Somatostatin-28 Encoded in a Cloned cDNA Obtained from a Rat Medullary Thyroid Carcinoma*

We have constructed and cloned in bacteria comple- mentary DNAs derived from a transplantable rat medullary thyroid carcinoma. Using a hybridization probe encoding an anglerfish islet pre-prosomatostatin, a pre- cursor of the tetradecapeptide somatostatin, we have identified and isolated a clone containing a 400-base pair complementary DNA encoding most of the rat carcinoma pre-prosomatostatin. The amino acid se- quence of the tetradecapeptide somatostatin and of the amino-terminally extended form, somatostatin-28 was deduced from the nucleotide sequence of the complementary DNA. Somatostatin-28 was found at the COOH terminus of a polypeptide of at least 80 amino acids indicating that somatostatin-28 arises by cleavage from a large precursor. The sequences of somatostatin-28 and somatostatin-14 are strictly conserved between the rat and other mammals. Such conservation of these sequences indicates strong selective pressures during evolution to maintain the sequence and suggests that somatostatin-28 may serve some essential biologic functions apart from, or in addition to, the important regulatory actions of somatostatin-14. Additionally, we found a high degree of homology in the amino acid sequences of the NHz-terminal extension peptides in the anglerfish islet and the rat carcinoma pre-proso-matostatins pointing further to a possible biologic func- tion of these described previously (22). The nucleotide sequences of overlapping end-labeled DNA fragments were determined by the method of Maxam and Gilbert (29).

We have constructed and cloned in bacteria complementary DNAs derived from a transplantable rat medullary thyroid carcinoma. Using a hybridization probe encoding an anglerfish islet pre-prosomatostatin, a precursor of the tetradecapeptide somatostatin, we have identified and isolated a clone containing a 400-base pair complementary DNA encoding most of the rat carcinoma pre-prosomatostatin. The amino acid sequence of the tetradecapeptide somatostatin and of the amino-terminally extended form, somatostatin-28 was deduced from the nucleotide sequence of the complementary DNA. Somatostatin-28 was found at the COOH terminus of a polypeptide of at least 80 amino acids indicating that somatostatin-28 arises by cleavage from a large precursor. The sequences of somatostatin-28 and somatostatin-14 are strictly conserved between the rat and other mammals. Such conservation of these sequences indicates strong selective pressures during evolution to maintain the sequence and suggests that somatostatin-28 may serve some essential biologic functions apart from, or in addition to, the important regulatory actions of somatostatin-14. Additionally, we found a high degree of homology in the amino acid sequences of the NHz-terminal extension peptides in the anglerfish islet and the rat carcinoma pre-prosomatostatins pointing further to a possible biologic function of these extension peptides.
Somatostatin is a tetradecapeptide that regulates the release of pituitary, pancreatic, and gastrointestinal hormones (1). Initially identified in the hypothalamus as an inhibitor of growth hormone secretion (2), somatostatin has subsequently been found in extrahypothalamic brain, spinal cord, retina, gastrointestinal tract, pancreatic islets, and thyroid (3-6). In addition to inhibiting the secretion of a number of peptide hormones, somatostatin has been proposed to act as a neurotransmitter and to modulate gastrointestinal motility (7,8).
The diverse functions and the widespread distribution of the tetradecapeptide somatostatin (somatostatin-14) have focused attention on the biosynthesis of the hormone. Several studies have shown that somatostatin-14 is synthesized as part of a larger precursor. A 28-amino acid form of the hormone (somatostatin-28) has recently been identified in extracts of porcine hypothalamus (9), gastrointestinal tract (IO), and ovine hypothalamus (1 1). Somatostatin-28 may have functions distinct from those of the tetradecapeptide (12-14).
* This work was supported in part by grants from the National Institutes of Health and the National Science Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Pulse and pulse-chase labeling studies in brain and pancreatic islets indicate that somatostatin-28 is also derived from a larger precursor (15,16). Cell-free translations of mRNA isolated from the pancreatic islets and gastrointestinal tissues of anglerfish (17)(18)(19), the pancreatic islets of channel catfish (20) and rat hypothalami (21) have confirmed the existence of large (Mr = 14,000 to 16,000) precursors of somatostatin (preprosomatostatins). Recently, nucleotide sequencing of cloned complementary DNAs (cDNAs) have provided the amino acid sequences of two anglerfish islet pre-prosomatostatins (22,23). These sequences show that the somatostatin-14 peptides are located at the COOH termini of 119 to 121 amino acid precursors and that the fish and mammalian somatostatin-14 peptides have identical sequences. However, only partial conservation of the somatostatin-28 sequence was observed between fish and mammals. Little is known about the structures of the peptides that lie NH2-terminal to the somatostatin-28 sequence of mammalian pre-prosomatostatins.
We now report the use of a cloned cDNA containing coding sequences for an anglerfish islet pre-prosomatostatin as a hybridization probe to identify a cloned somatostatin-related cDNA derived from a transplantable rat medullary carcinoma of the thyroid. Nucleotide sequencing of the cDNA revealed the complete sequence of rat somatostatin-28 and the partial sequence of the NH2-terminal precursor extension of the preprosomatostatin. We find that the amino acid sequence of somatostatin-14 and somatostatin-28 in the rat are identical with those found in the ovine and porcine species. In addition, comparison of the amino acid sequences of the NH2-terminal extensions encoded in the 5' regions of the fish and rat cDNAs show considerable homology (53% among nucleotides and 39% among amino acids). The latter observations point to the existence of strong selective pressures to conserve the structures of the NH2-terminal extensions over the 400 million years since fish and rat diverged in evolution. These findings suggest that these peptide extensions may have some, as yet unrecognized, biologic function other than simply to serve as protein "spacer" sequences.

