Proc. NatI. Acad. Sci. USA
Vol. 88, pp. 8327-8331, October 1991
Expression of biologically active heterodimeric bovine
follicle-stimulating hormone in milk of transgenic mice
NORMAN M. GREENBERG*, JOSEPH W. ANDERSON*t, AARON J. W. HSUEH*, K. NISHIMORIt,
JERRY J. REEVES§, DAVID M. DEAVILA§, DARRELL N. WARD$, AND JEFFERY M. ROSEN*II
*Department of Cell Biology, Baylor College of Medicine, Houston, TX 77030; *Department of Gynecology and Obstetrics, Stanford University School of
Medicine, Stanford, CA 94305; §Department of Animal Sciences, Washington State University, Pullman, WA 99164; and IDepartment of Biochemistry
and MolecularBiology,M.D. Anderson Cancer Center, Houston, TX 77030
Communicated by Neal L. First, June 6, 1991 (receivedfor review April 15, 1991)
litropin) is a pituitary glycoprotein composed of two post-
translationally modified subunits, which must properly assem-
ble to be biologically active. FSH has been difficult to purify
and to obtain in quantities sufficient for detailed biochemical
studies. We have targeted FSH expression to the mammary
gland of transgenic mice by using cDNAs encoding the bovine
a andFSHP subunits and a modified rat f-casein gene-based
expression system. Lines of bigenic mice expressing both
subunits have been generated either by coinection of the
subunit transgenes or by mating mice that acquired and
expressed transgenes encoding an individual subunit. Up to 60
international units (15 ,sg) of biologically active FSH per ml
was detected in the milk ofthe bigenic mice. These lines provide
a model system for studying the post-transcriptional mecha-
nisms that effect the expression and secretion of this het-
Follicle-stimulating hormone (FSH; fol-
Follicle-stimulating hormone (FSH; follitropin) is amember of
the glycoprotein family ofpituitary hormones, which includes
thyroid-stimulating hormone (TSH), luteinizing hormone
(LH), and chorionic gonadotropin (CG). Like LH and CG,
FSH is a gonadotropin and is composed of a common a
subunit that is noncovalently linked to a hormone-specific P
subunit (1, 2). FSH has been difficult to purify and to obtain
in sufficient quantities for detailed biochemical studies (for a
review, see ref. 3). The a and FSHI
translationally modified, and the nature and extent of such
modifications can exertaprofound effect on subunit assembly,
secretion, and stability (4-6). Only heterodimers with appro-
priately glycosylated subunits exhibit significant biological
and receptor-binding activity (5, 7, 8). Targeting FSH to the
mammary gland oftransgenic animals would, therefore, serve
as a model system in which to study glycoprotein processing
and secretion as well as a means to produce large quantities of
FSH. A standardized source of recombinant FSH would be
useful to both human and livestock fertilization programs to
achieve the reproducible development of ovarian follicles.
Several different milk protein-based constructs have been
employed to express diverse heterologous proteins in the milk
ofa variety oftransgenic animals (forreviews, see refs. 9-11).
We have demonstrated previously that a -524/+490 minimal
rat f3-casein promoter fragment can direct the expression of
chloramphenicol acetyltransferase to the mammarygland (12).
To determine whether the mammary gland could be used to
secrete large quantities of a bioactive heterodimeric protein
into milk, we have used a modified rat ,8-casein-based vector
to target and express bovine FSH (bFSH) to the mammary
gland and into the milk of transgenic mice.
subunits are post-
MATERIALS AND METHODS
Construction of the Transgenes. The FSH subunit cDNAs
were obtained from Genzyme; a as a 730-base-pair (bp)
EcoRI fragment and FSH(3 as a 560-bp EcoRI/BamHI frag-
ment. The cDNA fragments were inserted into pUC19 (13)
with the rat 13-casein -524/+490fragment (12) and an 850-bp
EcoRI fragment carrying the simian virus 40 small tumor
antigen intron with transcript cleavage and polyadenylylation
signals (kindly provided by S. Berget, Baylor College of
Medicine). A 408-bp HindIll fragment of the mouse mam-
mary tumor virus long terminal repeat (LTR) carrying four
glucocorticoid response element (GRE) sequences (kindly
provided by M. Parker, Imperial Cancer Research Fund
Laboratories) was placed at -330 in the rat 13-casein fragment
of the a construct.
