Novel Mode of Action of Angiotensin I Converting Enzyme Inhibitors

Angiotensin I converting enzyme (kininase II; ACE) inhibitors are important therapeutic agents widely used for treatment in cardiovascular and renal diseases. They inhibit angiotensin II release and bradykinin inactivation; these actions do not explain completely the clinical benefits. We found that enalaprilat and other ACE inhibitors in nanomolar concentrations activate human bradykinin B1 receptors directly in the absence of ACE and the B1 agonist des-Arg10-Lys1-bradykinin. These inhibitors activate at the Zn2+-binding consensus sequence HEXXH (195–199) in B1, which is present also in ACE but not in the B2 receptor. Activation elevates [Ca2+] i and releases NO from endothelial or transfected cells expressing the B1 receptor but is blocked by Ca-EDTA, a B1 receptor antagonist, the synthetic undecapeptide sequence (192–202) of B1, and the mutagenesis of His195 to Ala195. Except for the B1 antagonist, these agents and manipulations did not block activation by a peptide ligand. Thus, Zn2+ is essential for B1 receptor activation by ACE inhibitors at the zinc-binding consensus sequence. Ischemia or cytokines induce abundant B1 receptor expression. B1 receptor activation by ACE inhibitors, a novel mode of action reported here first, can contribute to their therapeutic effects by releasing NO in the heart and to some side effects.

Angiotensin I converting enzyme inhibitors (ACEIs) 1 are used for treatment in conditions such as hypertension, congestive heart failure, diabetic nephropathy, and others (1)(2)(3)(4)(5)(6)(7)(8). For instance, the administration of an ACEI after a myocardial infarction and in the absence of any thrombolytic therapy reduced the incidence of death or the development of severe congestive heart failure (3,4). ACEIs were also reported to inhibit neointima formation after endothelial injury (7). Despite the beneficial effects of ACEI therapy proven in many millions of patients world-wide, the modes of action of ACEIs have not been fully characterized (5). The inhibition of ACE or kininase II blocks angiotensin II release or bradykinin (BK) inactivation (9 -11), but these actions alone do not fully explain their effectiveness (5). ACEIs also potentiate the effects of BK and its ACE-resistant analogs on their B 2 receptor by inducing an enzyme/receptor protein-protein interaction (12)(13)(14), a heterodimer formation (14).
Of the two BK receptors B 1 and B 2 , B 2 is widely expressed and primarily mediates the actions of kinins under physiological conditions (11). Normally, few cell types express the B 1 receptor, but various pathologic conditions such as ischemia, atheromatous disease, or exposure to inflammatory cytokines rapidly induce expression (15,16). The elimination of the B 2 receptor gene in knockout mice also up-regulated the B 1 receptor (17).
The ligands of the two receptors differ, as plasma carboxypeptidase N or tissue carboxypeptidase M cleave the Cterminal Arg of the B 2 receptor agonists kallidin (Lys-BK) and BK to generate B 1 agonists des-Arg-kinins (18). Of the products, des-Arg 10 -kallidin (des-Arg 10 -Lys 1 -bradykinin) is about three orders of magnitude more potent than des-Arg 9 -BK on the B 1 receptor (15).
The contributions of the kinin B 2 receptor to the effects of ACEIs have been established (10,13), and a possible but not previously explored role for the B 1 receptor could be deduced from the fact that many patients treated with ACEIs suffer from conditions that lead to B 1 -receptor induction (15).
Although ACEIs do not directly affect the BK B 2 receptor, they augment its function when ACE is also expressed on the cell surface (12)(13)(14). Here we report that ACEIs in nanomolar concentrations directly activate the BK B 1 receptor in cells without an intermediate peptide ligand and in the absence of ACE. We also characterized the site and the mode of action of ACEIs on the human B 1 receptor, resulting in the generation of nitric oxide (NO).
CHO cells were stably transfected with cDNA of the human B 1 receptor inserted into pcDNA3 (donated by Dr. F. Leeb-Lundberg of the University of Texas, San Antonio) using SuperFect essentially as described for the human B 2 receptor (14). HEK 293 or COS-7 cells were transiently transfected with the wild type or H195A mutant human B 1 receptor using SuperFect or LipofectAMINE 2000 as described by the manufacturer (Qiagen, Valencia, CA and Invitrogen). Experiments were done 24 h after transfection.
