摘要:

Polyamino acids, such as polyleucine, are able to catalyse the asymmetric epoxidation of α,β-unsaturated enones1by basic H2O2, a transformation known as the Juliá–Colonna reaction (Scheme 1).1To date, most of the epoxides2prepared have been of the type R1 = aryl or alkyl and R2 = aryl or vinyl, for which ees are generally >80%.2The first part of this communication shows that enones1where R1 = aryl and R2 = alkyl can also be epoxidised with good stereocontrol, under the previously described biphasic reaction conditions.3Reagents and conditions: (a) urea–H2O2, DBU, poly-L-leucine or SiO2–poly-L-leucine, THF, rt.Grignard reagents add to the carbonyl group of α,β-epoxyketones in a highly stereoselective manner; diastereoselectivities >99 ∶ 1 in favour of the product of Cram-chelate addition are typically observed in the generation of the tertiary alcohols.4Such enantiomerically enriched epoxyalcohols have, in the past, been formed by a less efficient, five-step Sharpless asymmetric epoxidation–oxidation–alkylation–oxidation–alkylation strategy.5,6It has been shown that when treated with stoichiometric quantities of certain Lewis acids such epoxyalcohols undergo a semi-pinacol rearrangement to afford aldols. A range of Lewis acids including TiCl4,5BF3·OEt2,6–8alumina 7and SnCl4 9has been used to this effect. The corresponding silyl ethers exhibit similar reactivity when treated with BF3·OEt2,10TiCl4 5and SnCl4.10In the case of less reactive substrates, Ti(i-PrO)2Cl2 6and Ti(i-PrO)3Cl 10have been used. Suzuki has reported a catalytic variant of the silyl protected epoxyalcohol rearrangement using 2–5 mol% TMSI or TMSOTf.11When this work was initiated no catalytic variant applicable to unprotected epoxyalcohols had been reported. However, recently Tuet al.have shown that ZnBr2can catalyse such semi-pinacol rearrangements.12In the second part of this paper we show that rare earth triflatescan be employed as Lewis acid catalysts in a similar manner.Rare earth triflateshave been shown to be viable alternatives to more conventional Lewis acids for a number of transformations, including a variety of epoxide opening reactions.13The principal advantages of such triflates are their stability to moisture, low toxicity, ease of handling, recyclable nature and ability to function in a catalytic capacity.14Poor reactivity, under previously reported rearrangement conditions, in the case of some of our substrates, prompted an investigation into the utility of rare earth triflates as catalysts for epoxyalcohol semi-pinacol rearrangements.The IUPAC name for triflate is trifluoromethanesulfonate.A number oftrans-β-alkylenones were prepared in order to provide substrates for developing the alkylation–rearrangement methodology. The β-ethylenone3was preparedviaa Wittig reaction between propionaldehyde and benzoylmethylene triphenylphosphorane and the β-benzylenone5was synthesised as previously described.15The γ-benzyloxyenone4was prepared by application of chemistry previously utilised for the ethyloxy-substituted analogue.16Thus, 2-benzyloxyethanol was oxidised under Swern conditions and the resulting aldehyde was treated with benzoylmethylene triphenylphosphorane to afford the desired enone4in 38% yield over the two steps.Oxidation of the β-ethylenone3under biphasic conditions,3using urea–H2O2and DBU in THF, catalysed by silica-immobilised 17poly-L-leucine (416 mg mmol−1enone), gave the epoxide6in 64% yield and 60% ee after 17 hours (Scheme 2). The use of 2-tert-butylimino-2-diethylamino-1,3-dimethylperhydro-1,3,2-diazaphosphorine (BEMP) 18instead of DBU allowed a sub-stoichiometric amount of base to be employed; under optimum conditions, using 3 mol% BEMP, the epoxide6was obtained in 86% yield and 77–79% ee. Further reduction in the amount of BEMP led to a significant loss in stereoselectivity, for example with 0.1 mol% the epoxide6was generated in only 74% ee. Treatment of the enone4under similar conditions (20 mol% BEMP) gave the epoxide7in 79% yield and 93% ee. On the other hand, epoxidation of the enone5proceeded in a lower yield due to the formation of a dimeric side product, presumablyviaMichael addition of the anion generated by deprotonation of one molecule of5to a second molecule of the enone. However, when the reaction was performed with additional silica supported catalyst (2 g mmol−1enone) none of the dimeric material was generated, leading to an improved yield of the epoxide8(75%) with an ee > 90%.Complete resolution of the two enantiomers of7using chiral HPLC, chiral GC or chiral shift NMR was not achieved. The quoted ee of >90% is an estimate from the partly resolved HPLC analysis.Reagents and conditions: (a) urea–H2O2, DBU, SiO2–poly-L-leucine, THF, rt; (b) urea–H2O2, BEMP, SiO2–poly-L-leucine, THF, rt; (c) MeMgBr, THF, Et2O, −78 °C; (d) PhMgBr, THF, Et2O, −78 °C.Treatment of the epoxyketones6–8with methylmagnesium bromide or phenylmagnesium bromide furnished the corresponding alcohols9–12in 90 to 96% yield; in each case only a single diastereomer was observed in the1H NMR spectrum. The stereochemistry of the products was assigned by comparison with a previously reported example in which the configuration was determined by X-ray crystallography.19Using epoxyalcohol9, three readily available metal triflates, Sc(OTf )3, La(OTf )3and Yb(OTf )3, were tested as rearrangement catalysts employing dichloromethane and acetonitrile as solvents. Initial reactions were performed using 20 mol% of the triflate catalysts. All three of the Lewis acids were found to catalyse the migration of the aryl group of9to afford the aldol13(Scheme 3). Dichloromethane was found to be the better solvent, with the product proving clean enough to obviate the need for additional purification (Table 1). The rate of reaction catalysed by the three Lewis acids was found to decrease in the order Yb(OTf )3 > Sc(OTf )3 > La(OTf )3,with the former proving to be the most synthetically useful, generating the α-phenyl-β-hydroxy ketone13in 99% yield after 3 hours.The semi-pinacol rearrangement of9catalysed by various rare earth triflates.CatalystMol% catalystTimeYield (%)Sc(OTf )3206 h90La(OTf )32010 h90Yb(OTf )3203 h99Yb(OTf )31515 h97Yb(OTf )3103 days87Yb(OTf )356 days87Reagents and conditions: (a) Yb(OTf )3(20 mol%), CH2Cl2, rt.Further experiments were performed in order to determine the effect of reducing the molar percentage of Yb(OTf )3. When less than 15 mol% Yb(OTf )3was employed the reaction took three or more days to run to completion and the crude product contained significant impurities. The optimised conditions of 20 mol% Yb(OTf )3in DCM were then applied to the enones10,11and12with the corresponding aldols14,15and16being formed in 82, 100 and 97% yield respectively.Previous semi-pinacol rearrangements of epoxyalcohols have been reported to proceed with inversion of stereochemistry at the site of epoxide opening. X-Ray analysis of a single crystal of the aldol16demonstrated that this is also the case in the present study (Fig. 1).X-Ray crystal structure of 3-hydroxy-1,2-diphenylpentan-1-one (16). Thermal ellipsoids drawn with 50% probability.Cyclic trisubstituted enones such astrans-2-benzylidene-1-tetralonecan be epoxidised with good enantioselectivity under polyleucine catalysis.20Treatment of17, generated by such an epoxidation, with methylmagnesium bromide gave the tertiary alcohol18as expected. Unexpectedly, treatment of18with 20 mol% Yb(OTf )3in DCM furnished benzaldehyde (20) (identified by1H NMR and TLC) and an α-methyltetralone (40% yield). Analysis of the latter by1H NMR spectroscopy showed the material to be 1-methyl-2-tetralone (19) by comparison with literature data 21and with an authentic sample. In particular, a quartet atδ3.53 (J7 Hz), due to the benzylic proton coupling to the methyl group, is a distinctive feature of the spectrum and clearly excludes the possibility of having formed 2-methyl-1-tetralone (the product of methyl group migration and retro-aldol reaction) (Scheme 4).The IUPAC name for 1-tetralone is 3,4-dihydronaphthalen-2(1H)-one.Reagents and conditions: (a) MeMgBr, THF, Et2O, −78 °C; (b) Yb(OTf )3(20 mol%), CH2Cl2, rt.The mechanism of this latter reaction is under investigation. At this point, we have established that open chain analogues behave differently; the epoxyalcohol21forms the enone22on treatment with rare-earth triflates (Scheme 5).In summary, we have demonstrated the applicability of modified Juliá–Colonna epoxidation conditions to β-alkyl substituted enones. The resultant epoxyketones have been converted into aldols by Grignard addition followed by semi-pinacol rearrangement catalysed by Yb(OTf )3. The newly developed rearrangement conditions constitute an efficient, convenient and mild alternative to previously reported procedures.

ISSN:1472-7781    

DOI:10.1039/b101838h

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

Trifluoromethyltrimethylsilane (Ruppert's reagent, CF3SiMe3) has recently attracted attention as a source of the trifluoromethyl functional group, and has been widely used for the conversions of aldehydes and ketones to trifluoromethyl alcohols,1of esters to trifluoromethyl ketones 2and of azirines to trifluoromethylaziridines.3In particular, the reactions of acylsilanes with CF3SiMe3have produced the versatile intermediate, difluoroenoxysilane, which is readily converted to difluoromethyl ketones, alcohols and their derivatives.4These effective trifluoromethylations are considered to be due to the strong nucleophilicity of the trifluoromethanide (CF3−) analog formed upon initiation by the fluoride ion. However, the trifluoromethanide liberated in these reaction systems is also astrong base (pKaof HCF3 = 31) and this usually makes it difficult to control these reactions, which give low yields of the products.5We have fortunately found that the reactions of acetylene with CF3SiMe3–Bu4NF effectively produce trimethylsilylacetylenesviathe corresponding acetylides. The trimethylsilyl group can be conveniently used as a good protecting group of terminal acetylenes.6Furthermore, they can then be converted to other useful compounds such as the trimethylsilylethynyl ketones,7ethynyl sulfones,8vinyl silanes,9and allenyl silanes.10However, the carbon–silyl bonds are easily cleaved upon usual work-up or purification of the products. If convenient methods for the trimethylsilylationof the various types of terminal acetylenes could be found, they would be useful as a good protection procedure in the syntheses of various complex targets. We now report the novel trimethylsilylation of acetylenes using CF3SiMe3–Bu4NF.

ISSN:1472-7781    

DOI:10.1039/b101537k

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionEnantioselective recognition remains a major challenge for host–guest chemists.1The ability to discriminate between enantiomers, using synthetic receptors, would, for example, allow separation of racemates by selective transport across a membrane.2We, and others, have previously utilised thioureas as a binding site for the carboxylate functionality,3and incorporation of thioureas into macrocyclic structures has produced selective receptors for carboxylate derivatives such as amino acids.4Such receptors are, however, structurally rather complex and demanding to synthesise. In an effort to produce simpler enantioselective receptors for carboxylate derivatives we have now prepared a series of acyclic thioureas and have examined their ability to bind a range of amino acids.5In these novel receptors the thiourea functionalityis linked to aryl amides to provide additional hydrogen bonds to the carboxylate moiety and it was anticipated that the formation of a well defined set of four hydrogen bonds around the carboxylate, binding to both thesynandantilone pairs of the carboxylate,6should also serve to create a cleft which, when chiral building blocks are incorporated, might discriminate effectively between enantiomeric guests (Fig. 1). The use of pyridyl amides might further help to preorganise the chiral cleft since the desired U-shaped conformation may be stabilised by weak hydrogen bonding of the amide and thiourea hydrogens to the pyridyl nitrogen.7In order to test this hypothesis we prepared the achiral receptors5and6. We also prepared two chiral pyridyl receptors7and15, the former incorporating a primary amide, derived from phenylalanine, to provide further hydrogen bonding functionality, and the latter incorporating an electron deficient aromatic amide to provide aromatic interactions with guests with aromatic sidechains.Proposed binding of carboxylates by a bispyridylthiourea. R*represents a chiral group.

