The serpins are a family of proteins of special interest to structural biologists because of their ability to spontaneously undergo a profound change in conformation.1 This conformational transition reflects the prime role of the serpins as protease inhibitors but it has also been adapted in individual serpins to allow the modulation of their function.2 A series of structures over the last 20 years has shown how this flexibility enables the serpins to act as controlling factors in key mechanisms in life, including coagulation, tissue growth, the immune response and hormone release. A particular structural challenge in this last group has been that of the most rudimentary member of the family, angiotensinogen, which is the initiating source of the angiotensin peptide that constricts blood vessels and hence principally controls blood pressure. The recent solving of the structure of angiotensinogen together with that of the complex it forms with renin shows that angiotensinogen is not, as had been believed, a passive source of the angiotensin peptide but rather that it interacts conformationally with renin to expose the otherwise inaccessible peptide cleavage site. The surprise finding is the dependence of this conformational change on a labile disulphide bridge in angiotensinogen that acts as a redox switch controlling the efficiency of angiotensin release. The finding opens an unexpected insight into the causation of high blood pressure and notably of pre-eclampsia - the hypertensive complication of pregnancy that threatens the health and survival of mother and child.
The crystallisation of angiotensinogen and the solving of its structure was the culmination of a twenty-year effort. Eventual success required the recombinant production of rat and mouse angiotensinogens as well as that of the human. Crystals were obtained for each with the collection of datasets at up to 2.1Å resolution at the Daresbury and Diamond synchrotron light sources. Solving the structures provided a daunting test of the advanced molecular replacement algorithms in Phaser,3 because there were no close homologues of known structure. A weak molecular replacement solution for human angiotensinogen, supplemented by a low resolution GdCl3 derivative, gave electron density that could be used to solve the C2 crystal form of rat angiotensinogen. After 2-fold averaging, electron density from the C2 crystal form of rat angiotensinogen was then used to solve the structure of the P3221 crystal form of the same protein. The electron density map sufficient for a reasonable model to be built was then obtained by 4-fold multicrystal averaging across the two crystal forms. Finally, the mouse angiotensinogen structure was solved by molecular replacement, placing four copies of the model of the rat protein. The endpoint was a high-resolution structure of human angiotensinogen (Fig. 1) showing that angiotensinogen essentially retains the serpin fold, with the 61-residue amino-terminal extension containing the angiotensin sequence forming an ordered superstructure.
Figure 1: Angiotensinogen and its complex with renin: Showing (left) the superstructure to the serpin framework of angiotensinogen in blue and the amino-terminal angiotensin peptide in green, with the buried cleavage site arrowed red. Interaction with renin, shown right, displaces the protruding CD peptide loop (yellow) with the concerted movement of the angiotensin-containing terminal into the active-site cleft of renin.
As is often so, the solving of the new structure provided both a surprise and posed a question. The surprise was that the terminal angiotensin sequence was not freely exposed as expected but rather was firmly bound to the body of the molecule with the critical peptide-cleavage site being inaccessibly buried. The question was: how does renin gain access to this cleavage site? The answer to this became clear with the crystallisation and solving of the structure of the initiating complex formed by human angiotensinogen with recombinantly inactivated renin. Although at low resolution, the complex clearly shows that the interaction with renin results in two major changes in angiotensinogen. The complementary binding of renin necessitates a 10Å displacement of a protruding loop on angiotensinogen with an accompanying 10-20Å extension and movement of the N-terminal substrate peptide of angiotensinogen into the active site cleft of renin. The two concerted movements being critically linked by a disulphide bridge that is conserved in all known angiotensinogens.
Figure 2: The redox-switch: The complex of renin (left) with angiotensinogen (shadowed), highlighting the labile (18-138) disulphide bond that bridges the two mobile elements: the CD loop in yellow and the amino-terminus in blue/green.
Previous work from our group had focused attention on this conserved 18-138 disulphide bridge, which the structure now showed to be externally sited. Confirmation that this is a labile disulphide bond was demonstrated by the incubation of angiotensinogen over a range of physiological redox-potentials. Unequivocal evidence of the presence in the circulation of the reduced unbridged-form as well as the oxidised bridged-form came from the examination of a series of normal plasma samples, showing the consistent presence of near 40% of angiotensinogen in the reduced form, apparently independent of age or gender. This finding immediately suggested that this potential redox-switch, between the bridged and unbridged forms, could act as a control mechanism modulating the release of angiotensin. To our puzzlement however, the differences in efficiency of release of angiotensin between the two forms was relatively small. A breakthrough in understanding came from previous studies indicating that physiologically the cleavage of angiotensinogen principally occurs with renin bound to a cell-membrane protein, the prorenin receptor.4 When the assays of angiotensin release by renin were repeated in the presence of the prorenin receptor a clear loss of activity was evident in the reduced unbridged-form, confirming that the 18-138 disulphide could indeed act as a redox-switch capable of fine-tuning the efficiency of release of angiotensin.
Our corroboration of the significance of the receptor-bound activation of renin is in keeping with the growing realisation that the critical release of angiotensin takes place at a cellular level in renal and other tissues rather than in the circulation. In particular, our identification of an activating redox-switch fitted with a body of observations by others indicating that oxidative stress can be a contributory cause to the onset of hypertension. Our suspicion that increases in blood pressure could result from the oxidative conversion of angiotensinogen to its more active bridged form was difficult to confirm, as changes taking place at a focal cellular level are likely to have little effect on the pooled circulating angiotensinogen. An exception however, is the more overt oxidative stress resulting from placental dysfunction. This is associated with the increase in blood pressure that in 2-7% of all pregnancies leads on to pre-eclampsia - the hypertensive crisis of pregnancy responsible globally for an estimated 50,000 maternal and 500,000 infant deaths each year. To test at a pilot level whether conversion of angiotensinogen to its more active oxidised form occurred in pre-eclampsia, 24 plasma samples provided by Professor Broughton Pipkin in Nottingham were examined - 12 from pre-eclamptics and 12 from matched normal pregnancies. Although mixed and with blinded-identifiers, electrophoresis subsequent to SH-labelling showed clear differences between the two cohorts, with some of the pre-eclamptic samples being immediately recognisable by a marked decrease in reduced as compared to oxidised angiotensinogen. Support for the inference from these findings, that the oxidative activation of angiotensinogen could be a contributory cause of pre-eclamptic hypertension, came from the previous observation by others of a similar activation of angiotensin-release in a patient with pre-eclampsia due to a unique mutation at the renin-cleavage site in angiotensinogen.
Our overall findings illustrate the way that the determination of a structure can underpin medical understandings2. There have been altogether some 95,000 papers centred on studies of angiotensin including 1700 reviews. Yet a starting point for this whole field has been the misapprehension that angiotensinogen, the core source of angiotensin, is a passive substrate - represented diagrammatically as a blob with the angiotensin peptide dangling from it! Our prime finding here is the demonstration of the active and precise contribution angiotensinogen makes to the release mechanism. Added to this, the structural findings immediately open insights not only to the control of angiotensin release but also to aberrations that lead to disease.
Zhou, A., Carrell, R.W., Murphy, M.P., Wei, Z, Yan, Y, Stanley, P.L., Stein, P.E., Broughton Pipkin, F., Read, R.J. A redox switch in angiotensinogen modulates angiotensin release. Nature. 468,108-11(2010)
The British Heart Foundation and the Wellcome Trust
Diamond Light Source is the UK's national synchrotron science facility, located at the Harwell Science and Innovation Campus in Oxfordshire.
Copyright © 2017 Diamond Light Source
Diamond Light Source Ltd
Harwell Science & Innovation Campus