The Salk Ihstitute for Biological Studies, La Jolla, CA 92037; and §Howard Hughes ..... Casanova, J., Copp, R. P., Janocko, L. & Samuels, H. H. (1985)J. Biol.
Proc. Natl. Acad. Sci. USA Vol. 85, pp. 4755-4759, July 1988 Developmental Biology
Identification of rat growth hormone genomic sequences targeting pituitary expression in transgenic mice SERGIO A. LIRA*t, E. BRYAN CRENSHAW III*, CHRISTOPHER K. GLASS*, LARRY W. SWANSONO§, AND MICHAEL G. ROSENFELD*§
*Eukaryotic Regulatory Biology Program, School of Medicine, M-013, University of California, San Diego, La Jolla, CA 92093; The Salk Ihstitute for Biological Studies, La Jolla, CA 92037; and §Howard Hughes Medical Institute, La Jolla, CA 92037
*Neural Systems Laboratory,
Communicated by Daniel Steinberg, March 7, 1988
ABSTRACT Cdnstructs containing different segments of the 5' flanking region of the rat growth hormone gene fused to the human growth hormone coding sequences were introduced into fertilized mouse oocytes. As few as 181 base pairs of the rat growth hormone promoter targeted reporter gene expression to the pituitary gland of the resulting transgenic mice. A construct containing only 45 base pairs of the promoter failed to target expression of the reporter to the pituitary, indicating that the pituitary expression is directed by information contained in the segment spanning positions -181 to -45. In the pituitary, immunohistochemistry showed transgene expression mainly in the growth hormone-producing cells (somatotrophs), in a subset of cells producing thyrotropin, and occasionally in prolactin-producing cells. These data establish that cis-active elements contained within the first 180 base pairs of the promoter are sufficient for transcriptional activation of the growth hormone gene in somatotrophs and suggest a functional relationship among growth hormone, prolactin, and thyrotropin gene activation.
tary cells as well. These authors suggested that a repressor element located between bp - 236 and - 550 of the rGH gene might restrict its expression to the pituitary. To analyze the relative contribution of these putative regulatory elements to the expression of GH in vivo, we used transgenic mice, which provide the potential for screening the expression of specific fusion genes in the whole organism. Here we report on the expression of the human GH (hGH) gene directed by different segments of the rGH 5' flanking region in transgenic mice. The hGH gene has been used as a reporter gene in several transgenic experiments (14), and its expression can be determined by sensitive and specific immunohistochemical and radioimmunological methods.
MATERIAL AND METHODS Constructions. Segments ofthe rGH 5' flanking region were fused to the structural part of the hGH gene (Fig. 1). The BamHI restriction site at bp + 2 of the hGH gene was con-
The anterior pituitary in many vertebrate species develops from a common primordium and contains five major endocrine cell types. These cells are phenotypically defined (somatotrophs, lactotrophs, corticotrophs, thyrotrophs, and gonadotrophs) by the hormones they produce: somatotropin (growth hormone, GH); prolactin; corticotropin; thyrotropin (thyroid-stimulating hormone, TSH); and the gonadotropins, lutropin and follitropin (1, 2). The anterior pituitary, therefore, provides a model to study the molecular mechanisms whereby phenotypically distinct cell types arise in an organ, We have analyzed the determinants of GH gene expression to address this developmental code. In rodents, there is a single copy of the GH gene, which is expressed exclusively in the pituitary (3, 4). As in dther eukaryotic genes (5), the molecular dissection of regulatory elements within this gene has revealed the existence of cisactive sequences involved in its tissue-specific expression. We (6) and others (7-9) have suggested that cell-specific as well as hormonal