ISSN:1472-7781    

DOI:10.1039/b102298a

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionPolycyclic aromatic hydrocarbons (PAHs),e.g.phenanthrene1and benzo[c]phenanthrene2, and the corresponding aza-, oxa- and thia-heterocyclic analogues are ubiquitous in the environment as a result of partial combustion of organic material including wood and fossil fuels. Their biodegradation in animals, plants and fungi by eucaryotic monooxygenases generally involves arene oxides,e.g.3and4, and in the presence of epoxide hydrolases these are in turn converted totrans-dihydrodiols,e.g.5and6,Scheme 1.1Bacteria commonly utilize a complementary biodegradation pathway for arenes, relative to animals, plants and fungi, involvingcis-dihydrodiolsas initial metabolites and procaryotic dioxygenases as biocatalysts,Scheme 1.2–4Although bacterial biodegradation has been identified as a major method for the removal of arenes from the environment, the majority of procaryotic metabolism reports have focused on the smaller members,e.g.mono-, bi-, and tri-cyclic arenes.2–4Studies of the bacterial biodegradation of larger PAHs,e.g.tetra- and penta-cyclic arenes are of importance in view of their increased recalcitrance and threat to health.Angular members of the PAH series may contain both bay (cf.phenanthrene,1) and fjord (cf.benzo[c]phenanthrene,2) regions. While the bacterial biotransformation of larger PAHs and their heterocyclic analogues, containing one or more bay regions has been investigated,3,4studies of arenes with fjord or pseudo-fjord regions (Schemes 1 and 2) have not previously been reported. The low solubility of larger PAHs in water, allied to the inability of the more widely used dioxygenase enzymes to accept such substrates,e.g.toluene dioxygenase (TDO) and naphthalene dioxygenase (NDO), are assumed to be limiting factors; isosteric heteroarenes being more water-soluble than carbocyclic arenes are expected to biodegrade more efficiently.Bacterial metabolism of the carcinogenic angular PAHs benz[a]anthracene,5,6chrysene 7,8and benzo[a]pyrene 5and the heterocyclic analogues benzo[b]naphtho[2,1-d]thiophene 8and benzo[c]phenanthridine 8has been studied in our laboratories usingSphingomonas yanoikuyaeB8/36, a mutant strain containing biphenyl dioxygenase (BPDO) and deficient incis-diol dehydrogenase activity. These studies have confirmed that bacterial dioxygenase-catalysedcis-dihydroxylation of angular PAHs occurs in a regioselective manner to yield bay-regioncis-dihydrodiols as major,5,6or exclusive,5,7,8bioproducts . A second sequentialcis-dihydroxylationwas also found to occur in the tetracyclic arenes chrysene and benzo[b]naphtho[2,1-d]thiophene at the other bay region in these substrates.8Angular PAHs, containing a fjord region, are more congested compared with arenes containing a bay region, and thus some are found to be non-planar in the crystalline state. This study indicates that bacterialcis-dihydroxylation of tetracyclic arenes and tetrahydroarenes occurs with optimal regio- and stereo-selectivity within the congested fjord (arenes2,8, and9) or pseudo-fjord (arenes12and13) regions using biphenyl dioxygenase (BPDO) fromS. yanoikuyaeB8/36 or a naphthalene dioxygenase (NDO) found to be present in a mutant strain (9816/11) ofPseudomonas putida. Furthermore, a novel tandem biotransformation pathway involving bis-desaturation andcis-dihydroxylation of the tetrahydrothiaarene13was found to yield a fjord-regioncis-dihydrodiol regioisomer17which was not detected during biotransformation of benzo[b]naphtho[1,2-d]thiophene9.

