Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites

Current Topics in Medicinal Chemistry 2004, 4, 403-429 403 Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites: Conformational Studies in Solution, Homology Models and Comparison with Other Zinc Metallopeptidases

Georgios A. Spyroulias*,1, Athanassios S. Galanis1, George Pairas1, Evy Manessi-Zoupa2 and Paul Cordopatis*,1

Departments of Pharmacy1 and Chemistry2, University of Patras, GR-26504, Patras, GREECE

Abstract: Angiotensin-I Converting Enzyme (ACE) is a Zinc Metallopeptidase of which the three-dimensional stucture was unknown until recently, when the Xray structure of testis isoform (C-terminal domain of somatic) was determined. ACE plays an important role in the regulation of blood pressure due to its action in the frame of the Renin-Angiotensin System. Efforts for the specific inhibition of the catalytic function of this enzyme have been made on the basis of the X- ray structures of other enzymes with analogous efficacy in the hydrolytic cleavage of peptide substrate terminal fragments. Angiotensin-I Converting Enzyme bears the sequence and topology characteristics of the well-known gluzincins, a sub-family of zincins metallopeptidases and these similarities are exploited in order to reveal common structural elements among these enzymes. 3D homology models are also built using the X-ray structure of Thermolysin as template and peptide models that represent the amino acid sequence of the ACE’s two catalytic, zinc-containing sites are designed and synthesized. Conformational analysis of the zinc-free and zinc-bound peptides through high resolution 1H NMR Spectroscopy provides new insights into the solution structure of ACE catalytic centers. Structural properties of these peptides could provide valuable information towards the design and preparation of new potent ACE inhibitors.

1. INTRODUCTION the last 30 years has been achieved through inhibitors based on the pioneering work of Ferreira S.H. [15] and Ondetti 1.1. Angiotensin-I Converting Enzyme and Hypertension M.A. [16,17]. These researchers showed that the venom of a Angiotensin-I Converting Enzyme (ACE), isolated in the Brazilian pit viper contained a factor that greatly enhanced mid 50’s, is a Zinc Metallopeptidase and one of the major the smooth-muscle-relaxing action of the nonapeptide BK components of the so-called Renin-Angiotensin System which also inhibiting ACE. All ACE inhibitors were (RAS) [1-3]. Renin is responsible for the liberation of prepared in the absence of the ACE’s three-dimensional Angiotensin I (AI) in blood, after renin’s catalytic action on structure and bear two main characteristics: (i) designed the angiotensinogen. ACE possesses a crucial role in the on the basis of venom peptide extracting structural regulation of blood pressure since it catalyzes the cleavage of information for the enzyme catalytic site from the crystal the C-terminal His-Leu dipeptide of the rather inactive structure of Carboxypeptidase A (CPA) [18], and (ii) high decapeptide Angiotensin I (AI), in the vasopressor biological activity strongly coupled with enhanced zinc octapeptide Angiotensin II (AII) [4-6] (Fig. (1)). However, binding ability [19]. ACE impact is not only focused on the generation of Angiotensin II but also extended to inactivation of the 1.2. ACE Inhibitors and their Impact in Medicine and vasodilator peptides Bradykinin (BK) and Kallidin [7-9]. Pharmacology ACE is encountered in two distinct forms in humans, the ACE inhibitors are considered among the most potent somatic and the testis form. These differ from the structural antihypertensive drugs and apart their major action, exhibit point of view, mainly in size and number of catalytic sites beneficial lateral effects in the prevention of cardiovascular [10,11]. According to its function, ACE is classified among disease in various classes of hypertensive patients. the peptidyl dipeptidases of zinc metallopeptidases super- Additionally ACE inhibitors have been proven more family due to its ability to remove C-terminal dipeptide [5] effective than other hypertensive substances in reducing from substrates. ACE can also exhibit activity of an proteinuria and retarding the progression of renal damage in endopeptidase against substrates such as Substance P, patients with various types of nephropathy. These features Cholecystokinin and Luliberin (LHRH), peptides with are probably among the reasons that two (International amidated C-end [9,12,13]. As far as the ACE role in blood Society of Hypertension-World Health Organization, ISH- pressure is concerned, the inhibition of ACE enzymatic WHO; Canadian Society of Hypertension) of the three health activity against AI was considered as one of the major organisations (the third is the British Hypertension Society) challenges against hypertensive disease and congestive heart affiliated with hypertension, recommend ACE inhibitors in failure [14]. Therapy, today and after extensive research for the first-line of antihypertensive drug treatment, after the results of the first trials of this type of inhibitors became *Address correspondence to these authors at Department of Pharmacy, available at 1996. University of Patras, GR-265 04, Greece; Tel: +30 2610 997 721; Fax: +30 An alternative treatment of hypertension is focused on 2610 997 714; e-mail: [email protected] and [email protected] the blockade of AII receptors, AT1R and AT2R (94% amino

Fig. (1). Schematic representation of the Renin-Angiotensin system and the role of the two enzymes, Renin and Angiotensin Converting Enzyme (ACE) on regulation of blood pressure through the generation and release of Angiotensin II vassopressor peptide. acid sequence identity), with appropriate antagonists. The the “spacers”. Zinc sites are also characterised by secondary class of AT1R which are currently under continuous interactions with neighbouring amino acids that position in development and trials, exhibit at the moment only one main space and conformational features strongly depend on the the advantage over ACE inhibitors, which is the absence of overall protein folding and three-dimensional structure. cough as a side effect. Interestingly, only ISH-WHO has These features are critical for the structure-function relationship of this class of metalloenzymes and dictate their recommended AT1R antagonists as first line antihypertensive drugs, and this is probably due to the absence of any classification into various families. published long-term trial results. For all these reasons, the The characteristic amino acid sequences, which contain design and preparation of new potent ACE inhibitors still the potential three zinc ligands in the zinc metallopeptidase remains one of the main challenges in the intersection of the family, comprise the binding motif sequences that are a fields of chemistry, pharmacology and medicine. diagnostic tool in enzyme classification. The first two of the protein ligands are found in the first three-, four- or five- 1.3. Zinc Catalytic Sites and their Characteristics residue binding motif, while the third is found in a second characteristic motif. These residue-ligands are generally Zinc sites in metalloenzymes and related biomolecules separated either by short amino acid “spacers” among the are classified according to their ligands and coordination first, second and third ligands, or by a short spacer between geometry into three types of zinc binding sites [20,21]: (i) the two first ligands and by a large spacer between the the (ii) the and (iii) the catalytic, cocatalytic structural. second and third, or fourth ligand, should one exist. The Hydrolases like ACE possess a zinc site which catalytic magnitude of a short spacer could vary from one to three usually coordinates with nitrogen, oxygen and sulphur amino acids between the first two ligands. On the other hand, donors of His, Glu, Asp and Cys residues while His is most the long spacers usually found in various metallopeptidase frequently encountered in the coordination sphere of zinc subfamilies could vary from 5 to over 100 amino acids. The metal ion. Water is also a zinc ligand in catalytic sites and is length of the spacer between the two first ligands belonging activated for ionisation, polarisation, or displacement by the to the same binding motif often characterizes the secondary identity and arrangement of ligands coordinated with zinc structure of this protein fragment. For example, a three- [22]. The zinc coordination number for this kind of sites residue spacer is characteristic of a a -helix conformation has been found to be four or five and the donor atoms of while a one-residue spacer indicates a b sheet conformation. residues define a distorted-tetrahedral or trigonal-bipyramidal coordination geometry. Ionisation and/or polarisation of the 1.4. Focus of this Article activated H2O is assisted by the base form of an active site- residue or in some cases by a “second-shell” residue This article aims to provide new structural insights into that yields hydroxide ions at neutral pH while water ACE, an enzyme whose role in hypertension has stimulated displacement results in Lewis acid catalysis on the part of the over the last 35 years extensive and continuous effort catalytic zinc metal. towards designing its potential inhibitors, even without the most important tool in the hands of biochemists, The structure of the zinc catalytic sites comprises: (i) the enzymologists and drug designers: the ACE three- zinc-bound residues, (ii) the characteristic sequence of the dimensional structure. What follows is an attempt to review amino acids around those ligated to the metal, (iii) the short the latest progress in structural biology of zinc or long amino acid “spacers” among the three (in some cases metallopeptidases and extract structural information through: four) protein ligands, and (iv) the conformational features of Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 405

(i) an investigation of ACE sequence and structure zinc-binding motif, HEXXH (X = any amino acid residue) in similarities/differences among representative members of order to classify the known zinc metalloproteases into five metallopeptidase family, (ii) homologous modeling based on families, where each one has its prototype: (i) Thermolysin known enzyme structure, and (iii) a solution conformational (TLN) [27,53-55] (ii) Astacin (AST) [34], (iii) Serratia [56- analysis through 1H NMR spectroscopy on 36-residue 58], (iv) Matrixin [59,60] and (v) Reprolysin/Adamalysin synthetic peptides whose sequences represent the two ACE [24,49,52,61]. Determination of the zinc metalloproteases catalytic sites and their zinc-binding properties. 3D crystal structure sheds new light on the structural features of these enzymes and new terms, such as zincins, metzincins, 2. ZINC METALLOPEPTIDASES CHARACTER- aspzincins, gluzincins and inverzincins, have been used in ISTICS AND CLASSIFICATION order to discriminate between them [24,51,52] (see Fig. (2)). In addition to ACE conformational features extracted through The super-family of zinc-containing bio-macromolecules theoretical and experimental studies, an update of the sequence has been enormously extended over the last years, with the and topology characteristics of zinc metalloproteases/ same occurring in the frame of the zinc metalloproteases/ peptidases through the structural analysis of known 3D peptidases family [23,24] (Fig. ( )). An attempt to classify 2 structures is presented. previously unknown biopolymers with distinct motifs and zinc-binding affinity at the beginning of the 90’s led 2.1. Zincins, Metzincins and Gluzincins: Where ACE researchers to identify diagnostic motifs within the belongs? polypeptide amino acid sequences through sequence and topology comparison [45]. Using this methodology, not only Overall, the zinc metallopeptidases can be divided into sequence but also structural similarities have been identified two categories according to the sequence of the first zinc among various zinc metalloproteases and thus enzyme binding motif: (i) the zinc enzymes with the characteristic groups into certain categories [46-52]. In 1992 Jiang and HEXXH motif, where the two histidines are the potential Bond [48] compare the sequences around the diagnostic, protein ligands, and (ii) the zinc enzymes without HEXXH

Fig. (2). Classification of Zinc Metalloproteases according to MEROPS Protease Database (July 2002; http://www.merops.co.uk/merops/ index.htm). Data concerning protein source and zinc ligands, amino acid spacer between the binding motifs, accession number and codes for sequence and coordinates are also given. The X-ray structures of the active site for some representative metalloenzymes are presented. The zinc protein ligands are labelled. The number of amino acids (aa) refers to the spacer length between the protein zinc ligands. X stands for any amino acid while, in bold letters, His, Glu and Met residues are noted. 406 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al. motif. The presence of HEXXH defines the so-called zincins sequence EXXXX. TLN is considered the prototype member family [49]. The amino acid sequence and topology close to of this category of zinc metallopeptidases, whose first X-ray the second binding motif or the motif sequence itself divides crystal structure was solved almost 30 years ago [27,53-55] the zincins family into metzincins [49,52] and gluzincins [24] (see also Fig. (2), (3)). The last residue of the EXXXX subfamilies (Fig. (2)). According to Bode et al. [49] the sequence for TLN (pdb: 1lnd) [27]) and Neprilysin (or crystal structure determination of AST (pdb: 1ast [34]) and Neutral Endopeptidase; NEP) (pdb: 1dmt [28]) crystal Adamalysin II (pdb: 1iag [61]; see also Fig. (3)) revealed structures has been found to be an aspartatic acid which is that these two proteinases exhibit low sequence similarity also found in the same position of the proposed second but significant topological equivalence. Their main feature is binding motif of ACE (EAIGD). According to the sequence the almost identical zinc environment with a methionine- identity of the key-residues in the two potential binding containing turn and, according to these findings, Bode has motif sequences, ACE is classified among the gluzincins suggested that these enzymes could be termed as metzincins. superfamily and specifically in the M2 family of MA(E) clan The third zinc ligand in the vast majority of metzincins is a of zinc metalloproteases (Table (1)). histidine residue, while in some members of this family a In 1994, the gluzincin term, according to Hooper, tyrosyl residue has been identified as the fourth coordinated spanned six families of the metalloprotease MA clan, M1, protein ligand [49,52]. However, the X-ray crystal structures M2, M3, M4, M13 and M27 with representative members of Snapalysin (M7 family, pdb : 1kuh [31], Fig. (3)) and Aminopeptidase A (EC 3.4.11.7), ACE (EC 3.4.15.1), Peptidyl-Lys Metalloendopeptidase (M35 family, pdb: 1g12 Thimet Oligopeptidase (EC 3.4.24.15), TLN (EC 3.4.24.27), [62], Fig. (3)) revealed that an aspartate residue is the third NEP (EC 3.4.24.11), and Botulinum neurotoxin A (EC 3.4. protein ligand, although these enzymes belong to the MA(M) 24.69) respectively. Until mid 2002, the gluzincin term clan and have been termed aspzincins [62]. encompasses eighteen metallopeptidase families, according

The term gluzincins was introduced in the mid 90’s by to the Protease Database (http: // www . merops . co.uk/merops/ merops.htm), while there are 9 3D X-ray structures out of 99 Hooper N. M. for the zincins whose third zinc ligand is a assigned peptidase sequences classified in gluzincins, glutamic acid [24] found in the consensus binding motif

Fig. (3). 3D X-ray structures of the zinc-containing active sites of the two superfamilies, metzincins MA(M) (left) and gluzincins MA(E) (right), of zinc metalloproteases, available until July 2002. Representative models, from each clan, which possesses at least one characteristic enzyme with available 3D structure, are presented and pdb codes are given. Prototype clan members with no 3D structure determined so far are also listed. Active site helices are shown in blue/gold ribbon representation while non-active helices are depicted in grey/light grey. Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 407

Table 1. Amino Acid Composition of Zinc Binding Motif Sequences for Representative Members of All Known Families of MA(E) Clan, Gluzincins, Zinc Metallopeptidases. Horizontal Black Shaded Boxes Indicate the Family of the MA(E) Clan. Dark- gray Columns Indicate the Metal Binding Ligand which have Either been Identified Through the Available X-ray Structure or Proposed in Analogy to other Enzymes. PDB Code, when Available, is Given (Last Column).