EXPERIMENTAL PROCEDURES
Construction and Cloning of cDNAs from a Rat Medullary Carcinoma of the Thyroid-The preparation of a cloned cDNA library from a rat medullary carcinoma of the thyroid kindly provided by N. H. Bell was described previously (24). In brief, cDNA was prepared from the polyadenylated RNA (25) by using an oligo(dT) primer and reverse transcriptase (26). Double-stranded DNA was prepared from the cDNA with polymerase I, inserted into the Pst I restriction endonuclease site of the plasmid pBR322 (26), and recombinant plasmids were introduced into Escherichia coli x1776 by the procedure of Villa Komaroff et al. (27).
Identification of Somatostatin-related Medullary Thyroid Carcinoma cDNA-Approximately 2,000 cloned cDNAs derived from the rat medullary thyroid carcinoma were screened using a modification

Somatostatin-28 Precursor
of the procedure of Grunstein and Hogness (28). Cloned cDNA encoding anglerfiih pre-prosomatostatin (22) was digested with the restriction endonuclease Xma 1 and Dde 1, labeled at the 3'-ends with :12P (22), and electrophoresed on 5% polyacrylamide gels containing Tris/borate/NanEDTA. A 58-base pair DNA fragment encoding the tetradecapeptide somatostatin was isolated from the gels and used as a hybridization probe.
Bacterial colonies containing recombinant cDNAs derived from rat medullary thyroid carcinoma were grown on nitrocellulose filters. Bacteria were lysed and the DNA was fixed onto the filters as described by Grunstein and Hogness (28). Filters were prehybridized for 18 h a t 68°C in a solution containing 6 x SSC (0.9 M NaCI, 0.09 M sodium citrate), 5 X Denhardt's reagent (0.1% w/v each of bovine serum albumin, polyvinylpyrrolidone, and Ficoll), 0.5% sodium dode-cy1 sulfate, 5 pg/ml of sonicated denatured salmon sperm and E. coli DNA, and 10% dextran sulfate.
After prehybridization, the buffer was discarded and replaced by an identical solution containing heat-denatured hybridization probe. Filters were allowed to hybridize for 2 h a t 68°C and were subsequently washed 10 times a t 68°C with a solution containing 5 X SSC, 1 X Denhardt's reagent, and 0.5% sodium dodecyl sulfate. Filters were then washed twice a t room temperature in 0.03 X SSC and air-dried. Autoradiograms were prepared using Kodak X-0-Mat film exposed for 2 to 3 days a t -90 "C using a Dupont Cronex intensifying screen.
Nucleotide Sequence Analysis of Cloned cDNAs-Restriction fragments of the thyroid carcinoma cDNA encoding somatostatin were labeled using T4 polynucleotide kinase or terminal transferase as described previously (22). The nucleotide sequences of overlapping end-labeled DNA fragments were determined by the method of Maxam and Gilbert (29).

