Tyrosine 547 Constitutes an Essential Part of the Catalytic Mechanism of Dipeptidyl Peptidase IV*

Human dipeptidyl peptidase IV (DPP-IV) is a ubiqui-tously expressed type II transmembrane serine protease. It cleaves the penultimate positioned prolyl bonds at the N terminus of physiologically important peptides such as the incretin hormones glucagon-like peptide 1 and glucose-dependent insulinotropic peptide. In this study, we have characterized different active site mutants. The Y547F mutant as well as the catalytic triad mutants S630A, D708A, and H740L showed less than 1% wild type activity. X-ray crystal structure analysis of the Y547F mutant revealed no overall changes compared with wild type apoDPP-IV, except the ablation of the hydroxyl group of Tyr 547 and a water molecule positioned in close proximity to Tyr 547 . To elucidate further the reaction mechanism, we determined the crystal structure of DPP-IV in complex with diisopropyl fluorophosphate, mimicking the tetrahedral intermediate. The kinetic and structural findings of the tyrosine residue are discussed in relation to the catalytic mechanism of DPP-IV and to the inhibitory mechanism

incretin hormones, and DPP-IV has therefore been proposed as an important regulator of different physiological and pathophysiological conditions (5). There is a considerable pharmaceutical interest in DPP-IV, because the enzyme has been shown to inactivate the incretin hormones glucagon-like peptide 1 and glucose-dependent insulinotropic peptide in vivo (6,7). This makes DPP-IV an important regulator of glucose homeostasis, as glucagon-like peptide 1 and glucose-dependent insulinotropic peptide have glucose-dependent insulinotropic as well as neogenetic effects on the pancreatic ␤-cells. The use of DPP-IV inhibitors in diabetes is being explored, and the first short term treatments of diabetes mellitus type 2 patients with DPP-IV inhibitors have demonstrated clinical proof of the concept (8,9).
The published crystal structure of human recombinant DPP-IV in complex with the substrate analog valine pyrrolidide (ValPyr) (Fig. 1) has revealed many important details regarding the inhibitor binding in the active site cavity, which by analogy illustrate substrate binding (10). The active site is positioned in a large cavity, formed at the interface of an ␣/␤ hydrolase domain and an eight-bladed ␤-propeller domain. The catalytic triad has been identified by site-directed mutagenesis in mouse DPP-IV, which by homology corresponds to Ser 630 , Asp 708 , and His 740 in human DPP-IV (11). The catalytic triad is arranged in a topological fold and sequential order, which defines the ␣/␤ hydrolase domain. Furthermore, the catalytically essential Ser 630 is located in a so-called "nucleophile elbow" consisting of the sequence Gly 628 -Trp 629 -Ser 630 -Tyr 631 -Gly 632 , a consensus sequence characteristic for all serine peptidases in the SC clan, i.e. GX1SX2G (12,13). Furthermore, the crystal structure determinations have suggested detailed information of the catalytic mechanism of DPP-IV. For example, we now understand the essential function of the residues Glu 205 and Glu 206 for coordination of the Nterminal amine of the substrate, which had been demonstrated by use of site-directed mutagenesis (14). In addition, the residues Arg 125 and Asn 710 appear essential for coordination of the carbonyl of the N-terminal amino acid residue of the substrate and, together with the two glutamates, align the substrate optimally for the nucleophilic attack by Ser 630 . A negatively charged tetrahedral oxyanion intermediate is generated in the transition state and is stabilized by a so-called oxyanion hole. This is a recognized mechanism among serine proteases. Based on analysis of the structure and sequence alignment to the homolog S9 protein family member prolyl oligopeptidase (POP), this oxyanion hole has been suggested to be formed via hydrogen bonding to Tyr 547 -OH and the backbone NH of Tyr 631 (10,(15)(16)(17)(18). In this study, we have investigated the oxyanion stabilization of DPP-IV catalysis by removing the hydroxyl group of Tyr 547 by means of site-specific mutagenesis to the phenylalanine equivalent. The kinetic data of the Y547F mutant show that Tyr 547 is essential * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The for DPP-IV catalysis. Determination and comparison of the crystal structures of (i) the mutant Y547F, (ii) a complex between DPP-IV and the covalent inhibitor diisopropyl fluorophosphate (DFP), and (iii) the apo forms of DPP-IV (in-house as well as previously published) confirm the oxyanion stabilizing role of the hydroxyl group of Tyr 547 in the catalytic mechanism of DPP-IV. In addition, these results suggest that the inhibitory mechanism of the pharmacologically important class of DPP-IV inhibitors, the 2-cyanopyrrolidines, is conducted via proton acceptance from Tyr 547 resulting in stable covalent complexes.

