Supporting Information A Maitland-Japp Inspired Synthesis of Dihydropyran-4-ones and their Stereoselective Conversion to Functionalised Tetrahydropyran-4-ones

All reactions were carried out under N2 unless otherwise specified. IR analyses were carried out on a ThermoNicolet IR100 FT-IR spectrometer using NaCl plates. Nuclear magnetic resonance spectra were recorded on a Jeol ECX-400 MHz, Jeol ECS-400 MHz or Bruker AV-500 MHz spectrometer at ambient temperature; chemical shifts are quoted in parts per million (ppm) and were referenced as follows: chloroformd, δH 7.26 ppm, δC 77.0 ppm; benzene-d, δH 7.15 ppm, δC 128.02 ppm. Mass spectrometry was performed by the University of York mass spectrometry service using ESI ionization. Microanalysis was performed by the microanalytical unit at the University of York. Melting points were determined using a Stuart SMP3 apparatus and remain uncorrected. Optical rotations were recorded at ambient temperature using a JASCO DIP-370 digital polarimeter. Thin layer chromatography was performed on glass-backed plates coated with Merck Silica gel 60 F254, and developed using ultraviolet light and ethanolic anisaldehyde, acidic aqueous ceric ammonium molybdate or basic aqueous potassium permanganate. Liquid chromatography was performed using forced flow (flash column) with the solvent systems indicated. The stationary phase was silica gel 60 (220-240 mesh) supplied by Fluorochem or silica gel Merck TLC grade 11695 supplied by Sigma-Aldrich. Dry solvents were obtained from a PureSlov PS-MD-7 solvent tower or distilled as follows: dichloromethane and toluene from calcium hydride, THF from sodium/benzophenone ketyl, diethyl ether from lithium aluminium hydride. All other solvents and reagents were used as received from commercial suppliers.


Introduction
Tetrahydropyran (THP) containing natural products, such as (−)-centrolobine, (+)-phorboxazole A and B, (−)-lasonolide A, and Civet cat secretion (Fig. 1), are an important class of synthetic targets because of their challenging architectural features, their biological activities and their limited availability from natural sources.In the cases of (−)-centrolobine, (+)-phorboxazole A and B, and (−)-lasonolide A, each has potent activity against a human disease, with (−)-centrolobine showing activity against the parasite responsible for leishmaniasis 1 and the phorboxazoles and lasonolide A showing potent anticancer activity. 2,3As such, these molecules have the potential to become the next generation of therapeutic agents if enough material can be provided to complete the required biological studies and satisfy the supply problem.The challenging molecular architectures of these compounds, coupled with their biological activities, have prompted many groups around the world to embark upon research programs aimed at the development of new methods for the construction of the tetrahydropyran rings found within them. 4 There have been significant developments in the formation of tetrahydropyrans by the Prins reaction 5 and the hetero-Diels-Alder reaction, 6 and these strategies have been applied with varying degrees of success to the synthesis of tetrahydropyran-containing natural products including (−)-centrolobine, (+)-phorboxazole A and B, and (−)-lasonolide A. 4 Over the last few years we have been interested in developing new methods for the synthesis of functionalised tetrahydropyran-4-ones 7 and the application of these methods to the total synthesis of tetrahydropyran containing natural pro- ducts such as (−)-centrolobine 8 and (+)-phorboxazole B. 9,10 Our work in this area focused on updating the venerable Maitland-Japp reaction, 11 initially as a two-pot process involving the addition of the Weiler dianion to an aldehyde in the first step, to be followed by the Lewis acid catalysed Knoevenagel reaction and oxy-Michael cyclisation in the second step. 12This in turn led to the development of a onepot procedure.When Chan's diene was used as the nucleophile, we found that we could effect a Lewis acid catalysed Mukaiyama aldol reaction and follow it with the Knoevenagel reaction and oxy-Michael cyclisation, without the need for isolation of the intermediate δ-hydroxy-β-ketoester adduct.This generated mixtures of 2,6-cis and 2,6-trans-tetrahydropyran-4-ones in good yields. 13Later we replaced Chan's diene with diketene and made the reaction pot, atom and step economic (PASE), 14 as well as asymmetric. 14,15owever, despite the utility of the Chan's diene and diketene versions of the Maitland-Japp reaction, it became apparent that there were a number of difficulties associated with them.Of primary concern was the formation of mixtures of the 2,6-cis and 2,6-trans diastereomers 1 and 2, which interconverted under the reaction conditions (Scheme 1).13b While these diastereomers could be separated via flash column chromatography and re-equilibrated to give the desired diastereomer, such a procedure was not ideal.Of secondary concern was the inherent difficulty in functionalising either the 3-or 5-positions of the tetrahydropyran-4-one ring.Treating the tetrahydropyran-4-one products with a base resulted in a retro-Michael reaction affording 3; 8 furthermore, after decarboxylation, it proved impossible to control the regioselectivity of enolate formation in the resulting decarboxylated tetrahydropyran-4-one 4 and hence formation of products 5 and 6 (Scheme 2).As such, the tetrahydropyran-4-one products from the Maitland-Japp reaction cannot be readily converted into the tetrahydropyrans found in the C20-C32 fragment of the phorboxazoles 16 or the A-ring of lasonolide A.
In order to overcome these problems we considered the possibility of developing a procedure to generate dihydropyran-4-ones 8, which would be more amenable to further functionalisation.Conjugate addition of a nucleophile to the double bond of the dihydropyran-4-one would generate an enolate which we hoped we could trap with an appropriate electrophile, thus generating a quaternary stereocentre.If the nucleophile was a hydride, then the resulting tetrahydropyran-4-one 9 would have the 2,6-cis relationship, and if the nucleophile was an organometallic reagent the resulting tetrahydropyran-4-one would have a tertiary stereocentre at C2 (Scheme 3).This paper builds on our earlier communication and fully details our studies in this area. 17