regulatory elements of the rat GH (rGH) gene are found in the 5' flanking region. Approximately 200 base pairs (bp) of information adjacent to the rGH transcription initiation site (bp +1) allows the expression of the reporter component of fusion genes transfected into pituitary-derived cell lines. This segment contains an enhancer element, located between bp - 47 and - 235, that directs the expression of reporter genes in pituitary cell lines but not in five non-pituitary lines (6). DNase I protection ("footprint") analysis of this enhancer revealed two tissue-specific elements that appear to bind a positive transcription factor (1012). Larsen et al. (13), however, reported that the rGH promoter directs expression of reporter genes in non-pitui-
verted to a Sal I site and fused in the first three constructions to a Xho I site present at bp + 8 of the rGH gene. Restriction sites located at bp - 181 (Sau3A) and - 311 (Kpn I) of the rGH gene were converted to BamHI sites, and the segment Was inserted into the pUC18 or pBR322 vector containing the hGH gene. The 1.1-GH plasmid was constructed by inserting the hGH gene into a pUC18 vector containing 2.2 kbp of the 5' flanking region of the rGH gene. The fourth construction, plasmid 47-GH, was made by annealing two oligonucleotides containing the first 45 bases of the rGH promoter; this fragment contained a HindIII site at its 5' end and a Sal I site at its 3' end and was cloned in a vector containing the hGH gene. The hGH gene used here extends from bp + 2 to an EcoRI site (bp + 2153) present in all constructions at the 3' end. This site is located 0.5 kbp 3' of the hGH poly(A) site. Gene Transfer into Embryos. Fragments for injection were excised by digestion of the plasmids with BamHI/EcoRI (320-GH and 180-GH) and HindIII/EcoRI (47-GH). The 1.7-GH fragment was excised by EcoRI digestion, using the endogenous EcoRI site present at kbp - 1.7. The resulting linearized fragments (devoid of vector sequences) were purified by agarose gel electrophoresis followed by electroelution and were dialyzed against injection buffer (5 mM Tris'HCI, pH 7.4/5 mM NaCl/0.1 mM EDTA) on dialysis filters (Millipore). Each fragment (1-5 ng/ml) was injected into the male pronuclei of fertilized eggs, (C57BL/6J x DBA/2J)F1, and the eggs were transferred to foster mothers by standard procedures (15). DNA Analysis. Transgenic mice were identified by dot blot analysis of DNA extracted from a segment of the tail (16). For the genomic DNA analysis, tail DNA was digested with Sst I and electrophoresed in agarose gel, transferred to nitrocel-
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Abbreviations: GH, growth hormone; rGH, rat GH; hGH, human GH; TSH, thyroid-stimulating hormone. tTo whom reprint requests should be addressed.
4755
4756
Proc. Natl. Acad. Sci. USA 85
Developmental Biology: Lira et al. SERIES
hGH
rGH
1 -1500 Eco RI
-500 Pvu 11
(Xhol/Sall)
Il
Pvu 11
Eco RI
-
AUG
Sst
UAG
Kpn
I__ Sau 3A
1.7-GH 1-7GH
Eco RI
320-GH Eco RI
180-GH Hind III
Eco RI
(1988)
FIG. 1. Constructs injected into fertilized oocytes. Fusion genes containing 1.7 kbp, 311 bp, 181 bp, and 45 bp of the rGH 5' flanking region fused to the structural part of the hGH gene are called 1.7-GH, 320-GH, 180-GH, iand 47-GH. Boxes represent hGH exons; hatched regions are untranslated sequences upstream of the AUG initiation codon and dowpstream of the UAG termination codoh. The BamHI site at bp +2 in the hGH gene was transformed to a Sal I site and fused in the 1.7-GH, 320-GH, and 180-GH constructs to the Xho I site at bp + 8 of the rGH gene. The native rGH Kpn I and Sau3A sites were converted to BamHI sites during the cloning procedures. The *HindIII site was created upstream of a 53-bp (bp -45 to + 8 of the rGH gene) oligonucleotide, and a Sal I site was created 3' to it. mouse
47-GH ' 'A H H ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
