On the ordeal of quinolone preparation via cyclisation of aryl-enamines; synthesis and structure of ethyl 6-methyl-7-iodo-4-(3-iodo-4-methylphenoxy)-quinoline-3-carboxylate

Synthesis

X-ray diffraction studies

X-ray diffraction data for compound 6 were collected at room temperature, using a small single crystal, on a Bruker APEX II diffractometer with graphite monochromatized Mo Kα radiation (λ=0.71073 Å). The single-crystal specimen was needle-shaped with dimensions 0.42×0.09×0.07 mm³. A monoclinic unit cell [a=4.4835(1), b=18.4261(3), c=24.6676(4) Å, β=99.731(1)°] was first derived from indexing the Bragg spots of the first 36 CCD frames and it was further refined by least-squares from the positions of 7814 measured Bragg spots at the end of the data-collection. Systematic absences pointed to the monoclinic P2₁/c space-group, and this assignment was confirmed from the structure solution. Data reduction and a multi-scan absorption correction were performed with the SAINT and SADABS [14] suite of programs.

The crystallographic structure was solved by direct methods using SHELXT-2014/4 [15]. Refinements were carried out with the SHELXL-2014/7 package [16]. All refinements were made by full-matrix least-squares on F² with anisotropic displacement parameters for all non-hydrogen atoms. Hydrogen atoms were placed at calculated idealized positions and refined as riding using SHELXL-2014/7 default parameters. The final refinement showed a significant positive residual density of 2.76 e/ų at position (0.0018, 0.3795, 0.0119) that was interpreted as due to an alternate position of the I2 atom of a minor rotational disorder of the 3-iodo-4-methylphenoxy group around the single C–O bonds. Accordingly, the position of the I2 atom was split into the two alternate positions and refined with site occupancies constrained to a sum of one for the two positions.

The refined structural model gave a final R₁ factor of 0.0435 for 3276 reflections of I>2σ and R_all of 0.0699 for all 4607 reflections and 243 parameters.

Crystallographic data for the structure in this paper have been deposited with the Cambridge Crystallographic Data Centre as supplementary publication no. CCDC 1495502. Copies of the data can be obtained, free of charge, on application to CCDC, 12 Union Road, Cambridge CB2 1EZ, UK, (fax: +44-(0)1223-336033 or e-mail: deposit@ccdc.cam.ac.Uk).

Infrared spectroscopy studies

The vibrational spectroscopy studies performed were centred on the characterization of the 6-methyl-7-iodo-4-oxo-quinoline-3-carboxylate 3 and its enol form, 6-methyl-7-iodo-4-hydroxy-quinoline-3-carboxylate 5. The room temperature infrared spectrum of 3 was obtained using the attenuated total reflection technique, with 1 cm⁻¹ resolution, in a Thermo Nicolet IR300 Fourier transform infrared (FTIR) spectrometer, equipped with a Smart Orbit ATR accessory, a deuterated triglycine sulfate (DTGS) detector and a Ge/KBr beam splitter. The compound was then sublimated under high-vacuum conditions using a specially designed thermoelectrically heatable mini-oven placed inside the chamber of a cryogenic system (closed-cycle helium refrigerator APD Cryogenics, with a DE-202A expander). The sublimate was deposited onto the CsI optical cold substrate of the cryostat (T=15±0.1 K) to form a glassy state that was then probed by infrared spectroscopy, or isolated in an inert argon matrix by co-deposition of the sublimate with a large excess of argon (N60, Air Liquide) coming from a separate line. The obtained matrix was then subjected to infrared spectroscopy analysis. The details of the post-processing (temperature variation) of the samples obtained as described above are provided later in this article. The low-temperature IR spectra were registered, with 0.5 cm⁻¹ resolution, using a Thermo Nicolet Nexus 670 FTIR spectrometer equipped with a DTGS detector and a Ge/KBr beam splitter, whose sample compartment was continuously purged by a flux of air free of H₂O vapor and CO₂.