Production and Screening of Transgenic Mice. Transgenic
mice were generated and mouse tail DNA was isolated as
described previously (12). The polymerase chain reaction
(PCR) was employed to screen for positive transgenic mice.
The sequences ofthe synthetic oligonucleotides used in PCR
reactions were as follows (5'
TCTCTTGTCCTCCGC; 2, ACAGAGACAAAATGGCCA-
GAATGAC; 3, GCYTl ATTGCTTTTCTCCTTATCCT; 4,
TCTCTGTAGGTAGTTTGTCCAATTA; 5, AGGCATTC-
GAACTGGACAGACT; and 7, TACTGACCTCTGCTCTC-
Transgene cointegration was analyzed by Southern blot-
ting tail DNA (10 ug) digested with EcoRI. Blots were
hybridized with 32P-labeled a- or FSHf3-specific probes pre-
pared by random oligonucleotide labeling.
RNA Isolation and Analysis. Total RNA was isolated from
mouse mammary gland tissue by the method of Chirgwin et
al. (14). For Northern blots, RNA (20,4g)was fractionated in
agarose gels containing formaldehyde (15). For slot blots,
RNA (1, 2, or 4 iug) was applied to ZetaProbe membrane
(Bio-Rad) and compared to known amounts ofthe a- orFSH,3
cDNAs included as standards. Quantitation was performed
by scanning with an LKB laser densitometer.
Collection ofMouse Milk. Mice were anesthetized with 1 ml
ofAvertin (20mg/ml) administered i.p. immediately prior to
milking, and 0.5 ml of oxytocin [200 international units
(IU)/ml; Sigma] was administered i.p. before milk samples
were harvested by gentle suction into tubes at 40C. The whey
fraction was prepared by centrifugation ofskim milk at 16,000
x g for 15 min at 40C.
-. 3'): 1, GAGCTTCATCTTC-
Abbreviations: FSH, follicle-stimulating hormone (follitropin);
bFSH, bovine FSH; rbFSH, recombinant bFSH; oFSH, ovine FSH;
GRE, glucocorticoid response element; IU, international units.
tPresent address: Animal Resource Center, University of Colorado
Health Sciences Center, Denver, CO 80262.
"To whom reprint requests should be addressed.
The publication costs ofthis 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.
Biochemistry: Greenberg et al.
Characterization of FSH in Mouse Milk. A heterologous
double-antibody radioimmunoassay (RIA) was performed as
described (16). For immunoblot analysis, whey protein (300
,ug) was fractionated by SDS/PAGE at room temperature
(15). Samples were not heated and did not contain 2-mer-
captoethanol. Ovine FSH (oFSH; NIADDK-oFSH-16; 20
National Institutes of Health units/mg) was added to non-
transgenic mouse milk for positive controls. A sample of
recombinant bFSH (rbFSH) made in Chinese hamster ovary
(CHO) cells (a gift of Genzyme) was used to assess the
cross-reactivity of the antibody in this assay. Blots were
probed with the JAD-17-689 antiserum (16) (1:5000), kindly
provided by J. Dias (State of New York Department of
Health), and developed with a goat anti-rabbit IgG-
horseradish peroxidase enhanced chemiluminescence (ECL)
detection scheme (Amersham).
For radioreceptor assays, samples were initially diluted
with an equal volume of assay buffer (100 mM TrisHCl/100
mM sucrose/5 mM MgCl2/0.1% bovine serum albumin, pH
7.4) and incubated with a chicken testis receptor preparation
(17). The standard was NIH-FSH-S9 (18). Data analysis was
by the ALLFIT(FLEXIFIT) program, version 2.6 (Laboratory of
Theoretical and Physical Biology, National Institute ofChild
Health and Human Development). Calculation of the ng of
rbFSH was based on the specific activity of bFSH (19).
Calculation ofFSH activity was from the ED50 ofthe assay.
Acid-dissociation radioreceptor assay experiments (20) mea-
sured FSH activity in 50 ,ul of milk surviving 1 M propionic
acid treatment for 1 hr at 37°C. Precipitated casein and other
milk proteins were removed by centrifugation. The granulosa
cell bioassay and chromatofocusing analysis were performed
as described (5, 21).
