J. Biochem. 110, 648-654 (1991)
Identification of Full-Sized Forms of Salivary (S-Type) Cystatins (Cystatin SN, Cystatin SA, Cystatin S, and Two Phosphorylated Forms of Cystatin S) in Human Whole Saliva and Determination of Phosphorylation Sites of Cystatin S Satoko Isemura,* Eiichi Saitoh," Kazuo Sanada,** and Kayoko Minakata'** 'Nippon Dental University Junior College at Niigata, Niigata, NOgata 951;**Department of Oral Biochemistry, School of Dentistry at Niigata, Nippon Dental University, Niigata, Niigata 951; and '"Department of Legal Medicine, Hamamatsu University School of Medicine, Hamamatsu, Shizuoka 431-31 Received for publication, March 22, 1991
Our recent work on the gene structures for human salivary (S-type) cystatins [Saitoh, E. et al. (1987) Gene 61, 329-338] has suggested that the structures of cystatins which we determined previously at the protein level lack N-terminal peptide portions of the fullsized intact forms. In the present study, attempts were made to isolate full-sized S-type cystatins by introducing methanol fractionation into the purification steps to suppress the enzymatic activity present in saliva. Full-sized cystatin SN and two phosphorylated forms of full-sized cystatin S were thus isolated. Analysis of one fraction indicated that this was a mixture of full-sized cystatin SA and non-phosphorylated cystatin S. The phosphorylation sites of cystatin S were determined to be Ser-Ser-Ser'(P)-Lys-Glu-Glu- for monophosphorylated cystatin S and Ser'(P)-Ser-Ser3(P)-Lys-Glu-Glu- for diphosphorylated cystatin S. Immunoblotting analysis with anti-cystatin S antiserum revealed that tears and seminal plasma also contained S-type cystatins, but diphosphorylated cystatin S was detected neither in tears nor in seminal plasma and no cystatin SN was found in seminal plasma. These data indicate that S-type cystatins are secreted into the oral cavity without significant degradation in salivary glands or ducts and that they are expressed tissue specifically.
Animal tissues and body fluids contain various kinds of endogenous proteinase inhibitors to regulate their protein metabolism or to protect tissues from attacks by bacteria or viruses (1-4). Cystatins are cysteine proteinase inhibitors belonging to the cystatin superfamily, supposedly evolved from the same ancestral gene. The cystatin superfamily is now grouped into three families (family 1, family 2, and family 3) according to the degree of structural homology (5). Family 2 cystatins are secretory inhibitors with two disulfide bonds and molecular weights of about 15,000 (5). We have isolated three cysteine proteinase inhibitors, cystatin S, cystatin SA, and cystatin SN, which belong to family 2 of the cystatin superfamily, from human whole saliva and determined their amino acid sequences (6-8). These three inhibitors exhibit about 90% sequence homology with one another. In addition to these S-type cystatins, human saliva contains cystatin C, which is also a family 2 cystatin member and has 50% homology with S-type cystatins (9). Human family 2 cystatin genes constitute a multigene family probably having seven members, six of which have been characterized: CST1 for cystatin SN, CST2 for cystatin SA, CST3 for cystatin C, CST4 for cystatin S and two pseudogenes, CSTP1, and CST5 (CSTP2) (10-12). A new cystatin corresponding to the yet unidentified seventh gene may be expressed somewhere in the human body. Abbreviations: TFA, trifluoroacetic acid ; N-, amino. 648
Electrophoresis of human saliva on basic gel consistently gives four bands immunoreactive with anti-cystatin S antiserum, suggesting the possibility of the presence of the S-type cystatin not characterized yet. Our recent results on nucleotide sequences of genes (2012) have indicated that the cystatins which we isolated and characterized previously (6-8) are not full-sized proteins, and suggest that they are probably formed by enzymatic degradation. This study was performed to see whether full-sized inhibitor proteins can be isolated from whole saliva by preventing enzymatic degradation, and also to find out whether there is a new S-type cystatin in human saliva or other body fluids. MATERIALS AND METHODS Materials—The following materials were purchased from the sources indicated: DE32, Whatman Separation; DEAE-Sepharose CL-6B and Sephacryl S-200, Pharmacia Fine Chemicals; Novapak Ci8, Waters Associates; lysylendopeptidase and malachite green oxalate, Wako Pure Chemical; TPCK-trypsin [EC 3.4.21.4], Cooper Biochemical; alkaline phosphatase [EC 3.1.3.1], Miles Scientific; Immun-blot assay kit and nitrocellulose membrane, BioRad Laboratories; Centriprep 10, Amicon. Rabbit polyclonal anti-cystatin S antiserum was prepared as described previously (13). J. Biochem.
649
S-Type Cystatins in Saliva Collection and Handling of Saliva—Human whole saliva was collected from students into ice cooled test tubes containing 0.5 ml of proteinase inhibitor solution (500 fig p-amidinophenylmethanesulfonyl fluoride, 25 fig phosphoramidon, 500 fig bestatin, 500 units Trasylol, 1 mg NaF, 10 mg EDTA 2Na in 0.5 ml of 0.1 M Tris-HCl, pH 7.5). Saliva amounting to 10 ml per person was collected and pooled. Four volumes of cold methanol were added to the pooled saliva. Precipitates formed by the addition of methanol were removed by filtration. Methanol was removed from the filtrate by evaporation under reduced pressure at 40*C, and the residue was lyophilized. The lyophilized material was dissolved in water and dialyzed against 0.02 M TrisHCl buffer (pH 7.5) containing 5% inhibitor solution. After removing insoluble materials by centrifugation, the supernatant was applied to a DE32 column equilibrated with the same buffer. After washing, elution with a gradient formed from 100 ml each of the initial buffer and 3% NaCl in the same buffer was performed (Fig. 1). Fractions eluted from the column were monitored by immunoblotting after polyacrylamide gel electrophoresis at pH 9.4. Samples for Electrophoresis—Submaxillary-sublingual saliva was sampled from the outlet of the submaxillarysublingual duct by using a micropipet. Parotid saliva was collected into an ice-chilled test tube by using a suction cup under sour candy stimulation. Tears were sampled with a micropipet from the outlet of the lachrymal gland. Seminal plasma was obtained as described previously (14). Ten microliters each of undiluted submaxillary-sublingual saliva, tears, and seminal plasma was used for each run of gel electrophoresis. Twenty microliters of parotid saliva was concentrated to 10//I by lyophilization and used for electrophoresis. Samples other than seminal plasma were analyzed immediately after sampling. Polyacrylamide Gel Electrophoresis and Western Blotting—Polyacrylamide gel electrophoresis was performed by the method of Davis (25) at pH 9.4 with 7% slab gel. Gels were stained with Coomassie Brilliant Blue G-250 or transblotted to nitrocellulose membrane by using a transblotting apparatus (Atto) with 0.7% acetic acid. Immunostaining was performed using an Immun-blot assay kit according to the manual issued by Bio-Rad. Anti-cystatin S antiserum was used as the first antibody. Amino Acid Analysis—Proteins or peptides were hydrolyzed in constant boiling hydrochloric acid at 108"C for 24 h in a Pico-tag work station (Waters) and analyzed by the Waters amino acid analysis system using the orthophthalic acid post column labeling method. Amino Acid Sequence Analysis—Amino acid sequence analysis was performed by the direct manual Edman degradation method (16) or by using a gas phase automatic amino acid sequence analyzer (Applied Biosystems). Phosphate Measurement—An aliquot was taken from each fraction eluted from the column, and the volume was adjusted to 390//I by adding 0.01 M NrL,HCO3. Alkaline phosphatase (0.2 unit) in 10//I of 0.01 M N^HCO-j was added. The mixture was incubated for 1 h at 37'C. Color was developed by adding 25 ft 1 of 7.5 M sulfuric acid and 75 //\ of a molybdate-malachite green reagent (17). Absorbance at 650 mn was measured. For quantitative analysis of phosphate, samples hydrolyzed in 6 N HC1 for 24 h at 100'C were used. Enzymatic Digestion—Purified proteins (10-100 nmol) Vol. 110, No. 4, 1991