ISSN:1472-7781    

DOI:10.1039/b101833g

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

Results and discussionβ-Lactam containing oxazolidinonesTo date, we have utilized two β-lactam based oxazolidinones4a/4b 13and7 14as sources of azomethine ylide reactivity for the synthesis of carbapenams, carbapenems, penams, penems, and oxapenams.2dThe chemistry described in this paper will focus on the more readily available derivatives4a/4b, which are prepared on a multigram scale starting from natural clavulanic acid. Oxazolidinones4a/4bare readily purified by crystallization but are sensitive to moisture and are not readily amenable to chromatography. Almost all of the chemistry described in this paper was carried out using the 4-nitrobenzyl ester derivative4a, although it should be appreciated that the corresponding benzyl ester4bbehaves in an essentially identical fashion.The enantiomerically pure oxazolidinone4givesracemiccycloadducts. In contrast, with theC(6)substituted variant7, theO-silyl protected (6S )-(1-hydroxyethyl) side chain is retained throughout the cycloaddition process, and this feature serves to maintain overall stereochemical integrity leading ultimately to enantiomerically pure cycloadducts. Nevertheless, the way in which oxazolidinones4a/4bformed racemic products proved to be highly significant in our mechanistic investigations.Synthetic and mechanistic experimentsIt became clear that the reactivity of4a(and also7) was not consistent with the general mechanistic pathway established by Grigg 10forN-alkyloxazolidinone derivatives (as outlined inScheme 2). In particular,4awas relatively stable towards thermolysis and when unreactive (usually electron rich) dipolarophiles were employed,4awas recovered even after prolonged reaction times. Furthermore, when dipolarophiles of different reactivitye.g.N-phenylmaleimide (NPM)vs. PhC&z.tbd;CH were used, a significantly longer reaction time was required for the less reactive PhC&z.tbd;CH to achieve (i ) a comparable conversion of4aand (ii ) a similar yield of cycloadduct.The higher reactivity of NPM as compared to PhC&z.tbd;CH was easily demonstrated by a direct competition experiment. Also, a number of solvents have been successfully employed in the β-lactam cycloaddition process, including MeCN, PhMe, EtOAc, a range of chlorinated and ether solvents, DMF, DMSO, and CF3CH2OH. Acetonitrile proved to give consistently cleaner and faster reactions and higher yields.This suggested that it was the reactivity rather than the actual generation of the dipolarophile that was important and, in turn, raised the possibility of cycloaddition as the rate-determining step. We never observed (e.g.by1H NMR) the build-up of any intermediate dipolar species in reactions involving less reactive dipolarophiles and these results prompted a series of more detailed experiments designed to probe this mechanistic enigma.A critical observation was made when thermolysis of4awas carried outin the absence of a dipolarophile trap:the anticipated decomposition of the oxazolidinone didnotoccur.Oxazolidinone4adoes decompose slowly on prolonged heating at 80 °C and more rapidly at temperatures in excess of 110 °C. Extended heating of4a(MeCN, >30 h) in the absence of a dipolarophile led to complete decomposition, and we have not been able to characterize any products (such as azomethine ylide dimer derived from the “parent” 1,3-dipole2) from this experiment.Instead4awas re-isolated (afterca. 17 h) in essentially quantitative yield but this recovered material was shown to be racemic. This observation was verified by using optical rotation measurements and chiral shift1H NMR experiments using Eu(hfc)3(hfc = heptafluoropropylhydroxymethylenecamphorate). A more controlled kinetic study, again using bothαDand1H NMR to follow the course of racemisation of4a, showed that this was a first order process witht1/2 = ca. 2.5 h (in MeCN at 80 °C). Observation of this racemisation process was crucial to the development of an alternative mechanistic rationale to account for the reactivity of the β-lactam based oxazolidinones4a/4b(and by inference theC(6)substituted variant7). Precedent for this type of racemisation process does exist within the β-lactam area,15and this is more appropriately discussed in detail in the accompanying paper.11We have also probed the kinetic properties of two cycloaddition reactions using MeCN as solvent. The first involved a reactive dipolarophile (NPM) and the other using a significantly less reactive trap (methyl dithiobenzoate) (Scheme 3). In the case of NPM, the isolation method 16was used in order to establish that the reaction leading to the racemic cycloadduct8was second order overall (k = 17.8 × 10−3mol dm−1s−1at 98 °C), and first order in each of oxazolidinone4aand NPM. We were also able to determine the thermodynamic parameters for this cycloaddition reaction and these are shown inTable 1. Data from two related reactions (those shown inEquations 1 and 2) 17are provided for purposes of comparison, and the values we have obtained, especially in respect of the strongly negative entropy of activation, are similar to these more typical 1,3-dipolar cycloaddition processes. There is a caveat that must be placed on the thermodynamic data obtained from4a. The fragmentation–cycloaddition reaction does not take place to a significant extent below 80 °C, at which temperature oxazolidinone4ais relatively stable. However, at higher temperatures (>110 °C), the rate of decomposition of4acomplicated the kinetic analysis and, as a consequence, we were limited to using a narrow temperature range (85 to 98 °C) to obtain the parameters shown inTable 1.4a + NPM→8Equation 1 Equation 2 Ref.17a and 17b.Ref.17b and 17c.Calculated based on reported data.ΔH °‡76.8 kJ mol−176.6 kJ mol−165.7 kJ mol−1ΔS °‡−76.8 J mol−1K−1−121.4 J mol−1K−1−133.9 J mol−1K−1ΔG °‡107.6 kJ mol−1112.7 kJ mol−1105.6 kJ mol−1EACT79.9 kJ mol−179.1 kJ mol−1 68.2 kJ mol−1 Similar results and limitations were associated with the reaction between4aand methyl dithiobenzoate to give the penem precursor9, again using MeCN as solvent. The reaction was first order with respect to both4aand the C&z.dbd;S based dipolarophile (second order overall;k = 1.1 × 10−3mol dm−1s−1at 80 °C). In this slower reaction, partial decomposition of both reactants over the prolonged reaction time required made more detailed analysis of this process impractical. Additionally, the computational studies described below point to a possible complication associated with the kinetic analysis of the NPM reaction (seeScheme 9).Useful conclusions can also be drawn from the stereochemical outcome of the cycloaddition process. Firstly, using dimethyl fumarate and dimethyl maleate with the benzyl ester derivative4b, we have demonstrated that the cycloaddition step is stereospecific (Scheme 4). A single cycloadduct10was isolated in 74% yield from dimethyl maleate, and dimethyl fumarate led to a major product11in 42% yield. A minor isomeric product (which wasnot10) was detected by NMR in this latter reaction but could not be isolated or further characterized. The relative stereochemistry of10has been determined by NOE experiments. We were not able to unambiguously assign the structure of11(the major adduct derived from fumarate) using NOE, and coupling constants (3JH(2)–H(3),3JH(3)–H(4)) do not, in our experience, provide a reliable means for establishing relative stereochemistry in this type of β-lactam derivative. The structure shown for11is consistent with that observed in most cases examined,i.e.atransrelationship between the ring substituents atC(2)andC(3), but this assignment must be regarded as provisional.Importantly, the lack of any crossover products from these two reactions supports the participation of an azomethine ylide intermediate in an otherwise conventional and concerted [4π + 2π] cycloaddition step leading to the cycloadducts that we have observed. This statement can, of course, apply only to C&z.dbd;C-based dipolarophiles, since it is conceivable that cycloadditions involving highly polarized dipolarophiles, such as those incorporating C&z.dbd;X (where X = NR, O, S, Se), could proceedviaa stepwise pathway.More significantly for our mechanistic proposals, we have only isolated products corresponding to the thermodynamically more stable stereochemicaltransrelationship 18betweenC(2)andC(5)—see8–11. This places theC(2)carboxy moiety on theexoface of the 1-azabicyclo[3.2.0]heptane framework, an observation that not only applies to these examples but to all of the cycloadducts we have isolated to date. This particular stereochemical outcome does not conflict with the Grigg 10mechanism (cf.Scheme 2) since a similar outcome would be expected if anantidipole was involved in the cycloaddition step (see Computational studies described below). However, the stereochemical outcome in the β-lactam series may also be interpreted in another way that is consistent with the other observations that we have made.Clearly, the experimental results described above indicate that the mechanism by which azomethine ylide reactivity is released from4differs fundamentally from that associated with better knownN-alkyloxazolidinones. 1. The simple β-lactam based oxazolidinone4adoes not undergo spontaneous decarboxylation when heated—in the presence of an effective dipolarophile,4ais consumed but when heated for a similar period of time either alone or with an unreactive trap,4ais recovered. 2. Oxazolidinone4adoes, however, undergo a first order racemisation under thermal conditions. 3. Kinetic analysis of the reaction of4awithN-phenylmaleimide has shown that this cycloaddition is a second order process overall, with a negative entropy of activation and we conclude that it is likely that the cycloaddition step is rate determining. 4. The stereochemistry of the dipolarophile is retained and we conclude from this that the cycloaddition involving alkenes is a concerted process. However, the (C2) carboxy stereochemistry is thermodynamically controlled and contains no information about the structure of reaction intermediates.The mechanism already established for the fragmentation ofN-alkyloxazolidinones, which would require the direct (concerted) decarboxylation of4to give azomethine ylide2, is not consistent with these observations, and an alternative explanation for the fate of the β-lactam-based oxazolidinones is required. Our proposal, shown inScheme 5, has two key features: (i ) the nature of the azomethine ylide species13involved and (ii ) the suggestion that the (concerted and stereospecific) cycloaddition stepprecedesdecarboxylation. The first phase of this process requires ring opening of oxazolidinone4to give12, followed by proton transfer to generate13, and it is this ‘carboxylated’ azomethine ylide intermediate—thecarboxylatedvariant of our initial target 1,3-dipole2—that is postulated to participate in the crucial cycloaddition reaction. The ring strain associated with4is a likely contributor to the ease of ring opening and, based on related structures reported by Bordwell,19the malonyl proton present in12is anticipated to be of comparable acidity to that of a carboxylic acid.Bordwell 19has measured the acidity of bothiii(pKHA11.8) andii(pKHA5.6). These measurements were made in DMSO (a non-hydrogen bond donor solvent), and it is of interest to compare these values to those obtained for carboxylic acidse.g.PhCO2H: pKHA11.1 (in DMSO). The zwitterionsivderived fromiiicontain structural features—enolate of a β-dicarbonyl adjacent to an iminium moiety (albeit notN-acyl)—similar to those found in13.Reagents and conditions: i, K2CO3, BrCH2CO2Me, DMF (68%); ii, AgNO3, py, MeOH, EtOAc, then 4-NO2C6H4OCOCl,i-Pr2NEt, DMAP, THF (83% overall); LiHMDS, THF (46%).To account for the racemisation of4, formation of the ‘carboxylated’ azomethine ylide13must be reversible. In the absence of an effective dipolarophile, this equilibrium would result in the observed racemisation, but in the presence of a suitable trap, the azomethine ylide13is intercepted and undergoes a cycloaddition reaction. The final transformation would then require decarboxylation of the initially formed adduct14. If this pathway is followed, then decarboxylation must take place under the reaction conditions since we have been unable to detect adducts retaining the carboxylic acid function. This final step is also expected to give the (observed) thermodynamically more stabletransrelationship betweenC(2)andC(5)in the final product15. Based on the kinetic data presented above, we suggest that the cycloaddition step is rate determining,i.e.k3is slow. An alternative scenario would involve decarboxylation of14to give15as the rate-limiting event (k3 > k4), but we regard this as less likely. We also consider decarboxylation of12to give2as unlikely (seeScheme 8). Under specific circumstances, this pathway can operate (seeScheme 7) but only when proton tautomerisation is unavailable.Reagents and conditions: i, NaHMDS, −78 °C, MeI, THF (29% as a 2 ∶ 1 mixture of isomers); ii, MeCN, 100 °C (sealed tube) 2 h; iii,N-phenylmaleimide, MeCN, 100 °C (sealed tube), 15 h (20: 78% yield).In order to probe further the proposal outlined inScheme 5, two additional substrates have been investigated. We evaluated the thio analogue18as a precursor to an azomethine ylide, and the preparation of racemic18is outlined inScheme 6, starting from the known S-protected 4-tritylthioazetidin-2-one16.19N-Alkylation of racemic16, followed by trityl cleavage andS-acylation gave the thiocarbonate17, which underwent base-mediated cyclization to give the thiazolidinone18as a crystalline solid. Thermolysis of18(MeCN, 100 °C, 23 h or 1,2-dichlorobenzene, 150 °C, 20 h) in the presence of NPM failed to produce a cycloadduct, and18was recovered unchanged. Subsequent heating of18and NPM in 1,2-dichlorobenzene at 200 °C (in a sealed tube) resulted in decomposition—no evidence for cycloaddition was detected. Loss of COS (vs. CO2) does (ultimately) represent a thermodynamically less favourable process (see below) and the involvement of a more acidic thiol acid (theS-analogue of12) may also influence (disfavourably) the key proton transfer step [cf.12to13]. However, a significant factor accounting for the reduced reactivity of18is the comparative lack of strain associated with a sulfur (as opposed an oxygen) containing five-membered ring; the ring strain associated with this type of substrate is dealt with in the computational section of this paper.The second substrate examined was the α-methylated oxazolidinone19. This substrate, which has proven to be particularly useful, lacks the acidic malonyl proton but otherwise closely mimics oxazolidinone4. The α-methylated variant19was prepared by direct alkylation of4aand isolated, albeit in low yield, as a 2 ∶ 1 mixture of diastereoisomers; the sensitivity of19precluded separation of these isomers by chromatography and the mixture was used in subsequent experiments (Scheme 7). Thermolysis of19(MeCN, 100 °C, sealed tube) was then studied. After 2 hours, the 2 ∶ 1 mixture of isomers of19had undergone clean conversion to a 5 ∶ 95 mixture (by1H NMR), favouring what had originally been the minor diastereomer. Heating19in the presence ofN-phenylmaleimide (MeCN, 100 °C, sealed tube, 15 h) gave a single cycloadduct20in 78% isolated yield, and while the gross structure of20has been established, the detailed stereochemistry of this adduct was not pursued.Thermal epimerisation of19is presumed to involveC(5)rather thanC(2)via21. We have not determined that the mixture corresponding to19is composed of enantiomerically pure isomers, and the structure of19(as shown inScheme 7) is not intended to imply an absolute