Asparatic Acid 221 Is Critical in the Calcium-induced Modulation of the Enzymatic Activity of Human Aminopeptidase A*

Aminopeptidase A (APA) plays an important role in the regulation of blood pressure by mediating angiotensin II degradation in the renin-angiotensin system. The Ca2+-induced modulation of enzymatic activity is the most characteristic feature of APA among the M1 family of aminopeptidases. In this study, we used site-directed mutagenesis for any residues responsible for the Ca2+ modulation of human APA. Alignment of sequences of the M1 family members led to the identification of Asp-221 as a significant residue of APA among the family members. Replacement of Asp-221 with Asn or Gln resulted in a loss of Ca2+ responsiveness toward synthetic substrates. These enzymes were also unresponsive to Ca2+ when peptide hormones, such as angiotensin II, cholecystokinin-8, neurokinin B, and kallidin, were employed as substrates. These results suggest that the negative charge of Asp-221 is essential for Ca2+ modulation of the enzymatic activity of APA and causes preferential cleavage of acidic amino acid at the N-terminal end of substrate peptides.

such as angiotensin (Ang) II, cholecystokinin-8 (CCK8), and neurokinin B (12,(15)(16)(17)(18). The extracellular domain of APA is composed of two subdomains, a 107-kDa catalytic domain containing an HEXXH motif and a 45-kDa C-terminal domain (19). The C-terminal domain is required for the dimerization of the enzyme and acts as an intramolecular chaperone necessary for the correct folding, intracellular trafficking, and activity of the enzyme (20,21).
APA is expressed in many tissues, including the brush border of intestinal and renal epithelial cells and the vascular endothelium. It has also been identified as a murine B lymphocyte differentiation antigen (BP-1/6C3) and a human kidney (gp160) differentiation antigen (16,17). Employing highly selective APA and APN inhibitors, Reaux et al. (12) elucidated the important roles of these two enzymes in the renin-angiotensin system. It was shown that APA is responsible for the conversion of Ang II to Ang III in the brain, and then Ang III mediates the increase in blood pressure. These results indicate that APA is an important regulator of blood pressure and a potential therapeutic target for the treatment of hypertension. To develop a targeting molecule, it is important to clarify the structural features of the enzyme.
It was shown that Ca 2ϩ up-or down-regulates the enzymatic activity of APA, depending on the substrates tested (22,23). Although the enzyme has rather broad specificity toward synthetic substrates in the absence of Ca 2ϩ , acidic amino acids are preferentially released in the presence of 1.0 mM Ca 2ϩ , indicating that Ca 2ϩ modulates the substrate specificity of APA, increasing its preference for the acidic amino acids, and thus functions as a blood pressure regulator through conversion of Ang II to Ang III (23). Among the M1 family of aminopeptidases, APA is the only enzyme whose activity is modulated by Ca 2ϩ .
It is well known that the extracellular concentration of Ca 2ϩ is strictly regulated and, in general, maintained at ϳ1.0 -2.0 mM. Since the catalytic domain of APA is exposed to and functions in extracellular fluids, it is reasonable to assume that the Ca 2ϩ -mediated preference of the enzyme to the N-terminal acidic amino acids of substrates has some physiological and/or pathological relevance (24).
In this study, we searched for an amino acid residue of human APA (hAPA) responsible for Ca 2ϩ modulation. We identified Asp-221 as a responsive residue. Replacement of Asp-221 with either Asn or Gln caused an almost complete loss of Ca 2ϩ modulation of hAPA. Our data indicate that Asp-221 is essential for Ca 2ϩ modulation and plays an important role in the enzymatic properties of hAPA.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis-The cDNAs encoding for mutant hAPAs were generated by two-step PCR. PCRs were carried out in 0.2-ml tubes with a 30-l reaction volume. First PCR was performed out for one cycle at 98°C for 3.5 min, followed by 15 cycles at 95°C for 0.5 min, at 55°C for 1 min, and at 72°C for 2 min using Pyrobest DNA polymerase (Takara, Kyoto, Japan). Sense primer A (5Ј-CACCAACTCCTAAAAAACCGCCAC-CATGAATCTACTTCTGATCC-3Ј), containing an initial CACC sequence for directional cloning and initiation ATG codon (underlined) and antisense primers complementary to the desired sequences were employed for the amplification of upstream fragments. Downstream fragments with the sequence for His 6 tag at the 3Ј-end were amplified using mutagenic sense primers and primer B (5Ј-TTAGTGATGG-TGATGGTGATGACCACTCTCAAGTAA-3Ј).