were reduced and carboxymethylated, and digested with 0.025 unit of lysylendopeptidase in 0.5ml of 0.01M Tris-HCl buffer (pH 8.5) for 2 h at 25'C. Trypsin digestion of the purified lysylendopeptidase peptide was performed in 0.01 M NR,HC0 3 for 16 h at 25#C. Phosphorylation Sites Determination—The direct Edman degradation method was slightly modified to apply it to the determination of phosphorylation sites of phosphopeptides. Each of the phosphopeptides (2 nmol) was dissolved in 30 //I of dimethylallylamine buffer containing 1.5 //I of phenylisothiocyanate. After 1 h reaction at 40'C, the reaction mixture containing the phenylthiocarbamyl peptide was dried up without extraction with benzene. Ten microliters of TFA was added to the dried sample and incubated at 40"C for 20 min to release the phenylthiazolinone derivative of an amino acid. TFA was removed from the reaction mixture by a nitrogen gas stream. Inorganic phosphate released during the TFA-treatment, when a peptide had carried an N-terminal phosphoserine residue (18), was dissolved in 0.4 ml of 0.01 M NH,HC0 3 . The color was developed by adding 25 ft\ of 7.5 M sulfuric acid and 75 //I of a molybdate-malachite green reagent (17). RESULTS Immunoblotting Analysis of Human Body Fluids—Figure 2 shows immunoblotting analysis of human submaxillary-sublingual saliva, parotid saliva, tears, and seminal plasma after electrophoresis on anionic gel. Human serum gave no positive bands. Submaxillary-sublingual saliva, parotid saliva, and tears from four individuals were tested. In all submaxillary-sublingual saliva samples tested, four bands (1, 2, 3, and 4) were detected and the intensity of band 2 relative to band 1 or band 3 was low. In parotid saliva, four bands were also detected, but the intensity of band 2 was the highest in all samples. Band 4 was not detected in any sample of tears tested. Neither band 1 nor band 4 was detected in seminal plasma. Isolation of the Protein Corresponding to Band 1—The fractions not adsorbed on a DE32 column (fraction A in Fig. 1) were lyophilized, dissolved in water, dialyzed against 0.01 M Tris-HCl buffer (pH8.5), and then applied to a DEAE-Sepharose CL-6B column equilibrated with the same buffer. The elution profile is shown in Fig. IS. The protein contained in the fractions indicated by the bar had the same electrophoretic mobility as band 1 of fresh saliva. Isolation of the Protein Corresponding to Band 2—Fraction B in Fig. 1 was lyophilized, then dissolved in water, dialyzed against 0.05 M Tris-HCl buffer (pH8.0) and applied to a column of DEAE-Sepharose CL-6B equilibrated with the same buffer containing 0.075 M NaCl. The elution profile is shown in Fig. 2S. Fraction BI in Fig. 2S contained proteins giving two bands electrophoretically: one corresponding to band 2 and reactive with anti-cystatin S antiserum, and the other having slightly lower mobility and unreactive with anti-cystatin S antiserum. This fraction was then concentrated in a Centriprep 10, and insoluble materials formed during concentration were removed by centrifugation. The concentrate was applied to a column of Sephadex G-50 (IX90cm) (Fig. 3S). The electrophoretically homogeneous protein corresponding to band 2 was obtained from fractions indicated by the bar. Isolation of the Protein Corresponding to Band 3—The
S. Isemura et al. protein in fraction BII in Fig. 2S was homogeneous electrophoretically and gave the band corresponding to band 3. Salts were removed by gel filtration through a column of Sephadex G-50 (1x90 cm). Isolation of the Protein Corresponding to Band 4—Although the protein in fraction BUI in Fig. 2S was homogeneous electrophoretically, its amino acid analysis suggested that this fraction contained a protein mixture having
the same electrophoretic mobility as band 4. This fraction was concentrated and applied to a column of Sephacryl S-200 (Fig. 4S). The second peak in Fig. 4S contained the protein immunoreactive with anti-cystatin S antiserum and having the same electrophoretic mobility as band 4. N'-Terminal Amino Acid Sequence of the Purified Proteins—N-Terminal amino acid sequences of the purified proteins were as follows. Band 1 protein:
Band 3 protein: Band 4 protein:
20
40
Fraction
60
80
Number
Fig. 1. Separation of methanol-treated whole saliva on a DE 32 column. The solid material (1.15 g) from 500 ml of whole saliva was applied to the column (1.7 x 15 cm) equilibrated with 0.02 M TrisHC1 (pH 7.5). After washing with equilibration buffer, elution was performed with a gradient formed from 100 ml each of initial bufFer and 3% NaCl in the same buffer, and 5 ml fractions were collected. The arrow indicates the position where gradient elution was started. The protein corresponding to band 1 was detected in fraction A, and the proteins corresponding to band 2, band 3, and band 4, in fraction B.
2
3 4
2 34 e f
2 34 Fig. 2. Immunoblotting analysis of human body fluids. Electrophoresis was performed at pH 9.4. Lane a, tears; lane c, seminal plasma; lane e, parotid saliva; lanes b, d, and f, submaxillary-sublingual saliva. The arrow indicates the direction of migration.
Trp-Ser-Pro-Lys-Glu-Glu-Asp-ArgIle-Ile-Pro-Gly-Gly-Ile-Tyr-Asn-AlaAsp-Leu-AsnSer-Ser-Ser-Lys-Glu-Glu-Asn-ArgIle-Ile-Pro-Gly-Gly-Ile-Tyr-AspSer-Ser-Ser-Lys-Glu-Glu-Asn-ArgIle-Ile-Pro-Gly-Gly-Ile-Tyr-Asp-
In the case of band 2 protein, results obtained by using an automated amino acid sequence analyzer indicated that the preparation contained two proteins in a ratio of about 7 : 3. The N-terminal sequence of the major protein was: TrpSer-Pro-Gln-Glu-Glu-Asp-Arg-Lle-Ile-Glu-Gly-Gly-Ile-TyrAsp-Ala-Asp-Leu-Asn-. The N-tenninal sequence of the minor one was identical with those of band 3 and band 4 proteins. When these sequences were compared with the amino acid sequences predicted from the gene structures {10, 12), it was found that band 1 and band 2 major proteins had the N-terminal amino acid sequences of full-sized cystatin SN and full-sized cystatin SA, respectively. The N-terminal sequences of band 3, band 4, and band 2 minor proteins were all identical with the N-terminal sequence of fullsized cystatin S. Amino Acid Composition and Phosphate Contents of the Purified Proteins—The amino acid composition of band 1 protein coincided with that of full-sized cystatin SN within the range of error. The amino acid compositions of both band 3 and band 4 proteins corresponded well with that of full-sized cystatin S. The amino acid composition of the fraction corresponding to band 2 was compatible with the composition, if we assume that the fraction contained full-sized cystatin SA and cystatin S in an approximate ratio of 7 : 3. Their amino acid compositions are shown in
TABLE I. Phosphate content of purified proteins. Phosphate content (mol/mol protein) Band 1 protein Band 2 protein 1.32 Band 3 protein Band 4 protein 2.35
Fig. 3. Effect of alkaline phosphatase digestion on salivary cystatins. Whole saliva (1 ml) was digested with 0.2 unit of alkaline phoephatase for 1 h at 37"C. Electrophoresia and immunoblotting were performed as 2 34 described in the experimental section. Samples of 10//I each were analyzed. Lane a, whole saliva digested with alkaline phosphatase; lane b, whole saliva (control). The arrow indicates the direction of migration. J. Biochem.
S-Type Cystatins in Saliva
651 TABLE II. Axnino acid compositions and phosphate contents of lysylendopeptidase peptides derived from band 3 and band 4 proteins (mol/mol peptide). Ser Ly3 Phosphate 2.33 1.00 0.65 nP 2.86 1.00 0.58 nP 2.95 1.00 2.07
A 650 A
280
0.15
0.
TABLE III. Phosphate released by Edman degradation of phosphopeptides. The values indicate the ratio in percentage of phosphate released by Edman degradation to the phosphate content of the peptide used for Edman degradation. Cycle of degradation Peptide subjected to 3 Edman degradation* 1 2 nP 39.2 1.8 0.3 52.2 9.5 mp 82.8 'Edman degradation was performed for the peptide purified after every cycle of degradation.
