M)tO-711X/92 $5.00f0.00 Copyright 0 1992 Pergamon Press Ltd
inr. .I Biochem.Vol.24.No.9,pp.1487-1491, 1992 Printed in Great Britain. All rights reserved
PROTEOLYTIC PROCESSING SITES PRODUCING THE MATURE FORM OF HUMAN CATHEPSIN D TAKAHIKO KOBAYASHL’ KOICHI HONKE, I* SHINSEI GASA,' TETSUYA FUJII,’ SHIRO MAGUCHI,’ TAMOTSUMIYAZAKI’ and AKIRA MAKITA’
LBiochemistry Laboratory, Cancer Institute and 2The Third Department of Internal Medicine. Hokkaido University School of Medicine, Sapporo 060, Japan (Received 1I December 1991) Abstract-l. The proteolytic processing sites of human lysosomal aspartic protease cathepsin D at which the intermediate single-chain form was converted into the mature two-chain form were determined. 2. The two chains were isolated by reversed-phase HPLC in order to investigate the cleavage sites of the enzyme. 3. Protein sequencing of the heavy chain, which was presumed to be derived from the C-terminal side in the single-chain enzyme, gave an N-terminal Leu 105. In addition, it revealed that there were also minor sequences, which commenced with Gly 106 and Gly 107. 4. A small C-terminal peptide was isolated from the light chain, which had been digested with two kinds of exogenous proteases. Sequence deter~nation of this peptide, which was characterized as a nona~ptide by mass s~ctrometry, suggested that the C-terminus of the light chain was Ser 98. 5. These results indicate that a Ser 9%Ala 99 bond and an Ala 104-Leu IO5 bond are cleaved to release 6 amino acid residues between the two chains.
Cathepsin D (EC 3.4.23.5) is a lysosomal aspartic protease that is present in all mammalian cells, and its
biosynthesis and protein structure have been studied (for review, Barret, 1977; Tang and Wong, 1987; Erickson, 1989). Recently, the primary structure of porcine cathepsin D was determined by protein sequencing (Shewale and Tang, 1984), and those of human (Faust et al., 1985) and rat (Fujita et al., 1991) cathepsin D were deduced from the cDNA sequence. Human cathepsin D, like other lysosomal hydrolases, is synthesized as a preproenzyme that undergoes subsequent proteolytic processing to produce the mature enzyme (Hasilik and Neufeld, 1980; Erickson et a!., 198 1). The first cleavage is co-translational and removes the signal peptide to yield procathepsin D. The proenzyme is processed to remove 44 amino acid residues from the N-terminal, producing an active single-chain enzyme with a relative molecular mass of 44,000. The single-chain enzyme is eventually cleaved into a M, 15,000 light chain and a IV, 30,000 heavy chain. The sequences of the last processing region were analyzed in porcine and bovine cathepsin D (Yonezawa et al., 1988) and rat cathepsin D (Fujita et al., 1991). When human cathepsin D was aligned by sequence homology against porcine cathepsin D, it was suggested that 7 amino
acid residues were removed in the processing region during the conversion of the single-chain form to the *To whom correspondence
should be addressed.
two-chain form of human cathepsin D (Faust et al., 1985). However, the actual processing sites where this occurred remained undefined. The present study was undertaken to determine the processing sites of human cathepsin D that produce the mature enzyme.
MATERIALS AND METHODS Materiais Achromobacter Protease I (Lysyl Endopeptidase) was purchased from Wako Chemicals, and endoproteinase Asp-N from Boehringer Mannheim Yamanouchi. Other chemicals were of analytical grade. Cathepsin D was purified from human liver as described previously (Magu~hi et al., 1988). Normal human liver tissues were obtained at autopsy, and stored immediately at -80°C until use. Enzymatic digestion Digestion of the light chain (80 nmol) with Achromobacter Protease I with a SO:1 (moi/mol) ratio of substrate to enzyme, was carried out in 50 mM Tris-HCI, pH 9.0, for 20 hr at 30°C. The peptide termed Ly4 was treatd with endoproteinase Asp-N with a 100: 1 (w/w) ratio of substrate to enzyme, in 50 mM sodium phosphate buffer, pH 8.0, with 1 M urea for 18 hr at 37°C. Reversed-phase HPLC Nine milligrams of purified cathepsin D was reduced and S-pyridylethylated (Friedman et al., 1970; Hermodson et al., 1973). The treated enzyme was applied to a reversedphase column @BONDERSPHERE, 5 p, C8-300 A, 0.39 x 15 cm, Waters), and eluted with a linear gradient of from 0 to 80% acetonitrile in 0.05% trifluoroacetic acid. The Aeh~omobucfe~ Pro&ease I hydrolysate of the light chain
1487
TAKAHIKO KOBAYASHI et al.
1488
was chromatographed on the same column under the same conditions. Endoproteinase Asp-N hydrolysate of the C-terminal peptide of the light chain was subjected to another reversed-phase column (p BONDERSPHERE, 5 p. C18100 A, 0.39 x 15cm, Waters), and eluted with the solvent system described above. Determination sequencing
:
Mrxlo-s
2 -42.7 e
-31.0 -21.6
6
m-l
Seven nanomoles of each separated Achromobacter Protease I digest was hydrolyzed under vacuum in a 6 M HCI at 110°C for 24 hr and studied with an amino acid analyzer (Hitachi 835). The N-terminal sequence analysis was performed by automated Edman degradation using a protein sequencer (Applied Biosystem 477A).