lulose filter, washed in prehybridization buffer, and hybridized with a 1050-bp nick-translated Pvu II-Pvu II probe from the hGH gene (Fig. 1). Filters were washed with 0.3 M NaCl/0.03 M sodium citrate, pH 7/0.1% NaDodSO4 at room temperature for 5 min and then with 0.03 M NaCl/0.003 M sodium citrate, pH 7/0.1% NaDodSO4 at 650C for 1 hr. RNA Analysis. Animals from at least two pedigrees in the 1.7-GH, 320-GH, and 180-GH series were killed by cervical dislocation, and their tissues were rapidly dissected and frozen ini liquid nitrogen. After homogenization with a Polytron (Bfinkman)j total nucleic acids from brain, lung, kidney, intestine, liver, spleen, skeletal muscle, and gonads were isolated by the method of Shields and Blobel (17). Total RNA was separated from DNA by precipitation with LiCl. In brief, total nucleic acids were incubated with 2.5 M LiCI/20 mM sodium acetate, pH 5.0, for at least 3 hr on ice. Samples were spun ih a microcentrifuge at 40C for 10 min, and the precipitates were resuspended in water. Aliquots (15-20 ,g) were denatured and electrophoresed in 1% agarose gels containing 2.2 M formaldehyde. RNA was transferred to nitrocellulose, washed in prehybridization buffer, and hybridized as described above. Total nucleic acids (2-6 pug) extracted from the pituitary glands of transgenic and control animals served as positive controls. Immunohistochemical Analysis. Mice were anesthetized with chloral hydrate and perfused through the heart with paraformaldehyde/glutaraldehyde solutions (18). The pituitary glands were removed and postfixed in 4% (wt/vol) paraformaldehyde/0.1% glutaraldehyde in 0.2 M sodium borate buffer (pH 9.5) with 10% (wt/vol) sucrose overnight at 40C. Free-floating sections (20 ,um thick) were processed for indirect immunofluorescence (18). Stained sections were mounted on glass slides, coverslipped with 50%o (vol/vol) glycerol in 0.4 M potassium bicarbonate buffer (pH 8.6), and viewed with a Leitz Dialux 20 fluorescence microscope. Antiserum against hGH was generated in guinea pig (Arnel Products, New York) and used at a 1:20,000 dilution. Immunostaining in transgenic pituitaries was completely blocked by preabsorption of the antiserum with hGH. To detect endogenous GH in mouse somatottophs, we used a monkey antimouse GH antibody (provided by W. P. Vanderlaan, Whittier Institute, La Jolla, CA) at a 1:25,000 dilution. Antiserum to human TSH was generated in rabbit (Bioproducts, Westbury, NY) and used at a 1:5000 dilution. This antiserum does not crossreact with hGH even when used at a 1:500 dilution. Rabbit anti-rat lutropin antiserum (A. F. Parlow, National Institute of Arthritis, Diabetes, and Digestive and Kidney Diseases, Bethesda, MD) and anti-human corticotropin antiserum (Wylie Vale, Salk Institute) were used at 1:10,000 dilution. Rabbit anti-rat prolactin antiserum (A. F. Parlow) was used at a 1:20,000 dilution. Secondary antisera raised in goats (and labeled with fluorescein or rhodamine) against
guinea pig, human, or monkey IgG (Cappel, Malvern, PA) or against rabbit IgG (Tago, Burlingame, CA) were used at a dilution of 1:100 (anti-human IgG) or 1:200. Hormone Analysis. Hormone determinations were performed with an immunoradiometric assay specific for hGH (Hybritech, La Jolla, CA). Three to eleven mice of each series were rapidly removed from their cages and decapitated. Blood was collected from the trunk, allowed to clot, and spun in a microcentrifuge for 15 min. Supernatants were collected and stored at - 200C until the time of assay. Pituitaries were rapidly dissected, immersed in 100 mM potassium phosphate buffer (pH 7.4), and dispersed with a 25-gauge syringe. Extracts were frozen and thawed three times and spun in a microcentrifuge for 10 min at 40C. Supernatants were stored at - 700C. Protein concentrations in the extracts were determined by the method of Peterson (19).
