The Structure of the l3ovine Pancreatic Secretory Trypsin Inhibitor-Kazal’s Inhibitor DETERMINATION OF THE DISULFIDE BONDS AND PROTEOLYSIS BY THERMOLYSIN”

Abstract The three disulfide bonds of bovine pancreatic secretory trypsin inhibitor were assigned on the basis of the structures of cystine peptides isolated from thermolysin hydrolysates of the native inhibitor prepared at pH 6.5. The peptides were isolated by chromatographic procedures with Sephadex G-75, Sephadex G-25, and Dowex 50-X2. Five major cystine peptides were oxidized with performic acid and the cysteic acid fragments were separated by chromatography on Dowex 50-X2 and by high voltage paper electrophoresis. The cysteic acid peptides were located in the amino acid sequence of the inhibitor (Greene, L. J., and Bartelt, D. C., J. Biol. Chem., 244, 2646 (1969)) on the basis of amino acid composition and end group determination. Cystine peptides corresponding to disulfide bonds I–V, II–IV, and III–VI were isolated in 55, 44, and 91% yield, respectively. Intermediates in the degradation of the inhibitor by thermolysin having either one or two peptide bonds hydrolyzed per molecule were isolated. The sites of hydrolysis were identified as the peptide bonds Ile-Leu (residues 2 to 3) and Glu-Val (residues 12 to 13).

The three disulfide bonds of bovine pancreatic secretory trypsin inhibitor were assigned on the basis of the structures of cystine peptides isolated from thermolysin hydrolysates of the native inhibitor prepared at pH 6.5. The peptides were isolated by chromatographic procedures with Sephadex G-75, Sephadex G-25, and Dowex 50-X2. Five major cystine peptides were oxidized with performic acid and the cysteic acid fragments were separated by chromatography on Dowex 50-X2 and by high voltage paper electrophoresis.
The cysteic acid peptides were located in the amino acid sequence of the inhibitor (GREENE, L. J., AND BARTELT, D. C., J. Biol. Chem., 244,2646 (1969)) on the basis of amino acid composition and end group determination.
Cystine peptides corresponding to disulfide bonds I-V, II-IV, and III-VI were isolated in 55, 44, and 91% yield, respectively.
Intermediates in the degradation of the inhibitor by thermolysin having either one or two peptide bonds hydrolyzed per molecule were isolated.
The sites of hydrolysis were identified as the peptide bonds Ile-Leu (residues 2 to 3) and Glu-Val (residues 12 to 13).
The study of the interaction of polypeptide trypsin inhibitors with trypsin and trypsin-like enzymes affords the opportunity to examine specific protein-protein interactions of physiological and perhaps clinical significance (4,5). Knowledge of the amino acid sequence and three-dimensional structure of the enzymes and their inhibitors is a prerequisite to the elucidation of the chemistry of these reactions (cf. References 6 to 8).
Bovine pancreatic tissue contains two trypsin inhibitors. The * This research was carried out at Brookhaven National Laboratory under the auspices of the United States Atomic Energy Commission.
Preliminary reports of parts of this work have been presented at the First International Conference on Proteinase Inhibitors, Munich, November 1970 (l), and the 62nd Annual Meeting of American Society of Biological Chemists, San Francisco, 1971 (2). Paper II of this series is Reference 3. 1 Visiting Biochemist, 1969to 1971. Recipient of North Atlantic Treaty Organization Postdoctoral Fellowship, 1969. Permanent address, Unite de Recherche de Pathologie Digestive, 46, Chemin de la Gaye, 13 Marseille (9), France. basic pancreatic trypsin inhibit'or (Kunitz' inhibitor) described by Kunitz and Northrop (9) also inhibits chymotrypsin and porcine kallikrein.
The trypsin inhibitor first isolated by Kazal, Spicer, and Brahinsky in 1948 (13) was later shown to be a component of the pancreatic exocrine secretion (14,15).
In previous reports from this laboratory we described the isolation of bovine PSTII from pancreatic juice and determined the sequence of the reduced S-aminoethylated inhibitor which contains 56 amino acid residues arranged in a single polypeptide chain (3). We wish t,o report; the disposition of the three disulfide bonds in bovine PSTI.
The cystine peptides were prepared by extensive hydrolysis of the inhibitor by thermolysin at pH 6.5 and t'hey were isolated by column chromatographic procedures.
