Competitive inhibitors of primate renin (for example compound XXIV in Table 2, KT = 1-25 x 1(T6 M) were developed by Burton and coworkers by replacing one or both of the Leu residues flanking the cleavage site in pig angiotensinogen (6-13) octapeptide I (Table 1) with Phe residues10'11. Szelke and colleagues obtained potent competitive inhibitors of canine renin (IC50 = 2-3 x 108 M) by replacing the Leu 10-Leu 11 peptide bond in pig angiotensinogen (6-13) octapeptide I with a reduced isostere as in compound XXV (Table 2), that is -CH2-NH- replaces -CO-NH-12. These same workers recently reported a similarly potent, selective inhibitor of human renin, in which the reduced isostere is incorporated into the analogous human substrate sequence and extended to a decapeptide (compound XXVI, Table 2)13. The general aspartyl-protease inhibitor pepstatin, isovaleryl-Val-Val-Sta-Ala-Sta (compound XXIII, Table 2), a naturally occurring pentapeptide containing two units of the unusual amino acid statine (Sta), also inhibits renin and other acid proteases14'15. It has been proposed that the 35-hydroxyl of this internal Sta residue is an analogue of the transition-state or tetrahedral-intermediate for the enzyme reaction, thereby enabling the efficient binding to renin, even though pepstatin bears no close structural relationship to the minimal renin substrate sequence I16. While pepstatin is indeed a potent inhibitor of renin (pig kidney) with KI = 3.9 x 107 M at pH 7.3 (see below), it is much more effective against other aspartyl proteases, for example KI for pepsin is 4.6 x 1CT11 M17. Considering the extraordinary substrate specificity exhibited by renin, for which the minimum kinetically-competent substrate is an octapeptide portion (6-13) around the cleavage site (compound I, Fig. I)18, we postulated that incorporation of statine into a peptide sequence more closely resembling that of angiotensinogen would yield improved inhibitors of renin.Several analogues of renin substrate (6-13) were prepared to explore the effects of statine position and peptide chain length on inhibitory potency. It was assumed that the 35 hydroxyl group of statine mimicked some intermediate in the hydrolysis of the amide bond between Leu 10 and Leu 11 and that the isobutyl side chain of statine could be isosteric with either Leu side chain. Furthermore the rate-enhancing effects of peptide substituents at points distal to the peptide cleavage site had been demonstrated previously for both pepsin19 and renin18 so that peptide chain-length was expected to influence inhibitor potency if the statine 35-hydroxyl group formed a mechanistically derived complex with renin. Fig. 1 Renin inhibitor sequence analogy. Portions of the structures of pig renin substrate I, the substrate-based renin inhibitor II, and the naturally occurring inhibitor pepstatin XXIII. Bold portions of I and II show areas where sequence analogy between substrate and inhibitor is apparent and confirmed by structure activity relationships. The lower potency of pepstatin XXIII as a renin inhibitor may be a consequence of poor sequence analogy to I in this region. The statine residue in II is thought to serve as an analogy of the transition state in the hydrolysis of I.Table 1 lists some of the synthetic analogues used to determine the minimal sequence for potent binding to renin. All peptides were prepared either in solution or by solid phase syntheses. Analogues were fully characterized by ^nuclear magnetic resonance, microanalysis, amino acid analysis, thin-layer chromatography and high pressure liquid chromatography. Details of these