Download Pleiotropic effects of the mouse lethal yellow (Ay) mutation

1695
Development 120, 1695-1708 (1994)
Printed in Great Britain © The Company of Biologists Limited 1994
Pleiotropic effects of the mouse lethal yellow (Ay) mutation explained by
deletion of a maternally expressed gene and the simultaneous production of
agouti fusion RNAs
David M. J. Duhl1, Mary E. Stevens1, Harry Vrieling1*, Paul J. Saxon2†, Miles W. Miller1, Charles J. Epstein2
and Gregory S. Barsh1‡
1Department
of Pediatrics and Howard Hughes Medical Institute, Stanford University School of Medicine, Stanford, California
94305-5428, USA
2Department of Pediatrics, University of California, San Francisco, California, 94143, USA
*Present address: MGC Department of Radiation Genetics and Chemical Mutagenesis, Wassenaarseweg 72, 2333 AL Leiden, The Netherlands
†Present address: Department of Laboratory Medicine, Sutter Memorial Hospital, Sacramento, California 95819, USA
‡Author for correspondence
SUMMARY
Heterozygosity for the mouse lethal yellow (Ay) mutation
leads to obesity, increased tumor susceptibility and
increased activity of the agouti coat color gene; homozygosity for Ay results in embryonic death around the time of
implantation. Although these pleiotropic effects have not
been separated by recombination, previous studies have
suggested that the dominant and recessive effects result
from distinct genetic lesions. Here we use a combination of
genomic and cDNA cloning experiments to demonstrate
that the Ay mutation is caused by a 120 kb deletion which
lies centromere-proximal to the agouti coat color gene. The
deletion removes coding but not 5′ untranslated sequences
for a ubiquitously expressed gene predicted to encode a
protein similar in sequence to an RNA-binding protein,
which we named Merc, for maternally expressed hnRNP Crelated gene, but have renamed Raly, since the gene is
nearly identical to one reported recently by Michaud et al.
(Gene Dev. 7, 1203-1213, 1993). The Ay deletion results in
the splicing of Merc/Raly 5′ untranslated sequences to
agouti protein-coding sequences, which suggests that
ectopic expression of the normal agouti protein by the Ay
fusion RNA is responsible for the pleiotropic effects associated with heterozygosity for Ay.
We find that Merc/Raly RNA is present in the unfertilized egg and is also transcribed in preimplantation
embryos. Using a PCR-based assay to determine the
genotype of individual embryos from an Ay/a × Ay/a intercross, we show that, in the absence of zygotic Merc/Raly
expression, Ay/Ay embryos develop to the blastocyst stage,
but do not hatch from the zona pellucida or form trophoblastic outgrowths. Injection of a Merc/Raly antisense
oligonucleotide into non-mutant embryos blocks development prior to the blastocyst stage, and can be rescued by
coinjection of a Merc/Raly transgene. These results suggest
that maternal expression of Merc/Raly plays an important
role in preimplantation development and that its deletion
of is sufficient to explain Ay-associated embryonic lethality.
INTRODUCTION
Pedersen, 1974), an inability of embryos to form outgrowths
in culture (Papaioannou, 1988) and failure to hatch from the
zona pellucida when placed into diapause (Papaioannou and
Gardner, 1992). Although the variety of defects observed may
be due to a combination of factors including different genetic
backgrounds, variable expressivity of the mutation, and the
lack of an independent marker to determine the genotype of
individual embryos, these studies suggest that the effects of Ay
are most evident around the time of implantation. Because
abnormalities are observed in both the TE and the inner cell
mass (Papaioannou, 1988), and because Ay/Ay cells cannot be
rescued in aggregation chimeras (Barsh et al., 1990), Ay may
affect a critical developmental step required prior to implantation.
In 1905, mice carrying the lethal yellow mutation (Ay) were
used to provide the first evidence that heritable traits could be
used to study vertebrate development (Cuénot, 1905). Inviability of Ay/Ay animals, revealed initially by a deviation from
Mendelian proportions in the offspring of Ay/− × Ay/− intercrosses, has been the subject of many embryologic studies over
the last several decades [reviewed in Magnuson (1986); Silvers
(1979)]. Morphologic and ultrastructural abnormalities attributed to the effects of the Ay mutation include degeneration of
trophectoderm in the peri-implantation blastocyst (Eaton and
Green, 1963), an increased frequency of excluded blastomeres
in cleavage-stage embryos (Calarco and Pedersen, 1976;
Key words: agouti, lethal yellow, hnRNP, mouse development
1696 D. M. J. Duhl and others
Ay is also one of the oldest recognized mutations of the
agouti coat color locus, in which alleles associated with the
synthesis of yellow pigment in hair follicles are dominant to
those associated with the synthesis of black pigment
[(reviewed in Silvers (1979)]. Ay in combination with all other
agouti locus alleles leads to the exclusive production of yellow
pigment and, in addition, has pleiotropic effects, including
adult onset obesity and increased somatic growth (Carpenter
and Mayer, 1958; Castle, 1941; Danforth, 1927), increased susceptibility to tumor formation (Heston and Deringer, 1947;
Heston and Vlahakis, 1961; Vlahakis and Heston, 1963) and
premature infertility (Granholm and Dickens, 1986; Granholm
et al., 1986).
The agouti gene has been cloned recently and is thought to
encode a signaling molecule that directs follicular melanocytes
to switch from the synthesis of black pigment, eumelanin, to
yellow pigment, phaeomelanin (Bultman et al., 1992; Miller et
al., 1993). For most agouti genotypes, agouti RNA is
expressed only in the skin during the time of phaeomelanin
synthesis. However, Ay is associated with the ubiquitous
expression of a chimeric RNA that fuses a novel 5′ end, ‘exon
1Ay’, to agouti-coding sequences at its 3′ end (Miller et al.,
1993). This suggests that some or all of the pleiotropic effects
of Ay are due to ectopic expression of the normal agouti protein.
Three separate arguments suggest that lethality of Ay/Ay
animals may not result from altered expression of the agouti
coat color gene. First, the effects of Ay on coat color are a gainof-function, yet embryonic lethality is most likely a loss-offunction (Barsh and Epstein, 1989; Barsh et al., 1990). Second,
the viable yellow (Avy) mutation has similar effects to Ay with
regard to obesity and tumor susceptibility, but is not lethal
when homozygous (Dickie, 1962; Wolff et al., 1986). Finally,
genetic complementation analysis suggests that there are
multiple genes close to or within the agouti locus required for
embryonic development [(Barsh and Epstein, 1989; Lyon et
al., 1985; Russell et al., 1963); reviewed in Siracusa (1991)].
To understand the molecular basis for Ay-associated lethality
better, we have performed a combination of molecular and
embryologic studies. We find that the Ay chromosome contains
a 120 kilobase (kb) deletion that lies centromere-proximal to
all previously identified agouti exons. The deletion removes
most of the coding sequences for a gene that is likely to encode
an RNA-binding protein that we have named Merc, and results
in a novel mutational mechanism in which the non-deleted first
exon of the Merc gene is spliced to protein-coding sequences
of agouti. Based on the developmental defects that we observe
both in Ay/Ay embryos and in normal embryos in which
expression of the Merc gene has been inhibited experimentally,
we conclude that loss of Merc function is responsible for Ayassociated embryonic lethality.
MATERIALS AND METHODS
Mouse strains and mutations
Mice carrying the Ay, AW and A alleles are propagated in our laboratory as C57BL/6J-Ay/a, 129/SvJ-AW/AW and FVB/N-A/A, respectively. The C57BL/6J and 129/SvJ strains were obtained from The
Jackson Laboratory and the FVB/N strain was obtained from Taconic
(Germantown, NY). Mice carrying the ax mutation were generously
provided by Dr Virginia Papaioannou (Tufts University School of
Medicine, Boston, MA) as a balanced lethal stock Ay/ax; the ax allele
is propagated in our laboratory by backcrossing to C57BL/6J-a/a
animals.
Genomic cloning, cDNA cloning and DNA probes
We have previously described the isolation and structure of cosmid
clones that constitute a 60 kb contig containing agouti exons 1B, 1C,
2, 3 and 4 (Miller et al., 1993) (The agouti exon nomenclature used
in the present work differs from that used previously; the rationale is
summarized in Fig. 1). As it became apparent that alternative isoforms
of the agouti cDNAs had 5′ ends that were not contained within these
cosmid clones, the contig was extended further 5′ with regard to the
direction of agouti transcription by screening a commercial bacteriophage P1 library (Genome Systems, St Louis, MO) with primer pairs
derived from agouti cDNAs or genomic clones. The commercial P1
library was constructed from a mouse fibroblast cell line, C127,
derived from the RIII mouse strain. The RIII strain carries the A allele,
which is consistent with our observation that, in regions of overlap,
the P1 clones that we obtained are identical to genomic clones derived
from DNA of mice that carry the A allele.