10
0.05
Fraction Number Fig. 4. Comparison of lysylendopeptidase digests of bands 3 and 4 proteins. The digests were fractionated on a column of DEAESepharose CL-6B (0.9x25 cm) equilibrated with 0.01 M NrLHCO, and elution was performed with a gradient formed from 100 ml each of 0.01 M NH.HCO, and 1 M NH,HC03 . Fractions of 2 ml each were collected. Solid and dotted lines indicate the absorbance values at 230 and 650 nm, respectively, (a) Chromatography of band 3 protein digests, (b) Chromatography of band 4 protein digests. Peptides (P)
assigned to each peak are: 1, 1', K; 2, 2', SSSK, ATEDEYYRRPLQVLRAREQTFGGVNYFFDVEVGRTICTK; 3, SQPNLDTCAFHEQPELQK; 3', SQPNLDTCAFHEQPELQK, SSSK; 4,4', EENRIIPGGIYDADLNDEWVQRALHFAISEYNK; 5,5', QLCSFEIYEVPWEDRMSLVNSRCQEA
Table IS. Phosphate contents of the purified protein fractions are listed in Table I. The data suggested that band 3 and 4 proteins were phosphorylated. Alkaline Phosphatase Digestion of Whole Saliva—Human whole saliva was digested with alkaline phosphatase and analyzed by electrophoresis and immunoblotting (Fig. 3). Band 3 and band 4 disappeared, and band 2 was intensified. This result indicated that both band 3 and band 4 represented phosphorylated proteins and that they migrated to the position of band 2 upon alkaline phosphatase digestion. Based on these results, band 3 and band 4 proteins were Vol. 110, No. 4, 1991
tentatively assigned as mono- and diphosphorylated fullsized cystatin S, respectively. Full-sized cystatin SN (band 1 protein) and cystatin SA (band 2 major protein) appeared not to be phosphorylated. Comparison of Peptide Maps of Band 3 and Band 4 Proteins—In order to confirm that band 3 and band 4 proteins are the same proteins with different degrees of phosphorylation, lysylendopeptidase digests of the two proteins were compared. Each protein was digested with lysylendopeptidase after reduction and carboxymethylation, and the digests were separated on a column of DEAESepharose CL-6B. Elution profiles of band 3 and band 4 protein digests are shown in Fig. 4, a and b, respectively. Elution profiles monitored by measuring the absorbance at 230 nm of the two chromatograms showed no apparent difference, but the profiles monitored in terms of phosphate content were clearly distinguishable. The peptides assigned to each peak are shown in the legend to Fig. 4. These assignments were based on the amino acid composition and N-terminal sequence analysis of each peak, and HPLC of peptides derived by trypsin digestion of each peak (not shown). As HPLC reference peptides, we used tryptic peptides prepared from sequenced cystatin S (6). There has been a discrepancy between the sequences of cystatin S determined by protein chemistry and those predicted from the gene structure. The 107th residue (115th in the fullsized sequence) in cystatin S is Asp (6), while it is Asn from the gene structure of CST4 [12). Sequence analysis of peak 5 and peak 5' indicated that both band 3 and band 4 proteins had Asn at this position. Therefore, we correct here the previously determined sequence of cystatin S (6) by placing Asn at the 107th residue. The phosphate peaks designated I and I' in Fig. 4, a and b, were considered to represent inorganic phosphate since these phosphate peaks could be detected even without alkaline phosphatase digestion. In a control run, inorganic phosphate was confirmed to be eluted at this position. The phosphate-containing fractions designated II, II', and EH' were separated on a column of Novapak Cis with a gradient of 0-60% acetonitrile in 0.1% TFA. In each case, phosphate was detected in the fraction not adsorbed on the column. These unadsorbed fractions obtained from II, II', and HI' were designated as lip, lip', and Hip', respectively. Amino acid analysis data
652
S. Isemura et al.
listed in Table IS suggest that these three peptides are practically pure and are derived from the amino terminal portion of full sized cystatin S. Phosphate contents relative to amino acid components of lip, lip', and Hip' are listed in Table II. On the basis of the results in Table II, lip, and lip' are considered to correspond to the monophosphorylated N-terminal tetrapeptide of full-sized cystatin S, and Dip' to the diphosphorylated N-terminal tetrapeptide of full-sized cystatin S. The appearance of a high inorganic phosphate peak (I') in Fig. 4b may be due to partial dephosphorylation of the purified band 4 protein during enzyme digestion. Phosphorylotion Sites of lip and nip'—In the case of Edman degradation of phosphopeptide, inorganic phosphate has been shown to be released during TFA treatment when a phosphoserine residue was located at the Nterminus (28). Therefore we applied this method to the determination of phosphorylation sites of phosphopeptides isolated in this study. lip and Hip' were subjected to Edman degradation. In the case of Kip', the peptide was purified after each cycle of degradation by a DEAE-Sepharose CL-6B column using a gradient from 0.01 to 1 M NK,HC0 3 , while in the case of lip, peptide purification was not performed during 3 cycles of degradation. Table III lists the ratio of phosphate released after each cycle of degradation to the phosphate content of the original peptide as a percentage. The results indicated that the structures of Up and Hip' were Ser-SerSer(P)-Lys and Ser(P)-Ser-Ser(P)-Lys, respectively. lip' in Pig. 4b was considered to be identical with lip, since their amino acid compositions, ratio of phosphorylation and chromatographic behavior were the same. lip' was probably formed artificially by dephosphorylation of purified band 4 protein during handling These data lead us to conclude that the protein corresponding to band 3 was full-sized cystatin S monophosphorylated at the 3rd serine residue and that the protein corresponding to band 4 was full-sized cystatin S diphosphorylated at the 1st and the 3rd serine residues. DISCUSSION In the early stage of our study on cyatatins, we isolated several forms of S-type cystatins (29). The occurrence of multimolecular forms could be explained by the presence of proteins having closely related but distinct amino acid sequences such as cystatin S, cystatin SN, and cystatin SA, and their degradation products. Even when conditions were used under which enzymatic degradation during purification was largely suppressed, there was still a possibility that degradation was not prevented completely, since cystatin SN and cystatin S A obtained by this method lacked the N-terminal 3 and 4 residues of the full-sized proteins predicted from the gene structures, respectively (20). In the present study, we introduced the addition of methanol to the concentration of 80%. This resulted in precipitation of most of the proteins in saliva, but significant amounts of cystatins could be recovered from the supernatant. Although the yield of cystatins (4.4 mg/liter for monophosphorylated cystatin S, 0.98 mg/liter for diphosphorylated cystatin S, 3.35 mg/liter for cystatin SN) was a little less as compared with the previous method, the present method provided full-sized forms of S-type cystatins. Inclusion of methanol precipitation appeared to be very effective to
TABLE IV. The structures around phosphoserine residues of salivary proteins. Phosphoprotein Structure Reference Human cystatin S (monophosphorylated) Ser-Ser-Ser(P)-Lys-GIu-GluThis study (diphosphorylated) Ser(P)-Ser-Ser(P)-Lys-Glu-Glu This study Human statherin Asp-Ser(P)-Ser(P)-Glu-Glu-Lys22 Macaca fascicularis Asp-Ser(P)-Ser(P)-Glu-Glu-Lys23 statherin Human histidine-rich Asp-Ser(P)-Hi8-Glu-Lys-Arg24 polypeptide Human proUne-rich -Val-Ser(P)-Gln-Glu-Asp-Val25 phosphoprotein
suppress proteolytic enzyme activities. Fresh whole saliva consistently gave four bands on immunoblotting after anionic gel electrophoresis, and the four proteins corresponding to these bands were studied in this work. The fact that full-sized forms of cystatins were isolated indicates that S-type cystatins are secreted into the oral cavity without suffering significant degradation in salivary glands or in ducts. Previous studies (19, 20) have shown that enzymes present in saliva can cleave cystatins at their N-terminal portions. Enzymes responsible for the cleavage are now under investigation. In our previous method, only cystatin S was isolated as the full-sized form (20). It is possible that this is because enzymes cleaving cystatin S are less active under the previous purification conditions, where no methanol fractionation was included, as compared with those cleaving cystatin SN and cystatin SA. Alternatively, a subtle structural difference may explain the difference in susceptibility, if we assume that the same enzyme(s) is responsible for degradation of all S-type cystatins. Two phosphorylated forms of cystatin S were found to be present in fresh saliva. It is not clear at the moment whether monophosphorylated cystatin S is formed by differential phosphorylation of cystatin S or by phosphatase digestion of the diphosphorylated form. Mammary gland casein kinase and casein kinase II are proposed to recognize Ser*-X-Glu/Ser(P) and Ser*-X-XGlu/Ser(P), respectively, where Ser' indicates potential acceptor serine and X can be any amino acid residue (21). Since the phosphoserine residue of all phosphorylation sites of cystatin S, Le., the third residue of monophosphorylated cystatin S and the first and third residues of diphosphorylated cystatin S, have glutamic acid or phosphoserine at two residues to the right of them, all phosphorylation sites of cystatin S are in agreement with the recognition sequence of mammary gland casein kinase. Phosphoserine of the third residue of mono- or diphosphorylated cystatin S has glutamic acids at two and three residues toward the right. Therefore, this site of phosphorylation is in agreement with not only the proposed recognition sequence of the mammary gland casein kinase but also that of casein kinase II. Phosphorylation sites which are in agreement with the recognition sites of both casein kinase II and mammary gland casein kinase have been found in salivary proteins such as human statherin (22) and Macaca fascicularis statherin (23). Phosphorylation sites only compatible with the recognition sequence of mammary J. Biochem.
