Absolute Side-chain Structure at Position 13 Is Required for the Inhibitory Activity of Bromein*

Bromelain isoinhibitor (bromein), a cysteine proteinase inhibitor from pineapple stem, has a unique double-chain structure. The bromein precursor protein includes three homologous inhibitor domains, each containing an interchain peptide between the light and heavy chains. The interchain peptide in the single-chain precursor is immediately processed by bromelain, a target proteinase. In the present study, to clarify the essential inhibitory site of bromein, we constructed 44 kinds of site-directed and deletion mutants and investigated the inhibitory activity of each toward bromelain. As a result, the complete chemical structure of Leu13 in the light chain was revealed to be essential for inhibition. Pro12 prior to the leucine residue was also involved in the inhibitory activity and would control the location of the leucine side chain by the fixed ϕ dihedral angle of proline. Furthermore, the five-residue length of the interchain peptide was strictly required for the inhibitory activity. On the other hand, no inhibitory activity against bromelain was observed by the substitution of proline for the N terminus residue Thr15 of the interchain peptide. In summary, these mutational analyses of bromein demonstrated that the appropriate position and conformation of Leu13 are absolutely crucial for bromelain inhibition.

Cysteine proteinases are involved in specific processing or more general degradation of proteins in a wide variety of organisms, including viruses, fungi, plants, and animals (1). Their activity is regulated by limited proteolysis of inactive precursors, by the pH of the surroundings (2), and by tight binding with proteinaceous inhibitors (3). With regard to proteinaceous inhibitors, six structurally different families of cysteine proteinase inhibitors have been reported so far: bromelain isoinhibitors (bromein) 2 (4), cystatins (5), soybean trypsin inhibitor-like inhibitors (6), thyropins (7), inhibitors homologous to the propeptide regions of cysteine proteinases (8), and clitocypins (9).
Bromelain is known as cysteine proteinases in the stem and fruit of Ananas comosas, while the inhibitors exist only in the stem and have been classified into eight isoforms based on their amino acid sequences (10). The major component of bromelains, stem bromelain, has been sequenced (11) and shown to be a member of the papain superfamily (12). On the other hand, the presence of inhibitory fractions has also been confirmed in pineapple stem (13), and the amino acid sequence of the seventh bromein (bromein-7) was the first sequence determined among the inhibitory fractions (4). Hatano et al. (14) revealed the complete primary structures of all eight bromein isoforms: each isoform is composed of a light chain (10 -11 residues) and a heavy chain (40 -41 residues), which are cross-linked by five disulfide bridges.
The three-dimensional solution structure of the sixth bromein with the two chains (bromein-6 N ) is characterized by inhibitory and stabilizing domains, each of which is formed by a three-stranded antiparallel ␤-sheet (15,16). As shown in Fig. 1, A and B, the inhibitory domain consists of the light chain and two parts of the heavy chain (Glu 20 -Cys 26 and Asp 51 -Lys 60 ). This domain is thought to be the major bromelain inhibitory site, and it has a relatively flexible structure that appears to allow itself to fit well into the active site cleft of the target proteinase (17). On the other hand, the structure of the stabilizing domain (Thr 29 -Ile 48 ) is thought to contribute mainly to the conformational stability of the inhibitory one, because the NMR structures calculated were well converged (17). Surprisingly, bromein-6 N shares the same fold and disulfide bridge connectivity as the Bowman-Birk serine proteinase inhibitor (BBI) (17). For instance, BBI from soybeans is a 71-residue inhibitor that has two independent inhibitory sites for the serine proteinases trypsin and chymotrypsin (18). It is noteworthy that bromein-6 N exhibits relatively weak inhibitory activity against these serine proteinases (19).
The genomic DNA of a bromein precursor protein (27.5 kDa) was found to encode three homologous isoinhibitor domains, each of which contains an interchain peptide (five residues) between the light chain and heavy chain, two interdomain peptides (19 residues each), and a C-terminal pro-peptide (18 residues) (20). The precursor protein would be converted into mature isoinhibitors (6 kDa) by proteolytic processing. Moreover, we constructed a recombinant single-chain sixth bromein with the interchain peptide (bromein-6 R , Fig. 1B) and revealed that it shows almost the same inhibitory activity and secondary structure as bromein-6 N (20). Interestingly, bromein-6 R is processed between the light chain and interchain peptide by stem bromelain, the target proteinase (19). To date, the essential inhibitory site of bromein has been examined using several different recombinant inhibitors. For example, another single-chain inhibitor without the interchain peptide (bromein-6 r ) shows no inhibition against bromelain (20). It is worth noting that bromelain-digested bromein-6 R (bromein-6 RP ) exhibited much more bromelain inhibitory activity than bromein-6 R (19). However, the essential inhibitory site has not been precisely specified so far. In the present report, to identify the inhibitory site of bromein, we prepared 44 kinds of sitedirected and deletion mutants and investigated the inhibitory activity of each toward bromelain. For the mutants with low inhibitory activity, the secondary structures were examined using their circular dichroism (CD) spectra.