The two products of these reactions were used as templates for the second round of PCR. Secondary PCR was carried out with primers A and B for 1 cycle at 98°C for 3.5 min, followed by 25 cycles at 95°C for 0.5 min, at 55°C for 1 min, and at 72°C for 2 min. The resultant products were inserted into the entry vector, pENTR-D-TOPO, using a TOPO-cloning system (Invitrogen). The products were confirmed by automated sequencing on an Applied Biosystems model 377 DNA sequencer.

Expression and Purification of a Soluble Form of Wild-type and Mutant Human APAs in a Baculovirus
System-Expression of soluble forms of wild-type and mutant APAs were performed by the method described previously (23). In brief, the cDNA for soluble forms of wild-type and mutant hAPAs was transferred into a baculovirus transfer vector, pDEST8, using LR clonase (Invitrogen). Then the plasmid was introduced into DH10Bac cells to produce recombinant bacmid DNA containing the cDNA for the soluble form of APA. Next, Sf-9 cells were transfected with the bacmid using Cellfectin reagent (Invitrogen), and after 72 h, recombinant baculoviruses were harvested. For the expression of the soluble forms of hAPAs, Sf-9 cells (2.0 ϫ 10 6 /ml) infected with the recombinant baculovirus (multiplicity of infection of ϳ1-3) were cultured for 72 h in 100 ml of SFM-900 III medium (Invitrogen) at 27°C.
The conditioned medium was collected, dialyzed against 25 mM Tris/HCl buffer (pH 7.5), and then applied to a DEAE-Toyopearl (TOSOH, Tokyo, Japan) column (2.5 ϫ 10 cm) equilibrated in the same buffer and eluted with 200 mM NaCl. The hAPA-containing fractions were pooled, applied to a Co 2ϩ -charged chelating Toyopearl (TOSOH, Tokyo, Japan) column (1.0 ϫ 10 cm), and then eluted with 200 mM imidazole. The active fractions were collected, concentrated, and subjected to further analysis.
Measurement of Aminopeptidase Activity of APA-The aminopeptidase activity of the recombinant hAPA was determined with various fluorogenic aminoacyl-4-metheylcoumaryl-7-amides (aminoacyl-MCAs) as substrates. Typically, the reaction mixture containing various concentrations of aminoacyl-MCA and the enzyme in 0.5 ml of 25 mM Tris/HCl buffer (pH 7.5) with or without 1.0 mM Ca 2ϩ ion was incubated at 37°C for 5 min. The reaction was terminated by adding 2.5 ml of 0.1 M sodium acetate buffer (pH 4.3) containing 0.1 M sodium monochloroacetate. The amount of 7-amino-4-methylcoumarin released was measured by spectrofluorophotometry (F-2000; Hitachi) at an excitation wavelength of 360 nm and an emission wavelength of 460 nm and shown as arbitrary units. The kinetic parameters were calculated from Lineweaver-Burk plots. The results are represented by K m , k cat , and k cat /K m values. K i values were calculated from the Dixon plot using Glu-MCA as a substrate. In this study, each enzyme was used at either 0.5 or 1.0 g/ml. Enzymatic activity was not affected by dilution, since there was no difference between activity measured in the presence or absence of 10 g/ml of bovine serum albumin. All measurements were performed in triplicate.