A mixture of partially hydrolyzed forms of bovine PST1 having one or two peptide bonds cleaved per molecule was isolated and characterized.

EXPERIMENTAL PROCEDURE
Column Chromatography-Columns of Sephadex G-25, G-50, and G-75 were prepared and operated as indicated in Greene and Giordano (16). Gradient elution chromatography on Dowex 50-X2 by the method of Schroeder (17) was carried out under the conditions given in Ferreira, Bartelt,and Greene (18). Chromatographic elution patterns were determined by subjecting aliquots of column effluent to alkaline hydrolysis followed by reaction with ninhydrin (19).
Amino Acid Analysis-Samples containing 0.05 to 0.1 pmole of peptide were hydrolyzed in evacuated sealed tubes with 1 ml of twice distilled constant boiling HCl for 22 hours at 110". Amino acid analyses were performed by the method of Spackman, Stein, and Moore (20) on an automatic instrument with provisions for multiple sample application (21). No corrections were applied for the destruction of serine, threonine, or tyrosine caused by acid hydrolysis.
Amino acids present in quantities less than 0.10 mole per mole of peptide have not been included in the tables.
Performic Acid Oxidation-Two procedures were used. The cystine content of peptides was determined as cysteic acid after performic acid oxidation by the method of Moore (22) prior to acid hydrolysis.
Cysteic acid peptides suitable for structural studies were prepared by performic acid oxidation of cystine peptides by the procedure of Hirs (23). The values reported for cysteic acid have not been corrected for incomplete oxidation in either case.
Edman Degradation-The procedures used for subtractive Edman degradation have been described in Bartelt and Greene (24).
Detection of Cystine Peptides-Column effluents were not monitored for cystine peptides by specific calorimetric procedures. The eflluent was combined on the basis of ninhydrin color after alkaline hydrolysis.
Duplicate aliquots of these pools were submitted to acid hydrolysis with and without prior performic acid oxidation by the method of Moore (22).
Bovine PSTI-These experiments were performed on two samples of inhibitor prepared from fractions of bovine pancreas by procedures developed for the isolation of inhibitor from bovine pancreatic juice (14). The results documented here were obtained on the same sample of inhibitor as was used for the amino acid sequence determination of the inhibitor (3,16). The second sample was carried through the Dowex 50-X2 fractionation step (cf. Fig. 1) with essentially the same results with respect to chromatographic elution patterns and amino acid composition of peptides.
Thermolysin Hydrolysis-In order to minimize the disulfide interchange reaction, the hydrolysis was conducted at pH 6.5, 37", and cystine peptides were isolated under acidic conditions (25-28). Thermolysin (10.4 mg of protein) was suspended in 0.02 M calcium acetate at 0" and solubilized by the dropwise addition of 0.2 N NaOII until the pH reached 11. 5 (29). The pH of the solution was immediately adjusted to pH 8.2 with 0.2 N acetic acid. The protein concentration was adjusted to 1.04 mg per ml by dilution with 0.1 M MES buffer (2-(N-morpholino)ethane sulfonic acid), pH 6.5, containing 0.002 M calcium chloride. The activity of thermolysin solubilized under these conditions (determined by a spectrophotometric procedure with FAGLA (3-(2-furylacryloyl)glycyl-L-leucine amide) (30) was comparable to that of thermolysin prepared at pH 8.5.
Five micromoles of bovine PST1 (32.5 mg) and 0.2 mg of thermolysin were incubated in 10 ml of 0.1 M MES buffer, pH 6.5, 0.002 M calcium chloride at 37" for 48 hours. The reaction was stopped by adjusting the solution to pH 3.0 with glacial acetic acid and the hydrolysate was submitted to gel filtration on Sephadex G-75. Cystine peptides were isolated by the procedures summarized in Fig. 1.
High Voltage Electrophoresis-Peptides were subjected to electrophoresis (Electrophorator D, Gilson Medical Electronics) on Whatman 3MM paper at pH 6.5 (25 ml of pyridine, 1 ml of acetic acid, and 225 ml of HzO) and at pH 3.5 (1 ml of pyridine, 10 ml of acetic acid, and 189 ml of HZO). Guide strips were developed with ninhydrin (0.5% in acetone). Acidic peptides were eluted from the paper with 50% aqueous pyridine and neutral peptides with 0.1 M ammonium hydroxide.