Quantum chemical calculations

Quantum chemical calculations on tautomers 3 and 5 were performed using Gaussian 09 [17], at the density functional theory (DFT) level of theory, with the 6-311++G(d,p) basis set [18], [19] for all atoms except iodine, for which the LANL2DZ (Los Alamos National Laboratory 2 double-ζ) effective core potential (ECP) was used [20], [21], [22]. The DFT calculations were performed using the Becke 3-parameters, Lee, Yang and Parr (B3LYP) hybrid functional [23], [24], [25]. Calculated vibrational frequencies were obtained at the same level of approximation and subsequently scaled down by the factor 0.978, to account for the approximations inherent to the used theoretical model and anharmonicity [13].

Synthesis

6-Methyl-7-iodo-4-oxo-quinoline-3-carboxylate (6Me7I-EOQ, 3) was synthesized from the α,β-unsaturated ester derivative 2 (Scheme 1, route a) by thermally driven intramolecular cyclisation, following the Gould-Jacobs methodology [10]. This 4-oxo-quinoline derivative (3) was characterized by mass spectrometry (provided as Supplementary Data – Fig. S7), elemental analysis and infrared spectroscopy, and its structure was investigated in detail through quantum chemistry calculations (see below). ¹H NMR and ¹³C NMR data could not be obtained for compound 3 due to its poor solubility in all the available solvents. A previously described methodology [12] was adapted for the preparation of the enamine precursor 2 from diethyl ethoxymethylenemalonate and 3-iodo-4-methylaniline [10] which, in turn, was obtained from 2-iodo-4-nitrotoluene by reduction on the surface of metallic iron [26]. Mass, ¹H NMR and ¹³C NMR spectra, of 3-iodo-4-methylaniline and of its enamine derivative 2, are provided as Supplementary Data (Fig. S1–S6).

Considering the low solubility exhibited by 4-oxo-quinoline 3 it was decided to attempt the preparation of its 4-chloro-quinoline analog (6Me7I-EClQ, 4). Chloroquinolines are generally more soluble than the corresponding oxo derivatives, so chloroquinoline 4 would be more easily isolated and could then be converted into the corresponding 4-oxo-quinoline (Scheme 1, route c). Additionally, chloroquinolines are known as versatile building blocks [12]. We proposed that compound 4 would be a suitable intermediate for the preparation of quinolones with chemical diversity at position 7, given the versatility provided by the iodine substituent in this position (4; Scheme 1).

Although chloroquinoline 4 could be prepared directly from 4-oxo-quinoline 3 (Scheme 1, route b), we decided to attempt its synthesis from the α,β-unsaturated ester derivative 2 (Scheme 1, route d). This approach involves cyclisation of the enamine 2 and subsequent chlorination at position 4 of the quinolone core, in one pot, to afford the chloroquinoline 4 [11]. Application of the procedure outlined (Scheme 1, route d) and subsequent isolation afforded an amorphous solid corresponding to a sole compound. However, analysis by NMR spectroscopy and mass spectrometry provided data that were not compatible with the expected structure for chloroquinoline 4. For example, by MS it was not possible to find a peak corresponding to the molecular ion (m/z 376 [(M)⁺] or 399 [(M+Na)⁺]) but, instead, a peak corresponding to m/z around 573 indicated the presence of an unexpected product (MS spectrum provided as Supplementary Data – Fig. S8). It was then envisaged to proceed with characterization by X-ray crystallography. After crystallization, a suitable crystal was isolated and the crystal structure was investigated, revealing the presence of ethyl 7-iodo-4-(3-iodo-4-methylphenoxy)-6-methyl-quinoline-3-carboxylate (6, Scheme 2 and Fig. 2). The low isolated yield (12%) presented for compound 6 may be due to difficulties on the nucleophilic attack by the 4-hydroxyl group in compound 5, due to steric hindrance.

Fig. 2:

ORTEP drawing with atom-labeling scheme of ethyl 7-iodo-4-(3-iodo-4-methylphenoxy)-6-methyl-quinoline-3-carboxylate (6) with thermal ellipsoids drawn at 50% probability level. For clarity, the alternate position corresponding to the minor disordered form of the I2 atom is omitted.