RESULTS AND DISCUSSION
Characterization of the a and FSHl Transgenes. The
cDNAs encoding a or FSH8 were placed into a rat f3-casein
expression vector (Fig. 1). Lines of transgenic mice were
generated by either individual or coinjection (12, 22) ofa and
FSHI3 constructs. A construct carrying four copies ofaGRE
from the mouse mammary tumor virus promoter (Fig. 1B)
was also employed to direct high-level a-subunit expression,
since in the normal pituitary a is expressed in excess overthe
dimeric hormone (23-27). Screening by PCR identified
founder animals carrying either the a or the FSHB construct
or both (Fig. 1 D and E).
Southern blot experiments were used to characterize the
architecture of the integrated transgenes. Since the trans-
genes carry a single EcoRI site, the detection of strongly
hybridizing species in the 2- to 3-kb range (Fig. 2A and B) is
diagnostic for transgene cointegration. When the a orFSHP
transgenes were coinjected, multiple copies were found to be
cointegrated in >85% of the positive lines. Only a few lines
carried individual transgenes [e.g., line 7905 carries a single
a transgene (Fig. 2A and B, lane 1)]. Some head-to-head and
tail-to-tail cointegration events occurred. Divergent PCR
confirmed the head-to-tail orientation (Fig. 2C).
Expression of a and FSHU mRNAs In Mammary Glnds of
Lactating Transgenic Mice. Northern blot analysis indicated
the presence ofmajor 1519-nucleotide a and 1340-nucleotide
FSH. mRNA species (Fig. 3) corresponding to the expected
transcript sizes. The smaller a mRNA species (Fig. 3A, lane
3) may arise from cleavage and polyadenylylation at signals
within the 3' untranslated region of the a cDNA (28, 29).
Examination of the transcripts by reverse transcriptase-
mediated PCR indicated that most a and FSH,
species encode unit-length proteins (N.M.G. and J.M.R.,
When Northern blots were rehybridized with a mouse
f-casein exon 7 probe, the 1-casein mRNA level wasfound to
be -5- to 10-fold greater than that observed for the a-subunit
transgenic mice by PCR. (A) Structure ofthe a cDNA transgene. The
elements are -524 to +1, the 5' flanking region of rat P-casein; +1 to
+43, thenoncoding firstexon ofratP-caseinand 5' splicedonor; +490,
a cDNA open reading frame; hatched region, simian virus 40 small
tumor antigen splice and polyadenylylation signals. The primers used
for PCR were 1, 2, 3, 4, and 5 (see Materils and Methods). (B)
Structure oftheGRE-enhaned acDNA transgene. ThefourGREs are
denoted by thin vertical bars. (C) Structure of the FSHj3 cDNA
transgene. ATG and TAA denote the bFSHI8 cDNA open reading
frame. TheprimersusedforPCRwere6and 7. (D)PCRanalysisforthe
a transgene. Primers 1 and 2 were used to screen tail DNA by PCR.
Lanes 1-10 representfounder mice 7905, 7502, 2038, 7398, 7485, 7919,
7389,7667,7668,and7904, respectively.Lane11, nontransgenicmouse
control. The sizes ofthe PCR productsare shown on therightand the
migration of the DNA markers (lane M) is shown on left in kilobases
(kb). (E) PCR analysis for the FSH(3 transgene. Primers 1 and 6 were
used to screen tail DNA by PCR. Lanes as in D.
mRNA (data not shown). Since ,3-casein mRNA has been
estimated to make up -20% of the total mRNA at day 10 of
lactation, the level ofthe a-subunitmRNA shouldcorrespond,
therefore, to -2% of the total mRNA, in agreementwith the
quantitative slot blot determination (see below).
Two-thirds of the mice carrying the GRE-enhanced a
construct expressedthetransgene,whereasonlyone-sixth of
those lackingtheGREexpresseda. Ofmice carryingminimal
a and FSHI3 constructs, 3 of 11 (27%) expressed both
transgenes, while 6 of 10 (60%6) expressed the cointegrated
GREa and FSH(3 constructs. Lines oftransgenicmice car-
rying the GRE-enhanced constructs expressed more fre-
quently (30, 31) and at higher levels (see below). Line 7905
(single copyofGREa)has been bred to line 7502(4to 6copies
of a tandemly arranged FSHI3) to establish line 2038 (Fig.1
D and E), which expressed both independent loci (Fig. 3).
