Using Epox to Cure a Pox

Whilst researching material for my recent webinar on the synthesis of drugs approved by the FDA in 2024,¹ I came across an interesting reaction used in the production of the broad-spectrum, 5^th generation, cephalosporin-based anti-bacterial agent ceftobiprole medocaril (Figure 1).² This iv administered prodrug was approved in April 2024 for the treatment of hospital-acquired/community pneumonia and complex skin/soft tissue infections. The active metabolite (ceftobiprole) works by targeting penicillin-binding proteins, inhibiting the transpeptidase activity essential for the synthesis of bacterial cell walls.³ Hydrolysis of ceftobiprole medocaril releases the active metabolite ceftobiprole (Figure 1).

The synthetic step I found particularly interesting is the Wittig reaction used to prepare the penultimate olefin intermediate (A, Figure 2). Early approaches to (A) used fairly standard Wittig chemistry- generation of a weakly-stabilised ylide from the phosphonium salt using a strong base (t-BuOK) at cryogenic temperatures (-78°C, DCM/THF/toluene).⁴ The aldehyde was prepared in two steps from commercially available 7-ACA (Figure 2), the latter being obtained from enzymatic (amide) hydrolysis of natural Cephlosporin C.⁵ For those interested in total synthesis, Cephlosporin C was first synthesised from L-cysteine by R. B. Woodward in 1965, and was a significant topic in his Nobel lecture on the art of chemical synthesis.⁶ The second component- the phosphonium bromide ylide precursor- was ultimately derived from protected (R)-asparigine.⁷

The totally unique approach, described in a 2015 US patent filed by Sandoz, enables the use of fully “unprotected” aldehyde (free amine and carboxylic acid) and generation of an ylide under non-basic conditions using a reagent combination I’d never seen before- a mixture of bis(trimethylsilyl)acetamide (BSA) and propylene oxide.⁸ BSA generates the trimethylsilyl ester of the carboxylic acid- an in-situ protection strategy. The amino group is also temporarily protected. The role of propylene oxide is less clear. It turns out that a little known phosphonium salt activation method, reported by Buddrus in 1968, uses an epoxide to generate an active phosphorane intermediate. Mechanistically this occurs by equilibrium dehydrohalogenation of the phosphonium salt (Figure 3).⁹ A US patent on “a process for preparing alkylidine phosphoranes” was filed in 1972.^9b

Interestingly, a patent filed by Biochemie Gesellschaft in 1995 describes olefination using very similar conditions.¹⁴ In fact the Sandoz parent refers explicitly to this document in their preamble- “suitable reaction conditions and solvent systems are described, for example, in WO95/29182 on page 26, second paragraph to page 27, to the last-but-one paragraph inclusive, and in Example 31 on page42 in WO95/29182”.

From an operational and sustainability perspective the fact that the reaction could be run at 1°C is a major advantage. Cooling a large vessel to cryogenic temperatures requires a lot of cooling. I assume the transient silicon protecting groups are hydrolysed during the work-up. The resulting silicate waste, although relatively benign, would need extracting or filtering in some way. Notwithstanding the generation of alkyl bromide waste (unless it re-generates epoxide in situ), this is a very interesting and potentially highly practical approach.

A citation search on the Buddrus paper returned only 8 hits. Perhaps this is an overlooked trick that could be very useful for base sensitive substrates. I would love to hear if anyone has applied this in their work. From a scale-up perspective it looks attractive. Perhaps one issue- possibly a major issue- is that it requires the use of the propylene oxide (and the silylating reagent) in large excess. Propylene oxide is used as a co-solvent in the process described in the Sandoz patent.⁸ I don’t have much of an appreciation of the properties this material, so I did some digging.

Described back in the 1860’s by Linnemann as a precursor to acetone, propylene oxide (PO) is a highly flammable, acutely toxic, volatile liquid (b.p. 34°C).¹⁰ Mixtures of PO and air form flammable mixtures with a lower explosive limit of 2.3%. Handling large volumes of PO requires high containment closed systems and may require a dedicated high potency facility, again perhaps a limitation to the scale-up of the process described above.¹¹ Heating the epoxide results in rapid decomposition via self-polymerisation.