S-Type Cystatins in Saliva gland casein kinase are present in-salivary proteins such as human and M. fascicularis statherin (22, 23), human hi8tidine-rich polypeptide (24) and human proline-rich protein (25). Sequences around the phosphorylation sites of salivary proteins are shown in Table IV. Since removal of phosphate by alkaline phosphatase had no effect on inhibitory activity for papain (unpublished data), phosphorylation on serine residues does not appear to be involved in the inhibitor activity of cystatin S. Chicken egg white cystatin, a family 2 cystatin member, is also partly phosphorylated, although the phosphorylation site (Ser-80) is different from that of cystatin S's (26). Hawke et al. isolated full-sized cystatin S (27) and full-sized cystatin SN (28) from human whole saliva by TFA fractionation and subsequent separation by HPLC. However, they only described one component of cystatin S and gave no information about phosphorylation. Al-Hashimi et al. isolated full-sized cystatin SN from human submaxillary-sublingual saliva without using methanol or TFA fractionation (29). Therefore, enzymes responsible for cleavage of cystatin SN may be derived from salivas other than submaxillary-sublingual saliva or from microbes. There has been no report on whether the full-sized form of cystatin SA is present in whole saliva. Therefore, this is the first report to describe the existence of the full-sized protein which has been predicted from the gene structure analysis (10). The S-type cystatins are tissue specifically expressed as shown in Fig. 2. Salivary glands may be different from lachrymal gland or seminal glands as regards the phosphorylation mechanism, since diphospho-cystatin S is detected only in saliva but not in tears or in seminal plasma. The intense band at the position of cystatin SA (band 2) in parotid saliva, tears and seminal plasma could be nonphosphorylated cystatin S, since non-phosphorylated cystatin S has the same electrophoretic mobility as that of full-sized cystatin SA. Differential expression by tissues may reflect differences in the physiological significance among the inhibitors of this type. The four proteins isolated in this study are encoded by three genes, CST1, CST2, and CST4. We could not detect S-type cystatin coded by an uncharacterized new gene (10). Whether there is an S-type gene product derived from the new gene is not clear at present, although there is a possibility that the weak band detected on immunoblotting (Fig. 2) could represent a new cystatin. REFERENCES 1. Katunuma, N. (1989) in Intraceilular Proteolysis (Katunuma, N. & Kominarai, E. eds.) pp. 3-23, Jpn. Sri. Soc. Press, Tokyo 2. Bjoi-ck, L., Akesson, P., Bohus, M., Ttojnar, J., Abrahamson, M.,
Vol. 110, No. 4, 1991
653 Olafuson, I., & Grubb, A. (1989) Nature 337, 385-386 3. Korant, B.O., Brain, J., & Turk, V. (1985) Biochem. Biophys. Res. Common. 127, 1072-1076 4. Bjflrck, L-, Grubb, A., & Kjellen, L. (1990) J. Virol. 64, 941-943 5. Barrett, A.J., Fritz, H., Grubb, A., Isemura, S., Jflrvinen, M., Katunuma, N., Machleidt, W., Mttller-Esterl, W., Sasaki, M., & Turk, V. (1986) Biochem. J. 236, 312 6. Isemura, S., Saitoh, E., & Sanada, K. (1984) J. Biochem. 96, 489-498 7. Isemura, S., Saitoh, E., & Sanada, K. (1986) FEBS Lett. 198, 145-149 8. Isemura, S., Saitoh, E., & Sanada, K. (1987) J. Biochem. 102, 693-704 9. Abrahamson, M., Barrett, A.J., Salvesen, G., & Grubb, A. (1986) J. BioL Chem. 261, 11282-11289 10. Saitoh, E., Kim, H.S., Smithies, O., & Maeda, N. (1987) Gene 61, 329-338 11. Saitoh, E., Sabatini, L.M., Eddy, R.L., Shows, T.B., Azen, E.A., Isemura, S., & Sanada, K. (1989) Biochem. Biophys. Res. Commun. 162, 1324-1331 12. Saitoh, E., Isemura, S., Sanada, K., & Ohnishi, K. (1991) Biomed. Biochim. Ada 60, 599-605 13. Isemura, S., Saitoh, E., Ito, S., Isemura, M., & Sanada, K. (1984) J. Biochem. 96, 1311-1314 14. Minakata, K. &Asano, M. (1985) BioL Chem. Hoppe-Seyler 366, 15-18 15. Davis, B.J. (1964) Ann. N.Y. Acad. Sci. 121, 404-427 16. Iwanaga, S., Wallen, P., Groendahl, N.J., Henschen, A., & Blombfick, B. (1969) Eur. J. Biochem. 8, 189-199 17. Matsubara, C, Yamamoto, Y., & Takamura, K. (1987) Analyst 112, 1257-1260 18. Proud, C.G., Rylatt, D.B., Yeaman, S.J., & Cohen, P. (1977) FEBS Lett. 80, 435-440 19. Isemura, S., Saitoh, E., Sanada, K., Isemura, M., &Ito, S. (1986) in Cysteine Proteinases and Their Inhibitors (Turk, V., ed.) pp. 497-505, Walter de Gruyter, Berlin, New York 20. Saitoh, E., Isemura, S., Sanada, K., Kim, H.-S., Smithies, O., & Maeda, N. (1988) Biol. Chem. Hoppe-Seyler 369, Suppl., 191197 21. Kemp, B.E. & Pearson, R.B. (1990) Trends Biochem. Sci. 16, 342-346 22. Schlesinger, D.H. & Hay, D.I. (1977) J. BioL Chem. 262, 16891695 23. Oppenheim, F.G., Offner, G.D., & Troxler, R.F. (1982) J. Biol. Chem. 267, 9271-9281 24. Oppenheim, F.G., Yang, Y.-C, Diamond, R.D., Hyslop, D., Offner, G.D., & Troxler, R.F. (1986) J. BioL Chem. 261, 11771182 25. Wong, R.S.C. & Bennick, A. (1980) J. Biol. Chem. 265, 59435948 26. Laber, B., Krieglstein, K., Henschen, A., Kos, J., Turk, V., Huber, R., & Bode, W. (1989) FEBS Lett. 248, 162-168 27. Hawke, D.H., Yuan, P.M., Wilson, K.J., & Hunkapiller, M.W. (1987) Biochem. Biophys. Res. Commun. 145, 1248-1253 28. Hawke, D.H., Yuan, P.M., Ohno, J., Kikuchi, M., Hirayama, K., & Fukuyama, K. (1989) in Intraceilular Proteolysis (Katunuma, N. & Kominami, E., eds.) pp. 389-390, Jpn. Sri. Soc. Press, Tokyo 29. Al-Hashimi, I., Dickinson, D.P., & Levine, M.J. (1988) J. BioL Chem. 263, 9381-9387
654
S. Isemura et al. Supplemental Materials Ami no acid composition! of purified proteins and phosphopeptides, and ami.no acid composition* of full-sized cystatin S, cystatin SA and cyst*tin SN, predicted frog the gone structures. values axe shown in molar ratios. Band 1 protein
Asp Thr Ser Glu Pro Gly Ala l/2Cys
val Hat 11* L«u Tyr Pho Trp His
13.0 4.8 6.4 19.7 7.4 6.1 6.0 2.5 7.6 4.7 7.0 5.4 6.0 nd* 1.9 6.6 10.4
Lys Arg
Band 2 protein
13.0 4.8 6.3 22.1 5.6 4.9 7.2 3.1 6.9 0.8 6.0 B.8 5.1 4.9 nd 2.6 4.8 10.6
Band 3 protein
13.1 4.7 7.7 21.9 5.1 5.6 6.8 3.8 7.0 0.7 4.8 7.9 5.1 5.5 nd 1.7 5.1 8.4
Band 4 protein
15.1 4.6 7.2 23.0 6.6 4.9 6.9 2.2 6.9
Cystatin SN
Cystatin SA
13 5 7 19 6 5 6 4 8
13 4 6 23 4 5 6 4
Cystatin S
13 5 8 22 5 5 7 4 7 1 6 8 6 6 2 2 5 9
a. 8 5.6 9.1 5.0 5.9 nd 1.5 5.1 8.4
6 8 6 S 3 2 7 11
9 6 5 2 2 4 11
:ip
Up1
IIIp'
0.05
0.26
2. 86 0.14
2 .95 0 .27
0.5 -
0 .25 0 .IS
0.18 0.22
1.00
1.00
20
1.00
40
60
Fraction Nuiiber
nd f not determined.
Pig. 23 Further fractlonation of fraction B in Pig. 1. Fraction B in Fig. 1 was applied to a coluan of DEAE-Sepharoee CL-6B(lx2Scm) equilibrated with 0.05M Tria-HCHpH 8.0) containing 0.075M NaCl. After washing, elution was performed with a gradient formed from 150ml each of equilibration buffer and 0.3M NaCl in 0.05M Trls-HCMpH 8.0). Fractions of SmL each were collected. BI, BII and Bill fractions contained band 2, band 3 and band 4 proteins, respectively. The arrow indicates the position where gradient elution was started.
Fig. 3S Purification of band 2 protein. Fraction BI in Fig. 28 was applied to a Sephadex G-5O(lx9Ocm) column. Elution wan performed with 0.01H NH4HCO3 , and 2ml fractions were collected. Band 2 protein was obtained from fractions indicated by the bar.