MATERIALS AND METHODS
Bioinformatics-Analyses of the structure of recombinant human DPP-IV were performed using the Quanta software (Accelrys Inc.). The Vector NTI Suite 6.0 (InforMax Inc.) was used for sequence analysis, gene alignments, and primer design.
Chemicals and Reagents-QIAprep Miniprep System and Qiaex Gel Extraction II kits were purchased from Qiagen (San Diego). The baculovirus transfer vector pBlueBac4.5 was from Invitrogen. Mouse anti-CD26 monoclonal antibody clones MA2600 and MA261 were from Endogen (Rockford, IL) and Bender MedSystems (Vienna, Austria), respectively. Horseradish peroxidase-conjugated rabbit anti-mouse IgG was from Dako (Glostrup, Denmark). Spodoptera frugiperda 9 (Sf9) and High5 insect cells were grown in Grace Insect medium supplemented with fetal calf serum ranging from 0 to 10%, yeastolate, 20 mM Lglutamine, and 0.25 g/ml gentamycin in either tissue culture flasks or glass spinner bottles at 28°C. Chromatographic columns and materials (CNBr-activated Sepharose 4B matrix, MonoQ ion-exchange column, and Q-Sepharose high performance resin) were from Amersham Biosciences. Adenosine deaminase (ADA) protein was from Roche Applied Science. DFP was from Sigma.
Site DNA Sequencing-DNA sequencing was performed on an ABI Prism 310 DNA Analyzer using ABI PRISM® BigDYE TM Terminator Cycle Sequencing Ready Reaction from Applied Biosystems.
Generation of Recombinant Baculovirus-Recombinant baculoviruses were constructed using the derived Autographa californica nuclear polyhedrosis virus Bac-N-Blue TM DNA and the transfer vector pBlueBac4.5 from Invitrogen. Purified transfer vectors with recombinant inserts were mixed with Bac-N-Blue TM DNA and transfected into Sf9 cells using either Lipofectin® (Invitrogen) or Cellfectin® (Invitrogen). Virus isolates were plaque-purified according to the manufacturer's instructions and amplified in Sf9 cells for production of 100% recombinant baculovirus high titer stocks.
Quantification of DPP-IV Protein-DPP-IV quantifications and interactions with ADA were analyzed by an enzyme-linked immunosorbent assay (ELISA). Maxisorp 96-well plates (Merck) were incubated overnight at 4°C with 3 g of ADA per well in PBS, washed twice in PBS, blocked for 1-2 h in PBS containing 3% bovine serum albumin and 0.05% Tween 20 (blocking buffer), and washed twice in PBS. The coated plates were then incubated with increasing amounts of DPP-IV containing insect cell supernatants for 1-2 h followed by two successive washes with PBS. Bound DPP-IV was detected by incubation for 1-2 h with a primary mouse anti-CD26 monoclonal antibody (either MA2600 or MA261 at a final concentration of 5 g/ml in blocking buffer). After three washes with PBS, bound antibody was detected by incubation for 1-2 h with a horseradish peroxidase-coupled rabbit anti-mouse IgG conjugate, diluted in PBS with 0.1% bovine serum albumin. The wells were washed six times with PBS and developed using ortho-phenylenediamine and hydrogen peroxide according to the manufacturer's instructions (DAKO, Glostrup, Denmark). Optical densities were read at 450 nm. Purified DPP-IV was used as a standard and displayed reproducible linearity in the range of 0 -2 g/ml (R 2 Ͼ 0.9).