Global production exceeds 6 million tons annually, most of which was used in the manufacture of polyether polyols (precursor to polyurethane foam), propylene glycol, and glycol ethers.¹² It’s application in the pharmaceutical industry is limited. The vast majority of PO is produced in racemic form. Volatility and high water-solubility (425g/L at 20°C) facilitates removal from process streams vide supra. PO hydrolyses in water with a half-life of 30 days.

As is common in the commodity chemical industry, production methods are designed to maximise atom efficiency. By-products are valorised to equally valuable industrial materials.¹² The original process, still run industrially, converts propene to the chlorohydrin with chlorine/water followed by conversion to epoxide using calcium (or potassium) hydroxide (Figure 4). The by-product (calcium chloride) is a stochiometric waste product with no real monetary value. A newer process (SMPO) utilises oxygen gas in benzylic per-oxidation of ethylbenzene and reaction of the hydroperoxide with propene. The phenethyl alcohol by-product is a precursor to styrene (via dehydration)- a valuable commodity material used in the polymer industry (Figure 4). In a similar fashion, using 2-methylpropane as the oxidation substrate, t-butanol is the by-product, this being a valuable solvent, and a precursor to another solvent- MTBE. A newer PO manufacturing process uses titanium-catalysed oxidation of propene with hydrogen peroxide (Figure 4).¹² An electrochemical process is also being developed.¹³

Going back to the propylene oxide/BSA reaction, I think these olefination conditions are interesting for a couple of reasons. Firstly, I believed the Wittig reaction had no more secrets to be uncovered and secondly, that generating a poorly stabilised phosphorous ylides without using stochiometric strong bases at mild temperatures is even possible. Add in the romance of re-vitalising an approach that hasn’t seen light of day for 60 years and is a lovely story to uncover.

Finally, for the purists out there, I know that pox relates to a viral not a bacterial infection and that ceftobiprole has no standing here. However, for the sake of something catchy and clever, the title remains.

See you next time.

References:

1. Synthesis of the new drugs approved in 2024: J. Studley, Jan. 2025, https://www.scientificupdate.com/webinar_events/synthetic-approaches-to-the-new-drugs-approved-in-2024/
2. Ceftobiprole medocaril: a new fifth-generation cephalosporin: P. Kale-Pradhan et al, Annals of Pharmacotherapy. 2024; doi:1177/10600280241293773
3. Penicillin binding proteins: key players in bacterial cell cycle and drug resistance processes: Dessen et al, FEMS Microbiology Reviews, 2006, 30, 673–691
4. Patent review of manufacturing routes to fifth-generation cephalosporin drugs. part 2, ceftaroline fosamil and ceftobiprole medocaril: D. Hughes Process Res. Dev. 2017, 21, 800-815; WO2001090111 (Basilea Pharmaceutica AG)
5. Cephalosporin C biosynthesis and fermentation in acremonium chrysogenum: L. Liu et al, Appl Microbiol Biotechnol. 2022, 106, 6413-6426
6. The total synthesis of cephalosporin C: R. B. Woodward et al, J. Am. Chem. Coc. 1966, 88, 852-853
7. US7511154; EP1067131; US6232306 (Basilea Pharmaceutica AG)
8. US9096610 (Sandoz AG)
9. Wittig reaction with the aid of ethylene oxide: a) J. Buddrus, Chem. Int. Ed. 1968, 7, 536-537; b) US3634518
10. Propylene oxide: H. Baer et al, Ullmann’s Encyclopedia of Industrial Chemistry 2012, https://doi.org/10.1002/14356007.a22_239.pub3
11. Carcinogenic effects of exposure to propylene oxide: https://www.cdc.gov/niosh/docs/89-111/
12. Efficient industrial organic synthesis and the principles of green chemistry: T. Schaub, Eur. J.2021, 27, 1865-1869
13. Electrification of the chemical industry- performing chemical synthesis with renewable electricity can reduce carbon emissions: J. Barton, Science 2020, 368, 1181-1182
14. WO9529182 (Biochemie Gesellschaft)