The products of performic acid oxidation of Peptide I-I (0.2 pmole) were separated by electrophoresis at pH 6.5 for 30 min at 52 volts per cm. The oxidation products of Peptide II-3 (0.2 pmole) were prepared by electrophoresis at pH 6.5 for 40 min at 73 volts per cm.
Peptides present in Fraction IV (Fig. 3) were separated into components IV-l, IV-2, and IV-3 by electrophoresis at pII 6.5 1. Flow diagram for the isolation of cystine peptides derived from thermolysin hydrolysate of bovine PSTI.
for 30 min at 52 volts per cm. Fraction IV-3 was resubmitted to electrophoresis at pH 3.5 for 150 min at 52 volts per cm to prepare Fractions IV-&A, IV-&B, and IV-3-C. Fraction V (Fig. 3) was purified by electrophoresis at pH 3.5 for 150 min at 52 volts per cm to prepare Fractions V-l, V-2, V-3, and V-4.
Calculation of Recovery of Pep&-The recoveries of peptides are based on the results of amino acid analysis.
They were corrected for material consumed for detection and amino acid analysis but not for chromatographic losses. The yield of cystine peptides have been calculated on the basis of the extensively hydrolyzed inhibitor (Fraction 2, Fig. 2) which represents 75% of the inhibitor treated with thermolysin for 48 hours. Similarly, the data for sites of and the extent of thermolysin hydrolysis of bovine PST1 summarized in Fig. 8 are based on peptides isolated from Fraction 2.

RESULTS
The hydrolysis of bovine PST1 by thermolysin leads to the formation of two classes of products which are separable by gel filtration on Sephadex G-75 (Fig. 2). Fraction 1 contains a mixture of partially hydrolyzed inhibitors, with one or two peptide bonds cleaved per molecule.
Fraction 2 is a mixture of small peptides resulting from the extensive hydrolysis of bovine PST1 by thermolysin.
The amino acid compositions of both fractions were essentially the same as the starting material (cf.   The assignment of the disulfide bonds based on the isolation of cystine peptides in Fraction 2 will be presented first, followed by the characterization of partially hydrolyzed inhibitor present in Fraction 1.

Assignment of Disulkde Bonds
Isolation of Cystine Peptides from Fraction 2 (Fig. 2)-The amino acid content of Fraction 2 accounted for all of the amino acids in bovine PST1 and was recovered in 75% yield relative to the amount of inhibitor incubated with thermolysin. After one cycle of subtractive Edman degradation the amino acid content of Fraction 2 decreased by approximately 12 to 14 residues per equivalent of inhibitor, showing that the mixture contained many small peptides with an average size of 4 to 5 residues per molecule.
The flow diagram given in Fig. 1 summarizes the procedures used for the isolation of cystine peptides from Fraction 2. The system of nomenclature used specifies t,he route of purification by identifying the elution diagram peaks from which the purified peptides have been prepared (cf. Fig. 1). Cystine peptides were effectively separated from smaller peptidcs when Fraction 2 was submitted to gel filtration on Sephadcx G-25 as shown in Fig. 3. The effluent corresponding to t,he filled bars I and II contained 88% of the cystine peptides applied to t,he column.
I and II were further fractionated on Dowex 50-X2 (Fig. 4). The Jilled bars indicate eluate containing cystinc Peptides I-l, I-3, I-4, and I-7 derived from Pool I (Fig. 4, top) and Peptides 11-3, 11-8, and II-9 derived from Pool II (Fig. 4, bottom). Peaks I-7 and II-9 contained less than 3% of the cystine present in Fraction 2. This material was not sufficiemly homogeneous nor present in high enough quantities to be further characterized.
Pept,ides I-l, I-3, I-4, 11-3, and II-8 were oxidized with performic acid and the resulting cyst& acid peptides were isolated either by chromatography on Dowes 50-X2 or by high voltage I II  Fig. 2). The sample contained peptides equivalent to 3.5 pmoles of inhibitor. The column, 0.9 X 400 cm, was equilibrated and developed with 0.2 M pyridine acetate buffer, pH 3.1, at 23". The effluent was collected in 1.9ml fractions.
Peptides were located by ninhydrin analysis after alkaline hydrolysis.
