Preparation of 2,2-dimethylchroman-4-ones from 5-alkyl-substituted resorcinols: microwave-assisted synthesis and theoretical calculations

The influence of different 5-alkyl-substituted resorcinols on the formation of 2,2-dimethyl-chroman-4-ones is examined experimentally and theoretically. Structures are fully assigned by means of experimental and theoretical 13 C and 1 H NMR chemical shifts. Based on experimental and theoretical calculations of Friedel-Crafts acylation, it is possible to explain the formation of 2,2-dimethyl-5-hydroxychroman-4-ones and/or 2,2-dimethyl-7-hydroxychroman-4-ones. Evaluation of their biological activity as cannabinoid receptor ligands is also reported.


Introduction
Chroman-4-ones are attractive compounds not only because of their therapeutic applications but also as intermediates in preparation of more complex structures.As reported by Saengchantara and Wallace 1 in a review covering chromanols, chromanones and chromones, many chroman-4ones are isolated from natural sources such as Dysophylla stellata, rhizomes of rhubarb, or Penicillium frequentans.Examples of these compounds include cannabichromanones identified as components of the Cannabis sativa plant. 2,3Moreover, 2,2-dimethyl-5-hydroxychroman-4ones have been described as intermediates in the synthesis of tetrahydrocannabinols 4 and, as we have reported, in the synthesis of cannabinoid chromenopyrazoles. 5A survey of the preparations of differently substituted 2,2-dimethylchroman-4-ones was published by Tímár et al. 6 twelve years ago.Most of them are prepared by reacting the appropriate phenol with 3,3-dimethyl-acryloyl chloride or its corresponding acid in presence of Lewis acid.However, it has been mentioned in the literature that the reactions are not always successful, providing poor yields or mixtures of different chromanones, or requiring long reaction times. 7,8On consideration of this, a study was undertaken to examine the influence of different 5-alkyl-substituted resorcinols on the formation of 2,2-dimethychroman-4-ones using both microwave (MW) assisted reactions and conventional heating (C.H.).Structures have been fully assigned by means of experimental and theoretical 13 C and 1 H NMR chemical shifts.Finally, the first step, Friedel-Crafts acylation reaction has been examined theoretically in order to correlate the experimental evidence using Density Functional Theory (DFT).