RESULTS Generation of Transgenic Mice. Fusion genes containing different segments of the 5' flanking region of the rGH gene linked to the structural part of the hGH gene (Fig. 1) were microinjected into mouse eggs. The injected DNA integrated into the genome of 10-30% of the animals generated from the injected eggs. We generated 13 positive animals in the 1.7-GH series, 6 in the 320-GH series, 9 in the 180-GH series, and 8 in the 47-GH series (Fig. 2A). Most of the positive animals in the four series had more than one copy of the transgene integrated into their genomes. The copies tended to arrange in tandem, in a head-to-tail orientation. This arrangement allowed the detection of fragments with the approximate size of the transgene when the genomic DNA was digested with Sst I, which cleaves at a single restriction site that is present in all the constructions. Positive founder mice were mated with nontransgenic mice to generate transgenic lines (or pedigrees). At least three pedigrees were established in each series, and their positive offspring were used in the studies described here. RNA Analysis. To determine the site of transgene expression, we first screened for hGH mRNA in several tissues from transgenic mice (Fig. 2B). We were not able to detect ectopic expression of hGH mRNA in nonpituitary tissues except ih one pedigree (320-GH/27), where hGH message was detected in liver, intestine, and kidney (data not shown). This pattern of expression was heritable and probably reflects an influence of the integration site (14). These mice did not display signs of gigantism up to 140 days of age, and they shared two phenotypic characteristics of other transgenic animals with unregulated production of GH: small pituitaries and infertility in females. Immunohistochemical Analysis. The expression of the hGH gene product in the pituitary was examined immunohisto-
Developmental Biology: Lira et al.
Proc. Natl. Acad. Sci. USA 85 (1988)
A 1 7 GH
320 GH
0
180 GH
- 2.4
47 GH
*-wS@SS2
-2
_
*...
ft"@
2
_..
*-22
B 1 234 5 6 7
1 234 5678
-28 S - 28 S - 18 S
, - 1 Kb
- 18S
I
- 1 Kb
FIG. 2. (A) Representative analysis of genomic DNA from animals in the four different series. In all cases the DNA (10 Jtg) was digested with Sst I, which cleaves at a single restriction site present in all constructions. A Pvu II-Pvu II fragment (1050 bp) from the hGH gene was radiolabeled and used as hybridization probe. The major hybridizing bands (3.9, 2.4, 2.3, and 2.2 kbp) correspond to the approximate sizes of the fragments injected. Pedigree numbers are above the lanes. (B) Representative analysis of size-fractionated total RNA (15-20 ,ug unless otherwise indicated) extracted from the tissues of transgenic animals (15). (Left) Tissues derived from a mouse of the 180-GH/1 pedigree. Lanes: 1, brain; 2, kidney; 3, intestine; 4, skeletal muscle; 5, spleen; 6, lung; 7, pituitary (4 ,ug). (Right) Tissues from an animal in the 1.7-GH/10 pedigree. Lanes: 1, brain; 2, kidney; 3, spleen; 4, testes; 5, intestine; 6, liver; 7, pituitary (2 ,ug); 8, lung. The hybridization probe was identical to that described for A; this probe does not distinguish between mouse and human GH mRNA at the stringency conditions used. The size of the RNA recognized by the probe [1 kilobase (kb)] was estimated by comparison with 28S and 18S rRNA.
chemically with a hGH-specific antibody that does not crossreact with any antigen in the mouse pituitary (Fig. 3). The assignment of cell type was accomplished by double labeling with antibodies against the major pituitary horWe analyzed 4 animals from 2 pedigrees in the 47-GH series, 11 animals from 3 pedigrees of the 180-GH series, 11 animals from 4 pedigrees of the 320-GH series, and 17 animals from 5 pedigrees of the 1.7-GH series. Immunostaining for hGH was restricted to the anterior lobe of the pituitary and was observed in all pedigrees of the 1.7-GH, 320-GH, and 180-GH series except one (320-GH/81). hGH immunostaining was not detected in the pituitaries of the 47-GH mice, control mice, or mice bearing a fusion of the herpes simplex mones.