syntheses will be reported elsewhere. Our initial hypothesis was that Sta could replace Leu 10 of the substrate sequence and might function as a dipeptide analogue of Leu 10-Leu 11, due to its two extra backbone atoms as compared to an a-amino acid. Heptapeptide II (Table 1) inhibits pig kidney renin with an IC50 more than 1, 000-fold lower than the KM value for the comparable octapeptide substrate I. Amino-terminal acylation gives inhibitors of comparable potency (compounds III, IV, XI) with possible increased enzymatic stability (aminopeptidase resistance). Extension from the minimally inhibitory tripeptide V to the full heptapeptide, for example V-"VI-"VII-"VIII-"II, establishes the importance of renin substrate derived substituents in positions 6-9, with a major increase in inhibition brought on by the Phe 8 residue (VI-VII). Stepwise shortening of the C-terminal portion of compound XI, as in X -> IX, establishes the importance of the two amino acid residues extending from the statine carboxyl. Further extension of the C-terminal peptide chain does not increase potency. Methyl esters, for example compound XVI, while less soluble, are slightly better inhibitors than C-terminal amides.Table 1 Inhibition of pig kidney renin by statine-containing substrate analogues Compound Structureno. 6 7 8 9 10 11 12 13 IC50(M) I His-Pro-Phe-His-Leu-Leu-Val-Tyr (5.5 xlO"5)II His-Pro-Phe-His-Sta - Leu - Phe-NH2 2.0xlO8 III Iva-His-Pro-Phe-His-Sta - Leu - Phe-NH2 3.1xlO8tIV Boc-His-Pro-Phe-His-Sta - Leu - Phe-NH2 2.7xlO8 V Sta - Leu - Phe-NH2 2.0xlO3VI His-Sta - Leu - Phe-NH2 3.7xlO4 VII Phe-His-Sta - Leu - Phe-NH2 1.3xlO6VIII Pro-Phe-His-Sta - Leu - Phe-NH2 2.0xlO7 IX Ibu-His-Pro-Phe-His-Sta-NH2 2.9xlO5X Ibu-His-Pro-Phe-His-Sta - Leu-NH2 6.7xlO7 XI Ibu-His-Pro-Phe-His-Sta - Leu - Phe-NH2 4.3xlO8XII Ibu-His-Pro-Phe-His-Sta - Ala - Phe-NH2 5.7xlO8 XIII Ibu-His-Pro-Phe-His-Sta - Val - Phe-NH2 1.2xlO7XIV Iva-His-Pro-Phe-His-Sta - He -- Phe-NH2 1.3xlO7 XV Boc-His-Pro-Phe-His-Sta - Leu - Tyr-NH2 2.6xlO8XVI Boc-His-Pro-Phe-His-Sta - Leu - Phe-OCH3 l.lxlO'8 XVII Iva-His-Pro-Phe-His-Sta-Leu-Val-Phe-NH2 4.6xlO8XVIII POA-Leu-Sta - Val - Phe-OCH3 4.5xlO6 XIX POA-Leu-Sta - Leu - Phe-OCH3 4.5 x 106XX POA-His-Sta - Leu - Phe-OCH3 2.9xlO6 XXI Iva-His-Pro-Phe-His-Leu-Sta-Val-Phe-NH2 1.5 xlO'4XXII Iva-His-Pro-Phe-His-Stat- Leu -- Phe-NH2 1.4 xlO"5 (3R)Pig kidney renin was assayed as previously described22, using a 14C-Leu labelled decapeptide substrate, except the pH was raised to pH 7.3 by use of 0.05 M citrate/phosphate buffer, 30 C. Enzyme concentration was determined to be 9 x 109 M, substrate concentration was 8.0 x 105 M, and substrate KM = 5 x 105 M. IC50 values (concentration for 50% inhibition) were determined by linear regression of logit against log concentration over a 15-90% inhibition range, using 3-6 concentrations. ICso values were found to be routinely reproducible within20%. Compounds were synthesized by standard peptide methods either by solid-phase techniques according to Merrifield 4 or by solution phase methods similar to those previously described22. The synthesis of suitably protected statine has been described25. Compounds were characterized by TLC (silica gel), HPLC (reversed phase),H NMR, amino acid analyses, mp and elemental analyses. Details of syntheses and characterization will be reported separately. Abbreviations: Iva, isovaleryl (3-methylbutyryl); Ibu, isobutyryl (2-methylpropanoyl); Boc, terf-butyloxycarbonyl; Sta, statyl ((35, 4S)-4-amino-3-hydroxy-5-methyl-heptanoyl); POA, phenoxyacetyl. The numbering scheme above the compound list is that of the pig renin substrate, the minimum kinetically competent fragment of which is I, and reflects the hypothesized analogy between inhibitors and substrate.KM value18. t 3.100.37 x 108 M (means.e. of 4 experiments using three different synthetic samples and assays on different days.