|
1. |
Design, synthesis and evaluation of a novel series of spiroketals basedon the structure of the antibacterial gyrase inhibitor novobiocin |
|
Journal of the Chemical Society, Perkin Transactions 1,
Volume 1,
Issue 18,
1997,
Page 2789-2802
William Bell,
Preview
|
|
摘要:
J. Chem. Soc. Perkin Trans. 1 1997 2789 Design synthesis and evaluation of a novel series of spiroketals based on the structure of the antibacterial gyrase inhibitor novobiocin William Bell Michael H. Block,* Calum Cook J. Andrew Grant and David Timms Zeneca Pharmaceuticals Mereside Alderley Park Macclesfield Cheshire UK SK10 4TG Molecular modelling has been used in conjunction with crystallographic and biological data in an attempt to design compounds that mimic the structure and activity of the coumarin antibiotic novobiocin 2. Calculations on four conformations of a 6,6-spiroketal system 8–11 suggest that whilst 9 should be the lowest energy 8 should also be readily accessible and this conformation overlays well with the coumarin and sugar rings of novobiocin. Incorporation of key hydroxy and carbamate groups onto the cyclohexane and a hydrogen bond acceptor onto the aromatic ring suggest 12a as the initial target.The crystal structure of a model compound 24 shows the conformation illustrated in 9 and thus supports the molecular modelling. However the crystal structure of novobiocin bound to a 24 kD fragment of gyrase B the target of the coumarin antibiotics reveals that it is the lactone carbonyl rather than the 4-oxy group which is responsible for a key interaction with an Arg residue and consequently the carboxy group in 12a is misplaced as a corresponding H-bond acceptor. Subsequently compounds incorporating an extra aromatic (as in 13a) or heteroaromatic ring (as in 45) bearing an H-bond acceptor along with the extra dimethyl groups on the cyclohexane ring have been designed.Models of these structures overlay convincingly with novobiocin bound to 24 kD gyrase B. Versatile synthetic routes have been developed allowing these compounds and the underlying hypotheses to be tested. Unfortunately none of the compounds demonstrates significant enzyme or antibacterial activity which we attribute to a combination of features including the lack of a replacement for the noviose methoxy group and the failure to achieve a good stacking (charge transfer) interaction with Arg-76. Introduction DNA Gyrase is a bacterial enzyme responsible for the process of supercoiling DNA.1 It is an enzyme that is vital to bacteria has no direct counterpart in mammals and consequently it represents an attractive antibacterial target. Structurally the enzyme has two discrete subunits (A and B) and it functions as an A2B2 tetramer in which the A subunits are responsible for the binding breaking and recombination of the DNA and the B subunits ‘power’ the process by the hydrolysis of ATP.Two discrete classes of clinically effective antibiotics are known to act at DNA gyrase.2 These are the quinolones and the coumarins typified by ciprofloxacin 1 and the natural product novobiocin 2 respectively. A third class of compounds currently under investigation include the natural products cyclothialidine3 3 a variety of synthetic analogues 4 and the closely related GR122222X5 4. Several other gyrase inhibitors have been reported.6 The quinolones bind to the A subunits forming a stable ternary complex between DNA enzyme and drug which is responsible for the subsequent cell death.1 These compounds have found very widespread use as antibiotics against a whole spectrum of bacteria.The coumarins such as novobiocin 2 bind to the B subunit of DNA gyrase causing inhibition of the ATPase activity.1 Novobiocin 2 itself has found limited clinical use mostly against Gram positive organisms but suffers from problems of toxicity and rapidly developing resistance.7 More potent coumarins such as coumermycin A1 8 and clorobiocin 59 have physical properties such as poor solubility which have limited their development. Cyclothialidine 3 and GR122222X 4 also inhibit DNA gyrase by competitive binding to the B subunit.3,5 The ever increasing incidence of infection resulting from bacteria resistant to currently available therapy has prompted the continuing need to develop new and improved antibacterial agents.The inhibition of DNA gyrase in this case specifically at the B subunit was seen as an attractive approach. In the past there has been considerable work both within Zeneca10 and elsewhere 11 to develop an improved antibacterial agent by modification of the coumarin natural products. The chemistry was highly complex and in trying to establish structure–activity relationships there was always a concern that weak levels of activity may simply be due to contamination with small amounts of the very potent parent compound. However the structural information available for novobiocin 2 and more recently for a complex between novobiocin and a 24 kD protein fragment derived from the B subunit of Escherichia coli DNA gyrase opened up the possibility of designing novel wholly synthetic compounds which mimic novobiocin.This paper outlines the design and synthesis of a novel series of spiroketals intended to mimic those structural features of novobiocin which are believed to be responsible for binding and inactivation of gyrase B. Results and discussion Design At the outset of this project the crystal structure of novobiocin 2 was available12 (see Fig. 1). A 2D 1H NMR experiment confirmed that the key conformational features seen in the solid state were also evident in solution.13 There was also some understanding of the structure–activity relationship within the coumarins which was relevant to the design phase. First and foremost there are features of both the aromatic portion and the sugar which seem essential for activity.Compounds having trivial replacements for either portion are inactive.10 For the aromatic part of the molecule it appears that replacing the terminal phenolic amide with simple amides such as isobutyroyl (as in 6) results in lower but still significant activity.10,11 Similarly the presence of the aromatic methyl substituent (or chloro substituent in the case of clorobiocin 5) is believed to enhance activity but is not essential.11 It has been shown for several bacteria that resistance to novobiocin 2 results from a variety 2790 J. Chem. Soc. Perkin Trans. 1 1997 of point mutations of a ‘key’ arginine residue (equivalent to Arg 136 in E. coli).14 Therefore the interaction with Arg 136 is important for novobiocin 2 and we assumed that the 4-oxy group on the coumarin (as its anion) 15 was the most likely group responsible for this interaction.With regard to the sugar apart from variations to the carbamate group the limited changes that have been investigated were detrimental.10 We set out to design compounds containing a cyclohexyl mimic of the noviose sugar that would carry similar functionality to the natural product and place an aromatic ring in the same vicinity as the hydroxycoumarin. The aromatic ring would bear a hydrogen bond acceptor to pick up the ‘key’ interaction with Arg 136. There is an inherent weakness in targeting a residue which is known to be capable of mutation. However the extent to which novobiocin is compromised in the clinic as a result of gyrase mutation is dependent on several factors which include potency and pharmacokinetics as well as the frequency and viability of the point mutation (equivalent to Arg 136 in E.N N CO2H O HN O O O OH O H2N O O O HN O –Na + O OH O O O OH O O Cl O HN OH O OH O NH O O O OH O H2N O O O HN OH O O HN S OH HO N H O O NH CO2H O O N O NH2 R HO F 1 2 3 R = OH 4 R = H 5 6 coli). We were optimistic that if we could identify activity in a series that had the potential for synthetic manipulation we might then be able to optimise potency and pharmacokinetics whilst enhancing the interactions to other parts of the enzyme thus reducing the effect of mutation of the arginine residue. We hoped that by rigidifying the structure in an appropriate conformation we would enhance binding by reducing the entropic penalty. One problem with using a cyclohexane ring to replace the pyran of the sugar is that in the absence of the anomeric effect and without other influences the O-aryl substituent will occupy an equatorial rather than axial position.Consequently we considered the possibility of having a second oxygen linked substituent in the form of a ketal to balance out this tendency. The use of a cyclic ketal to link a phenyl ring via two oxygen substituents to the cyclohexane ring would have the double advantage of increasing chemical stability and also freezing out the conformation. We first studied a 6,5-spiroketal in which one C]O aryl bond must inevitably be axial and the conformation would be rigidly constrained. Evaluation of a model of this structural type 7 against novobiocin 2 showed the expected overlap of the cyclohexane and pyran rings but also revealed that the phenyl ring was not at all close in space or orientation to the phenyl ring of the coumarin (Fig.2). We next investigated the possibility that a 6,6-spiroketal would have the correct orientation. For this system we considered four ‘low energy’ conformations (8–11) which represent the two ring flipped chair conformations in combination with the aromatic ring either bent towards or away from the oxygenated side of the cyclohexane ring. The relative energies of these conformations were obtained from an MM2 derived molecular mechanics potential 16 and confirmed by low level semiempirical calculations. From these calculations we immediately discount the likelihood of the conformation given in 11. This is because even given the inherent limitations of the potential function and an absence of a solvation model we compute a high relative conformational energy for 11 (also confirmed by the semiempirical calculations).The relative conformational energies of the other structures (8–10) differ only by 2.1 kJ mol21. In this case the limitations of the computational model do not lead to precise conclusions about the conformational populations. However we were satisfied that 8 is an acceptable conformation for the 6,6-spiroketal. Fig. 1 Crystal structure of novobiocin Fig. 2 Overlay of novobiocin with 7 J. Chem. Soc. Perkin Trans. 1 1997 2791 Comparison of 8 with novobiocin 2 showed a good overlap of both the cyclohexane and pyran rings and also the aromatic and coumarin rings (Fig. 3). The only difference observed was a slight tilt of the aromatic rings away from co-planarity.Based on these modelling results our initial target became the spiroketal 12a bearing oxyen functionality on the cyclohexane ring and a carboxy group on the phenyl ring to mimic the 4-oxy group on the coumarin.15 It was then our intention to extend our studies to amides and other hydrogen bond acceptors and also to include the gem-dimethyl group (as in 12b). During the course of this research further highly relevant structural information became available namely the X-ray crystal structure of a complex between novobiocin 2 and a 24 kD protein fragment derived from E. coli gyrase B.17 The conformation for clorobiocin 5 in the protein–ligand complex was also solved in solution by NMR and was in agreement with that observed in the solid state.18 Not only did the crystal structure provide information on the conformation of novobiocin 2 in a ‘biologically relevant’ environment but it also clarified the O O OH O OH O O O OH O O O O O O O H O O O O H 7 8 (+1.7 kJ mol-1) 9 (0 kJ mol-1) 10 (+2.1 kJ mol-1) 11 (+20 kJ mol-1) Fig.3 Overlay of novobiocin with 8 O O CO2H R R OH O H2N O O O Me Me OH O H2N O 12a R = H 12b R = Me CO2H 13a nature of the interactions of novobiocin 2 with those parts of the active site contained within the protein fragment. On the whole comparison of the two structures shows the conformations to be quite similar (Fig. 1 and Fig. 4) but the terminal phenolic group of the bound novobiocin appears to be bent back on itself and there is a slightly increased twist of the coumarin ring relative to the sugar. However the protein crystal structure did reveal that it is the lactone carbonyl (as its oxyanion) rather than the 4-oxy group that binds to Arg-136.This information implied that the carboxy group of 12a would not be in the correct position to pick up the key interactions with Arg-136. As indicated earlier the coumarin rings of novobiocin and the phenyl ring of 12a were not strictly coplanar in our modelling. Hence we did not believe that simply annealing a second aromatic ring would enable the correct positioning of a carboxy group. We considered whether the natural twist imparted by a biphenyl derivative would allow a second aromatic ring to be orientated in a similar manner to the coumarin ring and consequently position a hydrogen bond acceptor correctly. Modelling the biphenyl acid 13a in comparison to novobiocin bound to 24 kD gyrase showed that the natural twist imparted by the biphenyl put the terminal aromatic ring back into co-planarity with the coumarin and placed the carboxy group in a very favourable position to pick up the interaction with Arg-136 (as in Fig.5). As a result of this modelling and the lack of activity seen in the early compounds our synthetic targets became the acid 13a and related compounds. We were also interested in preparing other biaryl systems bearing hydrogen bond acceptor groups. Synthesis Our retrosynthetic analysis of the original target 12a/b is illustrated in Scheme 1. It was recognised that the final products prepared by this route would be racemic but during the phase Fig. 4 Conformation of novobiocin in novobiocin–gyrase B 24 kD complex Fig.5 Overlay of novobiocin from novobiocin–gyrase B 24 kD complex with 13a 2792 J. Chem. Soc. Perkin Trans. 1 1997 of lead identification this was deemed to be acceptable. We envisaged that 12a/b could be prepared from the unsaturated compound 14a/b by conversion of the double bond to the cisdiol and subsequent manipulation to the carbamate along with conversion of the bromo group to a carboxylic acid. We expected that 14a/b could be prepared directly from the corresponding unsaturated ketone 15a/b and 5-bromo-2-hydroxybenzyl alcohol 16. In the event we were unable to condense 15a and 16 directly under a variety of conditions and consequently an alternative sequence was developed (Scheme 2). Cyclohexanone was converted directly to the bromodimethyl ketal 17a using bromine in methanol according to the method of Garbisch.19 Transketalisation with 16 gave the bromo spiroketal 18a which was treated with sodium methoxide in Scheme 1 O O Br R R R R O HO HO Br 12a/b 14a R = H 14b R = Me 15a R = H 15b R = Me 16 + Scheme 2 Reagents a MeOH Br2; b 16 PTSA toluene; c NaOMe DMSO; d TMSI CH2Cl2; e PhMe3N1Br2 O O Br R R R R O O O R R Br O O Br R R Br O O Br R R 15a R = H 15b R = Me O O Br 18a R = H (68%) 18b R = Me 17a R = H 17b R = Me 19 (73%) TMS TMS 14a R = H (91%) 14b R = Me (84% from 19) a b c d e DMSO to give 14a in 62% yield over three steps.In order to establish the relative stereochemistry and to provide structural information the bromo derivatives themselves were progressed to final products (Scheme 3). Dihydroxylation of 14a gave a 1 1 mixture of the diastereomeric diols 20a and 21a in 54% yield which could be separated with difficulty by HPLC.Alternatively the mixture was converted to the carbonates 22a and 23a with triphosgene in 81% yield,20 which could be easily separated by flash chromatography. (Hydrolysis of the single diastereomers 22a and 23a gave the corresponding diols 20a and 21a.) Assigning the relative stereochemistry between these diastereomeric series by 1H NMR spectroscopy proved to be difficult and not entirely convincing. For the des-methyl series where the relevant peaks could be distinguished in both diastereomers (e.g. 22a and 23a) the assignment of relative stereochemistry was tentatively based on the presence of an NOE from a benzylic proton to the proton at position 39 on the cyclohexane ring (for 22a) and the absence of an equivalent result with the diastereomer (23a) (see Fig.6). Definitive proof of the relative stereochemistry came from a single crystal X-ray structure of diol 21a.21 In order to complete the synthesis of the model compound 24 the required carbonate 22a was reacted with ammonia to form a 1 1 mixture of the 2- and 3-carbamates 24 and 25. In this case the two isomers were separated by HPLC with dif- ficulty and we were able to get an X-ray crystal structure of 24.21 Undoubtedly it would be possible to develop a sequence of reactions that would allow selective functionalisation of the 3-hydroxy group. However as our purpose was to identify activity in a lead series we were happy to accept the benefits that the described sequence presented namely straightforward chemistry and the ability to form carbamates from a variety of amines at a late stage.Except where circumstances were particularly favourable we did not separate the 2- and 3-carbamate mixtures but tested the mixture directly for biological activity. For the synthesis of the first genuine target 12a the bromo derivative 14a was transmetallated and quenched with CO2 to form the corresponding acid 26a in 50% yield (Scheme 3). We were concerned about the base sensitivity of the final carbamate and also the acid sensitivity of the spiroketal and therefore chose the trichloroethyl ester as a protecting group for the acid that could be removed with zinc under buffered conditions.22 In retrospect our concerns about the acid sensitivity of the spiroketal were unnecessary. The acid 26a was converted to 27a in 96% yield.Subsequent dihydroxylation gave diol 28 and the corresponding diastereomer 29 in 59% yield (83% based on recovered starting material). This 1 1 mixture was taken through to the carbonate derivatives 30 and 31 in 74% yield and separated by flash chromatography. The assignment of relative stereochemistry was based on 1H NMR NOE experiments along with a comparison of the spectra and NOE results seen with the bromo series. Opening up of the carbonate with Fig. 6 Informative 1H NMR NOEs O O O O O H H H Br O O O O O H Me Me O O O O O H H Br Me Me J. Chem. Soc. Perkin Trans. 