IChO 2014 · Theory problem 7

Synthesis of artemisinin

Download PDFIChO 2014

(+)-Artemisinin, isolated from Artemisia annua L. (Qinghao, Compositae) is a potent antimalarial effective against resistant strains of Plasmodium. A simple route for the synthesis of Artemisinin is outlined below.

Figure 1

First, pyrolysis of (+)-2-Carene broke the cyclopropane ring forming, among other products, (1R)-(+)-trans-isolimonene A (\(\ce{C10H16}\)), which then was subjected to regioselective hydroboration using dicyclohexylborane to give the required alcohol B in 82% yield as a mixture of diastereoisomers. In the next step, B was converted to the corresponding \(\gamma,\delta\)-unsaturated acid C in 80% yield by Jones’ oxidation.

Figure 2

7.1 Draw the structures (with stereochemistry) of the compounds A - C.

The acid C was subjected to iodolactonization using \(\ce{KI}\), \(\ce{I2}\) in aqueous. \(\ce{NaHCO3}\) solution to afford diastereomeric iodolactones D and E (which differ in stereochemistry only at \(C_3\)) in 70% yield.

Figure 3

7.2 Draw the structures (with stereochemistry) of the compounds D and E.

The iodolactone D was subjected to an intermolecular radical reaction with ketone X using tris(trimethylsilyl)silane (TTMSS) and AIBN (azobisisobutyronitrile) in a catalytic amount, refluxing in toluene to yield the corresponding alkylated lactone F in 72% yield as a mixture of diastereoisomers which differ only in stereochemistry at \(C_7\) along with compound G (~10%) and the reduced product H, \(\ce{C10H16O2}\) (<5%).

Figure 4

7.3 Draw the structures (with stereochemistry) of compound H and the reagent X.

The keto group of F reacted with ethanedithiol and \(\ce{BF3.Et2O}\) in dichloromethane (DCM) at \(0\,^\circ\mathrm{C}\) to afford two diastereomers: thioketal lactones I and J in nearly quantitative yield (98%). The thioketalization facilitated the separation of the major isomer J in which the thioketal group is on the opposite face of the ring to the adjacent methyl group.

Figure 5

7.4 Draw the structures (with stereochemistry) of the compounds I and J.

The isomer J was further subjected to alkaline hydrolysis followed by esterification with diazomethane providing hydroxy methyl ester K in 50% yield. The hydroxy methyl ester K was transformed into the keto ester L using PCC (Pyridium ChloroChromate) as the oxidizing agent in dichloromethane (DCM).

A two-dimensional NMR study of the compound L revealed that the two protons adjacent to the newly-formed carbonyl group are cis to each other and confirmed the structure of L.

Figure 6

7.5 Draw the structures (with stereochemistry) of the compounds K and L.

The ketone L was subjected to a Wittig reaction with methoxymethyl triphenyl-phosphonium chloride and KHMDS (Potassium HexaMethylDiSilazid - a strong, non-nucleophilic base) to furnish the required methyl vinyl ether M in 45% yield. Deprotection of thioketal using \(\ce{HgCl2}\), \(\ce{CaCO3}\) resulted in the key intermediate N (80%). Finally, the compound N was transformed into the target molecule Artemisinin by photo-oxidation followed followed by acid hydrolysis with 70 % \(\ce{HClO4}\).

Figure 7

7.6 Draw the structures (with stereochemistry) of the compounds M and N.

Official English text of the problem.

Answers

The lactone names below use the benzofuran numbering shown in F. All specified stereocentres are encoded in the SMILES.

  • 7.1 — A, B and C

A: (1R,4R)-trans-isolimonene, or (1R,4R)-\(p\)-mentha-2,8-diene.

C[C@H]1C=C[C@H](C(=C)C)CC1

B: A mixture of the (2S) and (2R) forms of 2-[(1R,4R)-4-methylcyclohex-2-en-1-yl]propan-1-ol.

(2S): C[C@H]1C=C[C@H]([C@H](C)CO)CC1

(2R): C[C@H]1C=C[C@H]([C@@H](C)CO)CC1

C: A mixture of the (2S) and (2R) forms of 2-[(1R,4R)-4-methylcyclohex-2-en-1-yl]propanoic acid.

(2S): C[C@H]1C=C[C@H]([C@H](C)C(=O)O)CC1

(2R): C[C@H]1C=C[C@H]([C@@H](C)C(=O)O)CC1

  • 7.2 — D and E

D: (3R,3aS,6R,7R,7aR)-7-Iodo-3,6-dimethylhexahydrobenzofuran-2(3H)-one.