stereochemistry. Although the stereochemistry of cycloadduct20has not been pursued, it is noteworthy that the1H NMR spectrum of20is virtually superimposable (chemical shifts andJvalues) on the similar regions of the1H NMR spectrum of cycloadduct8, which has already been assigned as theendoadduct.2aThermolysis of the α-methylated oxazolidinone19(MeCN, 100 °C, sealed tube, 15 h) but in theabsenceof NPM did result in significant decomposition, and we have correlated the pathways associated with decomposition and cycloaddition (leading to20). Solutions of19(in MeCN containing a small quantity of DMSO as an internal standard) were heated with and without NPM. After 14 hours, the extent of decay of19(as judged by the amount of19remaining—ca. 20%) correlated almost exactly to the amount of cycloadduct20produced in the parallel experiment. This result indicates that the thermal decomposition of19and the subsequent cycloaddition reaction with NPM are both efficient processes.Based on these observations, the mechanism associated with conversion of19to20most reasonably involves a stepwise (reversible) fragmentationvia21and irreversible decarboxylation to yield azomethine ylide22. Isomerization of19(from a 2 ∶ 1 mixture to a 5 ∶ 95 mixture of isomers) under the conditions used provides support for this pathway.This result does raise the possibility that a similarnonconcertedpath is available to oxazolidinone4leading to azomethine ylide2, and these two processes have been compared computationally (seeScheme 8). However, given that we observe both racemisation of4(where both theC(2)andC(5)stereocentres have to be involved) and cycloadduct formation under conditions that otherwise do not lead to decomposition of4, this alternative stepwise decarboxylation can only play at best a minor role.Computational studiesIn order to gain further insight into the mechanistic possibilities outlined above, semi-empirical calculations were performed in order to estimate the relative energies of the intermediates and transition states involved in the generation of the two possible dipoles2and13as shown inScheme 5. For simplicity, the calculations were carried out on the methyl (rather than PNB) ester23, but using a simulated solvent dielectric in order to closely match the calculation results to the actual reaction conditions used. Both the oxazolidinone (X = O) and thiazolidinone (X = S) series have been compared, and the results are summarised below (Scheme 8andTables 2 and 3).Heats of formation and transition state imaginary frequencies forScheme 8StructureHf X = OHf X = Sνi X = Oνi X = SHeats of formation in kcal mol−1, obtained using AM1 (X = O) or PM3 (X = S) Hamiltonians and the COSMO 20model to simulate a solvent field equivalent to that of acetonitrile.All transition structures were characterized by observing them to have a single negative vibrational frequency corresponding to the reaction coordinate following a normal mode analysis (expressed in cm−1).Transition state could not be located in acetonitrile and energy is a single-point estimate (COSMO) on a gas-phase derived transition structure.Energies are those of 1,3-dipole plus calculated heats of formation (COSMO) of CO2(Hf(MeCN) = −88.89 kcal mol−1) and COS (Hf(MeCN) = −23.75 kcal mol−1) respectively.23−163.05−136.36——24−144.83−103.30——25−141.82 −76.68 ——26−149.93−106.50——TS1−136.87−89.47−355.93−146.04TS2(−129.49) (−80.68) (−330.23) (−472.94) TS3−128.67−85.28−568.50−572.22Activation energies forScheme 8value X = Ovalue X = SHeats of formation in kcal mol−1, obtained using AM1 (X = O) or PM3 (X = S) Hamiltonians and the COSMO 20model to simulate a solvent field equivalent to that of acetonitrile.Transition state could not be located in acetonitrile and energy is a single-point estimate (COSMO) on a gas-phase derived transition structure.Energies are those of 1,3-dipole plus calculated heats of formation (COSMO) of CO2(Hf(MeCN) = −88.89 kcal mol−1) and COS (Hf(MeCN) = −23.75 kcal mol−1) respectively.ΔE126.1846.89ΔE2(33.56) (55.68) ΔE3(−12.33) (−4.00) ΔE4−7.96−13.83ΔE516.1618.02ΔE6−12.338.60Three discrete fragmentation pathways have been considered. 1. Stepwise fragmentation followed by decarboxylation to give azomethine ylide25(viaTS1andTS3) (cf.19→22). 2. Stepwise fragmentation followed by tautomerisation (viaTS1andTS4) to give carboxylated azomethine ylide26(which represents our favoured mechanistic pathway). 3. Concerted decarboxylation (viaTS2) also to give azomethine ylide25which corresponds to the established 10mechanism for the formation of azomethine ylides fromN-alkyloxazolidinones.Clearly, the data shown inTables 2 and 3indicate that cleavage of the C–O bond in oxazolidinone23(X = O) to yield zwitterion24(X = O)viatransition stateTS1(X = O) is favoured and reversible (ΔE4 < ΔE1) compared to concerted loss of carbon dioxide from23(X = O)viatransition stateTS2(X = O) to yieldantidipole25(ΔE1 < ΔE2). Indeed, transition stateTS2(X = O) was not stable in a simulated acetonitrile environment when attempts at optimisation were made using the COSMO 20model, and all measurements are single-point estimates on a transition structure optimisedin vacuo.Decarboxylation of zwitterion24(X = O) to yield anti-dipole25(X = O)viatransition stateTS3(X = O) is, however, predicted to be energetically reasonable. However, an alternative pathway involving formal (but stepwise) proton tautomerism of zwitterion24(X = O)viatransition states approximatedviaTS4(X = O) to yield dipole26(X = O) would appear particularly favourable. Although the energetics and the structure(s) ofTS4(X = O) have not been calculated in the present study, a number of previous reports 19,21strongly indicate the extremely favourable nature of the transformation of zwitterion24(X = O) into dipole26(X = O). As mentioned above, Bordwell and co-workers 19have found that malonyl-based systems which contain an iminium moiety attached to the malonyl carbon as present in zwitterion24(X = O) show the expected enhanced degrees of acidity when compared to simpler malonate-derived carbon acids.Additionally for weak acids, including malonate derivatives, Dillon 21has demonstrated an approximately linear relationship between measured pKa’s and measured rates (and thus activation energies) of ionisation. A weak carbon acid, such as 1-chloronitroethane, has a pKaof around 7 with an activation energy for ionisation of 20 kcal mol−1. Substituting the inductively anion-stabilising chlorine atom for a conjugating substituent (as in ethyl nitroacetate) decreases the pKato approx. 6 with a corresponding decrease in activation energy required for ionisation to 16 kcal mol−1. It would be expected that C–H cleavage in zwitterion24(X = O) would be particularly favourable and this species is likely to have a pKaof 5.6 or lower. This would require an energy barrier probably appreciably lower than 16 kcal mol−1to induce C–H bond cleavage and carboxylate protonation to yield dipole26(X = O).These calculations indicate that generation ofantidipole25viaconcertedthermal loss of carbon dioxide from oxazolidinone23(X = O) does not occur in acetonitrile. As suggested by the α-methylated variant19, formation of azomethine ylide25in acetonitrile is possibleviastepwise decarboxylation with the intermediacy of zwitterion24(X = O).A parallel series of calculations relating to the transformations shown inScheme 8(X = O throughout) were performed based on the one diastereomer (of the possible two) of the α-methylated oxazolidinonev, and results are summarized here. Heats of formation and transition state imaginary frequencies for the methyl ester analogues of23−161.84;24−142.02;25−144.78;TS1−135.12 [−321.65];TS2(−129.50)a[(−337.71)a];TS3−128.52 [−535.36].Activation energies for equivalent transformations involvingvas shown inScheme 8: ΔE126.72; ΔE2(32.34)a; ΔE3(−15.28)a; ΔE4−6.9; ΔE513.5; ΔE6−16.26. (a Transition state could not be located in acetonitrile and the energy is a single-point estimate (COSMO) on a gas-phase derived transition structure.)However, tautomerism in zwitterion24(X = O) to yield dipole26(X = O) is even more favourable and, for this and the other reasons discussed above, remains the preferred reaction pathway for fragmentation of oxazolidinone23(X = O) to give an azomethine ylide.The data inTables 2 and 3also shed light on the observed lack of reactivity of the thio variant23(X = S). The activation energies for both C–S cleavage of this substrate and the concerted loss of COS are in excess of 20 kcal mol−1higher than those determined for23(X = O). Additionally, as with the oxazolidinone, the concerted fragmentation pathway could not be established in a simulated acetonitrile environment. As indicated previously, although a number of factors could contribute to the higher energies involved in the sulfur-containing substrates, the relative strain energies of oxygenversussulfur ring systems appears to be an important factor. Comparison of the calculated heats of formation (H) and differences associated with ring size (ΔH ) for a series of simple bicyclicO- andS-based heterocycles is shown inFig. 1. As can be seen, decreasing ring size in the oxygen-containing ring system results in a considerable increase in energy compared to that for the thia analogues.The viability of both 1,3-dipoles of the types25and26(Scheme 8) to undergo cycloaddition has also been investigated using a semi-empirical approach (Scheme 9andTables 4–6). Due to computational limitations, initial studies were performed in the gas phase as summarised below. These calculations were also extended to include solvent-based studies on the simplified (methyl ester containing) dipole species (seeTable 5).Heats of formation, transition state imaginary frequencies and dipole moments forScheme 9StructureStructureNo.RR′XHf (νi) [Dipole] No.RR′XHf (νi) [Dipole] Heats of formation (Hf ) and energy differences (ΔH ) in kcal mol−1, obtained using AM1 Hamiltonian.All transition structures were characterized by observing them to have a single negative vibrational frequency corresponding to the reaction coordinate following a normal mode analysis (expressed in cm−1).Expressed in Debyes.27PNB—H−2.3827Me—H−33.4728—Ph—5.8328—Me—−28.8829PNBPhH−66.7229MeMeH−134.6130PNBPhH−63.8730MeMeH−131.97TS5PNBPhH10.79TS5MeMeH−55.88(−334.5)(−319.67)[5.29][2.45]TS6PNBPhH14.45TS6MeMeH−52.64(−385.66)(−381.95)[6.63][3.31]27PNB—CO2H−88.1827Me—CO2H−119.4229PNBPhCO2H−137.0629MeMeCO2H−205.0130PNBPhCO2H−137.8330MeMeCO2H−206.85TS5PNBPhCO2H−68.70TS5MeMeCO2H−134.82(−506.43)(−499.24)[6.33][2.86]TS6PNBPhCO2H−67.38TS6MeMeCO2H−134.72(−503.28)(−463.69)[3.99][3.58]Heats of formation, transition state imaginary frequencies and dipole moments forScheme 9in MeCNStructure No.RR′XHf (νi) [Dipole] These entries refer to calculations of reactions in acetonitrile using the COSMO 20algorithm. Heats of formation (Hf ) and energy differences (ΔH ) in kcal mol−1, obtained using AM1 Hamiltonian.All transition structures were characterized by observing them to have a single negative vibrational frequency corresponding to the reaction coordinate following a normal mode analysis (expressed in cm−1).Expressed in Debyes.27Me—H−52.9328—Me—−42.8529MeMeH−163.3930MeMeH−161.69TS5MeMeH−83.03(−387.55)[4.46]TS6MeMeH−80.16(−375.07)[5.63]27H—CO2H−159.05297HMeCO2H−246.9730HMeCO2H−247.55TS5HMeCO2H−177.79(−575.07)[6.12]TS6HMeCO2H−176.83(−566.82)[4.52]Energies of cycloadditions inScheme 9EntryRR′XΔE9ΔE10ΔE11ΔE121PNBPhH7.3411.00−77.51−78.322PNBPhCO2H13.6514.97−68.36−70.453MeMeH6.479.71−78.73−79.334MeMeCO2H13.4813.58−70.19−72.135MeMeH12.7515.62−80.36−81.536HMeCO2H24.1125.07−69.18−70.72The energetics of these cycloaddition processes are shown inTable 6, and as can be seen, both the ‘decarboxylated’ 1,3-dipoles27(X = H), and the ‘carboxylated’ systems27(X = CO2H) are predicted to undergo efficient cycloadditions to maleimide derivatives, cycloadditionsin vacuogenerally possessing activation energies in the range 7–14 kcal mol−1(entries 1–4) and 12–24 kcal mol−1for the solution-phase reactions (entries 5 and 6) respectively.All cycloadditions favour formation of theendoadducts29(i.e.ΔE9 < ΔE10) with those resulting from cycloadditions of the ‘non-carboxylated’ dipoles (entries 1, 3, and 5) having around half the activation energy of the corresponding ‘carboxylated’ systems (entries 2, 4, and 6). In the ‘carboxylated’ series (X = CO2H), however, the calculated preference associated with theendopathway is marginal. It is particularly noteworthy that the value of the activation energy (ΔE9 = 24.11 kcal mol−1) for cycloaddition of the ‘carboxylated’ dipole model27(X = CO2H, R = H) in acetonitrile is comparable with the largest of the calculated activation energies in the oxazolidinone fragmentation sequence (that of ΔE1 = 26.18 kcal mol−1, corresponding to C–O cleavage) which suggests that the rates of both steps may influence the overall reaction kinetics. The heats of formation of the cycloadducts29and30are generally much lower than those for the corresponding transition statesTS5andTS6, such that the cycloadditions are essentially irreversible under these conditions (i.e., ΔE11 ≫ ΔE9, and ΔE12 ≫ ΔE10).Effect of Lewis acidsThe mechanistic proposals presented here raise a number of other interesting issues. While we cannot determine experimentally the relative positions of the equilibria linking12and13(Scheme 5) or rates of these interconversions (k1/k−1,k2/k−2), increasing the concentration of the azomethine ylide13should accelerate the rate of formation of the final cycloadduct15. One approach to this would be use of an oxaphilic Lewis acid that might serve to lower the barrier to the initial fragmentation of4(which corresponds to ΔE1inScheme 8and is the highest barrier in the formation of the carboxylated dipole) and also enhance the C–H acidity associated with12. Such interactions might then increase the effective concentration of the key azomethine ylide intermediate, and with this in mind a wide range of Lewis acids have been evaluated using the reaction between4aand NPM as a model system.A variety of Lewis acids were screened without success: Ti(Oi-Pr)4, AgX (X = OAc, NO3), MgBr2, ZnBr2, CeBr3, Sc(OTf )3. Lithium salts (LiClO4, LiBr, LiI) led to some acceleration, but LiBr was clearly superior. LiBr is similarly effective at accelerating the rate of reaction between the more substituted oxazolidinone7and NPM. We have also studied, without success, a range of acidic and basic catalysts with the aim of promoting, for example, the proton transfer step associated with the conversion of12to13. We have not attempted to promote ring opening–fragmentation of the thia analogue18or the α-methylated oxazolidinone19using a Lewis acid.The general sensitivity of oxazolidinone4aprecluded use of many of the more conventional Lewis acids, but LiBr did show a significant effect (Scheme 10). Using the same concentrations of reactants and the same reaction temperature, addition of 1.1 equivalents of LiBr resulted in a significant decrease in reaction time for a comparable level of conversion of4aand isolated yield of adduct (±)-8. A similar acceleration has also been observed when thiobenzophenone was employed, a process that leads directly to the diphenylpenam derivative (±)-31. In this case, however, use of the Lewis acid catalyst did cause some reduction in yield.The possibility that the carbonyl group of NPM or the thiocarbonyl moiety of thiobenzophenone interacts with LiBr and that this provides a mechanism for activation cannot be ignored. Attempts to catalyze the reaction of4awith PhC&z.tbd;CH using LiBr as additive failed because the resulting cycloadduct underwent rapid decomposition in the presence of LiBr, a fact that was supported by a control experiment.