1 1997 2793 ammonia in non-alcoholic solvents gave poor results and the use of ethanol resulted in transesterification. Consequently 30 was reacted with ammonia in a small quantity of trichloro- Scheme 3 Reagents and conditions a OsO4 O(CH2CH2)2(Me)N1O2 or Me3N1O2 (54% 1 1 20a 21a; 49% 1 1 20b 21b 66% based on recovered 14b; 59% 1 1 28 29 83% based on recovered 27a); b triphosgene (81% 1 1 22a 23a from 20a 1 21a; 76% 1 1 22b 23b from 20b 1 21b; 74% 1 1 30 31 from 28 1 29); c NH3 EtOH; d BunLi CO2; e DCCI DMAP HOCH2CCl3; f NH3 HOCH2CCl3 CAUTION!; g Zn THF pH 7 phosphate buffer O O Br R R O O X R R O O X R R O O O O X R R HO OH O O R R HO OH X O O O R R O O X O O O X O OH H2N O O O HO O 14a R = H 14b R = Me X 26a R = H X = CO2H (50%) 26b R = Me X = CO2H 27a R = H X = CO2CH2CCl3 (96%) 27b R = Me X = CO2CH2CCl3 22a R = H X = Br 22b R = Me X = Br 30 R = H X = CO2CH2CCl3 O NH2 20a R = H X = Br 20b R = Me X = Br 28 R = H X = CO2CH2CCl3 21a R = H X = Br 21b R = Me X = B r 29 R = H X = CO2CH2CCl3 23a R = H X = Br 23b R = Me X = Br 31 R = H X = CO2CH2CCl3 24 X = Br (16%) 32 X = CH2CO2CCl3 (35%) 12a X = CO2H (58%) 25 X = Br (21%) 33 X = CH2CO2CCl3 (23%) + + + a a b c or f d e g ethanol (CAUTION! potential hazard) to give 32 and 33 in 58% yield which were separated by chromatography.Deprotection with zinc in THF–phosphate buffer (pH 7) gave the target acid 12a in 58% yield. In tackling the synthesis of the dimethyl series (e.g. 13a) we attempted to prepare the key unsaturated spiroketal 14b using analogous chemistry to the des-methyl series (Scheme 2). However this proved completely unsuccessful. Formation of the bromo dimethyl ketal 17b went very poorly and under no circumstances could 17b be converted to the bromo spiroketal 18b even when 17b was freshly distilled.Consequently an alternative sequence was developed (also shown in Scheme 2). Ketalisation of 15b with the bis (TMS) ether derived from 16 using TMSI catalysis gave 19 in good yield (73%).23 Direct bromination could be achieved with phenyltrimethylammonium tribromide24 and the product 18b was converted without purifi- cation to 14b by treatment with sodium methoxide in DMSO (84% yield). The sequence has been performed on a 1 mol scale in 40% yield over the three steps. (The lower than expected isolated yields on the larger scale are a consequence of poor recovery from the final crystallisation. No attempt was made to chromatograph the residues because 14b proved to be fairly unstable to silica.) The unsaturated ketal 14b was progressed through a sequence of transmetallation and quenching with CO2 to give the acid 26b which was protected as its trichloroethyl ester 27b.However this compound proved to be even less stable to silica than the bromo analogue 14b and this instability made studies of the dihydroxylation difficult. In any event these studies were abandoned because of the change of focus towards targets such as 13a which followed on from the protein–novobiocin structural information. However the cis-hydroxylation of 14b was still seen as a critical step and it was with some relief that it was discovered that 14b could be converted to the diols 20b and 21b in 49% yield (66% based on recovered starting material) using slightly modified conditions. More importantly 20b and 21b and all subsequent derivatives of these compounds suffered from no problems of stability.As previously the diols were converted to the carbonates 22b and 23b and then easily separated by flash chromatography (76% yield). In this case the relative stereochemistry could be assigned unambiguously by 1H NMR NOE experiments (see Fig. 6). It was envisaged that the biphenyl linkage in the target spiroketal biphenyl acid 13a could be prepared by a palladium catalysed cross-coupling reaction (Scheme 4). The bromo derivative 22b could be coupled with the o-formyl boronic acid 34a using Pd0 and CsF catalysis25 to give 35 in 84% yield. Alternatively 22b could be coupled with the boronic ester 34b using Pd0 and Cs2CO3 catalysis 26 in 87% yield. The aldehyde 35 was oxidised to the acid 36 using sodium chlorite 27 in 98% yield and the carbonate group of 36 was opened with ammonia to give the target compound 13a along with the isomer 37a as a 1 1 mixture in 73% yield.In order to exploit potential lipophilic interactions the higher carbamates 13b–d and their isomers 37b–d were prepared by reaction with the corresponding amines. In these examples the ratio of isomers varied from 1 1 to 4 1 and in some cases the desired product could be isolated pure by chromatography following the preferential crystallisation of the minor isomer. The route illustrated in Scheme 4 for the synthesis of biaryl derivatives was somewhat limiting when it came to varying the terminal aryl group because of the need to prepare boronic acids (or an equivalent organometallic derivative) for each of the aryl or heteroaryl groups in question.Consequently we evaluated an alternative approach in which the organometallic functionality is contained within the spiroketal component and reacted with a variety of aryl or heteroaryl bromides or iodides (Scheme 5). The choice of appropriate protecting groups for the diol (22b/23b) was not an obvious one. Acyl groups would not be stable to the conditions of transmetallation and in a brief investigation we were only able to silylate the equatorial 3- 2794 J. Chem. Soc. Perkin Trans. 1 1997 hydroxy group. We considered the dimethyl ketal because of its stability toward organometallics and suspected that it might be more labile to acid hydrolysis than the spiroketal and this turned out to be the case. The diols 20b and 21b were converted to the dimethyl ketals 38 and 39 in 85% yield and separated easily by flash chromatography.The bromo group of 38 was converted by transmetallation capture with triisopropyl borate and transesterification to the cyclic boronate ester 40 which could also be purified by chromatography (80% yield). The boronate ester 40 could be coupled with a variety of bromo- or iodo-aromatics using Pd0 and Cs2CO3 catalysis 26 in high yield. For example coupling of 40 with 2-bromopyridine gave the biaryl derivative 41 which could be cleaved very selectively under acidic conditions to give the diol 42. This compound was converted to the carbonate 43 in an overall yield of 44% from 40. Oxidation to the corresponding pyridine N-oxide 44 was accomplished using MCPBA in 91% yield and subsequent reaction of 44 with Scheme 4 Reagents a 34a Pd(PPh3)4 CsF or 34b Pd(PPh3)4 Cs2CO3; b NaOCl2; c RNH2 EtOH O O Br Me Me O O O O O Me Me O O O X CHO M O O Me Me O OH CO2H O O Me Me HO O CO2H RNH O O RNH 34a M = B(OH)2 34b M = B[O(CH2)3O] 35 X = CHO (87%) 36 X = CO2H (98%) 13a R = H (73% 1:1 13a:37a) 13b R = Me (53%) 13c R = Et (57%) 13d R = Pri (68%) 22b 37a R = H 37b R = Me (15%) 37c R = Et (25%) 37d R = Pri (19%) a b c ammonia or methylamine gave the target carbamates 45 and 47 along with the corresponding isomers 46 and 48.The boronate ester 40 was also used to prepare a compound analogous to 13a but containing the 5-methylpyrrole-2- carboxylate instead of the carbamate (Scheme 5). Coupling of Scheme 5 Reagents a CH3C(OCH3)2CH3 H1; b i BunLi B(OPri)3 ii HO(CH2)3OH toluene; c 2-bromopyridine Pd(PPh3)4 Cs2CO3; d 1 M HCl MeOH; e triphosgene; f MCPBA; g RNH2 EtOH; h trichloroethyl 2-iodobenzoate Pd(PPh3)4 Cs2CO3; i 5-methyl-2-trichloroacetylpyrrole K2OBut; j Zn THF pH 7 phosphate buffer O O X Me Me O O O O Me Me O O Br O O Br Me Me HO OH O O Me Me HO OH Br O O Me Me RO OR' N O O Me Me RO OR' CO2CH2CCl3 O O Me Me RO OR' N+ –O O O Me Me O OH CO2R" 22b 23b 38 X = Br (43%) 40 X = B[O(CH2)3O] (80%) 39 (42%) 44 R = R' = CO (91%) 45 R = CONH2 R' = H (50% 1:1 45:46) 46 R = H R' = CONH2 47 R = CONHMe R' = H (82% ca.2:1 47:48) 48 R = H R' = CONHMe 49 R = R' = C(CH3)2 (53%) 50 R = R' = H O NH 41 R = R' = C(CH3)2 42 R = R' = H 43 R = R' = CO (ca. 44% from 40) 51 R"= CH2CCl3 (21% from 50) 52 R"= H (44%) a + + b h c d e d j g i J. Chem. Soc. Perkin Trans. 1 1997 2795 40 with the trichloroethyl ester of 2-iodobenzoic acid using Pd0 and Cs2CO3 as catalysts 26 gave the biphenyl derivative 49 in 53% yield.Again deprotection of the dimethyl ketal was achieved selectively and the resulting diol 50 could be selectively acylated on the 3-hydroxy group to give 51 albeit in poor yield using the trichloroacetylpyrrole and potassium tert-butoxide (21% 27% based on recovered 50). The trichloroethyl ester was then removed as previously using zinc in THF–buffer (pH 7) to give 52 in 44% yield. Evaluation The crystal structure 21 of the model compound 24 shows a conformation essentially identical to that illustrated for structure 9 and is consistent with the modelling results which predicted 9 to be the lowest energy of the four conformations modelled (8–11). Conformation 8 provides the best overlap with novobiocin (Fig.3) but as discussed previously we believe that the calculated energy difference between 8 and 9 if real is small enough to imply that 8 should still be an easily accessible conformation for our target compounds. We therefore believe that the 6,6-spiroketals do provide a scaffold on which to append both functionality equivalent to the noviose sugar and also an appropriately positioned aromatic ring albeit with a slight twist relative to the coumarin. All our synthetic products were tested for gyrase inhibitory activity in a standard supercoiling assay28 at 800 mg ml21 and 200 mg ml21 (or 1600 mg ml21 and 400 mg ml21 where a 1 1 mixture of carbamates was evaluated). In this assay the coumarin standards would have the following activity novobiocin (IC50 = 0.1 mg ml21) clorobiocin (IC50 = 0.01 mg ml21) and the semisynthetic isobutyryl derivative 6 (IC50 = 1 mg ml21).Very little activity was seen for any of the target compounds. The biphenyl acid 13b showed very weak activity in the supercoiling assay (IC50 = 200 mg ml21) but the lack of activity in closely related compounds (13a,c,d) and in particular for the acyl pyrrole derivative 52 was not encouraging. The compounds were also tested in standard antibacterial screens 29 at doses up to 128 mg ml21. The weak antibacterial activity which was seen for some of the intermediates is presumed not to be related to gyrase inhibition. Discussion Having seen the novobiocin–24 kD gyrase B crystal structure it is not surprising that the early compounds such as 12a showed no activity as the carboxy group is not correctly positioned to pick up any interaction with Arg-136.These compounds also lack the gem-dimethyl group which is likely to be important for binding. There are some explanations for the lack of activity seen in the biaryl series which we do not believe are responsible. The most fundamental of these is that our modelling was done with a structure derived from a small fragment of the gyrase B subunit which may not be relevant to the whole enzyme. We do not believe this explanation to be the case. Whilst the 24 kD protein may not contain all the binding sites for novobiocin microcalorimetry clearly demonstrates potent binding of novobiocin to 24 kD protein.30 Furthermore for selected examples of our target compounds (including 13b) we have demonstrated by microcalorimetry that there is no binding to 24 kD protein up to the experimental limits of solubility (250 mg ml21).We have also repeated our calculations on structures analogous to 8–11 but bearing the gem-dimethyl groups and the results are very similar to those for 8–11. The possibility that the oxymethylene bridge in our spiroketals is clashing with some part of the protein which is not observed in the novobiocin–24 kD structure is not supported by inspection of the 43 kD gyrB dimer structure in which the relevant section of protein is present.31 Furthermore if this were the fundamental cause of inactivity against the full enzyme you would still expect to see binding to 24 kD protein. There are several factors which we believe may contribute to the lack of activity. These include the absence of the methoxy group found on the noviose sugar which may be important.Also there appears to be a significant p-stacking (charge transfer) interaction between the hydroxycoumarin of novobiocin 2 and Arg-76 which could contribute up to 13 kJ mol21 of binding energy.32 This interaction may be significantly reduced for the target molecules as a consequence of the lack of co-planarity seen in the overlay of the aromatic ring of modelled target compounds and novobiocin. It is also possible that the terminal aromatic ring is not adequately mimicking the electronic features of the delocalised anion derived from the hydroxycoumarin. The penalty for de-solvating this delocalised anion may also be lower than that for a carboxylic acid. Compounds prepared in further studies which incorporated a double hydrogen bond acceptor in the form of a 3-hydroxypyridine N-Oxide as the terminal aromatic ring (and presumed to be partially ionised) also failed to show activity.In all likelihood it is a combination of features which are really quite subtle that result in the lack of activity and it remains a disappointment that structures that we believe have the correct conformation carry much of the functionality believed to be important for binding and model so well into the protein showed no activity. Conclusions It is an important premise of rational drug design that different ligands that bind to the same receptor site giving rise to a similar pharmacological response possess a degree of molecular shape similarity. This idea has been used in the design of a series of 6,6-spiroketal derivatives based on the structure of a known inhibitor ligand (novobiocin) and its conformation bound to a 24 kD fragment of DNA gyrase.We have developed synthetic routes to these targets and the compounds appear to exhibit much of the spatial (and structural) features of the inhibitor (novobiocin). However the compounds do not show any significant activity against DNA gyrase at concentrations of 104 higher than the IC50 of novobiocin. We believe that the lack of activity arises from a combination of features including the lack of a replacement for the noviose methoxy group and the failure to achieve a good stacking (charge transfer) interaction with Arg-76. Experimental All operations were carried out at room temperature unless otherwise stated. Tetrahydrofuran (THF) was dried by distillation from calcium hydride.Flash chromatography was performed on silica (Merck Kieselgel Art. 9385). Melting points were taken on a Buchi apparatus with the use of glass capillary tubes and are uncorrected. 1H NMR spectra were recorded on Bruker WM200 WM250 or WM400 instruments and are reported as d values (parts per million) relative to Me4Si as internal standard. J Values are given in Hz. Mass spectra were recorded on VG 12-12 quadrupole VG 70-250 SE VG ZAB 2- SE or a VG modified AE1/Kratos MS9 spectrometers. 6-Bromospiro[4H-1,3-benzodioxine-2,19-cyclohex-29-ene] 14a 5-Bromo-2-hydroxybenzyl alcohol 16a (13.65 g 67.3 mmol) 17a19 (10 g 44.8 mmol) and toluene-p-sulfonic acid (PTSA) (0.1 g) were heated at reflux in toluene passing the condensed solvent through a Soxhlet adapter filled with 4 Å molecular sieves.After 8 h the solution was cooled and concentrated under reduced vacuum to a light brown oil which was chromatographed on silica using 10% ethyl acetate in hexane as eluent to give 18a as a clear colourless oil which was used directly (10.95 g 68%) dH(250 MHz; CDCl3) 1.5–2.4 (8H m 4 × CH2) 4.3 (1H s CHBr) 4.74 (1H d J 14 CHAHB) 4.94 (1H d J 14 CHAHB) 6.74 (1H d J 8 ArH) 7.13 (1H d J 2 ArH) 7.25 2796 J. Chem. Soc. Perkin Trans. 1 1997 (1H dd J 3 8 ArH); m/z (EI1) 360 (M1 40%) 184 (C7H5OBr 100). 18a (10.95 g 30.2 mmol) was stirred in DMSO (30 ml) and NaOMe (4.77 g 90 mmol) was added keeping the temperature below 10 8C over the addition and for the following 10 min and then allowing the reaction mixture to warm to room temperature. After 3 h the mixture was poured into brine (150 ml) and extracted with 2-methylpentane (200 ml).The organic extract was dried (MgSO4) filtered and evaporated to give 14a as a clear oil (7.8 g 91%) dH(250 MHz; CDCl3) 1.83 (2H m CH2) 2.0 (2H m CH2) 2.13 (2H m CH2) 4.8 (1H d J 14 CHAHB) 4.9 (1H d J 14 CHAHB) 5.82 (1H d J 10 CHCHCH2) 6.05 (1H dt J 3 10 CHCHCH2) 6.73 (1H d J 8 ArH) 7.12 (1 H d J 2 ArH) 7.26 (1 H dd J 2 8 ArH); m/z (EI1) 280 (M1 75%) 184 (C7H5OBr 65) 96 (C6H8O 100). 6-Bromo-5959-dimethylspiro[4H-1,3-benzodioxine-2,19-cyclohex- 29-ene] 14b 5-Bromo-2-hydroxybenzyl alcohol 16a (4.54 g 22 mmol) was dissolved in THF (50 ml) and N-methyl-N-(trimethylsilyl)- trifluoroacetamide (10.2 ml 55 mmol) was added. The solution was stirred under argon at 55 8C for 30 min and then concentrated under vacuum.The residual material was diluted with 2- methylpentane and filtered through a sinter and the filtrate concentrated to give the bis-TMS ether (7.96 g 100%). A portion of this material (2.6 g 7.45 mmol) and 3,3- dimethylcyclohexanone (940 mg 7.46 mmol) were dissolved in CH2Cl2 (30 ml) and cooled to 278 8C under argon. Trimethylsilyl iodide (ca. 50 mg) was added and the reaction mixture was stirred at 278 8C for 50 min and warmed to 0 8C and stirred for a further 30 min. Saturated aqueous NaHCO3 (30 ml) was added followed by Et2O (60 ml). The organic layer was washed with NaHCO3 (sat. aq. 20 ml) Na2S2O5 (sat. aq. 20 ml) and brine (20 ml) dried over MgSO4–Na2CO3 (1 1) filtered and concentrated to give 19 as a colourless oil (1.69 g 73%) dH(250 MHz; CDCl3) 0.97 (3H s CH3) 0.99 (3H s CH3) 1.34 (2H m CH2) 1.54–1.74 (6H m 3 × CH2) 4.73 (1H d J 15 OCHAHB) 4.83 (1H d J 15 OCHAHB) 6.69 (1H d J 9 ArH) 7.08 (1H d J 2 ArH) 7.23 (1 H dd J 2 9 ArH); m/z (EI1) 310 (100% M1).To a stirred solution of 19 (7.8 g 25.16 mmol) in THF (166 ml) at 0 8C under argon was added rapidly phenyltrimethylammonium tribromide (97%; 9.75 g 25.16 mmol). The resulting solution was stirred for 2 h at 0 8C and then diluted with NaHCO3 (sat. aq. 12 ml) and stirred for a further 30 min at 0 8C. The mixture was concentrated under reduced pressure without heating to remove most of the THF and the residue was extracted with Et2O (2 × 100 ml). The combined organic layers were washed with brine (50 ml) dried over MgSO4 filtered and concentrated. The resulting oil was diluted with hexane (60 ml) filtered and the filtrate concentrated under vacuum to give 18b as a pink gum (9.225 g 94%) which was dissolved in DMSO (50 ml) and stirred at 0 8C under argon.NaOMe (3.85 g 71.3 mmol) was added and the reaction mixture was stirred for 3 h at room temperature and then poured into brine (250 ml) and extracted with hexane (2 × 150 ml). The combined organic layers were dried over MgSO4 filtered and concentrated under vacuum to give 14b as a yellow solid (6.56 g 89%). For analytical purposes a small sample (1 g) was recrystallised from 2-methylpentane (7 ml) to give 14b as yellow crystals (680 mg). (NB 14b is not completely stable to chromatography on silica) mp 111–114 8C (Found C 58.7; H 5.6. C15H17BrO2 requires C 58.3; H 5.5%); dH(200 MHz; CDCl3) 1.02 (3H s CH3) 1.04 (3H s CH3) 1.87 [d J 5 CH2C(OR)2] 1.97 (2H dd J 1.5 4 CHCH2) 4.85 (2H s OCH2) 5.81 (1H ddd J 10 1 1 CH2CHCH) 5.98 (1H ddd J 10 3 3 CH2CHCH) 6.70 (1H d J 8 ArH) 7.10 (1H d J 2 ArH) 7.25 (1H dd J 8 2 ArH); m/z (CI1) 309 (MH1 60%) 125 (C8H13O1 100).(2R*,29R*,39R*)- and (2S*,29R*,39R*)-6-Bromo-29,39- dihydroxyspiro[4H-1,3-benzodioxine-2,19-cyclohexane] 20a and 21a 14a (2.0 g 7.1 mmol) and N-methylmorpholine N-oxide (1.43 g 10.6 mmol) were dissolved in ButOH (40 ml) and OsO4 (4% aq. 0.56 ml 0.07 mmol) was added followed by water (1 ml). After stirring the solution for 3 days the reaction was quenched by the addition of solid sodium metabisulfite (ca. 300 mg) and stirring for 30 min. The mixture was evaporated to dryness under vacuum and chromatographed using ethyl acetate–hexane (1 1) as eluent to give a mixture of 20a and 21a as a colourless oil (1.2 g 54%).For synthetic purposes this mixture was used directly. For analytical purposes a sample of the mixture (246 mg) was purified by HPLC (Dynamax C18 column) using 60% methanol–40% water to give 20a (96 mg) and 21a (71 mg) as white solids. 