14.4
(FAB-MS)
FAB-MS in a positive mode was performed with a JMS-HXllO mass spectrometer (JEOL) equipped with a JMA-DA5000 data system (JEOL). Approximately I nmol of a peptide termed A3 was bombarded by Xe gas at 3 kV in a matrix of glycerol, and the fragments were accelerated at 8 kV. Polyacryamide gel elecrrophoresis Polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate and p-mercaptoethanol (SDS PAGE) was performed according to the method of Laemmli (1970). The gel was stained for protein with silver stain. RESULTS
Separation of human cathepsin D subunits
When purified mature cathepsin D was reduced, S-pyridylethylated, and subjected to reversed-phase HPLC, two major peptides, L and H, were separated (Fig. 1). The isolated peptides L and H were homogeneous on SDS-PAGE (Fig. 1, inset), corresponding to the molecular weights of the light chain and the heavy chain of liver cathepsin D, respectively (Barret, 1977; Maguchi et al., 1988). N-Terminal sequence of the heavy chain
Comparing the amino acid homology of human cathepsin D to that from other animal sources (Yonezawa et al., 1988), the heavy chain was presumed to be derived from the C-terminal portion of the single-chain form. When the N-terminal sequence of peptide H was analyzed, it was found that a major sequence (_ 55%) coincided with the peptides begin-
0
Time (min
)
Fig. I. Separation of human cathepsin D subunits. Purified cathepsin D was reduced, S-pyridylethylated and subjected to reversed-phase HPLC using a 0.39 x 15 cm Waters FBONDERSPHERE, 5 p, C8-300 A column. The peptides were eluted with a linear gradient of from 0 to 80% acetonitrile in 0.05% trifluoroacetic acid. Inset: SDS-PAGE of the isolated peptides L and H. A part (approx. 1 pg each) of the peaks L and H on the HPLC was subjected to SDS-PAGE under a reduced condition and stained with silver. Molecular weights of standard proteins are indicated on the right side.
ning with Leu 105 of human cathepsin D (Table 1). In addition, peptide H included other minor sequences which commenced with Gly 106 (~30%) and Gly 107 (- 10%). In the first cycle of the sequence analysis, a small amount of alanine was also detected in addition to leucine and glycine. However, it could not be determined whether another peptide starting with Ala 104 existed or not, since we should consider the influence of the carryover after the first cycle, which originated from the major peptide commencing with Leu 105. This N-terminal heterogeneity of the heavy chain was reproducible in three different heavy chain preparations. Based on these observations, it was concluded that the heavy chain was a mixture of polypeptides beginning with Leu 105, Gly 106 and Gly 107.
Table I. Seauence determination Data source
cDNA Peptide A3 Peptide H
H
‘I
- 66.2
of amino acid composition and protein
Fast atom bombardment-mass speciromeiry
I-
of the arocessinn reaion
Amino acid sequence 90
98
105
I
I
I
SQDTVSVPCQSASSASALGGVKVERQVFGE DTVSVPCQS LGGVKVERQV GGVKVEROVF GVKVERQVFG
(-55%) (- 30%) i- io%j
The deduced amino acid sequences from cDNA of human cathepsin D (Faust et al., 1985) and the determined ones are shown in the one letter code. Amino acid No. 1 is assigned to the first residue of the mature enzyme protein.
Human cathepsin D
1489
Table 2. Amino acid composition of peptide Ly4
Sequence of the C-terminal region of the light chain
Peptide Ly4
In order to determine the C-terminal of the light chain, peptide L was digested with Achromobacter Protease I and subjected to reversed-phase HPLC. Six peptides were separated (Fig. 2), and the amino acid composition of each peptide was analyzed. The amino acid composition of the isolated peptide termed Ly4 was consistent with that of the C-terminal region (Asn 70-Ser 98) deduced from cDNA (Table 2). To confirm the C-terminal sequence of peptide Ly4, the peptide was further digested with endoproteinase Asp-N. Although the hydrolysate of peptide Ly4 was expected to separate into three peptides, it was separated into four peaks by reversed-phase HPLC (Fig. 3). Through protein sequencing of each peptide, it was revealed that the peptide called A3 was coincident with a nonapeptide from Asp 90 to Ser 98 of human cathepsin D (Table I). Peptide Al was identical to peptide A2 with respect to the amino acid sequence, which coincided with a peptide from Asn 70 to Phe 74, though Asn 70 was not detected because of the addition of a carbohydrate chain (data not shown). Asn 70 is the only potential N-glycosylation site in the light chain. The light chain of human liver cathepsin D contained endo-fi-N-acetylglucosaminidase H sensitive and insensitive carbohydrate chains (Ohhira et al., 1991).
ASP Thr ser GilI GUY Ala Val Met Ile
Leu TYr Phe LYS His PEb-Cys Arg Pro
Am 7fKkr
3.3 (3) 2.0 (2) 6.6 (7) 2.3 (2) 4.6 (5) 0.1 (0) 2.2 (2) 0.0 (0) 1.0 (1) 2.1 (2) 2.0 (2) 1.1 (1) 0.1 (0) I.0 (1) 0.4 (0) 0.0 (0) 1.3 (1)
98*
3 2 7 2 4 0 2 0
1 2 2 1 0
I 1 0
I
Data
in the table are expressed in molar ratios relative to the vaIues for isoleucine as indicated. The figures in parentheses show the nearest integers, ‘Peptide from Asn 70 to Ser 98, deduced from human cathepsin D cDNA. bPE, S-pyridylethylated
Therefore, abIy due structure. Asp 75 to
peptides Al and A2 were separated probto the difference of their oligosa~haride Peptide A4 coincided with a peptide from Gln 89. Peptide A3 was also characterized
A4
R
a LYE
A3
A2 LY5 LY4 I
P 1
all.LY3
LYf
-II-
LY2
-e--
Fig. 2. Separation of Achromobacrer Protease I hydrolysate of the light chain. Peptide L in Fig. 1 was digested with Achromobacfer Protease I (peptide : enzyme ratio 50 : 1 (mol/mol) in 50 mM Tris-HCI, pH 9.0 for 20 hr at 30°C) and subjected to reversed-phase HPLC using the same column and the same conditions as described in Fig. 1.