Purification of Recombinant Human DPP-IV-Recombinant human DPP-IV protein secreted into High5 insect cell supernatants was purified using ADA affinity chromatography as described previously (19). Briefly, purified ADA was chemically coupled to a CNBr-activated Sepharose matrix according to the manufacturer's instructions (Amersham Biosciences). The ADA affinity column and a Q-Sepharose High Performance column were interconnected on an Ä kta Purifier flow pressure chromatographic system. High5 cell supernatant with expressed protein was applied for direct elution from the ADA affinity column onto the Q-Sepharose High Performance column. Before applying to columns, supernatants were centrifuged (15-20,000 ϫ g) and filtered (0.45 m), and pH and conductivity were adjusted to pH 8 and 17-20 mS cm Ϫ1 . A final purification was performed using a MonoQ ion-exchange column. Pooled fractions were concentrated on Centriprep YM10 and Centricon YM10 from Millipore Corp. Protein contents were determined by UV280 nm spectroscopy and instantly flash-frozen in liquid nitrogen or in a dry ice ethanol bath.
Enzymatic Activity Assays-Enzymatic activity was determined kinetically at room temperature in 50 mM Tris, pH 7.4, 150 mM NaCl, 0.1% Triton X-100 buffer using different p-nitroanilide (pNA) substrates, including Ala-Ala-Pro-pNa, succinyl-Ala-pNa, Arg-Pro-pNa, Asp-Pro-pNa, Ala-Pro-pNa, Val-Pro-pNa, Gly-Pro-pNa, Gly-Gly-pNa, Ala-Phe-pNa, and Ala-Ala-pNA (Sigma or Bachem, Bubendorf, Germany). Released pNA was determined at 450 or 405 nm, and incubations were performed for 60 min with absorbance measurements every 5 min. To determine steadystate Michaelis-Menten kinetic parameters, different concentrations of putative substrates were used. Enzyme rates (mOD/min) were used for determination of K m and V max values by direct fitting to the Michaelis- . At least triplicate measurements were used for all kinetic determinations.
Crystallization Conditions-DPP-IV crystals were grown essentially as described previously (10) by using the hanging drop vapor diffusion method. Purified recombinant human DPP-IV in 50 mM Tris, pH 8.0, and 150 mM NaCl was mixed with equal amounts of a reservoir containing 0.4 M sodium acetate, 16 -19% w/v PEG4000, 0.1 M Tris, pH 8.6. Prior to data collection, crystals were soaked in 0.4 M sodium acetate, 35% w/v PEG4000, 0.1 M Tris, pH 8.6, and flash-cooled in a nitrogen stream. Prior to crystallization of the DFP complex, DPP-IV was mixed with DFP in a molar ratio of 1:10.
Crystallographic Data Collection and Handling-Crystallographic data collection was performed at the synchrotron beamlines MaxLab 711 (Lund, Sweden), ESRF ID 14-4 (Grenoble, France) and on an in-house rotating anode Rigaku RU300. Data reductions were performed with the HKL2000 software package (20). The structures were solved by the molecular replacement method using wild type DPP-IV as a search model (Protein Data Bank code 1N1M). Model building was performed using Quanta software (Accelrys Inc.) and iterative refinement (initially performed as a rigid body refinement) using CNX (21). Structure validation and handling were performed with Procheck (22) and Moleman. 2

RESULTS
Structural and Sequence Analysis of the Active Site of DPP-IV-The hydroxyl group of the side chain of Tyr 547 is coordinated via a water molecule to the hydroxyl group of Ser 630 ( Fig.  1) and has, together with the main chain NH of Tyr 631 , been suggested as a stabilizer of the oxyanion intermediate during catalysis (10). To test the catalytic effect of the hydroxyl group, the Y547F mutant variant lacking the para-positioned hydroxyl group was constructed.