1.5
0.15
0.5
-
0.10 •
Identification of Full-Sized Forms of Salivary (S-Type) Cystatins (Cystatin SN, Cystatin SA, Cystatin S, and Two Phosphorylated Forms of Cystatin S) in Human Whole Saliva and Determination of Phosphorylation Sites of Cystatin S Satoko Isemura,* Eiichi Saitoh," Kazuo Sanada,** and Kayoko Minakata'** 'Nippon Dental University Junior College at Niigata, Niigata, NOgata 951;**Department of Oral Biochemistry, School of Dentistry at Niigata, Nippon Dental University, Niigata, Niigata 951; and '"Department of Legal Medicine, Hamamatsu University School of Medicine, Hamamatsu, Shizuoka 431-31 Received for publication, March 22, 1991
Our recent work on the gene structures for human salivary (S-type) cystatins [Saitoh, E. et al. (1987) Gene 61, 329-338] has suggested that the structures of cystatins which we determined previously at the protein level lack N-terminal peptide portions of the fullsized intact forms. In the present study, attempts were made to isolate full-sized S-type cystatins by introducing methanol fractionation into the purification steps to suppress the enzymatic activity present in saliva. Full-sized cystatin SN and two phosphorylated forms of full-sized cystatin S were thus isolated. Analysis of one fraction indicated that this was a mixture of full-sized cystatin SA and non-phosphorylated cystatin S. The phosphorylation sites of cystatin S were determined to be Ser-Ser-Ser'(P)-Lys-Glu-Glu- for monophosphorylated cystatin S and Ser'(P)-Ser-Ser3(P)-Lys-Glu-Glu- for diphosphorylated cystatin S. Immunoblotting analysis with anti-cystatin S antiserum revealed that tears and seminal plasma also contained S-type cystatins, but diphosphorylated cystatin S was detected neither in tears nor in seminal plasma and no cystatin SN was found in seminal plasma. These data indicate that S-type cystatins are secreted into the oral cavity without significant degradation in salivary glands or ducts and that they are expressed tissue specifically.
Animal tissues and body fluids contain various kinds of endogenous proteinase inhibitors to regulate their protein metabolism or to protect tissues from attacks by bacteria or viruses (1-4). Cystatins are cysteine proteinase inhibitors belonging to the cystatin superfamily, supposedly evolved from the same ancestral gene. The cystatin superfamily is now grouped into three families (family 1, family 2, and family 3) according to the degree of structural homology (5). Family 2 cystatins are secretory inhibitors with two disulfide bonds and molecular weights of about 15,000 (5). We have isolated three cysteine proteinase inhibitors, cystatin S, cystatin SA, and cystatin SN, which belong to family 2 of the cystatin superfamily, from human whole saliva and determined their amino acid sequences (6-8). These three inhibitors exhibit about 90% sequence homology with one another. In addition to these S-type cystatins, human saliva contains cystatin C, which is also a family 2 cystatin member and has 50% homology with S-type cystatins (9). Human family 2 cystatin genes constitute a multigene family probably having seven members, six of which have been characterized: CST1 for cystatin SN, CST2 for cystatin SA, CST3 for cystatin C, CST4 for cystatin S and two pseudogenes, CSTP1, and CST5 (CSTP2) (10-12). A new cystatin corresponding to the yet unidentified seventh gene may be expressed somewhere in the human body. Abbreviations: TFA, trifluoroacetic acid ; N-, amino. 648
Electrophoresis of human saliva on basic gel consistently gives four bands immunoreactive with anti-cystatin S antiserum, suggesting the possibility of the presence of the S-type cystatin not characterized yet. Our recent results on nucleotide sequences of genes (2012) have indicated that the cystatins which we isolated and characterized previously (6-8) are not full-sized proteins, and suggest that they are probably formed by enzymatic degradation. This study was performed to see whether full-sized inhibitor proteins can be isolated from whole saliva by preventing enzymatic degradation, and also to find out whether there is a new S-type cystatin in human saliva or other body fluids. MATERIALS AND METHODS Materials—The following materials were purchased from the sources indicated: DE32, Whatman Separation; DEAE-Sepharose CL-6B and Sephacryl S-200, Pharmacia Fine Chemicals; Novapak Ci8, Waters Associates; lysylendopeptidase and malachite green oxalate, Wako Pure Chemical; TPCK-trypsin [EC 3.4.21.4], Cooper Biochemical; alkaline phosphatase [EC 3.1.3.1], Miles Scientific; Immun-blot assay kit and nitrocellulose membrane, BioRad Laboratories; Centriprep 10, Amicon. Rabbit polyclonal anti-cystatin S antiserum was prepared as described previously (13). J. Biochem.
649
S-Type Cystatins in Saliva Collection and Handling of Saliva—Human whole saliva was collected from students into ice cooled test tubes containing 0.5 ml of proteinase inhibitor solution (500 fig p-amidinophenylmethanesulfonyl fluoride, 25 fig phosphoramidon, 500 fig bestatin, 500 units Trasylol, 1 mg NaF, 10 mg EDTA 2Na in 0.5 ml of 0.1 M Tris-HCl, pH 7.5). Saliva amounting to 10 ml per person was collected and pooled. Four volumes of cold methanol were added to the pooled saliva. Precipitates formed by the addition of methanol were removed by filtration. Methanol was removed from the filtrate by evaporation under reduced pressure at 40*C, and the residue was lyophilized. The lyophilized material was dissolved in water and dialyzed against 0.02 M TrisHCl buffer (pH 7.5) containing 5% inhibitor solution. After removing insoluble materials by centrifugation, the supernatant was applied to a DE32 column equilibrated with the same buffer. After washing, elution with a gradient formed from 100 ml each of the initial buffer and 3% NaCl in the same buffer was performed (Fig. 1). Fractions eluted from the column were monitored by immunoblotting after polyacrylamide gel electrophoresis at pH 9.4. Samples for Electrophoresis—Submaxillary-sublingual saliva was sampled from the outlet of the submaxillarysublingual duct by using a micropipet. Parotid saliva was collected into an ice-chilled test tube by using a suction cup under sour candy stimulation. Tears were sampled with a micropipet from the outlet of the lachrymal gland. Seminal plasma was obtained as described previously (14). Ten microliters each of undiluted submaxillary-sublingual saliva, tears, and seminal plasma was used for each run of gel electrophoresis. Twenty microliters of parotid saliva was concentrated to 10//I by lyophilization and used for electrophoresis. Samples other than seminal plasma were analyzed immediately after sampling. Polyacrylamide Gel Electrophoresis and Western Blotting—Polyacrylamide gel electrophoresis was performed by the method of Davis (25) at pH 9.4 with 7% slab gel. Gels were stained with Coomassie Brilliant Blue G-250 or transblotted to nitrocellulose membrane by using a transblotting apparatus (Atto) with 0.7% acetic acid. Immunostaining was performed using an Immun-blot assay kit according to the manual issued by Bio-Rad. Anti-cystatin S antiserum was used as the first antibody. Amino Acid Analysis—Proteins or peptides were hydrolyzed in constant boiling hydrochloric acid at 108"C for 24 h in a Pico-tag work station (Waters) and analyzed by the Waters amino acid analysis system using the orthophthalic acid post column labeling method. Amino Acid Sequence Analysis—Amino acid sequence analysis was performed by the direct manual Edman degradation method (16) or by using a gas phase automatic amino acid sequence analyzer (Applied Biosystems). Phosphate Measurement—An aliquot was taken from each fraction eluted from the column, and the volume was adjusted to 390//I by adding 0.01 M NrL,HCO3. Alkaline phosphatase (0.2 unit) in 10//I of 0.01 M N^HCO-j was added. The mixture was incubated for 1 h at 37'C. Color was developed by adding 25 ft 1 of 7.5 M sulfuric acid and 75 //\ of a molybdate-malachite green reagent (17). Absorbance at 650 mn was measured. For quantitative analysis of phosphate, samples hydrolyzed in 6 N HC1 for 24 h at 100'C were used. Enzymatic Digestion—Purified proteins (10-100 nmol) Vol. 110, No. 4, 1991