Formation of dihydropyran-4-ones
Our initial investigations focused on modifying the Maitland-Japp cyclisation to produce dihydropyran-4-ones 10.We achieved this by replacing the second aldehyde in the Maitland-Japp reaction sequence with the dimethyl acetal of a N,N-dimethyl amide (Table 1). 17As can be seen from Table 1, a wide range of δ-hydroxyβ-ketoesters 7 can be reacted with the dimethyl acetals of N,Ndimethyl acetamide or benzamide to generate dihydropyran-4-ones 10 in good to excellent yields.However, the scope of this approach is limited by the commercial availability and synthetic accessibility of such orthoamides.While the dimethyl acetal of N,N-dimethyl acetamide was commercially available, the corresponding dimethyl acetal of N,N-dimethyl benzamide required a two-step synthesis.This involved first reacting the N,N-dimethyl benzamide with dimethyl sulfate and then treating the resulting product with NaOMe in methanol. 18Thus, while unfunctionalised alkyl and aryl dimethyl acetals of N,N-dimethyl amides can be formed, this procedure cannot be used for any amides containing either Lewis acid or base sensitive functional groups.
In order to overcome this problem, we studied the use of orthoesters, which are more easily accessible than their orthoamide counterparts.We selected two commercially available orthoesters to study: trimethyl orthoacetate and trimethyl orthovalerate (Table 2).However, it is worth noting that functionalised orthoesters can be synthesised in two steps from the appropriate nitrile. 19e found that these orthoester Maitland-Japp reactions required heating under reflux, the presence of acetic anhydride as a dehydrating agent and a large excess of orthoester in order to achieve completion.However, the large excess of orthoester caused problems in the isolation of the dihydropyran-4-one products 10.We therefore investigated the use of microwave heating, 20 which enabled us to reduce the amount of orthoester to only 2 equiv.and still maintain reasonable yields.Microwave heating also reduced the reaction time from hours to a matter of minutes.

Conversion of dihydropyran-4-ones to 2,6-cis-tetrahydropyran-4-ones
Having developed the Maitland-Japp cyclisation to form dihydropyran-4-ones 10, we turned our attention to reduction of the double bond to form tetrahydropyran-4-ones 11.We predicted that addition of a hydride to the double bond would occur from a pseudo-axial trajectory, thus generating the 2,6cis-tetrahydropyran-4-one stereoselectively.A number of reducing agents were investigated, with L-Selectride® proving to be the best.Treatment of dihydropyran-4-ones 10 with L-Selectride® delivered tetrahydropyran-4-ones 11 as the sole products as mixtures of ketone and enol tautomers with excellent 2,6-cis-diastereoselectivity (Table 3). 17n the case of the 2-methyl tetrahydropyran-4-ones 11a-k a trace amount of the 2,6-trans-tetrahydropyran-4-one was formed, although this could be separated from the major 2,6cis-product by flash column chromatography using cyclohexane-ethyl acetate mixtures.We believe that the 2,6-trans products arose from a retro-Michael/Michael equilibration, rather than from pseudo-equatorial addition of a hydride.Indeed, we have seen this equilibration in these tetrahydropyran-4-ones previously, especially under Lewis or Brønsted acid conditions. 13With larger C2 substituents, the 2,6-transtetrahydropyran-4-ones were not observed; the 2-butyl tetrahydropyran-4-ones 11q and 11r were formed solely as the ketone tautomer.Interestingly, the 2-phenyl tetrahydropyran-4-ones 11l and 11m were formed exclusively as the enol tautomer.
The structures of the 2,6-cis ketone tautomers were elucidated by analysis of the coupling constants in 1 H NMR and nOe studies.Coupling constants of about 10 Hz were observed between H2/H3 and H5 ax /H6, indicating that the two pairs had trans-diaxial relationships and thus all of the protons occupied axial positions.Positive nOe correlations between H2 and H6 of 1.7-2.6%confirmed the 2,6-cis relationship.The 2,6-cis enol stereochemistry was also confirmed by positive nOe correlations between H2 and H6 of around 1.0-1.5%.

