Full text of DOI:10.1093/oxfordjournals.hmg.a018917

© 2000 Oxford University Press
Human Molecular Genetics, 2000, Vol. 9, No. 15 2263–2273
Two imprinted gene mutations: three phenotypes
Bruce M. Cattanach+, Josephine Peters, Simon Ball and Carol Rasberry
Medical Research Council, Mammalian Genetics Unit, Harwell, Didcot, Oxfordshire OX11 0RD, UK
Received 11 May 2000; Revised and Accepted 4 August 2000
Genetic modifications of imprinted genes have been
generated in the mouse to investigate the regulation
of their expression. They show classical imprinted
gene inheritances. Here we describe two imprinted
gene mutations deriving from mutagenesis experi-
ments. One is expressed only when transmitted
through males. It causes a prenatal growth retard-
ation which resembles that of the Igf2 knockout and
maps close to the locus on chromosome 7. Differ-
ences from the knockout, which include an abnormal
head phenotype, homozygous lethality, and an
inability to rescue a Tme (Igf2r-deficient) lethality,
suggest that Igf2 itself may not be directly affected.
The second mutation maps close to the Gnas cluster
of imprinted genes on distal chromosome 2. It gives
two distinct phenotypes according to parental origin,
a gross neonatal oedema with microcardia and a
postnatal growth retardation. The oedema phenotype
is effectively lethal and resembles that of mice with
paternal partial disomy for distal chromosome 2, as
well as that of mice having a maternally derived Gnas
exon 2 knockout. However, the second growth
retardation phenotype differs from that of the
maternal partial disomy and the paternal knockout. A
hypothesis to explain the phenotypes associated
with the three genotypes based on the Nesp/Nespas
sense/antisense and Gnasxl transcripts in the Gnas
cluster is offered.
Ube3a (7) knockouts and the targeted central chromosome 7
deletions which involve several imprinted genes (8,9). The T
maternal effect (Tme) lethality, associated with the long-
recognized Thp mutation (10,11) which is attributable to the
deletion of the maternally imprinted Igf2r locus (12), also
shows an imprinted gene inheritance.
In humans, two inherited and imprinted forms of pseudo
hypo-parathyroidism have been recognized. One, PHP-Ia, is
associated with Albright’s hereditary osteodystrophy (AHO)
and involves the GNAS locus (13) on chromosome 20. The
other, PHP-Ib, is not associated with AHO but maps to the
same region (14). Isolated examples of familial Prader–Willi
and Angelman syndromes resulting from mutations at putative
imprinting control loci have also been described (15).
However, up until now imprinted gene inheritances, other than
those of genetic modifications of imprinted genes (3–7) and
that attributable to the Igf2r deletion (12), have not been
described in the mouse. Here we report two mouse mutations
which show imprinted gene inheritances. The first mutant,
called minute (Mnt), lies within the distal chromosome 7
imprinting region and may involve the imprinted Igf2 locus.
The phenotype resembles the Igf2 knockout (3,4) but there are
differences. The second, provisionally called oedematous-
small (Oed-Sml), lies within the distal chromosome 2
imprinting region and uniquely shows two distinct phenotypes
according to the parental route of transmission. One of the
latter resembles a distal chromosome 2 imprinting effect (16).
RESULTS
Minute (Mnt)
INTRODUCTION
Genomic imprinting in mammals specifies a germ line
marking of parental alleles that results in their differential
expression in the zygote. This has been most widely demon-
strated in the mouse through the creation of uniparental diso-
mies and partial disomies, which have two copies of genes
within the chromosomes/regions from one parent and none
from the other (1). Such studies have shown imprinting to
affect genes in at least 11 regions of the mouse genome, with
the imprinting anomalies or phenotypes ranging from early
embryonic lethalities, through neonatal abnormalities, to post-
natal growth retardations. In a number of cases, the genes
responsible for the abnormalities have been identified and
following genetic modification, imprinted gene inheritances,
as first hypothesized by Hall (2), have been demonstrated.
These include those for Igf2 (3,4), Rasgrf1 (5), Gnas (6) and
Origin. The original mutant derived from one of a series of
specific locus mutation experiments (17) in which spermato-
gonially irradiated (6 Gy X-rays) F1 hybrid C3H/HeH × 101/H
males were mated to females homozygous for seven different
recessive mutations. All offspring of the cross were screened
not only for mutations at the tester loci but also for new domi-
nant mutations, many of which, with radiation, have been
found to represent chromosomal changes, notably deletions
and duplications (18). The Mnt mutation was detected in a
male through its severe growth retardation (60% of normal)
observed initially at birth, but this was accompanied by an
abnormal domed head effect (Fig. 1). No evidence of a
chromosomal change has been seen in detailed G-band
analysis (E.P. Evans, personal communication).
+To whom correspondence should be addressed. Tel: +44 1235 834393: Fax: +44 1235 834776; Email: b.cattanach@har.mrc.ac.uk

2264 Human Molecular Genetics, 2000, Vol. 9, No. 15
Table 1. Inheritance and viability of Mnt
Cross
No. of young
Litter size
Phenotypes at birth
Mnt postnatal
viabilitya (%)
Mnt live
45
Mnt dead
+
3H1 FF × Orig M
3H1 FF × Mnt MM
Mnt FF × 3H1 MM
M.cast FF × Mnt MM
158
471b
320
581
7.2
7.4
6.3
4.8
12
22
0
99
55
25
–
111
0
320
320
329
265
0
89
3H1 = F1(C3H/HeH × 101/H) hybrid; F, female, M, male; M.cast., M.m.castaneus.
aLive at weaning relative to live at birth.
bSample of data only.
Table 2. Breeding tests on phenotypically + male progeny of Mnt females
Deduced
genotype
No. tested No. of
young
Litter size No. classified at birth
MntP
612
0
+
MntM/+
37
42
2431
1121
7.0
8.3
1689
1121
+/+
MntM, maternally inherited Mnt allele; MntP, paternally inherited Mnt allele.
Figure 1. Mnt and + sib at birth. Domed heads are arrowed.