ISSN:1472-7781    

DOI:10.1039/b010046n

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

IntroductionIn the preceding paper,1we described the application of the β-lactam based oxazolidinone1as a source of azomethine ylide reactivity providing an efficient and flexible entry to bicyclic β-lactams, including carbapenams, carbapenems, penams and penems.2The generation of azomethine ylides by decarboxylation of oxazolidinones is well established for simpler systems,3and the formation of cycloadducts4can be accounted for in a formal sense in terms of the “parent” azomethine ylide2, the product of aconcerteddecarboxylation of1( path a,Scheme 1). However, evidence has now been presented in support of an alternative andstepwisefragmentation of1( path b,Scheme 1) leading to the carboxylated azomethine ylide3as the key dipolar species.1,4As a result, and unlike the simpler and more established oxazolidinone fragmentations, the cycloaddition event in the β-lactam series takes placebeforeloss of carbon dioxide. Provided thatpath bis available (see preceding paper which describes a 3,3-disubstituted oxazolidinone which does not have access to this path), then available evidence also disfavours the participation of an alternative stepwise process as a means of generating azomethine ylide2: fragmentation to form an iminium ion, followed by decarboxylation ( path c,Scheme 1). The important mechanistic features of the azomethine ylide strategy elucidated to date, which are based on the evidence presented in the preceding paper 1and elsewhere,4are summarized inScheme 1.There are a number of issues associated with the reactivity of the different azomethine ylides2vs.3. The stabilization of the 1,3-dipole intermediate inferred from the presence of the additional carboxy function present in3may place limitations on the scope of the cycloaddition process. Indeed,3reacts poorly with electron-rich dipolarophiles and use of even moderately hindered substrates can present problems. Access to a less stabilized (and more reactive) 1,3-dipole, such as the “parent” β-lactam-based azomethine ylide2could therefore enhance the potential of the basic azomethine ylide strategy, and the generation of2has remained a goal of this programme.In this paper we outline a number of approaches to both the azomethine ylide2and a series of structurally related 1,3-dipoles, which were designed to explore and extend the range of β-lactam cycloadducts that can be made available using the basic cycloaddition strategy. All of these approaches have led to reaction pathways other than 1,3-dipole formation, and results obtained serve to underscore those key features associated with the success of the oxazolidinone-mediated pathway. In particular, with respect to dipole generation, the scope of the cycloaddition strategy has now been more fully defined and the observations presented below reinforce the notion that the ring strain and the 1,3-dicarbonyl (malonyl) component associated with1play a critical role in the expression of azomethine ylide reactivity, as in 1,3-dipole3.Studies directed towards the generation of the “parent” azomethine ylide2Our initial approach to generating the parent azomethine ylide2was based on a similar process to that successfully exploited for the “carboxylated” azomethine ylide3. This involved positioning a suitable leaving group atC(4)of anN-alkylated azetidinone, which carries an acidic C–H adjacent to nitrogen. Substrates6–12and14, all of which incorporate this combination of structural features, were prepared starting from the commercially available 4-acetoxyazetidin-2-one5, and the routes employed, some aspects of which are based on earlier literature,5–8are outlined inScheme 2.Reagents and conditions: i, R1SNa, EtOH, H2O (R1 = Ph 88%; R1 = Me 71%); ii, BrCH2CO2Me, NaH, DMF (R1 = Ph 87%; R1 = Me 71%); iii, heat, NPM, see text; iv, NaIO4, MeOH, H2O (78%); v, 2-pySNa, EtOH, H2O (72%), then BrCH2CO2Me, LHMDS (67%); vi, NaIO4, MeOH, H2O (79%); vii, BrCH2COO(CH2)2CH&z.dbd;CH2, NaH, DMF (65%); viii, NaIO4, MeOH, H2O (60%); ix, Ac2O, NPM, reflux (done in both the presence and absence of EtNi-Pr2); x, BrCH(CO2Me)2, NaH, THF, −50 °C (33%).These substrates were then screened for their ability to generate a viable azomethine ylide under thermal conditions, in a variety of solvents (MeCN, PhMe, 1,2-dichlorobenzene) and usingN-phenylmaleimide (NPM) as a trap. The 4-thiosubstituted azetidinones6and7were thermally stable (up to approx. 180 °C), and attempts to induce iminium ion formation (using AgBF4as a thiophile) resulted only in decomposition. Enhancing the leaving group ability of sulfurviasulfoxide8was also unsuccessful;8was stable in both MeCN (at reflux) and PhMe (at reflux). Use of ZnI2(in MeCN at 50 °C) as a Lewis acid to promote ionization again led to decomposition of8. The azetidinyl 2-pyridyl sulfide9, which incorporates an internal base and the corresponding sulfoxide10(obtained as a 1 ∶ 1.5 mixture of diastereoisomers) were prepared using similar methods. Again, these substrates proved to be remarkably stable and no evidence was obtained under a variety of thermal conditions for azomethine ylide formation.Substrates carrying an intramolecular alkenyl trap were examined, but again, both sulfide11and sulfoxide12were thermally stable. When12was heated in acetic anhydride (with a view to promoting ionizationviaO-acylation of the sulfoxide entity), only the corresponding acetate13was isolated (as judged by1H NMR). In summary, all efforts to drive the formation of an azomethine ylide by initialN-acyl iminium ion formation from a simple but only monocyclic azetidinone precursor have been unsuccessful.The selenide and selenoxide corresponding to6and8respectively were also prepared by analogous routes. The selenide was thermally stable but the selenoxide proved to be too unstable to merit further study. The sulfoxide-based substrates [8,10,12] were prepared with a view to promoting a concerted rather than stepwise fragmentation analogous to the well-known sulfoxide elimination process used in alkene synthesis. The possibility of a base assisting an alternative concerted pathway prompted incorporation of the pyridyl unit into9and10.Enhancing the acidity of the proton adjacent to nitrogen provided an alternative, but ultimately equally fruitless line of investigation. TheN-malonyl derivative14was prepared and subjected to thermolysis (xylene at reflux, with and without a catalytic quantity of trifluoroacetic acid) in the presence ofN-phenylmaleimide (NPM); no cycloadduct formation was observed. Fragmentation of this substrate would have led to a 1,3-dipole15similar to3, but14proved to be stable and was recovered unchanged even after prolonged heating (xylene at reflux for 3 hours). These same conditions would have resulted in complete consumption of the bicyclic variant1, and this again raises the issue of the viability of a monocyclicvs.a (more strained) bicyclic precursor to a β-lactam-based azomethine ylide.Generation and reactivity of betaine17In the preceding paper, we discussed the ease with which the bicyclic oxazolidinone1undergoes thermal racemization, but not decarboxylation, when heated in the absence of a reactive dipolarophile. While this was a critical observation in terms of our mechanistic studies, it is not without precedent. Some years ago, Cherryet al.9reported that treatment of enantiomerically pure deoxyclavulanic acid16with triethylamine resulted in C&z.dbd;C migration to give theracemicoxapenem18. Further, these workers were able to isolate the betaine intermediate17and showed that this species was converted to18on heating. In addition, exposure of18to triethylamine resulted in re-formation of betaine17(Scheme 3). The most likely intermediate involved in the formation of betaine17(and its subsequent conversion to18) is the stablized azomethine ylide19, though participation of aC-protonated variant of19on the pathway between16and17cannot be excluded. Azomethine ylide19bears a striking similarity to azomethine ylide3and, given the close relationship between the two processes involved (1→3→4and16→18), it was felt that this process merited further investigation as a potential source of azomethine ylide reactivity.Reagents and conditions: i, Et3N, EtOAc (62%); ii, heat, EtOAc (78%); iii, NEt3, EtOAc (87%).We have repeated the Glaxo work and isolated betaine intermediate17. The structure of17(recrystallized from water) has now been confirmed by X-ray crystallographic analysis (Fig. 1andTable 1).Single crystals of C21H29N3O6·3H2O (17·3H2O) were obtained from water, coated in vacuum grease and mounted on a glass fibre.Crystal data. C21H35N3O9,M = 473.5, triclinic,a = 8.014(4),b = 8.438(4),c = 18.463(10) Å,α = 89.99(2),β = 82.22(4),γ = 78.06(4)°,U = 1209.8(10) Å3,T = 173 K, space groupP1&cmb.macr; (no. 2),Z = 2,μ(Mo-Kα) = 0.102 mm−1, 10221 reflections measured, 4224 unique (Rint = 4.1%). Final residuals:wR2 = 10.1% (all data),R1 = 4.3% (2719 observed data). Selected bond lengths and angles are shown inTable 1. CCDC 155747. Seehttp://www.rsc.org/suppdata/p1/b0/b010050l/for crystallographic files in .cif format.Further, betaine17appears to be racemic, based on the crystallographic analysis and a lack of an optical rotation.The Glaxo group 9verified the racemic nature of18by correlation to an enantiomerically pure derivative. Oxapenem18is also opened by other nucleophiles (e.g.thiols), chemistry which is especially significant for providing an early entry to penems.Selected bond lengths (Å) and angles (°) for betaine17N(1)–C(4)1.369(3)N(1)–C(5)1.441(2)N(1)–C(2)1.474(2)C(2)–N(18)1.527(3)C(2)–C(3)1.544(3)C(3)–C(4)1.518(3)C(4)–O(4)1.213(2)C(5)–C(6)1.393(3)C(5)–C(9)1.441(3)C(4)–N(1)–C(5)129.7(2)C(4)–N(1)–C(2)93.8(2)C(5)–N(1)–C(2)132.0(2)N(1)–C(2)–N(18)114.5(2)N(1)–C(2)–C(3)87.89(15)N(18)–C(2)–C(3)118.4(2)C(4)–C(3)–C(2)85.40(15)N(1)–C(4)–C(3)92.9(2)C(6)–C(5)–C(9)124.8(2)C(6)–C(5)–N(1)117.1(2)C(9)–C(5)–N(1)118.1(2)O(6)–C(6)–C(5)121.6(2)Solid state structure of17. Hydrogen atoms and water molecules have been omitted for clarity.Our aim was to attempt to intercept19using a reactive dipolarophile (NPM), but under a range of thermolysis conditions, betaine17gaveonlythe conjugated oxapenem18. It is interesting to compare the reactivity of1to that of17. The presence of a good (carboxylate) leaving group in1may provide a synthetically useful concentration of the reactive azomethine ylidei.e.3. However, in the case of17, the internal nucleophile (the oxygen of the ketone enolate) is an efficient trap for the iminium component of19, and an intermolecular cycloaddition pathway cannot effectively compete.Alternative oxazolidinone precursorsA series of other β-lactam based oxazolidinones22,23and33have also been prepared and evaluated as alternative sources of azomethine ylide reactivity. In the first instance the motivation to attempt to extend the scope of dipole precursors arose from problems that we had encountered in the application of the azomethine ylide strategy (based on1) for the synthesis of so-called Δ2-carbapenems. Thermolysis of oxazolidinone1in the presence of an alkyne does provide a versatile synthesis of the carbapenem isomers20(Scheme 4). However, the isomerization of20(a Δ1-carbapenem—classical penam numbering) to the biologically more important isomer21(a Δ2-carbapenem) is neither an efficient nor general process,10and a practical solution to this problem demanded an alternative approach.Reagents and conditions: i, R1C&z.tbd;CH; ii, DBU, CH2Cl2(under these thermodynamic conditions,ca. 1 ∶ 1 ratio of20to21is observed when R1 = SPh).Use of the sulfide or sulfoxide-based oxazolidinones22and23respectively in principle offered access to a number of interesting 1,3-dipoles. Decarboxylation of22or23would provide novel, functionalized β-lactam-based azomethine ylides (e.g.1,3-dipole35,seeScheme 6below). Alternatively, a stepwise fragmentation and proton shift would offer access to azomethine ylide24, which was anticipated to react with an alkene to give cycloadduct25. This cycloadduct could then undergo thiol or sulfoxide elimination, rather than decarboxylation, to provide the desired Δ2-carbapenem directly. In this way we aimed to retain the carboxy function present in the starting oxazolidinones22or23as the carboxylic acid of the final target (Scheme 4).The sulfide- and sulfoxide-containing oxazolidinones22and23respectively were both prepared starting from clavulanic acid26(Scheme 5). Using the procedures described by Hunt,11clavulanic acid was converted to diene27. Addition of PhSH to27gave a 2 ∶ 1 mixture of regioisomeric adducts28(40% yield as an inseparable 1 ∶ 5 mixture ofEandZisomers) and29(19% yield). Chromatographic separation followed by controlled ozonolysis of28gave the sulfide-containing oxazolidinone22as an oil. Oxidation (MCPBA) of28followed again by controlled ozonolysis provided the sulfoxide derivative23as a mixture of diastereoisomers.Though reported as such in the literature,11the stereochemistry at C(2) of22has not been rigorously determined. Placing the SPh residue on theexoface does, however, represent the thermodynamically more reasonable configuration. Nevertheless, and to avoid any confusion, it should be made clear that we have assumed the configurations shown inScheme 5and in the computational study associated withScheme 6.Reagents and conditions: i,N,N-dimethylformamide dimethyl acetal; ii, PhSH, AIBN, heat; iii, O3, EtOAc (99% for22; 97% for23); iv, MCPBA, CH2Cl2(78%); v, NPM, MeCN, reflux, 120 h; vi, DMAD, MeCN, reflux, 48 h.Thermolysis of22in the presence of NPM resulted in the formation of the thiol adduct31in 43% yield. A related fragmentation process was observed when23was heated in the presence of DMAD, and in this case the sulfenic acid adduct32was isolated in 35% yield. In each case, other sulfur-containing by-products (e.g.PhSSPh, PhSSO2Ph) were also isolated and characterized. In neither case could products retaining the β-lactam component be characterized and the fate of this unit remains unclear. Products31and32can be accounted for by loss of PhS−or PhSO−from the oxazolidinone precursors22and23respectively and subsequent capture of these nucleophiles by the electrophilic alkene/alkyne present. While no direct evidence has been obtained for the intermediacy of theN-acyl iminium component30(see also below, structure36,Scheme 6), it would appear that this pathway is preferred to formation of the alternative (and desired)N-acyl iminium species24(cf.Scheme 4) resulting from C–O bond scission. Further, a possible reason for the preference of22(and by extrapolation23) to undergo C–S cleavage to give30rather than C–O cleavage (which would lead to24) has been identified, which is discussed in more detail below.The isomeric thiol adduct29provided access to yet another oxazolidinone substrate. Careful oxidative cleavage of29gave the unsubstituted β-lactam-based oxazolidinone33(Scheme 5), a compound that has previously been prepared by a very similar procedure.12However, oxazolidinone33was very sensitive and could not be rigorously purified from the other ozone-derived product. We have examined the ability of33to generate an azomethine ylide, but without success: thermolysis of crude33(MeCN, 81 °C) in the presence of an excess NPM failed to generate an adduct and no characterizable products were isolated. The reactivity of33has also been examined computationally and compared to carboxylated variants (seeScheme 7).Computational studiesThe fragmentation of sulfide22and the fate of “parent” oxazolidinone33have also both been explored computationally. By analogy to the theoretical approach described in the preceding paper,1semi-empirical calculations were used to compare the hypothetical stepwise and concerted decarboxylation of sulfide22to yield an azomethine ylide35vs.a fragmentation pathway involving loss of PhS−(Scheme 6andTables 2 and 3).Heats of formation and transition state imaginary frequencies forScheme 6StructureHf /kcal mol−1 νi/cm−1 Heats of formation obtained using PM3 Hamiltonian using the COSMO model to simulate a solvent field equivalent to that of acetonitrile.All transition structures were characterized by having a single negative vibrational frequency.Transition state unable to be located in acetonitrile and energy is a single-point estimate (COSMO) 13on a gas-phase derived transition structure.22−71.82—34−43.26—35−49.18—36−57.34—TS1−33.09−402.58TS2−27.95−455.41TS3(−23.55) (−490.98) TS4−45.69−52.02Activation energies forScheme 6EnergyValue/kcal mol−1EnergyValue/kcal mol−1Transition state unable to be located in acetonitrile and energy is a single-point estimate (COSMO) 13on a gas-phase derived transition structure.ΔE138.73ΔE5(48.27) ΔE2−10.17ΔE6−25.63ΔE315.31ΔE726.13ΔE4−21.23ΔE811.65These data confirm that fragmentation of oxazolidinone22vialoss of phenylthiolate to yield cation36(activation energy ΔE7 = 26.13 kcal mol−1) is considerably more favourable than decarboxylation to yield azomethine ylide35by either the stepwise pathway (involving transition statesTS1andTS2with activation energies ΔE1 = 38.73 and ΔE3 = 15.31 kcal mol−1respectively) or a concerted pathway (involving transition stateTS3with an activation energy ΔE5 = 48.27 kcal mol−1) pathways. As with other calculations involving concerted decarboxylation to yield β-lactam based azomethine ylides (see preceding paper), the transition structureTS3could not be obtained using calculations which simulate the presence of solvent, and all energies are derived from single-point estimates on a transition structure obtained in the gas-phase. A particularly interesting aspect of these studies relates to the mechanism of elimination of thiolate from oxazolidinone22. Repeated attempts to identify the transition state for the cleavage of the C–S bond consistently produced structureTS4in which the endocyclic N–C bond present within the four-membered ring is partially cleaved. This bond is essentially antiperiplanar to the breaking C–S bond (assuming the C(2) stereochemistry shown) such that this transition structure results from an E2 type elimination pathway to yield an acyl cation36(Scheme 6). The relative ease of this reaction compared to the hypothetical decarboxylative entry to azomethine ylide35may be due, in part, to the relief of ring strain associated with the formation of cation36viaTS4. Clearly also, cation36is, in a broad sense, structurally equivalent to cation30(seeScheme 5).Similar calculations involving the hypothetical decarboxylation of parent oxazolidinone33also tend to support the limited experimental observations concerning the behaviour of this substrate (Scheme 7andTables 4 and 5). The stepwise decarboxylation of oxazolidinone33via(i)cleavage of the ring C–O bond (throughTS5) to yield zwitterion37followed by(ii)decarboxylation (throughTS6) to yield parent β-lactam-based azomethine ylide38is predicted to require 27.49 kcal mol−1(ΔE9) and 33.99 kcal mol−1(ΔE11) respectively. Although the energy associated with the initial fragmentation (ΔE9) is similar to that calculated for the corresponding transformation involving the ester-containing oxazolidinone1(R = Me), which requires 26.18 kcal mol−1, the energy barrier for the decarboxylation step (viaTS6) is over twice that calculated (33.99vs. 16.16 kcal mol−1) for the equivalent pathway involving1(R = Me).1Similarly, the concerted decarboxylation of33(viaTS7) is calculated to require nearly 10 kcal mol−1more energy than is calculated for1(R = Me).Heats of formation and transition state imaginary frequencies forScheme 7StructureHf /kcal mol−1 νi/cm−1 Heats of formation obtained using PM3 Hamiltonian using the COSMO model to simulate a solvent field equivalent to that of acetonitrile.All transition structures were characterized by having a single negative vibrational frequency.Transition state unable to be located in acetonitrile and energy is a single-point estimate (COSMO) 13on a gas-phase derived transition structure.33−83.63—37−66.10—38−41.02—TS5−56.14−334.97TS6−32.11−669.64TS7(−41.24) (−511.26) Activation energies forScheme 7EnergyValue/kcal mol−1EnergyValue/kcal mol−1Transition state unable to be located in acetonitrile and energy is a single-point estimate (COSMO) 13on a gas-phase derived transition structure.ΔE927.49ΔE12−8.91ΔE10−9.96ΔE13(42.39) ΔE1133.99ΔE14(−0.22) It would appear that for the simple, unsubstituted system33, the stepwise decarboxylative generation of the corresponding azomethine ylide38is unfavourable due to the large barrier (ΔE11) to loss of carbon dioxide. This is presumed to reflect the reduced ability of zwitterion37to stabilize the increasing electron density at the exocyclic methylene as the reaction enters into the transition state (TS6), when compared to the corresponding pathway starting with1(R = Me).The need for caution regarding the calculated energies for the concerted decarboxylation pathway is underlined in the case ofTS7. The single point estimate in a simulated solvent dielectric on the gas phase-derived structure yields an energy which is 0.22 kcal mol−1lower than the solution phase energy of the products (dipole38and CO2). However, comparing the gas phase energies for thiazolidinone33(Egas = −63.90 kcal mol−1),TS7(Egas = −8.65 kcal mol−1) and38 + CO2(Egas = −28.79 kcal mol−1) yields more reasonable relative energies.To summarize, a combination of factors seem to predispose the β-lactam based oxazolidinone1towards stepwise fragmentation to give the synthetically important azomethine ylide3. Ring strain in the bicyclic framework plays an important role, and this is evident from the monocyclic azetidinones shown inScheme 2that do not give azomethine ylides despite the presence of similar structural components to those present in1i.e.a leaving group at C(4) and an acidic C–H bond adjacent to the ring nitrogen.The Glaxo betaine17is interesting in this context because this species appears to undergo reversible ring closureviaan azomethine ylide19. However, we have not been able to capture this dipolar intermediate in a cycloaddition reaction, and it appears that the enolate oxygen associated with19is a highly efficient trap. In relation to the fragmentation of1to give3, then the differences in reactivity seen with17(and by implication19) may also involve differences in charge distribution—a ketone enolatevs. a carboxy-based variant.The sulfur-containing bicyclic oxazolidinones22and23again bear a close resemblance to1, and were designed with a complementary synthetic role in mind. However in these cases, the preferred iminium species appears to be associated with C–S rather than C–O cleavage, although the fate of the β-lactam moiety remains unclear. In the case of22, computational studies suggest that the ionization process involves participation and cleavage of the β-lactam ring, and this likely leads to even more relief of strain.Fragmentation—either concerted or stepwise—of the simple, unsubstituted derivative33failed to yield a cycloadduct, apparently because this system lacks the electronic stablization that is associated with the carboxy function found in3. Calculations indicate that both of these pathways are energetically highly demanding, and we have been unable to isolate any 1,3-dipolar cycloadducts arising from33.In conclusion, while the azomethine ylide strategy is of proven and general utility for the synthesis of a wide range of bicyclic β-lactams, oxazolidinones such as1remain the only viable source of the β-lactam based azomethine ylide to underpin this synthetic strategy. Less stabilized and more reactive variants, such as2, remain elusive, but we now know much more about those alternative reaction pathways that limit the range of oxazolidinones that are capable of generating synthetically useful β-lactam based azomethine ylides.