C[C@H]1[C@@H](I)[C@@H]2OC(=O)[C@H](C)[C@@H]2CC1

E: (3S,3aS,6R,7R,7aR)-7-Iodo-3,6-dimethylhexahydrobenzofuran-2(3H)-one.

C[C@H]1[C@@H](I)[C@@H]2OC(=O)[C@@H](C)[C@@H]2CC1

  • 7.3 — H and X

H: (3R,3aS,6R,7aS)-3,6-Dimethylhexahydrobenzofuran-2(3H)-one.

C[C@H]1C[C@@H]2OC(=O)[C@H](C)[C@@H]2CC1

X: Methyl vinyl ketone, or but-3-en-2-one.

C=CC(C)=O

  • 7.4 — I and J

I: (3R,3aS,6R,7R,7aS)-3,6-Dimethyl-7-[2-(2-methyl-1,3-dithiolan-2-yl)ethyl]hexahydrobenzofuran-2(3H)-one. Its side chain and the \(C_6\) methyl group are cis.

C[C@H]1[C@@H](CCC3(C)SCCS3)[C@@H]2OC(=O)[C@H](C)[C@@H]2CC1

J: (3R,3aS,6R,7S,7aS)-3,6-Dimethyl-7-[2-(2-methyl-1,3-dithiolan-2-yl)ethyl]hexahydrobenzofuran-2(3H)-one. Its side chain and the \(C_6\) methyl group are trans.

C[C@H]1[C@H](CCC3(C)SCCS3)[C@@H]2OC(=O)[C@H](C)[C@@H]2CC1

  • 7.5 — K and L

K: Methyl (2R)-2-[(1S,2S,3S,4R)-2-hydroxy-4-methyl-3-[2-(2-methyl-1,3-dithiolan-2-yl)ethyl]cyclohexyl]propanoate.

C[C@H]1[C@H](CCC2(C)SCCS2)[C@H](O)[C@H]([C@@H](C)C(=O)OC)CC1

L: Methyl (2R)-2-[(1S,3S,4R)-4-methyl-3-[2-(2-methyl-1,3-dithiolan-2-yl)ethyl]-2-oxocyclohexyl]propanoate.

C[C@H]1[C@H](CCC2(C)SCCS2)C(=O)[C@H]([C@@H](C)C(=O)OC)CC1

  • 7.6 — M and N

M: Methyl (2R)-2-[(1S,3S,4R)-2-(methoxymethylidene)-4-methyl-3-[2-(2-methyl-1,3-dithiolan-2-yl)ethyl]cyclohexyl]propanoate.

C[C@H]1[C@H](CCC2(C)SCCS2)C(=COC)[C@H]([C@@H](C)C(=O)OC)CC1

N: Methyl (2R)-2-[(1S,3S,4R)-2-(methoxymethylidene)-4-methyl-3-(3-oxobutyl)cyclohexyl]propanoate.

C[C@H]1[C@H](CCC(C)=O)C(=COC)[C@H]([C@@H](C)C(=O)OC)CC1

Solution

7.1 — Ring opening, selective hydroboration and oxidation

Opening the cyclopropane ring gives the \(p\)-menthane skeleton of A, with an internal cyclohexene double bond and an isopropenyl substituent. The specified trans relationship places the ring methyl and isopropenyl groups on opposite faces. Together with the stated \(1R\) configuration, this gives the (1R,4R) structure listed above.

To sketch A, use the cyclohexane orientation in F: put the methyl group at the top behind the page and the isopropenyl group at the opposite ring carbon in front. The internal double bond occupies the left-hand ring bond, corresponding to \(C_7=C_{7a}\) in the later lactone numbering.

Dicyclohexylborane preferentially hydroborates the less hindered, terminal double bond. Boron attaches to its terminal carbon, and oxidative workup replaces the C–B bond by C–O with retention. Thus the side-chain transformation is

\[ \ce{-C(CH3)=CH2 -> -CH(CH3)CH2OH}. \]

Hydrogen and boron add syn, but addition can create either configuration at the new side-chain stereocentre. Consequently, B is the stated mixture of two diastereomers; the existing ring stereocentres remain unchanged.