Cleavage of Peptide Hormones by APA-Peptide hormones (Peptide Institute, Osaka, Japan) (25 M) were incubated with the enzyme (1.0 g/ml) at 37°C in 25 mM Tris/HCl buffer (pH 7.5) containing 50 mM NaCl. The reaction was terminated by adding 2.5% (v/v) formic acid. The peptides generated were separated by reverse phase HPLC (AT-10; Shimazu) on a COSMOSIL (4.6 ϫ 250-mm) column (Nacalai Tesque, Kyoto, Japan) at a flow rate of 0.5 ml/min. The buffers used for the isocratic separation of the peptides were as follows: for the peptides from Ang II or kallidin, 19% acetonitrile containing 0.086% trifluoroacetic acid; for the peptides from CCK8 or neurokinin B, 35% acetonitrile containing 0.083% trifluoroacetic acid. The molecular masses of peptides were determined by matrix-assisted laser desorption ionization time of flight (MALDI-TOF) MS with a REFLEX mass spectrometer (Bruker-Franzen Analytik) using ␣-cyano-4-hydroxycinnamic acid as the matrix.
Molecular Modeling of Human APA-The recently published x-ray crystallographic structures of human leukotriene A 4 hydrolase (25) and Thermoplasma acidophilum TIFF3 (Tricorn-interacting factor F3) (26) were used as templates for modeling the catalytic site of hAPA using the SWISS-MODEL Internet server. The structure was displayed using the CueMol program (R. Ishitani, CueMol: Molecular Visualization Framework).

Identification of a Residue Responsive to Ca 2ϩ
Modulation of APA-Among the M1 family of aminopeptidases, only APA is modulated in its enzymatic activity by Ca 2ϩ . To identify a residue responsible for Ca 2ϩ modulation, we reasoned that the critical residue should be polar as a Ca 2ϩ -binding residue and unique among the family. We also postulated that the corresponding residue should be important for the enzymatic activity of the family members. Hence, we searched for a residue conserved among all of the M1 aminopeptidases except for APA. As shown in Fig. 1, alignment of the M1 family indicated that Asp-221 of hAPA is significant in that all other members of the family besides leukocyte-derived arginine aminopeptidase and aminopeptidase O carry Gln at the corresponding site (27). Therefore, we examined the role of this residue in the Ca 2ϩresponsive enzymatic activity of hAPA hereafter.
Preparation of Wild-type and Mutant Enzymes-To elucidate the significance of Asp-221 in the Ca 2ϩ modulation as well as the enzymatic activity of hAPA, wild-type, and mutant hAPAs carrying a single substitution of this residue (i.e. D221E, D221Q, D221N, and D221A hAPAs) were transiently expressed in a baculovirus system and purified to homogeneity on SDS-PAGE. As expected, all of the enzymes showed a single band with apparent molecular masses of ϳ115 kDa in the reducing condition (data not shown).