Sequence1 MA(E) - M2 PDB REF

ACEN H390 EMG H394 ---23--- E418 AIGD

ACEC H390 EMG H394 ---23--- E418 AIGD 1O8A [148-9]

MA(E) – M1

LTA4 Hydrolase H295 EIS H299 ---18--- E318 GHTV 1hs6 [25]

M3A

Neurolysin H474 EFG H478 ---24--- E503 VPSQ 1i1i [26]

M3B

Oligopeptidase F H387 ETG H476 ---23--- E415 IAST

Thermolysin H142 ELT H146 ---19--- E166 AISD 1lnd [27]

Mycolysin H201 EAG H205 ---33--- E239 GFAD

M9A

Microbial collagenase H401 EYV H405 ---23--- E429 GIAE

M9B

Collagenase colA (H) H414 EYT H418 ---27--- E446 GGAE

M13

Neprilysin H583 EIT H587 ---58--- E646 NIAD 1dmt [28]

M26

IgA1 protease H1494 EMT H1498 ---19--- E1518 FFAKFFAK2

M27

Botulinum neurotoxin H222 ELI H226 ---34--- E261 LRTF 3bta [29]

M30

Hyicolysin H246 EYQ H250 ---44--- E269 AFAM

M32

Carboxypeptidase Pfu H276 EMG H280 ---24--- E305 SQSRSQSR3 1k9x [150]

M34

Anthrax Lethal factor H719 EFG H723 ---44--- E768 FFAE 1j7n [30]

M36

Fungalysin H429 EYT H432 ---25--- E458 GWSDGWSD4

M41

FtsH endopeptidase H414 E_G H418 ---57--- E476 EIIYEIIY5

M47 408 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al.

Table 1. (Contd….)

Sequence1 MA(E) PDB REF

PRSM1 metallopeptidase H204 ELG H208 ------6

M48_

CAAX prenyl protease H335 ELG H339 ---75--- E415 FQAD

M48_

HtpX endopeptidase H139 E_S H143 ---78--- E222 FHAD4

M60

Enhacin H248 ELG H252 ---14--- E267 VWNN7

M61

Glycyl aminopeptidase H248 ELG H252 ---29--- E282 GQTQ8

Consensus Sequence H EXX H E XXXX 1 Residue numbering is according to the sequence record deposited at Swiss-Prot (http://www.expasy.ch/) when no 3D structure is available. With peptidases for which coordinates have been deposited at Protein Data Bank (http://www.rcsb.org/pdb/) the residue numbering follows the PDB record. Numbers between the 2nd and 3rd protein ligand indicate the magnitude of amino acid spacer. 2 Tentative assignment – E 1518 is the 1 st conserved glutamate in all known sequences of M26 family. The 2nd conserved E is found after 90 intervening residues in a pentapeptide that possesses the sequence EGNSI. 3 Tentative assignment – E305 is found in the consensus sequence of HESQSX (where X = R or L) of M32 family. This E has been assigned as 3rd zinc ligand according to the conserved residues of the motif and to the preceding histidyl residue, which is also found in the same position in ACE sequences. 4 Tentative assignment – E 458 has been assigned as 3rd zinc ligand since the 4th residue of the consensus sequence EXXXD is an Aspartic acid as has also been found in the same position in M4, M5, M13, M48A and ACE sequences. 5 Tentative assignment – E476 is the only conserved glutamate in all known sequences of M41 family. 6 No assignment is attempted since only one sequence for M47 family is available. First E residue is found after 19 intervening residues in the sequence EWPGG. 7 Tentative assignment – E267 is the 1st conserved glutamate in all known sequences of M60 family in the consensus sequence EXWZN (where X = I or V and Z= N or T). The 2nd conserved E is found after 50 intervening residues in the non-consensus motif sequence ERNIA. 8 Tentative assignment – E282 is the only conserved glutamate in all known sequences of M61 family in the consensus sequence EGXTZ (where X = T, F or Q and Z= S or Q). yielding representative 3D models for six of the families in all zincins’ first binding motif there is no other conserved (M1, M3, M4, M13, M27 and M34; see Fig. (3)). amino acid in the two binding motifs. However, among the The catalytic, zinc-containing, sites extracted from the 18 gluzincins families (plus three subfamilies), in 7 families available X-ray structures of representative members of zinc the fourth residue of the second, the “glutamate” binding motif, is an aspartic acid, as also happens in the ACE metalloproteases metzincins, MA(M), and gluzincins, MA(E) sequence. Of these families, only two, MA(E)-M4 and -M13, clans are illustrated in Fig. (3). possess sequences for which X-ray crystal structures are 2.2. Sequence and Structure Characteristics: Binding available (Fig. (3)). Representative members of these two Motifs and Active Sites of Zincins, Metzincins and families are TLN and NEP respectively, exhibiting the Gluzincins HEXTH and EXIXD consensus zinc binding motif sequences, which exhibit significant similarities with the HEMGH and Enzymes from each of the above mentioned EAIGD sequences of ACE. Additionally, TLN’s active site metallopeptidase clans and families are grouped below possesses a 19 amino acid spacer that is comparable to that according to three main characteristics concerning the of the 23 amino acid spacer attributed to ACE’s active site characteristic sequence of: (i) the first zinc binding motif, (ii) polypeptide, while NEP possesses a spacer consisting of 58 the second zinc binding motif and (iii) the amino acid spacer amino acids. between the two binding motifs and its magnitude. Despite the low overall sequence identity between The amino acid composition of their binding motif metzincins and gluzincins peptidases, such as for AST and sequences and the number of residues which constitute the TLN respectively, significant topological similarities in their spacer for gluzincin sequences for which both binding motifs active site have been implied by sequence comparison have been distinguished and characterized is illustrated at [24,29]. Crystal structure determination of metzincins Table (1) while analogous data for metzincins are presented representatives in the early 90’s, such as Adamalysin [pdb: at Table (2). 1iag, 61] and AST [pdb: 1ast, 34], revealed that the first binding consensus zincins sequence HEXXH constitutes a According to Hooper N. M. [24] the consensus sequence helix, which has also been found in the X-ray structure of of gluzincins is HEXXH-spacer-EXXXX and the magnitude TLN [49,27,53-55]. Crystal structures of metzincins and of the spacer varies between 18 and >80 amino acids. Except gluzincins peptidases solved over the last decade verify the for the three zinc ligands and the glutamic acid that is present Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 409

Table 2. Amino Acid Composition of Zinc Binding Motif Sequences for Representative Members of MA(M) Clan Families, Metzincins, Zinc Metallopeptidases which Posses at Least One Member with known Xray Crystal Structure. Horizontal Black Shaded Boxes Indicate the Family of the MA(M) Clan. Dark-gray Columns Indicate the Metal Binding Ligand According to Xray Structure. The Light-gray Column Indicates the Three-residue Sequence of the Characteristic ‘Met- turn’ while Tyrosyl Residue in bold Characters Indicates the Assigned Fourth Protein Zinc Ligand. PDB Code, when Available, is Given (Last Column).

Sequence1 MA(M) – M7 PDB REF

Snapalysin H83 ETG H87 ----5--- D93 HYSG --5- M103SG 1kuh [31]

Leishmanolysin H264 EMA H268 ---67--- H334 IKMR --6- M345AP 1lml [32]

M10A

Collagenase 12 H218 ELG H222 ----5--- H228 STDI --3- M236YP 1fbl [33]

M10B

Serralysin2 H176 EIG H180 ----5--- H186 PGDY -23- M214SY 1srp [58]

M12A

Astacin2 H92 EAM H96 ----5--- H102 EHTR -40- M147HY 1ast [34]

M12B

Adamalysin H142 ELG H146 ----5--- H152 DGKD --9- M166RP 1iag [61]

M35

Peptidyl-Lys H117 ESS H121 ----9--- D130 YAYG ------3 1g12 [62] endopeptidase

Consensus Seq. H EXX H H/D XXXX MXX 1 Residue numbering is according to the record with the structure coordinates (PDB record) which have been deposited at Protein Data Bank (http://www.rcsb.org/pdb/). Numbers between the 2nd and 3rd protein ligand, as well as before the “Met-turn” Methionine, indicate the magnitude of amino acid spacer. 2 Tyrosine in the Met-turn sequence is noted with bold characters and indicates the 4th zinc ligand. 3 No Met-turn motif has been identified for this peptidase either in identified sequence or in X-ray structure. Last residue in sequence (P81054; Swiss-Prot : http://www.expasy.ch/) and in X-ray structure is Ser167. assumption that the HEXXH first binding sequence for all sequence equivalence, (ii) conformation of first and second zincins is part of a helix, the so-called ‘active site helix’. The binding zinc motif sequence for both metzincins and second binding motif of metzincins and gluzincins varies gluzincins is remarkably similar among members of each structurally as implied by sequence comparison between family even when sequence identity of active site is low (15- various peptidases of the two MA(M) and MA(E) clans. The 25%) and intervening residues differ in nature and length. third protein zinc ligand is found in a helical fragment for all gluzincins structures solved, but this is not the case with 2.3. Functional and Structural Relationship of ACE with metzincins peptidases. The latter possess a histidine or other Zinc Metallopeptidases aspartate residue as a zinc third ligand, which has not arisen Structural and catalytic properties of ACE have been from a second helix but is located in an open coil segment. discussed in the literature in comparison either with a zincin, No matter what the amino acid composition, the length or the the gluzincin metallopeptidase TLN, or with a non-zincin conformation of the spacer between the two binding motifs metallopeptidase the CPA [63-69]. ACE is a dipeptidyl in gluzincins (Fig. (3)), the EXXXX sequence always carboxypeptidase while CPA is an exopeptidase [63,64] and comprises a second ‘active site helix’. Consequently, the TLN is an endopeptidase [66,69]. active site environment in gluzincins is characterised by two ‘active site helices’, which constitute the main component of Carboxypeptidase A (see Fig. (1)) has a HXXE as a first the zinc catalytic pocket. Additionally, the major structural binding motif sequence and HXX as a second. CPA first feature that characterises the conformation of the 18-58 binding motif is a four-residue motif in contrast to the five- amino acids spacer is the a -helix, especially when the residue zincins motif and possesses the glutamate and the magnitude of this spacer is longer than 20 residues. histidine as second and third protein ligand. In contrast, the first zincins motif sequence bears the two histidines and the Summarising the information drawn from the analysis of glutamate follows as the third ligand sited at the second zincins structure available so far are one should note that: binding motif. None of the CPA binding motifs comprises a (i) Topological equivalence may occur even without helical fragment while the amino acid spacer between them 410 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al. varies from 108 to 135 amino acids among the various hypothetical model of ACE based on the active center of carboxypeptidases sequences. CPA. (see Fig. (4C)) [72]. They also suggested that the catalytic mechanism of ACE is similar to that of CPA. Based 2.3.1. Carboxypeptidase A. on these studies several research groups designed and Research on the inhibition of ACE’s catalytic action synthesized numerous potential ACE inhibitors, arriving at started in the early 1970s [15-17], and has been based on the first pharmacologically promising antihypertensive function similarities of ACE with other enzymes with known compounds that bind at ACE’s active sites [19]. 3D structures [70-72]. One of these enzymes was CPA (M14 Analysis of 3D structure of CPA catalytic cavity [73] family of MC clan), the structure of which was determined revealed that a positively charged residue (noted by “+” in by X-ray crystallography (pdb: 1yme, Fig. (4)) [18,36] and Fig (4C)), Arg145, extends its side-chain NH3 groups its active site and catalytic mechanism has been extensively towards the metal site and has the role of forming ionic investigated. [63,73]. CPA possesses a zinc-binding motif bonds with the negatively charged carboxyl group of the with the characteristic HXXE sequence, and is able to cleave substrate’s C-terminal (Scheme (1A)). As an analogy to that, a single amino acid from the carboxy-terminal end of a a positively charged residue is believed to exist at ACE peptide substrate, in contrast with ACE, which hydrolyses active site and participate in the catalytic mechanism of the the COOH-terminal dipeptides. Ondetti and Cushman, who hydrolytic dipeptide cleavage [72]. Also, in CPA, the zinc were recently honored for their pioneering work and ion, which polarizes the carbonyl group of the scissile continuous effort in the design and study of potent and peptide bond, has been found in a substrate’s single-residue specific ACE inhibitors [19], proposed a late 1970’s far from Arg145, which interact with the substrate’s COO-

Fig. (4). (A ) X-ray crystal structure of CPA (pdb: 1yme; top) and TLN (pdb: 1lnd; bottom). (B) Conformation of structural and catalytical important residues of the zinc active site of CPA (B, top) and TLN (B, bottom). Side-chains of residues are presented in ball and stick. (C) Cartoon representation of active site of CPA based on its X-ray structure and a CPA-based hypothetical model of ACE active site where various known inhibitors are accommodated. Sub-sites noted as S1, S2, S1’ and S2’ are cavities or areas in an enzyme’s active site where amino acid groups interact with adjacent side-chains of the substrate’s (peptide or inhibitors) groups in a molecular recognition and complex- formation procedure. Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 411

