Identification of a glutamic acid and an aspartic acid residue essential for catalytic activity of aspergillopepsin II, a non-pepsin type acid proteinase.

Aspergillopepsin II from Aspergillus niger var. macrosporus is a non-pepsin type or pepstatin-insensitive acid proteinase. To identify the catalytic residues of the enzyme, all acidic residues that are conserved in the homologous proteinases of family A4 were replaced with Asn, Gln, or Ala using site-directed mutagenesis. The wild-type and mutant pro-enzymes were heterologously expressed in Escherichia coli and refolded in vitro. The wild-type pro-enzyme was shown to be processed into a two-chain active enzyme under acidic conditions. Most of the recombinant mutant pro-enzymes showed significant activity under acidic conditions because of autocatalytic activation except for the D123N, D123A, E219Q, and E219A mutants. The D123A, E219Q, and E219A mutants showed neither enzymatic activity nor autoprocessing activity under acidic conditions. The circular dichroism spectra of the mutant pro- and mature enzymes were essentially the same as those of the wild-type pro- and mature enzyme, respectively, indicating that the mutant pro-enzymes were correctly folded. In addition, two single and one double mutant pro-enzyme, D123E, E219D, and D123E/E219D, did not show enzymatic activity under acidic conditions. Taken together, Glu-219 and Asp-123 are deduced to be the catalytic residues of aspergillopepsin II.

The regular aspartic proteinase family (1)(2)(3)(4)(5) includes such enzymes as pepsin, chymosin, cathepsins D and E, renin, and fungal and retroviral pepsins. The family is homologous both in primary and tertiary structures so they are thought to share a common ancestor in evolution. Generally, members of the aspartic proteinase family are composed of two homologous domains, each containing a catalytic aspartyl residue in a consensus sequence of Asp-(Thr/Ser)-Gly, in close proximity to enable catalytic function (6). The proteinases are often characterized by susceptibility to specific inhibitors such as pepstatin A (7), 1,2epoxy-3-(p-nitrophenoxy)propane (8), and the diazoacetyl-DL-norleucine methyl ester in the presence of cupric ions (9).
Aspergillopepsin II is unique in the primary structure of known proteins with the exception of scytalidopepsin B (20) and the EapB and EapC proteins from C. parasitica (14). Aspergillopepsin II consists of two polypeptide chains, a 39-residue light chain and a 173-residue heavy chain (11), which are derived from a precursor polypeptide of 282 residues (21). Although this two-chain structure is different from the one-chain structures of scytalidopepsin B and the EapB and EapC proteins, a 44 -65% identity in amino acid sequence indicates that these proteins have been derived in evolution from the same ancestral protein and presumably share common catalytic residues and mechanisms. Recently Barrett et al. (22) have classified aspergillopepsin II and its homologs as members of the aspartic proteinases of family A4.
The catalytic mechanisms of these pepstatin-insensitive acid proteinases are not well understood. Recently, Oyama et al. (23) identified two aspartic acids as the catalytic residues of pseudomonapepsin and xanthomonapepsin that belong to the aspartic proteinases of family A7 (22). In the proteinases of family A4, aspartic and glutamic acid were previously proposed as the catalytic residues of scytalidopepsin B (24,25). However, it was later determined that neither residue was conserved among the enzymes of family A4 and that the putative glutamic acid residue is actually encoded as a glutamine in the cDNA of scytalidopepsin B, similar to other enzymes of family A4 (26). Thus, the catalytic residues of the proteinases of family A4 still remain to be identified.
Aspergillopepsin II has an optimum pH of 2.6 for milk casein (10) and of less than 2 for hemoglobin, 1 implicating acidic residues in the catalytic mechanism of the enzyme. Therefore, acidic residues conserved within the proteinases of family A4 may be candidates for these catalytic residues. Because neither the tertiary structure nor the specific inhibitor of aspergillopepsin II is known, analysis of the conserved residues by site-directed mutagenesis is the most practical approach for identifying the catalytic residues. In this study, all the conserved acidic residues were individually replaced with other amino acids by site-directed mutagenesis, and the mutant proenzymes were prepared by heterologous expression followed by in vitro refolding and analysis. The results have led us to conclude that Glu-219 (Glu-110 in the heavy chain) and Asp-123 2 (Asp-14 in the heavy chain) are the most probable catalytic residues of the enzyme.