Secretion of rbFSH Into Mouse Milk. rbFSH was detected
in milk collected at lactation (Fig. 4A), but not in milk from
nontransgenic littermates, by using a heterologous double-
Structure of the rbFSH transgenes and identification of
Proc. NatL Acad Sci. USA 88(1991)
Proc. Natl. Acad. Sci. USA 88 (1991)
antibody RIA. Proteins present in normal mouse milk did not
interfere with the assay.
The species of FSH present in milk were further charac-
terizedby immunoblotting (Fig. 4B). Preparations ofpituitary
oFSH (lanes B, C, and D in Fig. 4B) and rbFSH prepared
from transfected CHO cells (lane I in Fig. 4B) were included
as controls. The antiserum tooFSH detected a species of 38
kDa in the milk from bigenic mouse 8942 (lanes E and K in
Fig. 4B). The 38-kDa species corresponds in size to the
species detected in both the oFSH and rbFSH standards.
Some microheterogeneity in the post-translational modifica-
tions ofthe FSH may explain the broad bands observed (see
chromatofocusing results) (32, 33). One microgram of the
oFSH standard (lane B) gave amuch stronger signal at 38kDa
than an equivalent amount of the CHO rbFSH protein (lane
; 4 f3
at day 10 of lactation. (A) Hybridized with the a cDNA insert. Lane
1, RNA prepared from a nontransgenic mouse; lanes 2-6, RNA from
mouse (line number) 9667 (7905), 9415 (7398), 8611 (7667), 2038
(2038), and 9434 (7502), respectively. The relative migration of the
18S and 28S ribosomal RNAs is shown on the left. Length in
nucleotides is given on the right. Time of autoradiography was 30
min. (B) Hybridized with the FSH3 cDNA insert. Lanes as in A.
Time of autoradiography was 20.5 hr.
Northern analysis of FSH mRNA from transgenic mice
ies. (A) EcoRI-digested tail DNA was hybridized with the
.DNA insert (lanes as in Fig. 1 D and E). Sizes of DNA
rkers ate shown on the left. (B) Blot in A was re-probed
h the FSHp-cDNA insert. (C) Divergent PCR analysis of
IDNA from mouse 7667. PCR reactions were performed
th a- or FSH,-specific primers (Fig. 1) to detect cointegra-
n events. Lane 1, primers 2 and 3; lane 2, primers 3 and 7;
e 3, primers 3 and 6; lane 4, primers 2 and 7; lane 5, primers
ad 6; and lane 6, primers 6 and 7.
Southern blot analysis of integrated bFSH trans-
I), reflecting that the antiserum was raised against oFSH
rather than bESH.
A strongly immunoreactive species with a mass of 18 kDa
was detected in the milk from bigenic mice, as well as from
mice expressing only the a-subunit mRNA and may be free
a subunit. This was not detected in the control milk sample.
The immunoblot and RIA results suggest the polyclonal
anti-oFSH antiserum can crossreact with both heterodimer
and free a subunit, and it may contain species recognizing
free bovine a and heterodimer, but with different affinities.
Therefore, the immunoblot could not be used to quantita-
tively determine the relative abundance of a and rbFSH.
Steady-state a andFSH.3mRNAs were quantitated by slot
blot hybridization analysis. Summarized in Table 1, amRNA
levels were consistently higher, 7- to 17-fold, than those for
FSHf3 mRNA. Levels of both mRNAs were independent of
transgene copy number; line 7905 carries a single GREa
transgene yet expresses high levels of a mRNA. Consistent
with previous results, expression appeared to be highly
dependent on the site ofintegration (11, 12, 31), and the level
of mRNA was observed to vary as much as 3-fold between
littermates (e.g., a-FSH mRNA, line 7919). The relatively
high a and low FSH/3 mRNA levels suggest that post-
transcriptional mechanisms influence their steady-state lev-
els, supporting the hypothesis that the 3' untranslated region
ofFSH.8mRNA may impart instability (34).
Milk of bigenic mouse 8942 was capable of displacing an
125I-labeled purified porcine FSH preparation from chicken
testis FSH receptors (Fig. 4C) (2, 17). No displacement was
observedwhen milkfromanontransgenic littermate was used.