Characterization of the Hydrolytic Activities of Wild-type and Mutant Enzymes toward Synthetic
Substrates- Fig. 2 shows enzymatic activity of wild-type and mutant hAPAs toward various synthetic substrates measured in the presence or absence of Ca 2ϩ . The hydrolytic activity of wild-type enzyme toward Glu-MCA was up-regulated 2.1fold, and that toward Gln-MCA was reduced to 15% in the presence of 1.0 mM Ca 2ϩ , confirming Ca 2ϩ modulation of this enzyme as reported previously (23). When the enzymatic activity of D221E hAPA was tested, this mutant still retained some Ca 2ϩ modulation, although the overall hydrolytic activity was significantly lower. Namely, Ca 2ϩ caused a 30% enhancement of hydrolytic activity toward Glu-MCA and a 52% decrease toward Gln-MCA compared with the mutant activity in the absence of Ca 2ϩ . On the other hand, both D221Q and D221N hAPAs revealed no Ca 2ϩ modulation at all, although they retained Glu-MCA hydrolytic activity comparable with that of wild-type enzyme in the absence of Ca 2ϩ . The enzymatic activity of D221A hAPA was significantly reduced. This mutant also lost Ca 2ϩ modulation. These results strongly suggested that Asp is the most favorable residue at amino acid 221 of hAPA, both for its enzymatic activity and Ca 2ϩ modulation. Table 1 shows the kinetic parameters of wild-type and mutant hAPAs toward Glu-MCA measured in the presence or absence of 1.0 mM Ca 2ϩ . Ca 2ϩ decreased the K m values and thus enhanced the catalytic efficiency of both wild-type and D221E hAPAs. Two mutant enzymes, D221N and D221E hAPAs have values comparable with wild-type hAPA in all kinetic parameters in the absence of Ca 2ϩ . Moreover, Ca 2ϩ had little effect on the parameters to these mutants. D221E and D221A hAPAs had lower aminopeptidase activity due to the decrease in the affinity and the turnover numbers, causing a significant decrease in the catalytic efficiency to the substrates.   We then examined the effects of Ca 2ϩ and Zn 2ϩ on the enzymatic activities of wild-type, D221E, and D221N hAPAs. As shown in Fig. 3A, Ca 2ϩ enhanced the hydrolytic activity of wildtype enzyme in a dose-dependent manner, with the maximum activity obtained at around 0.1 mM, as reported previously (23). D221E hAPA activity was also enhanced clearly by Ca 2ϩ in a dose-dependent manner, although it was less responsive than that of wild-type. On the other hand, Ca 2ϩ had little influence on the enzymatic activity of D221N hAPA, even at concentrations as high as 2.5 mM.
Although the mechanism is unknown, Zn 2ϩ is a strong and general inhibitor of the M1 family of aminopeptidases (3). As shown in Fig. 3B, Zn 2ϩ inhibited the enzymatic activities of these three enzymes equally, indicating the specific role of Asp-221 in the Ca 2ϩ modulation.
Effects of Ca 2ϩ on the Hydrolytic Activities of Wild-type and D221N hAPAs toward Peptide Substrates-We compared hydrolytic activities of wild-type and D221N hAPAs toward natural hormones that were cleaved by the enzyme. It was shown that Ca 2ϩ enhances the hydrolytic activity of hAPA toward several peptides having acidic amino acids at their N-terminal ends, such as Ang II, CCK8, and neurokinin B (23). Fig. 4 compares the cleavage of Ang II by wild-type and D221N hAPAs. The wild-type enzyme cleaved 90% Ang II to Ang III within 20 min in this assay system in the presence of 1.0 mM Ca 2ϩ . In the absence of Ca 2ϩ , the conversion to Ang III was only 41%, indicating that the cleavage of Ang II to Ang III by the enzyme is Ca 2ϩ -responsive, as reported previously (23). On the other hand, conversion of Ang II to Ang III by Ca 2ϩ -irresponsive D221N hAPA was nearly the same both in the presence and absence of Ca 2ϩ , indicating that, as in the case with synthetic substrates, hydrolytic activity of this mutant enzyme toward Ang II was not modulated by Ca 2ϩ . However, to our surprise, irrespective of the presence of Ca 2ϩ , hydrolytic activity of D221N hAPA was comparable with that of wild-type enzyme measured in the presence of Ca 2ϩ . D221N hAPA cleaved 89% Ang II in the presence of Ca 2ϩ , and 98% conversion was observed in its absence. A similar tendency was observed when   the concentration of the mutant was reduced. Almost the same result was obtained with D221Q hAPA. It should be noted that probably because of decreased enzymatic activity, Ca 2ϩ -mediated enhancement of the hydrolytic activity of D221E hAPA toward Ang II was barely detectable in the same assay (data not shown). As expected, the hydrolytic activity of D221A hAPA was significantly reduced and also irresponsive to Ca 2ϩ (data not shown). The hydrolytic activities of wild-type and D221N hAPAs toward CCK8 and neurokinin B were then compared (Fig. 5). As shown previously (23), Ca 2ϩ enhanced the hydrolytic activity of wild-type enzyme toward these substrates. Again, D221N hAPA was rather irresponsive to Ca 2ϩ , with comparable activity observed both in the presence and absence of Ca 2ϩ . However, in contrast to Ang II, irrespective of the presence of Ca 2ϩ , the hydrolytic activities of this mutant enzyme were comparable with those of wild type measured in the absence of Ca 2ϩ , suggesting that the activity remained low despite the presence of Ca 2ϩ . Taken together, Asp-221 of the wild-type hAPA is crucial for Ca 2ϩ modulation of the enzymatic activity toward peptide substrates with an acidic amino acid at their N-terminal ends as well as toward synthetic substrates.
