Oncogene (1997) 14, 2259 ± 2264
 1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00
Gli family members are di€erentially expressed during the mitotic phase of
spermatogenesis
Stephan P Persengiev1, Ivanela I Kondova1,3, Clarke F Millette2 and Daniel L Kilpatrick1
1
Worcester Foundation for Biomedical Research, Neurobiology Group, 222 Maple Avenue, Shrewsbury, Massachusetts 01545;
Department of Cell Biology and Neuroscience, University of South Carolina School of Medicine, Columbia, South Carolina 29208,
USA
2
The Gli family of DNA binding proteins has been
implicated in multiple neoplasias and developmental
abnormalities, suggesting a primary involvement in cell
development and di€erentiation. However, to date their
speci®c roles and mechanisms of action remain obscure,
and a drawback has been the lack of a model system in
which to study their normal function. Here we
demonstrate that Gli family members are di€erentially
expressed during spermatogenesis in mice. Speci®cally,
Gli and Gli3 mRNAs were detected in mouse germ cells,
while Gli2 was not. Further, both Gli and Gli3 exhibited
stage-dependent patterns of expression selectively in
type A and B spermatogonia. Gli expression was
somewhat higher in type B spermatogonia while the
abundance of Gli3 transcripts was similar in type A and
B cells. Gel-shift analyses also demonstrated the
enrichment of DNA binding activity speci®c for the
Gli target sequence in spermatogonial cells. These
results indicate a selective role for Gli and Gli3 during
mitotic stages of male germ cell development. Spermatogenesis may thus provide a unique opportunity to
identify downstream targets and explore the normal
function of Gli family proteins.
Keywords: proto-oncogenes; spermatogonia;
binding proteins; male germ line
DNA
Introduction
Many proto-oncogenes appear to function as transcriptional regulators at various stages of cell development
and di€erentiation. The Gli genes constitute a family of
DNA binding proteins that have been implicated in
several neoplasias, including glioblastoma (Kinzler et
al., 1987; Roberts et al., 1989). Three distinct Gli genes
have been identi®ed in human and mouse, Gli, Gli2,
Gli3 (Kinzler et al., 1988; Ruppert et al., 1988, 1990;
Vortkamp et al., 1992; Hui et al., 1994). The Gli
protein has been shown to transform cells in
combination with the adenovirus E1A protein
(Ruppert et al., 1991), consistent with its presumptive
role as an oncogene. The close similarity of Gli proteins
to the developmentally important KruÈppel gene
products, their expression in embryonal carcinoma
Correspondence: DL Kilpatrick
3
Current address: Division of Infectious Diseases, TVDL, Tufts
University School of Veterinary Medicine, North Grafton, MA 01536
Received 9 September 1996; revised 27 January 1997; accepted 27
January 1997
cells (Ruppert et al., 1988), and the association of Gli3
gene abnormalities with developmental defects in mice
and humans (Vortkamp et al., 1991, 1992; Schimmang
et al., 1992; Hui and Joyner, 1993) suggested that this
gene family normally performs a critical role during
development. The identi®cation of the Gli homologues
ci (Orenic et al., 1990) and tra-1 (Zarkower and
Hodgkin, 1993), which regulate development in
Drosophila and Caenorhabditis elegans, respectively,
has provided additional support for this notion.
Further, more recent studies have shown that all three
mouse Gli genes are selectively expressed in developing
mesodermal and ectodermal tissues (Hui et al., 1994;
Walterhouse et al., 1993).
It is presumed that Gli gene products function
during development and cell transformation as
transcriptional regulators. This is based mainly on the
demonstration of DNA binding activity of the
expressed Gli proteins involving their C2H2-type zinc
®nger domains as well as their predominantly nuclear
localization within cells (Ruppert et al., 1990; Kinzler
and Vogelstein, 1990; Pavletich and Pabo, 1993).
However, how these proteins regulate gene transcription, either during cell development or in transformed
cells, is unknown and the lack of identi®ed target genes
for the Gli proteins is a signi®cant limitation to
understanding their function.
Spermatogenesis has served as a valuable model for
investigating cell development and di€erentiation.
