J. Biochem. 104, 22-29 (1988)

Primary Structure of Human Urinary Prokallikrein1

Saori Takahashi,* Akiko Irie,** and Yoshihiro Miyake*, 1*Department of Biochemistry , National Cardiovascular Center Research Institute and **Clinical Laboratory, National Cardiovascular Center Hospital, Suita, Osaka 565

Received for publication, January 26, 1988

The complete amino acid sequence of human urinary prokallikrein has been determined by amino acid analysis and sequence determination of peptide fragments obtained from chemical and enzymological cleavages of and by comparison of the N-terminal sequence of prokallikrein with that of kallikrein, the active form. Prokallikrein was a single chain polypeptide which comprised 238 amino acid residues of kallikrein and 7 amino acid residues of the propeptide. The sequence, Asn-X-Thr(Ser), which is a common glycosylation site was found at positions 78-80, 84-86, and 141-143. Two -suscep tible sites were identified. One is the Arg(-1)-Ile(1) bond and the other is the Arg (87)-Gln(88) bond. The sequence of human urinary kallikrein was identical with that of human pancreatic and kidney (Fukushima, D. et al. (1985) Biochemistry 24, 8037-8043; Baker, A.R. & Shine, J. (1985) DNA 4,445-459), which were predicted from the nucleotide sequences of cDNAs. The primary structure of human urinary kallikrein is homologous to those of the other animal kallikreins and kallikrein-related proteins. Key amino acid residues, His(41), Asp(96), and Ser(190), required for catalytic activity and Asp (184) required for kallikrein-type specificity are completely conserved. The results show that human urinary prokallikrein and kallikrein are of tissue type and they are excreted in urine without any modification.

Tissue-type kallikreins represent a group of closely related from human urine and characterized some of their prop serine proteases which liberate lysyl-bradikinin, a vaso erties (18). In addition, it has been shown from the active decapeptide, from kininogens. These are N-terminal amino acid sequences of the two proteins that found in many mammalian tissues (1-4). Recently, the the inactive form is a proenzyme, prokallikrein, having an primary structures of tissue kallikreins have been predict- additional seven amino acids at the amino terminus of ed from molecular cloning and sequence analysis of cDNAs kallikrein (19). However, the complete amino acid se encoding mouse submaxillary gland kallikrein (5), rat and quence of human urinary prokallikrein has not been human pancreatic kallikreins (6, 7), and human kidney determined. kallikrein (8). In the mouse, these enzymes are encoded by In the present study, we have determined the complete a closely linked multigene family on chromosome 7 (9). amino acid sequence of human urinary prokallikrein, and The urinary kallikrein is particularly interesting from the sequence was compared with those of kallikreins of both physiological and clinical aspects, since it may partici other mammalian species and kallikrein-like proteins. On pate in the regulation of water and electrolyte balance (10), the basis of the present results and our earlier data (18-20) and the amount of total kallikrein and the ratio of active- we identified trypsin-susceptible sites in human urinary to-total kallikrein in urine changed with essential hyperten prokallikrein, and proposed an activation mechanism of tion and high-renin Bartter's syndrome (11-13). Human prokallikrein by trypsin as a model for the physiological urinary kallikrein is a single-chain polypeptide and an activation. acidic glycoprotein, and the partial amino acid sequence

was reported (14, 15). Moreover, the existence of an MATERIALS AND METHODS inactive form of kallikrein in human urine has been demonstrated (16, 17). However, the physiological sig Materials-Human urinary kallikrein and prokallikrein nificance and the structural relationship between the two were purified by the method of Irie et al. (18). The purified forms of kallikrein remained to be settled. Recently, we preparations were homogeneous as judged by poly have purified both active and inactive forms of kallikrein acrylamide gel electrophoresis at pH 8.0 and sodium dodecyl sulfate polyacrylamide gel electrophoresis. Lysyl endopep 1 Preliminary results have been presented at the 5th International tidase from Achromobacter lyticus was purchased from Congress on Kinin held in Tokyo, Nov. 29-Dec. 3, 1987. This work Wako Pure Chemical Co., Tokyo, TPCK-treated bovine was supported in part by Grants-in-Aid for Special Project Research pancreatic trypsin from Cooper Biomedical, bovine pancre (No. 62122005) and for Scientific Research (No. 61570151) from the atic from Sigma, and a Pep RPC HR 5/5 Ministry of Education, Science and Culture of Japan. column (5•~50mm, 100A) for high-performance liquid 2 To whom correspondence should be addressed . chromatography (HPLC) from Pharmacia Fine Chemicals Abbreviations: HPLC, high-performance liquid chromatography; . PTH, phenylthiohydantoin; NGF, nerve growth factor; EGF, epider All other reagents were of guaranteed grade. mal growth factor. Reduction and Carboxymethylation•\Human urinary