The catalytic triad residues have been identified by sitedirected mutagenesis in mouse DPP-IV (11,12) and by homology to Ser 630 , Asp 708 , and His 740 in human DPP-IV. The mutant variants S630A, D708A, and H740L were included in this study as controls.
Expression and Characterization of DPP-IV Mutants-The DPP-IV mutants were generated using a PCR-based site-directed mutagenesis method directly in CD5/DPP-IV-pBlueBac. This construct encodes a soluble recombinant human form of DPP-IV lacking the cytosolic and transmembrane domains. The open reading frame of this construct is fused to the leader secretion signal of CD5. Thus, expressed protein is secreted to the cell supernatants after post-translational modifications. All DPP-IV mutated constructs were verified for PCR-introduced sequence errors by complete DNA sequencing.
High titer baculovirus stocks produced in Sf9 insect cells were used for expression studies in High5 insect cells (multiplicity of infection Ͼ1). Expression levels were analyzed for intracellular levels by use of SDS-PAGE-Coomassie ( Fig. 2A) and for secreted protein to cell supernatant by use of an ADAsandwich ELISA. From SDS-PAGE analysis of total cell lysates, protein bands with an electrophoretic mobility equal to purified DPP-IV were observed and interpreted as DPP-IV expressed protein. No protein bands with similar electrophoretic mobility could be observed with insect cell expression controls. Intracellular DPP-IV protein accounted for ϳ10 -30% of total cellular protein (ϳ20 -60 g/10 6 cells). The level of secreted protein in the cell supernatants was significantly lower compared with intracellular DPP-IV (i.e. 0.5-5 g per ml ϳ1-10% of intracellular amounts assuming ϳ10 6 cells/ml supernatant). Sandwich ELISA titration of the cell supernatants indicated that the secreted mutant DPP-IV proteins bound to ADA and anti-human CD26 monoclonal antibodies. Altogether, these data verified that structurally intact DPP-IV mutants had been expressed. Only cell-secreted DPP-IV was used for further studies.
All DPP-IV mutants were normalized directly in the cell supernatants to three different protein concentrations (i.e. 0.11, 0.16, and 0.54 M) by using ELISA, and enzymatic activity levels were characterized at these concentrations by using substrate analogs (Table I). The three catalytic triad mutants S630A, D708A, and H740L exhibited less than 1% specific activity compared with wild type DPP-IV. Surprisingly, the Y547F mutant showed equally low specific activity as the catalytic triad mutants.
Purification and Characterization of Wild Type and Y547F DPP-IV-Wild type DPP-IV and the Y547F variant were expressed in large scale (Ͼ2 liters of insect cell supernatants) and purified by a three-step procedure using interlinked ADA-coupled Sepharose affinity-and Q-Sepharose HP chromatography followed by MonoQ ion-exchange chromatography. The purified products were analyzed by SDS-PAGE and Coomassie staining, showing more than 99% purity (Fig. 2B).
Kinetically, k cat dropped for the Y547F mutant by ϳ50-fold by using the putative substrate Gly-Pro-pNA, whereas K m values increased ϳ30-fold compared with wild type, resulting in an overall drop of more than 1,500-fold for the second-order rate constant k cat /K m (Fig. 3 and Table II). Similar results were obtained with other substrate analogs, showing no differences in substrate specificity as a result of the mutation.
X-ray Crystallography Structure Determination-The x-ray crystal structures of the apoDPP-IV, the complex DFP⅐DPP-IV, and the DPP-IV mutant Y547F were determined. Diffraction data sets were collected at 2.0 Å resolution for the apoDPP-IV, 2.7 Å for the DFP⅐DPP-IV complex, and 2.2 Å for the Y547F mutant. Crystallographic data collection and refinement statistics are listed in Table III. All structures were solved by molecular replacement using the previously published DPP-IV structure (Protein Data Bank code 1N1M) as a search model excluding inhibitor and water molecules.