The calculated competitive binding displacement (single point
assay, Fig. 4C) of milk treated with 1 M propionic acid was
equivalent toonly82 ngofFSHascompared to2000ngofFSH
in 501d of untreated milk, representing 96% inactivation.
Biochemistry: Greenberget aL
., -6, 144
Biochemistry: Greenberg et al.
standard; (A), USDA BS standard in nontransgenic milk. Sample inhibition curves are (values obtained for FSH in milk calculated from the
dilution giving 50o inhibition in parentheses): (o), milk 8942 (2.3 mg/ml); (A), milk 8611(2.5 mg/mnl); o, milk 1262 (0.63 mg/ml). (B) Immunoblot
analysis of FSH in milk. Milk samples fractionated by SDS/PAGE were probed with JAD-17-679. Lanes A-D, normal mouse milk with 0, 1,
0.5, 0.25 ,ug of oFSH added, respectively. Lanes E-H are milk from mice (line number in parentheses) 8942 (7919), 9667 (7905), 9434 (7502),
and 2038 (2038), respectively. Lane I, CHO rbFSH (200 ,g of total protein; 1.25-,.gof FSH). Lanes J and K are 2-hr exposures of lanes D and
E. Lanes L and M show lanes equivalent to D and E from a gel run in parallel and stained with Coomassie blue. (C) Competitive binding
experiments for rbFSH in milk, using a chicken testis radioreceptor assay. Samples used were FSH (NIH-FSH-S9; o), milk from transgenic
mouse 8942 (m), and milk froim a nontransgenic mouse (e). The putative ng ofrbFSH has been plotted in comparison with the ng ofNIH-FSH-S9
used in the assay (to avoid the weight-to-dilution comparison). A comparable dilution for the control milk (0) is shown on the same scale. A
sample of milk 8942 treated in 1 M propionic acid for 1 hr at 37°C (*) was also analyzed. (D) Analysis of bioactive rbFSH by granulosa cell
aromatase assay. Granulosa cell cultures were treated with increasing aliquots ofmilk whey protein fractions from transgenic and control mice.
Symbols as in A; data are mean + SEM.
Characterization ofrbFSH in milk oftransgenic mice. (A) RIA using rabbit antiserum JAD-17-679. Standard curves: (0), USDA B5
These results, summarized in Table 1, indicate that the rbFSH
secreted into milk can interact with FSH receptors.
To measuie the biological activity of the rbFSH, rat
granulosa cell in vitro bioassays were utilized (Fig. 4D). In
granulosa cells, FSH stimulates both the conversion of
cholesterol to pregnenolone and the aromatization of the
estrogen precursor androstenedione (33). The results are
summarized in Table 1. Milk samples from independent
bigenic lines (mice 7994, 1262, 8611, and 8942) contained high
levels of biologically active FSH. High FSH activity was
detected in milk from line 2038, while milk from the parental
lines contained no detectable bioactive FSH. Therefore,
Summary of FSH subunit mRNA levels in the mammary gland and FSH activity in the milk of lactating
ng a mRNA/
ng FSH3 mRNA/
,ug total RNA
<0.1 (n = 1)
<0.1 (n = 1)
<0.1 (n = 1)
<0.1 (n = 1)
39.2 (n = 1)
39.2 ± 15.5 (n = 4)
66.2 + 14.8 (n = 5)
36 (n = 1)
GREa x FSH,3
4.2 (n = 6)
Values in columns A and B are rounded off. The A/B ratio is accurate to two significant figures. FSH subunit mRNA
levels are expressed as ng ofspecific mRNA per ,ug oftotal RNA per assay. To assay FSH in milk, samples harvested from
lactatingfemale mice were characterized by the radioreceptorand in vitrogranulosacellbioassays;results aregiven ±SEM.
ND, not determined.
Proc. Natl. Acad. Sci. USA 88(1991)
Proc. NatL Acad. Sci. USA 88 (1991) Download full-text
FIG.5.Analysis of rbFSH by chromatofocusing.
rbFSH from bigenic mouse milk ortransfected CHO cells (Genzyme)
were chromatographed on a PBE-94 column (Pharmacia). The pH (e)
and ability to stimulate estrogen synthesis by 8942 milk (-; ~800 ng
of FSH) and CHO FSH (v;
"2.5 ,ug of FSH) were determined from
production of estrogen
specifically related to the presence of heterodimeric FSH.
