Printer vriendelijke versie

Target: hydroxychloroquine

28 March 2020 - Antivirals

Hydroxychloroquine is now prescribed in corona treatments in Europe even though the effectiveness of the drug as an antiviral is unsure (link). Antimalarial drug chloroquine was identified as a possible cure as early as the beginning of February but cousin hydroxychloroquine is considered less toxic. Existing users of the drug (sufferers of Lupus are mentioned) now watch in alarm as supplies rapidly dwindle away.

(RS)-2-(4-((7-chloroquinolin-4-yl)amino)pentyl(ethyl)amino)ethanol was patented in 1951 by Sterling Drug Inc of New York as an antimalarial drug but describes only two steps (link). An improved procedure was patented in 2005 by IPCA Laboratories of Mumbay (Link).

Acetyl butyrolactone is commercially available and step one is a hydrolysis / chlorination:

Step two is a ketone protection to an acetal:

With protection in place step 3 is an amination with 2-(N-ethyl)-aminoethanol:

Step 4 is the acetal deprotection step (the 110% yield is a patent text error but we may assume the yield is quantitative):

Step 5 is a reductive amination

The final step is reaction with 4,7-dichloroquinoline (also commercially available)

Key difference with the 1951 patent is the elimination of the use of phenol in the final step. The yield is higher but a reaction time of up to 50 hours is excessive by any standard. With an overall yield of 51% the production costs in terms of the two raw materials can be kept below 1 euro per gram. In a 2010 patent the final step is carried out at 20 bar of nitrogen with a reported 78% yield at 4 hours and 100°C. (Link) In a 2015 patent (Link) chloro-2 pentanone and 2-(N-ethyl)-aminoethanol are reacted directly without the need for a protecting-deproting step. Most recently in 2019 a continuous flow procedure was published (DOI).

Wanted: the Roche lysis buffer formula

27 March 2020 - Trade secrets

Pharmaceutical company Roche is taking some heat in The Netherlands after a publication by investigative platform "Follow The Money" (Link). Roche has a market share of 80% for the equipment currently used in corovirus testing (real-time PCR) and the hospitals and laboratories that use this equipment complain that Roche is unable to supply key reagents (Link, Link). Problem is that the disposables, consumables (multiwell plates, tube strips) reagents and particularly a buffer solution are tied to the supplier. If you buy the instrument from Roche you rely on them for everything else, just like a printer where ink or toner is also from the same manufacturer.

The lysis buffer is in particular demand but Roche states it is doing the best it can. At the same time, according to the laboratories, the company is also unwilling to share the recipe for the buffer solution. This recipe is a trade secret. Not true, counters Roche, you can find buffer recipes on the internet but then the laboratories complain that their homemade formulations do not really work in Roche equipment.

The issue has been debated in the The Netherlands parliament. If need be, the government should force cooperation from Roche by inventing an emergency law (link). Things are turning nasty: this morning a The Netherlands Roche facility was vandalized with "how many dead?" in sprayed graffiti. The government has asked Former DSM CEO Feike Sijbesma to look into corona testing procedures and into upscaling testing capabilities.

In the meanwhile what is this crucial lysis buffer? It breaks open cells and liberates the biomolecules such as DNA or RNA. I am going to assume here that the Roche machine is a Magna Pure 96, a "high-throughput robotic workstation for fully automated purification of nucleic acids from up to 96 samples for in vitro diagnostic use" (Link). I cannot be sure but the FDA has this page that lists the Roche MagNA Pure LC as a certified extraction tool for use in corona virus testing (real time PRC), to use with the Total NA External_lysis protocol. Then the buffer is a "MagNA Pure 96 External Lysis Buffer". Then the material datasheet lists the following compounds: guanidinium thiocyanate (30-50%), a chaotropic agent and alpha-(4-(1,1,3,3-Tetramethylbutyl)phenyl)-omega-hydroxypoly(oxy-1,2-ethanediyl) (20-25%) which is a surfactant very much like Triton X-100 (link). A datasheet is not a recipe but literature on lysis buffer formulations is impossible to find (Roche?) as it turns out. One 1990 research article here is close: dissolve 120 g of guanidinium thiocyanate in 100 mL 0.1 M Tris hydrochloride (a buffer solution), add 22 mL 0.2 M EDTA, adjust with sodium hydroxide to pH 8 and then add 2.6 g of Triton X-100.

Update! Just in! Roche will share the formula after all. The 1990 publication mentioned above WAS the "you can find the recipe on the internet" Roche recipe (Link). Dutch news organisation RTL Nieuws talked to two of the authors of the 1990 publication who also happen to be Dutch and they explain that the recipe has been refined since and they are skeptical when it comes to the "Game of Words" played by Roche. To paraphrase, "they are in it for the money".

Update 06 April 2020: New research just in with up date 2020 recipe (Link) mentioned by Paul Cotter

Update 17 April: More home brewing here, Charles Swanton of The Francis Crick Institute.

Target favipiravir

21 March 2020 - X days to Doomsday

New wonder drug Favipiravir is making the rounds in Corona land. (link) Produced by Japanese Toyama Chemical, the drug killed of the virus within 4 days in a Chinese human trial. The firm is a subsidiary of Fujifilm a rare example of obsolete photographic film maker reinventing itself as a hip antiviral drug manufacturer. (How?) The drug has been around for a while and has been found effective for other viruses such as Ebola. The first patent can be traced back to 1999! The compound is surprisingly simple: 6-Fluoro-3-hydroxypyrazine-2-carboxamide and the synthetic procedure reads like an orgchem I exercise in aromatic substitutions.

Methyl 3-amino-2-pyrazinecarboxylate is commercially available, it takes 42 euro to call 1 gram your own. The 1999 patent mentions a 1949 publication for the first bromination step to methyl 6-bromo-3-amino-2-pyrazinecarboxylate (DOI) but as my local library has closed down (quarantine) it is unavailable to me (digital access yes but only through local wifi, works in peace time but obviously not now). As a workaround a 2004 patent describing the same step will have to do (Link) using N-Bromosuccinimide:

The next step converting the amino group to a methoxy group is a Sandmeyer reaction:

Then a Buchwald-Hartwig amination with benzophenone imine:

Then reaction with ammonia converting the ester to the amide to 6-amino-3-methoxy-2-pyrazinecarboxamide:

Then the Olah reagent replaces the amino group with a fluorine group:

Finally, methoxy cleavage with trimethylsilyl chloride / sodium iodide, another George Olah invention (DOI):

The total yield is disappointing: 1 x 0.84 x 0.35 x 0.43 x 0.96 x 0.85 x 0.15 = 0.0054 x 100% = 0.54%. It would set the favipiravir raw material costs for a gram at 7777 euro. A lot of patents have appeared since, next objective is to find out where the improvements (if any) are.