Inheritance. On crossing the original animal with normal (+)
females, a proportion of the progeny (of both sexes) expressed
the mutant phenotype (Table 1). A simple dominant inherit-
ance was therefore suggested and this appeared to be con-
firmed on crossing affected sons with + females. Elevated
levels of neonatal loss were encountered, notably in the second
generation, but survival was enhanced in Mus mus castaneus
crosses, perhaps because of the low litter size of this sub-
species (Table 1). Remarkably, none of the progeny of mutant
females were affected (Table 1), yet clearly the mutation was
transmitted as approximately half their sons proved capable of
producing Mnt young (Table 2). Mutant expression was there-
fore obtained only with transmission through males. This rep-
resents an imprinted gene inheritance as illustrated in Figure 2
and implies that the locus affected by Mnt is normally
expressed only when transmitted through the male germ line;
the maternal allele is silent.
Figure 2. Diagram showing inheritance of Mnt. Semi-hypothetical pedigree in
which the original affected male produces up to 50% affected sons and daugh-
ters in matings to + females. Affected sons breed like the father, producing up
to 50% affected male and female progeny in crosses with + females. Affected
females produce only + progeny in crosses with + males, but 50% of their
progeny inherit Mnt. When transmitted into the next generation through males,
the Mnt phenotype reappears.
declined from the earliest prenatal age at which the size differ-
ences could be distinguished (13 days gestation) up until the
time of birth when they approximated <50% of normal (Fig. 3).
The size was significantly different from that of the sibs
throughout (P < 0.1 × 10–9). Extrapolating the data backwards,
the straight line of best fit crossed the age axis at embryonic
day (E) 10.7 (data not shown), possibly defining the age at
which the size difference in the fetus might first be initiated.
Placental weight ratios were more affected than fetal weight
ratios (see legend to Fig. 3) and the straight line of best fit
crossed the age axis ∼3 days earlier. This might suggest that the
reduced fetal growth is at least in part mediated through the
placenta.
Mutant animals recovered from the phenotypically + males
that had inherited the mutation from their mothers (MntM) were
not detectably different from those (MntP) that had inherited
the mutation directly from their affected fathers (Table 2).
Further characterization of Mnt. Prenatal growth rates, as
assessed by ratios of weights of mutant to + sibs within litters

Human Molecular Genetics, 2000, Vol. 9, No. 15 2265
Figure 4. Mapping of Mnt within distal chromosome 7. Only recombinants are
shown. NT, not tested with D7Mit108; a, one animal of each marked genotype
was not tested with D7Mit108.
Figure 3. Growth rates of Mnt animals shown as weight ratios relative to their
+ sibs. All ratios are plotted on a logarithmic scale. Placental weight ratios
(data not shown) were significantly smaller than those for embryos (t = 4.41;
df = 82; P = 0.000031) and did not vary significantly with age [F(5,77) =
0.874; P = 0.50].
Close linkage with Mit markers located within the distal
chromosome 7 imprinting region (19) established that this was
the location of the mutation. Thus, with D7Mit108 and the
more distal D7Mit47, which span the imprinted Igf2 locus and
lie 12 units apart, 13 and 2 recombinants, respectively, were
detected among 100 backcross progeny, suggesting that Mnt
maps between these loci and therefore close to Igf2. This possi-
bility was confirmed in linkage tests with D7Mit291, which
lies between D7Mit108 and Igf2. Thus, using a further 96 back-
cross animals, two recombinants with the phenotype were
detected and a further four were found with D7Mit47. Finally,
when all 22 recombinants were tested with other markers in the
region (D7Mit167, D7Mit259, D7Nds4 and specifically
D7Mit46, which primes up part of the Igf2 gene itself), co-
segregation with D7Mit46 and D7Mit 167 was demonstrated
(Fig. 4). Mnt therefore involves, or lies very close to, Igf2
within the distal chromosome 7 imprinting region. All Mit
markers selected could be studied. No evidence of a deletion
on the Mnt chromosome was therefore indicated.
Postnatal Mnt growth rates suggested that the size difference
declined further to reach a minimum at ∼4 days of age (Fig. 3)
when most inviability occurred. It seems likely that the
lethality is only a consequence of difficulties that these small
animals have in competing with their larger, normal sibs, as
growth and survival improve when litter sizes are artificially or
naturally reduced, as in M.m.castaneus crosses which have low
litter sizes (Table 1). At later ages there appeared to be a
gradual recovery or ‘catch up’ (Fig. 3) but this could be due to
the surviving animals being the least affected. Thus, differ-
ences between the weight ratios at birth of non-survivors
(beyond 4 days) and survivors to later ages were highly signifi-
cant [F(5,11) = 6.14; P = 0.000046]. Mnt adults were still
∼60% of the weight of their + sibs. The domed head phenotype
was most clearly evident shortly after birth (Fig. 1) but could
be observed at all ages.
Fate of Mnt homozygotes. Investigations into the fate of the
Mnt homozygotes were initiated before the map positions were
known and did not provide a clear conclusion. The litter size of
Mnt × Mnt intercrosses were notably low (mean = 3.9, n = 78),
suggesting that the homozygote dies prenatally. On the other
hand, openings of pregnant females from intercross matings
failed to demonstrate a novel or lethal class at 17.5 days gesta-
tion, pre- and post-implantation losses approximating normal
levels (2.4 and 9.2%, respectively). Notably, litter size was
within the normal range at this time (mean = 6.5, n = 35) and
Mnt and + embryos were detected in near-equal frequencies
(124 and 102, respectively). It therefore seemed that at this
stage of development Mnt homozygotes are viable and indis-
tinguishable from heterozygotes. On this basis, the Mnt classes
would comprise MntP/MntM and MntP/+ and the + classes
MntM/+ and +/+.