We next compared the conversion of kallidin to bradykinin by wild-type and D221N hAPAs (Fig. 6). As shown previously (23), Ca 2ϩ reduced the activity of wild-type enzyme. Although 74% conversion was observed within 20 min in the absence of Ca 2ϩ , only 37% was observed with Ca 2ϩ . On the other hand, the enzymatic activity of D221N hAPA toward kallidin was rather high both in the presence (80% conversion) and absence (96% conversion) of Ca 2ϩ , indicating that D221N hAPA is also less responsive to the ion than wild-type in this conversion assay. The converting activity of D221N hAPA was slightly higher than that of wild type measured in the absence of Ca 2ϩ , suggesting that the mutant enzyme retained high activity irrespective of the presence of Ca 2ϩ . These results indicate that Asp-221 is also responsible for Ca 2ϩ down-regulation of hAPA activity toward kallidin, which has a basic amino acid at the N-terminal end. Table 2 shows the kinetic parameters of wild-type and D221N hAPAs toward Ang II, CCK8, neurokinin B, and kallidin measured either in the presence or absence of 1.0 mM Ca 2ϩ . As in the case with synthetic substrates, Ca 2ϩ affected mainly K m values of wild type to either increase or decrease the catalytic efficiency to the substrates. In the absence of Ca 2ϩ , D221N hAPA showed values in all kinetic parameters comparable with those of wild-type enzyme toward CCK8 and neurokinin B. On the other hand, this mutant had higher affinity toward Ang II and higher turnover numbers toward kallidin, causing higher catalytic efficiency than wild-type enzyme in the absence of Ca 2ϩ . As in the case with synthetic peptides, Ca 2ϩ had little effect on the kinetic parameters of the mutant enzyme toward all peptide substrates tested. Table 3 shows the effect of Ca 2ϩ on K i values of APA inhibitors amastatin and Ang IV to wild-type and D221N hAPAs. In the presence of 1.0 mM Ca 2ϩ , K i values of amastatin and Ang IV to wild-type hAPA increased by 8.0-and 3.3-fold, respectively, when compared with the values measured in the absence of Ca 2ϩ , indicating that the ion lowers the affinity of inhibitors to the enzyme. On the contrary, although the K i value of Ang IV to D221N hAPA increased only slightly, that of amastatin was significantly low. These results indicate that the effects of Ca 2ϩ on affinity of inhibitors to D221N hAPA were far less than those to wild-type APA. It is plausible that although Ca 2ϩ induces significant structural change in the substrate pocket of wild-type hAPA, Ca 2ϩ -induced change of the mutant hAPA should be less significant than wild-type.

DISCUSSION
In this study, we have identified Asp-221 of hAPA as a residue responsible for Ca 2ϩ modulation of the enzyme. Depending on the substrate tested, Ca 2ϩ up-or down-regulates the enzymatic activity, causing the preferential cleavage of N-terminal acidic amino acids by the enzyme (22,23). Since APA has been recognized as an important enzyme regulating blood pressure by degrading Ang II (12), it is important to elucidate the mechanism of the enzymatic action and identify an amino acid residue of the enzyme responsible for Ca 2ϩ modulation of the enzyme.