From analysis of the active site cavity and the overall structure of the Y547F mutant, it was clear that the overall structure of the mutant was conserved, and comparison of active site residues showed no conformational changes, neither main chain nor side chain, from wild type DPP-IV. Superimposition of C ␣ trace of the dimer structure of wild type DPP-IV and the Y547F mutant showed a root mean square of 0.86 Å 2 (1530 C ␣ atoms). Thus, the decreased enzyme activity of the Y547F mutant was not a result of an active site collapse and/or conformational changes within the active site or the whole protein as such. Inspection of the electron density of the mutated residue revealed a strongly defined phenylalanine, positioned exactly as the phenyl moiety of the tyrosine residue (Fig. 4, A  and B). Most interesting, coordination of a water molecule (Wat 123 , see  (10)) also has a water molecule at the same position.
The structure of the complex between DPP-IV and DFP showed that the irreversible organophosphorous inhibitor was covalently bound to the active serine. The C ␣ trace as well as the active site residues of the DFP-DPP-IV structure aligned completely to the comparable structural elements of the apo and Y547F structures. The complex forms a tetrahedral arrangement that mimics the negatively charged tetrahedral intermediate formed during substrate catalysis (Fig. 4C). The oxygen of the PϭO moiety forms hydrogen bonds to the hydroxyl group of Tyr 547 and to the main chain NH of Tyr 631 , thus positioning it spatially between the side chains of Tyr 547 and Ser 630 close to the position of the water molecule Wat 123 in the apo structure. DISCUSSION In this study, we have shown that the residue Tyr 547 of DPP-IV is essential for catalysis. The analysis of the mutant structure as well as the apo and DFP complex of DPP-IV supports the suggestion that the role of Tyr 547 is to stabilize the tetrahedral oxyanion intermediate. Exchanging Tyr 547 with phenylalanine resulted in a vast drop in activity of the same magnitude as alanine/leucine mutants of the catalytic triad residues (11). Structure determination of the Y547F mutant revealed a completely intact protein structure, and the only difference observed was the absence of the para-positioned hydroxyl group and a coordinating water molecule, Wat 123 . Structural alignment of the mutant, apo, and DFP complex structures showed no differences in the overall fold or side chain conformations. Most interesting, Wat 123 is found at a similar position in the DPP-IV⅐ValPyr complex, inbetween the Ser 630 -OH and the Tyr 547 -OH motif in close proximity (ϳ3.2 Å) to the proline-mimicking moiety of the inhibitor ValPyr (Fig. 1), thereby not excluding a direct catalytic role in the mechanism, e.g. functioning as a nucleophilic water molecule. Furthermore, the complex structure between human DPP-IV and the covalent inhibitor DFP showed that the phosphonate PϭO was within hydrogen bonding distance with the Tyr 547 -OH and the main chain NH of Tyr 631 . DFP is well accepted as a mimic of the tetrahedral enzyme-substrate intermediate, and taken together with the structures of the mutant and the apo form, this strongly implies that the oxyanion hole is indeed comprised of Tyr 547 -OH and the main chain NH of Tyr 631 . Note, structural alignments between the previously published porcine and human apo structures showed few differences, e.g. side chain positions of Tyr 105 , Phe 208 , Tyr 439 , Tyr 534 , Cys 551 , and Trp 639 . The structural deviations seem to be species related, because the previously published human apo structures are similar to what we observe. Only the human apo structure 1NU6 showed a difference, because this structure lacked the water molecule positioned between Tyr 547 and Ser 630 (26).