EXPERIMENTAL PROCEDURES
Materials-Restriction endonucleases, DNA amplification reagents, and Taq DNA polymerases were purchased from Toyobo (Tokyo, Japan), and a T4 DNA ligation kit was purchased from Takara (Otsu, Japan). The reagents for DNA sequencing were from Perkin-Elmer. All other reagents used were of analytical grade and obtained from Wako Pure Chemicals (Tokyo, Japan). The expression vector pAR2113 and Escherichia coli BL21 (DE3) were obtained from Dr. F. W. Studier at Brookhaven National Laboratory, NY.
Expression, Partial Purification, and Refolding of Aspergillopepsinogen II-The expression plasmid pAR-ANA was constructed in the same manner as that for aspergillopepsinogen I (27). The region coding for a putative pro-enzyme of aspergillopepsin II (residues 19 -282 of the prepro-enzyme) (21) was amplified by polymerase chain reaction using genomic DNA for a template and a set of primers as follows: sense, 5Ј-AGGATCATATGGCTCCTCTCACTGAGAAGCGC-3Ј; and antisense, 5Ј-TATATGGATCCTATTAGACGTAGGTGACAGTGAC-3Ј. The sense and antisense primers contain a NdeI and BamHI site, respectively. The amplified DNA was digested with NdeI and BamHI and was inserted into the NdeI/BamHI site of pAR2113. Expression was carried out according to the method of Studier et al. (28). For preculture, 0.1 ml of a glycerol stock of transformed E. coli BL21 (DE3) with pAR-ANA was seeded into 10 ml of M9ZB broth (28) containing 0.2 mg/ml ampicillin. After shaking at 37°C for a few hours, the culture became slightly turbid. It was then seeded into 1 liter of M9ZB broth containing 0.05 mg/ml ampicillin and incubated at 37°C with shaking. When the absorbance at 600 nm reached 0.6 -1.0, the expression of aspergillopepsinogen II was induced by the addition of isopropyl-␤-D-thiogalactopyranoside (final concentration, 1 mM). The culture was further incubated at 37°C for 150 min with shaking. The cells were harvested by centrifugation and suspended in 50 ml of a lysis buffer (50 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride). The suspension was frozen in liquid nitrogen and then thawed. To this was added lysozyme to a final concentration of 1 mg/ml, and the mixture was kept on ice for 30 min. The resulting solution was subjected to sonication (100 watts for 20 s, repeated four times) and then centrifuged at 17,000 ϫ g for 10 min. After removing the supernatant, the pellet was dissolved in 50 ml of a denaturation buffer (8 M urea, 100 mM 2-mercaptoethanol, 50 mM NaCl, 1 mM EDTA, 50 mM sodium phosphate, pH 6.2). After standing at room temperature for 1 h, the solution was centrifuged at 100,000 ϫ g for 30 min. The supernatant was loaded onto a DE52 (Whatman) column (1.5 ϫ 6 cm) and was eluted with a linear gradient of 0.05-0.35 M NaCl. The fractions containing denatured aspergillopepsinogen II were pooled. The solution of denatured aspergillopepsinogen II was diluted with 10 volumes of a refolding buffer (50 mM sodium acetate, pH 5.25) and left to stand overnight at 4°C. Then the solution was dialyzed against the refolding buffer at 4°C for 24 h.
Purification of Recombinant Aspergillopepsin II-The wild-type and potentially active mutant pro-enzymes were processed to various extents during dialysis. For full conversion into the mature form, each solution after the refolding step was dialyzed against a processing buffer (50 mM sodium acetate, pH 3.5) overnight at 4°C. Then the solution was dialyzed against 20 mM sodium citrate, pH 4.5, and applied to a DEAE-Toyopearl (TOSOH) column (2 ml). Elution was performed with a linear gradient in 20 mM sodium citrate from pH 4.5 to 2.5.