were reduced and carboxymethylated, and digested with 0.025 unit of lysylendopeptidase in 0.5ml of 0.01M Tris-HCl buffer (pH 8.5) for 2 h at 25'C. Trypsin digestion of the purified lysylendopeptidase peptide was performed in 0.01 M NR,HC0 3 for 16 h at 25#C. Phosphorylation Sites Determination—The direct Edman degradation method was slightly modified to apply it to the determination of phosphorylation sites of phosphopeptides. Each of the phosphopeptides (2 nmol) was dissolved in 30 //I of dimethylallylamine buffer containing 1.5 //I of phenylisothiocyanate. After 1 h reaction at 40'C, the reaction mixture containing the phenylthiocarbamyl peptide was dried up without extraction with benzene. Ten microliters of TFA was added to the dried sample and incubated at 40"C for 20 min to release the phenylthiazolinone derivative of an amino acid. TFA was removed from the reaction mixture by a nitrogen gas stream. Inorganic phosphate released during the TFA-treatment, when a peptide had carried an N-terminal phosphoserine residue (18), was dissolved in 0.4 ml of 0.01 M NH,HC0 3 . The color was developed by adding 25 ft\ of 7.5 M sulfuric acid and 75 //I of a molybdate-malachite green reagent (17). RESULTS Immunoblotting Analysis of Human Body Fluids—Figure 2 shows immunoblotting analysis of human submaxillary-sublingual saliva, parotid saliva, tears, and seminal plasma after electrophoresis on anionic gel. Human serum gave no positive bands. Submaxillary-sublingual saliva, parotid saliva, and tears from four individuals were tested. In all submaxillary-sublingual saliva samples tested, four bands (1, 2, 3, and 4) were detected and the intensity of band 2 relative to band 1 or band 3 was low. In parotid saliva, four bands were also detected, but the intensity of band 2 was the highest in all samples. Band 4 was not detected in any sample of tears tested. Neither band 1 nor band 4 was detected in seminal plasma. Isolation of the Protein Corresponding to Band 1—The fractions not adsorbed on a DE32 column (fraction A in Fig. 1) were lyophilized, dissolved in water, dialyzed against 0.01 M Tris-HCl buffer (pH8.5), and then applied to a DEAE-Sepharose CL-6B column equilibrated with the same buffer. The elution profile is shown in Fig. IS. The protein contained in the fractions indicated by the bar had the same electrophoretic mobility as band 1 of fresh saliva. Isolation of the Protein Corresponding to Band 2—Fraction B in Fig. 1 was lyophilized, then dissolved in water, dialyzed against 0.05 M Tris-HCl buffer (pH8.0) and applied to a column of DEAE-Sepharose CL-6B equilibrated with the same buffer containing 0.075 M NaCl. The elution profile is shown in Fig. 2S. Fraction BI in Fig. 2S contained proteins giving two bands electrophoretically: one corresponding to band 2 and reactive with anti-cystatin S antiserum, and the other having slightly lower mobility and unreactive with anti-cystatin S antiserum. This fraction was then concentrated in a Centriprep 10, and insoluble materials formed during concentration were removed by centrifugation. The concentrate was applied to a column of Sephadex G-50 (IX90cm) (Fig. 3S). The electrophoretically homogeneous protein corresponding to band 2 was obtained from fractions indicated by the bar. Isolation of the Protein Corresponding to Band 3—The
S. Isemura et al. protein in fraction BII in Fig. 2S was homogeneous electrophoretically and gave the band corresponding to band 3. Salts were removed by gel filtration through a column of Sephadex G-50 (1x90 cm). Isolation of the Protein Corresponding to Band 4—Although the protein in fraction BUI in Fig. 2S was homogeneous electrophoretically, its amino acid analysis suggested that this fraction contained a protein mixture having
the same electrophoretic mobility as band 4. This fraction was concentrated and applied to a column of Sephacryl S-200 (Fig. 4S). The second peak in Fig. 4S contained the protein immunoreactive with anti-cystatin S antiserum and having the same electrophoretic mobility as band 4. N'-Terminal Amino Acid Sequence of the Purified Proteins—N-Terminal amino acid sequences of the purified proteins were as follows. Band 1 protein:
Band 3 protein: Band 4 protein:
20
40
Fraction
60
80
Number
Fig. 1. Separation of methanol-treated whole saliva on a DE 32 column. The solid material (1.15 g) from 500 ml of whole saliva was applied to the column (1.7 x 15 cm) equilibrated with 0.02 M TrisHC1 (pH 7.5). After washing with equilibration buffer, elution was performed with a gradient formed from 100 ml each of initial bufFer and 3% NaCl in the same buffer, and 5 ml fractions were collected. The arrow indicates the position where gradient elution was started. The protein corresponding to band 1 was detected in fraction A, and the proteins corresponding to band 2, band 3, and band 4, in fraction B.
2
3 4
2 34 e f
2 34 Fig. 2. Immunoblotting analysis of human body fluids. Electrophoresis was performed at pH 9.4. Lane a, tears; lane c, seminal plasma; lane e, parotid saliva; lanes b, d, and f, submaxillary-sublingual saliva. The arrow indicates the direction of migration.
Trp-Ser-Pro-Lys-Glu-Glu-Asp-ArgIle-Ile-Pro-Gly-Gly-Ile-Tyr-Asn-AlaAsp-Leu-AsnSer-Ser-Ser-Lys-Glu-Glu-Asn-ArgIle-Ile-Pro-Gly-Gly-Ile-Tyr-AspSer-Ser-Ser-Lys-Glu-Glu-Asn-ArgIle-Ile-Pro-Gly-Gly-Ile-Tyr-Asp-
In the case of band 2 protein, results obtained by using an automated amino acid sequence analyzer indicated that the preparation contained two proteins in a ratio of about 7 : 3. The N-terminal sequence of the major protein was: TrpSer-Pro-Gln-Glu-Glu-Asp-Arg-Lle-Ile-Glu-Gly-Gly-Ile-TyrAsp-Ala-Asp-Leu-Asn-. The N-tenninal sequence of the minor one was identical with those of band 3 and band 4 proteins. When these sequences were compared with the amino acid sequences predicted from the gene structures {10, 12), it was found that band 1 and band 2 major proteins had the N-terminal amino acid sequences of full-sized cystatin SN and full-sized cystatin SA, respectively. The N-terminal sequences of band 3, band 4, and band 2 minor proteins were all identical with the N-terminal sequence of fullsized cystatin S. Amino Acid Composition and Phosphate Contents of the Purified Proteins—The amino acid composition of band 1 protein coincided with that of full-sized cystatin SN within the range of error. The amino acid compositions of both band 3 and band 4 proteins corresponded well with that of full-sized cystatin S. The amino acid composition of the fraction corresponding to band 2 was compatible with the composition, if we assume that the fraction contained full-sized cystatin SA and cystatin S in an approximate ratio of 7 : 3. Their amino acid compositions are shown in
TABLE I. Phosphate content of purified proteins. Phosphate content (mol/mol protein) Band 1 protein Band 2 protein 1.32 Band 3 protein Band 4 protein 2.35
Fig. 3. Effect of alkaline phosphatase digestion on salivary cystatins. Whole saliva (1 ml) was digested with 0.2 unit of alkaline phoephatase for 1 h at 37"C. Electrophoresia and immunoblotting were performed as 2 34 described in the experimental section. Samples of 10//I each were analyzed. Lane a, whole saliva digested with alkaline phosphatase; lane b, whole saliva (control). The arrow indicates the direction of migration. J. Biochem.
S-Type Cystatins in Saliva
651 TABLE II. Axnino acid compositions and phosphate contents of lysylendopeptidase peptides derived from band 3 and band 4 proteins (mol/mol peptide). Ser Ly3 Phosphate 2.33 1.00 0.65 nP 2.86 1.00 0.58 nP 2.95 1.00 2.07
A 650 A
280
0.15
0.
TABLE III. Phosphate released by Edman degradation of phosphopeptides. The values indicate the ratio in percentage of phosphate released by Edman degradation to the phosphate content of the peptide used for Edman degradation. Cycle of degradation Peptide subjected to 3 Edman degradation* 1 2 nP 39.2 1.8 0.3 52.2 9.5 mp 82.8 'Edman degradation was performed for the peptide purified after every cycle of degradation.