Scheme 1. Proposed catalytic mechanisms for the (A) Carboxypeptidase A and (B) Thermolysin catalyzed cleavage of peptides. terminal group. On the other hand, at the ACE catalytic site, His231, Tyr157 and Asp226 in the enzyme’s function has zinc metal ion has to be found in a position at a distance of a been also suggested, Fig. (4B) [75-80]. dipeptide-residue, which itself must be far from the The peptide substrate forms a Michaelis complex enzyme’s positively charged residue. Furthermore, Arg127, nd (Scheme (1B); 2 step) with an enzyme’s active site and its Glu270 orient their side-chains towards a zinc-catalytic site carbonyl oxygen is accommodated among His231, Tyr157 (Fig. ( ) & Scheme ( )) and are implicated in CPA 4B 1A and the coordinated H O, which however becomes slightly catalytic mechanism. They have the task of polarizing the 2 displaced from its original position after peptide binding. scissile carbonyl bond, while the Tyr248 OH group is Coordinated H O is then sited closer to Glu143, in a position hydrogen bonded to the terminal carboxylate of the substrate 2 favorable for its polarization, which is also assisted by zinc [63] (Scheme ( )). 1A cation. Glu143 and metal ion therefore enhance the 2.3.2. Thermolysin nucleophilicity of the H2O and promote its attack on the carbonyl carbon. Glu143 accept a proton which in turn is Thermolysin is a thermostable extracellular gluzincin donated to the substrate’s peptide bond nitrogen, forming a endopeptidase which belongs to the M4 family of MA(E) gem-diolate intermediate (Scheme (1B); 3rd step). At this zinc metallopeptidase clan and has been isolated from stage the zinc metal is in a five-coordination state and the Bacillus thermoproteolyticus. TLN catalyses the hydrolytic peptide carbon possesses a tetrahedral geometry. The cleavage of the peptide bond specifically on the imino side geometry of the metal site and the intermediate has been of large hydrophobic residues, and especially leucine, identified through transition-state analogue inhibitors [75- isoleucine and phenylalanine [74] His142, His146 and 80]. The formation of the intermediate product is believed to Glu166 together with a solvent water are found coordinated be assisted by a hydrogen bond network where the side- with the zinc ion in a distorted tetrahedral coordination [27]. chains of His231, Tyr157 and the carbonyl oxygen of Glu143 and Asp170 are positioned in the so-called “second substrate are involved. The role of Asp266 is focused on the coordination shell” of the zinc cation and both belong to the stabilization of the positive charge required for catalysis consensus HEXXH and EXIXD sequences of the first and through a salt-bridge between its negative carboxylate side- second binding zinc motifs, respectively (Fig (4B)). Glu143 chain and the protonated imidazole of His231 [74,80]. is highly possibly hydrogen-bonded with its side-chain Finally, the peptide C-N bond is cleaved and the protonated carboxylate to the coordinated H2O, while Asp170, which is product is released, while Glu143 is proposed to abstract the conserved in various gluzincins including ACE, is considered second water proton, shuttling it to the amine nitrogen a structurally and/or functionally important residue (Scheme (Scheme (1B); 4th step). (1B)). The Asp170 structural role arises from the fact that its charged side chain forms a salt link with the imidazole ring 2.3.3. Angiotensin-I Converting Enzyme and Inhibitors of His142, the first protein zinc-ligand [11,27,53]. According to extensive structural studies of enzyme-inhibitor The discovery of two active sites in somatic ACE has complexes, the catalytic role of the outer shell residues provoked many assays to establish functional or structural 412 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al. difference in these two highly homologous enzyme regions. interaction with mainly three subsites or pockets at the ACE Studies of an ACE somatic isoform fragment that contains active sites, named S1, S1' and S2 (Fig. (4C)). Substrates only the N-terminal active site has been performed with the containing phenyl group (R3) that interacts with the ACE S1 aim of identifying the functional differences or the selective subsite have proved to be significant for the binding properties inhibition between the two catalytic centers of ACE. For of inhibitors [94-97] like antihypertensive compounds example, whereas the most important peptide substrates such enalapril, lisinopril, trandolapril, fosinopril etc. (Fig. (4C)). as BK and AI [4,9, 81,82] are hydrolyzed at both sites, while Furthermore, S1' subsite has been found to have a weaker some others, like LHRH or Angiotensin1-7, are cleaved binding affinity toward the peptide substrate, while it may preferentially only by N-domain active site [83-86]. not be large enough to accommodate peptide substrates with Additionally, it has been reported that the phosphinic extended, long, side-chains [98,99]. Additionally, most compound RXP407 is an N-selective inhibitor of ACE, able substrates that specifically inhibit ACE active sites bear to differentiate between the two active sites [87,88], in proline or proline-like analogues at their carboxy terminus, similarity with some other peptide inhibitors (BBPs) which which interact with the S2' binding pocket [72,100] have been reported recently to selectively inhibit C-domain (captopril, enalapril, lisinopril etc.; see Fig. (4C)). Another active sites. Furthermore, it has been reported that ACE feature that might be essential in a potential enzyme inhibitor catalytic activity requires monovalent anions such as Cl- for is its free C-terminal carboxylate group [101]. According to maximal activity and that the extent of activation is substrate the zinc-binding group, inhibitors or potential inhibitors can and chloride dependent [89-91]. Data suggest that the C- be classified in three categories: (i) those with a sulfydryl domain active site is more sensitive to chloride concentration group, such as captopril [72,100,102], (ii) those with a for the hydrolysis of some substrates [81,92]. carboxyl group, bound to zinc ion as enalapril [102], The catalytic action of ACE has long been discussed in lisinopril [97, 103] etc and (iii) inhibitors with a phosphonate concert with that of CPA and/or TLN [74,80], whose 3D group, like fosinopril [104]. The structures of the above mentioned compounds accommodated at the proposed crystal structure is available. In the absence of an ACE structure, speculations are made on the basis of sequence hypothetical ACE catalytic cavity are presented in Fig. (4). homology, or topology similarities with the above mentioned 2.4. Structural Insights to ACE Active Sites Through enzymes and structure-function correlation studies of Site-directed Mutagenesis mutated enzymes. Between ACE and CPA/TLN, there is undoubtedly low sequence identity while the known Attempts to extract conformational characteristics essential structures of CPA and TLN differ considerably. Moreover, for an enzyme’s function have been also focused on the ACE binding motif and spacer resemble those of TLN while substitution of residues, which in analogy to other peptidases both differ remarkably from those of CPA. ACE binding are believed to be actively involved in ACE’s catalytic motif sequences consist of the residues HEMGH.….EAIGD mechanism. Such residues should be sited close to the zinc and between them 23 intervening residues have been site (Table (1)) and their role elucidated through site directed identified in analogy to the sequences of TLN, mutagenesis and structure-function relationship studies. HELTH…..EAISD separated by 19 residues. However, a When the two histidines of the first potential ACE zinc- profound analysis of the TLN and CPA zinc environment in binding motif have been mutually substituted by other amino their 3D structures suggest that there are some common acids, it is revealed that enzyme [11,92,103,105,106] elements. For example CPA Glu270 is relative to TLN completely abolishes its activity. These data indicate that Glu143 (see Fig. (4A) and (4B)), and in analogy to these two histidyl residues are essential for ACE catalytic Glu362/960 of the two ACE active sites considered vital for activity. Moreover, these amino acids are believed to be the enzyme activity [93]. On the other hand the two residues two first protein zinc-ligands. Further site-directed suggested as the proton donors in the catalytic mechanisms mutagenesis experiments yield a GluàAsp substitution, of CPA and TLN, Tyr248 and His231 respectively, are not where the replaced glutamate residue is that found in the found in a comparable position in their structures. This possible two-histidyl binding motif with the sequence structural variation has promoted the aspect that there was no HEMGH, which has resulted in suppression of enzyme absolute requirement for the involvement of a histidine or a catalytic activity [92]. Aspartate residue retains the negative tyrosine in zinc neutral proteases hydrolytic action. charge in the sequence, which, however, has been displaced Nevertheless, recently a histidyl residue, His1089 has been of approximately 1.4 Å. This glutamate is involved in a basic suggested to possess a similar role to that of TLN His231. attack of the substrate peptide bond and its role is of Additionally, sequence comparison of the amino acid spacer crucial importance to the catalytic efficacy of ACE. This that separates the second and the third zinc ligand in TLN strongly supports the role postulated for Glu143 of the and ACE reveals three and two tyrosines (368/965, 369 and HELTH in TLN [86]. 372/970; somatic form numbering) for the two ACE catalytic sites analogous to Tyr157 in TLN. As far as the nature and properties of the second ACE possible binding motif are concerned, the glutamate residue Since until recently ACE structure remained obscure, the in EAIGD sequence is conserved in many other zinc binding affinity of potential inhibitors to ACE and structure metalloproteases which highly possibly plays the role of the activity relationships have not been totally illuminated. third protein zinc ligand. This glutamate characterizes the Extensive studies have been performed towards a detailed gluzincins and is, indeed, coordinated with the zinc metal ion understanding its active centers specificity. Data indicate that according to the X-ray structures solved for other members the majority of ACE inhibitors exhibit specificity on of this family. When an aspartic residue or a valine replaces Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 413 this glutamic acid, ACE catalytic activity has been decreased coordination. The aspartate residue of EAISD motif orients by more than two orders of magnitude, or has been its side-chain parallel to that of glutamate and towards the completely extinguished, respectively [11]. Finally, mutation first histidine imidazole ring. studies performed by the same research group with aspartic The two antiparallel terminal helices in ACE homology acid in EAIGD motif as target residue, which is replaced by models are in great agreement with the “two active helices a glutamic, or an alanine residue, indicate a specific site” model adopted according to the two helices observed in functional role of its carboxylate side-chain. Asp170 role has the catalytic cavities of all the 9 X-ray gluzincin crystal been proposed to as the salt link formation with the structures available hitherto. These features appear imidazole ring of the first potential zinc coordinated histidine independently of the active site amino acid composition and [11], which is observed in TLN X-ray structure [27,53]. A magnitude. Additionally, theoretical prediction of secondary similar role in ACE function is expected for Asp393/991 structure of ACE structure in Network Protein Sequence (somatic isoform numbering). Analysis [116] through GOR protocols [117-119] reveals Site directed mutagenesis studies have been also two major helices, one towards the N-terminus and the other performed in order to probe the nature of a basic amino acid, towards the C-terminus, both of which include the HEMGH adjacent to an enzyme’s active site, which is believed to bind and EIAGD bonding motifs. a chloride ion. Concentration of chloride ions has been Having analysed all the above data we wish to make an reported among the crucial parameters that adjust ACE effort to elucidate experimentally the structural features of catalytic activity [90,107] and Arg1098 [108] has been ACE active sites. This attempt is performed through conform- reported critical for the chloride dependence of ACE ational analysis in solution, using NMR spectroscopy, of the catalytic activity and binds the anion. However, this is in synthetic 36-residue ACE model peptides. The results of contrast with the results obtained through earlier chemical NMR data analysis for apo- and zinc-bound forms of modification studies on lysyl residues [109], which had ACE [His361-Ala396] peptide and NMR solution structure proposed active involvement in chloride and other N of the free peptide are presented below. monovalent anion binding [109,110]. Furthermore, substitution of His1089 of human somatic 4. SYNTHESIS OF THE TWO ACE Zn-SITES – WHAT form, by an Ala or Leu, has prompted the researcher to COULD BE THE BENEFIT? propose that this histidyl residue stabilizes the transition state X-ray crystallography and NMR spectroscopy are the complex through hydrogen bonding with the tetrahedral two experimental methods, which could yield a high- intermediate product [111]. A similar role has been also resolution three-dimensional structure of biopolymers. attributed to His231 in the catalytic mechanism of TLN However, both of these techniques have to overcome some [74,80]. Finally, mutation studies of human and rabbit testis limitations. X-ray crystallography has the drawback of ACE isoform have demonstrated a functional role of Tyr200 single-crystal and heavy-atom derivative preparation. On the and Tyr236, respectively [112-113] in analogy to the Tyr198 other hand, NMR bypasses the time-consuming stage of in CPA. crystallization but is restricted by the size of the molecule. 3. 3D HOMOLOGY MODELS OF ACE ACTIVE SITES Larger NMR structures of biomolecules or biomolecular complexes reaches the limit of 50-60 kDa while sporadic are Sequence alignment [114] of ACE and other gluzincin the examples that complete sequence specific resonance active site fragments for which an X-ray crystal structure has assignment has been performed for proteins with more than been determined is performed and presented in Fig. (5A). 400 amino acid. But even in that case, NMR studies become The structure of TLN active site (pdb: 1lnd [27]) has been feasible only after the development and the successful chosen as a template in order to generate 3D homology applications of molecular biology and biotechnology models of the 36-residue peptide that represent the techniques which permits the selective labeling and 15 13 ACEN[His361-Ala396] and ACEC[His959-Ala994] zinc deuteration in concert with N and C labeling. Until active sites, as shown in Fig. (5B) [115]. TLN zinc site recently, ACE structure had not been elucidated, neither in fulfils the following criteria: (i) high sequence identity with solid-state, probably due to unsuccessful attempts to obtain ACE binding motifs, (ii) similar amino acid spacer high-quality crystals, nor in solution since its molecular magnitude, (iii) topological and conformational similarities weight exceeds 80 and 140 kDa, for testis and somatic resulting from theoretical prediction of ACE secondary isoform respectively. On the other hand, solid-phase structure at NPSA server (Fig. (5C)). Despite the low synthesis of peptides and polypeptides is able to produce primary structure similarity (<25%), the above mentioned peptides or polypeptides with any given sequence-bearing features indicate remarkable topological analogy. protein or even non-protein amino acids in satisfactorily high 3D homology ACE peptide models indicate that the yields and purity [120]. Additionally, synthetic peptides of backbone folds in two helical fragments, one at each metal-binding sequences are widely used in order to probe terminus, where the two binding motifs are sited. Therefore, the structure, the metal-binding affinity and transport the two histidyl side-chains of the HEMGH motif are pathways of heavy metals to target proteins, since their separated by a a -helix turn and their side-chains are parallel. structure closely resembles the native active site of a Therefore, they adopt the desired geometry in order to metallobiomolecule [121-123]. donate the imidazole nitrogens to the zinc coordination Zinc metal is essential for ACE function and substrate sphere. The oxygen atom of the third zinc protein ligand that binding and for this reason our approach to investigate the of glutamate is also found in favourable geometry for zinc conformational features of ACE catalytic cavity begins with 414 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al.