Assay of Proteolytic Activity-Proteolytic activity was determined essentially by the method of Anson (29). An aliquot of enzyme or pro-enzyme (0.2 ml) was mixed with 0.4 ml of 2% acid-denatured hemoglobin in 100 mM sodium phosphate, pH 2.0, and the mixture was incubated at 37°C for 30 min. To this was added 0.8 ml of 5% (w/v) trichloroacetic acid, and the mixture was allowed to stand for 30 min and then centrifuged at 8,000 ϫ g for 10 min. One unit of activity was defined as the amount of enzyme or pro-enzyme that gives an increase of 1.0 absorbance unit/min at 280 nm. In this assay, the activity of the pro-enzyme actually indicates the potential enzyme activity because the pro-enzyme is rapidly autoactivated under the assay conditions at pH 2.
Determination of Amino Acid Sequences-The amino acid sequences of the N-terminal region of the pro-enzyme and its activation products were determined with an automated protein sequencer model 477A, equipped with an on-line HPLC (PE Biosystems).
Site-directed Mutagenesis-Oligonucleotide-directed mutagenesis was performed by the method of Kunkel (30).
CD Measurements-The CD spectra were measured at a protein concentration of 0.1 mg/ml in buffer with a Jasco J-720 spectropolarimeter at room temperature using the water-jacketed quartz cell with a light path of 1 mm. The sample solutions were prepared by appropriately diluting the solutions of the refolded pro-enzyme and the activated mature enzyme with 50 mM sodium acetate buffer at pH 5.25 or 3.5, respectively, just before measurement. For all measurements, a 1.0-nm bandwidth and a 1.0-s time constant were used, and 32 scans were repeated from 200 to 250 nm at the speed of 50 nm/min with 0.1 nm/point resolution. Spectra were converted to mean residue molecular ellipticity prior to analysis.

Preparation of Recombinant Aspergillopepsinogen II-To
identify the catalytic residues by site-directed mutagenesis, aspergillopepsinogen II was expressed in E. coli. The region of the cDNA coding for the putative pro-enzyme (the residues 19 -282 of the prepro-enzyme) (21) was placed downstream of the T7 promoter in an expression vector pAR2113. The encoded polypeptide contained the putative pro-segment (residues 19 -59), light chain (residues 60 -98), intervening sequence (residues 99 -109), and heavy chain (residues 110 -282) (Fig. 1). Using this expression plasmid, the pro-enzyme was expressed in E. coli BL21 (DE3). Proteolytic activity toward acid-denatured hemoglobin was detected in the cytosol fraction by activity staining after non-denaturing PAGE. 3 Western blot analysis using an anti-aspergillopepsin II antiserum, however, showed that the major part of the expressed protein formed inclusion bodies (data not shown).
Because most of the expressed protein formed inclusion bodies, extraction took place under denaturing conditions, and then protein was refolded in vitro as described under "Experimental Procedures." The pro-enzyme was eluted as a single protein peak from the DE52 column (data not shown) and showed a single band on SDS-PAGE ( Fig. 2A, lane 1). The behavior on SDS-PAGE of recombinant aspergillopepsinogen II as well as native aspergillopepsin II was anomalous. The molecular masses of recombinant aspergillopepsinogen II and the heavy chain of native aspergillopepsin II as estimated by SDS-PAGE were 42 ( Fig. 2A, lane 1) and 38 kDa (Fig. 2B), respectively. This is approximately twice as large as the theoretical molecular masses of 28 and 18 kDa, respectively. The reason for this is not certain at present but may be at least partly because of the abundance of acidic residues.
The N-terminal sequence of the expressed protein was determined to be Ala-Pro-Leu-Thr-Glu, which is consistent with that of the recombinant sequence except for the lack of the initial methionyl residue. Because the protein extracted from the inclusion bodies was completely denatured, refolding in vitro was required to obtain a potentially active pro-enzyme.
