Crystal structure of D. melanogaster AnCE-peptide complexes
AnCE was co-crystallized with Ang II, BK, Thr6–BK, BPPb and their structures were determined at 2-Å resolution (Fig. 1 and Tables 1 and 2). The co-crystallization of Ang I (Asp–Arg–Val–Tyr–Ile–His–Pro–Phe–His–Leu) with AnCE resulted in conversion to Ang II (Asp–Arg–Val–Tyr–Ile–His–Pro–Phe) which can be observed in the substrate-binding channel. In the AnCE–Ang II peptide complex structure, clear electron density was observed for the tetrapeptide Tyr–Ile–His–Pro (Fig. 2A and Table 2). Ang II is resistant to hydrolysis by AnCE (Fig. S1) and repositions itself in the active site so that the penultimate C-terminal Pro residue shifts from S2 to the S2′ subsite after the hydrolysis of Ang I. Based on molecular modelling, we predict that the C-terminal Phe of Ang II could be accommodated in the binding pocket. It is likely that the side chain of Phe occupies the hydrophobic pocket surrounded by aromatic residues Tyr496, Phe127, Trp263 and Phe169 and the peptide main chain atoms extend into the solvent channel by displacing some of the bound water molecules towards a cluster of polar residues Asp360, Gln266, Asn261 up to Glu269. Unlike Ang II, BK (Arg–Pro–Pro–Gly–Phe–Ser–Pro–Phe–Arg) and Thr6–BK (Arg–Pro–Pro–Gly–Phe–Thr–Pro–Phe–Arg) undergo degradation by AnCE to BK1–7 and Thr6–BK1–7, respectively and then to BK1–5 (Fig. S2A). BK1–5 is further cleaved by AnCE to release the dipeptide Gly–Phe (Fig. S2B) and therefore under the conditions employed in the crystallization, it is expected that both BK and Thr6–BK will be sequentially hydrolysed to the final product, Arg–Pro–Pro (BK1-3). Therefore, it was not surprising that the structures of AnCE in complex with BK and Thr6–BK showed Arg–Pro–Pro bound in a similar fashion in the active site cleft (Fig. 2B,C and Tables 2 and 3). For the inhibitory BPPb peptide (pGlu–Gly–Leu–Pro–Pro–Arg–Pro–Lys–Ile–Pro–Pro; Ki, 107 μm) clear continuous electron density was observed for residues Arg–Pro–Lys–Ile–Pro–Pro (Fig. 2D and Table 2). In all four complex structures the catalytic zinc ion at the active site provides the anchor point through direct coordination with the peptide backbone. The peptide interactions with the active site residues are further stabilized by a string of bound water molecules (Figs 3A,C,E and S3). Optimal interaction of the substrate with residues in the active site leads to the displacement of a water molecule, previously in coordination with the zinc ion, towards the active site Glu. This displacement results in an enhancement of the nucleophilicity of the water molecule and positions it for nucleophilic attack on the substrate carbonyl carbon. The binding of the peptide did not introduce any conformational change in the active site of the protein.
Table 1. Crystallographic statistics
| ||AnCE–Ang II peptide complex||AnCE–BPPb peptide complex||AnCE–BK peptide complex||AnCE–Thr6–BK peptide complex|
|Cell dimensions (Å; a = b, c)||173.41, 102.24||173.21, 102.89||173.12, 101.67||173.29, 101.45|
|Angle (°; α = β, γ)||90, 120||90, 120||90, 120||90, 120|
|Completeness (%)||98.4 (91.9)||90.3 (85.3)||96.6 (82.3)||96.7 (83.2)|
| R symm a ||5.3 (23.2)||4.9 (23.0)||8.8 (42.1)||5.0 (16.9)|
|I/σ (I)||19.0 (4)||18.2 (3.5)||9.1 (2.2)||14.9 (5.1)|
| R cryst b ||18.8||20.3||18.1||18.4|
| R free c ||21.0||22.5||19.9||20.0|
|Rmsd in bond lengths (Å)||0.007||0.007||0.006||0.006|
|Rmsd in bond angles (°)||0.99||1.03||0.91||0.90|
|B-factor statistics (Å2)|
