12 and 20 and the non-covalent interactions formed in their crystal structure along the b-axis are shown in Figure 5 and 6.
In the crystal packing of 12, the oxygen from the methoxy group participates in two non conventional C-H … O hydrogen bonds.The molecules of 12 are joined via intermolecular hydrogen bonds (C10 … O3 3.491(4) Å, H10-O3 2.57 Å, 170.0° and C16 … O2 3.323(5) Å, H16C-O2 2.60 Å, 132.0°) giving rise to the formation of the zig-zag supramolecular chain, as shown in the Figure 5.The crystal packing of 20 also shows the existence of an intermolecular C-H … O hydrogen bond (C2 … O1 3.424(5) Å, H … O1 2.56 Å, C2-H2B … O1 148°).As can be observed in Figure 6, this hydrogen bond are responsible for the self-assembly into one-dimensional chain, formed by the interaction at the end of the neighboring molecules while in the 12 it was formed in the middle of adjacent molecules.

Theoretical calculations
The distance lengths and angles for both X-ray and B3LYP/6-311G++(d,p) optimized structures of 12 and 20 presented in Tables 4 and 5 are normal for this type of compound.The values of C-S and C-O bond distances calculated for compounds 12 and 20 are 1.875 and 1.795 Å, and 1.418 and 1.369 Å, respectively.The C1-C2-S-C9 torsion angle experimentally observed are 123.3(5)ºand 34.4(1)º, while the values calculated were of 123.3 o and 35.7 o for compounds 12 and 20, respectively.These results are in agreement with the experimental values.No significant differences were found between these structures (12 and 20) and the remaining molecules.The harmonic frequency showed that the molecular geometrics obtained by DFT calculations correspond to a local minimum energy structure, no imaginary frequencies.In addition, Table S1 (supplementary material) presents the results to conformational analysis carried out to molecules 12 and 20.These results indicate that the structures optimized with the 6-311++G(d,p) basis sets using X-ray diffraction data as starting points have the lowest energy, indicating global Before applying the SAR analysis to the nineteen compounds under study, each calculated property (variables or descriptors) was auto scaled.In the auto scaling method, each variable is scaled to a mean of zero and a variance of unity.This method is very important because each variable is weighted equally and this provides a measure of the ability of a descriptor to discriminate classes of compounds.With this method, we can compare all variables at the same level although presenting different units.
SAR analysis has been carried out using molecular descriptors that were selected by stepwise discriminant analysis (SDA).The main objective of SDA is to determine discriminant functions using the measured variables that separate the groups as distinctly as possible.In this study, we considered two groups: active and inactive molecules against L. donovani and L. amazonensis activities.The SDA linear function is based on the Fisher's test (F test) for the significance of the variables.In each step one variable is selected based on its significance and after several steps, the more significant variables are extracted from the whole data set under investigation.
From the twenty-two synthetic analogues of neolignans reported by Aveniente et al. 14 three of them were not included in the analysis, because they not have experimental values (values reported as "nd" not determined).Therefore, the equations presented in this study were constructed with nineteen moleculescompounds 1-19 -that present antileishmaniasis activities (values expressed as percent inhibition of parasite growth at 80 µg mL -1 concentration).The SDA results will be used for molecule 20 to verify if this new molecule would be active or inactive.The allocation rule derived from the SDA results, when the antileishmanials activity of a new neolignan compound is investigated, is: (i) initially one calculates, for the new neolignan compound, the values for the more important descriptors obtained with the SDA; (ii) substitute these values in the two discriminant functions obtained in this study; (iii) check out which discriminant function (active or inactive compounds) presents the higher value.The new neolignan compound is active if the higher value is related to the active discriminant function and vice versa.

