T H E JOURNAL OF BIOLOGICAL CHEMISTRY G 1989 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 264, No. 36, Issue of December 25, pp. 21529-21535,1989 Printed in U.S.A.

Thyrotropin-releasing Hormone (TRH) Precursor Processing CHARACTERIZATION OF MATURE TRH AND NON-TRH PEPTIDES SYNTHESIZED BY TRANSFECTED MAMMALIAN CELLS*

(Received for publication, July 27, 1989)

Kevin A. SevarinoS, Richard H. Goodman, Joachim Spiessg, Ivor M. D. Jacksonq, and Ping Wull From the Division of Molecular Medicine, Department of Medicine, New England Medical Center, Boston, Massachusetts 02111, the §Department of Molecular Neuroendocrinology, Max Planck Institute of Experimental Medicine, Hermann-Rein-Strasse 3, 3400 Gottingen, Federal Republic oj Germany, and the llDivision of Endocrinology, Brown University, Rhode Island Hospital, Providence, Rhode Island 02903

Prepro-thyrotropin-releasing hormone (TRH) con- tains five TRH progenitor sequences and at least six other potential peptides (Lechan, R. M., Wu, P., Jack- son, I. M. D., Wolf, H., Cooperman, s., Mandel, G., and Goodman, R. H. (1986a) Science 231, 159-161). Pre- vious studies using radioimmunoassays developed against discrete regions of prepro-TRH have demon- strated that several of the potential peptides are pres- ent in rat brain and pancreas (Wu, P., Lechan, R. M., and Jackson, I. M. D. (1987) Endocrinology 121,108- 115; Wu, P. and Jackson, I. M. D. (1988a) Bruin Res. 456, 22-28; Wu, P., and Jackson, I. M. D. (1988b) Regul. Pept. 22, 347-360). However, the low level of peptides present in intact tissues has made isolation of the peptides difficult. CA77 cells, a medullary thyroid carcinoma cell line, also express prepro-TRH and dis- play processing similar to that found in tissues. How- ever, peptide content in this tumor cell line is enhanced only %fold compared with normal tissues (Sevarino, K. A., Wu, P., Jackson, I. M. D., ROOS, B. A., Mandel, G., and Goodman, R. H. (1988) J. Biol. Chern. 263, 620-623). To achieve higher levels of expression for facilitating peptide sequencing studies and to see if alternate processing of prepro-TRH could be detected in different cell types, we transfected into 3T3, GH4, AtT20, and RIN 5F cells a cDNA vector under control of the cytomegalovirus immediate-early promoter. 3T3 and GH4 cells failed to process prepro-TRH be- yond cleavage of the signal sequence. Both AtT20 and RIN 5F cells efficiently cleaved the precursor at di- basic sites to generate mature TRH and the non-TRH peptides previously identified in vivo. Peptide content was up to 30 times greater than in hypothalamic ex- tracts and 10 times greater than in CA77 cells. Secre- tion experiments with transfected AtT2O cells demon- strated that both mature TRH and the non-TRH pep- tides were secreted via a regulated secretory pathway similar to that utilized by endogenously synthesized peptides. We isolated several of the non-TRH peptides synthesized by transfected AtT2O cells and character- ized these peptides by sequential Edman degradation. These studies identified the signal sequence cleavage site and determined that the non-TRH peptides are

* Support for these studies was provided through National Science Foundation Grant BNS 8706694 (to R. H. G. and K. A. S.), National Institutes of Health Grant DK 34540 ( to I. M. D. J. and P. W.), and the Max Planck Research Program (to J. S.). The costs ofpublication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Recipient of a postdoctoral fellowship from the Juvenile Diabetes Foundation.

generated by cleavage at the dibasic sites flanking the five TRH progenitor sequences. Further, we deter- mined that processing occurs at the ArgS1-ArgS2 site located in the amino-terminal portion of the precursor, the only dibasic site not flanking a TRH progenitor sequence.

The sequence of the precursor for thyrotropin-releasing hormone (TRH)’ has been deduced from the sequence of its cDNA (Lechan et al., 1986a). Five TRH progenitor sequences (Gln-His-Pro-Gly) flanked by paired basic residues are con- tained within the 255-amino acid precursor. The remainder of the precursor is comprised of a signal peptide, an amino- terminal flanking sequence containing a paired arginine se- quence, four intervening peptides, and a carboxyl-terminal flanking sequence (Fig. 1). Mature TRH, pyroGlu-His-Pro- amide, is formed by excising the progenitor sequence at the paired basic residues (Griffiths et al., 1983), trimming the basic residues with a carboxypeptidase B-like enzyme (Gainer et al., 1985), cyclizing the amino-terminal glutamine residue (Fischer and Spiess, 1987), and amidating the carboxyl-ter- minal proline residue (Bradbury et al., 1982; Kizer et al., 1984; Eipper and Mains, 1988). By generating antisera to discrete portions of the TRH precursor, several of the non-TRH- containing peptides have been detected in rat brain, pancreas, and thyroid (Wu et al., 1987; Wu and Jackson, 1988a, 1988b; Bulant et al., 1988; Gkonos et al., 1989). However, precise identification of the peptides derived from prepro-TRH has been difficult because of the low level of prepro-TRH biosyn- thesis and the diffuse distribution of TRH-producing cells in vivo.

The rat medullary thyroid carcinoma cell line CA77 has been shown to synthesize TRH and other prepro-TRH-de- rived products (Sevarino et al., 1988). These cells process prepro-TRH similarly to rat hypothalamus. However, the CA77 cell content of non-TRH-containing peptides is only three times that of intact tissues. To obtain higher levels of prepro-TRH expression and to see if variations in prepro- TRH processing could be detected, we introduced into heter- ologous cell lines the prepro-TRH cDNA under control of the

The abbreviations used are: TRH, thyrotropin-releasing hor- mone; CMV IE, cytomegalovirus immediate-early; ipYT, immuno- reactive pYT22; ipCC, immunoreactive pCC10; ipYE27, immuno- reactive ipYE27; ipYE17, immunoreactive ipYE17; ipEH24, immu- noreactive ipEH24; BSA, bovine serum albumin; CA, carbonic anhydrase; CC, bovine cytochrome c; Apr, aprotinin; RIA, radio- immunoassay; HPLC, high pressure liquid chromatography; pyroGlu, pyroglutamic acid.