Fig. (5). (A) Sequence alignment of the polypeptide fragments comprising the zinc active sites of ACE and representative gluzinicins of which 3D X-ray structures are available. (B) TLN zinc-containing active site polypeptide (32 residue X-ray, pdb: 1lnd) and 3D homology models of ACEC/ACEN 36-residue peptides. Side chains of zinc ligands and F/Y, M/L, A/S and E/R variable ACEC/ACEN residues are shown in ball and stick representation for all models while the four different residues in ACEC and ACEN sequences are also illustrated. (C) Prediction of secondary structure at NPSA server through GOR protocols. the solid-phase synthesis of 36-residue peptide whose amino binding sequences (ACEN and ACEC refer to zinc sites acid composition and sequence represent the ACE active site towards N- and C- terminus respectively; Fig. (5C)). fragment. ACE zinc-binding sequence comprises a 29- The amino acid sequence of the above ACE and ACE residue fragment which contains the three proposed protein N C peptides was built “step by step” on the acid-sensitive 2- ligands; His361/959, His365/963 and Glu389/987 for the chlorotrityl chloride resin (substitution 0.6 mmol/g) applying two zinc-sites, found at MG and AIG (somatic HE H E D the Fmoc strategy [124-126]. Final purification was achieved isoform numbering). 36-residue synthetic peptides represent by semipreparative HPLC on a RP C-18 support (Phase Sep the ACE [His361-Ala396] and ACE [His959-Ala994] zinc- N C C-18 S10 ODS2) eluted with a linear gradient 20% to 60% Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 415 acetonitrile (0.1% TFA) over 30 min at a 2 ml/min flow rate combined analysis of the TOSCY [128,129] (Fig. (6A-B)) [127]. The final products were determined to be at least 96% and NOESY [130,131] spectra. His1 is the only amino acid pure by analytical HPLC. The overall yield was 48% for where backbone and aliphatic chain protons have not been both 36-residue constructs (Scheme 2). identified. However, the characteristic cross peak between These peptides share 89% sequence identity and are the the non-exchangeable protons of His1 imidazole ring has main fragment of the ACE catalytic sites region (see Table been assigned. No HN proton resonance was identified for Glu2. Assignment and chemical shifts are given in Table ( ). (3)). Active site reconstitution has been performed through 4 addition of ZnCl2 in a peptide solution and metal binding Fig. (6C) shows the short and medium range NOE b properties are monitored in solution. observed for the backbone and C H protons in the ACEN NOESY maps. There are three regions with diagnostic 5. CONFORMATIONAL STUDY OF ACEN connectivities for helix conformation, such as HN-HN(i,i + CATALYTIC SITE IN SOLUTION 2), Ha -HN(i,i + 2), Ha -HN(i,i + 3) and Ha -Hb(i,i + 3), that 1 were observed. Met3-Tyr12, Val17-Gly22 and Gly26-Ala36 5.1. H NMR Spectroscopy of ACEN[His361-Ala396] 36- residue Zinc-Free Peptide define these regions in both peptides. On the other hand Ha - HN(i,i + 4) NOE are observed only in the first of the above- 5.1.1. NOE and Secondary Structure mentioned regions. This is due to signal overlapping which The complete spin-system of 35 out of 36 residues and does not allow unambiguous assignment of such type of sequential assignment has been accomplished through the NOE for the Val17-Ala36 region.

Scheme 2. Flow chart for solid phase, step-by-step, synthesis and purification procedure for the ACE 36-residue constructs (HOBt, 1- hydroxybenzotriazole; DIC, N,N’-diisopropylcarbodiimide; Pip, pideridine; EDT, 1,2-ethanedithiol, TFA, trifluoroacetic acid). 416 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al.

Table 3. Amino Acid Composition of the Active Sites’ 44 Residue Polypeptide Containing One or Two Zinc Binding Motifs of ANGIOTENSIN-I Converting Enzyme Known Sequences. Dark-gray Columns Indicate the Metal Binding Ligand According to the SWISS-PROT Record.

Sequence 1st B. M.1 Spacer 2nd B. M.1 AA2

ACE_HUMAN3 VAH H390 EMG H394 IQYYLQYKDLPVSLRRGANPGFH E418 AIGD VLALS 1306

“ --- H988 --- H992 ---FM------A--E------E1016 ------

ACE_MOUSE3 TVH H395 EMG H399 VQYYMQYKDLHVSLRRGANPGFH E423 AIGD VLALS 1312

“ IA- H993 --- H997 I--F------P-TF-E------E1021 ---- IM---

ACE_CHICK3 TVH H288 EMG H292 VQYYLQYKDQPVSFRGGANPGFH E316 AIGD VLSLS 1193

“ --- H886 --- H890 ---F---M-Q-I---D------E913 ---- -MA--

ACE_RABIT3,4 TVH H395 EMG H399 VQYYLQYKDQPVSLRR- ANPGFH E422 AIGD VLALS 1310

“ V-- H992 --- H996 I--FM----L--A--EG------E1020 ------

ACE_RAT3 TVH H396 EMG H400 VQYYLQYKDLHVSLRRGANPGFH E424 AIGD VLALS 1310

“ IA- H994 --- H998 I--FM-----P-TF------E1022 ------

ACE_DROME5 TVH H367 ELG H371 IQYFLQYQHQPFVYRTGANPGFH E395 AVGD VLSLS 615

ACE_HAEIE6 --- H367 -M- H371 ------E395 ------615

ACET_HUMAN VAH H414 EMG H418 IQYFMQYKDLPVALREGANPGFH E442 AIGD VLALS 732

ACET_MOUSE I-- H413 --- H417 ------TF------E441 ---- IM--- 732

ACET_RABIT VV- H419 --- H423 ------E447 ------737

Cons. Seq. XXH H EXG H XQYXXQYXXXXXXXRXGANPGFH E AXGD XXXLS 1 Amino acids belong to 1st and 2nd binding motif. Numbering is according to the sequence record deposited at Swiss-Prot (http://www.expasy.ch/). 2 Number of amino acids found in the sequence. 3 Somatic ACE isoforms possess two zinc-containing active sites, while testis isoform contains only one. 4 A residue is missing at the position of conserved Gly residue in all the other ACE sequences and noted by - 5 Drosophila melanogaster (Fruit fly).

3 5.1.2. JHNHa Coupling Constants configuration. There are three distinct regions where the Dd values are negative suggesting helical character (Ile6- Spin-spin coupling constant values could also be of great Ha importance for the diagnosis of the secondary polypeptide Lys13, Val17-Gly22 and Gly26-Val34) and three shorter regions where the Dd values are positive suggesting structure. The observed J depends on the average of the Js of Ha extended conformation (Glu2-His5, Asp14-Pro16 and each available conformation multiplied by its statistical Ala23-Pro25). The most negative Dd value is measured in weight, and a - and 3 - helical structures exhibit experimental Ha 10 both peptides for Val17 (~ 0.4 ppm) while Asn24 exhibits J values in the range of 4.8 to 5.6 [132], somewhat larger the larger positive Dd value (~ 0.3 ppm) (Fig. ( )). than the ideal values of 3.9 and 4.2 Hz respectively. On the Ha 6D 3 other hand, extended structures such as the parallel or anti- An overall evaluation of the observed NOE, JHNHa and parallel b sheet give rise to J values larger than 8.0 Hz. Most DdHa values, implies that in three fragments the skeleton of 3 (14 out of 18 JHNHa values) of the J coupling constants the 36-residue ACEN active site model peptide adopts a well- measured in ACE peptides exhibits values below 6.2 Hz (see defined a -helical structure even in the absence of zinc metal Fig. (6C)) [133,134]. [138]. The longer helices are anticipated for their C- and N- termini, and a smaller one for the intermediate region. The 5.1.3. Chemical Shift Index two prolines, at positions 17 and 25, act as helix-stop The strong relationship between the backbone residues between the three helices. The second binding motif conformation and chemical shift Ha values serves as a strong EAIGD seems to be part of the C-terminal helix while no indicator for the assignment of secondary structure in any safe conclusion could be extracted for the first binding motif, polypeptide sequence [135,136]. Chemical shift difference HEMGH, which comprise the N-terminal pentapeptide. analysis between the observed Ha shift values and the Taking into account the flexibility of any terminal peptide corresponding random coil values [137] is presented in Fig. fragment and the fact that resonances for all protons for His1 (6D) and provides strong evidence for the conformational and HN proton of Glu2 have not been observed the preference of the majority of amino acids towards the helical conformation of this pentapeptide should be the average over Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 417

1 1 Fig. (6). H- H 2D TOCSY 600-MHz NMR of ACEN (A & B) and ACEC (F & G) fingerprint region of H_-HN protons (H2O/TFE-d2 34%/66% v/v, at pH=5.0, T= 298 K). The number of the amino acid in the ACE sequences, to which the H_-HN and side-chain proton connectivities belong, is noted. (C) Sequential connectivities for ACEN are presented together with the predicted secondary structure 3 elements. JHNH_ coupling constants are illustrated by arrows (¯ for values in the range 4.2 –5.6, • for values equal or above 8.0 Hz) or by filled circles (values in the range of 6.0-7.0 Hz). (D) DdH_ chemical shift difference (in ppm) from random coil values according to CSI. The NMR data were acquired on a Bruker AVANCE 600 spectrometer. (E) Number of meaningful NOE constraints per residue for ACEN. White, grey, and dark grey bars respectively represent intraresidue, sequential, and medium-range connectivities. Long-range connectivities have not been observed. Schematic presentation of the predicted secondary structures according to the sequential connectivities is shown at the bottom of panel (C) and according to CSI at the top of panel (D). various conformers. However, the positive DdHa values are constraints. The resulting DYANA family of 30 structures relatively large for Glu2 and His5 (>0.10 ppm) and close to has rmsd values (calculated for residues 3-33) of 0.57 ± 0.24 the random coil values for Met3 and Gly4. The J values, Å and 1.25 ± 0.21 Å respectively, for backbone and heavy feasibly measured for Met3 and His5, found £ 6.2 Hz for atoms. The target function lies in the range 0.39-0.45 Å2 2 ACEC, the same holding for Met3. These data are (0.43 ± 0.018 Å ). The final REM [141,142] family exhibits controversial with DdHa values, but are in great agreement pairwise rmsd values for the 30 structures 0.57 ± 0.24 Å, with the NOE sequential connectivities illustrated at Fig. 1.27 ± 0.21 Å and to the mean structure 0.40 ± 0.18 Å, 0.89 (6C). According to the observed NOE, almost all the helix- ± 0.11 Å for backbone and heavy atoms respectively (Fig. diagnostic connectivities of Ha -HN(i,i + 2), Ha -HN(i,i + 3), (6E)). Restraint violations and structural and energetic Ha -Hb(i,i + 3), and Ha -HN(i,i + 4) type, involving the statistics for the ACEN[His361-Ala396] 36-residue Zinc-Free Met3-Gly4-His5 amino acids, have been observed in peptide are reported at Table (5). NOESY maps. These data suggest that these three residues The ACE polypeptide chain (Fig. (7)) is characterized of the N-terminal binding motif are in -helical structure and N a by the high content of helical structure, which is distributed possibly comprise the initial turn of the first 10-12-residue in three fragments; the a -helical N- and C- terminal together helix of ACEN/C. with the 310-helix observed in the center of the intermediate fragment between the two other helices. Two turns of the 5.2. High Resolution NMR Solution Structures of peptide skeleton, the first after the end of the N-terminal ACE [His361-Ala396] 36-residue Zinc-Free Peptide N helix in the region of Asp14-Pro16 and the second after the DYANA [139,140] structure calculations have been intermediate helix in the region of Asn24-Pro25 creates a U- performed using 22.0 NOE-derived distance constraints per shaped cavity in the middle of the peptide sequence. This U- residue (>15 meaningful) together with 20 H-bond distance turn structure is comprised of the residues Lys13-Pro25 and constraints (two distance limits for each H-bond) and 18 j brings the two terminal helices to a distance, which varies 418 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al.

Table 4. Chemical Shifts (ppm) of the Protons of the Residues in the [1-36]ACEN at 298K (H2O/TFE-d2 34%/66% v/v, pH=4.9).

Residue HN Ha Hb other

1 His Hd2 8.263; He1 7.298

2 Glu 4.457 2.184, 2.040 Hg 2.389

3 Met 8.692 4.539 2.150, 2.094 Hg 2.618

4 Gly 8.385 3.994

5 His 8.214 4.743 3.370, 3.323 Hd2 8.457; He1 7.246

6 Ile 8.032 4.071 2.039 Hg 1.628, 1.313; gCH3 1.010; dCH3 0.963

7 Gln 8.511 4.029 2.115, 2.081 Hg 2.392 ; dNH2 6.554, 7.141

8 Tyr 7.785 4.268 3.088 Hd 6.948; He 6.754

9 Tyr 7.823 4.311 3.253, 3.158 Hd 7.177; He 6.881

10 Leu 8.202 4.052 1.945, 1.874 Hg 1.526; dCH3 0.930

11 Gln 7.780 4.123 1.985, 1.890 Hg 2.191, 2.138 ; dNH2 6.389, 6.975 12 Tyr 7.850 4.426 3.135, 2.819 Hd 7.015; He 6.775

13 Lys 7.830 4.082 1.795, 1.751 Hg 1.398; Hd 1.624; eCH3 3.032

14 Asp 7.915 4.809 2.875, 2.657

15 Leu 7.778 4.532 1.779 Hg 1.597; dCH3 0.970, 0.943

16 Pro 4.450 2.316, 2.105 Hg 2.041, 1.967; Hd 3.862, 3.653

17 Val 7.688 3.815 2.179 gCH3 1.053, 0.999

18 Ser 7.999 4.240 3.980, 3.924

19 Leu 7.796 4.345 1.733 Hg 1.650; dCH3 0.936, 0.891

20 Arg 7.754 4.207 1.929 Hg 1.748, 1.628 ; Hd 3.169; He 7.063

21 Arg 8.016 4.278 1.895, 1.835 Hg 1.744, 1.647 ; Hd 3.174; He 7.153

22 Gly 7.953 3.959

23 Ala 7.780 4.380 1.416

24 Asn 7.956 5.020 2.931, 2.762 dNH2 6.613, 7.443

25 Pro 4.444 2.387, 2.319 Hg 2.054, 1.990; Hd 3.900, 3.804

26 Gly 8.331 3.931

27 Phe 7.816 4.512 3.153, 3.117 Hd 7.180; He 7.299; Hz 7.249

28 His 8.053 4.458 3.306, 3.223 Hd2 8.493; He1 7.303

29 Glu 8.244 4.228 2.123, 2.063 Hg 2.394

30 Ala 8.097 4.323 1.457

31 Ile 7.722 4.091 1.849 Hg 1.453, 1.196; gCH3 0.883; dCH3 0.830 32 Gly 7.952 3.954, 3.863

33 Asp 7.960 4.728 2.810

34 Val 7.759 4.099 2.226 gCH3 0.990

35 Leu 7.795 4.400 1.700 Hg 1.629; dCH3 0.918, 0.880

36 Ala 7.476 4.179 1.402 Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 419

a a Table 5. Statistical Analysis for the REM and Structures of [1-36]ACEN.

REM

RMS Violations per Experimental Distance Constraints (Å) b

intraresidue (128) 0.0213 ± 0.0030 0.0215

sequential (253) 0.0258 ± 0.0018 0.0311

medium-range (173) 0.0265 ± 0.0015 0.0237

total (544) 0.0251 ± 0.0013 0.0269

Average Number of Violations per Structure

intraresidue 7.00 ± 1.84 10.0

sequential 17.67 ± 2.05 17.0

medium-range 14.07 ± 1.39 14.0

total 38.73 ± 2.94 41.0

average no. of NOE violations > than 0.3 Å 0.000 ± 0.00 00.0

largest residual NOE distance violation (Å) 0.203 0.259

average distance penalty function (Å2) 0.415 ± 0.03 0.466

Statistics of Other Structural Constraints

3 j constraints from JHNHa (18) RMS violations per j constraint 0.0627 ± 0.24

average no. of j violations per structure 0.0667 ± 0.25

largest residual j violations 0.0062

average torsion penalty function (kJ mol-1) 0.0008 ± 0.003 0.0004

AMBER energy (kJ mol-1) -1078.44 ± 59.8 -1107.12 a REM indicates the energy-minimized family of 40 structures and the mean energy-minimized structure. b Numbers in parenthesis indicate the number of meaningful upper distance limits per class.