ISSN:1472-7781    

DOI:10.1039/b010050l

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

IntroductionThe catalysis and control of intermolecular radical addition to alkenes, carbonyl compounds and imines are a challenge and only recently have the first examples of diastereo- and enantioselective control in these transformations appeared.1Intermolecular radical addition to C&z.dbd;N bonds catalyzed by Lewis-acid complexes is a rarely studied reaction, although it is a powerful method for the formation of C–C bonds. Naitoet al. have shown, in a series of papers, that BF3in particular can be used to promote radical addition to oxime ethers and they used this approache.g.in diastereoselective reactions, employing mainly Oppolzer's camphorsultam as the chiral auxiliary, and in solid-phase reactions.2Addition of alkyl radicals to chiral glyoxylate imines has been studied in the presence of ZnEt2and depending on the reaction conditions addition to both the carbon atom and the nitrogen atom of the C&z.dbd;N bond was observed.3However, it should be noted that the addition to the nitrogen atom is not a radical process. More recently it has been shown that Lewis acids can catalyze the diastereoselective radicaladdition to chiral hydrazones givingN-acylhydrazines in moderate yield and with high stereocontrol.4This paper presents an investigation of the metal-catalyzed, and the metal-catalyzed, enantioselective intermolecular radical addition to C&z.dbd;N bonds of various alkyl halides. The influence of different Lewis acids on the reaction course is investigated and it is shown that it is possible to perform catalytic enantioselective radical addition to C&z.dbd;N bonds using chiral Lewis acids as the catalyst.

ISSN:1472-7781    

DOI:10.1039/b101762o

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

ExperimentalGeneralMps were determined on a Kofler hot-stage apparatus and are uncorrected, IR spectra on either a Perkin-Elmer 681 or a Nicolet 205 spectrophotometer; mass spectra (EI) and (FAB using 3-nitrobenzyl alcohol as matrix) on an AEI MS 902 spectrometer equipped with a MSS Data System for Windows (Data Version 2.03, Software Version 10.0). Optical rotations were recorded on a Perkin-Elmer 141 polarimeter and are reported in units of 10−1deg cm2g−1. Extracts were dried over Na2SO4. All chromatography columns, thick layer (PLC) and thin layer (TLC) glass plates were made up using Merck Kieselgel GF254. Column sizes and solvent systems used are specified in each case. Flash chromatography was carried out using Fluka Kiesel gel 60 (239–400 mesh).1H and13C NMR spectra, using tetramethylsilane as internal standard, were determined at 270 and 67.8 MHz respectively with a JEOL GX270 spectrometerfitted with a dual 5 mm C/H probe.1H NMR spectra were acquired with 32 K data points over a spectrum width of 3001.2 or 6002.4 Hz;Jvalues are given in Hz. Carbon atom types were established in the13C NMR spectrum by employing a combination of broad-band proton-decoupled and distortionless enhancement by polarisation transfer (DEPT) experiments with 32 K data points over a spectrum width of 17605.6 Hz. Assignments were established by employing a combination of 1-D and 2-D NMR experiments. 2-Dimensional spectra were acquired and processed by standard JEOL software;1H–1H correlations by double quantum-filtered COSY (VDQFN), resolution 2.93 Hz in thef1 andf2 domains, PW1 = PW2 = π/2; [1JC–H]13C–1H correlations (VCHSHF), resolutionf2 17.19 andf1 5.9 Hz, pulse delay 1, 2 or3 s,JC–H140 Hz; and [2JC–Hand3JC–H]13C–1H correlations were established using the FLOCK pulse sequence of Reynoldset al.,8resolutionf2 17.19 andf1 5.9 Hz, pulse delay 1, 2 or 3s,Δ186.5 andΔ246.5 ms orΔ144.0 andΔ224.0 ms.Isolation of the metabolitesX.  grammica(isolates no. KC 109, 12 EBV, P 114) was grown for eight weeks in subdued daylight at 23 °C in Thompson bottles (4 × 2 dm3) each containing malt media (1 dm3, 3%) containing additional glucose (6%). The mycelium was removed by filtration through muslin and the medium extracted with ethyl acetate (3 × 1 dm3). The dried extract was evaporated to yield a semi-solid gum (6.1 g), which was triturated with a little ethyl acetate and the mixture filtered to yield a pale brown solid (0.68 g), which after crystallisation either from a mixture of acetone–hexane, or ethyl acetate, gave4-oxo-4H-pyran-3-acetic acid(γ-pyrone-3-acetic acid)2as needles (0.56 g), mp 172–173 °C (sublimes at 150 °C),m/z154 (M+, 5%), 136 (41), 110 (100) (FoundC, 54.7; H, 3.8. Calculated for C7H6O4C, 54.55; H, 3.9%);νmax(KBr)/cm−13600–3300, 1725 and 1638;δH(C5D5N) 3.74 (2 H, d,J0.73 Hz, 7-H), 6.47 (1 H, d,J5.86 Hz, 5-H), 7.91 (1 H, dd,J5.86, 0.73 Hz, 6-H), 8.16 (1 H, d,J0.73 Hz, 2-H) and 10.75 (1 H, br s, OH);δC(C5D5N) 31.73 (7-CH2), 116.82 (5-CH), 125.63 (3-C), 154.54 (2-CH), 156.27 (6-CH), 172.94 (8-C) and 177.51 (4-C).The residue (5.33 g) remaining after evaporation of the triturating solvent was dissolved in the solvent mixture toluene–ethyl acetate–acetic acid (60 ∶ 40 ∶ 1) and applied to a column of silica gel (3.5 × 46 cm). Fractions of 2 cm3were collected.Tubes 113–214Evaporation of the solvent yielded4-hydroxy-4H-furo[2,3-b]pyran-2(7aH)-one(grammicin)1as a yellow oil (2.49 g); C7H6O4;m/z154 (M+, 4%), (CI[M + NH4]+172.0610, C7H10O4N requires 172.0610); [α]23D+23.89° (c0.98 in EtOH);νmax(CHCl3)/cm−13400–3500 and 1790;λmax(EtOH)/nm 213 (ϵ/dm3mol−1cm−19024) and 263 (ϵ1540);δHandδCdata are presented inTables 1and2.Tubes 381–499These gave a solid, which crystallised from a mixture of acetone and hexane, or sublimed under vacuum at 90 °C, to yieldmethyl 4-oxo-4H-pyran-3-acetate10as diamond-shaped crystals (0.13 g); C8H8O4; mp 92 °C;m/z168 (M+) (Found: C, 57.0; H, 4.8. C8H8O4requires C, 57.1; H, 4.8%);νmax(KBr)/cm−11738 and 1660;δH(CDCl3) 3.42 (2 H, d,J1.10 Hz, 7-H), 3.72 (3 H, s, CH3O), 6.39 (1 H, d,J5.86 Hz, 5-H), 7.76 (1 H, dd,J5.86, 1.10 Hz, 6-H) and 7.85 (1 H, d,J1.10 Hz, 2-H);δC(CDCl3) 30.48 (7-CH2), 52.29 (CH3O), 116.75 (5-CH), 124.06 (3-C), 153.97 (2-CH), 155.54 (6-CH), 170.68 (8-C)and 177.40 (4-C).The endophytic isolates no. 306, 307, and 536 were similarly cultured; all three gave the same products. Culture 307 gave the highest yield of grammicin1(2.49 g) and γ-pyrone-3-methyl acetate (methyl 4-oxo-4H-pyran-3-acetate)10(0.13 g). γ-Pyrone-3-acetic acid2was not produced by any of the three cultures.Acetylation of grammicin1Grammicin1(2.0 g) was mixed with acetic anhydride (5 cm3) and pyridine (3 drops). The mixture, which darkened and became warm, was set aside overnight and then poured on to ice. After 3 h at 5 °C the brown crystalline solid (1.6 g) was filtered and the filtrate neutralised with aqueous sodium hydrogen carbonate and extracted with ether (3×). Evaporation of the dried ether extract gave a gum, which crystallised immediately to yield additional solid (0.3 g). Recrystallisation of the combined solids from light petroleum (bp 80–100 °C) gave4-acetoxy-4H-furo[2,3-b]pyran-2(7aH)-one5as colourless needles (1.46 g); mp 81–83 °C;m/z196 (M+) (Found: C, 55.4; H, 4.0. C9H8O5requires C, 55.1; H, 4.1%); [α]23D+234° (c1.06 in EtOH);νmax(KBr)/cm−11786, 1736 and 1637;δH(CDCl3) 2.19 (3 H, s, 11-CH3), 4.96 (1 H, dd,J6.23, 2.75 Hz, 5-H), 6.08 (1 H, dd,J2.20, 0.73 Hz, 3-H), 6.15 (1 H, s, 8-H), 6.24 (1 H, ddd,J2.20, 2.20, 2.75 Hz, 4-H) and 6.45 (1 H, ddd,J6.23, 2.20, 0.73 Hz, 6-H);δC(CDCl3) 20.75 (11-CH3), 64.36 (4-CH), 97.38 (8-CH), 100.16 (5-CH), 116.02 (3-CH), 143.69 (6-CH), 159.51 (9-C), 168.76 (2-C)* and 170.03 (10-C)* (* indicates assignments that may be interchanged).3-(2-Hydroxy-5-oxo-2,5-dihydrofuran-3-yl)propenal7(a) A solution of grammicin1(500 mg) in aqueous hydrochloric acid (5 cm3, 10%) was heated for 30 min on a boiling water bath. The orange–brown solution was evaporated to dryness under vacuum and the red crystalline residue (500 mg) was purified by flash chromatography in the solvent system toluene–ethyl acetate (50 ∶ 50). Evaporation of the eluent gave a pale yellow, crystalline solid (200 mg), which after recrystallisation either from a mixture of acetone and light petroleum (bp 60–80 °C), or from water, yielded3-(2-hydroxy-5-oxo-2,5-dihydrofuran-3-yl)propenal7as needles (170 mg); mp 151–153 °C;m/z154 (M+) (Found: C, 54.7; H, 4.0. C7H6O4requires C, 54.55; H, 3.9%);νmax(KBr)/cm−13223,1768 and 1668;λmax(EtOH)/nm 214 (ϵ/dm3mol−1cm−121267), 263 (ϵ21668) and 361 (ϵ3234);δH[(CD3)2CO] 6.51 (1 H, s, 5-H), 6.61 (1 H, d,J0.73 Hz, 3-H), 6.72 (1 H, ddd,J16.12, 7.51, 0.73 Hz, 7-H), 7.14 (1 H, d,J8.43 Hz, OH), 7.60 (1 H, dd,J16.12, 0.73 Hz, 6-H) and 9.77 (1 H, d,J7.51 Hz, 8-H),δC[(CD3)2CO] 98.42 (5-CH), 125.04 (3-CH), 136.76 (7-CH), 139.89 (6-CH), 159.60 (4-C), 170.06 (2-C) and 194.21 (8-CHO).(b) A solution of grammicin acetate5(200 mg) in aqueous acetic acid (6 cm3, 5%) was refluxed for 1.5 h. The orange-coloured solution was evaporated to dryness under vacuum and the residue purified by PLC in the solvent system toluene–ethyl acetate–acetic acid (50 ∶ 49 ∶ 1). The products were detected by UV. Band 1 (Rf0.9 blue) gave unchanged acetate5(52 mg) and band 2 (Rf0.4 violet–blue) gave the propenal7(60 mg) as described above after recrystallisation from a light petroleum–acetone mixture.4-Oxo-4H-pyran-3-acetic acid2from grammicin acetate5A solution of grammicin acetate5(200 mg) in aqueous acetic acid (1 cm3, 2%) and tetrahydrofuran (5 cm3) was refluxed for 7 days. The deep orange solution was evaporated to yield a brown semi-solid. The crude product, which gave a blue and two violet spots (TLC toluene–ethyl acetate–acetic acid, 50 ∶ 49 ∶ 1) when visualised at 366 nm, was triturated with ethyl acetate to give a brown solid (79 mg). Recrystallistion of the latter from ethyl acetate gave 4-oxo-4H-pyran-3-acetic acid2as rosettes of small needles; mp 172–173 °C. Evaporation of the trituration solvent yielded a gum which was purified by PLC to give unchanged grammicin acetate5(top blue spot, 70 mg), 3-(2-hydroxy-5-oxo-2,5-dihydrofuran-3-yl)propenal7(middle violet spot, 20 mg) and additional 4-oxo-4H-pyran-3-acetic acid2(bottom violet spot close to baseline, 7 mg).