21529

21530 Prepro-TRH Processing in Transfected Cell Lines

cytomegalovirus immediate-early (CMV IE) promoter. Expression in fibroblastic 3T3 cells, which typically fail to

process transfected precursors (Thomas et al., 1988), was examined to rule out nonspecific breakdown of the precursor under cell culture conditions. Rat somatotrophic GH4 cells were studied because they are capable of very high levels of neuropeptide cDNA expression (Sevarino et al., 1987). How- ever, they display variable processing of transfected precur- sors. Preproparathyroid hormone and preproglucagon are processed by GH4 cells (Hellerman et al., 1984; Drucker et al., 1986), while preprosomatostatin and prepro-opiomelano- cortin are not (Sevarino et al., 1987; Thomas et al., 1988). Mouse corticotrophic AtT20 cells and rat insulinoma RIN 5F cells efficiently process foreign precursors including preproin- suiin, preprosomatostatin, and preproglucagon (Thomas et al., 1988). We examined the processing of prepro-TRH in both AtT20 and RIN 5F cells because AtT2O cells appear to recognize monobasic processing sites more efficiently than RIN 5F cells (Sevarino et al., 1987).

In this study we demonstrated that AtT20 and RIN 5F cells process prepro-TRH to mature TRH and to several non-TRH peptides detected i n uiuo. Processing appeared to occur at all dibasic sites, and processing at monobasic sites was not ob- served. Both mature TRH and the non-TRH peptides were targeted to a regulated secretory pathway. The cell contents of the non-TRH peptides in transfected cells were up to 30 times higher than in hypothalamic extracts (Wu et al., 1987). We were able to isolate and sequence three peptides derived from various portions of the TRH precursor to identify these peptides positively, to define the probable signal sequence cleavage site, and to characterize several internal proteolytic sites.

EXPERIMENTAL PROCEDURES~

RESULTS

Processing of Prepro-TRH to Mature TRH Is Restricted to Certain Cell Types-We expressed the prepro-TRH cDNA in GH4, 3T3, RIN 5F, and AtT2O cells. Approximately 25% of the G418-resistant clones contained a predominant prepro- TRH mRNA species of 1750 bases, 150 bases larger than native prepro-TRH mRNA isolated from rat hypothalamus or isolated from CA77 cells (Fig. 3). This difference reflects altered transcriptional termination and polyadenylation due to the substitution of the normal 3‘ end of the prepro-TRH gene by the SV40 3’-flanking signals. One AtT2O clone ex- pressed a prepro-TRH mRNA of approximately 3000 bases (Fig. 3, lane I ) , but there were no differences in the patterns of processing or secretion between this and any of the other AtT2O clones. Untransfected cells contained no detectable prepro-TRH mRNA. Transfected cells were also screened for secretion of products containing the sequence pYT22 (ipYT). All cells that expressed prepro-TRH mRNA also secreted ipYT. Wild-type cells secreted no ipYT.

To screen for accurate processing to mature products, clones were assayed for secretion of mature TRH (iTRH). Transfected 3T3 and GH4 cells synthesized trace levels of a material cross-reacting in the TRH assay which did not dilute in parallel to the TRH standard. This material was also detected in the media of wild-type 3T3 and GH4 cells and is unlikely to represent authentic TRH. In contrast, transfected

Portions of this paper (including “Experimental Procedures,” Tables 1-111, and Figures 2-7) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

1 2 4 Arg$rg TRH TRH TRH TRH TRH 255 EJmEJm m I

Signal

00 0 pYE27 pYT22 pEH24

mmmmrm 7kDa ipYT22 2.5 kDa ipEH24

mn pCC10 pYE17

6 kDa ipYE17

mmmmmmommmmm 4 kDa ipYE27 3 kDa ipYT22

FIG. 1. Schematic diagram of prepro-TRH deduced from its complimentary DNA sequence. The predicted signal peptide (hor- izontal stripes) and the five TRH progenitor sequences, Gln-His-Pro- Gly, flanked by paired basic amino acid sequences (oblique stripes), are marked (Lechan et al., 1986a). A further potential Arg-Arg cleav- age sequence is present in the amino-terminal sequence of pro-TRH (arrow). The positions of synthetic peptides used in this study are shown below the prepro-TRH sequence; black areas indicate substi- tutions or extensions to the cDNA-deduced sequence to allow radio- iodination or cyclization (Jackson et al., 1985). The peptides corre- spond to the prepro-TRH sequence as follows: pYT22 corresponds to (Tyr’) prepro-TRH 53-74, pYE27 corresponds to (Tyro) prepro-TRH 25-50, pEH24 corresponds to prepro-TRH 83-106, pYE17 corre- sponds to (Tyro) prepro-TRH 241-255, and pYS23 corresponds to (Tyro) prepro-TRH 165-186. The peptide pCCl0 corresponds to (Cys”.’) prepro-TRH 75-82, and also (Cyso.’) 200-207 and (Cys”, Lys’, Cy$) 107-114, 152-159, and 170-177 prepro-TRH. Furthest below the prepro-TRH schematic are boxes (uertical stripes) representing the non-TRH peptides synthesized by transfected AtT2O cells and identified in this study. The ipCC peptides are not shown as their exact location is unknown.

AtT2O and RIN 5F clones expressing prepro-TRH mRNA synthesized high levels of authentic iTRH. Untransfected AtT2O and RIN 5F cells did not secrete any material cross- reacting in the TRH assay.

