Preparation of tetrahydro-1 H -xanthen-1-one and chromen-1-one derivatives via a Morita-Baylis-Hillman/oxa-Michael/elimination cascade

We

The methodologies described so far for obtaining these moieties from 2-hydroxybenzaldehydes and activated olefins in a single step are summarized in Scheme 1. [17][18][19] In the first description of this transformation, Bräse et al. found that DABCO in water was able to provide in a single step tetrahydroxanthenone derivatives in yields ranging from poor to excellent. 17In this work, the authors were not able to intercept the MBH adduct, and thus proposed a mechanism in which the first step was the oxa-Michael addition followed by an aldol condensation step.In 2005 Shi et al. prepared tetrahydroxanthenones and chromenone derivatives by reacting salicyl N-tosylimines with cyclic enones using a phosphine as the catalyst. 18Contrasting with Bräse's results, the authors intercepted the aza-MBH adducts, indicating that the overall transformation might also proceed firstly by a MBH step.In their protocol, an additional step with DBU was necessary to promote elimination and the overall yields ranged from moderate to excellent.In 2010, 19 Bräse et al. employed the same catalytic system developed by Shi et al. with salicylaldehydes, obtaining the corresponding tetrahydroxanthenones in poor to quantitative yields (most examples in less than 65 % yield).The authors concluded that the MBH pathway could be an alternative with less basic catalysts such as phosphines.
In the present work we describe the use of a bicyclic imidazolyl alcohol (BIA) as a new bifunctional organocatalyst for the synthesis of tetrahydroxanthenones and tricyclic chromenones from the direct reaction of 2-hydroxybenzaldehydes and cyclic enones in water using sodium dodecyl sulfate (SDS) as an additive.

Results and Discussion
We began our investigation with the reaction of salicylaldehyde (1a) and 2-cyclohexen-1-one (2a) as model reaction, and screened a series of parameters.1][22] Given the evidence that the reaction might be proceeding through the MBH adduct, 19 we decided to test this catalytic system in the synthesis of the tetrahydroxanthenone 3a and to compare it with catalysts already known to perform this transformation (Table 1).
We first tested the use of equimolar quantities of salicylaldehyde and 2-cyclohexen-1-one, 50 mol% of BIA and 10 mol% of sodium dodecyl sulfate (SDS) in water, and, after 60 hours, the tetrahydroxanthenone 3a was obtained in 58 % yield (entry 1).In order to check if we could improve this yield, knowing that BIA provides best results in aqueous conditions, we tested different solvent combinations with water (entries 2-5).However, no improvement was observed and, in fact, all results were inferior to those in pure water.We also evaluated the neat transformation, but the yield was also low (entry 6).Thus, water was established as the optimal solvent for the reaction.
We also evaluated the impact of catalyst loading of BIA to the yield of 3a.The use of only 20 mol% of BIA led to an unsatisfactory yield of 30 % (entry 7).When 65 mol% of BIA was used, however, a slightly enhanced yield was observed (entry 8).A further increase in the catalyst loading to 100 mol% did not improve the isolated yield (entry 9).Having fixed 65 mol% of BIA as an optimal catalyst charge, we then evaluated the number of 2-cyclohexen-1-one equivalents in relation to salicylaldehyde (entries 10-12).The reaction performed better when excess of 2a was employed, and the use of 2.0 equivalents of the 2cycloenone led to a 68 % isolated yield of 3a (77 % yield considering the recovery of unreacted salicylaldehyde).Increasing the temperature of the reaction did not improve the yields (entries 13 and 14).
Thus, the optimal conditions were defined as 0.65 equiv. of BIA, 2.0 equiv. of 2-cycloenone and water as the solvent at room temperature (entry 11).In order to compare our protocol with the one reported in the literature by Bräse et al., we performed the reaction of 1a in the same conditions as reported by the authors (entry 15). 17Unfortunately, we were not able to reproduce the yield described by Bräse.We also tested the use of imidazole as catalyst, which is a structurally simple catalyst compared to BIA (entry 16).However, the isolated yield was lower (60 %), confirming the importance of the bifunctional nature of the BIA catalyst to furnish 3a in higher yields.With the optimized conditions, we moved on to evaluate the scope of the reaction by applying seven different salicylaldehydes and six different 2-cycloenones (Scheme 2).The yields reported are all isolated yields after complete consumption of the starting materials or stagnation of the reaction (as confirmed by 1 H NMR of a crude sample).Electron-rich salicylaldehydes gave generally better yields than electron-deficient ones with 2-cyclohexen-1-one.The lowest yield of the 2-cyclohexen-1-one series was when 2-formyl-βnaphthol was the coupling partner.
When 4-bromo-2-hydroxybenzaldehyde was employed, besides the expected product (3f), we also noticed significant formation of 4a, a product from the aldol condensation of 3f with a second molecule of aldehyde.When 2-cyclopenten-1-one was used as reactant, the obtained yields were generally lower and, with electron rich aldehydes, we also observed the formation of the aldol-condensation products (4b and 4c).With electron poor aldehydes, we did not observe aldol-condensation products, nevertheless, we were able to isolate the intermediates MBH adducts (5a and 5b).To the best of our knowledge, this is the first time these MBH adducts are isolated.In addition, in presence of BIA catalyst and in aqueous medium, MBH adduct 5a slowly converts to dihydrocyclopenta[b]chromen-1(2H)-one 3j (35 % yield of 3j after 5 days of reaction).These empirical observations are a strong evidence that the reaction proceeds via an MBH reaction followed by oxa-Michael/elimination steps under our conditions, and not the opposite way.With 2cyclohepten-1-one as the coupling partner, a low yield (13 %) of the expected product was observed.An extensive degradation of compound 3m was observed by thin-layer chromatography during the reaction, which might explain the reduced yield.Gem-disubstituted 2-cyclohexen-1-ones at position 4 or 5 did not give the desired products 3n, 3o and 3p after 7 days of reaction time.This is probably due to a combination of stereoelectronic and steric factors, which might hinder the 1,4-addition of the BIA catalyst to the cycloenone. 23o demonstrate the feasibility of this reaction, we ran an essay on a gram scale.The reaction between salicylaldehyde (1a, 8.46 mmol) and cyclohexenone (2 equiv.)provided the xanthenone 3a in 63 % yield (1.07 g), after 5 days at room temperature.
A single crystal of 4c could be obtained and analyzed by X-ray diffraction and its ORTEP diagram is shown in Figure 2 (CCDC 1980794).This compound co-crystallized with chloroform and is a definitive proof of the structure of the aldol condensation product and the geometry of the double bond.