Synthesis of a constituent of Civet cat secretion
Having developed methods for the synthesis of dihydropyran-4-ones 10 and for their conversion to 2,6-cis-tetrahydropyran-4-ones 11, we looked to apply them to the synthesis of the small 2,6-cis-tetrahydropyran natural product that is found in the glandular secretions of the Civet cat (Viverra civetta) and is used in the fragrance industry. 21ur synthesis began with the Maitland-Japp formation of dihydropyran-4-one 10g in 72% yield using the orthoamide procedure.This was then treated with L-Selectride® to furnish the 2,6-cis-tetrahydropyran-4-one 11g in 67% yield in a 1 : 0.19 ratio of ketone and enol tautomers.Microwave mediated decarboxylation in wet DMF provided 2,6-cis-tetrahydropyran-4-one 12 quantitatively.Tetrahydropyran-4-one 12 was converted into tetrahydropyran 13, quantitatively, by formation of the dithiolane and removal of the benzyl group with BCl 3 •SMe 2 in CH 2 Cl 2 .Reduction of the dithiolane with RANEY® Ni and H 2 gave alcohol 14 in 68% yield.Alcohol 14 was then oxidized with Jones reagent to give carboxylic acid in 63% yield, thus completing the total synthesis of the Civet cat secretion natural product 15 in 7 steps (Scheme 4).

Synthesis of tetrahydropyran-4-ones with quaternary stereocentres
As the addition of L-Selectride® to dihydropyran-4-ones 10 generated an enolate, we wondered whether it would be possible to trap the enolate with a carbon electrophile.We envisaged that the enolate trapping should occur anti to the incoming hydride nucleophile; thus if MeI were used as an electrophile this should lead to structures containing the substitution found on the A-ring of (−)-lasonolide A, specifically the quaternary stereocentre.L-Selectride® was added to a solution of the dihydropyran-4-ones 10 in THF at −78 °C and after an hour MeI was introduced and the reaction was warmed to room temperature.
The desired 2,6-cis-tetrahydropyran-4-ones 16 were formed in moderate to good yields with alkylation at C3 and with the methyl substituent in an axial position (Table 4).The exceptions to this were dihydropyran-4-ones 10l and 10m where R 3 was a phenyl group.In these cases alkylation occurred on the C4 oxygen to give enol ethers 17l and 17m.The 2,6-cis stereochemistry was again confirmed by trans-diaxial couplings between H5 ax and H6 of around 11.0-12.0Hz and positive nOe correlations between H2 and H6 of 2.8%.Positive nOe correlations between H5 ax and the C3 methyl substituent of 1.2% showed that the methyl quench occurred from the expected pseudo-axial trajectory, anti to the addition of a hydride.

Synthesis of the tetrahydropyran A-ring of lasonolide A
With a procedure in place for the alkylation of the C3 position we could focus on completing a synthesis of a model A-ring of lasonolide A (Scheme 5). 22Tetrahydropyran 16t with the quaternary stereocentre at C3 was treated with an excess of LiAlH 4 in THF to reduce both the ketone and ester functional groups.This furnished diol 18, where a hydride had been delivered to the ketone in a pseudo-axial manner to generate the equatorial alcohol.The stereochemistry of the new alcohol was confirmed by analysis of the coupling constants that H4 had with both H5 ax and H5 eq .The coupling constant between H4 and H5 ax was 12.0 Hz, indicating a trans-diaxial relationship, while that between H4 and H5 eq was only 4.8 Hz (Fig. 2).Reduction of 16t with L-Selectride® in THF resulted in the formation of 19, where delivery of a hydride occurred from the pseudo-equatorial trajectory placing the hydroxyl group in an axial position.Once again, 1 H NMR coupling constants confirmed the stereochemistry.Now H4 had a coupling constant of 2.7 Hz to H5 ax and 5.7 Hz to H5 eq , indicating that H4 was indeed equatorial (Fig. 2).Treatment of 19 with LiAlH 4 reduced the ester to the primary alcohol, thus generating tetrahydropyranol 20, which has the substitution and relative configuration present in the A-ring of lasonolide A.

