Ozonolysis of some 8-alkoxyquinolines, and synthesis of a precursor to the non-sedating antihistamine Claritin

3-Formyl-2-methoxycarbonylpyridine and isopropyl 3-formylpyridine-2-carboxylate have each been efficiently accessed in one step via the ozonolyses of 8-methoxy-or of 8-isopropoxy-quinoline under near-ambient conditions. The compounds can be utilized as intermediates for syntheses of the tricyclic ketone 8-chloro-6,11-dihydro-5 H -benzo[5,6]cyclohepta[1,2-b ]pyridin-11-one, a precursor to the important non-sedating antihistamine Claritin.


Introduction
Ethyl 4-(8-chloro-5,6-dihydro-11H-benzo [5,6]cyclohepta [1,2-b]pyridin-11-ylidene)piperidine-1carboxylate 1, is a widely-utilised non-sedating anti-histaminic, anti-allergenic drug that is structurally somewhat related to the tricyclic antidepressants.The compound 1 is sold under trade names such as Claritin, 1 Clarityn 2 and Loratadine.Sales of 1 during 2001 exceeded US $3 billion, making it the fourth-largest selling drug in the world for that year. 3The drug became a generic product in the USA at the end of 2002, but the $ value of global sales is now far lower than it was.An interesting history of the discovery of Claritin has been published. 4here is an extensive literature that describes synthetic routes to Claritin 1. Thus (Scheme 1), an initial disclosure 5 was followed by others, 6,7,8 each of which broadly described Claisen-like condensations between an alkyl 3-picolinate 2 and the anion derived from the benzylic nitrile 3 to give the keto-nitrile 4. Scheme 1. Early routes to the key tricyclic ketone 7. [5][6][7][8] This was then subjected to series of manipulations to yield the N-oxide 5, which underwent a Reissert-Henze reaction to deliver the pyridyl nitrile 6. Hydrolysis of 6 to the derived carboxylic A variant on this route 13,14 involves the intramolecular cyclisation of a secondary amide 13 using strongly acidic catalysts such as P2O5 -CF3SO3H to give an imine that is then hydrolysed to yield the ketone 7.
It has been reported 15 that 3-methylpyridine-2-carboxylic acid 14 reacts smoothly with two equivalents of LDA, even at ambient temperatures, to give a dianion that is C-alkylated by 3chlorobenzyl chloride to yield the acid 15, which can then be cyclised to give 7. Similarly, the dihydrooxazoline 16 can be mono-lithiated at the 3-methyl group and then alkylated using 3chlorobenzyl chloride to give 16 another precursor of the tricyclic ketone 7.

Results and Discussion
In devising possible alternative routes to Claritin 1, the financial and pharmaceutical potency of which was earlier apparent, it appeared to us that the regioselective ozonolysis of suitable quinoline derivatives under environmentally benign conditions might provide useful differentially functionalised pyridines that would be potentially valuable synthons for the key tricyclic ketone intermediate 7.