Mapping of Mnt. Almost all imprinted genes identified so far
have been mapped within the imprinting regions defined with
the use of chromosomal translocations (1). It was therefore
anticipated that the Mnt locus would lie within one of these
regions, several of which give growth retardation imprinting
effects (1). On the basis that the maternal allele is normally
silent and only the paternal allele is expressed, proximal
chromosome 11 was initially considered the most likely site of
the Mnt mutation but linkage tests (data not shown) with
appropriate genetic markers negated this possibility. Distal
chromosome 7 provided a further candidate region as a pater-
nally inherited knockout of one of the known imprinted genes
in this region, Igf2, also bringing about a growth retardation
(3,4) that closely resembles that of Mnt. This gene therefore
became the most likely candidate gene.
Confirmation that the homozygote indeed survived until at
least shortly before birth was finally obtained by opening

2266 Human Molecular Genetics, 2000, Vol. 9, No. 15
Table 3. Inheritance and viability of Oed and Sml phenotypes
Cross
No. of
young born
Litter size
at birth
No. classified at birth
No. classified at 5–7 days
Viability at
classification (%)
Oed
17
+
Sml
–
+
Orig. Sml F × 3H1 MM
Oed FF × 3H1 MM
Sml FF × 3H1 MM
Sml MM × 3H1 FF
73
77
6.7
3.9
5.1
7.4
56
–
Oed 0.30
Oed 0.31
Oed 0.53
Sml 0.57
18
59
402
–
–
–
615
1498
213
–
–
–
454
799
M, male; F, female.
females from Mnt F1 M.m castaneus intercrosses and screening
fetuses for the closely linked Mit46(Igf2) and/or Mit167
markers. At 16–17 days gestation, 21 MntM/MntP, 28 MntP, 21
MntM and 18 + fetuses were identified. Seven of the homo-
zygotes had recently died, and from the failure to find any of
this class at birth it may be concluded that Mnt is homozygous
lethal and that the loss of this class occurs late in gestation to
around the time of birth. The homozygotes did not show any
features that could enable them to be distinguished from heter-
ozygotes. While the abnormal Mnt head phenotype provided
the first indication that the mutant and the Igf2 knockout (3,4)
were not identical, the homozygous inviability illustrated a
second key difference.
compared (1.039 0.03; t35 = 0.437, P = 0.20). The latter
finding might suggest that some element of the T maternal
effect might be modified by the presence of Mnt, but it may
just be more difficult to validly assess weight differences in
these already severely compromised fetuses. The overall
inability of Mnt to rescue the Tme lethality suggests a third
difference from the Igf2 knockout (3,4).
Oedematous-small (Oed-Sml)
Origin. As with Mnt, the Oed-Sml mutation derived from a
specific locus mutation experiment and was detected because of
a growth retardation evident at birth and later ages. In this case
the mutagen used was ethylnitrosourea (ENU) (250 mg/kg body
wt; B.M. Cattanach, unpublished data) and the original variant
animal was female. As ENU is known to cause mutations by
AT→GC base pair transitions or AT→TA transversions (21),
Oed-Sml was not expected to be a chromosomal mutation. No
evidence of a chromosomal change has been seen in detailed
G-band analysis (E.P. Evans, personal communication).
Effect of Mnt on brachyury (T) deletion lethality. The combi-
nation of a paternally inherited Igf2 knockout, in which the
gene is not expressed, with maternally inherited brachyury
deletions, which encompass the Igf2r receptor locus, allows
partial rescue of the fetuses with the T maternal effect (Tme)
lethality (20). The lethality, which typically occurs late in
gestation (10,11), has been attributed to an excess of the Igf2
ligand in the absence of the receptor and the rescue being
achieved by the deficit of Igf2 (20). To determine whether Mnt
could similarly effect such a rescue, females heterozygous for
the brachyury mutation, T37H, which causes the typical, vari-
able Tme lethality when maternally inherited (B.M. Cattanach,
unpublished data), were crossed with Mnt males. Some T37H
survival to birth was found in control crosses of such females
with + males (five dead T37H among 92 + young born), as occa-
sionally found with the Thp deletion (10,11) on the same
genetic background (B.M. Cattanach, unpublished data).
However, when mated with Mnt males, no T37H young were
recovered among 61 offspring born to T37H mothers, although
Mnt young were obtained in the expected frequency. No
suggestion of a rescue was therefore identified.
Further investigation of T37H × Mnt crosses at 17.5 days
gestation indicated a high incidence of late deaths and early
(small mole) losses and, perhaps as a result, the segregation of
both T and Mnt were skewed. However, there was a clear
shortage of both T and T Mnt offspring relative to their + and
Mnt sibs (19:29 and 27:43, respectively), again showing no
indication of rescue. Similar results were obtained with a small
sample of equivalent crosses using Thp females (3:4 and 5:8,
respectively). Weight ratios, however, indicated that the T37H
fetuses were significantly heavier (due to oedema or over-
growth) than their + sibs (1.274 0.041; t33 = 7.53, P = 8.1 ×
10–9), but this was not found when T Mnt and Mnt fetuses were
Inheritance. On being subjected to breeding tests with a +
male, the variant female did not produce growth-retarded
young like herself. Instead, a new variant phenotype appeared
and clearly had a genetic basis as near-identical examples (n =
17) were found in 11 litters (Table 3). The novel phenotype
principally comprised a gross oedema, hereafter denoted Oed.
None of the affected young lived for more than a few days.
Ovary transplantation from neonates was used to try and
achieve survival of the mutant stock. This yielded further
affected animals, but the risk of losing the mutation remained
high. Only the survival of a single Oed male assured the future
of the mutation and this had to be accomplished using in vitro
fertilization and embryo transfer, the male seemingly being
unable to breed normally. Litters were produced by 16 normal
recipient females and, on classification, these yielded a further
unexpected finding. Of the 116 progeny born, none showed the
Oed phenotype but, within 5–7 days of birth, approximately
half became obviously growth retarded, like the original
variant female. Subsequent studies established that when
females showing this growth-retarded or small phenotype
(denoted Sml) were mated with + males, the Oed phenotype
reappeared among their offspring. On the other hand, when
their Sml male sibs, or surviving Oed males, were crossed with
+ females, the Sml phenotype was produced, appearing in
about half the offspring. Data establishing the inheritance of
the double, parental source-dependent phenotypes are

Human Molecular Genetics, 2000, Vol. 9, No. 15 2267
Figure 5. Diagram showing inheritance of Oed-Sml. Semi-hypothetical pedigree
in which the original Sml female produced up to 50% Oed progeny of both sexes
in crosses with a + male. Surviving Oed females breed like their Sml mother, pro-
ducing up to 50% mutant progeny which are Oed. Oed males, on the other hand,
produce a proportion of mutant progeny all of which are Sml like the original
female. Sml females breed like Oed females and produce a proportion of mutant
young all of which are Oed. Sml males produce a proportion of mutant young all
of which are Sml. + animals produce only + progeny.
presented in Table 3 and illustrated diagrammatically in Figure
5. The Oed-Sml mutation therefore presents a second type of
imprinting inheritance.