Expression and Purification of Recombinant Bromein-6 R and
Its Mutants-His-tagged bromein-6 R was expressed in Escherichia coli and was purified by Ni-NTA agarose and MonoQ chromatography as described previously (20). The expression vectors for the mutants were constructed from a pET32Ј-bromein-6 R plasmid by using the QuikChange mutagenesis kit (Stratagene, La Jolla, CA). The mutant proteins were expressed in E. coli and were purified according to the same method used for bromein-6 R . The identification and purity of the samples were confirmed by both matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry (MS) and Tricine SDS-PAGE (16.5% gels) without dithiothreitol using a Tris-Tricine buffer system (21).
Assay of Inhibitory Activity toward Bromelain-The inhibitory activity toward stem bromelain was measured as described previously (19). Enzyme (final concentration of 320 nM), inhibitor (584 nM), and the substrate p-nitrophenyl benzyloxycarbonyl-L-lysinate (20 M) were incubated together for 3 min in 50 mM sodium acetate buffer (pH 4.6) containing 100 mM KCl and 1 mM dithiothreitol at 25°C. The initial velocity (v 0 ) was calculated from the increase in absorbance at 317 nm of the released p-nitrophenol. The percent inhibition of stem bromelain was estimated by Equation 1.
The experiments were performed in triplicate, and the results were expressed as the mean values Ϯ S.D. The protein concentrations of the bromein-6 R derivatives were usually estimated from the absorbance at 275 nm by using an extinction coefficient of 6200 M Ϫ1 ⅐ cm Ϫ1 . For the C6A/C26A, C11A/C24A, L13W, L13Y, and bromein-6 BBI-2 mutants, the absorbance at 275 nm was also used to calculate the protein concentrations by using the extinction coefficients of 6080, 6080, 11410, 7600, and 4920 M Ϫ1 ⅐ cm Ϫ1 , respectively. The protein concentration of stem bromelain was estimated using E 280 1% values of 20.1 (22). CD Measurements-CD spectra were obtained at room temperature on a Jasco J-700 spectropolarimeter (Jasco, Tokyo). The sample was dissolved at a concentration of 15 M in 5 mM sodium-acetate buffer and was adjusted to pH 4.6. The path length of a quartz cell was 0.1 cm. Sixteen scans were accumulated at a bandwidth of 1.0 nm and a speed of 100 nm/min with 0.2-nm increments. The spectra thus obtained were submitted to a noise-reduction procedure and expressed in terms of the mean residue ellipticity [] in deg⅐cm 2 ⅐dmol Ϫ1 .
NMR Measurements-The intact BI-VI R and the mutants L13F, L13V, C6A/C26A, and C11A/C24A were dissolved at 130 -910 M in 90% H 2 O/10% D 2 O (pH 3.9). The spectra were recorded using 32768 data points with a spectral width of 9000 Hz on a Varian Unity INOVA 600 spectrometer at 30°C. The number of scans for each sample was set to 512, and the chemical shift values were referenced to the external sodium 2,2-dimethyl-2-silapentane-5-sulfonate. The water signal was suppressed by irradiation during the relaxation delay, and the chemical shifts of individual protons of the spectra were assigned by referring to the previous result (3).

RESULTS AND DISCUSSION
Inhibitory Activity of the Bromein-6 R Derivatives and Their CD Spectra- Fig. 1B shows the amino acid sequences of the recombinant forms that contain more than one mutation. The mutant constructs, in addition to bromein-6 R , were expressed in the soluble fraction of the cell lysate and were purified by two steps of chromatography to give a single band on Tricine SDS-PAGE (data not shown). Furthermore, the purity of the mutants was confirmed to be more than 95% by MALDI-TOF MS analysis (data not shown). As shown in Fig. 2, almost complete loss of the inhibitory activity was observed in the Pro 12and Leu 13 -substituted mutants as well as in the mutants without the interchain peptide. It is remarkable that the mutant T15P showed no inhibitory activity against bromelain. In the following sections, we will discuss the inhibitory activity of these mutants in greater detail.