C
Fig. 3. Separation of endoproteinase Asp-N hydrolysate of the C-terminal peptide Ly4. Peptide Ly4 (Fig. 2) was digested with endoproteinase Asp-N (peptide:enzyme ratio 100: 1 (w/w) in 50 mM sodium phosphate buffer, pH 8.0 containing 1 M urea for 18 hr at 37°C) and subjected to reversed-phase HPLC using a 0.39 x 15cm Waters FBONDERSPHERE, 5~ C18-lOO A column under the same condition as described in Fig. 1.
1490
TAKAHIKO
50
KOBAYASHI
et al.
115 207
x ,’ D 5 0 0 aI .?
0 *100 100
200
300
400
500
700
600
--
6 f
(M+H)+
50
1040
I
O-
800
900
I
1000
1100
1200
1300
1400
1500
m/z
Fig. 4. Positive fast atom bombardment-mass spectrum of peptide A3. Approximately 1 nmol of a peptide A3 (Fig. 3) was bombarded by Xe gas at 3 kV in a matrix of glycerol, and the fragments were accelerated at 8 kV. A major peak at m/z 1040 corresponds to the molecular ion (M + H)+ of the S-pyridylethylated nonapeptide, Asp-Thr-Val-Ser-Val-Pro-PE-Cys-Gln-Ser, where PE means S-pyridylethylated.
by positive
fast analysis
atom
bombardment-mass
spec-
(Fig. 4). One major peak at m/z 1040 corresponded to the molecular ion (M + H)+ of the S-pyridylethylated nonapeptide. Therefore, we concluded that the C-terminus of the light chain was Ser 98. These results indicated that a Ser 98-Ala 99 bond and an Ala 104Leu 105 bond were cleaved to release six amino acid residues between the two chains (Table 1). trometry
DISCUSSION
We determined the proteolytic processing sites of human cathepsin D where the single-chain form was converted into the two-chain form. The splitting site of human cathepsin D was close to those of bovine, porcine and rat cathepsin Ds (Shewale and Tang, 1984; Yonezawa et aI., 1988; Fujita et al., 1991) and the processing regions was located in a p-hairpin loop (Yonezawa et al., 1988). Comparing human cathepsin D with the enzyme from the other species, however, it was found that the amino acid sequences within the processing region and the proteolytic processing sites differ considerably. The susceptibility to the proteolytic processing varies with the species. Human and porcine cathepsin D are mostly composed of the two-chain form (Hasilik and Neufeld, 1980; Huang et al., 1979) the rat enzyme contains mostly the single-chain form and a small amount of the two-chain form (Yonezawa et al., 1988; Fujita et al., 1991), whereas the bovine enzyme contains almost equivalent amounts of each of the two forms (Huang et al., 1980). There are neither basic amino acid residues, which are targetted by the so-called processing proteases (Loh et al., 1984; Schwartz, 1986; Bond and Butler, 1987) nor an asparaginyl-X
bond, which is involved in the proteolytic processing of many lysosomal hydrolases (Takio et al., 1983; Mahuran et al., 1988; Shewale and Tang, 1984), in the processing region of human cathepsin D. It was predicted the cleavage would occur directly after Gln 97 in the human single chain sequence, based on the amino acid sequence homology between human and porcine cathepsin Ds (Faust et al., 1985). Since asparagine and glutamine have an amide side chain, the possibility that the amide side chain plays an important role in site recognition by the processing protease has been presented (Erickson, 1989). However, the cleavage was found to occur not after Gln 97 but after Ser 98 in this study. These observations indicate that the proteolytic processing of the single-chain enzyme of cathepsin D from various species is not carried out by a common specific processing protease which recognizes the primary structure. Alternatively, it was suggested by experiments using inhibitors of cysteine proteases (Hentze et al., 1984; Gieselmann et al., 1985) that lysosomal cystein proteases such as cathepsin B, H or L, which have a broad peptide bond specificity, act on the proteolytic processing, which converts the intermediate single-chain form of cathepsin D into the mature two-chain form. It is not clear whether these cathepsins recognize the unique conformation of the processing region. As in the heavy chain of human cathepsin D, heterogeneity at the N-terminus has been shown in the subunits of human /I-hexosaminidase (Mahuran et al., 1988). Therefore, N-terminal trimming may be found in many lysosomal hydrolases. Although the possibility that the heavy chain suffered from non-physiological proteolysis during the purification of the enzyme is not excluded, it is likely that the
1491
Human cathepsin D N-terminal heterogeneity results from the physiological processing by cystein proteases, which display aminopeptidase activity (Barrett and Kirschke, 1981), or an unknown protease within the lysosomes. Quite recently, during preparation of this manuscript, Horst et al. (1991) also reported the N-terminal heterogeneity of the heavy chain of cathepsin Ds which were isolated from BHK cells transfected with human cathepsin D cDNA and from human promonocytes U937. However, they did not examine the C-termini of the light chains. The physiological significance of the cleavage which forms the two-chain enzyme is unclear. It is difficult to distinguish clearly whether the proteolytic cleavage occurs during a maturation process or during a catabolic process. Considering that the single-chain and two-chain forms appear to be equally active (Huang et al., 1979), the cleavage may be the first step of degradation, although the nick might bring about stability in the conformation. We are now studying the proteolytic processing regions of other lysosomal enzymes. The accumulation of such data should give us a clue as to the physiological significance of proteolytic processing, and help to identify the processing proteases.
structure of three forms of mature enzymes. Biochem. biophys.
