| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
FJ
EXPRESS SUMMARY ARTICLE The Full-length version of this article is also available, published online August 7, 2002 as doi:10.1096/fj.01-1017fje. |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
2
* Department of Biochemistry, University of Stockholm, Arrhenius Laboratories A4, S-106 91 Stockholm, Sweden;
Department of Chemistry and The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, California, USA; and
Department of Medical Biochemistry and Biophysics, Karolinska Institutet, S-171 77 Stockholm, Sweden
2Correspondence: Department of Medical Biochemistry and Biophysics, Division of Chemistry II, Karolinska Institutet, S-171 77 Stockholm, Sweden. E-mail: jesper.haeggstrom{at}mbb.ki.se
SPECIFIC AIMS
Leukotriene (LT) A4 hydrolase/aminopeptidase (LTA4H) is a bifunctional zinc metalloenzyme that controls the biosynthesis of LTB4, a potent chemoattractant and immune-modulating lipid mediator implicated in the pathophysiology of inflammatory diseases. To probe the active site(s) and catalytic mechanisms of LTA4H and set the stage for structure-based drug design, we have determined three high-resolution crystal structures of the enzyme in complex with different classes of competitive inhibitors: a thioamine, an amino hydroxamic acid, and captopril.
PRINCIPAL FINDINGS
1. The thioamine binds to the zinc and occupies a putative LTA4 binding pocket
The thioamine (3-(4-benzyloxyphenyl)-2-(R)-amino-1-propane thiol) is a specific tight-binding inhibitor of LTA4H designed to integrate a transition state of an amide cleavage as well as hydrophobic properties of the carbon backbone of LTA4. The free thiol group interacts with the catalytic zinc, creating a tetrahedral coordination sphere. The thioamine makes extensive interactions with the protein that render the compound a strong inhibitor (Fig. 1
). Thus, the aminopropane thiol portion of the molecule is bound in a narrow binding site with the amine group anchored firmly to Gln136 and Glu271. At the opposite end of the inhibitor, the benzyloxyphenyl tail sticks deep into a narrow, L-shaped, hydrophobic pocket that has been proposed to bind the olefinic tail of LTA4.
|
2. The hydroxamic acid mimics the binding of LTA4 and makes an aberrant metal complex
The second structure contains a hydroxamic acid (N-hydroxy-N-[(2S)-2-amino-3-(benzyloxyphenyl)propyl]-5-carboxypentanamide). The benzyloxyphenyl tail, a mimic of the C7-C20 tail of LTA4, binds to the narrow hydrophobic pocket exactly as the corresponding structure of the thioamine inhibitor (Figs. 1
, 2)
. The ether linkage is positioned near a hydrophilic patch formed by Asp375, Gln134, and Tyr267. The hydroxamate makes a monodentate complex with the zinc involving only the hydroxyl oxygen in a tetrahedral coordination rather than the expected bidentate complex (Fig. 2
A). The other oxygen interacts with Glu296 and the backbone nitrogen of Gly269. The flanking pentanoic acid is slightly compressed, as if there is not enough space for this segment of the inhibitor, and the carboxylate interacts with Arg563.
|
3. Captopril exerts its inhibitory action via binding to the catalytic zinc
The third structure shows how captopril (D-3 mercapto-methylpropionyl-L-proline), a classical antihypertensive drug targeted against ACE, binds in an open and wide segment of the active center. Its thiol group is bound to the zinc, apparently accounting for most of its inhibitory action on LTA4H. Otherwise, the compound makes few interactions with the protein, which is probably why captopril is less potent (Fig. 1)
. The captopril molecule does not block the entrance to the hydrophobic pocket proposed to bind LTA4 and will not provide much resistance against this substrate. Hence, the position of captopril agrees with its inhibitor profile with a weak potency against the epoxide hydrolase activity (Fig. 1)
.
CONCLUSIONS AND SIGNIFICANCE
The thioamine and the hydroxamic acid are chemical mimics of LTA4, and previous studies of SAR have shown that the hydroxamate chelates the zinc and is the equivalent of the epoxide of LTA4. Furthermore, the flanking pentanoic acid is anchored to a basic residue and the benzyloxyphenyl tail is buried in a hydrophobic pocket. In our structure, the hydroxamate is bound almost exactly as predicted, thus acting as a template for the binding of LTA4 (Figs. 1
, 2)
. Hence, our structural data indicate that LTA4 binds with the C1 carboxylate interacting with Arg563 and/or Lys565, the epoxide positioned near the zinc atom, and the C7-C20 olefinic tail reaching deep into the narrow L-shaped hydrophobic cavity. In this way, C12 of LTA4 gets close to the hydrophilic patch in the pocket. This binding conformation of LTA4 supports a mechanism of the epoxide hydrolase activity in which the zinc ion, together with Glu271, participates in the activation and opening of the C5-C6 epoxide ring according to an SN1 reaction. A carbocation intermediate will be formed whose charge is delocalized over the conjugated triene system. In a second step, Asp375, Gln134, or Tyr267 could guide the stereo-specific introduction of the hydroxyl group at C12 of the substrate. Arg563 and/or Lys565 most likely act as a recognition site for the carboxylate of LTA4.
Binding of the hydroxamic acid is suboptimal and offers clues to structure-based improvements of the molecule. Thus, the aberrant monodentate zinc complex is probably due to the tight binding of the benzene rings in the hydrophobic pocket and the strong interaction of the amine group with Glu271 and Gln136 (Fig. 2A, B
). The amine and the hydroxamic acid moiety are separated by a methylene group, which is one carbon atom more than what is seen in a normal or modified peptide linkage (Fig. 2C
). As a result, one of the hydroxamate oxygens has been pushed away and can no longer interact with the Zn2+ ion (Fig. 2C
). Removal of this extra methylene carbon would be expected to improve the metal chelation and generate a compound with increased inhibitor potency. Changing the chirality of the amino-terminal amine carbon from S to R, to better mimic the stereochemistry of a peptide substrate, could be another strategy for inhibitor optimization.