Fig. (7). Backbone representation of: (A) Family of 30 ACEN REM models and (B) mean REM structure. 420 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al. from 13.5 to 24.5 Å (between the C_ atoms of Gln7 and As far as the structure of the two potential binding motifs His28, the carbon atom of -COO- and the nitrogen atom of - sited at the N- and C- terminus of the ACE model peptide is NH3, respectively), one above the other, while their dipoles concerned, they are found in helix-like or helical fragments form an angle of almost 90°. The helical fragments (Fig. (8)). The helical conformation of the C- terminus has comprised by 12 (Glu2-Lys13) and 8 (Phe27-Val34) residues been unambiguously determined and supported by all NMR for the N- and C- termini, respectively, and by 3 residues experimental data, such as helix-diagnostic sequential (Arg20-Gly22) for the intermediate [138]. NOE-connectivities (dNN(i,i+2), daN(i,i+2), daN(i,i+3) and da(i,i+3), see Fig. (6C)), J-coupling constants and chemical The structure of the C-terminal fragment has been shift analysis (Fig. (6D)). On the other hand helix-like determined with higher resolution than that of the N- conformation for the N- terminus resulted for the mean, terminal, due to the higher number of NOE and angle energy-minimized, NMR structure supported by some constraints. No proton resonances have been identified for sequential NOE typical for helix conformation, observed for the backbone and side-chain aliphatic protons of His1 and the N-terminus starting from Met3 (see Fig. (6C)). Apart amide proton of Glu2, probably due to the conformational from loss of NOE information, the only contradiction for a averaging and/or exchange with solvents of labile NH helix structure in this region arises with CSI, and the positive protons. The two prolines, in position 16 and 25 of the DdHa values in the Glu2-His5 (Fig. (6D)). Helix structure has sequence, found in the intervening residues between the been calculated for the Ile6-Lys13 residues and all NMR helices probably interrupt the helicity of the backbone as data is fully consistent with this conformation. suggested also by sequential NOE and DdHa values. Addit- At this point we should note that this N-terminal motif ionally, next to Pro25 is found Gly26 and these two residues HEMGH sequence comprise the well-known zincins’ first comprise a notorious residue-pair with helix-destabilizing zinc-binding motif which is always found in all 3D crystal ability, since both one characterized by the lowest helix structures to be a helix, the so-called “active-site helix”. This propensity among the 20 natural amino acids [143,144]. ambiguity in determination of this binding motif conformation

Fig. (8). Colour-coded chemical shift perturbation mapping in identification of conformational changes in free ACEN mean NMR structure. Shift differentiation is illustrated: (A) between the ACEN and ACEC sequences which differs in 4 residues: Tyr9Phe, Leu10Met, Ser18Ala and Arg21Glu, (B) when zinc metal is added to the solution of free peptide and (C) the DdH_ (CSI) chemical shift differences are indicative of the zinc-bound peptide’s secondary structure (grey for DdH_ > 0; dark grey for DdH_ < 0 suggesting _-helix conformation). Chemical shift differences in bar diagrams are also presented at the bottom. The NMR data were acquired on a Bruker AVANCE 600 spectrometer. Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 421 strongly depends on the fact that the first histidine of the through 1H 1D and 2D NMR Spectroscopy and many proton motif is the first peptide residue. However, even in this case resonances in the NH region shift considerably and their the mean NMR solution structure of the free peptides fits half-width is considerably increased when Zn(II) is added well with the “two active-site helices” model suggested for (Fig. (9A) & (9D)). Additionally, the previously degenerated gluzincins. Both binding motifs are in helix or helix-like resonance of the two geminal Gly Ha protons has been split conformation. The side-chains of the two histidines are into two well-defined resonances (Fig. (9B) & (9C). This parallel, the same occurring with side-chains of the third resonance different-iation for the previously degenerated Ha protein zinc-ligand, the glutamate, and the aspartate four glycine protons suggests a metal-coupled structural change residues after glutamate towards the C-terminus, whose of the N-terminal motif and indicates a different magnetic conformation plays a crucial role in the catalytic activity of environment for those nuclei in metal-peptides from those in some gluzincins. their free state. NMR data suggest similar elements of secondary structure between free and zinc-bound peptides 5.3. Analysis of the Tyr9Phe, Leu10Met, Ser18Ala and [146]. Arg21Glu Differentiation Between ACE and ACE N C The Ha values of residue fragment with negative or Through Chemical Shift Perturbation Mapping positive DdHa values in free peptides exhibit the same nature The two ACE 36-residue peptides designed and and magnitude of deviation from their corresponding random synthesized differ, as actually happens in native sequence, in coil values. In the presence of Zn(II) the majority of residues four residues (89% sequence identity), found in positions 9, (23 residues) still exhibit negative deviation (larger than 0.05

10, 18 and 21. The aromatic and aliphatic character of amino ppm) from random coil values (Fig. (8C)). DdHa values in acids in positions 9 and 10 are in general maintained ACEN-Zn peptide define three possibly helical regions : (i) a between the two sequences, due to Tyr/Phe and Leu/Met 7-residue fragment close to the peptide N-terminus (Ile6- substitution. On the other hand, the polar Ser18 which is Lys13), (ii) another 6-residue fragment in the middle of the considered a “surface” residue is replaced by the non-polar, sequence (Val17-Arg21), and (iii) a third comprising of 10 “internal”, Ala. Additionally, the basic, positively charged residues close to peptide C-terminus (Gly26-Val34). These Arg21 has been replaced by the acidic, negatively charged, data strongly suggest that in the zinc-bound peptide, the glutamic acid without perturbing the hydrophilic properties second proposed binding motif (EAIGD) retains its helical at this point of the sequence. These residue variations conformation, also observed for the free peptides. As far as between the two ACE peptides do not seem to provoke any the two histidyl motif is concerned, no definite conclusion noticeable conformational change in the secondary structure could be reached [146]. In the TLN [27] crystal structure, of the studied biopolymers, since according to NMR data that binding motif, which possesses similar sequence to that (sequential connectivities and DdHa values) the helical of ACE, has been found in helical conformation. character of the fragments where the four residues reside is However, helix-diagnostic ( + 3) type connectivity not disturbed. da N i between Gly4 and Gln7, present in the NOESY spectrum of However, comparative analysis of proton resonances free peptides disappears when Zn(II) is added (Fig. (9B)). through Chemical Shift Perturbation Mapping [145], provides Nevertheless, typical helix da N(i + 4) and da N(i + 3) connect- DdHa -HN values between 0.15 and 0.20 ppm. These differences ivities between Ile6-Met10 and Gln7-Met10 respectively, in are mapped in the backbone of the 3D average NMR free peptides are still present after zinc addition [146]. These structure of ACEN peptide and presented in different colors data strongly suggest that the HEMGH motif conformation at the top and center of Fig. (8A). Chemical shift changes might undergo a conformational transition from helix-like or around Tyr/Phe9 and Leu/Met10 are moderate and slightly partial helix to non-helix structure when zinc coordinates exceed 0.05 ppm, while the relative changes for the peptide with the donor atoms of the peptide ligands. This could fragment Leu15-Arg/Glu21 fluctuate between 0.07 and 0.17 probably be due to the fact that the first peptide zinc ligand is ppm. Additionally, Asn24 present a chemical shift difference also the first residue in the peptide sequence. Upon the lack value of around 0.075 ppm. Differences around 0.05 ppm of other residue(s), the His1 backbone and side-chain possess have also been estimated for the His28-Ala30 tripeptide. a remarkably high degree of conformational freedom. Thus, Plots of the Ha and HN shift differences are also presented when zinc binds the peptide bonds of the pentapeptide at the bottom of Fig. (8A). Assignment and chemical shifts binding sequence HEMGH, it could be accommodated in a for the 36-residue peptide ACEC are given in Table (6). random, open coil segment. Such conformation justifies the absence of the typical sequential NOE for a helix and has 1 5.4. H NMR Spectroscopy, Zn(II) Binding Properties resulted after preliminary DYANA structure calculation for and Conformational Features of ACEN[His361-Ala396] ACEN 36-residue peptide. On the other hand, no changes for 36-residue Peptide the fragment Ile6-Lys13 are implied by NMR data. 5.4.1. NOE and Secondary Structure – Differences with Moreover, the same type as for free peptides sequential NOE Free Peptides has been detected and structure calculation indicates that the first HEMGH open coil structure in ACEN follows a helical Color-coded chemical shift changes, upon Zn(II) addition, fragment which is extended from Ile6 to Tyr12/Lys13. are presented in Fig. (8B) while chemical shift difference Another difference between free and zinc-bound peptides, analysis between the observed Ha shift values and their which has also been identified through sequential connect- corresponding random coil values are presented in Fig. (8C). ivities diagram and preliminary structure calculations, is the Zinc binding properties of ACE peptides are monitored absence of the helical fragment in the area Ser18-Arg21 in 422 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al.

Table 6. Chemical Shifts (ppm) of the Protons of the Residues in the [1-36]ACEC at 298K (H2O/TFE-d2 34%/66% v/v, pH=4.9).

Residue HN Ha Hb other

1 His Hd2 8.349 ; He1 7.342

2 Glu 4.467 2.188, 2.050 Hg 2.407

3 Met 8.668 4.557 2.154, 2.095 Hg 2.628

4 Gly 8.362 3.994

5 His 8.252 4.719 3.355, 3.315 Hd2 8.485; He1 7.232

6 Ile 8.043 4.063 2.026 Hg 1.628, 1.301; gCH3 1.006; dCH3 0.954

7 Gln 8.443 4.099 2.098, 2.079 Hg 2.397 ; dNH2 6.564, 7.160

8 Tyr 7.841 4.292 3.077 Hd 6.936; He 6.748

9 Phe 7.950 4.377 3.312, 3.229 Hd 7.322; He 7.360; Hz 7.344

10 Met 8.283 4.225 2.200, 2.104 Hg 2.743, 2.692

11 Gln 7.804 4.149 1.977, 1.903 Hg 2.170 ; dNH2 6.384, 6.987

12 Tyr 7.869 4.413 3.129, 2.819 Hd 7.010; He 6.751

13 Lys 7.796 4.069 1.790, 1.745 Hg 1.398; Hd 1.639; eCH3 3.040

14 Asp 7.881 4.822 2.906, 2.690

15 Leu 7.772 4.451 1.773 Hg 1.646; dCH3 0.971, 0.938 16 Pro 4.384 2.394, 2.130 Hg 2.048, 1.890; Hd 3.874, 3.650

17 Val 7.579 3.690 2.169 g CH3 1.062, 0.981

18 Ala 7.920 4.083 1.452

19 Leu 7.900 4.240 1.766 Hg 1.595; d CH3 0.917, 0.884

20 Arg 7.686 4.150 1.950 Hg 1.754, 1.633; Hd 3.156; He 7.090

21 Glu 8.224 4.294 2.096, 1.985 Hg 2.418, 2.387

22 Gly 7.936 3.938

23 Ala 7.766 4.390 1.434

24 Asn 7.888 5.028 3.002, 2.798 dNH2 6.598, 7.498

25 Pro 4.450 2.343 Hg 2.048, 1.999; Hd 3.908, 3.835

26 Gly 8.341 3.930

27 Phe 7.818 4.488 3.156 Hd 7.151; He 7.276; Hz 7.234

28 His 8.040 4.415 3.314, 3.255 Hd2 8.534; He1 7.343

29 Glu 8.198 4.212 2.125, 2.082 Hg 2.418

30 Ala 8.059 4.292 1.455

31 Ile 7.718 4.069 1.826 Hg 1.417, 1.172; CH3 0.856; d CH3 0.803

32 Gly 7.925 3.939, 3.867

33 Asp 7.961 4.728 2.827

34 Val 7.752 4.090 2.225 g CH3 0.999

35 Leu 7.758 4.397 1.707 Hg 1.638; d CH3 0.917, 0.878 36 Ala 7.491 4.194 1.409 Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 423

Fig. (9). Downfield regions of 600 MHz 1H 2D NMR NOESY (upper panel) and 1H 1D NMR spectra with characteristic His imidazole ring and Ha -HN cross-peaks of ACE peptide before (A & B) and after Zn(II) addition (C & D) in solution. Intraresidue Ha -HN cross-peaks of selected N-terminal residues are illustrated in dark grey and black, while interesidue d_N(i + 1,3,4) cross-peaks are illustrated in grey.