The P1 clones P1-24 and P1-25 were isolated using oligonucleotide
primers from a subclone of cosD3 (see Fig. 1), 5′-CCTAGGTTTCTCTGTGTCCCC-3′
and
5′-CAGAAGTCCCTGGTAGCTGC-3′, that amplify a 225 base pair fragment containing
agouti exon 1B. The P1 clones P1-74 and P1-75 were isolated using
oligonucleotide primers from exon 1A, 5′-AGTCTGAGTCCTTGAGCCTC-3′ and 5′-TGGGACCCCCGGTGGTTC-3′, that
amplify a 78 base pair fragment. A P1 clone that contains the 5′ end
of the Ay-specific cDNA, P1-62, was isolated using the oligonucleotide primers 5′-CCGAGGGGGCGGAAGCGG-3′ and 5′CCGCGCCGAGGTCTGGAG-3′, that lie at the ends of exon 1Ay
(see Fig. 1 and Fig. 4) and amplify a 147 base pair fragment. The
BamHI restriction maps of the P1 clones were determined by partial
digestion and indirect end labeling. Comparison of restriction maps
allowed a 160 kb contig to be generated that included all the cosmids
and P1-24, P1-25, P1-74, and P1-75 (see Fig. 1). The P1-62 clone is
approximately 100 kb in length but does not overlap with any of the
cosmid or bacteriophage P1 clones.
A 1.0 kb BamHI-EcoRI genomic fragment that contains exon 1Ay
was subcloned from P1-62 and is described below as the ‘exon 1Ay
probe’. A 1.3 kb XbaI-BamHI genomic fragment that contains exon
1A was subcloned from P1-75 and is described below as the ‘exon
1A probe’. A 1.3 kb XbaI-BamHI fragment subcloned from one of
the Merc cDNA clones contains the entire open reading frame and
corresponds to residues 403-1720 in Fig. 4A; it is described below as
the ‘Merc cDNA probe’. In some experiments, an internal 340 bp
SmaI-ClaI fragment was used instead of the 1.3 kb XbaI-BamHI
fragment. A 1.0 kb XbaI-EcoRI genomic fragment that contains
agouti exon 4 is described below as the ‘exon 4 probe’ [The same
probe was described in Miller et al., 1993 as ‘probe d’].
To isolate the Merc cDNA, a commercial cDNA library prepared
from the teratocarcinoma cell line PCC4 (Stratagene, San Diego, CA)
was screened initially with the exon 1Ay probe. Two clones were
recovered, 1.7 kb and 0.8 kb in length; each clone contained exon1Ay
sequences followed by an open reading frame at the 3′ end, but neither
clone contained a poly(A) sequence or an obvious polyadenylation
signal. A 340 basepair (bp) fragment from the 3′ end of the 1.7 kb
clone was then used to rescreen the library, and a third clone was
isolated, 1.5 kb in length, that contained the entire Merc open reading
frame followed by a consensus polyadenylation sequence and a
poly(A) tail.
RNA expression studies
Expression of the Merc RNA in adult tissues was determined by
northern hybridization using the 147 bp exon 1Ay fragment as a
hybridization probe. For PCR-based studies of embryos, total RNA
extracted from approximately 50 mouse eggs, 2-cell embryos, or from
Molecular basis of the mouse Ay mutation 1697
25 blastocysts as previously described (Andria et al., 1992) was
reverse transcribed using an oligonucleotide primer complementary
to sequences in the 3′ end of the Merc open reading frame (see Fig.
4A), 5′-ACCATCATCTCGAATTTG-3′, and then PCR-amplified
using the oligonucleotide primers 5′-CAGGGCCGCCTCTTC-3′ and
5′-TGTGCACAGAGCAACCAG-3′. The same group of primers
were used to analyze RNA from neonatal skin of AW/AW mice. These
primers span at least one Merc intron that separates 5′ untranslated
from protein-coding sequences, and produce fragments of 342 bp or
259 bp depending on the presence or absence, respectively, of an alternatively spliced 83 bp exon (see Figs 4, 5). The identity of the
products was confirmed both by direct sequencing and by Southern
hybridization with the Merc cDNA probe.
For embryo expression studies of the Merc/agouti fusion RNA,
total RNA was extracted from approximately 50 a/a eggs, or from 25
blastocysts derived from Ay/a males mated to a/a females, in which
fusion RNA must arise from the paternal chromosome. The RNA was
reverse transcribed using an oligonucleotide primer, 5′-GAGGAATTCACCTTGCCACCTTCTTCATCGA-3′, complementary to
sequences within the agouti protein-coding region, and then PCRamplified using the oligonucleotide primers 5′-CAGGGCCGCCTCTTC-3′ and 5′-CGAGTTCATGGAGGAGTTACTCCGC-3′.
These primers are located in Merc 5′ untranslated and agouti proteincoding sequence, respectively, and amplify a 386 bp or 229 bp
fragment depending on the presence or absence of agouti exons 1A
and 1A′ (see Fig. 5). The identity of the products has been confirmed
by hybridization with an internal oligonucleotide probe from agouti
exon 2, 5′-CAGGAATTCACCATGGATGTCACCCGCCTACTC-3′.
Embryologic studies
To determine the developmental potential of Ay/Ay embryos,
C57BL/6J-Ay/a animals were intercrossed in natural matings, blastocysts were recovered by uterine flushing at 3.5 days post coitum (dpc),
and placed into in vitro microdrop culture with DMEM supplemented
with 10% fetal calf serum. After 3-4 days, the development of each
embryo was assessed with regard to hatching from the zona pellucida,
attachment and trophoblastic outgrowth. After phenotypic scoring, the
agouti locus genotype of each embryo was determined using a PCRbased assay as previously described (Frohman et al., 1993). The assay
determines inheritance of a variant HindIII site, present on the Ay
allele and absent from the a allele, located ≤0.2 cM distal to Ay
(Siracusa et al., 1987).
To determine whether inhibition of Merc gene expression affected
development of non-mutant embryos, a sense or antisense
oligodeoxynucleotide that spanned the predicted Merc translational
initiation site was injected into fertilized 1-cell embryos obtained from
superovulated FVB/N or C57BL/6J mice. The oligodeoxynucleotides,
5′-TTCAAGGACATGGTGTTCAC-3′ (antisense) and 5′-GTGAACACCATGTCCTTGAA-3′ (sense), were synthesized with phosphorothioate linkages at the first and last 5 residues to decrease susceptibility to intracellular degradation. An additional control ‘scrambled’
oligodeoxynucleotide with a base composition similar to the antisense
oligonucleotide, 5′-TTCTCAGGATGGATGTCACC-3′, was used in
some experiments. Based on the experience of Strickland and colleagues with injection of antisense oligodeoxynucleotides into
oocytes (Salles et al., 1993), the oligonucleotides were dissolved at 1
µg/µl in 10 mM Tris, pH 7.5; 0.1 mM EDTA, and 3-5 picoliters of
the undiluted solution or two serial 50-fold dilutions were microinjected under direct visualization into each embryo, corresponding
approximately to 5×108, 1×107, and 2×105 molecules of oligonucleotide injected per embryo. In experiments designed to examine
whether expression of Merc could rescue developmental arrest
induced by the antisense oligodeoxynucleotide, a Merc or control
transgene was coinjected with oligonucleotide at a concentration of 1
µg/ml, which corresponds approximately to 10,000 molecules of 2 kb
fragment injected per embryo. The Merc transgene was constructed
by fusing the PGK promoter from the vector pNT (Tybulewicz et al.,
1991) to a 1.3 kb XbaI-EcoRI fragment that contained the coding
sequences of the Merc gene (residues 403-1726 of the Merc cDNA;
see Fig. 4); the transgene was then excised as a 1.9 kb EcoRI
fragment. A 2 kb control transgene was also used that contained
agouti-coding sequences. In a first series of experiments, the embryos
were injected with undiluted sense or antisense oligonucleotide, transferred into pseudopregnant foster mothers for 3 days, and then placed
into microdrop culture as described above. In subsequent experiments,
development of cleavage stage embryos was assessed by transferring
injected embryos into microdrop culture and scoring their phenotype
with regard to cell number and/or cell lysis on subsequent days.
Additional techniques
High molecular weight DNA for pulsed field gel electrophoresis was
prepared from homogenized spleen tissue and digested with ‘rarecutting’ restriction enzymes as previously described. Contour
clamped homogenous electric field (CHEF) electrophoresis was
performed using a commercial apparatus (CBS Scientific, San Diego,
CA) at 150 V, in 0.5× TBE (45 mM Tris, pH 8, 45 mM boric acic, 1
mM EDTA) and 0.1 µg/ml ethidium bromide, at 6°C. Direction of the
electric field was changed according to a ramped program from 40 to
140 sec, using 1 second intervals, and total electrophoresis time was
≈48 hours. Molecular weight estimations were based on a yeast
subline derived from strain AB1380, which carries 245 kb, 290 kb,
370 kb, 460 kb, 580 kb, 630 kb, 700 kb, 770 kb, 800 kb, 850 kb and
945 kb chromosomes, and a mouse yeast artificial chromosome
(YAC) of 480 kb, which served as a convenient registration marker.