During this process, stem cells (spermatogonia)
di€erentiate in a series of mitotic events into primary
spermatocytes. These cells in turn undergo meiosis to
yield haploid spermatids, which subsequently differentiate into mature sperm through a ®nal set of steps
termed spermiogenesis. Coordination of these events is
critical for normal sperm development and involves a
complex set of control mechanisms in which stagedependent gene transcription is essential. In fact,
multiple transcriptionally active proto-oncogenes are
expressed in distinct stage-dependent patterns during
spermatogenesis (Propst et al., 1988; Wolfes et al.,
1989). We previously identi®ed a Gli consensus binding
sequence within a region of the rat and mouse
proenkephalin germ line promoters that formed
speci®c DNA/protein complexes with germ cell
nuclear factors (Galcheva-Gargova et al., 1993). This
prompted us to investigate whether the Gli gene family
is expressed during mouse spermatogenesis. Our results
indicate that a subset of these putative transcription
factors are preferentially expressed during the initial
stages of spermatogenesis, where they are likely to be
important in the mitotic proliferation of early
precursor cells.
Gli family expression during spermatogenesis
SP Persenglev et al
2260
Results
Expression of Gli genes during spermatogenesis
Previous studies suggested that Gli-related proteins
were present in rodent testes and spermatogenic cells
(Galcheva-Gargova et al., 1993). To address this
directly, Northern analysis was performed on puri®ed
mouse cells representing varying stages of sperm
maturation (Figure 1). Expression of Gli and Gli3
was detected within the male germ line, while Gli2
transcripts were not (Figure 2). This expression was
restricted to spermatogonial cells, with Gli mRNA
somewhat enriched in type B spermatogonia, and Gli3
transcripts of similar abundance in type A and B cells.
Expression of both genes was very low or undetectable
in later developmental stages. The sizes of the Gli and
Gli3 transcripts detected in spermatogonia were
identical to those in mouse embryos (&4 kb and
9 kb, respectively). The fact that two distinct Gli
E
A
B
PL L/Z
EP LP RS
C
transcripts (4 and 4.4 kb) were not resolved, as
previously reported for mouse embryos (Hui et al.,
1994), likely re¯ects band compression due to adjacent
28S ribosomal RNA in our experiments. The expression pattern for Gli and Gli3 during spermatogenesis is
consistent with their very low transcript levels in adult
mouse testis (Figure 1), since spermatogonial cells are
present in much smaller numbers than later spermatogenic stages in this tissue.
Analysis of Gli and Gli3 expression during postnatal
development of the mouse testis further substantiated
this pattern of regulation in spermatogenic cells.
Transcripts for both Gli family members were readily
detectable by Northern analysis in total RNA from
testes of day 8 mice, in which spermatogonial cells are
the predominant germ cell stages (Figure 2). However,
these mRNAs were already undetectable in mouse
testis by postnatal day 17 when spermatocytes have
become much more abundant than spermatogonia
(Bellve et al., 1977b). (Figure 2). This is consistent
with restriction of Gli family expression mainly due to
spermatogonial germ cell stages.
T
Polysome analysis of Gli mRNAs in mouse testis
GLI
GLI2
GLI3
GAPDH
Figure 1 Northern blot analysis of Gli family mRNAs in mouse
embryos, testes and spermatogenic cells. E: 14-day old embryo; A:
type A spermatogonia; B: type B spermatogonia; PL: preleptotene
spermatocytes; L/Z: leptotene/zygotene spermatocytes; EP: early
pachytene spermatocytes; LP: late pachytene spermatocytes; RS:
round spermatids; C: cytoplasts; T: testes. Twenty mg of total
RNA was examined in each lane for Gli, Gli2, Gli3 and GAPDH
mRNAs
8d
17d
S
Gli
Gli3
GAPDH
Figure 2 Developmentally regulated expression of Gli and Gli3
mRNAs in mouse testes. Total RNA (20 mg) was examined from
8 day and 17 day old mouse testes and 14 day mouse Sertoli
cells. Northern blot hybridizations were preformed sequentially
with mouse Gli, Gli3 and GAPDH cDNAs
Since many mRNAs expressed during spermatogenesis
are ineciently translated (Kleene, 1996), we examined
the distribution of Gli and Gli3 transcripts on
polysomes from mouse testis using RNase protection.