ACEN zinc peptides. The majority of HN-HN and Ha -HN His5 protons resonance exhibits a remarkably large sequential connectivities identified in free ACEN have not difference while the largest one is identified for Met3. been detected after zinc addition. However, Ha -Hb(i + 3) Unfortunately, no proton resonance was identified for His1 connectivity, present in free peptide is still detectable and for the amide proton of Glu2. between Asp14 and Val17. Other helix-indicative sequential According to the data presented above, it is highly connectivities such as H -HN( + 2) between Val17-Leu19, a i possible that zinc coordination has a dual effect on the Arg21-Ala23 and Gly22-Asn24 together with H -HN( + 3) a i peptides’ N-terminal free-to-bound structure differentiation: type NOE between Arg20-Ala23 have also been detected for (i) it provokes a structural transition from free peptide the intervening fragment between the two terminus helices conformation, helix or helix-like, to a well-defined but non- [146]. The limited number of sequential, helix-type helical structure of zinc-bound peptide, and (ii) diminishes connectivities, could be related to the fact that many of these the conformational flexibility of this fragment leading to an protons have been found to resonate in similar field values ordered structure where residues adopt the desired geometry possessing almost identical chemical shifts. prior zinc coordination. Other significant chemical shift differences are identified in the region covered by Leu15- 5.4.2. Metal-induced Chemical Shift Perturbation & Zinc- Gly22, which constitute the intermediate fragment of 23- binding Implication residue spacer between the two proposed binding motifs. Chemical shift perturbation mapping [145] provides This HN and Ha proton shift value differentiation is valuable insight for Zn(II) induced structural changes when probably due to the fact that zinc coordination forces the two the metal is added to a peptide solution. The largest changes peptide termini to approach one another in order the donor were identified for the N- and C- terminal decapeptide atoms of the potential peptide ligands sited at N- and C- ends containing the HEMGH and EAIGD sequences where to achieve a four-ligand distorted-tetrahedral coordination potential zinc ligands are located. The proton chemical shift geometry. This kind of geometry is that most frequently differences for Glu29 are the largest throughout the peptide encountered in other members of the zinc metallopeptidase Gln7-Ala36 region, suggesting its coordination to zinc. The family and bend polypeptide fragments have been identified 424 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al. in other gluzincs, such as TLN. Such a model is consistent TLN and the two helices approach each other. The with a bend peptide skeleton of a U-like shape, as has been NMR structure of the ACEN free peptide reveals that observed in the case of TLN X-ray crystal structure [146]. the potential zinc ligands could bind the metal without severely disturbing the elements of its tertiary and 6. CONCLUSIONS secondary structure. The conformation of the amino 6.1. Common and Variable Structural Features Between acid spacer (310 helix fragment) and the relative Free ACE Peptides, 3D Homology ACE Models and orientation of the two helices could be substantially Catalytic Sites of Known Gluzincin Structures differentiated upon zinc binding to the polypeptide, and the peptide structure could further resemble that Having gone through theoretical and experimental studies of the models and TLN. in order to elucidate the solution structure of the ACE zinc- containing active sites various valuable structural viii. The conformation of the 23-residue spacer could information for these biologically interesting catalytic either retain or perturb the 310 helix conformation of centres has been extracted: the Arg20-Gly22 fragment observed in free peptide NMR structure. The small negative value exhibited by i. Despite the low sequence similarities among known Leu19 (see Fig. (8C) bottom) could be considered sequences of gluzincins, all the available X-ray crystal within the error of the CSI method and if that happens structures possess a characteristic “two helix active- would indicate a helix-stop signal for this peptide site”, no matter what the length of the inter-helical fragment. Nevertheless, a non-helix, coil-like structure amino acid spacer (Fig. (3)). is consistent with TLN active site and helix structure ii. Theoretical study of the secondary structure of the 36- could be also consistent with the Neurolysin structure, residue active site model ACE peptides predicts which possess a spacer similar to ACE 24-residue helical conformation for the N- and C- terminus, with a small helical fragment (see also Fig. (3)). where the two zinc-binding motif sequences are sited ix. One of the main structure-property related differences (HEMGH and EAIGD, respectively, Fig. (5C)). between the TLN and ACE active centre is the lack of iii. 3D homology models of these 36-residue peptides any positively-charged residue in the TLN possess two helices, one at each peptide terminus, polypeptide fragment, which comprise its zinc- suggesting that the peptide bonds could be containing catalytic site (Fig. (10), lower panel). accommodated in conformation with high helical Electrostatic potentials not only in the surface but also degree (Fig. (5B)). These helices are found to form an in the substrate channel and cavity are crucial for the angle similar to that of TLN helices (Fig. (10)). molecular recognition and complex formation process between enzyme and substrate. iv. Experimental study of the free peptides in solution 1 through high resolution H NMR spectroscopy, 6.2. Outlook and Perspectives followed by structure calculation using NMR-derived 3 According to the above data, free and zinc peptides structural constraints (NOEs, JHNHa ) in concert with CSI suggest helix conformation for the N- and C- exhibit conformational features highly similar to the zinc- terminus (Fig. (6C-E), (7) and (8)). containing active sites of various known zinc metallopeptidases that belong to the same super-family as ACE, the v. Minor conformational differences between free and gluzincins. The structural calculation based on NMR-derived zinc-bound peptides could be identified for the constraints for the two ACE-Zn peptides is under way and secondary structure of the first binding motif the models of both zinc-containing ACE catalytic sites will sequence, constituting the first pentapeptide fragment. become available in the form of 36-residue peptides. These The conformational freedom of the first peptide zinc- ligand, His1, could be responsible for this structural models could be used in enzyme-substrate complex variation, for which however the helical structure has simulation/docking studies in an effort to understand the also been predicted (Fig. (5C)). factors governing the association of the two molecules. vi. Remarkable backbone conformational similarities The structural information acquired for ACE active sites could be identified even between the free model through NMR spectroscopy and structure calculation indicate peptides studied in solution and the structure of that the multidimensional process based on design, solid- TLN’s active site in solid state. Among them the two phase synthesis and conformational analysis of ACE metal- helices in homologous regions of the polypeptide binding sequences seems to be able to provide a suitable chains such as the binding motifs (the N- and C- maquette of the actual ACE catalytic sites. Valuable inform- ACEN peptide termini) are of immense interest (Fig. ation could be acquired and would be exploited in structure- (10), upper panel). based design of biologically interesting substances against hypertension, until the three-dimensional structure of the vii. The NMR solution structure of the free ACEN peptide, the 3D homology ACE-Zn models and the TLN active entire Angiotensin-I Converting Enzyme becomes available. site structure exhibit striking similarities. Data suggest Note added in proof that even in the absence of metal ion the synthesised polypeptide folds in a tertiary structure where the While this paper was in press, the crystal structures of human helical content is comparable to that of the model and testicular [148] and Drosophila [149] ACE were reported. Additionally, shortly before the release of ACE model, the Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 425

Fig. (10). Comparison between the ribbon representation of the 3D structures of zinc active sites of (A) Thermolysin, (B), ACEC (C) ACEN 3D homology models, and (D) NMR structure is presented in upper panel. The helix angles calculated in terms of helix dipoles are also given. The distribution of the electrostatic potential on the surface of the models is also presented for the peptides in lower panel. Figures were generated with the program MOLMOL [147].

Xray structure of another gluzincin metalloprotease, that of EC’s Access to Research Infrastructures Action of the Carboxypeptidase Pyrococcus Furiosus (M32 clan), had Improving Human Potential Program (PARABIO, Contract been determined [150]. No. HPRI-CT-1999-00009) for further support (G.A.S., P.C. and A.S.G.). ACKNOWLEDGEMENTS. REFERENCES University of Patras for a K. Karatheodoris Research Grant (P.C., G.A.S.) and General Secretariat of Research and [1] Inagami, T. The renin–angiotensin system. Essays Biochem. 1994, 28, 147–164. Technology of Greece-Pened 99 Program (G.P., E.M.-Z.) [2] Ondetti, M. A.; Cushman, D. W. Enzymes of the renin-angiotensin are acknowledged for financial support. We also thank system and their inhibitors. Annu. Rev. Biochem. 1982, 51, 283-308. 426 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al.

[3] Roks, A.; Buikema, H.; Pinto, Y. M.; van Gilst, W. H. The renin- [27] Holland, D. R.; Hausrath, A. C.; Juers, D.; Matthews, B. W. angiotensin system and vascular function. The role of angiotensin Structural analysis of zinc substitutions in the active site of II, angiotensin-converting enzyme, and alternative conversion of thermolysin. Protein Sci. 1995, 4, 1955-1965. angiotensin I. Heart Vessels 1997, 12, 119-124. [28] Oefner, C.; D'Arcy, A.; Hennig, M.; Winkler, F. K.; Dale, G. E. [4] Skeggs, L. T.; Kahn, J. R.; Shumway, N. P. Preparation and Structure of Human Neutral Endopeptidase (Neprilysin) Complexed function of the hypertension converting enzyme. J. Exp. Med. 1956, with Phosphoramidon. J. Mol. Biol. 2000, 296, 341-349 103, 295-299. [29] Lacy, D. B.; Tepp, W.; Cohen, A. C.; Dasgupta, B. R.; Stevens, R. [5] Soffer, R. L. Angiotensin-converting enzyme and the regulation of C. Crystal Structure of Botulinum Neurotoxin Type a and vasoactive peptides. Annu. Rev. Biochem. 1976;45, 73-94. Implications for Toxicity. Nat. Struct. Biol. 1998, 5, 898-902. [6] Ehlers, M. R.; Riordan, J. F. Angiotensin-converting enzyme: new [30] Pannifer, A. D.; Wong, T. Y.; Schwarzenbacher, R.; Renatus, M.; concepts concerning its biological role. Biochemistry 1989, 28, Petosa, C.; Bienkowska, J.; Lacy, D. B.; Collier, R. J.; Park, S.; 5311-5318. Leppla, S. H.; Hanna, P.; Liddington, R. C. Crystal Structure of the [7] Dorer, F. E.; Kahn, J. R.; Lentz, K. E.; Levine, M.; Skeggs, L. T. Anthrax Lethal Factor. Nature 2001, 414, 229-233 Hydrolysis of bradykinin by angiotensin-converting enzyme. Circ. [31] Kurisu, G.; Kinoshita, T.; Sugimoto, A.; Nagara, A.; Kai, Y.; Kasai, Res. 1974, 24, 824–827. N.; Harada, S. Structure of the zinc endoprotease from [8] Kokkonen, J. O.; Lindstedt, K. A.; Kuoppala, A.; Kovanen, P. T. Streptomyces caespitosus. J. Biochem. 1997, 121, 304-308. Kinin-degrading pathways in the human heart. Trends Cardiovasc. [32] Schlagenhauf, E.; Etges, R.; Metcalf, P. The crystal structure of the Med. 2000, 10, 42-45. Leishmania major surface proteinase leishmanolysin (gp63). [9] Yang, H. Y. Y.; Erdös, E. G.; Levin, V. A dipeptidyl Structure 1998, 6, 1035-1046. carboxypeptidase that converts angiotensin I and inactivates [33] Li, J.; Brick, P.; O'Hare, M. C.; Skarzynski, T.; Lloyd, L. F.; Curry, bradykinin. Biochem. Biophys. Acta 1970, 214, 374-376. V. A.; Clark, I. M.; Bigg, H. F.; Hazleman, B. L.; Cawston, T. E.; [10] Soubrier, F.; Alhenc-Gelas, F.; Hubert, C.; Allegrini, J.; John, M.; Blow, D. M. Structure of full-length porcine synovial collagenase Tregear, G.; Corvol, P. Two putative active centers in human reveals a C-terminal domain containing a calcium-linked, four- angiotensin I-converting enzyme revealed by molecular cloning. bladed beta-propeller. Structure 1995, 3, 541-549. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 9386-9390. [34] Bode, W.; Gomis-Ruth, F. X.; Huber, R.; Zwilling, R.; Stocker, W. [11] Williams, T. A.; Corvol, P.; Soubrier, F. Identification of two active Structure of astacin and implications for activation of astacins and site residues in human angiotensin I-converting enzyme. J. Biol. zinc-ligation of collagenases. Nature 1992, 358, 164-167. Chem. 1994, 269, 29430-29434. [35] Mcauley, K. E.; Jia-Xing, Y.; Dodson, E. J.; Lehmbeck, J.; [12] Skidgel, R. A.; Erdös, E. G. Novel activity of human angiotensin I Ostergaard, P. R.; Wilson, K. S. A Quick Solution: Ab Initio converting enzyme: release of the NH2- and COOH-terminal Structure Determination of a 19 kDa Metalloproteinase Using tripeptides from the luteinizing hormone-releasing hormone. Proc. Acorn Acta Crystallogr., Sect. D 2001, 57, 1571-1578. Natl. Acad. Sci. U.S.A. 1985, 82, 1025-1029. [36] Greenblatt, H. M.; Feinberg, H.; Tucker, P. A.; Shoham, G. [13] Ehlers, M. R.; Riordan, J. F. Angiotensin-converting enzyme: zinc- Carboxypeptidase A: native, zinc-removed and mercury-replaced and inhibitor-binding stoichiometries of the somatic and testis forms. Acta Crystallogr., Sect. D. 1998, 54, 289-305. isozymes. Biochemistry 1991, 30, 7118-7126. [37] 1lbu is the pdb code for the “Refined Structure of a Zn2+ D-Alanyl- [14] Gavras, H. Angiotensin converting enzyme inhibition and its impact D-Alanine Carboxypeptidase of S. Albus G at 1.8 Angstroms on cardiovascular disease. Circulation 1990, 81, 381-388. Resolution” (Wery, J.-P.; Charlier, P.; Dideberg, O.) which refers to [15] Ferreira, S. H.; Bartelt, D. C.; Greene, L. J. Isolation of bradykinin- the original enzyme structure with pdb code 0zgp : Dideberg, O.; potentiating peptides from Bothrops jararaca venom . Biochemistry Charlier, P.; Dive, G.; Joris, B.; Frere, J. M.; Ghuysen, J. M. 1970, 9, 2583-2593. Structure of a Zn2+-containing D-alanyl-D-alanine-cleaving [16] Ondetti, M. A.; Williams, N. J.; Sabo, E. F.; Pluscec, J.; Weaver, E. carboxypeptidase at 2.5 A resolution. Nature 1982 299, 469-470. R.; Kocy, O. Angiotensin-converting enzyme inhibitors from the [38] Taylor, A. B.; Smith, B. S.; Kitada, S.; Kojima, K.; Miyaura, H.; venom of Bothrops jararaca. Isolation, elucidation of structure, and Otwinowski, Z.; Ito, A.; Deisenhofer, J. Crystal Structures of synthesis. Biochemistry 1971, 10, 4033-4039. Mitochondrial Processing Peptidase Reveal the Mode for Specific [17] Cheung, H. S.; Cushman, D. W. Inhibition of homogenous Cleavage of Import Signal Sequences. Structure 2001, 9, 615-625. angiotensin-converting enzyme of rabbit lung by synthetic venom [39] Burley, S. K.; David, P. R.; Taylor, A.; Lipscomb, W. N. Molecular peptides of Bothrops jararaca. Biochim. Biophys. Acta 1973, 293, structure of leucine aminopeptidase at 2.7-A resolution. Proc. Natl. 451-463. Acad. Sci. U.S.A. 1990, 87, 6878-6882. [18] Steitz, T. A.; Ludwig, M. L.; Quiocho, F. A.; Lipscomb, W. N. The [40] Roderick, S. L.; Matthews, B. W. Structure of the cobalt-dependent structure of carboxypeptidase A. V. Studies of enzyme-substrate methionine aminopeptidase from Escherichia coli: a new type of and enzyme-inhibitor complexes at 6 Å resolution. J. Biol. Chem. proteolytic enzyme. Biochemistry 1993, 32, 3907-3912. 1967, 242, 4662-4668. [41] Rowsell, S.; Pauptit, R. A.; Tucker, A. D.; Melton, R. G.; Blow, D. [19] Cushman, D. W.; Ondetti, M. A. Design of angiotensin converting M.; Brick, P. Crystal structure of carboxypeptidase G2, a bacterial enzyme inhibitors. Nat. Med. 1999, 5, 1110-1113. enzyme with applications in cancer therapy. Structure 1997, 5, 337- [20] Vallee, B. L.; Auld, D. S. Zinc Coordination, Function, and 347. Structure of Zinc Enzymes and Other Proteins. Biochemistry 1990, [42] Chevrier, B.; Schalk, C.; D'Orchymont, H.; Rondeau, J. M.; Moras, 29, 5647-5659. D.; Tarnus, C. Crystal structure of Aeromonas proteolytica [21] Vallee, B. L.; Auld, D. S. New Perspective on Zinc Biochemistry: aminopeptidase: a prototypical member of the co-catalytic zinc Cocatalytic Sites in Multi-Zinc Enzymes. Biochemistry 1993, 32, enzyme family. Structure 1994, 2, 283-291. 6493-6500. [43] Pearson, M. A.; Michel, L. O.; Hausinger, R. P.; Karplus, P. A. [22] Vallee, B. L; Auld, D. S. Active-site zinc ligands and activated H2O Structures of Cys319 variants and acetohydroxamate-inhibited of zinc enzymes. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 220-224. Klebsiella aerogenes urease. Biochemistry 1997, 36, 8164-8172. [23] Auld, D. S. Zinc sites in metalloenzymes and related proteins. In [44] Fritsche, E.; Paschos, A.; Beisel, H.-G.; Bock, A.; Huber, R. Crystal Handbook on Metallopreoteins; Bertini, I.; Sigel, A.; Sigel, H., Eds, Structure of the Hydrogenase Maturating Endopeptidase Hybd from Marcel Dekker, 2001, pp. 881-959. Escherichia Coli. J. Mol. Biol. 1999, 288, 989-998. [24] Hooper, N. M. Families of zinc metalloproteases. FEBS Lett. 1994, [45] Laszlo, P.; Blundell, T. L. Sequences and topology. Curr. Opin. 354, 1-6. Struct. Biol. 1995, 5, 357-359 [25] Thunnissen, M. M. G. M.; Nordlund, P. N.; Haeggstrom, J. Z. [46] Jongeneel, C. V.; Bouvier, J.; Bairoch, A. A unique signature Crystal Structure of Human Leukotriene A4 Hydrolase, a identifies a family of zinc-dependent metallopeptidases. FEBS Lett. Bifunctional Enzyme in Inflammation. Nat. Struct. Biol. 2001, 8, 1989, 242, 211-214. 131-135. [47] Vallee, B. L.; Auld, D. S. Short and long spacer sequences and [26] Brown, C. K.; Madauss, K.; Lian, W.; Beck, M. R.; Tolbert, W. D.; other structural features of zinc binding sites in zinc enzymes. Rodgers, D. W. Structure of Neurolysin Reveals a Deep Channel FEBS Lett. 1989, 257, 138-140. that Limits Substrate Access. Proc. Nat. Acad. Sci.U.S.A. 2001, 98, [48] Jiang, W.; Bond, J. S. Families of metalloendopeptidases and their 3127-3132. relationships. FEBS Lett. 1992, 312, 110-114. Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 427