SDA for the Leishmania donovani activity
In this analysis, we considered two groups: active molecules (1, 2, 4-6, 8-12 and 16) and inactive molecules (3, 7, 13-15, 17-19) against L. donovani.The SDA indicated that the descriptors: electronic affinity (AE), charges on C6, O or S and C5' atoms were the most important in order to get the separation of active and inactive compounds.The discriminant functions for L. donovani activity obtained with nineteen compounds are given in equation 1 and 2: Inactive compounds = -1.01+ 1.17 Through the discriminant functions above (equation 1 and 2) and the values of each variable for the compounds studied (Table 3); we obtain the classification matrix by using all compounds in the analysis (Table 6).The SDA allowed correct classification scores of 100% (active compounds), 87.5% (inactive compounds) and 94.7% (total), resulting in a better performance in the separation of the two groups (Table 6).Derivative 18 was incorrectly classified in the group of active compounds.Probably this occurred because this compound showed a sulfur bond in position 8 that increases activity if compared to compounds bearing oxygen bond, like previous results shown by Aveniente et al. 14 Furthermore, the molecule 18 shows a meta chlorine substituents in the B ring, a methyl substituents in position 7, and a meta methoxyl substituents in the A ring, similar to molecule 13 which is active.
In accordance with equation 1 and Table 3, in general, active neolignans against L. donovani have more positive charges on atoms 6 and 5' and less positive charges on heteroatom.The charge is electronic descriptor; therefore, we can conclude that electronic effects have a very important role when one is trying to understand the activity of neolignan derivatives.The charges on the atoms S or O in compounds 14 and 15 have more positive values, because the inductive effect on the B ring, which have a Cl substituent in the molecule 14 and a methyl substituent in the molecule 15, in general, molecules with more positive values for S or O charge are inactives.In Figure 7 is shown the box plot for Q6 (see supplementary material Figure S2 for the Q5' charges).
In order to verify if a new molecule would be active or inactive against L. donovani, we had to apply the results obtained with the discriminant functions for the compound 20.From the results obtained, we can see that this molecule was classified as active, and from it we can conclude that the model obtained with SDA can be applied to new neolignan compounds whose biological activity is unknown.
Recently, we successfully used three-dimensional MEP surfaces to define the most probable sites of protonation of dipyridamole 66 , aparisthman, 67 cordatin, 8-epicordatin; 68 on the study of the molecular mechanisms the Diels-Alder reaction 69 and to get some clues about the transition state of the catalyzed reaction. 40The MEP surfaces of compound 1 (active) and compound 14 (inactive) in terms of total electron density show that the lowest electronic potential is in the proximity of oxygen atoms of the carbonyl (O1), and oxygen heteroatom (Figure 8).The large negative potential of oxygen atoms may be regarded to a nucleophilic suction pump, acting as a possible magnet for electrophilic attack of a biological receptor.The surfaces of 1 and 14 are different and compound 1 (active) provides a much more intense region of negative electrostatic potential than compound 14 (inactive).

SDA for the Leishmania amazonensis activity
For this activity the molecules were also classified in two groups: active molecules (1-8, 10-13, 15-18) and inactive molecules (9, 14 and 19).The most significant descriptors selected by SDA were obtained with nineteen compounds and is given in equation 3 and equation 4: Inactive compounds = -6.21-3.40 ClogP + 2.87 Q1 + 5.25 Q2 -4.20 Q1' Active compounds = -0.22+ 0.64 ClogP -0.54 Q1 -0.98 Q2 + 0.79 Q1' (4)    In such case, the four descriptors (ClogP and charges on 1, 2 and 1'substituent in the carbon atoms) represent the strength of a molecular association by electronic interaction.By using the quantities given in the discriminant functions above, we can obtain the classification summary showed in Table 7.The classification error rate was zero, resulting in a satisfactory separation of the two groups.
In accordance with equations 3 and 4 and Table 3, we can observe that, in general, molecules with more negative charge on C1 and C2 atoms are actives, while more positive charge on C1' atom are inactives.In addition, Figure 9 shows that, in general, molecules with more positive values for Q1' charge are actives, and the molecules with more negative values for Q1' charge are inactives.The ClogP is a measure of hydrophobicity; molecules with large value of ClogP have higher hydrophobicity and consequently better transport through cell membranes.In general, we can observe that molecules with high ClogP values are active, while the molecules with low values of ClogP are inactive (see supplementary material Figure S3).In other words, the hydrophobic character of molecules improves to L. amazonensis activity, what can indicate that the active compounds must interact with a target system such as an enzyme or receptor where the binding site is usually hydrophobic.
In order to verify if a new molecule would be active or inactive against L. amazonensis, we need to apply the results obtained with discriminant functions for the molecule 20.From the results obtained in this work we can conclude that the models attained here with SDA can be applied to new neolignans compounds whose biological activity is unknown.
In Figure 10 we can observe that compound 2 (active) provides a much more intense region of negative electrostatic potential than compound 9 (inactive), thus compound 2 has a more attractive cation-binding site.In general, we observed that active compounds have more intense region of negative electrostatic potential than inactive ones.