22 J. Biochem. Human Urinary Prokallikrein; Complete Amino Acid Sequence 23

kallikrein and prokallikrein were reduced and carboxy quenator (Applied Biosystems, Model 430A). The obtained methylated by the method of Lottspeich et al. (15). The PTH-amino acid were identified by a PTH-analyzer (Ap- preparations (500ƒÊg) were lyophilized and then dissolved plied Biosystems, Model 120A PTH analyzer) directly in 500,u 1 of 6 M guanidine hydrochloride containing 0.2M connected to the sequenator. All peptide solutions were acetic acid, 0.1M Tris, and 5mM EDTA, pH 4.7. The made in 25% trifluoroacetic acid containing 0.1% SDS solution was flushed with nitrogen gas, and 25 pl of 2- before being applied to the sequenator. mercaptoethanol was added. The solution was incubated for Amino Acid Analysis S-Carboxymethylated proteins 20 h at 40•Ž, and then 500,u 1 of 2M Tris-HCl, pH 9.0, and peptides were hydrolyzed under HCl vapor at 110•Ž for containing 60 mg of iodoacetate and 6M guanidine hy 20 h. The hydrolysates were analyzed by the phenylthio- drochloride was added. The reaction mixture was incubated carbamyl method with the Waters Pico-Tag system (21). for 15 min at room temperature. The solution was acidified to about pH 3.0 with 50% formic acid, then immediately RESULTS applied to a Sephadex G-15 column (1 x 50 cm) previously equilibrated with 50 % acetic acid. The protein fractions Amino Acid Compositions of Human Urinary Kallikrein were collected and lyophilized. and Prokallikrein-Table I shows the amino acid composi- Lysyl Digestion•\S-Carboxymethylated tions of S-carboxymethylated human urinary kallikrein kallikrein (490ƒÊg) was dissolved in 300 pl of 50mM and prokallikrein. Both proteins showed a very similar Tris-HCl, pH 9.0, and then 5 pl of lysyl endopeptidase amino acid composition. However, amino acids of the solution (10 units/ml of distilled water) was added. The propeptide could not be distinguished under the conditions solution was incubated for 2 hat 37•Ž and then 20 hat room employed. The amino acid composition of kallikrein was temperature. The reaction was terminated by the also found to be very similar to that reported earlier (15). addition of 200 pl of 1% trifluoroacetic acid. Sequencing Strategy-The sequencing strategy for hu- Chymotryptic Digestion of Peptide L9•\Chymotrypsin man urinary kallikrein is shown in Fig. 1. Two sets of was used for fragmentation of peptide L9, a peptide from fragments of S-carboxymethylated kallikrein were pre- the lysyl endopeptidase-digested sample. The peptide (5 pared. One is lysyl endopeptidase-digested fragments and the nmol) was dissolved in 100 pl of 0.1M Tris-HCl, pH 7.4, other is cyanogen bromide-cleaved fragments. Subfrag- and digested with 0.5ƒÊg of chymotrypsin at 37•Ž for 4 h. ments of some peptides from lysyl endopeptidase digestion The resulting peptides were purified by reversed-phase and cyanogen bromide cleavage were obtained by digestion HPLC. with trypsin or chymotrypsin. Tryptic Digestion of Peptide L10•\Peptide L10, a pep- Amino Acid Sequences of Peptides from Lysyl Endopep- tide from the lysyl endopeptidase-digested sample, was tidase Digestion-Peptides from lysyl endopeptidase diges- further digested with trypsin. The peptide (4nmol) was tion of S-carboxymethylated kallikrein were purified by dissolved in 300ƒÊl of 50mM Tris-HCl, pH 8.0 and then 5 reversed-phase HPLC. The elution profile is shown in Fig. ls. With small fractions (Ll, L2, L3, L4, L5, L6, and L7), pl of trypsin (0.1mg/ml) was added to the solution. The the amino acid sequences were directly determined by reaction was carried out at 25•Ž overnight. Chemical Cleavage with Cyanogen Bromide-For the automated Edman degradation and the amino acid composi- cleavage of methionine residues, 450ƒÊg of S-carboxy tion of each peptide was analyzed. Amino acid analysis and methylated sample was dissolved in 300ƒÊl of 50% formic acid, and then 2mg of cyanogen bromide was added. The reaction was allowed to proceed for 20 h at 37•Ž. The TABLE I. Amino acid compositions of human urinary kalli- krein and prokallikrein.' reaction mixture was diluted 5-fold with distilled water and lyophilized. The lyophilized material was dissolved in 500 p 1 of 50% acetic acid for chromatography. Chymotryptic Digestion of Cyanogen Bromide-Cleaved Peptide-Proteolytic digestion of cyanogen bromide- cleaved peptides (CB3, CB5, and CB6) with chymotrypsin was carried out in 0.1M Tris-HCl, pH7.5, at 37•Ž for 4 h. The molar ratio of chymotrypsin to peptide was about 1 to 200 Purification of Peptides by Reversed-Phase HPLC-Pep tides from cyanogen bromide cleavage and protease diges tion of S-carboxymethylated kallikrein were separated by reversed-phase HPLC immediately after digestion. HPLC was carried out on the fast protein liquid chromatography (FPLC) system on a reversed-phase column of Pep RPC HR5/5. The aqueous mobile phase was 0.1% trifluoroacetic acid in water and organic mobile phase was 0.1% tri fluoroacetic acid in acetonitrile. The flow rate was 0.5 ml