ISSN:1472-7781    

DOI:10.1039/b101708j

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionHexa-3,5-dienoic acid methyl esters are important synthons in organic synthesis. The unsubstituted diene is one of the key intermediates used for the synthesis of the “southern” half of vitamin B12.1It has also been used as an intermediate in the synthesis of the alkaloid lycorine.2Although there is a vast amount of literature on the synthesis of its conjugated analogs, our literature survey showed that there are only few reports that describe the synthesis of this class of compounds. Some of the reported methods include the following: (i) the Rh-catalysed C–C coupling of allene with but-3-enoic acid to give a mixture of esters of hexa-3,5-dienoic acid in a 9 ∶ 1 ratio;3(ii) the transition-metal-catalysed reaction of styrene with but-3-enoic acid;4and (iii) the reaction of allylic alcohols with acetylene, CO andMeOH (as described in an Italian patent) in the presence of an NiBr2–phosphine complex to yield two products, one of which is the title compound.5Other methods include the use of Li–diisopropylamide–HMPA in THF to deconjugate methyl sorbate to the corresponding diene.6Previously we have illustrated the successful use of lead(IV) acetate in combination with Lewis acids to effect a 1,2-carbonyl transposition in acetophenones 7and acyclic α,β-unsaturated ketones 8and a ring contraction in cyclic α,β-unsaturated ketones and related systems.9a–cAs part of our continuing work to test the effectiveness of this reagent combination, we inserted one more double bond between the carbonyl and the benzylidene system to see whether the presence of the doubly conjugated moiety in the ketone (1) would also facilitate a 1,2-carbonyl shift.