The degree to which mRNAs are speci®cally associated
with larger polysomal fractions provides a measure of
their active translation in a given tissue. The majority
of Gli and Gli3 transcripts were localized to the larger
polysomal fractions (Figure 3; HKMG gradient). This
association was speci®c (i.e., required polysomal
integrity) since these mRNAs migrated with lighter,
monosomal and post-polysomal fractions in the
presence of EDTA (HKE gradient). These results
indicated that Gli and Gli3 mRNAs are translated
with relatively high eciency in spermatogonia.
Characterization of Gli DNA binding activities in testis
and male germ cells
Nuclear proteins were previously detected in spermatogenic cell extracts that formed cell-speci®c complexes
with a region of the rat and mouse proenkephalin
promoters containing a Gli consensus binding element
in reverse orientation (Galcheva-Gargova et al., 1993).
Gel shift studies were therefore performed to verify
whether authentic Gli DNA binding activities were
present in testis, and in particular, in spermatogonial
cells. Extracts prepared from 14 day mouse embryos
were analysed as a positive control. Three slowly
migrating complexes (A, B, C) were detected in embryo
extracts using a Gli consensus oligonucleotide probe
(B1) that were speci®cally competed by the unlabeled
homologous oligonucleotide but not by one in which
the Gli binding sequences were mutated (B1-mut)
(Figure 4a ± c). A fourth complex of slightly faster
mobility (N) was also detected in all extracts examined
that appeared to be unrelated since it was competed by
the mutated Gli oligonucleotide (Figure 3c). Interestingly, only two of the speci®c complexes (A, C) were
detected in extracts from adult mouse testis (Figure
4b), consistent with the absence of Gli2 expression in
Gli family expression during spermatogenesis
SP Persenglev et al
mouse germ cells. The identities of the individual Gli
family complexes could not be de®nitively established
by supershift assays due to the lack of appropriate
antisera.
Examination of enriched and puri®ed spermatogenic
cells showed that the two testicular Gli DNA-protein
complexes were highly enriched in spermatogonial cells
a
(Figure 4b). Complex A was increased in type B
spermatogonia relative to type A cells, which correlates
with the increased abundance of Gli mRNA observed
in these cells (see Figure 1). Gli-related complexes were
substantially reduced in early pachytene spermatocytes
and were barely detectable in a crude germ cell fraction
consisting mainly of late pachytene spermatocytes,
HKMG
nt
1
2
3
4
b
5
HKMG
nt
7 t
6
1
1200
2
3
4
5
6
t
7
730
HKE
HKE
730
1200
bottom
top
top
bottom
Figure 3 Polysome pro®le of (a) Gli and (b) Gli3 mRNAs in mouse testes. Polysomes are intact in sucrose gradients containing
HEPES-KCl-Mg2+ (HKMG) bu€er and are dissociated in HEPES-KCl-EDTA (HKE) gradients (containing EDTA in place of
Mg2+). The direction of sedimentation is from right to left as shown. Polysomes are localized in fractions 1 ± 3, and monosomes and
postpolysomal material are localized in fractions 4 ± 7. Following puri®cation, RNA from each fraction was analysed by RNase
protection
a
b
A—
B—
C—
N—
E
T
G
A
B
P
c
1
2
3
4
5
6
P
A—
B—
C—
N—
Figure 4 Gel-shift analysis of Gli DNA binding activity in mouse testis and germ cells. (a) Comparison of optimal binding
sequences for Gli (Gli-B1) and Gli3 as previously determined (Vortkamp et al., 1995). The sequences of the oliognucleotide probes
Gli-B1mut and the rat proenkephalin Gli consensus element (Gli-proenk) are also shown. Note that the bottom strand sequence is
shown in the case of Gli-proenk for comparison. (b) Gel-shift complexes obtained using the Gli-B1 probe and whole-cell extracts
from 14 day mouse embryo (E), adult testis (T), enriched germ cells (G) and type A and B spermatogonia (A,B). Complexes A,B,C
are Gli-speci®c while N represents a complex not dependent on Gli sequences. (c) Competition analysis of Gli complexes. Extracts
from mouse embryo (lanes 1, 3 and 5) and B type spermatogonia (lanes 2, 4 and 6) were incubated with labeled Gli-B1 probe and no
excess oligonucleotide (lanes 1 and 2), excess Gli-B1 competitor (lanes 3 and 4) and excess Gli-B1mut competitor (lanes 5 and 6). P:
probe alone. Fifteen mg of extract was used for each sample in (b) and (c) except for mouse testes (45 mg)
2261
Gli family expression during spermatogenesis
SP Persenglev et al
2262
round and condensing spermatids (Figures 4b and 5a).