The fraction containing the denatured pro-enzyme was diluted with the refolding buffer, pH 6.0, without urea and 2-mercaptoethanol to allow refolding. Then, the refolding mixture was dialyzed against the same buffer to completely remove the reagents. The resulting solution showed proteolytic activity toward acid-denatured hemoglobin at pH 2.0.
Optimum pH for the Refolding of Aspergillopepsinogen II-The efficiency of refolding was examined by assaying proteolytic activity toward hemoglobin at pH 2 after refolding of the proteinase at several different pH values between 3.25 and 8.5 followed by dialysis against 20 mM sodium acetate buffer, pH 3.5. The pH of the refolding buffer used for dilution of the denatured pro-enzyme solution had a marked effect on the refolding efficiency (Fig. 3). The optimum pH was around 5.3. When the pH was shifted by as little as 1 pH unit in either direction, the refolding efficiency was greatly decreased. At pH  (14). The alignment was generated using Clustal W1.7 (39). Identical residues are boxed, and identical acidic residues are shaded. Residues are numbered according to aspergillopepsinogen II. The prepro sequence of scytalidopepsin B has not been reported. Dashes indicate deletions. 6 -8.5, a low but significant extent of refolding was found to occur. At pH 5.25, the refolding efficiency was estimated to be 38%. The pH of the buffer used for dialysis after dilution had no effect on the refolding efficiency when it was performed in the range of pH 3-7.
Activation of Aspergillopepsinogen II-During the refolding under favorable conditions (for example at pH 5.3), the proenzyme was gradually converted to the processed form ( Fig. 2A,  lane 3), whereas the polypeptide that was allowed to refold at pH 3.7 or 7.3 retained a molecular size identical to that of the pro-enzyme ( Fig. 2A, lanes 2 and 4). The pH dependence of the processing was investigated with a sample refolded at pH 6 where the majority of the protein remained as the pro-enzyme. When this sample was acidified and incubated at 37°C, the pro-enzyme was converted to a form indistinguishable in size from the native enzyme. The analysis of the time course of the processing by SDS-PAGE showed that the optimum pH for the processing was in the range of pH 3-4 (data not shown). At above pH 6, the processing was very slow, and no processing was observed at this pH within 4 h as examined by SDS-PAGE.
Purification of Recombinant Aspergillopepsin II-Denatured aspergillopepsinogen II was purified in the presence of 8 M urea to give a single strong band on SDS-PAGE. However, autolysis occurring in the subsequent refolding and purification processes was a major problem affecting the purification of the refolded pro-enzyme. We have not been successful in com-pletely separating the pro-enzyme from truncated forms. Thus, after refolding, the refolded sample was dialyzed at pH 3.5 to allow simultaneous processing to the mature enzyme. After the processing, active aspergillopepsin II was purified by column chromatography on DEAE-Toyopearl using a pH gradient of 4.5-3.0. The purified enzyme gave two bands on SDS-PAGE corresponding to the heavy (H) and light (L) chains (Fig. 2B), and its specific activity toward hemoglobin was essentially the same as that of the native enzyme. The yield of activity at this purification step was about 80%.
Properties of the Recombinant Aspergillopepsin II-The Nterminal sequences of the light and heavy chains of the recombinant aspergillopepsin II were determined to be Glu-Glu-Tyr-Ser-Ser and Lys-Arg-Gln-Ser-Glu-Glu-Tyr, respectively. The N terminus of the light chain was Glu-60, which was identical to that of the native enzyme. On the other hand, the N terminus of the heavy chain was Lys-108, which was different from the native heavy chain pyroglutamyl residue that corresponds to Gln-110 (Fig. 1). Despite the difference in the N-terminal structure of the heavy chain and some heterogeneity in the C terminus of the light chain, no difference was observed in enzymatic properties between the recombinant and the native enzymes. The dependence on pH and temperature of their activity toward hemoglobin was indistinguishable, and both enzymes showed maximal activity at 55-60°C and a large decrease in activity at 60 -65°C. CD Spectra of Recombinant Aspergillopepsinogen II and Aspergillopepsin II-The CD spectrum of aspergillopepsinogen II at pH 5.25 after refolding at the same pH (Fig. 4A, curve a) was different from that of native aspergillopepsin II at pH 3.5 (Fig.  4B). The spectrum of the recombinant pro-enzyme was changed by acidification (Fig. 4A, curve b). After incubation at pH 3.5, the spectrum was indistinguishable from that of the native enzyme (Fig. 4B). This change in spectrum was irreversible; the readjustment of pH to 5.25 did not change the spectrum (Fig. 4A, curve c).