(3R, 45)-4-amino-3-hydroxy-5-methyl heptanoyl, the Sta isomer epimeric at the 3-hydroxyl. Table 2 Species specificity of renin inhibitorsRenin, IC50, M Compound Structure Human Human Dog Pig Ratno. 6 7 8 9 10 11 12 13 plasma kidney plasma kidney plasma IllIva-His-Pro-Phe-His-Sta - Leu - Phe-NH2 1.6 xlO"8 1.9xlO8t 4.2xlO8 3.1xlO8 6.0xlO7 XIV Iva-His-Pro-Phe-His-Sta - He -- Phe-NH2 1.9xlO9 1.9xlO9t 3.8xlO8 1.3xlO7 7.1xlO7XXIII Iva-Val-Val-Sta - Ala - Sta 2.2xlO5 1.3xlO5t 1.3xlO6 1.0xlO6 2.6xlO5 XXIV Pro-His-Pro-Phe-His-Phe-Phe-Val-Tyr 7.0xlO6t 1.3xlO6t 7.0 xlO5 3.0xlO5 7.0xlO5xxv|| xxvm D-His-Pro-Phe-His-LeuLeu-Val-Tyr Pro-His-Pro-Phe-His-LeuVal-Ile-His-Lys 1.0 xlO"6 1.0xlO8 NRl.OxlO'8 2.4xlO8 1.0xlO5 NR NR 6.3xlO7 NR For compounds III, XIV and XXIII: Pig kidney renin was assayed as described in Table 1. Human, dog, rat, and other plasma renin assays were performed by radioimmunoassay for angiotensin I as described by Haber et a/.26 using a commercial kit (Clinical Assays), at pH 7.4 (phosphate, 37 C). The optimum angiotensinase inhibitor for assays at pH 7.4 was found to be 8-hydroxyquinoline, superior to phenylmethylsulphonyl fluoride (PMSF) in rat plasma27 and in dog plasma, and equal to PMSF in human plasma at this pH. Plasma samples (pooled) were collected on ice in EDTA from mongrel dogs, Sprague-Dawley rats, cebus monkeys, and cats. High plasma renin samples were obtained when required from animals pretreated with furosemide, hydrochlorothiazide, or by maintenance on a low-sodium diet. Lyophilized human plasma was obtained from Ortho Diagnostics (low renin) or Clinical Assays (high renin). IC50 values were obtained as described and should be considered to be reproducible to30%. The 95% confidence limits obtained by data analysis of a single experiment were routinely10%. Human kidney renin assay was performed with synthetic tetradecapeptide renin substrate by a fluorometric method28, pH 7.20 (0.10 M, citrate/phosphate), 37 C. Human kidney renin was International Standard human kidney renin29 that was further purified by affinity chromatography on pepstatin-aminohexyl-Sepharose2 . Inhibition data were consistent with competitive inhibition, and KI values were determined from linear I/V versus I/S plots23, with an estimated error of20%. Inhibition data for compounds XXIV, XXV and XXVI were obtained from refs 11, 12 and 13, respectively. Assays for these compounds were performed in similar conditions to those described above, but data comparisons should be made cautiously. NR, not reported. For abbreviations see Table 1. Additionally, 'Leu^Leu' or 'Leu^Val' indicates a reduced isosteric bond, that is -CH2NH- in place of the peptide bond -CO-NH-.The ^following additional IC50 values were obtained for plasma renin inhibition at pH 6.0 (maleate buffer): human, 1.9 x 108 M; cebus monkey, 5.3 x 109 M; mongrel dog, 1.8 x 108 M; beagle dog, 1.3 x 108 M; cat, 4.1 x 108 M; and Sprague-Dawley rat, 1.3 x 10"7 M. t JTi value, M.Estimated from data in ref. 11: 30% inhibition at 4 x 106 M; 60% inhibition at 1 x 10"5 M. Ref. 11.