After postdoctoral times with Prof. Baer in Toronto and Prof. R. B. Woodward at Harvard Fritz Eckstein in 1964 joined the Max-Planck-Institute for Experimental Medicine in Goettingen in the Dept. of Chemistry headed by Prof. Cramer as the Director. The focus of this Dept. was the new area of the chemistry of nucleic acids. Eckstein’s interest very soon became concentrated primarily on the phosphorothioate modification of nucleic acids but also on 2’-modifications
The first compound of this series to be synthesised was adenosine 5’-phosphorothioate. The key experiment, in 1966, was the observation that this phosphorothioate was resistant to alkaline phosphatase (8). It was later shown that this resistance holds for all phosphorothioate monoesters such as ATPaS and GTPaS as examples. This surprising result stimulated the synthesis of phosphorothioate diesters such as uridine 2’, 3’-cyclic phosphorothioate. Such diesters have a chiral phosphorus and these compounds thus exist as pairs of diastereomers. The Rp-diastereomer of this compound, whose configuration had been determined by X-ray structural analysis (25), permitted to determine the stereochemistry of cleavage by RNase A (22, 34, 39), thus establishing an in-line mechanism. This was the first example of approximately 50 enzymes involved in phosphate or nucleotidyl transfer to be characterised by the phosphorothioate approach (reviewed in ref. 151), clarifying an-in-line mechanism for most. A prerequsite for such studies is that the diastereomers of the substrates can be separated. Some examples for such studies are illustrated in fig. 1.
Of particular interest and importance are the stereospecificity of all RNA and DNA polymerases which accept only the Sp diastereomer of the phosphorothioate triphosphate as substrate such as ATPaS, resulting in the Rp diastereomer of the internucleotidic phosphorothioate (92).
A general characteristic of the phosphorothioates is their reduction in degradation, differing on the diastereomer.
One polynucleotide obtained by polymerising ATP and UTPaS (a mixture of diastereomers at the time), resulting in poly r(A-psU), showed considerable resistance to degradation by snake venom phosphodiesterase (17). As a consequence this led to a collaboration with Prof. T. Merigan and his postdoc E. De Clercq at Stanford in 1969. They were interested at the time in the stimulation of interferon. Poly r(A-U) containing phosphorothioate throughout was examined as an Interferon inducer. It showed 2- to 20-fold stimulation and most importantly cellular resistance to viral infection by 100- to 10,000-fold (26). Increase of interferon was also observed in the rabbit (33). Properties of the phosphorothioates observed in vitro thus also were valid in vivo. These results were presumed to be due to the high stability of the polynucleotide to degrading enzymes.
Inhibition of gene expression
This observation then became the basis for the antisense-methodology for the inhibition of gene expression with phosphorothioate oligonucleotides. It was first applied in 1987/1988 by an NIH group and that of Prof. P. Zamecnik after the oligonucleotide chemical synthesis by the Caruther’s method on polymer support had become available. There oxidation is replaced by addition of sulfur, resulting in a mixture of diastereomers, and has been wide-spread since. Nuclease resistance was the reason for introducing this modification but time has shown that the phosphorothioate also facilitates cellular uptake. Most therapeutic oligonucleotides, of different types, now in clinical trials are phosphorothioates. Two of these, prepared by ISIS, Fomivirsen (1998) and Kynamro (2013) have been approved by the FDA.
Cause of resistance
There is no reason from a chemical point of view that phosphorothioates should be resistant to degrading enzymes. Certainly the difference in activity of the diastereomers points to another explanation. A convincing one has been elucidated by Brautigam and Steitz in 1998 for the 3’-5’ nuclease of the Klenow fragment. Two metal ions are at the active site which are involved in catalysis. The Rp isomer of the oligonucleotide is a substrate whereas the Sp is not. X-ray structural analysis shows that in the latter one of the metal ions is displaced because of the somewhat larger van-der-Waals radius of the sulfur thus preventing catalysis. Even though this is demonstrated here for the displacement of a metal ion one can easily imagine for other enzymes a protein functional group to be dislocated and thus preventing activity. Another explanation for the inactivity of the phosphorothioates in enzymatic reactions might be that in the transition state there is no place for the sulfur but this has not been rigorously shown.
Interaction with metal ions
Phosphorothioates differ from phosphates in their coordination to metal ions. Prof. M. Cohn in 1978 showed with ATPaS that soft metal ions such as Mn2+ or Cd2+ for example coordinate to phosphorothioates whereas hard metal ions such as Mg2+ do not but coordinate of course to phosphate. This difference in property permits the identification of metal ions essential for an enzyme activity when a phosphorothioate substrate is used by switching between such two metal ions. Such analysis has been performed for a number of enzymes by several groups, e. g. for the group II ribozyme.
Phosphorothioate in bacterial DNA
The Eckstein results were instrumental in identifying phosphorothioates in bacterial DNA where they are present in the Rp configuration (315).
Working on the hammerhead ribozyme the Eckstein group demonstrated that the 2’-fluoro modification stabilises it against nuclease degradation without interfering with its activity (206). Subsequently it was shown that T7 RNA polymerase can use the triphosphate as substrate (219). These two observations made this 2’-modification an ideal one for the incorporation into aptamers and has been adopted by many colleagues for this purpose. The 2’-azido-CDP had offered a new entry into unravelling the mechanism of ribonucleotide reductase (70, 135).
The references quoted refer to the publication list.