A patent was issued this month that used “tetrabutylammonium pyrophosphate” as a reagent to convert the primary alcohol of derivatives of adenosine, thymidine, guanosine and cytidine into the corresponding triphosphate derivatives (such as a derivative of ATP in which the triphosphate is at the primary alcohol of adenosine) also using a chlorophosphine as a reactant that apparently provides the third phosphorous. Details of the reaction are shown in numerous examples in the patent, such as in columns 527 and 528 of Ju; J., Li; X., Chen; X., Li; Z., Kumar; S., Shi; S., Guo; C., Ren; J., Hsieh; M., Chien; M., Tao; C., Erturk; E., Kalachikov; S., Russo; J. (The Trustees of Columbia University in the City of New York) US Patent 11,085,076, 10-Aug-2021.
In order to understand the reaction and its stoichiometry, it would be valuable to understand the composition of tetrabutylammonium pyrophosphate since the pyrophosphate anion (PO3-O-PO3; sometimes called diphosphate) can have 1, 2, 3 or 4 negative charges and that means that, in principle, a quat pyrophosphate could contain 1, 2, 3 or 4 quat cations.
One of the reactions reports using 180 mg (0.33 mmol) of “tetrabutylammonium pyrophosphate” with 0.10 mmol of the adenosine derivative. That and other reactions in the patent suggest a molecular weight for tetrabutylammonium pyrophosphate of about 540-545 g/mole. However, tetrabutylammonium triihydrogen pyrophosphate would have a MW of 419 g/mole. Commercially available bis(tetrabutylammonium) dihydrogen pyrophosphate has a MW of 661 g/mole. Commercially available tris(tetrabutylammonium) hydrogen pyrophosphate has a MW of 902 g/mole. So, while we were hoping to provide insight into this reaction, we are not sure what is the composition of tetrabutylammonium pyrophosphate used in this (non-PTC) reaction.
The reaction was performed in DMF so it is possible that not much of the organophilic tetrabutylammonium cation is required to solubilize the pyrophosphate anion in the reaction phase.
We welcome any input any of our readers may have about this reaction, especially ideas about the composition of what the inventors used that they call tetrabutylammonium pyrophosphate.
Ok, you just made me pull out my thesis from 1999. Here’s what I have.
In 1959, Cramer and Bohm reported reaction conditions suitable for phosphorylating geraniol and farnesol to their respective diphosphates by condensation with bis(triethylammonium) phosphate, trichloroacetonitrile, and substrate alcohol in acetonitrile at rt. Reaction times were 4-24h. Diphosphate yields varied from 15-25%. (Cramer, F.; Bóhm, W. Angew Chem. 1959, 71, 775).
In 1988, Danilov modified the Cramer reaction by using Bu4NH2PO4 as a phosphate source. This reaction took 10-30 minutes, and gave ~45% yields of diphosphate for primary alcohols, and 20-30% for secondary alcohols. This was the method I used in my thesis. (Danilov, L.; Mal’tsev, S.; Shibaev, V. Soviet J. Bioorg. Chem. 1988, 14, 712).
In 1985, Poulter introduced another modification that used SN2 displacement of a tosylate or halide with tris-tetrabutylammonium pyrophosphate (ie: diphosphate) in acetonitrile. Yields were 60-90%, with no monophosphate or polyphosphate byproducts. This procedure was only useful for primary alcohols (secondary/tertiary gave elimination and/or rearrangement). (Davisson, V.; Woodside, A.; Neal, T.; Stremler, K.; Muehlbacher, M.; Poulter, C. J. Org. Chem. 1986, 51, 4768).
It looks like the patent you reference is based off of the Poulter method. My guess is that they are making the mono-quat-trihydrogen reagent themselves in the presence of 6 equivalents of water (7 after the reaction) using pyrophosphoric acid and tetrabutylammonium hydroxide. Pyrophosphoric acid is available in 94% purity. The sticky part is that this would require 73% tetrabutylammonium hydroxide (2.8 M), which does not appear to be readily available.
I hope this gives you some insight.