Clones of each cell type expressing the CMV-TRH-SV40 vector were expanded for further study. The intracellular contents and the products secreted over 24 h were extracted and assayed for ipYT, peptides containing the TRH progen- itor sequence pCCl0 (ipCC), and iTRH. The results for one representative clone of each cell type are displayed in Table 1. The media from both 3T3 and GH4 clones displaced the tracer in the ipYT and ipCC radioimmunoassays. These im- munoreactivities did not dilute in parallel to their respective standards and were not quantifiable. As noted above, trace levels of material cross-reactive in the TRH assay were de- tected as in untransfected cells. AtT20 cells synthesized high levels of iPYT and mature TRH and stored both to a signif- icant degree (greater than 80% of the immunoreactivity pres- ent after 24 h was intracellular). Low levels of ipCC were detected in the medium, but this immunoreactivity did not dilute in parallel to the standard curve. Three separate AtT2O clones with a wide range of prepro-TRH expression were examined. Each clone displayed a similar ratio of ipYT to iTRH and the same degree of intracellular storage. RIN 5F clones were similar to AtTXO clones but synthesized lower levels of iTRH. Further, RIN 5F clones stored prepro-TRH- derived peptides less efficiently than AtT20 clones. Four individual RIN 5F clones with a wide range of prepro-TRH expression were examined, and the clones displayed very similar ratios of ipYT and iTRH and levels of intracellular storage.

Transfected 3T3 and GH4 Cells Synthesize and Secrete Unprocessed Pro-TRH-The products secreted by transfected GH4 and 3T3 cells were fractionated on Sephadex G-75 columns. Both cell types synthesized a high molecular mass species of approximately 30,000 daltons reactive in the pYT22, pCC10, pEH24, and the carboxyl-terminal pYE17

Prepro-TRH Processing in Transfected Cell Lines 21531

radioimmunoassays (data not shown). The predicted amino terminus of pro-TRH (Lechan et al., 1986a) contains the pYE27 sequence. However, the pYE27 assay requires the free carboxyl terminus of pYE27 for reactivity (Wu and Jackson, 1988a). The free carboxyl terminus would be generated by cleavage at the Ar$l-Ar$* dibasic site in prepro-TRH, and thus the predicted intact pro-TRH would not react in the pYE27 assay. The high molecular weight species isolated from the media of transfected GH4 and 3T3 cells was negative for ipYE27, and no smaller species of ipYE27 was detected. We conclude that the high molecular weight product secreted by transfected GH4 and 3T3 cells represents the intact prohor- mone. The same species was detected in transfected GH4 cell extracts.

Transfected AtT20 and RIN 5F Cells Synthesize Pro-TRH- derived Products Identical to Those Synthesized in Vivo-The immunoreactive products from clones of transfected RIN 5F and AtT20 cells were fractionated on Sephadex G-50 columns. RIN 5F (Fig. 4) and AtT2O (Fig. 5) clones produced the ipYT peptides (3-kDa ipYT and 7-kDa ipYT), ipCC peptides (6- kDa ipCC and 15-kDa ipCC), and ipYE17 peptide (6-kDa ipYE17) detected in rat hypothalamus (Wu et al., 1987). Further characterization demonstrated that transfected AtT2O cells synthesized 2.5-kDa ipEH24 (Fig. 5, panel D ) and 4-kDa ipYE27 (Fig. 7, panel A ) , peptides previously detected in rat brain (Wu and Jackson, 1988a). Unprocessed pro-TRH was not detected in either RIN 5F or AtT2O extracts.

No significant differences were found between intracellular and secreted peptides except for the ipYT peptides synthe- sized by transfected RIN 5F cells (Fig. 4, panels B and D). For each of the three clones analyzed (rKS6-4, rKS6-9, and rKS6-24), the ratio of 7-kDa ipYT to 3-kDa ipYT was less than 0.7 for the intracellular products. In contrast, this ratio was greater than 2.0 for the secreted products.

Mature TRH and Non-TRH Peptides Are Stored in Regu- lated Intracellular Pools-To determine if any of the pro- TRH-derived peptides was secreted via a different pathway from that utilized by the other pro-TRH-derivedpeptides, the basal (Table 2) and pharmacologically stimulated (Fig. 6) secretory rates of each peptide were measured. Under unstim- ulated conditions, the molar amounts of intracellular and secreted ipYE27, ipEH24, and ipYEl7 were approximately equal. Compared with the other peptides, lower molar amounts of ipYT were detected in the media and cells. This deficit was likely due to adsorption losses to which the ipYT peptides are prone.3 Cell extracts contained approximately twice as much iTRH on a molar basis (assayed before car- tridge purification) as the other pro-TRH-derived peptides, while basal secretion of iTRH was nearly equal to that of the intervening peptides. The release of TRH and other pro- TRH-derived peptides from transfected AtT20 cells was stim- ulated 4-fold by potassium over 1 h (Fig. 6). This release was calcium-dependent. Norepinephrine increased the release of TRH and non-TRH peptides 3-fold in 1 h. The degree of secretogogue stimulation was similar for each of the non- TRH peptides and mature TRH. Equivalent results were obtained in secretion experiments conducted over 3-h epochs.

Isolation and Sequencing of Non-TRH Peptides-Three peptides, 4-kDa ipYE27, 3-kDa ipYT, and 6-kDa ipYE17, were isolated and sequenced to define the cleavage sites used to generate these peptides. The purification of 4-kDa ipYE27 is summarized in Table 3 and Fig. 7. Similar results were obtained for the other two peptides. The determined sequence of each peptide is aligned with the predicted sequence of prepro-TRH in Fig. 8. The 26 residues of 4-kDa ipYE27 are

' P. Wu, unpublished observations.

identical to the amino-terminal 26 residues predicted for the sequence of pro-TRH. Although there was a trace amount of the arginyl derivative detected in the 27th cycle, the ipYE27 radioimmunoassay requires a free carboxyl-terminal gluta- mate residue for recognition (Wu and Jackson, 1988a). Thus, the large majority of the isolated peptide does not contain this final arginine residue. We conclude that 4-kDa ipYE27 is generated by cleavages at the predicted signal sequence cleavage site AlaZ4-Leuz5 and preceding the dibasic Arg5'-Arg52 site. Sequence analyses of purified 3-kDa ipYT and 6-kDa ipYE17 further indicate that transfected AtT20 cells cleave prepro-TRH carboxyl-terminal to the dibasic Ar$'-Ar$* and the LysZo6-Arp207 sites to generate these non-TRH peptides.