The amounts of rbFSH in milk from mouse 8942 determined
by the radioreceptor and granulosa cell assays were quite
similar. This value of 66 IU/ml corresponds to
FSH per mi, assuming 1 ,ug of FSH = 4 IU of FSH (19). No
"45.3 Zg of
adverse reproductive consequences were observed in the
expressing bigenic lines, and rbFSH was not detected in the
serum collected from lactating animals (D. Bolt, personal
communication), suggesting that rbFSH is secreted vectori-
ally into milk. Normal patterns of transgene transmission and
expression have been observed in several litters of offspring
for all our expressing lines.
Since isoforms of FSH can be separated on the basis of
their isoelectric properties, which are, in part, related to
terminal sialic acid content (32, 33) chromatofocusing anal-
ysis was performed. Fractions from 8942 milk (4 IU total) and
the CHO rbFSH (10 IU total) were collected, the pH was
measured, and FSH activities were determined by granulosa
cell bioassay. The 8942 rbFSH had one major peak of activity
between pH values 6.1 and 4.2 (Fig. 5) and the CHO rbFSH
profile exhibited a similar single peak (pH 5.2 to 4.0). Both
profiles are consistent with the observations of Galway et al.
(5) for FSH with appropriate N-linked carbohydrate struc-
tures. The broader peak observed for transgenic rbFSH
probably reflects the capacity of the mammary gland to add
terminal sialic acid residues to these proteins. The lack of
terminal sialic acid residues does not affect FSH receptor
binding or in vitro bioactivity but may, however, be related
to enhanced clearance rates for FSH in blood plasma (5, 32).
By several independent criteria we have demonstrated that
rbFSH can be produced
/3-casein expression system. As post-transcriptional mecha-
nisms are probably responsible for the differences observed
in the relative levels of the a and FSH~subunit mRNAs, it
may be possible with appropriate engineering to express a
more stable FSHf3 mRNA, thereby increasing the levels of
a "second generation"
transgenes has been constructed to determine whether higher
precise exchange of FSHA 3 and a cDNA open reading f ames.
transgenes provide useful models for the study of the mech-
anisms regulating the post-translational processing of both
the individual subunits and the heterodimer. Finally, these
levels of FSHj3 mRNA
willresult from the
individual a,and FSHo3
studies have demonstrated that the mammary gland can be
used as abioreactor to direct the high expression and vectoral
secretion into milk of heterodimeric proteins requiring ex-
tensive post-translational modifications. Although the quan-
tities ofglycosylated hormone produced in mice are sufficient
for further biochemical analysis, the introduction of such
transgenes into livestock (9) will be required to provide
sufficient quantities for both research and commercial pur-
We thank T. Duffy, F. Farzam, C. Nichols, and Hyun Nahm for
technical assistance. Hormone standards were gifts of the U.S.
Department of Agriculture Animal Hormone Program. This work
was supported in part by the U.S. Department ofAgriculture (Grant
88-37266-3951 to J.M.R.) and Granada Biosciences (J.M.R.) and the
National Institutes of Health (Grant DK-09801 to D.N.W.).
Pierce, J. G. & Parsons, T. F. (1981) Annu. Rev. Biochem. 50, 466-495.
Ward, D. N., Bousfield, G. R. & Moore, K. H. (1991) in Reproduction
in Domestic Animals, ed. Cupps, P. T. (Academic, New York), pp.
Liu, W.-K. & Ward, D. N. (1975) Pharmacol. Ther. B. 1, 545-570.
Matzuk, M. M., Kornmeier, C. M., Whitfield, G. K., Kourides, I. A. &
Boime, I. (1988) Mol. Endocrinol. 2, 95-100.
Galway, A. B., Hsueh, A. J. W., Keene, J. L., Yamoto, M., Fauser,
B. C. J. M. & Boime, I. (1990) Endocrinology 127, 93-100.
Kaetzel, D. M., Virgin, J. B., Clay, C. M. & Nilson, J. H. (1989) Mol.
Endocrinol. 3, 1765-1774.