Site-directed Mutagenesis of the Conserved Acidic Residues-To investigate the importance of the acidic residues to catalysis, all the conserved Asp and Glu residues, i.e. Asp-123, -125, -137, -148, -160, -170, and -220, and Glu-152, -189, -215, -219, and -222 (Fig. 1), were individually replaced using sitedirected mutagenesis with Asn and Gln, respectively. In the exploratory experiments, the mutant pro-enzymes were prepared in a small scale. Without purification of the denatured pro-enzyme extracted from the inclusion bodies, they were allowed to refold at pH 5.25, and the activity in the crude sample solution was assayed (Fig. 5A). After refolding, the mutants D125N, D137N, D148N, E152Q, D160N, D170N, E189Q, D220N, and E222Q were found to be markedly active toward acid-denatured hemoglobin. In contrast, the activities of the D123N and E219Q mutants were weak. The activity of E215Q was significant but also very weak. To further examine the importance of Asp-123, Glu-215, and Glu-219, the residues were individually replaced with Ala. The resulting mutant E215A was active whereas the activity of the mutants D123A and E219A was almost at background levels. In addition, additional mutants for the unconserved Asp and Glu residues D208N, D225N, D254N, E128Q, E141Q, E177Q, E206Q, and E211Q were expressed and were all active (data not shown).
To investigate whether Asp-123 and Glu-219 are essential for catalytic activity, the mutant pro-enzymes D123N, D123E, D123A, E219Q, E219D, E219A, and D123E/E219D (double mutant) were prepared in a large scale, purified under denaturing conditions, allowed to refold at various pH values and activated at pH 3.5. However, the refolded mutants at each of the tested pH values had little activity (see Fig. 3 for some of the mutants). The activities of the mutants refolded at pH 5.25 are shown in Fig.  5B. The activities of the D123E, D123A, E219D, E219A, and D123E/E219D mutants were not detectable (Ͻ0.2%), and the activity of E219Q was very low (0.3%) whereas the activity of the D123N mutant was low but significant (5.2%). The activity of the D123N mutant was not significantly changed by incubation at pH 2.0 at 37°C for 2 h before assay.
Autoprocessing of the Mutant Pro-enzymes-After refolding of the purified mutant pro-enzymes D123N, D123A, E219Q, and E219A, they were incubated under acidic conditions and then analyzed on SDS-PAGE. In the case of the D123A, E219A, and E219Q mutants, no processing was observed (Fig. 6, lanes  2, 9, and 12) whereas D123N was processed to a form with a reduced molecular size (Fig. 6, lane 6) whose N terminus was identified as Ala-31 with protein sequencing. When D123A, D123N, E219A, and E219Q were incubated with a small amount (one-thirtieth by weight) of native aspergillopepsin II, the fully processed inactive enzyme was produced (Fig. 6, lanes  3, 7, 10, and 13). This mature inactive enzyme was resistant to further proteolysis.
CD Spectra of the Mutant Pro-enzymes-To confirm the correct folding of the recombinant mutant pro-enzymes, the CD spectra were measured. As shown in Fig. 7A, the spectra of the mutant pro-enzymes D123A and E219A were indistinguishable from that of the recombinant pro-enzyme. After incubation at pH 3.5, the spectrum of the recombinant pro-enzyme was changed into essentially the same spectrum as that of the native enzyme (Fig. 4B) whereas the spectra of the mutant pro-enzymes were not changed with incubation at pH 3.5 (Fig.  7B). However, after incubation with a small amount (one-thirtieth by weight) of native aspergillopepsin II at pH 3.5, the CD spectra were transformed similar to that of the mature enzyme (Fig. 7C).

