Type a reaction class (ex: alkylation) or name (ex: Lossen rearrangement)

Total Synthesis proposal (batch mode) of VX-661 a CFTR modulator against Cystic fibrosis disease, actually in Phase III


Route proposal


Actually in Phase III, this molecule is planed to be manufactured commercially in a continuous flow chemistry type. More information here: Manufacturing Trends: In Continuous Mode

I didn’t find any publication about a synthetic route for this molecule, but i didn’t make a deep search. So i decided on paper to elaborate from scratch a total synthesis in a batch mode with the goal of industrial scale exploitation, with starting materials as cheap as possible, and the simpler chemistry as possible.

There is always numerous way which lead to the targeted molecule, this proposal is one among others. 

I think the building block (5: indole) and (11: cyclopropyl phenyl carboxylic acid derivative) could be manufactured by a third party, and the convergence by flow chemistry with the constraint of absence of suspension (which is not the case in this proposal) to avoids a clogging of the installation.

Total steps: 15 with 10 isolations

Update (i have finally found a document describing the route)
I have finally found the patent (CA2796642A1), here the original route with 15 steps. I prefer my version about the indol part synthesis because it is not necessary to make a cross-coupling, or use a Grignard, perchlorate and dihydrogen. This is globally the same numbers of steps (7) and starting materials cost is similar.

About the cyclopropyl moiety, their starting materials for the two versions are expensive, it would be better to use the benzodioxole and make the bromination (2nd version not showed, 1 step shorter). My version is very explorative with the lactone opening, also the cross-coupling method must be tested (i have seen in a publication an analog of (8) without the Cl, cross-coupled by a Pd catalyst, so the cross-coupling is probably tolerated by (8)).

Original route published by Vertex (15 steps)


Indole preparation

2-fluoro-1,4-dinitrobenzene (1) 

Unfortunately, this product seems to not be commercially available or difficult to find, so i have chosen to use the 3-fluoro-4-nitro-aniline and oxidize it with Oxone® to have a dinitro compound which enhance the ArSNH.

The reaction is made in water/acetone. The reference has tried the oxidation on para-deactivated aromatic ring (carboxylic acid). The nitro in para position should have a greater EWG effect than a carboxylic group, the yield could be lower.

A Mild Oxidation of Aromatic Amines, SmithKlineBeecham Pharmaceuticals, Synthetic Chemistry Department, Tetrahedron Letters,Vol. 36, No. 14, pp. 2377-2378, 1995.

To give an idea of a protocol at large scale: Practical and Efficient Large-Scale Preparation of Dimethyldioxirane, Hannes Mikula, Dennis Svatunek, Daniel Lumpi, Florian Glöcklhofer, Christian Hametner, and Johannes Fröhlich, Org. Process Res. Dev., 2013, 17 (2), pp 313–316.

ArSNH with 2-chloro-malonic diethyl ester (2)

The two more electrophilic sites are the 5-position, by 2-F and 4-NO2, and the 3-position by the same substituent, but this one suffer of a little steric hindrance.

Also, for an ortho from NO2 ArSNH, the control must be kinetic, so a strong base must be used, preferentially in THF due to the ion pairing involving a counter ion interaction with the NO2 group (but due to the molecule configuration, NMP could be also considered) and with a suitable reaction temperature to have a good selectivity.

Nucleophilic Substitution of Hydrogen in Arenes andHeteroarenes, Mieczysław Mąkosza and Krzysztof Wojciechowski, Top HeterocyclChem (2014) 37: 51–106.

Aldolization with isobutyric aldehyde (3) (one-pot possible)

Since this is a malonic derivative, conditions could be similar for the deprotonation, most probably, by using a weaker base this time like potassium carbonate and same solvent, the reaction occur easily.

Personal feelings.

Reduction of Nitro groups (4) (one-pot possible)

Since there is only mineral compounds excepting (3) and impurities, water could be added with sodium dithionite and potassium carbonate.

An attention should be taken about the isolation method, especially if NMP is used.


The reduction of aromatic nitro groups on solidsupports using sodium hydrosulfite (Na2S2O4),Tetrahedron Letters, 41, (2000), p6531-6535.

Oxidation of secondary alcohol to ketone and cyclization (5)

Oppenauer oxidation is the method of choice. But due to the proximity of the NH2 group, the aluminum catalyst maybe replaced by the couple benzophenone / potassium ter-butoxide in toluene.