Membrane Preparation-HEK 293 cells were transiently transfected with the wild type cDNA of the human B 1 receptor, as above, and the plasma membrane fraction was obtained with a slight modification of the technique in Ref. 20.
Radioligand Binding-Binding assays were performed at room temperature with [ 3 H]des-Arg 10 -kallidin in the presence and absence of enalaprilat concentrations ranging from 0.1 nM to 10 M (modified from Ref. 20).
Use of Inhibitors-(Ethylenedinitrilo)tetraacetic acid dicalcium salt (Ca-EDTA) was added to the cells for 30 min at 1 mM concentration to bind Zn 2ϩ , then cells were washed with zinc-free medium. The undecapeptide (LLPHEAWHFAR) was synthesized by the Protein Sciences Facility (University of Illinois, Champaign) and tested as an inhibitor (10 or 100 M) after 20 min of preequilibration.
Site-directed Mutagenesis-The H195A mutation of the human B 1 receptor was done by the PCR method using a QuikChange site-directed mutagenesis kit from Stratagene (La Jolla, CA). The human wild type B 1 pcDNA3 was used as the template with the two mutagenic primers shown in the following sequences: B1HA 1 , 5Ј-CTGCTCCTCCCCGCTG-AGGCCTGGCACTTT and B1HA 2 , 5Ј-GTGCCAGGCCTCAGCGGGGA-GGAGCAGGAT. The sequence of the construct was confirmed by automatic sequencing at the DNA core facility of the University of Illinois, Chicago.
Detection of Nitric Oxide-NO was measured using either a porphyrinic microsensor (21, 22) or with a fluorescence assay (23) in BPAE cells. The microsensor consists of carbon fibers which are electroplated with a highly conductive polymeric porphyrin to facilitate the electron transfer on or from NO to the sensor. Cells were preincubated for a few minutes at 37°C until a stable baseline was established. Ligands were added, and the responses (current versus time) were recorded continuously. Current generated on porphyrinic electrode was proportional to the NO released and was quantitated with a known standard NO solution.
For the fluorescence assay we used diamino fluorescein diacetate. Cellular fluorescence was measured at 515 nm following excitation at 490 nm (23).
Statistics-Statistical analysis was performed using one-way analysis of variance test.

Effect of Enalaprilat on IMR-90 Human Fetal
Lung Fibroblasts-To investigate whether ACEIs can activate the BK B 1 receptor, we tested them on IMR-90 fibroblasts that constitutively express both the B 1 and B 2 receptors. As a prototype of an active ACEI we used enalaprilat and measured the increase in [Ca 2ϩ ] i from cells (Fig. 1). The receptors were activated with either BK for the B 2 receptor or des-Arg 10 -kallidin as a ligand of the B 1 receptor. BK (10 nM) and des-Arg 10 -kallidin (10 nM) raised [Ca 2ϩ ] i in distinctly different patterns (Fig. 1, A and B). B 1 receptor activation led to a very prolonged, sustained increase in [Ca 2ϩ ] i , whereas BK stimulated a more transient response. Enalaprilat (1 nM) in the absence of a peptide agonist significantly enhanced the [Ca 2ϩ ] i level (Fig. 1C). Enalaprilat mediated a response which clearly differed temporally from the one caused by BK, but was very similar to that produced by des-Arg 10 -kallidin. These experiments indicated that enalaprilat in nanomolar concentrations directly activated the B 1 receptor within seconds in the absence of added kinins. Using B 1 and B 2 receptor antagonists, des-Arg 10 -Leu 9 -kallidin for B 1 and HOE 140 for B 2 (Fig. 1, D and E), we confirmed this conclusion because the response to enalaprilat was completely inhibited by the B 1 antagonist, whereas HOE 140 had no effect.
Effect of Enalaprilat on the Transfected Human B 1 Receptor in CHO Cells-The data with IMR-90 cells strongly suggested that the B 1 receptor was mediating the direct effect of enalaprilat. However, because IMR-90 cells have low but detectable ACE activity, we had to exclude the possibility that ACE is required for the activation of the B 1 receptor, similar to the B 2 receptor (12). Therefore, CHO cells, having no ACE (12), were transfected to express human B 1 receptors (CHO/B 1 ). Adding enalaprilat (10 nM) caused immediately, within seconds, a typical B 1 receptor agonist response as indicated by the prolonged shape of the elevated [Ca 2ϩ ] i curve (Fig. 2, A and B). As control, we also tested transfected CHO cells expressing ACE only (CHO/ACE) or the B 2 receptor (CHO/B 2 ) with negative results; thus, enalaprilat, in the absence of the B 1 receptor, was inactive in these cells (Fig. 2, C-E).