The mode of enzymatic action of murine APA (mAPA) has been extensively characterized by mutational analysis. It was initially shown that His-385, His-389, and Glu-408 in the 385 HEXXHX 18 E 408 gluzincin motif function as zinc ligands and are essential for the catalytic activity of the enzyme (28,29). Presumably, the zinc atom is coordinated by three ligands and a water molecule. Glu-386 in the HEXXH motif polarizes the water molecule and promotes the nucleophilic attack of the carbonyl carbon of the peptide bond, forming a tetrahedral intermediate. It was also shown that Glu-352 and Asn-353 in the GAMEN motif play roles in the exopeptidase specificity of mAPA through interaction with the N-terminal amino acid of the substrate (30,31). Tyr-471 of mAPA contributes to the catalytic activity of the enzyme by stabilizing the transition state complex through interaction of the tyrosine hydroxy group with the oxyanion of the tetrahedral intermediate via a hydrogen bond (32).
Ca 2ϩ modulation makes APA unique among the M1 family of aminopeptidases. Iturrioz et al. (33) reported that His-450 of mAPA is responsible for the modulation. Replacement of His-450 of mAPA, which is conserved among most of the M1 family, with Phe (H450F mAPA) caused a significant decrease in the enzymatic activity. More importantly, compared with the wildtype enzyme, the mutant enzyme required a higher concentration of Ca 2ϩ for maximal enzymatic activity. We confirmed these results in that corresponding H458F hAPA was also less responsive to the ion than was wild-type. However, Ca 2ϩ regulation of the enzymatic activity was still apparent, with a 2.5fold increase in activity toward Glu-MCA and an 85% decrease toward Gln-MCA observed in the presence of 1.0 mM Ca 2ϩ (data not shown).
We have identified in this study Asp-221 as another residue responsible for Ca 2ϩ modulation of the enzyme. Replacement of Asp-221 with Asn or Gln caused almost complete loss of modulation, whereas hydrolytic activities of the mutant enzymes toward Glu-MCA in the absence of Ca 2ϩ were  retained comparable with wild-type. Although hydrolytic activity was reduced significantly, D221E hAPA retained some Ca 2ϩ modulation, suggesting strongly that a negative charge at this site is required for Ca 2ϩ modulation. Considering that D221A hAPA showed a significant decrease in the enzymatic activity and a complete loss of Ca 2ϩ sensitivity, the length of the side chain of this residue is critical in the enzymatic activity of the enzyme. In fact, D221L hAPA had considerable enzymatic activity toward synthetic substrates (data not shown). Taken together, our data indicate for the first time that Asp-221 is required for the maximum activity and Ca 2ϩ modulation of the enzyme. Fig. 7 shows the possible catalytic mechanisms of hAPA based on the data discussed above. As for the role of Ca 2ϩ in the enzymatic activity of hAPA, it is plausible that the ion bridged the interaction between Asp-221 at the S1 site and an N-terminal acidic residue at the P1 site of the substrates (Fig.  7A). In the absence of Ca 2ϩ , loss of the interaction might make the P1 residue more flexible, causing the relatively broad substrate specificity of and the observed increase in the affinity of inhibitors to wild-type enzyme. On the other hand, it is also possible that the interaction of Asp-221 and Ca 2ϩ may induce the conformational change of another unidentified residue of the S1 site to facilitate the interaction of a substrate's N-terminal residue with the putative residue (Fig. 7B). It is noteworthy that human leukocyte-derived arginine aminopeptidase, which also has Asp at the corresponding site of hAPA, is not Ca 2ϩresponsive (27). At present, however, it is still difficult to explain all of the aspects observed in this study, including the Ca 2ϩ -mediated decrease in the affinity toward substrates having an N-terminal nonacidic amino acid, such as kallidin. APA cleaved Ang II, CCK8, and neurokinin B, all of which have acidic amino acids at their N-terminal ends, more effectively in the presence of Ca 2ϩ than in its absence (23). Therefore, we initially expected that cleavage of these substrate peptides by Ca 2ϩ -irresponsive mutant enzymes should be maximal in the presence of saturation levels of Ca 2ϩ (i.e. 1.0 mM) and minimal in its absence. However, we found that two Ca 2ϩ -irresponsive mutants, D221Q and D221N hAPAs, revealed activities toward Ang II comparable with wild-type enzyme measured in the presence of Ca 2ϩ , indicating that these two mutants retain high activities irrespective of the presence of Ca 2ϩ . In contrast, the hydrolytic activity of D221N hAPA toward CCK8 and neurokinin B both in the presence and absence of Ca 2ϩ were comparable with that of wild-type hAPA measured in the absence of Ca 2ϩ , indicating that CCK8 and neurokinin B hydrolytic activity of the mutant enzyme remains low even in the presence of Ca 2ϩ . These results suggest that, depending on the substrates tested, the Ca 2ϩ -irresponsive mutant enzymes exert either maximum or minimum activities constantly toward peptide substrates having acidic amino acids at their N-terminal ends irrespective of the presence of Ca 2ϩ . Wildtype enzyme exerts its maximum activity only in the presence of a saturation level of Ca 2ϩ .