In some experiments, ligated concatemers of bacteriophage λ also
served as size markers. For the results shown in Table 1, the sizes of
individual fragments were first estimated by comparison to yeast
markers in adjacent lanes, which provided a ‘window’ of approximately ±40 kb. Exact sizes within this window were then estimated
by construction of an internally consistent restriction map for the a
and Ay chromosomes. These maps differ by virtue of (1) a 120 kb
deletion present only in the Ay allele which is described below; and
(2) an approximately 20 kb insertion present only in the a allele
between agouti exons 1C and 2 [(Bultman et al., 1992); unpublished
observations].
Northern hybridizations, Southern hybridizations and DNA
sequence analysis using modified T7 polymerase and dideoxy chain
termination were performed according to standard techniques
(Sambrook et al., 1989) using radiolabeled nucleotides and hybridization in the presence of 10% dextran sulfate. Reverse transcription
and PCR protocols are described in Miller et al. (1993).
RESULTS
Identification of an Ay-specific first exon of agouti
In the course of cloning cDNAs from Ay/ax mice, a chimeric
RNA was identified in which a 147 bp Ay-specific sequence
(exon 1Ay) was spliced to other exons that contained agouticoding sequences (Miller et al., 1993). The analysis of these
chimeric RNAs was complicated by the normally heterogeneous nature of the 5′ end of agouti mRNA, in which we have
found three different first exons (1A, 1B, 1C) and one alternatively spliced exon (1A′) in cDNAs from A/A and Aw/Aw mice
(Fig. 1). Of 17 cDNA clones isolated from Ay/ax skin RNA,
exon 1Ay was fused to agouti exon 1A in six clones, and exon
1Ay was fused to agouti exon 2 in four clones. The remaining
seven cDNA clones did not contain exon 1Ay, but instead
began with agouti exons 1B or 1C. Fusion of exon 1Ay to
agouti exon 2 uses the same splice acceptor used by agouti
exons 1A, 1A′, 1B, and 1C. However, fusion of exon 1Ay to
agouti exon 1A appears to use a cryptic splice acceptor, 5′-
1698 D. M. J. Duhl and others
ccaacatgcag|GAA-3′, since this entire sequence is present
normally in exon 1A (see below, Fig. 4).
of agouti genotypes Ay/a, Ay/ax, a/a, or a/ax were digested with
BssHII, EagI, or SmaI, fractionated on an agarose gel in the 50750 kb size range, transferred to a nylon filter, and hybridized separately with genomic probes that contained exon 1Ay or agouti
exons (Fig. 2). The results obtained with a probe that contained
agouti exon 1A were identical to those obtained with a probe that
contained agouti exon 4 (data not shown). For some enzyme/probe
combinations, two fragments were observed with DNA from at
The Ay chromosome contains a 120 kb deletion
between exon 1Ay and agouti exon 1A
To investigate if a chromosomal or subvisible rearrangement
caused the Ay mutation, we first determined whether the normal
chromosomal location of exon 1Ay lay close to the agouti gene
on the distal portion of mouse
chromosome 2. Oligonucleotide primers at the ends of
exon 1Ay were used to
amplify DNA from laboratory
mice, a hamster cell line (GM
459), and a mouse/hamster
hybrid cell line that carries
most of mouse chromosome 2
(ADCT-25). DNA was also
amplified from 4 irradiationreduced somatic cell hybrids
which have been shown previously to retain mouse DNA
markers that lie within
approximately 5 cM of the
agouti gene (Ollmann et al.,
1992). A PCR product
specific for exon 1Ay was
observed with the laboratory
mouse DNA, ADCT-25 DNA,
and one radiation hybrid
DNA, RH33, but not with GM
459 parent hamster DNA (Fig.
1C). These results indicate
that exon 1Ay is derived from
a region of mouse chromosome 2 that lies close to the
normal agouti gene.
Hybridization of exon 1Ay
to cloned genomic DNA
indicated that exon 1Ay was
not located within a 160 kb
cosmid and bacteriophage P1
Fig. 1. Genomic and cDNA structure of alternative agouti isoforms and Ay-specific cDNAs. (A) Physical
contig (Fig. 1A) that contains
location
and splicing pattern for agouti isoforms from the wild-type AW allele. Generation of the 160 kb
the entire 125 kb agouti gene
contig
comprising
four bacteriophage P1 clones and three cosmid clones is described in Materials and
(data not shown). In addition,
Methods.
Agouti
isoforms
that begin with exon 1A are expressed only in the ventrum of AW/AW animals
a 100 kb mouse genomic P1
but throughout the entire hair growth cycle. In contrast, isoforms that begin with exon 1B or exon 1C are
clone that contains exon 1Ay
expressed in the dorsum and the ventrum but only during the mid-portion of the hair growth cycle. The
did not overlap with the
total length of the agouti gene from the beginning of exon 1A to the end of exon 4 is 125 kb. BamHI and
agouti contig, suggesting that
EcoRI restriction sites are indicated with hash marks above and below the thick horizontal line,
exon 1Ay was located at least
respectively. Experiments that support this rationale for agouti exon nomenclature including an analysis
150 kb away from agouti exon
of cDNAs recovered from AW/AW skin and expression of individual exons in animals that carry different
agouti alleles are described in Vrieling et al. (1994). The sequence described as ‘exon 1’ in Miller et al.
1A. To determine more
(1993) is described as ‘exon 1C’ in the current work. (B) Exon structure of cDNA clones isolated from
precisely the physical relaAy/ax skin. Ten of the seventeen clones began with a 147 bp sequence, ‘exon 1Ay’, specific for the Ay
tionship between exon 1Ay
mutation.
(C) Chromosome mapping of exon 1Ay sequences from DNA of somatic cell hybrid lines.
and the agouti gene, we used
Genomic
DNA
from mouse, hamster, a somatic cell hybrid line that contains most of mouse chromosome
genomic probes and pulsed
2 (ADCT-25), or different radiation-reduced hybrid lines (RH1, RH33, RH49, and RH150) was PCRfield gel electrophoresis to
amplified with exon 1Ay primers as described in Materials and Methods. An aliquot of the amplified
construct a long-range restricmaterial was electrophoresed on a 1% agarose gel, which was then stained with ethidium bromide. The
tion map of both the Ay and
147 bp exon 1Ay fragment is present in mouse, ADCT-25, and two of the four radiation hybrid lines,
non-mutant chromosomes.
which demonstrates that exon 1Ay lies on the distal portion of mouse chromosome 2. Characterization of
DNA samples from animals
the hybrid cell lines is described in Ollman et al. (1992).
Molecular basis of the mouse Ay mutation 1699
Fig. 2. Long-range restriction map of the Ay chromosome based on CHEF
gel electrophoresis. (A) Restriction map based on the results shown in Fig.
2B and summarized in Table 1. The 120 kb Ay deletion removes an EagI site,
but placement of the deletion breakpoints with regard to this site is arbitrary.
The locations of exon 1Ay, agouti exon 1A, and agouti exon 4 are indicated
with arrows above the corresponding genomic probes, which are described in
Materials and Methods. Results obtained for the exon 1A probe and the exon 4 probe were identical. The exon 1Ay probe and the exon 1A probe
hybridize to the same Ay-specific fragments whose sizes are listed in Table 1. An EagI site present by sequence analysis in the exon 1Ay probe is
not susceptible to digestion in genomic DNA. (B) DNA from Ay/a, Ay/ax, ax/a, or a/a mice was digested with the indicated restriction enzyme,
fractionated by CHEF gel electrophoresis as described in Materials and Methods, and a Southern blot was hybridized with the indicated probes.
The autoradiogram shown for the Merc cDNA probe was obtained with an internal 340 bp fragment as described in Materials and Methods.
Removal of residual probe between hybridizations was checked by autoradiography. The approximately 20 kb insertion present in the a allele
between agouti exons 1C and exon 2 does produce a distinguishable difference in mobility compared to fragments from the other alleles.
least one heterozygote but not from a/a homozygotes, and comparison of the results allowed us to identify the allelic origin of
each fragment (Table 1, Fig. 2). The exon 1Ay and agouti exon
1A probes detect the same a-specific 320 kb and Ay-specific 180
kb BssHII fragments. Similarly, these probes detect the same aspecific 345 kb and Ay-specific 205 kb SmaI fragments. These
results indicate that exon 1Ay and agouti exon 1A are located a
maximum of 320 kb apart in the a chromosome and a maximum
Table 1. High-molecular-weight restriction fragments
detected by probes that span the Ay deletion*
BssHII
probe
exon 1Ay
Merc cDNA
exon 1A;
exon 4
Fig. 3. Expression of exon 1Ay sequences in tissues of A/A and Ay/ax
mice. Total RNA (30 µg) from the indicated tissues was fractionated
by formaldehyde-agarose gel electrophoresis, transferred to a nylon
filter, and hybridized with a probe prepared from the 147 bp exon 1Ay
sequences. The filter was also hybridized with a Gapd ‘housekeeping’
probe to control for the amount and integrity of the RNA.