per min. Fractions were collected and stored at -20•Ž until aNumbers of amino acid residues per mol were obtained from amino use. Amino Acid Sequence Analysis -Automated sequence acid analysis and values in parentheses are the numbers from sequence analysis. •Žys was identified as S-carboxymethylcysteine. `Not analyses of kallikrein, prokallikrein, and peptide fragments determined. isolated by HPLC were performed with a gas-phase se

Vol. 104, No. 1, 1988 24 S. Takahashi et al.

Fig. 1. The alignment of peptides derived from S-carboxymethylated human urinary kallikrein by enzymatic digestion or chemical cleavage. The peptides which were obtained by degradation with cyanogen bromide (CB), lysyl endopeptidase (L), trypsin (T), and chymotrypsin (C) are shown. Solid lines show the sequences determined by automated Edman degradation. Broken lines show undetermined sequences.

Edman degradation of L8 and L9 also revealed that each analyzed. Among them, the amino acid compositions and fraction contained a single peptide (Tables Is and Ills). L10 sequences of L10-T5, L10-T6, and L10-T7 are shown in was from the amino terminus by comparison with the Tables Is and IVs. As a lysine residue was detectable only amino-terminal sequence of intact kallikrein (data not in L10-T2 among the L10 fragments analyzed, the frag shown). ment was placed at the carboxyl terminus of L10 (Fig. 1). L9 and L10 were further digested with chymotrypsin or Edman degradation of L10-T3, L10-T5, L10-T6, and trypsin and the resulting peptides were separated by L10-T7 disclosed that each fraction contained a single reversed-phase HPLC (Figs. 2s and 3s). Two major and one peptide. With L10-T7, no PTH-amino acid was detectable minor fractions (L9-C1, L9-C2, L9-C3 in Fig. 2s) from the at the 25th or 31st cycle of Edman degradation (Table IVs). chymotryptic digest of L9 were then sequenced. As shown These two residues are probably also asparagine from in Fig. 1, the sequences of L9-C3 and L9-C2 covered almost Table Is and IVs, and may be glycosylation sites (Figs. 1 and all of the sequence of L9. L9-C1 was a tetrapeptide of the 3). C-terminal portion of L9-C3 (data not shown). In the Amino Acid Sequences of Peptides from Cyanogen Bro sequence analysis of L9-C2, no PTH-amino acid was mide Cleavage•\S-Carboxymethylated kallikrein was detectable at the 7th cycle of Edman degradation (Table cleaved with cyanogen bromide. Reversed-phase HPLC of Ills). This residue was probably asparagine as judged from cyanogen bromide-cleaved kallikrein on Pep RPC HR5/5 the amino acid analysis (Table Is) and sequencing of L9 and gave six peaks (Fig. 4s). Sequencing and amino acid L9-C2 (Table Ills), and it may be a glycosylation site analysis of the peptides revealed that four fractions (CB1, having the sequence Asn-Phe-Ser (Figs. 1 and 3). CB2, CB3, and CB5) contained a single peptide (Tables IIs, L10 was digested with trypsin and separated by re- IVs, and Vs), and CB4 was intact kallikrein (data not versed-phase HPLC, and eight fractions (L10-T1-L10-T8 shown). The large fragments, CB3, CB5, and CB6, were in Fig. 3s) from L10 were sequenced. Amino acid compo further digested with chymotrypsin and the resulting sitions of six peptides (L10-T2, to L10-T7) were also peptides were separated by reversed-phase HPLC (Figs.