20a mp 114–116 8C (Found C 49.4; H 4.6. C13H15BrO4 requires C 49.5; H 4.8%); dH(400 MHz; CDCl3) 1.4–1.9 (6H m 3 × CH2) 2.32 (1H d J 8 CH2CHOH) 2.54 (1H d J 3 CH2CHOHCHOH) 3.92 (1H m CH2CHOHCHOH) 3.97 (1H br m CH2CHOHCHOH) 4.87 (2H s OCH2) 6.74 (1H d J 8 ArH) 7.10 (1H d J 2 ArH) 7.27 (1H dd J 2 8 ArH); dC(100 MHz; CDCl3) 17.76 (CH2CH2CH2) 27.34 [CH2C(OR)2] 28.46 (CH2CHOH) 59.88 (OCH2) 69.59 (CH2CHOH) 71.83 (CH2CHOHCHOH) 101.05 [C(OR)2] 113.05 (CBr) 118.95 (CHCHCBr) 121.51 (CCH2O) 127.47 (CHCBr) 131.28 (CHCBr) 149.36 (COCR2O); m/z (CI1) 314 (M1 50%) 297 (M1 2 OH 50) 185 (C7H6OBr 100).21a mp 160–161 8C (Found C 49.8; H 4.8. C13H15BrO4 requires C 49.5; H 4.8%); dH(400 MHz; CDCl3) 1.35–1.9 (6H m 3 × CH2) 2.28 (1H br d J 8 CH2CHOH) 2.49 (1H s CH2CHOHCHOH) 3.9 (2H br s CH2CHOHCHOH) 4.76 (1H d J 14 OCHAHB) 4.83 (1H d J 14 OCHAHB) 6.79 (1H d J 8 ArH) 7.10 (1H d J 2 ArH) 7.29 (1H dd J 2 8 ArH); dC(100 MHz; CDCl3) 17.68 (CH2CH2CH2) 26.33 [CH2C(OR)2] 28.44 (CH2CHOH) 59.62 (OCH2) 69.40 (CH2CHOH) 72.77 (CH2CHOHCHOH) 101.38 [C(OR)2] 113.11 (CBr) 118.93 (CHCHCBr) 121.56 (CCH2O) 127.53 (CHCBr) 131.28 (CHCBr) 149.36 (COCR2O); m/z (CI1) 314 (M1 45%) 297 (M1 2 OH 50) 185 (C7H6OBr 100). (2R*,29R*,39R*)- and (2S*,29R*,39R*)-6-Bromo-29,39- dihydroxy-59,59-dimethylspiro[4H-1,3-benzodioxine-2,19-cyclohexane] 20b and 21b These compounds were prepared from 14b in an approximately 1 1 ratio [49% (66% based on recovered 14b)] in a similar manner as described for 20a and 21a with the exception that trimethylamine N-oxide was used as co-oxidant.For synthetic purposes this mixture was used directly. For analytical purposes careful chromatography on silica using EtOAc–hexane (2 3) gave samples of 20b and 21b as the predominant isomers which were crystallised from EtOAc–hexane to give pure samples. 20b mp 138–139 8C (Found C 52.4; H 5.7. C15H19O4Br requires C 52.5; H 5.6%); dH(400 MHz; CDCl3) 0.97 (3H s CH3ax) 0.99 (3H s CH3eq) 1.54 (1H dd J 5 13 CHeqHaxCHOH) 1.605 [1H d J 14 CHeqHaxC(OR)2] 1.635 (1H d J 13 CHeqHax- CHOH) 1.70 [1H d J 14 CHeqHaxC(OR)2] 2.08 (1H d J 9 CH2CHOH) 2.46 [1H s CHOHC(OR)2] 3.91 [1H s CHOH- (OR)2] 4.14 (1H m CH2CHOH) 4.82 (1H d J 15 OCHAHB) 4.88 (1H d J 15 OCHAHB) 6.70 (1H d J 8 ArH) 7.11 (1H d J 2 ArH) 7.27 (1H dd J 2 8 ArH); m/z (CI1) 360 (M 1 NH4 1 100%) 342 (M 1 H1 20).21b mp 148–149 8C (Found C 52.1; H 5.7. C15H19O4Br requires C 52.5; H 5.6%); dH(400 MHz; CDCl3) 0.98 (3H s CH3eq) 1.09 (3H s CH3ax) 1.53 (1H dd J 5 13 CHeqHaxCHOH) 1.64 [1H d J 14 CHeqHaxC(OR)2] 1.67 (1H d J 13 CHeqHaxCHOH) 1.76 [1H d J 14 CHeqHaxC(OR)2] 2.11 (1H d J 9 CH2CHOH) 2.46 [1H s CHOHC(OR)2] 3.92 [1H s CHOH(OR)2] 4.03 (1H m CH2CHOH) 4.74 (1H d J 17 OCHAHB) 4.88 (1H d J 17 OCHAHB) 6.75 (1H d J 8 ArH) 7.10 (1H d J 2 ArH) 7.27 J. Chem. Soc. Perkin Trans. 1 1997 2797 (1H dd J 2 8 ArH); m/z (CI1) 360 (M 1 NH4 1 100%) 342 (M 1 H1 20). (2R*,3a9R*,7a9R*)- and (2S*,3a9R*,7a9R*)-6-Bromospiro[4H- 1,3-benzodioxine-2,49-perhydro[1,3]benzodioxol]-29-one 22a and 23a The mixture of 20a and 21a (ca.1 1; 1.2 g 3.8 mmol) was dissolved in CH2Cl2 (10 ml) and pyridine (1.8 g 22.8 mmol). The solution was cooled to 278 8C under an argon atmosphere and a solution of triphosgene (564 mg 1.9 mmol) in CH2Cl2 (10 ml) was added dropwise. The mixture was allowed to warm to room temperature and then quenched with ammonium chloride (sat. aq. 10 ml) and diluted with CH2Cl2. The organic layer was washed with 2 M HCl (10 ml) NaHCO3 (sat. aq. 20 ml) and brine (20 ml). The solution was dried (MgSO4) filtered and concentrated. The residue was chromatographed on silica using 30% EtOAc–hexane as eluent to give 22a (550 mg Rf 1:1 EtOAc–hexane 0.6) and 23a (510 mg Rf 1 1 EtOAc–hexane 0.4) as colourless oils which crystallised on standing (yield 81%).22a mp 151–153 8C (Found C 49.7; H 3.7. C14H13O5Br requires C 49.3; H 3.8%); dH(250 MHz; CDCl3) 1.65–2.15 (6H m 3 × CH2) 4.70 (1H d J 6 CHOCO) 4.83 (1H d J 16 OCHAHB) 4.88 (1H d J 6 CHOCO) 4.95 (1H d J 16 OCHAHB) 6.75 (1H d J 8 ArH) 7.15 (1H d J 2 ArH) 7.3 (1H dd J 2 8 ArH); m/z (Cl1) 358 (M 1 NH4 1 100%). 23a mp 94–99 8C (Found C 49.3; H 3.7. C14H13O5Br requires C 49.3 H 3.8%); dH(250 MHz; CDCl3) 1.6–2.2 (6H m 3 × CH2) 4.58 (1 H d J 6 CHOCO) 4.84 (2H s OCH2) 4.92 (1H m CH2CHOCO) 6.82 (1H d J 8 ArH) 7.11 (1H d J 2 ArH) 7.3 (1H dd J 2 8 ArH); m/z (CI1) 358 (M 1 NH4 1 100%). (2R*,3a9R*,7a9R*)- and (2S*,3a9R*,7a9R*)-6-Bromo-59,59- dimethylspiro[4H-1,3-benzodioxine-2,49-perhydro[1,3]- benzodioxol]-29-one 22b and 23b A mixture of diols 20b and 21b (3.83 g 11.1 mmol) was converted to a mixture of the carbonates 22b and 23b using the same procedure as described above for 22a and 23a.The crude product started to crystallise on standing and CH2Cl2 (20 ml) was added and 23b was filtered off as a white solid (1.165 g 29%). The filtrate was concentrated under vacuum and chromatographed on silica using 10–50% EtOAc in hexane as eluent to give 22b as an oil which crystallised on the addition of Et2O (5 ml) and hexane (10 ml) (1.45 g 37%) along with a further quantity of 23b (0.4 g 10% total 39%). 22b mp 124–125 8C (Found C 52.1; H 4.7. C16H17BrO5 requires C 52.1; H 4.6%); dH(400 MHz; CDCl3) 1.09 (3H s CH3ax) 1.12 (3H s CH3eq) 1.79–1.93 [4H m CH2C(CH3)2CH2] 4.69 (1H dd J 1.5 7.5 CH2CHOCHO) 4.84 (1H d J 16 OCHAHB) 4.96 (1H d J 16 OCHAHB) 4.98 (1H m CH2CHOCHO) 6.71 (1H d J 8 ArH) 7.15 (1H d J 2 ArH) 7.29 (1H dd J 2 8 ArH); irradiation of CH3ax (1.09) gives an NOE enhancement of ArH (6.71); m/z (CI1) 386 (M 1 NH4 1 100%) 368 (M 1 H1 10.23b mp 178–181 8C (Found C 51.8; H 4.6. C16H17BrO5 requires C 52.1; H 4.6%) dH(400 MHz; CDCl3) 1.07 (3H s CH3ax) 1.11 (3H s CH3eq) 1.705 (1H d J 15 CHeqHaxCHO) 1.89 [1H dd J 5.5 14 CHeqC(OR)2] 1.96 (1H d J 15 CHeqHaxCHO) 1.97 [1H d J 14 CHaxHeqC(OR)2] 4.64 (1H d J 7 CH2CHOCHO) 4.83 (1H d J 16 OCHAHB) 4.87 (1H d J 16 OCHAHB) 4.98 (1H m CH2CHOCHO) 6.79 (1H d J 8 ArH) 7.11 (1H d J 2 ArH) 7.30 (1H dd J 2 8 ArH); irradiation of CH3ax (1.07) gives an NOE enhancement of OCHAHB (4.85); m/z (CI1) 386 (M 1 NH4 1 100%) 368 (M 1 H1 10).(2R*,29R*,39R*)-6-Bromo-29-hydroxy-39-carbamoyloxyspiro- [4H-1,3-benzodioxine-2,19-cyclohexane] 24 and (2S*,29R* 39R*)-6-bromocarbamoyloxy-39-hydroxyspiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 25 22a (490 mg) was dissolved in sat. ethanolic ammonia (5 ml) and the solution stirred for 16 h at room temperature and then evaporated under reduced pressure to give 24 and 25 in an approximately 1 1 mixture. A sample of the mixture (400 mg) was separated by HPLC on a 5 mm silica preparative Dynamax column using 0.1% TFA–15% propan-2-ol–hexane as eluent to give pure samples of the components as gums along with the remainder of the material as a mixture. The gums crystallised on the addition of 30% diethyl ether–hexane (5 ml) to give white crystals of 24 (85 mg 16%) and 25 (110 mg 21%). 24 was recrystallised from ethyl acetate (2 ml) and hexane (10 ml) which gave single crystals of a 1 1 EtOAc adduct suitable for X-ray diffraction.24 mp 78–80 8C (Found C 48.3; H 5.5; N 3.2. C14H16NO5Br?C4H8O2 requires C 48.4; H 5.4; N 3.1%); dH(250 MHz; CDCl3) 1.25 (3H t J EtOAc) 1.4–1.9 (6H m 3 × CH2) 2.0 (3H s EtOAc) 4.08 (1H dd J 0.5 3 CHOH) 4.25 (2H q EtOAc) 4.73 (2H br s NH2) 4.83 (2H s OCH2) 5.0 (1H m CHOCONH2) 6.76 (1H d J 8 ArH) 7.08 (1H d J 2 ArH) 7.26 (1H dd J 2 8 ArH); m/z (CI1) 358 (MH1 60%) 174 (C7H12NO4 1 100). 25 mp 140–144 8C (Found C 46.6; H 4.4; N 3.6. C14H16NO5Br requires C 46.9; H 4.5; N 3.9%); dH(250 MHz; CDCl3) 1.4–1.9 (6H m 3 × CH2) 3.98 (1H m CHOH) 4.67 (1H d J 14 OCHAHB) 4.7 (2H br s NH2) 4.87 (1H d J 14 OCHAHB) 5.12 (1H d J 3.5 CHOCONH2) 6.67 (1H d J 8 ArH) 7.04 (1H d J 2 ArH) 7.21 (1H dd J 2 8 ArH); m/z (CI1) 358 (MH1 30%) 174 (C7H12NO4 1 100) 148 (C7H10NO3 1 100).Spiro[4H-1,3-benzodioxine-2,19-cyclohex-29-ene]-6-carboxylic acid 26a ButLi (1.5 M in pentane; 18 ml 27 mmol) was placed in a dry three-necked flask fitted with a thermometer and under an atmosphere of argon. The solution was cooled in a dry ice– acetone bath to 260 8C and diethyl ether (75 ml) was added. A solution of 14a (3 g 10.7 mmol) in diethyl ether (3 ml) was added keeping the temperature below 260 8C. After 5 min carbon dioxide was bubbled into the solution and the reaction mixture was allowed to warm to room temperature quenched with NH4Cl (sat. aq. 50 ml) and partitioned. The aqueous layer was further extracted with diethyl ether (3 × 40 ml) then adjusted to pH 5 with conc.HCl and extracted with diethyl ether (40 ml) and the combined organic layers were dried (MgSO4) filtered and concentrated. The resultant solid was slurried with 20% ethyl acetate–hexane and filtered to give 26a as white crystals (1.31 g 50%) mp 252–254 8C (decomp.) (Found C 68.3; H 5.8. C14H14O4 requires C 68.3; H 5.7%); dH(250 MHz; [2H6]DMSO) 1.75 (2H m CH2) 1.93 (2H m CH2) 2.10 (2H m CH2) 4.93 (2H s OCH2) 5.81 (1H d J 10 CH2CHCH) 6.08 (1H dt J 3 10 CH2CHCH) 6.89 (1H d J 8 ArH) 7.72 (1H d J 2 ArH) 7.76 (1H dd J 2 8 ArH); m/z (CI1) 247 (MH1 20%) 97 (C6H9O1 100). 1,1,1-Trichloroethyl spiro[4H-1,3-benzodioxine-2,19-cyclohex- 29-ene]-6-carboxylate 27a To a stirred mixture of 26a (1.2 g 4.87 mmol) 1,3-dicyclohexylcarbodiimide (DCC) (1.11 g 5.4 mmol) and DMAP (61 mg 0.5 mmol) in CH2Cl2 was added dropwise 1,1,1- trichloroethanol (800 mg 5.35 mmol) maintaining the temperature below 5 8C with an ice bath.The mixture was stirred for 18 h at room temperature filtered and the filtrate concentrated under vacuum to a gum which was chromatographed on silica using 20% EtOAc in hexane as eluent to give 27a as a colourless oil which solidified on standing (1.86 g 96%) mp 78–81 8C (Found C 50.9; H 4.2. C16H15C13O4 requires C 50.9; H 4.0%); dH(200 MHz; CDCl3) 1.6–2.2 (6H m CH2CH2CH2) 4.95 (4H s OCH2 1 CO2CH2) 5.63 (1H dt J 1 10 CHCHCH2) 6.08 (1H dt J 3 10 CHCHCH2) 6.93 (1H d J 9 ArH) 7.80 (1H d J 2 ArH) 7.96 (1H dd J 2 9 Ar); m/z (FAB1) 377 (100% MH1). 1,1,1-Trichloroethyl (2R*,3a9R*,7a9R*)- and (2S*,3a9R* 7a9R*)-spiro[4H-1,3-benzodioxine-2,49-perhydro[1,3]benzodioxole]- 6-carboxylate 30 and 31 These compounds were prepared from 27a by dihydroxylation 2798 J.Chem. Soc. Perkin Trans. 1 1997 as described for 20b/21b followed by reaction with triphosgene as described for 22a/23a. The resulting oil was chromatographed on silica using 30% EtOAc–hexane as eluent to give 30 and 31 (57% from 27a) as colourless oils and 30 crystallised on standing. 30 Less polar mp 180–182 8C (Found C 47.1; H 3.6. C17H15O7Cl3 requires C 46.7; H 3.45%); dH(400 MHz; C6D6) 0.9–1.5 (6H m CH2CH2CH2) 3.79 (2H m CHOCHO) 4.04 (1H d J 15 OCHAHB) 4.20 (1H d J 15 OCHAHB) 4.665 (1H d J 12 CO2CHAHB) 4.705 (1H d J 12 CO2CHAHB) 6.60 (1H d J 8 Ar) 7.59 (1H d J 1 Ar) 8.00 (1H dd J 1 8 Ar); dC(100 MHz; CDCl3) 15.30 (CH2CH2- CH2CHO) 25.67 (CH2CH2CH2CHO) 28.27 (CH2CH2CH2- CHO) 60.31 (OCH2) 74.35 (CO2CH2) 75.83 (CH2CHOCHO) 76.30 (CH2CHOCHO) 95.09 (CCl3) 98.60 (COROR9) 117.40 (Ar-CH) 119.20 (Ar-CCO2) 121.86 (Ar-CCH2) 127.55 (Ar-CH) 130.81 (Ar-CH) 154.27 (Ar-CO) 154.38 (OCO2) 164.19 (CO2CH2); m/z (FAB1) 459 (M 1 Na1 70%) 436 (M1 30) 289 (M 2 OCH2CCl3 1 100).31 More polar (Found M1 435.9851. C17H15O7Cl3 requires M 435.9883); dH(400 MHz; C6D6) 0.8–1.7 (6H m CH2CH2CH2) 3.67 (1H d J 6 CH2CHOCHO) 3.81 (1H m CH2CHOCHO) 4.11 (1H d J 15 OCHAHB) 4.24 (1H d J 15 OCHAHB) 4.67 (2H s CO2CH2) 6.77 (1H d J 8 Ar) 7.66 (1H d J 1 Ar) 7.96 (1H dd J 1 8 Ar); dC(100 MHz; CDCl3) 15.16 (CH2CH2CHO) 25.51 (CH2CH2CH2CHO) 30.01 (CH2CH2CH2CHO) 60.89 (OCH2) 74.36 (CO2CH2) ca. 76 (hidden CH2CHOCHO) 78.45 (CH2CHOCHO) 95.14 (CCl3) 98.65 (COROR9) 118.10 (Ar-CH) 119.21 (Ar-CCO2) 12.56 (Ar-CCH2) 127.24 (Ar- CH) 130.83 (Ar-CH) 154.07 (Ar-CO) 155.00 (OCO2) 164.34 (CO2CH2); m/z (FAB1) 459 (M 1 Na1 80%) 436 (M1 50) 289 (M 2 OCH2CCl3 2 100).1,1,1-Trichloroethyl (2R*,29R*,39R*)-29-hydroxy-39-carbamoyloxyspiro[ 4H-1,3-benzodioxine-2,19-cyclohexane]-6-carboxylate 32 and 1,1,1-trichloroethyl (2S*,29R*,39R*)-29-carbamoyloxy- 39-hydroxyspiro[4H-1,3-benzodioxine-2,19-cyclohexane]-6-carboxylate 33 The cyclic carbonate 30 (200 mg 0.45 mmol) was dissolved in 1,1,1-trichloroethanol (2 ml) and liquid ammonia (conc. 1 ml) was added with CAUTION and the reaction mixture was then stirred at room temperature overnight. The mixture was partitioned between CH2Cl2 (20 ml) and brine (20 ml). The organic layer washed with brine (3 × 10 ml) dried over MgSO4 filtered and concentrated under vacuum.The resulting oil was chromatographed on silica using 50–70% EtOAc–hexane as eluent to afford 32 (72 mg 35%) and 33 (48 mg 23%) as colourless oils. 32 Less polar (Found M1 453.0134. C17H18NO7Cl3 requires M 453.0149); dH(400 MHz; C5D5N 1 CD3CO2D) 1.6–2.2 (6H m CH2CH2CH2) 4.61 (1H s CHOH) 4.85 (1H d J 15 OCHAHB) 5.08 (1H d J 15 OCHAHB) 5.11 (2H s CO2CH2) 5.31 (1H m CHOCONH2) 6.99 (1H d J 8 Ar) 7.82 (1H d J 1 Ar) 7.97 (1H dd J 1 8 Ar); m/z (CI1) 454 (MH1 100%). 33 More polar (Found M1 453.0131. C17H18NO7Cl3 requires M 453.0149); dH(400 MHz; C5D5N 1 CD3CO2D) 1.57–2.04 (6H m CH2CH2CH2) 4.32 (1H m CHOH) 4.89 (1H d J 15 OCHAHB) 5.18 (1H d J 11 CO2CHAHB) 5.29 (1H d J 11 CO2CHAHB) 5.53 (1H d J 15 OCHAHB) 5.85 (1H d J 1 CHOCONH2) 6.99 (1H d J 8 Ar) 7.91 (1H d J 1 Ar) 8.01 (1H dd J 1 8 Ar); m/z (CI1) 454 (MH1 100%).(2R*,29R*,39R*)-29-Hydroxy-39-carbamoyloxyspiro[4H-1,3- benzodioxine-2,19-cyclohexane]-6-carboxylic acid 12a The trichloroethyl ester 32 (60 mg 0.13 mmol) was dissolved in THF (3 ml) and zinc dust (100 mg) followed by NaH2PO4 (1 M aq.; 0.65 ml) and H3PO4 (1 M aq; 0.65 ml). The mixture was stirred for 18 h and then a further portion of zinc dust (100 mg) was added. The mixture was filtered through Celite diluted with water (10 ml) and extracted with EtOAc (10 ml). The aqueous layer was adjusted to pH 4 with conc. HCl and extracted with EtOAc (3 × 10 ml). The combined organic layers were dried over MgSO4 filtered and concentrated to give 12a as a white solid (32 mg 58%) mp 203–205 8C (Found M1 323.0983. C15H17NO7 requires M 323.1005); dH(250 MHz; [2H6]DMSO 1 CD3CO2D) 1.5–1.7 (6H m CH2CH2CH2) 3.74 (1H dd J 2 5 CHOH) 4.67 (1H m CHOCONH2) 4.81 (1H d J 15 OCHAHB) 4.99 (1H d J 15 OCHAHB) 5.33 (1H d J 5 OH) 6.36 (2H br s NH2) 6.72 (1H d J 8 Ar) 7.70 (1H d J 1 Ar) 7.75 (1H dd J 1 8 Ar) 12.5 (1H br s CO2H); m/z (CI1) 324 (MH1 100%).(2R*,3a9R*,7a9R*)-6-(2-Formylphenyl)-59,59-dimethylspiro- [4H-1,3-benzodioxine-2,49-perhydro[1,3]benzodioxol]-29-one 35 A mixture of the boronic ester 34b (917 mg 4.83 mmol) 22b (1.19 g 3.22 mmol) and caesium carbonate (1.63 g 5. mmol) was stirred in DMF (25 ml) under an atmosphere of argon. Palladium tetrakis(triphenylphosphine) (75 mg 0.06 mmol) was added and the reaction mixture heated at 80 8C for 4 h. The mixture was cooled to room temperature diluted with CH2Cl2 (30 ml) and water (30 ml) and separated. The organic layer was dried over MgSO4 and then concentrated under vacuum and chromatographed on silica using 20–30% EtOAc–hexane as eluent to give 35 as an off white gum (1.1 g 87%) mp.105–112 8C (Found C 70.1; H 5.4. C23H22O6 requires C 70.0; H 5.6%); dH(200 MHz; [2H6]DMSO) 1.06 (3H s CH3) 1.11 (3H s CH3) 1.5–2.0 [4H m CH2C(CH3)2CH2] 5.01 (2H s OCH2) 5.04 (1H d J 6 CH2CHOCHO) 5.19 (1H m CH2CHOCHO) 6.99 [1H d J 8 ArC(6)H] 7.23 [1H d J 2 ArC(3)H] 7.29 [1H dd J 2 8 ArC(5)H] 7.51 [1H dt J 1 7 ArC(49)H] 7.58 [1H dd J 1 7 ArC(69)H] 7.75 [1H dt J 2 7 ArC(59)H] 7.91 [1H dd J 2 7 ArC(39)H] 9.92 (1H s CHO); m/z (CI1) 395 (MH1 100%). (2R*,3a9R*,7a9R*)-6-(2-Carboxyphenyl)-59,59-dimethylspiro- [4H-1,3-benzodioxine-2,49-perhydro[1,3]benzodioxol]-29-one 36 The aldehyde 35 (159 mg 0.4 mmol) was dissolved in THF (3 ml) and sulfamic acid (52 mg 0.53 mmol) was added along with water (2 ml).The mixture was stirred for 5 min and then sodium chlorite (47 mg 0.52 mmol) was added and the reaction mixture was stirred for 30 min at room temperature. EtOAc (5 ml) and water (5 ml) were added and the organic layer separated dried over MgSO4 filtered and concentrated under vacuum to give the acid 36 as a colourless gum (162 mg 98%) dH(200 MHz; [2H6]DMSO) 1.08 (3H s CH3) 1.12 (3H s CH3) 1.5–2.0 [4H m CH2C(CH3)2CH2] 4.99 (2H s OCH2) 4.96 (1H d J 6 CH2CHOCHO) 5.18 (1H m CH2CHOCHO) 6.90 [1H d J 8 ArC(6)H] 7.12 [1H d J 2 ArC(3)H] 7.19 [1H dd J 2 8 ArC(5)H] 7.37 [1H dd J 2 7 ArC(69)H] 7.43 [1H dt J 1 7 ArC(49)H] 7.55 [1H dt J 2 7 ArC(59)H] 7.73 [1H dd J 2 7 ArC(39)H]; m/z (CI1) 411 (MH1 100%). (2R*,29R*,39R*)-39-Carbamoyloxy-6-(2-carboxyphenyl)-29- hydroxy-59,59-dimethylspiro[4H-1,3-benzodioxine-2,19-cyclohexane] 13a and (2R*,29R*,39R*)-29-carbamoyloxy-6-(2-carboxyphenyl)- 39-hydroxy-59,59-dimethylspiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 37a The carbamates 13a 1 37a were prepared from 36 (158 mg 0.38 mmol) and ethanolic ammonia in an analogous fashion to that described for 13b and 37b except that the 1 1 mixture of isomers that were produced was not separated (120 mg 73%).Satisfactory analyses were obtained with spectral properties similar to those of 13b and 37b. (2R*,29R*,39R*)-6-(2-Carboxyphenyl)-29-hydroxy-59,59- dimethyl-39-methylcarbamoyloxyspiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 13b and (2R*,29R*,39R*)-6-(2-carboxyphenyl)- 39-hydroxy-59,59-dimethyl-29-methylcarbamoyloxyspiro[4H-1,3- benzodioxine-2,19-cyclohexane] 37b The carbonate 36 (250 mg 0.61 mmol) was stirred in a saturated solution of methylamine (30% in industrial methylated spirits) for 18 h and then diluted with water (10 ml) adjusted to pH 2 with conc.HCl and extracted with EtOAc (10 ml). The J. Chem. Soc. Perkin Trans. 1 1997 2799 organic layer was dried over MgSO4 filtered and concentrated under vacuum to give 13b and 37b in a 2 1 mixture as an oil. Et2O was added and 37b was precipitated as a white solid which was collected by filtration (39 mg 15%). The filtrate which still contained a mixture of 13b and 37b was chromatographed on silica using 0.1% HOAc–70–100% EtOAc–hexane as eluent to give a pure sample of 13b (144 mg 53%). 13b (Rf 0.1% HOAc– EtOAc 0.2) (Found M1 441.1760.C24H27NO7 requires M 441.1787); dH(250 MHz; [2H6]DMSO) 0.93 (3H s CH3) 1.09 (3H s CH3) 1.3–1.9 [4H m CH2C(CH3)2CH2] 3.88 (1H m CHOH) 4.77 (1H d J 15 OCHACHB) 4.91 (1H d J 15 OCHACHB) 4.99 (1H m CHOCONHMe) 5.27 (1H d J 6 CHOH) 6.82 [1H d J 8 ArC(6)H] 6.90 (1H m NH) 7.03 [1H d J 2 ArC(3)H] 7.12 [1H dd J 2 8 ArC(5)H] 7.36 [1H dd J 2 7 ArC(69)H] 7.40 [1H dd J 2 7 ArC(49)H] 7.53 [1H dt J 1 7 ArC(59)H] 7.68 [1H dt J 2 7 ArC(59)H]; m/z (ESP1) 442 (MH1 100%). 37b (Rf 0.1% HOAc–EtOAc 0.15) (Found M1 441.1745. C24H27NO7 requires M 441.1787); dH(250 MHz; [2H6]DMSO) 0.95 (3H s CH3) 1.10 (3H s CH3) 1.3–1.7 [4H m CH2C(CH3)2CH2] 2.58 (3H d J 4 NHCH3) 4.00 (1H m CHOH) 4.70 (1H d J 6 CHOCONHCH3) 4.73 (1H d J 15 OCHACHB) 4.94 (1H d J 15 OCHACHB) 5.06 (1H m CHOH) 6.84 [1H d J 8 ArC(6)H] 6.97 (1H m NH) 7.03 [1H d J 2 ArC(3)H] 7.13 [1H dd J 2 8 ArC(5)H] 7.33 [1H dd J 2 7 ArC(69)H] 7.41 [1H dd J 2 7 ArC(49)H] 7.53 [1H dt J 1 7 ArC(59)H] 7.67 [1H dt J 2 7 ArC(59)H]; m/z (ESP1) 442 (MH1 70%) 216 (MH1 2 C10H17NO4 100).