DISCUSSION

Dynamic studies of TRH biosynthesis have been hampered by the low levels of TRH production in intact tissues and primary culture systems (McKelvy, 1975). In addition these sources contain heterogeneous cell populations, only a small proportion of which may be synthesizing TRH. The nonclonal cell line CA77 also synthesizes prepro-TRH (Sevarino et al., 1988). These cells process prepro-TRH similarly to the pat- tern seen i n vivo, and their content of pro-TRH-derived peptides is only about three times that found in rat hypothal- amus. The availability of the cDNA for prepro-TRH has allowed us to establish lines of cultured cells expressing high levels of prepro-TRH. These studies provide the first defini- tive evidence that the isolated prepro-TRH cDNA (Lechan et al., 1986a) is the authentic TRH precursor. The intracellular levels of ipYT and iTRH in transfected AtT2O and RIN 5F cells were 1-2 orders of magnitude greater than in tissue extracts and CA77 cells. These levels of expression were stable for over 50 cell passages.

The processing of prepro-TRH in the cell types used here was similar to that reported for other neuropeptide precursors (Thomas et al., 1988). 3T3 and GH4 cells failed to process prepro-TRH beyond signal sequence cleavage, while AtT2O and RIN 5F cells efficiently processed prepro-TRH to the same products found in intact rat tissues. While AtT20 cells were better able than RIN 5F cells to recognize the monobasic Arg-I5 processing site in preprosomatostatin (Sevarino et al., 1987), no processing following single arginine or lysine resi- dues was observed for prepro-TRH expressed in AtT20 cells. A main difference between RIN 5F and AtT2O cells was the lower level of mature TRH synthesized by RIN 5F cells. This difference may reflect an actual difference in cell content of TRH maturation enzymes or may reflect a higher metabolic requirement of RIN 5F cells for a TRH maturation substrate, such as ascorbate required for a-amidation (Eipper and Mains, 1988). We also observed that transfected RIN 5F cells synthesized both 15-kDa ipCC and 6-kDa ipCC peptides while transfected AtT2O cells synthesized only 6-kDa ipCC. It is not known whether this difference results from 15-kDa ipCC being a short-lived intermediate in AtT20 cells or whether AtT20 cells process prepro-TRH differently than RIN 5F cells such that 15-kDa ipCC is never produced. Despite these differences, our data indicate that in the main AtT20 and RIN 5F cells process prepro-TRH similarly to rat hypothal- amus (Wu et al., 1987; Wu and Jackson, 1988a).

The secretion studies indicate that TRH and other pro- TRH-derived peptides are stored in regulated secretory pools in transfected cells and have similar rates of secretion. The only indication of differential targeting was seen for the 3- kDa ipYT and 7-kDa ipYT peptides synthesized by trans- fected RIN 5F cells. Under basal conditions, 7-kDa ipYT appeared to be secreted more efficiently than 3-kDa ipYT.

21532 Prepro-TRH Processing in Transfected Cell Lines

2 4 v 5 3 0 35 4 0 4 5

PreproTRH H L P E A A Q E E G A V T P D L P G L E N V Q V R P E D 5 1 52

ipYE27 L P E A A Q E E G A V T P D L P G L E N V Q V R P E

pmol 153 113 108 125 165 84 72 89 45 59 40 10 40 23 29 34 23 27 22 19 12 14 13 12 11 3

60 65

PreproTRH

ipYT22 F L L Q V G L G A

pmol 30 17 5 6 4 5 3 7 3

ipYE17 A L G H P C G P Q G T Q T G L Q L L D L S

pmol 66 75 79 5 20 19 41 17 17 26 9 24 11 23 10 8 12 14 5 9 2

FIG. 8. Sequences of purified 4-kDa ipYE27, 3-kDa ipYT22, and 6-kDa ipYE17. Above each deter- mined sequence a portion of the cDNA-derived sequence of prepro-TRH is given in one-letter amino acid code, numbered with respect to prepro-TRH in italics above the sequence. Adjacent paired basic residue sequences are shown in shaded boxes, with the processing sites denoted by arrows. The amino-terminal sequences of the purified peptides, 4-kDa ipYE27 ( top) , 3-kDa ipYT22 (middle), and 6-kDa ipYE17 (bot tom), are given below each predicted sequence. Cycles where yields were insufficient to provide definite identifications are left blank. Yields in picomoles (pmol) are provided below each identified residue. The peptides were isolated from the AtT2O clone aKS6-19 and sequenced as described under "Experimental Procedures." The predicted signal sequence cleavage site follows Alaz4. The sequencing confirms that the dibasic Ar$'-Arg62 site i s the carboxyl-terminal processing site for 4-kDa ipYE27 and the amino-terminal processing site for 3-kDa ipYT22. The dibasic sequence Lyszo6-Argo7 is the amino- terminal processing site used to generate 6-kDa ipYE17.

However, the half-lives of the two peptide species in the medium and their individual responses to secretogogues will have to be determined before this difference can be deemed significant. We also found that secretogogues caused an equiv- alent stimulation of TRH, non-TRH peptide, and ACTH secretion from transfected AtT2O cells (data not shown). These observations indicate that TRH and the non-TRH peptides are secreted via a similar pathway to that used by endogenous neuropeptide products. The highly efficient intra- cellular storage of prepro-TRH-derived peptides by AtT2O clones and the less efficient storage displayed by RIN 5F clones parallel that found for these cell types when they express foreign preprosomatostatin (Sevarino et al., 1987).

