Synthesis of Medium Ring and Macrocyclic Acetylenic Lactones by the Ring Expansion of Oxabicycloalkenones

Lactonas acetilênicas de tamanho médio e macrocíclico 15a-e [6-decin-9-olídeo ( 15a), 7-undecin-10-olídeo ( 15b), 8-dodecin-11-olídeo ( 15c), 12-hexadecin-15-olídeo ( 15d) e 5-decin-9olídeo (15e)] foram preparadas a partir de oxabiciclo-alquenonas 7a-d e 2, respectivamente, pela expansão de anel das tosil-hidrazonas 16a-e, efetuada pela reação com N-bromo-succinimida, sob condições rigorosamente controladas. A hidrogenação (Pd-C, H 2) completa da ligação tripla forneceu lactonas racêmicas saturadas: 9-decanolídeo (foracantolídeo I, 6, de 15a e 15e), 10-undecanolídeo, 11-dodecanolídeo (di-hidro-recifeiolídeo) e 15-hexadecanolídeo. As tentativas de converter di-hidropironas 7a,d nas respectivas lactonas acetilênicas 15 ,d, via cloro-hidrinas 8a,d e/ou clorocetolactonas 9a,d, foram apenas parcialmente bem sucedidas.


Introduction
Medium ring and macrocyclic lactones are important natural products and we have developed several methods for their synthesis 1,2 .Thus, in 1990, we described the preparation of desoxydiplodialide D (5) and phoracantholide I (6) from cyclohexane-1,3-dione 3 (1) by a novel ring expansion of the vinylogous lactone intermediate 2, as illustrated in Scheme 1.
The success of this new ring enlargement protocol encouraged us to test its scope with several other oxabicycloalkenone systems shown in Scheme 2. While the process worked well with the dihydropyrones 7a,d, affording the corresponding chlorohydrins 8a,d and subsequently the chloroketolactones 9a,d, it was not satisfactory with the phenolic substrates 11 to 14, which failed either to undergo a clean reaction with HOCl or satisfactory ring expansion to the desired chloroketolactone 4 .Apart from this lack of wide applicability, even in the successful cases (7a,d), the chloroketolactones 9a,d were contaminated with the corresponding chlorohydrins (8a,d) and did not undergo the reductive dechlorination (Zn, AcOH, ∆ or ultrasound) 4,5 to the desired ketolactones 10a,d, reverting instead to the starting vinylogous lactones 7a,d.

Aims & Objectives
In the face of this failure, instead of looking for alternative methods for the removal of the chlorine atom, we thought it better to exploit its presence for the generation of a regiospecific triple bond according to a method, also under development in our laboratories 6 , whereby an α-chloroketo function is transformed into an acetylenic linkage, as depicted in Scheme 3.Moreover, it appeared mechanistically feasible to prepare the same acetylenic lactones 15a-d from the chlorohydrin 8a-d, as shown in item C of Scheme 3. Thus, we hoped to obtain the acetylenic lactones 15a-d from the respective mixtures, contain- 604  Mahajan & Resck J. Braz.Chem.Soc.ing both the chlorohydrin (8a-d) and the chloroketolactone (9a-d).
Apart from this route, we also envisaged preparing the acetylenic lactones 15a-e from the heteroannular and homoannular oxabicycloalkenones (2, 7a-d) via their tosylhydrazones (16a-e), involving a fragmentation reaction patterned after that of the bicycloalkenones 7 , as illustrated in Scheme 4. These expectations have been largely fulfilled 5,[8][9][10] now and herein we describe the details of these investigations.