EagI
SmaI
a
Ay
a
Ay
a
Ay
320
320
320
180
−
180
180†
180
220
180†
−‡
180
345
345
345
205
−
205
*Estimated sizes are in kilobases based upon the results shown in Figure 2,
and are based on the mobility of yeast chromosomes in adjacent lanes. Probes
are described fully in the Materials and Methods and in Figure 2. Comparison
of the fragments observed in the a/a and Ay/a DNA samples allowed us to
determine their chromosomal origin as “a” or “Ay” with one exception as
noted below.
†The exon 1Ay probe detects a single EagI fragment of 180 kb in Ay/a
DNA. The results obtained with BssHII and SmaI indicate that the probe is
present in both alleles.
‡Absence of the 180 kb EagI fragment from the Ay allele is based on the
results obtained with BssHII and SmaI.
1700 D. M. J. Duhl and others
of 180 kb apart in the Ay chromosome. Furthermore, the difference
between the a-specific and Ay-specific fragments is the same for
BssHII and SmaI, suggesting that a deletion of 120 kb has occurred
in the Ay chromosome. (The deletion is 20 kb less than the difference between the two fragments, 140 kb, due to a 20 kb insertion
in the a chromosome that is not present in the Ay chromosome).
Similar results were obtained for the enzyme combinations
MluI+SalI and NotI+SalI (data not shown).
Using a combination of Southern hybridization and DNA
Fig. 4. Sequence and splicing pattern of
the Merc gene. (A) Composite sequence
obtained from three overlapping cDNAs
isolated from a teratocarcinoma cell line.
The 147 bp exon 1Ay sequence and a
consensus polyadenylation signal have a
single underline, and the alternatively
spliced 83 bp exon has a double
underline. The predicted protein sequence
as translated from the first methionine
codon is shown from residues 483-1370.
(B) Diagram of alternative splicing and
exon usage in the Ay and non-mutant
chromosome as determined by PCRbased studies. The Ay-specific pattern of
splicing was determined by comparison
of the composite Merc cDNA sequence to
the sequences of cDNA clones shown in
Fig. 1. Ay-specific cDNAs that contain
agouti exon 1A do not contain its entire 5′
end, suggesting that inclusion of exon 1A
in Merc/agouti fusion RNAs is due to a
cryptic splice acceptor. The non-mutant
pattern of splicing was determined by
using oligonucleotide primers in 5′
untranslated and potential Merc-coding
regions to PCR-amplify cDNA from
neonatal AW/AW skin RNA as described in
Materials and Methods. Direct
sequencing of the gel-purified products
(see Fig. 5 for example) revealed two
patterns of splicing that differ in their
inclusion of an 83 bp region. The
genomic sequence following Merc exon 1
is identical to the first 4 nucleotides of
this 83 bp region, referred to as Merc
exon 1′ in the figure, and contains two
potential splice donor sites (see text).
(C) Alignment of potential Merc proteincoding sequences with those of the
human hnRNP C protein. The RNAbinding domain in hnRNP C consists of a
four-stranded beta-pleated sheet with two
alpha helices whose positions are
indicated.
sequence analysis, positions have been determined for the
BssHII, EagI, and SmaI sites that are susceptible to cleavage
in genomic DNA. The BssHII and EagI sites are located at the
3′ end of agouti exon 4, and the SmaI site is located 25 kb
further 3′ (Fig. 2). Applying this information to construct a
long-range restriction map indicates that the 120 kb Ay
deletion removes a EagI site that lies 200 kb 5′ to agouti exon
4, and that normally separates exon 1Ay and agouti exon 1A
(Fig. 2).
Molecular basis of the mouse Ay mutation 1701
Exon 1Ay represents the 5′ end of a ubiquitously
expressed cDNA that is deleted from the Ay
chromosome
A simple model to explain both the 120 kb deletion and the
chimeric exon 1Ay-agouti mRNA associated with the Ay
mutation postulates that the deletion removes the 3′ portion of
a gene that begins with exon 1Ay, and juxtaposes the first
intron of this gene with 5′ flanking sequences of agouti exon
1A. To determine if exon 1Ay might be contained normally in
a gene other than the Ay-specific chimeric mRNA, we used
exon 1Ay as a probe to examine RNA from animals of different
agouti genotypes by northern hybridization analysis. In most
tissues examined of Ay heterozygotes, the exon 1Ay probe
detected two RNAs, approximately 1.0 kb and 1.7 kb in length
(Fig. 3). In all tissues examined of A/A animals, however, only
the 1.7 kb RNA was present. Probes from the agouti cDNA
detect only the 1.0 kb RNA in Ay heterozygotes, indicating that
the 1.0 kb RNA represents the chimeric Ay-specific RNA, and
that the 1.7 kb RNA represents a different gene.
To isolate cDNAs representative of the 1.7 kb RNA, we
used an exon 1Ay probe to screen a teratocarcinoma cDNA
library. Two cDNA clones were identified among 6×105
plaques screened; subsequently, an internal probe from one of
these clones was used to rescreen the library and a third clone
was isolated. A composite sequence from these three cDNA
clones is 1760 bp long including a 35 bp region of poly(A) at
the 3′ end (Fig. 4A). The sequence contains a 296 amino acid
open reading frame with putative 5′ and 3′ untranslated
regions 482 bp and 355 bp in length, respectively. Based on
the sequence and expression pattern of the cDNA (see below),
we refer to it as Merc, for maternally expressed hnRNP Crelated gene.
To investigate the physical relationship of the Merc cDNA
to the 120 kb Ay deletion, we hybridized a cDNA fragment that
contained the Merc open reading frame to nylon filters in
which Ay-specific large restriction fragments had been defined
previously (Fig. 2). The Merc open reading frame probe did
not hybridize to any Ay-specific fragments, but did detect the
same a-specific BssHII, EagI, and SmaI fragments as the exon
1Ay probe, indicating that most or all of the Merc open reading
frame was contained in the 120 kb Ay deletion (Fig. 2).
The Merc cDNA sequence is alternatively spliced
and predicts an RNA-binding protein
The entire exon 1Ay sequence, 147 bp in length, is colinear with
both the Merc cDNA (residues 244-390) and the corresponding
genomic sequence. Surprisingly, the Merc cDNA and the
genomic sequence are identical for an additional 4 nucleotides
beyond the end of exon 1Ay, at which point they diverge from
each other (Fig. 4B). To investigate whether alternative usage
of splice donors used normally by the Merc gene might account
for this difference, cDNA produced from Aw/Aw skin was PCRamplified with oligonucleotide primers that spanned the interval
from exon 1Ay to the Merc open reading frame, and the
products were characterized by gel electrophoresis and direct
sequencing. Two products were obtained in an approximate 3:1
ratio (see Fig. 5A for example), gel purified and sequenced
using an internal primer. The major (≈75%) product was
identical to the Merc cDNA; however, the minor (≈25%)
product was missing an 83 bp region (described as Merc exon
1′, see Fig. 4B) that began at the end of exon 1Ay and that did
not affect the Merc open reading frame. Although further
genomic sequence analysis will be required to distinguish
whether the major Merc cDNA product is produced by an 83
nt exon that uses the same splice donor as that used by exon
1Ay, or by a 79 nt exon that uses a different splice donor from
the one used by exon 1Ay, these results indicate that the Merc
gene normally produces two isoforms, at least one of which
uses the same splice donor used by exon 1Ay.
The methionine codon that initiates the 296 amino acid Merc
open reading frame at nucleotide residue 483 is the first ATG
in the Merc cDNA sequence, and lies within a region, 5′ACCATGTCC-3′, predicted to function as a site for eukaryotic
translational initiation. Sequence similarity searches of computerized databases revealed 43% identity with the human
protein heterogeneous nuclear ribonucleoprotein particle C
(hnRNP C) (Swanson et al., 1987). A comparison (Fig. 4C)
indicates that most of the similarity lies within a 90 residue
amino terminal region (80% identity), which includes the RNAbinding domain (Gorlach et al., 1992; Wittekind et al., 1992).
In both Merc and hnRNP C, the predicted RNA-binding domain
is followed by a proline-rich domain and the carboxyl terminus
contains a high proportion of acidic residues. The most striking
difference between the two sequences is their predicted isoelectric points, 4.84 for hnRNP C, and 10.26 for Merc.