Possible optimization:
If the amino group does not disturb the oxidation, an optimization could be a nitro reduction before aldolization, and the use of an excess of 2,2-dimethyl acetaldehyde to make the aldolization and oxidation (see 2nd reference). But a study must be done on the solvent to chaining the aldolization, the oxidation and imine hydrolysis followed by cyclization. Moreover, if a suitable solvent is found for the cited previous reactions, it must be tried for the ArSNH which give a 3in1 pot.

Considering the previous proposition, if the amino group is a problem, the nitro reduction could be placed as the last reaction before the cyclization with a suitable method which not reduces the ketone and the indol.

The cyclization most probably occurs naturally.Depending the behavior of the molecule towards the aluminum catalyst, some works should be done for the cyclization.

Standard Oppenauer method or

Other method which avoid a drying, but need a good wash to eliminate the residual NMP if used (attention should be taken about the catalyst, since it could be denaturized by the amine, also the process must be elaborated to reuse the catalyst):
Dehydrogenative Oxidation of Alcohols in Aqueous MediaUsing Water-Soluble and Reusable Cp*Ir Catalysts Bearing a FunctionalBipyridine Ligand, R. Kawahara, K.-i. Fujita, R. Yamaguchi, J. Am. Chem. Soc.,2012, 134, 3643-3646.

Cyclopropyl moiety preparation 

Bromation of 2,2-difluoromethylène-1,3-dioxobenzene (6)

Since the commercially available halo derivative compound is nearly ten times more expensive, again, Oxone® could be used with NH4Br as brominating agent in water/methanol mixture or 100% water or 100% methanol. These different solvent combinations should be tested for the best yield.

If the substrate is not sufficiently reactive, NBS could be the bromide source in presence of BF3-H2O.

For the coupling step, the product must be well dried.


N-Halosuccinimide/BF3-H2O, Efficient ElectrophilicHalogenating Systems for Aromatics, G. K. S. Prakash, T. Mathew, D. Hoole, P.M. Esteves, Q. Wang, G. Rasul, G. A. Olah, J. Am. Chem. Soc., 2004, 126,15770-15776.

Alkylation of 2-chloro-malonate diethyl ester (7)

The chloro ester derivatives of the lactone is not commercially available, also the carbonyl ester derivative is very expensive. Considering these facts, it is better to synthesize the lactone from the chloro malonic diethyl ester which is at an affordable price.

The alkylation could probably be made in ethanol with potassium carbonate and a PTC, or in water with a PTC too. In this last case, the conditions are “neat” since the alkylation proceeds in the 2-bromoethanol/product micro bubbles. Also the product could be extracted for the next step. This method has my preference.

Personal experience.

Lactonization of diester (8)

In this case this is a trans-esterification which could be made with catalytic amount of I2 in refluxing toluene.

To prepare the coupling step, the mixture could be treated with aqueous sodium thiosulfate and potassium iodide to reduce iodine to iodide.

A Simple and Efficient Method for Transesterificationof β-Ketoesters Catalysed by Iodine, S. P. Chavan, R. R. Kale, K. Shivasankar,S. I. Chandake, S. B. Benjamin, Synthesis, 2003, 2695-2698.

Cross-coupling reaction (9)

As in a previous “report”, i have chosen a cross-coupling reaction of a Grignard reagent with FeX3. I didn’t find any literature about FeX3 catalyzed cross-coupling reaction with a deactivated Grignard, but i think it could works.

  • In FeX3 catalyzed reaction with a Grignard, if FeCl3 is used, Tetramethyléthylene diamine must be added to avoid homo-coupling and for a yield increasing. Moreover, deactivated Grignard reagents are inactive versus common reactive functional group like ester or nitrile in these conditions.

  • When Grignard reagents are deactivated with bis[2-(N,N-dimethylamino)-ethyl] ether, they are inactive versus reactive functional groups like esters or nitriles.

  • Lastly, if Fe(acac)3 is used, there is no need of Tetramethyléthylene diamine, since the diketone act as stabilizing agent.

The in-situe generated active reagent is a not well defined “inorganic Grignard” [Fe(MgX)2]. It is doesn’t known if the stabilizing agent form a complex with the “inorganic-Grignard” or not. If not, maybe there is a stabilized organic Grignard with the Tetramethyléthylene diamine like with the bis[2-(N,N-dimethylamino)-ethyl] ether, which will then explain previous enumerated facts. But the kinetic of cross-coupling could explain the absence of reactions versus ester or nitrile, since the reaction is very fast.