Results and Discussion
When we tried to prepare oxabicyclodecenone 7a by the known procedure 11 , involving the acylation of 1-morpholinocyclohexene with crotonyl chloride in the presence of triethylamine, followed by hydrolysis and basic isomerization of the resulting probable mixture (Scheme 5; brackets), we obtained only 30% of the desired product 7a, instead of the 73% reported earlier 11 .However, on carrying out the isomerization/cyclization under acidic conditions (AcOH:H2O:conc.HCl, 1:1:1), we could raise the yield to 75-80% (see Experimental).
Following this slightly modified procedure, we were also able to improve the yield of oxabicyclohexadecenone 7d from the reported 12 16-19% to 30%.Moreover, the melting point of our product, 54-56 °C, is much higher than that described earlier 12 : 34 °C.
However, even this improved procedure gave only poor yields in the case of cycloheptanone and cyclooctanone [7b (10%), 7c (20%)], and required chromatographic purification.Both 7b and 7c are new compounds, having spectral (IR, 1 H-NMR) absorptions characteristic of the known oxabicycloalkenones 7a and 7d.Moreover, both gave the corresponding tosylhydrazone (16b, 16c), as described in the experimental part of this work.
The conversion of the vinylogous lactones 7a,c,d into the corresponding chlorohydrins 8a,c,d proceeded in almost quantitative yield, employing the household sodium hypochlorite solution under slightly acidic conditions: Qboa, AcOH, EtOAc.In the case of 7a, a solid chlorohydrin fraction 8a, free of chloroketolactone 9a, could be isolated in 50% yield, but the other substrates furnished only gummy products, containing both the chlorohydrin and the corresponding chloroketolactone (TLC, IR, 1 H-NMR).The latter (9a,c,d) arise from the former (8a,c,d) by a retro-aldol reaction provoked, probably, during the basic washing of the reaction mixture with aq.Na2CO3.Isomerization of the relatively pure chlorohydrin 8a, or that of the mixtures 8c,d, to the corresponding chloroketolactones 9a,c,d was conducted under mild conditions (DABCO, CHCl3, reflux) in order to avoid undesirable side reactions, such as the Favorskii rearrangement.The final product was slightly contaminated with the respective chlorohydrin and some unidentified byproducts (TLC, IR, 1 H-NMR).
The attempted reductive dechlorination (Zn, AcOH) of two such crude chloroketones (9a,d), using heat or ultrasound 4,5 , resulted in the recovery of the starting dihydropyrones 7a,d, instead of the desired ketolactones 10a,d.The regeneration of the vinylogous lactones 7a,d involves, most probably, the intramolecular (transannular) acylation of the organo-zinc intermediate, in preference to its protonation, as encountered earlier 3 , albeit as a minor process, in our first example outlined in Scheme 1.
In the face of this failure, we tried to convert these chlorohydrins, or their mixture containing the respective chloroketolactones, into the corresponding acetylenic compounds 15a,c,d, as pointed out earlier in the introductory section and illustrated in Scheme 3.However, the procedure was successful, only moderately, with the chlorohydrin 8a, affording the acetylenic lactone 15a in ~50% isolated yield, after chromatographic purification.In other cases, the reaction mixture showed a number of byproducts and only a small quantity of the desired acetylenic compound (TLC), whose laborious purification was deemed unnecessary, as, taking into account the fact that, in a concurrent study we had developed 8 a novel and high-yield protocol for the preparation of acetylenic lactones 15a-e from tosylhydrazones 16a-e (Scheme 4 and vide infra).
Fortunately, the heteroannular and homoannular vinylogous lactones (2, 7a-d) furnished the corresponding tosylhydrazones (16a-e) under the standard conditions usually employed in the case of ordinary ketones: MeOH, TsNHNH2, cat.H + , reflux.As mentioned earlier and illustrated in Scheme 4, their fragmentation reaction was patterned after that of the tosylhydrazones of some bicycloalkenones, described by the Swiss workers 7 in R=H *: A bulky alcohol (t-butanol) favors the formation of hemi-ketal D, which is converted by ring expansion directly into the lactone 15a.1979.The success of the desired ring expansion depends vitally on the 1,4-addition of the nucleophile (ROH, H2O) to the intermediate chemical species A/B, instead of the 1,2-addition which reverts them to the starting ketone, as illustrated for enone 7a in Scheme 6; incidently, the use of NBS for the regeneration of ketones from their tosylhydrazones was reported originally by Rosini 16 in 1974.
We would like to highlight here our preliminary but very instructive experiments with tosylhydrazone 16a.On carrying out its fragmentation reaction according to the recommended procedure 7 (16a, 2-BuOH, acetone, -15 °C; NBS, -15 °C; aq.NaHSO3, 55 °C, etc.), we obtained the desired lactone 15a, contaminated with the expected starting enone 7a and some other minor impurities (TLC, IR, 1 H-NMR).To improve the reaction, we next resorted to freshly crystallized and dessicator-dried NBS.To our surprise, the major product in this experiment turned out to be the 2-butyl hydroxy-ester ( 17) of the ring-opened lactone.Subsequently, chromatographic separation and spectral (IR, 1 H-NMR) identification of the reaction products proved them to be the starting enone 7a, the desired lactone 15a, the hydroxy-ester 17, and 2-butyl tosylate 18.In fact, the last mentioned compound (18) was admixed with the acetylenic product 15a and could not be separated by chromatography or distillation.However, the spectral absorptions left no doubts about its identity.Nevertheless, we prepared an authentic sample of 2-butyl tosylate (2-BuOH, TsCl, pyridine) and found out that, apart from having the expected spectral signals, it had the same Rf as that of the acetylenic lactone 15a.
The formation of 2-butyl tosylate requires some explanation.It could conceivably arise by the attack of 2-BuOH on the sulphonyl group of the reactive species A/B, or, most probably, by its tosylation with tosyl bromide (TsBr), generated from the byproduct p-toluenesulphinic acid (TsH) on reaction with the excess NBS present in the reaction mixture (Scheme 7).
The isolation of 2-butyl hydroxy-ester 17 proves inequivocally the 1,4-addition of the nucleophile (2-BuOH) to the reactive species A/B and the intermediacy of the adduct C, which after ring expansion (fragmentation) and the capture of a molecule of water, during the aqueous treatment, would lead to the hemi ortho-ester E. The latter can then afford either the desired lactone 15a or the unwanted hydroxy-ester 17 (Scheme 6).
Moreover, the formation of the 2-butyl hydroxy-ester 17 to suggested us that if we employed in the above reaction a small amount of water incorporated in a bulky alcohol (t-BuOH), there should result the intermediate hemi-ketal D, instead of the mixed ketal C, which upon ring expansion would transform directly into the desired acetylenic lactone 15a (Scheme 6).Consequently, we carried out the aforementioned reaction using a mixture of t-BuOH:H2O (9:1).
After the usual work-up and chromatographic purification, we obtained the product 15a in 90-95% yield.This improved procedure was then successfully employed for the preparation of other acetylenic lactones 8 : 15b-d (85-95%) and 15e (68%).
To our surprise and disappointment, the tosylhydrazone 17 of the phenolic subtrate 12 did not undergo the above fragmentation reaction satisfactorily; affording the recovered tosylhydrazone (mixed m.p.) along with a complex mixture of unidentified products.
The details regarding the conversion of these acetylenic lactones into the corresponding ethylenic compounds, containing either the Z or E double bond, has been reported 10 recently as well as their transformation into the acyclic insect pheromones 25 .
In conclusion, it is very satisfying to record that we have been able to develop an efficient procedure for the preparation of medium to macrocyclic acetylenic lactones 15a-e from both the heteroannular and homoannular oxabicycloalkenones (2, 7a-d), via their tosylhydrazones 16a-e, involving a fragmentation/ring expansion reaction provoked by NBS, under strictly controlled conditions.The experimental details are given below.