|Protein all atoms||34.3||36.7||28.3||25.4|
|Protein main chain atoms||34.0||36.4||28.0||25.0|
|Protein side chain atoms||34.6||37.0||28.5||25.7|
|Glycosylated carbohydrate atoms||52.7||56.3||40.7||37.3|
|PDB code|| 4AA1 || 4AA2 || 4ASQ || 4ASR |
Table 2. Substrates bound to Drosophila melanogaster AnCE in the crystal structures. Amino acids left after degradation are underlined
|Substrate||Peptide sequence||Resolution of the crystal structure (Å)||Ordered visible peptide observed in the structure|
Table 3. Hydrogen bond interactions of Drosophila melanogaster AnCE with Ang II, BK, Thr6–BK and BPPb peptides
|Ang II peptide||BPPb peptide||BK peptide||Thr6–BK peptide|
|Ligand atom||Interacting atom from AnCE (and Zn ion)||Distance (Å)||Ligand atom||Interacting atom from AnCE (and Zn ion)||Distance (Å)||Ligand atom||Interacting atom from AnCE (and Zn ion)||Distance (Å)||Ligand atom||Interacting atom from AnCE (and Zn ion)||Distance (Å)|
|Y4 N||A340 O||3.1||K3 N||A340 O||2.9|| || || || || || |
|Y4 O||A340 N||2.8||K3 O||A340 N||2.9|| || || || || || |
|Y4 OH||T387 OG1||2.7|| || || || || || || || || |
| || || || || || ||R1 O||H367 NE2||3.0||R1 O||H367 NE2||3.1|
|I5 O||Y507 OH||2.7||I4 O||Y507 OH||2.6||R1 O||Y507 OH||2.6||R1 O||Y507 OH||2.6|
|I5 O||Zinc ion||2.2||I4 O||Zinc ion||2.4||R1 O||Zinc ion||2.5||R1 O||Zinc ion||2.5|
| || || || || || ||R1 N||H371 NE2||3.1||R1 N||H371 NE2||3.2|
| || || || || || ||R1 NH1||Y496 OH||3.0||R1 NH1||Y496 OH||3.0|
|H6 N||A338 O||3.2|| || || || || || || || || |
|H6 O||H337 NE2||2.6||P5 O||H337 NE2||2.8||P2 O||H337 NE2||2.7||P2 O||H337 NE2||2.8|
|H6 O||H497 NE2||2.9||P5 O||H497 NE2||3.0||P2 O||H497 NE2||3.1||P2 O||H497 NE2||3.2|
|P7 O||Y504 OH||2.6||P6 O||Y504 OH||2.6||P3 O||Y504 OH||2.5||P3 O||Y504 OH||2.5|
|P7 O||K495 NZ||2.7||P6 O||K495 NZ||2.7||P3 O||K495 NZ||2.7||P3 O||K495 NZ||2.7|
|P7 O||Q265 NE2||3.2||P6 O||Q265 NE2||3.0|| || || || || || |
Table 4. Kinetic constants for Drosophila melanogaster AnCE. App Km, apparent or observed Km; SE, standard error
|Substrate||Inhibitor||App Km (μm)||Ki (μm)||SE|
|Ang I|| ||1040a|| ||210|
|HHL||Ang IIb|| ||75||9.6|
Figure 1. Substrate-bound Drosophila melanogaster AnCE crystal structure. AnCE (cyan) in cartoon representation, with Ang II as red sticks, glycosylation carbohydrates as yellow sticks. The catalytic zinc ion is shown as an olive green sphere.
Download figure to PowerPoint
Figure 2. Portions of the difference electron density map for the bound peptide in the active site of AnCE. Electron density map is contoured at 1σ level. The picture was created using a Fourier difference density map in which the peptide atoms were omitted (A) Ang II, (B) BK, (C) Thr6–BK and (D) BPPb in the crystal structure of their respective complex with AnCE.
Download figure to PowerPoint
Figure 3. AnCE–peptide interactions. (A) Ang II-bound AnCE crystal structure. AnCE (cyan), with Ang II in red sticks. The catalytic zinc ion is shown as a green sphere. Bound water molecules as small spheres. (B) Schematic view of Ang II binding. Hydrophobic interactions, hydrogen bonds and distances cited (grey). (C) BK-bound AnCE crystal structure. AnCE (cyan), with BK in pink sticks. Citrate ion in grey. (D) Schematic view of BK binding. Hydrophobic interactions, hydrogen bonds and distances cited (grey). (E) BPPb-bound AnCE crystal structure. AnCE (cyan) with BPPb in orange sticks. (F) Schematic view of BPPb binding. Hydrophobic interactions, hydrogen bonds and distances cited (grey).