REFERENCES
Barret A. J. (1977) Cathepsin D and other carboxyl proteinases. In Proteinases in Mammalian Cells and Tissues (Edited by Barret A. J.), pp. 209-248. North-Holland, Amsterdam. Barret A. J. and Kirschke H. (1981) Cathepsin B, cathepsin H, and cathepsin L. Meth. Enzym. 80, 535-561. Bond J. S. and Butler P. E. (1987) Intracellular proteases. A. Rev. Biochem.
56, 333-364.
Biophys.
230, 375-382.
12, 31463153.
Horst M. and Hasilik A. (1991) Expression and maturation of human cathepsin D in baby-hamster kidney cells. Biochem.
J. 273, 355-361.
Huang J. S., Huang S. S. and Tang J. (1979) Cathepsin D isozymes from porcine spleens. Large scale purification and polypeptide chain arrangements. J. biol. Chem. 254, 11,405-11,417. Huang J. S., Huang S. S. and Tang J. (1980) Enzyme Regulation and Mechanism of Action (Edited by Mildner P. and Ries B.), pp. 289-306. Pergamon Press, Oxford. Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Lond. 227, 680-685.
Loh Y. P., Brownstein M. J. and Gainer H. (1984) Proteolysis in neuropeptide processing and other neural functions. A. Rev. Neurosci. 7, 189-222. Maguchi S., Taniguchi N. and Makita A. (1988) Elevated activity and increased mannose-6-phosphate in the carbohydrate moiety of cathepsin D from human hepatoma. Cancer Res. 48, 362-367. Mahuran D. H., Neote K., Klavins M. H., Leung A. and Gravel R. A. (1988) Proteolytic processing of pro-a and pro-/J precursors from human p-hexosaminidase. J. biol. Chem. 263, 46124618. Ohhira M., Gasa S., Makita A., Sekiya C. and Namiki M. (1991) Elevated carbohydrate phosphotransferase activity in human hepatoma and phosphorylation of cathepsin D. Br. J. Cancer 63, 905-908.
Erickson A. H. (1989) Biosynthesis of lysosomal endopepti dases. J. cell. Biochem. 40, 3141. Erickson A. H., Conner G. E. and Blobel G. (1981) Biosynthesis of a lysosomal enzyme. Partial structure of two transient and functionally distinct NH,-terminal sequences in cathepsin D. J. biol. Chem. 256, 11,22411,231. Faust P. L., Kornfeld S. and Chirgwin J. M. (1985) Cloning and sequence analysis of cDNA for human cathepsin D. Proc. natn. Acad. Sci. U.S.A.
179, 190-196.
Hermodson M. A., Ericsson L. H., Neurath H. and Walsh K. A. (1973) Determination of the amino acid sequence of porcine trypsin by sequenator analysis. Biochemistry
Nature,
thank Dr T. Kumazaki, Faculty of Pharmaceutical Science and Dr A. Kimura, Faculty of Agriculture, Hokkaido University, for their valuable suggestions during this work. Acknowledgemenrs-We
Res. Commun.
Gieselmann V., Hasilik A. and von Figura K. (1985) Processing of human cathepsin D in lysosomes in vitro. J. biol. Chem. 260, 3215-3220. Hasilik A. and Neufeld E. F. (1980) Biosynthesis of lysosoma1 enzymes in fibroblasts. Synthesis as precursors of higher molecular weight. J. biol. Chem. 235, 4937-4945. Hentze M., Hasilik A. and von Figura K. (1984) Enhanced degradation of cathepsin D synthesized in the presence of the threonine analog b-hydroxynorvaline. Archs Biochem.
82, 491@4914.
Friedman M., Krull L. H. and Cavins J. F. (1970) The chromatographic determination of cystine and cysteine residues in proteins as S-b-(4-pyridylethyl)cysteine. J. biol. Chem. 245, 3868-3871. Fujita H., Tanaka Y., Noguchi Y., Kono A., Himeno M. and Kato K. (1991) Isolation and sequencing of a cDNA clone encoding rat liver lysosomal cathepsin D and the
Schwartz T. W. (1986) The processing of peptide precursors. ‘Proline-directed arginyl cleavage’ and other monobasic processing mechanisms. FEBS Lett. 200, l-10. Shewale J. G. and Tang J. (1984) Amino acid sequence of porcine spleen cathepsin D. Proc. natn. Acad. Sci. U.S.A. 81, 3703-3707. Takio K., Towatari T., Katunuma N., Teller D. C. and Titani K. (1983) Homology of amino acid sequences of rat liver cathepsin B and H with that of papain. Proc. natn. Acad. Sci. U.S.A.
80, 3666-3670.
Tang J. and Wong R. N. S. (1987) Evolution in the structure and function of aspartic proteases. J. cell. Biochem.
3,
5363.
