Purification and Amino Acid Sequence of a Noncalcitonin Secretory Peptide Derived from Preprocalcitonin*

A previous report from this laboratory (Birnbaum R. S., O’Neil, J. A., M., D. C., and ROOS, B. A. (1982) J. BioL Chem. 257, 241-244) provided im- munochemical and biochemical evidence for the exist-ence of a secretory peptide derived from the noncalci- tonin region of rat preprocalcitonin. By a variety of criteria, we demonstrated that this naturally occurring peptide was similar, if not identical, to a synthetic peptide which consisted of the NH2-terminal 16 residues of the calcitonin mRNA translation product. We have now purified this peptide from rat medullary thyroid carcinoma and sequenced it. A rat tumor of the 1-2-4 tumor series was extracted in 0.1 N HC1 yielding 900 pg of immunoreactive peptide. The peptide was purified to homogeneity by: 1) trichloroacetic acid precipitation of contaminating protein; 2) gel filtration; and finally, 3) reverse phase high pressure liquid chromatography. Overall yield was approximately 24%. Amino acid analysis and sequencing of the peptide yielded a composition and sequence identical with that of the synthetic peptide. Calcitonin, a 32-residue amidated peptide hormone, is pro-duced in C cells of the thyroid by post-translational processing of larger precursors (1-6). While the complete amino acid sequence of the precursor awaits elucidation, an inferred

A rat tumor of the 1-2-4 tumor series was extracted in 0.1 N HC1 yielding 900 pg of immunoreactive peptide. The peptide was purified to homogeneity by: 1) trichloroacetic acid precipitation of contaminating protein; 2) gel filtration; and finally, 3) reverse phase high press u r e liquid chromatography. Overall yield was approximately 24%. Amino acid analysis and sequencing of the peptide yielded a composition and sequence identical with that of the synthetic peptide.
Calcitonin, a 32-residue amidated peptide hormone, is produced in C cells of the thyroid by post-translational processing of larger precursors (1)(2)(3)(4)(5)(6). While the complete amino acid sequence of the precursor awaits elucidation, an inferred amino acid sequence has been proposed (6,7). Based upon information obtained from the nucleic acid sequence of a cDNA derived from calcitonin mRNA, it appears that the mature calcitonin hormone is nestled within a 136-residue precursor. Within this precursor, the first residue of mature calcitonin is preceded by the pair of amino acids, Lys-Arg, which are in turn preceded by 82 additional amino acids. On the distal side of the mature hormone is a 20-residue sequence beginning with the tetrapeptide Gly-Lys-Lys-Arg (Fig. 1). This tetrapeptide sequence is also found in precursors of a-MSH' (8) where it separates a-MSH (ACTHI1-13)amide) ' The abbreviations are: a-MSH, melanocyte-stimulating hormone; ACTH, corticotropin; HPLC, high pressure liquid chromatography; Pth, phenylthiohydantoin; y-MSH, the peptide sequence corresponding to residues -55 to -44 of the predicted sequence of bovine preproopiomelanocortin (8); CCAP, calcitonin-carboxyl-adjacent peptide.
from corticotropin-like intermediate lobe protein (ACTH( 18-39)). By analogy to a-MSH synthesis, we suggested that a peptide corresponding to the fiial hexadecapeptide constitutes a noncalcitonin secretory peptide derived from preprocalcitonin (9). In our initial studies (9) to detect this hexadecapeptide, termed CCAP for calcitonin-carboxyl-adjacent peptide, we showed that a peptide with sue and isoelectric point indistinguishable from synthetic CCAP was detected by an immunoassay based on the synthetic peptide. We also showed that the CCAP and calcitonin content of calcitonin-containing tissues were approximately equimolar. Higher molecular weight immunoreactive forms could also be demonstrated.
However, the possibility existed that the low molecular weight immunoreactive peptide detected in calcitonin-producing tissues was not identical with the synthetic peptide. Posttranslational modification of a slightly larger peptide, for example acetylation of an Arg-CCAP (Fig. l), could result in chromatographic and electrophoretic behavior similar to synthetic CCAP. In addition, there was a need to confirm the nucleic acid sequence for this important portion of preprocalcitonin. We also noted reports in the literature of discrepancies between the cDNA-derived amino acid sequence of human P-lipotropin in preproopiomelanocortin (1 1) and the actual sequence of human pituitary ,&lipotropin (12, 13). Accordingly, we decided to purify the peptide from tumor tissue and determine its amino acid sequence.