Using transfected cells as a source of pro-TRH-derived peptides we purified and sequenced three non-TRH peptides. The complete sequence of 4-kDa ipYE27 aligned with residues 25-50 of the TRH precursor. Thus, cleavage occurs at the predicted signal sequence processing site AiaZ4-LeuZ5 (von Heijne, 1983) and preceding the internal Arg51-Arg5' sequence to generate 4-kDa ipYE27. Alternatively, we could have uti- lized the classic approach of in vitro translation of prepro- TRH mRNA in the presence and absence of microsomal membranes to define the signal sequence cleavage site. Fur- ther evidence that the internal Arg-Arg processing site is utilized was provided by the partial sequence of the peptide 3-kDa ipYT, whose amino terminus was found to be generated by cleavage immediately carboxyl-terminal to this site. It is likely that 7-kDa ipYT is an amino-terminally extended form of 3-kDa ipYT which has not undergone processing at the Arg-Arg site. Finally, it was determined that the carboxyl-

terminal 6-kDa ipYEl7 peptide is generated by cleavage fol- lowing the Arg206-L~~"~ dibasic site. Taken with the observed sizes of the non-TRH peptides, these sequence data support the hypothesis that processing of prepro-TRH occurs exclu- sively at dibasic sites (Wu et al., 1987).

Complete processing of pro-TRH at paired basic amino acid sequences would generate five TRH molecules/precursor. The ratios of mature TRH to each non-TRH peptide were less than 5:l in cell extracts and secreted products. There are several possible explanations for this discrepancy. The TRH assay is specific for mature TRH, needing both the amino- terminal pyroGlu and carboxyl-terminal Pro-amide for rec- ognition (Jackson and Reichlin, 1974). Thus, the relatively low amount of iTRH observed may result from partial proc- essing of TRH progenitor sequences. These nonimmunoreac- tive TRH progenitor sequences could retain untrimmed basic amino acid residues, uncyclized amino-terminal glutaminyl residues, or carboxyl-terminal glycine residues. It is also pos- sible that not all TRH progenitor sequences are cleaved from the precursor, resulting in extended peptides containing TRH progenitor sequences linked to amino and/or carboxyl-ter- minal extensions. Such extended TRH sequences would not react in the TRH assay. Indeed, 6-kDa and 15-kDa ipCC probably represent extended peptides that retain TRH pro- genitor sequences. Bulant et al. (1988) have also detected a prepro-TRH-derived intervening peptide which retains an uncleaved TRH progenitor sequence at one terminus. I t is intriguing to speculate upon the possible biological role of such extended sequences. By analogy with the extended en- kephalin sequences produced by selective cleavage of prepro-

Prepro-TRH Processing in Transfected Cell Lines 21533

dynorphin (Zamir et al., 1984), such peptides may increase the diversity of biologically active molecules derived from the TRH precursor. Other explanations for the low level of TRH observed might be that TRH is degraded more rapidly than the other pro-TRH-derived peptides or that TRH is extracted less efficiently from cells than these other peptides.

Quantitation of the ipCC peptides was not possible because of nonparallel dilution of the immunoreactivity relative to the standard curve. The low levels of ipCC detected may result from extended TRH progenitor sequences being present at low concentrations or because extended forms of the progen- itor sequence react poorly in the pCCl0 assay. From fraction- ation studies it was clear that the actual progenitor species Gln-His-Pro-Gly-Lys/Arg-Lys, or species nearly as small, were not present. Further, ipCC is not detectable immunohis- tochemically in the median eminence or in the axons and peripheral regions of neuron perikarya in the hypothalamus (Jackson et al., 1985; Lechan et al., 1986b; Wu and Jackson, 1988a). These data indicate that processing at paired basic residue sites and subsequent maturation of progenitor se- quences occurs relatively early in the passage of pro-TRH through the neural secretory pathway.

Our work and that of others (Bulant et al., 1988; Gkonos et al., 1989) have demonstrated that all of the regions flanking the TRH sequences in prepro-TRH are processed to stable peptides in tissues (brain, pancreas, and thyroid) and in cultured cells. These peptides have unique distributions in uiuo which suggests that they possess distinct biological roles (Lechan et al., 1987; Bulant et al., 1988). Thus, it will be important to elucidate the dynamics of TRH and non-TRH peptide biosynthesis from prepro-TRH. Clonal cell lines ex- pressing prepro-TRH are at present the most tractable models for studies of prepro-TRH processing and secretion. Process- ing studies using pulse-chase experiments with native and mutated cDNAs coding for prepro-TRH may provide a means to define the maturation pathways of prepro-TRH.

Acknowledgments-We gratefully acknowledge the technical as- sistance of R. Felix, A. Ronco, s. Patrick, and C. Melucci. We thank Drs. K. Stewart and S. Grubman for assistance with the CMV-TRH- SV40 construction and Drs. J. Kopchick and M. Tocci, Merck, Sharp, and Dohme, Rahway, NJ, for the gift of CMV-BglII.

REFERENCES

Bradbury, A. F., Finnie, M. D. A,, and Smyth, D. G. (1982) Nature

Bulant, M., Delfour, A., Vaudry, H., and Nicolas, P. (1988) J. Biol. 298,686-688

Chem. 263,17189-17196

Drucker, D. J., Mojsov, S., and Habener, J. F. (1986) J . Biol. Chem.

Eipper, B. A., and Mains, R. E. (1988) Annu. Reu. Physiol. 50, 333-

Fischer. W. F.. and SDiess, J . (1987) Proc. Natl. Acad. Sci. U. S. A.

26 1,9637-9643

344 - .