(2R*,29R*,39R*)-6-(2-Carboxyphenyl)-29-hydroxy-59,59- dimethyl-39-ethylcarbamoyloxyspiro[4H-1,3-benzodioxine-2,19- cyclohexane] 13c and (2R*,29R*,39R*)-6-(2-carboxyphenyl)-39- hydroxy-59,59-dimethyl-29-ethylcarbamoyloxyspiro[4H-1,3- benzodioxine-2,19-cyclohexane] 37c The carbamates 13c 1 37c were prepared from 36 (250 mg 0.61 mmol) and ethylamine in ethanol in an analogous fashion to that described for 13b and 37b to give 13c (158 mg 57%) as a gum and 37c as a white solid (68 mg 25%). Satisfactory analyses were obtained with spectral properties similar to those of 13b and 37b.(2R*,29R*,39R*)-6-(2-Carboxyphenyl)-29-hydroxy-59,59- dimethyl-39-isopropylcarbamoyloxyspiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 13d and (2R*,29R*,39R*)-6-(2-carboxyphenyl)- 39-hydroxy-59,59-dimethyl-29-isopropylcarbamoyloxyspiro[4H- 1,3-benzodioxine-2,19-cyclohexane] 37d The carbamates 13d 1 37d were prepared from 36 (250 mg 0.61 mmol) and ethylamine in ethanol in an analogous fashion to that described for 13b and 37b to give 13d (195 mg 68%) as a gum and 37d as a white solid (55 mg 19%). Satisfactory analyses were obtained with spectral properties similar to those of 13b and 37b. (3R*,3a9R*,7a9R*)- and (2S*,3a9R*,7a9R*)-6-Bromo- 29,29,69,69-tetramethylspiro[4H-1,3-benzodioxine-2,49-perhydro- [1,3]benzodioxole] 38 and 39 A mixture of the diols 22b and 23b (6.6 g 19.2 mmol) was dissolved in DMF (50 ml) and 2,2-dimethoxypropane (10 ml) was added along with conc.HCl (0.2 ml). The solution was stirred for 18 h at room temperature and then partitioned between EtOAc (150 ml) and water (150 ml). The organic layer was washed with water (2 × 100 ml) and brine (100 ml) dried over NaSO4 filtered and concentrated. The residual oil was chromatographed on silica using 3–6% EtOAc–hexane as eluent to give 38 (3.2 g 43%) as a colourless oil and 39 (3.1 g 42%) as a white solid. 38 (Found C 56.9; H 5.9. C18H23BrO4 requires C 56.4; H 6.05%); dH(400 MHz; CDCl3) 1.015 (3H s CH2- CCH3axCH3eq) 1.08 (3H s CH2CCH3axCH3eq) 1.34 (3H s CH3) 1.55 (3H s CH3) 1.635 (1H dd J 7 14 CHeqCHax- CHO) 1.75 (1H dd J 5 14 CHeqCHaxCHO) 1.78 [2H s CH2C(OR)2] 4.185 (1H d J 5.5 CH2CHOCHO) 4.50 (1H dd J 2 6 CH2CHOCHO) 4.82 (1H d J 15 OCHAHB) 4.98 (1H d J 15 OCHAHB) 6.70 (1H d J 9 ArH) 7.125 (1H d J 2 ArH) 7.265 (1H dd J 2 9 ArH); irradiation of CH3ax (1.015) gives an NOE enhancement of ArH (6.70); m/z (CI1) 383 (MH1 70%) 202 (C7H5OBr 1 NH4 1 100).39 Mp 66– 75 8C (Found C 56.7; H 6.2. C18H23BrO4 requires C 56.4; H 6.05%); dH(400 MHz; CDCl3) 1.04 (3H s CH2CCH3axCH3eq) 1.07 (3H s CH2CCH3axCH3eq) 1.35 (3H s CH3) 1.50 (3H s CH3) 1.67–1.85 [4H m CH2C(CH3)2CH2] 4.155 (1H dd J 1 6 CH2CHOCHO) 4.485 (1H dd J 7 13 CH2CHOCHO) 4.83 (2H s OCH2) 6.815 (1H d J 9 ArH) 7.075 (1H d J 2 ArH) 7.25 (1H dd J 2 9 ArH); irradiation of CH3ax (1.04) gives an NOE enhancement of OCH2 (4.83); m/z (CI1) 383 (MH1 20%) 216 (C11H18O3 1 NH4 1 100). (2R*,3a9R*,7a9R*)-6-(1,3,2-Dioxaborinan-1-yl)-29,29,69,69- tetramethylspiro[4H-1,3-benzodioxine-2,49-perhydro[1,3]- benzodioxole] 40 38 (3.1 g 8.1 mmol) was dissolved in THF (80 ml) and cooled under argon in a dry ice–acetone bath.BunLi (1.6 M in hexane; 5.6 ml 8.9 mmol) was added over 10 min and the reaction mixture stirred for a further 5 min before adding triisopropyl borate (1.82 g 9.7 mmol) rapidly. The mixture was allowed to warm to room temperature diluted with saturated aqueous ammonium chloride (100 ml) and extracted with EtOAc (100 ml). The organic layer was washed with water (2 × 50 ml) and concentrated under vacuum. Toluene (100 ml) was added followed by propane-1,3-diol (5 ml) and the mixture stirred for 1 h at room temperature diluted with water and the organic layer concentrated under vacuum to give 40 (2.6 g 80%) as a pale yellow oil which crystallised on standing mp 125–127 8C (Found C 65.2; H 7.5.C21H29BO6 requires C 65.0; H 7.5%); dH(200 MHz; CDCl3) 1.03 (3H s CH3ax) 1.07 (3H s CH3eq) 1.33 (3H s CH3) 1.57 (3H s CH3) 1.6–1.8 (2H m CH2CHO) 1.81 [2H s CH2C(OR)2] 2.03 [2H quintet J 5 CH2(CH2O)2B] 4.14 [4H t J 5 CH2(CH2O)2B] 4.20 (1H d J 6.5 CH2CHOCHO) 4.49 (1H m CH2CHOCHO) 4.85 (1H d J 15 OCHAHB) 5.02 (1H d J 15 OCHAHB) 6.78 (1H d J 8 ArH) 7.41 (1H d J 2 ArH) 7.56 (1H dd J 2 8 ArH); m/z (EI1) 388 (M1 20%) 330 (M1 2 C3H6O 20) 190 (C11H18O3 1 100). (2R*,29R*,39R*)-29,39-Dihydroxy-59,59-dimethyl-6-(2-pyridyl)- spiro[4H-1,3-benzodioxine-2,19-cyclohexane] 42 Boronic ester 40 (339 mg 0.87 mmol) and 2-bromopyridine (138 mg 0.87 mmol) were dissolved in DMF (10 ml). Cs2CO3 (425 mg 1.3 mmol) was added and the mixture placed under an atmosphere of argon.Pd(PPh3)4 (20 mg 0.017 mmol) was added and the mixture stirred at 100 8C for 2 h and then cooled to room temperature. The reaction mixture was partitioned between EtOAc (40 ml) and water (40 ml) and the organic layer was washed with water (2 × 20 ml) and brine (20 ml) dried over Na2SO4 filtered and concentrated to give crude 41 (385 mg) as an oil which was dissolved in MeOH (15 ml) and 1 M HCl (3 ml) was added. The solution was stirred for 3 h at room temperature left standing at 0 8C for 15 h and then diluted with water 20 ml) and adjusted to pH 11 with 1 M NaOH. The aqueous solution was extracted with EtOAc (2 × 30 ml) and the combined organics were washed with brine (20 ml) dried over Na2SO4 filtered and concentrated to an oil.Trituration with Et2O gave a small quantity of 42 as a white crystalline solid (55 mg 18%) along with the filtrate which was concentrated to an oil (250 mg) containing mostly 42 which was progressed without further purification. For the solid mp 123–129 8C (Found C 69.7; H 6.8; N 3.7; H2O 0.6. C20H23NO4?0.1H2O requires C 70.0; H 6.8; N 4.1; H2O 0.5%); dH(250 MHz; CDCl3) 0.98 (3H s CH3ax) 1.02 (3H s CH3eq) 1.50–1.72 [4H m CH2C- (CH3)2CH2] 2.08 (1H d J 9 CH2CHOH) 2.45 [1H s CHOHC- (OR)2] 3.97 [1H s CHOH(OR)2] 4.18 (1H m CH2CHOH), 2800 J. Chem. Soc. Perkin Trans. 1 1997 4.92 (1H d J 15 OCHAHB) 5.02 (1H d J 15 OCHAHB) 6.90 [1H d J 8 Ar(6)H] 7.19 [1H m pyr(5)H] 7.6–7.8 [4H m Ar(3 1 5)H 1 pyr(3 1 4)H] 8.65 [1H m pyr(6)H]; m/z (ESP1) 342 (MH1 90%) 184 (100 MH1 2 C8H14O3).(2R*,3a9R*,7a9R*)-69,69-Dimethyl-6-(2-pyridyl)spiro[4H-1,3- benzodioxine-2,49-perhydro[1,3]benzodioxol]-29-one 43 The compound was prepared from crude 42 in a similar manner as described for 22a (ca. 44% yield from 40) which was used directly dH(250 MHz; CDCl3) 1.11 (3H s CH3ax) 1.13 (3H s CH3eq) 1.8–2.05 [4H m CH2C(CH3)2CH2] 4.75 (1H dd J 1 8 CHO) 4.95 (1H d J 15 OCHAHB) 5.0 (1H m CHO) 5.08 (1H d J 15 OCHAHB) 6.91 [1H d J 8 Ar(6)H] 7.20 [1H m pyr(5)H] 7.6–7.8 [4H m Ar(3 1 5)H 1 pyr(3 1 4)H] 8.65 [1H m pyr(6)H]. 2R*,3a9R*,7a9R*)-69,69-Dimethyl-6-(1-oxido-2-pyridinio)- spiro[4H-1,3-benzodioxine-2,49-perhydro[1,3]benzodioxol]-29- one 44 50% m-Chloroperbenzoic acid (125 mg 0.36 mmol) was dissolved in CH2Cl2 (2 ml) and the organic layer added to a solution of 43 (110 mg 0.3 mmol) in CH2Cl2 (10 ml).The solution was stirred at room temperature for 15 h and then concentrated and chromatographed on silica using 5% MeOH–CH2Cl2 as eluent to give 44 as a colourless oil (105 mg 91%) which was used directly dH(250 MHz; CDCl3) 1.12 (3H s CH3ax) 1.13 (3H s CH3eq) 1.75–2.0 [4H m CH2C(CH3)2CH2] 4.75 (1H dd J 1 8 CHO) 4.91 (1H d J 15 OCHAHB) 5.0 (1H m CHO) 5.08 (1H d J 15 OCHAHB) 6.92 [1H d J 8 Ar(6)H] 7.21 [1H dt J 1 6 pyr(4)H] 7.28 [1H t J 6 pyr(5)H] 7.60 [1H dd J 1 8 Ar(5)H] 7.65 [1H d J 1 Ar(3)H] 8.30 [1H d J 6 pyr(6)H]; m/z (ESP1) 384 (MH1 100%). (2R*,29R*,39R*)39-Carbamoyloxy-29-hydroxy-59,59-dimethyl-6- (1-oxido-2-pyridinio)spiro[4H-1,3-benzodioxine-2,19-cyclohexane] 45 and (2R*,29R*,39R*)-29-carbamoyloxy-39-hydroxy- 59,59-dimethyl-6-(1-oxido-2-pyridinio)spiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 46 The compounds were prepared as a 1 1 mixture from 44 in a similar manner as described for 24/25 (26 mg 50%) (Found M1 400.1625.C21H24N2O6 requires M 400.1634); dH(200 MHz; CDCl3) (45 major isomer) 1.04 (3H s CH3ax) 1.14 (3H s CH3eq) 1.55–1.9 [4H m CH2C(CH3)2CH2] 4.1 (1H br s CHOH) 4.73 (1H br s OH) 4.9 (2H m OCHAHB) 5.25 (1H m CHOCONH2) 6.90 [1H d J 8 Ar(6)H] 7.3–7.75 (7H m 5 × ArH 1 NH2) 8.55 [1H dd J 0.5 8 pyr(6)H]; (46 minor isomer) 1.01 (3H s CH3ax) 1.11 (3H s CH3eq) 1.5–1.9 [4H m CH2C(CH3)2CH2] 4.55 (1H m CHOH) 4.74 (1H br s OH) 4.9 (2H m OCHAHB) 5.24 (1H br s CHOCONH2) 6.95 [1H d J 8 Ar(6)H] 7.3–7.75 (7H m 5 × ArH 1 NH2) 8.55 [1H dd J 0.5 8 pyr(6)H]; m/z (ESP1) 401 (MH1 95%) 200 (100 M1 2 C9H15NO4). (2R*,29R*,39R*)-2-Hydroxy-59,59-dimethyl-39-methylcarbamoyloxy- 6-(1-oxido-2-pyridinio)spiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 47 and (2R*,29R*,39R*)-3-hydroxy-59,59- dimethyl-29-methylcarbamoyloxy-6-(1-oxido-2-pyridinio)- spiro[4H-1,3-benzodioxine-2,19-cyclohexane] 48 The compounds were prepared as a 2 1 mixture from 44 in a similar manner as described for 24/25 except using methylamine (44 mg 82%) (Found M1 414.1794.C22H26N2O6 requires M 414.1791); dH(400 MHz; CDCl3) (47 major isomer) 1.01 (3H s CH3ax) 1.12 (3H s CH3eq) 1.5–1.9 [4H m CH2C(CH3)2CH2] 2.81 (3H d J 5 NHCH3) 4.1 (1H br s CHOH) 4.79 (1H br s OH) 4.92 (1H d J 15 OCHAHB) 4.99 (1H d J 15 OCHAHB) 5.29 (1H m CHOCONHMe) 6.98 [1H d J 8 Ar(6)H] 7.23 1 7.31 1 7.44 [3 × 1H m pyr(3,4,5)H] 7.55 [1H dd J 2 7 Ar(5)H] 7.77 [1H d J 2 Ar(3)H] 8.33 [1H d J 8 pyr(6)H]; (48 minor isomer) 1.04 (3H s CH3ax) 1.15 (3H s CH3eq) 1.5– 1.9 [4H m CH2C(CH3)2CH2] 2.87 (3H d J 5 NHCH3) 4.3 (1H m CHOH) 4.79 (1H br s OH) 4.85 (1H d J 15 OCHAHB) 4.95 (1H d J 15 OCHAHB) 5.24 (1H br s CHOCONHMe) 6.94 [1H d J 8 Ar(6)H] 7.23 1 7.31 1 7.44 [3 × 1H m pyr(3,4,5)H] 7.63 [1H dd J 2 7 Ar(5)H] 7.77 [1H m Ar(3)H] 8.33 [1H d J 8 pyr(6)H]; m/z (ESP1) 415 (MH1 85%) 200 (100 M1 2 C10H17NO4).(2R*,3a9R*,7aR*)-29,29,69,69-Tetramethyl-6-[2-(1,1,1-trichloroethoxycarbonyl) phenyl]spiro[4H-1,3-benzodioxine-2,49- perhydro[1,3]benzodioxole] 49 This compound was prepared from 40 and 1,1,1-trichloroethyl 2-iodobenzoate in a similar manner as described for 41 (420 mg 53%) mp 57–58 8C (Found C 58.6; H 5.5. C27H29C13O6 requires C 58.3; H 5.3%); dH(200 MHz; CDCl3) 1.05 (6H s CH3ax 1 CH3eq) 1.26 (3H s CH3) 1.43 (3H s CH3) 1.55–- 1.8 [4H m CH2C(CH3)2CH2] 4.15 (1H dd J 1 6.5 CH2CHOCHO) 4.46 (1H m CH2CHOCHO) 4.76 (1H d J 15 OCHAHB) 4.90 (2H s CO2CH2) 4.94 (1H d J 15 OCHA HB) 6.82 (1H d J 8 ArH) 7.05 (1H d J 2 ArH) 7.13 (1H dd J 2 8 Ar9H) 7.43 (1H dd J 2 8 ArH) 7.51 (1H dt J 2 8 Ar9H) 7.65 (1H dt J 2 8 Ar9H) 7.83 (1H dd J 2 8 Ar9H); m/z (CI1) 555 (M1 100%).(2R*,29R*,39R*)-(29,39-Dihydroxy)-59,59-dimethyl-6-[2-(1,1,1- trichloroethoxycarbonyl)phenyl]spiro[4H-1,3-benzodioxine-2,19- cyclohexane] 50 The acetonide 49 (360 mg 0.65 mmol) was dissolved in acetonitrile (10 ml) and HCl (1 M; 3 ml) was added. The solution was stirred for 2 h at room temperature and then quenched by the addition of solid NaHCO3. Water (10 ml) was added and the aqueous solution was extracted with EtOAc (30 ml) and the organic layer washed with brine (20 ml) dried over Na2SO4 filtered and concentrated under vacuum to give 50 (350 mg) as a slightly impure colourless oil which was used directly dH(200 MHz; CDCl3) 0.92 (3H s CH3ax) 1.05 (3H s CH3eq) 1.25 (1H m CHeqHaxCHOH) 1.46–1.74 [3H m CH2C(CH3)2CHeqHax- CHOH] 3.70 (1H m CH2CHOHCHOH) 3.85 (1H m CH2CHOHCHOH) 4.38 (1H d J 7 OH) 4.49 (1H d J 6 OH) 4.75 (1H d J 15 OCHAHB) 4.89 (1 H d J 15 OCHAHB) 4.90 (2H s CO2CH2) 6.80 (1H d J 8 ArH) 7.02 (1H d J 2 ArH) 7.12 (1H dd J 2 8 Ar9H) 7.43 (1H dd J 2 8 ArH) 7.51 (1H dt J 2 8 Ar9H) 7.65 (1H dt J 2 8 Ar9H) 7.84 (1 H dd J 2 8 Ar9H); m/z (ESP1) 532 (M 1 NH4 1 100%).(2R*,29R*,39R*)-29-Hydroxy-59,59-dimethyl-39-(5-methylpyrrol- 2-ylcarbonyloxy)-6-[2-(1,1,1-trichloroethoxycarbonyl)- phenyl]spiro[4H-1,3-benzodioxine-2,19-cyclohexane] 51 The diol (50 284 mg 0.55 mmol) and 5-methyl-2-trichloroacetylpyrrole (375 mg 1.65 mmol) were dissolved in THF (30 ml) and potassium tert-butoxide (74 mg 0.66 mmol) was added.The mixture was stirred at reflux for 2 h and further portions of the pyrrole (200 mg 0.88 mmol) and potassium tert-butoxide (80 mg 0.71 mmol) were added. After stirring at reflux for a further 2 h more potassium tert-butoxide (100 mg 0.89 mmol) was added and the mixture stirred at reflux for 15 h cooled and diluted with NH4Cl (sat. aq. 30 ml) and extracted with EtOAc (30 ml). The organic layer was washed with brine (20 ml) dried over Na2SO4 filtered and concentrated under vacuum to a yellow oil which was chromatographed on silica using 1 2 EtOAc–hexane as eluent to give 51 as a colourless gum (73 mg 21%) along with recovered 50 (57 mg 27% yield of 51 based on recovered starting material).51 (Found M1 621.1034. C30H30NO7Cl3 requires M 621.1088); dH(250 MHz CDCl3) 1.02 (3H s CH3ax) 1.15 (3H s CH3eq) 1.2–1.75 [4H m CH2C(CH3)2CH2] 2.30 (3H s pyrrCH3) 4.14 (1H m CHOH) 4.75 (2H s OCH2) 4.90 (2H s CO2CH2) 5.52 (1H m CHOCO) 5.95 (1H m pyrrC3-H) 6.81 (1H m pyrrC4-H) 6.90 (1H d J 8 ArH) 6.95 (1H d J 2 ArH) 7.14 (1H dd J 2 8 Ar9H) 7.35 (1H dd J 2 8 ArH) 7.42 J. Chem. Soc. Perkin Trans. 1 1997 2801 (1H dt J 2 8 Ar9H) 7.60 (1H dt J 2 8 Ar9H) 7.93 (1H dd J 2 8 Ar9H) 9.0 (1H br s NH); m/z (CI1) 622 (M 1 H1 100%). (2R*,29R*,39R*)-6-(2-Carboxyphenyl)-29-hydroxy-59,59- dimethyl-39-(5-methylpyrrol-2-carbonyloxy)spiro[4H-1,3- benzodioxine-2,19-cyclohexane] 52 The trichloroethyl ester 51 (73 mg 0.117 mmol) was dissolved in THF (5 ml).Zinc dust (250 mg) was added along with NH4OAc (1M; 6 ml) and acetic acid (5 ml). The mixture was stirred at room temperature for 2 h and then filtered through Celite washing the residue with water (5 ml) and EtOAc (10 ml). A further portion of EtOAc (20 ml) was added and the organic layer separated and washed with water (10 ml) and brine (10 ml) dried over Na2SO4 filtered and concentrated under vacuum. The residue was chromatographed on silica using 7–12% MeOH–CH2Cl2 as eluent to afford 52 (25 mg 44%) as a colourless glassy solid (Found M1 491.1906. C28H29NO7 requires M 491.1944); dH(250 MHz; CDCl3) 0.98 (3H s CH3ax) 1.11 (3H s CH3eq) 1.2–1.95 [4H m CH2C(CH3)2CH2] 2.20 (3H s pyrrCH3) 4.15 (1H m CHOH) 4.85 (2H s OCH2) 5.46 (1H m CHOCO) 5.93 (1H s pyrrC3-H) 6.75–6.85 (3H m pyrrC4-H 1 2 × ArH) 7.13 (1H d J 8 Ar9H) 7.30–7.40 (2H m ArH 1 Ar9H) 7.50 (1H t J 8 Ar9H) 7.86 (1H d J 8 Ar9H) 9.4 (1H br s NH); m/z (ESP1) 492 (M 1 H1 100%).Acknowledgements We thank S. Brightwell Dr K. J. Barrett-bee Mrs L. Dawson Mrs C. M. Morton and Dr R. N. Illingworth for the biological testing of the compounds; Mr G. A. Holdgate and Dr W. H. J. Ward for differential scanning calorimetry studies with gyrase B 24kD fragment; Mr B. Wright Mr H. A. Beeley and Mr P. A. McNally for NMR support; Mrs M. F. Vickers for MS support and Dr M. C. McPartlin for crystal structures of 21a and 24. References 1 For a review of Gyrase structure and function see R. J. Reece and A. Maxwell Crit. Rev. Biochem. Mol. Biol. 1991 26 335. 2 K.Orlica and S. Coughlin Pharmacol. Ther. 1989 44 107; S. Radl Pharmacol. Ther. 1990 48 1. 3 N. Nakada H. Shimada T. Hirata Y. Aoki T. Kamiyama J. Watanabe and M. Arisawa Antimicrob. Agents Chemother. 1993 37 2656; E. Goetschi P. Angehrn H. Gmünder P. Hebeisen H. Link R. Masciadri and J. Nielsen Pharmacol. Ther. 1994 60 367; J. Watanabe N. Nakada S. Sawaira H. Shimada S. Ohshima T. Kamiyama and M. Arisawa J. Antibiot. 1994 47 32; T. Kamiyama N. Shimma T. Ohtsuka N. Nakayama Y. Itezono N. Nakada J. Watanabe and K. Yokose J. Antibiot. 1994 47 37; N. Nakada H. Gmünder T. Hirata and M. Arisawa Antimicrob. Agents Chemother. 1994 38 1966. 4 M. Arisawa E. Gotshi T. Kamiyama R. Masciadri H. Shimada J. Watanabe P. Hebeisen and H. Link EP-106105 1991; E. Goschi P. Hebeisen H. Link and T. Lubbers EP-0-675-122-A2 1995.5 M. Oram B. Dosanjh N. A. Gormley C. V. Smith L. M. Fisher A. Maxwell and K. Duncan Antimicrob. Agents Chemother. 1996 40 473. 6 R. H. Chen D. N. Whittern A. M. Buko and J. B. McAlpine J. Antibiotics 1989 42 533; M. S. Osbourne W. M. Maiese and Mgreenstein Antimicrob. Agents Chemother. 1990 34 1450; C. Hubschwerlen I. Kompis and J.-L. Specklin EP 0 433 648 A1 1990; J. L. Vizan C. Hernandez-Chico I. del Castillo and F. Moreno EMBO J. 1991 10 467; S. Sakemi T. Inagaki K. Kaneda H. Hiraia E. Iwata T. Sakakibara Y. Yamauchi M. Norcia L. M. Wondrack J. A. Sutcliffe and N. Kojima J. Antibiot. 1995 48 134. 7 L. P. Garrod H. P. Lambert and F. O’Grady Antibiotic and Chemotherapy Churchill Livingstone Edinburgh 5th edn. 1981 pp. 225–229. 8 H. Kawaguchi H. Tsukiura M.Okanishi T. Miyaki T. Ohmori K. Fujisawa and H. Koshiyama J. Antibiot. 1965 18 1; J. Berger A. J. Schochter A. D. Batcho B. Pecherer O. Keller J. Maricq A. E. Karr B. P. Vaterlaus A. Furlenmeier and H. Speigelberg Antimicrob. Agents Chemother. 1965 9 778. 9 L. Linet F. Benazet Y. Chapentie M. Dubost J. Florent D. Mancy J. Preud’Homme T. L. Threlfall B. Vuillemin D. E. Wright A. Abraham M. Cartier N. de Chezelles C. Godard and J. Theilleux C. R. Acad. Sci. Ser. C. 1972 275 455. 10 J. P. Poyser K. J. Rogers N. Taylor D. Hollinshead A. J. Platt J. Culshaw A. Glen D. Heaton G. Kay and L. Feron unpublished results Zeneca Pharmaceuticals. 11 For reviews of the chemical modifications to the coumarin antibiotics and the resultant SAR see M. J. Ryan Antibiotics vol. V ed. F. E. Hahn Springer Verlag 1979 214; J.G. Keil J. C. Godfrey M. J. Cron I. R. Hooper D. E. Nettleton K. E. Price and H. Schmitz; J. Berger and A. D. Batcho J. Chromatogr. 1978 15 101. 12 M. O. Boles and D. J. Taylor Acta Crystallogr. Sect. B 1975 31 1400. 13 B. Wright unpublished results Zeneca Pharmaceuticals. 14 M. L. Holmes and M. L. Dyall-Smith J. Bacteriol. 1991 173 642; A. Contreras and A. Maxwell Mol. Microbiol. 1992 6 1617; M. Stieger P. Angehrn B. Wohlgensinger and H. Gmünder Antimicrob. Agents. Chemother. 1996 40 1060. 15 The pKa of Novobiocin has been measured as 4.3 (D. Newton and R. Kluza Drug Intell. Clin. Pharm. 1978 12 546). 16 N. L. Allinger J. Am. Chem. Soc. 1977 99 8127. 17 R. J. Lewis O. M. P. Singh C. V. Smith A. Maxwell T. Skarzynski A. J. Wonacott and D. Wigley J. Mol. Biol.1994 241 128; R. J. Lewis O. M. P. Singh C. V. Smith T. Skarzynski A. Maxwell A. J. Wonacott and D. B. Wigley EMBO J. 1996 15 1412. 18 F. T. F. Tsai O. M. P. Singh T. Skarzynski A. J. Wonacott S. Weston A. Tucker R. A. Pauptit A. L. Breeze J. P. Poyser R. O’Brien J. E. Ladbury and D. B. Wigley Proteins Struct. Funct. Genet. 1997 28 41. 19 E. Garbisch J. Org. Chem. 1965 30 2109. 20 R. M. Burk and M. B. Roof Tetrahedron Lett. 1993 34 395. 21 The X-ray crystal structures for 21a and 24 were solved by Dr M. C. McPartlin at the University of North London unpublished results. 22 G. Just and K. Grozinger Synthesis 1976 457. 23 T. Tsunoda M. Suzuki and R. Noyori Tetrahedron Lett. 1980 21 1357. 24 B. D. Mookherjee R. R. Patel and W. O. Ledig J. Org. Chem. 1971 36 4124. 25 S. W. Wright D. L.Hageman and L. D. McLure J. Org. Chem. 1994 59 6095. 26 T. Watanabe N. Miyaura and A. Suzuki Synlett. 1992 207. 27 F. M. Hauser and S. R. Ellenberger Synthesis 1987 723. 28 M. Gellert K. Mizuuchi M. H. O’Dea and H. A. Nash Proc. Natl. Acad. Sci. USA 1976 73 3872. 29 National Committee for Clinical Laboratory Standards (1993) Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically 3rd edn.; approved standard. Document M7-A3 13 (25) Villanova PA. 30 J. A. Holdgate and W. H. J. Ward unpublished results Zeneca Pharmaceuticals. 31 D. B. Wigley G. J. Davies E. J. Dodson A. Maxwell and G. Dodson Nature 1991 351 624. 