Captopril is a classical inhibitor of ACE, a zinc protease composed of two domains, each of which harbors a catalytic zinc binding site structurally similar to the one present in LTA4H. Thus, both enzymes are members of the MA clan of metallopeptidases, which contain a canonical HEXXH signature also present in thermolysin. In the absence of structural data for ACE, it has generally been assumed that captopril chelates the catalytic zinc atoms. Our structure of the captopril-LTA4H complex, the first structure of this inhibitor in complex with any protein, corroborates this hypothesis and suggests that captopril may bind to a number of other MA metalloproteases in a similar manner.
LTA4H has a strong preference for arginyl di- and tripeptides. Since captopril and bestatin are peptide mimics (Fig. 1)
, the structures of the corresponding inhibitor-LTA4H complexes can be used to define the binding site for a natural peptide substrate. The predicted binding mode for an arginyl tripeptide, as deduced from these structures, is in excellent agreement with a zinc-assisted general base mechanism for the peptide cleavage. This mechanism involves Glu296 as the base catalyst and Tyr383 as the proton donor. The amino and carboxyl termini are anchored to Glu271 and Arg563, respectively.
Thermolysin is a typical representative of the MA clan of zinc metallopeptidases to which LTA4H also belongs. Despite minimal sequence identity (7%), the catalytic domain of LTA4H is structurally very similar to thermolysin, illustrating the significant functional similarities that exist between these two enzymes. Yet thermolysin is an endoprotease whereas LTA4H is an exopeptidase. The deduced binding mode for peptide substrates in LTA4H and previous crystallographic data for thermolysin allow examination of the molecular basis for the differences in substrate specificity. In both enzymes, the peptide substrate is bound between the 
-helical and the mixed /ß lobe of the catalytic domain via hydrogen bonding to one strand of the ß-sheet. However, in LTA4H the neighboring ß-strand is shorter and has two glycines (268 and 269) within binding distance to the substrate. These Gly residues belong to a GXMEN motif (GGMEN in LTA4H) conserved within the M1 family of the MA clan of metallopeptidases, and the penultimate Glu residue (Glu271 in LTA4H) of this motif has been suggested to act as an amino-terminal recognition site for peptide substrates. Our structural data thus indicate a more general role for the GXMEN signature in substrate binding and alignment, a conclusion that pertains to the entire enzyme family. In thermolysin, Trp115 contributes to its function as an endoprotease, whereas in LTA4H, the corresponding position is occupied by Glu271. Indeed, Glu271 was recently shown to be important for the enzyme’s exopeptidase specificity.
In conclusion, the three structures of the bifunctional LTA4H/aminopeptidase in complex with competitive inhibitors offer detailed insights to substrate binding and catalysis, especially the molecular mechanisms for generation of the proinflammatory LTB4. Conclusions regarding the aminopeptidase active site and reaction mechanism most likely pertain to all members of the M1 family of the MA clan of ‘thermolysin-like’ metallopeptidases. In particular, our data indicate that the GXMEN motif plays an important role in the alignment and binding of peptide substrates. Moreover, the present structure of captopril in complex with a zinc peptidase provides new insights into its mode of action against ACE, and together the three inhibitor-LTA4H complexes highlight the potentials of structure-based drug design.
|
FOOTNOTES
1 To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.01-1017fje; to cite this article, use FASEB J. (August 7, 2002) 10.1096/fj.01-1017fje 
This article has been cited by other articles:
|
|
 |
|
  C. Bauvois, L. Jacquamet, A. L. Huston, F. Borel, G. Feller, and J.-L. Ferrer Crystal Structure of the Cold-active Aminopeptidase from Colwellia psychrerythraea, a Close Structural Homologue of the Human Bifunctional Leukotriene A4 Hydrolase J. Biol. Chem., August 22, 2008; 283(34): 23315 - 23325. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Liang, E. Claret, J. Ouled-Diaf, A. Jean, D. Vogel, D. R. Light, S. W. Jones, W. J. Guilford, J. F. Parkinson, and R. M. Snider Development of a Homogeneous Time-Resolved Fluorescence Leukotriene B4 Assay for Determining the Activity of Leukotriene A4 Hydrolase J Biomol Screen, June 1, 2007; 12(4): 536 - 545. [Abstract] [PDF]
|
|
|
|
|
|
F. Tholander, F. Kull, E. Ohlson, J. Shafqat, M. M. G. M. Thunnissen, and J. Z. Haeggstrom Leukotriene A4 Hydrolase, Insights into the Molecular Evolution by Homology Modeling and Mutational Analysis of Enzyme from Saccharomyces cerevisiae J. Biol. Chem., September 30, 2005; 280(39): 33477 - 33486. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Z. Haeggstrom Leukotriene A4 Hydrolase/Aminopeptidase, the Gatekeeper of Chemotactic Leukotriene B4 Biosynthesis J. Biol. Chem., December 3, 2004; 279(49): 50639 - 50642. [Full Text] [PDF]
|
|
|
|
|
|
P. C. Rudberg, F. Tholander, M. Andberg, M. M. G. M. Thunnissen, and J. Z. Haeggstrom Leukotriene A4 Hydrolase: IDENTIFICATION OF A COMMON CARBOXYLATE RECOGNITION SITE FOR THE EPOXIDE HYDROLASE AND AMINOPEPTIDASE SUBSTRATES J. Biol. Chem., June 25, 2004; 279(26): 27376 - 27382. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