Conclusions
We have carried out a synthesis and X-ray crystallographic investigation of the molecular structures of compounds 12 and 20.No significant differences were found in the bond distances and angles of the two molecules.The most relevant structural difference is that the structure of 20 is almost planar, while the structure of 12 is itself twisted out of the plane of the aromatic rings.The results show that the agreement between theory and X-ray diffraction data is excellent.In fact the structure geometrics used in theoretical studies are more stable conformers; suggest that calculated structures for molecules 12 and 20 are the global minimums.In addition, the SDA method is quite efficient to classify the nineteen neolignans studied here in two groups, actives and inactives, according to their antileishmanials activities, and only some descriptors as atomic charges in positions 1 (Q1), 2 (Q2), 6 (Q6), O or S, 1' (Q1') and 5' (Q5') atoms, the electronic affinity (EA) and ClogP are responsible for the separation between the active and inactive compounds.Four different sets of descriptors   were found to correlate with the two different biological activities, which may indicate that the interaction between the receptor and the binding site and/or mode of action must depend on the type of biological activity.2

Figure 1 .
Figure 1.Structural skeleton numbering (a) and ORTEP view of molecule 12 (b), with the atom labelling scheme and thermal ellipsoids vibration with 50% of probability.

Figure 2 .
Figure 2. Structural skeleton numbering (a) and ORTEP view of molecule 20 (b), showing the atom labelling scheme and the 50% probability thermal vibration ellipsoids.
The compounds were purified by PTLC and purity done by GC-MS.Compound 20 was isolated as a colorless crystals and molecular formula was determined by elemental analysis and MS M + C 14 H 11 O 2 Cl [246 (40%)].Its 1 H NMR spectrum showed typical signals of an aromatic compound.The singlet signal at d 5.20 ppm was assigned to the two H-8 protons.The two doublets at d 6.90 and 7.20 ppm with orto coupling (J7.0 Hz) belong to the pairs H-3´/H-5´ and H-2´/H-6´ protons of the aromatic ring-B, respectively.Multiplet signals at d 7.35-7.60ppm are due to H-3, H-4 and H-5 protons of the aromatic ring-A and the signal at d 7.95 ppm (dd, J 6.8 Hz and J 1.5 Hz) was assigned to the H-2 and H-6 protons.

Figure 5 .
Figure 5.View along the crystallographic b-axis of supramolecular arrangement supported by the intermolecular hydrogen bonds between the adjacent molecules in compound 12, indicated as dashed lines.

Figure 6 .
Figure 6.Packing view of supramolecular arrangement formed by the intermolecular hydrogen bonds between the adjacent molecules in compound 20, indicated as dashed lines (online version in color).

Figure 8 .
Figure 8. MEP surfaces of (a) compound 1 (active) and (b) compound 17 (inactive).The increase of negative charges goes from positive (dark gray) to negative.

Figure 10 .
Figure 10.MEP surfaces of (a) compound 2 (active) and (b) compound 9 (inactive).The increase of negative charges goes from positive (dark gray) to negative.

Figure S1 .
Figure S1.Box plot of L. donovani activity for O or S atom considering nineteen neolignan compounds.

Table 1 .
Biological response for the neolignans studied.See Figure1for the positions of substituents 'R' listed in the first line of the table Figure 3. Strucutal skeleton and numbering of twenty neolignan derivatives.

Table 2 .
Crystallographic data for compounds 12 and 20

Table 3 .
The eight most important descriptors that classified the twenty neolignans used in the SDA study

Table 6 .
Classification matrix obtained using the SDA method for L. donovani activity

Table 7 .
Classification matrix obtained using the SDA method for L. amazonensis activity Figure 9. Box plot of L. amazonensis activity for Q1' considering nineteen neolignan compounds.