The fractions indicated by the bars were combined.
Solid bars indicate efhucnt pools containing cystine peptides. The column was developed at 15 ml per hour and the effluent was collected in 2-ml fractions.
Peptides were located by ninhydrin reaction after alkaline hydrolysis.
The fractions indicated by the bars were combined.
The solid bars indicate efiluent containing cystine peptides.
paper electrophoresis. The identification of the peptides is based on the amino acid composition, determinat'ion of amino terminal residue by subtractive Edman degradation, and is interpreted in terms of the sequence of the inhibitor.
No further information was required for the unique assignment of these peptides within the amino acid sequence of bovine PSTI. The 6 half-cystine residues of the inhibitor have been assigned Roman numerals starting from the amino-terminal portion of the molecule.
DisulJide Bond I-V-Cystine Peptide II-8 was coeluted with the tetrapeptide Leu-Gly-Arg-Glu from the Dowex 50-X2 column (Fig. 4, bottom). However, this contaminating peptide was effectively separated on Dowex 50 from cysteic acid peptides after performic acid oxidation of the mixture (Fig. 5,  bottom). The elution position of the tetrapeptide was unchanged because it was not susceptible to performic acid oxidation.
The yield of Peptide II-8 was 55%. The amino acid compositions and results of subtractive Edman degradation are given in Table II. These data show that a disulfide bond links residues 9 and 38.

II
Disulfide bond I-V of bovine PSTI Amino acid composition and subtractive Edman degradation. of cystine Peptide II-8 (Fig. 4, bottom) and Peptides 11-8-A and II-S-B by performic acid oxidation (Fig. 5, bottom). The numbers in parentheses are theoretical and were obtained by sequence analysis (3). Boldface type is used to show the residue released by one cycle of subtractive Edman degradation. formic acid oxidation of I-4 by chromatography on Dowex 50-X2 (Fig. 5, top). The analytical data for these peptides are given in Table III. The results shorn that a disulfide bond connects residues 16 and 35.
A mixture of related peptides, I-3, was also isolated in 19% yield (Fig. 4, top). I-3 had exactly the same amino acid composition as Peptide I-4. The cysteic acid fragments had also the same amino acid composition as I-4-A and I-4-B but had different elution positions from the Dowex 50-X2 column. These peptides probably represent. a mixture of products result, Structure of Bovine Pancreatic Xecretory Trypsin Inhibitor. III Vol. 246,No. 24 ing from deamidation, imide or /?-peptide bond formation in-acid composition and results of subtractive Edman degradation volving residues 14 to 15 (Asn-Gly) and residues 33 to 34 (Asn-are given in Table IV. The data show that a disulfide bond Glu) (cf. References 31, 32).
connects residues 24 and 56 in the inhibitor. Dim&de Bond III-VI-Two cystine peptides, I-l and 11-3, A two-dimensional schematic diagram of the structure of corresponding to bond III-VI were released by the action of bovine PSTI showing the arrangement of the disulfide bonds and thermolysin on bovine PSTI.
The peptides were homogeneous the sequence of the amino acids is given in Fig. 6. The amino after chromatography on Dowex 50-X2 (Fig. 4) and were iso-acid residues given in the mottled circles indicate the cystine peplated in 31 and 60% yield, respectively. After oxidation of the tides characterized in this study. The yields of these peptides, cystine peptides with performic acid the cysteic acid peptides based on the amount of extensively hydrolyzed inhibitor (Fracwere purified by high voltage paper electrophoresis. The amino tion 2, Fig. 2) are: I-V, 55%; II-IV, 44%; and III-VI, 91%.  The amino acid composition of Fraction 1 (Fig. 2) corresponded to that of bovine PST1 less -0.7 eq of aspartic acid and isoleucine because of the loss of the amino-terminal dipeptide Asn-Ile (cf. Table I). The sites of cleavage in this mixture of peptides were determined by oxidation with performic acid followed by separation of the fragments by gel filtration on a 200-cm column of Sephadex G-50 (Fig. 7, left panel).