Further characterization of the Oed phenotype. The Oed pheno-
type was readily detectable at birth but the recovered frequency of
affected young at this age was only half that expected (Table 3).
Prenatal studies indicated an elevated level of resorption sites
(9.9%) and dying Oed fetuses (15.3%) at 17.5–19.5 days gesta-
tion, accounting for the shortage.
At birth, Oed animals resembled mice with the oedematous
imprinting phenotype caused by paternal duplication/maternal
deficiency (partial disomy) for distal chromosome
2
(PatDp.dist2) (13). However, in addition, brown fat accumula-
tions could be seen through the skin, localized at the scapular
region on the back (Fig. 6A) and around the throat. The
oedema, as seen at birth [36% difference between wet and dry
weights (Fig. 6B)], was similar to that of the PatDp genotype at
its peak at 16 days gestation (40% difference) (22), but was
notably more evident before birth (56% difference). With both
genetic conditions the oedema appears to decline just before
and also subsequent to birth (Fig. 7A) and the few adult
Oed survivors (obtained only following crosses with
M.m.castaneus) were not overtly oedematous. However, in
sharp contrast to the PatDp class, Oed neonates showed no
indication of the hyperkinetic behaviour that characterizes the
PatDp class (16).
The ongoing loss of Oed young after birth was accompanied
by a decline in the weight ratios (clearly observable beyond the
loss of oedema). This reached minimum values over the
second and third weeks and then, among the few that survived
longer, appeared to rise (Fig. 7A). In view of the small
numbers of survivors, the significance of the rise is not clear.
Pathological examination of those Oed young that survived
for several days after birth showed inflammatory reactions in
the larynx and pharynx and, in the lungs, there was a haemor-
aghagic reaction with an infiltration of blood cells. Perhaps
Figure 6. (A) Oed mice showing brown fat at scapular regions and around base
of neck (arrowed), with + sib. (B) Grossly oedematous Oed mouse at birth with
+ sib. (C) Declining Oed mice (a and b) at 5 days of age, + sib (c) and growth-
retarded Sml sib (d) deriving from an Sml × Sml intercross.
consistent with the latter finding was the observation of
breathing difficulties in the few animals that survived to such
ages. The aetiology of these effects was not clear; they may be
only secondary consequences of a primary defect. Of more
significance was the finding that heart size was severely

2268 Human Molecular Genetics, 2000, Vol. 9, No. 15
Figure 7. (A) Pre- and postnatal body weight ratios of Oed mice relative to their + sibs. Placental weight ratios were only marginally greater than 1 (mean ratio =
1.084 0.047; t = 1.86; df = 13; P = 0.085). (B) Heart weight ratios of Oed mice relative to their + sibs. (C) Postnatal body weight ratios of Sml mice relative to
their + sibs. All ratios are plotted on a logarithmic scale.
reduced, typically being less than half that of normal. This was
evident from 16 days gestation to 10 days after birth during
which time heart weight ratios declined further (Fig. 7B).
However, two surviving adults had normal heart sizes. The
early effect appeared that of a heart-specific growth retardation
as no other abnormality within the heart was detected. Other
organs appeared normal.
and rough in adults. Adult Oed animals also looked shorter
bodied than their sibs, but body weights were not notably
affected. Similar effects have been noted with those
PatDp.dist2 mice that survive to adulthood (B.M. Cattanach,
unpublished data). Oed mice of both sexes can be fertile.
Further characterization of the Sml phenotype. Weight ratios
based on whole litters indicated a statistical difference [P < 0.1 ×
10–8] from one over the ages tested [birth to 8 weeks (Fig. 7C)].
Further anomalies noted in Oed mice that survived beyond 2
weeks of age included sparse, wiry coats that became dense

Human Molecular Genetics, 2000, Vol. 9, No. 15 2269
Figure 8. Mapping of Oed-Sml relative to Mit markers and the breakpoints of translocations defining the imprinting region.
Some Sml individuals were visibly growth retarded shortly after
birth but classification became reliable only at ∼5–7 days.
Weight ratios were minimum at 10–14 days after birth. Subse-
quently there was some indication of recovery, but comparisons
of surviving and non-surviving Sml mice at earlier ages showed
that the survivors were marginally bigger [F(14,121) = 1.64; P =
0.07]. The apparent recovery may therefore be attributable to
survivors being the least affected. Losses of the Sml class were
substantial with most occurring around the ages when weight
ratios were at their lowest. The size difference appeared to be a
whole animal effect as weight ratios of their organs (heart,
kidney, liver), as a proportion of body weight, did not differ
from 1 (data not shown).
of Sml and Oed classes in relation to + were in accord with
expectation (Table 3). Prenatal lethality of the homozygote
was thus suggested. Consistent with this, the mean litter size of
intercrosses was notably low (4.13/female, n = 75), especially
in comparison with that of outcrosses of Sml females (5.13/
female) (Table 3) in which the low viability Oed class is twice
as numerous.
Openings of females derived from Sml M.m.castaneus F1
intercrosses at 17–19 days gestation again failed to identify a
novel class that could represent the homozygote. On typing 90
fetuses using the two closely linked Mit markers, D2Mit25 and
D2Mit174, 13 were recognized as Oed, 46 were identified as
heterozygotes and 31 were deduced to be +. None were
homozygous for the M.m.musculus alleles, indicating that the
homozygote must die early in development. Post-implantation
losses were marginally higher than that in Sml female × + male
crosses (28 cf. 22%). Given that the frequency of Oed fetuses
in the intercrosses should be only half that in the outcrosses, it
would seem likely that much of the post-implantation loss in
intercross matings is attributable to lethality of the homo-
zygotes.