Experimental
Reagent grade chemicals and solvents were used as received from the commercial suppliers, unless noted otherwise.All reactions were monitored routinely by thin layer chromatography (TLC: silica gel, revealed by I2 vapours).Organic extracts were dried over anhydrous Na2SO4 and evaporated under reduced pressure on a rotary evaporator.Bransonic ultrasonic cleaner (Model 1210 or 2210; 47 ± 6 KHz) was used to conduct some heterogeneous reactions.Temperatures in the short path distillations refer to the air bath.Chromatographic purifications were conducted by dry-column flash chromatography 26 on silica gel (Merck, 60 Å, 230-400 mesh).Melting points were determined on a Kofler block and are uncorrected.IR spectra of liquid samples (neat films) and solids (KBr disks) were recorded on a Nicolet 5ZDX-FT spectrometer.Raman spectra were obtained on Jarrel Ash spectrometer, model 25-300, or Jobin-Yvon instrument, model 1000, both using argon laser.Routine 1 H-NMR spectra, reported in the experimental text were obtained on a Varian EM-390 (90 MHz) instrument as CCl4 solutions, unless noted otherwise.The 13 C spectra of the acetylenic lactones were recorded in CDCl3 either at 300/75 or 200/50 MHz, as shown in the Table 1.Gas chromatographic (GC) analyses were carried out on a Varian Aerograph, Model 1440, using 15% FFAP column (3 mm x 2 m), at 210 to 230 °C, swept with N2 (40 mL/min).Other experimental details are given below.