Scheme 4. Selective ozonolysis of an 8-alkoxyquinoline
We considered (Scheme 4) that in an ether 17 derived from 8-hydroxyquinoline, the electronrich aryl ring would undergo preferential oxidative cleavage by ozone leaving the more electrondeficient pyridine ring intact and provide, via secondary cleavage of the (Z)-enal 18, a 2-carboxyalkylpyridine-3-carboxaldehyde 19 that might then be converted into the key tricyclic ketone 7.
Some early studies on the exhaustive degradative ozonolysis of alkylquinolines were carried out during the 1940s by Schenck and Bailey. 17,18More controlled experiments 19 with a number of substituted quinolines that included 8-hydroxyquinoline 20 showed that the initial reaction with ozone took place on the benzenoid ring, and that oxidative work-up yielded pyridine 2,3dicarboxylic acid 21 (5-95%).Pyridine-2,3-dicarboxaldehyde 22 has 20 also been obtained via ozonolysis of quinoline, but yields were very low (<3%).It was later shown [20][21][22] that the pyridyl ring of quinoline was also attacked by ozone, but only to a minor extent.
Quinoline has been converted into pyridine-3-carboxylic acid 23 by a procedure involving ozonolysis in aqueous nitric acid, 23 but ozonolysis of pyridine in neutral aqueous tert-butanol gives 24 the derived N-oxide and this type of reaction represents a potential threat to yields where ozonolysis of quinolines to give pyridine free bases is concerned.However, it has been reported 25 that a number of vinylpyridines can be ozonised at -40 o C in methanol to give modest yields of the derived pyridine carboxaldehydes after reduction of the ozonides with sodium sulfite.
Surprisingly, the literature is almost totally silent regarding the ozonolysis of 8alkoxyquinolines.In a solitary example it has been reported 26 that 3-ethyl-8-methoxyquinoline is ozonised to yield 5-ethylpyridine-1,2-dicarboxylic acid.The oxidative work-up used in this case clearly precluded the isolation of any aldehyde as product.More recently, Taddei et al. have described 27 the successful ozonolysis of three substituted 5-alkoxyquinolines to yield keto-esters as cleavage products in yields ranging from 39-45%.
We initially carried out the ozonolysis of 8-methoxyquinoline 24 in methanol at 0 o C, using triethylamine to reduce the ozonide that was formed.However, only minor amounts of the desired aldehydo-ester 25 were produced, together with some of the diester 26.The outcome (Scheme 5) was completely different if dimethyl sulfide (which has been recommended 28 as being a superior reagent for the reduction of ozonides) was used instead of triethylamine, when the aldehyde 25 was isolated in good (81%) yield.A by-product formed during the ozonolysis of 8-methoxyquinoline 24 was the dimethyl acetal 27.This may be the source of another minor by-product, the diester 26, since ozone is known to oxidise acetals to esters. 29 the ozonolysis of 8-methoxyquinoline 24 at 0 o C in methanol was interrupted before the calculated amount of ozone had been passed into the reaction mixture the (E)-and (Z)-isomers of the unsaturated aldehyde 28 could be isolated.These could not be separated by column chromatography because whenever this was attempted the (Z)-form of 28 underwent conversion into the (E)-isomer, a process that may occur because of traces of acid in the silica gel that was used.The formation of 28 under ozone-limiting conditions is not entirely unexpected, since this reflects initial attack by ozone at the most electron-rich C-C bond of 8-methoxyquinoline 24.Similar partial ozonolysis reactions of naphthalene have been reported. 30,31 Isopropoxyquinoline 29 was also successfully ozonised in methanol at 0 o C to give the expected aldehydo-ester 30 in excellent yield after reductive work-up using dimethyl sulfide.An alternative multi-step synthesis of this compound from quinolinic anhydride has been described, 32 but the overall yield obtained using that route was only ca.20%.With easy access to large quantities of the aldehydo-esters 25 and 30 in hand we next explored olefination reactions of the aldehyde 25.A Horner-Emmons reaction that was attempted between the aldehydo-ester 25 and the anion of the benzylic phosphonate 31 afforded a complex mixture of products from which the desired stilbazole 32 could not be isolated.However (Scheme 6), Wittig olefination of aldehyde 25 using the ylide derived from the phosphonium salt 33 gave a separable mixture of the (E)-and (Z)-isomers of 32.
If unpurified aldehyde 25, obtained directly from the ozonolysis of 8-methoxyquinoline 24, was used in this Wittig reaction, methyl (E)-3-(3-chlorophenyl)prop-2-enoate 34 could be isolated as an additional minor component during chromatography of the product mixture.The precursor to this must be methyl glyoxylate, most likely formed via further reaction of ozone with the initially-formed mono-ozonide that leads from 8-methoxyquinoline to the ,unsaturated aldehyde 28.
Hydrogenation of (E)-/(Z)-32 led to the ethane derivative 35, which was readily hydrolysed to give the pivotal target acid 15.
The isopropyl ester 30 was similarly converted by Wittig olefination into the corresponding stilbazole (E)/(Z)-36.Hydrogenation of 36 gave 37, which was also hydrolysed to yield the acid 15.Conventional intramolecular cyclisation of the acid 15, via its acyl chloride, then delivered the targeted tricyclic ketone 7.

