Isolation and Characterization of a Complementary DNA (Galanin) Clone from Estrogen-induced Pituitary Tumor Messenger RNA*

The administration of high levels of estrogen is a well established method for producing prolactin-se-creting pituitary tumors in rodents but the mechanism of tumor induction is not clear. In this paper we de-scribe a cDNA clone (pEIC) which has been isolated from an estrogen-induced pituitary tumor cDNA library. The mRNA transcript corresponding to the pEIC clone is 0.9 kilobase in length and is not detectable in normal pituitaries but is expressed as early as 3 h after estrogen stimulation. Nucleotide sequence analysis of two 700-base pair recombinant clones shows that they encode a 124-amino acid protein which is 70% identical to the porcine galanin precursor. The sequence of 29 amino acid residues coded for by the pEIC cDNA clone is 88% identical with porcine galanin with only three amino acid substitutions near the C terminus. This extensive homology suggests that the pEIC cDNA clone codes for rat galanin or a protein belonging to the galanin gene family. These results provide the first evidence of a physio- logical regulator (estrogen) of the expression of the galanin gene. They also imply that galanin is secreted by prolactin-secreting tumors. Because intracerebroventricular injection of galanin can stimulate prolactin secretion and galanin inhibits hypothalamic dopamine release, it is conceivable that galanin may play a role in the induction of prolactin-secreting tumors.


Recipient of a Manitoba Health Research Council fellowship and
to whom reprint requests should be addressed.
The abbreviations used are: PRL, prolactin; E2, estrogen. (Holmgren et al., 1982;Melmed et al., 1986) that are secreted during pregnancy. At present there is no way to identify those PRL-secreting tumors that will grow rapidly under the influence of E2 from those that will not, nor is it clear whether the effects of E2 on pituitary growth are direct or indirect. Evidence supporting the view that the proliferative effect of E2 is mediated by estromedins has been suggested by Sirbasku (1978) and Biro (1986). In order to explore the mechanism by which estrogen induces pituitary tumors we have examined the cDNA clones that are induced by estrogen in Fisher rats, a rat strain in which estrogen leads to the rapid development of pituitary tumor in a high proportion of the animals. In this paper we report the isolation and characterization of a cDNA clone that is induced by estrogens in normal pituitary glands and which is expressed at a high level in pituitary tumors induced by estrogens.

EXPERIMENTAL PROCEDURES
Materials-Radioisotopes were purchased from Du Pont-New Eng- Induction of Pituitary Tumors-Female Fisher 344 rats (200-g weight) were obtained from Charles River (Quebec, Canada). Pituitary tumors were induced in rats using the silastic tubes containing the synthetic estrogen preparation, diethylstilbestrol, as reported previously (Vrontakis et al., 1987). For studies on the induction of an estrogen-inducible mRNA (pEIC) female rata were given 10 pg of 17B-estradiol subcutaneously.
Poly (A) mRNA Isolation-Pituitaries were removed from Fisher 344 rats and frozen in liquid nitrogen. Total RNA was isolated by the guanidine thiocyanate/cesium chloride method (Chirgwin et al., 1979). Poly(A)+ mRNA was isolated by two cycles of oligo(dT)cellulose chromatography (Aviv and Leder, 1972).
Construction of cDNA Library-A cDNA library was prepared from estrogen-induced pituitary tumor mRNA. The library was constructed in phage Xgt 10, as described by Huynh et al. (1985) except that the second strand was made according to the protocol of Okayama and Berg (1982). DNA was packaged using an in vitro packaging kit from Amersham Corp. according to the supplier's instructions. Recombinant phage were plated on high frequency lysogen strain E. coli MA150 (Young and Davis, 1983). Phage plates were blotted onto nitrocellulose by the method of Benton and Davis (1977).
All recombinant DNA research was prepared under the current Medical Research Council (Canada) guidelines for biohazardous materials.
Screening of Pituitary cDNA Library-To screen initially for cDNAs which were more strongly expressed in the pituitaries of E2treated rats, an enriched cDNA probe from the E2-induced pituitary tumor mRNA was obtained by subtractive hybridization according to Davis et al. (1984). Plaques which hybridized strongly to this probe were picked and rescreened at a lower density. Differential screening of the library was followed with a [32P]cDNA probe made from control pituitary mRNA (-1 and E2-induced pituitary tumor mRNA (+), RNA Blot Analysis-Total RNA samples extracted for Northern blot analysis (Alwine et al., 1977) from pituitaries of control or estrogen-treated rats were sized-fractionated on 1.3% agarose, 2.2 M formaldehyde gel (Lehrach et al., 1977) and blotted onto nitrocellulose 16755 Cloning of Rat Galanin from Pituitary Tumors paper (Thomas, 1980). Hybridizations were performed at 42 'C in 50% formamide using lo6 dpm/ml of 32P-labeled pEIC cDNA insert nick-translated to a specific activity of 3-5 X 10' dpmlpg. Final washing conditions were 0.2 X SSC, 0.1% sodium dodecyl sulfate at 65 'C for 30 min. DNA Sequence Deterrninatwn-The pEIC clone was digested by restriction enzymes using the three-core buffer as described by Maniatis et ol. (1982). Restriction enzyme fragments were sized either on agarose or polyacrylamide gels using HaeIII-cut $X174 fragments as markers. Appropriate restriction fragments were subcloned with the m13 vectors mp18 or mp19 (Vieira and Messing, 1982). Recombinant phage DNA was propagated in E. coli J M 101. DNA sequencing was by the dideoxy chain termination method (Sanger et al., 1977) using 36S-labeled dATP and the gradient gel modification (Biggin et al., 1983). The gels were fixed in 10% acetic acid and dried before exposure to Kodak XAR film.