The low molecular weight material, Peak B, was a mixture of peptides corresponding to residues 1 to 12 and 3 to 12 in the molar ratio 1:2 (cf. Table V). The peptides in Fraction A were partially resolved by a second passage through a column of Sephadex G-50 with an effective length of 400 cm (cj. Fig. 7, right panel). The data given in Table V show that A-l corresponds to residues 3 to 56 plus a small amount of peptide corresponding to residues 1 to 56, and that A-2 contains residues 13 to 56. These assignments were confirmed by the results of subtractive Edman degradation where the amino-terminal residues of A-l and A-2 were shown to be primarily leucine and valine.
On the basis of the amounts of each peptide recovered, the following composition can be given for the mixture of partially hydrolyzed inhibitors: 20% corresponds to a cleavage only at residues 12 to 13; 35% is inhibitor with only the bond at residues 2 to 3 hydrolyzed; 40% of the molecules have both bonds, residues 2 to 3 and 12 to 13 cleaved; and approximately 5% is intact inhibitor. On the basis of the relative simplicity of the elution diagrams and the analytical data we conclude: (a) the partially hydrolyzed inhibitor is a mixture of peptides with bonds cleaved at residues 2 to 3 (Ile-Leu) and residues 12 to 13 (Glu-Val); and (b) no other peptide bonds are cleaved in the early stages of hydrolysis.  Fig. 2). Left panel, the sample contained 0.45 pmole of inhibitor. The column, 0.9 X 200 cm, was equilibrated and developed with 0.2 M pyridine acetate buffer, pH 3.1. Right panel, Peak A (left panel) was submitted to gel filtration a second time with a column with an effective length of 400 cm. All other conditions were the same as given for left panel. Peptides were located by ninhydrin reaction after alkaline hydrolysis.
Fractions indicated by the bars were combined. order to determine whether the partially hydrolyzed inhibitors were intermediates in the production of extensively hydrolyzed inhibitor.
On the basis of the Sephadcx G-75 elution diagram, amino acid composition and results of one cycle of subtractive Edman degradation, the extensively hydrolyzed material, obtained in 50~~ yield, corresponded to Fraction 2 of the first hydrolysis.
These results show that Fraction 1 is an int'ermediate in the production of Fraction 2 and indicate that the initial event in the hydrolysis of intact bovine PST1 is the cleavage of two bonds at residues 2 to 3 and 12 to 13, followed by the slow cleavage of one or more bonds leading to the extensive degradation of the molecule.

Peptides
Peptides devoid of cystine were present in every fraction of the Sephadex G-25 effluent (Fig. 3). Most of these peptides in Fractions I and II were obtained in homogeneous form during the purification of cystine peptides (Fig. 4) and cysteic acid peptides (Fig. 5, bottom).
Fraction III contained a tetrapeptide and the peptides in Fractions IV and V were isolated by high voltage paper electrophoresis.
These peptides and their yields are given in Table VI. The peptides marked with asterisks had integral molar ratios of the constituent amino acids. The others, although somewhat contaminated, could be identified on the basis of the sequence of bovine PSTI.
The free amino acids  (Fraction 2,Fig. 2) Peptides were identified on the basis of amino acid composition in terms of the amino acid sequence of the inhibitor (3). The asterisk indicates peptides having integral molar ratio of constituent amino acid. The remaining peptides, although present in mixtures were identifiable because of differences in concentration. The fraction number indicates the method of preparation of the peptide and the numbers in parentheses give the electrophoretic mobility relative to glycine except for Peptides IV-l and IV-2 which are related to aspartic acid. The yields of peptide are based on the amount of extensively hydrolyzed inhibitor listed in the table mere deterrnincd directly on the amino acid analyzer on a sample which had not been hydrolyzed.

Sites of Thermolysin Hydrolysis
The peptide bonds cleaved by thermolysin at plI 6.5, 37", for 48 hours are given in Fig. 8. The extent of cleavage is based on extensively hydrolyzed inhibitor (Fraction 2). These numbers must be considered minimum values because they were calculated on the basis of the yield of recovered and identifiable peptides. DISCUGSION The most striking feature of these results is that thermolysin, acting on bovine PSTI, produced a small number of cystine peptides in high yield: 55, 44, and 91% for disulfide bridges I-V, II-IV, and III-VI, respectively. The use of column chromatographic procedures was advantageous not only for the preparation of the cystine peptides but also for the recognition and isolation of partially hydrolyzed inhibitor as well as the identification and characterization of noncystine peptides. The cysteic acid peptides were located in the amino acid sequence of bovine PST1 on the basis of amino acid composition and end group determination.