Scheme 2. Preparation of 5-n-butyl-1,3-dihydroxybenzene (1c).
Upon treatment of resorcinols (1a-1e) with 3,3-dimethylacrylic acid in methanesulfonic acid in presence of phosphorus pentoxide, the Friedel-Crafts acylation was expected to take place at either the 2 or 6 position of 1,3-dihydroxybenzene (Scheme 1).To study the effect of the alkyl substituents on the resorcinol, reaction parameters (solvent, molar ratio of reactants, catalyst, and temperature) were at first investigated.The reactions were carried out under the most advantageous reactant ratio: resorcinol / phosphorus pentoxide / 3,3-dimethylacrylic acid [1/1.2-1.5/1.5-2] using methanesulfonic acid as solvent at 70 ºC.Under these conditions the reaction was performed under microwave irradiation for 10, 20 or 60 minutes or with conventional heating for 12 hours.The reactions were monitored by HPLC/MS (C-18 reverse phase column, eluent: MeCN/H2O) that allowed the ratio in which regioisomers 2 and 3 were formed to be determined.
As reported in Table 1, 2,2-dimethyl-7-hydroxychroman-4-one was isolated as the only regioisomer in the case of 3a and 3b (R 1 = H and R 1 = Me), whereas the regioisomer 2,2dimethyl-5-hydroxychroman-4-one was isolated as the unique isomer for 2e (R 1 = 1,1dimethylheptyl).The reaction of 3,3-dimethylacrylic acid with 5-butyl-1,3-dihydroxybenzene (R 1 = n-butyl) and olivetol (R 1 = n-pentyl) resulted in the formation of both isomers 2 (2c, 2d) and 3 (3c, 3d).The ratio in the crude product mixture could be easily determined by HPLC/MS.These ratios are reported in Table 1.Regioisomers 2 and 3 could be isolated by column chromatography.Two-dimensional NMR experiments were of particular value for the identification of the regioisomers 2 and 3, as shown in Figure 1.Moreover, the chemical shifts of All the molecules 2 and 3 obtained experimentally have also been calculated using GIAO-DFT methods at B3LYP/6-311++G(d,p) computational level.The 13 C-NMR theoretical data nicely resemble the experimental ones, showing an excellent linear correlation with the experimental findings (R 2 = 0.993) and an absolute average error of 3.1 ppm.The 1 H-NMR theoretical results show an absolute average error of 0.18 ppm and a R 2 value of 0.99.Altering the reaction conditions did not significantly affect the regiochemical outcome.However, by increasing the microwave exposure to 1 hour, diacylated product (4) was detected in the reaction mixture except for 1,1-dimethylheptyl derivative that was not observed (Table 1  entry 19).These compounds (4a, 4b, 4c and 4d) were identified by HPLC/MS that also allowed the determination of the ratios in the crude reaction.The structures of 4a and 4d, which were isolated by chromatography, were unambiguously determined by MS and NMR experiments.Note that the diacylated product 4 could be formed from both 2,2-dimethyl-5-hydroxychroman-4-one (2c, 2d) and 2,2-dimethyl-7-hydroxychroman-4-one (3c, 3d).
It is interesting to note the difference of reactivity between the microwave-assisted reactions and the reaction carried out by conventional heating.Compounds 3a, 3b and 3d (10 min) formed faster than 4a, 4b and 4d (1h) are the kinetic products of this reaction.Compounds 4a, 4b and 4d obtained in lower amount are thermodynamically more stable than the corresponding compounds 3.Under conventional heating conditions, the reaction seems to be fully kinetically controlled since there is no formation of 4.
In an attempt to explain the reactivity of different resorcinols with 3,3-dimethylacrylic acid, theoretical calculations have been performed.Thus, the reactivity of three systems has been considered (Scheme 3); resorcinol, orcinol and 5-tert-butyl-1,3-dihydroxybenzene have been selected.Hydrogen and methyl groups in R 1 are the corresponding substituents of resorcinol and orcinol.The tert-butyl group in R 1 was chosen to mimic the 1,1-dimethylheptyl moiety.

Scheme 3. Schematic representation of the reaction.
For each of the three starting molecules, two possibilities have been considered: the reaction on the C2 and the one in C4.Two possible transition state (TS) paths have been explored for the Friedel-Crafts acylation depending on the disposition of the attacking carbonyl molecule with respect to the aromatic ring (Figure 2).

TS(in)-I3
TS(out)-I3 Small values of the barriers are found with respect to the isolated reactants (Table 2).The differences between the most favorable TS that yields I3-H + and I2-H + are small being in all the cases the one that yields I3-H + the smaller.However, the relative energies of the charged species obtained in this first step of the reaction are different depending on the R 1 group.Thus, for R 1 = H and Me, the most stable molecule correspond to the I3-H + while for R 1 The relative energies of the neutral intermediate molecules I3 and I2, as well as the bicyclic structures 3 and 2 obtained after a subsequent intramolecular Michael addition are gathered in Table 3.Only in the case of X = H is I3 more stable than I2.In the rest of the cases, I2 is more stable than I3 and the same happened between 2 and 3.In addition, the energetic differences increase with the size of the R 1 group.Based on the experimental and the computational results shown here, the Friedel-Crafts acylation that yields the formation of I3-H + and I2-H + seems to be the one responsible of the percentage of final products obtained.Thus, the calculated relative energies of the I3-H + and I2-H + intermediates are able to explain the experimental results.In other terms, the influence of different 5-alkyl-substituted resorcinols on the formation of 2,2-dimethychroman-4-ones does not depend on the energetic barrier of the first step involved in Friedel-Crafts acylation.But it is rather the difference of relative energies of the charged species at the transition state that depends on the nature of the alkyl substituent.Based on these theoretical considerations, it has been possible to explain the formation of 2,2-dimethyl-5-hydroxychroman-4-ones and/or 2,2dimethyl-7-hydroxychroman-4-ones obtained experimentally.