Enalaprilat Stimulates Nitric Oxide Release in Endothelial Cells-Enalaprilat also elevated [Ca 2ϩ ] i in BPAE cells, which constitutively express both the B 1 and the B 2 receptor (24) (not shown). We determined whether the enalaprilat-induced elevation of [Ca 2ϩ ] i levels would stimulate NO production. We monitored the release of NO from cultured BPAE cells using either a fluorescence assay or an electrochemical method with a porphyrinic microsensor (see "Materials and Methods"). Enalaprilat (10 nM) did indeed stimulate NO release, and this response was more prolonged than the transient one that followed B 2 receptor activation (Fig. 3A). Its pattern resembled the prolonged elevation of [Ca 2ϩ ] i (Fig. 1C) and mimicked the response of the cells to the B 1 agonist des-Arg 10 -kallidin (Fig.  3A). The dose-response curve showed that enalaprilat in log M concentrations linearly enhanced the release of NO from 10 nM to 10 M (Fig. 3B). We repeated these experiments with another technique to measure NO release based on fluorescence detection and obtained similar results (not shown).
Enalaprilat stimulated NO production by BPAE cells via the B 1 receptor. This conclusion is based on findings that enalaprilat and des-Arg 10 -kallidin released NO similarly (242 nM Ϯ 5 S.E. per well and 288 Ϯ 7 nM NO), and the B 1 receptor antagonist des-Arg 10 -Leu 9 -kallidin (100 nM) reduced both effects Search for the Site of Activation-To establish where ACEIs activate the B 1 receptor, we compared the amino acid sequence of ACE with the BK B 1 and B 2 receptors (15,(25)(26)(27). Although there is little overall homology, the human B 1 receptor contains in its second extracellular loop (residues 195-199) an HEAWH sequence (Fig. 4A), and this is similar to the one in the active centers of the two domains of ACE (HEMGH) (26,28) and matches the HEXXH zinc-binding consensus sequence (29) in other members of the zinc metalloproteases. This motif is absent from the B 2 receptor. The ACEIs combine with the active centers of ACE via the Zn 2ϩ cofactor with their SH or COO Ϫ group, and mutation of the HEXXH motif eliminates inhibitor binding (30); thus, this region in B 1 receptor was very likely the site for activation by inhibitors.
[  7) in the absence of the B 1 receptor did not respond to either des-Arg 10 -kallidin or enalaprilat (not shown). Consequently, the HEAWH sequence in the second extracellular loop is essential for the direct activation of the B 1 receptor by enalaprilat but not for des-Arg 10 -kallidin, which acts on other epitopes in the third extracellular loop (20,31).
To explore the role of Zn 2ϩ on receptor response, we preincubated BPAE cells for 30 min with 1 mM Ca-EDTA. This heavy metal sequestering agent blocked the effect of 10 nM enalaprilat completely (n ϭ 4), whereas 10 nM des-Arg 10 -kallidin (n ϭ 5) raised the level of [Ca 2ϩ ] i uninhibitedly. Enalapril (the inactive prodrug of enalaprilat), which has an esterified carboxyl group that does not bind Zn 2ϩ , did not activate the B 1 receptor in IMR-90 cells at 1 M, (n ϭ 3) or 10 nM concentration. We concluded that for activation by a peptide ligand zinc is not needed, but it is essential for activation by enalaprilat.
Blockade by B 1 -(192-202) Undecapeptide-To further confirm the importance of our finding obtained with the [H195A]B 1 mutant, we used the synthetic undecapeptide (LL-PHEAWHFAR) corresponding to residues 192-202 of the B 1 receptor, the putative site of activation by ACEIs. The presence of 10 or 100 M of this peptide blocked the effect of the ACEI completely, whereas des-Arg 10 -kallidin was not affected (Fig. 5;  n ϭ 3). As control, the same concentration of another peptide (AIKLGTGRRFTTC) of similar size but unrelated sequence affected neither enalaprilat nor des-Arg 10 -kallidin responses in the BPAE cells (n ϭ 3, not shown).