Merc is normally expressed in the unfertilized egg
and the blastocyst
Because the Merc open reading frame is contained within the
Ay-associated 120 kb deletion (see above), we considered
1702 D. M. J. Duhl and others
whether failure to express Merc might be responsible for Ayassociated embryonic lethality. Although hnRNP C and other
hnRNPs are well characterized biochemically, little is known
regarding their expression in development. To investigate
whether Merc was expressed at or prior to the time at which
Ay/Ay embryos are thought to exhibit morphologic abnormalities, RNA from 25-50 eggs, 2-cell embryos, or blastocysts was
reverse transcribed and PCR-amplified with the same oligonucleotide primers used to analyze Merc expression in skin (Fig.
4B). Gel electrophoresis and Southern hybridization of the
amplified products indicated that both forms of the Merc RNA
(with and without Merc exon 1′, Fig. 4B) were present at relatively high levels in unfertilized eggs and relatively low levels
in blastocysts (Fig. 5A). To determine whether the blastocyst
RNA was likely to represent embryonic synthesis and/or degradation of maternal stores, expression of the Merc/agouti fusion
RNA was examined in embryos derived from an Ay/a male
crossed to an a/a female, in which the only source of the fusion
RNA is the Ay chromosome provided by the male genome. The
fusion RNA was detectable in Ay/a blastocysts but not in a/a
blastocysts or in a/a unfertilized eggs (Fig. 5B), suggesting that
Merc RNA is expressed embryonically as well as maternally.
Ay/Ay embryos fail to hatch in
vitro
Previous observations of embryos
derived from Ay/− × Ay/− intercrosses
have reported a wide range of developmental defects in 25% of the animals
compared to control crosses (Calarco
and Pedersen, 1976; Eaton and Green,
1963; Johnson and Granholm, 1978;
Papaioannou, 1988; Papaioannou and
Gardner, 1992; Pedersen, 1974).
However, most of these studies are complicated by the inability to determine the
genotype of individual embryos. To
determine if Ay/Ay embryos exhibited
morphologic or developmental abnormalities prior to implantation, 3.5 dpc
blastocysts were recovered from Ay/a ×
Ay/a matings, placed into culture and
their development was scored after 2-3
days. The genotype of each embryo was
determined using a PCR-based assay to
evaluate the presence of a variant
HindIII site at the Emv-15 locus, which
is located less than 0.2 cM distal to Ay
(Siracusa et al., 1987). Of 92 embryos
evaluated, 65 hatched and formed trophoblastic outgrowths, while 27 never
hatched (Table 2). Based on the identity
of the variant HindIII site, 14/27 that did
not hatch were Ay/Ay, while no Ay/Ay
embyros were found among the 65
embryos that did hatch.
Inhibition of Merc expression
blocks development of cleavagestage embryos
Although the expression pattern of the
Merc gene is consistent with a role in Ay-associated embryonic
lethality, the 120 kb deletion in the Ay chromosome may
remove additional genes that contribute to the failure of Ay/Ay
embryos to develop past the blastocyst stage. To separate
effects due to loss of Merc expression from other effects due
to the Ay-associated deletion, we examined the development of
non-mutant embryos injected with Merc antisense
oligodeoxynucleotides.
In a first series of experiments, approximately 5×108
molecules of sense or antisense oligonucleotides that spanned
the predicted translational initiation site of Merc were injected
into 1-cell fertilized embryos and the embryos were transferred into the oviducts of pseudopregnant females. At 3.5
dpc, embryos were recovered from the uterus, placed into in
vitro culture and their development compared to each other
and to uninjected embryos. Of 169 1-cell embryos injected
with the Merc antisense oligonucleotide, only 76 (45%) were
recovered from the uterus compared to 87% and 77% for the
uninjected and sense controls, respectively (Fig. 6). Nearly all
of the embryos injected with the antisense oligonucleotide
appeared morphologically abnormal at 3.5 dpc; most had not
developed beyond the 4-cell stage and many appeared to
Fig. 5. Embryonic expression of the Merc and the Merc/agouti fusion RNAs. (A) After
reverse transcription of RNA from approximately 25-50 C57BL/6J-a/a unfertilized eggs, 2cell embryos, or blastocysts, Merc cDNA was PCR-amplified using oligonucleotide primers
as indicated in the diagram and as described in Materials and Methods. Three fragments were
evident after ethidium bromide staining or hybridization with the Merc cDNA probe as
shown. Direct sequencing of the gel-purified products indicated that the lower fragment was
the 255 bp product (which lacked the 83 bp exon shown in Fig. 4B), and that both upper
fragments were different forms of the 342 bp product (which contained the 83 bp exon).
(B) After reverse transcription of RNA from approximately 25-50 C57BL/6J-a/a unfertilized
eggs, blastocysts or Ay/a blastocysts in which the Ay chromosome was provided by the male
parent, Merc/agouti fusion cDNA was PCR-amplified using oligonucleotide primers as
indicated in the diagram and as described in Materials and Methods. The predominant
fragment detected by hybridization to an internal oligonucleotide probe from agouti exon 2 is
a 386 bp product that contains agouti exons 1A and 1A′.
Molecular basis of the mouse Ay mutation 1703
Table 2. Development in vitro of embryos derived from an
Ay/a×Ay/a intercross*
Genotype†
Hatched
Unhatched
Total
a/a
Ay/a
Ay/Ay
Total
24
4
28 (30%)
41
9
50 (54%)
0
14
14 (15%)‡
65
27
92
*3.5 dpc blastocysts were obtained as described in Materials and Methods.
After 3-4 days in culture, individual embryos were scored with regard to
hatching from the zona pellucida and the formation of trophoblastic
outgrowths.
†The agouti genotype of each embryo was inferred from the identity of a
variant HindIII site at the closely linked Emv-15 locus as described in
Frohman et al. (1993).
‡The recovery of Ay/Ay embryos is not significantly different from a 1:2:1
distribution at the 95% confidence level (χ2=4.95, 2 degrees of freedom).
molecules of antisense oligonucleotide, coinjection of the
Merc transgene did not significantly affect the proportion of
embryos developing into blastocysts, 9.5%, when compared
to antisense oliogonucleotide alone, 2.8% (P=0.24, Table 3).
However, with 1×107 molecules of antisense oligonucleotide,
coinjection of the Merc transgene resulted in partial rescue;
53% of coinjected embryos developed into blastocysts
compared to 27% injected with antisense oligonucleotide
alone (P=0.017, Table 3). Finally, coinjection of the Merc
transgene completely rescued embryos injected with 2×105
molecules of antisense oligonucleotide; 76% of coinjected
embryos developed into blastocysts compared to 12%
injected with antisense oligonucleotide alone (P=0.0001,
Table 3).
DISCUSSION
exhibit lysis of individual blastomeres compared to embryos
injected with the sense oligonucleotide (Fig. 7). By 5.5 dpc,
only 3 of the 76 embryos (4%) injected with the Merc
antisense oligonucleotide and recovered from the uterus had
developed into blastocysts, compared to 61% and 23% for the
uninjected and sense, respectively (Fig. 6). The 3 blastocysts
that did develop from embryos injected with the Merc
antisense oligonucleotide were probably abnormal, since none
of them hatched from the zona pellucida by 6.5 dpc, compared
to 87% and 69% for the uninjected and sense controls, respectively (Fig. 6).
In a second series of experiments designed to examine the
time course and concentration dependence of developmental
arrest induced by injection of the Merc antisense oligonucleotide, 1-cell fertilized embryos were injected with approximately 5×108, 1×107, or 2×105 molecules of sense or antisense
oligonucleotide, transferred immediately into microdrop
culture, and their development was scored with regard to cell
number and/or blastomere lysis for the next 4 days. At the two
higher concentrations, an effect of the antisense compared to
the sense oligonucleotide was apparent 1 day after injection—
blastomere lysis was observed in 22% or 19% of the embryos
injected with 5×108 or 1×107 molecules of antisense oligonucleotide, respectively, but in none of the embryos injected with
the same amount of sense oligonucleotide (Fig. 8). By 4.5 dpc,
an effect of the antisense compared to the sense oligonucleotide was apparent in embryos injected with all three
oligonucleotide concentrations—only 2.8%, 29%, or 12% of
embryos injected with 5×108, 1×107, or 2×105 molecules of
antisense oligonucleotide, respectively, had developed into
blastocysts compared to 18%, 90%, or 73%, of embryos
injected with the corresponding amount of sense oligonucleotide (Fig. 8; Table 3). The effect of the antisense oligonucleotide at 4.5 dpc was also apparent when compared to
embryos injected with a control ‘scrambled’ oligonucleotide
(Table 3).
In a third series of experiments designed to examine
whether developmental arrest induced by the Merc antisense
oligonucleotide could be rescued by expression of Merc,
embryos were coinjected with different amounts of antisense
oligonucleotide and a Merc transgene, and their development
was compared by scoring the proportion of injected embryos
that had developed into blastocysts by 4.5 dpc. With 5×108
The Ay mutation has intrigued developmental biologists for 90
years due to its pleiotropic effects on coat color, regulation of
body weight, tumor susceptibility and embryonic development.