10
0.05
Fraction Number Fig. 4. Comparison of lysylendopeptidase digests of bands 3 and 4 proteins. The digests were fractionated on a column of DEAESepharose CL-6B (0.9x25 cm) equilibrated with 0.01 M NrLHCO, and elution was performed with a gradient formed from 100 ml each of 0.01 M NH.HCO, and 1 M NH,HC03 . Fractions of 2 ml each were collected. Solid and dotted lines indicate the absorbance values at 230 and 650 nm, respectively, (a) Chromatography of band 3 protein digests, (b) Chromatography of band 4 protein digests. Peptides (P)
assigned to each peak are: 1, 1', K; 2, 2', SSSK, ATEDEYYRRPLQVLRAREQTFGGVNYFFDVEVGRTICTK; 3, SQPNLDTCAFHEQPELQK; 3', SQPNLDTCAFHEQPELQK, SSSK; 4,4', EENRIIPGGIYDADLNDEWVQRALHFAISEYNK; 5,5', QLCSFEIYEVPWEDRMSLVNSRCQEA
Table IS. Phosphate contents of the purified protein fractions are listed in Table I. The data suggested that band 3 and 4 proteins were phosphorylated. Alkaline Phosphatase Digestion of Whole Saliva—Human whole saliva was digested with alkaline phosphatase and analyzed by electrophoresis and immunoblotting (Fig. 3). Band 3 and band 4 disappeared, and band 2 was intensified. This result indicated that both band 3 and band 4 represented phosphorylated proteins and that they migrated to the position of band 2 upon alkaline phosphatase digestion. Based on these results, band 3 and band 4 proteins were Vol. 110, No. 4, 1991
tentatively assigned as mono- and diphosphorylated fullsized cystatin S, respectively. Full-sized cystatin SN (band 1 protein) and cystatin SA (band 2 major protein) appeared not to be phosphorylated. Comparison of Peptide Maps of Band 3 and Band 4 Proteins—In order to confirm that band 3 and band 4 proteins are the same proteins with different degrees of phosphorylation, lysylendopeptidase digests of the two proteins were compared. Each protein was digested with lysylendopeptidase after reduction and carboxymethylation, and the digests were separated on a column of DEAESepharose CL-6B. Elution profiles of band 3 and band 4 protein digests are shown in Fig. 4, a and b, respectively. Elution profiles monitored by measuring the absorbance at 230 nm of the two chromatograms showed no apparent difference, but the profiles monitored in terms of phosphate content were clearly distinguishable. The peptides assigned to each peak are shown in the legend to Fig. 4. These assignments were based on the amino acid composition and N-terminal sequence analysis of each peak, and HPLC of peptides derived by trypsin digestion of each peak (not shown). As HPLC reference peptides, we used tryptic peptides prepared from sequenced cystatin S (6). There has been a discrepancy between the sequences of cystatin S determined by protein chemistry and those predicted from the gene structure. The 107th residue (115th in the fullsized sequence) in cystatin S is Asp (6), while it is Asn from the gene structure of CST4 [12). Sequence analysis of peak 5 and peak 5' indicated that both band 3 and band 4 proteins had Asn at this position. Therefore, we correct here the previously determined sequence of cystatin S (6) by placing Asn at the 107th residue. The phosphate peaks designated I and I' in Fig. 4, a and b, were considered to represent inorganic phosphate since these phosphate peaks could be detected even without alkaline phosphatase digestion. In a control run, inorganic phosphate was confirmed to be eluted at this position. The phosphate-containing fractions designated II, II', and EH' were separated on a column of Novapak Cis with a gradient of 0-60% acetonitrile in 0.1% TFA. In each case, phosphate was detected in the fraction not adsorbed on the column. These unadsorbed fractions obtained from II, II', and HI' were designated as lip, lip', and Hip', respectively. Amino acid analysis data
652
S. Isemura et al.
listed in Table IS suggest that these three peptides are practically pure and are derived from the amino terminal portion of full sized cystatin S. Phosphate contents relative to amino acid components of lip, lip', and Hip' are listed in Table II. On the basis of the results in Table II, lip, and lip' are considered to correspond to the monophosphorylated N-terminal tetrapeptide of full-sized cystatin S, and Dip' to the diphosphorylated N-terminal tetrapeptide of full-sized cystatin S. The appearance of a high inorganic phosphate peak (I') in Fig. 4b may be due to partial dephosphorylation of the purified band 4 protein during enzyme digestion. Phosphorylotion Sites of lip and nip'—In the case of Edman degradation of phosphopeptide, inorganic phosphate has been shown to be released during TFA treatment when a phosphoserine residue was located at the Nterminus (28). Therefore we applied this method to the determination of phosphorylation sites of phosphopeptides isolated in this study. lip and Hip' were subjected to Edman degradation. In the case of Kip', the peptide was purified after each cycle of degradation by a DEAE-Sepharose CL-6B column using a gradient from 0.01 to 1 M NK,HC0 3 , while in the case of lip, peptide purification was not performed during 3 cycles of degradation. Table III lists the ratio of phosphate released after each cycle of degradation to the phosphate content of the original peptide as a percentage. The results indicated that the structures of Up and Hip' were Ser-SerSer(P)-Lys and Ser(P)-Ser-Ser(P)-Lys, respectively. lip' in Pig. 4b was considered to be identical with lip, since their amino acid compositions, ratio of phosphorylation and chromatographic behavior were the same. lip' was probably formed artificially by dephosphorylation of purified band 4 protein during handling These data lead us to conclude that the protein corresponding to band 3 was full-sized cystatin S monophosphorylated at the 3rd serine residue and that the protein corresponding to band 4 was full-sized cystatin S diphosphorylated at the 1st and the 3rd serine residues. DISCUSSION In the early stage of our study on cyatatins, we isolated several forms of S-type cystatins (29). The occurrence of multimolecular forms could be explained by the presence of proteins having closely related but distinct amino acid sequences such as cystatin S, cystatin SN, and cystatin SA, and their degradation products. Even when conditions were used under which enzymatic degradation during purification was largely suppressed, there was still a possibility that degradation was not prevented completely, since cystatin SN and cystatin S A obtained by this method lacked the N-terminal 3 and 4 residues of the full-sized proteins predicted from the gene structures, respectively (20). In the present study, we introduced the addition of methanol to the concentration of 80%. This resulted in precipitation of most of the proteins in saliva, but significant amounts of cystatins could be recovered from the supernatant. Although the yield of cystatins (4.4 mg/liter for monophosphorylated cystatin S, 0.98 mg/liter for diphosphorylated cystatin S, 3.35 mg/liter for cystatin SN) was a little less as compared with the previous method, the present method provided full-sized forms of S-type cystatins. Inclusion of methanol precipitation appeared to be very effective to
TABLE IV. The structures around phosphoserine residues of salivary proteins. Phosphoprotein Structure Reference Human cystatin S (monophosphorylated) Ser-Ser-Ser(P)-Lys-GIu-GluThis study (diphosphorylated) Ser(P)-Ser-Ser(P)-Lys-Glu-Glu This study Human statherin Asp-Ser(P)-Ser(P)-Glu-Glu-Lys22 Macaca fascicularis Asp-Ser(P)-Ser(P)-Glu-Glu-Lys23 statherin Human histidine-rich Asp-Ser(P)-Hi8-Glu-Lys-Arg24 polypeptide Human proUne-rich -Val-Ser(P)-Gln-Glu-Asp-Val25 phosphoprotein
suppress proteolytic enzyme activities. Fresh whole saliva consistently gave four bands on immunoblotting after anionic gel electrophoresis, and the four proteins corresponding to these bands were studied in this work. The fact that full-sized forms of cystatins were isolated indicates that S-type cystatins are secreted into the oral cavity without suffering significant degradation in salivary glands or in ducts. Previous studies (19, 20) have shown that enzymes present in saliva can cleave cystatins at their N-terminal portions. Enzymes responsible for the cleavage are now under investigation. In our previous method, only cystatin S was isolated as the full-sized form (20). It is possible that this is because enzymes cleaving cystatin S are less active under the previous purification conditions, where no methanol fractionation was included, as compared with those cleaving cystatin SN and cystatin SA. Alternatively, a subtle structural difference may explain the difference in susceptibility, if we assume that the same enzyme(s) is responsible for degradation of all S-type cystatins. Two phosphorylated forms of cystatin S were found to be present in fresh saliva. It is not clear at the moment whether monophosphorylated cystatin S is formed by differential phosphorylation of cystatin S or by phosphatase digestion of the diphosphorylated form. Mammary gland casein kinase and casein kinase II are proposed to recognize Ser*-X-Glu/Ser(P) and Ser*-X-XGlu/Ser(P), respectively, where Ser' indicates potential acceptor serine and X can be any amino acid residue (21). Since the phosphoserine residue of all phosphorylation sites of cystatin S, Le., the third residue of monophosphorylated cystatin S and the first and third residues of diphosphorylated cystatin S, have glutamic acid or phosphoserine at two residues to the right of them, all phosphorylation sites of cystatin S are in agreement with the recognition sequence of mammary gland casein kinase. Phosphoserine of the third residue of mono- or diphosphorylated cystatin S has glutamic acids at two and three residues toward the right. Therefore, this site of phosphorylation is in agreement with not only the proposed recognition sequence of the mammary gland casein kinase but also that of casein kinase II. Phosphorylation sites which are in agreement with the recognition sites of both casein kinase II and mammary gland casein kinase have been found in salivary proteins such as human statherin (22) and Macaca fascicularis statherin (23). Phosphorylation sites only compatible with the recognition sequence of mammary J. Biochem.