EXPERIMENTAL PROCEDURES
Materials-Synthetic CCAP was obtained as previously reported (9). Sephadex G-25 (Fine) was obtained from Pharmacia. Acetonitrile (UV grade) was purchased from Burdick and Jackson Laboratories, Inc. Sequanal-grade trifluoroacetic acid, Quadrol, and Polybrene were obtained from Pierce Chemical Co.
Radioirnmunoassuy-The assay for CCAP was performed as described previously (9) except that phase separation was accomplished using a formalin-treated heat-denatured preparation of Staphylococcus aureus, Cowan I strain (Iggsorb, The Enzyme Center) (14). The assay standards were based on amino acid analysis of synthetic CCAP.
Tissue Extraction-The starting material for the purification of tumor CCAP was 29 g of nonnecrotic tissue of the 1-2-4 rat medullary thyroid carcinoma tumor series (15). The tissue was cut into large chunks and disrupted in 6 volumes of ice-cold 0.1 N HCI using a Brinkmann Polytron. All subsequent steps were performed at 4 "C. The extract was centrifuged for 10 min a t 2200 X g. Trichloroacetic acid was added slowly to the supernatant to a final concentration of 5%. The suspension was centrifuged as before and the resulting supernatant extracted 3 times with equal volumes of water-saturated diethyl ether. The solution was then lyophilized.
Gel Filtration-A column (0.9 X 114 cm) of Sephadex G-25 (Fine) CCAP was abbreviated CAP in a previous communication from this laboratory (9); it has also been called CCP, for COOH-terminal designated PDN-21 (10).
cleavage peptide (6)  was equilibrated in 0.05 M ammonium carbonate, pH 8.0, at 4 "C. A lyophilized residue containing immunoreactive CCAP was dissolved in 1% trifhoroacetic acid (16) and applied to the column. 2-ml fractions were collected and assayed for immunoreactive CCAP and protein by the method of Lowry et al. (17).
HPLC-For the purification of immunoreactive CCAP a modular system consisting of a Spectra-Physics 8700 solvent delivery system with a dynamic mixer, a Waters Associates U6K injector, a Kratos SF769Z detector, and a Hewlett-Packard 3390A reporting integrator was assembled. Samples of about 200 p1 were chromatographed on either a Waters CM pBondapak analytical column (46 x 250 m m ) or a Whatman Protesil 300 Ca analytical column. Peptides were separated by a 1525% acetonitrile gradient in 0.1% trifluoroacetic acid at a flow rate of 1 ml/min. Pth derivatives were identified by HPLC using a Beckman 345 ternary liquid chromatograph equipped with a Beckman 501 autosampler, a Beckman 165 variable wavelength detector, and a Hewlett-Packard 3390A reporting integrator.
Amino Acid Analysis-Samples were dissolved in 6 M HC1, degassed, and heated in a sealed ampoule to 110 "C for 24 h. The sample was lyophilized and the residue was dissolved in sodium citrate buffer and applied to a Durmm model 500 amino acid analyzer.
Sequence Analysis-CCAP (19 nmol) was degraded in a Beckman 890D Sequencer according to the method of Edman and Begg (18), using the 0.1 M Quadrol program of Beckman (No. 345801). Sequenator chemicals were prepared as previously described (19,20) with the exception that 0.1 M Quadrol was used as buffer. Prior to the addition of the peptide, 2 mg of Polybrene were dissolved in 0.7 ml of 50% acetic acid and applied to the cup of the sequenator, dried under vacuum, and subjected to three complete cycles of automated Edman degradation (21). CCAP was then introduced into the cup and sequentially degraded under the direction of the same program. The resulting samples from the sequenator were converted to the corresponding Pth derivatives as previously described (19).
Pth Derivative Identification and Quantitation-The mobile phase employed for the separation and identification of the Pth derivatives was a modification of the method of Zimmerman et al. (22). A two-solvent system was employed consisting of an aqueous 40 mM solution of sodium acetate (pH 4.90) to which 100% acetonitrile was added to a concentration of 22.5% (by volume) and a solution of 40 mM triethylamine acetate (pH 4.5) in 80% acetonitrile. A Beckman Ultrasphere ODS column (46 X 250 mm) was used at a temperature of 50 "C and at a flow rate of 0.8 ml/min. Detection of the Pth derivatives was performed at 269 nm except for the time period between 15.5 and 18.5 min. During this 3-min interval, the wavelength was automatically changed to 313 nm for the purpose of detecting the dithioerythritol adduct of dehydrothreonine.
Yields of Pth derivatives were determined as previously described (19). For the peptide described herein, at least 80% of the expected amount was degraded. In no cycle did the ratio of signal to noise fall below 3, stepwise yields ranged from 95 to 97%, and only one sequence was observed.