Isolation of the pEIC Clone"
cDNA library was constructed using mRNA from estrogen-induced pituitary tumors. The library was constructed in the phage vector X g t l O as described under "Experimental Procedures." The library was initially screened with an enriched cDNA probe also made from estrogen-induced pituitary tumor mRNA as described under "Experimental Procedures." Sixty clones that strongly hybridized to this probe were selected for further screening by differential hybridization with ["PIcDNA probes made from control pituitary mRNA and estrogen-induced pituitary tumor mRNA. The differential screening resulted in the isolation of three estrogen-induced and two estrogensuppressed clones. From the three estrogen-induced clones, the pEIC was not detectable a t all with the cDNA probe made from control pituitary mRNA. When the pEIC clone was used to rescreen a further 2200 recombinant plaques of the X g t l O library 11 more clones were isolated suggesting that the pEIC clone represents approximately 0.5% of the estrogen-induced pituitary tumor mRNA.
Expression of the pEIC in the Rat Pituitary- Fig. 1 shows autoradiograms of a Northern blot of RNA from normal female pituitaries and estrogen-induced pituitary tumors. Only the estrogen-induced pituitary tumor RNA hybridizes to the pEIC insert indicating that the RNA is not detectable in normal pituitaries. The transcript of the pEIC cDNA clone has a size of approximately 900 base pairs. As an internal control for the quality of the blotted RNA, hybridization of the same blot with the rat PRL cDNA probe indicates hybridization to RNA in both experimental conditions. Time course of the induction of pEIC RNA by estrogen shown in Fig. 2 indicates that the transcript is detectable as early as 3 h after a single injection of the 178-estradiol with the peak around 9 h and after 24 h has declined. The induction of the pEIC transcript is dose-dependent (data not shown).
DNA Sequence Analysis-Restriction enzyme analysis and sequencing strategy for the pEIC cDNA clones are shown in Fig. 3. Two clones, pEICA and pEICB, were analyzed to obtain the nucleotide sequence. The nucleotide sequence and the deduced amino acid sequence of pEIC clones are given in Fig.  4. The pEIC clone encodes a 124-residue protein which has a molecular weight of 13,313. The 3"untranslated region is 184 base pairs in length and contains an AATAAA polyadenylation signal (Proudfoot and Brownlee, 1976) 12 nucleotides upstream from the poly(A) tail. Comparison of the pEIC cDNA clone sequence with the sequences in Genbank revealed that 70% of this sequence at the amino acid level is identical to the porcine galanin precursor (Rokaeus and Brownstein, 1986). Fig. 5 shows a comparison of the porcine amino acid galanin sequence with the predicted amino acid sequence of the pEIC cDNA clone. The 29-amino acid single porcine galanin sequence is 88% identical with the pEIC sequence 20 pg of total RNA from normal pituitaries ( l a n e 1 ) and estrogen-induced pituitary tumors ( l a n e 2) were electrophoresed in a 1% formaldehyde agarose gel, blotted onto nitrocellulose paper, and hybridized in 50% formamide at 42 "C to the 32P-labeled pEIC insert (left panel) and then the same blot to the 32P-labeled rat PRL cDNA (right panel). Only the estrogen-induced pituitary tumor RNA hybridizes to the pEIC insert indicating that the RNA is not detectable in normal pituitaries. As an internal control for the quality of the blotted RNA, hybridization of the same blot with a rat PRL cDNA probe indicated hybridization to RNA in both experimental conditions. Northern blot analysis of 20 pg of total RNA from rat pituitaries 3 h ( l a n e 2), 9 h ( l a n e 3 ) . 24 h ( l o n e 4 ) , and 48 h ( l a n e 5 ) after a single injection of 10 pg of 178-estradiol in ovariectomized Fisher 344 rats. The hybridization conditions with 32P-labeled pEIC insert were the same as in Fig. 1. with a serine instead of histidine a t residue 55, histidine instead of tyrosine a t residue 58, and threonine instead of alanine a t residue 61. The pEIC sequence has an extra valine a t residue 87, so the total amino acid sequence consists of 124 residues instead of 123 as is the case for the porcine galanin  underlined. The closed box corresponds to the galanin sequence.