84, 3628-3632 Gainer, H., Russell, J. T., and Loh, Y. P. (1985) Neuroendocrinology

40,176-184 Gkonos, P. J., Tavianini, M. A., and Roos, B. A. (1989) 71st Annual

Meeting of the Endocrine Society, Seattle, WA, The Endocrine Society, Bethesda, MD, Abstract, p. 337

Griffiths, E. C., Millar, R. P., and McKelvey, J . F. (1983) in Thyro- tropin Releasing Hormone (Griffiths, E. C., and Bennett, G. W., eds) pp. 45-49, Raven Press, New York

Hellerman, J. G., Cone, R. C., Potts, J., Rich, A., Mulligan, R. C., and Kronenberg, H. M. (1984) Proc. Natl. Acad. Sci. U. S. A. 8 1 ,

Jackson, I. M. D., and Reichlin, S. (1974) Endocrinology 95,854-862 Jackson, I. M. D., Wu, P., and Lechan, R. M. (1985) Science 229 ,

Kizer, J. S., Busby, W. H., Cottle, C., and Youngblood, W. W. (1984) Proc. Natl. Acad. Sci. U. S. A. 8 1 , 3228-3232

Lechan, R. M., Wu, P., Jackson, I. M. D., Wolf, H., Cooperman, S., Mandel, G., and Goodman, R. H. (1986a) Science 231,159-161

Lechan, R. M., Wu, P., and Jackson, I. M. D. (1986b) Endocrinology

Lechan, R. M., Wu, P., and Jackson, I. M. D. (1987) Endocrinology

McKelvy, J. F., Sheridan, M., Joseph, C. , Phelps, C. H., and Perrie,

Pasleau, F., Tocci, M. J., Leung, F., and Kopchick, J. J. (1985) Gene

Petrarca, M. A., Reiss, L., Huang, A. S., Burakoff, S. S., and Faller, D. V. (1989) Microbiol. Pathogenesis, in press

Sevarino, K. A., Felix, R., Banks, C. M., Low, M. J., Montminy, M. R., Mandel, G., and Goodman, R. H. (1987) J . Biol. Chem. 262 ,

Sevarino, K. A., Wu, P., Jackson, I. M. D., Roos, B. A., Mandel, G., and Goodman, R. H. (1988) J. Biol. Chem. 263.620-623

Spiess, J . (1986) in Advanced Methods in Protein Sequence Analysis (Wittmann-Liebold, B., Salnikoff, J., and Erdmann, V. A., eds) pp. 302-315, Springer-Verlag, Berlin

Thomas, G., Thorne, B. A., and Hruby, D. E. (1988) Annu. Reu. Physiol. 50, 323-332

Vale, W., Vaughan, J., Yamamoto, G., Bruhn, T., Douglas, C., Dalton, D., Rivier, C., and Rivier, J . (1983) Methods Enzymol. 103 , 565- 577

5340-5344

1097-1099

119,1210-1216

121,1879-1891

S. (1975) Endocrinology 97, 908-918

(Amst.) 38,227-232

4987-4993

von Heijne, G. (1983) Eur. J. Biochem. 133 , 17-21 Wu, P., and Jackson, I. M. D. (1988a) Brain Res. 456, 22-28 Wu, P., and Jackson, I. M. D. (1988b) Regul. Pept. 22,347-360 Wu, P., Ackland, J. F., Ling, N., and Jackson, I. M. D. (1986) Regul.

Pept. 15, 311-321 Wu, P., Lechan, R. M., and Jackson, I. M. D. (1987) Endocrinology

121,108-115 Zamir, N., Weber, E., Palkovits, M., and Brownstein, M. (1984) Proc.

Natl. Acad. Sci. U. S. A. 81, 6886-6889

Continued on next page.

21534 Prepro-TRH Processing in Transfected Cell Lines SUPPLEMENTARY MATERIAI

m Table 1 Immunoreactivity profiles of transfected cell lines

THYROTROPIN RELEASING b0RMONE PRECURSOR PROCESSING, CHARACTERUATKNOF MATURE TRH AND NONTRH PEPTIDES SYNTHESIZED BY TRANSFECTED MAMMALIAN CELLS

Kevtn A Sevanno, Rlchard H. Goodman, Joachlm Spless. lvor M D. Jackson. and Pmg Wu

EXPERIMENTAL PROCEDURES

by Inserting a 650 bp ECO RI fragment of the plasmid pLW-4.2 (Lechan el a/., 1966a), - The expresslon vector CMV-TRH-SV40 (Fog. 2) was constructed

adapted with Bam HI linkers, and an 850 bp Bgl I1 - Bam HI fragment of pxf-MoA

COnStrUCtlOn places the TRH coding sequence and SV40 small t antigen polyadenylation (Petrarca et a/., 1989) into the plasmld CMV-BgI I1 (Pasleau et a/., 1985). Thls

and transcrlptlonal termination sognals downstream from the CMV IE promoter

and TranSfection - Cell 11nes were malntatned as prevlously described (Sevarlno et a/., 1987) except that AtT20 cells were mamtalned following translectlon In Dulbecco's-modifled Eagle's essential medlum/lO% fetal calf serum Transfections were performed by calclum phosphate co-preclpitation of 20 pg of the TRH expressfon vector CMV-TRH-SV4O and 2 Fg of pRSVneo per lo6 cells (Sevanno eta/, 1987). G418- resistant Clones Were screened by Northern biot analyses of cytoplasmic RNA probed wlth a 32P-labelled antisense RNA probe speclfic for preproTRH mRNA sequences (Sevarlno e1 a/.. 1988). Clones were also screened for secretion of immunoreactwe TRH

peptldes contatnlng the amlno-termmal regoon of the TRH precursor (Wu et a/., 1987; see (ITRH) and lmmunoreactlve pYT22 (I~YT) as described below. The ipYT assay detects

Flg 1)

Other ProTRH-- Jackson and Relchlin (1974) wlth mcdficalions (Wu el a/., 1987). Radioimmunoassays

' - TRH was assayed as described by

usmg antlsera raised agalnst synthetic peptides were used to identify Other ~ ~ O T R H . derived products (Wu elal.. 1987; Wu and Jackson, 1988a) These pepttdes included a TRH progenitor sequence flanked by cysteine residues (pcclo), two peptides dertved from the amino-terminal flanking sequence 01 proTRH (pYE27 and pYT22). a connecting peptide (pEH24), and the carboxy-terminal 16 resdues of proTRH (pyE17) The posltlons of these sequences are Illustrated in Figure 1.