Here, we examined the secondary structures of the mutants with low inhibitory activity by using their CD spectra. As a result, the spectra of the mutants V10A, P12A, L13A, L13I, R14D, and R14I were roughly identical to the spectrum of bromein-6 R (Fig. 3). This indicates that the secondary structures of these mutants and bromein-6 R are similar to the secondary structure of bromein-6 N , because bromein-6 R has almost the same secondary structure as bromein-6 N (20). Interestingly, the CD spectrum of the non-inhibitory mutant T15P was almost identical to that of bromein-6 R (Fig. 3). On the other hand, the spectrum of the non-inhibitory mutant L13V was different from that of bromein-6 R (Fig. 3), indicating that the L13V mutant did not retain the same secondary structure as bromein-6 R .
Site-directed Mutagenesis of the Light Chain-In a previous study (14), we constructed a binding model between bromein and papain by computer modeling. According to this model, the C-terminal region of the light chain, especially Arg 14 , protrudes into the solvent from one edge of the inhibitor and fits well into the active site cleft region of the target proteinase. Therefore, we focused the mutational analyses on the C-terminal region of the light chain (Fig. 2). As a result, the inhibitory activities of the mutants S7A, E8A, and V10A were only slightly decreased as compared with that of bromein-6 R . On the other hand, the activity of the P12A and P12R mutants fell to less than 20 and 3%, respectively. This indicates that the Pro 12 residue, the angle of which is fixed, plays an important role in restricting the backbone movement in the C-terminal region of the light chain and the conformation of Leu 13 .
With the exception of the L13Y mutant, the inhibitory activity for each of the Leu 13 -substituted mutants fell to less than about 5% (Fig. 2). Interestingly, even the substitution of leucine to isoleucine or valine, whose chemical structure is similar to that of leucine, resulted in about 95% inhibitory activity loss. The secondary structure of these mutants, with the exception of L13V, was almost identical to that of Bromein-6 R as judged by CD analysis (Fig. 3). It seems likely that the secondary structure of the L13V mutant is disrupted (Fig. 3), and thus may be misfolded. However, the NMR analysis indicates that the whole structures of the mutants L13F and L13V were almost the same as the structure of bromein-6 R (Fig.  4). Taken together, these findings demonstrated that bromelain strictly recognizes the chemical structure of Leu 13 in bromein. Furthermore, the fixed dihedral angle of Pro 12 was also found to be important for keeping the proper side-chain conformation of Leu 13 for bromelain inhibition.
Previously, we proposed that the Arg 14 residue is a putative reactive site for bromelain (14,24), since leupeptin, another cysteine proteinase inhibitor, has the chemical structure of N-acetyl-Leu-Leu-argininal and exhibits trypsin inhibitory activity (23). Furthermore, the peptide bond of Arg-X is reported as a preferable cleavage site for bromelain (22). However, the present study revealed that the Arg 14 -substituted mutants, especially the mutant R14E, did not lose a substantial amount of inhibitory activity (Fig. 2). This indicates that the positive charge at this position is not important for bromelain B, sequence alignments of the recombinant forms that are not single-point mutants. The recombinant single-chain inhibitor, native inhibitor, bromein-6 R , and bromein-6 N are also shown for comparison. The amino acid sequences are shown using single-letter notation for amino acids. The numbering system corresponds to that of bromain-6 R , while the numbering of BBI corresponds to the original sequence. The BBI sequence is aligned as proposed previously (17), and the N and C termini are indicated by N and C, respectively. Pink and cyan boxed sequences represent the light and heavy chains, respectively. The symbol Ϫ indicates a deletion inserted to optimize the homology. In the sequence of BBI, the asterisks at positions 13 and 44 indicate reactive sites for trypsin and chymotrypsin, respectively (32). The symbol ϩ is indicative of residues whose NH resonances are shifted abnormally to the lower magnetic field (3). inhibition. On the other hand, the bromelain inhibitory activity of the R14D and R14I mutants were 40 and 28%, respectively. The results suggested that the different sizes of the side chain of aspartic acid or isoleucine might influence the side-chain conformation of Leu 13 , thereby affecting the inhibitory activity.
Site-directed Mutagenesis of the Interchain Peptide-We previously revealed that stem bromelain performs stepwise processing of the interchain peptide, and that the 50% inhibitory (IC 50 ) value of bromein-6 R was ϳ10-fold higher than that of bromein-6 RP (19). Accordingly, the full inhibitory activity might require cleavage between the light chain and interchain peptide or the removal of some residue(s) in the interchain peptide. In this study, we prepared several recombinants, the mutants T15A, T15D, T15P, S16A, S17A, S18A, D19A, and D19S, which are mutated on the region of the interchain peptide. These mutants except for the T15P mutant did not show any activity loss (Fig. 2), indicating that the removal of some residue(s) in the interchain peptide is not important for bromelain inhibition.