Affinity towards cannabinoid receptors
The affinity of chromanones 2d-e, 3a-d, 4a and 4d at CB1 and CB2 cannabinoid receptors were evaluated using radiolabeled binding assays.  Standard cannabinoid ligand WIN55,212-2 was also assessed for comparison with the new compounds.The tested compounds displayed much lower affinity than the reference cannabinoid ligand.Although these affinities are low, these data suggest some relation structure-activity.In the 2,2dimethyl-7-hydroxychroman-4-one series (3), the lipophilic methylenic chain is necessary for CB2 receptor binding (3d).However, this is not the case for the tricyclic scaffold (4) for which the moderate CB2 affinity was independent of the lipophilic nature of the alkyl chain (4a and 4d).2,2-Dimethyl-5-hydroxychroman-4-one 2d showed weak but selective affinity for the CB2 cannabinoid receptors whereas the dimethylheptyl derivative 2e binds weakly to both receptors.

Conclusions
The influence of different 5-alkyl-substituted resorcinols on the formation of 2,2dimethychroman-4-ones does not depend on the energetic barrier of the first step involved in Friedel-Crafts acylation.But it is rather the difference of relative energies of the charged species at the transition state that depends on the nature of the alkyl substituent.Based on these theoretical considerations, it has been possible to explain the formation of 2,2-dimethyl-5hydroxychroman-4-ones and/or 2,2-dimethyl-7-hydroxychroman-4-ones obtained experimentally.From an experimental point of view, the use of microwave radiation considerably reduced reaction times but it did not increase reaction yields.At longer microwave irradiation, diacetylated products started to form in the reaction mixture.With respect to their biological activity, some of the chromanones described here showed moderate but selective affinity for the CB2 cannabinoid receptor.

Experimental Section
Chemistry General.All commercially available reagents and solvents were used as supplied without further purification.Microwave-assisted synthesis was carried out with a Biotage Initiator Microwave (0-400W at 2.45 GHz).Column chromatography was performed using silica gel 60 (230-400 mesh).The purity of final compounds was determined by HPLC-MS and elemental analyses.HPLC-MS analysis was performed on a Waters 2695 HPLC system equipped with a Photodiode Array 2996 coupled to Micromass ZQ 2000 mass spectrometer (ESI-MS), using a reverse-phase column SunFire TM (C-18, 4.6 x 50 mm, 3.5 μm) and 5 min gradient A) MeCN/0.08% formic acid, B) H2O/1% formic acid visualizing at λ = 254 nm.Elemental analyses were performed using a LECO CHNS-932 apparatus. 1H and 13 C NMR spectra were recorded on a Bruker 300 (300 and 75 MHz respectively) at 25ºC with CDCl3 as solvent and TMS as internal standard.Chemical shifts were expressed in ppm (δ) and coupling constants were reported in Hz.Melting points were determined in open capillary tubes with a MP70 Mettler Toledo apparatus.5-(1,1-Dimethylheptyl)-1,3-dihydroxybenzene (1e) was prepared as described previously by some of us. 5

Computational method
The geometry of the systems has been optimized with the B3LYP/6-31G(d) computational level 15,16 within the Gaussian-09 program. 17In order to include the effect of the solvent, methanesulfonic acid, the Polarizable Continuum Method (PCM) model 18 has been considered with a dielectric constant of 80 which is an intermediate value of the ones described in the literature for this solvent. 19Frequency calculations have been carried out to confirm that the structures obtained correspond to energy minima or true transition states (TS).The NMR theoretical chemical shielding of the synthesized molecules has been calculated using the GIAO method 20 at the B3LYP/6-311++G(d,p) computational level using the geometries optimized at the same level.The effect of the solvent used in the experimental NMR measurements (CDCl3) has been simulated using the PCM model of this solvent.The calculated absolute chemical shieldings have been transformed into chemical shifts using empirical relationships previously described. 21,22

Figure 2 .
Figure 2. The two TS found for the Friedel-Crafts acylation that yield the formation of I3-H + , R 1 = H.

Table 1 .
Reaction of 5-alkyl-substituted resorcinols with 3,3-dimethylacrylic acid in methanesulfonic acid under given heating conditions and the ratios of the two isomers and the diacylated product determined by HPLC/MS areas analysis

Table 2 .
Relative energy (kJ/mol) of the TS and Friedel-Crafts acylation with respect to the isolated reactants.In parenthesis, the population based on the energies of I3-H + and I2-H + is indicated.Values at B3LYP/6-311++G(d,p) computational level in PCM, ε = 80