As further veri®cation of this developmental pro®le in
male germ cells, Gli DNA binding activities were
compared in extracts from testes at di€erent developmental stages (Figure 5a). The two Gli complexes were
most abundant in testes from early postnatal (8 day
old) mice, in which spermatogonial stages are most
abundant and markedly declined in 17 day old and
adult testes when later spermatogenic stages (spermatocytes and spermatids) become predominant. These
results are consistent with spermatogonia being the
primary stages in which Gli DNA binding proteins are
expressed.
The optimal binding sites determined for human
GLI based on selection of genomic fragments contain a
nonamer core with slightly di€ering adjacent sequences
(Kinzler and Vogelstein, 1990) (see the B1 consensus
element in Figure 4a). The optimal binding sequence
for GLI3 closely resembles the GLI B1 site and
includes the nonamer core sequence (Figure 4a)
(Vortkamp et al., 1995). In contrast, while the Glilike binding site within the rat proenkephalin promoter
contains the consensus nonamer core, it is more
divergent in the adjacent 5'- and 3'-regions (Figure
4a). Since the Gli-like sequence was a potential
regulatory site within the proenkephalin germ line
promoter, it was of interest to determine whether it in
fact bound to the Gli-related proteins present in
spermatogenic cells. The proenkephalin Gli sequence
speci®cally competed for factor binding to the Gli B1
probe using extracts from 8 day mouse testes (Figure
4b). Thus, the altered nature of the sequences
bracketing the proenkephalin Gli nonamer core did
not substantially reduce its ability to interact with Gli
DNA binding proteins present in spermatogenic cells.
a
T8
T17
Ta
A
Pa
G
Discussion
Previous studies have demonstrated selective expression
of Gli family genes during mouse embryogenesis, and
disruption of Gli3 expression is associated with the extratoes mutation in mice and Grieg cephalopolysyndactyly
syndrome in humans (Hui and Joyner, 1993; Schimmang
et al., 1992; Vortkamp et al., 1991, 1992). These ®ndings,
together with the fact that the invertebrate developmental regulators tra-1 and ci are Gli homologues
(Orenic et al., 1990; Zarkower and Hodgkin, 1993),
indicates that the Gli proteins perform primary roles in
cell growth and di€erentiation. The association of
enhanced Gli gene expression with numerous forms of
cancer (Kinzler et al., 1987; Roberts et al., 1989) is also
consistent with a critical role in regulating cell
proliferation. The present ®ndings thus extend the
implicated role for Gli genes in developing cells to
spermatogenesis as well. The selective association of Gli
family gene expression with mitotically active male germ
cells is similar to the situation during somatic cell
development, in which all three Gli genes are expressed
in actively dividing mesodermal and ectodermal cells, but
are not detectable in post-mitotic cells (Hui et al., 1994).
One important di€erence, however, is the absence of
signi®cant Gli2 expression during development of the
male germ line, which appears to be a predominant Gli
family member in day 14 mouse embryos (see Figure 1).
This suggests either a unique function for Gli2 in somatic
cell development and/or redundancy in Gli family
activities during mitosis of spermatogenic cells. It is
noteworthy that in the developing neural tube Gli2 and
Gli3 exhibit overlapping patterns of expression in ventral
regions while Gli expression is mainly dorsal (Hui et al.,
1994).