Preparation of Oxabicycloalkenones 7a-d. General Procedure
To a solution of 1-morpholinocycloalkene (50 mmol) and triethylamine (10.50 mL, 7.60 g, 75 mmol) in chloroform (100 mL), stirred magnetically and kept around 35 °C (tap water), and under an anhydrous atmosphere of N2 (CaCl2), was added, during 1 h, a solution of crotonyl chloride (6.23 mL, 6.79 g, 65 mmol) in chloroform (50 mL).The resulting reddish-brown mixture was kept on a warm water bath (38-40 °C) for 24-36 h, when dil. HCl (50 mL) and 95% ethanol (10 mL) were added and the mixture refluxed for 8-10 h, under vigorous stiring, to effect the hydrolysis.After cooling, the organic layer was separated and the aqueous portion (pH 1) was extracted with chloroform (3 x 50 mL).The combined organic extract was washed succesively with distilled water (3 x 50 mL), satd.solution of sodium bicarbonate (50 mL) and brine (50 mL).Drying and evaporation of solvent gave a reddish-brown liquid, which upon short path distillation, 110-120 °C/0,5 Torr, afforded a yellowish liquid, showing several spots on TLC.Thus, it was subjected to isomerization/cyclization reactions by refluxing, under stirring, for 4-6 h in a mixture of AcOH:Conc.HCl:H2O (1:1:1) (1 mL of each component for 1 g of the distillate).The cooled reaction mixture was diluted with water (20-30 mL) and extracted with ethyl acetate (3 x 50 mL).The combined extract was washed with distilled water (3 x 50 mL), satd.solution of sodium bicarbonate (50 mL) and brine (50 mL).Drying and evaporation of solvent gave the crude product, which was purified as described for the individual members.
Purification of the crude product by its conversion into the corresponding tosylhydrazone (vide infra), reproduced the same yield.

Preparation of Chlorohydrins 8a,c,d. General Method
A solution of household sodium hypochlorite (Q-boa; 0.65-0.70molar, 6 mL) was added to a well-stirred solution of the oxabicycloalkenone 7a,c,d (1 mmol) in acetic acid (1.5 mL) and ethyl acetate (15-20 mL).After stirring at room temperature for 30-40 min, the organic phase was separated and washed successively with water (2 x 10 mL), sodium carbonate solution (10 mL) and brine (10-15 mL).After drying and evaporation of the solvent, there was obtained a white solid (8a) or a slightly yellowish gum (8c,d), in a quantitative yield, which was characterized as follow.
Procedure B: The reaction mixture described in Procedure A was subjected to ultrasound irradiation for 1 h, there being liberation of gas and the mixture turned orange colored.It was worked-up and purified just as described in Procedure A. The lactone 15a was obtained in 50% yield.

Attempted Preparation of Acetylenic Lactones 15c,d from Chlorohydrins 8c,d
Chlorohydrins 8c or 8d, contaminated with the corresponding chloroketolactone 9c or 9d, when subjected to the ring expansion protocol described in Procedure A or B above, resulted in a complex mixture of products, containing only a small quantity of the desired lactone 15c or 15d (TLC, IR, 1 H-NMR), whose purification was deemed unworthy in face of their much better preparation from the respective tosylhydrazones, described earlier.

Attempted Preparation of Acetylenic Lactone 15a from Chloroketolactone 9a
The crude chloroketolactone 9a, obtained by the isomerization of chlorohydrin 8a (vide supra), when allowed to react according to Procedure A or B gave poor results in comparison to chlorohydrin 8a, thus discouraging the extension of the methodology to other chloroketolactones.

Catalytic Hydrogenation of the Acetylenic Lactones 15a-e to the Saturated Lactones. General Procedure
The acetylenic lactones 15a-e (1 mmol) dissolved in hexane (10 mL), containing 10% Pd-C (20-30 mg), were hydrogenated in a Parr apparatus (2-3 atm) for 4-6 h, when there was no more starting material (TLC).After filtering the catalyst and evaporation of hexane, the saturated lactones were obtained as colorless liquids, in 95-100% yield.Their other characteristics are described below.

:Scheme 6 .
Scheme 6. Mechanism for the ring expansion of tosylhydrazones of oxabicycloalkenones8  .