4757
virus thymidine kinase promoter with the hGH gene (data not shown). The transgenes were expressed in somatotrophs in all 11 positive pedigrees (Fig. 3). Although the expression levels varied between pedigrees of a given series, it did not appear to vary between animals of the same pedigree. No expression of the transgenes was detected in cells producing lutropin or corticotropin in any series, but, unexpectedly, all three transgenes were expressed in a subset of thyrotrophs (Fig. 4). Expression of hGH in lactotrophs was not observed in 9 pedigrees; however, mice from 2 pedigrees ofthe 1.7-GH series and from 1 pedigree of the 320-GH series also expressed the transgene in a subset of lactotrophs. Hormone Analysis. To establish further that authentic hGH is produced in the transgenic mice, we analyzed hGH serum levels with a specific radioimmunoassay. Immunoreactive hGH was present in the serum of transgenic mice from all three expressing series (1.7-GH, 320-GH, and 180-GH) but was not detected in mice from the 47-GH group, in control mice, or in transgenic mice containing a metallothionein-rGH fusion gene that express rGH at levels 410-fold higher than the endogenous mouse GH. The levels of hGH varied between the series and integration events, but mice derived from the 1.7-GH and 320-GH series tended to have higher serum levels than those in the 180-GH series. Three pedigrees were analyzed in the 180-GH series; their serum hGH levels ranged from 0 to 2 ng/ml. In the four lines of the 320-GH series that we analyzed, the hGH levels ranged from 0.7 to 16 ng/ml. In the 1.7-GH series, three pedigrees were analyzed, and the levels ranged from 0.4 to 48 ng/ml. In parallel, we measured hGH in the pituitaries of transgenic animals by radioimmunoassay. hGH was detected in the pituitaries of four of the 180-GH pedigrees analyzed, at 1:1000 dilutions of the pituitary extracts. No hGH was detected in 1:10 dilutions of pituitary extracts from 47-GH (three pedigrees) or control mice (Fig. 5).
DISCUSSION We have shown that marker genes can be targeted to the pituitary of transgenic mice by sequences contained in the 5' flanking region of the rGH gene, indicating the existence of a common set of signals in the mouse and rat genes. Previous attempts to target the expression of marker genes by using flanking regions of the rGH gene or the hGH gene were unsuccessful (27), perhaps due to the presence of vector sequences in the fragments injected, which can reduce the expression of transgenes (14). Because tissue-specific expression can be achieved with as few as 180 bp of rGH-gene 5' flanking information, the data suggest that the cis-active elements that are responsible for transcriptional activation in cell culture experiments (6) are actually critical for developmental activation of the GH gene. The lack of expression of the 47-GH construction rules out the possibility that signals contained in the coding region of the hGH gene or 3' to it could be responsible for the expression of the transgenes in the pituitary. The data therefore suggest that the genetic information necessary for the pituitary expression of the transgenes is located between bp - 181 and - 45 of the rGH gene. This 136-bp segment in the rat (12) and human (20) genes includes two cis-active sequences that bind trans-acting factors isolated from pituitary cell lines expressing GH. A thyroid hormone receptor binding site has also been identified in this region in the rat gene (21). Our results indicate that not all sequences of the enhancer element described previously (6) are necessary for the appropriate expression of the gene in vivo. Additional components of the enhancer, such as the segment between bp - 180 and - 235, for example, can be removed without affecting the tissue-specific expression. However, we cannot rule out the possibility that this or other segments contribute to the
4758
Proc. Natl. Acad Sci. USA 85
Developmental Biology: Lira et al. human GH
(1988)
mouse GH
CON.