|| Ref. 12. f Ref. 13.Table 3 Acid protease specificity of renin inhibitors Compound KI , 109Mno. Structure Human renin Pig renin Cathepsin D PepsinXIV Iva-His-Pro-Phe-His-Sta-Ile-Phe-NH2 1.9 50 134 43 III Iva-His-Pro-Phe-His-Sta-Leu-Phe-NH2 19 10 210 27II His-Pro-Phe-His-Sta-Leu-Phe-NH2 50 6 900 40 XX POA-His-Sta-Leu-Phe-OCH3 6, 300 1, 100 3, 300 67XIX POA-Leu-Sta-Leu-Phe-OCH3 27, 000 1, 700 35 12 XXIII Iva-Val-Val-Sta-Ala-Sta 13, 000 390 0.50.05t Human kidney renin inhibition constants (Ki) were determined as described in Table 2. Pig kidney renin IC50 values were determined as described in Table 1, and KI values were calculated, assuming competitive inhibition, using the equation of Cha30: Ki = (ICs0-Et/2)(l + So/KM)1, wheret = 9xlO9M (enzyme concentration), S0 = 8.0x 105M (substrate concentration), and KM = 5xW5M. Rabbit liver cathepsin D inhibition constants were determined as previously described31, using [14C-methyl]glycinated bovine haemoglobin substrate, pH4.0 (citrate buffer), with data analysis by Dixon plots23, giving KI values. Porcine pepsin (Sigma) was assayed at pH 4.0 (formate buffer) using a heptapeptide substrate Phe-Gly-His-Phe(4-NO2)-Phe-Ala-Phe-OCH3 followed at 310nm3 . Using several inhibitor concentrations, IC50 values were calculated as for pig kidney renin assay (Table 1). Using the Cha equation30, as described above, KI values could be calculated using the values: Et = 1.2 x 108 M; S0 = 1.0 x 10"4 M; and KM = 1.25 x 105 M. Abbreviations: see Table 1.Value from ref. 33. The measured value in the assay used here is 2.5 x 109 M, likely an overestimation due to tight-binding of the inhibitor and enzyme depletion34. t Value from ref. 17. The measured value of IC50 = 5.9 x 109 M is the enzyme concentration limiting value (EJ2).The analogy between substrate I and inhibitors IX-XVII with respect to substrate positions 11 to 13 remains unclear. The inhibition data establish that addition of Leu and Phe to Sta 10 increases binding 700-fold, but it is uncertain whether these amino acids are analogous to positions 11-12 or 12-13 in renin substrate, or bind in other positions not exactly analogous to substrate. Statine can be considered to be a transition-state analogue replacing only Leu 10 (as in XVII) or as a dipeptide analogue of the transition-state complex replacing both Leu 10-Leu 11. The data presented here do not distinguish between these models. The single Leu 10 replacement hypothesis cannot be exact, since XVII is no better an inhibitor than III. The dipeptide analogy must also be imprecise since XIII is a weaker inhibitor than XI. Since position 13 is known to be required for substrate binding18, we tentatively propose that the terminal Phe in inhibitors related to III, IV and XI is bound in a position analogous to substrate position 13, and that the residue following Sta roughly corresponds to position 12. Molecular modelling studies and human substrate analogues provide additional evidence for this approximate dipeptide analogy hypothesis and will be reported separately. A similar analogy between pepstatin and pepsin substrates has been proposed based upon analysis of pepsin subsite specificity20.Tetrapeptides XVIII and XIX were prepared to test the possibility that Sta might replace Leu 11 rather than Leu 10 as in the inhibitors discussed above. These tetrapeptide analogs are now recognized to be analogous to substrate residues 8-13 (in I) with Sta replacing Leu 10, as in XX. Indeed, incorporation of Sta as a replacement for Leu 11 in the full octapeptide inhibitor XXI gives a very poor inhibitor. The identity of the renin substrate and inhibitor structures in positions 6-10 and their comparable effects on binding to renin provide strong evidence that these inhibitors act in a mechanistically related fashion. As the substrate positions 9 to 6 are filled with amino acids found at these sites in angiotensinogens, the tetrapeptide