S-Type Cystatins in Saliva gland casein kinase are present in-salivary proteins such as human and M. fascicularis statherin (22, 23), human hi8tidine-rich polypeptide (24) and human proline-rich protein (25). Sequences around the phosphorylation sites of salivary proteins are shown in Table IV. Since removal of phosphate by alkaline phosphatase had no effect on inhibitory activity for papain (unpublished data), phosphorylation on serine residues does not appear to be involved in the inhibitor activity of cystatin S. Chicken egg white cystatin, a family 2 cystatin member, is also partly phosphorylated, although the phosphorylation site (Ser-80) is different from that of cystatin S's (26). Hawke et al. isolated full-sized cystatin S (27) and full-sized cystatin SN (28) from human whole saliva by TFA fractionation and subsequent separation by HPLC. However, they only described one component of cystatin S and gave no information about phosphorylation. Al-Hashimi et al. isolated full-sized cystatin SN from human submaxillary-sublingual saliva without using methanol or TFA fractionation (29). Therefore, enzymes responsible for cleavage of cystatin SN may be derived from salivas other than submaxillary-sublingual saliva or from microbes. There has been no report on whether the full-sized form of cystatin SA is present in whole saliva. Therefore, this is the first report to describe the existence of the full-sized protein which has been predicted from the gene structure analysis (10). The S-type cystatins are tissue specifically expressed as shown in Fig. 2. Salivary glands may be different from lachrymal gland or seminal glands as regards the phosphorylation mechanism, since diphospho-cystatin S is detected only in saliva but not in tears or in seminal plasma. The intense band at the position of cystatin SA (band 2) in parotid saliva, tears and seminal plasma could be nonphosphorylated cystatin S, since non-phosphorylated cystatin S has the same electrophoretic mobility as that of full-sized cystatin SA. Differential expression by tissues may reflect differences in the physiological significance among the inhibitors of this type. The four proteins isolated in this study are encoded by three genes, CST1, CST2, and CST4. We could not detect S-type cystatin coded by an uncharacterized new gene (10). Whether there is an S-type gene product derived from the new gene is not clear at present, although there is a possibility that the weak band detected on immunoblotting (Fig. 2) could represent a new cystatin. REFERENCES 1. Katunuma, N. (1989) in Intraceilular Proteolysis (Katunuma, N. & Kominarai, E. eds.) pp. 3-23, Jpn. Sri. Soc. Press, Tokyo 2. Bjoi-ck, L., Akesson, P., Bohus, M., Ttojnar, J., Abrahamson, M.,
Vol. 110, No. 4, 1991
653 Olafuson, I., & Grubb, A. (1989) Nature 337, 385-386 3. Korant, B.O., Brain, J., & Turk, V. (1985) Biochem. Biophys. Res. Common. 127, 1072-1076 4. Bjflrck, L-, Grubb, A., & Kjellen, L. (1990) J. Virol. 64, 941-943 5. Barrett, A.J., Fritz, H., Grubb, A., Isemura, S., Jflrvinen, M., Katunuma, N., Machleidt, W., Mttller-Esterl, W., Sasaki, M., & Turk, V. (1986) Biochem. J. 236, 312 6. Isemura, S., Saitoh, E., & Sanada, K. (1984) J. Biochem. 96, 489-498 7. Isemura, S., Saitoh, E., & Sanada, K. (1986) FEBS Lett. 198, 145-149 8. Isemura, S., Saitoh, E., & Sanada, K. (1987) J. Biochem. 102, 693-704 9. Abrahamson, M., Barrett, A.J., Salvesen, G., & Grubb, A. (1986) J. BioL Chem. 261, 11282-11289 10. Saitoh, E., Kim, H.S., Smithies, O., & Maeda, N. (1987) Gene 61, 329-338 11. Saitoh, E., Sabatini, L.M., Eddy, R.L., Shows, T.B., Azen, E.A., Isemura, S., & Sanada, K. (1989) Biochem. Biophys. Res. Commun. 162, 1324-1331 12. Saitoh, E., Isemura, S., Sanada, K., & Ohnishi, K. (1991) Biomed. Biochim. Ada 60, 599-605 13. Isemura, S., Saitoh, E., Ito, S., Isemura, M., & Sanada, K. (1984) J. Biochem. 96, 1311-1314 14. Minakata, K. &Asano, M. (1985) BioL Chem. Hoppe-Seyler 366, 15-18 15. Davis, B.J. (1964) Ann. N.Y. Acad. Sci. 121, 404-427 16. Iwanaga, S., Wallen, P., Groendahl, N.J., Henschen, A., & Blombfick, B. (1969) Eur. J. Biochem. 8, 189-199 17. Matsubara, C, Yamamoto, Y., & Takamura, K. (1987) Analyst 112, 1257-1260 18. Proud, C.G., Rylatt, D.B., Yeaman, S.J., & Cohen, P. (1977) FEBS Lett. 80, 435-440 19. Isemura, S., Saitoh, E., Sanada, K., Isemura, M., &Ito, S. (1986) in Cysteine Proteinases and Their Inhibitors (Turk, V., ed.) pp. 497-505, Walter de Gruyter, Berlin, New York 20. Saitoh, E., Isemura, S., Sanada, K., Kim, H.-S., Smithies, O., & Maeda, N. (1988) Biol. Chem. Hoppe-Seyler 369, Suppl., 191197 21. Kemp, B.E. & Pearson, R.B. (1990) Trends Biochem. Sci. 16, 342-346 22. Schlesinger, D.H. & Hay, D.I. (1977) J. BioL Chem. 262, 16891695 23. Oppenheim, F.G., Offner, G.D., & Troxler, R.F. (1982) J. Biol. Chem. 267, 9271-9281 24. Oppenheim, F.G., Yang, Y.-C, Diamond, R.D., Hyslop, D., Offner, G.D., & Troxler, R.F. (1986) J. BioL Chem. 261, 11771182 25. Wong, R.S.C. & Bennick, A. (1980) J. Biol. Chem. 265, 59435948 26. Laber, B., Krieglstein, K., Henschen, A., Kos, J., Turk, V., Huber, R., & Bode, W. (1989) FEBS Lett. 248, 162-168 27. Hawke, D.H., Yuan, P.M., Wilson, K.J., & Hunkapiller, M.W. (1987) Biochem. Biophys. Res. Commun. 145, 1248-1253 28. Hawke, D.H., Yuan, P.M., Ohno, J., Kikuchi, M., Hirayama, K., & Fukuyama, K. (1989) in Intraceilular Proteolysis (Katunuma, N. & Kominami, E., eds.) pp. 389-390, Jpn. Sri. Soc. Press, Tokyo 29. Al-Hashimi, I., Dickinson, D.P., & Levine, M.J. (1988) J. BioL Chem. 263, 9381-9387
654
S. Isemura et al. Supplemental Materials Ami no acid composition! of purified proteins and phosphopeptides, and ami.no acid composition* of full-sized cystatin S, cystatin SA and cyst*tin SN, predicted frog the gone structures. values axe shown in molar ratios. Band 1 protein
Asp Thr Ser Glu Pro Gly Ala l/2Cys
val Hat 11* L«u Tyr Pho Trp His
13.0 4.8 6.4 19.7 7.4 6.1 6.0 2.5 7.6 4.7 7.0 5.4 6.0 nd* 1.9 6.6 10.4
Lys Arg
Band 2 protein
13.0 4.8 6.3 22.1 5.6 4.9 7.2 3.1 6.9 0.8 6.0 B.8 5.1 4.9 nd 2.6 4.8 10.6
Band 3 protein
13.1 4.7 7.7 21.9 5.1 5.6 6.8 3.8 7.0 0.7 4.8 7.9 5.1 5.5 nd 1.7 5.1 8.4
Band 4 protein
15.1 4.6 7.2 23.0 6.6 4.9 6.9 2.2 6.9
Cystatin SN
Cystatin SA
13 5 7 19 6 5 6 4 8
13 4 6 23 4 5 6 4
Cystatin S
13 5 8 22 5 5 7 4 7 1 6 8 6 6 2 2 5 9
a. 8 5.6 9.1 5.0 5.9 nd 1.5 5.1 8.4
6 8 6 S 3 2 7 11
9 6 5 2 2 4 11
:ip
Up1
IIIp'
0.05
0.26
2. 86 0.14
2 .95 0 .27
0.5 -
0 .25 0 .IS
0.18 0.22
1.00
1.00
20
1.00
40
60
Fraction Nuiiber
nd f not determined.
Pig. 23 Further fractlonation of fraction B in Pig. 1. Fraction B in Fig. 1 was applied to a coluan of DEAE-Sepharoee CL-6B(lx2Scm) equilibrated with 0.05M Tria-HCHpH 8.0) containing 0.075M NaCl. After washing, elution was performed with a gradient formed from 150ml each of equilibration buffer and 0.3M NaCl in 0.05M Trls-HCMpH 8.0). Fractions of SmL each were collected. BI, BII and Bill fractions contained band 2, band 3 and band 4 proteins, respectively. The arrow indicates the position where gradient elution was started.
Fig. 3S Purification of band 2 protein. Fraction BI in Fig. 28 was applied to a Sephadex G-5O(lx9Ocm) column. Elution wan performed with 0.01H NH4HCO3 , and 2ml fractions were collected. Band 2 protein was obtained from fractions indicated by the bar.
1.5
0.15
0.5
-
0.10 •