Yonezawa S., Takahashi T., Wang X., Wong R. N. S., Hartsuck J. A. and Tang J. (1988) Structures at the proteolytic processing region of cathepsin D. J. biol. Chem. 263, 16,50416,511.
inr. .I Biochem.Vol.24.No.9,pp.1487-1491, 1992 Printed in Great Britain. All rights reserved
PROTEOLYTIC PROCESSING SITES PRODUCING THE MATURE FORM OF HUMAN CATHEPSIN D TAKAHIKO KOBAYASHL’ KOICHI HONKE, I* SHINSEI GASA,' TETSUYA FUJII,’ SHIRO MAGUCHI,’ TAMOTSUMIYAZAKI’ and AKIRA MAKITA’
LBiochemistry Laboratory, Cancer Institute and 2The Third Department of Internal Medicine. Hokkaido University School of Medicine, Sapporo 060, Japan (Received 1I December 1991) Abstract-l. The proteolytic processing sites of human lysosomal aspartic protease cathepsin D at which the intermediate single-chain form was converted into the mature two-chain form were determined. 2. The two chains were isolated by reversed-phase HPLC in order to investigate the cleavage sites of the enzyme. 3. Protein sequencing of the heavy chain, which was presumed to be derived from the C-terminal side in the single-chain enzyme, gave an N-terminal Leu 105. In addition, it revealed that there were also minor sequences, which commenced with Gly 106 and Gly 107. 4. A small C-terminal peptide was isolated from the light chain, which had been digested with two kinds of exogenous proteases. Sequence deter~nation of this peptide, which was characterized as a nona~ptide by mass s~ctrometry, suggested that the C-terminus of the light chain was Ser 98. 5. These results indicate that a Ser 9%Ala 99 bond and an Ala 104-Leu IO5 bond are cleaved to release 6 amino acid residues between the two chains.
Cathepsin D (EC 3.4.23.5) is a lysosomal aspartic protease that is present in all mammalian cells, and its
biosynthesis and protein structure have been studied (for review, Barret, 1977; Tang and Wong, 1987; Erickson, 1989). Recently, the primary structure of porcine cathepsin D was determined by protein sequencing (Shewale and Tang, 1984), and those of human (Faust et al., 1985) and rat (Fujita et al., 1991) cathepsin D were deduced from the cDNA sequence. Human cathepsin D, like other lysosomal hydrolases, is synthesized as a preproenzyme that undergoes subsequent proteolytic processing to produce the mature enzyme (Hasilik and Neufeld, 1980; Erickson et a!., 198 1). The first cleavage is co-translational and removes the signal peptide to yield procathepsin D. The proenzyme is processed to remove 44 amino acid residues from the N-terminal, producing an active single-chain enzyme with a relative molecular mass of 44,000. The single-chain enzyme is eventually cleaved into a M, 15,000 light chain and a IV, 30,000 heavy chain. The sequences of the last processing region were analyzed in porcine and bovine cathepsin D (Yonezawa et al., 1988) and rat cathepsin D (Fujita et al., 1991). When human cathepsin D was aligned by sequence homology against porcine cathepsin D, it was suggested that 7 amino
acid residues were removed in the processing region during the conversion of the single-chain form to the *To whom correspondence
should be addressed.
two-chain form of human cathepsin D (Faust et al., 1985). However, the actual processing sites where this occurred remained undefined. The present study was undertaken to determine the processing sites of human cathepsin D that produce the mature enzyme.
MATERIALS AND METHODS Materiais Achromobacter Protease I (Lysyl Endopeptidase) was purchased from Wako Chemicals, and endoproteinase Asp-N from Boehringer Mannheim Yamanouchi. Other chemicals were of analytical grade. Cathepsin D was purified from human liver as described previously (Magu~hi et al., 1988). Normal human liver tissues were obtained at autopsy, and stored immediately at -80°C until use. Enzymatic digestion Digestion of the light chain (80 nmol) with Achromobacter Protease I with a SO:1 (moi/mol) ratio of substrate to enzyme, was carried out in 50 mM Tris-HCI, pH 9.0, for 20 hr at 30°C. The peptide termed Ly4 was treatd with endoproteinase Asp-N with a 100: 1 (w/w) ratio of substrate to enzyme, in 50 mM sodium phosphate buffer, pH 8.0, with 1 M urea for 18 hr at 37°C. Reversed-phase HPLC Nine milligrams of purified cathepsin D was reduced and S-pyridylethylated (Friedman et al., 1970; Hermodson et al., 1973). The treated enzyme was applied to a reversedphase column @BONDERSPHERE, 5 p, C8-300 A, 0.39 x 15 cm, Waters), and eluted with a linear gradient of from 0 to 80% acetonitrile in 0.05% trifluoroacetic acid. The Aeh~omobucfe~ Pro&ease I hydrolysate of the light chain
1487
TAKAHIKO KOBAYASHI et al.
1488
was chromatographed on the same column under the same conditions. Endoproteinase Asp-N hydrolysate of the C-terminal peptide of the light chain was subjected to another reversed-phase column (p BONDERSPHERE, 5 p. C18100 A, 0.39 x 15cm, Waters), and eluted with the solvent system described above. Determination sequencing
:
Mrxlo-s
2 -42.7 e
-31.0 -21.6
6
m-l
Seven nanomoles of each separated Achromobacter Protease I digest was hydrolyzed under vacuum in a 6 M HCI at 110°C for 24 hr and studied with an amino acid analyzer (Hitachi 835). The N-terminal sequence analysis was performed by automated Edman degradation using a protein sequencer (Applied Biosystem 477A).