Crystal structure of ethyl 7-iodo-4-(3-iodo-4-methylphenoxy)-6-methyl-quinoline-3-carboxylate (6): The X-ray data obtained for compound 6 shows that the quinoline moiety is planar, with an rms deviation of the ring atoms from the least-squares plane of just 0.024 Å. The ethyl-carboxylate substituent is almost within the plane of the quinoline ring, the O2–C17–C3–C4 torsion angle being –2.45°. The 3-iodo-methy-phenoxy substituent forms an angle of 59.62(13)° with the quinoline least-squares plane. There is some rotational flexibility of this substituent for rotations around the C4–O1 and O1–C11 bonds, as shown by the torsion angles C10–C4–O1–C11 and C4–O1–C11–C12 of 41.95 and 23.13°, respectively. It appears that such rotational flexibility might allow for the presence in the crystal of two alternate positions of this substituent, with occupancies of 96% and 4%, responsible for the presence of a small “ghost” peak of the iodine I2 atom in the alternate position corresponding to the minor disordered form.

Bond lengths and angles in the molecule are within the expected range of values. Packing of the molecules in the crystal appears to be dictated mainly by weak interactions between the π-clouds of the aromatic rings, resulting in a stacking of molecules along the a-axis (Fig. 3). In addition, there is a short C–H···Cg contact with the π-electron cloud of one of the aromatic rings.

Fig. 3:

Crystal packing of ethyl 7-iodo-4-(3-iodo-4-methylphenoxy)-6-methyl-quinoline-3-carboxylate (6) showing the network of Cg···Cg and C–H···Cg weak intermolecular interactions.

The NMR and MS data previously obtained for compound 6 proved to be compatible with the structure obtained. The molecular weight of 6 (572.93 g/mol) fits in with the peak observed by mass spectrometry (MS spectrum as Supplementary Data – Fig. S8), m/z around 573. Also, the NMR spectra confirm the presence of structure 6 (Fig. S9 and S10 and Table S1 – Supplementary Data). Signals due to one additional aromatic ring, corresponding to resonances of the 4-aryloxy group (atom positions 11 to 16), are clearly observed in the ¹H NMR and in the ¹³C NMR spectra. The presence of two methyl groups in compound 6 (atom positions 17 and 20) is also proved by the characteristic singlet peaks at 2.53 ppm and 2.37 ppm, in the ¹H NMR, and the peaks at 27.6 ppm and 26.8 ppm in the ¹³C NMR spectra. Furthermore, it is important to note the absence of a signal due to the N–H group (present for oxo-quinoline compounds, but absent for 6) and the presence of signals due solely to one ethyl ester group.

Isolation of the unexpected ether 6 raised some questions, and one of them is the mechanism underlying its synthesis. Our proposal, depicted in Scheme 2, considers initial formation of the 4-oxo-quinoline 3-ester 3, as could be expected under the conditions outlined (Scheme 1, route d). The oxoquinoline may then undergo tautomerization to the corresponding 4-hydroxy-quinoline 3-ester 5. Once formed, the hydroxy-quinoline tautomer 5 may act as nucleophile (through its 4-hydroxyl group), reacting with the enamine precursor 2 and subsequently affording the 4-aryloxyquinoline 6 (steric hindrance renders nucleophilic attack difficult, resulting in a low yield for the synthesis of compound 6).

The quinolone/hydroxyquinoline tautomerism was previously observed for other quinolone 3-esters and has been investigated [13]. The relative amounts of both tautomeric forms depend mostly on the structure of both species and are also affected by the medium conditions. The mechanism proposed for the formation of ether 6 assumes the presence of a fair amount of 4-hydroxy-quinoline 5 in the reaction medium, this implicating that compound 5 is energetically favored, compared to its 4-oxo-quinoline tautomer 3. To assess this aspect and gather more information to support a mechanistic proposal for the formation of ether 6, we decided to proceed with a deeper investigation of the molecular structure of tautomers 3 and 5, theoretically, using molecular orbital calculations, and experimentally, using matrix isolation coupled to FTIR spectroscopy.

Molecular structure of ethyl 4-oxo-6-methyl-7-iodo-quinoline-3-carboxylate (3) and its enol form (5): The 4-oxo-quinoline 3-ester 3 and its enol form, 4-hydroxy-quinoline 5, were studied in more detail using infrared spectroscopy, complemented by quantum chemistry DFT calculations. The molecule of 3 has three conformationally flexible coordinates, defined by the torsions about the C–C bond connecting the ethyl carboxyester to the oxo-quinoline ring and the two C–O bonds (carboxylic and ester). It is well known [27], [28] that in carboxylic esters conformations with a cis arrangement (O=C–O–C dihedral ~0°) are by far more stable than those having a trans (O=C–O–C dihedral ~180°) geometry. Keeping the cis arrangement about the C–O carboxylic bond, the performed DFT calculations yield four different conformers for the molecule of the oxo-quinoline 3, with conformer III being the most stable form (Fig. 4).