RESULTS
Immunoreactive somatostatin has been identified in extracts of transplantable rat medullary thyroid carcinomas (30,31). Preliminary studies using immunoprecipitation and hybridization selection techniques were unsuccessful, however, in identifying somatostatin precursors from the cell-free translation products programed by medullary thyroid carcinoma mRNAs.' We had previously constructed a cDNA library from a rat medullary thyroid carcinoma for the purpose of isolating and sequencing CDXA encoding a precursor of calcitonin (24). By using a cloned cDNA encoding a somatostatin precursor from anglerfish islets we were able to identify a rat somatostatin-related cDNA in the carcinoma cDNA library by the colony hybridization method. Only one clone, 400 base pairs in length, was identified out of approximately 2,000 clones which were screened. The paucity of clones containing somatostatin-related cDNAs reflects the low levels of somatostatin found in these tumors (30).
The identification of sites which are cleaved by the restriction endonucleases Xma 1, Pst 1, and Rsa 1 proved particularly useful in determining the nucleotide sequence of the cDNA. Restriction fragments were sequenced on both the sense and nonsense strands (Fig. 1).
The nucleotide sequence of the cloned cDNA and the corresponding amino acid sequence is shown in Fig. 2. This cDNA encodes the sequence of the tetradecapeptide somatostatin and somatostatin-28 in addition to a 51-amino acid NH2-terminal extension. As predicted by Patzelt et al. (32), the tetradecapeptide somatostatin is located at the COOH terminus of the precursor, followed by a stop codon TAG. The 14 amino acids predicted by the nucleotide sequence which precede the tetradecapeptide are identical with those found in ovine and porcine somatostatin-28 (9-11).
Somewhat surprising was the nucleotide sequence coding for Leu-Gln-Arg which separates somatostatin-28 from the remainder of the precursor (Figs. 2 and 3). This sequence is reminiscent of the sequence Leu-Glu-Arg found at the analogous position in the anglerfish pre-prosomatostatins (23). In-' R. H. Goodman, unpublished data.

C C T C C C T T * C *~\~T T C~A C I C T C T C T~~~T~C~A -A T T A T~= T C~~~T T~T C~~~~~, ( A ) .
FIG. 2. Nucleotide sequence of a cDNA encoding a rat prosomatostatin. The corresponding amino acid sequence is shown below the nucleotide sequence. The region of the cDNA encoding somatostatin-I4 and somatostatin-28 is indicated by the enclosed sequence preceding the stop codon shown enclosed in the hexagon. The sequence Leu-Gln-Arg (underlined) is similar to the Leu-Glu-Arg sequence at the analogous position in the anglerfish somatostatin precursors (see Fig. 4) and may represent an atypical cleavage site. The argininyl lysine sequence that separates somatostatin-14 from the NH2-terminal region of somatostatin-28 is shown enclosed in circles. Sequence Asn-Gln-Thr (underlined) is a potential N-glycosylation site. AATAAA sequence (underlined) within the 3' untranslated region of the cDNA is characteristic of eukaryotic mRNAs.
FIG. 3. Nucleotide sequence analysis of cloned cUNA coding for the junction between somatostatin-Zd and the NHAerminal extension of prosomatostatin. cDNA was labeled at the 5'-end and subjected to four base-specific chemical cleavage reactions as described by Maxam and Gilbert (29). The cleavage products were analyzed on an 8% polyacrylamide sequencing gel. Sequence depicted on the nonsense strand encodes the amino acids Glu-Leu-Gln-Arg. Nucleotide sequence was confirmed on the opposite strand and also by demonstration of the specific cleavage of the sequence CTGCAG by the restriction endonuclease Pst I. by guest on March 17, 2020 http://www.jbc.org/ Downloaded from asmuch as the sites of cleavages typically found in prohormones consist of combinations of two of three of the basic amino acids arginine and lysine (33), our sequence raises the question of whether somatostatin-28 is produced by posttranslational cleavages of prosomatostatin in the rat medullary thyroid carcinoma.
The sequence Asn-Gln-Thr within the NH2-terminal extension of the rat somatostatin precursors represents a potential N-glycosylation site (34). Patzelt et al. have suggested that rat prosomatostatin may be glycosylated (32). No analogous potential glycosylation sites are present within the two anglerfish pre-prosomatostatins.
Inasmuch as there were differences between the sequences of anglerfish pre-prosomatostatin I reported by our laboratory (22) and that of Hobart and co-workers (23), we have extensively reanalyzed the sequence of the anglerfish somatostatin precursor. The revised sequence, which is slightly different from either of the two sequences reported previously is shown in Fig. 4. Fig. 4 additionally compares the nucleotide and amino acid sequences of the anglerfish and rat somatostatin precursors. Within the coding sequence for the tetradecapeptide somatostatin, 35 of the 42 nucleotides (83%) are conserved between rat and anglerfish. Within the coding sequence for somatostatin-28, 22 of the 28 amino acids (79%), and 58 of the 84 nucleotides (69%) are maintained. The six amino acid substitutions between the fish and rat somatostatin-28 sequences appear to be conservative in nature. It is necessary to add or delete specific codons to maintain the homology between the NH2-terminal extensions of the fish and rat sequences. If these deletions and additions are made, 20 of the 51 amino acids within the extension (39%) are strictly conserved and an additional 15 amino acid changes are conservative in nature. Eighty-one of 153 nucleotides (53%) within this region are conserved. With the exception of the AATAAA sequence and nil?