J. Biochem. Human Urinary Prokallikrein; Complete Amino Acid Sequence 25

Fig. 2. N-Terminal amino acid sequence of human urinary kallikrein and prokallikrein. S- Carboxymethylated kallikrein (1.2nmol) and pro kallikrein (0.95nmol) were used for automated sequence analysis. The propeptide sequence is un derlined. cm-Cys, S-carboxymethyleysteine.

Fig. 3. Primary structure of human urinary prokallikrein. The sequence between Ala(-7) and Arg(-1) is the propeptide. Glycosylated asparagine residues are boxed. Open arrowhead, the bond that is cleaved on activation and the very rapidly trypsin -susceptible site; solid arrowhead, the second slowly trypsin-susceptible site; solid circles, the charge relay amino acid residues; CHO, oligo-saccharide.

5s-7s). The chymotryptic digest of CB3 showed more than quences of CB5-C2, CB5-C5, and CB5-C9 were also deter- mined, and amino acid compositions of these seven peptides ten fractions (Fig. 5s). Among them, CB3-C1, CB3-C2, were also analyzed. The amino acid compositions and CB3-C3, and CB3-C4 were sequenced and the amino acid sequences of CB5-C6 and CB5-C8 are shown in Tables Its compositions were analyzed. The results with CB3-C2, are and Vs. Digestion of CB6 with chymotrypsin produced shown in Tables Its and Vs (data with CB3-C1, CB3-C3, about seven fractions as found by reversed-phase HPLC and CB3-C4 not shown). Six major and 3 minor peptides were obtained from CB5 (Fig. 6s). The sequences of (Fig. 7s). The amino acid compositions and sequence of CB6-C1, CB6-C2, and CB6-C3 were determined, and the CB5-C1, CB5-C6, CB5-C7, and CB5-C8 could be deter- results with CB6-C2 and CB6-C3 are shown in Tables Its mined completely by Edman degradation. Partial se

Vol. 104, No. 1, 1988 26 S . Takahashi et al.

Fig. 4. Amino acid sequences of kallikreins and related proteins. The regions of identical amino acid residues among the proteins are boxed. Sequence sources: (a) this study, (b) Fukushima et al. (7), (c) Fiedler et al. (23), (d) Swift et al. (6), (e) Mason et al. (9), (f) van Leeuwen et al. (24), (g) Evans and Richards (25), (h) Lundgren et al. (26).