Here we show that a 120 kb deletion in the Ay chromosome
removes potential protein-coding sequences for a maternally
expressed gene, Merc, that we have isolated from a teratocarcinoma cDNA library, and which is highly similar to the
human hnRNP C gene. The deletion also results in the production of a chimeric RNA in which ubiquitously expressed
Merc 5′ untranslated sequences are fused to protein-coding
sequences of the agouti gene (Fig. 9). During the preparation
of this manuscript, isolation of a gene nearly identical to Merc
was reported by Michaud et al. (1993). These investigators
used exon 1Ay sequences to isolate a 1.4 kb cDNA from a
postimplantation mouse embryo cDNA library which had the
potential to encode a gene they named Raly, for RNA-binding
Fig. 6. Development of embryos injected with Merc antisense or
sense oligonucleotides between 3.5 dpc and 6.5 dpc. Fertilized 1-cell
embryos were injected with approximately 5×108 molecules of
oligonucleotide as described in Materials and Methods. After 3 days
in a pseudopregnant foster mother, the injected embryos and a
population of uninjected control embryos were flushed from the
uterus, placed into in vitro culture, and their development was scored
for the next 3 days. The exact number of embryos in each class are
indicated in parentheses.
1704 D. M. J. Duhl and others
A (Sense)
B (Antisense)
Fig. 7. Morphology of 3.5 dpc embryos injected with Merc antisense or sense oligonucleotides at 0.5 dpc. (A) Embryos injected with the sense
oligonucleotide exhibit a range of normal development from compacted morula to blastocysts. Approximately 30% of the embryos exhibit
developmental abnormalities that may be a non-specific result of the injection (see Table 3). (B) Developmental abnormalities are present in
nearly all the embryos injected with the antisense oligonucleotide.
protein associated with the lethal yellow mutation. Two differences between the Merc and the Raly cDNA sequences
would not produce significant changes in the encoded protein
and may be due to the different sources used to isolate the
cDNA clones; therefore, we will refer to this gene as Raly, subsequently. Our results confirm those of Michaud et al. (1993)
and, in addition, provide a model for understanding the origin
of the Merc/agouti fusion (Fig. 9). Finally, our expression and
embryologic studies suggest that Ay-associated recessive
lethality is caused by loss of Merc gene expression.
Relationship of Ay-specific cDNAs to agouti
genomic structure and origin of the Ay mutation
Initial studies of agouti cDNAs suggested the potential for a
complex genomic structure in which alternative isoforms were
produced with different 5′ ends, but their significance was not
clear until recently. Molecular cloning and expression studies
using RNA from animals that carry the wild-type light-bellied
agouti (AW) allele have now shown that agouti exons 1A and
1A′ initiate isoforms expressed only in the ventrum but
throughout the entire hair growth cycle, and are located 100 kb
5′ of agouti exons 1B and 1C, which initiate isoforms
expressed in the dorusm and the ventrum but only during the
active phase (mid-anagen) of the hair cycle (see Figs 1, 9;
Vrieling et al., 1994). Isoforms produced by the agouti (A)
allele always begin with exons 1B or 1C because the ventralspecific promoter is inactive. Of seventeen cDNA clones that
we isolated from Ay/ax skin RNA, seven began with agouti
exons 1B or 1C. Since there is very little RNA detectable by
northern hybridization analysis from the ax allele, these seven
cDNA clones are likely to represent normal activity of the hair
cycle-specific promoter from the Ay allele.
In contrast, the remaining ten cDNA clones that we isolated
from Ay/ax skin RNA were Merc/agouti fusions; of these, six
contained agouti exon 1A or agouti exons 1A and 1A′ due to
an apparent cryptic splice acceptor in exon 1A, while four
resulted from splicing of Merc 5′ sequences directly to agouti
exon 2. Because none of the cDNA clones that we isolated
began with agouti exon 1A, the ventral-specific promoter is
probably not active in the Ay allele. The Ay mutation was one
of many coat color variants bred by mouse fanciers in the 19th
century; therefore its origins and relationships to current inbred
strains are somewhat obscure. However, haplotype analysis of
closely linked molecular markers indicates that the Ay mutation
is related more closely to the A and AW alleles than to the
nonagouti (a) allele (Winkes et al., 1993). Thus, absence of
Molecular basis of the mouse Ay mutation 1705
ventral-specific promoter activity from the Ay allele may be
explained because (1) the 120 kb Ay deletion has removed cisregulatory sequences required for activity of the ventralspecific promoter; or (2) the allele of origin for the 120 kb Ay
deletion was A; therefore the ventral-specific promoter was
already inactive. Functional analysis of the ventral-specific
promoter in combination with physical mapping of the Ay
deletion breakpoints will help to distinguish between these
possibilities.
We detected the Merc RNA in northern hybridization experiments reported here using a probe that contained exon 1Ay.
In a similar experiment described in Miller et al. (1993), the
1.7 kb Merc RNA was not evident because the ‘Ay-specific
probe’ used in these experiments comprised agouti exons 1A
and 1A′. As described above, these exons are found normally
only in ventral-specific agouti isoforms; therefore, exons 1A
and 1A′ appear specific for the Ay cDNA when analyzed
opposite agouti alleles where the ventral-specific promoter is
inactive such as A or a. Exons 1A and 1A′ were also present
in several Ay-specific cDNAs isolated by Michaud et al. (1993).
Pleiotropy associated with heterozygosity for Ay
The Merc/agouti fusion RNAs and the normal Merc RNA are
both expressed in every adult tissue that we have examined,
suggesting that the 120 kb Ay deletion has not altered regulatory elements which normally direct Merc expression. The
fusion RNAs are capable of encoding a normal agouti protein,
suggesting that ectopic expression of this protein is responsible for the pleiotropic effects associated with heterozygosity
for Ay (Carpenter and Mayer, 1958; Castle, 1941; Danforth,
1927; Granholm and Dickens, 1986; Granholm et al., 1986;
Heston and Deringer, 1947; Heston and Vlahakis, 1961;
Vlahakis and Heston, 1963). Alternative explanations for the
pleiotropic effects, including the production of a novel fusion
protein or the action of closely linked genes, are unlikely since
the pleiotropic phenotype is observed in mice that carry the
viable yellow (Avy) mutation (Dickie, 1962; Wolff et al., 1986),
in which an agouti fusion RNA is also expressed ubiquitously
(Yen et al., 1994), and in which mutant animals can be
compared to coisogenic non-mutant littermates.
There are several possibilities to explain how ectopic
expression of the normal agouti protein could lead to obesity
and increased tumor susceptibility, including antagonism of
melanocortin receptors sequestration of melanocortins, or activation of an as yet unidentified agouti receptor (Conklin and
Bourne, 1993; Jackson, 1993) [reviewed in Yen et al. (1994)].
Experiments with adipose tissue transplants (Meade et al.,
1979) and aggregation chimeras (Barsh et al., 1990) suggest
Table 3. Development in vitro of 1-cell embryos injected
with Merc antisense or control oligonucleotides, with or
without a Merc expression vector*
Dilution of Oligonucleotide†
Injected DNA
1
(1) Control “scrambled”
5/12 (42%)
(2) Sense
4/22 (18%)
(3) Antisense
1/36 (2.8%)
Chi-square vs. (1+2)‡ 6.20; p=0.013
(4) Antisense + Merc
2/21(9.5%)
transgene
Chi-square vs. (1+2)
1.4; p=0.24
Chi-square vs. (3)
0.24; p=0.63
(5) Antisense + control transgene
Fig. 8. Development of embryos injected with Merc antisense or
sense oligonucleotides between 0.5 dpc and 4.5 dpc. Fertilized 1-cell
embryos were injected with approximately 5×108, 1×107, or 2×105
molecules of sense or antisense oligonucleotide and placed into
microdrop culture as described in Materials and Methods. Embryos
were scored daily and categorized within the five conditions shown
on the right. Bars represent percentages (shown on the abscissa) of
embryos in each condition on each day. The results shown represent
the sum of two different experiments, in which a total of 150
embryos were injected with 5×108, 1×107, or 2×105 molecules of
antisense or sense oligonucleotide.
1:50
1:2500
8/14 (57%)
19/21 (90%)
14/51 (27%)
18.6; p=0.0001
25/47 (53%)
10/12 (83%)
11/15 (73%)
3/25 (12%)
20; p=0.0001
16/21 (76%)
4.0; p=0.046
5.7; p=0.017
7/21 (33%)
0.05; p=0.83
16.8; p=0.0001
*Data are given as No. of embryos developing into blastocysts by 4.5
dpc/No. of embryos injected (percent). The data shown include the results
depicted in Figure 8 and represent the results of four separate experiments.