P
b
C
PE
B1
P
A—
C—
N—
Figure 5 Developmental expression of testicular Gli DNA binding proteins. (a) Binding of whole-cell extracts (15 mg) to the Gli-B1
probe was examined for 8-day mouse testes (T8), 17-day testes (T17) and adult testes (Ta). Extracts from type A spermatogonia (A),
late pachytene spermatocytes (Pa) and enriched germ cells (G) were also examined. P: probe alone. (b) Competition using the rat
proenkephalin competitor; tissue extracts from 8 day mouse testes. C: control; PE: 100-fold excess of Gli-proenk competitor; B1:
100-fold excess Gli-B1 competitor; P: probe with no extract
Gli family expression during spermatogenesis
SP Persenglev et al
Several proto-oncogenes have been implicated in the
regulation of spermatogenic cell growth and differentiation and exhibit distinct patterns of stagedependent expression (Propst, 1988; Torry, 1992;
Alcivar et al., 1991; Wolfes et al., 1989). For
example, c-myc, c-jun and c-fos are predominantly
expressed in spermatogonia with the highest levels
occurring in type B cells, while c-mos, c-abl and pim-1
are mainly expressed in meiotic and/or post-meiotic
stages. In this study, we have demonstrated that the Gli
and Gli3 genes continue to be expressed following
maturation of the adult mouse testes exclusively by
spermatogonial cells. The pattern of Gli family gene
expression is similar to that for c-myc, c-jun and c-fos
(Wolfes et al., 1989) and indicates that Gli proteins
function as part of a transcriptional program regulating mitotic proliferation/di€erentiation of spermatogenic cells. The elevated expression of Gli relative to
Gli3 in type B cells suggests an enhanced role for Gli
during this speci®c stage of development.
Our ®ndings also demonstrate that mouse spermatogenic cells express DNA binding proteins that
speci®cally recognize the Gli consensus sequence. To
our knowledge, this represents the ®rst demonstration
of binding by endogenous nuclear factors to the Gli
consensus element in any cell type. We have also
identi®ed a variant Gli binding site within the rat and
mouse proenkephalin germ line promoter regions that
retains high anity for testicular Gli DNA binding
proteins. The presence of a Gli target sequence within
the proenkephalin germ line promoter raises the
possibility that the latter may be a speci®c target of
the Gli protein family in spermatogenic cells. For
example, since the proenkephalin germ line promoter is
not active in spermatogonial cells, this site could
function to repress promoter activity during mitosis.
However, recent ®ndings from transgenic analysis of
proenkephalin promoter constructs lacking the Gli site
indicate that this element is not critical for promoter
regulation (F Liu and D Kilpatrick, unpublished
observations). This emphasizes the importance of
functional identi®cation of transcription factor target
genes.
Essentially nothing is known about the speci®c role
of Gli proteins and their mechanism of action during
mitosis in mammalian cells. They presumably function
as transcriptional regulators based on their ability to
bind speci®cally to DNA sequences harboring the core
sequence TGGGTGGTC. However, the nature of their
target genes, whether they function as activators or
repressors and their potential interactions with other
transcriptional regulators are unde®ned. Recent findings indicate that the Gli homologue ci controls
anterior cell fate during Drosophila embryogenesis by
virtue of its ability to repress hedgehog expression and
induce hedgehog responsiveness in these cells (Dominguez et al., 1996). Whether parallel mechanisms
operate in mammalian cell development and spermatogenesis speci®cally is uncertain. The present ®ndings
thus form the basis for potential examination of these
various questions by demonstrating Gli expression in a
de®ned lineage for which cells of di€erent developmental stages can be puri®ed in amounts sucient for
biochemical analysis. This may assist in the identification of potential target genes as well as in studying
speci®c functions of the Gli proteins. Extra-toes mutant
mice are not informative as to the role of Gli3 in
spermatogenesis since they die just prior to or after
birth, and data are not available on Gli knockout mice,
to our knowledge. However, expression within the
germ cell lineage provides a potential opportunity to
study in vivo function without the complicating factor
of embryonic lethality. For example, using appropriate
cell-speci®c promoter constructs, it should be possible
to express intact and modi®ed Gli proteins selectively
in di€erent spermatogenic stages of transgenic mice in
order to establish the physiological impact of their
altered expression on germ cell development. One
interesting question is whether forced expression of
Gli2 would have a disruptive e€ect on spermatogenesis.
This approach can also be used to enhance or inhibit
the expression of Gli family members in order to
identify potential downstream target genes.
Materials and methods
Preparation of mouse spermatogenic cells
Enriched spermatogenic cells were prepared from adult
mouse testes by enzymatic digestion as previously
described (Bellve et al., 1977a). These preparations
consisted mainly of pachytene spermatocytes, round
spermatids and cytoplasts. These cell types were further
puri®ed to greater than 90% purity by unit gravity
sedimentation through 2 ± 4% linear gradients of bovine
serum albumin in Krebs-Ringer Bicarbonate Glucose
bu€er (Bellve et al., 1977a). Spermatogenic cells were also
puri®ed from immature mouse testes for RNA and protein
analyses as previously described (Bellve et al., 1977b).