and Vs (data with CB6-C1 not shown). hydrolyzed on activation of prokallikrein, and the active The sequence and yields of amino acid residues in Tables form consisting of 238 amino acid residues is formed by Ills, IVs, and Vs and the amino acid compositions in Tables release of the propeptide (open arrowhead in Fig. 3). Two Is and Its are those of the main peptides among those trypsin-susceptible sites were also identified. One is the analyzed. The results described above permitted alignment bond that is cleaved on activation and the other is the of all the peptides, and the complete amino acid sequence of Arg(87)-Gln(88) bond (solid arrowhead in Fig. 3). The S-carboxymethylated kallikrein is shown in Fig. 1. sequence of kallikrein was identical with that of human N- Terminal Sequence of Human Urinary Kallikrein and pancreatic and kidney kallikreins predicted from the Prokallikrein•\When intact kallikrein and prokallikrein nucleotide sequences of cDNAs (7, 8). The amino acid were subjected to automated sequence analysis, the se residues of the charge relay system were present at quences of the first 25 and 30 residues, respectively, were His(41), Asp(96), and Ser(190) (solid circles in Fig. 3). obtained, except that cycle 7 of kallikrein and cycle 14 of Therefore, the present results indicate strongly that both prokallikrein were empty (16). Therefore, prokallikrein human urinary kallikrein and prokallikrein are of tissue and kallikrein were S-carboxymethylated, and the N- type, and are excreted into the urine from some tissue as terminal sequence was analyzed. As shown in Fig. 2, the native form and a single-chain polypeptide. N-terminal sequences consisting of 37 and 28 amino acid residues of kallikrein and prokallikrein were determined, DISCUSSION respectively. In the present analysis, cycle 7 and 14 in Edman degradation of kallikrein and prokallikrein were We have determined the complete amino acid sequence of identified as S-carboxymethylcysteine, and it was con- human urinary prokallikrein (Figs. 1 and 3). A large part of firmed again that the sequence of the propeptide was the sequence could be determined by analyzing lysyl (NH2)-Ala-Pro-Pro-Ile-Gln-Ser-Arg-(COOH) as reported endopeptidase-digested peptides and their trypsin or chy already (16). motrypsin digests. Cyanogen bromide digestion was ef On the Primary Structure of Human Urinary Pro fective to supplement and connect the sequences of pep- kallikrein•\From the sequence analyses of kallikrein and a tides that could not be determined from lysyl endopep comparison of the N-terminal sequence of kallikrein with tidase digestion. The overlapping peptides used to establish that of prokallikrein, we proposed the complete amino acid the sequence are indicated in Fig. 1. With the overlapping sequence of human urinary prokallikrein, which is shown in peptides, two or more residues were overlapped between Fig. 3. Prokallikrein contained 245 amino acid residues. On peptides from lysyl endopeptidase digestion and cyanogen the basis of our previous and present results (19), it has bromide cleavage. However, positions 154 and 229 were been shown that the bond between Arg (-1) and Ile (1) is overlapped by only one residue. Placement of position 154