†A solution of 10 mM Tris, pH7.5; 0.1mM EDTA that contained either
oligonucleotide or oligonucleotide + linear DNA was injected into each
embryo. Therefore, dilutions of 1, 1:50, and 1:2500 correspond approximately
to 5×108, 1×107, and 2×105 molecules of oligonucleotide injected per
embryo. The Merc and control transgenes were maintained at a concentration
of 1 µg/ml, which corresponds approximately to 10,000 molecules of 2 kb
fragment injected per embryo.
‡The differences between development of embryos injected with control
(1) and sense (2) oligonucleotides were not significant at a 95% confidence
level, and have therefore been combined in comparison to the antisense (3),
and antisense + Merc transgene (4) injections. The Chi-square values are
corrected for continuity.
1706 D. M. J. Duhl and others
that Ay-induced obesity is cell non-autonomous; however,
parabiosis experiments demonstrate that the factors that
mediate Ay-induced obesity are not transmitted through the circulation (Wolff, 1963). Thus, like the action of agouti in hair
follicles (Silvers, 1958, 1961), agouti-induced obesity may be
mediated by a diffusible factor with a limited radius of action.
1988; Papaioannou and Gardner, 1979, 1992; Pedersen, 1974).
Although these previous studies were complicated by an
inability to determine the genotype of morphologically
abnormal embryos, our results confirm that Ay/Ay embryos can
develop to the blastocyst stage yet fail to hatch when placed
into culture or into diapause.
Ay-associated embryonic lethality
Function of the Merc gene
In contrast to the pleiotropic effects associated with heterozyEmbryos injected with Merc antisense oligonucleotides rarely
gosity for Ay, several observations have suggested that lethality
develop to the blastocyst stage, and thus are more severely
of Ay/Ay embryos is not due to altered expression of the agouti
affected than Ay/Ay embryos. Injection of control sense or
gene (Barsh and Epstein, 1989; Barsh et al., 1990; Lyon et al.,
‘scrambled’ oligonucleotides did not prevent blastocyst devel1985). Our results provide a molecular explanation for these
opment, which suggests that the effects that we observed after
observations, since the 120 kb Ay deletion produces two
injection with Merc antisense oligonucleotides are likely to be
distinct genetic lesions; a neomorphic effect caused by ubiqdue to loss of Merc gene expression. Although expression of
uitous expression of the Merc/agouti fusion RNA, and an
RNA in eggs and in embryos is difficult to quantitate, our
amorphic effect due to the loss of Merc. Although the
results clearly indicate that Merc RNA is present in unfertilexpression of other genes besides Merc and agouti may be
ized eggs, possibly at much higher amounts than in blastocysts.
affected by the Ay deletion, injection of 1-cell embryos with
Thus, maternal Merc RNA and/or protein in the Ay/Ay embryos
Merc antisense oligonucleotides prevents development to the
may provide a ‘sparing’ effect compared to non-mutant
blastocyst stage, suggesting that loss of Merc gene expression
embryos injected with Merc antisense oligonucleotides. These
is sufficient to explain Ay-associated lethality. Of three other
results are also consistent with an inability to rescue Ay/Ay cells
recessive lethal agouti mutations that have been studied, only
in aggregation chimeras (Barsh et al., 1990), and suggest that
one, nonagouti lethal (al), fails to complement Ay (Lyon et al.,
loss of Merc gene expression in preimplantation embryos is a
1985). Genetic and molecular studies of al indicate it is a
cell-lethal. It remains to be determined, however, whether
deletion that removes several genes (Barsh and Epstein, 1989;
Merc expression is required for cell viability later in developLyon et al., 1985), and mapping of the al proximal breakpoint
ment, especially given the evidence that Ay/Ay inner cell mass
may help to refine further the physical boundaries of the Ay
can be rescued in blastocyst injection chimeras (Papaioannou
recessive lethal complementation group, and to delimit regions
and Gardner, 1979). It is possible that the Merc gene product
required for normal Merc gene expression.
plays a specialized role in metabolism of maternally stored
Functional and/or immunologic assays for Merc will be
RNAs, and that, later in development, general roles for hnRNA
required to determine the degree to which its expression is
processing are assumed by other hnRNPs. Although oogenesis
affected by injection of antisense oligonucleotides but, in
proceeds differently in mammals and invertebrates, it is
general, inhibition of gene expression using antisense
intriguing that a recently described mutation that interferes
approaches is usually not absolute, and preimplantation develwith dorsal/ventral axis formation during Drosophila
opment may therefore be sensitive to dosage of the Merc
oogenesis, squid (Kelley, 1993), encodes one of the major
protein. Heterozygosity for Ay has no obvious effect on
Drosophila hnRNPs (Kelley, 1993; Matunis et al., 1992).
viability;
however,
increased
expression of agouti in these
animals may mask subtle effects on
growth caused by a 50% reduction in
Merc gene expression since a/al
animals are slightly reduced in size
compared to a/a animals (unpublished observations). Subtle effects
of reduced Merc gene dosage on
growth and development may also
account for the observation that Ay/a
cells are recovered less frequently
than expected in aggregation
chimeras between A/A embryos and
those derived from an Ay/a × Ay/a
intercross (Barsh et al., 1990).
Previous studies of embryos
derived from Ay/− × Ay/− intercrosses have suggested that homozygosity for Ay may lead to a wide
Fig. 9. Diagram of the 120 kb Ay deletion and its effects on Merc and agouti expression. The exon
variety of developmental defects
structure of Merc has not yet been determined and is indicated with dotted lines. As described in
(Calarco and Pedersen, 1976; Eaton
the text, the Ay allele can give rise to normal agouti mRNAs initiated from the hair cycle-specific
and Green, 1963; Johnson and
promoter. However, it is not clear whether the absence of agouti mRNAs initiated with the
ventral-specific promoter is due to the effects of the deletion or to the allele of origin.
Granholm, 1978; Papaioannou,
Molecular basis of the mouse Ay mutation 1707
The predicted Merc protein is 80% identical to human
hnRNP C over the first 90 amino acids. This region contains a
well-characterized RNA-binding domain found in many
hnRNPs which assembles into a four-stranded beta-pleated
sheet and binds to homopolymeric uridine tracts in vitro (Burd
et al., 1989; Gorlach et al., 1992; Wittekind et al., 1992).
Although the carboxyl terminus of the predicted Merc protein
is very different from hnRNP C, it contains several regions rich
in glycine and serine, which are found in other hnRNPs (Burd
et al., 1989), and the Merc protein product is likely to function
as an hnRNP in cells. In vitro and cell culture experiments
suggest that hnRNPs mediate processing and/or transport of
newly synthesized nuclear RNA in most eukaryotic cells.
Embryonic lethality due to absence of Merc expression in mice
is somewhat surprising given the large number of hnRNPs in
most organisms, and the potential for functional redundancy.
Experiments directed at identification and isolation of Merc
homologs in other organisms may help shed light on the possibility that Merc encodes an hnRNP used specifically for the
processing and/or transport of maternally expressed RNA.
This work was supported in part by HG-00377 from the National
Institutes of Health. H. V. is supported by the J.A. Cohen Institute,
Leiden, The Netherlands. G. S. B. is an Assistant Investigator of the
Howard Hughes Medical Institute.
REFERENCES
Andria, M. L., Barsh, G. S. and Levy, S. (1992). Expression of TAPA-1 in
preimplantation mouse embryos. Biochem. Biophys. Res. Commun. 186,
1201-1206.
Barsh, G. S. and Epstein, C. J. (1989). Physical and genetic characterization
of a 75-kilobase deletion asssociated with al, a recessive lethal allele at the
mouse agouti locus. Genetics 121, 811-818.
Barsh, G. S., Lovett, M. and Epstein, C. J. (1990). Effects of the lethal yellow
(Ay) mutation in mouse aggregation chimeras. Development 109, 683-690.
Bultman, S. J., Michaud, E. J. and Woychik, R. P. (1992). Molecular
characterization of the mouse agouti locus. Cell 71, 1195-1204.
Burd, C. G., Swanson, M. S., Gorlach, M. and Dreyfuss, G. (1989). Primary
structures of the heterogeneous nuclear ribonucleoprotein A2, B1, and C2
proteins: a diversity of RNA-binding proteins is generated by small peptide
inserts. Proc. Nat. Acad. Sci. USA 86, 9788-9792.
Calarco, P. G. and Pedersen, R. A. (1976). Ultrastructural observations of
lethal yellow (Ay/Ay) mouse embryos. J. Embryol. Exp. Morph. 35, 73-80.
Carpenter, K. J. and Mayer, J. (1958). Physiologic observations on yellow
obesity in the mouse. Am. J. Physiol. 193, 499-504.
Castle, W. E. (1941). Influence of certain coat color mutations on body size in
mice, rats, and rabbits. Genetics 26, 177-191.
Conklin, B. R. and Bourne, H. R. (1993). Mouse coat colour reconsidered.