32 E. M. Duffy P. J. Kowalczyk and W. L. Jorgenson J. Am. Chem. Soc. 1993 115 9271. Paper 7/00647K Received 28th January 1997 Accepted 2nd June 1997 J.Chem. Soc. Perkin Trans. 1 1997 2789 Design synthesis and evaluation of a novel series of spiroketals based on the structure of the antibacterial gyrase inhibitor novobiocin William Bell Michael H. Block,* Calum Cook J. Andrew Grant and David Timms Zeneca Pharmaceuticals Mereside Alderley Park Macclesfield Cheshire UK SK10 4TG Molecular modelling has been used in conjunction with crystallographic and biological data in an attempt to design compounds that mimic the structure and activity of the coumarin antibiotic novobiocin 2. Calculations on four conformations of a 6,6-spiroketal system 8–11 suggest that whilst 9 should be the lowest energy 8 should also be readily accessible and this conformation overlays well with the coumarin and sugar rings of novobiocin.Incorporation of key hydroxy and carbamate groups onto the cyclohexane and a hydrogen bond acceptor onto the aromatic ring suggest 12a as the initial target. The crystal structure of a model compound 24 shows the conformation illustrated in 9 and thus supports the molecular modelling. However the crystal structure of novobiocin bound to a 24 kD fragment of gyrase B the target of the coumarin antibiotics reveals that it is the lactone carbonyl rather than the 4-oxy group which is responsible for a key interaction with an Arg residue and consequently the carboxy group in 12a is misplaced as a corresponding H-bond acceptor. Subsequently compounds incorporating an extra aromatic (as in 13a) or heteroaromatic ring (as in 45) bearing an H-bond acceptor along with the extra dimethyl groups on the cyclohexane ring have been designed.Models of these structures overlay convincingly with novobiocin bound to 24 kD gyrase B. Versatile synthetic routes have been developed allowing these compounds and the underlying hypotheses to be tested. Unfortunately none of the compounds demonstrates significant enzyme or antibacterial activity which we attribute to a combination of features including the lack of a replacement for the noviose methoxy group and the failure to achieve a good stacking (charge transfer) interaction with Arg-76. Introduction DNA Gyrase is a bacterial enzyme responsible for the process of supercoiling DNA.1 It is an enzyme that is vital to bacteria has no direct counterpart in mammals and consequently it represents an attractive antibacterial target. Structurally the enzyme has two discrete subunits (A and B) and it functions as an A2B2 tetramer in which the A subunits are responsible for the binding breaking and recombination of the DNA and the B subunits ‘power’ the process by the hydrolysis of ATP.Two discrete classes of clinically effective antibiotics are known to act at DNA gyrase.2 These are the quinolones and the coumarins typified by ciprofloxacin 1 and the natural product novobiocin 2 respectively. A third class of compounds currently under investigation include the natural products cyclothialidine3 3 a variety of synthetic analogues 4 and the closely related GR122222X5 4. Several other gyrase inhibitors have been reported.6 The quinolones bind to the A subunits forming a stable ternary complex between DNA enzyme and drug which is responsible for the subsequent cell death.1 These compounds have found very widespread use as antibiotics against a whole spectrum of bacteria.The coumarins such as novobiocin 2 bind to the B subunit of DNA gyrase causing inhibition of the ATPase activity.1 Novobiocin 2 itself has found limited clinical use mostly against Gram positive organisms but suffers from problems of toxicity and rapidly developing resistance.7 More potent coumarins such as coumermycin A1 8 and clorobiocin 59 have physical properties such as poor solubility which have limited their development. Cyclothialidine 3 and GR122222X 4 also inhibit DNA gyrase by competitive binding to the B subunit.3,5 The ever increasing incidence of infection resulting from bacteria resistant to currently available therapy has prompted the continuing need to develop new and improved antibacterial agents.The inhibition of DNA gyrase in this case specifically at the B subunit was seen as an attractive approach. In the past there has been considerable work both within Zeneca10 and elsewhere 11 to develop an improved antibacterial agent by modification of the coumarin natural products. The chemistry was highly complex and in trying to establish structure–activity relationships there was always a concern that weak levels of activity may simply be due to contamination with small amounts of the very potent parent compound. However the structural information available for novobiocin 2 and more recently for a complex between novobiocin and a 24 kD protein fragment derived from the B subunit of Escherichia coli DNA gyrase opened up the possibility of designing novel wholly synthetic compounds which mimic novobiocin.This paper outlines the design and synthesis of a novel series of spiroketals intended to mimic those structural features of novobiocin which are believed to be responsible for binding and inactivation of gyrase B. Results and discussion Design At the outset of this project the crystal structure of novobiocin 2 was available12 (see Fig. 1). A 2D 1H NMR experiment confirmed that the key conformational features seen in the solid state were also evident in solution.13 There was also some understanding of the structure–activity relationship within the coumarins which was relevant to the design phase. First and foremost there are features of both the aromatic portion and the sugar which seem essential for activity.Compounds having trivial replacements for either portion are inactive.10 For the aromatic part of the molecule it appears that replacing the terminal phenolic amide with simple amides such as isobutyroyl (as in 6) results in lower but still significant activity.10,11 Similarly the presence of the aromatic methyl substituent (or chloro substituent in the case of clorobiocin 5) is believed to enhance activity but is not essential.11 It has been shown for several bacteria that resistance to novobiocin 2 results from a variety 2790 J. Chem. Soc. Perkin Trans. 1 1997 of point mutations of a ‘key’ arginine residue (equivalent to Arg 136 in E. coli).14 Therefore the interaction with Arg 136 is important for novobiocin 2 and we assumed that the 4-oxy group on the coumarin (as its anion) 15 was the most likely group responsible for this interaction.With regard to the sugar apart from variations to the carbamate group the limited changes that have been investigated were detrimental.10 We set out to design compounds containing a cyclohexyl mimic of the noviose sugar that would carry similar functionality to the natural product and place an aromatic ring in the same vicinity as the hydroxycoumarin. The aromatic ring would bear a hydrogen bond acceptor to pick up the ‘key’ interaction with Arg 136. There is an inherent weakness in targeting a residue which is known to be capable of mutation. However the extent to which novobiocin is compromised in the clinic as a result of gyrase mutation is dependent on several factors which include potency and pharmacokinetics as well as the frequency and viability of the point mutation (equivalent to Arg 136 in E.N N CO2H O HN O O O OH O H2N O O O HN O –Na + O OH O O O OH O O Cl O HN OH O OH O NH O O O OH O H2N O O O HN OH O O HN S OH HO N H O O NH CO2H O O N O NH2 R HO F 1 2 3 R = OH 4 R = H 5 6 coli). We were optimistic that if we could identify activity in a series that had the potential for synthetic manipulation we might then be able to optimise potency and pharmacokinetics whilst enhancing the interactions to other parts of the enzyme thus reducing the effect of mutation of the arginine residue. We hoped that by rigidifying the structure in an appropriate conformation we would enhance binding by reducing the entropic penalty. One problem with using a cyclohexane ring to replace the pyran of the sugar is that in the absence of the anomeric effect and without other influences the O-aryl substituent will occupy an equatorial rather than axial position.Consequently we considered the possibility of having a second oxygen linked substituent in the form of a ketal to balance out this tendency. The use of a cyclic ketal to link a phenyl ring via two oxygen substituents to the cyclohexane ring would have the double advantage of increasing chemical stability and also freezing out the conformation. We first studied a 6,5-spiroketal in which one C]O aryl bond must inevitably be axial and the conformation would be rigidly constrained. Evaluation of a model of this structural type 7 against novobiocin 2 showed the expected overlap of the cyclohexane and pyran rings but also revealed that the phenyl ring was not at all close in space or orientation to the phenyl ring of the coumarin (Fig.2). We next investigated the possibility that a 6,6-spiroketal would have the correct orientation. For this system we considered four ‘low energy’ conformations (8–11) which represent the two ring flipped chair conformations in combination with the aromatic ring either bent towards or away from the oxygenated side of the cyclohexane ring. The relative energies of these conformations were obtained from an MM2 derived molecular mechanics potential 16 and confirmed by low level semiempirical calculations. From these calculations we immediately discount the likelihood of the conformation given in 11. This is because even given the inherent limitations of the potential function and an absence of a solvation model we compute a high relative conformational energy for 11 (also confirmed by the semiempirical calculations).The relative conformational energies of the other structures (8–10) differ only by 2.1 kJ mol21. In this case the limitations of the computational model do not lead to precise conclusions about the conformational populations. However we were satisfied that 8 is an acceptable conformation for the 6,6-spiroketal. Fig. 1 Crystal structure of novobiocin Fig. 2 Overlay of novobiocin with 7 J. Chem. Soc. Perkin Trans. 1 1997 2791 Comparison of 8 with novobiocin 2 showed a good overlap of both the cyclohexane and pyran rings and also the aromatic and coumarin rings (Fig. 3). The only difference observed was a slight tilt of the aromatic rings away from co-planarity.Based on these modelling results our initial target became the spiroketal 12a bearing oxyen functionality on the cyclohexane ring and a carboxy group on the phenyl ring to mimic the 4-oxy group on the coumarin.15 It was then our intention to extend our studies to amides and other hydrogen bond acceptors and also to include the gem-dimethyl group (as in 12b). During the course of this research further highly relevant structural information became available namely the X-ray crystal structure of a complex between novobiocin 2 and a 24 kD protein fragment derived from E. coli gyrase B.17 The conformation for clorobiocin 5 in the protein–ligand complex was also solved in solution by NMR and was in agreement with that observed in the solid state.18 Not only did the crystal structure provide information on the conformation of novobiocin 2 in a ‘biologically relevant’ environment but it also clarified the O O OH O OH O O O OH O O O O O O O H O O O O H 7 8 (+1.7 kJ mol-1) 9 (0 kJ mol-1) 10 (+2.1 kJ mol-1) 11 (+20 kJ mol-1) Fig.3 Overlay of novobiocin with 8 O O CO2H R R OH O H2N O O O Me Me OH O H2N O 12a R = H 12b R = Me CO2H 13a nature of the interactions of novobiocin 2 with those parts of the active site contained within the protein fragment. On the whole comparison of the two structures shows the conformations to be quite similar (Fig. 1 and Fig. 4) but the terminal phenolic group of the bound novobiocin appears to be bent back on itself and there is a slightly increased twist of the coumarin ring relative to the sugar.However the protein crystal structure did reveal that it is the lactone carbonyl (as its oxyanion) rather than the 4-oxy group that binds to Arg-136. This information implied that the carboxy group of 12a would not be in the correct position to pick up the key interactions with Arg-136. As indicated earlier the coumarin rings of novobiocin and the phenyl ring of 12a were not strictly coplanar in our modelling. Hence we did not believe that simply annealing a second aromatic ring would enable the correct positioning of a carboxy group. We considered whether the natural twist imparted by a biphenyl derivative would allow a second aromatic ring to be orientated in a similar manner to the coumarin ring and consequently position a hydrogen bond acceptor correctly. Modelling the biphenyl acid 13a in comparison to novobiocin bound to 24 kD gyrase showed that the natural twist imparted by the biphenyl put the terminal aromatic ring back into co-planarity with the coumarin and placed the carboxy group in a very favourable position to pick up the interaction with Arg-136 (as in Fig.5). As a result of this modelling and the lack of activity seen in the early compounds our synthetic targets became the acid 13a and related compounds. We were also interested in preparing other biaryl systems bearing hydrogen bond acceptor groups. Synthesis Our retrosynthetic analysis of the original target 12a/b is illustrated in Scheme 1. It was recognised that the final products prepared by this route would be racemic but during the phase Fig. 4 Conformation of novobiocin in novobiocin–gyrase B 24 kD complex Fig.5 Overlay of novobiocin from novobiocin–gyrase B 24 kD complex with 13a 2792 J. Chem. Soc. Perkin Trans. 1 1997 of lead identification this was deemed to be acceptable. We envisaged that 12a/b could be prepared from the unsaturated compound 14a/b by conversion of the double bond to the cisdiol and subsequent manipulation to the carbamate along with conversion of the bromo group to a carboxylic acid. We expected that 14a/b could be prepared directly from the corresponding unsaturated ketone 15a/b and 5-bromo-2-hydroxybenzyl alcohol 16. In the event we were unable to condense 15a and 16 directly under a variety of conditions and consequently an alternative sequence was developed (Scheme 2). Cyclohexanone was converted directly to the bromodimethyl ketal 17a using bromine in methanol according to the method of Garbisch.19 Transketalisation with 16 gave the bromo spiroketal 18a which was treated with sodium methoxide in Scheme 1 O O Br R R R R O HO HO Br 12a/b 14a R = H 14b R = Me 15a R = H 15b R = Me 16 + Scheme 2 Reagents a MeOH Br2; b 16 PTSA toluene; c NaOMe DMSO; d TMSI CH2Cl2; e PhMe3N1Br2 O O Br R R R R O O O R R Br O O Br R R Br O O Br R R 15a R = H 15b R = Me O O Br 18a R = H (68%) 18b R = Me 17a R = H 17b R = Me 19 (73%) TMS TMS 14a R = H (91%) 14b R = Me (84% from 19) a b c d e DMSO to give 14a in 62% yield over three steps.In order to establish the relative stereochemistry and to provide structural information the bromo derivatives themselves were progressed to final products (Scheme 3). Dihydroxylation of 14a gave a 1 1 mixture of the diastereomeric diols 20a and 21a in 54% yield which could be separated with difficulty by HPLC.Alternatively the mixture was converted to the carbonates 22a and 23a with triphosgene in 81% yield,20 which could be easily separated by flash chromatography. (Hydrolysis of the single diastereomers 22a and 23a gave the corresponding diols 20a and 21a.) Assigning the relative stereochemistry between these diastereomeric series by 1H NMR spectroscopy proved to be difficult and not entirely convincing. For the des-methyl series where the relevant peaks could be distinguished in both diastereomers (e.g. 22a and 23a) the assignment of relative stereochemistry was tentatively based on the presence of an NOE from a benzylic proton to the proton at position 39 on the cyclohexane ring (for 22a) and the absence of an equivalent result with the diastereomer (23a) (see Fig.6). Definitive proof of the relative stereochemistry came from a single crystal X-ray structure of diol 21a.21 In order to complete the synthesis of the model compound 24 the required carbonate 22a was reacted with ammonia to form a 1 1 mixture of the 2- and 3-carbamates 24 and 25. In this case the two isomers were separated by HPLC with dif- ficulty and we were able to get an X-ray crystal structure of 24.21 Undoubtedly it would be possible to develop a sequence of reactions that would allow selective functionalisation of the 3-hydroxy group. However as our purpose was to identify activity in a lead series we were happy to accept the benefits that the described sequence presented namely straightforward chemistry and the ability to form carbamates from a variety of amines at a late stage.Except where circumstances were particularly favourable we did not separate the 2- and 3-carbamate mixtures but tested the mixture directly for biological activity. For the synthesis of the first genuine target 12a the bromo derivative 14a was transmetallated and quenched with CO2 to form the corresponding acid 26a in 50% yield (Scheme 3). We were concerned about the base sensitivity of the final carbamate and also the acid sensitivity of the spiroketal and therefore chose the trichloroethyl ester as a protecting group for the acid that could be removed with zinc under buffered conditions.22 In retrospect our concerns about the acid sensitivity of the spiroketal were unnecessary.The acid 26a was converted to 27a in 96% yield. Subsequent dihydroxylation gave diol 28 and the corresponding diastereomer 29 in 59% yield (83% based on recovered starting material). This 1 1 mixture was taken through to the carbonate derivatives 30 and 31 in 74% yield and separated by flash chromatography. The assignment of relative stereochemistry was based on 1H NMR NOE experiments along with a comparison of the spectra and NOE results seen with the bromo series. Opening up of the carbonate with Fig. 6 Informative 1H NMR NOEs O O O O O H H H Br O O O O O H Me Me O O O O O H H Br Me Me J. Chem. Soc. Perkin Trans. 