Mapping of the Oed-Sml mutation. The similarity between the
Oed and PatDp.dist2 phenotypes suggested that the Oed-Sml
locus might lie within the distal chromosome 2 imprinting
region. Linkage back-crosses using Sml males and Mit markers
which lie close to the region (D2Mit226 and D2Mit52 prox-
imal and D2Mit174 distal) or within it (D2Mit25 and
D2Mit504) (19, and J. Peters, unpublished data) confirmed that
the mutation lies distally on chromosome 2. Thus, in the first
screen of 30 animals there was complete linkage with
D2Mit174, although there was a remarkably high recom-
bination frequency (3/30, 10%) with D2Mit25 (Fig. 8) consid-
ering the distal location of these markers on the chromosome.
In a second screen with 35 further animals, complete linkage
was found with D2Mit504, but a high recombination frequency
(8/35, 22.8%) was again found with D2Mit25 (Fig. 8). The
high recombination frequencies are puzzling but the data
nevertheless establish that the Oed-Sml mutation lies distally in
chromosome 2 and probably at the distal end of the imprinting
region. As such it becomes the candidate locus for at least the
oedema component of the PatDp.dist2 imprinting phenotype.
DISCUSSION
Mnt provides a classic example of an imprinted gene inherit-
ance as postulated by Hall (2). Thus, in some crosses a clear
dominant inheritance is evident and yet in others the phenotype
disappears as though inherited as a recessive. The parental
origin of the mutation provides the explanation. Expression is
only seen with transmission through males; when maternally
transmitted the phenotype is wild-type. The only other
example in the mouse is seen with brachyury (T) mutations that
are associated with deletions that encompass the paternally
imprinted/repressed Igf2r locus (12). Described long before
imprinting in mammals was recognized (10,11) these give the
Tme lethality when transmitted though females; a functional
maternal Igf2r copy is required (12). Knockouts of imprinted
genes have in more recent times also shown imprinted gene
inheritances (3–7), as have targeted deletions (8,9). The Oed-
Sml mutation provides a variation on the theme. However,
rather than giving a mutant phenotype when transmitted
through one sex and wild-type or normal when transmitted
Fate of the Oed-Sml homozygote. Intercrosses of phenotypi-
cally Sml animals generated 310 young among which, at birth,
there were 54 Oed and 256 +. Of 224 + progeny that were
reclassified at 5–7 days of age, 92 proved to be Sml and 132
were +. Twenty of the + class were test-mated and all proved
to be genotypically +. A novel phenotype that might identify
the homozygote was not therefore evident, and the proportions

2270 Human Molecular Genetics, 2000, Vol. 9, No. 15
through the other, two different mutant phenotypes are
produced according to parental route of transmission. The
double phenotype therefore resembles the seemingly ‘oppo-
site’ phenotypes of distal chromosome 2 and proximal
chromosome 11 reciprocal uniparental duplications/partial
disomies (16).
Consistent with their imprinting inheritances, both Mnt and
Oed-Sml map within known imprinting regions of mouse
chromosomes (1). Mnt maps close to Igf2 within the distal
chromosome 7 imprinting domain (1,19), and Oed-Sml maps
within the distal chromosome 2 imprinting domain in the
vicinity of the Gnas imprinted gene cluster (22–24).
Igf2 expression that is affected. Ongoing Igf2 expression
studies and detailed analysis of the region are providing
evidence that the mutation is complex. The findings will be the
subject of a further paper.
In seeking interpretation of the double Oed-Sml phenotype,
it may be noted that similar parent of origin-dependent pheno-
types are brought about by uniparental duplications/partial
disomy of the distal chromosome 2 imprinting region (16) and
also with a knockout of exon 2 of Gnas (6). Comparison of the
three phenomena suggests that the effects of all three condi-
tions can be interpreted in terms of an absence of functional
maternal/paternal alleles of distal chromosome 2 loci.
In the case of Mnt, it is possible that the mutation involves
the imprinted Igf2 locus itself. Thus, recombinants with
D7Mit46, which is intrinsic to Igf2, have not yet been found,
and the growth retardation phenotype closely resembles that of
the Igf2 knockout (3,4). Furthermore, the time in development
when the growth retardation is first seen accords with that
observed with the Igf2 knockout (25). Placental weight growth
retardation was seen with Mnt fetuses and this has also been
found with the knockout (25). The knockout data show the
growth effects to appear within the 24 h immediately prior to
E13.5 and remain almost constant thereafter (25), whereas
with Mnt both the fetal and placental growth ratios showed a
significant decrease with increasing age (slopes = –0.0832
With absence of a functional maternal allele, oedema is
found with all three genotypes. Microcardia is characteristi-
cally seen with the Oed-Sml mutation and we have since
detected it in lower degree in paternal duplication (PatDp)
mice derived in studies using the T30H translocation (weight
ratio = 0.796 0.050; P = 0.0024 compared with 0.542 0.03
for Oed mice). Microcardia has not, however, been reported
with the maternally inherited knockout (6). Postnatal viability
is poor to zero with all three classes. On the other hand, no
trace of the abnormal hyperkinetic behaviour seen with the
PatDp genotype (16) is found in Oed mice, whereas some
possibly different form of ‘ataxia’ has been described with the
maternally inherited knockout (6).
0.0077 and –0.0570
0.0149, respectively). Whether this
In the reciprocal situation with functional paternal allele
absence, clear similarities are seen for two of the three geno-
types. With the MatDp (16) and the paternally derived
knockout (6), neonates show thin flat-sided bodies and typi-
cally fail to suckle well if at all. They are hypokinetic. The only
recognized difference between the two phenotypes is that the
maternal duplication (MatDp) class invariably dies within a
few hours, whereas a significant number of the knockouts
survive despite the feeding difficulties.