[49] Bode, W.; Gomis-Rüth, F.-X.; Stöckler, W. Astacins, serralysins, physiological systems involved in regulation of blood pressure. snake venom and matrix metalloproteinases exhibit identical zinc- Clin. Chim. Acta. 1971, 31, 413-419. binding environments (HEXXHXXGXXH and Met-turn) and [72] Ondetti, M. A.; Rubin, B.; Cushman, D. W. Design of specific topologies and should be grouped into a common family, the inhibitors of angiotensin-converting enzyme: new class of orally `metzincins'. FEBS Lett. 1993, 331, 134-140. active antihypertensive agents. Science 1977, 196, 441-444. [50] Rawlings, N. D.; Barrett, A. J. Evolutionary families of peptidases. [73] Quiocho, F. A.; Lipscombe, W. N. Carboxypeptidase A: a protein Biochem. J. 1993, 290, 205-218. and an enzyme. Adv. Protein Chem. 1971, 25, 1-78. [51] Blundell, T. L. Metalloproteinase superfamilies and drug design. [74] Pelmenschikov, V.; Blomberg, M. R. A.; Siegbahn, P. E. M. A Nat. Struct. Biol. 1994, 1, 73-75. theoretical study of the mechanism for peptide hydrolysis by [52] Bode, W.; Stöckler, W. Structural features of a superfamily of zinc- thermolysin. J. Biol. Inorg. Chem. 2002, 7, 284-298. endopeptidases: the metzincins. Curr. Opin.. Struct. Biol. 1995, 5, [75] Kester, W. R.; Matthews, B. W. Crystallographic study of the 383-390. binding of dipeptide inhibitors to thermolysin: implications for the [53] Colman, P. M; Jansonius, J. N.; Matthews, B. W. The structure of mechanism of catalysis. Biochemistry 1977, 16, 2506-2516. thermolysin : an electron density map at 2-3 A resolution. J. Mol. [76] Holmes, M. A.; Matthews, B. W. Binding of hydroxamic acid Biol. 1972, 70, 701-24. inhibitors to crystalline thermolysin suggests a pentacoordinate zinc [54] Matthews B.W., Weaver L.H., Kester W.R. The conformation of intermediate in catalysis. Biochemistry 1981, 20, 6912-6920. thermolysin. J. Biol. Chem. 1974, 249, 8030-8044. [77] Hangauer, D. G.; Monzingo A. F.; Matthews, B. W. An interactive [55] Holmes MA, Matthews BW. Structure of thermolysin refined at 1.6 computer graphics study of thermolysin-catalyzed peptide cleavage A resolution. J. Mol. Biol. 1982, 160, 623-639. and inhibition by N-carboxymethyl dipeptides. Biochemistry 1984, [56] Baumann U. Crystal structure of the 50 kDa metallo protease from 23, 5730-5741. Serratia marcescens. J. Mol. Biol. 1994, 242, 244-251. [78] Holden, H. M.; Tronrud, D. E.; Monzingo, A. F.; Weaver, L. H.; [57] Katsuya, Y.; Hamada, K.; Hata, Y.; Tanaka, N.; Sato, M., Katsube, Matthews, B. W. Slow- and fast-binding inhibitors of thermolysin Y.; Kakiuchi, K.; Miyata, K. Preliminary X-ray studies on Serratia display different modes of binding: crystallographic analysis of protease. J. Biochem. 1985, 98, 1139-1142. extended phosphonamidate transition-state analogues. Biochemistry [58] pdb code 1srp is linked to the reference : Hamada, K.; Hiramatsu, 1987, 26, 8542-8553. H.; Katsuya, Y.; Hata, Y.; Matsuura, Y.; Katsube, Y. Structural [79] Tronrud, D. E.; Holden, H. M.; Matthews, B. W. Structures of two Analysis of Serratia Protease. Acta Crystallogr., Sect. A 1993, 49, thermolysin-inhibitor complexes that differ by a single hydrogen 153. However, this citation seems to be a misprint, as it does not bond. Science 1987, 30, 571-574. exist to the corresponding journal. A relative article to serralysin [80] Matthews, B. W. Structural Basis of the Action of Thermolysin and structure should be the following : Katsuya, Y.; Hamada, K.; Hata, Related Zinc Peptidases. Acc. Chem. Res. 1988, 21, 333-340. Y.; Tanaka, N.; Sato, M.; Katsube, Y.; Kakiuchi, K.; Miyata, K.; [81] Jaspard, E.; Wei, L.; Alhenc-Gelas, F. Differences in the properties Preliminary X-ray studies on Serratia protease. J. Biochem. 1985 and enzymatic specificities of the two active sites of angiotensin I- 98, 1139-1142. converting enzyme (kininase II). Studies with bradykinin and other [59] Spurlino, J. C.; Smallwood, A. M.; Carlton, D. D.; Banks, T. M.; natural peptides. J. Biol. Chem. 1993, 268, 9496-9503. Vavra, K. J.; Johnson, J. S.; Cook, E. R.; Falvo, J.; Wahl, R. C.; [82] Erdös, E. G.; Deddish, P. A.; Marcic, B. M. Potentiation of Pulvino, T. A.; Wendoloski, J. J.; Smith, D. L. 1.56 Å structure of Bradykinin Actions by ACE Inhibitors. Trends Endocrinol. Metab. mature truncated human fibroblast collagenase. Proteins 1994, 19, 1999, 10, 223-229. 98-109. [83] Skidgel, R. A.; Erdos, E. G. Novel activity of human angiotensin I [60] Browner M. F.; Smith W. W.; Castelhano A. L. Matrilysin-inhibitor converting enzyme: release of the NH2- and COOH-terminal complexes : common themes among metalloproteases. tripeptides from the luteinizing hormone-releasing hormone. Proc. Biochemistry 1995, 34, 6602-6610. Natl. Acad. Sci. U. S. A. 1985, 82, 1025-1029. [61] Gomis-Ruth, F. X.; Kress, L. F.; Bode, W. First structure of a snake [84] Ehlers, M. R.; Riordan, J. F. Angiotensin-converting enzyme: zinc- venom metalloproteinase: a prototype for matrix and inhibitor-binding stoichiometries of the somatic and testis metalloproteinases/collagenases. EMBO J. 1993, 12, 4151-4157. isozymes. Biochemistry 1991, 30, 7118-7126. [62] Hori, T.; Kumasaka, T.; Yamamoto, M.; Nonaka, T.; Tanaka, N.; [85] Casarini, D. E.; Plavinik, F. L.; Zanella, M. T.; Marson, O.; Hashimoto, Y.; Ueki, T.; Takio, K. Structure of a New 'Aspzincin' Krieger, J. E.; Hirata, I.Y.; Stella, R. C. Angiotensin converting Metalloendopeptidase from Grifola Frondosa: Implications for the enzymes from human urine of mild hypertensive untreated patients Catalytic Mechanism and Substrate Specificity Based on Several resemble the N-terminal fragment of human angiotensin I- Different Crystal Forms. Acta Crystallogr., Sect. D 2001, 57, 361- converting enzyme. Int. J. Biochem. Cell Biol. 2001, 33, 75-85. 368. [86] Deddish, P. A.; Marcic, B.; Jackman, H. L.; Wang, H. Z.; Skidgel, [63] Christianson, D. W.; David, P. R.; Lipscomb, W. N. Mechanism of R. A.; Erdös, E. G. N-domain-specific substrate and C-domain carboxypeptidase A: hydration of a ketonic substrate analogue. inhibitors of angiotensin-converting enzyme: angiotensin-(1-7) and Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 1512-1515. keto-ACE. Hypertension 1998, 31, 912-917. [64] Christianson, D. W.; Lipscomb, W. B. Carboxypeptidase-A. Acc. [87] Dive, V.; Cotton, J.; Yiotakis, A.; Michaud, A.; Vassiliou, S.; Chem. Res. 1989, 22, 62-69. Jiracek, J.; Vazeux, G.; Chauvet, M. T.; Cuniasse, P.; Corvol, P. [65] Osterman, A. L.; Grishin, N. V.; Smulevitch, S. V.; Matz, M. V.; RXP 407, a phosphinic peptide, is a potent inhibitor of angiotensin I Zagnitko, O. P.; Revina, L. P.; Stepanov, V. M. Primary structure of converting enzyme able to differentiate between its two active sites. carboxypeptidase T: Delineation of functionally relevant features in Proc. Natl. Acad. Sci. U. S. A. 1999, 96, 4330-4335. Zn-carboxypeptidase family. J. Protein Chem. 1992, 11, 561-570. [88] Junot, C.; Gonzales, M. F.; Ezan, E.; Cotton, J.; Vazeux, G.; [66] Lipscomb, W. N.; Sträter N. Recent Advances in Zinc Enzymology. Michaud, A.; Azizi, M.; Vassiliou, S.; Yiotakis, A.; Corvol, P.; Chem. Rev. 1996, 96, 2375-2433. Dive, V. RXP 407, a selective inhibitor of the N-domain of [67] Makarova, K. S.; Grishin, N. V. The Zn-peptidase superfamily: angiotensin I-converting enzyme, blocks in vivo the degradation of functional convergence after evolutionary divergence. J. Mol. Biol. hemoregulatory peptide acetyl-Ser-Asp-Lys-Pro with no effect on 1999, 292, 11-17. angiotensin I hydrolysis. J. Pharmacol. Exp. Ther. 2001, 297, 606- [68] Greenblatt, H. M.; Feinberg, H.; Tucker, P. A.; Shoham, G. 611. Carboxypeptidase A: native, zinc-removed and mercury-replaced [89] Corvol, P.; Williams, T. A. In Handbook of Proteolytic Enzymes; forms. Acta Crystallogr., Sect. D 1998, 54, 289-305. Barrett, A. J.; Rawlings, N. D.; Woessner, J. F., Eds.; Academic [69] Blundell T. L. Metalloproteinase superfamilies and drug design. Press: New York, 1998; pp 1066-1076. Nat. Struct. Biol. 1992, 2, 73-75. [90] Shapiro, R.; Holmquist, B.; Riordan, J. F. Anion activation of [70] Cushman, D. W.; Pluscec, J.; Williams, N. J.; Weaver, E. R.; Sabo, angiotensin converting enzyme: dependence on nature of substrate. E. F.; Kocy, O.; Cheung, H. S.; Ondetti, M. A. Inhibition of Biochemistry 1983, 22, 3850-3857. angiotensin-coverting enzyme by analogs of peptides from [91] Lamango, N. S.; Sajid, M.; Isaac, R. E. The endopeptidase activity Bothrops jararaca venom. Experientia 1973, 29, 1032-1035. and the activation by Cl- of angiotensin-converting enzyme is [71] Elisseeva, Y. E.; Orekhovich, V. N.; Pavlikhina, L. V.; Alexeenko, evolutionarily conserved: purification and properties of an L. P. Carboxycathepsin--a key regulatory component of two 428 Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 Spyroulias et al.