1 1997 2793 ammonia in non-alcoholic solvents gave poor results and the use of ethanol resulted in transesterification. Consequently 30 was reacted with ammonia in a small quantity of trichloro- Scheme 3 Reagents and conditions a OsO4 O(CH2CH2)2(Me)N1O2 or Me3N1O2 (54% 1 1 20a 21a; 49% 1 1 20b 21b 66% based on recovered 14b; 59% 1 1 28 29 83% based on recovered 27a); b triphosgene (81% 1 1 22a 23a from 20a 1 21a; 76% 1 1 22b 23b from 20b 1 21b; 74% 1 1 30 31 from 28 1 29); c NH3 EtOH; d BunLi CO2; e DCCI DMAP HOCH2CCl3; f NH3 HOCH2CCl3 CAUTION!; g Zn THF pH 7 phosphate buffer O O Br R R O O X R R O O X R R O O O O X R R HO OH O O R R HO OH X O O O R R O O X O O O X O OH H2N O O O HO O 14a R = H 14b R = Me X 26a R = H X = CO2H (50%) 26b R = Me X = CO2H 27a R = H X = CO2CH2CCl3 (96%) 27b R = Me X = CO2CH2CCl3 22a R = H X = Br 22b R = Me X = Br 30 R = H X = CO2CH2CCl3 O NH2 20a R = H X = Br 20b R = Me X = Br 28 R = H X = CO2CH2CCl3 21a R = H X = Br 21b R = Me X = B r 29 R = H X = CO2CH2CCl3 23a R = H X = Br 23b R = Me X = Br 31 R = H X = CO2CH2CCl3 24 X = Br (16%) 32 X = CH2CO2CCl3 (35%) 12a X = CO2H (58%) 25 X = Br (21%) 33 X = CH2CO2CCl3 (23%) + + + a a b c or f d e g ethanol (CAUTION! potential hazard) to give 32 and 33 in 58% yield which were separated by chromatography.Deprotection with zinc in THF–phosphate buffer (pH 7) gave the target acid 12a in 58% yield. In tackling the synthesis of the dimethyl series (e.g. 13a) we attempted to prepare the key unsaturated spiroketal 14b using analogous chemistry to the des-methyl series (Scheme 2). However this proved completely unsuccessful. Formation of the bromo dimethyl ketal 17b went very poorly and under no circumstances could 17b be converted to the bromo spiroketal 18b even when 17b was freshly distilled.Consequently an alternative sequence was developed (also shown in Scheme 2). Ketalisation of 15b with the bis (TMS) ether derived from 16 using TMSI catalysis gave 19 in good yield (73%).23 Direct bromination could be achieved with phenyltrimethylammonium tribromide24 and the product 18b was converted without purifi- cation to 14b by treatment with sodium methoxide in DMSO (84% yield). The sequence has been performed on a 1 mol scale in 40% yield over the three steps. (The lower than expected isolated yields on the larger scale are a consequence of poor recovery from the final crystallisation. No attempt was made to chromatograph the residues because 14b proved to be fairly unstable to silica.) The unsaturated ketal 14b was progressed through a sequence of transmetallation and quenching with CO2 to give the acid 26b which was protected as its trichloroethyl ester 27b.However this compound proved to be even less stable to silica than the bromo analogue 14b and this instability made studies of the dihydroxylation difficult. In any event these studies were abandoned because of the change of focus towards targets such as 13a which followed on from the protein–novobiocin structural information. However the cis-hydroxylation of 14b was still seen as a critical step and it was with some relief that it was discovered that 14b could be converted to the diols 20b and 21b in 49% yield (66% based on recovered starting material) using slightly modified conditions. More importantly 20b and 21b and all subsequent derivatives of these compounds suffered from no problems of stability.As previously the diols were converted to the carbonates 22b and 23b and then easily separated by flash chromatography (76% yield). In this case the relative stereochemistry could be assigned unambiguously by 1H NMR NOE experiments (see Fig. 6). It was envisaged that the biphenyl linkage in the target spiroketal biphenyl acid 13a could be prepared by a palladium catalysed cross-coupling reaction (Scheme 4). The bromo derivative 22b could be coupled with the o-formyl boronic acid 34a using Pd0 and CsF catalysis25 to give 35 in 84% yield. Alternatively 22b could be coupled with the boronic ester 34b using Pd0 and Cs2CO3 catalysis 26 in 87% yield. The aldehyde 35 was oxidised to the acid 36 using sodium chlorite 27 in 98% yield and the carbonate group of 36 was opened with ammonia to give the target compound 13a along with the isomer 37a as a 1 1 mixture in 73% yield.In order to exploit potential lipophilic interactions the higher carbamates 13b–d and their isomers 37b–d were prepared by reaction with the corresponding amines. In these examples the ratio of isomers varied from 1 1 to 4 1 and in some cases the desired product could be isolated pure by chromatography following the preferential crystallisation of the minor isomer. The route illustrated in Scheme 4 for the synthesis of biaryl derivatives was somewhat limiting when it came to varying the terminal aryl group because of the need to prepare boronic acids (or an equivalent organometallic derivative) for each of the aryl or heteroaryl groups in question.Consequently we evaluated an alternative approach in which the organometallic functionality is contained within the spiroketal component and reacted with a variety of aryl or heteroaryl bromides or iodides (Scheme 5). The choice of appropriate protecting groups for the diol (22b/23b) was not an obvious one. Acyl groups would not be stable to the conditions of transmetallation and in a brief investigation we were only able to silylate the equatorial 3- 2794 J. Chem. Soc. Perkin Trans. 1 1997 hydroxy group. We considered the dimethyl ketal because of its stability toward organometallics and suspected that it might be more labile to acid hydrolysis than the spiroketal and this turned out to be the case. The diols 20b and 21b were converted to the dimethyl ketals 38 and 39 in 85% yield and separated easily by flash chromatography.The bromo group of 38 was converted by transmetallation capture with triisopropyl borate and transesterification to the cyclic boronate ester 40 which could also be purified by chromatography (80% yield). The boronate ester 40 could be coupled with a variety of bromo- or iodo-aromatics using Pd0 and Cs2CO3 catalysis 26 in high yield. For example coupling of 40 with 2-bromopyridine gave the biaryl derivative 41 which could be cleaved very selectively under acidic conditions to give the diol 42. This compound was converted to the carbonate 43 in an overall yield of 44% from 40. Oxidation to the corresponding pyridine N-oxide 44 was accomplished using MCPBA in 91% yield and subsequent reaction of 44 with Scheme 4 Reagents a 34a Pd(PPh3)4 CsF or 34b Pd(PPh3)4 Cs2CO3; b NaOCl2; c RNH2 EtOH O O Br Me Me O O O O O Me Me O O O X CHO M O O Me Me O OH CO2H O O Me Me HO O CO2H RNH O O RNH 34a M = B(OH)2 34b M = B[O(CH2)3O] 35 X = CHO (87%) 36 X = CO2H (98%) 13a R = H (73% 1:1 13a:37a) 13b R = Me (53%) 13c R = Et (57%) 13d R = Pri (68%) 22b 37a R = H 37b R = Me (15%) 37c R = Et (25%) 37d R = Pri (19%) a b c ammonia or methylamine gave the target carbamates 45 and 47 along with the corresponding isomers 46 and 48.The boronate ester 40 was also used to prepare a compound analogous to 13a but containing the 5-methylpyrrole-2- carboxylate instead of the carbamate (Scheme 5). Coupling of Scheme 5 Reagents a CH3C(OCH3)2CH3 H1; b i BunLi B(OPri)3 ii HO(CH2)3OH toluene; c 2-bromopyridine Pd(PPh3)4 Cs2CO3; d 1 M HCl MeOH; e triphosgene; f MCPBA; g RNH2 EtOH; h trichloroethyl 2-iodobenzoate Pd(PPh3)4 Cs2CO3; i 5-methyl-2-trichloroacetylpyrrole K2OBut; j Zn THF pH 7 phosphate buffer O O X Me Me O O O O Me Me O O Br O O Br Me Me HO OH O O Me Me HO OH Br O O Me Me RO OR' N O O Me Me RO OR' CO2CH2CCl3 O O Me Me RO OR' N+ –O O O Me Me O OH CO2R" 22b 23b 38 X = Br (43%) 40 X = B[O(CH2)3O] (80%) 39 (42%) 44 R = R' = CO (91%) 45 R = CONH2 R' = H (50% 1:1 45:46) 46 R = H R' = CONH2 47 R = CONHMe R' = H (82% ca.2:1 47:48) 48 R = H R' = CONHMe 49 R = R' = C(CH3)2 (53%) 50 R = R' = H O NH 41 R = R' = C(CH3)2 42 R = R' = H 43 R = R' = CO (ca. 44% from 40) 51 R"= CH2CCl3 (21% from 50) 52 R"= H (44%) a + + b h c d e d j g i J. Chem. Soc. Perkin Trans. 1 1997 2795 40 with the trichloroethyl ester of 2-iodobenzoic acid using Pd0 and Cs2CO3 as catalysts 26 gave the biphenyl derivative 49 in 53% yield.Again deprotection of the dimethyl ketal was achieved selectively and the resulting diol 50 could be selectively acylated on the 3-hydroxy group to give 51 albeit in poor yield using the trichloroacetylpyrrole and potassium tert-butoxide (21% 27% based on recovered 50). The trichloroethyl ester was then removed as previously using zinc in THF–buffer (pH 7) to give 52 in 44% yield. Evaluation The crystal structure 21 of the model compound 24 shows a conformation essentially identical to that illustrated for structure 9 and is consistent with the modelling results which predicted 9 to be the lowest energy of the four conformations modelled (8–11). Conformation 8 provides the best overlap with novobiocin (Fig.3) but as discussed previously we believe that the calculated energy difference between 8 and 9 if real is small enough to imply that 8 should still be an easily accessible conformation for our target compounds. We therefore believe that the 6,6-spiroketals do provide a scaffold on which to append both functionality equivalent to the noviose sugar and also an appropriately positioned aromatic ring albeit with a slight twist relative to the coumarin. All our synthetic products were tested for gyrase inhibitory activity in a standard supercoiling assay28 at 800 mg ml21 and 200 mg ml21 (or 1600 mg ml21 and 400 mg ml21 where a 1 1 mixture of carbamates was evaluated). In this assay the coumarin standards would have the following activity novobiocin (IC50 = 0.1 mg ml21) clorobiocin (IC50 = 0.01 mg ml21) and the semisynthetic isobutyryl derivative 6 (IC50 = 1 mg ml21).Very little activity was seen for any of the target compounds. The biphenyl acid 13b showed very weak activity in the supercoiling assay (IC50 = 200 mg ml21) but the lack of activity in closely related compounds (13a,c,d) and in particular for the acyl pyrrole derivative 52 was not encouraging. The compounds were also tested in standard antibacterial screens 29 at doses up to 128 mg ml21. The weak antibacterial activity which was seen for some of the intermediates is presumed not to be related to gyrase inhibition. Discussion Having seen the novobiocin–24 kD gyrase B crystal structure it is not surprising that the early compounds such as 12a showed no activity as the carboxy group is not correctly positioned to pick up any interaction with Arg-136.These compounds also lack the gem-dimethyl group which is likely to be important for binding. There are some explanations for the lack of activity seen in the biaryl series which we do not believe are responsible. The most fundamental of these is that our modelling was done with a structure derived from a small fragment of the gyrase B subunit which may not be relevant to the whole enzyme. We do not believe this explanation to be the case. Whilst the 24 kD protein may not contain all the binding sites for novobiocin microcalorimetry clearly demonstrates potent binding of novobiocin to 24 kD protein.30 Furthermore for selected examples of our target compounds (including 13b) we have demonstrated by microcalorimetry that there is no binding to 24 kD protein up to the experimental limits of solubility (250 mg ml21).We have also repeated our calculations on structures analogous to 8–11 but bearing the gem-dimethyl groups and the results are very similar to those for 8–11. The possibility that the oxymethylene bridge in our spiroketals is clashing with some part of the protein which is not observed in the novobiocin–24 kD structure is not supported by inspection of the 43 kD gyrB dimer structure in which the relevant section of protein is present.31 Furthermore if this were the fundamental cause of inactivity against the full enzyme you would still expect to see binding to 24 kD protein. There are several factors which we believe may contribute to the lack of activity. These include the absence of the methoxy group found on the noviose sugar which may be important.Also there appears to be a significant p-stacking (charge transfer) interaction between the hydroxycoumarin of novobiocin 2 and Arg-76 which could contribute up to 13 kJ mol21 of binding energy.32 This interaction may be significantly reduced for the target molecules as a consequence of the lack of co-planarity seen in the overlay of the aromatic ring of modelled target compounds and novobiocin. It is also possible that the terminal aromatic ring is not adequately mimicking the electronic features of the delocalised anion derived from the hydroxycoumarin. The penalty for de-solvating this delocalised anion may also be lower than that for a carboxylic acid. Compounds prepared in further studies which incorporated a double hydrogen bond acceptor in the form of a 3-hydroxypyridine N-Oxide as the terminal aromatic ring (and presumed to be partially ionised) also failed to show activity.In all likelihood it is a combination of features which are really quite subtle that result in the lack of activity and it remains a disappointment that structures that we believe have the correct conformation carry much of the functionality believed to be important for binding and model so well into the protein showed no activity. Conclusions It is an important premise of rational drug design that different ligands that bind to the same receptor site giving rise to a similar pharmacological response possess a degree of molecular shape similarity. This idea has been used in the design of a series of 6,6-spiroketal derivatives based on the structure of a known inhibitor ligand (novobiocin) and its conformation bound to a 24 kD fragment of DNA gyrase.We have developed synthetic routes to these targets and the compounds appear to exhibit much of the spatial (and structural) features of the inhibitor (novobiocin). However the compounds do not show any significant activity against DNA gyrase at concentrations of 104 higher than the IC50 of novobiocin. We believe that the lack of activity arises from a combination of features including the lack of a replacement for the noviose methoxy group and the failure to achieve a good stacking (charge transfer) interaction with Arg-76. Experimental All operations were carried out at room temperature unless otherwise stated. Tetrahydrofuran (THF) was dried by distillation from calcium hydride.Flash chromatography was performed on silica (Merck Kieselgel Art. 9385). Melting points were taken on a Buchi apparatus with the use of glass capillary tubes and are uncorrected. 1H NMR spectra were recorded on Bruker WM200 WM250 or WM400 instruments and are reported as d values (parts per million) relative to Me4Si as internal standard. J Values are given in Hz. Mass spectra were recorded on VG 12-12 quadrupole VG 70-250 SE VG ZAB 2- SE or a VG modified AE1/Kratos MS9 spectrometers. 6-Bromospiro[4H-1,3-benzodioxine-2,19-cyclohex-29-ene] 14a 5-Bromo-2-hydroxybenzyl alcohol 16a (13.65 g 67.3 mmol) 17a19 (10 g 44.8 mmol) and toluene-p-sulfonic acid (PTSA) (0.1 g) were heated at reflux in toluene passing the condensed solvent through a Soxhlet adapter filled with 4 Å molecular sieves.After 8 h the solution was cooled and concentrated under reduced vacuum to a light brown oil which was chromatographed on silica using 10% ethyl acetate in hexane as eluent to give 18a as a clear colourless oil which was used directly (10.95 g 68%) dH(250 MHz; CDCl3) 1.5–2.4 (8H m 4 × CH2) 4.3 (1H s CHBr) 4.74 (1H d J 14 CHAHB) 4.94 (1H d J 14 CHAHB) 6.74 (1H d J 8 ArH) 7.13 (1H d J 2 ArH) 7.25 2796 J. Chem. Soc. Perkin Trans. 1 1997 (1H dd J 3 8 ArH); m/z (EI1) 360 (M1 40%) 184 (C7H5OBr 100). 18a (10.95 g 30.2 mmol) was stirred in DMSO (30 ml) and NaOMe (4.77 g 90 mmol) was added keeping the temperature below 10 8C over the addition and for the following 10 min and then allowing the reaction mixture to warm to room temperature.After 3 h the mixture was poured into brine (150 ml) and extracted with 2-methylpentane (200 ml). The organic extract was dried (MgSO4) filtered and evaporated to give 14a as a clear oil (7.8 g 91%) dH(250 MHz; CDCl3) 1.83 (2H m CH2) 2.0 (2H m CH2) 2.13 (2H m CH2) 4.8 (1H d J 14 CHAHB) 4.9 (1H d J 14 CHAHB) 5.82 (1H d J 10 CHCHCH2) 6.05 (1H dt J 3 10 CHCHCH2) 6.73 (1H d J 8 ArH) 7.12 (1 H d J 2 ArH) 7.26 (1 H dd J 2 8 ArH); m/z (EI1) 280 (M1 75%) 184 (C7H5OBr 65) 96 (C6H8O 100). 6-Bromo-5959-dimethylspiro[4H-1,3-benzodioxine-2,19-cyclohex- 29-ene] 14b 5-Bromo-2-hydroxybenzyl alcohol 16a (4.54 g 22 mmol) was dissolved in THF (50 ml) and N-methyl-N-(trimethylsilyl)- trifluoroacetamide (10.2 ml 55 mmol) was added. The solution was stirred under argon at 55 8C for 30 min and then concentrated under vacuum.The residual material was diluted with 2- methylpentane and filtered through a sinter and the filtrate concentrated to give the bis-TMS ether (7.96 g 100%). A portion of this material (2.6 g 7.45 mmol) and 3,3- dimethylcyclohexanone (940 mg 7.46 mmol) were dissolved in CH2Cl2 (30 ml) and cooled to 278 8C under argon. Trimethylsilyl iodide (ca. 50 mg) was added and the reaction mixture was stirred at 278 8C for 50 min and warmed to 0 8C and stirred for a further 30 min. Saturated aqueous NaHCO3 (30 ml) was added followed by Et2O (60 ml). The organic layer was washed with NaHCO3 (sat. aq. 20 ml) Na2S2O5 (sat. aq. 20 ml) and brine (20 ml) dried over MgSO4–Na2CO3 (1 1) filtered and concentrated to give 19 as a colourless oil (1.69 g 73%) dH(250 MHz; CDCl3) 0.97 (3H s CH3) 0.99 (3H s CH3) 1.34 (2H m CH2) 1.54–1.74 (6H m 3 × CH2) 4.73 (1H d J 15 OCHAHB) 4.83 (1H d J 15 OCHAHB) 6.69 (1H d J 9 ArH) 7.08 (1H d J 2 ArH) 7.23 (1 H dd J 2 9 ArH); m/z (EI1) 310 (100% M1).To a stirred solution of 19 (7.8 g 25.16 mmol) in THF (166 ml) at 0 8C under argon was added rapidly phenyltrimethylammonium tribromide (97%; 9.75 g 25.16 mmol). The resulting solution was stirred for 2 h at 0 8C and then diluted with NaHCO3 (sat. aq. 12 ml) and stirred for a further 30 min at 0 8C. The mixture was concentrated under reduced pressure without heating to remove most of the THF and the residue was extracted with Et2O (2 × 100 ml). The combined organic layers were washed with brine (50 ml) dried over MgSO4 filtered and concentrated. The resulting oil was diluted with hexane (60 ml) filtered and the filtrate concentrated under vacuum to give 18b as a pink gum (9.225 g 94%) which was dissolved in DMSO (50 ml) and stirred at 0 8C under argon.NaOMe (3.85 g 71.3 mmol) was added and the reaction mixture was stirred for 3 h at room temperature and then poured into brine (250 ml) and extracted with hexane (2 × 150 ml). The combined organic layers were dried over MgSO4 filtered and concentrated under vacuum to give 14b as a yellow solid (6.56 g 89%). For analytical purposes a small sample (1 g) was recrystallised from 2-methylpentane (7 ml) to give 14b as yellow crystals (680 mg). (NB 14b is not completely stable to chromatography on silica) mp 111–114 8C (Found C 58.7; H 5.6. C15H17BrO2 requires C 58.3; H 5.5%); dH(200 MHz; CDCl3) 1.02 (3H s CH3) 1.04 (3H s CH3) 1.87 [d J 5 CH2C(OR)2] 1.97 (2H dd J 1.5 4 CHCH2) 4.85 (2H s OCH2) 5.81 (1H ddd J 10 1 1 CH2CHCH) 5.98 (1H ddd J 10 3 3 CH2CHCH) 6.70 (1H d J 8 ArH) 7.10 (1H d J 2 ArH) 7.25 (1H dd J 8 2 ArH); m/z (CI1) 309 (MH1 60%) 125 (C8H13O1 100).(2R*,29R*,39R*)- and (2S*,29R*,39R*)-6-Bromo-29,39- dihydroxyspiro[4H-1,3-benzodioxine-2,19-cyclohexane] 20a and 21a 14a (2.0 g 7.1 mmol) and N-methylmorpholine N-oxide (1.43 g 10.6 mmol) were dissolved in ButOH (40 ml) and OsO4 (4% aq. 0.56 ml 0.07 mmol) was added followed by water (1 ml). After stirring the solution for 3 days the reaction was quenched by the addition of solid sodium metabisulfite (ca. 300 mg) and stirring for 30 min. The mixture was evaporated to dryness under vacuum and chromatographed using ethyl acetate–hexane (1 1) as eluent to give a mixture of 20a and 21a as a colourless oil (1.2 g 54%).For synthetic purposes this mixture was used directly. For analytical purposes a sample of the mixture (246 mg) was purified by HPLC (Dynamax C18 column) using 60% methanol–40% water to give 20a (96 mg) and 21a (71 mg) as white solids. 