J. Biochem. Human Urinary Prokallikrein; Complete Amino Acid Sequence 27

was made by overlapping L9-C2 and CB3-C3. Namely, ogy to mouse ƒÁ-NGF and EGF-binding protein is 61 and alignment of (L9)-(L3)-(L1)-(L7) was established by over- 57%, respectively. As already pointed out (7), His(41), lapping each sequence with those of CB3-C2 and CB3-C3. Asp(96), and Ser(190), the active site, and Asp(184), which Placement of position 229 was made by overlapping the is thought to be the substrate , exist in human sequences of L5 and L6 with that of CB6-C2 and by urinary kallikrein also. These important residues are well establishing alignment of (L6)-(L5) using the sequences of conserved with all the proteins listed in Fig. 4 C136-C1, CB6-C2, and CB6-C3. Human urine contains both kallikrein and prokallikrein Three asparagine residues having a characteristic se in the ratio of about 2 : 3 (17). However, the significance of quence, Asn-X-Thr(Ser), as a potential glycosylation site these proteins in human urine is not understood and the (22) were identified at positions 78, 84, and 141 (boxes in physiological activator of prokallikrein is also unknown. To Fig 3). In the case of hog, rat, and mouse kallikreins, solve these problems, further detailed studies are neces however, only one possible glycosylation site per mol of sary. kallikrein has been found (Fig. 4). The present work has clarified in detail the properties of REFERENCES human urinary prokallikrein and kallikrein. It has been reported that the enzymological properties, isoelectric 1. Erfors, T.O., Riekkinen, P.J., Malmiharju, T., & Hopsu-Havu, V.K. (1967) Hoppe-Seyler's Z. Physiol. Chem. 348, 111-118 point, and molecular weight of kallikrein are very similar to 2. Habermann, E. (1962) Hoppe-Seyler's Z. Physiol. Chem. 328, those of the trypsin-activated prokallikrein (18), and the 15-23 immunochemical properties of the two are also indistin 3. Dietl, T., Kruch, J., & Fritz, H. (1978) Hoppe-Seyler's Z. Physiol. guishable (28). Moreover, prokallikrein was activated by Chem. 359, 499-505 4. Hojima, Y., Maranda, B., Moriwaki, C., & Schachter, M. (1977) trypsin very rapidly by releasing a small peptide having the J. Physiol. 268, 793-801 sequence Ala-Pro-Pro-Ile-Gln-Ser-Arg, maintaining the 5. Richards, R.I., Catanzaro, D.F., Mason, A.J., Morris, B.J., molecular weight almost unchanged (19). These results are Baxter, J.D., & Shine, J. (1982) J.Biol. Chem. 257, 2758-2761 explained by considering that trypsin hydrolyzed the 6. Swift, G.H., Dagorn, J.-C., Ashley, P.L., Cummings, S.W., & MacDonald, R.J. (1982) Proc. Natl. Acad. Sci. U.S. 79, 7263- Arg(-1).IIe(1) bond selectively to produce the active form 7267 of kallikrein. The results in Table I and Fig. 2 also support 7. Fukushima, D., Kitamura, N., & Nakanishi, S. (1985) Biochemis this mechanism. In this connection, it has been observed try 24, 8037-8043 that two-chain kallikrein was gradually produced on pro- 8. Baker, A.R. & Shine, J,, (1985) DNA 4, 445-450 9. Mason, A.J., Evans, B.A., Cox, D.R., Shine, J., & Richards, R.I. longed reactiiiiionof human urinary prokallikrein with tryp (1983) Nature 303, 300-307 sin, and the N-terminal sequence of one of the two polypep 10. Carrtero, O.A. & Scicli, A.G. (1980) Am. J. Physiol. 238, F247- tides was determineded edas Gln-Ala-Asp-Glu-Asp-Tyr-Ser- F255 His-Asp-Leu (19). This sequence was identical with that 11. Margolius, H.S., Geller, R., Pisano, J.J., & Sjoerdsma, A. (1971) Lancet 11, 1063-1065 from Gln(88) to Leu(97) in the sequence shown in Figs. 1 12. Lieberthal, W., Arbeit, L., Oza, N.B., Bernard, D.B., & Levins and 3. Therefore, the second, weakly trypsin-susceptible ky, N.G. (1983) Hypertension 5, 603-609 site is the Arg(87)-Gln(88) bond, and two-chain kallikrein is 13. Vinci, J.M., Gill, J.R., Jr., Bowden, R.E., Pisano, J.J., Izzo, J.L., Jr., Radfar, N., Taylor, A.A., Zusman, R.M., Bartter, F.C., & produced by the cleavage of this bond. This bond may be Keiser, H.R. (1978) J. Clin. Invest. 61, 1671-1682 located at the surface of the kallikrein molecule, because 14. ole-MoiYoi, 0., Spragg, J., & Austen, K.F. (1979) Proc. Natl. the sequences of human urinary kallikrein and porcine Acad. Sci. U.S. 76, 3121-3125 pancreatic kallikrein are highly homologous, and when the 15. Lottspeich, F., Geiger, R., Henschen, A., & Kutzbach, C. (1979) Hoppe-Seyler's Z. Physiol. Chem. 360,1947-1950 sequences between the third and fourth cysteinyl residues 16. Irie, A., Kushiro, H., Kodama, J., Ohota, M., & Miyake, Y. of the two kallikreins were overlapped, the highest se (1982) in Recent Progress on Kinins (Fritz, H., Dietze, G., quence homology was obtained by shifting Ala(81) of Fiedler, F., & Haberland, G. L., eds.) pp. 131-136, Birkhauser porcine kallikrein to the position of Ala(89) of human Verlag, Basel kallikrein and leaving eight residues at this portion of 17. Irie, A., Katayama, Y., Ito, K., Takahashi, S., & Miyake, Y. (1985) Jpn. J. Clin. Chem. 14, 27-33 porcine kallikrein open (data not shown). Moreover, this 18. Irie, A., Takahashi, S., Katayama, Y., Ito, K., & Miyake, Y. disrupted portion of two-chain porcine kallikrein lies at the (1986) Biochem. Int. 13,375-382 surface of the molecule (27). 19. Takahashi, S., Irie, A., Katayama, Y., Ito, K., & Miyake,Y. (1987) Biochem. Int. 14,467-474 A comparison of the amino acid sequence of human 20. Takahashi, S., Irie, A., Katayama, Y., Ito, K., & Miyake,Y. urinary kallikrein in Figs. 1 and 3 with those reported for (1986) J. Biochem. 99, 989-992 kallikrein of various animals and kallikrein-related pro 21. Bidlingmeyer, B.A., Cohen, S.A., & Tarvin, T.L. (1984) J. teins is shown in Fig. 4. The amino acid sequence of human Chromatogr. 336, 93-104 22. Bause, E. (1983) Biochem. J. 209, 331 336 urinary kallikrein was identical with that of human pancre 23. Fiedler, F., Fink, E., Tschesche, H., & Fritz, H. (1981) Methods atic and kidney kallikreins predicted from the nucleotide Enzymol. 80, 493-532 sequence of cDNAs (7, 8), suggesting that urinary and 24. van Leeuwen, B.H., Evans, B.A., Tregear, G.W., & Richards, R. tissue kallikreins are the same gene products. The sequence 1. (1986) J. Biol. Chem. 261, 5529-5535 25. Evans, B. A. & Richards, R.I. (1985) EMBO J. 4, 133-138 of human urinary kallikrein was also highly homologous to 26. Lundgren, S., Ronne, H., Rask, L., & Peterson, P.A. (1984) J. those of kallikrein in other mammalian species, such as hog, Biol. Chem. 259,7780-7784 rat, and mouse, as well as those of mouse y-nerve growth 27. Bode, W., Chen, Z., Bartels, K., Kutzbach, C., & Schmidt- Kastner, G. (1983) J. Mol. Biol. 164, 237-282 factor (NGF) and epidermal growth factor (EGF) -binding 28. Irie, A., Takahashi, S., Katayama, Y., Shibata, Y., & Miyake, Y. protein. In fact, the sequence of the kidney enzyme is 65% (1988) Clin. Chim. Acta 173, 289-298 homologous to that of hog pancreatic kallikrein and 60- 63% homologous to those of rat and mouse kallikreins. Homol