The formal net hydration balance is

\[ \underbrace{\ce{C10H16}}_{\mathbf A}+\ce{H2O} \longrightarrow \underbrace{\ce{C10H18O}}_{\mathbf B}. \]

Jones’ reagent oxidizes the primary alcohol to a carboxylic acid:

\[ \ce{-CH(CH3)CH2OH -> -CH(CH3)CO2H}. \]

A balanced equation using the supplied chromium reagent is

\[ \ce{3C10H18O + 4CrO3 + 6H2SO4 -> 3C10H16O2 + 2Cr2(SO4)3 + 9H2O}. \]

This gives C, retaining the two side-chain configurations. Counting outward from its carboxyl carbon gives

\[ \ce{CO2H-C_{\alpha}-C_{\beta}-C_{\gamma}=C_{\delta}}, \]

so the remaining ring alkene is correctly \(\gamma,\delta\)-unsaturated.

7.2 — Iodolactonization

The carboxylate attacks the nearer carbon of the iodine-activated ring double bond. This forms a five-membered \(\gamma\)-lactone, with oxygen bonded to the former \(\gamma\)-carbon and iodine bonded to the former \(\delta\)-carbon.

Opening of the iodine bridge occurs from the opposite face, so the new C–O and C–I bonds are anti. The side chain already fixes the face from which the intramolecular oxygen attack occurs.

Using the orientation of F, the required drawing has:

  • The \(C_6\) methyl group behind the page.
  • Both bridgehead hydrogens, at \(C_{3a}\) and \(C_{7a}\), behind the page: the ring fusion is cis.
  • The lactone C–O bond at \(C_{7a}\) toward the viewer.
  • Iodine at \(C_7\) behind the page.
  • The \(C_3\) methyl group toward the viewer in D, and behind the page in E.

The identity of D is fixed by the \(C_3\) stereochemistry subsequently shown in F. Thus D is the \(3R\) isomer and E the \(3S\) isomer.

For either epimer, the balanced net reaction is

\[ \ce{C10H16O2 + I2 + NaHCO3 -> C10H15IO2 + NaI + CO2 + H2O}. \]

7.3 — Radical coupling and reduction

The side chain drawn in F is

\[ \ce{-CH2CH2C(=O)CH3}. \]

It results from addition of the lactone radical to the terminal carbon of methyl vinyl ketone, X, C=CC(C)=O.

A silyl radical abstracts iodine from D, producing a radical at \(C_7\). Addition to X creates a radical adjacent to its carbonyl group; hydrogen-atom transfer from TTMSS then gives F. The approximately planar \(C_7\) radical explains why both configurations at that carbon are obtained.

Writing \(Q=\ce{(Me3Si)3Si}\), the balanced coupling is

\[ \underbrace{\ce{C10H15IO2}}_{\mathbf D} +\underbrace{\ce{C4H6O}}_{\mathbf X} +\ce{Q-H} \longrightarrow \underbrace{\ce{C14H22O3}}_{\mathbf F} +\ce{Q-I}. \]

The two forms of F are the \(7R\) and \(7S\) epimers of (3R,3aS,6R,7aS)-3,6-dimethyl-7-(3-oxobutyl)hexahydrobenzofuran-2(3H)-one:

  • \(7R\): C[C@H]1[C@@H](CCC(C)=O)[C@@H]2OC(=O)[C@H](C)[C@@H]2CC1
  • \(7S\): C[C@H]1[C@H](CCC(C)=O)[C@@H]2OC(=O)[C@H](C)[C@@H]2CC1

Direct hydrogen-atom transfer to the lactone radical instead produces H:

\[ \underbrace{\ce{C10H15IO2}}_{\mathbf D}+\ce{Q-H} \longrightarrow \underbrace{\ce{C10H16O2}}_{\mathbf H}+\ce{Q-I}. \]

Thus H is simply D with iodine replaced by hydrogen. Carbon \(C_7\) becomes a methylene group, and all other stereocentres retain their spatial configurations.

The side product G independently confirms X: it is 4-[tris(trimethylsilyl)silyl]butan-2-one,

CC(=O)CC[Si]([Si](C)(C)C)([Si](C)(C)C)[Si](C)(C)C

formed by net addition of the silane across the double bond:

\[ \ce{CH2=CHC(=O)CH3 + Q-H -> Q-CH2CH2C(=O)CH3}. \]

The change from \(7aR\) in D to \(7aS\) in H reflects a change in CIP priorities after iodine removal, not inversion at that stereocentre.