Furthermore,
we have not found evidence for heterogeneity of disulfide bonds either in homogeneous peptides or mixtures of peptides.
We estimate that we would have detected 10% of disulfide heterogeneity, if present. Evidence for amide heterogeneity was found for asparagine, residues 14 and 33 (Peptide I-3) but this may have been the result of the acidic conditions used for gel filtrat'ion and Dowex 50 chromatography (31). The sites of peptide bond cleavage and the extent of hydrolysis by thermolysin acting on bovine PST1 are summarized in Fig. 8. Peptide bonds involving the amino group of leucine, isoleucine, valine, alanine, serine, threonine, and methionine were readily hydrolyzed whereas tyrosine , glutamic acid, and glycine were cleaved less rapidly.
It is noteworthy that the peptide bond Pro-Val was hydrolyzed to the extent of 91 and 52% for residues 22 to 23 and 47 to 48, respectively.
The hydrolysis of sequence Cys-Leu-Leu-Cys (residues 35 to 38) was crucial for the determination of the disulfide bonds. Thermolysin hydrolyzed the Cys-Leu and Leu-Leu bonds to the extent of at least 68 and 55%, respectively.
In preliminary experiments we attempted t,o digest intact inhibitor with pepsin at pH 2.2 followed by trypsin at pH 6.5. Only one disulfide bond, III-VI, could be identified, presumably because of the lack of hydrolysis of Cys-Leu-Leu-Cys sequence by pepsin and the difficulty of separating closely related peptides resulting from incomplete hydrolysis (27). The specificity of thermolysin hydrolysis of intact bovine PST1 at pH 6.5 is similar to that reported by other investigators who used thermolysin at pH 8 to 8.5 on cytochrome c (33), azurin (34), Kunitz' inhibitor (35), and peptides (35-38). The first application of the use of thermolysin for the determination of disuhide bonds was reported by Jentsch for bovine insulin (39). An important difference in the experimental procedure is that we conducted the thermolysin hydrolysis at pH 6.5 to reduce disulfide interchange (25-28). Jent'sch (39) digested insulin at pH 8.0 but did not present data on disulfide interchange.
The hydrolysis of intact bovine PST1 by thermolysin leads to the formation of high molecular weight, intermediates having only one or two peptide bonds cleaved per molecule and a mixture of small peptides.
Presumably the initial cleavages take place at exposed or partially unfolded species and hydrolysis at the third or subsequent bonds destabilizes the structure, leading by guest on July 9, 2020 http://www.jbc.org/ to exposure of more bonds to proteolytic activity, finally resulting in the formation of extensively degraded peptides (cf. Reference 40). On this basis, residues 2 to 3 and 12 to 13 of bovine PST1 are exposed and accessible to the action of thermolysin (cj. Fig. 6). Similar limited proteolysis reactions occur when ribonuclease is treated with subtilisin or elastase leading to the formation of ribonuclease S (41) and ribonuclease E (42, 43). We are attempting t.o separate the modified forms of bovine PST1 present in Fraction 1 for further chemical, physical, and inhibitor studies.
In the absence of crystallographic information, an analysis of the significance of the two-dimensional structure given in Fig. 6 must necessarily be incomplete.
However, certain features deserve comment.
The carboxyl-terminal cysteine residue VI, is linked to the cysteine III at position 24. The other two disulfide bonds, I-V and II-IV, form a ring containing 12 amino acid residues.
The disposition of the disulfide bonds appears to give the molecule a compact structure which may be responsible in part for its stability at acid pH and elevated temperature (13). The reactive site (44) of the molecule, Arg-Ile (residues 18 to 19), is located within a disulfide loop of 9 amino acid residues between cystines II and III (45). After hydrolysis of the Arg-Ile bond, both fragments remain attached by two other disulfide bonds, I-V and II-IV.
This structure supports the generalization of Laskowski and Sealock (6) and Ozawa and Laskowski (44) t'hat reactive site of inhibitors is located between cystine residues. The 6 half-cystine residues of cow, pig (24, 46) and sheep (47) pancreatic secretory trypsin inhibitors occupy the same positions in the amino acid sequences of these polypeptides.
On this basis it would be expected lhat t'he disposition of the disulfide bonds of pig and sheep PST1 will be the same as found for bovine PSTI.2