Competition Binding-Our experiments strongly suggested that ACEIs and des-Arg 10 -kallidin bind at different sites of the B 1 receptor. We investigated this further in competition binding assays. Membrane preparations, obtained from homogenized HEK 293 cells and transiently transfected with the wild type B 1 receptor, were exposed to 1 nM [ 3 H]-des-Arg 10 -kallidin in the absence or presence of increasing concentrations of enalaprilat. Enalaprilat competed with des-Arg 10 -kallidin binding and replaced the labeled ligand at the B 1 receptor but only at a relatively high (Ͼ1 M) concentration (Fig. 6). This is in contrast with the ability of nanomolar enalaprilat to release NO or raise [Ca 2ϩ ] i but is in agreement with the conclusion that des-Arg 10 -kallidin and enalaprilat activate the B 1 receptor at different sites.
Other ACEIs-We also tested other ACEIs to determine whether the direct B 1 receptor activation is specific to enalaprilat or is rather a group-related effect. Like enalaprilat, captopril (10 nM, n ϭ 4 and 1 M, n ϭ 2) and ramiprilat (1 nM, 10 nM and 1 M, n ϭ 3) elevated [Ca 2ϩ ] i levels by activating the B 1 receptor in BPAE cells. On the other hand, another active but structurally different ACEI, lisinopril (10 nM, n ϭ 5; 1 M, n ϭ 3), did not activate the B 1 receptor in these cells (data not shown).
We also tested D-penicillamine (␣-amino-␤-methyl-␤-mercaptobutyric acid), which is structurally related to captopril. Dpenicillamine was inactive on the B 1 receptor in BPAE cells (10 nM, n ϭ 3; 1 M, n ϭ 3; 100 M, n ϭ 2). Also, bradykininpotentiating pentapeptide (BPP5a), an ACEI (32, 33) and a slowly cleaved substrate (9), was inactive in 100 nM and 10 M concentrations (n ϭ 2, 5; not shown). ACEIs (1-7). Because ACE has a dual action, it activates angiotensin I and inactivates BK by cleaving C-terminal dipeptides (9, 18, 28), and its inhibitors block the release of the potent vasoconstrictor and mitogen angiotensin II and aug-ment activation of the B 2 receptor by BK. Besides preventing kinin inactivation by ACE (10), these inhibitors indirectly potentiate B 2 because they induce an ACE/B 2 receptor cross-talk (12)(13)(14). Des-Arg 10 -kallidin is the endogenous ligand that potently activates the second kinin receptor, B 1 , at a low concentration (15). The role of the B 1 receptor in contributing to ACEI therapy has not been explored systematically. We showed above that ACEIs in nanomolar concentrations directly activate the human or bovine B 1 receptor in cultured cells in the absence of ACE or peptide ligands at a different extracellular domain than the peptide. ACEIs were inactive on cells that did not express the BK B 1 receptor.

Tens of millions of patients worldwide are treated with
ACEIs release NO (34,35), which contributes to their beneficial therapeutic effects. The mechanism of NO release by ACEIs is usually attributed to blocking BK inactivation and thereby potentiating the effect of BK on B 2 receptors, which stimulate endothelial NO synthesis (10). Activation of the B 1 receptors by peptide ligand also releases NO (16). Enalaprilat in nanomolar concentrations did indeed stimulate NO production from pulmonary arterial endothelial cells, and the prolonged release pattern was similar to that caused by des-Arg 10kallidin but not by BK (Fig. 3). The effect of enalaprilat was suppressed by des-Arg 10 -Leu 9 -kallidin, a B 1 receptor antagonist, and NO production increased linearly with the log M concentration of the inhibitor.
We found the sequence in the B 1 receptor where enalaprilat activated directly after comparing the amino acid sequences of human ACE and BK B 2 and B 1 receptors. The second extracellular loop of the human B 1 receptor contains the HEAWH (195-199) sequence (15,25). This HEXXH motif represents a Zn 2ϩ -binding sequence of the two active centers in the N-and C-domains of ACE and in other members of the zincin family of zinc-metalloenzymes (29), but it is absent in the B 2 receptor (27).