In the structurally and functionally related endopeptidase POP (member of the S9a family), a similar mechanism using a tyrosine moiety for oxyanion stabilization has been suggested based on mutagenesis and analysis of kinetic data. This protein consists of an ␣/␤ hydrolase fold, encompassing the catalytic triad composed of Ser 554 , His 680 , and Asp 641 , and a sevenbladed tunnel-forming ␤-propeller, i.e. notably different from the homologue's eight-bladed domain in DPP-IV (18,28). Furthermore, POP is an endopeptidase with a different substrate specificity profile compared with DPP-IV (29). Tyr 473 in the active site cavity of POP, which is homologous to Tyr 547 of DPP-IV, has been suggested as an oxyanion coordination site using a similar mutagenesis strategy as employed in this study of DPP-IV (30). Szeltner et al. (30) demonstrated that the POP mutant Y473F had at least an 8-fold reduction in the kinetic specificity rate constant k cat /K m , depending on which putative substrate was used. They concluded on this basis that Tyr 473 contributes to the transition state stabilization via an oxyanion binding capacity and speculated on whether the hydroxyl group of Tyr 473 also interacts with the substrate carbonyl oxygen in the Michaelis-Menten complex formation. We observed a drop of more than 1,500-fold in k cat /K m for the DPP-IV mutant of the homologues compared with wild type, i.e. a 50-fold drop and a 30-fold increase for k cat and K m , respectively. Thus the POP mutant retained a significantly higher intrinsic activity compared with the homologue's mutant of DPP-IV, an intrinsic activity that was suggested to be the result of bulk water occupying and substituting the position of the ablated hydroxyl group. This notion is not confirmed by the crystallographic FIG. 2. A, SDS-PAGE analysis of DPP-IV High5 cell lysates. 1st lane, control expression of a B12 protein (ϳ70 kDa). 2nd to 5th lanes, DPP-IV mutants. 6th lane, purified DPP-IV. All DPP-IV mutants could be expressed with a molecular weight equal to wild type. B, threestep purification of wild type DPP-IV and Y547F mutants to Ͼ95% purity using first an ADA-Sepharose-coupled affinity resin, second a Q-Sepharose HP ion-exchange column, and third a MonoQ-Sepharose ion-exchange column. 1st lane, molecular weight marker. 2nd lane, purified Y547F mutant. 3rd lane, purified wild type.
Significantly lower when compared with wild type DPP-IV values as determined by Student's t test (p Ͻ 0.005).
results presented here on DPP-IV, where the crystal structure determination of the Y547F mutant on the contrary revealed less bound water. A spatial structural search in the Protein Data Base using the program SPASM (31) revealed that the Tyr 547 residue together with the catalytic triad constitute a unique structure-activity relationship, because the only serine proteases with three-dimensional structures known to date having the spatial arrangement Tyr 547 , Ser 630 , Asp 708 , and His 740 (i.e. numbering according to DPP-IV) were identified as DPP-IV and POP. Thus, only peptidases of the S9 subfamilies seem to contain this catalytic motif, and from the kinetic data obtained in this study on DPP-IV the motif seems even more important for optimal activity than the homologous motif in POP. In addition, the main chain Tyr 631 -NH coordination of the   oxyanion is not in itself sufficient to stabilize the tetrahedral intermediate, and coordination by a tyrosine OH might be needed specifically for this family because of its rare catalytic ability to cleave a peptide bond next to a proline residue, with its particular rigidity and steric constraints. There are other examples of serine proteases that use a side chain for creation of the oxyanion hole, for example subtilisin which beside a main chain NH uses the side chain NH 2 of Asn 155 for the oxyanion stabilization (32). However, subtilisins are not able to cleave peptide bonds around proline residues and constitute a completely different kind of peptidases (33).
