The total synthesis of quinine, a naturally-occurring antimalarial drug, was developed over a 150-year period. The development of synthetic quinine is considered a milestone in organic chemistry although it has never been produced industrially as a substitute for natural occurring quinine. The subject has also been attended with some controversy: Gilbert Stork published the first stereoselective total synthesis of quinine in 2001, meanwhile shedding doubt on the earlier claim by Robert Burns Woodward and William Doering in 1944, claiming that the final steps required to convert their last synthetic intermediate, quinotoxine, into quinine would not have worked had Woodward and Doering attempted to perform the experiment. A 2001 editorial published in Chemical & Engineering News sided with Stork, but the controversy was eventually laid to rest once and for all when Williams and coworkers successfully repeated Woodward's proposed conversion of quinotoxine to quinine in 2007.[1]
The aromatic component of the quinine molecule is a quinoline with a methoxy substituent. The amine component has a quinuclidine skeleton and the methylene bridge in between the two components has a hydroxyl group. The substituent at the 3 position is a vinyl group. The molecule is optically active with five stereogenic centers (the N1 and C4 constituting a single asymmetric unit), making synthesis potentially difficult because it is one of 16 stereoisomers.
The first step in this sequence is sodium hypobromite addition to quinotoxine to an N-bromo intermediate possibly with structure 2. The second step is organic oxidation with sodium ethoxide in ethanol. Because of the basic conditions the initial product quininone interconverts with quinidinone via a common enol intermediate and mutarotation is observed. In the third step the ketone group is reduced with aluminum powder and sodium ethoxide in ethanol and quinine can be identified. Quinotoxine is the first relay molecule in the Woodward/Doering claim.
The key step in the assembly of quinotoxine is a Claisen condensation:
Woodward and Doering argue that Rabe in 1918 already proved that this compound will eventually give quinine but do not repeat Rabe's work. In this project 27-year-old assistant professor Woodward is the theorist and postdoc Doering (age 26) the bench worker. According to William, Bob was able to boil water but an egg would be a challenge. As many natural quinine resources were tied up in the enemy-held Dutch East Indies, synthetic quinine was a promising alternative for fighting malaria on the battlefield and both men become instant war heroes making headlines in the New York Times, Newsweek and Life.
The starting materials are trans-2-butene-1,4-diol and ethyl orthoacetate and the key step is a Claisen rearrangement
In this process the racemic lactone reacts in aminolysis with (S)-methylbenzylamine assisted by triethylaluminum to a diastereomeric pair of amides which can be separated by column chromatography. The S-enantiomer is converted back to the S-lactone in two steps by hydrolysis with potassium hydroxide and ethylene glycol followed by azeotropic ring closure.
The Stork quinine synthesis starts from chiral (S)-4-vinylbutyrolactone 1. The compound is obtained by chiral resolution and in fact, in the subsequent steps all stereogenic centers are put in place by chiral induction: the sequence does not contain asymmetric steps.
The lactone is ring-opened with diethylamine to amide 2 and its hydroxyl group is protected as a tert-butyldimethyl silyl ether (TBS) in 3. The C5 and C6 atoms are added as tert-butyldiphenylsilyl (TBDPS) protected iodoethanol in a nucleophilic substitution of acidic C4 with lithium diisopropylamide (LDA) at −78 °C to 4 with correct stereochemistry. Removal of the silyl protecting group with p-toluenesulfonic acid to alcohol 4b and ring-closure by azeotropic distillation returns the compound to lactone 5 (direct alkylation of 1 met with undisclosed problems).
The lactone is then reduced to the lactol 5b with diisobutylaluminum hydride and its liberated aldehyde reacts in a Wittig reaction with methoxymethylenetriphenylphosphine (delivering the C8 atom) to form enol ether 6. The hydroxyl group is replaced in a Mitsunobu reaction by an azide group with diphenylphosphoryl azide in 7 and acid hydrolysis yields the azido aldehyde 8.
The methyl group in 6-methoxy-4-methylquinoline 9 is sufficiently acidic for nucleophilic addition of its anion (by reaction with LDA) to the aldehyde group in 8 to form 10 as a mixture of epimers. This is of no consequence for stereocontrol because in the next step the alcohol is oxidized in a Swern oxidation to ketone 11. A Staudinger reaction with triphenylphosphine closes the ring between the ketone and the azide to the tetrahydropyridine 12. The imine group in this compound is reduced to the amine 13 with sodium borohydride with the correct stereospecificity. The silyl protecting group is removed with hydrogen fluoride to alcohol 14 and then activated as a mesyl leaving group by reaction with mesyl chloride in pyridine which enables the third ring closure to 15. In the final step the C9 hydroxyl group was introduced by oxidation with sodium hydride, dimethylsulfoxide and oxygen with quinine to epiquinine ratio of 14:1.
The 1944 Woodward–Doering synthesis starts from 7-hydroxyisoquinoline 3 for the quinuclidine skeleton which is somewhat counter intuitive because one goes from a stable heterocyclic aromatic system to a completely saturated bicyclic ring. This compound (already known since 1895) is prepared in two steps.
The first reaction step is condensation reaction of 3-hydroxybenzaldehyde 1 with (formally) the diacetal of aminoacetaldehyde to the imine 2 and the second reaction step is cyclization in concentrated sulfuric acid. Isoquinoline 3 is then alkylated in another condensation by formaldehyde and piperidine and the product is isolated as the sodium salt of 4.
Hydrogenation at 220 °C for 10 hours in methanol with sodium methoxide liberates the piperidine group and leaving the methyl group in 5 with already all carbon and nitrogen atoms accounted for. A second hydrogenation takes place with Adams catalyst in acetic acid to tetrahydroisoquinoline 6. Further hydrogenation does not take place until the amino group is acylated with acetic anhydride in methanol but by then 7 is again hydrogenated with Raney nickel in ethanol at 150 °C under high pressure to decahydroisoquinoline 8. The mixture of cis and trans isomers is then oxidized by chromic acid in acetic acid to the ketone 9. Only the cis isomer crystallizes and used in the next reaction step, a ring opening with the alkyl nitrite ethyl nitrite with sodium ethoxide in ethanol to 10 with a newly formed carboxylic ester group and an oxime group. The oxime group is hydrogenated to the amine 11 with platinum in acetic acid and alkylation with iodomethane gives the quaternary ammonium salt 12 and subsequently the betaine 13 after reaction with silver oxide.
Quinine's vinyl group is then constructed by Hofmann elimination with sodium hydroxide in water at 140 °C. This process is accompanied by hydrolysis of both the ester and the amide group but it is not the free amine that is isolated but the urea 14 by reaction with potassium cyanate. In the next step the carboxylic acid group is esterified with ethanol and the urea group replaced with a benzoyl group. The final step is a Claisen condensation of 15 with ethyl quininate 16, which after acidic workup yields racemic quinotoxine 17. The desired enantiomer is obtained by chiral resolution with the chiral dibenzoyl ester of Tartaric acid. The conversion of this compound to quinine is based on the Rabe–Kindler chemistry discussed in the timelime.