Purity of collected fractions as monitored by Nomarski
optics were as follows: type A spermatogonia, 92 ± 94%,
with type B spermatogonia being the main contaminant;
type B spermatogonia, 95 ± 96%, with red blood cells the
major contaminant; various spermatocyte stages, 92 ± 97%,
with the only signi®cant contaminants being either red
blood cells or other spermatocyte stages. Testicular somatic
cells represnted considerably less than 1% contamination
in all cases.
RNA isolation and Northern analysis
Total RNA was isolated from mouse testis, embryo, liver
and spermatogenic cells by the guanidinium isothiocyanate/CsCl method (Chirgwin et al., 1979) and Northern
blot analyses were performed as previously described (Kew
and Kilpatrick, 1990). The Gli, Gli2 and Gli3 probes used
for Northern analysis were generated by digestion of the
original vectors containing the mouse cDNAs (Hui et al.,
1994) with EcoRI.
Polysome analysis
The procedure for polysome pro®le evaluation of Gli3
mRNA was performed as in previous studies (Kew et al.,
1989). Cytoplasmic fractions were prepared from puri®ed
mouse testes (1.5 g/gradient) and were separated on 38 ml
gradients of 10 ± 40% sucrose. Gradient fractions were
precipitated with isopropanol in the presence of 50 mg
yeast RNA as a carrier, and RNA was extracted and
puri®ed further by the method of Chomczynski and Sacchi
(1987) and reprecipitated. Equivalent percentages of each
fraction were analysed by RNase protection as previously
described (Kilpatrick et al., 1990). The plasmid pGli-3a was
linearized by HindIII digestion for use as a template for
mouse Gli3 antisense RNA (Hui et al., 1994).
2263
Gli family expression during spermatogenesis
SP Persenglev et al
2264
Protein extraction and gel-shift assays
Nuclear and whole-cell extracts from mouse embryo, testes
and spermatogenic cells were prepared as described
elsewhere (Cereghini et al., 1987). All procedures were
performed at 48C. Brie¯y, nuclei were initially suspended in
bu€er consisting of 20 mM HEPES (pH 7.9), 25% glycerol,
1.5 mM MgCl2, 0.25 mM EDTA and 0.14 M NaCl and
protease inhibitors. An additional amount of the same
bu€er containing 0.7 M NaCl was then added to bring the
®nal concentration to 0.37 M. After gentle mixing for
30 min, extracted proteins were separated by centrifugation
and dialyzed twice against bu€er containing 20 mM
HEPES (pH 7.9), 60 mM NaCl, 20% glycerol, 0.25 mM
EDTA, 0.125 mM EGTA, and protease inhibitors. Protein
concentrations ranged from 5 ± 10 mg/ml.
For the preparation of whole-cell extracts, tissues were
homogenized in lysis bu€er (20 mM HEPES (pH 7.9), 0.4 M
NaCl, 1 mM EDTA, 10 mM DTT, 1 mM PMSF, 25%
glycerol). Cells were kept on ice for 20 min and lysed in
the same bu€er by two freeze/thaw cycles. After centrifugation for 10 min at 15 000 r.p.m. and 48C, supernatants were
stored at 7808C.
Binding reactions for gel-shift assays were performed in a
®nal volume of 20 ml as previously described (GalchevaGargova et al., 1993). Competition with unlabeled oligonucleotides was generally performed with a 100-fold molar
excess. DNA-protein complexes were resolved on nondenaturing, 4% polyacrylamide gels. Double-stranded oligonucleotides were labeled with [32P]dCTP by ®ll-in reactions
using Klenow. Whole-tissue or -cell extracts were generally
used in these studies and initial experiments con®rmed that
identical protein complexes were also detected using nuclear
extracts.
Acknowledgements
We thank Dr Alexandra Joyner for providing plasmids
containing mouse Gli, Gli2 and Gli3 cDNAs. We also
thank Ms Cathy Warren for the preparation of this
manuscript. This work was supported by PHS grants
DK36468 to DLK and HD15269 to CFM.
References
Alcivar AA, Hake LE, Kwon YK and Hecht NB. (1991).
Mol. Reprod. Dev., 30, 187 ± 193.