14.4
(FAB-MS)
FAB-MS in a positive mode was performed with a JMS-HXllO mass spectrometer (JEOL) equipped with a JMA-DA5000 data system (JEOL). Approximately I nmol of a peptide termed A3 was bombarded by Xe gas at 3 kV in a matrix of glycerol, and the fragments were accelerated at 8 kV. Polyacryamide gel elecrrophoresis Polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate and p-mercaptoethanol (SDS PAGE) was performed according to the method of Laemmli (1970). The gel was stained for protein with silver stain. RESULTS
Separation of human cathepsin D subunits
When purified mature cathepsin D was reduced, S-pyridylethylated, and subjected to reversed-phase HPLC, two major peptides, L and H, were separated (Fig. 1). The isolated peptides L and H were homogeneous on SDS-PAGE (Fig. 1, inset), corresponding to the molecular weights of the light chain and the heavy chain of liver cathepsin D, respectively (Barret, 1977; Maguchi et al., 1988). N-Terminal sequence of the heavy chain
Comparing the amino acid homology of human cathepsin D to that from other animal sources (Yonezawa et al., 1988), the heavy chain was presumed to be derived from the C-terminal portion of the single-chain form. When the N-terminal sequence of peptide H was analyzed, it was found that a major sequence (_ 55%) coincided with the peptides begin-
0
Time (min
)
Fig. I. Separation of human cathepsin D subunits. Purified cathepsin D was reduced, S-pyridylethylated and subjected to reversed-phase HPLC using a 0.39 x 15 cm Waters FBONDERSPHERE, 5 p, C8-300 A column. The peptides were eluted with a linear gradient of from 0 to 80% acetonitrile in 0.05% trifluoroacetic acid. Inset: SDS-PAGE of the isolated peptides L and H. A part (approx. 1 pg each) of the peaks L and H on the HPLC was subjected to SDS-PAGE under a reduced condition and stained with silver. Molecular weights of standard proteins are indicated on the right side.
ning with Leu 105 of human cathepsin D (Table 1). In addition, peptide H included other minor sequences which commenced with Gly 106 (~30%) and Gly 107 (- 10%). In the first cycle of the sequence analysis, a small amount of alanine was also detected in addition to leucine and glycine. However, it could not be determined whether another peptide starting with Ala 104 existed or not, since we should consider the influence of the carryover after the first cycle, which originated from the major peptide commencing with Leu 105. This N-terminal heterogeneity of the heavy chain was reproducible in three different heavy chain preparations. Based on these observations, it was concluded that the heavy chain was a mixture of polypeptides beginning with Leu 105, Gly 106 and Gly 107.
Table I. Seauence determination Data source
cDNA Peptide A3 Peptide H
H
‘I
- 66.2
of amino acid composition and protein
Fast atom bombardment-mass speciromeiry
I-
of the arocessinn reaion
Amino acid sequence 90
98
105
I
I
I
SQDTVSVPCQSASSASALGGVKVERQVFGE DTVSVPCQS LGGVKVERQV GGVKVEROVF GVKVERQVFG
(-55%) (- 30%) i- io%j
The deduced amino acid sequences from cDNA of human cathepsin D (Faust et al., 1985) and the determined ones are shown in the one letter code. Amino acid No. 1 is assigned to the first residue of the mature enzyme protein.
Human cathepsin D
1489
Table 2. Amino acid composition of peptide Ly4
Sequence of the C-terminal region of the light chain
Peptide Ly4
In order to determine the C-terminal of the light chain, peptide L was digested with Achromobacter Protease I and subjected to reversed-phase HPLC. Six peptides were separated (Fig. 2), and the amino acid composition of each peptide was analyzed. The amino acid composition of the isolated peptide termed Ly4 was consistent with that of the C-terminal region (Asn 70-Ser 98) deduced from cDNA (Table 2). To confirm the C-terminal sequence of peptide Ly4, the peptide was further digested with endoproteinase Asp-N. Although the hydrolysate of peptide Ly4 was expected to separate into three peptides, it was separated into four peaks by reversed-phase HPLC (Fig. 3). Through protein sequencing of each peptide, it was revealed that the peptide called A3 was coincident with a nonapeptide from Asp 90 to Ser 98 of human cathepsin D (Table I). Peptide Al was identical to peptide A2 with respect to the amino acid sequence, which coincided with a peptide from Asn 70 to Phe 74, though Asn 70 was not detected because of the addition of a carbohydrate chain (data not shown). Asn 70 is the only potential N-glycosylation site in the light chain. The light chain of human liver cathepsin D contained endo-fi-N-acetylglucosaminidase H sensitive and insensitive carbohydrate chains (Ohhira et al., 1991).
ASP Thr ser GilI GUY Ala Val Met Ile
Leu TYr Phe LYS His PEb-Cys Arg Pro
Am 7fKkr
3.3 (3) 2.0 (2) 6.6 (7) 2.3 (2) 4.6 (5) 0.1 (0) 2.2 (2) 0.0 (0) 1.0 (1) 2.1 (2) 2.0 (2) 1.1 (1) 0.1 (0) I.0 (1) 0.4 (0) 0.0 (0) 1.3 (1)
98*
3 2 7 2 4 0 2 0
1 2 2 1 0
I 1 0
I
Data
in the table are expressed in molar ratios relative to the vaIues for isoleucine as indicated. The figures in parentheses show the nearest integers, ‘Peptide from Asn 70 to Ser 98, deduced from human cathepsin D cDNA. bPE, S-pyridylethylated
Therefore, abIy due structure. Asp 75 to
peptides Al and A2 were separated probto the difference of their oligosa~haride Peptide A4 coincided with a peptide from Gln 89. Peptide A3 was also characterized
A4
R
a LYE
A3
A2 LY5 LY4 I
P 1
all.LY3
LYf
-II-
LY2
-e--
Fig. 2. Separation of Achromobacrer Protease I hydrolysate of the light chain. Peptide L in Fig. 1 was digested with Achromobacfer Protease I (peptide : enzyme ratio 50 : 1 (mol/mol) in 50 mM Tris-HCI, pH 9.0 for 20 hr at 30°C) and subjected to reversed-phase HPLC using the same column and the same conditions as described in Fig. 1.