Nature 364, 110.
Cuénot, L. (1905). Les races pures et leurs combinasisons chez les souris (4me
note). Arch. Zool. Exp. Gen., 4e ser. 3, cxxiii-cxxxii.
Danforth, C. H. (1927). Hereditary adiposity in mice. J. Hered. 18, 153-162.
Dickie, M. M. (1962). A new viable yellow mutation in the house mouse. J.
Hered. 53, 84-86.
Eaton, G. J. and Green, M. M. (1963). Giant cell differentiation and lethality
of homozygous yellow mouse embryos. Genetica 34, 155-161.
Frohman, M. A., Martin, G. R., Cordes, S. P., Halamek, L. P. and Barsh, G.
S. (1993). Altered Rhombomere-Specific gene expression and hyoid bone
differentiation in the mouse segmentation mutant, kreisler (kr). Development
117, 925-936.
Gorlach, M., Wittekind, M., Beckman, R. A., Mueller, L. and Dreyfuss, G.
(1992). Interaction of the RNA-binding domain of the hnRNP C proteins
with RNA. EMBO J. 11, 3289-3295.
Granholm, N. H. and Dickens, G. A. (1986). Effects of reciprocal ovary
transplantation on reproductive performance of lethal yellow mice (Ay/a;
C57BL/6J). J. Reprod. Fert. 78, 749-753.
Granholm, N. H., Jeppesen, K. W. and Japs, R. A. (1986). Progressive
infertility in female lethal yellow mice (Ay/a; strain C57BL/6J). J. Reprod.
Fert. 76, 279-287.
Heston, W. E. and Deringer, M. K. (1947). Relationship between the lethal
yellow (Ay) gene of the mouse and susceptibility to spontaneous pulmonary
tumors. J. Nat. Cancer Inst. 7, 463-465.
Heston, W. E. and Vlahakis, G. (1961). Influence of the Ay gene on
mammary-gland tumors, hepatomas, and normal growth in mice. J. Nat.
Cancer Inst. 26, 969-982.
Jackson, I. J. (1993). Molecular genetics. Colour-coded switches [news].
Nature 362, 587-588.
Johnson, L. L. and Granholm, N. H. (1978). In vitro analysis of pre- and early
postimplantation development of lethal yellow (Ay/Ay) mouse embryos. J.
Exp. Zool. 204, 381-390.
Kelley, R. L. (1993). Initial organization of the Drosophila dorsoventral axis
depends on an RNA-binding protein encoded by the squid gene. Genes Dev.
7, 948-960.
Lyon, M. F., Fisher, G. and Glenister, P. H. (1985). A recessive allele of the
mouse agouti locus showing lethality with yellow, Ay. Genet. Res. 46, 95-99.
Magnuson, T. (1986) Mutations and chromosomal abnormalities: How are
they useful for studying genetic control of early mammalian development? In
Experimental Approaches to Mammalian Embryonic Development. (ed. J.
Rossant and R. A. Pedersen). pp 437-474. Cambridge: Cambridge University
Press.
Matunis, E. L., Matunis, M. J. and Dreyfuss, G. (1992). Characterization of
the major hnRNP proteins from Drosophila melanogaster. J. Cell Biol. 116,
257-269.
Meade, C. J., Ashwell, M. and Sowter, C. (1979). Is genetically transmitted
obesity due to an adipose tissue defect? Proc. R. Soc. Lond. B. 205, 395-410.
Michaud, E. J., Bultman, S. J., Stubbs, L. J. and Woychik, R. P. (1993). The
embryonic lethality of homozygous lethal yellow mice (A(y)/A(y)) is
associated with the disruption of a novel RNA-Binding protein. Gene Dev. 7,
1203-1213.
Miller, M. W., Duhl, D. M. J., Vrieling, H., Cordes, S. P., Ollmann, M. M.,
Winkes, B. M. and Barsh, G. S. (1993). Cloning of the mouse agouti gene
predicts a secreted protein ubiquitously expressed in mice carrying the
Lethal-Yellow mutation. Gene Dev. 7, 454-467.
Ollmann, M. M., Winkes, B. M. and Barsh, G. S. (1992). Construction,
Analysis, and Application of a Radiation Hybrid Mapping Panel
Surrounding the Mouse Agouti Locus. Genomics 13, 731-740.
Papaioannou, V. E. (1988). Investigation of the tissue specificity of the lethal
yellow (Ay) gene in mouse embryos. Dev. Genet. 9, 155-165.
Papaioannou, V. E. and Gardner, R. L. (1979). Investigation of the lethal
yellow (Ay/Ay) embryo using mouse chimaeras. J. Embryol. Exp. Morph. 52,
153-163.
Papaioannou, V. E. and Gardner, R. L. (1992). Effects of diapause on lethal
yellow (Ay/Ay) mouse embryos. J. Exp. Zool. 263, 309-315.
Pedersen, R. A. (1974). Development of lethal yellow (Ay/Ay) mouse
embryos in vitro. J. Exp. Zool. 188, 307-320.
Russell, L. B., McDaniel, M. N. C. and Woodiel, F. N. (1963). Crossing over
within a ‘locus’ of the mouse (abstr.). Genetics 48, 907.
Salles, F. J., Richards, W. G., Huarte, J., Gubler, P., Vassalli, J. D. and
Strickland, S. (1993) Microinjecting antisense sequences into oocytes. In
Methods in Enzymology. Guide to Techniques in Mouse Development. (ed. P.
M. Wassarman and M. L. DePamphilis). pp 351-360. San Diego: Academic
Press.
Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989) Molecular Cloning: A
Laboratory Manual. Cold Spring Harbor, New York: Cold Spring Harbor
Laboratory Press.
Silvers, W. K. (1958). An experimental approach to action of genes at the
agouti locus in the mouse. II. Transplants of newborn aa ventral skin to ata,
Awa, A-, and aa hosts. J. Exp. Zool. 137, 181-188.
Silvers, W. K. (1961). Genes and the pigment cells of mammals. Science 134,
368-373.
Silvers, W. K. (1979) The agouti and extension series of alleles, umbrous and
sable. In The Coat Colors of Mice. pp. 6-44. New York: Springer-Verlag.
Siracusa, L. D. (1991). Genomic organization and molecular genetics of the
agouti locus in the mouse. Ann. N. Y. Acad. Sci. 642, 419-430.
Siracusa, L. D., Russell, L. B., Eicher, E. M., Corrow, D. J., Copeland, N. G.
and Jenkins, N. A. (1987). Genetic organization of the agouti region of the
mouse. Genetics 117, 93-100.
Swanson, M. S., Nakagawa, T. Y., LeVan, K. and Dreyfuss, G. (1987).
Primary structure of human nuclear ribonucleoprotein particle C proteins:
1708 D. M. J. Duhl and others
conservation of sequence and domain structures in heterogeneous nuclear
RNA, mRNA, and pre-rRNA-binding proteins. Mol. Cell Biol. 7, 1731-1739.
Tybulewicz, V. L., Crawford, C. E., Jackson, P. K., Bronson, R. T. and
Mulligan, R. C. (1991). Neonatal lethality and lymphopenia in mice with a
homozygous disruption of the c-abl proto-oncogene. Cell 65, 1153-1163.
Vlahakis, G. and Heston, W. E. (1963). Increase of induced skin tumors in the
mouse by the lethal yellow gene (Ay). J. Natl. Cancer Inst. 31, 189-195.
Vrieling, H., Duhl, D. M. J., Millar, S. E., Miller, K. A. and Barsh, G. S.
(1994). Differences in dorsal and ventral pigmentation result from regional
expression of the mouse agouti gene. Proc. Natn. Acad. Sci. USA (In press).
Winkes, B.M., Ollmann, M. M. and Barsh, G. S. (1993) Association os Xmv10 and the nonagouti (a) mutation explained by close linkage instead of
causality. Mammalian Genome, in press.
Wittekind, M., Gorlach, M., Friedrichs, M., Dreyfuss, G. and Mueller, L.
(1992). 1H, 13C, and 15N NMR assignments and global folding pattern of
the RNA-binding domain of the human hnRNP C proteins. Biochemistry 31,
6254-6565.
Wolff, G. L. (1963). Growth of inbred yellow (Aya) and non-yellow (aa) mice
in parabiosis. Genetics 48, 1041-1058.
Wolff, G. L., Roberts, D. W. and Galbraith, D. B. (1986). Prenatal
determination of obesity, tumor susceptibility, and coat color pattern in
viable yellow (Avy/a) mice. The yellow mouse syndrome. J. Heredity 77,
151-158.
Yen, T. T., Gill, A. M., Frigeri, L. G., Barsh, G. S. and Wolff, G. L. (1994).
Obesity, diabetes, and neoplasia in yellow Avy/- mice: Ectopic expression of
the agouti gene. The FASEB J. (in press).
(Accepted 25 March 1994)