DISCUSSION
Refolding of Aspergillopepsinogen II-Aspergillopepsin II is known to be unfolded above neutral pH mainly because of the electrostatic repulsion inside the molecule (31). The pro-peptide of aspergillopepsinogen II is rich in basic residues whereas the mature enzyme moiety is abundant in acidic residues. Therefore, as in the cases of ordinary aspartic proteinase zymogens such as pepsinogen (32) and aspergillopepsinogen I FIG. 5. Relative specific activities of recombinant wild-type and mutant aspergillopepsin II toward acid-denatured hemoglobin. After refolding, aliquots of the sample solutions were assayed. The activity of the wild-type enzyme is taken as 100%. A, crude denatured aspergillopepsin II was used for refolding and then assayed without any purification. B, purified denatured aspergillopepsinogen II was used for refolding. (27), the propeptide may interact electrostatically with the enzyme moiety. Such an interaction is thought to be responsible for correct folding of the polypeptide chain and control of the activation of the pro-enzyme. The optimum pH for refolding of aspergillopepsinogen II was around 5.3, and the refolding efficiency decreased at both higher and lower pH values (Fig. 3). The decrease in refolding efficiency at pH below 5.3 may be in part attributed to the weakening of the interaction between the propeptide and the enzyme moiety, which presumably decreases the refolding efficiency of the denatured pro-enzyme. Although the refolding efficiency was also decreased at higher pH, a low but significant proteolytic activity was observed after refolding at pH 6 -8.5. Thus, the pH profile of the refolding efficiency above pH 5.3 appears to be biphasic, and the reason for this is not clear at present.
Although the CD spectrum of the recombinant pro-enzyme at pH 5.25 after refolding at the same pH was different from that of the native enzyme at pH 3.5, acidification of the solution of the recombinant pro-enzyme to pH 3.5 resulted in an irreversible change in spectrum into one that was indistinguishable from that of the native enzyme (Fig. 4). These results indicate that the recombinant enzyme has the same folded structure as the native enzyme and that the change in spectrum can be attributed to activation and processing of the pro-enzyme.
Autocatalytic Processing of Aspergillopepsinogen II-The recombinant pro-enzyme was converted to the two-chain structure enzyme similar to the native enzyme under acidic conditions. Because the pro-enzyme, refolded at pH 3.7 or 7.3 ( Fig.  2A), as well as some mutant pro-enzymes showed no processing (Fig. 6), it is implausible that another proteinase from E. coli caused the processing of the wild-type pro-enzyme, indicating that the pro-enzyme itself has the proteolytic activity responsible for this processing.
The processing generated two new N termini, Glu-60 and Lys-108, on the light and heavy chains, respectively. The processing site between the pro-segment and the light chain, i.e. Asn-592Glu-60, is compatible with the cleavage specificity of aspergillopepsin II, which is known to cleave among others Asn-X bonds fairly preferentially (12,33). On the other hand, the major processing site between the intervening sequence and the heavy chain of the recombinant pro-enzyme was Asn-1072Lys-108. This site is two residues ahead of the major processing site, Arg-1092Gln-110, in the native pro-enzyme. Such a discrepancy in the processing site between the native and recombinant pro-enzymes was observed also in the activation of recombinant aspergillopepsinogen I (27). The cleavage site Asn-1072Lys-108 is compatible with the cleavage specificity of aspergillopepsin II, but the cleavage at Arg-1092Gln-110 is not. So far as examined with various protein and peptide substrates, no cleavage by the enzyme has been observed at Arg-X bonds. Therefore, the complete maturation to the native enzyme might require another peptidase, such as an aminopeptidase, to remove Lys-108 and Arg-109 from the N terminus of the heavy chain or a Kex2-like endopeptidase that is known to occur in A. niger (34) to cleave the Arg-1092Gln-110 bond.