The mutation of this Zn 2ϩ -binding domain to remove an essential His residue abolished the effect of enalaprilat but not that of des-Arg 10 -kallidin. In agreement with this finding, enalapril, the inactive prodrug with an esterified carboxyl group that does not bind zinc, did not activate the B 1 receptor. This was further supported by experiments where Ca-EDTA or a synthetic undecapeptide (LLPHEAWHFAR) that corresponds to sequence 192-202 of B 1 and incorporates the suggested site of activation blocked the effect of enalaprilat. Therefore, the HEAWH sequence in the second extracellular loop of the B 1 receptor is essential for the direct effect of enalaprilat on the B 1 receptor. The absence of a similar sequence in the B 2 receptor explains the lack of direct action of ACEIs on it (11)(12)(13). The B 1 receptor may belong to a group of receptors that is modulated by zinc ions (36).
The peptide inhibitor of ACE, BPP5a, is essentially a slow substrate (9) and did not activate the B 1 receptor. This is not surprising, as it is not expected of the B 1 receptor to hydrolyze ACE substrates. Of the ACEIs tested for a direct effect on the B 1 receptor, only lisinopril did not activate it. This consequently shows that ACE inhibition alone is not a sufficient attribute for a molecule to bind to the B 1 receptor. However, for this class of clinically used ACEIs, the ability to inhibit ACE seems to be necessary for B 1 receptor binding. The inactive prodrug enalapril, which only differs in structure from enalaprilat by an esterified carboxyl group, is inactive on the B 1 receptor. ACEIs are highly active (nanomolar concentration) agonists of B 1 receptors, also indicating that many of the features necessary to inhibit ACE potently are required of agonists that bind at the HEAWH receptor site. It is unlikely that a common structural element, unrelated to ACE inhibition, would link ACEIs to B 1 receptors. This conclusion is also supported by the fact that lisinopril was inactive on the B 1 recep-tor. The structure of lisinopril is identical to that of enalaprilat except for the presence of a lysine side chain in the P1Ј position; enalaprilat and most of the other ACEIs have only a methyl group at this position in contrast to the -(CH 2 ) 4 -NH 2 of lisinopril. Thus, the presence of this large charged substituent is likely to be the reason why lisinopril does not activate the B 1 receptor; it could prevent access to a recessed zinc-binding site, or the positively charged side chain could be repelled by a positive charge on the receptor (37).
ACEIs are used extensively with few side effects (1)(2)(3)(4)(5)(6)(7)(8). The B 1 receptor can be rapidly induced under various pathological conditions (for example, by cytokines, ischemia and atherosclerosis) (15,16) undoubtedly encountered by many subjects treated with ACEIs. A recent report (38), published after submission of this manuscript, showed that ACEIs themselves can induce expression of the B 1 receptor in rodent kidney, heart, and vasculature and that B 1 receptor stimulation plays a role in the hypotensive effect of ACEIs. Furthermore, the up-regulation is dependent on stimulation of the B 1 receptor itself because B 1 receptor antagonists blocked it (38). The mode of activation of the receptor at the cellular or molecular level was not investigated, but our results indicate it could be due to direct stimulation of the B 1 receptor by ACEIs. Activation of B 1 receptor by ACEIs could mediate some of the beneficial effects of these drugs by stimulating NO production. For example, administration of an ACEI shortly after acute myocardial infarction reduced the incidence of death or development of severe congestive heart failure (3,4). Acute myocardial infarction also induces B 1 receptor expression (39). The induction and activation of the B 1 receptor could prove advantageous under various stressful conditions, ranging from infection to cardiovascular disorders (15). The B 1 receptor may protect the heart in ischemic preconditioning (40). Activation of B 1 receptors in an ischemic heart inhibited norepinephrine outflow and prevented potentially lethal ventricular fibrillation (41). Neointima formation after balloon angioplasty was suppressed by the B 1 receptor (42) and also by ACEIs after endothelial injury (7). The novel mode of action of ACEIs via B 1 receptor, described above, may add to the therapeutic effects of ACEIs (43) in other pathological conditions (44), and furthermore it may contribute to some side effects.
In conclusion, using cultured cells which express B 1 receptor either constitutively or after transfection we demonstrated a direct activation of the BK B 1 receptor by ACEIs and identified the site of action. To our knowledge, this is the first report showing a direct effect of ACEIs, apart from their effects on ACE, that relates to the therapeutic efficacy of this important class of drugs.