Fig. 4:

B3LYP/6-311++G(d,p)/LANL2DZ optimized geometries for the 4 conformers of 6Me7I-EOQ (3) with a cis carboxylic-ester group (O=C–O–C≈0°). Relative energies, in italic style, are in kJ mol⁻¹ and are relative to the most stable conformer (III); the values in parentheses are relative to the most stable conformer of 6Me7I-EHQ (5). For conformers of C₁ symmetry, the C–O–C–C and C(O)–C–C–O dihedral angles are given, in this order, in parentheses. Note that all C₁ conformers shown in the figure have an equivalent-by-symmetry related form.

Calculations performed at the same level of theory on the hydroxyquinoline tautomer 5, which has one additional conformationally relevant torsion (defining the orientation of the OH group), revealed the existence, in this case, of eight different conformers with a cis arrangement about the C–O carboxylic bond. The two lowest energy forms (I and II) are stabilized by a strong O–H⋅⋅⋅O= intramolecular hydrogen bond and are considerably more stable than the remaining forms. The next two conformers of 5 in order of energy have a weaker intramolecular H-bond of type O–H⋅⋅⋅O_(ester) (this type of H-bond is well-known to be weaker than the O–H⋅⋅⋅O= one because the carbonyl oxygen has a better ability to act as H-bond acceptor) [29], [30] and are more than 15 kJ mol⁻¹ higher in energy than the most stable conformer I (Fig. 5). Very interestingly, the lowest energy conformers of the hydroxyquinoline 5 are lower in energy than the most stable conformer of its oxoquinoline tautomer 3 by more than 27 kJ mol⁻¹. Such piece of information called for a detailed investigation on the nature of the tautomer (3 or 5) present in the synthesized solid sample through matrix isolation and FTIR spectroscopy studies.

Fig. 5:

B3LYP/6-311++G(d,p)/LANL2DZ optimized geometries for the eight conformers of 6Me7I-EHQ (5) with a cis carboxylic-ester group (O=C–O–C≈0°). Relative energies, in italic style, are in kJ mol⁻¹ and are relative to the most stable conformer (I). For conformers of C₁ symmetry, the C–O–C–C, C(OH)–C–C–O and C–C–O–H dihedral angles are given, in this order, in parentheses. Note that all C₁ conformers shown in the figure have an equivalent-by-symmetry related form.

Matrix-isolation and infrared spectroscopy studies: To obtain further insight on the 3/5 tautomerism, a series of experiments were designed. The first one was the collection of the infrared spectrum of the synthesized solid at room temperature. The examination of such spectrum, and its comparison with the calculated spectra for the more stable conformer of 3 and 5, suggested the sole presence of the oxoquinoline tautomer in the sample. However, since in absolute terms the predicted spectra for the two tautomers have many similarities and also because the calculations were performed for the isolated molecules, so that the comparison of the calculated spectra with the experimental one for the solid sample could be expected a priori to have some limitations, we decided to obtain the infrared spectrum of the matrix-isolated species resulting from sublimation of the synthesized solid. Matrix isolation infrared spectroscopy is a very powerful technique for structure elucidation and is particularly sensitive to subtle structural modifications like those resulting from changes in conformation or tautomerization.