Despite the difference in the N terminus of the heavy chain between the recombinant and native enzymes, the enzymatic properties appeared to be essentially identical with each other. Therefore, the present expression and refolding system of aspergillopepsinogen II is thought to be generally useful to study the structure/function relationships of the enzyme.
Catalytic Residues of Aspergillopepsin II-Based on the assumption that certain acidic residues should contribute to catalysis and that the catalytic residues should be conserved among the proteinases of family A4, we replaced each conserved acidic residue with the other amino acids. The enzymatic activity was assayed by examining the digestion of aciddenatured hemoglobin and the autoprocessing capability. Whether the mutants have the correctly folded structure was analyzed by comparing CD spectra and by examining resistance to proteolysis by native aspergillopepsin II. There are seven Asp and five Glu residues completely conserved among the proteinases of family A4 listed in Fig. 1. In preliminary experiments, the mutant pro-enzymes were prepared in a small scale, and their activities were assayed (Fig. 5A). The relative specific activity reflects not only the catalytic activity of the refolded protein but also the efficiency of refolding. Ten of the 12 conserved acidic residues were mutated without much loss of activity and therefore were ruled out as candidates for the catalytic residues. Glu-219 was shown to be the most likely candidate for the catalytic residue. In addition, Asp-123 was also suggested to be an important residue although the D123N mutant showed small activity. Further analyses were therefore carried out with Asp-123 and Glu-219 mutants.
The refolded D123A, E219A, and E219Q mutant pro-enzymes showed no processing at pH 3.5 (Fig. 6). When the refolded mutants were incubated with a small amount of native aspergillopepsin II at pH 3.5, all mutants were processed similar to the wild-type pro-enzyme, and the resulting mature mutant enzymes were resistant to further proteolysis. These results suggest that the mutants were correctly folded although they had no autoprocessing ability. In addition, the CD spectra indicated that the mutant pro-enzymes had the same secondary structure as the wild-type pro-enzyme (Fig. 7). On the other hand, the D123N mutant was converted at pH 3.5 to a slightly smaller molecule with Ala-31 at the N terminus. This  1, 4, 8, and 11, purified denatured D123A, D123N, E219A, and E219Q mutants, respectively, before refolding are shown. Lane 5, D123N mutant is shown after refolding by dilution and dialysis against the refolding buffer (50 mM sodium acetate, pH 5.25). Lanes 2, 6, 9, and 12, D123A, D123N, E219A, and E219Q mutants, respectively, after refolding followed by incubation at pH 3.5 and 37°C for 4 h are shown. Lanes 3, 7, 10, and 13, D123A, D123N, E219A and E219Q mutants, respectively, after refolding followed by incubation at pH 3.5 and 37°C for 4 h in the presence of a small amount of the native enzyme are shown. Lane M, molecular size markers. truncated form was also produced during the processing of the wild-type pro-enzyme as an intermediate, 4 but further processing of this intermediate from the D123N mutant was very slow. Thus, the D123A mutant was inactive whereas D123N showed a weak proteolytic activity. In this connection, it is interesting to note that the mutants D123E, E219D, and D123E/E219D had no activity, indicating that these two acidic residues can-not be mutually substituted for the enzyme to be active.
Taken together, Glu-219 is concluded to be the catalytic residue of the enzyme. If aspergillopepsin II has a catalytic mechanism similar to that of the pepsin-type aspartic proteinases, it would also have two acidic catalytic residues. In that case, Asp-123 would be the second catalytic residue. It seems to be noteworthy that the regions containing Asp-123 and Glu-219 are among the most conserved regions in the enzyme (Fig.  1). Because Cys-127 and Cys-210, which are proximal to Asp-4 X.-P. Huang, H. Inoue, and K. Takahashi, unpublished data. FIG. 7. CD spectra of recombinant aspergillopepsinogen II and aspergillopepsin II and mutants. A, recombinant aspergillopepsinogen II and its mutant proenzymes at pH 5.25. B, recombinant mutant pro-enzymes at pH 5.25 and pH 3.5. C, recombinant aspergillopepsin II and its mutants that were incubated with a small amount of native aspergillopepsin II at pH 3.5. Each spectrum of the mutants was corrected for the contribution of the added native enzyme. 123 and Glu-219, respectively, are linked by a disulfide bond (11), it is possible that Asp-123 and Glu-219 are closely located in space to each other to function cooperatively in catalysis. The corresponding two cysteine residues are also conserved in scytalidopepsin and the EapC protein and are the most conserved cysteine residues among the four homologous enzymes (Fig. 1). It is probable that, as in the present enzyme, the cysteines form a disulfide bond in homologous enzymes.