C
Fig. 3. Separation of endoproteinase Asp-N hydrolysate of the C-terminal peptide Ly4. Peptide Ly4 (Fig. 2) was digested with endoproteinase Asp-N (peptide:enzyme ratio 100: 1 (w/w) in 50 mM sodium phosphate buffer, pH 8.0 containing 1 M urea for 18 hr at 37°C) and subjected to reversed-phase HPLC using a 0.39 x 15cm Waters FBONDERSPHERE, 5~ C18-lOO A column under the same condition as described in Fig. 1.
1490
TAKAHIKO
50
KOBAYASHI
et al.
115 207
x ,’ D 5 0 0 aI .?
0 *100 100
200
300
400
500
700
600
--
6 f
(M+H)+
50
1040
I
O-
800
900
I
1000
1100
1200
1300
1400
1500
m/z
Fig. 4. Positive fast atom bombardment-mass spectrum of peptide A3. Approximately 1 nmol of a peptide A3 (Fig. 3) was bombarded by Xe gas at 3 kV in a matrix of glycerol, and the fragments were accelerated at 8 kV. A major peak at m/z 1040 corresponds to the molecular ion (M + H)+ of the S-pyridylethylated nonapeptide, Asp-Thr-Val-Ser-Val-Pro-PE-Cys-Gln-Ser, where PE means S-pyridylethylated.
by positive
fast analysis
atom
bombardment-mass
spec-
(Fig. 4). One major peak at m/z 1040 corresponded to the molecular ion (M + H)+ of the S-pyridylethylated nonapeptide. Therefore, we concluded that the C-terminus of the light chain was Ser 98. These results indicated that a Ser 98-Ala 99 bond and an Ala 104Leu 105 bond were cleaved to release six amino acid residues between the two chains (Table 1). trometry
DISCUSSION
We determined the proteolytic processing sites of human cathepsin D where the single-chain form was converted into the two-chain form. The splitting site of human cathepsin D was close to those of bovine, porcine and rat cathepsin Ds (Shewale and Tang, 1984; Yonezawa et aI., 1988; Fujita et al., 1991) and the processing regions was located in a p-hairpin loop (Yonezawa et al., 1988). Comparing human cathepsin D with the enzyme from the other species, however, it was found that the amino acid sequences within the processing region and the proteolytic processing sites differ considerably. The susceptibility to the proteolytic processing varies with the species. Human and porcine cathepsin D are mostly composed of the two-chain form (Hasilik and Neufeld, 1980; Huang et al., 1979) the rat enzyme contains mostly the single-chain form and a small amount of the two-chain form (Yonezawa et al., 1988; Fujita et al., 1991), whereas the bovine enzyme contains almost equivalent amounts of each of the two forms (Huang et al., 1980). There are neither basic amino acid residues, which are targetted by the so-called processing proteases (Loh et al., 1984; Schwartz, 1986; Bond and Butler, 1987) nor an asparaginyl-X
bond, which is involved in the proteolytic processing of many lysosomal hydrolases (Takio et al., 1983; Mahuran et al., 1988; Shewale and Tang, 1984), in the processing region of human cathepsin D. It was predicted the cleavage would occur directly after Gln 97 in the human single chain sequence, based on the amino acid sequence homology between human and porcine cathepsin Ds (Faust et al., 1985). Since asparagine and glutamine have an amide side chain, the possibility that the amide side chain plays an important role in site recognition by the processing protease has been presented (Erickson, 1989). However, the cleavage was found to occur not after Gln 97 but after Ser 98 in this study. These observations indicate that the proteolytic processing of the single-chain enzyme of cathepsin D from various species is not carried out by a common specific processing protease which recognizes the primary structure. Alternatively, it was suggested by experiments using inhibitors of cysteine proteases (Hentze et al., 1984; Gieselmann et al., 1985) that lysosomal cystein proteases such as cathepsin B, H or L, which have a broad peptide bond specificity, act on the proteolytic processing, which converts the intermediate single-chain form of cathepsin D into the mature two-chain form. It is not clear whether these cathepsins recognize the unique conformation of the processing region. As in the heavy chain of human cathepsin D, heterogeneity at the N-terminus has been shown in the subunits of human /I-hexosaminidase (Mahuran et al., 1988). Therefore, N-terminal trimming may be found in many lysosomal hydrolases. Although the possibility that the heavy chain suffered from non-physiological proteolysis during the purification of the enzyme is not excluded, it is likely that the
1491
Human cathepsin D N-terminal heterogeneity results from the physiological processing by cystein proteases, which display aminopeptidase activity (Barrett and Kirschke, 1981), or an unknown protease within the lysosomes. Quite recently, during preparation of this manuscript, Horst et al. (1991) also reported the N-terminal heterogeneity of the heavy chain of cathepsin Ds which were isolated from BHK cells transfected with human cathepsin D cDNA and from human promonocytes U937. However, they did not examine the C-termini of the light chains. The physiological significance of the cleavage which forms the two-chain enzyme is unclear. It is difficult to distinguish clearly whether the proteolytic cleavage occurs during a maturation process or during a catabolic process. Considering that the single-chain and two-chain forms appear to be equally active (Huang et al., 1979), the cleavage may be the first step of degradation, although the nick might bring about stability in the conformation. We are now studying the proteolytic processing regions of other lysosomal enzymes. The accumulation of such data should give us a clue as to the physiological significance of proteolytic processing, and help to identify the processing proteases.
structure of three forms of mature enzymes. Biochem. biophys.