Download figure to PowerPoint
In the case of AnCE–Ang II peptide complex, the observable part of the Ang II peptide docked with residues P2, P1, P1′ and P2′, occupying the S2, S1, S1′ and S2′ subsites, respectively (Table 3 and Fig. 3A,B). In this configuration, the peptide could not be cleaved, suggesting that Ang II is a true competitive inhibitor of AnCE-mediated conversion of Ang I. The C-terminal Pro residue (P2′) is clearly visible and anchors the peptide at the S2′ subsite by hydrogen bonds with three residues (Gln265, Tyr504 and Lys495), whereas the pyrrolidine ring is stabilized through hydrophobic interaction with the surrounding aromatic residues (in particular Tyr507 and Phe441). The main chain of the His (P1′) residue from the Ang II peptide makes hydrogen bond interactions with His337 and His497. At the P1 position, the Ile residue is involved in a tetrahedral coordination with the zinc ion and stabilized by a hydrogen bond with Tyr507. The visible electron density terminates at the P2 position with the main chain of Tyr strongly interacting with the main chain of Ala340 (through two hydrogen bonds). The hydroxyl group of Tyr interacts with Thr387 of AnCE.
The detailed structures of AnCE co-crystallized with BK and Thr6–BK are presented in Figs 3C,D and S3A,B and exhibit similar mode of binding for the final cleavage product, BK(1-3). Strong hydrogen bond interactions with AnCE anchor the main chain of the N-terminal Arg (P1) at the S1 site (His367, His371 and Tyr507), which is also involved in coordination with the zinc ion. Furthermore, the side chain for this residue is stabilized through contact with Tyr496. Additionally, both structures bind a citrate ion (from the crystallization medium) at the same position, making contact with the N-terminus of the bound BK (1-3) peptide. The citrate ion also interacts directly with the AnCE main chain and the surrounding water molecules. The two Pro residues at positions P1′ and P2′ form a strong interaction with AnCE through hydrogen bonds with residues His497, His337 and Tyr504, Lys495 respectively. The general mode of binding for BK (1-3) appears similar to that of Ang II, particularly through the various hydrogen bonds with AnCE. Experimental evidence suggests a preferred affinity for Pro at the S2′ binding site and highlights the importance of the hydrophobic pocket (formed by residues Phe441, Phe511, Tyr504 and Tyr507) for peptide recognition, which allows the products of BK and Thr6–BK hydrolysis to shift position and ‘register’ in the active site subsites.
The structure of the AnCE–BPPb complex revealed interactions between the pyrrolidine ring of the two Pro residues in the penultimate and C-terminal positions (P1′ and P2′, respectively) and the amino acid side chains that form the S1′ and S2′ enzyme subsites, respectively (Table 3 and Fig. 3E,F). A cluster of aromatic residues forms this binding pocket (Phe441, Phe511, Tyr504, Tyr507), and Tyr507 ‘stacks’ against the C-terminal Pro, thus enhancing the interaction with AnCE. The two Pro residues (P1′ and P2′) are strongly anchored in the prime binding sites (S1′ and S2′) by hydrogen bonding with multiple residues. The P2′ Pro, in particular, has strong interactions with Gln265, Tyr504 and Lys495. The Ile residue from the BPPb peptide forms a direct coordination with the catalytic zinc ion by replacing the usual water molecule observed in the native AnCE structure . This residue is further stabilized by interaction with the hydroxyl group of Tyr507. The main chain of the next Lys residue of the peptide in the P2 position interacts with Ala340 through two hydrogen bonds. Of the two visible N-terminal residues of BPPb, namely Pro and Arg, Pro is stabilized by hydrophobic interaction with a bulky Trp341 residue, whereas the Arg side chain forms a weak interaction with AnCE via a water-mediated hydrogen bond with Asp501. This part of the structure shows a clear solvent network involving many charged residues and provides room for the accommodation of longer peptide substrates. The homology between AnCE and ACE, and the presence of ‘unique’ N- or C-domain residues provides opportunities for these subsites to be exploited further for the rational design of new domain-selective ACE inhibitors.
Comparison of the four structures clearly shows a common mode of peptide binding to AnCE. First, the P2′ Pro appears as an essential element in the binding with strong hydrogen bonds and hydrophobic interactions at the S2′ subsite. Second, the backbone interactions at the P1′ and P1 sites are identical in all structures with His337 and His497 stabilizing P1′, and Tyr507 and the zinc ion interacting with the P1 main chain. Finally, Ala340 stabilizes the main chain of P2 when the peptides are visible at that position. This arrangement in the orientation of the peptide backbone makes the P1–P1′ linkage resistant to cleavage.
The common mode of recognition in the four complex structures involves the main chain of the peptide taking control at the catalytic site by replacing a water molecule involved in the proteolysis mechanism. Furthermore, Glu368 (potentially acting as the catalytic base that protonates the amine product) does not seem to be involved in any peptide interaction. Altogether, the peptides presented here show strong interactions with AnCE via their backbone atoms and through direct coordination with the zinc ion, thus preventing proteolysis.