Bellve AR, Millette CF, Bhatnager YM and O'Brien DA.
(1977a). J. Histochem. Cytochem., 25, 480 ± 494.
Bellve AR, Cavicchia JC, Millette CF, O'Brien DA,
Bhatnagar Y and Dym M. (1977b). J. Cell Biol., 74, 68 ±
85.
Cereghini S, Raymondjean M, Carranca AG, Herbomel P
and Yaniv M. (1987). Cell, 50, 627 ± 638.
Chirgwin JM, Przbyla AE, MacDonald RJ and Rutter WJ.
(1979). Biochemistry, 18, 5294 ± 5299.
Chomczynski P and Sacchi N. (1987). Anal. Biochem., 162,
156 ± 159.
Dominguez M, Brunner M, Hafen E and Basler K. (1996).
Science, 272, 1621 ± 1625.
Galcheva-Gargova Z, Tokeson JP, Karagyosov LK, Ebert
KM and Kilpatrick DL. (1993). Mol. Endocrinol., 7, 979 ±
991.
Hui C-C and Joyner AL. (1993). Nat. Genet., 3, 241 ± 246.
Hui C-C, Slusarski D, Platt KA, Holmgren R and Joyner
AL. (1994). Dev. Biol., 162, 402 ± 413.
Kew D, Jin DF, Kim F, Laddis T and Kilpatrick DL. (1989).
Mol. Endocrinol., 3, 1191 ± 1196.
Kew D and Kilpatrick DL. (1990). Mol. Endocrinol., 4, 337 ±
340.
Kilpatrick DL, Zin SA, FitzgeraldnM, Higuchi H, Sabol SL
and Meyerhardt J. (1990). Mol. Cell Biol., 10, 3717 ± 3726.
Kinzler KW, Bigner SH, Bigner DD, Trent JM, Law ML,
O'Brien SJ, Wong AJ and Vogelstein BV. (1987). Science,
23, 70 ± 72.
Kinzler KW, Ruppert JM, Bigner SH and Vogelstein BV.
(1988). Nature, 332, 371 ± 374.
Kinzler KW and Vogelstein B. (1990). Mol. Cell. Biol., 10,
634 ± 642.
Kleene KC. (1996). Mol. Reprod. Dev., 43, 268 ± 281.
Orenic T, Slusarski DC, Kroll KL and Holmgren RA.
(1990). Genes Dev., 4, 1053 ± 1067.
Pavletich NP and Pabo CO. (1993). Science, 261, 1701 ±
1707.
Propst F, Rosenberg MP and Vande Woude GF. (1988).
Trends in Genet., 4, 183 ± 187.
Roberts WM, Douglass EC, Peiper SC, Houghton PJ and
Look AT. (1989). Cancer Res., 49, 5407 ± 5413.
Ruppert JM, Kinzler KW, Wong AJ, Bigner SH, Kao F-T,
Law ML, Seuanez HN, O'Brien SJ and Vogelstein B.
(1988). Mol. Cell. Biol., 8, 3104 ± 3113.
Ruppert JM, Vogelstein B, Arheden K and Kinzler KW.
(1990). Mol. Cell. Biol., 10, 5408 ± 5415.
Ruppert JM, Vogelstein B and Kinzler KW. (1991). Mol.
Cell. Biol., 11, 1724 ± 1728.
Schimmang T, Lemaistre M, Vortkamp A and RuÈther U.
(1992). Development, 116, 799 ± 804.
Torry DS. (1992). Am. J. Reprod. Immunol., 27, 167 ± 170.
Vortkamp A, Gessler M and Grzeschik K-H. (1991). Nature,
352, 539 ± 540.
Vortkamp A, Franz T, Gessler M and Grzeschil K-H. (1992).
Mamm. Genome, 3, 461 ± 463.
Vortkamp A, Gessler M and Grzeschik K-H. (1995). DNA
and Cell Biology, 14, 629 ± 634.
Walterhouse D, Ahmed M, Slusarski D, Kalamaras J,
Boucher D, Holmgren R and Iannaccone P. (1993). Dev.
Dyn., 196, 91 ± 102.
Wolfes H, Kogawa K, Millette CF and Cooper GM. (1989).
Science, 245, 740 ± 743.
Zarkower D and Hodgkin J. (1993). Nucleic Acids Res., 21,
3691 ± 3698.