20a mp 114–116 8C (Found C 49.4; H 4.6. C13H15BrO4 requires C 49.5; H 4.8%); dH(400 MHz; CDCl3) 1.4–1.9 (6H m 3 × CH2) 2.32 (1H d J 8 CH2CHOH) 2.54 (1H d J 3 CH2CHOHCHOH) 3.92 (1H m CH2CHOHCHOH) 3.97 (1H br m CH2CHOHCHOH) 4.87 (2H s OCH2) 6.74 (1H d J 8 ArH) 7.10 (1H d J 2 ArH) 7.27 (1H dd J 2 8 ArH); dC(100 MHz; CDCl3) 17.76 (CH2CH2CH2) 27.34 [CH2C(OR)2] 28.46 (CH2CHOH) 59.88 (OCH2) 69.59 (CH2CHOH) 71.83 (CH2CHOHCHOH) 101.05 [C(OR)2] 113.05 (CBr) 118.95 (CHCHCBr) 121.51 (CCH2O) 127.47 (CHCBr) 131.28 (CHCBr) 149.36 (COCR2O); m/z (CI1) 314 (M1 50%) 297 (M1 2 OH 50) 185 (C7H6OBr 100).21a mp 160–161 8C (Found C 49.8; H 4.8. C13H15BrO4 requires C 49.5; H 4.8%); dH(400 MHz; CDCl3) 1.35–1.9 (6H m 3 × CH2) 2.28 (1H br d J 8 CH2CHOH) 2.49 (1H s CH2CHOHCHOH) 3.9 (2H br s CH2CHOHCHOH) 4.76 (1H d J 14 OCHAHB) 4.83 (1H d J 14 OCHAHB) 6.79 (1H d J 8 ArH) 7.10 (1H d J 2 ArH) 7.29 (1H dd J 2 8 ArH); dC(100 MHz; CDCl3) 17.68 (CH2CH2CH2) 26.33 [CH2C(OR)2] 28.44 (CH2CHOH) 59.62 (OCH2) 69.40 (CH2CHOH) 72.77 (CH2CHOHCHOH) 101.38 [C(OR)2] 113.11 (CBr) 118.93 (CHCHCBr) 121.56 (CCH2O) 127.53 (CHCBr) 131.28 (CHCBr) 149.36 (COCR2O); m/z (CI1) 314 (M1 45%) 297 (M1 2 OH 50) 185 (C7H6OBr 100). (2R*,29R*,39R*)- and (2S*,29R*,39R*)-6-Bromo-29,39- dihydroxy-59,59-dimethylspiro[4H-1,3-benzodioxine-2,19-cyclohexane] 20b and 21b These compounds were prepared from 14b in an approximately 1 1 ratio [49% (66% based on recovered 14b)] in a similar manner as described for 20a and 21a with the exception that trimethylamine N-oxide was used as co-oxidant.For synthetic purposes this mixture was used directly. For analytical purposes careful chromatography on silica using EtOAc–hexane (2 3) gave samples of 20b and 21b as the predominant isomers which were crystallised from EtOAc–hexane to give pure samples. 20b mp 138–139 8C (Found C 52.4; H 5.7. C15H19O4Br requires C 52.5; H 5.6%); dH(400 MHz; CDCl3) 0.97 (3H s CH3ax) 0.99 (3H s CH3eq) 1.54 (1H dd J 5 13 CHeqHaxCHOH) 1.605 [1H d J 14 CHeqHaxC(OR)2] 1.635 (1H d J 13 CHeqHax- CHOH) 1.70 [1H d J 14 CHeqHaxC(OR)2] 2.08 (1H d J 9 CH2CHOH) 2.46 [1H s CHOHC(OR)2] 3.91 [1H s CHOH- (OR)2] 4.14 (1H m CH2CHOH) 4.82 (1H d J 15 OCHAHB) 4.88 (1H d J 15 OCHAHB) 6.70 (1H d J 8 ArH) 7.11 (1H d J 2 ArH) 7.27 (1H dd J 2 8 ArH); m/z (CI1) 360 (M 1 NH4 1 100%) 342 (M 1 H1 20).21b mp 148–149 8C (Found C 52.1; H 5.7. C15H19O4Br requires C 52.5; H 5.6%); dH(400 MHz; CDCl3) 0.98 (3H s CH3eq) 1.09 (3H s CH3ax) 1.53 (1H dd J 5 13 CHeqHaxCHOH) 1.64 [1H d J 14 CHeqHaxC(OR)2] 1.67 (1H d J 13 CHeqHaxCHOH) 1.76 [1H d J 14 CHeqHaxC(OR)2] 2.11 (1H d J 9 CH2CHOH) 2.46 [1H s CHOHC(OR)2] 3.92 [1H s CHOH(OR)2] 4.03 (1H m CH2CHOH) 4.74 (1H d J 17 OCHAHB) 4.88 (1H d J 17 OCHAHB) 6.75 (1H d J 8 ArH) 7.10 (1H d J 2 ArH) 7.27 J. Chem. Soc. Perkin Trans. 1 1997 2797 (1H dd J 2 8 ArH); m/z (CI1) 360 (M 1 NH4 1 100%) 342 (M 1 H1 20).(2R*,3a9R*,7a9R*)- and (2S*,3a9R*,7a9R*)-6-Bromospiro[4H- 1,3-benzodioxine-2,49-perhydro[1,3]benzodioxol]-29-one 22a and 23a The mixture of 20a and 21a (ca. 1 1; 1.2 g 3.8 mmol) was dissolved in CH2Cl2 (10 ml) and pyridine (1.8 g 22.8 mmol). The solution was cooled to 278 8C under an argon atmosphere and a solution of triphosgene (564 mg 1.9 mmol) in CH2Cl2 (10 ml) was added dropwise. The mixture was allowed to warm to room temperature and then quenched with ammonium chloride (sat. aq. 10 ml) and diluted with CH2Cl2. The organic layer was washed with 2 M HCl (10 ml) NaHCO3 (sat. aq. 20 ml) and brine (20 ml). The solution was dried (MgSO4) filtered and concentrated. The residue was chromatographed on silica using 30% EtOAc–hexane as eluent to give 22a (550 mg Rf 1:1 EtOAc–hexane 0.6) and 23a (510 mg Rf 1 1 EtOAc–hexane 0.4) as colourless oils which crystallised on standing (yield 81%).22a mp 151–153 8C (Found C 49.7; H 3.7. C14H13O5Br requires C 49.3; H 3.8%); dH(250 MHz; CDCl3) 1.65–2.15 (6H m 3 × CH2) 4.70 (1H d J 6 CHOCO) 4.83 (1H d J 16 OCHAHB) 4.88 (1H d J 6 CHOCO) 4.95 (1H d J 16 OCHAHB) 6.75 (1H d J 8 ArH) 7.15 (1H d J 2 ArH) 7.3 (1H dd J 2 8 ArH); m/z (Cl1) 358 (M 1 NH4 1 100%). 23a mp 94–99 8C (Found C 49.3; H 3.7. C14H13O5Br requires C 49.3 H 3.8%); dH(250 MHz; CDCl3) 1.6–2.2 (6H m 3 × CH2) 4.58 (1 H d J 6 CHOCO) 4.84 (2H s OCH2) 4.92 (1H m CH2CHOCO) 6.82 (1H d J 8 ArH) 7.11 (1H d J 2 ArH) 7.3 (1H dd J 2 8 ArH); m/z (CI1) 358 (M 1 NH4 1 100%). (2R*,3a9R*,7a9R*)- and (2S*,3a9R*,7a9R*)-6-Bromo-59,59- dimethylspiro[4H-1,3-benzodioxine-2,49-perhydro[1,3]- benzodioxol]-29-one 22b and 23b A mixture of diols 20b and 21b (3.83 g 11.1 mmol) was converted to a mixture of the carbonates 22b and 23b using the same procedure as described above for 22a and 23a.The crude product started to crystallise on standing and CH2Cl2 (20 ml) was added and 23b was filtered off as a white solid (1.165 g 29%). The filtrate was concentrated under vacuum and chromatographed on silica using 10–50% EtOAc in hexane as eluent to give 22b as an oil which crystallised on the addition of Et2O (5 ml) and hexane (10 ml) (1.45 g 37%) along with a further quantity of 23b (0.4 g 10% total 39%). 22b mp 124–125 8C (Found C 52.1; H 4.7. C16H17BrO5 requires C 52.1; H 4.6%); dH(400 MHz; CDCl3) 1.09 (3H s CH3ax) 1.12 (3H s CH3eq) 1.79–1.93 [4H m CH2C(CH3)2CH2] 4.69 (1H dd J 1.5 7.5 CH2CHOCHO) 4.84 (1H d J 16 OCHAHB) 4.96 (1H d J 16 OCHAHB) 4.98 (1H m CH2CHOCHO) 6.71 (1H d J 8 ArH) 7.15 (1H d J 2 ArH) 7.29 (1H dd J 2 8 ArH); irradiation of CH3ax (1.09) gives an NOE enhancement of ArH (6.71); m/z (CI1) 386 (M 1 NH4 1 100%) 368 (M 1 H1 10.23b mp 178–181 8C (Found C 51.8; H 4.6. C16H17BrO5 requires C 52.1; H 4.6%) dH(400 MHz; CDCl3) 1.07 (3H s CH3ax) 1.11 (3H s CH3eq) 1.705 (1H d J 15 CHeqHaxCHO) 1.89 [1H dd J 5.5 14 CHeqC(OR)2] 1.96 (1H d J 15 CHeqHaxCHO) 1.97 [1H d J 14 CHaxHeqC(OR)2] 4.64 (1H d J 7 CH2CHOCHO) 4.83 (1H d J 16 OCHAHB) 4.87 (1H d J 16 OCHAHB) 4.98 (1H m CH2CHOCHO) 6.79 (1H d J 8 ArH) 7.11 (1H d J 2 ArH) 7.30 (1H dd J 2 8 ArH); irradiation of CH3ax (1.07) gives an NOE enhancement of OCHAHB (4.85); m/z (CI1) 386 (M 1 NH4 1 100%) 368 (M 1 H1 10).(2R*,29R*,39R*)-6-Bromo-29-hydroxy-39-carbamoyloxyspiro- [4H-1,3-benzodioxine-2,19-cyclohexane] 24 and (2S*,29R* 39R*)-6-bromocarbamoyloxy-39-hydroxyspiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 25 22a (490 mg) was dissolved in sat. ethanolic ammonia (5 ml) and the solution stirred for 16 h at room temperature and then evaporated under reduced pressure to give 24 and 25 in an approximately 1 1 mixture. A sample of the mixture (400 mg) was separated by HPLC on a 5 mm silica preparative Dynamax column using 0.1% TFA–15% propan-2-ol–hexane as eluent to give pure samples of the components as gums along with the remainder of the material as a mixture. The gums crystallised on the addition of 30% diethyl ether–hexane (5 ml) to give white crystals of 24 (85 mg 16%) and 25 (110 mg 21%).24 was recrystallised from ethyl acetate (2 ml) and hexane (10 ml) which gave single crystals of a 1 1 EtOAc adduct suitable for X-ray diffraction. 24 mp 78–80 8C (Found C 48.3; H 5.5; N 3.2. C14H16NO5Br?C4H8O2 requires C 48.4; H 5.4; N 3.1%); dH(250 MHz; CDCl3) 1.25 (3H t J EtOAc) 1.4–1.9 (6H m 3 × CH2) 2.0 (3H s EtOAc) 4.08 (1H dd J 0.5 3 CHOH) 4.25 (2H q EtOAc) 4.73 (2H br s NH2) 4.83 (2H s OCH2) 5.0 (1H m CHOCONH2) 6.76 (1H d J 8 ArH) 7.08 (1H d J 2 ArH) 7.26 (1H dd J 2 8 ArH); m/z (CI1) 358 (MH1 60%) 174 (C7H12NO4 1 100). 25 mp 140–144 8C (Found C 46.6; H 4.4; N 3.6. C14H16NO5Br requires C 46.9; H 4.5; N 3.9%); dH(250 MHz; CDCl3) 1.4–1.9 (6H m 3 × CH2) 3.98 (1H m CHOH) 4.67 (1H d J 14 OCHAHB) 4.7 (2H br s NH2) 4.87 (1H d J 14 OCHAHB) 5.12 (1H d J 3.5 CHOCONH2) 6.67 (1H d J 8 ArH) 7.04 (1H d J 2 ArH) 7.21 (1H dd J 2 8 ArH); m/z (CI1) 358 (MH1 30%) 174 (C7H12NO4 1 100) 148 (C7H10NO3 1 100).Spiro[4H-1,3-benzodioxine-2,19-cyclohex-29-ene]-6-carboxylic acid 26a ButLi (1.5 M in pentane; 18 ml 27 mmol) was placed in a dry three-necked flask fitted with a thermometer and under an atmosphere of argon. The solution was cooled in a dry ice– acetone bath to 260 8C and diethyl ether (75 ml) was added. A solution of 14a (3 g 10.7 mmol) in diethyl ether (3 ml) was added keeping the temperature below 260 8C. After 5 min carbon dioxide was bubbled into the solution and the reaction mixture was allowed to warm to room temperature quenched with NH4Cl (sat. aq. 50 ml) and partitioned. The aqueous layer was further extracted with diethyl ether (3 × 40 ml) then adjusted to pH 5 with conc.HCl and extracted with diethyl ether (40 ml) and the combined organic layers were dried (MgSO4) filtered and concentrated. The resultant solid was slurried with 20% ethyl acetate–hexane and filtered to give 26a as white crystals (1.31 g 50%) mp 252–254 8C (decomp.) (Found C 68.3; H 5.8. C14H14O4 requires C 68.3; H 5.7%); dH(250 MHz; [2H6]DMSO) 1.75 (2H m CH2) 1.93 (2H m CH2) 2.10 (2H m CH2) 4.93 (2H s OCH2) 5.81 (1H d J 10 CH2CHCH) 6.08 (1H dt J 3 10 CH2CHCH) 6.89 (1H d J 8 ArH) 7.72 (1H d J 2 ArH) 7.76 (1H dd J 2 8 ArH); m/z (CI1) 247 (MH1 20%) 97 (C6H9O1 100). 1,1,1-Trichloroethyl spiro[4H-1,3-benzodioxine-2,19-cyclohex- 29-ene]-6-carboxylate 27a To a stirred mixture of 26a (1.2 g 4.87 mmol) 1,3-dicyclohexylcarbodiimide (DCC) (1.11 g 5.4 mmol) and DMAP (61 mg 0.5 mmol) in CH2Cl2 was added dropwise 1,1,1- trichloroethanol (800 mg 5.35 mmol) maintaining the temperature below 5 8C with an ice bath.The mixture was stirred for 18 h at room temperature filtered and the filtrate concentrated under vacuum to a gum which was chromatographed on silica using 20% EtOAc in hexane as eluent to give 27a as a colourless oil which solidified on standing (1.86 g 96%) mp 78–81 8C (Found C 50.9; H 4.2. C16H15C13O4 requires C 50.9; H 4.0%); dH(200 MHz; CDCl3) 1.6–2.2 (6H m CH2CH2CH2) 4.95 (4H s OCH2 1 CO2CH2) 5.63 (1H dt J 1 10 CHCHCH2) 6.08 (1H dt J 3 10 CHCHCH2) 6.93 (1H d J 9 ArH) 7.80 (1H d J 2 ArH) 7.96 (1H dd J 2 9 Ar); m/z (FAB1) 377 (100% MH1). 1,1,1-Trichloroethyl (2R*,3a9R*,7a9R*)- and (2S*,3a9R* 7a9R*)-spiro[4H-1,3-benzodioxine-2,49-perhydro[1,3]benzodioxole]- 6-carboxylate 30 and 31 These compounds were prepared from 27a by dihydroxylation 2798 J.Chem. Soc. Perkin Trans. 1 1997 as described for 20b/21b followed by reaction with triphosgene as described for 22a/23a. The resulting oil was chromatographed on silica using 30% EtOAc–hexane as eluent to give 30 and 31 (57% from 27a) as colourless oils and 30 crystallised on standing. 30 Less polar mp 180–182 8C (Found C 47.1; H 3.6. C17H15O7Cl3 requires C 46.7; H 3.45%); dH(400 MHz; C6D6) 0.9–1.5 (6H m CH2CH2CH2) 3.79 (2H m CHOCHO) 4.04 (1H d J 15 OCHAHB) 4.20 (1H d J 15 OCHAHB) 4.665 (1H d J 12 CO2CHAHB) 4.705 (1H d J 12 CO2CHAHB) 6.60 (1H d J 8 Ar) 7.59 (1H d J 1 Ar) 8.00 (1H dd J 1 8 Ar); dC(100 MHz; CDCl3) 15.30 (CH2CH2- CH2CHO) 25.67 (CH2CH2CH2CHO) 28.27 (CH2CH2CH2- CHO) 60.31 (OCH2) 74.35 (CO2CH2) 75.83 (CH2CHOCHO) 76.30 (CH2CHOCHO) 95.09 (CCl3) 98.60 (COROR9) 117.40 (Ar-CH) 119.20 (Ar-CCO2) 121.86 (Ar-CCH2) 127.55 (Ar-CH) 130.81 (Ar-CH) 154.27 (Ar-CO) 154.38 (OCO2) 164.19 (CO2CH2); m/z (FAB1) 459 (M 1 Na1 70%) 436 (M1 30) 289 (M 2 OCH2CCl3 1 100).31 More polar (Found M1 435.9851. C17H15O7Cl3 requires M 435.9883); dH(400 MHz; C6D6) 0.8–1.7 (6H m CH2CH2CH2) 3.67 (1H d J 6 CH2CHOCHO) 3.81 (1H m CH2CHOCHO) 4.11 (1H d J 15 OCHAHB) 4.24 (1H d J 15 OCHAHB) 4.67 (2H s CO2CH2) 6.77 (1H d J 8 Ar) 7.66 (1H d J 1 Ar) 7.96 (1H dd J 1 8 Ar); dC(100 MHz; CDCl3) 15.16 (CH2CH2CHO) 25.51 (CH2CH2CH2CHO) 30.01 (CH2CH2CH2CHO) 60.89 (OCH2) 74.36 (CO2CH2) ca. 76 (hidden CH2CHOCHO) 78.45 (CH2CHOCHO) 95.14 (CCl3) 98.65 (COROR9) 118.10 (Ar-CH) 119.21 (Ar-CCO2) 12.56 (Ar-CCH2) 127.24 (Ar- CH) 130.83 (Ar-CH) 154.07 (Ar-CO) 155.00 (OCO2) 164.34 (CO2CH2); m/z (FAB1) 459 (M 1 Na1 80%) 436 (M1 50) 289 (M 2 OCH2CCl3 2 100).1,1,1-Trichloroethyl (2R*,29R*,39R*)-29-hydroxy-39-carbamoyloxyspiro[ 4H-1,3-benzodioxine-2,19-cyclohexane]-6-carboxylate 32 and 1,1,1-trichloroethyl (2S*,29R*,39R*)-29-carbamoyloxy- 39-hydroxyspiro[4H-1,3-benzodioxine-2,19-cyclohexane]-6-carboxylate 33 The cyclic carbonate 30 (200 mg 0.45 mmol) was dissolved in 1,1,1-trichloroethanol (2 ml) and liquid ammonia (conc. 1 ml) was added with CAUTION and the reaction mixture was then stirred at room temperature overnight. The mixture was partitioned between CH2Cl2 (20 ml) and brine (20 ml). The organic layer washed with brine (3 × 10 ml) dried over MgSO4 filtered and concentrated under vacuum.The resulting oil was chromatographed on silica using 50–70% EtOAc–hexane as eluent to afford 32 (72 mg 35%) and 33 (48 mg 23%) as colourless oils. 32 Less polar (Found M1 453.0134. C17H18NO7Cl3 requires M 453.0149); dH(400 MHz; C5D5N 1 CD3CO2D) 1.6–2.2 (6H m CH2CH2CH2) 4.61 (1H s CHOH) 4.85 (1H d J 15 OCHAHB) 5.08 (1H d J 15 OCHAHB) 5.11 (2H s CO2CH2) 5.31 (1H m CHOCONH2) 6.99 (1H d J 8 Ar) 7.82 (1H d J 1 Ar) 7.97 (1H dd J 1 8 Ar); m/z (CI1) 454 (MH1 100%). 33 More polar (Found M1 453.0131. C17H18NO7Cl3 requires M 453.0149); dH(400 MHz; C5D5N 1 CD3CO2D) 1.57–2.04 (6H m CH2CH2CH2) 4.32 (1H m CHOH) 4.89 (1H d J 15 OCHAHB) 5.18 (1H d J 11 CO2CHAHB) 5.29 (1H d J 11 CO2CHAHB) 5.53 (1H d J 15 OCHAHB) 5.85 (1H d J 1 CHOCONH2) 6.99 (1H d J 8 Ar) 7.91 (1H d J 1 Ar) 8.01 (1H dd J 1 8 Ar); m/z (CI1) 454 (MH1 100%).(2R*,29R*,39R*)-29-Hydroxy-39-carbamoyloxyspiro[4H-1,3- benzodioxine-2,19-cyclohexane]-6-carboxylic acid 12a The trichloroethyl ester 32 (60 mg 0.13 mmol) was dissolved in THF (3 ml) and zinc dust (100 mg) followed by NaH2PO4 (1 M aq.; 0.65 ml) and H3PO4 (1 M aq; 0.65 ml). The mixture was stirred for 18 h and then a further portion of zinc dust (100 mg) was added. The mixture was filtered through Celite diluted with water (10 ml) and extracted with EtOAc (10 ml). The aqueous layer was adjusted to pH 4 with conc. HCl and extracted with EtOAc (3 × 10 ml). The combined organic layers were dried over MgSO4 filtered and concentrated to give 12a as a white solid (32 mg 58%) mp 203–205 8C (Found M1 323.0983. C15H17NO7 requires M 323.1005); dH(250 MHz; [2H6]DMSO 1 CD3CO2D) 1.5–1.7 (6H m CH2CH2CH2) 3.74 (1H dd J 2 5 CHOH) 4.67 (1H m CHOCONH2) 4.81 (1H d J 15 OCHAHB) 4.99 (1H d J 15 OCHAHB) 5.33 (1H d J 5 OH) 6.36 (2H br s NH2) 6.72 (1H d J 8 Ar) 7.70 (1H d J 1 Ar) 7.75 (1H dd J 1 8 Ar) 12.5 (1H br s CO2H); m/z (CI1) 324 (MH1 100%).(2R*,3a9R*,7a9R*)-6-(2-Formylphenyl)-59,59-dimethylspiro- [4H-1,3-benzodioxine-2,49-perhydro[1,3]benzodioxol]-29-one 35 A mixture of the boronic ester 34b (917 mg 4.83 mmol) 22b (1.19 g 3.22 mmol) and caesium carbonate (1.63 g 5. mmol) was stirred in DMF (25 ml) under an atmosphere of argon. Palladium tetrakis(triphenylphosphine) (75 mg 0.06 mmol) was added and the reaction mixture heated at 80 8C for 4 h. The mixture was cooled to room temperature diluted with CH2Cl2 (30 ml) and water (30 ml) and separated.The organic layer was dried over MgSO4 and then concentrated under vacuum and chromatographed on silica using 20–30% EtOAc–hexane as eluent to give 35 as an off white gum (1.1 g 87%) mp. 105–112 8C (Found C 70.1; H 5.4. C23H22O6 requires C 70.0; H 5.6%); dH(200 MHz; [2H6]DMSO) 1.06 (3H s CH3) 1.11 (3H s CH3) 1.5–2.0 [4H m CH2C(CH3)2CH2] 5.01 (2H s OCH2) 5.04 (1H d J 6 CH2CHOCHO) 5.19 (1H m CH2CHOCHO) 6.99 [1H d J 8 ArC(6)H] 7.23 [1H d J 2 ArC(3)H] 7.29 [1H dd J 2 8 ArC(5)H] 7.51 [1H dt J 1 7 ArC(49)H] 7.58 [1H dd J 1 7 ArC(69)H] 7.75 [1H dt J 2 7 ArC(59)H] 7.91 [1H dd J 2 7 ArC(39)H] 9.92 (1H s CHO); m/z (CI1) 395 (MH1 100%). (2R*,3a9R*,7a9R*)-6-(2-Carboxyphenyl)-59,59-dimethylspiro- [4H-1,3-benzodioxine-2,49-perhydro[1,3]benzodioxol]-29-one 36 The aldehyde 35 (159 mg 0.4 mmol) was dissolved in THF (3 ml) and sulfamic acid (52 mg 0.53 mmol) was added along with water (2 ml).The mixture was stirred for 5 min and then sodium chlorite (47 mg 0.52 mmol) was added and the reaction mixture was stirred for 30 min at room temperature. EtOAc (5 ml) and water (5 ml) were added and the organic layer separated dried over MgSO4 filtered and concentrated under vacuum to give the acid 36 as a colourless gum (162 mg 98%) dH(200 MHz; [2H6]DMSO) 1.08 (3H s CH3) 1.12 (3H s CH3) 1.5–2.0 [4H m CH2C(CH3)2CH2] 4.99 (2H s OCH2) 4.96 (1H d J 6 CH2CHOCHO) 5.18 (1H m CH2CHOCHO) 6.90 [1H d J 8 ArC(6)H] 7.12 [1H d J 2 ArC(3)H] 7.19 [1H dd J 2 8 ArC(5)H] 7.37 [1H dd J 2 7 ArC(69)H] 7.43 [1H dt J 1 7 ArC(49)H] 7.55 [1H dt J 2 7 ArC(59)H] 7.73 [1H dd J 2 7 ArC(39)H]; m/z (CI1) 411 (MH1 100%).(2R*,29R*,39R*)-39-Carbamoyloxy-6-(2-carboxyphenyl)-29- hydroxy-59,59-dimethylspiro[4H-1,3-benzodioxine-2,19-cyclohexane] 13a and (2R*,29R*,39R*)-29-carbamoyloxy-6-(2-carboxyphenyl)- 39-hydroxy-59,59-dimethylspiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 37a The carbamates 13a 1 37a were prepared from 36 (158 mg 0.38 mmol) and ethanolic ammonia in an analogous fashion to that described for 13b and 37b except that the 1 1 mixture of isomers that were produced was not separated (120 mg 73%). Satisfactory analyses were obtained with spectral properties similar to those of 13b and 37b. (2R*,29R*,39R*)-6-(2-Carboxyphenyl)-29-hydroxy-59,59- dimethyl-39-methylcarbamoyloxyspiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 13b and (2R*,29R*,39R*)-6-(2-carboxyphenyl)- 39-hydroxy-59,59-dimethyl-29-methylcarbamoyloxyspiro[4H-1,3- benzodioxine-2,19-cyclohexane] 37b The carbonate 36 (250 mg 0.61 mmol) was stirred in a saturated solution of methylamine (30% in industrial methylated spirits) for 18 h and then diluted with water (10 ml) adjusted to pH 2 with conc.HCl and extracted with EtOAc (10 ml). The J. Chem. Soc. Perkin Trans. 1 1997 2799 organic layer was dried over MgSO4 filtered and concentrated under vacuum to give 13b and 37b in a 2 1 mixture as an oil. Et2O was added and 37b was precipitated as a white solid which was collected by filtration (39 mg 15%). The filtrate which still contained a mixture of 13b and 37b was chromatographed on silica using 0.1% HOAc–70–100% EtOAc–hexane as eluent to give a pure sample of 13b (144 mg 53%). 13b (Rf 0.1% HOAc– EtOAc 0.2) (Found M1 441.1760.C24H27NO7 requires M 441.1787); dH(250 MHz; [2H6]DMSO) 0.93 (3H s CH3) 1.09 (3H s CH3) 1.3–1.9 [4H m CH2C(CH3)2CH2] 3.88 (1H m CHOH) 4.77 (1H d J 15 OCHACHB) 4.91 (1H d J 15 OCHACHB) 4.99 (1H m CHOCONHMe) 5.27 (1H d J 6 CHOH) 6.82 [1H d J 8 ArC(6)H] 6.90 (1H m NH) 7.03 [1H d J 2 ArC(3)H] 7.12 [1H dd J 2 8 ArC(5)H] 7.36 [1H dd J 2 7 ArC(69)H] 7.40 [1H dd J 2 7 ArC(49)H] 7.53 [1H dt J 1 7 ArC(59)H] 7.68 [1H dt J 2 7 ArC(59)H]; m/z (ESP1) 442 (MH1 100%). 37b (Rf 0.1% HOAc–EtOAc 0.15) (Found M1 441.1745. C24H27NO7 requires M 441.1787); dH(250 MHz; [2H6]DMSO) 0.95 (3H s CH3) 1.10 (3H s CH3) 1.3–1.7 [4H m CH2C(CH3)2CH2] 2.58 (3H d J 4 NHCH3) 4.00 (1H m CHOH) 4.70 (1H d J 6 CHOCONHCH3) 4.73 (1H d J 15 OCHACHB) 4.94 (1H d J 15 OCHACHB) 5.06 (1H m CHOH) 6.84 [1H d J 8 ArC(6)H] 6.97 (1H m NH) 7.03 [1H d J 2 ArC(3)H] 7.13 [1H dd J 2 8 ArC(5)H] 7.33 [1H dd J 2 7 ArC(69)H] 7.41 [1H dd J 2 7 ArC(49)H] 7.53 [1H dt J 1 7 ArC(59)H] 7.67 [1H dt J 2 7 ArC(59)H]; m/z (ESP1) 442 (MH1 70%) 216 (MH1 2 C10H17NO4 100).(2R*,29R*,39R*)-6-(2-Carboxyphenyl)-29-hydroxy-59,59- dimethyl-39-ethylcarbamoyloxyspiro[4H-1,3-benzodioxine-2,19- cyclohexane] 13c and (2R*,29R*,39R*)-6-(2-carboxyphenyl)-39- hydroxy-59,59-dimethyl-29-ethylcarbamoyloxyspiro[4H-1,3- benzodioxine-2,19-cyclohexane] 37c The carbamates 13c 1 37c were prepared from 36 (250 mg 0.61 mmol) and ethylamine in ethanol in an analogous fashion to that described for 13b and 37b to give 13c (158 mg 57%) as a gum and 37c as a white solid (68 mg 25%). Satisfactory analyses were obtained with spectral properties similar to those of 13b and 37b.