The Sml phenotype obtained with paternally Oed-Sml trans-
mission differs considerably. They do not have detectably
abnormal body shapes, nor the hypokinetic behaviour with
near-absent suckling ability. Generally they are normal at birth
although there appears to be some early postnatal inviability
that is detectable by a shortage in their numbers at 5–7 days of
age. They cannot be reliably identified earlier and the only
recognized effect is the growth retardation. The paternally
inherited knockout is also reported to show growth retardation,
but it is possible that this is due to the suckling difficulties.
Overall, the Sml phenotype differs considerably from the
hypokinetic MatDp and paternally derived knockout classes
but, like the homozygous knockout (6), the Oed-Sml homo-
zygote is an embryonic lethal.
represents a real difference between the mutant and the
knockout is not clear. Extrapolation of the Mnt data backwards
showed that the lines of best fit to cross the age axis at ∼E10.7
and ∼E14.0, respectively, providing estimates of when the fetal
and placental growth differences might first be initiated. The
approach assumes linearity of growth, which in view of the
knockout data (25) may not be valid.
In contrast, at least two and possibly three observations
clearly distinguish the mutant from the knockout. The first is
the head phenotype. This has not been described with Igf2
knockout mice (3,4). The Mnt mutation, however, occurred
following parental radiation exposure. It is therefore likely to
represent a chromosomal change. A deletion could therefore
account for the compound growth retardation and head
anomaly phenotype, just as the Thp and many other brachyury
deletions bring about the Tme lethality attributable to the
absence of Igf2r (12), as well as the tail effect (10,11).
The second difference between Mnt and the Igf2 knockout is
that the mutant is lethal in the homozygous condition. Here
again, chromosomal deletion could be responsible as most
deletions are homozygous lethal, suggesting that genes within
them are haplo-insufficient (18).
It has recently been shown (6) that the loss of exon 2 of Gnas
disrupts expression of the neighbouring imprinted genes, Nesp
and Gnasxl. Both genes could therefore have roles in
producing the phenotypes brought about in the uniparental
duplication (16) and in the knockout (6) mice. For the Oed-Sml
mutation, the expectation is that it represents a point mutation,
but it also gives two parent of origin-dependent phenotypes.
However, taking everything together, a hypothesis can be
constructed that can explain the clearly established findings
(Fig. 9).
The third and less firmly established difference between the
mutant and the knockout is that Mnt appears to be unable to
rescue a Tme lethality as does the knockout (20). Genetic back-
ground is unlikely to be responsible for the difference as the
brachyury deletion mutant used (T37H) is not distinguishable
from that employed with the knockout (Thp) on the same
genetic background; with both, affected mice also occasionally
survive to birth without rescue. A failure to effect Tme rescue
suggests that at least some level of Igf2 expression remains in
Mnt mice. If verified, this would rule out deletion of the Igf2
locus itself as the basis for the mutation and would suggest that
either some other paternally expressed locus in the region is
responsible for the growth effects, or that it is the regulation of
If Nesp and the newly identified imprinted gene Nespas,
which runs antisense to Nesp (23), are taken as candidate loci
for the Oed-Sml mutation, a single point mutation could disrupt

Human Molecular Genetics, 2000, Vol. 9, No. 15 2271
Figure 9. Hypothesis to account for the phenotypes produced by the Oed-Sml, the Gnas exon 2 knockout, and MatDp.dist2 and PatDp.dist2 phenotypes (maternal
and paternal duplication of distal chromosome 2, respectively) based on Nesp/Nespas being the loci mutated in Oed-Sml. Developed from Figure 4 of Wroe et al.
(24). Blocks show, from left to right, exon-specific Nesp, Gnasxl and Gnas. The maternally expressed Nesp transcript is alternatively spliced with exon 2 of Gnas,
as is the paternally expressed transcript of Gnasxl. The directions of transcription are shown with arrows.
expression of both to give two phenotypes that are parent of
origin dependent. The absence of a maternal Nesp transcript in
Oed mice, in PatDp and in maternally inherited knockout mice,
could then account for the oedema and the absence of the
paternal Nespas transcript could cause the Sml growth retard-
ation phenotype (Fig. 9). The Sml phenotype would also be
expected in MatDp mice as these animals can have no Nespas
transcript, but these hypokinetic non-suckling mice die before
this effect can be detected. In contrast, no such growth retard-
ation would be expected with the paternally inherited
knockout, as the Nespas transcript would be expressed
normally. The hypothesis could be equally well applied to any
gene within the region which has sense and anti-sense tran-
scripts and which is disrupted by Gnas exon 2 knockouts.
If Gnasxl is taken as the candidate locus for the behavioural
effects the transcript, which splices onto exon 2 of Gnas,
would be disrupted in paternally inherited knockout mice. The
absence of this transcript could then account for the hypo-
kinetic second phenotype of the knockout and MatDp animals.
On this basis, the hyperkinetic behaviour of the PatDp class,
which is absent in Oed mice, could be explained by the pres-
ence of two expressed copies of Gnasxl (Fig. 9). The hypoth-
esis ignores the ambiguous growth retardation suggested for
the maternally inherited knockout and also the possibility that
the late onset ‘ataxia’ described for the paternally inherited
knockout (6) might correspond to the neonatal hyperkinetic
behaviour of PatDp animals (16). Further studies with knock-
outs of the Nesp, Gnasxl and other genes in the region should
elucidate the genetic complexities of this imprinting domain in
the mouse. Further such studies in the mouse may provide a
better understanding of imprinting in the homologous human
chromosome domain.
Appraisal of the imprinting effects attributable to genes
within the imprinting regions has suggested from the first that
imprinted genes have important roles in development and
commonly affect growth (26). Whether or not Mnt affects Igf2,
the prenatal growth effects of the mutation accords with this
observation. In the case of Oed-Sml, two such growth effects
are evident. The first, seen in Oed mice, is a growth retardation
of the heart. It might be expected that this organ will be one site
of the responsible mutant gene expression. The second, seen in
Sml mice, is the postnatal growth retardation, which no doubt
is growth hormone-dependent.