7.4 — Thioketal formation

Ethane-1,2-dithiol, SCCS, converts the side-chain ketone into a cyclic 1,3-dithiolane. Lewis-acid activation of the carbonyl allows sulfur attack, followed by ring closure and loss of water.

The side-chain change is

\[ \ce{-CH2CH2C(=O)CH3 -> -CH2CH2C(CH3)(SCH2CH2S)}, \]

where both sulfur atoms in the product are bonded to the same former carbonyl carbon, closing the five-membered ring.

The balanced condensation is

\[ \underbrace{\ce{C14H22O3}}_{\mathbf F} +\ce{HSCH2CH2SH} \longrightarrow \underbrace{\ce{C16H26O2S2}}_{\mathbf I\text{ or }\mathbf J} +\ce{H2O}. \]

The thioketal carbon is not a new stereocentre because the two paths through the unsubstituted dithiolane ring are equivalent. I and J therefore retain the \(C_7\) epimerism already present in F.

The stated relationship identifies J:

  • In J, draw the protected side chain at \(C_7\) toward the viewer, opposite the backward \(C_6\) methyl group.
  • In I, draw that side chain behind the page, on the same face as the \(C_6\) methyl group.

These are respectively the \(7S\) and \(7R\) structures given in the answers.

7.5 — Lactone opening and oxidation

Alkaline hydrolysis cleaves the lactone’s acyl–oxygen bond. It therefore gives a hydroxy carboxylate without breaking the bond between oxygen and the stereogenic ring carbon.

After acidification, diazomethane, C=[N+]=[N-], converts the carboxylic acid into its methyl ester. The net organic balance is

\[ \underbrace{\ce{C16H26O2S2}}_{\mathbf J} +\ce{H2O}+\ce{CH2N2} \longrightarrow \underbrace{\ce{C17H30O3S2}}_{\mathbf K} +\ce{N2}. \]

To draw K, open the five-membered lactone of J:

  • Replace the lactone oxygen connection at the former \(C_{7a}\) by an OH group pointing toward the viewer.
  • Replace the lactone carbonyl terminus by \(\ce{CO2Me}\).
  • Retain the configuration of the adjacent methyl-bearing side-chain carbon and all ring stereocentres.

PCC oxidizes this secondary alcohol to a ring ketone, giving L. Using \([O]\) for one formal oxygen-atom equivalent supplied by the oxidant,

\[ \underbrace{\ce{C17H30O3S2}}_{\mathbf K}+\ce{[O]} \longrightarrow \underbrace{\ce{C17H28O3S2}}_{\mathbf L}+\ce{H2O}. \]

Only the stereocentre bearing OH disappears.

The hydrogens on the two carbons flanking the new carbonyl—the former \(C_{3a}\) and \(C_7\)—both point behind the page. They are therefore cis, exactly as required by the NMR observation. No epimerization is needed to explain the result.

7.6 — Wittig reaction and thioketal deprotection

KHMDS removes a proton from the methoxymethylphosphonium salt to form (methoxymethylene)triphenylphosphorane, represented in its charge-separated form by

CO[CH-][P+](c1ccccc1)(c1ccccc1)c1ccccc1.

Its carbon attacks the ring ketone of L. Collapse of the Wittig intermediate replaces the carbonyl oxygen by \(\ce{=CHOMe}\), producing M:

\[ \ce{R2C=O + Ph3P=CHOMe -> R2C=CHOMe + Ph3P=O}. \]

The phosphorus-containing product is triphenylphosphine oxide, O=P(c1ccccc1)(c1ccccc1)c1ccccc1.

Thus, when drawing M, replace only the ring \(\ce{C=O}\) of L by an exocyclic \(\ce{C=CH-OCH3}\) group. The methyl ester, protected side-chain ketone and existing stereocentres remain as shown in L.

Mercury(II)-assisted deprotection then restores the side-chain ketone:

\[ \ce{-CH2CH2C(CH3)(SCH2CH2S) -> -CH2CH2C(=O)CH3}. \]

This gives N. The formal hydrolysis balance, before accounting for mercury binding to the sulfur-containing product, is

\[ \underbrace{\ce{C19H32O3S2}}_{\mathbf M} +\ce{H2O} \longrightarrow \underbrace{\ce{C17H28O4}}_{\mathbf N} +\ce{HSCH2CH2SH}. \]

The exocyclic enol ether and methyl ester are retained in N, ready for the final photo-oxidation and acid treatment.

The answers were checked against the official answers published by the organisers.