2.0 M acetlc acid1HCI. pH 2.0, contalning 0.1 pM phenylmethylsulfonyl fluorode, too V - Cells were grown to 30% wnfluence in 100 mm plates, scraped Into

nglml pepstatln. 1 pglml leupeptln, and 1.0 mM Na2EDTA. and processed as described (Sevarino et a/.. 1987). Secreted products Were collected for 24 hours and the medla

above were added before lyophhzation. Before assay of the nonTRH peptides, cell clarlfled by centrifugatlon. Protease inhibitors In the same concentrations as noted

triethylamine formatelacetonltrile solvent system (Vale 81 a/., 1983). extracts and media were purified on C18 Bondelute cartridges (Analytichem) us~ng a

n of Peo and Jackson, 1988a) on Sephadex G-50 and Sephadex G-75 superfine columns (1.6 x 85

licks . Gel filtration was carried out as previously described (wu

cm) equllibrated with 1M acetic acid, Columns were calibrated wlth the molecular wetght standards bovlne serum albumin (BSA). 68kD: carbontc anhydrase (CA), 2gkD: cytochrome c (CC). 12.4kD: and aprotinin (Apr), 6.5kD. Fractions for RIA contained 10 mg bovine Serum albumin (BSA) added before lyophll lzation.

- Transfected AtT20 cells were grown to confluence In 75 cm2 culture flasks. Growth media was removed and the cells were washed twlce with 10 ml Hank's balanced salt so lu tm buffered to pH 7.4 wlth 10 mM HEPES (HHBS). The cells were incubated for the indcated times with 5 ml of test medla. High potassium HHBS (Hlgh Kc) media contained 50 mM potassium wlth the sodium concentratlon reduced to mainfmn isotonicity. Cells exposed to high potassium, no Calcium HHBS were washed with HHBS without Calcdum for the second wash. Medlurn containing norepinephrine was treated wlth 0.1 mM ascorbate to prevent oxidatton. The protease inhibitors phenylmethylsulfonyl fluoride, aprotinin, pepstatin A and bestatln (Slgma) were added to medta samples in final concentrations of 0 001% prior to assay.

ton of - Purlflcation of the pepttde 4kDa ipYE27 was monitored by RIA. An lmmunoafflnity column prepared uslng an immunoglobulin fractton 01 anti-pYE27 was prepared as described prevlously (Wu el a/.. 1986). The column was primed with 0.5 g BSA and 0.lg Polypep (Slgma) before application of extract. The

phosphate pH 7.4, 150 mM NaCI) in a total volume of 350 ml and pumped through the lyophilized extract of transfected AtT2O cells was dissolved in PBS (20 mM sodium

wlumn at 35 mUhr. The column was washed wtth 100 ml of PBS and eluted wlth 2M acetlc acid at 25 mllhr. The column was then re-equlllbrated with PES and the immunoaffinity extraction was repeated. The eluants containing ipYE27 were pooled. concentrated, and fractionated by G-50 Sephadex chromatography. The peak fractions of

were adjusted to 0.2% trlfluoroacetlc acid (TFA) and were pumped directly onto a C18 ~mmunoreactwty, eluting at the same volume as 4kDa ipYE27 extracted from rat brain.

semi-preparatlve HPLC column (Vydac 30nm pore. 1 x 25 cm) at 3 mllmlnute. In all HPLC procedures solvent A was 0.1 56 aqueous TFA and solvent B 199% water. 80% acetonltrlle, 0.1 % TFA (vlv). Adsorbed material was eluted with an acetonitrile gradient as shown (Fig 7). This material was purifled further using biphenyl and C4 30nm pore analytlcal (Vydac. 0.42 x 25 cm) HPLC columns as descrlbed In the text. Purificatm was assessed by UV absorbance at 225 nm and by amino acid analysis. Slmilar procedures were used for the puriflcation of 3kDa ipYT22 and 6kDa ipYE17.

- The protein content of purified peptldes was determlned by ammo acid analysis (Spiess. 1986) or was estlmated by UV absorbance. Ten to 400 pmol 01 each peptide was subjected to automated Edman degradation In the presence of 1.5 mg

of May 8. 1987. Repetitive ylelds ranged from 85% - 95%. Polybrene using an Applied Bmystems Sequencer and the Programs NORMAL-1 and PRO-1

3T3 Media

Cells

w Media

cel ls

RIN SF Medla

Cells

At120 Medla

Cells

+

+

540.5

830.4

246.8

1635.4

+

+

+

+ 18.1

68 6

+ 195.4

- 1022.3

The preproTRH-derived peptldes were extracted from a representative clone of each cell type and the immunoreactwitles were assayed as described In the text.

only as positive. As explained in the text. the low levels of TRH tracer displacement Immunoreactivity that did not dllute in parallel to synthetic standards is expressed

seen for both Untransfected and transfected 3T3 and GH4 cells are due to an unknown cross-reacting materlal and these values are indicated as negative in the table. Untransfected cell contents and secreted material were negatlve for ipYT. lpCC and

extracted cellular protein. The media samples represent collections made over 24 iTRH except as described lor iTRH in 3T3 and GH4 cells. Values are expressed as pglmg

hours.

Table 2 PreproTRH-derived peptide content and release from transfected AtT20 cells

Cell exnact Basal release

eepticlaamol/moorotelnmolarratloomol/hr/maoroleinmolarratlo

pYE27 10.62 1 .oo 0.323 1.00

pYT22 7.32 0.69 0.135 0.46

pEH24 10.99 1.03 0.278 0.86

pYE17 9 16 0.86 0.280 0.87

TRH 21.71 2.04 0.357 1.10

Results were calculated using the respectlve synthetic peptlde standard CuNes. Three flasks were separately assayed: cdl proteinlflask was 9.65 mg f 0 25 (mean * S.E.M ). Molar r a t m are expressed retatlve to pYE27.

Table 3 Purification table for 4kDa ipYE27

1 lmmunoafffnlty (2 passes) 27.8 870

2 Sephadex G-50 26.5 17

3 C18 semipreparative HPLC 15.3 nd

4 Biphenyl HPLC 10 8 nd

5 C4 analyhcal HPLC 4.5 (1.15 nmoles) nd

nd - not determined

- Constructlon of the TRH cDNA expression vector pCYV-TRH-SV40 .

site. Plasmid pCMV-TRH-SV40 was constructed by inserting the preproTRH coding The plasmid pCMV IE . 0gl II contains the CMV IE promoter adjacent to a 0gl II cloning

sequence and the SV40 small t antigen 3' flanking sequences Into pCMV IE . B g l II as shown (see text). From 5 to 3'. the stippled area represents the CMV promoter, the solld black region represents the TRH cDNA sequence, and the striped area repesents the polyadenylatmn signals from tne SV40 small t antigen. Dashed lines indicate ligatlon sites.