Vol. 104, No. 1, 1988 28 S. Takahashi et al.

Supplemental Materials

Table is. Amino acid compositions of peptide fragments from kallikreina .

Fig. Is. Separation of lysyl endopeptidase digest of S-carboxymethylated

values a the numbers of amino acid residue per mot. and those to human urinary kallikrein by reversed-phase HPLC. Lysyl endopeptidase digest b parentheses identifiedare the as numbersS -carboxymethylcysteine from sequence analysis.b Cys was .c Not determined.d was injected onto a Pep RPC HR5/5 column and eluted with a linear gradient Position of peptide fragments in the sequence of kallikrein. of 0-50% acetonitrile. Table III, Amino acid compositions of peptide fragments from kallikreina.

a Valves a the Dumber. of amino acid residue per mol parenthesesare thenumbers from sequence analysis.b , and Cysthose wasin Notdetermined. identifiedas S-carboxymethylcysteine.c d Position of a tide fragments in the a quanta of ka11ikrein.

Table IIIs. Automated sequence analysis of peptide fragments from kallikrein. Fig. 2s. Separation of chymotryptic peptide from peptide L9 by reversed-

phase HPLC. Conditions for chromatography were the same as described in Fig.

I..

Fig. 3s. Separation of cryptic digest of peptide L10 by reversed-phase HPLC.

The digest was chromatographed on a Pep RPC HR5/5 column and eluted with a

linear gradient of 0-40% acetonitrile.

a Cm -Cys was eluted Sn HPLC between the PTH-derivatives of Ser and Gln. b Not determined .

Fig. 5s. Separation of ehymotryptic peptides from peptide C83 by reversed- Fig. 4s. Separation of cyanogen bromide-cleaved peptides by reversed-phase phase HPLC.The experimental conditions were the same as described in Fig. HPLC. Peptides were eluted with a linear gradient of 0-55% acetonitrile. Is. The sample (3nmol) was applied to a Pep RPCHR5/S column. J. Biochem. Human Urinary Prokallikrein; Complete Amino Acid Sequence 29

Table IVs. Automated sequence analysis of peptide fragments from kallikrein.

Fig. 6s. Separation of chymotryptic peptides from peptide CB5 by reversed-

phase HPLC. The chromatography was performed under the same conditions as

described in Pig. 1s. The sample (4nmol) was applied to a Pep RPC HR5/5

column.

L Cm-Cysdetermined waseluted. in HPLCbetween the PTN-derivativesof Set end Gln.b not

Table Vs. Automated sequence analysts of peptide fragments Eros kallikrein.

Fig. 7s. Separation of chymotryptic peptides from peptide CB6 by reversed

phase HPLC. Experimental conditions for the chromatography were the same as

described in Fig. 1s. The sample (4nmol) was applied to a Pep RPC HR5/5

column.

a Cm-Cys was eluted in HPLC between the PTH-derivatives of Ser and Gln.

Vol. 104, No. 1, 1988