The synthesized solid was then sublimated (at T~100°C) under high-vacuum, as described in the Section “Infrared spectroscopy studies”, mixed with a large excess of argon, and deposited onto a CsI substrate cooled at 15 K, for IR spectra recording. The obtained spectrum is shown in Fig. 6, and fits very well the simulated spectrum for a mixture population of the two lowest energy conformers (I and II) of hydroxyquinoline 5 in the proportions expected to exist in the thermodynamic equilibrium at T~100°C (67%: 33%; ratio estimated considering the calculated relative energy of the two conformers and assuming the Boltzmann law). Moreover, annealing of the matrix up to 40 K led to changes in relative intensities which clearly reveal that upon increase of the temperature the less stable conformer II converts into the most stable form I (Fig. 7), as it could be anticipated considering that in a matrix isolation experiment the initially trapped conformers’ populations are those existing in the vapor of the compound immediately before deposition (in this case those characteristic of the thermodynamic equilibrium at T~100°C) and should convert into those characteristic of the low temperature matrix (15 K) upon annealing of the matrix if the energy barrier separating the two conformers is low (as in the present case, where it amounts to ~2 kJ mol⁻¹; Fig. 8). The results of this experiment are then full consistent with the theoretical predictions, both in relation to the relative stability of the two tautomers (5 is in fact more stable than 3 in the gas-phase) and also in relation to the existence of two experimentally relevant low-energy conformers of the hydroxyquinoline 5, with I being slightly more stable than II.

Fig. 6:

Infrared spectra of (top): a glassy state resulting from deposition of vapors of the synthesized sample of 6Me7I-EOQ (3) onto an IR-transparent CsI windows mounted at the cold-tip of the cryostat (T=15 K); (middle): an argon matrix prepared by co-deposition of argon and vapors of the synthesized sample of 6Me7I-EOQ (3) onto an IR-transparent CsI windows mounted at the cold-tip of the cryostat (T=15 K); (bottom): simulated IR spectrum of a sample containing 67% and 33% of conformers I and II of 6Me7I-EHQ (5), built using the B3LYP/6-311++G(d,p)/LANL2DZ calculated IR spectra of these conformers with peaks broadened by convolution with Lorentzian functions centred at the calculated (scaled by 0.978) frequencies and peak areas equal to the calculated IR intensities. Sublimation of the 6Me7I-EOQ (3) crystal leads to conversion of the oxo-quinoline to the hydroxyl form (6Me7I-EHQ, 5). (see text for detailed explanation).

Fig. 7:

Infrared difference spectra. (top): Annealed argon matrix (40 K) minus as-deposited matrix (15 K); (bottom): simulated IR difference spectrum (6Me7I-EHQ, 5: I minus II), built using the B3LYP/6-311++G(d,p)/LANL2DZ calculated IR spectra of conformers I and II 6Me7I-EHQ (5) with peaks broadened by convolution with Lorentzian functions centred at the calculated (scaled by 0.978) frequencies and peak areas equal to the calculated IR intensities. Annealing of the matrix containing conformers I and II of 6Me7I-EHQ (5) at 40 K leads to the II→I conversion. (see text for detailed explanation).

Fig. 8:

Potential energy profile for rotation about the O–C ester bond leading to conformational interconversion between forms I and II of 6Me7I-EHQ (5).

The next experiment consisted in preparing a sample in an identical way as used for the argon matrix preparation, but depositing just the vapor of the synthesized solid. In this case, we obtained a glassy state of the neat compound, since the fast deposition of the molecules present in the gas phase in the cold substrate precluded any molecular rearrangement to take place. The obtained spectrum of this sample is shown in Fig. 6, and it can clearly be seen from the comparison of this spectrum with that of the matrix-isolated species (the hydroxyquinoline 5) that hydroxyquinoline 5 is the sole constituent in the low temperature glassy state prepared. The glassy state was then warmed stepwise, and at a temperature of about 115 K the spectrum starts to change. At 285 K the crystallization of the glass was complete (Fig. 9). The obtained spectrum of the crystal grown from the glass was found to be identical to that of the room temperature solid (Fig. 10), and the full set of data obtained in the described series of experiments doubtlessly indicated the sole presence of the oxoquinoline tautomer 3 in these samples.

Fig. 9:

Infrared spectra. (top): of the glassy state obtained by sublimation of vapors of the synthesized 6Me7I-EOQ (3) sample onto an IR-transparent CsI windows mounted at the cold-tip of the cryostat (T=15 K) [red trace] and of the crystal resulting from warming of this glass [blue trace]; (bottom): simulated IR spectra of 6Me7I-EHQ (5) [red trace] and 6Me7I-EOQ (3) [blue trace], built using the B3LYP/6-311++G(d,p)/LANL2DZ calculated IR spectra of conformers I and II (67%, 33%, respectively) of 6Me7I-EHQ (5) and of conformer III of 6Me7I-EOQ (3), with peaks broadened by convolution with Lorentzian functions centred at the calculated (scaled by 0.978) frequencies and peak areas equal to the calculated IR intensities. Upon heating, the glassy state formed by 6Me7I-EHQ (5) gives rise to crystalline 6Me7I-EOQ (3) (crystallization starts to take place at ~115 K and is complete at 285 K).