In the pepsin-type aspartic proteinases, the two catalytic aspartyl residues are proposed to work in concert, one (with a lower pK a ) as a deprotonated form and the other (with a higher pK a ) as a protonated form. In the case of porcine pepsin, which is most active at around pH 2 similar to the present enzyme, the pK a values of the two catalytic residues, Asp-32 and Asp-215, were estimated to be 1.50 and 3.95, respectively (35). Although the catalytic mechanism of the present enzyme is not known, it may be assumed by analogy with pepsin (36) that one of the catalytic residues (most likely Glu-219, with a similar lower pK a value caused by the specific microenvironment) deprotonates the nucleophilic water attacking the carbonyl carbon of the peptide bond and that the other catalytic residue (most likely Asp-123, with a similar higher pK a value) protonates the carbonyl oxygen of the peptide bond.
The D123N mutant exhibited weak but not negligible activity, and the pro-enzyme was partially processed under acidic conditions. The Asn residue at position 123 may function in part in the catalytic mechanism to substitute for Asp. The possibility cannot be completely excluded that a small amount of the wild-type enzyme might have been regenerated from the D123N mutant by unexpected hydrolysis of the amide group of Asn-123 during preparation of the mutant and/or during the assay at pH 2.0 toward hemoglobin. However, this possibility seems to be rather implausible for the following reason. As described in the preceding section, when the D123N mutant pro-enzyme was incubated with a small amount of native aspergillopepsin II, the fully processed inactive enzyme was formed (Fig. 6, lane 7). Therefore, if the Asn-123 residue of the D123N mutant pro-enzyme was partially deamidated, the resulting wild-type pro-enzyme should be autoprocessed to give a small amount of the wild-type enzyme, which would then convert the remaining mutant pro-enzyme to the fully processed inactive enzyme. This was actually not the case for the D123N mutant pro-enzyme; it was cleaved at the Glu-302Ala-31 bond, but further processing was very slow (Fig. 6, lane 6). Furthermore, when the D123N mutant was incubated at pH 2.0 and 37°C for 2 h before assay, no significant increase in activity was observed, indicating that Asn-123 was not hydrolyzed to Asp under such conditions. In this connection, it is interesting to refer to our recent finding that the mutants of aspergillopepsinogen I in which one of the catalytic Asp residues was replaced with an Asn had practically no activity. 5 Among the various families of proteinases, the zinc metalloendopeptidase family is known to possess a catalytic glutamic acid residue (37). To our knowledge, however, the catalytic glutamic acid residue has not been reported for other families of proteinases except for tetravirus endopeptidase. This enzyme, classified as an endopeptidase of family A21 by Barrett et al. (22), was reported to have a catalytic glutamic acid residue, which, however, is thought to function without turnover only once to intramolecularly cleave a specific Asn-X bond in the enzyme for autoactivation (38). Therefore, aspergillopepsin II appears to be rather unique as a more general endopeptidase with a catalytic glutamic acid residue because it is known to hydrolyze various peptide bonds in peptides and proteins (12,33). This may be called a glutamic proteinase or a glutamic-aspartic proteinase although it may be included in a wider sense in the family of aspartic proteinases because Asp-123 appears to be one of the catalytic residues. To obtain a more definitive conclusion on the catalytic residues and mechanism of aspergillopepsin II, further studies are necessary including precise kinetics using suitable peptide substrates and x-ray crystallographic analysis of the three-dimensional structure of the enzyme.