ISSN:1472-7781    

DOI:10.1039/b101242h

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

Results and discussionReagent variationWith the aim of reducing reaction times and, where necessary, increasing yields, we examined the conversion of 2,5-diphenylfuran1aand 2,3,5-triphenylfuran3ainto the isothiazoles2aand4a, respectively, under more vigorous conditions (Table 1). Replacing the solvent benzene by toluene (Method B) decreased the reaction time from 24 to 6 h for1athough with a drop in yield, and from 18 to 1 h for3a, with a large increase in yield. This trend continued with chlorobenzene as solvent and with pyridine optionally replaced by isoquinoline as the base, the amount of thionyl chloride being doubled to take account of its volatility (Method C); both1aand3agave high yields of the respective isothiazoles (80 and 95%) in a mere 30 min of heating under reflux. These hotter reactions in toluene and in chlorobenzene, with premixing of the carbamate, thionyl chloride and base at room temperature before the addition of the furan, became our standard procedures for most of the later reactions. Quinoline and 2,6-lutidine were less efficient than isoquinoline in chlorobenzene reactions, though the more expensive DMAP was equally good. There was no reaction in the absence of added base though 4 Å molecular sieves, which absorb hydrogen chloride, did give very slow reaction in boiling benzene. Replacement of thionyl chloride by thionyl bromide gave much lower yields. No advantage was gained by using methyl, allyl or phenyl carbamate in place of urethane.Treatment of furans1aand3awith ethyl carbamate, thionyl chloride and a base in a boiling solventFuranUrethane/mmolSOCl2/mmolBaseSolventReaction time/hYield of2a/4a(%)1a4.34.3PyridineBenzene242a(98)3a4.34.3PyridineBenzene184a(56)3a4.34.3PyridineNone14a(77)3a4.34.34 Å mol. sievesBenzene74a(5)1a4.34.3PyridineToluene62a(70)3a4.34.3PyridineToluene14a(85)3a2.154.3PyridineToluene3.54a(71)1a4.38.6PyridinePhCl0.52a(77)3a4.38.6PyridinePhCl0.54a(88)3a4.38.6NonePhClNo reaction3a4.38.6QuinolinePhCl484a(26)3a4.38.62,6-LutidinePhCl124a(56)3a4.38.6DMAPPhCl0.54a(82)1a4.38.6IsoquinolinePhCl0.52a(80)1a4.38.6IsoquinolinePhCl0.52a(95)These higher temperature conditions were then tested on 3-bromo-2,5-diphenylfuran3b, 3,4-dibromo-2,5-diphenylfuran5and 3-bromo-2,4,5-triphenylfuran6readily made by bromination of1aand3a. With pyridine in benzene, the trisubstituted furan3bgave the isothiazole4bin virtually quantitative yield, though in a relatively long time (16 h); with isoquinoline in chlorobenzene it gave4bin 80% yield in 10 min.The tetrasubstituted furans5and6were not expected to react with the Katz reagent since they had been found not to react with (NSCl)3.1bThis seemed reasonable since the reaction stoichiometry involved the loss of hydrogen chloride, with the hydrogen being displaced from the furan ring.1bThese compounds were reported not to react in boiling benzene,2though a reinvestigation has shown that5does react very slowly, giving, after 48 h, 13% of the same isothiazole4bas from the monobromo furan3b. When this reaction was repeated with isoquinoline in chlorobenzene,4bwas isolated in 68% yield after1 h. Furan6also reacted with the Katz reagent to give low yields of the same isothiazole4aas from triphenylfuran3a; with pyridine in toluene consumption of6was slow giving 15% of4ain 24 h, with starting material remaining, and with isoquinoline in chlorobenzene it was fast but complex, giving 12% of4ain 45 min.In this unexpected conversion of bromofurans5and6into isothiazoles with one fewer bromine atom, bromine is lost presumably as Br+by electrophilic displacemente.g.by acid or NSCl. However, a brief attempt to intercept Br+, or BrCl, with anisole added to the reaction mixture was unsuccessful.The lack of reaction of5and6with (NSCl)3in boiling tetrachloromethane1bmay result from the lower temperature and/or the absence of acid, which is generated in the ethyl carbamate–thionyl chloride reaction.2,5-Di-tert-butylfuran1cgave the isothiazole2cin moderately good yield (66%) with pyridine in benzene, but much more slowly (6 days) than the diaryl compounds. With isoquinoline in chlorobenzene the reaction time was reduced to 12 h but with some decomposition and a lower yield (40%).All of these results indicate the much enhanced reactivity of the isoquinoline–chlorobenzene modification of the Katz reagent.Deactivated furansWe next examined a series of 5-phenylfurans7bearing an electron-withdrawing group, X, at the 2-position (X = CO2H, CO2Et, CH&z.dbd;NOH, CH&z.dbd;NOMe, CN, COMe, COPh and SO2Ph) to see if these deactivated rings would undergo the same reaction.With the exception of the oxime none of these compounds reacted significantly with the benzene–pyridine reagent, but they did react under the more vigorous conditions. 5- Phenylfuran-2-carboxylic acid7adecomposed but its ethyl ester7bgave the ethyl 5-benzoylisothiazole-3-carboxylate8bin 25% yield with pyridine in toluene after 3 days and in 52% yield in chlorobenzene after 48 h.Similar furan to isothiazole conversions were obtained in chlorobenzene with X = CN (24%, 2 days), X = CH&z.dbd;NOMe (27%, 2 days), X = COPh (51%, 2.5 h) and X = SO2Ph (60%, 6 h and 55% with pyridine in toluene for 48 h) to give8c, d, fandgrespectively. With X = Ac however, rapid decomposition ensued in toluene and chlorobenzene, possibly involving the enol tautomer. Treatment of the isomeric oximes7, X = CH&z.dbd;NOH, with the thionyl chloride reagent in boiling benzene, or better, toluene not surprisingly gave the cyanide7cvery efficiently (92%), but the isomeric oxime ethers7dresulted in the isothiazole8d, presumably as the more stableanti-isomer. The structures of the isothiazoles8followed from their spectroscopic properties, compared with other 5-benzoylisothiazoles made in the same way, and were confirmed by X-ray crystallography for the carboxylic acid8a, prepared by hydrolysis of its ethyl ester8b, and for the sulfone8g. Thus the regiochemistry of these reactions is as expected for the initial substitution of the polarised furan ring at the more nucleophilic β-position,via10which is thought to rearrange by ring-opening and ring-closing (Scheme 1) to give the 5-benzoylisothiazole exclusively.1aThe structure of the carboxylic acid8a(Fig. 1) shows the carboxylic acid group to lie almost coplanar with the isothiazole ring, the torsional twist about the C(3)–C(13) bond being onlyca.6°. The phenyl ring is fairly steeply inclined, byca.49°, to the isothiazole ring plane, the inclination being produced by torsional twists of 22 and 33° about the C(5)–C(6) and C(6)–C(7) bonds respectively. The O(1) carbonyl group is orientedsynto the isothiazole ring sulfur atom. The isothiazole ring is almost perfectly planar (maximum deviation of <0.001 Å), but C(6) lies 0.106 Å out of this plane whereas C(13) deviates by only 0.043 Å in the opposite sense. The bonding within the isothiazole ring is fairly typical of that reported previously for this ring system, there being a distinct pattern of bond ordering with the C(3)–N(2) andC(4)–C(5) bonds each displaying significant double bond character (Table 2). Adjacent centrosymmetrically related pairs of molecules pack with their isothiazole rings overlying each other, this stacking interaction (ainFig. 2) being supplemented by a pair of carbonyl⋯carboxylic Oδ−⋯Cδ+interactions (binFig. 2). These “dimer pairs” are then linkedviacarboxylic⋯carboxylic hydrogen bonds (cinFig. 2) to form continuous chains. Neighbouring chains are weakly cross-linked by π–π interactions between the phenyl rings in one chain and the opposite face of the carboxylic group involved in the δ+⋯δ−interaction (b) in the next andvice versa; the Ph(ring centroid)⋯C(13)distance is 3.48 Å. There are no short intermolecular S⋯N contacts.The molecular structure of8a.Part of one of the continuous chains of molecules present in the structure of8a. The respective mean interplanar and ring centroid⋯ring centroid separations are (a) 3.59, 3.84 Å. The O⋯C separation (b) is 3.15 Å. The O–H⋯O hydrogen bond (c) has O⋯O, H⋯O distances 2.65, 1.75 Å, and O–H⋯O angle 175°.Isothiazole bond lengths (Å) for structures8a,8g,8j,9jand138a8g8j9j13S(1)–N(2)1.635(2)1.642(2)1.644(2)1.635(3)1.639(2)N(2)–C(3)1.327(2)1.311(3)1.335(3)1.298(3)1.323(3)C(3)–C(4)1.404(3)1.410(3)1.424(3)1.432(3)1.422(3)C(4)–C(5)1.368(3)1.368(3)1.364(3)1.381(3)1.365(3)C(5)–S(1)1.707(2)1.714(2)1.701(2)1.704(3)1.708(2)The solid state structure of the sulfone8gis illustrated inFig. 3. The isothiazole phenyl sulfone adopts an open-book conformation analogous to that invariably observed for unstrained diphenyl sulfones,4the torsional twists about the C(3)–S(6) and S(6)–C(14) linkages beingca.89 and 85° respectively. The terminal phenyl ringBis less steeply inclined to the isothiazole ring than in8a, the 40° inclination being comprised of torsional twists of 18 and 25° about C(5)–C(15) and C(15)–C(21) respectively. As observed in8a, the carbonyl oxygen is orientedsynto the isothiazole sulfur. The pattern of bonding in the isothiazole ring does not differ markedly from that observed in8a. Unlike in8a, the C(15) and S(6) substituents lie only 0.056 and 0.008 Å respectively above and below the isothiazole ring plane (which is planar to within 0.007 Å).C2related molecules stack with the isothiazole ringAof one molecule overlaying the phenyl ringBof the next andvice versa(ainFig. 4). This stacking interaction is supplemented by a pair of C–H⋯π interactions between one of themetaC–H hydrogen atoms of ringBin one molecule and the phenyl sulfone ringCin the next (binFig. 4). These “dimer pairs” are linked by C–H⋯O hydrogen bonds between the isothiazole methine hydrogen in one molecule and one of the sulfone oxygen atoms O(8) in the next (cinFig. 4). These interactions combine to form continuouschains of molecules, though here there are no obvious interchain interactions of note.The molecular structure of8g.Part of one of the continuous chains of linked ``dimers'' present in the structure of8g. The respective mean interplanar and ring centroid⋯ring centroid separations are (a) 3.68, 3.91 Å. The C–H⋯π interaction (b) has H⋯π 2.82 Å and C–H⋯π 147 °. The C–H⋯O hydrogen bond (c) has C⋯O, H⋯O distances 3.31, 2.40 Å, and C–H⋯O angle 157°.As with the 5-phenyl derivatives, 5-tert-butylfuran-2-carboxylic acid also decomposed under the standard conditions, but its ethyl ester reacted cleanly to give ethyl 5-pivaloylisothiazole-3-carboxylate in 57 and 80% yield respectively in pyridine–toluene and in isoquinoline–chlorobenzene. A limited range of activated 2-phenylfurans was also examined. 2-Phenylfuran itself did not give any isothiazole with the Katz reagent in toluene, but instead a low yield (15%) of bis(5-phenylfuran-2-yl) sulfide11together with much decomposition. The structure of11was confirmed by a more obvious synthesis from 2-phenylfuran and sulfur dichloride in the presence of pyridine. The formation of11with the Katz reagent is not too surprising since the free α-position of 2-phenylfuran is highly susceptible to electrophilic attack, forexample by thionyl chloride to give the furansulfinyl chloride which can react again with starting furan to give the bis-furanyl sulfoxide;5this could be reduced by more thionyl chloride, a known process,6or by a related intermediate species, to the sulfide11. Substitution of the free α-position of the 2-phenylfuran by thiazyl chloride, NSCl, is also possible but this does not lead to a reasonably low energy ring-opening and ring-closing sequence to give an isothiazole, as the β-thiazyl derivatives do. With both α-positions blocked, 2-methyl-5-phenylfuran did give the expected 5-benzoyl-3-methylisothiazole8, X = Me in 40 and 50% yield in toluene and chlorobenzene respectively.Finally and more interestingly, 5-phenyl-2-phenylthiofuran12reacted with the Katz reagent in toluene (8 h) and in chlorobenzene (30 min) to give both possible isomeric 5-acylisothiazoles13and14in the yields shown (Table 3). The structures of13and14followed from the spectroscopic data, and structure13was confirmed by X-ray crystallography. In the starting furan12the β-position adjacent to the phenylthio group is the more nucleophilic and preferential attack here leads to the major isothiazole13. This result should be contrasted with the same reaction of the corresponding sulfone7gwhich gave exclusively the isothiazole8gresulting from attack at the β-position non-adjacent to the phenylsulfonyl group, as expected. This illustrates, once more, that the more electron-withdrawing α-substituent of the furan becomes the isothiazole 3-substituent and the more electron-releasing one becomes part of the isothiazole 5-acyl substituent.Treatment of 5-phenyl-2-phenylthiofuran12with ethyl carbamate, thionyl chloride and a base in a boiling solventBaseSolventReaction time/hYield (%)13Yield (%)14PyridineToluene83417IsoquinolineChlorobenzene0.53012The X-ray analysis of13(Fig. 5) shows the isothiazole and its adjacent phenyl ring (B) to be in almost perfect conjugation with each other, the torsional twist about the C(3)–C(11) linkage being onlyca.2°. The carbonyl oxygen is again orientedsynto the ring sulfur atom, though the group lies much closer to the isothiazole ring plane [torsional twist ofca.8° about C(15)–C(12)]. The terminal phenyl ringCis oriented almost orthogonally to the remainder of the molecule, there being aca.80° rotation about the S(13)–C(19) bond. Within the isothiazole ring, the bond lengths show only minor variations compared with those in8aand8g(Table 2). The molecules are linked by π–π stacking interactionsto form continuous chains (Fig. 6). There is (i) a centrosymmetric reverse-stacking of the phenylisothiazole ring systems, ringAoverlaying ringBandvice versa(ainFig. 6) and (ii) a parallel stacking of ringsCabout an independent inversion centre (binFig. 6). This latter interaction is supplemented by a pair of possible weak C–H⋯O hydrogen bonding interactions across the same inversion centre between theparahydrogen on ringCin one molecule and the carbonyl oxygen O(12) of the next (cinFig. 6). Adjacent chains are loosely linked by aromatic edge-to-face interactions between the edges of ringsCin one chain and the faces of ringsAin the next (the centroid⋯centroid separations are 5.14 Å).The molecular structure of13.Part of one of the continuous chains of molecules in the structure of13. The mean interplanar and centroid⋯centroid separations are (a) 3.60, 3.93 Å and (b) 3.58, 3.98 Å respectively. The C–H⋯O hydrogen bond (c) has C⋯O, H⋯O distances 3.30, 2.56 Å, and C–H⋯O angle 135°.Silylated furansBecause of the well known replaceability of a silyl group by a wide range of other substituents, we finally investigated the transformation of silylated furans into the corresponding silylated isothiazoles. This would be of particular value in that, for example, simple acylisothiazoles are hardly known, and could be made by appropriate protodesilylation.Both 5-silylated-2-phenylfurans7h–7jand 2,5-di-silylated furans were readily synthesised by reaction of the corresponding organolithium derivative with a trialkylsilyl chloride. The silyl groups investigated were trimethylsilyl, dimethyl-tert-hexylsilyl, and triisopropylsilyl.Reaction of the 5-silylated-2-phenylfurans7h–7jwith the Katz reagent in refluxing toluene with pyridine as the base, generated the expected 5-benzoyl-3-silylated isothiazoles8h–8jin 38, 47 and 35% respectively (Table 4). However, in boiling chlorobenzene with pyridine or isoquinoline as the base the trimethylsilylfuran7hgave instead the chlorodesilylated product,9m(38%). The more heavily substituted silylfuran derivatives7iand7jin hot chlorobenzene gave the expected products8iand8jin slightly lower yields, together with an analogous isothiazolebearing an extra chlorine substituent in place of the isothiazole ring hydrogen. Comparative spectral data together with X-ray crystallography on both of the isothiazoles derived from7j, revealed, surprisingly, that the chlorinated product was the rearranged 3-chloro-4-silylated-5-benzoylisothiazole9j. Furthermore, the isolated isothiazole8jwas not transformed, as expected, into the rearranged product by further Katz treatment in chlorobenzene! The starting furans were also not chlorinated by treatment with thionyl chloride in refluxing chlorobenzene, or by pyridine hydrochloride in the same solvent. The yield of the rearranged products is greatest when technical grade thionyl chloride is utilised, suggesting possibly that sulfur chlorides are the key chlorinating agents bringing about the rearrangement. Thus with technical grade thionyl chloride using pyridinein chlorobenzene, after 6 h reflux the yields of8jand9jwere 58 and 31% respectively; using purified thionyl chloride under the same conditions, the same products were isolated in 70 and 21% yields.Conversion of furans into 5-acylisothiazolesMethod AMethod BMethod CFuranIsothiazoleTime/hYield (%)Time/hYield (%)Time/hYield (%)Isoquinoline used as base.Pyridine used as base.Half of the thionyl chloride added at the start of reaction and the other half added after half reaction time.Only 4.3 mmol thionyl chloride used.Thionyl chloride added dropwise throughout the reaction.1a2a24986700.5 a801c2c666—12 a403a4a18561850.5 a953b4b1699—10 min a8054b4813—1 a6864a—24150.75 a127b8b—32548 a527c8c—2142 a247d8d—5202 a277f8f—1222.5 a517g8g—48556 a607h8h/9m—248h: 381.5 b9m: 357i8i/9i—488i: 472 b8i: 6;9i: 257i8i/9i——2 b,c8i: 38;9i: 117j8j/9j—248j: 356 a8j: 12;9j: 417j8j/9j——6 a8j: 31;9j: 477j8j/9j——6 a,d8j: 24;9j: 367j8j/9j——6 a,e8j: 80;9j: 131213—8340.5 a30+14—8170.5 a122-Me-5-Ph5-PhCO-3-Me—54072 a505-But-2-CO2Et3-EtO2C-5-COBut—245724 a80At present we have no evidence for the mechanism of this novel silyl rearrangement and chlorination process. Optimisation of the conditions allowed the triisopropylsilylfuran7jto give 80% of the unrearranged isothiazole8jtogether with 13% rearranged product9jby addition of thionyl chloride dropwise during the 6 h reflux in chlorobenzene with pyridine as base. Alternatively, using the standard mode of reaction 31% unrearranged together with 47% rearranged isothiazole were isolated.These silylated isothiazoles have so far proved impervious to the action of a variety of electrophiles including mineral acid, bromine or NBS and iodine and silver trifluoroacetate. Tetrabutylammonium fluoride in refluxing THF caused decomposition of the substrates.In the structure of8j, the replacement of the carboxylic acid in8aby a triisopropylsilyl unit is accompanied by a dramatic change in the orientation of the Ph–C&z.dbd;O substituent with respect to the isothiazole ring. As in8a, the phenyl and isothiazole ring planes are fairly steeply inclined (byca.60°), the inclination being a result of torsional twists ofca.34 and 32° about C(5)–C(6) and C(6)–C(7) respectively. However, here in8jthe carbonyl oxygen isantito the ring sulfur atom (Fig. 7)cf. synorientations in8a, 8gand13. The bond length within the isothiazole ring are again little changed compared with those in thepreceding structures (Table 2). The only intermolecular packing interaction of note is a parallel π–π stacking of centrosymmetrically related pairs of phenyl rings for which the mean interplanar and centroid⋯centroid separations are 3.54 and 3.87 Å respectively. The formation of linked chains is here prevented by the steric bulk and hydrophobic exterior of the triisopropylsilyl substituent.The molecular structure of8j.The X-ray structure of9jis shown inFig. 8. Here, the presence of the bulky triisopropylsilyl group on the 4-position forces the Ph–C&z.dbd;O unit to adopt an orientation nearly orthogonal to the isothiazole ring plane; the torsional twists about the C(5)–C(6) and C(6)–C(7) bonds are 82 and 5° respectively. Again there are only small differences in the pattern of bonding within the isothiazole ring compared with those in the other four structures (Table 2). There are no intermolecular packing interactions of note, again presumably because of the congested environment of the ring systems.The molecular structure of9j.Although the differences in the patterns of bonding within the isothiazole rings throughout the five structures investigated are small, there are, over the range, noticeable differences in the relative values for the N(2)–C(3) and C(3)–C(4) distances; within statistical significance, those for C(4)–C(5), C(5)–S(1) and S(1)–N(2) are unchanged (Table 2). There are, however, no obvious correlations between the small changes in the N(2)–C(3) and C(3)–C(4) bond lengths and the electron-withdrawing or -donating nature of the substituents on C(3) and C(4). The marginally longer value for C(4)–C(5) in9jis probably due to the absence of any possible conjugation with the Ph–C&z.dbd;O group attached to C(5),vide supra.The transformation of furans reported in this paper provides a novel and very simple route to a new range of functionalised isothiazoles of considerable interest as components of natural products and as pharmaceutical and agrochemical compounds.

ISSN:1472-7781    

DOI:10.1039/b101157j

出版商:RSC    

年代:2001

数据来源: RSC