Fig. 10:

Infrared spectra. (top): ATR IR spectrum of the synthesized 6Me7I-EOQ (3) sample; (middle): IR spectrum of the crystal (at 298 K) formed upon warming of the glassy state obtained by sublimation of vapors of the synthesized 6Me7I-EOQ (3) sample onto an IR-transparent CsI windows mounted at the cold-tip of the cryostat (T=15 K); (bottom): simulated IR spectrum (6Me7I-EOQ, 3), built using the B3LYP/6-311++G(d,p)/LANL2DZ calculated IR spectra of conformer III of 6Me7I-EOQ (3) with peaks broadened by convolution with Lorentzian functions centred at the calculated (scaled by 0.978) frequencies and peak areas equal to the calculated IR intensities. Upon heating, the glassy state formed by 6Me7I-EHQ (5) gives rise to crystalline 6Me7I-EOQ (3)-crystallization starts to take place at ~115 K and is complete at 285 K.

In conclusion, in the gas phase the hydroxyquinoline tautomer 5 is the sole present species (in agreement with the much lower energy of the isolated molecule of this tautomer compared to that of 3), while the crystalline phase is formed by the oxoquinoline 3 tautomer, as dictated by the crystal packing forces. Interestingly, though the intermolecular interactions are strong enough to induce the 5→3 tautomerization, the IR spectrum of the crystal is rather well reproduced by the theoretically calculated spectrum of the most stable monomer of 3, indicating that the intramolecular vibrational potential is not very much perturbed by the intermolecular potential (see Fig. 10).

Although the 4-oxo-quinoline tautomer 3 is favored in the crystal phase, the results obtained from the structural studies point to a relatively easy tautomerization between 4-oxo-quinoline and 4-hydroxy-quinoline forms, compounds 3 and 5, respectively. Thus, in the reaction conditions used (route d: POCl₃; 97°C) a fair amount of the energetically more stable 4-hydroxy-quinoline tautomer 5 should be present in the reaction medium. It is worth noting that the reaction involves production of HCl as by-product, also favoring the production of the tautomer 5, proposed as intermediate in the preparation of 4-aryloxy-quinoline derivative 6.

The results obtained support our mechanistic proposal for the formation of the 4-aryloxy-quinoline derivative 6 through route d (Scheme 2) and re-enforce previous observations regarding the impact of the oxo/hydroxyquinoline tautomerization in the fate of available synthetic routes for the preparation of quinolones.

Biological activity and cLogP calculations: The 4-oxo-quinolone 3 and its 4-aryloxy-quinoline derivative 6 were evaluated in vitro for their antiplasmodial activity against two strains of Plasmodium falciparum: a sensitive (3D7) and a multiresistant strain (Dd2). As shown in Table 1, results of antiplasmodial activity against the strain 3D7, measured by IC₅₀ values, fall in the high nanomolar range for both compounds, while for the multiresistant strain Dd2 the IC₅₀ values increase to the low micromolar range. Also 4-oxo-quinoline 3 appears to be more active than its 4-aryloxy derivative, in both parasite models (e.g. IC₅₀ values of 0.323 μM and 0.743 μM, respectively, against the strain 3D7).

In conclusion, both compounds are active against both CQ-sensitive and CQ-resistant malaria models. The better antimalarial activity demonstrated for compound 3, compared to compound 6, may be expected, considering the previous docking studies (importance of the N–H and 4-oxo groups in the quinolone core for enzyme inhibition). Thus, the 4-oxo-quinoline structure is expected to perform better from a pharmacodynamic viewpoint. Additionally, compound 6 presents a high cLogP value (6.18), which might mean that, as a drug, 6 is too hydrophobic and also too lipophilic. However, the possibility of easy 4-oxo-quinoline/4-hydroxy-quinoline tautomerization inside the enzyme pocket may impact in activity.