Leu-Leu-Val-Tyr (or Phe) is converted from a weak inhibitor (Ki = 3 x 104 M) to the good substrate I (KM = 5.5 x 105). In the inhibitor series these structural changes cause a 10, 000-fold decrease in dissociation constant as V is extended to II (ATi = 6x 109M for II assuming competitive inhibition). Furthermore when the chirality of the 3-hydroxyl group transition-state feature in statine is reversed (from 35 to 3R), as in heptapeptide XXII, there is a 500-fold loss in inhibition. The net gain in binding energy of II, relative to substrate I or inhibitor XXII, is greater than 4-5 kcal. This remarkable increase in binding is of the order of magnitude expected for a transition-state analogue21.Fig. 2 Inhibition of pig kidney renin pressor response in anaesthetized rats by Boc-His-Pro-Phe-His-Sta-Leu-Phe-NH2 (IV, O) and by pepstatin A, Iva-Val-Val-Sta-Ala-Sta (XXIII, o). The ability of compounds IV and XXIII to inhibit the pressor response to renin was determined in mecamylamine (1.25 mg per kg i.v.)- pretreated anaesthetized (Dial Urethane, 1 ml per kg i.p.) male rats (CRCDI strain). After mecamylamine treatment, mean blood pressure (carotid artery) fell from about 120mm Hg to about 50 mm Hg. After this pressure stabilized, hog kidney renin (same preparation used for in vitro inhibition as described in Table 1), dissolved in isotonic saline, was infused (right jugular vein) at 0.11GU per min (O.llml per min) for 5min, followed by a sustaining infusion of 0.028 GU per min. This sustaining infusion maintained blood pressure (bp) at about 120mm Hg. Injection of IV, in saline containing 0.1% acetic acid, into the left femoral vein caused a rapid (1 mg per kg, where ID50 is defined as that quantity of compound required to lower the measured blood pressure by 50% toward the pre-renin-infusion level (typically, 50 mg Hg). The ID50 for IV is 0.030 mg per kg, more than 50 times more potent on a molar basis (see Fig. 2). When blood pressure was maintained by angiotensin I, angiotensin II, or noradrenaline, neither IV nor XXIII lowered blood pressure, consistent with the antihypertensive action being due to inhibition of renin. The duration of action of the ID50 dose, defined as the time from half-maximal effect to half-recovery, is approximately 15 min for IV compared to<2 min for pepstatin XXIII. Larger doses of IV give a flat return to pre-renin-infusion levels of blood pressure, with a 2 mg per kg dose having a duration of almost 90 min. Further biological characterization of the mode of action of these inhibitors is in progress. The renin inhibitors described above are the most potent such compounds yet reported. While problems remain to be solved before medicinally useful agents result, these inhibitors possess already the inherent potency characteristic of medicinally useful compounds. Further study of the inhibition of renin in vivo using the compounds reported here should serve to elucidate the contribution of the renin system to various disease states. The design of these renin inhibitors, using abstraction of a transition-state feature from an inhibitor of a related enzyme and mechanistically related incorporation of this feature into the substrate sequence, suggests a general approach to the design of protease inhibitors.We thank Mr C. F. Homnick, Ms S. L. Fitzpatrick, Ms J. S. Murphy, Mr J.-P. Moreau, Dr D. W. Cochran, Ms B. M. Ashe, Dr M. Zimmerman and Dr B. H. Arison for analytical support, and Professor E. Haas for the gift of human kidney renin standard. We gratefully acknowledge the support and encouragement of Drs P. S. Anderson and R. F. Hirschmann.