Conclusions
The useful pyridine aldehydo-esters 25 and 30 have been efficiently obtained from inexpensive 8-alkoxyquinolines via simple and relatively "green" ozonolysis reactions carried out under nearambient conditions.These accessible and differentially-functionalised pyridines have been utilised in syntheses of the tricyclic ketone 7, which is a precursor of the non-sedating antihistamine Claritin, and should find other applications in the field of heterocyclic chemistry.

Experimental Section
General.Unless otherwise stated, 1 H and 13 C NMR spectra were recorded for solutions in CDCl3 using a Bruker Avance DPX 400 MHz spectrometer.Coupling constants are recorded in Hz.Assignments were verified where appropriate by 1 H-1 H COSY, 1 H- 13 C COSY, DEPT and HMBC experiments.IR spectra were recorded for Nujol mulls (N) or liquid films (L) between sodium chloride plates using a Mattson FT-IR spectrometer.Mass spectra were obtained under electrospray conditions using a Micromass time-of-flight instrument.Melting points (uncorrected) were measured in unsealed capillary tubes using an Electrothermal IA9100 apparatus.Thin layer chromatography was carried out using Merck Kieselgel 60 F254 0.2 mm silica gel plates.Column chromatography was carried out using Merck Kieselgel 60 (70-230 mesh) silica gel.Ozonolysis was carried out using a BOC Mark 2 apparatus.All solvents were dried and distilled before use.Organic extracts of reaction products were dried over anhydrous magnesium sulfate.(24).8-Hydroxyquinoline (4.6 g; 31.7 mmol) and anhydrous potassium carbonate (10.1 g; 73.1 mmol) in DMF (60 mL) were stirred under N2 at 90 °C for 1 h after which time dimethyl sulfate (TOXIC! 3 mL; 31.7 mmol) was added and the mixture was stirred for a further 2 h at 90 o C. The cooled mixture was then diluted with water (300 mL) and extracted using dichloromethane.The combined organic layers were washed with aqueous potassium hydroxide solution (5%; 100 mL) in portions, and then with water until the washings were no longer alkaline.The extract was dried, filtered and evaporated to give 8methoxyquinoline 24 (2.67 g; 53%), obtained as an oily solid (lit. 33

Ozonolysis of 8-methoxyquinoline 24: 3-formyl-2-methoxycarbonylpyridine (25).
A threenecked flask (250 mL) fitted with a stirring bar and a dropping funnel was placed in an ice-bath and connected to the ozoniser.8-Methoxyquinoline 24 (5.7 g; 35.8 mmol), dissolved in methanol (100 mL), was added and the solution was subjected to a stream of ozonised O2 (containing ~1.5% O3) for 2 h (O2-flow rate of 1 L/min).Ozone production was discontinued and the system was flushed with O2 for 20 min to purge excess reagent.Dimethyl sulfide (6.6 mL; 90 mmol) was slowly added via the dropping funnel while the stirred mixture was continuously cooled in an ice-bath.After a further 30 min, solvents were removed under reduced pressure, the viscous brown oil obtained was taken up in ethyl acetate (100 mL) and the extract was washed with brine and dried.Evaporation of the solvent gave crude 3-formyl-2-carbomethoxypyridine 25 (4.8 g; 81%) which could be used without further purification in the following Wittig reaction.An analytical sample of the aldehyde 25 was obtained by extracting the crude oily product using petroleum ether (bp 40-60 °C).In some cases where 1 H NMR analysis of the crude product revealed significant amounts of impurities, the mixture was chromatographed over silica gel (EtOAc/hexane).In this way, samples of the by-products 26 and 27 were isolated and characterised.Dimethyl pyridine-2,3-dicarboxylate 26 had mp 54-56 °C (EtOAc) ( lit. 34