angiotensin-converting enzyme from the housefly, Musca [111] Fernandez, M.; Liu, X.; Wouters, M. A.; Heyberger, S.; Husain, A. domestica. Biochem. J. 1996, 314, 639-646. Angiotensin I-converting enzyme transition state stabilization by [92] Wei, L.; Alhenc-Gelas, F.; Corvol, P.; Clauser, E. The two His1089: evidence for a catalytic mechanism distinct from other Homologous Domais of Human Angiotensin I-converting Enzyme gluzincin metalloproteinases. J. Biol. Chem. 2001, 276, 4998-5004. Are Both Catalically Active. J. Biol. Chem. 1991, 266, 9002-9006. [112] Chen, Y. N.; Ehlers, M. R.; Riordan, J. F. The functional role of [93] Kester, W. R.; Matthews, B. W. Comparison of the structures of tyrosine-200 in human testis angiotensin-converting enzyme. carboxypeptidase A and thermolysin. J. Biol. Chem. 1977, 252, Biochem. Biophys. Res. Commun. 1992, 184, 306-309. 7704-7710. [113] Sen, I.; Kasturi, S.; Abdul Jabbar, M.; Sen, G. C. Mutations in two [94] Mayer, M.; Meyer, B. Mapping the active site of angiotensin- specific residues of testicular angiotensin-converting enzyme converting enzyme by Transferred NOE Spectroscopy. J. Med. change its catalytic properties. J. Biol. Chem. 1993, 268, 25748- Chem 2000, 43, 2093-2097. 25754. [95] Bünning, P.; Holmquist, B.; Riordan, J. F. Substrate specificity and [114] Thompson, J. D.; Gibson, T. J.; Plewniak, F.; Jeanmougin, F.; kinetic characteristics of angiotensin converting enzyme. Higgins, D. G. The CLUSTAL_X windows interface: flexible Biochemistry 1983, 22, 103-110. strategies for multiple sequence alignment aided by quality analysis [96] Patchett, A. A.; Cordes, E. H. The design and properties of N- tools. Nucleic Acids Res. 1997, 25, 4876-4882. carboxyalkyldipeptide inhibitors of angiotensin-converting enzyme. [115] Martí-Renom, M. A.; Stuart, A. C.; Fiser, A.; Sánchez, R.; Melo, > Adv. Enzymol. Relat. Areas Mol. Biol. 1985, 57, 1-84. F.; S ali, A. Comparative protein structure modeling of genes and [97] Patchett, A. A.; Harris, E.; Tristram, E. W.; Wyvratt, M. J.; Wu, M. genomes. Annu. Rev. Biophys. Biomol. Struct. 2000, 29, 291-325. T.; Taub, D. Peterson, E. R.; Ikeler, T.; ten Broeke, J.; Payne, L. G.; [116] Combet, C.; Blanchet, C.; Geourjon, C.; Deléage, G. NPS@: Ondeyka, D. L.; Thorsett, E. D.; Greenlee, W. J.; Lohr, N. S.; Network Protein Sequence Analysis. TIBS 2000, 25, 147-150. Hoffsommer, R. D.; Joshua, H.; Ruyle, W. V.; Rothrock, J. W.; [117] Garnier, J.; Osguthorpe, D. J.; Robson, B. Analysis of the accuracy Aster, S. D.; Maycock, A. L.; Robinson, F. M.; Hirschmann, R.; and implications of simple methods for predicting the secondary Sweet, C. S.; Ulm, E. H.; Gross, D. M.; Vassil, T. C.; Stone, C. A. structure of globular proteins. J. Mol. Biol. 1978, 120, 97-120. A new class of angiotensin-converting enzyme inhibitors. Nature [118] Gibrat, J. F.; Garnier, J.; Robson, B. Further developments of 1980, 288, 280-283. protein secondary structure prediction using information theory. [98] Moore, V. A.; Irwin, W. J.; Timmins, P.; Chong, S.; Dando, S. A.; New parameters and consideration of residue pairs. J. Mol. Biol. Morrison, R. A. A rapid screening system to determine drug 1987, 198, 425-443. affinities for the intestinal dipeptide transporter 1 : system [119] Garnier, J.; Gibrat, J. F.; Robson, B. GOR method for predicting characterisation. Inter. J. Pharm. 2000, 210, 15-27. protein secondary structure from amino acid sequence. Methods [99] Moore, V. A.; Irwin, W. J.; Timmins, P.; Chong, S.; Dando, S. A.; Enzymol. 1996, 266, 540-553. Morrison, R. A. A rapid screening system to determine drug [120] Merrifield R. B. Solid Phase Synthesis. I. The synthesis of a affinities for the intestinal dipeptide transporter 2: affinities of ACE Tetrapeptide. J. Am. Chem. Soc. 1963, 85, 2149-2154. inhibitors. Inter. J. Pharm. 2000, 210, 29-44. [121] Veglia, G.; Porcelli, F.; DeSilva, T.; Prantner, A.; Opella, S.J. The [100] Cushman, D. W.; Cheung, H. S.; Sabo, E. F.; Ondetti, M. A. Design structure of the metal-binding motiff GMTCAAC is similar in an of potent competitive inhibitors of angiotensin-converting enzyme. 18-residue linear peptide and the mercury binding protein Merp. J. Carboxyalkanoyl and mercaptoalkanoyl amino acids Biochemistry Am. Chem. Soc. 2000, 122, 2389-2390. 1977, 16, 5484-5491. [122] Opella, S. J.; DeSilva, T. M.; Veglia, G. Structural biology of [101] Petrillo, E. W.; Ondetti, M. A. Angiotensin-converting enzyme metal-binding sequence. Curr. Opin. Chem. Biol. 2002, 6, 217-223. inhibitors: medicinal chemistry and biological actions. Med. Res. [123] DeSilva, T. M.; Veglia, T.; Porcelli, F.; Prantner, A. M.; Opella, S. Rev. 1982, 2, 1-41. J. Selectivity in heavy metal-binding to peptides and proteins. [102] Cushman, D. W.; Cheung, H. S.; Sabo, E. F.; Ondetti, M. A. Biopolymers 2002, 64, 189-197. Development and design of specific inhibitors of angiotensin- [124] Carpino, L. A.; Han, G. Y.; 9-Fluorenylmethoxycarbonyl function, converting enzyme. Am. J. Cardiol. 1982, 49, 1390-1394. a new base-sensitive amino-protecting group J. Am. Chem. Soc. [103] Wu, M. T.; Douglas, A. W.; Ondeyka, D. L.; Payne, L. G.; Ikeler, 1970, 92, 5748-5749. T. J. Joshua, H.; Patchett, A. A.; Synthesis of N2-[(S)-1-carboxy-3- [125] Barlos, K.; Gatos, D.; Kallitsis, J.; Papaphotiou, G.; Sotiriou, P.; phenylpropyl]-L-lysyl-L-proline (lisinopril). J. Pharm. Sci. 1985, Wenquing, Y.; Shafer, W. Darstellung geshützter peptid-fragmente 74, 352-354. unter einsatz substituierter triphenylmethyl-harze, Tetraedron Lett. [104] Krapcho, J.; Turk, C.; Cushman, D. W.; Powell, J. R.; DeForrest, J. 1989, 30, 3943-3946. M.; Spitzmiller, E. R.; Karanewsky, D. S.; Duggan, M.; Rovnsak, [126] Papazacharias, S.; Pairas, G.; Kouimtzoglou, E.; Dermitzaki, E.; G.; Schwartz, J.; Natarajan, S.; Godfrey, J. D.; Ryono, D. E.; Manessi-Zoupa, E.; Gravanis, A.; Cordopatis, P. Synthesis and Neubeck, R.; Atwal, K. S.; Petrillo, E. W. Jr. Angiotensin- Biological Evaluation of New h/rCRH Peptide Analogs. In Peptides converting enzyme inhibitors. Mercaptan, carboxyalkyl dipeptide, 2000 ; Martinez, J.; Fehrentz, J.-A., Eds, EDK, 2001, pp. 933-934. and phosphinic acid inhibitors incorporating 4-substituted prolines. [127] King, D. S.; Fields, C. G.; Fields G. B. A cleavage method which J. Med. Chem. 1988, 31, 1148-1160. minimizes side reactions following Fmoc solid phase peptide [105] Michaud, A.; Chauvet, M. T.; Corvol, P. N-domain selectivity of synthesis. Int. J. Pept. Protein Res. 1990, 36, 255-266 angiotensin I-converting enzyme as assessed by structure-function [128] Braunschweiler, L.; Ernst, R. R. Coherence transfer by isotropic studies of its highly selective substrate, N-acetyl-seryl-aspartyl- mixing : application to proton correlation spectroscopy. J. Magn. lysyl-proline. Biochem. Pharmacol. 1999, 57, 611-618. Reson. 1983, 53, 521-528. [106] Araujo, M. C.; Melo, R. L.; Cesari, M. H.; Juliano, M. A.; Juliano, [129] Bax, A.; Davis, D. G. MLEV-17-based two-dimensional L.; Carmona, A. K. Peptidase specificity characterization of C- and homonuclear magnetization transfer spectroscopy. J. Magn. Reson. N-terminal catalytic sites of angiotensin I-converting enzyme. 1985, 65, 355-360. Biochemistry 2000, 39, 8519-8525. [130] Marion, D.; Wüthrich, K. Application of phase sensitive two- [107] Ehlers, M. R. W.; Kirsch, R. E. Catalysis of Angiotensin I dimensional correlated spectroscopy (COSY) for measurements of Hydrolysis by Human Angiotensin-Converting Enzyme : Effect of 1H-1H spin-spin coupling constants in proteins. Biochem. Biophys. Chloride and pH. Biochemistry, 1988, 27, 5538-5544. Res. Commun. 1983, 113, 967-974. [108] Liu, X.; Fernandez, M.; Wouters, M. A.; Heyberger, S.; Husain, A. [131] Jeener, J., Meier, B. H., Bachmann, P.; Ernst R.R. Investigation of Arg(1098) is critical for the chloride dependence of human exchange processes by two-dimensional NMR spectroscopy. J. angiotensin I-converting enzyme C-domain catalytic activity. J. Chem. Phys. 1979, 71, 4546-4553. Biol. Chem. 2001, 276, 33518-33525. [132] Wüthrich, K. In NMR of Proteins and Nucleic Acids; Wiley: New [109] Bünning, P.; Holmquist, B.; Riordan, J. F. Functional residues at York, 1986. the active site of angiotensin converting enzyme. Biochem. Biophys. [133] Wang, Y.; Nip, A.M.; Wishart D. S. A simple method to Res. Commun. 1978, 83, 1442-1449. quantitatively measure polypeptide JHNHa coupling constants from [110] Shapiro, R.; Riordan, J. F. Critical Lysine Residue at the Chloride TOCSY or NOESY spectra. J. Biomol. NMR 1997, 10, 373-382. Binding Site of Angiotensin Converting Enzyme. Biochemistry [134] Smith, L. J.; Bolin, K. A.; Schwalbe, H.; MacArthur, M. W.; 1983, 22, 5315-5321. Thorton, J. M.; Dobson, C. M. Analysis of main chain torsion Structural Features of Angiotensin-I Converting Enzyme Catalytic Sites Current Topics in Medicinal Chemistry, 2004, Vol. 4, No. 4 429

angles in proteins: prediction of NMR coupling constants for native [142] Pearlman, D. A.; Case, D. A.; Caldwell, G. C.; Siebel, G. L.; Singh, and random coil conformations. J. Mol. Biol. 1996, 255, 494-506. U. C.; Weiner, P.; Kollman, P. A. AMBER 4.0. University of [135] Wishart, D.; Sykes, B.; Richards, F. Relationship between nuclear California, San Francisco, 1991. magnetic resonance chemical shift and protein secondary structure. [143] Shortle, D.; Clarke, N. Alpha helix propensity of amino acids. J. Mol. Biol. 1991, 222, 311-333. Science 1993, 260, 1637-1640. [136] Wishart, D.; Sykes, B.; Richards, F. The chemical shift index: a fast [144] Shortle, D.; Clarke, N. Alpha helix propensity of amino acids. and simple method for the assignment of protein secondary Science 1993, 262, 917-918. structure through NMR spectroscopy. Biochemistry 1992, 31, 1647- [145] Zuiderweg, E. R. P. Mapping Protein-Protein Interactions in 1651. Solution by NMR Spectroscopy. Biochemistry, 2002, 41, 1-7. [137] Merutka, G.; Dyson, H. J.; Wright, P. E. 'Random coil' 1H chemical [146] Galanis, A. S.; Spyroulias, G. A.; Pierattelli, R.; Tzakos, A.; shifts obtained as a function of temperature and trifluoroethanol Troganis, A.; Gerothanassis, I. P.; Pairas, G.; Manessi-Zoupa, E.; concentration for the peptide series GGXGG. J. Biomol. NMR Cordopatis, P. Zinc-Binding in Peptide Models of ACE Active Sites 1995, 5, 14-24. Studied Through 1H NMR & Chemical Shift Perturbation Mapping. [138] Galanis, A. S.; Spyroulias, G. A.; Pairas, G.; Manessi-Zoupa, E.; Biopolymers, 2003, 69, 244-252. Cordopatis, P. NMR solution structures of apoform ACEN and [147] Koradi, R.; Billeter, M.; Wüthrich, K. MOLMOL: a program for ACEC active sites model peptides. In preparation. display and analysis of macromolecular structures. J. Mol. Graph. [139] Güntert, P.; Braun, W.; Wüthrich, K. Efficient computation of 1996, 14, 51-55. three-dimensional protein structures in solution from nuclear [148] Natesh, R.; Schwager, S. L. U.; Sturrock, E. D.; Acharya, K. R. magnetic resonance data using the program DIANA and the Crystal structure of the human angiotensin-converting enzyme- supporting programs CALIBA, HABAS and GLOMSA. J. Mol. lisinopril complex. Nature 2003, 421, 551-554 Biol. 1991, 217, 517-530. [149] Kim, H. M.; Shin, D. R.; Yoo, O. J.; Lee, H.; Lee, J. O. Crystal [140] Güntert, P.; Mumenthaler, C.; Wüthrich, K. Torsion angle dynamics structure of Drosophila angiotensin I-converting enzyme bound to for NMR structure calculation with the new program DYANA. J. captopril and lisinopril. FEBS Lett. 2003, 538, 65-70 Mol. Biol. 1997, 273, 283-298. [150] Arndt, J. W.; Hao, B.; Ramakrishnan, V.; Cheng, T.; Chan, S. I.; [141] Pearlman, D. A.; Case, D. A.; Caldwell, J. W.; Ross, W. S.; Chan, M. K. Crystal structure of a novel carboxypeptidase from the Cheatham, T. E.; Ferguson, D. M.; Seibel, G. L.; Singh, U. C.; hyperthermophilic archaeon Pyrococcus furiosus. Structure 2002, Weiner, P. K.; Kollman, P. A. AMBER 5.0. University of 215-224. California, San Francisco, 1997.