It seems remarkable that all three phenotypes described,
Mnt, Oed and Sml, show similarities with imprinting effects
associated with other chromosomal regions. Thus, the prenatal
growth reduction of the paternally inherited Mnt (and the Igf2
knockout) closely resembles the prenatal growth effect caused
by an absence of a paternal copy of proximal chromosome 11
(27). An oedema resembling that caused by the maternally

2272 Human Molecular Genetics, 2000, Vol. 9, No. 15
inherited Oed-Sml mutation, with congenital heart abnormality
(although with enlargement rather than reduction), also occurs
with absence of a maternal copy of Igf2r (20,27). Postnatal
growth retardation effects that are similar to that illustrated
with the paternally inherited Oed-Sml mutation occur with
both a paternally transmitted knockout of the imprinted
Rasgrf1 locus on chromosome 9 (5) and with an absence of a
maternal copy of central chromosome 7 (with mouse AS) (28).
This duplication of phenotypes is also seen with some other
chromosome regions. Thus, suckling problems are seen with
absences of a paternal copies of both distal chromosome 2 (16)
and central chromosome 7 (mouse PWS) (28) and neurological
abnormalities are seen with absence of maternal copies of both
distal chromosome 2 (16) and central chromosome 7 (mouse
AS) (28). More detailed characterization of these imprinting
effects are needed to establish how well these similarities
really correspond. Most of them fit with the ‘conflict hypo-
thesis’ of Moore and Haig (29), but elucidation of this seeming
phenotype duplication may help to resolve further the
intriguing puzzle of the role of imprinting in mammalian
development.
date regions. DNA was extracted from tail tissue using
standard procedures. For PCR analysis, 50 ng of DNA was
added to a final volume of 25 µl of PCR master mix (Advanced
Biotechnologies, Epsom, UK), including 2.5 mM MgCl2 and
0.4 µM of relevant primers. PCR was performed at a 55°C
annealing temperature and with 35 cycles. Products were run
out on 3% agarose gels with TBE buffer.
ACKNOWLEDGEMENTS
We thank Peter Glenister for achieving the rescue of the Oed-
Sml mutation by in vitro fertilization and embryo transfer. We
are also grateful to Dr Leon Cobb for helping with the
pathology, Colin Beechey for preparing the diagrams, David
Papworth for his detailed statistical analysis of the data and
also Lynnette Hobbs, Anne-Marie Woodward and Vicky
Savage for their animal husbandry.
REFERENCES
1. Cattanach, B.M. and Beechey, C.V. (1997) Genomic imprinting in the
mouse: possible final analysis. In Reik, W. and Surani, A. (eds), Genomic
Imprinting: Frontiers in Molecular Biology, Vol. 18. IRL Press, Oxford
University Press, Oxford, NY, Tokyo, pp. 118–145.
2. Hall, J.G. (1990) Genomic imprinting: review and relevance to human dis-
eases. Am. J. Hum. Genet., 46, 857–873.
3. DeChiara, T.M., Efstratiadis, A. and Robertson, E.J. (1990) A growth-
deficiency phenotype in heterozygous mice carrying an insulin-like
growth factor II gene disrupted by targeting. Nature, 345, 78–80.
4. DeChiara, T.M., Robertson, E.J. and Efstratiadis, A. (1991) Parental
imprinting of the mouse insulin-like growth factor II gene. Cell, 64, 849–
859.
5. Plass, C., Shibata, H., Kalcheva, I., Mullins, L., Kotelevtseva, N., Mullins,
J., Kato, R., Sasaki, H., Hirotsune, S., Okazaki, Y. et al (1966) Identifica-
tion of Grf1 on mouse chromosome 9 as an imprinted gene by RLGS-M.
Nature Genet., 14, 106–109.
6. Yu, S., Yu, D., Lee, E., Eckhaus, M., Lee, R., Corria, Z., Accli, D.,
Westphal, H. and Weinstein, L.S. (1998) Variable and tissue-specific
hormone resistance in heterotrimeric Gs protein α-subunit (Gsα) knockout
mice is due to tissue-specific imprinting of the Gsα gene. Proc. Natl Acad.
Sci. USA, 95, 8715–8720.
7. Yang, T., Adamson, T.E., Resnick, J.L., Leff.S, Wevrick, R., Francke, U.,
Jenkins, N.A., Copeland, N.G. and Brannan, C.I. (1998) A mouse model
for Prader-Willi syndrome imprinting-centre mutations. Nature Genet.,
19, 25–31.
8. Jiang, Y-H., Armstrong, D., Albrecht, U., Atkins, CM., Noebels, J.L.,
Eichele, G., Sweatt, J.D. and Beaudet, A.L. (1998) Mutation of the
Angelman ubiquitin ligase in mice causes increased cytoplasmic p53 and
deficits of contextual learning and long-term potentiation. Neuron, 21,
799–811.
9. Gabriel, J.M., Merchant, M., Ohta, T., Ji, Y., Caldwell, R.G., Ramsey,
M.J., Tucker, J.D., Longnecker, T. and Nicholls, R.D (1999) A transgene
insertion creating a heritable chromosome deletion mouse model of
Prader-Willi and Angelman syndromes. Proc. Natl Acad. Sci. USA, 96,
9258–9263.
MATERIALS AND METHODS
Breeding and experimental procedures
The specific locus tests from which the mutants derived
utilized F1 hybrid males deriving from the cross of the C3H/
HeH and 101/H inbred strains. The tester females used were
PT random breeds which are homozygous for seven different
recessive mutations: non-agouti (a), brown (b), pink-eyed dilu-
tion (p), chinchilla (cch), short-ear (se), dilute (d) and piebald
spotting (s). All genetic analyses of the two mutants, other than
mapping, was carried out in crosses using the F1 hybrid, or with
animals of an equivalent but mixed genetic background.