In addition, although conversion of kallidin to bradykinin by wild-type enzyme was maximal in the absence of Ca 2ϩ , the mutant enzymes retained nearly the full activity even in the presence of Ca 2ϩ . Therefore, as in the case with Ang II, the mutant retains maximum activity irrespective of the FIGURE 7. Possible mechanism of the enzymatic action of hAPA. A, Asp-221 of hAPA acts as the S1 site of the enzyme, bridging the P1 site of the substrate through interaction with Ca 2ϩ . B, interaction of Ca 2ϩ with Asp-221 induces the conformational change of the unidentified S1 site, which in turn mediates preferential cleavage of substrates having acidic amino acids at the N terminus. C, modeling of the catalytic site of hAPA using human LTA4H and T. acidophilum TIFF3 as templates. Glu-360 in the GAMEN motif, His-393, Glu-394, His-397, and Glu-416 in the HEXXHX 18 E motif and Tyr-479 are shown as residues in the catalytic site of the enzyme. presence of Ca 2ϩ toward kallidin, the N-terminal amino acid of which is lysine. It is noteworthy here that both Ang II and kallidin, of which activities were enhanced by Ca 2ϩ , have an arginine residue at the P2 position.
Taken together, it is plausible that Ca 2ϩ induces structural change around the catalytic site of the enzyme, enhancing preference for the acidic amino acids of the substrates. In fact, Ca 2ϩ induced significant changes of K i values of the inhibitors to wild-type hAPA but not to D221N hAPA, suggesting that Ca 2ϩ induces a significant structural change of wild-type enzyme around its catalytic site but not of irresponsive mutants. Modeling of the catalytic site of the enzyme based on the structure of leukotriene A 4 hydrolase and TIFF3 as templates suggests that Asp-221 is located near the catalytic HEXXHX 18 E motif (Fig.  7C) (25,26). Looking at the spatial relationship between Asp-221 and Glu-360, interaction of Asp-221 with the side chain of the P1 residue of substrate via Ca 2ϩ seems plausible. To elucidate the molecular basis of the Ca 2ϩ -induced modulation of the enzymatic activity of hAPA, it is essential to determine the three-dimensional structure of the enzyme exactly.
In this study, we have shown that Asp-221 of hAPA is important for Ca 2ϩ modulation as well as enzymatic activity of the enzyme. The Ca 2ϩ modulation is the most characteristic and unique feature of APA among the M1 family of aminopeptidases. Interestingly, the Ca 2ϩ -irresponsive mutant enzymes that have activities comparable with wild-type enzyme toward Glu-MCA in the absence of Ca 2ϩ retained either maximum or minimum activities, depending on the peptide substrates tested. Our data will be useful in clarifying the molecular basis of the Ca 2ϩ -mediated preference of the enzyme for acidic amino acids at the N-terminal end of substrates.