Prepro-TRH Processing in Transfected Cell Lines 21535 v,CA CCA r pYT22 S S S P S

A B C D E F G H I .~ . 1

A . 6 kDa ipCCl0 c. 6 kD: ipYEt7

t -1750

D.

-7 2.5 kDa ipEH24

c 0. 3 kDa ipYT22

FIGURE 3- Northern blot analysis 01 preproTRH mRNA levels In transfected cells . Cytoplasmic mRNA was isolated, fractmated. and probed tor preproTRH-specific species as descrlbed in 'Experimental Procedures'. Messenger RNA from transfected cells was compared with untranstected cell mRNA and with mRNA from sources expressing the endogenous pteproTRH gene as follows: Lane A - Poly A-selected mRNA from rat hypothalamus; Lane B - Cytoplasmic mRNA from CA77 cells. a rat medullary thyroid carcinoma cell line that expresses preproTRH endogenously; Lane C . Untranstected 3T3 cells; Lane D - Transfected 3T3 clone; Lane E - Untranstected RIN 5F celis: Lane F - Transfected RIN 5F clone: Lane G - Untransfected AtT20 cells; Lane H . Transtected AtT20 clone; Lane I - Transfected AtT20 clone expressing a preproTRH mRNA 01 atypical size. Sizes of mRNA species (marked by arrows) were determined relative to ethidium-bromide stained RNA size markers (Bethesda Research Laboratories). Fwe pg of cytopiasmlc RNA were fractionated per lane, except for the poly A-selected material used as a marker whlch was not quantitated. Hybridlzatlons were performed wlth probes 01 stmilar specific actwity and autoradlographs were exposed tor 24 hours under ldenttcal conditions except for Lane I which was exposed for 2 hours.

Volume (ml) Volume (ml)

- Sephadex G-50 chromatography of transfected AtT20 preproTRH- derived peptides - Intracellular products from the transfected clone aKS6-19 Were extracted and fractionated as described under 'Experimental Methods'. Each frame dlsplays the fractions assayed for a dlfferent immunoreactivity. A. ipCC10; 6. ipYT22; C. ipYE17: D. ipEH24 . The elution positions of preproTRH-derived peptides extracted from rat brain are marked above each frame by soltd arrows (Wu 8 Jackson, 1988a). Striped arrows at the top of the figure indicate the positions of the molecular welght markers carbonic anhydrase (CA). cytochrome c (CC). aprotinin (Apr), and Synfhetlc pYT22. Vo indicates the void volume. $ S S $ S

VoCA CC Apr pYT22

C. 6kDa 1pYEt7

4 6 kDa ipCC10

E. 3kDa ipYT22 D* 7kDa ipYT22 C CII K . K'mcd' AstC11 l y M N E l y M N E

Stimulus FIGURE 6 - Regulated secretion of preproTRH peptides from transfected AtT20 cells - Basal and stimulated release of preproTRH-derived products was

was normaltzed to basal release. The key lo the measured immunoreactivities is measured over one hour from the clonal line aKS6-19. Each assayed immunoreactlvlfy

displayed in the box at the right. Ctl - contml; K+ - high (50 mM) potassium; K+ no Cat+ - high potassium. no calcium; Asc Ctl - control containing 10' M ascorbic acid NE - lo'' M ascorbate containing norepinephrine at the two doses indicated. Error bars represent the SEM, n - 3. Basal release data are given In Table 2.

Volume (ml)

m4- Sephadex G-50 chromatography of transfected RIN 5F preproTRH- derived peptides - A. 24-hour Secretion products from clone rKS6-24. fractionated and assayed for ipCC10. B. intracellular products extracted from clone rKS6-9. fractlonated

fractionated and assayed for ipYE17. D. 24-hour secretion products from clone rKS6-9. and assayed for ipYT22. C. Intracellular products extracted from clone rKS6-9.

fractionated and assayed for ipYT22. The elution volumes of preproTRH-derived peptides isolated from rat hypothalamus are denoted by solrd arrows above each chromatogram

01 the molecular welght markers carbonic anhydrase (CA). cytochrome c (CC). aprotlnin (Wu and Jackson, 1988a). Striped arrows at the top of the flgure indicate the positions

(Apr). and synthetic pYT22. V, lnd~cates the vold volume.

C.

Volume (ml)

0.

Fnnion

D.

Fnnion Fnnion

FlGURE- Purification of 4kDa lpVE27 - A. G-50 chromatography on a 1.6 x 95 cm column with 1 M acetic acid as the eluant. The void and salt volumes WBI0 50 and 200 ml. The elution positton of ipYE27 isolated from rat hypothalamus is denoted by the

described. 6. C18 semipreparative HPLC on a 1 x 25 cm CIS. 5 pm particle. 30 nm pore arrow. Fractions from volumes 100 to 150 ml were pooled and purified further as

Vydac column. Solvent A - 0.1% aqueous tritluomacetlc acid (TFA). solvent B - 0.1% TFA. 19.9% water, 80% acetonitrile. The gradient profile is shown; flow - 3 mllmin. The rectangle above the chromatogram denotes the fractions pooled and purified further. C.

column. Solvents used were as in 8. The gradient profile is shown; flow - 1 mllmin. The Biphenyl analytical HPLC on a 0.42 x 25 cm (C6H5)2. 5 pm particle. 30 nm pore Vydac

rectangle above chromatogram indicates the fractions pooled for further purification. D. C4 analytical HPLC on a 0.42 x 25 cm C4. 5 pm particle. 30 nm pore Vydac column. Solvents used were as in B. The gradient profile is shown: flow - 1 mvmin.