(2R*,29R*,39R*)-6-(2-Carboxyphenyl)-29-hydroxy-59,59- dimethyl-39-isopropylcarbamoyloxyspiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 13d and (2R*,29R*,39R*)-6-(2-carboxyphenyl)- 39-hydroxy-59,59-dimethyl-29-isopropylcarbamoyloxyspiro[4H- 1,3-benzodioxine-2,19-cyclohexane] 37d The carbamates 13d 1 37d were prepared from 36 (250 mg 0.61 mmol) and ethylamine in ethanol in an analogous fashion to that described for 13b and 37b to give 13d (195 mg 68%) as a gum and 37d as a white solid (55 mg 19%). Satisfactory analyses were obtained with spectral properties similar to those of 13b and 37b. (3R*,3a9R*,7a9R*)- and (2S*,3a9R*,7a9R*)-6-Bromo- 29,29,69,69-tetramethylspiro[4H-1,3-benzodioxine-2,49-perhydro- [1,3]benzodioxole] 38 and 39 A mixture of the diols 22b and 23b (6.6 g 19.2 mmol) was dissolved in DMF (50 ml) and 2,2-dimethoxypropane (10 ml) was added along with conc.HCl (0.2 ml). The solution was stirred for 18 h at room temperature and then partitioned between EtOAc (150 ml) and water (150 ml). The organic layer was washed with water (2 × 100 ml) and brine (100 ml) dried over NaSO4 filtered and concentrated. The residual oil was chromatographed on silica using 3–6% EtOAc–hexane as eluent to give 38 (3.2 g 43%) as a colourless oil and 39 (3.1 g 42%) as a white solid. 38 (Found C 56.9; H 5.9. C18H23BrO4 requires C 56.4; H 6.05%); dH(400 MHz; CDCl3) 1.015 (3H s CH2- CCH3axCH3eq) 1.08 (3H s CH2CCH3axCH3eq) 1.34 (3H s CH3) 1.55 (3H s CH3) 1.635 (1H dd J 7 14 CHeqCHax- CHO) 1.75 (1H dd J 5 14 CHeqCHaxCHO) 1.78 [2H s CH2C(OR)2] 4.185 (1H d J 5.5 CH2CHOCHO) 4.50 (1H dd J 2 6 CH2CHOCHO) 4.82 (1H d J 15 OCHAHB) 4.98 (1H d J 15 OCHAHB) 6.70 (1H d J 9 ArH) 7.125 (1H d J 2 ArH) 7.265 (1H dd J 2 9 ArH); irradiation of CH3ax (1.015) gives an NOE enhancement of ArH (6.70); m/z (CI1) 383 (MH1 70%) 202 (C7H5OBr 1 NH4 1 100).39 Mp 66– 75 8C (Found C 56.7; H 6.2. C18H23BrO4 requires C 56.4; H 6.05%); dH(400 MHz; CDCl3) 1.04 (3H s CH2CCH3axCH3eq) 1.07 (3H s CH2CCH3axCH3eq) 1.35 (3H s CH3) 1.50 (3H s CH3) 1.67–1.85 [4H m CH2C(CH3)2CH2] 4.155 (1H dd J 1 6 CH2CHOCHO) 4.485 (1H dd J 7 13 CH2CHOCHO) 4.83 (2H s OCH2) 6.815 (1H d J 9 ArH) 7.075 (1H d J 2 ArH) 7.25 (1H dd J 2 9 ArH); irradiation of CH3ax (1.04) gives an NOE enhancement of OCH2 (4.83); m/z (CI1) 383 (MH1 20%) 216 (C11H18O3 1 NH4 1 100). (2R*,3a9R*,7a9R*)-6-(1,3,2-Dioxaborinan-1-yl)-29,29,69,69- tetramethylspiro[4H-1,3-benzodioxine-2,49-perhydro[1,3]- benzodioxole] 40 38 (3.1 g 8.1 mmol) was dissolved in THF (80 ml) and cooled under argon in a dry ice–acetone bath.BunLi (1.6 M in hexane; 5.6 ml 8.9 mmol) was added over 10 min and the reaction mixture stirred for a further 5 min before adding triisopropyl borate (1.82 g 9.7 mmol) rapidly. The mixture was allowed to warm to room temperature diluted with saturated aqueous ammonium chloride (100 ml) and extracted with EtOAc (100 ml). The organic layer was washed with water (2 × 50 ml) and concentrated under vacuum. Toluene (100 ml) was added followed by propane-1,3-diol (5 ml) and the mixture stirred for 1 h at room temperature diluted with water and the organic layer concentrated under vacuum to give 40 (2.6 g 80%) as a pale yellow oil which crystallised on standing mp 125–127 8C (Found C 65.2; H 7.5.C21H29BO6 requires C 65.0; H 7.5%); dH(200 MHz; CDCl3) 1.03 (3H s CH3ax) 1.07 (3H s CH3eq) 1.33 (3H s CH3) 1.57 (3H s CH3) 1.6–1.8 (2H m CH2CHO) 1.81 [2H s CH2C(OR)2] 2.03 [2H quintet J 5 CH2(CH2O)2B] 4.14 [4H t J 5 CH2(CH2O)2B] 4.20 (1H d J 6.5 CH2CHOCHO) 4.49 (1H m CH2CHOCHO) 4.85 (1H d J 15 OCHAHB) 5.02 (1H d J 15 OCHAHB) 6.78 (1H d J 8 ArH) 7.41 (1H d J 2 ArH) 7.56 (1H dd J 2 8 ArH); m/z (EI1) 388 (M1 20%) 330 (M1 2 C3H6O 20) 190 (C11H18O3 1 100). (2R*,29R*,39R*)-29,39-Dihydroxy-59,59-dimethyl-6-(2-pyridyl)- spiro[4H-1,3-benzodioxine-2,19-cyclohexane] 42 Boronic ester 40 (339 mg 0.87 mmol) and 2-bromopyridine (138 mg 0.87 mmol) were dissolved in DMF (10 ml). Cs2CO3 (425 mg 1.3 mmol) was added and the mixture placed under an atmosphere of argon.Pd(PPh3)4 (20 mg 0.017 mmol) was added and the mixture stirred at 100 8C for 2 h and then cooled to room temperature. The reaction mixture was partitioned between EtOAc (40 ml) and water (40 ml) and the organic layer was washed with water (2 × 20 ml) and brine (20 ml) dried over Na2SO4 filtered and concentrated to give crude 41 (385 mg) as an oil which was dissolved in MeOH (15 ml) and 1 M HCl (3 ml) was added. The solution was stirred for 3 h at room temperature left standing at 0 8C for 15 h and then diluted with water 20 ml) and adjusted to pH 11 with 1 M NaOH. The aqueous solution was extracted with EtOAc (2 × 30 ml) and the combined organics were washed with brine (20 ml) dried over Na2SO4 filtered and concentrated to an oil.Trituration with Et2O gave a small quantity of 42 as a white crystalline solid (55 mg 18%) along with the filtrate which was concentrated to an oil (250 mg) containing mostly 42 which was progressed without further purification. For the solid mp 123–129 8C (Found C 69.7; H 6.8; N 3.7; H2O 0.6. C20H23NO4?0.1H2O requires C 70.0; H 6.8; N 4.1; H2O 0.5%); dH(250 MHz; CDCl3) 0.98 (3H s CH3ax) 1.02 (3H s CH3eq) 1.50–1.72 [4H m CH2C- (CH3)2CH2] 2.08 (1H d J 9 CH2CHOH) 2.45 [1H s CHOHC- (OR)2] 3.97 [1H s CHOH(OR)2] 4.18 (1H m CH2CHOH), 2800 J. Chem. Soc. Perkin Trans. 1 1997 4.92 (1H d J 15 OCHAHB) 5.02 (1H d J 15 OCHAHB) 6.90 [1H d J 8 Ar(6)H] 7.19 [1H m pyr(5)H] 7.6–7.8 [4H m Ar(3 1 5)H 1 pyr(3 1 4)H] 8.65 [1H m pyr(6)H]; m/z (ESP1) 342 (MH1 90%) 184 (100 MH1 2 C8H14O3).(2R*,3a9R*,7a9R*)-69,69-Dimethyl-6-(2-pyridyl)spiro[4H-1,3- benzodioxine-2,49-perhydro[1,3]benzodioxol]-29-one 43 The compound was prepared from crude 42 in a similar manner as described for 22a (ca. 44% yield from 40) which was used directly dH(250 MHz; CDCl3) 1.11 (3H s CH3ax) 1.13 (3H s CH3eq) 1.8–2.05 [4H m CH2C(CH3)2CH2] 4.75 (1H dd J 1 8 CHO) 4.95 (1H d J 15 OCHAHB) 5.0 (1H m CHO) 5.08 (1H d J 15 OCHAHB) 6.91 [1H d J 8 Ar(6)H] 7.20 [1H m pyr(5)H] 7.6–7.8 [4H m Ar(3 1 5)H 1 pyr(3 1 4)H] 8.65 [1H m pyr(6)H]. 2R*,3a9R*,7a9R*)-69,69-Dimethyl-6-(1-oxido-2-pyridinio)- spiro[4H-1,3-benzodioxine-2,49-perhydro[1,3]benzodioxol]-29- one 44 50% m-Chloroperbenzoic acid (125 mg 0.36 mmol) was dissolved in CH2Cl2 (2 ml) and the organic layer added to a solution of 43 (110 mg 0.3 mmol) in CH2Cl2 (10 ml).The solution was stirred at room temperature for 15 h and then concentrated and chromatographed on silica using 5% MeOH–CH2Cl2 as eluent to give 44 as a colourless oil (105 mg 91%) which was used directly dH(250 MHz; CDCl3) 1.12 (3H s CH3ax) 1.13 (3H s CH3eq) 1.75–2.0 [4H m CH2C(CH3)2CH2] 4.75 (1H dd J 1 8 CHO) 4.91 (1H d J 15 OCHAHB) 5.0 (1H m CHO) 5.08 (1H d J 15 OCHAHB) 6.92 [1H d J 8 Ar(6)H] 7.21 [1H dt J 1 6 pyr(4)H] 7.28 [1H t J 6 pyr(5)H] 7.60 [1H dd J 1 8 Ar(5)H] 7.65 [1H d J 1 Ar(3)H] 8.30 [1H d J 6 pyr(6)H]; m/z (ESP1) 384 (MH1 100%). (2R*,29R*,39R*)39-Carbamoyloxy-29-hydroxy-59,59-dimethyl-6- (1-oxido-2-pyridinio)spiro[4H-1,3-benzodioxine-2,19-cyclohexane] 45 and (2R*,29R*,39R*)-29-carbamoyloxy-39-hydroxy- 59,59-dimethyl-6-(1-oxido-2-pyridinio)spiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 46 The compounds were prepared as a 1 1 mixture from 44 in a similar manner as described for 24/25 (26 mg 50%) (Found M1 400.1625.C21H24N2O6 requires M 400.1634); dH(200 MHz; CDCl3) (45 major isomer) 1.04 (3H s CH3ax) 1.14 (3H s CH3eq) 1.55–1.9 [4H m CH2C(CH3)2CH2] 4.1 (1H br s CHOH) 4.73 (1H br s OH) 4.9 (2H m OCHAHB) 5.25 (1H m CHOCONH2) 6.90 [1H d J 8 Ar(6)H] 7.3–7.75 (7H m 5 × ArH 1 NH2) 8.55 [1H dd J 0.5 8 pyr(6)H]; (46 minor isomer) 1.01 (3H s CH3ax) 1.11 (3H s CH3eq) 1.5–1.9 [4H m CH2C(CH3)2CH2] 4.55 (1H m CHOH) 4.74 (1H br s OH) 4.9 (2H m OCHAHB) 5.24 (1H br s CHOCONH2) 6.95 [1H d J 8 Ar(6)H] 7.3–7.75 (7H m 5 × ArH 1 NH2) 8.55 [1H dd J 0.5 8 pyr(6)H]; m/z (ESP1) 401 (MH1 95%) 200 (100 M1 2 C9H15NO4). (2R*,29R*,39R*)-2-Hydroxy-59,59-dimethyl-39-methylcarbamoyloxy- 6-(1-oxido-2-pyridinio)spiro[4H-1,3-benzodioxine- 2,19-cyclohexane] 47 and (2R*,29R*,39R*)-3-hydroxy-59,59- dimethyl-29-methylcarbamoyloxy-6-(1-oxido-2-pyridinio)- spiro[4H-1,3-benzodioxine-2,19-cyclohexane] 48 The compounds were prepared as a 2 1 mixture from 44 in a similar manner as described for 24/25 except using methylamine (44 mg 82%) (Found M1 414.1794.C22H26N2O6 requires M 414.1791); dH(400 MHz; CDCl3) (47 major isomer) 1.01 (3H s CH3ax) 1.12 (3H s CH3eq) 1.5–1.9 [4H m CH2C(CH3)2CH2] 2.81 (3H d J 5 NHCH3) 4.1 (1H br s CHOH) 4.79 (1H br s OH) 4.92 (1H d J 15 OCHAHB) 4.99 (1H d J 15 OCHAHB) 5.29 (1H m CHOCONHMe) 6.98 [1H d J 8 Ar(6)H] 7.23 1 7.31 1 7.44 [3 × 1H m pyr(3,4,5)H] 7.55 [1H dd J 2 7 Ar(5)H] 7.77 [1H d J 2 Ar(3)H] 8.33 [1H d J 8 pyr(6)H]; (48 minor isomer) 1.04 (3H s CH3ax) 1.15 (3H s CH3eq) 1.5– 1.9 [4H m CH2C(CH3)2CH2] 2.87 (3H d J 5 NHCH3) 4.3 (1H m CHOH) 4.79 (1H br s OH) 4.85 (1H d J 15 OCHAHB) 4.95 (1H d J 15 OCHAHB) 5.24 (1H br s CHOCONHMe) 6.94 [1H d J 8 Ar(6)H] 7.23 1 7.31 1 7.44 [3 × 1H m pyr(3,4,5)H] 7.63 [1H dd J 2 7 Ar(5)H] 7.77 [1H m Ar(3)H] 8.33 [1H d J 8 pyr(6)H]; m/z (ESP1) 415 (MH1 85%) 200 (100 M1 2 C10H17NO4).(2R*,3a9R*,7aR*)-29,29,69,69-Tetramethyl-6-[2-(1,1,1-trichloroethoxycarbonyl) phenyl]spiro[4H-1,3-benzodioxine-2,49- perhydro[1,3]benzodioxole] 49 This compound was prepared from 40 and 1,1,1-trichloroethyl 2-iodobenzoate in a similar manner as described for 41 (420 mg 53%) mp 57–58 8C (Found C 58.6; H 5.5. C27H29C13O6 requires C 58.3; H 5.3%); dH(200 MHz; CDCl3) 1.05 (6H s CH3ax 1 CH3eq) 1.26 (3H s CH3) 1.43 (3H s CH3) 1.55–- 1.8 [4H m CH2C(CH3)2CH2] 4.15 (1H dd J 1 6.5 CH2CHOCHO) 4.46 (1H m CH2CHOCHO) 4.76 (1H d J 15 OCHAHB) 4.90 (2H s CO2CH2) 4.94 (1H d J 15 OCHA HB) 6.82 (1H d J 8 ArH) 7.05 (1H d J 2 ArH) 7.13 (1H dd J 2 8 Ar9H) 7.43 (1H dd J 2 8 ArH) 7.51 (1H dt J 2 8 Ar9H) 7.65 (1H dt J 2 8 Ar9H) 7.83 (1H dd J 2 8 Ar9H); m/z (CI1) 555 (M1 100%).(2R*,29R*,39R*)-(29,39-Dihydroxy)-59,59-dimethyl-6-[2-(1,1,1- trichloroethoxycarbonyl)phenyl]spiro[4H-1,3-benzodioxine-2,19- cyclohexane] 50 The acetonide 49 (360 mg 0.65 mmol) was dissolved in acetonitrile (10 ml) and HCl (1 M; 3 ml) was added. The solution was stirred for 2 h at room temperature and then quenched by the addition of solid NaHCO3. Water (10 ml) was added and the aqueous solution was extracted with EtOAc (30 ml) and the organic layer washed with brine (20 ml) dried over Na2SO4 filtered and concentrated under vacuum to give 50 (350 mg) as a slightly impure colourless oil which was used directly dH(200 MHz; CDCl3) 0.92 (3H s CH3ax) 1.05 (3H s CH3eq) 1.25 (1H m CHeqHaxCHOH) 1.46–1.74 [3H m CH2C(CH3)2CHeqHax- CHOH] 3.70 (1H m CH2CHOHCHOH) 3.85 (1H m CH2CHOHCHOH) 4.38 (1H d J 7 OH) 4.49 (1H d J 6 OH) 4.75 (1H d J 15 OCHAHB) 4.89 (1 H d J 15 OCHAHB) 4.90 (2H s CO2CH2) 6.80 (1H d J 8 ArH) 7.02 (1H d J 2 ArH) 7.12 (1H dd J 2 8 Ar9H) 7.43 (1H dd J 2 8 ArH) 7.51 (1H dt J 2 8 Ar9H) 7.65 (1H dt J 2 8 Ar9H) 7.84 (1 H dd J 2 8 Ar9H); m/z (ESP1) 532 (M 1 NH4 1 100%).(2R*,29R*,39R*)-29-Hydroxy-59,59-dimethyl-39-(5-methylpyrrol- 2-ylcarbonyloxy)-6-[2-(1,1,1-trichloroethoxycarbonyl)- phenyl]spiro[4H-1,3-benzodioxine-2,19-cyclohexane] 51 The diol (50 284 mg 0.55 mmol) and 5-methyl-2-trichloroacetylpyrrole (375 mg 1.65 mmol) were dissolved in THF (30 ml) and potassium tert-butoxide (74 mg 0.66 mmol) was added.The mixture was stirred at reflux for 2 h and further portions of the pyrrole (200 mg 0.88 mmol) and potassium tert-butoxide (80 mg 0.71 mmol) were added. After stirring at reflux for a further 2 h more potassium tert-butoxide (100 mg 0.89 mmol) was added and the mixture stirred at reflux for 15 h cooled and diluted with NH4Cl (sat. aq. 30 ml) and extracted with EtOAc (30 ml). The organic layer was washed with brine (20 ml) dried over Na2SO4 filtered and concentrated under vacuum to a yellow oil which was chromatographed on silica using 1 2 EtOAc–hexane as eluent to give 51 as a colourless gum (73 mg 21%) along with recovered 50 (57 mg 27% yield of 51 based on recovered starting material).51 (Found M1 621.1034. C30H30NO7Cl3 requires M 621.1088); dH(250 MHz CDCl3) 1.02 (3H s CH3ax) 1.15 (3H s CH3eq) 1.2–1.75 [4H m CH2C(CH3)2CH2] 2.30 (3H s pyrrCH3) 4.14 (1H m CHOH) 4.75 (2H s OCH2) 4.90 (2H s CO2CH2) 5.52 (1H m CHOCO) 5.95 (1H m pyrrC3-H) 6.81 (1H m pyrrC4-H) 6.90 (1H d J 8 ArH) 6.95 (1H d J 2 ArH) 7.14 (1H dd J 2 8 Ar9H) 7.35 (1H dd J 2 8 ArH) 7.42 J. Chem. Soc. Perkin Trans. 1 1997 2801 (1H dt J 2 8 Ar9H) 7.60 (1H dt J 2 8 Ar9H) 7.93 (1H dd J 2 8 Ar9H) 9.0 (1H br s NH); m/z (CI1) 622 (M 1 H1 100%). (2R*,29R*,39R*)-6-(2-Carboxyphenyl)-29-hydroxy-59,59- dimethyl-39-(5-methylpyrrol-2-carbonyloxy)spiro[4H-1,3- benzodioxine-2,19-cyclohexane] 52 The trichloroethyl ester 51 (73 mg 0.117 mmol) was dissolved in THF (5 ml).Zinc dust (250 mg) was added along with NH4OAc (1M; 6 ml) and acetic acid (5 ml). The mixture was stirred at room temperature for 2 h and then filtered through Celite washing the residue with water (5 ml) and EtOAc (10 ml). A further portion of EtOAc (20 ml) was added and the organic layer separated and washed with water (10 ml) and brine (10 ml) dried over Na2SO4 filtered and concentrated under vacuum. The residue was chromatographed on silica using 7–12% MeOH–CH2Cl2 as eluent to afford 52 (25 mg 44%) as a colourless glassy solid (Found M1 491.1906. C28H29NO7 requires M 491.1944); dH(250 MHz; CDCl3) 0.98 (3H s CH3ax) 1.11 (3H s CH3eq) 1.2–1.95 [4H m CH2C(CH3)2CH2] 2.20 (3H s pyrrCH3) 4.15 (1H m CHOH) 4.85 (2H s OCH2) 5.46 (1H m CHOCO) 5.93 (1H s pyrrC3-H) 6.75–6.85 (3H m pyrrC4-H 1 2 × ArH) 7.13 (1H d J 8 Ar9H) 7.30–7.40 (2H m ArH 1 Ar9H) 7.50 (1H t J 8 Ar9H) 7.86 (1H d J 8 Ar9H) 9.4 (1H br s NH); m/z (ESP1) 492 (M 1 H1 100%).Acknowledgements We thank S. Brightwell Dr K. J. Barrett-bee Mrs L. Dawson Mrs C. M. Morton and Dr R. N. Illingworth for the biological testing of the compounds; Mr G. A. Holdgate and Dr W. H. J. Ward for differential scanning calorimetry studies with gyrase B 24kD fragment; Mr B. Wright Mr H. A. Beeley and Mr P. A. McNally for NMR support; Mrs M. F. Vickers for MS support and Dr M. C. McPartlin for crystal structures of 21a and 24. References 1 For a review of Gyrase structure and function see R. J. Reece and A. Maxwell Crit. Rev. Biochem.Mol. Biol. 1991 26 335. 2 K. Orlica and S. Coughlin Pharmacol. Ther. 1989 44 107; S. Radl Pharmacol. Ther. 1990 48 1. 3 N. Nakada H. Shimada T. Hirata Y. Aoki T. Kamiyama J. Watanabe and M. Arisawa Antimicrob. Agents Chemother. 1993 37 2656; E. Goetschi P. Angehrn H. Gmünder P. Hebeisen H. Link R. Masciadri and J. Nielsen Pharmacol. Ther. 1994 60 367; J. Watanabe N. Nakada S. Sawaira H. Shimada S. Ohshima T. Kamiyama and M. Arisawa J. Antibiot. 1994 47 32; T. Kamiyama N. Shimma T. Ohtsuka N. Nakayama Y. Itezono N. Nakada J. Watanabe and K. Yokose J. Antibiot. 1994 47 37; N. Nakada H. Gmünder T. Hirata and M. Arisawa Antimicrob. Agents Chemother. 1994 38 1966. 4 M. Arisawa E. Gotshi T. Kamiyama R. Masciadri H. Shimada J. Watanabe P. Hebeisen and H. Link EP-106105 1991; E. Goschi P.Hebeisen H. Link and T. Lubbers EP-0-675-122-A2 1995. 5 M. Oram B. Dosanjh N. A. Gormley C. V. Smith L. M. Fisher A. Maxwell and K. Duncan Antimicrob. Agents Chemother. 1996 40 473. 6 R. H. Chen D. N. Whittern A. M. Buko and J. B. McAlpine J. Antibiotics 1989 42 533; M. S. Osbourne W. M. Maiese and Mgreenstein Antimicrob. Agents Chemother. 1990 34 1450; C. Hubschwerlen I. Kompis and J.-L. Specklin EP 0 433 648 A1 1990; J. L. Vizan C. Hernandez-Chico I. del Castillo and F. Moreno EMBO J. 1991 10 467; S. Sakemi T. Inagaki K. Kaneda H. Hiraia E. Iwata T. Sakakibara Y. Yamauchi M. Norcia L. M. Wondrack J. A. Sutcliffe and N. Kojima J. Antibiot. 1995 48 134. 7 L. P. Garrod H. P. Lambert and F. O’Grady Antibiotic and Chemotherapy Churchill Livingstone Edinburgh 5th edn. 1981 pp. 225–229.8 H. Kawaguchi H. Tsukiura M. Okanishi T. Miyaki T. Ohmori K. Fujisawa and H. Koshiyama J. Antibiot. 1965 18 1; J. Berger A. J. Schochter A. D. Batcho B. Pecherer O. Keller J. Maricq A. E. Karr B. P. Vaterlaus A. Furlenmeier and H. Speigelberg Antimicrob. Agents Chemother. 1965 9 778. 9 L. Linet F. Benazet Y. Chapentie M. Dubost J. Florent D. Mancy J. Preud’Homme T. L. Threlfall B. Vuillemin D. E. Wright A. Abraham M. Cartier N. de Chezelles C. Godard and J. Theilleux C. R. Acad. Sci. Ser. C. 1972 275 455. 10 J. P. Poyser K. J. Rogers N. Taylor D. Hollinshead A. J. Platt J. Culshaw A. Glen D. Heaton G. Kay and L. Feron unpublished results Zeneca Pharmaceuticals. 11 For reviews of the chemical modifications to the coumarin antibiotics and the resultant SAR see M. J. Ryan Antibiotics vol.V ed. F. E. Hahn Springer Verlag 1979 214; J. G. Keil J. C. Godfrey M. J. Cron I. R. Hooper D. E. Nettleton K. E. Price and H. Schmitz; J. Berger and A. D. Batcho J. Chromatogr. 1978 15 101. 12 M. O. Boles and D. J. Taylor Acta Crystallogr. Sect. B 1975 31 1400. 13 B. Wright unpublished results Zeneca Pharmaceuticals. 14 M. L. Holmes and M. L. Dyall-Smith J. Bacteriol. 1991 173 642; A. Contreras and A. Maxwell Mol. Microbiol. 1992 6 1617; M. Stieger P. Angehrn B. Wohlgensinger and H. Gmünder Antimicrob. Agents. Chemother. 1996 40 1060. 15 The pKa of Novobiocin has been measured as 4.3 (D. Newton and R. Kluza Drug Intell. Clin. Pharm. 1978 12 546). 16 N. L. Allinger J. Am. Chem. Soc. 1977 99 8127. 17 R. J. Lewis O. M. P. Singh C. V. Smith A. Maxwell T. Skarzynski A. J. Wonacott and D.Wigley J. Mol. Biol. 1994 241 128; R. J. Lewis O. M. P. Singh C. V. Smith T. Skarzynski A. Maxwell A. J. Wonacott and D. B. Wigley EMBO J. 1996 15 1412. 18 F. T. F. Tsai O. M. P. Singh T. Skarzynski A. J. Wonacott S. Weston A. Tucker R. A. Pauptit A. L. Breeze J. P. Poyser R. O’Brien J. E. Ladbury and D. B. Wigley Proteins Struct. Funct. Genet. 1997 28 41. 19 E. Garbisch J. Org. Chem. 1965 30 2109. 20 R. M. Burk and M. B. Roof Tetrahedron Lett. 1993 34 395. 21 The X-ray crystal structures for 21a and 24 were solved by Dr M. C. McPartlin at the University of North London unpublished results. 22 G. Just and K. Grozinger Synthesis 1976 457. 23 T. Tsunoda M. Suzuki and R. Noyori Tetrahedron Lett. 1980 21 1357. 24 B. D. Mookherjee R. R. Patel and W. O. Ledig J. Org. Chem. 1971 36 4124.25 S. W. Wright D. L. Hageman and L. D. McLure J. Org. Chem. 1994 59 6095. 26 T. Watanabe N. Miyaura and A. Suzuki Synlett. 1992 207. 27 F. M. Hauser and S. R. Ellenberger Synthesis 1987 723. 28 M. Gellert K. Mizuuchi M. H. O’Dea and H. A. Nash Proc. Natl. Acad. Sci. USA 1976 73 3872. 29 National Committee for Clinical Laboratory Standards (1993) Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically 3rd edn.; approved standard. Document M7-A3 13 (25) Villanova PA. 30 J. A. Holdgate and W. H. J. Ward unpublished results Zeneca Pharmaceuticals. 31 D. B. Wigley G. J. Davies E. J. Dodson A. Maxwell and G. Dodson Nature 1991 351 624. 32 E. M. Duffy P. J. Kowalczyk and W. L. Jorgenson J. Am. Chem. Soc. 1993 115 9271. Paper 7/00647K Received 28th January 1997 Accepted 2nd June 1997
ISSN:1472-7781
DOI:10.1039/a700647k
出版商:RSC
年代:1997
数据来源: RSC
|
|