180GH
3 20GH FIG. 3. Colocalization of huand mouse GH in individual cells of the anterior lobe of the pituitary from mice in the 180-GH, 320-GH, and 1.7-GH series. Secondary antibodies were labeled with either fluorescein (green) or rhodamine (red). Specificity of the hGH antiserum is demonstrated by the absence of labeling in a section from a control (CON.) mouse. Sections were from mice in the following pedigrees: 180GH/1, 320-GH/27, and 1.7-GH/2. (x240.) man
1.7GH
overall expression of the rGH gene in vivo. Finally, our studies fail to support the postulated existence of a repressor, located between - 260 and -550, that would restrict the expression of this gene to the pituitary gland (13). Although the results of the in vitro analyses indicate that the 180-bp region is sufficient for tissue-specific expression, the more rigorous in vivo approach indicates that flanking sequences of 180 bp or even 1.7 kbp do not confer strict cell-type specificity within the pituitary. Although the transgenes are expressed predominantly in somatotrophs, expression is also consistently found in a subset of thyrotrophs and with poor penetrance in lactotrophs. The ectopic expression could be the result of differences between the 5' flanking regions of the mouse and rat genes, which are not known at the present; the result of a combination of signals contained in the rat and human GH genes; or the result of a different physical assembly of the transgene versus the endogenous gene. Alternatively, this unexpected pattern of expression could reflect a developmental relationship between these pituitary cell types. Some experimental observations are consistent with this latter possibility. First, coexpression of GH, TSH, and prolactin has been reported in certain pituitary tumors (22). Second, coexpression of GH and prolactin has been observed in pituitary-derived cell lines (23), in primary pituitary cultures (24), and more recently in rat pituitaries in situ (25). Third, the Snell dwarf mutation in mice results in loss of somatotrophs, lactotrophs, and thyrotrophs (26).
Finally, competition analysis of DNase I footprints and in vitro transcriptional studies indicate that regulatory elements in the rGH, prolactin, and jB-TSH genes bind a common trans-acting factor required for activation of gene transcription (12). The existence of a common transcriptional mechanism among these three genes could, therefore, account for the expression of the GH transgenes in thyrotrophs and lactotrophs. This hypothesis implies that the largely cellspecific expression of the rGH gene in the normal pituitary involves positive (activating) as well as negative (or restrictive) mechanisms. Our study has shown that a 180-bp sequence in the 5' flanking region of the rGH gene contains sufficient information for correct tissue and cell type expression, but that a more complex code must operate to restrict its expression to somatotrophs. Further investigation of a common developmental and functional linkage of GH, prolactin, and TSH genes, as well as of the nature of restriction within the normal pituitary, will require the biochemical characterization of their trans-acting factors and additional analysis of cis-active regions in transgenic mice. We thank Kristin Kalla and Jodi Harrold for their excellent technical work. We thank Stacey Dillon, Mark Stamnes, Sally Durgerian, and Donna Simmons for their contributions to this project, and we thank Marian Waterman for the drawings. S.A.L. is on leave of absence from Universidade Federal de Pernambuco,
Developmental Biology: Lira et al.
Proc. Natl. Acad. Sci. USA 85 (1988)
human GH
4759
TSH
1.7 GH
3 2 0GH FIG. 4. Immunohistochemical colocalization of hGH and TSH in some individual cells of the antenor pituitary of transgenic mice bearing different lengths of the rGH promoter fused to the structural part of the hGH gene. Secondary antibody to hGH was labeled with rhodamine (red) and that against TSH with fluorescein (green). Sections were derived from animals in the 1.7-GH/10, 320-GH/27, and 180-GH/39 pedi-
1 80-
GH
grees. ( x 240.)
40 r~~
~
~
30 C
a) 0
a. 0
20
I c
0) c
,v
in.
o 1 L-
11
22
180 GH
43
12
18
47 GH
29
Controls
C
FIG. 5. Immunoreactive hGH in extracts from pituitaries of transgenic mice in the 180-GH and 47-GH series and of control mice. Bars represent individual animals, and the numbers indicate the pedigrees. The absolute amounts of hGH in the extracts were corrected to the amounts of extracted protein. The 47-GH and control samples contained 100 times more protein than the 180-GH samples.