Growth of the mutants relative to their + sibs was made on
the basis of weight ratios to allow for variation between litters
and was assessed by Student’s t-tests. Wet and dry weights of
neonates were similarly investigated. Dry weights were meas-
ured after freeze-drying chopped-up carcasses for 30 h. The
brachyury mutation, T37H, used in crosses with Mnt, derived
from the same series of specific locus mutation radiation
experiments as Mnt (17). Like the well-studied Thp allele, T37H
represents a chromsome 17 deletion (E.P. Evans, personal
communication) and shows a Tme lethality when transmitted
through females. Affected young display an oedema and over-
growth with occasional omphalocoele or hare-lip (B.M. Cattanach,
unpublished data) and most die prior to birth. Mice with
PatDp.dist2 used for comparison with Oed-Sml mutants were
generated using the T(2;11)30H translocation.
10. Johnson, D.R. (1974) Hairpin-tail: a case of post reductional gene action
in the mouse egg? Genetics, 76, 795–805.
11. Johnson, D.R. (1975) Further observations on the hairpin-tail (Thp) muta-
tion in the mouse. Genet. Res., 24, 207–213.
12. Barlow, D.P., Stoger, R., Herrmann, B.G., Saito, K. and Schweifer, N.
(1991) The mouse insulin-like growth factor type-2 receptor is imprinted
and closely linked to the Tme locus. Nature, 349, 84–87.
All animal studies were carried out under the guidance
issued by the Medical Research Council in The Use of Animals
for Medical Research (July 1993) and Home Office Project
Licence nos. 30/00875 and 30/01518.
13. Wilson, L.C. and Trembath, R.C. (1994) Albright’s Hereditary Osteodys-
trophy. J. Med. Genet., 31, 779–794.
Mapping procedures
14. Juppner, H., Schipani, E., Bastepe, M., Cole, D.E.C., Lawson, M.L.,
Mannstadt, M., Hendy, G.N., Plotkin, H., Koshyami, H., Koh, T. et al
(1998) The gene responsible for pseudohypoparathyroidism type Ib is
paternally imprinted and maps in four unrelated kindreds to chromosome
20q13.3. Proc. Natl Acad. Sci. USA, 95, 11798–11803.
Mnt and Oed-Sml mapping was carried by crossing mutant
males with M.m.castaneus females, backcrossing to C3H/HeH
and typing the backcross progeny for M.musculus–
M.m.castaneus polymorphisms at Mit marker loci in the candi-

Human Molecular Genetics, 2000, Vol. 9, No. 15 2273
15. Nicholls, R.D., Saitoh, S. and Horsthemke, B. (1998) Imprinting in
Prader-Willi and Angelman syndromes. Trends Genet., 14, 194–200.
16. Cattanach, B.M. and Kirk, M. (1985) Differential activity of maternally
and paternally derived chromosome regions in mice. Nature, 315, 496–
498.
17. Cattanach, B.M., Peters, J. and Rasberry, C. (1989) Induction of specific
locus mutations in mouse spermatogonial stem cells by combined chemi-
cal X-ray treatments. Mutat. Res., 212, 91–101.
18. Cattanach, B.M., Burtenshaw, M.D., Rasbery, C. and Evans, E.P. (1993)
Large deletions and other gross forms of chromosome imbalance compat-
ible with viability and fertility in the mouse. Nature Genet., 3, 56–61.
19. Mouse Genome Database (MGD), Mouse Genome Informatics Web Site,
The Jackson Laboratory, Bar Harbor, ME (http://www.informat-
ics.jax.org/May2000 ).
20. Filson, A.J., Louvi, A., Efsratiadis, A. and Robertson, E.J. (1993) Rescue
of the T-associated maternal effect in mice carrying null mutations in Igf-
2 and Igf2r, two reciprocally imprinted genes. Development, 118, 731–
736.
23. Peters, J., Wroe, S.F., Wells, C.A., Miller, H.J., Bodle, D., Beechey, C.V.,
Williamson, C.M. and Kelsey, G. (1999) A cluster of oppositely imprinted
transcripts at the Gnas locus in the distal imprinting region of mouse chro-
mosome 2. Proc. Natl Acad. Sci. USA, 96, 3830–3835.
24. Wroe, S.F., Kelsey, G., Skinner, J.A., Bodle, D., Ball, S.T., Beechey,
C.V., Peters, J. and Williamson, C.M. (2000) An imprinted transcript,
antisense to Nesp, adds complexity to the cluster of imprinted genes at the
mouse Gnas locus. Proc. Natl Acad. Sci. USA, 97, 5203–5208.
25. Baker, J., Liu, J.-P., Robertson, E.J. and Efstratiadis, A. (1993) Role of
insulin-like growth factors in embryonic and postnatal growth. Cell, 75,
73–82.
26. Cattanach, B.M. (1991) Gene expression and its control. In Davies, K. and
Tighleman, S. (eds), ChromosomeIimprinting and its Significance for
Mammalian Development. Genome Analysis, Vol. 2. Cold Spring Harbor
Laboratory Press, Cold Spring Harbor, NY, pp. 41–71.
27. Miyabara, S, Bogart, M.H., Winking, H. and Sugihara, H. (1993) Over-
growth and enlarged heart in mouse fetuses with maternally inherited Thp
and twLub2. An example of genomic imprinting in animals. Cong. Aonom.,
33, 63–75.
21. Justice, M.J., Noveroska, J.K., Weber, J.S., Zheng, B. and Bradley, A.
(1999) Mouse ENU mutagenesis. Hum. Mol. Genet., 8, 1855–1963.
22. Williamson, C.M., Beechey, C.V., Papworth, D., Wroe, S.F., Wells, C.A.,
Cobb, L. and Peters, J. (1998) Imprinting of distal mouse chromosome 2
is associated with phenotypic anomalies in utero. Genet. Res. Camb., 72,
255–265.
28. Cattanach, B.M., Barr, J.A., Beechey, C.V., Martin, J., Noebels, J. and
Jones, J. (1997) A candidate model for Angelman syndrome in the mouse.
Mamm. Genome, 8, 472–478.
29. Moore, T. and Haig, D. (1991) Genomic imprinting in mammalian
development: a parental tug-of-war. Trends Genet., 7, 45–49.