Inhibition of thermolysin by phosphonamidate transition-state analogs: measurement of phosphorus-31-nitrogen-15 bond lengths and chemical shifts in two enzyme-inhibitor complexes by solid-state nuclear magnetic resonance

Publisher Website
Google Scholar
Abstract
31P and15N chemical shifts and 31P-15N bond lengths have been measured with solid-state NMR techniques in two inhibitors of thermolysin, carbobenzoxy-Glyp-L-Leu-L-Ala (ZGpLA) and carbobenzoxy-L-Phep-L-Leu-L-Ala (ZFpLA), both as free lithium salts and when bound to the enzyme. Binding of both inhibitors to thermolysin results in large changes in the 31P chemical shifts. These changes are more dramatic for the tighter binding inhibitor ZFpLA, where a .apprx. 20 ppm downfield movement of the 31P isotropic chemical shift (.sigma.iso) is observed. This shift is due to changes in the shift tensor elements .sigma.11 and .sigma.22, while .sigma.33 remains essentially constant. We observed a similar pattern for ZGpLA, but only a .apprx. 5 ppm change occurs in .sigma.iso. The changes in the 15N chemical shifts for both inhibitors are small upon binding, amounting to downfield shifts of 2 and 4 ppm for ZGpLA and ZFpLA, respectively. This indicates that there are no changes in the protonation state of the 15N in either the ZFpLA- or the ZGpLA-thermolysin complex. NMR distance measurements yield a P-N bond length rP-N = 1.68 .+-. 0.03 .ANG. for the tight binding inhibitor ZFpLA both in its free lithium salt form and in its thermolysin-ZFpLA complex, a distance that is much shorter than the 1.90-.ANG. distance reported by X-ray crystallography studies [Holden et al. (1987) Biochemistry 26, 8542-8553]. For the ZGpLA inhibitor, we measure rP-N = 1.60 .+-. 0.05 .ANG., a distance which, within the experimental errors of our measurements, we consider to be similar to that of ZFpLA. Moreover, we observe no change in P-N bond length upon binding of ZGpLA to thermolysin. It is clear that the large changes in 31P chemical shift observed in the thermolysin-ZFpLA complex are not correlated with changes in either the 15N protonation state or P-N bond length. The 31P NMR data suggest that the tight binding of ZFpLA may be due to the formation of a strong bidentate complex between the inhibitor''s phosphorus oxygen atoms and the protein zinc ion, and/or to favorable hydrophobic interactions between the inhibitor and the enzyme active site.