Synthesis of 2,2,6-substituted tetrahydropyran-4-ones from dihydropyran-4-ones
We next turned our attention to extending the scope of the nucleophile we could employ in the conjugate addition reac-tion.Gilman cuprates had been previously reported in the conjugate addition reaction to dihydropyran-4-ones. 23 In addition to these we extended the scope of the investigation to other nucleophiles (Table 5).
As can be seen from Table 5, when MeMgBr was used, the reaction generated essentially equal amounts of the 1,4-and 1,2-addition products 21 and 22.The inclusion of a CuBr 2 •SMe 2 additive did bias this in favour of the 1,4-addition product 21, but also resulted in the formation of 23 which presumably arose from an elimination reaction.Gilman cuprate (Me 2 CuLi) also resulted in both 1,4-addition and elimination products.However, when TMSCl was added to the reaction, 24 an increase in rate and selectivity for the 1,4-addition product 21 was seen.Finally, the use of a higher order cuprate was investigated but this did not lead to any further improvements and actually gave a sizable amount of the 1,2-addition product 22.As a result of these studies we opted for the use of Gilman cuprates.
We chose to investigate the reactions of Me 2 CuLi, Bu 2 CuLi, (H 2 CvCH) 2 CuLi and Ph 2 CuLi with a representative number of dihydropyran-4-ones 10 (Table 6).The Gilman cuprates all added from a pseudo-axial trajectory to form products with a 2,6-cis relationship between the new C2 substituent and H6, which was shown by positive nOe correlations of 4% in the case of the butyl, vinyl and phenyl substituents.Interestingly the tetrahydropyran-4-ones were actually formed as mixtures of three tautomers: the enol tautomer 25 and two ketone tautomers 21 and 24 which resulted from protonation of the intermediate enolate from either face.The product of the pseudo-axial protonation 21 had a positive nOe correlation of 1.6% between H3 and H5 ax , confirming the stereochemistry, whilst in the product of pseudo-equatorial protonation, H5 ax was shifted about 0.5 ppm downfield in 1 H NMR due to an interaction with the nearby axial ester substituent (Fig. 3).When Ph 2 CuLi was used as the nucleophile the enol-tautomer 25 was the only product.Where mixtures of tautomers occurred they could be converted into single enol acetate products 26 in good yields by treatment with acetic anhydride in pyridine at 40 °C.This conversion provided further support for our assignment of these 1,4-addition products as compounds 21, 24 and 25.

Conclusions
We have developed a new modification of the Maitland-Japp reaction using orthoamides and orthoesters, which provides access to a range of dihydropyran-4-ones 10 in good yields.These dihydropyran-4-ones 10 can be converted to 2,6-cis tetrahydropyran-4-ones 11 by the stereoselective addition of L-Selectride®.The intermediate enolate resulting from this addition can be trapped stereoselectively with either a proton or MeI to form tetrahydropyran-4-ones with a quaternary stereocentre at C3 16.The utility of these procedures was demonstrated by their use in the total synthesis of a constituent of the Civet cat secretion and for the synthesis of a model A-ring of lasonolide A. Treatment of the dihydropyran-4-ones 10 with a Gilman cuprate has led to the development of a procedure for the stereoselective formation of tetrahydropyran-4-ones 25 that are doubly substituted at the C2 position.Hence, we have overcome the difficulties inherent in the functionalisation of 2,6-cis-tetrahydropyran-4-one products of the Maitland-Japp reaction 1, and provided a route to the stereoselective construction of highly functionalised tetrahydropyran rings.

General methods
For general experimental details, including information on solvent purification and the spectrometers used in this research, as well as for procedures and spectroscopic and crystallographic data not reported below, see ESI. † General procedure for the synthesis of 2-methyl dihydropyrans N,N-Dimethylacetamide dimethyl acetal (0.16 mL, 1.08 mmol) was added to a stirred solution of δ-hydroxy-β-ketoester (0.54 mmol) in dry toluene (4 mL) at room temperature.The solution was stirred at room temperature and monitored by TLC.Upon completion of the reaction, the solvent was removed in vacuo.Purification by flash column chromatography ( petroleum ether-ethyl acetate) afforded the product.
Methyl 2-methyl-4-oxo-6-phenyl-      General procedure for Gilman cuprate addition to dihydropyran-4-ones An organolithium solution (0.41 mmol) was added to a suspension of copper iodide (38.7 mg, 0.20 mmol) in THF (1.17 mL) at 0 °C.The mixture was stirred at this temperature for 20 minutes and then cooled to −78 °C.Addition of TMSCl (0.08 mL, 0.64 mmol) was followed by addition of DHP (0.13 mmol) in THF (1.17 mL) at −78 °C.The reaction mixture was stirred at this temperature for 4 hours, then quenched with sat.aq.NH 4 Cl (1 mL) and allowed to warm to rt with vigorous stirring.The mixture was diluted with sat.aq.NH 4 Cl (10 mL) and extracted with EtOAc (4 × 15 mL).The combined organic extracts were washed with H 2 O (15 mL) and brine (15 mL), then dried over MgSO 4 and concentrated in vacuo.Flash column chromatography ( petroleum ether-ethyl acetate) afforded the products.