This is then an explorative chemistry in this case, with reagents which meet the conditions for a good reaction: compound (7) a chloro alkyl with EWG, on one side, aryl bromide with EDG on the other side. However, cross-coupling reaction with substituted aryl Grignard seems to lead to a lower yield than with an alkyl Grignard.

The procedure could be a Grignard bis[2-(N,N-dimethylamino)-ethyl] ether complex exchange with the aryl bromide at room temp, distillation at a reduced pressure and at a constant volume with a co-solvent or reaction solvent to remove the haloalkane, add the catalyst FeCl3 or Fe(acac)3, and add dropwise the chloro lactone, or reversely, prepare the deactivated Grignard, remove the haloalkane, and add dropwise on the chloro lactone / catalyst mixture. The suitable solvent must be determined. If a mix toluene / THF don’t work, instead of using toluene, use Me-THF for the lactonization, which is non-miscible with water.


Iron-Catalyzed Reactions in Organic Synthesis, CarstenBolm,* Julien Legros, Jacques Le Paih, and Lorenzo Zani, Chem.Rev.2004,104,6217−6254. 

Decarboxylative cyclopropanation (10) (11)

This is here again an explorative chemistry about the lactone ring opening. The decarboxylative cyclopropanation is a well known method for cyclopropyl synthesis, but it is realized with keto-lactone, here, this is with an ester. The ring opening must be done via a nucleophilic substitution on the C-O by a halide, followed by a decarboxylation and a cyclization.

Conditions are met to make a carboxonium, since there is not a labile H in alpha of the carbonyl moiety.

The ring opening with chlorination and decarboxylation is made in aqueous HCl. This is where the problem is, the ester will be hydrolyzed, and consequently, the deprotonation will be much harder, if occurs. There is also another method which involve NaCl in polar aprotic solvent at high temperature.

Otherwise, there is a mild ring opening method for aminolysis using sodium 2-ethyl-hexanoate in THF.

By combining these methods, maybe the desired reaction could occur, with 2-ethyl-hexanoïc acid (or PTSA ?), potassium or sodium chloride, a polar solvent and a PTC. THF does not probably have an enough high boiling point, MeTHF or dimethoxyethane. Once the lactone is opened, if not decarboxylated, make the decarboxylation by adding a strong organic acid. Once the decarboxylation is done, make the cyclization with a suitable base (K2CO3 / CTP ?), and finalize by saponification with aqueous KOH

There is another way, less explorative but with less friendly reagents, to obtain this molecule (see below):
(16) most probably gives regioisomere which must be separated. 
(17) process to avoid a self condensation and a Canizzaro 
(19) process to remove the phosphine.

Some one pot could be made: 2in1 (16) and (17) / 2*2in1 or other combination from (18) to (11). 

There is an optimization on this route shortening the synthesis of 2 steps with a Friedel & Craft to obtain directly (18) from the benzodioxole and ethyl oxalate chloride. In the cited reference benzodioxole was used. Instead of using AlCl3, Zeolite could be tried. Also, to remove the phosphine oxide, and if it does not perturb the Johnson-Corey- Chaykovsky reaction, and not be a problem during the saponification, it could be removed by treatment with oxalyl chloride, which could serve also to prepare the acid chloride by a Vilsmeier-Haak.

A mild method forring-opening aminolysis of lactones, Process R&D, Chemical and AnalyticalDevelopment, Novartis Institute for Biomedical Research, 59 Route 10, EastHanover, NJ 07936, USA, Tetrahedron Letters, 42 (2001) 2439-2441.

A novel synthesis of cyclopropyl ketones via decarboxylative ring contractions of a-acyl-y-butyrolactones catalyzed by halide ions in dipolar aprotic solvents, Saburo Takei* and

Wittig reaction: Isoflavones:Chemistry, Analysis, Function and Effects page 64-65.

Johnson–Corey–Chaykovsky reaction: Twenty-fiveyears of dimethylsulfoxonium ethylide (corey's reagent), Yu.G. Gololobov, A.N.Nesmeyanov V.P. lysenko, I.E. Boldeskul, Volume 43, Issue 12, 1987, Pages2609–2651.

Reference for optimized alternate route:
Friedel & Craft: Reliable and Versatile Synthesis of 2-Aryl-Substituted Cinnamic Acid Esters, A. Ianni, S. R. Waldvogel, Synthesis, 2006, 2103-2112.