RESULTS
Two factors facilitated the purification of CCAP. The availability of large transplantable medullary thyroid carcinomas allowed US to begin purification of the peptide from a larger amount of tissue than could conveniently be obtained using rat thyroids. The 1-2-4 series tumor selected as starting material contained greater than 10-fold more immunoreactive CCAP per g of tissue than normal thyroid (9). T h u s 29 g of tissue from a single 1-2-4 series tumor was extracted in hydrochloric acid yielding 900 pg of immunoreactive CCAP in 1.44 g of protein. The second factor was the solubility of the peptide in 5% trichloroacetic acid. Approximately two-thirds of the immunoreactive CCAP remained in the supernatant after acid treatment, whereas 75% of the protein was precipitated.
The resulting lyophilized residue was applied to a Sephadex G-25 column (Fig. 2). Immunoreactive CCAP eluted with a K d of 0.2 and was separated from the bulk of the protein which eluted at t h e void volume. The specific activity of t h e peak tubes of immunoreactive CCAP was about 700 pg/mg of protein. Even though synthetic CCAP gives a value of 840 pg/ mg of protein in the Lowry assay, the protein values, except for the void volume fractions, probably did not reflect the true amount of contaminating protein. The impurity of this material became apparent when sequencing attempts yielded essentially all Pth derivatives in the first sequenator cycle.
The peak tubes of immunoreactive CCAP from Sephadex gel filtration were pooled and lyophilized.
The residue was dissolved in 1% trifluoroacetic acid (16) and subjected to reverse phase chromatography. Initially we used the 60 Apore size CIX pBondapak column; later the 300 A-pore size CX Protesil column was substituted (23, 24). The protein and immunoassay profiles were very similar for the two columns; however, only a run using the Clx column is depicted in Fig.  3. The only significant difference between the two columns was the yield of immunoreactive CCAP. Approximately 50-60% was recovered using the CIS column while recovery was essentially 100% for the Cx column.
Two peaks of immunoreactive CCAP were observed, which corresponded to two of the UV-absorbing peaks seen with synthetic CCAP (not shown).
The major immunoreactive peak seen in Fig. 3 eluted with the same retention time as that of the major peak of the synthetic material, 25 min. The minor peak eluting at about minute 18 has been tentatively identified as methionine-sulfoxide CCAP based upon: 1) its shorter retention time; 2) its similarity in size to the species eluting at minute 25; 3) its reduction with 2.9 M mercaptoethylamine at 37 "C for 42 h which shifted the retention time

5465
to that of the major CCAP peak; and 4) its amino acid analysis following acid hydrolysis showed no differences from the major peak. Rechromatography of the major peak under isocratic conditions (17.5% acetonitrile) revealed no contaminating material (not shown). Overall recovery of CCAP from tumor extract to HPLC-purified material was about 24%. The purification is summarized in Table I. The amino acid content was determined by analysis of an acid hydrolysate of the major CCAP peak (Table 11). No differences in amino acid content were noted from that predicted from the cDNA sequence or from amino acid analysis of synthetic CCAP. In particular, there was only a single lysine and no arginine, ruling out the possibility that the naturally occurring peptide contained the sequence of the synthetic material with an NHB-terminal extension.  Sequence analysis was performed on 19 nmol of purified CCAP. The identity and yield of the 16 amino acids found are shown in Table 111. No P t h derivative was detected at Cycle 17. Extrapolation of the yield to the initial amount (or Cycle 0) gives approximately 19 nmol, indicating that the sequence of only one peptide form of CCAP was present.

DISCUSSION
The CCAP-like peptide isolated from the calcitonin-producing rat medullary thyroid carcinoma is identical with the synthetic hexadecapeptide and the amino acid sequence predicted from the cDNA structure. Nothing was observed during purification or analysis to suggest that CCAP was either larger than the synthetic peptide (due to an NHa-terminal extension) or had undergone such post-translational modifications as phosphorylation, glycosylation, acetylation, or sulfation. We cannot, however, rule out the possibility that certain minor forms of CCAP might be so modified.
The detection, characterization, and subsequent isolation of CCAP resulted from predictions of the processing of preprocalcitonin based upon a sequence inferred from analysis of a cDNA (6,7). Our success confirms this approach to identifying novel secretory peptides, first demonstrated with the y-MSH region of preproopiomelanocortin (8,25,26). However, the   " The residue at the NH2 terminus after enzymatic removal of the "codon." significance of a glycine residue preceding the two arginine residues was not initially appreciated as an indication of a n amidated peptide (27). Extension of this approach to other hormones or secretory peptides is likely to reveal many more novel secretory products. The sequencing of tumor CCAP also confirms our original working hypothesis that processing of the calcitonin precursor is analogous to the processing of a-MSH precursors in the intermediate lobe of the pituitary. From the known or predicted structures of amidated secretory peptide precursors (Table IV), certain generalizations can be made. A glycine residue must follow the carboxyl-terminal residue of the amidated peptide. Then either a di-or tribasic segment follows. The residue which is amidated and the first amino acid of the cleaved product do not appear to be restricted. Proteolytic cleavage of the peptide chain appears to precede amidation since Bradbury et al. (34) have reported that D-Tyr-Val-Gly could be converted to D-Tyr-Val-CoNHZ by pituitary homogenates. However, this finding has not yet been confirmed.
Finally, it is apparent from differences in the processing of ACTH precursors in the anterior and intermediate lobes of the pituitary that the endopeptidase which cleaves the Gly-Lys-Lys-Arg segment is distinct from the trypsin-like enzyme which acts at the Lys-Arg dipeptides flanking ACTH.
CCAP bears no homology to any known bioact,ive peptide. The sequence, therefore, contains no obvious clues to a biological role. However, CCAP's coelaboration with calcitonin does suggest the possibility of some function associated with skeletal and mineral homeostasis. Recently, a report has appeared that t,he analogous human synthetic peptide, whose sequence was also based upon analysis of a cDNA derived from human medullary thyroid carcinoma, decreased blood calcium in rats but had no effect on phosphate excretion (10).