Recife, Brazil, and is partially supported by the Brazilian Research Council. E.B.C. acknowledges support from the Department of Biology, University of California, San Diego. C.K.G. is an L. B. Markey Foundation Scholar. 1. Farquhar, M. G., Skutelsky, E. H. & Hopkins, C. R. (1975) in The Anterior Pituitary, eds. Tixier-Vidal, A. & Farquhar, M. G. (Academic, New York), pp. 83-135. 2. Dada, M. O., Campbell, G. T. & Blake, C. A. (1984) J. Endocrinol. 101, 87-94. 3. Barta, A., Richards, R. I., Baxter, J. D. & Shine, J. (1981) Proc. Natl. Acad. Sci. USA 78, 4867-4871. 4. Linzer, D. I. H. & Talamantes, F. (1985) Proc. Natl. Acad. Sci. USA 260, 9574-9579.
5. Maniatis, T., Goodburn, S. & Fischer, J. A. (1987) Science 236, 12371245. 6. Nelson, C., Crenshaw, E. B., III, Franco, R., Lira, S. A., Albert, V. R., Evans, R. M. & Rosenfeld, M. G. (1986) Nature (London) 322, 557-562. 7. Cattini, P. A., Peritz, L. N., Anderson, T. R., Baxter, J. D. & Eberhardt, N. L. (1986) DNA 5, 503-509. 8. Crew, M. D. & Spindler, S. R. (1986) J. Biol. Chem. 261, 5018-5022. 9. Casanova, J., Copp, R. P., Janocko, L. & Samuels, H. H. (1985) J. Biol. Chem. 260, 11744-11748. 10. Ye, Z. & Samuels, H. H. (1987) J. Biol. Chem. 262, 6313-6317. 11. West, B. L., Catanzaro, D. F., Mellon, S. H., Cattini, P. A., Baxter, J. D. & Reudelhuber, T. L. (1987) Mol. Cell. Biol. 7, 1193-1197. 12. Nelson, C., Albert, V., Elsholtz, H. P., Lu, L. I. W. & Rosenfeld, M. G. (1988) Science 239, 1400-1405. 13. Larsen, P. R., Harney, J. W. & Moore, D. D. (1986) Proc. NatI. Acad. Sci. USA 83, 8283-8287. 14. Palmiter, R. D. & Brinster, R. L. (1986) Annu. Rev. Genet. 20, 465-499. 15. Hogan, B., Costantini, F. & Lacy, E. (1986) Manipulating the Mouse Embryo: A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Habor, NY). 16. Palmiter, R. D., Chen, H. Y. & Brinster, R. L. (1982) Cell 29, 701-710. 17. Shields, D. & Blobel, G. (1977) Proc. Natl. Acad. Sci. USA 74, 20592063. 18. Swanson, L. W., Sawchenko, P. E., Rivier, J. & Vale, W. W. (1983) Neuroendocrinology 36, 165-183. 19. Peterson, G. L. (1983) Methods Enzymol. 91, 95-119. 20. Lefevre, C., Imagawa, S., Dana, S., Gridley, J., Bodner, M. & Karin, M. (1987) EMBO J. 6, 971-981. 21. Glass, C. K., Franco, R., Weinberger, C., Albert, V. R., Evans, R. M. & Rosenfeld, M. G. (1987) Nature (London) 329, 738-741. 22. Hall, R., Peters, J. R., Foord, S., Dieguez, C., Scanlon, M. F., Arnao, M. D. R. & Gomez-Pan, A. (1984) in Pituitary Hyperfunction: Physiopathology and Clinical Aspects, eds. Camanni, F. & Muller, E. E. (Raven, New York), pp. 303-314. 23. Hinkle, P. M. (1984) in Secretory Tumors of the Pituitary Gland, eds. Black, P. McL., Zervas, N. T., Ridgway, E. C. & Martin, J. B. (Raven, New York), pp. 25-43. 24. Frawley, L. S., Boockfoor, F. R. & Hoeffler, J. P. (1985) Endocrinology 116, 734-737. 25. Nikitovich-Winer, M. B., Atkin, J. & Maley, B. E. (1987) Endocrinology 121, 625-630. 26. Roux, M., Bartke, A., Dumont, F. & Dubois, M. P. (1982) Cell Tissue Res. 223, 415-420. 27. Hammer, R. E., Palmiter, R. D. & Brinster, R. L. (1984) Nature (London) 311, 65-67.