Acetylation of aromatics over acid zeolites: Seeking a viable alternative to Friedel–Craftscatalysts*, Matteo Guidotti1, Jean-Marie Coustard, Patrick Magnoux and Michel Guisnet, Pure Appl. Chem., Vol. 79, No. 11, pp. 1833–1838, 2007.

Acylation of 2-Methoxynaphthalene over Ion-Exchanged Beta Zeolite, İsmail Cem Kantarli, Dissertation Msc, 2002.

Catalytic activity of the beta zeolite with enhanced textural  properties in the Friedel-Crafts  acylation of aromatic compounds, R.A. García*, D.P. Serrano, G. Vicente, D. Otero and M. Linares, Studies in Surface Science and Catalysis Volume 174, Part 2, 2008, Pages 1091-1094.

Zeolite coated structures for the acylation of aromatics, A.E.W Beers, T.A. Nijhuis, F. Kapteijn, J.A. Moulijin, Microporous and Mesoporous Materials, 48, 2001, 279-284.

Phosphine oxide removal: A convenient and mild chromatography-free method for the purification of the products of Wittig and Appel reactions, Peter A. Byrne, Kamalraj V. Rajendran, Jimmy Muldoona and Declan G. Gilheany, Org. Biomol. Chem., 2012,10, 3531-3537.


Amidification (12) 

Activation of the acid in THF or Me-THF or dimethoxyethane with ethyl chloroformate at 0°C or below and add the compound (5). The base must not to be too strong to avoid an activation of the N from the indol by a deprotonation. Another activation method could be with the carbonyldiimidazole. 

Personal experience 

Alkylation by epichlorhydrine (13) (one-pot possible depending the solvent) 

THF and heating or K2CO3 with PTC or a stronger base and heating.

For the pKa of the base which could be used:
Enhanced imine synthesis in water: from surfactant mediatedcatalysis to host guest mechanisms, Kamel Meguellati, Ali Fallah-Araghi, Jean-ChristopheBaret, Abdeslam El Harrak, Thomas Mangeat, Carlos M. Marques, Andrew D.Griffiths and Sylvain Ladame, Chem.Commun., 2013, 49, 11332. 

Hydroxymethylation (14)

Probably, in a mix miscible water solvent (THF (?))/water (water brought by the formaldehyde solution) by activating the formaldehyde with LiBr and / or weak acid as source of proton to avoid the presence of OH- and a premature decarboxylation depending the conditions.

Alternate method:
Fromylation with DMF/(COCl)2 and reduction to alcohol with sodium dithionite.

Reference for a draft of procedure:

There is another publication, but i cannot remember the reference, where EtOH was used and heating to reflux.

References for alternate method:

Reduction of Aldehydes and Ketones by SodiumDithionite, Johannes G. de Vries and Richard M. Kellogg*, J. Org. Chem.1980,45,4126-4129.

Final decarboxylation and nucleophilic substitution (15) (one pot possible)

By adding water, KOH and heating.

Science of Synthesis: Houben-Weyl Methods of MolecularTransformations Vol 10, Hetarenes with one heteroatom, p581.

Costing with starting material only: 948,064/kg (1609.34$/mol) cheaper for best of each route

Original route 2678.67$/kg (404.47$/mol)

Alternate route 2569.22$/kg (366.76$/mol)

Cyclopropyl moiety
Original route Ver.1a 4022.23$/kg (1597.53$/mol)

Original route Ver.1b 1883.18$/kg (400.2$/mol)

Alternate route Ver.1 1659,60$/kg (297.12$/mol))

Alternate route Ver.2a 1152.51$/kg (182$/mol)

Alternate route Ver.2b 1044.57$/kg (165.06$/mol)

This is some personal works on paper only, i have no responsibility in any way if somebody would try this route and has all sort of troubles, including but not limited to: injuries and money loss. This is for experienced chemists only, and tests must be conducted in a suitable lab only.

But if my work is used to synthesize the targeted molecule described here, please, send a word, even if it fails, chemistry is always an experimental science. This will make me pleased, thank you.

© David Le Borgne, 2015, specialist in chemical process development and optimization.


  1. Update: I have finally found a patent describing the route used: CA2796642A1 (http://www.google.is/patents/CA2796642A1?cl=en)

    Honestly, about the indole route, i prefer mine.

  2. Update: Added the route published by Vertex and improved the readability

  3. Update: Added an optimized alternate route for the cyclopropyl moiety, shortening the route by 2 steps with a Friedel&Craft on benzodioxole derivative.

  4. Update: Added price comparison between the two routes.