One can imagine several possible explanations as to why the S9b family members use a tyrosine hydroxyl group and a backbone NH for the oxyanion stabilization. First of all, the use of tyrosine may simply have evolved to address the particular geometrical requirement for hydrolysis after a proline residue, which may not be possible with the more rigid arrangement of the two donor NHs. Furthermore, Tyr 547 -OH may add conformational flexibility as a result of the functionality of the side chain, whereas the use of two specific main chain NHs in close proximity to the active serine (34,35), as known from most serine proteases, is a structurally more rigid motif. Such increased flexibility of the active site could compensate for the higher inherent rigidity of the "proline substrates." Finally, the tyrosine hydroxy group is capable of acting both as a hydrogen bond donor and acceptor, and it is sufficiently acidic to allow actual transfer of the OH-proton to the developing tetrahedral oxyanion, thereby participating as an anion sink in the charge relay system. Of relevance for this mechanistic possibility is the unique capability of DPP-IV and POP among serine proteases to be inhibited by amino acid 2-cyanopyrrolidine amides such as 1-[({2-[(5-iodopyridin-2-yl)amino]-ethyl}amino)-acetyl]-2-(S)-cyano-pyrrolidine (24), wherein the cyano group of the inhibitor has replaced the peptide bond of the putative peptide substrates. The cyano group of this type of DPP-IV inhibitor when attacked by Ser 630 will give initially an imidate anion, which being very basic could accept a proton from Tyr 547 (23).
This would lead to the inhibitor being bound to Ser 630 as an imidate, which is a stable entity. Intuitively, the location of a protonated species in an oxyanion hole would seem disfavored. However, stable covalent complexes with imidate geometry have been observed in crystal structures of DPP-IV with such inhibitors, both by us 3 and others (24), under conditions where stable acyl-enzyme intermediates have not been observed previously. The observation of such stable acyl-enzyme mimetics, presumed to be neutral imidates, thus supports the possibility that Tyr 547 may function both as a hydrogen bond donor forming the anion hole, and as a hydrogen bond acceptor stabilizing a covalently bound imidate, thereby providing a plausible explanation as to why the therapeutically very promising cyano-pyrrolidines are specific for the S9b family among serine proteases.
The conserved and structurally well defined Wat 123 may be involved in catalysis but could also just be located in the very favorable oxyanion site in the absence of ligands that utilize this functionality. The observation that Wat 123 is present in the DPP-IV⅐Val-Pyr structure and the apo structure but absent in the DPP-IV⅐DFP structure and in the acyl-enzyme mimetic structures points to the latter possibility.
In conclusion, the Y547F mutation decreased activity to less than 1% of wild type, a decrease of the same magnitude as knock-out mutants of the catalytic triad residues Ser 630 , Asp 708 , and His 740 . Structure analysis of the Y547F mutant revealed an intact active site with only a single water molecule absent, which leads us to conclude that the residues Tyr 547 , Ser 630 , Asp 708 , and His 740 are equally important for the catalytic mechanism of DPP-IV. The crystal structure of DPP-IV in complex with DFP, mimicking the tetrahedral intermediate,  , and complex DFP⅐DPP-IV (C, slightly different view, relative to A and B) contoured at 2 (cyan), 3 (red), 5 (purple, only contoured in the apo structure), and 8 (blue, only contoured in the DFP structure). The initial 2F o Ϫ F c electron density map is overlaid the complex DFP⅐DPP-IV contoured at 1 (gray). Structural inspections of the active site of the Y547F mutant reveals a missing water molecule, clearly seen in the wild type apo structure (i.e. hydrogen bonds between Tyr 547 -OH, Ser 630 -OH, and Tyr 631 -NH are indicated). The mutated residue (Phe 547 ) is positioned exactly as the tyrosine residue. The water molecule designated Wat 258 and Wat 421 in the apo and the Y547F mutant structure, respectively, is moved 0.5 Å away from the 547 residue and 0.3 Å (2.9 versus 3.2 Å) closer to the neighboring Tyr 666 -OH (not shown) in the mutant structure. The complex between DFP and DPP-IV showed that the organophosphorous inhibitor was covalently bound to Ser 630 , mimicking the tetrahedral intermediate.
Tyr 547 is not only important but is essential for cleavage of the prolyl peptide bond.