0 30 Galanin A L S A T L G L G S P V K E K R pEIC T L S A T L G L G Y P Z K E K R
precursor. The 5'-and 3"untranslated regions of the pEIC clone differ substantially from the 5'-and 3"noncoding regions of the porcine galanin precursor.

DISCUSSION
We have isolated and sequenced a cDNA clone from estrogen-induced pituitary tumor mRNA. The sequence of the pEIC clone indicates that 70% of the amino acid sequence is identical to porcine galanin precursor, with all but 3 of 29 amino acid residues (88%) corresponding to the porcine galanin sequence, indicating that the pEIC cDNA clone is either the rat equivalent of galanin or less probably a sequence belonging to the galanin gene family.
The mRNA coding for the pEIC cDNA clone is approximately 900 base pairs, the same size that has been reported for galanin (Rokaeus and Brownstein, 1986). The transcript of the pEIC cDNA clone is not detectable at all in normal pituitaries and is strongly induced by estrogen in the pituitary. The induction of the transcript occurs as early as 3 h after estrogen administration and is a primary event since administration of cycloheximide 10 min before the estrogen administration does not inhibit the induction of the transcript.* After 7 weeks of treatment with estrogen the pEIC cDNA clone transcript is quite abundant in the pituitary representing 0.5% of the pituitary mRNA.
Galanin was first identified as a gastrointestinal neuropeptide (Tatemoto et al., 1983), and subsequent studies demonstrated that it is localized to nerves of the myenteric and submucous plexus (Melander et al., 1985). Galanin-like immunoreactivity has also been found in the central nervous system (Ching et al., 1985) with the highest concentration in the median eminence and hypothalamus. The role of galanin in the central nervous system is unknown. Its presence in the median eminence suggests that it may play a role in the regulation of anterior pituitary function. It has been reported that intracerebroventricular injection of galanin into the rat brain stimulates PRL and growth hormone release (Ottlecz et al., 1986;Koshiyama et al., 1987) possibly by stimulating vasoactive intestinal polypeptide secretion. Nordstrom et al. (1987) reported that galanin has an inhibitory effect on dopamine release from the rat median eminence, a finding which would support a role for galanin in the regulation of PRL secretion as dopamine itself is a potent inhibitor of PRL secretion.
Since administration of estrogen in rodents induces prolactin-secreting pituitary tumors and as we have shown in this paper estrogen administration induces a high level of expression of galanin mRNA in the pituitary, it is conceivable that estrogen regulates PRL secretion and or pituitary tumor formation through a step that involves galanin.
Further studies are in progress to elucidate the exact mechanism by which estrogen regulates galanin and PRL synthesis and the role of galanin for the formation of PRL-secreting pituitary tumor by estrogen.