Here, a question arises as to whether or not the cleavage at Arg 14 -Thr 15 is essential for inhibitory activation. We examined the bromelain cleavage site of the non-inhibitory mutants L13I and T15P by SDS-PAGE and N-terminal sequencing analyses. The results showed that bromelain hydrolyzed easily at Arg 14 -Thr 15 of L13I within 2 h of incubation, while Arg 14 -Pro 15 of T15P was not cleaved by bromelain even after 20 h of incubation (data not shown). We propose that the fixed dihedral angle of Pro 15 likely restricts the conformation of the putative cleavage site of T15P. Considering that bromein-6 R was processed at Arg 14 -Thr 15 within 4 h (19), we concluded that this cleavage is not essential for the activation of inhibition and the free C-terminal motility in the light chain region would be required for full inhibitory activity.
Deletion Analysis of the Interchain Peptide-In the previous section, we confirmed that the amino acid sequence of the interchain peptide is not particularly important for the inhibitory activity. Here, we investigated whether or not the length of the interchain peptide can affect the activity. We constructed deletion mutants with various lengths of the interchain peptide and examined the activity of each mutant. The deletion mutant lacking one residue in the interchain peptide, the dD 19 mutant, showed ϳ30% loss of activity, and the mutants lacking more than two residues, the dSD 19 and dSSSD 19 mutants and bromein-6 r , lost almost all inhibitory activity (Fig. 2). This indicates that at least four residues of the interchain peptide were necessary for the inhibitory activity. The five-residue length in the interchain peptide thus would maintain the proper conformation of the Leu 13 side chain for a good fit into the catalytic site of the target proteinase. Considering that the doublechain inhibitor bromein-6 N showed higher inhibitory activity than the single-chain inhibitor bromein-6 R , the free movement of the Leu 13 side chain appears to be very important for inhibition.
Site-directed Mutagenesis of the Heavy Chain-We next performed site-directed mutagenesis for the charged residues and terminal regions in the heavy chain. Only the mutant D28A showed a more than 30% loss of inhibitory activity by this analysis (Fig. 2), indicating that the other residues in the heavy chain are not directly involved with proteinase inhibition. In previous studies (3,25), the carboxyl groups of the side chains of Asp 28 and Asp 51 were revealed to form hydrogen bonds with the FIGURE 3. Far-ultraviolet CD spectra of the bromein-6 R mutants with considerably lower inhibitory activity than bromein-6 R . The CD spectrum of bromein-6 R is drawn as a solid line for comparison. FIGURE 4. One-dimensional NMR spectra of the bromein-6 R mutants. The NH chemical shifts (pH 3.9, 40°C) of Ser 7 , Cys 26 , Lys 38 , and Cys 49 in bromein-6 N were reported to be 10.75, 10.13, 11.13, and 9.86 ppm, respectively (15). Cys 26 and Cys 49 are located on the three-stranded antiparallel ␤-sheet of the inhibitory and stabilizing domains in bromein-6 N , respectively. The insets are indicative of the magnified spectra of the downfield regions. similar to that of papain, although the structure of bromelain so far has not been solved. On the other hand, the crystal structure of a papain-leupeptin complex has been determined by x-ray crystallography (31); therefore, we can discuss the inhibitory mechanism of bromein-6 N using papain instead of bromelain, since papain is known to be inhibited by bromein-6 N (3).
In the three-dimensional structure of the papain-leupeptin complex, the Leu 1 side chain in leupeptin is involved in hydrophobic interactions with the side chains of Tyr 61 and Tyr 67 in papain (31). The Leu 2 side chain in leupeptin firmly binds the hydrophobic pocket enclosed by the side chains of Tyr 67 , Pro 68 , Val 133 , and Val 157 in papain. On the other hand, the guanidino group of leupeptin does not interact with any residue in the complex. Now, we can discuss a putative papain-bromein complex based on these interactions in the papain-leupeptin complex. In this model, the Leu 13 side chain in bromein-6 N would bind to the hydrophobic pocket of papain and should be involved in hydrophobic interactions with the side chains of Val 132 and Leu 156 in bromelain. The Pro 12 residue in bromein-6 N probably interacts with Tyr 61 and Trp 67 in bromelain. In the solution structure of bromein-6 N (16), the C-terminal region in the light chain protrudes into the solvent from one edge of the inhibitor molecule and thus could be placed at a favorable location to serve as the inhibitory site.
In conclusion, this study demonstrated that the mutation of Leu 13 in bromein has the largest effect on bromelain inhibition. This indicates that the chemical structure of the leucine residue is strictly required for inhibition, and that the Leu 13 side chain binds stereospecifically to the active site cleft in the target proteinase. In addition, a similar sensitivity to mutation at Pro 12 was noted with the inhibition of bromelain. In consideration of the leupeptin-papain complex, the hydrophobic interaction between the hydrophobic pocket of the proteinase and Pro 12 and Leu 13 of bromein appears to be the most important factor in bromelain inhibition.