FIG. 4. Comparison of the anglerfish and rat somatostatin precursors.
Sequences corresponding to somatostatin-14 and -28 are indicated. Boxes denote amino acids strictly conserved between the fish and rat precursors. Dots refer to nucleotides conserved between the two cDNAs. To maintain homology between the NHz-terminal positions of the precursors, it was necessary to delete three codons from the rat cDNA and one codon from the fish cDNA. the polyadenylate tail, there appears to be little conservation of nucleotide sequence within the 3' untranslated regions.

DISCUSSION
A cloned cDNA of 400 base pairs coding for pre-prosomatostatin, a precursor of rat somatostatin, was identified by colony hybridization using a 32P-labeled cDNA encoding an anglerfish islet pre-prosomatostatin. Previous studies have indicated that rat hypothalamic and pancreatic somatostatin-14 are identical in amino acid composition and chromatographic properties with those isolated from ovine and porcine hypothalamus and pigeon and anglerfish pancreas (35, 36). The nucleotide sequence of our cDNA indicates that the primary structure of somatostatin-14 in rat is identical with that of these other species. This sequence differs from that of catfish islet somatostatin (37) and anglerfih islet somatostatin I1 (23). The amino acid sequence derived from the nucleotide sequence of our cDNA indicates that the somatostatin-like immunoreactive material isolated from the medullary thyroid carcinoma is authentic somatostatin.
The cDNA which we have characterized encodes the entire 3' untranslated region, the sequence of somatostatin-28 at the COOH terminus of the precursor, and a portion of the NH2terminal peptide extension. Estimates of the size of prosomatostatin isolated from rat pancreatic islets (32) and of preprosomatostatin from cell-free translations of rat hypothalamic mRNA (21), indicate that rat pre-prosomatostatin contains approximately 110 amino acids. Inasmuch as the NH2terminal leader or signal sequence of the pre-prohormone probably includes 24 to 28 amino acids, our partial cDNA encoding 79 amino acids represents nearly the entire rat prosomatostatin sequence.
Evidence regarding the processing of prosomatostatin in the rat medullary thyroid carcinoma has been conflicting. Berelowitz et al. (30) have suggested that the predominant form of somatostatin in their tumor was the tetradecapeptide. Benoit et al. (38) concluded that somatostatin-28 constituted the major immunoreactive form of the hormone. It was of interest therefore to examine the peptide sequence adjacent to somatostatin-28. The sequence Gln-Arg predicted from our cDNA differs from that of a typical prohormone cleavage site. It is possible that such a change in amino acid sequence at this site might affect the relative efficiency of processing to somatostatin-28 or somatostatin-14. It is conceivable in fact that the regulation of the processing of multigenic hormones such as somatostatin may involve production of separate prohormones with different cleavage sites. Further studies correlating the sequences of somatostatin cDNAs derived from particular tissues with the somatostatin-28 to somatostatin-14 ratios should elucidate this hypothesis. The possibility of a reverse transcriptase error during the construction of the cDNA library must also be considered. Although we were unable to c o n f i i the nucleotide sequence on an independently cloned cDNA due to the low abundance of the preprosomatostatin cDNAs in the cDNA library prepared from the carcinoma, eventual determination of the genomic sequence should provide such confiiation.
The revised sequence of the anglerfish pre-prosomatostatin cDNA depicted in Fig. 4 more closely resembles that of Hobart et al. (23) than the sequence which we initially reported (22). This difference is primarily due to a sequencing error resulting in a frame shift involving 17 of the 121 codons. Nonetheless, there remain a few codons which differ from those reported by Hobart et al. (23). These differences could result from reverse transcription errors or from the sequencing of distinct polymorphic cDNAs.
The sequence of porcine and ovine somatostatin-28 has previously been determined (9-1 1). Recent evidence suggests that the biologic activities of somatostatin-28 may be greater than, and perhaps different from, those of somatostatin-14 (12)(13)(14). Strict conservation of the amino acid sequence of somatostatin-28 between the rat and these other mammals, species which diverged over 75 million years ago (39), is strong evidence for the existence of evolutionary pressures to maintain this sequence. It is likely, therefore, that somatostatin-28 has an important biologic function in mammals. Conservation between anglerfiih and mammals of the six amino acids adjacent to the tetradecapeptide somatostatin suggests a particular importance of this portion of somatostatin-28 (Fig. 5).
Examination of the NHz-terminal portions of the fish and rat somatostatin precursors reveals several additional regions of considerable homology. This observation raises the possibility that the NHz-terminal portion of prosomatostatin may also have some specific biologic functions. Comparison of somatostatin cDNAs from other species should be useful in understanding the importance of this region. We are currently synthesizing by chemical methods peptide fragments of the NHn-terminal region of rat prosomatostatin for the preparation of antisera to be used in studies of intracellular transport, secretion, and potential biologic activity of the precursor region of the prosomatostatin. Such studies should further our understanding of neuropeptide biosynthesis and physiology. 14.