AUTHORS Conclusions
In this work, we developed a new catalytic system to obtain tetrahydro-1H-xanthen-1-ones and fused chromen-1-ones derivatives directly from the corresponding 2-hydroxybenzaldehydes and 2-cycloenones.The reaction proceeds under catalysis of a bifunctional, bicyclic imidazole alcohol (BIA) using water as solvent, and several examples were synthetized in low to good yields.We believe that under our optimized conditions this transformation involves a Morita-Baylis-Hillman step followed by cyclization (oxa-Michael/elimination steps), which is in contrast to what has previously been published in the literature.Strong evidence for this proposition is the isolation of the MBH adducts 5a and 5b.The synthesized products will be screened for potential biological applications and we are currently exploring the possibility of an asymmetric version of this transformation by employing enantioenriched bifunctional catalysts.

Experimental Section
General.All reagents were used from commercial suppliers without further purification.BIA catalyst was readily prepared according to a previously reported procedure. 24The reaction progress was monitored by thin layer chromatography on silica gel-coated aluminium foils.The products were revealed under UV light (254 nm), followed by staining with 25 % phosphomolybdic acid solution in ethanol or with sulfuric vanillin and heating with a heat gun.Reaction products were purified by flash column chromatography using silica gel (230-400 mesh). 1 H NMR and 13 C NMR spectra were acquired on a Bruker Avance 250 (250 MHz for 1 H NMR and 63 MHz for 13 C NMR); Bruker Avance 400 (400 MHz for 1 H NMR and 101 MHz for 13 C NMR) or Bruker Avance 500 (500 MHz for 1 H and 126 MHz for 13 C NMR).Chemical shifts (δ) were reported in ppm and the coupling constants (J) in Hertz (Hz).Signal multiplicity was assigned as singlet (s), broad singlet (brs), doublet (d), double doublet (dd), double triplet (dt), double double doublet (ddd), double double triplet (ddt), triplet (t), triple doublet (td), quartet (q).High resolution mass spectrometry (HRMS) was performed using electrospray ionization (ESI) on a Waters Synapt mass spectrometer.Melting points were obtained using a Gehaka equipment model PF 1500 FARMA and were corrected.The compounds were named according to IUPAC rules using the software MarvinSketch version 16.11.21.

Figure 2 .
Figure 2. ORTEP diagram of compound 4c co-crystallized with chloroform, with 50 % probability displacement ellipsoids.The crystallographic details are available in the Supplementary Material.