Strickland, T. W., Thomason, A. R., Nilson, J. H. & Pierce, J. G. (1985)
J. Cell. Biochem. 29, 225-237.
Sairam, M. R. & Bhargavi, G. N. (1985) Science 229, 65-67.
Hennighausen, L. (1990) Protein Expression Purification 1, 3-8.
Harris, S., McClenaghan, M., Simons, J. P., Ali, S. & Clark, A. J. (1990)
J. Reprod. Fertil. 88, 707-715.
Rosen, J. M., Bayna, E. M. & Lee, K.-F. (1989) Mol. Biol. Med. 6,
Lee, K.-F., DeMayo, F. J., Atiee, S. H. & Rosen, J. M. (1988) Nucleic
Acids Res. 16, 1027-1041.
Yanisch-Perron, C. J., Vieira, J. & Messing, J. (1985) Gene 33, 103-119.
Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter, W. J.
(1979) Biochemistry 24, 5294-5299.
Sambrook, J., Fritsch, E. & Maniatis, T. (1989) Molecular Cloning: A
Laboratory Manual(Cold SpringHarborLab., Cold Spring Harbor, NY),
Acosta, B., Tarnavsky, G. K., Platt, T. E., Hamernik, D. L., Brown,
J. L., Schoenemann, H. M. & Reeves, J. J. (1983) J. Anim. Sci. 57,
Gordon, W. L., Bousfield, G. R. & Ward, D. N. (1989) J. Endocrinol.
Invest. 12, 383-392.
Reichert, L. E., Jr., & Wilhelmi, A. E. (1973) Endocrinology 92, 1301-
Cheng, K.-W. (1976) Biochem. J. 159, 651-659.
Cheng, W. K. (1978) J. Biochem. 17S, 29-34.
Hsueh, A. J., Bicsak, T. A., Jia, X. C., Dahl, K. D., Fauser, B. C.,
Galway, A. B., Czekala, N., Pavlou, S. N., Papkoff, H. & Keene, J.
(1989) Recent Prog. Horm. Res. 45, 209-273.
Palmiter, R. D. & Brinster, R. L. (1986) Annu. Rev. Genet. 20, 465-499.
Paul, S. J., Ortolando, G. A., Haisenleder, D. J., Stewart, J. M., Shup-
nick, M. A. & Marshall, J. C. (1990) Mol. Endocrinol. 4, 1943-1955.
Dalkin, A. C., Haisenleder, J., Ortolando, G. A., Suhr, A. & Marshall,
J. C. (1990) Endocrinology 127, 798-806.
Bielinska, M. & Boime, I. (1979) Proc. Nati. Acad. Sci. USA 76,
Ortolando, G. A., Haisenleder, D. J., Dalkin, A. C., Iliffe-Sizemore,
S. A., Landefeld, T. D., Maurer, R. A. & Marshall, J. C. (1988) Endo-
crinology 123, 2946-2948.
Workewych, J. & Cheng, K. (1979) Endocrinology 104, 1075-1082.
Jameson, J. L., Becker, C. B., Lindell, C. M. & Habener, J. F. (1988)
Mol. Endocrinol. 2, 806-815.
Esch, F. S., Mason, A. J., Cooksey, K., Mercado, M. & Shimasaki, S.
(1986) Proc. Natl. Acad. Sci. USA 83, 6618-6621.
Greenberg, N. M. & Rosen, J. M. (1990) Mouse Molecular Genetics
(Cold Spring Harbor Lab., Cold Spring Harbor, NY), p. 203.
Greenberg, N. M., Reding, T. R., Duffy, T. & Rosen, J. M. (1991) Mol.
Endocrinol., in press.
Robertson, D. M. (1989) in Microheterogeneity of Glycoprotein Hor-
mones, eds. Keel, B. A. & Grotjan, H. E., Jr. (CRC, Boca Raton, FL),
Chappel, S. C., Ulioa-Aguirre, A. & Coutifaris, C. (1983) Endocrinol.
Rev. 4, 179-211.
Keene, J. L., Matzuk, M., Otani, T., Fausser, B. C. J. M., Galway,
A. B., Hsueh, A. J. W. & Boime, I. (1989) J. Biol. Chem. 264, 4769-
Biochemistry: Greenberget aL