REFERENCES
Barret A. J. (1977) Cathepsin D and other carboxyl proteinases. In Proteinases in Mammalian Cells and Tissues (Edited by Barret A. J.), pp. 209-248. North-Holland, Amsterdam. Barret A. J. and Kirschke H. (1981) Cathepsin B, cathepsin H, and cathepsin L. Meth. Enzym. 80, 535-561. Bond J. S. and Butler P. E. (1987) Intracellular proteases. A. Rev. Biochem.
56, 333-364.
Biophys.
230, 375-382.
12, 31463153.
Horst M. and Hasilik A. (1991) Expression and maturation of human cathepsin D in baby-hamster kidney cells. Biochem.
J. 273, 355-361.
Huang J. S., Huang S. S. and Tang J. (1979) Cathepsin D isozymes from porcine spleens. Large scale purification and polypeptide chain arrangements. J. biol. Chem. 254, 11,405-11,417. Huang J. S., Huang S. S. and Tang J. (1980) Enzyme Regulation and Mechanism of Action (Edited by Mildner P. and Ries B.), pp. 289-306. Pergamon Press, Oxford. Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Lond. 227, 680-685.
Loh Y. P., Brownstein M. J. and Gainer H. (1984) Proteolysis in neuropeptide processing and other neural functions. A. Rev. Neurosci. 7, 189-222. Maguchi S., Taniguchi N. and Makita A. (1988) Elevated activity and increased mannose-6-phosphate in the carbohydrate moiety of cathepsin D from human hepatoma. Cancer Res. 48, 362-367. Mahuran D. H., Neote K., Klavins M. H., Leung A. and Gravel R. A. (1988) Proteolytic processing of pro-a and pro-/J precursors from human p-hexosaminidase. J. biol. Chem. 263, 46124618. Ohhira M., Gasa S., Makita A., Sekiya C. and Namiki M. (1991) Elevated carbohydrate phosphotransferase activity in human hepatoma and phosphorylation of cathepsin D. Br. J. Cancer 63, 905-908.
Erickson A. H. (1989) Biosynthesis of lysosomal endopepti dases. J. cell. Biochem. 40, 3141. Erickson A. H., Conner G. E. and Blobel G. (1981) Biosynthesis of a lysosomal enzyme. Partial structure of two transient and functionally distinct NH,-terminal sequences in cathepsin D. J. biol. Chem. 256, 11,22411,231. Faust P. L., Kornfeld S. and Chirgwin J. M. (1985) Cloning and sequence analysis of cDNA for human cathepsin D. Proc. natn. Acad. Sci. U.S.A.
179, 190-196.
Hermodson M. A., Ericsson L. H., Neurath H. and Walsh K. A. (1973) Determination of the amino acid sequence of porcine trypsin by sequenator analysis. Biochemistry
Nature,
thank Dr T. Kumazaki, Faculty of Pharmaceutical Science and Dr A. Kimura, Faculty of Agriculture, Hokkaido University, for their valuable suggestions during this work. Acknowledgemenrs-We
Res. Commun.
Gieselmann V., Hasilik A. and von Figura K. (1985) Processing of human cathepsin D in lysosomes in vitro. J. biol. Chem. 260, 3215-3220. Hasilik A. and Neufeld E. F. (1980) Biosynthesis of lysosoma1 enzymes in fibroblasts. Synthesis as precursors of higher molecular weight. J. biol. Chem. 235, 4937-4945. Hentze M., Hasilik A. and von Figura K. (1984) Enhanced degradation of cathepsin D synthesized in the presence of the threonine analog b-hydroxynorvaline. Archs Biochem.
82, 491@4914.
Friedman M., Krull L. H. and Cavins J. F. (1970) The chromatographic determination of cystine and cysteine residues in proteins as S-b-(4-pyridylethyl)cysteine. J. biol. Chem. 245, 3868-3871. Fujita H., Tanaka Y., Noguchi Y., Kono A., Himeno M. and Kato K. (1991) Isolation and sequencing of a cDNA clone encoding rat liver lysosomal cathepsin D and the
Schwartz T. W. (1986) The processing of peptide precursors. ‘Proline-directed arginyl cleavage’ and other monobasic processing mechanisms. FEBS Lett. 200, l-10. Shewale J. G. and Tang J. (1984) Amino acid sequence of porcine spleen cathepsin D. Proc. natn. Acad. Sci. U.S.A. 81, 3703-3707. Takio K., Towatari T., Katunuma N., Teller D. C. and Titani K. (1983) Homology of amino acid sequences of rat liver cathepsin B and H with that of papain. Proc. natn. Acad. Sci. U.S.A.
80, 3666-3670.
Tang J. and Wong R. N. S. (1987) Evolution in the structure and function of aspartic proteases. J. cell. Biochem.
3,
5363.
Yonezawa S., Takahashi T., Wang X., Wong R. N. S., Hartsuck J. A. and Tang J. (1988) Structures at the proteolytic processing region of cathepsin D. J. biol. Chem. 263, 16,50416,511.
